The Centrosome in Cells and Organisms

See allHide authors and affiliations

Science  27 Jan 2012:
Vol. 335, Issue 6067, pp. 422-426
DOI: 10.1126/science.1209037


The centrosome acts as the main microtubule-nucleating organelle in animal cells and plays a critical role in mitotic spindle orientation and in genome stability. Yet, despite its central role in cell biology, the centrosome is not present in all multicellular organisms or in all cells of a given organism. The main outcome of centrosome reproduction is the transmission of polarity to daughter cells and, in most animal species, the sperm-donated centrosome defines embryo polarity. Here I will discuss the role of the centrosome in cell polarity, resulting from its ability to position the nucleus at the cell center, and discuss how centrosome innovation might have been critical during metazoan evolution.

The centrosome of animal cells is a cytoplasmic organelle formed around a core structure made of two microtubule-based cylinders of defined length and diameter, the centrioles, with a highly conserved ninefold radial symmetry (1). The centriole pair displays structural and functional asymmetry due to the generational difference between each member of the pair: The old, fully mature, mother centriole is distinguished by two sets of nine appendages at its distal end while the young, immature, daughter centriole, assembled during the previous cell cycle, is about 80% the length of the mother centriole (Fig. 1). Both centrioles nucleate microtubules in their vicinity, but only the mother centriole can anchor microtubules at the subdistal appendages (2). In resting cells, the mother centriole can turn into a so-called basal body, by docking to the plasma membrane through the distal appendages (Fig. 1), where it templates a nonmotile primary cilium that serves as a sensory organelle (3). The basal body from the sperm flagellum triggers the assembly of the zygote centrosome during fertilization.

Fig. 1

The centrosome of human cells. In nonadherent lymphoblasts (left), the centrosome at the cell center deforms the nucleus. It contains a structurally and functionally asymmetric pair of centrioles, the mother centriole (red) and a daughter centriole (green). When isolated, the centrosome displays ultrastructural features shown in (A) and indicated in the cartoon at the top. The mother centriole (MC) is distinguished by two sets of nine appendages at its distal end, which are required for anchoring microtubules (MTs) and for docking the mother centriole at the plasma membrane during ciliogenesis in quiescent cells (inset). The proximo-distal differentiation of the MC is demonstrated on the left and at the bottom by serial sections according to the lettering shown in (A) [adapted from (40)]. The proximal part of centrioles displays nine microtubule triplets, which turn clockwise when viewed from the distal end, whereas the distal part displays nine microtubule doublets oriented in a tangential manner, like ciliary doublets. DC, daughter centriole. Bar: 0.2 μm. (B) The centrosome in situ observed by immunofluorecence light microscopy shows microtubules (white) and ninein (red), which marks the subdistal appendages of the MC in a cell line stably expressing centrin–green fluorescent protein (GFP). Bar: 3 μm. (B′) Movements of MC (red) and DC (green) over a period of 20 min. Bar: 1 μm. [Adapted from (2)] (C) Ultrastructural view of a mother centriole in situ showing numerous microtubules abutting at the tip of subdistal appendages where ninein (red) localizes. Centrin (green) localizes into the distal lumen of both centrioles. Bar: 0.2 μm.

The centriole/basal body associated with a (9+2) flagellum is present in the common unicellular ancestor of eukaryotic cells (3, 4). It participates in three major cell functions: locomotion, sensory reception, and division. The innovation of the centrosome as a cytoplasmic organelle, no longer permanently associated with an axoneme, has taken place essentially in the unikont lineage. Centrosome evolution among divergent eukaryotes must reflect cellular variations in locomotion, sensory reception, or division underlying the adaptive evolution of organisms.

The Centrosome, Yesterday and Today

The centrosome, described as a “polar corpuscle,” was suspected from the start to have a role in cell symmetry breaking and the maintenance of cell polarity. Early experiments exploiting centrosome activity paved the way to the chromosomal theory of heredity. The fertilization of sea urchin eggs with two sperm cells results in the formation of a quadripolar mitotic spindle and the division of the egg into four blastomeres with an uneven segregation of chromosomes. When dissociated from each other at the four-cell stage, normal blastomeres form a dwarf embryo. Analyzing the development of blastomeres obtained after dispermy, Boveri could conclude that both the individuality of each chromosome and their continuity in the interphasic nucleus were important. The centrosome was seen as the organ coordinating karyokinesis and cytokinesis. The absence of centrosome in the unfertilized egg was seen as an efficient way to avoid parthenogenetic development. This contention was experimentally supported much later (5).

In the last decade, the molecular composition of the isolated centrosome in the human and fly has revealed a number of proteins conserved across eukaryotes (3, 4). Functional genomics in nematodes and flies has identified a small set of conserved proteins required for the initiation of centriole/basal body assembly and for centrosome reproduction. Recently, the structural basis of the highly conserved ninefold radial symmetry of the centriole/basal body has been elucidated (6, 7). Unexpectedly, it rests on the oligomerization of a single coiled-coil protein, SAS-6/Bld12p, which forms a cartwheel structure acting as a scaffold for centriole/basal body assembly.

The structural complexity of the centrosome is reflected in its biochemical composition: More than 100 proteins are localized either in the centrioles or in the centrosomal matrix (8, 9), among which several are disease gene products. For example, all genes involved in microcephaly syndromes so far encode centrosomal proteins (10).

The Centriole/Basal Body Defines Embryo Polarity at Fertilization

Besides restoring diploidy, egg fertilization in animals involves the resetting of egg polarity through the conversion of the sperm head–associated basal body into the male pronucleus-associated centrosome.

Egg symmetry breaking. The polarization of the zygote by the sperm-donated centrosome has been studied in great detail in the nematode Caenorhabditis elegans. The sperm centrosome triggers the cortical flow that establishes the anterior and posterior cortical polarity domains through two redundant mechanisms (11). One acts on the contractility of the cortex itself, via local inactivation of the small GTP-binding protein Rho. The other acts by localizing the polarity-associated protein PAR-2 to the cortex and antagonizing PAR-3–dependent recruitment of myosin. These findings exemplify the centrosome’s ability to trigger global spatial control by concentrating effectors locally.

Centrosome positioning at the egg’s center. In large eggs, the centrosome centers itself independently of aster-cortex contact (12), and the spindle assembles between a pair of centrosomes prepositioned by the preceding interphase aster centers (13). The microtubule-dependent orientation of nuclear stretching and division axis in sea urchin eggs forced into microfabricated chambers of defined geometry has been quantitatively modelized (14). Pulling forces appear to be nearly proportional to the cubic length of microtubules. This result suggests a “volume sensing” model in which the probability for a motor encountering a microtubule of a given length is proportional to a cone-shaped volume around the microtubule. The demonstration in C. elegans egg of a dynein-dependent cytoplasmic pulling anchored on moving organelles could provide a physical basis to this model (15). Alternatively, cytoplasmic actin could exert pulling forces on astral microtubules, as shown in cultured mitotic cells (16).

Blastomere individuation. Unfertilized Xenopus egg cleavage can be triggered by the injection of a somatic centrosome, whether from the same or from divergent vertebrate species, whereas activation of the egg by pricking triggers cell cycle resumption but does not lead to cleavage despite an equivalent complement of microtubules in the egg (Fig. 2) (17). Centrosome-dependent astral organization of the microtubule array is thus necessary for egg cleavage. It is also sufficient in eggs where growth requirements have been met during oogenesis: Enucleation experiments on starfish eggs after fertilization lead to individuation of the control number of enucleated blastomeres due to the sequential duplication of the sperm centrosome (Fig. 2) (18). This result supports the contention that the coevolution of the centrosome able to nucleate a microtubule aster, and of a cortical acto-myosin network, has been critical for the selection of cell division by fission (19).

Fig. 2

Centrosome-dependent blastomere individuation. (A) Centrosome-induced parthenogenesis in the frog Xenopus laevis. Oocytes, arrested in metaphase II of meiosis, are activated by pricking: 13 embryonic cell cycles ensue in this species, monitored by maturation promoting factor (MPF) activity, but no cleavage occurs in the absence of a sperm cell. Parthenogenetic cleavage and development can be triggered by the injection of a heterologous centrosome, isolated from mammalian cells [see (17)]. (B) Complete cleavage of enucleated early embryos in the star fish Astropecten aranciacus. Pronuclei have been removed after fertilization, leaving the sperm-donated centrosome. Cartoon depicting control (left) and enucleated embryos (right) after the nine embryonic cell cycles in this species: The control number of enucleated blastomeres is observed in the enucleated embryo. At the bottom, cartoon depicting single blastomeres in M phase: the sequential reproduction of the sperm-donated centrosome is sufficient to trigger complete cleavages. [Adapted from (18)]

Centrosome generational asymmetry and embryo development. Centrosome reproduction overlaps two cell-division cycles and gives rise to a generational asymmetry so that a centrosome always consists of a mother and a daughter centriole (Fig. 1). These features are exploited to segregate mRNAs or transcription factors localizing to the centrosomal matrix in a microtubule- and cell cycle–dependent manner. For example, in spiralian mollusc embryos, mRNAs of patterning genes are centrosomally located, which leads to the production of cells that differ in size, cleavage pattern, and fate (20). In the C. elegans embryo, PIE-1 protein localizes to the germline blastomeres throughout early development and maintains their totipotency by repressing the somatic state. PIE-1 initially associates with both centrosomes of the mitotic spindle (21), then rapidly disappears from the centrosome destined for the somatic daughter after spindle rotation, and persists in the centrosome of the daughter that becomes the next germline blastomere. A similar centrosome-based asymmetric degradation has been documented in dividing human embryonic stem cells in which the duration of signaling by the Smad1 transcription factor is terminated by proteasomal degradation (22). Centrosome-based asymmetric division is widely used in maintaining polarized tissue morphogenesis (23) and has been particularly well documented in the developing neocortex (24).

The Centriole/Basal Body in Somatic Cell Polarity

The three functional facets of cell polarity associated to the ancestral basal body/axoneme—namely, sensation, motion, and division—were preserved during metazoan multicellularity through the innovation of the centrosome organelle (19). The centrosome controls the maintenance of cell polarity during migration, tissue growth, and homeostasis. In many tissues, the primary cilium is a sensory organelle channeling signal transduction in several major pathways (8). A great range of stimuli seem to be collected in this way, and the primary cilium is thus critical for embryo development. Dysfunction of the primary cilium is responsible for many human pathologies previously thought to be unrelated (8). The molecular pathways that regulate the conversion of centrioles to basal bodies and vice versa, as well as the transcriptional controls involved, are being actively investigated (9).

Centrosome positioning in migrating cells and in tissues. Mitotic centrosomes are compact, with the typical replicative orthogonal configuration between parental and daughter centrioles at each pole. By contrast, centrosomes in somatic cells have a considerable intrinsic flexibility during interphase (2): The two centrioles are disengaged and held together by a flexible linker between their proximal ends, and the intercentriolar distance continually changes (Fig. 1). Only the mother centriole can center the centrosome (2). How the intercentriolar linker length is regulated is not known. More is known about the separation of the two centrosomes at the onset of mitosis: The linker contains not only Nek2 kinase substrates such as C-Nap1 and Rootletin but also β-catenin, an effector of the Wnt pathway, which participates in the reorganization of the linker upon Nek2 phosphorylation (25). Nek2 association with the centrosome at this step is also regulated by components of the Hippo signaling pathway involved in growth control (26).

How do somatic cells sense their shape and size? There is evidence in many cell systems that interaction of dyneins with the lateral surface of microtubules is an important component of the centrosome positioning (27). Through its physical association with the nucleus, the centrosome controls nuclear positioning, thus creating an internal asymmetry in the distribution of organelles: the nuclear volume is left out from the cell space in which microtubule-dependent centrosome positioning is taking place (Fig. 3). In nonadherent lymphocytes, or in enucleated adherent cells, the centrosome sits precisely at the cell’s center. The mutual positioning of the centrosome and nucleus of adherent cells during migration indicates where the center of cytoplasmic mass is repositioning to. The control of centrosome positioning is tightly regulated in individual cultured cells when adhesion is controlled (Fig. 3) (28), which should also be the case for cells in tissues. Adhesion proteins can shuttle to the centrosome and regulate its activity in a cell cycle–dependent manner and thus could participate in centrosome positioning. Centrosome repositioning is also involved in the repositioning of the pericentrosomal Golgi ribbon. Breaking the polarity axis by disconnecting the Golgi ribbon from the centrosome has a more pronounced effect on directional cell migration than disrupting the Golgi ribbon (29), supporting the idea that centrosome repositioning could also depend on lateral interaction of microtubules with the three-dimensional (3D) network of intracellular membranes.

Fig. 3

The nucleus–centrosome axis as a marker of cell symmetry breaking. (A) A drawing from Van Beneden in 1883. (B) The centrosome positioning in cultured cells is tightly controlled as can be demonstrated by controlling cell adhesion. The figure shows the superimposition of 75 cells adhering to a cross-bow–shaped pattern: All centrosomes appearing as yellow-white dots are found in a 2-μm-diameter area. The Golgi apparatus appears in red. The nucleus positioning is more variable as judged by the blurring of the blue staining. The cortical labeling corresponds to regions of actin assembly on adhesive edges (cortactin in red decorates lamellipodia) and of acto-myosin contraction on nonadhesive edges (phalloidin in green decorates stress fibers). [Adapted from (28)] (C) In an enucleated cell, here adhering to a triangular micropattern, the centrosome sits precisely at the cell center and the microtubule network is symmetrically distributed. [From (41)] Bars: 10 μm

The switch to primary cilium–dependent cell polarity. Primary cilium formation is correlated with cell cycle exit, but has long been thought to depend also on cell-cell contacts because ciliogenesis in culture is dependent on cell density. The use of adhesive micropatterns of increasing size to control individual cell spreading has demonstrated instead that cell spreading is a major parameter of ciliogenesis (Fig. 4) (30). Only cells confined on small patterns express a robust polarity, and a full-length primary cilium. The switch between nonciliated and ciliated states in animal cells appears to be a global cell response integrating different types of information. This is not an all-or-none transition, but rather a progressive transition in which the cilium length itself can be smoothly tuned (30).

Fig. 4

Cell spreading controls ciliogenesis in serum-starved human pigment epithelial retinal cells. The transition from a cytoplasmic centrosome located at the cell center to a plasma membrane–associated centrosome in which the mother centriole acts as a basal body is modulated by the balance between assembly and contraction of the actin network. From left to right: fully adhesive discs with an area of 3500, 1500, and 500 μm2. Cells were stained for F-actin (red) and for actetylated tubulin (green) to show the primary cilium. Bar: 10 μm. [Adapted from (30)]

Actin contractility could modulate ciliogenesis in many ways. About 50 gene products can facilitate or inhibit primary cilium growth (31). These are proteins regulating vesicle trafficking and actin dynamics, including components of focal adhesion that participate in cell signaling. A compact pericentrosomal structure revealed by recycling endocytic markers apparently stores transmembrane proteins required for the early steps of ciliogenesis. This pericentrosomal compartment appears to be an important regulator of pre-ciliogenesis, but its relationship to the pericentrosomal Golgi apparatus is unclear. Disrupting Golgi ribbon integrity at the centrosome or disconnecting the Golgi ribbon from the centrosome inhibits ciliogenesis (29).

Centriole/basal body and tissue differentiation. We are far from understanding how the centriole/basal body switch is regulated in the different cell lineages of animal organisms. Some lineages never make primary cilia—like the immune system, which instead uses centrosome activity to build transient immune synapses (32). Kidney epithelia grow primary cilia that have a critical role in the physiology of the organ, whereas gut epithelia do not grow primary cilia but instead develop conspicuous actin-based microvilli. Myoblast-derived multinucleated myotubes cannot make primary cilia because centrosomes are eliminated during the differentiation process, while monocyte-derived multinucleated osteoclasts keep all centrosomes associated with their respective nuclei but do not make cilia (33). How cell polarity is controlled, or lost, in each case is unclear.

The Evolution of the Centrosome

Conserved and yet different centrosomes among divergent species. The centrosome has experienced many secondary variations that correlate with the fate of the axoneme. When the axoneme is no longer fully functional, the basal body structure becomes less constrained. When axonemes are absent, basal bodies/centrioles are also absent and different centrosomal structures can be observed, such as the nucleus-associated body in cellular slime molds, or the spindle pole body in higher fungi. In the metazoan ecdysozoa (arthropods and nematodes), the basal body/centriole structure can depart largely from the canonical structure, showing short centrioles with doublets or singlets and a correlative loss of a large number of centrosomal proteins (3, 4). Accordingly, cilia are modified and involved only in sensory reception. The only beating axoneme in Drosophila is the sperm flagellum. In C. elegans, the sperm has centrioles but no flagellum. Centrosomes are robust microtubule-nucleating organelles during embryonic development and reproduce via the conserved orthogonal centriole duplication. However, somatic centrosomes are either absent or poor microtubule-nucleating centers during interphase. A fly acentriolar cell line, although prone to chromosome loss, was derived years ago (34), and more recently it has been demonstrated that ablating genes necessary for centriole or centrosome assembly is compatible with the adult fly development (35). Thus, constraints on the centrosome may be reduced in the larval and adult organisms of this taxon, as if the evolution of an exoskeleton would have reduced the need of a robust cell polarity.

The flagellar apparatus–nucleus connection of the unicellular ancestor has been proposed to be a critical feature in evolving the centrosome in opisthokonts (19). The requirement of centrin, a very ancient protein closely related to calmodulin, for centrosome reproduction in many species seems to be due to its participation in mechanisms that ensure nucleus–centrosome connection through cell division. As observed for the centrosome, evolutionary constraints on centrin proteins have also been relaxed in Ecdysozoa (19).

Can we learn from the lower animal organisms? Organisms that exhibit variations from the usual pattern of centrosome behavior or reproduction could help to demonstrate cellular and developmental implications of centrosomes. Animals among the early branching taxa—cnidaria, ctenophora, and bilateria—in which one can hope to trace the origin of features defining animal body plans are good candidates. For example, planarians have lost the orthogonal duplication mode of centriole/basal body assembly and the centrosome organelle as well (36).

Planarians move and feed using multiciliated ventral and pharyngeal cells. As expected from the presence of canonical basal body/(9+2) axonemes, the genome of the freshwater planarian Schmidtea mediterranea contains most of the genes associated with centriole/basal body in other species. The vast majority of these genes are apparently required in planarians for the assembly and function of the basal bodies in multiciliated cells (36). Depletion of planarian homologs of proteins essential for centriole duplication such as Sas-4 and Plk4 abolished basal body assembly in multiciliated cells, but the depleted animals showed no observable regeneration defect. Accordingly, wild-type animals do not possess centrioles in cells other than multiciliated cells, including in dividing neoblasts (36). The centrosome was apparently lost in the lineage leading to planarians. A small set of centrosome signature proteins absent in planarians are present in lower organisms, including in the early flatworm Macrostomum lignano, which has centrioles at the mitotic poles of neoblasts and some, although limited, regenerative ability (36).

These results raise a burning question: Does the complete absence of a centrosome in animals really matter? Multicellularity is believed to have started by the transition from a colonial stage of unicellular flagellated cells to a true multicellular assembly of cells nonpermanently ciliated due to the constraints of synthesizing and maintaining a flagellum (37). Perhaps during this transition, individual cell polarity had to be preserved to control the stability of the novel multicellular organism, as in tissues of modern organisms. Selective pressure to maintain individual cell polarity could have triggered the transition from one type of “polar corpuscle,” the plasma membrane–associated basal body/flagellum, to another, the cytoplasmic centrosome, which could switch back to the former type of cell polarity by growing a primary cilium.

What could be the developmental consequences of the absence of centrosome-based polarity in individual cells of an animal organism? Planarians depart from most animal multicellular organisms, including lower flatworms expressing centrosomes, like M. lignano, which undergo exclusive sexual reproduction: S. mediterranea can exist as sexual and asexual strains. The latter reproduce by transverse fission, do not differentiate germline or the somatic copulatory apparatus, and can generate abundant clonal progeny in controlled conditions (38). Possibly, reproduction by transverse fission of the body is incompatible with robust centrosome-dependent polarity of individual cells. Similarly, the long-range antero-posterior polarity system observed during asexual regeneration of planarians, or after experimental transection, to initiate head or tail formation could reflect the lack of polarity in individual cells.

If this view is correct, the absence of centrosomes in planarians would not conflict with, but rather support, the conclusion drawn from other organisms, in which the centrosome is an organ required for the stable individuation of polarized cells and for the transmission of polarities through cell division. Indeed, planarian embryonic cleavage has not retained the stereotypical pattern of cell division orientation of the ancestral spiral cleavage (36). Another possible consequence of evolving a centrosome in metazoa may be to preclude asexual reproduction. If further supported, this conclusion could have an important evolutionary implication as it would suggest that the centrosome’s innovation has contributed to the very early differentiation and sequestration of the germ line observed in animal development (39).

Concluding Remarks

It is likely that the transition to multicellularity in the animal lineage has preserved cell polarity and cell individuation by internalizing the cortical microtubule network of the unicellular choanoflagellate-like ancestor and by evolving the centrosome. Actin-dependent amoeboid cells able to interact with each other, in which centrosome positioning through microtubules allows cells to sense their shape and size, are the real innovation allowing the unfolding of animal morphogenesis.

The evolutionary success of the centriole/basal body structure is striking. Centriole duplication according to orthogonal budding is observed in all ancient unicellular organisms. Did this puzzling mechanism evolve to control cell polarity, as a structural mechanism integrating 3D spatial information? The proximo-distal polarity and circumferential anisotropy of the chiral centriole/basal body structure could also specify connectivity with other cell compartments.

The selective pressure for the ninefold symmetry could come from the axoneme beating constraints. When the axoneme is no longer functional, the basal body structure becomes less constrained. Are there other functions for basal body, besides growing the axoneme and controlling the entry of flagellar complexes? Does it also participate in the beating or in the signaling control? Is there an intracentriole/basal body dynamics required for such a control or for centriole functions in the centrosome? These are questions for future work.

References and Notes

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 40.
  42. Acknowledgments: I thank J. Sillibourne, R. M. Rios, and two reviewers for critical reading of the manuscript, and M. Théry for providing some of the figures. This work is funded by both CNRS and Institut Curie. M.B. is acting as the chief scientific officer for “CYTOO Cell Architects,” Grenoble, France.
View Abstract

Navigate This Article