Information

Are centrioles really absent in human neurons?

Are centrioles really absent in human neurons?


We are searching data for your request:

Forums and discussions:
Manuals and reference books:
Data from registers:
Wait the end of the search in all databases.
Upon completion, a link will appear to access the found materials.

A booklet (issued by my school) claims,

Centrioles, formerly believed to be absent in neurons, have been described in neurons and may be associated with the production and maintenance of neuro(micro)tubules in them.

I was under the impression that mature human neurons are totally devoid of centrioles.

Being something circulated at a school, it lacks niceties such as citations/references. I've done a spot of Googling, but that didn't yield anything confirming the claim in the booklet.


Q- As per latest studies/info. do (mature) human neurons really possess centrioles?


You asked for "latest studies/info"… Funnily enough, mature neurons with centrioles were described by Cajal (who needs no introduction) in 1911.

If you want a more modern source, you can see Jacobson (1978):

Microtubules associated with the cilia and centriole are present in all young neurons and many mature neurons and glia. Neuronal cilia and centrioles can be seen very easily by light microscopy in sections impregnated with silver by the Nauta method (H. A. Dahl, 1963). Centrioles can also be shown by light microscopy in neurons stained by other histological methods. (emphasis mine)


Sources:

  • CAJAL, S. RAMON Y. (1911). Histologie du système nerveux de l'Homme et des Vertébrés (trans. L. Azoulay, 1952), vol. i. Madrid: Consejo Superior de Investigaciones Cientificas.
  • Jacobson, M. (1978). Developmental Neurobiology. Boston, MA: Springer US.

Centrioles are organelles that have two critical functions. In dividing cells, they recruit a collection of proteins (known as pericentriolar material) to form larger organelles called centrosomes that nucleate microtubules and organize the spindle poles during cell division (Fu et al., 2015 Figure 1A) In non-dividing cells, centrioles are involved in the production of cilia, the tiny hair-like projections that cells use for signaling, sensing and moving extracellular fluid (Drummond, 2012).

How centrioles duplicate in C. elegans.

(A) Schematic of a metaphase centrosome containing a mature mother centriole that can recruit pericentriolar material and reproduce, and her immature daughter centriole. The pericentriolar material recruited by the mother centriole nucleates microtubules and organizes the pole of the mitotic spindle. (B) The steps in the assembly and maturation of the daughter centriole are illustrated, along with the proteins required for each step. Schematics on the bottom show a cross-sectional view of the daughter. Assembly begins when a cartwheel (grey) forms at a right angle to the mother centriole. In the second step, an outer wall made up of nine symmetrically-arranged microtubules (grey) forms around the cartwheel. Assembly of the paddlewheel (a set of protrusions that run along the length of each microtubule red) and acquisition of the ability to reproduce requires SAS-7. (C) The phenotypes observed when a sperm cell containing a wild-type pair of centrioles fertilizes a wild-type egg cell (left column), an egg cell lacking a component essential for daughter centriole formation (middle column), or an egg cell lacking a component required for daughter centrioles to acquire the ability to reproduce (right column).

An individual centriole consists of a central hub called the cartwheel surrounded by an outer wall that contains a nine-fold symmetric array of stabilized microtubules (Figure 1B Fu et al., 2015 Gönczy, 2012). When a cell is born, it contains two mature centrioles. Concurrent with DNA replication, the centrioles also begin to duplicate, with each centriole giving rise to a new daughter that forms at a right angle to the outer wall of its mother (Figure 1B). By metaphase, the new daughter centriole has a cartwheel and an outer wall. However, while it remains attached to its mother, the daughter centriole is immature because it lacks the ability to recruit its own pericentriolar material and to give rise to its own daughter. As the cell divides into two daughter cells, the new daughter centriole acquires these abilities when it separates from its mother (Figure 1B).

In vertebrates and insects, a pathway for centriole maturation has been identified that requires a specific protein called Cep295/Ana1 (Fu et al., 2016 Izquierdo et al., 2014 Tsuchiya et al., 2016). However, nematodes do not have a Cep295/Ana1 homolog, raising the question of how centrioles mature in these organisms. Now, in eLife, Bruce Bowerman and colleagues – including Kenji Sugioka of the University of Oregon and Danielle Hamill of Ohio Wesleyan University as joint first authors – report the results of experiments on the model nematode C. elegans that begin to answer this question (Sugioka et al., 2017). In particular, they have identified a C. elegans protein called SAS-7 that is required for centrioles to acquire the ability to reproduce.

The core centriole assembly pathway was discovered in C. elegans because the depletion of proteins required for centriole assembly from egg cells leads to a characteristic phenotype (Figure 1C). During fertilization, the sperm cell brings a pair of centrioles into the egg cell, which lacks centrioles. These sperm centrioles duplicate so that the centrosome at each pole of the mitotic spindle contains a mother-daughter centriole pair. After the first round of cell division, each cell of the two-cell embryo inherits two mature centrioles, a mother and a newly mature daughter from the first cell cycle, which both have the ability to reproduce and recruit pericentriolar material to form centrosomes (Figure 1C, left column). In contrast, when a protein required for daughter centriole formation is absent in the egg, the wild-type sperm still brings in a pair of centrioles, but new daughter centrioles fail to form during the first cell cycle, so each cell of the two-cell embryo inherits a single mature centriole, rather than the normal pair of centrioles. Consequently, both cells assemble spindles that have just one pole rather than the normal two (Figure 1C, middle column).

Screens in C. elegans identified four proteins whose inhibition leads to monopolar spindles in two-cell stage embryos, indicating that they are essential for the formation of daughter centrioles (Figure 1B): a kinase called Plk4 or ZYG-1 that initiates centriole assembly (O'Connell et al., 2001) SAS-5 and SAS-6, which are required to form the cartwheel (Dammermann et al., 2004 Delattre et al., 2004 Leidel et al., 2005) and SAS-4, which is a structural component of the outer wall of the centriole (Kirkham et al., 2003 Leidel and Gönczy, 2003). A fifth essential component, SPD-2 has two functions: it is required for centrioles to recruit pericentriolar material to form centrosomes and also for daughter centriole formation (Kemp et al., 2004 Pelletier et al., 2004). SPD-2 is the most upstream component in the assembly pathway because it recruits Plk4 kinase to the mother centriole to initiate daughter centriole formation (Delattre et al., 2006 Pelletier et al., 2006). All of these proteins are conserved in vertebrates and are being extensively studied to understand their roles in centriole assembly.

Sugioka et al. study centrioles in C. elegans embryos with a mutation in the gene encoding SAS-7. Whereas removing proteins essential for centriole assembly in egg cells leads to monopolar spindles in two-cell embryos, monopolar spindles were not observed until the four-cell stage in sas-7 mutant embryos fertilized by wild-type sperm. This new phenotype arises because daughter centrioles are able to form during the first cell cycle. The new daughter centrioles separate from their mothers as the first division completes and recruit pericentriolar material to form centrosomes. Thus, both cells of the two-cell embryo have normal bipolar spindles. However, the new centrioles formed in the first cell cycle lack the ability to reproduce and fail to form daughters. Consequently, in the four-cell embryo, the two cells that inherit the sperm centrioles and their daughters assemble normal bipolar spindles, whereas the two cells that inherit the centrioles assembled during the first cell cycle in the embryo have monopolar spindles (Figure 1C, right column). More research is needed to assess whether inhibition of other proteins can cause a similar phenotype (which would indicate that they have a role in daughter centrioles acquiring the ability of to reproduce), indicating a role in acquisition by daughter centrioles of the ability of to reproduce, because the genome-wide RNAi screen that identified the majority of the centriole assembly pathway only monitored one and two-cell embryos (Sönnichsen et al., 2005).

Sugioka et al. also used transmission electron microscopy to visualize centrioles from wild-type and mutant embryos. They found that the wild-type centrioles had a 'paddlewheel' structure that was absent from centrioles assembled in the sas-7 mutant (Figure 1B). Their results suggest that SAS-7 is required for the formation of this structure.

Sugioka et al. further show that SAS-7 localizes to centrioles and is recruited to them independently of SPD-2. SAS-7 interacts with SPD-2 via a small C-terminal region missing in the mutant protein, and recruitment of SPD-2 to centrioles during interphase, when the daughter centrioles form, is severely compromised in the sas-7 mutant. Interestingly, assembly of pericentriolar material in mitosis, which also requires SPD-2, is relatively normal, which explains why normal spindles form in two-cell sas-7 mutant embryos (Figure 1C).

Collectively, the findings of Sugioka et al. indicate that maturation of daughter centrioles involves two events: (1) acquisition of a paddlewheel and the ability to recruit SPD-2 during interphase, which confers on the centriole the ability to reproduce (2) acquisition of the ability to recruit SPD-2 and assemble pericentriolar material during mitosis to form a centrosome that can organize the spindle pole. SAS-7 is essential for the first event, but not the second, which is why mutations in the gene for SAS-7 affect the competence of centrioles to duplicate, without preventing formation of the spindle pole.

SAS-7 appears to be the functional analog of Cep295 in vertebrates. Like SAS-7, Cep295 recruits the SPD-2 homolog, Cep192, to daughter centrioles during their maturation through a direct interaction with its C-terminus (Tsuchiya et al., 2016). Although Sugioka et al. do not report any sequence homology with Cep295, they do report limited homology between SAS-7, a Drosophila protein called Chibby and a human protein called Cby2. Chibby and a paralog of Cby2 (Cby1) are implicated in centriole-to-basal body conversion (Enjolras et al., 2012 Lee et al., 2014), a process that has a central role in the production of cilia. This similarity raises the possibility that the maturation of daughter centrioles and the participation of centrioles in cilia formation may have similar mechanistic underpinnings.


What are centrioles?

The cell is composed of many different organelles. To understand these concepts better, consider the cell like the human body, and the organelles are like the organs. Just like every organ inside of us, the organelles present in cells each have their own unique functions and characteristics. These micro-units are highly specialized and their structure directly corresponds to the tasks being performed by them.

Centrioles (Photo Credit : Aldona Griskeviciene/Shutterstock)

Centrioles are one of the many organelles present inside a eukaryotic animal cell. They are cylindrical in shape with ridges present all over their surface. They are mainly composed of a globular protein called tubulin, which has two major components&mdash&alpha- and &beta-tubulins.

These two types of proteins bundle around each other and polymerize in a specific manner to form microtubules, which basically constitute centrioles. Microtubules, as their name suggests, are small tubes that group together and form centrioles. There are a total of nine pairs of microtubules in each centriole and as with every other cell organelle, the structure of centrioles is closely related to its function. We&rsquoll discuss more on that shortly.

Every cell has two centrioles just before cell division begins. These are always stationed at right angles to each other. Once the cell divides, each centriole occupies the space inside one of the new cells and later divides to once again form a pair.

This cycle continues throughout the life of the organism. Centrioles are considered to be a part of a larger organelle called centrosomes. The centrosome is the main microtubule organizing center and it regulates the cycle of cell division.


Introduction

Neuronal polarity is established by a series of highly coordinated processes, starting with the formation of the future axon. During axon specification, the first step of axon formation, one of the multiple unpolarized neurites of a neuron displays extensive growth. Axonal outgrowth critically relies on local cytoskeleton reorganization and growth cone dynamics (Dotti et al, 1988 Witte et al, 2008 ). Next, the newly developed axon undergoes significant reorganization as it matures, thereby adopting axon-specific hallmarks required for its function. An essential component of mature axons is the axon initial segment (AIS), a specialized compartment at the base of the axon where specific proteins (e.g., AnkG scaffolds, microtubule-organizing protein Trim46, and voltage-gated sodium and potassium channels) assemble in a highly organized manner (Leterrier, 2018 Freal et al, 2019 ). The AIS is crucial for maintaining neuronal polarity and generating action potentials (APs). The characteristic shaping and subsequent propagation of APs is facilitated by the local clustering of voltage-gated channels at the AIS (Kole et al, 2008 ). Another particularly important aspect of mature axons is their unique microtubule organization. In growing axons, the microtubule network undergoes extensive remodeling, as it shifts from a mixed microtubule polarity to a uniform plus-end out microtubule organization (Yau et al, 2016 ). Trim46 proteins targeted to the AIS act as regulators of these axonal microtubule rearrangements by forming parallel microtubule bundles in proximal axons (van Beuningen et al, 2015 ). In contrast, dendrites contain a microtubule organization of mixed polarities and gain additional minus-end out microtubules during development. This prominent difference in microtubule organization between axons and dendrites is essential for neuronal development and function, as it contributes to polarized cargo transport and the characteristic neuronal morphology (Baas et al, 1988 Yau et al, 2016 ). However, while microtubule remodeling in growing axons is important for axon specification and development, the mechanisms driving these microtubule cytoskeletal rearrangements remain largely unresolved.

Centrosomes, the main microtubule-organizing center (MTOC) in most animal cells, are essential for organizing the microtubule network in unpolarized neurons (Tsai & Gleeson, 2005 Stiess et al, 2010 Meka et al, 2020 ). These small, membrane-less, and centrally localized organelles are composed of two centrioles surrounded by the pericentriolar material (PCM) (Moritz et al, 2000 ). The majority of the microtubules are typically nucleated from γ-tubulin Ring Complexes (γTuRCs) embedded in the PCM (Moritz et al, 2000 ). During neuronal development, centrosomes gradually lose their function as MTOC while cilia, the major signaling hubs in polarized cells, are being formed (Stiess et al, 2010 Ishikawa & Marshall, 2011 ). In dissociated rodent neurons, this process was reported to occur during axon development, but the exact temporal relation between axon specification and the declining MTOC function of centrosomes is unclear (Stiess et al, 2010 ). The importance of centrosome function in early neurodevelopment is illustrated by the increasing number of identified mutations in centrosomal proteins causing microcephaly and other neurodevelopmental disorders (Nano & Basto, 2017 ). However, the precise function of centrosomes as MTOC for different processes of early axon development is still under debate.

Progress in understanding the role of centrosomes during axon specification has been hindered due to a number of technical challenges. In particular, centrosomes are found to display different functions in neurodevelopment in different species, resulting in conflicting findings. This is mostly illustrated by the poor recapitulation of human neurodevelopmental disorders caused by centrosome dysfunction in Drosophila and mice, whereas ferrets robustly model these diseases (Basto et al, 2006 Castellanos et al, 2008 Pulvers et al, 2010 Johnson et al, 2018 ). Axon outgrowth can also be differently affected by centrosomes, as this process is perturbed with centrosome dysfunction in mice and peripheral axons of zebrafish, but not in dissociated rodent neurons and central axons of zebrafish (de Anda et al, 2010 Stiess et al, 2010 Andersen & Halloran, 2012 ). The molecular mechanisms that underlie the observed species-specific differences remain largely unknown. Another technical challenge is presented by dissociated rodent neurons in culture, which are classically used to study axon developmental processes, as they likely undergo repolarization after being polarized in vivo rather than de novo polarization (Barnes & Polleux, 2009 ). Together, this highlights the importance of studying the role of centrosomes in de novo polarization using human neurons. The development of human-induced pluripotent stem cells (iPSCs) now enables studying the molecular mechanisms that drive the transition of an unpolarized human neuronal stem cell to a polarized human neuron (Lancaster et al, 2013 ). An additional important feature of human neurons is their significant protracted development, which allows for a more detailed investigation of the temporal processes underlying neuronal polarity (Otani et al, 2016 Linaro et al, 2019 ). To illustrate this, neurogenesis occurs after

1 week in rodents, whereas this takes about

3 months in humans, both in vivo and in vitro (Shi et al, 2012 Espuny-Camacho et al, 2013 Otani et al, 2016 Sousa et al, 2017 ). This profound slower development of human iPSC-derived neurons increases the temporal resolution to study axonal processes, which recently has led to the identification of an additional axon developmental stage in human neurons (Espuny-Camacho et al, 2013 Lindhout et al, 2020 Otani et al, 2016 Linaro, 2019 #9). Altogether, this emphasizes the relevance of studying centrosome functions during axon development in human iPSC-derived neurons as a model system.

In this study, we used a multidisciplinary approach, by combining human iPSC-derived neuron cultures with live-cell imaging, electrophysiology, and mass spectrometry analysis, to examine the role of centrosomes during early axon development. We found that centrosomes display microtubule-organizing functions during axon specification, similar to non-human neurons, and this function is gradually lost during further axon development. In human neurons specifically, Trim46 localization shifts from a pericentriolar region to the AIS during neuron development, coinciding with the developmental decline of centrosomal microtubule-organizing functions. Differentiation of centriole-depleted neuronal stem cells (NSCs) results in various axonal developmental defects in human neurons, including immature action potential firing, mislocalization of Trim46 proteins, growth cone perturbations, and impaired axonal microtubule remodeling. Together, these data imply that centrosomes mediate microtubule remodeling during axon specification in human iPSC-derived neurons, which is necessary for correct axon formation during further development.


Results

Centrosomes display microtubule-organizing functions during axon specification

The centrosome is the primary site for microtubule nucleation and anchoring during early neurodevelopment, and this function gradually declines as neurons mature (Fig 1A) (Stiess et al, 2010 ). To understand the role of centrosomes during early axon development in human neurons, we set out to explore if centrosomes display microtubule-organizing functions during this developmental process in human iPSC-derived neurons. To address this, we tested the microtubule nucleating capacity of centrosomes during development by measuring the endogenous levels of local γ-tubulin, an essential microtubule nucleating protein, at different developmental stages. The neurodevelopmental stages were defined as follows: stage 1 (day 0) as Ki67-positive NSCs stage 2 (day 7) as differentiated unpolarized MAP2-positive neurons with Trim46-negative processes stage 3 (day 12) as differentiated polarized MAP2-positive neurons with a Trim46-positive axon (Fig 1B). We identified centrosomes with centriole marker Centrin and quantified the intensity levels of γ-tubulin co-localizing with Centrin (Fig 1B). The γ-tubulin levels at centrosomes were consistently high in stage 1 and stage 2 neurons, and markedly reduced by

50% in stage 3 neurons, in line with previous findings in dissociated rat neurons (Fig 1B and C) (Stiess et al, 2010 ). To verify these results, we next performed high-resolution 3D STED imaging to resolve the dense neuronal microtubule network and observed that the microtubule network transformed from a radial organization toward a more non-radial organization during initial neurodevelopment (Fig 1D). This transition marks the decline of microtubule-organizing functions of centrosomes and correlates with the onset of axon development. Together, these results suggest that centrosomes display microtubule-organizing functions during the early developmental stages of human iPSC-derived neurons, which includes the process of axon specification, but not during later stages.

Figure 1. Centrosomes display MTOC function and Trim46 appearance during axon specification

  • A. Schematic illustration of centrosomal MTOC function in neurodevelopmental stages 1, 2, and 3 in human iPSC-derived NSCs/neurons.
  • B. Typical examples of stage 1 (day 0), stage 2 (day 7), and stage 3 (day 12) human iPSC-derived NSCs/neurons immunostained for γ-tubulin and Centrin, marked by arrows. Cells were co-immunostained with Ki67 and DAPI (day 0) or Trim46 and MAP2 (day 7 and 12) to define their neuronal stages. Arrowheads mark the AIS. Scale bar = 10 µm in overview, 2 µm in zooms.
  • C. Quantifications of normalized γ-tubulin fluorescent intensities at centrosomes in stage 1, 2, and 3 human iPSC-derived neurons. n = 46–51 cells in two independent experiments.
  • D. 3D STED imaging of α-tubulin in stage 1 (day 0), stage 2 (day 7), and stage 3 (day 13) hiPSC-derived neurons. Scale bar = 5 µm in overview, 1 µm in zoom.
  • E. Human iPSC-derived NSCs (day 0) immunostained for Trim46, γ-tubulin, and Centrin. Zoom represents centrosome structure. Scale bar = 5 µm in overview, 2 µm in zooms.
  • F. Typical examples of centrosomes of human iPSC-derived NSCs (day 0) with STED imaging of Trim46 and Centrin immunostaining, and confocal imaging of γ-tubulin immunostaining. Scale bar = 1 µm.
  • G. Typical examples of stage 2 and stage 3 human iPSC-derived neurons immunostained for Trim46 and MAP2. Inserts represent centrosomes. Zooms on the right represent a non-polarized neurite or a developing axon in stage 2 or stage 3 neurons, respectively. Scale bar = 20 µm in overview, 1 µm in insert, 10 µm in zoom.

Data information: Data represent mean ± SEM. One-way ANOVA including post hoc analysis with Bonferroni correction (C). ***P < 0.001, ns P ≥ 0.05.

Centrosome-associated localization of AIS protein Trim46 in early-stage human neurons

An important hallmark of axon development is the assembly of the AIS in the proximal axon, which occurs after the developmental decline of the microtubule-organizing function of centrosomes in dissociated rat hippocampal neurons (Stiess et al, 2010 ). In human neurons, we observed that the AIS protein Trim46 was localized at centrosomes during early neurodevelopment (stage 1–2), as shown by co-localization of Trim46 with the centrosome markers Centrin and γ-tubulin in hiPSC-derived NSCs (Fig 1E). Similarly, centrosome-associated localization of Trim46 was observed in human HeLa cells, but not in dissociated rat hippocampal neurons or mouse IMCD3 cells (Fig EV1A). The human-specific enrichment of Trim46 at centrosomes is confirmed by two additional antibodies, which detected Trim46 at centrosomes in human iPSC-derived neurons but not in primary rat neurons (Fig EV1B). Previous studies showed that Trim46 is a microtubule-binding protein however, a possible association with centrosomal proteins or structures has not been reported (van Beuningen et al, 2015 ). We sought to identify which centrosome substructure coincided with centrosomal Trim46 structures. Localization experiments by confocal microscopy showed that Trim46 appeared as an oval structure which only partially overlapped with the centriolar and pericentriolar structures marked by Centrin and γ-tubulin, respectively (Fig 1E). To gain more in-depth structural insights, we resolved centrosomal Trim46 structures by STED microscopy and observed that Trim46 appeared as a cloud of punctae surrounding but not overlapping with γ-tubulin structures (Fig 1F). In all cells, we observed a consistent alignment of Trim46 punctae surrounding γ-tubulin, which in turn surrounded Centrin punctae representing the centriolar core. The γ-tubulin structures mark the outer layer of the pericentriolar material as well as the minus-end nucleation sites of microtubules, suggesting that Trim46 localizes near the starting points of centrosomal microtubules (Mennella et al, 2012 ). Together, these data suggest that specifically in human cells, Trim46 localizes to centrosome-associated structures.

Figure EV1. Trim46 localization shifts from centrosomes to axonal and peripheral microtubules in human neurons and glia cells, respectively

  • A. Typical examples of human iPSC-derived NSCs (day 0), HeLa cells, primary dissociated rat neurons (day 0), and mouse IMCD3 cells immunostained for Trim46 and Centrin, and co-immunostained with nestin or α-tubulin if indicated. Inserts represent centrosomes. Scale bar = 10 µm in overview, 2 µm in inserts.
  • B. Typical examples of Trim46 antibodies SySy337003 (#03) and SySy337005 (#05) in hiPSC-derived neurons (day 7) and primary rat neurons (day 7). Inserts represent centrosomes. Scale bar = 20 µm in overview, 1 µm in inserts.
  • C. Typical examples of different maturation stages of human iPSC-derived glia cells (day 17) immunostained for Trim46, Centrin, and α-tubulin. Zooms represent centrosomes. Scale bar = 20 µm in overview, 1 µm in zooms.

Trim46 localization shifts from centrosomes to axonal microtubules during development

Studying the localization of Trim46 over time in human neurons revealed a clear decline of Trim46 staining at the pericentriolar region while the intensity of axonal Trim46 levels increased during development. More specifically, the pericentriolar Trim46 levels remained high in stage 2 neurons and were markedly decreased in stage 3 neurons, following the same trend as found with γ-tubulin (Fig 1G). Consistently, we observed a similar shift of Trim46 staining from centrosomes to dense peripheral microtubules arrays in maturing human iPSC-derived glia cells, which are present at low abundance in the human iPSC-derived neuron cultures (Fig EV1C). Together, these data suggest that Trim46 localization shifts from the pericentriolar region to peripheral axonal microtubule arrays during neurodevelopment.

Centrinone-B treatment depletes centrioles in neuronal stem cells

To study the effect of centrosome dysfunction on axon specification, we next aimed to remove centrioles in human iPSC-derived NSCs by using the pharmacological PLK4 inhibitor Centrinone-B. The efficient and robust centriole loss by Centrinone-B treatment has previously been validated in various other cell types (Wong et al, 2015 ). By inhibiting PLK4, Centrinone-B blocks centriole duplication during cell division in proliferating cells, thereby generating a mixed population of cells containing 0, 1, or 2 centrioles. We treated human iPSC-derived NSCs for 0, 2, or 5 days with Centrinone-B prior to neuronal induction and quantified the number of centrioles per cell. Centrioles were defined as puncta with overlapping staining of Pericentrin and Centrin (Fig 2A). We observed successful removal of either 1 or 2 centrioles in

50% of the NSCs after 2 days of Centrinone-B treatment (Fig 2B). This was not significantly enhanced after a prolonged 5 days of Centrinone-B treatment, likely because cells underwent premature neuronal differentiation and thus terminal cell cycle exit during this time. Centrinone-B treatment increased the number of cells with a characteristic neuron-like morphology, even before inducing neuronal differentiation. This premature neuronal differentiation phenotype is a well-described hallmark of centrosome dysfunction in various in vivo or 3D in vitro systems, and it underlies microcephaly and other neurodevelopmental disorders (Lancaster et al, 2013 ). We found significantly more neurons upon 2 days of Centrinone-B treatment, measured as the relative number of cells that were positive for neuron differentiation markers MAP2 or β3-tubulin, or proliferation marker Ki67 (Fig EV2A–E-EV2A–E). Together, these data show that Centrinone-B treatment results in successful depletion of centrioles in human iPSC-derived neuronal cells, and that it recapitulates neuronal developmental phenotypes that are broadly associated with centrosome defects.

Figure 2. Centriole loss in NSCs perturbs subsequent axonal Trim46 targeting and action potential maturation

  • A. Typical examples of Centrinone-B-treated or control human iPSC-derived NSCs immunostained for Pericentrin and Centrin. Inserts represent centriole(s). Scale bar = 5 µm in overview, 2 µm in inserts.
  • B. Quantifications of the percentage of cells with 0, 1, or 2 centrioles per cell after 0 (control), 2 or 5 days Centrinone-B treatment and prior to neuronal induction. n = 48–51 cells in two independent experiments.
  • C. Typical examples of Centrinone-B-treated or control human iPSC-derived neurons (12–15 days) immunostained for AnkG, Trim46, MAP2, and Centrin. Arrowheads mark AIS structures, arrows mark centrosomes. Inserts represent centrosomes, zooms on the right represent AIS structures. Scale bar = 20 µm in overview, 2 µm in insert, 10 µm in zooms.
  • D. Quantifications of percentage of human iPSC-derived neurons (12–15 days) treated with Centrinone-B containing a Trim46-positive or Trim46-negative process. Neurons are subdivided in populations containing 2 centrioles, or less than 2 centrioles, based on Centrin immunostaining. n = 31–53 cells in three independent experiments.
  • E. Quantifications of percentage of human iPSC-derived neurons (12–15 days) treated with Centrinone-B containing an AnkG-positive or AnkG-negative process. Neurons are subdivided in populations containing 2 centrioles, or less than 2 centrioles, based on Centrin immunostaining. n = 27–32 cells in two independent experiments.
  • F. Quantifications of percentage human iPSC-derived neurons (12–15 days) treated with Centrinone-B containing AnkG-positive processes that are Trim46-positive or Trim46-negative. Neurons are subdivided in populations containing 2 centrioles, or less than 2 centrioles, based on Centrin immunostaining. n = 27 cells in two independent experiments.
  • G. Top: Schematic illustration of the experimental electrophysiology setup. To determine action potential (AP) frequency, somatic current injections from −10 pA to 50 pA (steps of 5 pA, 400 ms) were applied. Bottom: Representative example of evoked AP firing in a Centrinone-B-treated human iPSC-derived neuron, response to hyperpolarizing and first two depolarizing current steps, recorded at day 13.
  • H. Neuronal excitability was recorded in 54/61 control cells and 53/54 Centrinone-B-treated cells. Percentage of cells firing zero, one or multiple APs in control (4 independent experiments no AP: n = 2, single AP: n = 32, multiple APs: n = 22) versus Centrinone-B-treated cultures (3 independent experiments no AP: n = 9, single AP: n = 34, multiple APs: n = 10).
  • I. Representative examples of evoked AP firing in Centrinone-B-treated human iPSC-derived neurons recorded at day 11. Shown is the response to a single depolarizing current step of a neuron that fires no APs, a neuron that fires a single AP and a neuron that fires multiple APs. The offset current (Ihold) was adjusted to keep the baseline membrane potential at approximately −60 mV (dashed lines).
  • J. Scatter plots of AP amplitude versus AP half-width grouped by days after plating for Centrinone-B-treated (10–11 days: n = 14 cells in one independent experiment, 13–14 days: n = 28 cells in three independent experiments) and control (7 days: n = 7 cells, 10–11 days: n = 15 cells in two independent experiments, 13–14 days: n = 36 cells in four independent experiments) human iPSC-derived neurons.
  • K. Phase plots of a representative AP of a human iPSC-derived neuron treated with Centrinone-B and a control neuron of 13 and 14 days, respectively.
  • L. AP half-width recorded in Centrinone-B-treated (n = 43 cells in three independent experiments) and control human iPSC-derived neurons (n = 59 cells in four independent experiments).
  • M. AP threshold, amplitude and after-hyperpolarization recorded in Centrinone-B-treated (n = 43 cells in three independent experiments) and control human iPSC-derived neuron cultures (n = 59 cells in four independent experiments).
  • N. Left top: Schematic representation of the voltage ramp protocol used to determine maximum sodium current membrane potential was changed from −100 mV to 200 mV in 400 ms. Left bottom: Representative example of maximum sodium peak recorded of a control neuron at day 13. Right: Maximum sodium peak in Centrinone-B-treated (n = 52 cells in three independent experiments) and control human iPSC-derived neurons (n = 48 cells in three independent experiments).

Data information: Data represent mean ± SEM. Chi-square test including post hoc analysis with Bonferroni correction (B, D, E, F) Mann–Whitney test (L, M: after-hyperpolarization, N), Student’s t-test (M: threshold, AP amplitude). ***P < 0.001, **P < 0.01, *P < 0.05, ns P ≥ 0.05.

Figure EV2. Centriole loss perturbs axonal Trim46 targeting and action potential maturation

  • A. Typical examples of human iPSC-derived NSCs/neurons (day 3) with 0 (control) or 3 days Centrinone-B treatment, and immunostained for Ki67, β3-tubulin and MAP2. Scale bar = 15 µm in overview.
  • B. Quantifications of the ratio MAP2-positive/Ki67-positive Centrinone-B-treated or control human iPSC-derived NSCs/neurons at day 1, 3, and 5. n = 54–110 cells in two independent experiments.
  • C. Quantifications of the ratio β3-tubulin-positive/Ki67-positive Centrinone-B-treated or control human iPSC-derived NSCs/neurons at day 1, 3, and 5. n = 49–104 cells in two independent experiments.
  • D. Quantifications of the percentage of MAP2-positive Centrinone-B-treated or control human iPSC-derived neurons at day 1, 3, and 5. n = 121–165 cells in two independent experiments.
  • E. Quantifications of the percentage β3-tubulin-positive Centrinone-B-treated or control human iPSC-derived neurons at day 1, 3, and 5. n = 121–165 cells in two independent experiments.
  • F. Quantifications of the percentage of control and Centrinone-B-treated human iPSC-derived neurons (12–14 days) with 0, 1, or 2 centriole(s) and subdivided into populations of neurons containing a Trim46-positive or a Trim46-negative process. n = 30 cells in two independent experiments.
  • G. Typical examples of human iPSC-derived NSCs transduced with CRISPR/Cas9 control or SAS6 KO lentivirus and immunostained for γ-tubulin and Centrin. Inserts represent centriole(s). Scale bar = 5 µm in overview, 2 µm in inserts.
  • H. Quantifications of the percentage of cells with 0, 1, or 2 centrioles per cell after transduction with CRISPR/Cas9 control or SAS6 KO lentivirus at day 5 and prior to neuronal induction. n = 65–147 cells in two independent experiments.
  • I. Typical examples of human iPSC-derived neurons (15 days, neuronal induction at day 5) transduced with CRISPR/Cas9 control or SAS6 KO lentivirus and immunostained for Trim46. Scale bar = 5 µm.
  • J. Quantifications of percentage of human iPSC-derived neurons (day 15, neuronal induction at day 5) containing a Trim46-positive or Trim46-negative process after transduction with CRISPR/Cas9 control or SAS6 KO lentivirus. n = 280–354 cells in two independent experiments.
  • K. Western blot of cell lysates from human iPSC-derived control neurons or Centrinone-B-treated neurons stained for TRIM46 and Actin.
  • L. Quantifications of percentage of human iPSC-derived neurons (day 11) containing a Trim46-positive or Trim46-negative process with or without Centrinone-B treatment after neuronal differentiation (day 5). n = 51 cells in two independent experiments.
  • M. Quantifications of percentage of human iPSC-derived neurons (day 11) containing AnkG-positive processes that are Trim46-positive or Trim46-negative. n = 51–53 cells in two independent experiments.
  • N. AP amplitude recorded in Centrinone-B-treated and control human iPSC-derived neurons of 10–11 days (control: n = 15 cells in two independent experiments +Centrinone-B: n = 14 cells in one independent experiment) and 13–14 days (control: n = 36 cells in four independent experiments, +Centrinone-B: n = 27 cells in three independent experiments).
  • O. AP half-width recorded in Centrinone-B-treated and control human iPSC-derived neurons of 10–11 days (control: n = 15 cells in two independent experiments +Centrinone-B: n = 14 cells in one independent experiment) and 13–14 days (control: n = 36 cells in four independent experiments, +Centrinone-B: n = 27 cells in three independent experiments).
  • P. Resting membrane potential of Centrinone-B-treated (n = 54 cells in three independent experiments) and control (n = 64 cells in four independent experiments) human iPSC-derived neurons.
  • Q. Input resistant of Centrinone-B-treated (n = 54 cells in three independent experiments) and control (n = 63 cells in four independent experiments) human iPSC-derived neurons.
  • R. AP amplitude recorded in human iPSC-derived neurons with or without Centrinone-B treatment after neuronal differentiation (control: n = 17 cells in three independent experiments, +Centrinone post-differentiation: n = 16 cells in three independent experiments).
  • S. AP half-width recorded in human iPSC-derived neurons with or without Centrinone-B treatment after neuronal differentiation (control: n = 17 cells in three independent experiments, +Centrinone post-differentiation: n = 16 cells in three independent experiments).

Data information: Data represent mean ± SEM. Chi-square test including post hoc analysis with Bonferroni correction (D, E, H, J, L, M) Student’s t-test (N: control vs Centrinone-B, 10–11 vs 13–14 days, S) Mann–Whitney test (O: control vs centrinone, 10–11 vs 13–14 days, P, Q, R). ***P < 0.001, **P < 0.01, *P < 0.05, ns P ≥ 0.05.

Figure EV3. Developmental proteome dynamics upon centriole loss

  • A. Correlative matrix of biological replicates of control neurons or Centrinone-B (Cent-B)-treated neurons used for proteome analysis.
  • B. Correlative analysis of relative protein expression in control neurons at day 7 to day 3 and day 1. Specific stem cell markers (blue), neuron markers (yellow), and axon markers (red) are highlighted.
  • C. Correlative analysis of relative protein expression in Centrinone-B-treated neurons at day 7 to day 3 and day 1. Specific stem cell markers (blue), neuron markers (yellow), and axon markers (red) are highlighted.
  • D. Ratios of relative protein expression at day 1, 3, and 7 and calibrated to day 1 for stem cell proteins, neuron proteins, and axon proteins.
  • E. Quantification of the total phalloidin intensity in a fan-like growth cone at day 5 and day 9. n = 71–80 growth cones in two independent experiments.

Data information: Data represent mean ± SEM.

Figure EV4. Centriole loss restrains microtubule remodeling during early axon development

  • A. Kymographs and schematic representations of time-lapse recordings of the proximal axon, for different time points (day 7 day 13) and conditions (control Centrinone-B). Scale bar = 5 µm.
  • B. Quantifications of the percentage of neurons exhibiting uniform, or non-uniform comet orientations in the anterograde direction in the proximal axon at day 7. n = 13–14 cells in four independent experiments.
  • C. Quantifications of the percentage of neurons exhibiting uniform, or non-uniform comet orientations in the anterograde direction in the proximal axon at day 13. n = 12–17 cells in four independent experiments.
  • D. Quantifications of the number of comets per minute moving in the anterograde direction in the proximal axon at day 7 and day 13. n = 12–17 cells in four independent experiments.
  • E. Quantifications of the number of comets per minute moving in the retrograde direction in the proximal axon at day 7 and day 13. n = 12–17 cells in four independent experiments.
  • F. Kymographs and schematic representations of time-lapse recordings of the dendrite, for different time points (day 7 day 13) and conditions (control Centrinone-B). Scale bar = 5 µm.
  • G. Quantifications of the percentage of neurons exhibiting uniform, or non-uniform comet orientations in the anterograde direction in the dendrite at day 7. n = 12–13 cells in four independent experiments.
  • H. Quantifications of the percentage of neurons exhibiting uniform, or non-uniform comet orientations in the anterograde direction in the dendrite at day 13. n = 13–16 cells in four independent experiments.
  • I. Quantifications of the number of comets per minute moving in the anterograde direction in the dendrite at day 7 and day 13. n = 12–16 cells in four independent experiments.
  • J. Quantifications of the number of comets per minute moving in the retrograde direction in the dendrite at day 7 and day 13. n = 12–16 cells in four independent experiments.
  • K. Kymographs and schematic representations of time-lapse recordings of the proximal axon following LS, for different time points (day 7 day 13) and conditions (control Centrinone-B). Red line and red arrowhead denote location and time of LS. Scale bar = 5 µm.
  • L. Quantifications of the percentage of neurons exhibiting uniform, or non-uniform comet orientations in the anterograde direction in the proximal axon following LS at day 7. n = 20–21 neurons in three independent experiments.
  • M. Quantifications of the percentage of neurons exhibiting uniform, or non-uniform comet orientations in the anterograde direction in the proximal axon following LS at day 13. n = 22–25 neurons in three independent experiments.
  • N. Quantifications of the number of comets per minute moving in the anterograde direction in the proximal axon following LS at day 7 and day 13. n = 20–25 cells in three independent experiments.
  • O. Quantifications of the number of comets per minute moving in the retrograde direction in the proximal axon following LS at day 7 and day 13. n = 20–25 cells in three independent experiments.
  • P. Kymographs and schematic representations of time-lapse recordings of the dendrite following LS, for different time points (day 7 day 13) and conditions (control Centrinone-B). Red line and red arrowhead denote location and time of LS. Scale bar = 5 µm.
  • Q. Quantifications of the number of comets per minute moving in the anterograde direction in the dendrite following LS at day 7 and day 13. n = 20–26 cells in three independent experiments.
  • R. Quantifications of the number of comets per minute moving in the retrograde direction in the dendrite following LS at day 7 and day 13. n = 20–26 cells in three independent experiments.
  • S. Quantifications of the percentage of neurons exhibiting uniform, or non-uniform comet orientations in the anterograde direction in the distal axon at day 13 for different conditions (control Centrinone-B addition post-differentiation at day 5). n = 21–23 cells in three independent experiments.
  • T. Quantifications of the number of comets per minute pointing in the anterograde direction in the distal axon at day 13 for different conditions (control Centrinone-B addition post-differentiation at day 5). n = 21–23 cells in two independent experiments.
  • U. Quantifications of the number of comets per minute pointing in the retrograde direction in the distal axon at day 13 for different conditions (control Centrinone-B addition post-differentiation at day 5). n = 21–23 cells in two independent experiments.

Data information: Data represent mean ± SEM. Chi-square test (B, C, G, H, L, M, S), unpaired t-test (D, E, I, J, N, O, Q, R, T, U) **P < 0.005, *P < 0.05, ns P ≥ 0.05.

Centriole loss perturbs axonal targeting of Trim46

We next asked whether these centriole-deprived and prematurely differentiated neurons upon Centrinone-B treatment follow normal developmental timing and grow functional axons. To investigate the role of centrosomes during axon specification, we assessed if the development of early-stage axons was affected by centriole loss. An important aspect of early-stage axon development is the specific sorting of axonal proteins, including the AIS protein Trim46. Thus, we tested if the axon-specific localization of Trim46 was affected by Centrinone-B-induced centriole removal. Neurons were immunostained with Centrin and Trim46 to correlate centriole number and the presence of axonal Trim46 for each cell (Fig 2C). We observed a marked

50% reduction of cells with a Trim46-positive process in the subpopulation of cells containing ≤ 1 centriole(s) compared to cells still containing 2 centrioles upon Centrinone-B treatment (Fig 2D). Accordingly, Trim46-negative processes were more often observed in centriole-depleted neurons compared to neurons still containing both centrioles upon Centrinone-B treatment, whereas this distinction was not observed in control (Fig EV2F). To control for possible off-target effects of Centrinone-B, we used an additional approach to induce centriole loss based on genetic manipulation of SAS6, an important regulator of centriole duplication and a component of the centriole cartwheel structure (Nakazawa et al, 2007 ). Transducing human iPSC-derived NSCs with CRISPR/Cas9 SAS6 gRNA knockout lentivirus (SAS6 KO) for five days resulted in a significant reduction of centriole numbers (Fig EV2G and H). Axonal appearance of Trim46 was significantly reduced in neurons transduced with SAS6 KO lentivirus, thereby confirming our previous finding (Fig EV2I and J). Centriole depletion did not seem to affect the overall Trim46 expression levels in neurons, as no clear differences were found upon Centrinone-B treatment with Western blot analysis (Fig EV2K). Another important protein that undergoes axonal sorting is the major AIS scaffold AnkG. In contrast to Trim46, we observed no changes in axonal appearance of AnkG upon Centrionone-B-induced centriole depletion (Fig 2E). Consistently, cells containing ≤ 1 centriole(s) showed reduced Trim46 co-localization at AnkG-positive axonal structures (Fig 2F). To control for possible post-differentiation effects of PLK4 inhibition unrelated to centriole number, we investigated if axonal targeting of Trim46 was affected when applying Centrinone-B treatment after neuronal differentiation. Treating neurons with Centrinone-B three days after neuronal induction, which we previously reported as a time point in which most neurons are differentiated but not yet polarized, did not affect Trim46 appearance at axons or at AnkG-positive structures at later stages (Fig EV2L and M) (Lindhout et al, 2020 ). Together, these data suggest that centrosomes are important for the targeting of Trim46, but not AnkG, to axons during early stages of neuronal development.

Centriole loss leads to immature action potential firing

The axonal targeting of Trim46 and AnkG is required to assemble the AIS, the highly specialized structure essential for mature and efficient AP firing. Thus, we assessed if the observed differential effects on axon protein targeting upon centriole depletion correlate with changes in AP properties. We performed whole-cell patch clamp recordings of control or Centrinone-B-treated neurons of 7–14 days, which coincides with early axon development. To measure neuronal excitability, we determined the number of APs fired with increasing somatic current injection (steps of 5 pA 400 ms) (Fig 2G). In Centrinone-B-treated cultures,

17% of neurons did not fire APs, whereas this was only

3% in control cultures (Fig 2H and I). Of the firing neurons, there were less Centrinone-B-treated neurons that fired multiple APs compared to control. Neurons that did not fire APs did generate small peaks upon current stimulation, indicating the opening of sodium channels, but did not generate a positive feedback to rapidly increase the membrane potential as is characteristic of APs. In addition, neurons treated with Centrinone-B did not display a progressive maturation of AP properties from day 10 to day 14, as was observed in control neurons (Figs 2J, and EV2N and O). APs fired by Centrinone-B-treated neurons appeared more immature, as they were wider, had smaller amplitudes and smaller after-hyperpolarizations (Figs 2K–M, and EV2N and O). In Centrinone-B-treated neurons, the input resistance was also higher, but membrane potential did not differ from control (Fig EV2O–Q-EV2O–Q). Although AP threshold was not affected by Centrinone-B treatment (Fig 2M), maximum sodium currents were significantly smaller in Centrinone-B-treated neurons (Fig 2N). To control for centriole-unrelated effects of PLK4 inhibition, we tested the effect of Centrinone-B treatment after differentiation on action potential firing. When Centrinone-B was added post-differentiation, we did not observe a difference in AP amplitude and half-width in neurons of 13–14 days (Fig EV2R and S). Together, the electrophysiology recordings from Centrinone-B-treated neurons show more immature AP firing and reduced sodium currents, thereby highlighting the functional relevance of centrosome-mediated control mechanisms during early stages of neuronal development.

Centriole depletion results in reduced expression of growth cone proteins

Our observations showed that the centriole-depleted NSCs develop into neurons with structural and functional perturbations in axon development. Next, we aimed to quantify effects of centriole depletion during axon specification with unbiased profiling. Therefore, we performed mass spectrometry-based quantitative proteomics analysis on days 1, 3, and 7, which generally corresponds with developmental stage 1, onset of stage 2, and onset of stage 3, respectively (Lindhout et al, 2020 ). We compared the proteome dynamics during early neurodevelopment of replicates of Centrinone-B-treated and control neurons (Fig EV3A, DATASET EV1). Centrinone-B treatment did not markedly alter the relative protein expression over time (Fig 3A). The protein expression profile of control neurons showed a developmental shift, which corresponds to the transitions from stage 1 to stage 2, and from stage 2 to stage 3 (Fig EV3B). Protein expression of centriole-depleted neurons largely follows the same trend (Fig EV3C). In both populations, proteins considered specific for NSCs are downregulated at days 3 and 7 (e.g., Ki67, nestin, Otx, Notch1), whereas neuronal proteins are upregulated (e.g., Stathmin1, Map2, doublecortin, Tubß3) (Fig EV3D). The onset of stage 3 marks axon specification, and indeed, we observed strong upregulation of the axonal proteins Trim46 and Tau at day 7, which was not affected by Centrinone-B treatment (Fig 3B). Expression of the growth cone proteins Basp1, Gap43, and Marcks was increased at day 3 in controls as well as Centrinone-B-treated neurons. This upregulation was even stronger at day 7 in control neurons, but was markedly suppressed in the centriole-depleted neurons (Figs 3C and EV3D). Together, the quantitative proteome analysis shows that treatment with Centrinone-B does not dramatically alter the protein expression profile during early stages of neurodevelopment, providing a successful global quality control of Centrinone-B-treated cells. However, these results indicate a specific effect on growth cone proteins upon depletion of centrioles.

Figure 3. Centriole loss in NSCs is accompanied with changes in neurite growth cone morphology

  • A. Correlative plot of changes in protein abundance between control neurons and neurons treated with Centrinone-B (day 7/day 3). Pearson’s correlation R = 0.5716, P < 0.001. Specific, significantly changing growth cone proteins highlighted in red.
  • B. Protein abundance profile over time for axon-related proteins Trim46 and Tau in control neurons and neurons treated with Centrinone-B (+C).
  • C. Protein abundance profile over time for growth cone-related proteins Basp1, Gap43, and Marcks in control neurons and neurons treated with Centrinone-B (+C).
  • D. Representative images of fan-like growth cones at day 5 of control and Centrinone-B-treated neurons with or without Nocodazole treatment. Growth cones are visualized by immunostaining for phalloidin. Scale bar = 5 µm.
  • E. Quantifications of the average area (µm 2 ) of growth cones of control and Centrinone-B-treated neurons, with or without Nocodazole treatment, at different time points. n = 25–83 growth cones in three independent experiments.
  • F. Representative images of different growth cone morphological categories: fan-like, torpedo-like and bulb-like. Scale bar = 5 µm.
  • G. Quantifications of the ratios of different subtypes (fan-like, torpedo-like, bulb-like) of growth cones of control and Centrinone-B-treated neurons, with or without Nocodazole treatment, at different time points.
  • H. Quantifications of the average area (µm 2 ) of different subtypes (fan-like, torpedo-like, bulb-like) of growth cones of control and Centrinone-B-treated neurons, with or without Nocodazole treatment, at different time points. n = 5–55 growth cones in three independent experiments.

Data information: Data represent mean ± SEM. One-way ANOVA including Tukey’s post hoc analysis (E), Chi-square test including post hoc analysis with Bonferroni correction (G), One-way ANOVA including Sidak’s post hoc analysis (H). ***P < 0.001, **P < 0.005, *P < 0.05, ns P ≥ 0.05.

Centriole depletion affects neurite growth cone morphology

Next, we examined if the reduced expression of growth cone proteins resulted in defects of axonal growth cone morphology. We found that in control neurons the size of growth cones is relatively large early in development and decreases over time (Fig 3D and E). Growth cones of neurons treated with Centrinone-B remained smaller at day 5, and their size decreased even further later in development. As microtubules are essential components to shape growth cones, we investigated whether manipulation of the microtubule cytoskeleton mimicked the effect of centriole depletion by Centrinone-B treatment on growth cone size (Dent et al, 2011 ). Indeed, treatment with Nocodazole, a microtubule destabilizer, also resulted in smaller growth cones already at day 5 (Fig 3D and E). Interestingly, Nocodazole treatment did not show an additional effect on growth cone size in centriole-depleted neurons, which suggests similar underlying mechanisms. To study the effect of centriole depletion on growth cones more specifically, we categorized three subtypes: fan-like, torpedo-like, and bulb-like (Fig 3F) (van der Vaart et al, 2013 ). We found a majority of fan-like growth cones at day 5, which shifted to more torpedo-like and bulb-like growth cones as axons matured (Fig 3G). Neither Centrinone-B nor Nocodazole treatment altered the relative abundance of these types of growth cones. However, Centrinone-B as well as Nocodazole treatment did result in significantly smaller fan-like growth cones at day 5, which in control neurons are considerably larger than torpedo- and bulb-like growth cones (Fig 3H). This effect was distinct for fan-like growth cones, as the sizes of torpedo-like and bulb-like growth cones were unaffected, which suggests that the observed decrease of growth cone size is specifically due to affected fan-like growth cones (Fig 3E and H). The actin cytoskeleton is another important cytoskeletal component at growth cones and was previously found to be controlled by microtubules as well as centrosome activity in dissociated rodent neurons (Zhao et al, 2017 Meka et al, 2019 ). Here, we observed no changes in local levels of F-actin at growth cones, suggesting that the reduced size of fan-like growth cones is not caused by changes in F-actin levels in growth cones (Fig EV3E). Together, these results suggest that centriole depletion causes growth cone morphology defects during axon outgrowth, presumably through microtubule-mediated mechanisms.

Centriole depletion perturbs microtubule reorganization in developing axons

To gain more insight into the potential role of centrosomes on the unique organization of the axonal microtubule cytoskeleton, marked by a uniform organization of plus-end out microtubules, we studied the effect of centriole loss on the microtubule remodeling processes during development. We systematically analyzed plus-end dynamics and orientations of microtubules in axons and dendrites on day 7 and 13, which coincide with the onset of stage 3 and late stage 3, respectively (Lindhout et al, 2020 ). For the analysis at axons, we measured both proximal and distal regions, as we previously reported that axon development follows a distal-to-proximal reorganization during this time window in human iPSC-derived neurons (Lindhout et al, 2020 ). We used two-color live-cell imaging to visualize neurite morphology and microtubule plus-end tracking proteins (MT+TIPs) in control neurons and Centrinone-B-treated neurons (Fig 4A, MOVIE EV1). The direction of moving MT+TIPs is used as read-out for microtubule orientations, as anterograde and retrograde movement of MT+TIPs mark plus-end out and minus-end out microtubules, respectively. On day 7, the percentage of Centrinone-B-treated neurons with a uniform microtubule polarity, which is a hallmark of mature axons, is similar to control neurons in both proximal and distal axons (Figs 4B and C, and EV4A and B). Similarly, the numbers of anterogradely and retrogradely growing MT+TIPs are not changed upon Centrinone-B treatment (Figs 4D and E, and EV4D and E). However, on day 13 the proportion of neurons with uniform microtubules in the distal axon was significantly reduced by Centrinone-B treatment compared to control neurons (Fig 4B and F). Microtubule polarity was specifically affected in the distal axon, which represents the most mature axonal stage, as no difference was observed in the proximal axons and the dendrites of control neurons and Centrinone-B neurons (Fig EV4, EV4, EV4, EV4, EV4, EV4, EV4, EV4A–C,F–H). This indicates that centriole depletion prevents neurons from retaining the characteristic uniform plus-end out microtubule organization during axonal development. Furthermore, Centrinone-B-treated neurons show more MT+TIP movement, in particular for anterograde moving MT+TIPs, which suggests a global increase of microtubule dynamics (Fig 4G and H). An increase in the number of MT+TIPs is also observed in the proximal axon at day 13, but not in the dendrites (Fig EV4D,E,I,J). The speed and run length of growing MT+TIPs was consistent between control neurons and Centrinone-B-treated neurons at dendrites as well as proximal and distal axons across both time points (DATASET EV2). Imaging of MT+TIPs only provides information about the dynamic ends of microtubules, but does not account for stabilized microtubules. Thus, we next aimed to analyze the microtubule orientations of the total axonal microtubule network, including both stable and dynamic microtubules. This was addressed by combining our approach with laser severing to generate new microtubule ends by cutting microtubules with a short-pulsed laser, which triggers newly formed MT+TIPs (Fig 4I and J, MOVIE EV2) (Yau et al, 2016 ). Consistent with previous findings, we observed a marked

40% reduction of neurons with a uniform plus-end microtubule organization in distal axons with Centrinone-B treatment following laser severing of microtubules, whereas no significant changes were observed in proximal axons (Figs 4K,L,O, and EV4K,L,M). At distal axons, the number of retrograde comets was significantly increased upon Centrinone-B treatment, whereas anterograde comets were unaffected, suggesting that centriole loss is accompanied with more stable minus-end out microtubules at developing axons (Fig 4M,N,P,Q). As expected, the number of anterograde and retrograde comets was not changed at proximal axons and dendrites with Centrinone-B treatment (Fig EV4N,O,P,Q,R). To control for PLK4 inhibition effects in postmitotic neurons unrelated to centriole number, we analyzed the microtubule organization in distal axons at day 13 in neurons treated with Centrinone-B after differentiation and observed no significant differences compared to control (Fig EV4S,T,U). Together, these data suggest that centrosomes are important for the unique axon-specific reorganization toward uniform plus-end out microtubules (Fig 4R).

Figure 4. Centriole loss restrains microtubule remodeling during early axon development

  • A. Example stills from a spinning-disk time-lapse recording of a neurite transfected with mRFP and GFP-MT+TIP. The top panel is a still of a neurite in mRFP, showing neurite morphology. The other panels show moving GFP-MT+TIP comets pointing in either an anterograde direction (green arrowheads) or retrograde direction (blue arrowheads). P indicates the proximal direction and D the distal direction of the neurite. Timestamp in minutes:seconds given on bottom right. Scale bar = 5 µm.
  • B. Kymographs and schematic representations of time-lapse recordings of the distal axon as shown in (A), for different time points (day 7 day 13) and conditions (control Centrinone-B). Scale bar = 5 µm.
  • C. Quantifications of the percentage of neurons exhibiting uniform, or non-uniform comet orientations in the anterograde direction in the distal axon at day 7. n = 15–17 cells in four independent experiments.
  • D. Quantifications of the number of comets per minute moving in the anterograde direction in the distal axon at day 7. n = 15–17 cells in four independent experiments.
  • E. Quantifications of the number of comets per minute moving in the retrograde direction in the distal axon at day 7. n = 15–17 cells in four independent experiments.
  • F. Quantifications of the percentage of neurons exhibiting uniform, or non-uniform comet orientations in the anterograde direction in the distal axon at day 13. n = 11–18 cells in four independent experiments.
  • G. Quantifications of the number of comets per minute moving in the anterograde direction in the distal axon at day 13. n = 11–18 cells in four independent experiments.
  • H. Quantifications of the number of comets per minute moving in the retrograde direction in the distal axon at day 13. n = 11–18 cells in four independent experiments.
  • I. Schematic representation of microtubule laser-severing (LS) experiments.
  • J. Example stills from a spinning-disk time-lapse recording of a neurite transfected with mRFP and GFP-MT+TIP. Red line denotes the location of LS. Scale bar = 5 µm.
  • K. Kymographs and schematic representations of time-lapse recordings of the distal axon as shown in (J) following LS, for different time points (day 7 day 13) and conditions (control Centrinone-B). Red line and red arrowhead denote location and time of LS. Scale bar = 5 µm.
  • L. Quantifications of the percentage of neurons exhibiting uniform, or non-uniform comet orientations in the anterograde direction in the distal axon following LS at day 7. n = 20–22 neurons in three independent experiments.
  • M. Quantifications of the number of comets per minute moving in the anterograde direction in the distal axon following LS at day 7. n = 20–22 cells in three independent experiments.
  • N. Quantifications of the number of comets per minute moving in the retrograde direction in the distal axon following LS at day 7. n = 20–22 cells in three independent experiments.
  • O. Quantifications of the percentage of neurons exhibiting uniform, or non-uniform comet orientations in the anterograde direction in the distal axon following LS at day 13. n = 24–27 neurons in three independent experiments.
  • P. Quantifications of the number of comets per minute moving in the anterograde direction in the distal axon following LS at day 13. n = 24–27 cells in three independent experiments.
  • Q. Quantifications of the number of comets per minute moving in the retrograde direction in the distal axon following LS at day 13. n = 24–27 cells in three independent experiments.
  • R. Schematic representation of the proposed orientation of microtubules in the distal axon in control conditions, and following Centrinone-B treatment.

Data information: Data represent mean ± SEM. Chi-square test (C, F, L, O), unpaired t-test (D, E, G, H, M, N, P, Q) **P < 0.005, *P < 0.05, ns P ≥ 0.05.


Centrosome maintenance

Historically, centrioles have been regarded as exceptionally stable structures. They are resistant to drug- and cold-induced MT depolymerization (Kochanski and Borisy, 1990), to forces and MT instability in mitosis (Belmont et al., 1990). Furthermore, fluorescence recovery after photobleaching (FRAP) of the basal body α-tubulin in Tetrahymena shows little turnover (Pearson et al., 2009). An elegant experiment in C. elegans demonstrated that, upon fertilization with sperm containing a single paternally contributed centriole that has been labeled with SAS4 tagged to GFP to mark the centriolar walls, could be detected up to the ∼350-cell stage after fertilization (Balestra et al., 2015), suggesting centrioles are stably inherited through many divisions.

However, centrosomes are lost from oocytes of most metazoan species (Delattre and Gönczy, 2004 Manandhar et al., 2005) and are known to be inactivated (i.e. loss of their MTOC capacity) in some cell types that undergo differentiation, such as neuronal, muscle and epithelial cells (Sanchez and Feldman, 2017). Upon neuronal differentiation in mammals and Drosophila, centrosomes lose PCM proteins and, consequently, their MTOC capacity (Stiess et al., 2010 Nguyen et al., 2011). Axon extension can occur in the absence of active centrosomes in mammals and in Drosophila (Tassin et al., 1985 Stiess et al., 2010 Nguyen et al., 2011). In myocyte differentiation, centrosomes lose PCM proteins and were described as absent from muscle fibers (Srsen et al., 2009 Przybylski, 1971). At the same time, PCM proteins accumulate at the nuclear periphery from which MTs are nucleated (Srsen et al., 2009 Przybylski, 1971). Several epithelial cell types also inactivate MT nucleation and/or abolish their anchoring from centrosomes and so generate MTs along the apical–basal axis (Sanchez and Feldman, 2017 Muroyama and Lechler, 2017). These lines of evidence suggest that the centrosome is under a homeostatic maintenance program that can be regulated, thereby giving rise to different MT arrays.

Moreover, several centriole components are dynamic, such as centrin in the lumen of the centriole (Bahmanyar et al., 2010), spindle assembly abnormal protein 6 (SAS6) in the cartwheel (Keller et al., 2014) and centrosomal protein 120 (CEP120) at the centriolar wall (Mahjoub et al., 2010). Therefore, a picture is emerging whereby a general homeostatic maintenance program (Box 1) exists for centrosomes that might underlie both their stability in cycling cells (Izquierdo et al., 2014) and their instability to a certain extent in some tissues, such as oocytes, neurons and epithelial cells (Pimenta-Marques et al., 2016 Yonezawa et al., 2015 Muroyama et al., 2016).

The centrosome homeostatic maintenance program depends on critical aspects of its structure, such as the PCM, the centriole walls and the centriole cartwheel, and is under the control of cell cycle regulators, such as CDKs and PLKs (Yang and Feldman, 2015 Muroyama et al., 2016). These kinases are known to regulate centrosome biogenesis, maturation and function they are degraded or inactivated, respectively, by the anaphase-promoting complex, also known as the cyclosome (APC/C) at the end of mitosis (Ferris et al., 1998 Glotzer et al., 1991 Lindon and Pines, 2004 Murray, 1989). A high activity of CDKs and PLKs is needed for an active mitotic centrosome, whereas a lower activity of those kinases is often associated with centrosomes of cells in interphase or those that have exited the cell cycle. Finally, no activity of these kinases is observed when the centrosome has been fully inactivated, which is often associated with centriole loss (Fig. 2A).

We next discuss in more detail different components of the centrosome maintenance program.

In human cultured cells, newly formed centrioles that have been blocked from maturing into centrosomes by removal of CEP295 disassemble at the end of the cycle upon cartwheel loss (Izquierdo et al., 2014). Inhibition of PLK1 retained the cartwheel (Wang et al., 2011) and rescued the loss of non-matured centrioles (Izquierdo et al., 2014). Matured centrioles, however, even though they normally lose the cartwheel at the end of the cell cycle, do not disassemble because they have PCM. This suggested that both the PCM and cartwheel are redundant in conferring stability to centrioles and compensate for each other in centrosome protection. Such a redundancy might not exist in all species and/or tissues. In S-phase-arrested Drosophila cells, depletion of four major PCM proteins (SPD-2, CNN, Asl and D-PLP) or of Polo (the ortholog of PLK1), was sufficient to lead to a reduction in centriole number, demonstrating that centrosomes are maintained homeostatically through the renewal of their components (Pimenta-Marques et al., 2016).

Centriole wall components and their post-translational modifications are also important for centriole stability. Injection or electroporation with an antibody against tubulin glutamylation (α-GT335) resulted in the disappearance of centrioles and centrosomes (Bobinnec et al., 1998). In this case, centrioles and discrete centrosomes ultimately reappeared in the cell population however, some centrioles exhibited loss of the MT triplets (Bobinnec et al., 1998), which are characteristic of normal centrioles (Fig. 1A). It is possible that glutamylation itself stabilizes the centriolar MT structure or promotes the binding of stabilizers. On that note, it was shown that downregulation of ATF5, which interacts with both glutamylated tubulin and pericentrin, thereby linking the centriole to the PCM, blocks the accumulation of PCM at the centrosome and causes the fragmentation of centrioles (Madarampalli et al., 2015). Centrobin, another factor that binds to centriolar tubulin and is normally associated with daughter centrioles (Gudi et al., 2011 Zou et al., 2005), promotes centriole elongation and prevents PCM recruitment in cultured cycling cells. Expression of dominant-negative centrobin led to an increase in the number of cells without centrioles (Gudi et al., 2011). Finally, in certain species, δ- and ε-tubulins found on centrioles contribute to the formation and/or stability of the triplet MTs (reviewed in Winey and O'Toole, 2014). All together, these studies suggest that there is an interplay between the cartwheel, the centriole walls (including the post-translational modifications of tubulin and centrobin function) and the PCM in supporting centriole stability.

Recent studies show that inactivation of the PCM leads to scheduled centrosome inactivation (e.g. in neurons, muscle and epithelial cells) or even their entire disappearance (e.g. in oocytes, Fig. 2A). Originally observed by Huettner in 1933, it is now known that oocytes of multiple species lose their centrosomes during meiosis (Huettner and Rabinowitz, 1933 Manandhar et al., 2005), which is achieved in different ways in different organisms (see Box 2). In Drosophila, during early oogenesis, a cyst of 16 interconnected cells is formed of these, one becomes the oocyte and inherits all centrioles by intercellular centriole migration, which results in a large MTOC consisting of 64 centrioles. Polo and some PCM components such as SPD-2 are transcriptionally downregulated before centrioles disappear (Jambor et al., 2015). This is correlated with loss of Polo and PCM proteins from the oocyte MTOC, followed by centriole disappearance before the egg divides (Pimenta-Marques et al., 2016). Forced localization of Polo to centrioles in oogenesis resulted in PCM maintenance and, consequently, persistence of the centrioles (Pimenta-Marques et al., 2016) (Fig. 2A). These findings in oocytes point to the importance of the active recruitment of newly synthesized Polo and PCM components to the centrosome, further supporting the existence of a regulated homeostatic maintenance program that can be switched off.

Centrosomes are eliminated from the oocytes of the majority of metazoan species (Manandhar et al., 2005), allowing for a correct number of centrosomes to be attained after fertilization. In fruit flies, worms and humans, centrioles are eliminated prior to meiotic division, one of the few acentriolar divisions in those species (Delattre and Gönczy, 2004 Manandhar et al., 2005 Cunha-Ferreira et al., 2009 Mikeladze-Dvali et al., 2012). In Drosophila, a centrosome maintenance program was elucidated that depends on Polo and PCM, which is turned off during oogenesis (Pimenta-Marques et al., 2016 see main text for discussion). In some species, centrioles are eliminated together with DNA through their extrusion inside the polar bodies during meiotic divisions, such as in snail oocytes, which only contain a single pair of centrioles (Krioutchkova et al., 1994). Interestingly, echinoderms (sea urchin, starfish and sea-cucumber) use both strategies to eliminate centrioles: extrusion and elimination. These species enter meiosis with two pairs of centrioles, one at each pole of meiosis I spindle. One pair is extruded through the polar body I (PBI). Subsequently, single centrioles are present on the spindle poles of meiosis II of these, the mother, which has MT nucleation capacity, is extruded with PBII (Borrego-Pinto et al., 2016), leaving a single centriole in the mature egg (Kato et al., 1990 Miyazaki et al., 2005 Nakashima and Kato, 2001). The remaining centriole is eventually eliminated, but it is not known exactly how this is achieved. If centrioles are artificially retained, they cannot be inactivated, resulting in multipolar zygotic spindles (Borrego-Pinto et al., 2016). The retained daughter centriole does not nucleate MT, perhaps because of insufficient levels of PCM, which may cause centriole destabilization. Perhaps mother and daughter centrioles have to be eliminated in different ways, because these cells have no mechanism to actively remove PCM. Future studies are needed to understand whether different species use similar mechanisms to inactivate centrosome maintenance programs and achieve centrosome elimination.

In the human body, there are several other examples of cells, in which centrosomes are either partially or completely inactivated. In some of these cases centrioles persist, while in others there is no conclusive evidence. For example, during differentiation of skeletal muscle, centrioles are inactivated upon fusion of myoblasts to give rise to the syncytial myotubes. Here, proteins such as γ-tubulin, pericentrin and ninein are captured by nesprin at the nuclear envelope, from which MTs are nucleated and extend, resulting in the formation of longitudinal MT bundles along the long axis of the cell (Tassin et al., 1985 Espigat-Georger et al., 2016). In differentiating hippocampal neurons, centrosome inactivation is associated with loss of γ-tubulin, pericentrin and centrin from the centrosome (Stiess et al., 2010). MTs are generated by augmin- and γTuRC-dependent nucleation from existing MTs. This ensures a uniform plus-end-out MT polarity in axons (Sánchez-Huertas et al., 2016). However, in both muscles and neurons, is not clear whether centrioles eventually disappear and what would happen if centrosome activity was maintained.

During epithelial differentiation, centrosomes often cease to be the major MTOC in the cell when acentrosomal MTOCs are established. Loss of CDK1 activity appears to be a major trigger for this change (Muroyama et al., 2016). An interesting example is proliferative basal cells of the mammalian epidermis here, MTs are recruited to the cell cortex upon differentiation. Loss of CDK1 activity upon exit from the cell cycle results in several changes at centrosomes and in the cell. MTs continue to be nucleated by CDK5RAP2–γ-TuRC complexes at the centrosome, whereas Nedd1–γ-TuRC complexes, which are required for MT anchoring in this system, rapidly delocalize from centrosomes, leading to a loss of astral MT configuration (Fig. 2A) (Muroyama, et al., 2016). This study suggests that different populations of γ-TuRCs have distinct functions and are regulated differently. Loss of centrosomal MTOC activity in these cells is associated with loss of pericentrin and γ-tubulin from centrosomes, but centrioles are not completely eliminated (Fig. 2A) (Muroyama et al., 2016). Similarly, during cell differentiation of C. elegans embryonic intestinal cells, MTOC function is reassigned to the apical membrane after downregulation of CDK-1. Interestingly, in this case, cells can divide after differentiation. Reactivation of the centrosomal MTOC is dependent on the conserved centrosome protein spindle-defective protein 2 (SPD-2 CEP192 in humans) and mitotic CDK activity (Yang and Feldman, 2015).

Taken together, it is possible that kinases, such as PLK1 and CDK1, function both as regulators of centrosome activity and maintenance, which makes it difficult in some cases to establish clear boundaries between both processes (Muroyama et al., 2016). Upregulation of PLK1 and CDK1 in mitosis leads to centrosome maturation and increased MT nucleation. However, their presence in interphase and in many differentiated cells is necessary for centrosome maintenance and even nucleation, with their absence leading to centrosome inactivation and disappearance in Drosophila (Fig. 2A). Future work will hopefully unravel how this program is regulated at the transcriptional and post-transcriptional level and how dynamic the PCM and centriole proteins actually are.


Cilia maintenance

Different organisms and cell types are likely to differently regulate the structural and functional maintenance of cilia. Although certain long-lived differentiated cells harbor stable cilia such as photoreceptors, cilia are known to assemble and disassemble in cycling cells (Fig. 2B). The photoreceptor outer segment (a modified cilium) is completely replaced in a matter of days and hence requires continuous maintenance (Besharse and Hollyfield, 1979 Hsu et al., 2017). The integrity of the cilium, both as a structural and signaling compartment, is critical for its function (Fig. 2B). It is likely that modifications that make the structure more robust, as well as the synthesis of ciliary components and their transport into the cilia, are all important factors for their maintenance (see also Box 3). As for centrioles, post-translational modifications of tubulin play a role in stabilizing axonemal MTs in cilia. Mutation in the tubulin deglutamylase CCPP-1 in C. elegans leads to a progressive degeneration phenotype of the axonemal MTs (O'Hagan et al., 2011). These worms appear to form normal cilia in early larval stages, but cilia defects arise over time (O'Hagan et al., 2011). Tubulin glycylation is not required for retina development in the mouse, but the photoreceptors degenerate with age if tubulin glutamylation and glycylation are not properly balanced (Bosch Grau et al., 2017) (Fig. 2B). The authors propose that this phenotype is linked to the inability of the photoreceptor to appropriately adapt to mechanical load (Bosch Grau et al., 2017).

Important insights into homeostatic cilia maintenance came from the bi-flagellated green alga Chlamydomonas reinhardtii. Chlamydomonas flagella (hereafter referred to as cilia) can be biochemically isolated (Witman et al., 1972) and since the late 1970s, a collection of temperature-sensitive mutants has been available that allow to acutely disassemble cilia (Huang et al., 1977). These tools led to the identification of genes and mechanisms involved in cilia maintenance. In particular, it was shown that a previously described intra-ciliary ‘motility’ (Kozminski et al., 1993) is dependent on kinesin-driven motor movement (Kozminski et al., 1995). This motility is now called intraflagellar transport (IFT) and is known to consist of a protein complex of 22 subunits (Piperno et al., 1998 Taschner and Lorentzen, 2016 Vashishtha et al., 1996). Acute removal of IFT leads to cilia resorption, which demonstrated that this complex is important for cilia maintenance. Live-imaging and modeling revealed that cilia length is dynamically controlled through tubulin turnover (Marshall and Rosenbaum, 2001 Marshall et al., 2005). Blocking IFT prevents the addition of new tubulin at the tip and causes cilia shortening (Marshall and Rosenbaum, 2001). However, only 12 years later, it was demonstrated that tubulin is in fact a bona fide IFT cargo in several species (Bhogaraju et al., 2013). Core IFT components are conserved in other species, and when mutated, can cause phenotypes similar to those seen in human pathologies (Pazour et al., 2000, 2002). Maintenance of protein composition in cilia, however, goes beyond tubulin transport: 20% of all axonemal and membrane proteins turn over within 6 hours in Chlamydomonas (Song and Dentler, 2001). At least some of these could be IFT cargoes, suggesting that IFT maintains additional ciliary properties apart from length. More recently, however, it was demonstrated that not all Chlamydomonas proteins depend on IFT for their ciliary localization (Harris et al., 2016). This opens the possibility that ciliary properties are maintained through both IFT-dependent and IFT-independent mechanisms.

The maintenance of cilia, as for centrioles, is also likely to require continuous transcription and translation of ciliary components. Evidence for this comes from a study on the role of miRNAs in photoreceptor maintenance in mice here, animals in which miRNA182 and miRNA183 were deleted formed fully functional photoreceptors in the first weeks after birth however, they exhibited specific defects in the maintenance of those cells (Busskamp et al., 2014). In the same study, a number of gene targets of miRNA182 and miRNA183 were annotated as associated with cilia and/or centrosomes, and could thus be relevant for maintenance of photoreceptors, and more generally, cilia (Fig. 2B). Additionally, the nuclear activity of the tumor suppressor gene VHL was linked to cilia maintenance in the kidney (Thoma et al., 2007). Here, loss of VHL and the subsequent reabsorption of cilia allows cells to re-enter the cell cycle and leads to cyst formation (Fig. 2B). However, it is unclear through which genes VHL controls cilia maintenance.

IFT is likely to have important roles in the maintenance of both the ciliary structure and the integrity of the cilium as a signaling compartment (Box 3). IFT occurs constantly, even in fully formed cilia, as shown by FRAP experiments and other live-cell imaging techniques (Hu et al., 2010 Milenkovic et al., 2015 Trivedi et al., 2012 Ye et al., 2013). In green algae, inhibition of IFT led to shortening of cilia that initially functioned normally (Marshall and Rosenbaum, 2001 Marshall et al., 2005). This raised the notion that ciliary proteins need to be replenished constantly in order to maintain flagella length (Marshall et al., 2005). Indeed, removal of IFT by gene deletion specifically in the retina of adult mice leads to photoreceptor degeneration (Jiang et al., 2015). Amongst the IFT cargoes, tubulin, the main component of the ciliary axoneme, was identified to be transported by the IFT complex proteins IFT81 and IFT74 in different species (Bhogaraju et al., 2013), and directly by kinesin-2 in Drosophila, which does not have these IFT subunits (Girotra et al., 2017 van Dam et al., 2013). However, not all systems may depend equally on IFT for cilia maintenance, as fully formed flagella of Trypanosoma show IFT-mediated motility, but, in this case, the IFT complex is only required to maintain flagellar function and not their structural integrity (Fort et al., 2016). In addition, Drosophila sperm does not need IFT either for biogenesis or for maintenance (Han et al., 2003 Sarpal et al., 2003) and IFT is absent from mature mouse sperm and hence is also not required for its maintenance (San Agustin et al., 2015). Further insights into how IFT regulates the maintenance cilia function can be gained from hedgehog signaling. Like many other signaling pathways, hedgehog signaling requires cilia (Bangs and Anderson, 2017). The G-protein-coupled receptor smoothened initiates hedgehog signaling when it localises to the ciliary membrane. In mice, IFT27 was shown to be required to keep unstimulated cilia in an off state by exporting smoothened out of the cilium (Eguether et al., 2014) however, IFT27 is frequently lost from the genomes of ciliated species (van Dam et al., 2013). Since IFT27 is required for hedgehog function (Eguether et al., 2014), it is unlikely that IFT maintains hedgehog signaling capacity in those systems. In fact, in Drosophila, hedgehog signaling is mostly independent of cilia (Han et al., 2003 Sarpal et al., 2003), with the exception of its role in olfaction (Kuzhandaivel et al., 2014 Sanchez et al., 2016). This suggests that the properties that are maintained by IFT depend on the species and perhaps also vary among the cilia within one organism.

The role of IFT in ciliary maintenance may also be regulated by the ciliary rootlet. In C. elegans rootletin mutants, which do not form rootlets but assemble cilia and IFT complexes (Mohan et al., 2013), IFT particles move at a lower speed and the cilia eventually degenerate. In Drosophila, no structural defects were found in rootletin mutants, but ciliary dysfunction was observed as well (Chen et al., 2015). These findings indicate that the ciliary rootlet is required for homeostatic maintenance of cilia function, but not for its structural integrity. In the future, it will be important to understand which IFT cargoes are important for maintaining cilia structure compared with its signaling capacities, as well as how transport is regulated by the rootlet.

Finally, more recently, cilia were shown to produce extracellular vesicles (often referred to as exosomes) in an actin-dependent manner (Nager et al., 2017). These exosomes are biologically active and are required for hatching in Chlamydomonas (Wood et al., 2013) or are used in communication between individual C. elegans (Wang et al., 2014). Exosomes can also return cilia to an inactive state after activation (Nager et al., 2017) and might be involved in maintaining other ciliary properties.


Coevolution of Nek kinases and centrioles

Several deflagellation-defective mutant strains of the unicellular biflagellate Chlamydomonas carry mutations in FA2 (for `flagellar autotomy'), a gene encoding a Nek (Mahjoub et al., 2002). Deflagellation (also known as deciliation) is a calcium-mediated stress response in which the cilia are severed at their base and shed into the environment (reviewed by Quarmby, 2004). Why is a member of a family known for its cell-cycle functions responsible for an important ciliary pathway? The fa2-null mutation produces a cell-cycle delay at G2-M (Mahjoub et al., 2002) rather than lethality, which suggests that a compensatory pathway exists, possibly involving another member of the Nek family. Indeed, whereas several microbial eukaryotes, including unicellular fungi, have only one or two members of the family (reviewed by O'Connell et al., 2003), Chlamydomonas express 10 Neks (Bradley et al., 2004) (B. Bradley and L.M.Q., unpublished). Since the fungi have neither centrioles nor cilia, we have hypothesized that the Nek family is expanded in ciliated organisms and coordinates cilia with the cell cycle (Bradley et al., 2004 Quarmby and Parker, 2005).

Examination of the genomes of several organisms reveals a correlation between the number of Neks in a particular organism and whether or not it has ciliated cells (Fig. 2). Drosophila and Caenorhabditis elegans have ciliated cells and thus might be expected to have more Neks than the higher plants, represented by Arabidopsis, which do not have ciliated cells. However, we posit that, because the only ciliated cells in Drosophila and C. elegans are terminally differentiated, these organisms do not have to coordinate cilia and the cell cycle (i.e. they do not have centrioles that serve as both basal bodies and microtubule-organizing centers). Indeed, employing a binary approach in which `1' is assigned to organisms that possess ciliated cells that re-enter the cell cycle and `0' is assigned to organisms that do not, we have tested this hypothesis by using the evolutionary software Continuous v1.0d13 PPC (Pagel, 1994). The result of comparing a model in which the number of Nek genes and the ability of ciliated cells to divide evolve independently with an alternative model in which the two traits evolve in a correlated fashion strongly supports our hypothesis (P=0.01). We propose that the large number of Neks in the higher plants is a consequence of an independent expansion involving substantial sub- and neo-functionalization. In other words, in higher plants, this family of kinases may have been coopted and expanded in the service of cellular activities unrelated to their ancestral functions. Consistent with this is the finding that higher-plant Neks fall into a clade distinct from the Neks of other organisms (Fig. 1C).


Author information

Giulia Pollarolo and Joachim G Schulz: These authors contributed equally to this work.

Affiliations

Department of Molecular and Developmental Genetics, Vlaams Instituut voor Biotechnologie, Campus Gasthuisberg, Leuven, Belgium

Giulia Pollarolo, Joachim G Schulz, Sebastian Munck & Carlos G Dotti

Katholieke Universiteit Leuven Center for Human Genetics, Campus Gasthuisberg, Leuven, Belgium

Giulia Pollarolo, Joachim G Schulz, Sebastian Munck & Carlos G Dotti

Centro de Biologìa Molecular Severo Ochoa, Universitad Autónoma de Madrid, Madrid, Spain

You can also search for this author in PubMed Google Scholar

You can also search for this author in PubMed Google Scholar

You can also search for this author in PubMed Google Scholar

You can also search for this author in PubMed Google Scholar

Contributions

G.P.: experimental design, data collection and assembly, data interpretation, manuscript writing. J.G.S.: experimental design, data assembly and interpretation, manuscript writing. S.M.: technical and imaging assistance, data analysis. C.G.D.: leading and coordinating the project, manuscript writing and editing. J.G.S. and C.G.D.: supervision of the project.

Corresponding authors


Materials and Methods

Fly stocks, antibodies and immunofluorescence

Fly stocks used in this study were described previously: wild-type Oregon R, GFP-Sas6 (Peel et al., 2007), GFP-Ana2 (Stevens et al., 2010a), and bld10c 04199 (Cep135Δ) (Mottier-Pavie and Megraw, 2009). All mutant tissues analyzed were both maternally and zygotically mutant (i.e. they were taken from homozygous mutants derived from homozygous mutant mothers) with the exception of mature adult testes, where the testes were taken from flies derived from a mixture of homozygous and heterozygous females. All transgenic lines generated here expressed GFP- or RFP-fusions of Cep135/Bld10 under the control of the Ubiquitin promoter, which drives expression at moderate levels in all tissues (Lee et al., 1988). For immunofluorescence analysis we used the following antibodies: Rabbit anti-Cep135 (this study – raised against amino acids 401–700 of the Drosophila Cep135/Bld10 coding sequence) Guinea-Pig anti-Asl (this study – raised against amino acids 1–333 of Asterless coding sequence) Rabbit anti-Cnn (Lucas and Raff, 2007). Secondary antibodies used were Alexa 488 anti-Rabbit and Alexa 568 anti-Guinea Pig. Testes of adult flies were dissected and fixed as previously described (Dix and Raff, 2007). Testes were imaged on an Olympus Flouview FV1000 with a 100×/1.40 Oil UPlanSApo objective. Images were processed and centrioles measured using ImageJ.

Sample preparation, electron tomography and immuno-electron microscopy

Sample preparation for testes was previously described (Stevens et al., 2010b). Wing-discs from 3rd instar larvae were dissected in PBS and prepared similarly to testes. Embryos were collected in fruit juice plates for 1 h and aged for 1 h. Early syncytial embryos were processed as previously described (Dzhindzhev et al., 2010). Immunolabeling and electron tomography were performed as previously described (Stevens et al., 2010b). Primary antibodies for immunolabeling were Rabbit anti-Asterless (Stevens et al., 2009) or Rabbit anti-Cep135 (this study). The secondary antibody used was 10 nm Gold Conjugated Goat Anti-Rabbit (Invitrogen). Tilt-series were acquired using SerialEM (Mastronarde, 2005) and tomograms reconstructed using the IMOD package (Kremer et al., 1996).

3D-structured illumination microscopy

Samples of testis and wingdiscs of 3rd instar larvae were prepared for 3D-SIM similarly to preparation for immunofluorescence with minor alterations. Testes and wingdiscs were squashed onto coverslips and fixed. The antibodies used were Rabbit anti-Asl (Stevens et al., 2009), Rabbit anti-DSpd2 (Dix and Raff, 2007), Rabbit anti-PLP (Martinez-Campos et al., 2004). Secondary antibodies used were Alexa 590 anti-Rabbit (Invitrogen) and GFP-Booster (Chromotek). Image stacks for super-resolution imaging were acquired in an OMX microscope (Applied Precision) with a 100×, 1.4 NA oil objective (Olympus) and processed using SoftWorx software (Applied Precision).

Data analysis

GFP protein diameters were measured by fitting a Guassian to the profile intensity and extracting the Full Width Half Maximum value. Asterless tube diameters were measured by fitting a Double Gaussian curve to the intensity profile and measuring the distance between both peaks. Fitting was done using Prism 5d software (GraphPad). Distributions were tested for Guassian distributions by the D’Agostino & Pearson omnibus test. Significance between distributions was tested by an unpaired t test for Gaussian distributions and the Mann–Whitney test for non-Gaussian distributions.


Watch the video: Neurons under microscope (June 2022).


Comments:

  1. Amenhotep

    In my opinion, mistakes are made. Let us try to discuss this. Write to me in PM.

  2. Bakari

    In my opinion, this is obvious. I would not like to develop this topic.

  3. Burdette

    and something analogous is?

  4. Scanlon

    New items are always cool !!!

  5. Ralston

    Between us speaking, in my opinion, it is obvious. I have found the answer to your question in google.com

  6. Dhruv

    Agree, this is the funny information

  7. Kikasa

    Sympathetic idea



Write a message