The Interactive Fly
Asymmetric inheritance of mother versus daughter centrosome in stem cell division
Adult stem cells often divide asymmetrically to produce one self-renewed stem cell and one differentiating cell, thus maintaining both populations. The asymmetric outcome of stem cell divisions can be specified by an oriented spindle and local self-renewal signals from the stem cell niche. Developmentally programmed asymmetric behavior and inheritance of mother and daughter centrosomes underlies the stereotyped spindle orientation and asymmetric outcome of stem cell divisions in the Drosophila male germ line. The mother centrosome remains anchored near the niche while the daughter centrosome migrates to the opposite side of the cell before spindle formation (Yamashita, 2007).
Adult stem cells maintain populations of highly differentiated but short-lived cells throughout the life of the organism. To maintain the critical balance between stem cell and differentiating cell populations, stem cells have a potential to divide asymmetrically, producing one stem and one differentiating cell. The asymmetric outcome of stem cell divisions can be specified by regulated spindle orientation, such that the two daughter cells are placed in different microenvironments that either specify stem cell identity (stem cell niche) or allow differentiation (Yamashita, 2007).
Drosophila male germline stem cells (GSCs) are maintained through attachment to somatic hub cells, which constitute the stem cell niche. Hub cells secrete the signaling ligand Unpaired, which activates the Janus kinase-signal transducer and activator of transcription (JAK-STAT) pathway in the neighboring germ cells to specify stem cell identity. Drosophila male GSCs normally divide asymmetrically, producing one stem cell, which remains attached to the hub, and one gonialblast, which initiates differentiation. This stereotyped asymmetric outcome is controlled by the orientation of the mitotic spindle in GSCs: The spindle lies perpendicular to the hub so that one daughter cell inherits the attachment to the hub, whereas the other is displaced away (Yamashita, 2007).
The stereotyped orientation of the mitotic spindle is set up by the positioning of centrosomes during interphase. GSCs remain oriented toward the niche throughout the cell cycle. In G1 phase, the single centrosome is located near the interface with the hub. When the duplicated centrosomes separate in G2 phase, one stays next to the hub, whereas the other migrates to the opposite side of the cell. Centrosomes in the GSCs separate unusually early in interphase, rather than at the G2-prophase transition, so it is common to see GSCs with fully separated centrosomes without a spindle (Yamashita, 2007).
Differences between the mother and daughter centrosomes underlie the stereotyped behavior of the centrosomes in Drosophila male GSCs. The mother centrosome normally remains anchored to the hub-GSC interface and is inherited by the GSC, whereas the daughter centrosome moves away from the hub and is inherited by the cell that commits to differentiation. Mother and daughter centrosomes were differentially labeled by transient expression of green fluorescent protein-pericentrin/AKAP450 C-terminus (GFP-PACT) from the Drosophila pericentrin-like protein under heat shock-Gal4 control. The PACT domain, which is necessary and sufficient for centriolar localization, is incorporated into centrioles only during centrosome duplication and does not exchange with the cytoplasmic pool. Both the mother and daughter centrosomes are labeled by GFP-PACT in the first cell cycle after heat shock. In the second cell cycle, the daughter centrosome retains GFP-PACT, whereas the mother centrosome is not labeled, thus distinguishing the mother and daughter centrosomes. After a short burst of GFP-PACT expression induced by a 2.5-hour heat shock, 20% to 30% of the GSCs had GFP-labeled centrosomes, indicating the duplication of centrosomes during the window of GFP-PACT expression. By 12 hours after heat shock, >90% of the labeled GSCs had two GFP-positive centrosomes, indicating that they had progressed to the G2 phase of the first cell cycle after GFP-PACT incorporation (Yamashita, 2007).
By 18 to 24 hours after heat shock, the number of GSCs with two GFP-positive centrosomes had decreased, whereas the number of GSCs with one GFP-positive and one GFP-negative centrosome had increased, suggesting progression into the second cell cycle. Generally, the centrosome distal to the hub was labeled, whereas the centrosome proximal to the hub was GFP-negative, indicating that the daughter centrosomes migrate away from the hub-GSC interface during asymmetric GSC divisions (Yamashita, 2007).
Labeling the mother rather than the daughter centrosomes confirmed that the male GSCs in the niche preferentially retain mother centrosomes over time. Centrioles assembled during early embryogenesis were labeled using the NGT40 Gal4 driver to drive the expression of GFP-PACT in blastoderm-stage embryos, shutting off after germband extension. In the first cell cycle after the depletion of the cytoplasmic pool of GFP-PACT in the GSCs, both the mother and daughter centrosomes should be labeled. In subsequent cell cycles, only the mother centrosomes should be labeled (Yamashita, 2007).
In most GSCs in the second or later cell cycle after the depletion of cytoplasmic GFP-PACT, the labeled centrosome was positioned next to the hub-GSC interface, and the unlabeled centrosome had moved away from the hub. The frequency of GSCs that had the proximal, but not distal, centrosome labeled remained constant over time for 10 days (L3 larvae to day-3 adults), suggesting that the mother centrosomes are reliably retained by the GSCs, even through multiple rounds of GSC divisions. Some GSCs maintained cytoplasmic GFP-PACT, especially in L3 larvae, suggesting that the GFP-PACT had not yet been diluted out. Some GSCs were also observed with two labeled centrosomes, suggesting that they are in the first cell cycle after the depletion of cytoplasmic GFP-PACT (Yamashita, 2007).
The mother centrosomes in GSCs appeared to maintain robust interphase microtubule arrays. Ultrastructural analysis of the GSCs revealed that the centrosome proximal to the hub was commonly associated with many microtubules throughout the cell cycle. Nineteen centrosomes in GSCs were scored in serial sections of the apical tips of five wild-type testes. Eleven centrosomes were localized close to the adherens junctions between the hub and the GSCs. Nine of these proximal centrosomes appeared to be in interphase cells, based on nuclear morphology and microtubule arrangement. Typically, these interphase centrosomes proximal to the hub were associated with numerous microtubules. In some samples, microtubules appeared to extend from the centrosome toward the adherens junctions. The other two proximal centrosomes appeared to be in cells in mitotic prophase, based on their robust microtubule arrays containing bundled microtubules running parallel to or piercing the nuclear surface (Yamashita, 2007).
In contrast, of the five distal centrosomes in the apparently interphase cells that were scored, four had few associated microtubules. The remaining three distal centrosomes appeared to be in cells in mitotic prophase, based on microtubule arrays containing bundled microtubules. Thus, the mother centrosomes may maintain interphase microtubule arrays that anchor them to the hub-GSC interface, whereas the daughter centrosomes may initially have few associated microtubules and be free to move, establishing a robust microtubule array only later in the cell cycle (Yamashita, 2007).
Consistent with the idea that astral microtubules anchor the mother centrosomes to the hub-GSC interface, mother versus daughter-centrosome positioning was randomized in GSCs that were homozygous mutant for centrosomin (cnn), an integral centrosomal protein required to anchor astral microtubules to centrosomes. Analysis of mother and daughter centrosomes after transient expression of GFP-PACT revealed that, for cnn homozygous mutant GSCs where one of the two centrosomes was positioned next to the hub, it was essentially random whether the mother or the daughter centrosome stayed next to the hub. In addition, in >25% of total labeled GSCs, neither of the two centrosomes was next to the hub (Yamashita, 2007).
These results indicate that the two centrosomes in Drosophila male GSCs have different characters and fates. The mother centrosome stays next to the junction with the niche and is inherited by the cell that self-renews stem cell fate. Thus, GSCs can maintain an old centriole assembled many cell generations earlier. In contrast, the daughter centrosome migrates away from the niche and is inherited by the cell that will initiate differentiation. It is postulated that the mother centrosomes in male GSCs may remain anchored to the GSC-niche interface throughout the cell cycle by attachment to astral microtubules connected to the adherens junction, whereas the daughter centrosomes may initially have few associated microtubules and thus can move away from the niche. Microtubule-dependent differential segregation of mother and daughter spindle-pole bodies (equivalent to centrosomes in higher organisms) is observed in budding yeast (Pereira, 2001). In cultured vertebrate cells, the centrioles mature slowly over the cell cycle, and the mother centrosomes (containing a mature centriole) attach astral microtubules more effectively and are more stationary than daughter centrosomes in interphase (Piel, 2000). The unusually early separation of centrosomes in interphase male GSCs may provide a way to move the daughter centrosome out of range of the stabilizing influence of the adherens junction complex before it becomes competent to hold a robust microtubule array (Yamashita, 2007).
Developmentally programmed anchoring of the mother centrosome may provide a key mechanism to ensure the stereotyped orientation of the mitotic spindle and thus the reliably asymmetric outcome of the male GSC divisions. Although it is tempting to speculate that determinants associated with the mother or daughter centrosome may play a role in specifying stem cell or differentiating-cell fates, such determinants are yet to be identified. Rather, the asymmetric inheritance of mother and daughter centrosomes in male GSCs may be a consequence of the cytoskeletal mechanisms that are imposed as part of the stem cell program to anchor one centrosome next to the niche throughout the interphase, ensuring a properly oriented spindle (Yamashita, 2007).
Functionally unequal centrosomes drive spindle orientation in asymmetrically dividing Drosophila neural stem cells
Stem cell asymmetric division requires tight control of spindle orientation. To study this key process, Drosophila larval neural stem cells (NBs) engineered to express fluorescent reporters for microtubules, pericentriolar material (PCM), and centrioles, were examined. Early in the cell cycle, the two centrosomes become unequal: one organizes an aster that stays near the apical cortex for most of the cell cycle, while the other loses PCM and microtubule-organizing activity, and moves extensively throughout the cell until shortly before mitosis when, located near the basal cortex, it recruits PCM and organizes the second mitotic aster. Upon division, the apical centrosome remains in the stem cell, while the other goes into the differentiating daughter. Apical aster maintenance requires the function of Pins. These results reveal that spindle orientation in Drosophila larval NBs is determined very early in the cell cycle, and is mediated by asymmetric centrosome function (Rebollo, 2007).
Immediately after cytokinesis, the single dot revealed by both PCM and centriole reporters splits in two, strongly suggesting that centrosome duplication has taken place. The YFP-Asl marker, like other centriolar markers in Drosophila, does not allow for resolution of individual centrioles within a centrosome in larval NBs. Therefore, timing of centriole duplication in these cells remains uncertain. The two resulting centrosomes migrate together within the single major aster of the cell to the apical cortex. Later on, one of the centrosomes loses PCM and starts to migrate. At this early time point in the cell cycle, unequal centrosome fate is already established: one, apical, will remain in the stem cell; the other will go into the differentiating daughter. Migration of the downregulated centrosome (revealed by the centriolar marker), initially within the apical side of the cell and more basally later on, occupies most of the cell cycle and is the most variable stage, its duration being dependent on cell-cycle length. The apical centrosome organizes the only aster found in the NB for most of the cell cycle. As mitosis onset approaches, the moving downregulated centrosome becomes stabilized at the basal side and starts to accumulate PCM and organize the second aster. As a direct consequence, the spindle is assembled already in alignment with the polarity axis of the cell. In Drosophila male germline stem cells (Yamashita, 2007), one of the centrosomes is also consistently located adjacent to the hub from early interphase onward. Only this centrosome maintains a robust aster through the cell cycle. The other, associated with only a few microtubules, moves away from the hub and is inherited by the gonialblast. In these cells, the oldest centriole is always in the centrosome that is proximal to the hub and is therefore retained by the stem cell (Yamashita, 2007). It has not yet been possible to determine which of the two centrosomes contains the oldest centriole in larval neural stem cells. In pins NBs, unequal centrosome fate and function are established, but, eventually, the stable, aster-forming apical centrosome is downregulated and starts to behave like the other, migrating across the cell. Like the other too, it organizes an aster only shortly before NEB. The place of assembly of the two asters in pins mutant NBs is not fixed and consequently spindle orientation is randomized, and so is the size difference between the two daughter cells (Rebollo, 2007).
It is still unclear how NB polarity is maintained from one cycle to the next; a distinct Baz apical crescent is only assembled at prophase. The permanent positioning of the NB centrosome in the apical side of the cell, through the cell cycle, suggests that it could be contributing to specifying the apical cortex after mitosis (Rebollo, 2007).
In summary, four main conclusions can be derived from these observations: (1) the two centrosomes of asymmetrically dividing Drosophila larval NBs become unequal early in the cell cycle in terms of mobility, MTOC activity, and fate; (2) such elaborated unequal centrosome regulation provides a means to position the asters, thus ensuring spindle alignment along the polarity axis of the cell; (3) Pins contributes to spindle orientation in NBs by preventing the downregulation of the MTOC capability of the apical centrosome, thus maintaining the apical aster in place, and (4) spindle orientation is predetermined and can be accurately predicted as soon as the aster reaches the apical cortex during the initial stages of the cell cycle. Altogether, these observations reveal that asymmetry in Drosophila neural stem cells goes beyond the polarized localization of a number of protein complexes during mitosis and may affect entire organelles such as the centrosome, which exerts a major effect on cell architecture and function throughout the cell cycle (Rebollo, 2007).
Tissue stem cells play a key role in tissue maintenance. Drosophila melanogaster central brain neuroblasts are excellent models for stem cell asymmetric division. It has been shown that their mitotic spindle orientation is established before spindle formation. This study investigated the mechanism by which this occurs, revealing a novel centrosome cycle. In interphase, the two centrioles separate, but only one is active, retaining pericentriolar material and forming a 'dominant centrosome.' This centrosome acts as a microtubule organizing center (MTOC) and remains stationary, forming one pole of the future spindle. The second centriole is inactive and moves to the opposite side of the cell before being activated as a centrosome/MTOC. This is accompanied by asymmetric localization of Polo kinase, a key centrosome regulator. Disruption of centrosomes disrupts the high fidelity of asymmetric division. A two-step mechanism is proposed to ensure faithful spindle positioning: the novel centrosome cycle produces a single interphase MTOC, coarsely aligning the spindle, and spindle-cortex interactions refine this alignment (Rusan, 2007).
By prophase, NBs contain two MTOCs that are almost fully separated and aligned along the NB/GMC axis, but analysis of fixed NBs revealed a single MTOC positioned opposite the GMCs before mitotic entry. Thus MTOC behavior was examined throughout the cell cycle as an initial approach to test the hypothesis that fixing the position of one MTOC through successive divisions helps ensure persistent spindle orientation. Live NBs expressing GFP-G147, an MT-associated protein were examined, revealing a striking temporal difference in MTOC activity. During interphase, a single detectable MTOC persists opposite the previous division site; this is referred to as the dominant MTOC. As NBs approach mitosis, this MTOC increases activity (matures; empirically judged by size and MT fluorescence intensity), forming an MT basket around the nucleus. This is referred to as preprophase; this stage is also seen in fixed samples stained for tubulin. Soon after, sometime before the dominant MTOC fully matures, something striking happens: a second MTOC appears distant from the first. This is the second MTOC and this stage is referred to as prophase onset. The second MTOC increases activity, maturing ~10 min before nuclear envelope breakdown (NEB). Using 4D imaging, the possibility was excluded that the second MTOC was present earlier in another focal plane. To further assess this, forming spindles were imaged end on. It is clear that the second MTOC did not emerge from the dominant MTOC or travel around the nucleus. Instead, the second MTOC appeared roughly opposite the dominant MTOC. MTOC separation began immediately, and by NEB, they were 146 ± 20° apart. Thus, NBs form two distinct MTOCs: an MTOC persisting from the previous division and another only activated at mitotic entry (Rusan, 2007).
This distant activation of the second MTOC raised questions about the centrosome cycle. One possibility is that NBs have two MT nucleating centrosomes, but only one can retain MTs and act as an MTOC during interphase, whereas the second acquires MT retention ability during mitotic entry, explaining the second MTOC's sudden appearance. There is precedent for this: mouse L929 cells have two γtub-bearing centrosomes that can nucleate MTs, but only one contains Ninein and can retain MTs to form an MTOC (Rusan, 2007).
To test this hypothesis in NBs, EB1-GFP was used. This binds growing MT plus ends and reliably identifies MT nucleation sites. Only one nucleation site was present in interphase and preprophase, and a new nucleation site appeared distant from the first, consistent with spatially and temporally distinct second MTOC activation. Thus, NBs regulate MT nucleation and not just MT retention (Rusan, 2007).
To examine how the new nucleation center forms, centrosomes were imaged using a PCM protein, GFP-Cnn. NBs contain a single detectable centrosome during interphase. When NBs reenter mitosis, a second Cnn-positive centrosome appears distant from the first, mimicking activation of the second MTOC. To verify that these occur simultaneously, NBs expressing mCherry-Tubulin (chTub) and GFP-Cnn were imaged. This revealed perfect temporal and spatial correlation between the appearance of the second centrosome and activation of the second MTOC. Physical separation of two centrosomes/MTOCs was never seen. This is the first example of asynchronous and physically distant centrosome maturation, suggesting that NBs use a novel centrosome cycle (Rusan, 2007).
Higher temporal/spatial resolution imaging revealed that two GFP-Cnn spots separate during mitotic exit. One GFP-Cnn spot persists as the NB interphase centrosome, forming the dominant MTOC, whereas the other spot disappears. The persistent Cnn spot (centrosome) remains relatively stationary in interphase, consistent with the hypothesis that coarse spindle alignment begins in interphase by anchoring the dominant centrosome (Rusan, 2007).
Centrosome fate was examined in the two daughters (new NB and GMC). They differ dramatically in PCM retention, in contrast to mammalian cells, where both daughters' centrosomes retain PCM. The GMC centrosome sheds all PCM. The new NB centrosome (that becomes the dominant MTOC) retains PCM throughout interphase and further accumulates PCM during the next mitosis. The complete shedding of PCM in GMCs appears to be the normal behavior of interphase centrosomes in most fly cells, whereas in syncytial early embryos, both daughters retain PCM foci through the cell cycle. In contrast to both cell types, the NB daughters exhibit differential PCM retention (Rusan, 2007).
These data suggest that NBs have a novel centrosome cycle in which the second centrosome matures distant from the dominant centrosome/MTOC. One hypothesis to explain this would be the distal positioning of a differentially regulated centriole that is blocked from recruiting PCM in interphase and thus cannot form an MTOC until 'activated' during mitosis. If this centriole is always inherited by the GMC, it might also explain complete PCM loss as GMCs exit mitosis (Rusan, 2007).
Centrioles were examined live to test this hypothesis, using the centriole marker GFP-PACT and Histone-GFP Mother/daughter centrioles disengaged in late telophase, as in mammalian cells and fly embryos. Thus, two NB centrioles are present throughout interphase despite the presence of only one MTOC (Rusan, 2007).
The two centrioles then exhibit different behaviors. One remains fairly stationary, whereas the second moves to roughly the other side of the nucleus. Disengagement perfectly correlates with separation of Cnn spots, suggesting that the stationary centriole retains PCM to form the dominant MTOC and the mobile centriole completely sheds PCM. To test this, NBs were imaged expressing chTub and GFP-PACT. The stationary centriole retained MTs, whereas the mobile centriole did not. Upon reentering mitosis, the mobile centriole regained nucleation activity, forming the second MTOC (Rusan, 2007).
This suggests that full separation of the MTOCs that organize the spindle is biphasic. It begins in interphase, when one centriole retains PCM, remains stationary, and forms the dominant MTOC, whereas the second centriole sheds PCM and becomes mobile. Movement of the second centriole away from the dominant MTOC in interphase accounts for ~70% of the separation needed to form a spindle. Mechanisms of transporting the mobile centriole remain to be identified, but it is nonrandom, as in 26/30 NBs, the second MTOC emerged more than 90° from the dominant MTOC. After the second MTOC is activated, the two separate the last 30%, most likely via MT sliding forces. This might explain defects in lissencephaly1 mutants, where MTOCs are only separated by 124° at NEB. Perhaps interphase centriole movement is normal, but MT-based MTOC separation is defective (Rusan, 2007).
These data suggest that NBs differentially regulate the activity of their two centrioles within the same cytoplasm. Interestingly, a similar observation was made in clam eggs, which have three centrosomes just after fertilization. The sperm centrosome is functionally inactivated, whereas female centrosomes organize the meiotic spindle (Rusan, 2007).
Next NB centrosome regulation was examined. In preprophase, one centriole (marked by anti-DPLP) formed the dominant MTOC, whereas the second centriole had no associated MTs and was randomly positioned, confirming the live-cell data. Thus whether γtub is recruited asymmetrically was examined. Fixed preprophase NBs had two centrioles; only that opposite the GMCs accumulated γtub. Further, both γtub and Cnn are absent from the NB centriole nearest the GMCs in interphase/preprophase (Rusan, 2007).
Polo kinase promotes centrosome maturation by promoting γtub recruitment during mitotic entry. Differences in Polo localization/activity might underlie differences in timing of NB centrosome maturation. NBs expressing Polo-GFP and the centriole marker mCherry-DSAS-6 were examined. Only the centriole pair that forms the dominant centrosome was Polo-GFP positive during preprophase. Polo-GFP accumulated on the mobile centriole pair as the NB entered mitosis, increased on both centriole pairs through prophase, and moved on to kinetochores. When Polo-GFP was examined in cells exiting mitosis, it was retained at low levels on the dominant centrosome. In the future, it will be interesting to examine the localization of Aurora A, another centrosome regulator (Rusan, 2007).
Unlike the distal appendages of mammalian mother centrioles, fly mother and daughter centrioles have been thought to have no known ultrastructural or molecular differences. These data suggest that differences exist. It is unlikely that this differential regulation is a result of location, as both centrioles are initially adjacent after disengagement. The differences may be due to centriole age or procentriole maturation state (Rusan, 2007).
The NB spindle is largely aligned by NEB. Based on the current data, the hypothesis was tested that the dominant centrosome helps define one spindle pole before prophase. The angle between the dominant centrosome/MTOC axis and the anaphase axis was calculated, using the nuclear centroid as a fixed reference.
This revealed two phases in defining the future spindle axis. Through prophase onset, the dominant centrosome remains fairly stationary roughly opposite the GMCs, agreeing with fixed images, whereas the second centriole moves to a distal position (to within 46 ± 33° of the anaphase axis). This is consistent with the hypothesis. The dominant centrosome may be immobilized by aster–cortex interactions or by absence of an active displacement mechanism. In the second phase, alignment is refined in prophase and prometaphase (the angle between the NB centrosome and anaphase axes decreases from 31 ± 29° to 15 ± 12° (Rusan, 2007).
To further test whether anchoring the dominant centrosome helps roughly align the spindle, asl mutant NBs were imaged live. They lack functional centrosomes and astral MTs. Mutant NBs lack a dominant interphase centrosome, allowing assessment of its role in spindle orientation and asymmetric cell division. Live imaging revealed robust chromatin-mediated MT nucleation and spindle assembly producing fairly normal spindles. Spindle poles emerge from a disorganized MT array near the chromosomes that focuses as the spindle lengthened. Spindles do not rotate during formation, always forming along the initial pole separation axis, but do rotate during metaphase, suggesting that rotation can occur without astral MTs or that asl mutants have a reduced astral array sufficient for rotation. Surprisingly, consecutive divisions in asl mutants usually produce adjacent or near-adjacent daughters, as in wild type. In a few cases, however, spindles form parallel to the GMC cap and, presumably, the polarity axis (2/13; ~15%); these NBs divide symmetrically. This suggests that the second phase of spindle alignment can occur without a dominant centrosome and can rescue misalignment, as long as it is not too extreme, but occasional atypical symmetric divisions occur. This results in defective brain anatomy, with ectopic paired, smaller NBs, presumably progeny of symmetric divisions (Rusan, 2007).
These data reveal two new aspects of asymmetric division in this stem cell model. First, cells can adjust the canonical centrosome cycle to allow novel cell behaviors, as was observed during clam meiosis. Central brain NBs also alter this cycle: rather than both centrosomes maturing in synchrony and proximity, the two centriole pairs are differentially regulated, maturing asynchronously and distant from one another. One retains MT nucleating activity throughout the cell cycle, forming the dominant MTOC during interphase, whereas the second is initially inactive, only forming a functional centrosome and nucleating MTs at mitotic entry. One speculative possibility is that these are mother and daughter centrioles and that one is preferentially retained in the stem cell, a hypothesis that will now be tested. It is also of interest to ask whether other stem cells use this mechanism (Rusan, 2007).
Second, the data suggest that this novel centrosome cycle helps ensure high-fidelity spindle positioning and thus asymmetric division. A model is proposed in which NB mitotic spindles are aligned in two phases to ensure that GMC daughters are born next to the previous GMC. Rough alignment is achieved by confining the dominant MTOC to a relatively fixed position from the previous division and moving the second centriole to the other side of the cell. As spindles form, a second process refines this initial alignment. In asl mutants, without centrosomes, the first mechanism is inactive, but the second mechanism can align the spindle unless initial alignment is wildly off axis. In mud mutants, centriole separation must occur normally, as prophase MTOCs are nearly fully separated, but alignment of spindle poles to cortical polarity cues is defective. The normal two-step process is a robust mechanism ensuring successful asymmetric divisions and reproducible brain anatomy (Rusan, 2007).
Asymmetric division of adult stem cells generates one self-renewing stem cell and one differentiating cell, thereby maintaining tissue homeostasis. A decline in stem cell function has been proposed to contribute to tissue ageing, although the underlying mechanism is poorly understood. This study shows that changes in the stem cell orientation with respect to the niche during ageing contribute to the decline in spermatogenesis in the male germ line of Drosophila. Throughout the cell cycle, centrosomes in germline stem cells (GSCs) are oriented within their niche and this ensures asymmetric division. GSCs containing misoriented centrosomes accumulate with age, and these GSCs are arrested or delayed in the cell cycle. The cell cycle arrest is transient, and GSCs appear to re-enter the cell cycle on correction of centrosome orientation. On the basis of these findings, it is proposed that cell cycle arrest associated with centrosome misorientation functions as a mechanism to ensure asymmetric stem cell division, and that the inability of stem cells to maintain correct orientation during ageing contributes to the decline in spermatogenesis. It was also shown that some of the misoriented GSCs probably originate from dedifferentiation of spermatogonia (Cheng, 2008).
GSCs with misoriented centrosomes accumulate as flies age. Since such misoriented GSCs divide less frequently as compared to oriented GSCs, accumulation of misoriented GSCs contributes to the decline in spermatogenesis that occurs with age. Although misoriented GSCs rarely divide, they are not permanently arrested (or senescent) and are correctly oriented when they divide. Whether correction of GSC orientation is an active process that is part of the acquisition of stem cell identity remains to be determined. The low cell cycle activity of misoriented GSCs may also suggest that mechanisms are in place to detect misorientation and induce cell cycle arrest in response to this change, although the underlying mechanisms remain to be identified (Cheng, 2008).
It was also demonstrated that misoriented GSCs originate, at least in part, from dedifferentiation of spermatogonia. Although dedifferentiated GSCs have high frequency (>40%) of centrosome misorientation, they can function as stem cells by resuming the cell cycle, with correctly oriented mitotic spindles just like as constitutive GSCs. GSC numbers do not decrease as quickly as expected from the calculated GSC half-life, suggesting that a mechanism to compensate for the loss of GSCs exists. Since misoriented spindles, or symmetric stem cell division, was rarely observed, it is speculated that dedifferentiation is the major mechanism to replace stem cells over time in the Drosophila male germ line (Cheng, 2008).
A decline in GSC number in older males (day 50) was reported recently (Boyle, 2007) This decrease in stem cell number is likely due to failure of the niche function (via decreased signal from the niche as well as decreased E-cadherin-based attachment between the niche and GSCs. However, the decrease in the production of spermatogonia and testis involution precede the loss of GSCs such that decreasing GSC numbers cannot explain the testis involution that is observed at younger ages (Cheng, 2008).
The present results provide a novel mechanistic link between the control of stem cell polarity and the age-related decline in tissue regenerative capacity. Mechanisms responsible for monitoring stem cell orientation with respect to the niche not only prevent overproliferation of stem cells by ensuring the asymmetric outcome of the stem cell division, but they contribute to the decline in tissue regenerative capacity during aging. Many of the misoriented GSCs originate from the dedifferentiation of spermatogonia, a mechanism thought to be responsible for maintaining the stem cell population over extended periods of time. Therefore, although GSCs produce less progeny over time, the system appears to maximize the number of progeny produced throughout life, while maintaining asymmetric stem cell division (Cheng, 2008).
In summary, it is proposed that the GSCs with misoriented centrosome divide less frequently and that a combination of such a decreased stem cell division and a higher frequency of the GSC misorientation in aged testes leads to a decline in spermatogenesis with age (Cheng, 2008).
Asymmetric divisions of radial glia progenitors produce self-renewing radial glia and differentiating cells simultaneously in the ventricular zone (VZ) of the developing neocortex. Whereas differentiating cells leave the VZ to constitute the future neocortex, renewing radial glia progenitors stay in the VZ for subsequent divisions. The differential behaviour of progenitors and their differentiating progeny is essential for neocortical development; however, the mechanisms that ensure these behavioural differences are unclear. This study shows that asymmetric centrosome inheritance regulates the differential behaviour of renewing progenitors and their differentiating progeny in the embryonic mouse neocortex. Centrosome duplication in dividing radial glia progenitors generates a pair of centrosomes with differently aged mother centrioles. During peak phases of neurogenesis, the centrosome retaining the old mother centriole stays in the VZ and is preferentially inherited by radial glia progenitors, whereas the centrosome containing the new mother centriole mostly leaves the VZ and is largely associated with differentiating cells. Removal of ninein, a mature centriole-specific protein, disrupts the asymmetric segregation and inheritance of the centrosome and causes premature depletion of progenitors from the VZ. These results indicate that preferential inheritance of the centrosome with the mature older mother centriole is required for maintaining radial glia progenitors in the developing mammalian neocortex (Wang, 2009).
Radial glia cells constitute a major population of neural progenitor cells that occupy the proliferative VZ in the developing mammalian neocortex. In addition to their well-characterized function as a scaffold in supporting neuronal migration, radial glia cells display interkinetic nuclear oscillation and proliferate extensively at the luminal surface of the VZ (that is, the VZ surface). During the peak phase of neurogenesis [around embryonic day 13-18 (E13-E18) in mice] they predominantly undergo asymmetric division to self-renew while simultaneously giving rise either directly to a neuron, or to an intermediate progenitor cell which subsequently divides symmetrically to produce neurons. Whereas differentiating progeny progressively migrate away from the VZ to form the cortical plate (CP) (the future neocortex) renewing radial glia progenitors remain in the VZ for subsequent divisions. The distinct migratory behaviour of radial glia progenitors and their differentiating progeny is fundamental to the proper development of the mammalian neocortex; however, little is known about the basis of these behavioural differences (Wang, 2009).
Centrosomes, the main microtubule-organizing centres in animal cells, have an important role in many cell processes, particularly during cell division. All normal animal cells initially inherit one centrosome, consisting of a pair of centrioles surrounded by an amorphous pericentriolar material. The two centrioles differ in their structure and function. The older 'mother' centriole, which is formed at least one-and-a-half generations earlier, possesses appendages/satellites that bear specific proteins, such as cenexin (also known as Odf2), and anchor microtubules and support ciliogenesis. In contrast, the younger 'daughter' centriole, which is formed during the preceding S phase, lacks these structures. Full acquisition of appendages/satellites by the daughter centriole is not achieved until at least one-and-a-half cell cycles later. During each cell cycle, the centrosome replicates once in a semi-conservative manner, resulting in the formation of two centrosomes: one of which retains the original old mother centriole (that is, the mother centrosome) while the other receives the new mother centriole (that is, the daughter centrosome). This intrinsic asymmetry in the centrosome has recently been demonstrated to be important for proper spindle orientation during the division of male germline stem cells and neuroblasts in Drosophila, although female germline stem cells appear to divide normally in the absence of centrioles/centrosomes. These studies indicate a critical role for the differential behaviour of centrosomes with differently aged mother centrioles in asymmetric division of the progenitor/stem cells, although it remains unclear whether proper behaviour and development of the progenitor/stem cells and their differentiating daughter cells depend on centrosome asymmetry. Asymmetric division of radial glia progenitors accounts for nearly all neurogenesis in the developing mammalian neocortex. Three out of four autosomal recessive primary microcephaly (MCPH) genes identified so far encode centrosomal components, suggesting that proper neocortical neurogenesis and development entail a tight regulation of the centrosome, which is so far poorly understood. To address these issues, centrosome regulation during the peak phase of mammalian neocortical neurogenesis was investigated (Wang, 2009).
To examine centrosome behaviour, a plasmid encoding centrin 1, a central component of the centriole, fused with enhanced green fluorescent protein (EGFP-CETN1) was introduced into the developing neocortex of E13.5 mouse embryos by in utero electroporation. As expected, EGFP-CETN1 formed pairs of dots that co-localized with γ-tubulin, a centrosomal marker, suggesting that transient expression of EGFP-CETN1 reliably labels the two centrioles of individual centrosomes in the developing neocortex in vivo. Moreover, it was observed that at the onset of peak neurogenesis (E13-E14), EGFP-CETN1-labelled centrosomes were predominantly located at the VZ surface with a small subset located in the subventricular zone (SVZ) and the intermediate zone (IZ) (Wang, 2009).
To identify the cell types harbouring EGFP-CETN1-labelled centrosomes, a plasmid encoding DsRedexpress (DsRedex), a red fluorescent protein that diffuses throughout cells and thereby reveals their morphology, was co-electroporated . In bipolar radial glia progenitors in the VZ the centrosome was located in their ventricular endfeet at the VZ surface, whereas in multipolar cells in the IZ and the SVZ the centrosome was harboured in their cell bodies. Moreover, dividing radial glia progenitors were observed that possess a pair of centrosomes together with condensed chromosomes at the VZ surface. Consistent with this differential centrosome localization between radial glia progenitors and their differentiating daughter cells, a progressive increase was observed in the appearance of EGFP-CETN1-labelled centrosomes in the IZ and the CP as development proceeded, in addition to some that remained at the VZ surface. This gradual increase in centrosome localization in the IZ and the CP coincided with the production and migration of differentiating cells such as neurons to these regions during this period (Wang, 2009).
The distinct positioning of the centrosome in radial glia progenitors versus their differentiating progeny prompted asking whether the centrosomes in these two cell populations/types are different. To explore this, a plasmid encoding ninein, a mature centriole-specific protein that localizes to appendages/satellites, fused with EGFP (EGFP-Nin) together with a plasmid encoding CETN1 fused to DsRedex (DsRedex-CETN1) were electroporated into the developing mouse neocortex at E13.5. As expected, both EGFP-Nin and DsRedex-CETN1 formed dot-like structures and co-localized to the centrosomes, especially those at the VZ surface, as identified by an antibody to the integral centrosomal protein pericentrin 1. Notably, EGFP-Nin was preferentially concentrated at one of the two centrioles marked by DsRedex-CETN1 in individual centrosomes, suggesting that the two centrioles in interphase radial glia progenitors are not identical. Given that Nin specifically associates with mature centrioles, these results indicate that the centriole with abundant EGFP-Nin is the more mature mother centriole, whereas the one with little EGFP-Nin is the less mature daughter centriole. A similar inequity in the recruitment of EGFP-Nin by the duplicated centrosomes was also observed in dividing radial glia progenitors at the VZ surface, indicating that the duplicated centrosomes are not identical during late mitosis. The centrosome with abundant EGFP-Nin is probably the centrosome that retains the mature old mother centriole and the centrosome with little EGFP-Nin is probably the centrosome that bears the relatively immature new mother centriole (Wang, 2009).
Having found that the centrosomes in dividing radial glia progenitors exhibit asymmetry in their maturity, it was next asked whether this centrosome asymmetry is related to the distinct behaviour of radial glia progenitors and their differentiating progeny in the developing neocortex during neurogenesis. To address this, the relative distribution of centrosomes labelled by DsRedex-CETN1 versus those labelled by EGFP-Nin in the developing neocortex were compared as development proceeded. Interestingly, whereas DsRedex-CETN1-labelled centrosomes progressively occupied the IZ and the CP, where differentiating cells are situated, EGFP-Nin-labelled centrosomes were mostly found in the VZ, where radial glia progenitors are located. Given that DsRedex-CETN1 labels all centrosomes whereas EGFP-Nin selectively labels mature centrosomes, these results point to an intriguing possibility that the duplicated centrosomes in dividing radial glia cells are differentially inherited depending on their age and maturity during neocortical neurogenesis. It is known that during each cell division one centrosome retains the old mature mother centriole and the other bears the new less mature mother centriole. Thus, these results suggest that centrosomes with differently aged mother centrioles are differentially inherited by the two daughter cells of asymmetrically dividing radial glia progenitors (Wang, 2009).
To test this, an assay was developed to distinguish explicitly between the centrosome containing the old mother centriole and the centrosome containing the new mother centriole in the developing neocortex in vivo. The assay takes advantage of the photoconvertible fluorescent protein Kaede, which changes from green to red fluorescence on exposure to violet light. Centrioles in the developing neocortex were labelled by transient expression of CETN1 fused with Kaede (Kaede-CETN1). Photoconversion was then performed to switch labelled centrioles from green to red fluorescence. This green-to-red fluorescence conversion of Kaede proteins is irreversible and the red protein is very stable, thus allowing long-term tracking of the existing photoconverted proteins and the structures with which they are associated. Moreover, all newly synthesized Kaede proteins are green fluorescent. It is known that centriole duplication requires new protein synthesis of centrin. As a result, newly duplicated centrioles that contain newly synthesized Kaede-CETN1 are green fluorescent, whereas previously existing centrioles are red fluorescent. Hence, in the first cell cycle after photoconversion, both centrosomes contain a red fluorescent mother centriole and a green fluorescent daughter centriole. However, in the second and subsequent cell cycles, centrosomes with the new mother centriole contain only green fluorescent centrioles, whereas centrosomes retaining the original old mother centriole harbour both red and green fluorescent centrioles, thus distinguishing between centrosomes with differently aged mother centrioles (Wang, 2009).
To carry out this assay in the developing neocortex in vivo, an in utero photoconversion procedure was developed and and combined with in utero electroporation. Kaede-CETN1, which localized specifically to the centrosomes, was introduced into the developing mouse neocortex at E13.5. One day later, that is, E14.5, the forebrain of electroporated embryos was treated with a short exposure of violet light while still in the uterus, which effectively converted nearly all Kaede-CETN1 proteins and their labelled centrosomes from green to red fluorescence [E13.5-E14.5(PC)]. The uterus was replaced and the embryos continued to develop in vivo. The localization and inheritance of centrosomes were analysed at different developmental stages thereafter (Wang, 2009).
It was found that one day after photoconversion [E13.5-E14.5(PC)-E15.5], around 95% of centrosomes contained both red and green fluorescent centrioles (indicated by yellow colour in the merged image), consistent with the notion that the labelled cells have undergone one round of division and have duplicated their centrioles during the 24-h period after photoconversion. This was shown directly by imaging centrosomes at high magnification, revealing that each centriole was mostly only red or green fluorescent. This also demonstrates that there is little diffusion of centrin proteins between duplicated centrioles, or between the centrioles and a cytoplasmic pool which was confirmed by fluorescence recovery after photobleaching (FRAP) experiments. Moreover, it was found that more than 30% of centrosomes possessed only green fluorescent centrioles 2 days after photoconversion [E13.5-E14.5(PC)-E16.5]. The appearance of the solely green fluorescent centrosomes 48 h after photoconversion indicates that the initially labelled radial glia progenitors have undergone two rounds of division during this period; this is consistent with the previous observation that the duration of the neocortical progenitor cell cycle is about 12 to 20 h around this developmental stage. The ongoing division of labelled radial glia cells at a normal rate suggests that expression of Kaede-CETN1 and the photoconversion procedure had no effect on their cell cycle. In addition, no obvious DNA damage or cell death was induced by the photoconversion treatment. Besides the green and yellow fluorescent centrosomes, about 4% of solely red fluorescent centrosomes were observed, indicating that a few labelled cells do not undergo cell division during this period (Wang, 2009).
Having successfully distinguished the centrosomes with differently aged mother centrioles in the developing neocortex in vivo, their distribution was examined to determine whether they are asymmetrically segregated. Remarkably, it was found that more than 76% of centrosomes with the new mother centriole (that is, only green fluorescent) were located in the IZ and the CP, whereas around 78% of centrosomes with the old mother centriole (that is, both green and red fluorescent) were located in the VZ in addition to the SVZ. These results demonstrate that the centrosomes with differently aged mother centrioles are asymmetrically segregated in the developing neocortex during the peak phase of neurogenesis. It is worth noting that a small fraction of both green and red (that is, yellow) fluorescent centrosomes was found in the IZ and the CP and that these centrosomes probably originated from the first cell cycle after photoconversion (Wang, 2009).
The asymmetric segregation of centrosomes suggests differential regulation of the duplicated centrosomes in dividing radial glia progenitors. To gather further evidence for this, time-lapse imaging experiments were carried out to monitor the behaviour of centrosomes with differently aged mother centrioles in dividing radial glia progenitors at the VZ surface in situ. Kaede-CETN1 was introduced into radial glia cells together with mPlum, a far-red fluorescent protein, to label cell morphology. Around 24 h later, cortical slices were prepared. Photoconversion of the existing Kaede-CETN1 proteins was then performed in individual slices, which were then cultured for another 24 h before being subjected to time-lapse imaging. Labelled dividing radial glia cells with enlarged and rounded cell bodies possessing a pair of centrosomes at the VZ surface were monitored at 10-min intervals over a period of 5 to 8 h. In six out of seven dividing radial glia cells that proceeded through mitosis at the VZ surface and reached the two-cell stage, the centrosome retaining the old mother centriole in both red and green fluorescence stayed at the VZ surface, whereas the centrosome containing the new mother centriole in solely green fluorescence migrated away from the VZ surface. These results demonstrate that the centrosomes with differently aged mother centrioles in dividing radial glia progenitors exhibit distinct behaviour during the peak phase of neurogenesis (Wang, 2009).
The distinct behaviour of the centrosomes suggests that they are differentially inherited by the two daughter cells embarking on different routes of fate specification and development. On the basis of their behaviour, it is postulated that the centrosome with the new mother centriole is largely inherited by differentiating cells, such as neurons, whereas the centrosome with the old mother centriole that remains located at the VZ is mostly inherited by radial glia progenitors. Indeed, it was found that 2 days after photoconversion [E13.5-E14.5(PC)-E16.5] the centrosomes with the new mother centriole, marked by green fluorescence alone, were mostly associated with cells expressing TUJ1, a differentiating neuronal marker, in the CP and the IZ. In contrast, the centrosomes that retained the old mother centriole in yellow (that is, both green and red) fluorescence at the VZ were largely associated with cells expressing Pax6, a radial glia progenitor marker. These results show that the centrosomes with differently aged mother centrioles in dividing radial glia cells are asymmetrically inherited by the two daughter cells: whereas the renewing radial glia progenitor inherits the centrosome with the old mother centriole, the differentiating daughter cell inherits the centrosome with the new mother centriole (Wang, 2009).
These data thus far show that centrosomes with differently aged mother centrioles are differentially inherited by the two daughter cells of asymmetrically dividing radial glia progenitors in the developing neocortex. Next, whether the selective inheritance of the centrosome with the old mature mother centriole by radial glia progenitors is necessary for their maintenance at the VZ was tested. Should this be the case, given that Nin is an essential component of the appendage/satellite structures specific to the mature centriole, it was predicted that removal of Nin, which prevents centriole maturation, would disrupt asymmetric segregation of centrosomes with differently aged mother centrioles and impair the maintenance of radial glia progenitors in the developing neocortex (Wang, 2009).
To test this, short hairpin RNA (shRNA) sequences were developed against Nin, that effectively suppressed its expression (Nin shRNAs). Consistent with the prediction, expression of Nin shRNA, but not control shRNA, disrupted asymmetric segregation of centrosomes with differently aged mother centrioles labelled with Kaede-CETN1 in the developing neocortex, suggesting that Nin is necessary for centriole maturation, thereby generating asymmetry between duplicated centrosomes. The presence of solely green fluorescent centrosomes in Nin shRNA-expressing cortices indicates that centrosome duplication and segregation and cell division are not severely affected by removal of Nin. More importantly, it was found that removal of Nin caused a premature depletion of cells from the VZ, where radial glia progenitors reside. This effect of Nin shRNAs correlated with their efficacy in suppressing Nin protein expression and was rescued by a shRNA-insensitive Nin plasmid, suggesting that the effect of the Nin shRNA is due to a specific depletion of the endogenous Nin protein. A similar reduction in cells in the VZ was observed when Nin expression was suppressed using small interfering RNA (siRNA) (Wang, 2009).
To characterize further the extent to which removal of Nin leads to a depletion of radial glia progenitors, the fate specification was examined of cells expressing either control or Nin shRNA. When compared with the control, expression of Nin shRNA led to a marked reduction in the percentage of cells positive for Pax6 and glutamate transporter (GLAST), two radial glia progenitor markers, and a significant increase in the percentage of cells positive for TUJ1, a differentiating neuronal marker. These results suggest that removal of Nin leads to a depletion of radial glia progenitors and a concomitant increase in differentiating neurons. Consistent with this, a significant reduction was observed in phospho-histone 3 (P-H3)-labelled mitotic cells at the VZ surface and a marked increase in cell cycle exit. No obvious change in the cleavage plane orientation of late stage mitotic cells at the VZ surface was observed (Wang, 2009).
Previous studies showed that the carboxy-terminus of Nin is responsible for its localization to the centriole and expression of this region displaces endogenous protein at the centriole. Interestingly, it was found that, similar to removal of Nin, expression of the carboxy-terminus of Nin (Nin-Cter) led to a premature depletion of radial glia progenitor cells from the VZ, suggesting that centriolar Nin is critical for maintaining radial glia progenitor cells in the VZ. Taken together, these results indicate that preferential inheritance of a centrosome containing the mature mother centriole is required for the maintenance of radial glia progenitors in the proliferative VZ of the developing neocortex (Wang, 2009).
The results presented in this study suggest that the centrosomes with differently aged centrioles in asymmetrically dividing radial glia progenitors exhibit different behaviour and are differentially inherited by the two daughter cells during the peak phase of mammalian neocortical neurogenesis. Whereas the centrosome with the less mature new mother centriole migrates away from the VZ surface and is largely inherited by differentiating cells, the centrosome with the more mature old mother centriole stays at the VZ surface and is predominantly inherited by renewing radial glia progenitors. Recently, asymmetric behaviour of centrosomes has been observed during asymmetric division of Drosophila male germline stem cells and neuroblasts. The findings of this study suggest that this type of asymmetric centrosome regulation may be a general feature of asymmetric cell division across species. Furthermore, the findings provide new insight into centrosome regulation in the developing mammalian neocortex, which has been linked to the pathogenesis of human microcephaly (Wang, 2009).
Centrosomes with differently aged mother centrioles differ in their protein composition and thereby in their biophysical properties, such as microtubule anchorage activity. In this study, it was found that Nin, an appendage/satellite-specific protein required for centriole maturation, localized differently to the duplicated centrosomes in radial glia progenitors in late mitosis. Notably, another appendage/satellite-specific protein cenexin was recently found to be asymmetrically localized to centrosomes in sister cells after mitosis; moreover, the cell receiving the more mature old mother centriole usually grew a primary cilium first (Anderson, 2009). The asymmetric inheritance of centrosomes with distinct biophysical properties may thereby differentially regulate the behaviour and development of the daughter cells that receive them. For example, given that primary cilia have essential roles in a number of signal transduction pathways, including Sonic hedgehog (Shh) and platelet-derived growth factor (PDGF) signalling, the asynchrony in cilium formation could differentially influence the ability of the two daughter cells to respond to environmental signals and thereby their behaviour and fate specification. Furthermore, the strong microtubule anchorage activity associated with the centrosome retaining the older mother centriole would facilitate its anchorage to a specific site (for example, the VZ surface), thereby tethering the cell that inherits it. Indeed, it was found that disruption of centriole maturation by removing Nin not only impairs asymmetric segregation of centrosomes, but also depletes radial glia progenitors from the VZ, a proliferative niche in the developing mammalian neocortex. Aside from their participation in microtubule organization and ciliogenesis, centrosomes associate with messenger RNAs (mRNAs) and membrane-bound organelles such as the Golgi and recycling endosomes and regulate protein degradation, thereby raising the possibility that asymmetric centrosome inheritance might contribute to proper segregation of cell fate determinants to the two daughter cells of asymmetrically dividing progenitor/stem cells (Wang, 2009).
In Caenorhabditis elegans, five proteins are required for centriole duplication: SPD-2, ZYG-1, SAS-5, SAS-6, and SAS-4. Functional orthologues of all but SAS-5 have been found in other species. In Drosophila and humans, Sak/Plk4, DSas-6/hSas-6, and DSas-4/CPAP-orthologues of ZYG-1, SAS-6, and SAS-4, respectively-are required for centriole duplication. Strikingly, all three fly proteins can induce the de novo formation of centriole-like structures when overexpressed in unfertilized eggs. This study finds that of eight candidate duplication factors identified in cultured fly cells, only two, anastral spindle 2 (Ana2) and Asterless (Asl), share this ability. Asl is now known to be essential for centriole duplication in flies, but no equivalent protein has been found in worms. This study shows that Ana2 is the likely functional orthologue of SAS-5 and that it is also related to the vertebrate STIL/SIL protein family that has been linked to microcephaly in humans. It is proposed that members of the SAS-5/Ana2/STIL family of proteins are key conserved components of the centriole duplication machinery (Stevens, 2010).
The centriole is composed of a radial array of nine microtubule (MT) triplets, doublets, or singlets depending on species and cell type. Centrioles are required to make two important cellular structures: centrosomes and cilia. The centrosome consists of a pair of centrioles surrounded by pericentriolar material (PCM) and is the major MT organizing center in many animal cells. Cilia are formed when the centriole pair migrates to the cell cortex, and the older, mother, centriole forms a basal body that nucleates the ciliary axoneme. Many different cell types possess cilia, and they have multiple roles in development (Stevens, 2010).
To ensure their inheritance by each daughter cell, centrioles duplicate precisely once per cell cycle. This process must be tightly regulated. Failure in centriole duplication leads to catastrophic errors during embryogenesis in both worms and flies, and an increasing number of human diseases have been linked to defects in centrosome and/or cilia function. Centriole overduplication can be equally damaging, as excess centrioles are frequently observed in human tumors, and there appears to be a direct causative relationship between centriole overduplication and tumorigenesis in flies (Stevens, 2010).
In canonical centriole duplication, a new daughter centriole grows at a right angle to the mother centriole. A series of genome-wide RNAi and genetic screens in worms have found just five proteins essential for centriole duplication: SPD-2, ZYG-1, SAS-5, SAS-6, and SAS-4. SPD-2 is required to recruit the kinase ZYG-1 to the centriole, and both proteins then recruit a complex of SAS-5 and SAS-6. SAS-5 and SAS-6 are mutually dependent for their centriolar localization and are in turn needed to recruit SAS-4 (Stevens, 2010 and references therein).
Although DSpd-2 is not essential for centriole duplication in flies, and no SAS-5 homologues have been identified outside worms, proteins related to ZYG-1, SAS-6, and SAS-4 have a conserved role in centriole duplication in other systems. In Drosophila, for example, the kinase Sak, which is related to ZYG-1, and the homologues of SAS-6 (DSas-6) and SAS-4 (DSas-4) are required for centriole duplication. Recently, however, several additional proteins have been identified in cultured fly cells that are potentially involved in centriole duplication. This study set out to identify which of these potential duplication factors are likely to function as upstream regulators of centriole formation (Stevens, 2010).
Genome-wide RNAi screens in cultured fly cells identified just 18 proteins that, when depleted, gave a reduced number of centrioles. This list includes Sak, DSas-6, and DSas-4, as well as eight other proteins that specifically localize to centrosomes (Ana1, Ana2, Ana3, Asl, DCP110, DCep135/Bld10, DCep97, and Rcd4): these eight are therefore good candidates to play a direct role in centriole duplication (Stevens, 2010).
GFP-Sak, GFP-DSas-6, and DSas-4-GFP share the unique ability to drive de novo formation of centriole-like structures in unfertilized eggs when highly overexpressed from the upstream activation sequence (UAS) promoter. UAS-GFP-Sak and UAS-GFP-DSas-6 induce these structures in ~95% of unfertilized eggs, whereas UAS-DSas-4-GFP does so in ~60% of unfertilized eggs. It was asked if this assay could be used to identify other components likely to function upstream in the centriole duplication pathway. Transgenic lines were generated carrying GFP fusions to all eight potential duplication factors under the control of the UAS promoter, which allowed overexpression in unfertilized eggs. Strikingly, only Ana2 (in 97% of eggs) and Asl (in 33% of eggs) were able to drive de novo formation of centriole-like structures (Stevens, 2010).
Asl has recently been shown to be essential for centriole duplication in flies, whereas, of the six proteins unable to induce de novo centriole formation, two, DCep135/Bld10 and Ana3, are now known not to be essential for centriole duplication in flies. These findings indicate that the overexpression assay can identify those proteins likely to be most intimately involved in centriole duplication. Since Asl has already been shown to be required for centriole duplication, focus was placed on investigating the function of Ana2 (Stevens, 2010).
Ana2 can drive de novo formation of centriole-like structures as efficiently as DSas-6 and Sak. It was important to verify, however, that Ana2 also has a role in canonical centriole duplication. Overexpressing GFP-Sak or GFP-DSas-6 from the ubiquitin (Ubq) promoter induces centriole overduplication in brains and embryos, respectively. Surprisingly, however, overexpression of Sak, DSas-6, or DSas-4 cannot drive centriole overduplication in primary spermatocytes, which suggests that another duplication protein is limiting. To test if Ana2 might be this limiting factor, Ubq-GFP-Ana2 transgenic lines were generated. Strikingly, it was found that in spermatocytes expressing Ubq-GFP-Ana2, in addition to the normal centriole pairs (doublets), centriole triplets, quadruplets, and even quintets were observed. The extra centrioles in these clusters appeared to be fully functional; they separated from one another by the end of meiosis I (as centriole doublets normally do), and the extra centrioles inherited by secondary spermatocytes recruited PCM and nucleated MT asters, and so formed multipolar spindles during meiosis II (Stevens, 2010).
It was important to compare the localization of Ana2 with that of the other Drosophila centriole duplication factors. DSas-4-GFP, GFP-DSas-6, and GFP-Sak are all enriched at the proximal and distal ends of the large spermatocyte centrioles. It was found that, likewise, Ana2-GFP localized preferentially to the proximal and distal centriole tips. Strikingly, however, Ana2-GFP (and GFP-Ana2) also exhibited a unique asymmetric distribution, consistently localizing preferentially along one centriole barrel (Stevens, 2010).
In primary spermatocytes, it is possible to distinguish mother and daughter centrioles, as the daughter can often be observed associating end-on with the side of the mother. In 25 centriole pairs where mother and daughter centrioles could unambiguously distinguished, Ana2-GFP was always enriched on the daughter. Mother and daughter centrioles can show important differences in their behavior in vertebrate cells and during asymmetric stem cell divisions in Drosophila. Although mother and daughter centrioles are morphologically and molecularly distinguishable in vertebrates, this is not the case in Drosophila. Ana2-GFP is the first fly protein shown to localize asymmetrically to mother and daughter centrioles in this manner (Stevens, 2010).
Interestingly, as spermatocytes progressed through meiosis I, this centriolar asymmetry became less pronounced, and this appeared to reflect the selective loss of GFP-Ana2 from the daughter centriole, bringing its levels down to that of the mother. Since overexpression of Ana2 can lead to centriole overduplication, Ana2 levels presumably must normally be tightly regulated to prevent the formation of extra centrioles (Stevens, 2010).
After exit from meiosis II, each spermatid inherits a single centriole, which acts as a basal body to nucleate the flagellar axoneme. Structural components of the centriole, like Ana3 and Drosophila pericentrin-like protein (D-PLP), continue to localize along the basal body. In contrast, Ana2, like the conserved duplication proteins, was undetectable along the basal body. Ana2 did, however, colocalize with GFP-DSas-6 at the proximal centriole-like structure, a small nodule adjacent to the basal body that has been proposed to be an early intermediate in centriole formation (Stevens, 2010).
Intriguingly, Drosophila homologues have been identified for all the C. elegans centriole duplication factors except SAS-5, which has no clear homologues outside worms. Ana2 and SAS-5 are similar in size and have a single central coiled-coil domain, leading to a suggeston that Ana2 could be the Drosophila equivalent of SAS-5. As SAS-5 interacts with SAS-6 in worms, genetic interaction between Ana2 and DSas-6 was tested in flies (Stevens, 2010).
A small percentage of eggs laid by mothers carrying two copies of a Ubq-GFP-DSas-6 transgene assemble centriole-like structures. To see if this effect could be enhanced, flies were generated carrying one copy of Ubq-GFP-DSas-6 and one copy of Ubq-Ana2-GFP, neither of which alone (as a single copy) induces the assembly of centriole-like structures. Strikingly, almost all the unfertilized eggs laid by these females contained hundreds of large structures that stained for centriole markers, recruited PCM, and nucleated asters. Importantly this interaction was specific to Ana2 and DSas-6. In eggs from mothers carrying one copy of either Ubq-Ana2-GFP or Ubq-GFP-DSas-6 together with one copy of either Ubq-GFP-Sak, Ubq-Asl-GFP, or Ubq-DSas-4-GFP, at most a very small number of asters were observed in very few eggs (Stevens, 2010).
Interestingly, the centriole-like structures produced by overexpressing UASp-GFP-DSas-6 differ significantly from those resulting from the overexpression of GFP-Sak, DSas-4-GFP, Asl-GFP, or Ana2-GFP in that they are much larger and often appear ring-shaped, and that only one structure is contained within each aster. The structures in the eggs from females expressing both Ubq-GFP-DSas-6 and Ubq-Ana2-GFP were similar to this DSas-6 type. This suggests that Ana2-GFP acts to promote the assembly of GFP-DSas-6 into these structures (Stevens, 2010).
Having shown that Ana2 functionally interacts with DSas-6, physical interaction was sought. Using a yeast two-hybrid (Y2H) assay, it was found that Ana2 and DSas-6 interact and that the N-terminal region of DSas-6 and the C-terminal region of Ana2 are necessary and sufficient for this interaction. Moreover, like SAS-5, Ana2 also interacts with itself. Attempts to test whether Ana2 and DSas-6 associate in vivo were hindered by their low abundance. However, it was found that DSas-6 antibodies coimmunoprecipitated Ana2-GFP from S2 cells overexpressing Ana2-GFP. Collectively, the evidence of a specific functional and physical interaction between Ana2 and DSas-6 indicates that Ana2 likely represents the Drosophila functional orthologue of SAS-5 (Stevens, 2010).
Having shown that Ana2 is the likely SAS-5 functional orthologue in Drosophila, Ana2/SAS-5 orthologues were sought in other species. Using an iterative BLAST search, significant homology was found between Ana2 and the STIL or SIL protein family. Moreover, the reciprocal iterative BLAST search starting with zebrafish STIL identified Ana2 as the most similar Drosophila protein. Although vertebrate STIL family members are larger than Ana2 or SAS-5, all of these proteins share a short, central, coiled-coil domain. In addition, a particularly conserved region of ~90 aa toward the C terminus of Ana2 and STIL was identified, that was called the STil/ANa2 (STAN) motif. The STAN motif of Ana2 is 31% identical (48% similar) to that of zebrafish STIL. A divergent STAN motif can be detected in SAS-5, which is 12% identical (26% similar) to that of zebrafish STIL. Importantly, the STAN motif is within the regions of SAS-5 and Ana2 that interact with SAS-6 and DSas-6, respectively (Stevens, 2010).
Data from studies of STIL in mice, zebrafish, and humans are consistent with a function in centriole duplication, although this was not appreciated at the time of these studies. First, mitotic spindles often lack centrosomes in stil mutant zebrafish. Second, STIL mutant mice show defects characteristic of aberrant cilia function, such as randomized left-right asymmetry and neural tube abnormalities. Most importantly, it has recently been shown that mutations in human STIL cause primary microcephaly (MCPH), a congenital disorder characterized by reduced brain size. Mutations in four other genes, MCPH1, CDK5RAP2, ASPM, and CPAP/CENPJ, are known to cause MCPH, and all are centrosomal proteins, strongly suggesting that STIL is required for efficient centrosome function in humans (Stevens, 2010).
This study has show that of eight centrosomal proteins identified as potential duplication factors in Drosophila tissue culture cells, only two, Asl and Ana2, appear to be able to induce de novo formation of centriole-like structures in unfertilized eggs. Asl has recently been shown to be essential for centriole duplication, and this study provides evidence that Ana2 is also a key centriole duplication factor. Thus, Ana2 and Asl join Sak, DSas-6, and DSas-4 to make up a module of just five proteins known to drive centriole duplication in flies (Stevens, 2010).
These data strongly suggest that the Ana2/STIL family of centrosomal proteins are the long-sought functional orthologues of SAS-5. Thus, four of these five components (Sak/ZYG-1, DSas-6/SAS-6, Ana2/SAS-5, and DSas-4/SAS-4) are functionally conserved between flies and worms. Moreover, three of these proteins are required for centriole duplication in humans, whereas the fourth, SAS-5/Ana2/STIL, also appears likely to be required for this process in vertebrates (Stevens, 2010).
Both flies and worms have an additional protein (SPD-2 in worms, Asl in flies) that appears to be essential for centriole duplication. Intriguingly, both SPD-2 are not only required for centriole duplication, but also for PCM recruitment. There is evidence that the PCM promotes centriole duplication, so SPD-2 and Asl could play a more indirect role in centriole duplication via their ability to recruit PCM. Alternatively, both proteins may act directly in centriole duplication, with the function of SPD-2 in worms perhaps being performed by Asl in flies (Stevens, 2010).
In summary, this study shows that Ana2 acts as a centriole duplication factor in Drosophila and is likely to have a conserved role in other species. Overall, centriole duplication appears to be a highly conserved process, at the heart of which is a small number of key proteins. The challenge will now be to tease apart how these components cooperate to build a centriole of the right size, in the right place, and at the right time (Stevens, 2010).
Klp10A is a kinesin-13 of Drosophila melanogaster that depolymerizes cytoplasmic microtubules. In interphase, it promotes microtubule catastrophe; in mitosis, it contributes to anaphase chromosome movement by enabling tubulin flux. This study shows that Klp10A also acts as a microtubule depolymerase on centriolar microtubules to regulate centriole length. Thus, in both cultured cell lines and the testes, absence of Klp10A leads to longer centrioles that show incomplete 9-fold symmetry at their ends. These structures and associated pericentriolar material undergo fragmentation. It was also show that in contrast to mammalian cells where depletion of CP110 leads to centriole elongation, in Drosophila cells it results in centriole length diminution that is overcome by codepletion of Klp10A to give longer centrioles than usual. How loss of centriole capping by CP110 might have different consequences for centriole length in mammalian and insect cells is discussed, and also these findings are related to the functional interactions between mammalian CP110 and another kinesin-13, Kif24, that in mammalian cells regulates cilium formation (Delgehyr, 2012).
To study Klp10A's roles at the spindle poles, flies with reduced Klp10A expression were examined that showed male sterility resulting from a P element insertion. Meiotic cells in Klp10A testes had supernumerary asters and abnormal, frequently multipolar meiotic spindles. Consistently, 79% of elongating spermatid bundles contained fewer than the normal 64 nuclei, and flagella were immotile and shorter than in control flies. Electron microscopy revealed that elongating spermatids had missing or incomplete axonemes. Moreover, Klp10A adults were uncoordinated and needed increased time to recover from mechanical shock, a hallmark of centriole defects (Delgehyr, 2012).
These findings led to an examination of centrioles in 16-cell cysts of primary spermatocytes in late G2 or in meiosis. Most such cells (>70%) were found to have more than two centrosomes together with additional rod- and dot-like structures that stained with anti-Dplp (Drosophila pericentrin-like protein). Moreover, whereas wild-type spermatocytes had only a few centrioles (2.2%) with no detectable Sas6 (spindle assembly abnormal protein 6) at their distal part, Klp10A spermatocytes showed an increase of such centrioles, suggesting they were incomplete or fragmented. Spd2 (spindle defective 2), which is closely associated with spermatocyte centrioles otherwise lacking pericentriolar material (PCM), also associated with centriolar fragments in Klp10A mutants, often fraying out from their ends. By the end of G2, Klp10A mutant centrioles displayed discontinuities in the outer microtubules of both basal body-like structures and the short primary cilia that they nucleate. Finally, Klp10A was present along the length of wild-type centrioles, being more concentrated near both ends, but was barely detectable in mutant centrioles (Delgehyr, 2012).
To understand how defective centrioles and centrosomes might arise in Klp10A mutant testes, earlier mitotic stages of spermatogenesis were examined. These showed significantly more cells with either more than (>20% of the cells) or fewer than two Dplp (centriole marker) bodies. Not only the number but also the size of Dplp bodies was compromised: whereas in wild-type spermatogonia, Dplp bodies were of uniform size, Klp10A mutants showed a significant increase in both shorter and longer Dplp bodies. Similar findings were obtained using anti-Spd2 labeling. Examination of spermatogonia by electron microscopy confirmed that Klp10A centrioles were both longer and shorter than wild-type; longer centrioles were generally not uniformly elongated, and shorter ones often appeared tenuously attached to them. This excessive elongation did not, however, seem to prevent formation of procentrioles, indicating that centriole duplication could occur. To better correlate the immunofluorescence observations with electron microscopy, structured illumination microscopy was carried out on anti-Spd2-stained spermatogonial cells. This revealed both shorter and longer centrioles 'frayed' at their ends in Klp10A spermatogonia. Apparent segments of centriolar walls were found both associated with the frayed ends or free in the cytoplasm, suggesting that the elongated structures could become fragmented (Delgehyr, 2012).
To test Klp10A's role in somatic centriole biogenesis, RNA interference (RNAi) was used to efficiently deplete the protein from cultured Dmel cells. Previous studies of the functional requirements for Klp10A for spindle microtubules have not examined centrosome behavior per se. After Klp10A RNAi, cells either lacking or having numerous Dplp bodies were detected. The increase in cells without Dplp bodies was however quite modest, around twice the control. Moreover, the proportion of such cells did not increase following several rounds of Klp10A depletion. This was a quite different outcome from that following depletion of Plk4, a core component of the centrosome duplication pathway, where virtually 100% of cells lose their centrosomes. Accordingly, it was found that centrosomes could reform after Plk4 RNAi treatment in either the presence or absence of Klp10A. Under these conditions, depletion of Sas6, which is known to be essential for centriole duplication, would block de novo centriole formation. Together, these results indicate that the centriole duplication pathway is independent of Klp10A function (Delgehyr, 2012).
It was noticed that Klp10A-depleted cells contained both weakly and brightly fluorescing Dplp bodies. Because it was difficult to measure the length of these bodies (centrioles are less than 0.2 μm in these cells), their fluorescence intensity was measured. Dplp is found at both centrioles and PCM, giving a combined indication of PCM size and centriole length. It was found that Klp10A RNAi resulted in a doubling of both very bright and very weak dots. By contrast, cells depleted for the microtubule depolymerases Klp59C or Klp59D (kinesin-13 family) or Klp67A (kinesin-8) did not show an increase in brightly fluorescent Dplp bodies, suggesting that changes in centrosome size are not a general consequence of cytoplasmic microtubule depolymerization (Delgehyr, 2012).
To understand how these different centrosomal bodies might be generated, video microscopy was carried out to follow centrosomes in cells constitutively expressing GFP-Spd2 from a weak promoter. Because centrosome number is notoriously variable in Dmel cells, cells with two GFP-Spd2 punctae were examined at interphase. In control cells, the two punctae became brighter on mitotic entry as centrosomes recruited PCM and separated to form the spindle poles. Centriole disengagement occurred in telophase, and the daughters had two centrosomal punctae. In Klp10A-depleted cells, the two centrosomes coalesced upon mitotic entry with the apparent collapse of the spindle and bipolarity was recovered; the coalesced centrosomes split into several punctae at one pole, whereas there were none at the other. Centrosomes then dispersed into numerous scattered dots. Thus, in the absence of Klp10A, fragmentation in M phase appears to account for the weak Dplp bodies (Delgehyr, 2012).
It is considered that the weak Dplp bodies could arise by PCM dispersion or centriole fragmentation per se. The former possibility was addressed by staining to reveal Spd2, which makes a greater contribution to PCM, in addition to Dplp. Klp10A depletion led to an increase in small punctae positive for Spd2 but not Dplp, consistent with some PCM fragmentation. In the absence of good antibodies to label centrioles, the possibility of centriole fragmentation was addressed by applying structured illumination immunofluorescence microscopy and electron microscopy to examine centrioles in control and Klp10A-depleted cells. Structured illumination microscopy resolved the punctae staining in control cells into circular structures of ∼0.4 μm outer diameter that, when rotated through 90°, revealed the rod-like structures of centrioles. In Klp10A-depleted cells, these rods could be the same length or longer than in control cells. What appeared to be segments of the centriolar wall was also observed, suggesting that centriole fragmentation could also contribute to the weak Dplp bodies seen by conventional immunofluorescence (Delgehyr, 2012).
Electron microscopy revealed an increase in centrioles longer than 0.18 μm from 15.4% in control cells to 72.1% in Klp10A-depleted cells and a less pronounced increase in centrioles shorter than 0.15 μm (from 12.8% to 17.6%). Moreover, 10.5% of centrioles in Klp10A-depleted cells had an incomplete complement of microtubules. It is possible that the small centrioles represent intermediates in centriole duplication. If so, the relatively small increase in their number suggests that centriole duplication cannot account for the almost doubling of weak Dplp bodies after Klp10A RNAi. Together, these results therefore suggest that Klp10A depletion leads to centriole elongation and fragmentation associated with dispersion of the PCM (Delgehyr, 2012).
Centrosome separation during mitosis is mainly asymmetric in Klp10A-depleted cells: one cell inherits both centrosomes, and the other inherits none. Because the latter cell might be expected to engage in de novo centriole formation, it was wondered whether formation of longer centrioles following Klp10A depletion could be a consequence of this. Therefore centriole reformation was examined after inducing their loss by extensive Plk4 depletion, and it was found that such centrioles were no longer than in control cells. This accords with the normal morphology of centrioles formed de novo in unfertilized eggs overexpressing Plk4. Thus, the long fragmented centrioles seen after Klp10A depletion are unlikely to be a consequence of de novo centriole formation (Delgehyr, 2012).
It was previously found that injection of anti-polyglutamylated tubulin antibody into HeLa cells in G2 resulted in centriole loss and scattering of PCM in a dynein-dependent manner. This has been widely taken to indicate that polyglutamylation might stabilize centrosomes to forces exerted upon them in mitosis. Unlike in mammalian cells, tubulin is not polyglutamylated in most Drosophila tissues, and accordingly this modification could not be detected in Drosophila centrosome preparations or on centrioles of primary spermatocytes or Dmel cells. It was therefore wondered whether centrosomes might scatter following Klp10A downregulation in a dynein-dependent process. To test this, it was decided to partially release tension on centrosomes by depleting dynein (Dhc64, dynein heavy chain 64) in the presence or absence of Klp10A. Because dynein depletion leads to inequitable heritance of centrosomes, their detachment from spindle poles, and cell death, around 50% of the protein was depleted. This did not lead to defects in centrosome number or staining intensity. When Klp10A was extensively depleted and dynein was partially depleted, numbers of weak Dplp bodies decreased at the expense of an increase in bright Dplp bodies as compared to cells depleted for Klp10A alone. Thus, partial depletion of dynein apparently rescues centrosome fragmentation resulting from Klp10A depletion. Electron microscopy revealed that codepletion of Klp10A and dynein resulted in long centrioles, some exceeding the size observed after Klp10A depletion alone. Although the possibility cannot be completely excluded that dynein depletion directly affects the centriole per se, it is more likely that dynein depeletion results in reduced forces upon the centriole, thus protecting the long centrioles of Klp10A-depleted cells from fragmentation (Delgehyr, 2012).
To determine whether Klp10A's microtubule-depolymerizing properties are needed for its centrosomal functions, mutations were generated in two conserved class-specific motifs of its motor domain shown to be essential for microtubule depolymerization in other organisms. These mutations changed residues KVD (aa 317–319) or KEC (aa 546–548) of Klp10A each to three alanines. In accord with the known properties of Klp10A, it was found that high levels of wild-type protein depolymerized interphase microtubules, whereas cells expressing similar levels of the mutant forms had interphase microtubules of normal length. Thus, both mutations abolish Klp10A's microtubule depolymerization activity. Lines expressing either wild-type or mutant forms of GFP-tagged Klp10A were also generated from the endogenous Klp10A promoter that had the 5′ but not the 3′ untranslated region (UTR) of Klp10A. Both wild-type and mutant forms were expressed at levels comparable to the endogenous protein and were associated with microtubules and centrosomes or spindle poles in interphase and mitosis. Expression of the tagged wild-type protein at these low levels did not result in depolymerization of cytoplasmic microtubules. Then endogenous Klp10A was depleted using RNAi directed at the 3′ UTR and the effects upon centrosomes were assessed. This led to an increase in cells either without centrosomes or with both weak and bright Dplp bodies in RNAi-treated cells expressing GFP (control) or GFP-tagged mutant forms of Klp10A, but not in cells expressing GFP-tagged wild-type protein. Thus, generation of both larger centrosomes and centrosome fragments requires loss of the microtubule-depolymerizing activity of Klp10A (Delgehyr, 2012).
Because human CP110 or Centrobin (required for CP110's centriolar localization) results in centriole elongation, it was intriguing to find Drosophila CP110 enriched at the distal ends of the centrioles of primary spermatocytes in a region similar to Klp10A. Moreover, CP110 colocalized with Klp10A in interphase Dmel cells, and other evidence suggested that CP110 and Klp10A can physically interact. It was found that several rounds of CP110 depletion led to an increase in Dmel cells lacking centrosomes, CP110 depletion affects centrosome biogenesis. However, centrosomes could reform after their elimination by Plk4 RNAi treatment, even in the absence of CP110, although this was prevented by further depletion of Sas6, required for bona fide centriolar duplicatio. Thus, there is no absolute requirement for CP110 for centriole duplication. Interestingly, however, there was an increase in weakly fluorescing Dplp bodies in CP110-depleted cells, and electron microscopy revealed that 50% of centrioles were shorter than 0.15 μm, compared to 12.8% in control cells, with most being about 0.11 μm long. This is unlikely to represent an increase in procentrioles, because total centriole number decreased. Centrioles of about 0.11 μm are similar in length to the unit cartwheel made mainly of three or four tiers in Chlamydomonas reinhardtii and may represent the smallest stable centriole structures. Thus, it is concluded that, in contrast to depletion of CP110 in mammalian cells , depletion of CP110 in Drosophila leads to centriole shortening and destabilization (Delgehyr, 2012).
It was then asked whether centriole shortening following CP110 depletion might be rescued by sequentially codepleting Klp10A. It was found that this resulted in the reappearance of long centrioles: 62.5% of the centrioles exceeded 0.18 μm, compared to 15.4% in control cells. When Klp10A was depleted first, followed by codepletion with CP110, the reappearance of longer centrioles was observed. Thus, in the absence of Klp10A, centrioles increase in length regardless of whether CP110 is present or absent. The destabilization of Dplp bodies after CP110 depletion appeared to be enhanced by overexpression of GFP-Klp10A that led to an increased proportion of cells lacking centrosomes after 3 days of CP110 RNAi. Together, these results suggest that CP110 might provide a barrier to prevent Klp10A-mediated depolymerization of centriolar microtubules, and indeed, it forms a plug-like structure at the distal part of the centriole. However, CP110 has no effect on microtubule elongation in the absence of Klp10A. Thus, Klp10A can restrict centriole length regardless of whether or not CP110 is present, and so their physical interaction is not required for Klp10A's recruitment and/or microtubule-depolymerizing activity. Indeed Klp10A may be recruited directly on the centrioles, because it is known to have affinity for the microtubule lattice in vitro (Delgehyr, 2012).
Klp10A is the first kinesin-13 demonstrated to regulate centriole length. Other family members have, however, been shown to regulate the length of flagella and more recently formation of cilia. In contrast to Klp10A, however, the mammalian kinesin-13 Kif24 is further limited to acting from the mother centriole on microtubules of cilia. Thus, Kif24 depletion leads to aberrant cilia formation in cycling cells but does not promote growth of centrioles in nonciliated cells. Although Drosophila Klp10A and mammalian Kif24 act differently, both are able to interact with CP110. However, the interaction of CP110 and Kif24 in mammalian cells affects only the process of cilium formation. These functional differences between kinesin-13:CP110 complexes could reflect a regulative adaptation by mammals associated with evolution of the ability to generate primary cilia in many cell types. Thus, whereas in Drosophila, where few cells are ciliated, the capping function of CP110 might help to fix the length of the centriole by blocking Klp10A-mediated microtubule depolymerization, in mammals it has the additional function of blocking cilium formation until the correct phase of the cell cycle. Whether in mammals another kinesin-13 might play the role of Klp10A in regulating centriole length remains an open question (Delgehyr, 2012).
Basal bodies are freed from cilia and transition into centrioles to organize centrosomes in dividing cells. A mutually exclusive centriole/basal body existence during cell-cycle progression has become a widely accepted principle. Contrary to this view, this study shows that cilia assemble and persist through two meiotic divisions in Drosophila spermatocytes. Remarkably, all four centrioles assemble primary cilia-centriole complexes that transit from the plasma membrane encased in a packet of membrane, recruit centrosomal material into microtubule-organizing centers, and persist at the spindle poles through division. Thus, spermatocyte centrioles organize centrosomes and cilia simultaneously at cell division. These findings challenge the prevailing view that cilia antagonize cell-cycle progression and raise the possibility that cilium retention at cell division may occur in diverse organisms and cell types (Riparbelli, 2012; see graphical abstract).
The elegant 9-fold radially symmetric design of the centriole is truly one of nature's aesthetic wonders. Centrioles are found at the core of microtubule-organizing centers in most animal cells. Despite their requirement for the spatial organization of the centrosomal material to nucleate cytoplasmic and spindle microtubules, centrioles are dispensable for somatic cell division where acentrosomal pathways of microtubule organization can suffice. Centrioles have an additional and perhaps more critical function: as templates for cilia and flagella, a context in which they are called basal bodies. The ability of centrioles to function both as centrosomal organizers at the spindle poles and also as basal bodies to template primary cilia was originally proposed in the Henneguy-Lenhossek hypothesis in 1898, which predicted the functional equivalence of these organelles (Riparbelli, 2012).
The cilium is an essential antenna-like projection that consists of the microtubule-based axoneme and a surrounding membrane that is continuous but distinct from the cell's plasma membrane. The axoneme is the structural backbone of the cilium and mirrors the 9-fold radial symmetry conveyed by the basal body. Generally, motile cilia contain a pair of axial singlet microtubules at the center of the axoneme surrounded by the nine doublet microtubules of the axoneme: a '9+2' structure. Nonmotile primary cilia lack the central pair ('9+0'). Primary cilia are sensory organelles for intercellular signaling involving hedgehog, Wnt, and PDGF pathways. Defects in cilia have pleiotropic effects on development and physiology, depending on the molecules affected and the alleles involved, and lesions in affected human genes are collectively called ciliopathies (Riparbelli, 2012).
Centrioles have a double life: as centrioles within the centrosomes, and as basal bodies at the cell membrane in differentiated or resting cells. In cycling cells, centrioles are found within the centrosomal material during cell division but then move to the cell periphery in quiescent cells to nucleate a ciliary axoneme after docking at the plasma membrane. Upon cell-cycle entry the centriole is released from the axoneme and/or the cilium resorbs, and then the centriole migrates into the cytoplasm and recruits pericentriolar material to organize the centrosome. Although cell-cycle-dependent cilium assembly and disassembly are typical of vertebrate cells, it remains unclear how this process is executed and regulated. The switch between centriole and basal body is accompanied by morphological and molecular changes. Moreover, there is considerable evidence that the presence of a cilium is incompatible with cell division and that cells will not divide until their centrioles are freed from cilia. It is generally assumed that the centriole cannot function simultaneously to assemble cilia and mitotic centrosomes. The vast majority of studies of cilium dynamics in the cell cycle were done in cell culture (Riparbelli, 2012).
Contrasting evidence suggested that cilia and flagella might be maintained at cell division. It was reported that 'flagellar' structures were maintained during meiotic divisions of some Lepidopteran, Neuropteran, and Dipteran species. Cilia and flagella may also be retained during the division of some protozoa. The seminal electron microscopy (EM) work by Tates showed the detailed changes that centrioles undergo during spermatogenesis and in retrospect revealed the presence of short cilia in spermatocytes during division (Riparbelli, 2012).
This study investigated the timing and ultrastructure of cilium biogenesis during Drosophila spermatogenesis using molecular markers for basal bodies and cilia, coupled with unequivocal ultrastructural imaging using EM. It was shown that cilia assemble in premeiotic prophase and remain intact through two rounds of cell division in vivo in spermatocytes, a cell type with a discrete developmental program involving two rounds of division followed by differentiation (Riparbelli, 2012).
These findings challenge the principle that cilia are inconsistent with progression through the cell cycle and provide a system to study this regulation. Remarkably, rather than dissociating basal bodies from the cilia, spermatocyte basal body-cilia complexes are internalized with an associated 'pocket' of plasma membrane as their centrioles assemble PCM into centrosomes that participate in spindle assembly. The detailed dynamic changes in the centriole-cilium structures from spermatocytes to spermatid will serve as vital features for future analysis of cilium regulation during cell division (Riparbelli, 2012).
The results clearly define two opposite scenarios in cilia biogenesis: (1) that of most known animal cells in which the primary cilium dynamics appear governed by cell cycle and are exclusive of dividing cells, and (2) that of Drosophila spermatocytes in which the primary cilium is stable through meiotic divisions. These observations shift the general point of view that the centriole must be freed from the cilium to organize a functional centrosome for spindle organization and that dividing cells are unable to form primary cilia. Nevertheless, the ability of spermatocytes to assemble cilia and then maintain this structure at the spindle poles during division could be a specialized feature of spermatocytes and perhaps other cell types. On the other hand, this could be a feature of progenitor cells that are programmed to divide a discrete set of rounds before differentiating into a ciliated cell type. There are few such studies performed in vivo to support or counter this possibility at present (Riparbelli, 2012).
Primary cilia in vertebrate cells are antenna-like projections that detect mechanical and chemical cues from the environment to signal differentiation, proliferation, and other responses. Drosophila spermatocytes develop within a cyst that is separated by the environment by a thin cellular envelope. Cytoplasmic bridges that ensure communication and synchrony interconnect the 16 spermatocytes within the cyst. In this condition the function of primary cilia as sensory organs may be redundant or their role unclear. Although centrosomes are essential for accurate division of spermatocytes, the importance or requirement of the cilia for division is unclear. Such a requirement is difficult to test at present because there are no known mutants that completely disrupt spermatocyte cilia that also leave the centrioles intact (Riparbelli, 2012).
A surprising property of spermatocyte cilia is their independence from intraflagellar transport (IFT) for their assembly. This could therefore be a special class of cilia, similar to the sperm flagellum with its independence on IFT for assembly. However, whereas the spermatid axoneme assembles in the cytoplasm, it is actually encased in membrane that invaginates from the plasma membrane to form a 5- to 10-nanometer-long cilium that retains this length at the distal tip of the axoneme as it grows. A requirement of IFT for assembly of this distal cilium has not been examined. On the other hand, the spermatocyte cilium could be an elaborated transition zone, which can assemble independent of IFT (Riparbelli, 2012).
Questions on the function of spermatocyte cilia are also raised by some of their peculiarities. Spermatocyte cilia are nucleated by all four centrioles, whereas in vertebrate cells only the mother centriole is able to do this. Vertebrate cells assemble a single primary cilium during G1 phase, but Drosophila spermatocytes assemble cilia during G2 phase. Moreover, the ciliary axonemes are nucleated by centrioles that continue to grow while the cilium grows, whereas in vertebrate cells the mother centriole has reached its definitive length at maturity before it organizes the primary cilium. How can the centriole and the cilium grow simultaneously? It can be supposed that the cilium elongates by addition of subunits at its distal tip. But how can the centriole elongate if its distal extremity is committed to form the ciliary axoneme? Although a necessary function is not demonstrated, spermatocyte cilia could be precursors for building the flagellar axoneme for the sperm, despite axonemal reorganization in early spermatids (Riparbelli, 2012).
Collectively, these results show that cilia are not incompatible with cell division as a general rule. Although cilium resorption initiates at cell-cycle entry in most cell culture models, it is not clear that disassembly of cilia is necessary for cell division to occur. In spermatocytes, on the other hand, it is possible that a potential cilium block to cell-cycle progression is repressed. Alternatively, the assembly of cilia in these cells in G2 phase might be a unique mechanism to circumvent a G1-specific cell-cycle block imposed by cilia in quiescent cells. Yet another possibility is that spermatocyte cilia have distinct properties from other primary cilia that are permissive for cell-cycle progression. Consistent with this possibility is the finding that these cilia appear to assemble independent of IFT. These intriguing clues provide a framework to further probe the interdependence of cilia with cell-cycle control (Riparbelli, 2012).
Cilia and flagella are involved in a variety of processes and human diseases, including ciliopathies and sterility. Their motility is often controlled by a central microtubule (MT) pair localized within the ciliary MT-based skeleton, the axoneme. This study characterized the formation of the motility apparatus in detail in Drosophila spermatogenesis. Assembly of the central MT pair starts prior to the meiotic divisions, with nucleation of a singlet MT within the basal body of a small cilium, and the second MT of the pair only assembles much later, upon flagella formation. BLD10/CEP135, a conserved player in centriole and flagella biogenesis, can bind and stabilize MTs and is required for the early steps of central MT pair formation. This work describes a genetically tractable system to study motile cilia formation and provides an explanation for BLD10/CEP135's role in assembling highly stable MT-based structures, such as motile axonemes and centrioles (Carvalho-Santos, 2012).
Cilia are microtubule (MT)-based organelles involved in a variety of processes, such as cell motility, fluid flow, and sensing mechanical stimuli and signaling molecules. At the base of each cilium there is at least one modified centriole, the basal body (BB), which templates the growth of the axoneme, the MT-based structure of cilia. Centrioles are also essential for the formation of centrosomes, the primary MT organizer of the cell. Cilia can exist as motile or immotile structures. Most motile cilia have a pair of central MTs within the lumen of their axoneme to coordinate motility. Defects in ciliary motility are associated with a variety of human disorders including infertility, respiratory problems, hydrocephalus and situs inversus, commonly found in patients with primary ciliary dyskinesia. Mice mutant for Hydin, a component of the central MT pair apparatus, show motility defects and develop hydrocephalus (Carvalho-Santos, 2012).
Much is known about axonemal components, mostly from work developed in the green algae. In fully assembled axonemes, the central MT pair starts at the most proximal part of the axoneme, a region called the transition zone. MT regulators are likely to control the assembly of the central MT pair: γ-tubulin and heterotrimeric kinesin-2 are required for this process in protozoa and sea urchin, respectively. However, little is known about the molecular mechanisms that govern the switch of centrioles to BBs and, in particular, when and how central MT pair assembly starts and is coordinated with other cellular processes (Carvalho-Santos, 2012).
Drosophila mutants for Bld10, a conserved player in centriole biogenesis, have short centrioles and the majority of their sperm flagella lack the central MT pair. In Chlamydomonas and Paramecium, BLD10/CEP135 localizes close to the centriolar MTs at the spoke tips of the cartwheel, a nine-fold symmetric structure present at the base of centrioles/BBs that enforces their symmetry. Depletion of BLD10/CEP135 in those organisms severely impairs BB assembly. In human cultured cells, BLD10/CEP135 is required for centriole assembly and localizes to the cartwheel, the centriolar walls, and the lumen of the centriole distal region. Altogether, these data strongly indicate that BLD10/CEP135 has a MT-related function that underlies its role in centriole/BB assembly. Moreover, its localization in the centrioles but not in the axoneme in human and Drosophila suggests a role in the initiation of central MT pair assembly (Carvalho-Santos, 2012).
Drosophila spermatogenesis was used to understand how the central MT pair is assembled and the role played by BLD10/CEP135 in that process. One of the MTs of the pair is assembled prior to the meiotic divisions, within the basal body of a small cilium. This MT is maintained through meiosis, after which the flagellum and the second MT of the pair within it are formed. BLD10/CEP135 directly binds and stabilizes MTs and that in Bld10 mutants the singlet MT does not form and consequently the central MT pair is not assembled (Carvalho-Santos, 2012).
The central MT pair complex is an essential and highly specialized structure present in motile cilia that is required for coordinated motility. The morphological and molecular changes that occur during central MT pair assembly are yet to be characterized. Building on influential morphological work that described spermatogenesis in the fruit fly, this study has established Drosophila spermatogenesis as a genetically tractable system to study central MT pair formation. The assembly of this structure initiates prior to the meiotic divisions, much earlier than previously thought. During that stage, a singlet MT forms within the BB of a small cilium in G2 spermatocytes. This MT is likely very stable as it is present throughout meiotic divisions and until axoneme extension. In early spermatids, the stage at which flagellum assembly takes place, a second MT appears close to the singlet, and completes central MT pair formation. Bld10 emerged as an ideal candidate to regulate central MT pair formation as mutants lack this structure and the protein localizes to both the lumen and distal regions of the centriole/BB. Accordingly, this study demonstrates that spermatogenesis in Bld10 mutants is carried out without assembly of the singlet MT, thus impacting on central MT pair biogenesis. Bld10 is a MAP whose overexpression leads to cytoplasmic MT stabilization in culture cells. Finally, this study directly links MT stabilizing activity of Bld10 to central MT pair assembly as (1) its N terminus, which binds and stabilizes MTs, contributes to this process, and (2) exposure to colchicine during central MT pair formation accentuates Bld10 mutant phenotype (Carvalho-Santos, 2012).
Morphological findings of central MT pair assembly during Drosophila spermatogenesis are summarized in Schematic Representation of Cilium and Flagellum Structures in Drosophila melanogaster Spermatogenesis (Figure 7). The identification and analysis of different stages of this process using TEM single and serial sections, and tomography, suggest this assembly is much more dynamic than was anticipated. It is proposed that central MT pair biogenesis starts with the formation of a singlet MT within the lumen of the BB. The detailed analysis of this process raises several important questions, including where the first singlet MT is nucleated from, and whether it actively participates in the assembly of the second MT of the pair. The identification of singlet MT or central MT pair markers that do not localize to other BB and axoneme structures will allow further mechanistic studies of central MT pair assembly in a less laborious fashion than as it is with electron microscopy (Carvalho-Santos, 2012).
The observation that a singlet MT forms within the BB as a precursor for the biogenesis of the central MT pair of the motile axoneme, implies a broader role for the BB in templating cilia than currently thought. This raises an important question on the significance and conservation of this process. Until now, the study of this process using stills of TEM sections has prevented the discovery and characterization of intermediate steps. Drosophila spermatogenesis proved to be a valuable system to study these fast intermediate stages since central pair assembly takes a few days to be completed. It is proposed that the presence of the singlet MT is an intermediate state until the second MT of the pair is nucleated. In mature sperm both MTs have equal length and therefore is not obvious that one MT was assembled first. In model organisms often used to study flagella assembly, such intermediate stages might have not been observed due to their transitory nature. In Nephrotoma suturalis, an insect that is phylogenetically close to Drosophila melanogaster, a singlet MT has also been observed. The presence of stages where a singlet MT is easily found can be a consequence of a slower central MT pair assembly in these insects (Carvalho-Santos, 2012).
It is intriguing that the central MT pair starts to assemble so early within the BB. An extensive literature search was conducted for ultrastructural studies on sperm flagella assembly in other species to address whether this was a conserved phenomenon. Surprisingly, in both vertebrate and invertebrate species, central MT pairs have been observed in cilia/flagella that assemble during spermatocyte stages, before meiosis I or II . Why the central MT pair assembles at that stage, and in particular whether spermatocyte cilia motility is needed for spermatogenesis, is an important question that deserves further study. It is still unclear whether in these different species the central MT pairs found in spermatocyte cilia/flagella are precursors of the central MT pair of sperm flagella, as is describe in this study for Drosophila (Carvalho-Santos, 2012).
Although much is known about the components of the central MT pair machinery, little is understood about the molecular regulation of its nucleation. In light of the current findings, BLD10/CEP135 is proposed to regulate the initiation of central MT pair biogenesis through MT stabilization. Bld10 conserved N-terminal domain is proposed to play an important role in this process; however, given that removal of that domain does not completely abolish Bld10 activity, it cannot be excluded that other Bld10 protein regions also stabilize the central MT pair. It is common for MAPs to have different MT binding domains. Since the M- and C-Bld10 truncations localize independently to centrioles, it is possible they also stabilize MTs associated with the centriole and the central MT pair, a hypothesis that should be investigated in the future (Carvalho-Santos, 2012).
How general is BLD10/CEP135 MT-stabilizing function? It is possible that BLD10/CEP135 ancestral function is stabilizing special sets of MTs including the centriole triplets and the singlet MT during axoneme central MT pair assembly. BLD10/CEP135 is present in the genome of most organisms that assemble centrioles and flagella but absent from those that lack these organelles, such as higher plants. While a role for BLD10/CEP135 in central MT pair assembly has not been investigated in vertebrates, TSGA10, a BLD10/CEP135 paralog only present in vertebrates, localizes to the sperm flagella and has been linked to male sterility, further corroborating a role for this family of proteins in flagella assembly. BLD10/CEP135 loss-of-function in several species generates phenotypes associated with centriolar MT defects, including MT triplet loss, and shorter centrioles. The link between Bld10 and stabilization of these specific MT sets is also supported by its localization to the cartwheel spokes and BB/centriolar MT triplets (in Chlamydomonas, human cells and Drosophila). The complete lack of BBs in Chlamydomonas BLD10 mutants might reflect defects both in cartwheel assembly and in the recruitment and/or stabilization of the BB MTs onto the cartwheel. This is not the case in Drosophila, as centrioles, albeit shorter, are observed in Bld10 mutants. It is possible that, in the fruit fly, other molecules are redundant with Bld10 in its role of stabilizing cartwheel-associated MTs. In the future, it will be important to understand how Bld10 might stabilize MTs, how its function is regulated in different centriole compartments (cartwheel, lumen, walls) and at different time points, such as during centriole or central MT pair assembly (Carvalho-Santos, 2012).
Centrioles and axonemes are very special organelles, being much more stable than any other MT-based structures. They not only withstand complex MT remodeling environments but in the case of centrioles they also perdure for several cell generations. Additionally, the assembly and stabilization of centrioles and axonemes involves particular MT regulators and posttranslational modifications. This study has shown that BLD10/CEP135 is a special MT regulator specifically involved in the formation of highly stable and specialized MTs. The discovery of these specialized MAPs, where SAS4/CPAP could also be included, opens the door to an exciting new biology of MT regulation. Moreover, given the importance of centrioles and cilia in development and homeostasis, this work will allow further contextualization of these cellular structures in human disease (Carvalho-Santos, 2012).
Anderson, C. T. and Stearns, T. (2009) Centriole age underlies asynchronous primary cilium growth in mammalian cells. Curr. Biol. 19: 1498-1502. PubMed Citation: 19682908
Boyle, M., Wong, C., Rocha, M., and Jones, D. L. (2007). Decline in self-renewal factors contributes to aging of the stem cell niche in the Drosophila testis. Cell Stem Cell 1(4): 470-8. PubMed Citation: 18371382
Carvalho-Santos, Z., Machado, P., Alvarez-Martins, I., Gouveia, S. M., Jana, S. C., Duarte, P., Amado, T., Branco, P., Freitas, M. C., Silva, S. T., Antony, C., Bandeiras, T. M. and Bettencourt-Dias, M. (2012). BLD10/CEP135 is a microtubule-associated protein that controls the formation of the flagellum central microtubule pair. Dev Cell 23: 412-424. Pubmed: 22898782
Cheng, J., et al. (2008). Centrosome misorientation reduces stem cell division during ageing. Nature 456(7222): 599-604. PubMed Citation: 18923395
Delgehyr, N., et al. (2012). Klp10A, a microtubule-depolymerizing kinesin-13, cooperates with CP110 to control Drosophila centriole length. Curr Biol. 22(6): 502-9. PubMed Citation: 22365849
Rebollo, E., Sampaio, P., Januschke, J., Llamazares, S., Varmark, H. and Gonzalez, C. (2007). Functionally unequal centrosomes drive spindle orientation in asymmetrically dividing Drosophila neural stem cells. Dev. Cell 12(3): 467-74. PubMed Citation: 17336911
Riparbelli, M. G., Callaini, G. and Megraw, T. L. (2012). Assembly and persistence of primary cilia in dividing Drosophila spermatocytes. Dev. Cell 23(2): 425-32. PubMed Citation: 22898783
Rusan, N. M. and Peifer, M. (2007). A role for a novel centrosome cycle in asymmetric cell division. J. Cell Biol. 177(1): 13-20. PubMed Citation: 17403931
Stevens, N. R., Dobbelaere, J., Brunk, K., Franz, A. and Raff, J. W. (2010). Drosophila Ana2 is a conserved centriole duplication factor. J Cell Biol 188: 313-323. Pubmed: 20123993
Wang, X., Tsai, J. W., Imai, J. H., Lian, W. N., Vallee, R. B. and Shi, S. H. (2009). Asymmetric centrosome inheritance maintains neural progenitors in the neocortex. Nature 461(7266): 947-55. PubMed Citation: 19829375
Yamashita, Y. M., Mahowald, A. P., Perlin, J. R. and Fuller, M. T. (2007). Asymmetric inheritance of mother versus daughter centrosome in stem cell division. Science 315(5811): 518-21. PubMed Citation: 17255513
Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.
The Interactive Fly resides on the
Society for Developmental Biology's Web server.