Centrosomes play an important part in many aspects of cell organisation, and they form when a mother centriole recruits pericentriolar material (PCM) around itself (Conduit et al., 2015). The PCM contains several hundred proteins (Alves-Cruzeiro et al., 2014), allowing the centrosome to function as a major microtubule (MT) organising centre, and also as an important coordination centre and signalling hub (Chavali et al., 2014; Vertii et al., 2016). Centrosome dysfunction has been linked to several human diseases and developmental disorders, including cancer, microcephaly and dwarfism (Bettencourt-Dias et al., 2011; Nigg and Raff, 2009).

During interphase, the mother centriole recruits a small amount of PCM that is highly organised (Lawo et al., 2012; Mennella et al., 2012; Sonnen et al., 2012; Fu and Glover, 2012). As cells prepare to enter mitosis, however, the PCM expands dramatically around the mother centriole in a process termed centrosome maturation (Palazzo et al., 2000). Electron microscopy (EM) studies suggest that centrioles organise an extensive ‘scaffold’ structure during mitosis that surrounds the mother centriole and recruits other PCM components such as the γ-tubulin ring complex (γ-TuRC) (Moritz et al., 1995; Moritz et al., 1998; Schnackenberg et al., 1998).

In the fruit fly Drosophila melanogaster and the nematode Caenorhabditis elegans, a relatively simple pathway seems to govern the assembly of this mitotic PCM scaffold. The conserved centriole/centrosome protein Spd-2/SPD-2 (fly/worm nomenclature) cooperates with a large, predominantly predicted-coiled-coil, protein (Cnn in flies, SPD-5 in worms) to form a scaffold whose assembly is stimulated by the phosphorylation of Cnn/SPD-5 by the mitotic protein kinase Polo/PLK-1 (Conduit et al., 2014a; Conduit et al., 2014b; Feng et al., 2017; Woodruff et al., 2017; Woodruff et al., 2015; Wueseke et al., 2016). Mitotic centrosome maturation is abolished in the absence of this pathway, and some aspects of Cnn and SPD-5 scaffold assembly have recently been reconstituted in vitro (Feng et al., 2017; Woodruff et al., 2015; Woodruff et al., 2017). Vertebrate homologues of Spd-2 (Cep192) (Gomez-Ferreria et al., 2007; Zhu et al., 2008), Cnn (Cdk5Rap2/Cep215) (Barr et al., 2010; Choi et al., 2010; Fong et al., 2008; Kim and Rhee, 2014; Lizarraga et al., 2010) and Polo (Plk1) (Haren et al., 2009; Lane and Nigg, 1996; Lee and Rhee, 2011) also have important roles in mitotic centrosome assembly, indicating that elements of this pathway are likely to be conserved in higher metazoans. In vertebrate cells another centriole and PCM protein, Pericentrin, also has an important role in mitotic centrosome assembly that is dependent upon its phosphorylation by Plk1 (Haren et al., 2009; Lee and Rhee, 2011). Pericentrin can interact with Cep215/Cnn (Buchman et al., 2010; Kim and Rhee, 2014; Lerit et al., 2015), but in flies the Pericentrin-like-protein (Plp) has a clear, but relatively minor, role in mitotic PCM assembly when compared to Spd-2 and Cnn (Lerit et al., 2015; Martinez-Campos et al., 2004; Richens et al., 2015).

Although most of the main players in mitotic centrosome-scaffold assembly appear to have been identified, several fundamental aspects of the assembly process remain mysterious. Cells entering mitosis, for example, contain two mother centrioles that assemble two mitotic centrosomes of equal size. It is unclear how this is achieved, as even a slight difference in the initial size of the two growing centrosomes would be expected to lead to asymmetric centrosome growth—as the larger centrosome would more efficiently compete for scaffolding subunits (Conduit et al., 2015; Zwicker et al., 2014). The centrioles in fly embryos appear to overcome this problem by constructing the PCM scaffold from the ‘inside-out’: Spd-2 and Cnn are only incorporated into the scaffold close to the mother centriole, and they then flux outwards to form an expanded scaffold around the mother centriole (Conduit et al., 2014b; Conduit et al., 2010). In this way, the growing PCM scaffold could ultimately attain a consistent steady-state size—where incorporation around the mother centriole is balanced by loss of the scaffold at the centrosome periphery—irrespective of any initial size difference in the PCM prior to mitosis (Conduit et al., 2015; Raff, 2019).

A potential problem with this ‘inside-out’ mode of assembly is that the rate of centrosome growth is limited by the very small size of the centriole. Mathematical modelling indicates that the incorporation of a crucial PCM scaffolding component only around the mother centriole cannot easily account for the high rates of mitotic centrosome growth observed experimentally (Zwicker et al., 2014). To overcome this problem, it has been proposed that centrosome growth is ‘autocatalytic’, with the centriole initially recruiting a key scaffolding component that can subsequently promote its own recruitment (Woodruff et al., 2014; Zwicker et al., 2014). It has been proposed that Spd-2 and Cnn could form a positive feedback loop that might serve such an autocatalytic function: Spd-2 helps recruit Cnn into the scaffold, and Cnn then helps to maintain Spd-2 within the scaffold, thus allowing higher levels of Spd-2 to accumulate around the mother centriole, which in turn drives higher rates of Cnn incorporation (Conduit et al., 2014b; Conduit et al., 2015; Raff, 2019).

In worms (Decker et al., 2011) and vertebrates (Joukov et al., 2010; Joukov et al., 2014; Meng et al., 2015), SPD-2/Cep192 can help recruit PLK1/Plk1 to centrosomes and Cep192 also activates Plk1 in vertebrates, in part through recruiting and activating Aurora A, another mitotic protein kinase implicated in centrosome maturation. We suspected, therefore, that in flies Spd-2 might recruit Polo into the centrosome-scaffold to phosphorylate Cnn and so help to generate a positive feedback loop that drives the expansion of the mitotic PCM. In flies, however, no interaction between Polo and Spd-2 has been reported. Indeed, an extensive Y2H screen for interactions between key centriole and centrosome proteins identified interactions between Spd-2 and the mitotic kinases Aurora A and Nek2, and between Polo and the centriole proteins Sas-4, Ana1 and Ana2, but not between Polo and Spd-2 (Galletta et al., 2016). A possible explanation for this result is that Polo/Plk1 is believed to be largely recruited to its many different locations in the cell, including centrosomes, through its Polo-Box-Domain (PBD), which binds to phosphorylated S-S/T(p) motifs (Elia et al., 2003a; Elia et al., 2003b; Lee et al., 1998; Liu et al., 2004; Reynolds and Ohkura, 2003; Seong et al., 2002; Song et al., 2000). Perhaps any such Polo binding sites in fly Spd-2 were simply not phosphorylated in the Y2H experiments. In support of this possibility, phosphorylated S-S/T(p) motifs in SPD-2/Cep192 have previously been shown to help recruit PLK1/Plk1 to centrosomes in worms (Decker et al., 2011), frogs (Joukov et al., 2010) and humans (Meng et al., 2015).

Here, we examine the potential role of Spd-2 in recruiting Polo to centrosomes in Drosophila embryos. We find that, like Cnn, Spd-2 is largely unphosphorylated in the cytosol, but is highly phosphorylated at centrosomes, where Spd-2 and Polo extensively co-localise within the pericentriolar scaffold. We show that a Spd-2 fragment containing 19 S-S/T motifs exhibits enhanced binding to the PBD in vitro when it has been phosphorylated by Plk1, but no enhancement is seen if these S-S/T motifs are mutated to T-S/T—a mutation that strongly perturbs PBD binding (Elia et al., 2003b). We express forms of Spd-2 in vivo in which either all 34 S-S/T motifs, or the 16 most conserved S-S/T motifs, have been mutated to T-S/T to perturb PBD-binding. These mutant Spd-2 proteins are still recruited to mother centrioles, as are Polo and Cnn, and these proteins assemble a PCM scaffold around the mother centriole. Strikingly, however, this PCM scaffold can no longer expand outwards, and centrosome maturation fails. These observations provide strong support for the hypothesis that Spd-2, Polo and Cnn cooperate to form a positive feedback loop that is required to drive the rapid expansion of the mitotic PCM in fly embryos.

Centrosome maturation appears to be a near-universal feature of the metazoan cell cycle. Although many of the key proteins required for centrosome maturation have been identified, how these proteins drive this process is unclear. We have previously proposed that three proteins—Spd-2, Polo and Cnn—together form a scaffold that expands around the mother centriole to recruit other PCM components to the mitotic centrosome (Conduit et al., 2014b; Conduit et al., 2015). The data presented here suggests that these three proteins cooperate to form a positive feedback loop that drives the dramatic expansion of the mitotic PCM scaffold in fly embryos (Figure 11A).

Figure 11

Download asset Open asset

We propose the following model (Figure 11B). In interphase cells, Spd-2, Polo and Cnn are recruited around the surface of the mother centriole (Fu and Glover, 2012), but Polo is inactive and Spd-2 and Cnn are not phosphorylated—so no scaffold is assembled (Figure 11Bi). As cells prepare to enter mitosis (Figure 11Bii), centrosomal Spd-2 becomes phosphorylated. Our in vitro data suggests that Polo is involved in this phosphorylation (via a ‘self-priming and binding’ mechanism), but other mitotic kinases may also be involved. Phosphorylation allows Spd-2 to form a scaffold (red, Figure 11Bii) that fluxes outwards (red arrows, Figure 11Bii) and that can recruit both Polo (via phosphorylated S-S/T(p) motifs) and Cnn (Conduit et al., 2014b). The active Polo (blue dots, Figure 11Bii) phosphorylates Cnn (orange arrow, Figure 11A), allowing it to also form a scaffold (green, Figure 11Bii) (Conduit et al., 2014a; Feng et al., 2017). The Spd-2 scaffold is inherently unstable (Conduit et al., 2014b), so it can only accumulate around the mother centriole if it is stabilised by the Cnn scaffold (dotted arrow, Figure 11A). The Cnn scaffold therefore allows the Spd-2 scaffold to expand outward, increasing Spd-2 levels within the PCM scaffold and allowing Spd-2 to recruit more Cnn and more Polo into the scaffold. This is a classical positive feedback loop in which the Output (the PCM scaffold in toto) directly increases the Input (the Spd-2 scaffold).

If Spd-2 cannot efficiently recruit Polo, as appears to be the case with the Spd-2-ALL and Spd-2-CONS mutants, it can still recruit Cnn, and this is, at least initially, phosphorylated by the pool of Polo that is still present around the mother centriole (Figure 11C). Our data suggests that this centriolar pool of Polo is not recruited by Spd-2 (at least not via the PBD), and we suspect that S-S/T(p) motifs in other centriole proteins, such as Sas-4 (Novak et al., 2016), normally recruit Polo to centrioles. As a result, mutant Spd-2 proteins can still support the assembly of a ‘mini-scaffold’ around the mother centriole, and this can recruit some PCM and organise some MTs (Figure 11Cii). The mutant Spd-2 scaffold that fluxes outwards from the mother centriole, however, cannot recruit Polo. Therefore the Cnn recruited by the expanding Spd-2 network cannot be phosphorylated, and it cannot form a scaffold to support the expanding Spd-2 network. As a result, the expanding mitotic PCM scaffold rapidly dissipates into the cytosol (Figure 11Ciii).

Although this mechanism is autocatalytic—as the expanding Spd-2 scaffold allows Polo and Cnn to be recruited into the PCM at an increasing rate—crucially, the mother centriole remains the only source of Spd-2 (Figure 11A). This potentially explains the conundrum of how mitotic PCM growth is autocatalytic (Zwicker et al., 2014), but at the same time requires the mother centriole (Basto et al., 2006; Cabral et al., 2019; Kirkham et al., 2003). This requirement for centrioles can also potentially explain how two spatially separated centrosomes usually grow their mitotic PCM to the same size (Conduit et al., 2015; Raff, 2019), as PCM size may ultimately be determined by how much Spd-2 can be provided by the centrioles, rather than how much PCM was present in the centrosome when maturation was initiated.

A key feature of this proposed mechanism is that Cnn cannot recruit itself or Spd-2 or Polo into the scaffold (although it helps to maintain the Spd-2 scaffold recruited by the centriole; Figure 11A). If it could do so, mitotic PCM growth would no longer be constrained by the centriole as Cnn could catalyse its own recruitment. Interestingly, although Spd-2 and Cnn are of similar size in flies (1146aa and 1148aa, respectively) Spd-2 has >5X more conserved potential PBD-binding S-S/T motifs than Cnn (Figure 2—figure supplement 1). Moreover, a similar ratio of conserved sites is found when comparing human Cep192 (1941aa) to human Cep215/Cdk5Rap2 (1893aa) (Figure 2—figure supplement 1), even though the human and fly homologues of both proteins share only limited amino acid identity. Perhaps, these two protein families have evolved to ensure that phosphorylated Spd-2/Cep192 can efficiently recruit Polo/Plk1, whereas phosphorylated Cnn/Cep215 cannot.

Our data indicates that multiple S-S/T(p) motifs in Spd-2 may be involved in Polo recruitment to the PCM. When only the most conserved motifs are mutated, other motifs in Spd-2 appear to be able to help recruit Polo, as evidenced by the additive effect of the Spd-2-ALL mutant compared to the Spd-2-CONS mutant. This mechanism of multi-site phosphorylation and recruitment could help amplify the maturation process (as the additional Polo recruited would allow Cnn to be phosphorylated at a higher rate) and so contribute to the establishment of the positive feedback loop.

Another important feature of this proposed mechanism is that Spd-2 is incorporated into the mitotic PCM at the centriole surface and then fluxes outwards (Conduit et al., 2014b). This Spd-2-flux has so far only been observed in Drosophila embryos and mitotic brain cells (Conduit et al., 2015; Conduit et al., 2010; Conduit et al., 2014b). In fly embryos, Cnn also fluxes outwards but, unlike Spd-2, this flux requires MTs and is only observed in embryos (Conduit and Raff, 2015). In C. elegans embryos, SPD-5 behaves like Cnn in somatic cells: it does not flux outwards and is incorporated isotropically throughout the volume of the PCM (Laos et al., 2015). Moreover, a very recent study found no evidence for an outward centrosomal flux of SPD-2 in worm embryos (Cabral et al., 2019). Clearly, it will be important to determine whether Spd-2/Cep192 homologues flux outwards in other species and, if so, whether this flux provides the primary mechanism by which the mother centriole influences the growth of the expanding mitotic PCM.

In vertebrates, Cep192 serves as a scaffold for Plk1 and also Aurora A (Joukov et al., 2010; Joukov et al., 2014; Meng et al., 2015)—another mitotic protein kinase that plays an important part in centrosome maturation in many species (Barr and Gergely, 2007). There appears to be a complex interplay between Cep192, Plk-1 and Aurora A in vertebrates, with Cep192 acting as a scaffold that allows these two important regulators of mitosis to influence each other’s activity and centrosomal localisation. Spd-2 clearly plays an important part in recruiting Aurora A to centrosomes in fly cells (Conduit et al., 2014b; Dobbelaere et al., 2008)—although it is unclear if this is direct, as fly and worm Spd-2/SPD-2 both lack the N-terminal region in vertebrate Cep192 that recruits Aurora A (Meng et al., 2015). How Aurora A might influence the assembly of the Spd-2, Polo/PLK-1 and Cnn/SPD-5 scaffold remains to be determined, although in worms AIR-1 (the Aurora A homologue) is required to initiate centrosome maturation, but is not required for subsequent PCM growth (Cabral et al., 2019).

Finally, there has been great interest recently in the idea that many non-membrane bound organelles like the centrosome may assemble as ‘condensates’ formed by liquid-liquid phase separation (Banani et al., 2017; Boeynaems et al., 2018; Raff, 2019). In support of this possibility for the centrosome, purified recombinant SPD-5 can assemble into condensates in vitro that have transient liquid-like properties, although they rapidly harden into a more viscous gel- or solid-like phase (Woodruff et al., 2017). Moreover, a mathematical model that describes centrosome maturation in the early worm embryo treats the centrosome as a liquid, and it is from this model that the importance of autocatalysis was first recognised (Zwicker et al., 2014). In vivo, however, the Cnn and SPD-5 scaffolds do not appear to be very liquid-like (Conduit et al., 2010; Conduit et al., 2014b; Laos et al., 2015) and fragments of Cnn can assemble into micron-scale assemblies in vitro that are clearly solid- or very viscous-gel-like (Feng et al., 2017). Our data suggests that the incorporation of Spd-2 into the PCM only at the surface of the centriole, coupled to an amplifying Spd-2/Polo/Cnn positive feedback loop, could provide an ‘autocatalytic’ mechanism that functions within the conceptual framework of a non-liquid-like scaffold that emanates from the mother centriole.

No major datasets have been generated for this study. All data generated or analysed during this study are included in the manuscript and supporting files. As part of this study, we consulted the publicly available iProteinDB database; the corresponding papers have been cited in the main text of this article and further details are provided below. Source data files have been provided for Figure 1 and Figure 1-figure supplement 1D; Figure 2 and Figure 2-figure supplement 1; Figure3B,C; Figure 4; Figure 5; Figure 5-figure supplement 1; Figure 6; Figure 7C,D; Figure 7-figure supplement 1A,B and Figure 7-figure supplement 1E,F; Figure 8, Figure 8-figure supplement 1 and Figure 8-figure supplement 2; Figure 9B-E; Figure 9-figure supplement 1 and Figure 10.

1. 1. Hu Y
   2. Sopko R
   3. Chung V
   4. Foos M
   5. Studer RA
   6. Landry SD
   7. Liu D
   8. Rabinow L
   9. Gnad F
   10. Beltrao P
   11. Perrimon N

   (2019) iProteinDB

   ID FBpp0075122. iProteinDB: An Integrative Database of Drosophila Post-translational Modifications (FBpp0075122).

   https://www.flyrnai.org/tools/iproteindb/web/protein/FBpp0075122/

1. 1. Alves-Cruzeiro JM
   2. Nogales-Cadenas R
   3. Pascual-Montano AD

   (2014) CentrosomeDB: a new generation of the centrosomal proteins database for human and Drosophila melanogaster

   Nucleic Acids Research 42:D430–D436.

   https://doi.org/10.1093/nar/gkt1126

   • PubMed
   • Google Scholar
2. 1. Aydogan MG
   2. Wainman A
   3. Saurya S
   4. Steinacker TL
   5. Caballe A
   6. Novak ZA
   7. Baumbach J
   8. Muschalik N
   9. Raff JW

   (2018) A homeostatic clock sets daughter centriole size in flies

   The Journal of Cell Biology 217:1233–1248.

   https://doi.org/10.1083/jcb.201801014

   • Google Scholar
3. Preprint

   1. Aydogan MG
   2. Steinacker TL
   3. Mofatteh M
   4. Gartenmann L
   5. Wainman A
   6. Saurya S
   7. Wong SS
   8. Zhou FY
   9. Boemo MA
   10. Raff JW

   (2019) An autonomous oscillator times and executes centriole biogenesis

   bioRxiv.

   https://doi.org/10.1101/510875

   • Google Scholar
4. 1. Ball G
   2. Demmerle J
   3. Kaufmann R
   4. Davis I
   5. Dobbie IM
   6. Schermelleh L

   (2015) SIMcheck: a toolbox for successful Super-resolution structured illumination microscopy

   Scientific Reports 5:15915.

   https://doi.org/10.1038/srep15915

   • PubMed
   • Google Scholar
5. 1. Banani SF
   2. Lee HO
   3. Hyman AA
   4. Rosen MK

   (2017) Biomolecular condensates: organizers of cellular biochemistry

   Nature Reviews Molecular Cell Biology 18:285–298.

   https://doi.org/10.1038/nrm.2017.7

   • PubMed
   • Google Scholar
6. 1. Barr AR
   2. Kilmartin JV
   3. Gergely F

   (2010) CDK5RAP2 functions in centrosome to spindle pole attachment and DNA damage response

   The Journal of Cell Biology 189:23–39.

   https://doi.org/10.1083/jcb.200912163

   • PubMed
   • Google Scholar
7. 1. Barr AR
   2. Gergely F

   (2007) Aurora-A: the maker and breaker of spindle poles

   Journal of Cell Science 120:2987–2996.

   https://doi.org/10.1242/jcs.013136

   • PubMed
   • Google Scholar
8. 1. Basto R
   2. Lau J
   3. Vinogradova T
   4. Gardiol A
   5. Woods CG
   6. Khodjakov A
   7. Raff JW

   (2006) Flies without centrioles

   Cell 125:1375–1386.

   https://doi.org/10.1016/j.cell.2006.05.025

   • PubMed
   • Google Scholar
9. 1. Bettencourt-Dias M
   2. Hildebrandt F
   3. Pellman D
   4. Woods G
   5. Godinho SA

   (2011) Centrosomes and cilia in human disease

   Trends in Genetics 27:307–315.

   https://doi.org/10.1016/j.tig.2011.05.004

   • PubMed
   • Google Scholar
10. 1. Bodenmiller B
    2. Malmstrom J
    3. Gerrits B
    4. Campbell D
    5. Lam H
    6. Schmidt A
    7. Rinner O
    8. Mueller LN
    9. Shannon PT
    10. Pedrioli PG
    11. Panse C
    12. Lee HK
    13. Schlapbach R
    14. Aebersold R

    (2007) PhosphoPep--a phosphoproteome resource for systems biology research in Drosophila Kc167 cells

    Molecular Systems Biology 3:.

    https://doi.org/10.1038/msb4100182

    • PubMed
    • Google Scholar
11. 1. Boeynaems S
    2. Alberti S
    3. Fawzi NL
    4. Mittag T
    5. Polymenidou M
    6. Rousseau F
    7. Schymkowitz J
    8. Shorter J
    9. Wolozin B
    10. Van Den Bosch L
    11. Tompa P
    12. Fuxreiter M

    (2018) Protein phase separation: a new phase in cell biology

    Trends in Cell Biology 28:420–435.

    https://doi.org/10.1016/j.tcb.2018.02.004

    • PubMed
    • Google Scholar
12. 1. Buchman JJ
    2. Tseng HC
    3. Zhou Y
    4. Frank CL
    5. Xie Z
    6. Tsai LH

    (2010) Cdk5rap2 interacts with pericentrin to maintain the neural progenitor pool in the developing neocortex

    Neuron 66:386–402.

    https://doi.org/10.1016/j.neuron.2010.03.036

    • PubMed
    • Google Scholar
13. 1. Cabral G
    2. Laos T
    3. Dumont J
    4. Dammermann A

    (2019) Differential requirements for centrioles in mitotic centrosome growth and maintenance

    Developmental Cell 50:355–366.

    https://doi.org/10.1016/j.devcel.2019.06.004

    • PubMed
    • Google Scholar
14. 1. Chavali PL
    2. Pütz M
    3. Gergely F

    (2014) Small organelle, big responsibility: the role of centrosomes in development and disease

    Philosophical Transactions of the Royal Society B: Biological Sciences 369:20130468.

    https://doi.org/10.1098/rstb.2013.0468

    • Google Scholar
15. 1. Choi Y-K
    2. Liu P
    3. Sze SK
    4. Dai C
    5. Qi RZ

    (2010) CDK5RAP2 stimulates microtubule nucleation by the γ-tubulin ring complex

    The Journal of Cell Biology 191:1089–1095.

    https://doi.org/10.1083/jcb.201007030

    • Google Scholar
16. 1. Conduit PT
    2. Brunk K
    3. Dobbelaere J
    4. Dix CI
    5. Lucas EP
    6. Raff JW

    (2010) Centrioles regulate centrosome size by controlling the rate of cnn incorporation into the PCM

    Current Biology 20:2178–2186.

    https://doi.org/10.1016/j.cub.2010.11.011

    • PubMed
    • Google Scholar
17. 1. Conduit PT
    2. Feng Z
    3. Richens JH
    4. Baumbach J
    5. Wainman A
    6. Bakshi SD
    7. Dobbelaere J
    8. Johnson S
    9. Lea SM
    10. Raff JW

    (2014a) The centrosome-specific phosphorylation of cnn by polo/Plk1 drives cnn scaffold assembly and centrosome maturation

    Developmental Cell 28:659–669.

    https://doi.org/10.1016/j.devcel.2014.02.013

    • PubMed
    • Google Scholar
18. 1. Conduit PT
    2. Richens JH
    3. Wainman A
    4. Holder J
    5. Vicente CC
    6. Pratt MB
    7. Dix CI
    8. Novak ZA
    9. Dobbie IM
    10. Schermelleh L
    11. Raff JW

    (2014b) A molecular mechanism of mitotic centrosome assembly in Drosophila

    eLife 3:e03399.

    https://doi.org/10.7554/eLife.03399

    • PubMed
    • Google Scholar
19. 1. Conduit PT
    2. Wainman A
    3. Raff JW

    (2015) Centrosome function and assembly in animal cells

    Nature Reviews Molecular Cell Biology 16:611–624.

    https://doi.org/10.1038/nrm4062

    • PubMed
    • Google Scholar
20. 1. Conduit PT
    2. Raff JW

    (2015) Different Drosophila cell types exhibit differences in mitotic centrosome assembly dynamics

    Current Biology 25:R650–R651.

    https://doi.org/10.1016/j.cub.2015.05.061

    • PubMed
    • Google Scholar
21. 1. Decker M
    2. Jaensch S
    3. Pozniakovsky A
    4. Zinke A
    5. O'Connell KF
    6. Zachariae W
    7. Myers E
    8. Hyman AA

    (2011) Limiting amounts of centrosome material set centrosome size in C. elegans embryos

    Current Biology 21:1259–1267.

    https://doi.org/10.1016/j.cub.2011.06.002

    • PubMed
    • Google Scholar
22. 1. Dix CI
    2. Raff JW

    (2007) Drosophila Spd-2 recruits PCM to the sperm centriole, but is dispensable for centriole duplication

    Current Biology 17:1759–1764.

    https://doi.org/10.1016/j.cub.2007.08.065

    • PubMed
    • Google Scholar
23. 1. Dobbelaere J
    2. Josué F
    3. Suijkerbuijk S
    4. Baum B
    5. Tapon N
    6. Raff J

    (2008) A genome-wide RNAi screen to dissect centriole duplication and centrosome maturation in Drosophila

    PLOS Biology 6:e224.

    https://doi.org/10.1371/journal.pbio.0060224

    • PubMed
    • Google Scholar
24. 1. Elia AE
    2. Cantley LC
    3. Yaffe MB

    (2003a) Proteomic screen finds pSer/pThr-binding domain localizing Plk1 to mitotic substrates

    Science 299:1228–1231.

    https://doi.org/10.1126/science.1079079

    • PubMed
    • Google Scholar
25. 1. Elia AE
    2. Rellos P
    3. Haire LF
    4. Chao JW
    5. Ivins FJ
    6. Hoepker K
    7. Mohammad D
    8. Cantley LC
    9. Smerdon SJ
    10. Yaffe MB

    (2003b) The molecular basis for phosphodependent substrate targeting and regulation of plks by the Polo-box domain

    Cell 115:83–95.

    https://doi.org/10.1016/S0092-8674(03)00725-6

    • PubMed
    • Google Scholar
26. 1. Feng Z
    2. Caballe A
    3. Wainman A
    4. Johnson S
    5. Haensele AFM
    6. Cottee MA
    7. Conduit PT
    8. Lea SM
    9. Raff JW

    (2017) Structural basis for mitotic centrosome assembly in flies

    Cell 169:1078–1089.

    https://doi.org/10.1016/j.cell.2017.05.030

    • PubMed
    • Google Scholar
27. 1. Foe VE
    2. Alberts BM

    (1983)

    Studies of nuclear and cytoplasmic behaviour during the five mitotic cycles that precede gastrulation in Drosophila embryogenesis

    Journal of Cell Science 61:31–70.

    • PubMed
    • Google Scholar
28. 1. Fong KW
    2. Choi YK
    3. Rattner JB
    4. Qi RZ

    (2008) CDK5RAP2 is a pericentriolar protein that functions in centrosomal attachment of the gamma-tubulin ring complex

    Molecular Biology of the Cell 19:115–125.

    https://doi.org/10.1091/mbc.e07-04-0371

    • PubMed
    • Google Scholar
29. 1. Fu J
    2. Glover DM

    (2012) Structured illumination of the interface between centriole and peri-centriolar material

    Open Biology 2:120104.

    https://doi.org/10.1098/rsob.120104

    • PubMed
    • Google Scholar
30. 1. Galletta BJ
    2. Fagerstrom CJ
    3. Schoborg TA
    4. McLamarrah TA
    5. Ryniawec JM
    6. Buster DW
    7. Slep KC
    8. Rogers GC
    9. Rusan NM

    (2016) A centrosome interactome provides insight into organelle assembly and reveals a non-duplication role for Plk4

    Nature Communications 7:12476.

    https://doi.org/10.1038/ncomms12476

    • PubMed
    • Google Scholar
31. 1. Gomez-Ferreria MA
    2. Rath U
    3. Buster DW
    4. Chanda SK
    5. Caldwell JS
    6. Rines DR
    7. Sharp DJ

    (2007) Human Cep192 is required for mitotic centrosome and spindle assembly

    Current Biology 17:1960–1966.

    https://doi.org/10.1016/j.cub.2007.10.019

    • PubMed
    • Google Scholar
32. 1. Habermann K
    2. Mirgorodskaya E
    3. Gobom J
    4. Lehmann V
    5. Müller H
    6. Blümlein K
    7. Deery MJ
    8. Czogiel I
    9. Erdmann C
    10. Ralser M
    11. von Kries JP
    12. Lange BM

    (2012) Functional analysis of centrosomal kinase substrates in Drosophila melanogaster reveals a new function of the nuclear envelope component otefin in cell cycle progression

    Molecular and Cellular Biology 32:3554–3569.

    https://doi.org/10.1128/MCB.00814-12

    • PubMed
    • Google Scholar
33. 1. Haren L
    2. Stearns T
    3. Lüders J

    (2009) Plk1-dependent recruitment of gamma-tubulin complexes to mitotic centrosomes involves multiple PCM components

    PLOS ONE 4:e5976.

    https://doi.org/10.1371/journal.pone.0005976

    • PubMed
    • Google Scholar
34. 1. Hu Y
    2. Sopko R
    3. Chung V
    4. Foos M
    5. Studer RA
    6. Landry SD
    7. Liu D
    8. Rabinow L
    9. Gnad F
    10. Beltrao P
    11. Perrimon N

    (2019) iProteinDB: an integrative database of Drosophila Post-translational modifications

    G3: Genes|Genomes|Genetics 9:1–11.

    https://doi.org/10.1142/S0219720017400054

    • Google Scholar
35. 1. Joukov V
    2. De Nicolo A
    3. Rodriguez A
    4. Walter JC
    5. Livingston DM

    (2010) Centrosomal protein of 192 kDa (Cep192) promotes centrosome-driven spindle assembly by engaging in organelle-specific aurora A activation

    PNAS 107:21022–21027.

    https://doi.org/10.1073/pnas.1014664107

    • PubMed
    • Google Scholar
36. 1. Joukov V
    2. Walter JC
    3. De Nicolo A

    (2014) The Cep192-organized aurora A-Plk1 cascade is essential for centrosome cycle and bipolar spindle assembly

    Molecular Cell 55:578–591.

    https://doi.org/10.1016/j.molcel.2014.06.016

    • PubMed
    • Google Scholar
37. 1. Kim S
    2. Rhee K

    (2014) Importance of the CEP215-pericentrin interaction for centrosome maturation during mitosis

    PLOS ONE 9:e87016.

    https://doi.org/10.1371/journal.pone.0087016

    • PubMed
    • Google Scholar
38. 1. Kirkham M
    2. Müller-Reichert T
    3. Oegema K
    4. Grill S
    5. Hyman AA

    (2003) SAS-4 Is a C. elegans Centriolar Protein that Controls Centrosome Size

    Cell 112:575–587.

    https://doi.org/10.1016/S0092-8674(03)00117-X

    • Google Scholar
39. 1. Koenig T
    2. Menze BH
    3. Kirchner M
    4. Monigatti F
    5. Parker KC
    6. Patterson T
    7. Steen JJ
    8. Hamprecht FA
    9. Steen H

    (2008) Robust Prediction of the MASCOT Score for an Improved Quality Assessment in Mass Spectrometric Proteomics

    Journal of Proteome Research 7:3708–3717.

    https://doi.org/10.1021/pr700859x

    • Google Scholar
40. 1. Lane HA
    2. Nigg EA

    (1996) Antibody microinjection reveals an essential role for human polo-like kinase 1 (Plk1) in the functional maturation of mitotic centrosomes

    The Journal of Cell Biology 135:1701–1713.

    https://doi.org/10.1083/jcb.135.6.1701

    • Google Scholar
41. 1. Laos T
    2. Cabral G
    3. Dammermann A

    (2015) Isotropic incorporation of SPD-5 underlies centrosome assembly in C. elegans

    Current Biology 25:R648–R649.

    https://doi.org/10.1016/j.cub.2015.05.060

    • PubMed
    • Google Scholar
42. 1. Lawo S
    2. Hasegan M
    3. Gupta GD
    4. Pelletier L

    (2012) Subdiffraction imaging of centrosomes reveals higher-order organizational features of pericentriolar material

    Nature Cell Biology 14:1148–1158.

    https://doi.org/10.1038/ncb2591

    • Google Scholar
43. 1. Lee KS
    2. Grenfell TZ
    3. Yarm FR
    4. Erikson RL

    (1998) Mutation of the polo-box disrupts localization and mitotic functions of the mammalian polo kinase Plk

    PNAS 95:9301–9306.

    https://doi.org/10.1073/pnas.95.16.9301

    • Google Scholar
44. 1. Lee K
    2. Rhee K

    (2011) PLK1 phosphorylation of pericentrin initiates centrosome maturation at the onset of mitosis

    The Journal of Cell Biology 195:1093–1101.

    https://doi.org/10.1083/jcb.201106093

    • PubMed
    • Google Scholar
45. 1. Lehmann V
    2. Müller H
    3. Lange BM

    (2006) Immunoisolation of centrosomes from Drosophila melanogaster

    Current Protocols in Cell Biology 3:.

    https://doi.org/10.1002/0471143030.cb0317s29

    • PubMed
    • Google Scholar
46. 1. Lerit DA
    2. Jordan HA
    3. Poulton JS
    4. Fagerstrom CJ
    5. Galletta BJ
    6. Peifer M
    7. Rusan NM

    (2015) Interphase centrosome organization by the PLP-Cnn scaffold is required for centrosome function

    The Journal of Cell Biology 210:79–97.

    https://doi.org/10.1083/jcb.201503117

    • PubMed
    • Google Scholar
47. 1. Liu J
    2. Lewellyn AL
    3. Chen LG
    4. Maller JL

    (2004) The polo box is required for multiple functions of Plx1 in mitosis

    Journal of Biological Chemistry 279:21367–21373.

    https://doi.org/10.1074/jbc.M400482200

    • PubMed
    • Google Scholar
48. 1. Lizarraga SB
    2. Margossian SP
    3. Harris MH
    4. Campagna DR
    5. Han AP
    6. Blevins S
    7. Mudbhary R
    8. Barker JE
    9. Walsh CA
    10. Fleming MD

    (2010) Cdk5rap2 regulates centrosome function and chromosome segregation in neuronal progenitors

    Development 137:1907–1917.

    https://doi.org/10.1242/dev.040410

    • Google Scholar
49. 1. Martinez-Campos M
    2. Basto R
    3. Baker J
    4. Kernan M
    5. Raff JW

    (2004) The Drosophila pericentrin-like protein is essential for cilia/flagella function, but appears to be dispensable for mitosis

    The Journal of Cell Biology 165:673–683.

    https://doi.org/10.1083/jcb.200402130

    • PubMed
    • Google Scholar
50. 1. Meng L
    2. Park JE
    3. Kim TS
    4. Lee EH
    5. Park SY
    6. Zhou M
    7. Bang JK
    8. Lee KS

    (2015) Bimodal interaction of mammalian Polo-Like kinase 1 and a centrosomal scaffold, Cep192, in the regulation of bipolar spindle formation

    Molecular and Cellular Biology 35:2626–2640.

    https://doi.org/10.1128/MCB.00068-15

    • PubMed
    • Google Scholar
51. 1. Mennella V
    2. Keszthelyi B
    3. McDonald KL
    4. Chhun B
    5. Kan F
    6. Rogers GC
    7. Huang B
    8. Agard DA

    (2012) Subdiffraction-resolution fluorescence microscopy reveals a domain of the centrosome critical for pericentriolar material organization

    Nature Cell Biology 14:1159–1168.

    https://doi.org/10.1038/ncb2597

    • PubMed
    • Google Scholar
52. 1. Moritz M
    2. Braunfeld MB
    3. Sedat JW
    4. Alberts B
    5. Agard DA

    (1995) Microtubule nucleation by gamma-tubulin-containing rings in the centrosome

    Nature 378:638–640.

    https://doi.org/10.1038/378638a0

    • PubMed
    • Google Scholar
53. 1. Moritz M
    2. Zheng Y
    3. Alberts BM
    4. Oegema K

    (1998) Recruitment of the gamma-tubulin ring complex to Drosophila salt-stripped centrosome scaffolds

    The Journal of Cell Biology 142:775–786.

    https://doi.org/10.1083/jcb.142.3.775

    • PubMed
    • Google Scholar
54. 1. Nigg EA
    2. Raff JW

    (2009) Centrioles, centrosomes, and cilia in health and disease

    Cell 139:663–678.

    https://doi.org/10.1016/j.cell.2009.10.036

    • PubMed
    • Google Scholar
55. 1. Novak ZA
    2. Conduit PT
    3. Wainman A
    4. Raff JW

    (2014) Asterless licenses daughter centrioles to duplicate for the first time in Drosophila embryos

    Current Biology 24:1276–1282.

    https://doi.org/10.1016/j.cub.2014.04.023

    • PubMed
    • Google Scholar
56. 1. Novak ZA
    2. Wainman A
    3. Gartenmann L
    4. Raff JW

    (2016) Cdk1 phosphorylates Drosophila Sas-4 to recruit polo to daughter centrioles and convert them to centrosomes

    Developmental Cell 37:545–557.

    https://doi.org/10.1016/j.devcel.2016.05.022

    • PubMed
    • Google Scholar
57. 1. Palazzo RE
    2. Vogel JM
    3. Schnackenberg BJ
    4. Hull DR
    5. Wu X

    (2000) Centrosome maturation

    Current Topics in Developmental Biology 49:449–470.

    https://doi.org/10.1016/S0070-2153(99)49021-0

    • PubMed
    • Google Scholar
58. 1. Raff JW
    2. Kellum R
    3. Alberts B

    (1994) The Drosophila GAGA transcription factor is associated with specific regions of heterochromatin throughout the cell cycle

    The EMBO Journal 13:5977–5983.

    https://doi.org/10.1002/j.1460-2075.1994.tb06943.x

    • PubMed
    • Google Scholar
59. 1. Raff JW

    (2019) Phase separation and the centrosome: a fait accompli?

    Trends in Cell Biology 29:612–622.

    https://doi.org/10.1016/j.tcb.2019.04.001

    • PubMed
    • Google Scholar
60. 1. Reynolds N
    2. Ohkura H

    (2003) Polo boxes form a single functional domain that mediates interactions with multiple proteins in fission yeast polo kinase

    Journal of Cell Science 116:1377–1387.

    https://doi.org/10.1242/jcs.00314

    • PubMed
    • Google Scholar
61. 1. Richens JH
    2. Barros TP
    3. Lucas EP
    4. Peel N
    5. Pinto DM
    6. Wainman A
    7. Raff JW

    (2015) The Drosophila Pericentrin-like-protein (PLP) cooperates with cnn to maintain the integrity of the outer PCM

    Biology Open 4:1052–1061.

    https://doi.org/10.1242/bio.012914

    • PubMed
    • Google Scholar
62. Book

    1. Roberts EB

    (1998)

    Drosophila: a Practical Approach

    Oxford University Press.

    • Google Scholar
63. 1. Schnackenberg BJ
    2. Khodjakov A
    3. Rieder CL
    4. Palazzo RE

    (1998) The disassembly and reassembly of functional centrosomes in vitro

    PNAS 95:9295–9300.

    https://doi.org/10.1073/pnas.95.16.9295

    • PubMed
    • Google Scholar
64. 1. Seong YS
    2. Kamijo K
    3. Lee JS
    4. Fernandez E
    5. Kuriyama R
    6. Miki T
    7. Lee KS

    (2002) A spindle checkpoint arrest and a cytokinesis failure by the dominant-negative polo-box domain of Plk1 in U-2 OS cells

    Journal of Biological Chemistry 277:32282–32293.

    https://doi.org/10.1074/jbc.M202602200

    • PubMed
    • Google Scholar
65. 1. Song S
    2. Grenfell TZ
    3. Garfield S
    4. Erikson RL
    5. Lee KS

    (2000) Essential function of the polo box of Cdc5 in subcellular localization and induction of cytokinetic structures

    Molecular and Cellular Biology 20:286–298.

    https://doi.org/10.1128/MCB.20.1.286-298.2000

    • PubMed
    • Google Scholar
66. 1. Sonnen KF
    2. Schermelleh L
    3. Leonhardt H
    4. Nigg EA

    (2012) 3D-structured illumination microscopy provides novel insight into architecture of human centrosomes

    Biology Open 1:965–976.

    https://doi.org/10.1242/bio.20122337

    • PubMed
    • Google Scholar
67. 1. Stevens NR
    2. Dobbelaere J
    3. Brunk K
    4. Franz A
    5. Raff JW

    (2010) Drosophila Ana2 is a conserved centriole duplication factor

    The Journal of Cell Biology 188:313–323.

    https://doi.org/10.1083/jcb.200910016

    • PubMed
    • Google Scholar
68. 1. Tinevez J-Y
    2. Perry N
    3. Schindelin J
    4. Hoopes GM
    5. Reynolds GD
    6. Laplantine E
    7. Bednarek SY
    8. Shorte SL
    9. Eliceiri KW

    (2017) TrackMate: an open and extensible platform for single-particle tracking

    Methods 115:80–90.

    https://doi.org/10.1016/j.ymeth.2016.09.016

    • Google Scholar
69. 1. Vertii A
    2. Hehnly H
    3. Doxsey S

    (2016) The centrosome, a multitalented renaissance organelle

    Cold Spring Harbor Perspectives in Biology 8:a025049.

    https://doi.org/10.1101/cshperspect.a025049

    • PubMed
    • Google Scholar
70. 1. Woodruff JB
    2. Wueseke O
    3. Hyman AA

    (2014) Pericentriolar material structure and dynamics

    Philosophical Transactions of the Royal Society B: Biological Sciences 369:20130459.

    https://doi.org/10.1098/rstb.2013.0459

    • Google Scholar
71. 1. Woodruff JB
    2. Wueseke O
    3. Viscardi V
    4. Mahamid J
    5. Ochoa SD
    6. Bunkenborg J
    7. Widlund PO
    8. Pozniakovsky A
    9. Zanin E
    10. Bahmanyar S
    11. Zinke A
    12. Hong SH
    13. Decker M
    14. Baumeister W
    15. Andersen JS
    16. Oegema K
    17. Hyman AA

    (2015) Centrosomes. Regulated assembly of a supramolecular centrosome scaffold in vitro

    Science 348:808–812.

    https://doi.org/10.1126/science.aaa3923

    • PubMed
    • Google Scholar
72. 1. Woodruff JB
    2. Ferreira Gomes B
    3. Widlund PO
    4. Mahamid J
    5. Honigmann A
    6. Hyman AA

    (2017) The centrosome is a selective condensate that nucleates microtubules by concentrating tubulin

    Cell 169:1066–1077.

    https://doi.org/10.1016/j.cell.2017.05.028

    • PubMed
    • Google Scholar
73. 1. Wueseke O
    2. Zwicker D
    3. Schwager A
    4. Wong YL
    5. Oegema K
    6. Jülicher F
    7. Hyman AA
    8. Woodruff JB

    (2016) Polo-like kinase phosphorylation determines Caenorhabditis elegans centrosome size and density by biasing SPD-5 toward an assembly-competent conformation

    Biology Open 5:1431–1440.

    https://doi.org/10.1242/bio.020990

    • PubMed
    • Google Scholar
74. 1. Zhai B
    2. Villén J
    3. Beausoleil SA
    4. Mintseris J
    5. Gygi SP

    (2008) Phosphoproteome analysis of Drosophila melanogaster embryos

    Journal of Proteome Research 7:1675–1682.

    https://doi.org/10.1021/pr700696a

    • PubMed
    • Google Scholar
75. 1. Zhu F
    2. Lawo S
    3. Bird A
    4. Pinchev D
    5. Ralph A
    6. Richter C
    7. Müller-Reichert T
    8. Kittler R
    9. Hyman AA
    10. Pelletier L

    (2008) The mammalian SPD-2 ortholog Cep192 regulates centrosome biogenesis

    Current Biology 18:136–141.

    https://doi.org/10.1016/j.cub.2007.12.055

    • PubMed
    • Google Scholar
76. 1. Zwicker D
    2. Decker M
    3. Jaensch S
    4. Hyman AA
    5. Jülicher F

    (2014) Centrosomes are autocatalytic droplets of pericentriolar material organized by centrioles

    PNAS 111:E2636–E2645.

    https://doi.org/10.1073/pnas.1404855111

    • PubMed
    • Google Scholar

[Editors’ note: a previous version of this study was rejected after peer review, but the authors submitted for reconsideration. The first decision letter after peer review is shown below.]

Thank you for submitting your work entitled "A positive feedback loop drives centrosome maturation in flies" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by a Reviewing Editor and a Senior Editor. The reviewers have opted to remain anonymous.

Our decision has been reached after consultation between the reviewers. Based on these discussions and the individual reviews below, we regret to inform you that your work will not be considered further for publication in eLife.

As you can see in the individual reviewers' comments, all appreciated how well the study is done addressing a critical question in the field of centrosome biology. However, all remained concerned that the study did not provide a proof of Spd-2-Polo interaction being critical. They felt that proving this point is not necessarily straightforward in the scope of a simple revision (leaving the possibility open that those experiments disprove the model, or it might be not straightforward to design the experiments to prove this point).

With that said, all reviewers are enthusiastic about this study overall: although we cannot consider publication of your manuscript at this point per eLife's policy to invite revision only if the revision experiments are straightforward and the results are unlikely to change the major conclusions, we would be happy to reconsider your work as a new submission, if you're able to address the major points of reviewers.

Reviewer #1:

The manuscript "A positive feedback loop drives centrosome maturation in flies" by Alvarez-Rodrigo et al. investigates the regulation of centrosome maturation when cells prepare for mitosis, by focusing on the expansion of the pericentriolar material (PCM). In particular, the authors analyze the involvement of Spd-2 and Polo in this process and study potential phosphorylation sites in Spd-2 that may help recruit Polo and that seem to be important for pericentriolar scaffold assembly and expansion, including the recruitment of other essential PCM proteins such as γ-tubulin to the expanded scaffold. They propose that Spd-2 is an essential component of a feedback mechanism that involves recruitment of Polo to drive incorporation and phosphorylation of Cnn near centrioles and subsequent outward movement of Cnn-Spd-2, thereby contributing to assembly of an expanded PCM. This process is driven by continuous Spd-2-Polo-dependent incorporation of additional Cnn near centrioles and its outward flux.

Overall this is a very nice story that puts together various previous and new observations into an appealing model to explain the mechanism by which centrosomes mature, at least in flies.

The paper is very well written and the data are presented clearly.

However, I found one major issue with this work that needs to be addressed. In my opinion the authors do not provide convincing evidence for a core element of their model: the role of Spd-2 in Polo recruitment.

Specific points:

1) The authors state "Based on our observations below, we refer to these mutant proteins as being unable to recruit Polo." However, the recruitment of Polo to Spd-2 is never demonstrated in the first place. I understand that the interaction is supposed to happen specifically at the centrosome, so immunoprecipitation from soluble extract is most likely not an option, but can the author provide any evidence for a direct interaction of Polo with Spd-2?

2) The observations that follow the authors' statement do not demonstrate the inability of the mutants to recruit Polo: it is shown that the mutants partially rescue embryos lacking Spd-2, that mutants are less concentrated at centrosomes, that these centrosomes have less MTs, and that mutants and Polo localize around centrioles but do not assemble into an expanded pericentriolar scaffold. This latter finding is the only piece of data in support of the proposed role of Spd-2 in Polo binding. However, as the authors also state, in these experiments the scaffold may simply not assemble. In this case the reduced Polo signal could simply be the result of the absence of an expanded scaffold and not necessarily due to lack of binding to Spd-2.

3) The demonstration that Spd-2 is phosphorylated at centrosomes is very convincing (Figure 1). However, it was not tested whether this phosphorylation is at least partially Polo-dependent. This could indicate interaction with Polo at centrosomes.

4) Figure 2: Are the mutants still phosphorylated at centrosomes? This should be revealed by an analysis as in Figure 1.

Reviewer #2:

In the manuscript by Alvarez-Rodrigo et al. entitled "A positive feedback loop drives centrosome maturation in flies" the authors aim to show that the interactions between Spd2-Polo-Cnn drive an autocatalytic expansion of the PCM to ensure synchronized growth of sister centrosomes. Based on a minimal Polo box binding consensus sequence, the authors mutated Spd2 to prevent Polo binding and then characterized the effects on fly viability, Spd2 localization, and PCM expansion. They found that increasing the mutational burden on Spd2 reduces PCM expansion and localization of centrosome components during centrosome maturation. If the authors were able to show the mechanism in Figure 10, this would constitute a major contribution to the centrosome field. Unfortunately, the lack of justification or biological relevance behind the authors' approach prevent them from achieving this goal. Therefore, I do not believe the manuscript should be published without extensive changes to the approach or substantial work to justify and validate their approach.

The authors perform initial MS/MS to identify the phosphorylated residues on Spd2 which may be a prerequisite for Polo binding. They then mutated these residues and found mild Spd2 defects that did not recapitulate the phenotypes found in the CONS or ALL mutants. From these data, it appears that mutating the biologically relevant sites was not sufficient to produce a phenotype, so the authors moved on to mutating a substantial number of serine residues in Spd2 only because it appears in a minimal consensus sequence. This approach led to studying the effects of a massively mutated protein without regard to the biology of the system.

Furthermore, the authors do not provide evidence that the Spd2 mutants result in an inability of Polo to bind Spd2. Perhaps Polo can bind Spd2 through non-canonical sites? That may be the case because the authors present evidence that Polo is still recruited to the centriole and functions properly, as Cnn is phosphorylated and centrioles are still able to disengage properly. Are the Spd2 mutants still being phosphorylated, as in Figure 1? This can be answered with their established method in Figure 1 and by mass spectrometry. Is mutant Spd2 still able to homodimerize and interact with Cnn? The lab has previously established a yeast 2-hybrid assay and has used SEC-MALS, both of which could be used to test this. Spd2 also interacts with Asterless and PLP. Do the Spd CONS or ALL mutants disrupt these interactions, which may explain the mutant phenotypes? Further biochemical analysis of mutant Spd2 should be performed to show that the mutants do not cause protein misfolding.

Although their discussion and conclusions are intriguing and could provide insight into how centrosome maturation may be regulated, I do not feel they have provided sufficient evidence to support these conclusions. Additionally, while there is a striking phenotype in the manuscript, the authors have only shown that this is caused by mutating Spd2. They have not shown that this is due to disrupting the interaction with Polo, let alone disrupting a potential positive feedback loop, as implied by the title of their manuscript, which is misleading.

Additional concerns:

1) All experiments are performed in Drosophila embryos, whereas the authors claim that the mechanism is universal to the entire fly "in vivo". This should be toned down.

2) The manuscript would benefit from more details of the mitotic defects in Figure 2A.

3) Are Spd2ALL, Spd2^-/- heterozygotes viable or can reduced dosage explain the phenotypes in Figure 5—figure supplement 1?

4) The images of centrosomes in Figures 4-6 appear to be S-phase centrosomes with flares, not mitotic centrosomes. The authors reference mitosis and maturation when, in fact, they show images of interphase (S-phase) centrosomes. The authors should show low magnification images of mitotic spindles and then high magnification insets of centrosomes to demonstrate that they are, in fact, examining mitotic centrosomes.

Reviewer #3:

In this manuscript, the authors show that Spd-2 is partially phosphorylated at centrosomes, allowing the recruitment of Polo to the centrosomes. Disabling the recruitment leads to outward pericentriolar scaffold expansion failure, and prevents centrosome maturation. The authors also demonstrate that Polo, Spd-2 and Cnn together can drives mitotic centrosome expansion by forming a positive feed-back loop. This is an interesting paper with data of high quality. I note the following issue the authors need to consider:

1) The authors should provide direct evidence about how much Spd-2-CONS and Spd-2-ALL are affects their binding with the Polo PBD compared to WT, which is critical for most of their conclusions. I'm concerned that mutating all 34 Serine sites may affect protein structure and this needs to be ruled out.

2) Do the authors now if the phosphorylation status is altered during the cell cycle?

3) Figure 4: authors conclude that the inability of mutant proteins to recruit Polo reduces the assembly and/or maintenance of this mixed scaffold, but the data shown can't exclude the possibility that mutant proteins are efficiently incorporated into the scaffold. The author should show a comparison of endogenous Spd-2 scaffold in all three conditions.

4) Figure 5: I'm not convinced by the authors' conclusion that the pericentriolar scaffold cannot expand because Spd-2-CON can't recruit Polo. First, they co-localize at the mother centriole, and second, Spd-2-CON itself is affecting scaffold formation in Spd-2 mutant background (Figure 4), so how can they be certain it's a Polo recruitment issue? I would potentially suggest a FRAP experiment to showcase the recruitment of Polo to the scaffold. In addition, apart from co-expressing with Polo-GFP, what's the difference between Figure 5B and Figure 4A (middle lane)? Why is there's a huge difference between the ratios (64/36 vs 9/91)? This was confusing to me.

5) Figure 7: what is the component of the scaffold? If the scaffold is made of Spd-2, when Spd-2 CONS is unable to form the scaffold, you won't be able to see other proteins there no matter how they are recruited. Is there any other marker for the scaffold?

6) Figure 9: The C-terminus of Spd-2 contains most of the conserved Serine sites, but the authors show that mutation N-terminal sites have the biggest effect? What do the authors think the molecular basis for this is?

[Editors’ note: what now follows is the decision letter after the authors submitted for further consideration.]

Thank you for submitting your article "Evidence that a positive feedback loop drives centrosome maturation in fly embryos" for consideration by eLife. Your article has been reviewed by three peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Anna Akhmanova as the Senior Editor. The reviewers have opted to remain anonymous.

The reviewers have discussed the reviews with one another and the Reviewing Editor has drafted this decision to help you prepare a revised submission.

Summary:

In the manuscript by Alvarez-Rodrigo et al., the authors show that the interactions between Spd2-Polo-Cnn drive an autocatalytic expansion of the PCM to ensure synchronized growth of sister centrosomes.

This is a resubmission of previously rejected manuscript, where all the reviewers appreciated the rigor of the work and an elegant model to explain the process of centrosome maturation but found that the direct evidence to support the importance of Spd-2-Polo interaction was lacking.

This revised and newly submitted version was reviewed by original reviewers, and now they all agree that this work now has strong evidence that supports the conclusions. Some reviewers raised a few minor comments, which the authors may want to address prior to formal acceptance.

Overall this is a very nice story that puts together various previous and new observations into an appealing model to explain the mechanism by which centrosomes mature, which may be conserved in other species as well.

Reviewer #1:

In the manuscript by Alvarez-Rodrigo et al. entitled "A positive feedback loop drives centrosome maturation in flies" the authors show that the interactions between Spd2-Polo-Cnn drive an autocatalytic expansion of the PCM to ensure synchronized growth of sister centrosomes. I feel that the authors have addressed my original concerns with one remaining issue. Regarding the immunoprecipitation (IP) experiments between the MBP-phospho-Spd2 fragment and GST-PBD (new Figure 1C), the resolution of these immunoblots appear low and the brightness/contrast altered such that it is difficult to gauge the load between the samples. It appears that there is more GST-PBD in the MBP-Spd-2 WT immunoprecipitate, but how much more? Replacing these with higher resolution blots with accompanying quantitation would be appropriate. This is an important experiment/result because it demonstrates that the authors have in fact generated a Spd-2 Polo-binding mutant. Also, is the kinase being used human Plk1, whereas the Spd-2 and PBD proteins are fly? Whatever the species, the fact that the authors used proteins from different organisms should be stated in the Results section.

I also recommend a couple of text changes in this section. The authors state that binding of the 19T mutant was reduced to background levels. How do they know this is background? This could be real binding and, therefore, would remain at WT levels, not a reduced level. The authors end this section stating, "Thus, a fragment of Spd-2 can bind directly to the PBD when phosphorylated, and this binding is prevented when the S-S/T motifs are mutated to T-S/T". Based on the data, it looks to me that binding is not prevented. Instead, they should say that phosphorylation of Spd-2 increases PDB binding which, in my opinion, is what the data shows. The data also suggest that Polo activity primes it's binding to Spd-2 (as they state in the Discussion) but this important finding should also be emphasized here at the end of this Results section.

Reviewer #2:

In their revised manuscript the authors have addressed my concerns by providing new experiments and additional discussion/explanation.

My main concerns were the lack of evidence for interaction between Spd-2 and Polo and the conundrum that the reduced Polo signal at centrosomes containing Spd-2 mutants may be due to specific loss of interaction with Polo or simply due to the lack of any scaffold where Polo could be recruited.

The first concern was dealt with by providing in vitro biochemical interaction data (Figure 1C) and the second by performing a new type of quantification of the centrosomal Polo signal in cells expressing a mix of WT and mutant Spd-2, at a time point when Polo signal is just starting to become reduced (Figure 10). This approach allowed the authors to demonstrate reduced Polo signal even in cells where a Spd-2 PCM scaffold is still present. Although the data supports the authors' conclusion that Spd-2 mutants cause a specific loss of PCM associated Polo, the quantification would be more convincing if actual intensities would have been quantified for the Spd-2 scaffold and Polo signals rather than subjective classification into "weak" and "strong". This is because in contrast to a similar quantification in Figure 6, where an "all or nothing" effect is observed, here the relative intensities of PCM scaffold-associated Spd-2 vs Polo are crucial for the result. I would suggest to also adding quantification of intensities to the figure.

My other concerns were also addressed, not in all cases with experiments, but I recognize the technical challenges that have led to this decision.

I would now support publication.

Reviewer #3:

The revised version of the paper is much improved and the authors have addressed most of the issues raised in my initial review to the best of their abilities. The manuscript in my opinion is now suitable for publication in eLife.

[Editors’ note: the author responses to the first round of peer review follow.]

Reviewer #1:

[…] Overall this is a very nice story that puts together various previous and new observations into an appealing model to explain the mechanism by which centrosomes mature, at least in flies.

The paper is very well written and the data are presented clearly.

However, I found one major issue with this work that needs to be addressed. In my opinion the authors do not provide convincing evidence for a core element of their model: the role of Spd-2 in Polo recruitment.

Specific points:

1) The authors state "Based on our observations below, we refer to these mutant proteins as being unable to recruit Polo." However, the recruitment of Polo to Spd-2 is never demonstrated in the first place. I understand that the interaction is supposed to happen specifically at the centrosome, so immunoprecipitation from soluble extract is most likely not an option, but can the author provide any evidence for a direct interaction of Polo with Spd-2?

To address this question, we have now purified an MBP-fusion containing a fragment of Spd-2 (aa352-758; the only recombinant fragment of Spd-2 that is relatively stable in our hands). This fragment contains 19 S-S/T residues and we show that it can bind to recombinant GST-PBD, but only when it has first been phosphorylated by recombinant Plk1 (New Figure 1C). Mutating the 19 S-S/T residues in this Spd-2 fragment to T-S/T abolishes PBD-binding. These findings provide strong support for our hypothesis that Drosophila Spd-2 can directly bind to the Polo-PBD when phosphorylated, and that this binding is prevented when the S-S/T sites are mutated to T-S/T.

2) The observations that follow the authors' statement do not demonstrate the inability of the mutants to recruit Polo: it is shown that the mutants partially rescue embryos lacking Spd-2, that mutants are less concentrated at centrosomes, that these centrosomes have less MTs, and that mutants and Polo localize around centrioles but do not assemble into an expanded pericentriolar scaffold. This latter finding is the only piece of data in support of the proposed role of Spd-2 in Polo binding. However, as the authors also state, in these experiments the scaffold may simply not assemble. In this case the reduced Polo signal could simply be the result of the absence of an expanded scaffold and not necessarily due to lack of binding to Spd-2.

To address this problem we have now injected mRNA encoding either WT or mutant forms of Spd-2-mKate2 into embryos expressing Polo-GFP. The mRNA is gradually translated, and so the Spd-2-mKate2 fusions gradually overwhelm the endogenous untagged Spd-2. We then looked for embryos injected with the mutant mRNA where the loss of Polo from the centrosome was first becoming apparent, and compared the centrosomal recruitment of the mutant Spd-2-mKate2 and Polo-GFP to similarly staged controls (injected with WT Spd-2-mKate2 mRNA). This analysis, scored blind, revealed that mutant Spd-2-mKate2 could still be detected in an expanded PCM even when the amount of Polo recruited to the scaffold was severely reduced (New Figure 10). This is presumably because there is enough WT Spd-2 present in these embryos to form a scaffold, but the mutant Spd-2 protein present in the scaffold cannot recruit Polo. This strongly supports our conclusion that the mutant Spd-2 proteins cannot recruit Polo to the expanded PCM. Nevertheless, we agree that we cannot formally exclude the possibility that the mutant Spd-2 proteins can still bind Polo in some way, and we have toned down our statements on this point throughout the manuscript.

3) The demonstration that Spd-2 is phosphorylated at centrosomes is very convincing (Figure 1). However, it was not tested whether this phosphorylation is at least partially Polo-dependent. This could indicate interaction with Polo at centrosomes.

This is a good question, but one that is difficult to address: a lack of Spd-2 phosphorylation after Polo inhibition might be a direct consequence of Polo failing to phosphorylate Spd-2, or an indirect consequence of there being no expanded PCM (as Spd-2 is mostly phosphorylated in the PCM). We do now provide evidence that Polo can phosphorylate a fragment of Spd-2 in vitro to create PBD binding sites (New Figure 1C), consistent with the possibility that at least some of the phosphorylation of Spd-2 at centrosomes is Polo-dependent.

4) Figure 2: Are the mutants still phosphorylated at centrosomes? This should be revealed by an analysis as in Figure 1.

This is a good question, but we have not attempted to answer it as it would be a lot of work to collect enough mutant material to do the biochemistry, and we think the answer may not be very informative: the S-S/T to T-S/T substitutions may or may not interfere with the phosphorylation of these motifs, but will perturb PBD-binding regardless.

Reviewer #2:

[…] If the authors were able to show the mechanism in Figure 10, this would constitute a major contribution to the centrosome field. Unfortunately, the lack of justification or biological relevance behind the authors' approach prevent them from achieving this goal. Therefore, I do not believe the manuscript should be published without extensive changes to the approach or substantial work to justify and validate their approach.

The authors perform initial MS/MS to identify the phosphorylated residues on Spd2 which may be a prerequisite for Polo binding. They then mutated these residues and found mild Spd2 defects that did not recapitulate the phenotypes found in the CONS or ALL mutants. From these data, it appears that mutating the biologically relevant sites was not sufficient to produce a phenotype, so the authors moved on to mutating a substantial number of serine residues in Spd2 only because it appears in a minimal consensus sequence. This approach led to studying the effects of a massively mutated protein without regard to the biology of the system.

This statement is based on the premise that we initially identified all the biologically relevant phosphorylation sites in our MS screen but that, when mutating these sites gave only a mild phenotype, we simply mutated many more sites (based only on a minimal S-S/T consensus) to generate a stronger phenotype. We disagree that these experiments were performed “without regard to the biology”. First, it is widely accepted that MS approaches may not identify all of the relevant phosphorylation sites in a protein, and this is demonstrably the case for Spd-2, where 4 independent MS screens (3 from embryos) have identified 41 phosphorylation sites in fly Spd-2, some of which were identified in multiple screens, but many of which were not (as summarised only for the sites that are potential PBD binding motifs in Table 2). Second, we considered a large body of evidence when devising this strategy: (1) The S-S/T(P) motif is widely accepted as a PBD-binding motif and the PBD is widely considered to be essential for targeting Polo/Plk1 to centrosomes^1-7; (2) All of the previously identified sites in worm, human and frog SPD-2/Cep192 that recruit Plk1 to centrosomes conform to this consensus^8-11; (3) The S-S/T to T-S/T substitution is widely accepted to prevent PBD binding^5,6; (4) Ser-to-Thr substitutions are considered to be very conservative in nature, and so unlikely to lead to protein misfolding.

Furthermore, the authors do not provide evidence that the Spd2 mutants result in an inability of Polo to bind Spd2. Perhaps Polo can bind Spd2 through non-canonical sites? That may be the case because the authors present evidence that Polo is still recruited to the centriole and functions properly, as Cnn is phosphorylated and centrioles are still able to disengage properly. Are the Spd2 mutants still being phosphorylated, as in Figure 1? This can be answered with their established method in Figure 1 and by mass spectrometry. Is mutant Spd2 still able to homodimerize and interact with Cnn? The lab has previously established a yeast 2-hybrid assay and has used SEC-MALS, both of which could be used to test this. Spd2 also interacts with Asterless and PLP. Do the Spd CONS or ALL mutants disrupt these interactions, which may explain the mutant phenotypes? Further biochemical analysis of mutant Spd2 should be performed to show that the mutants do not cause protein misfolding.

In the third paragraph the reviewer raises several points. First, they ask for evidence that the Spd-2 mutants cannot bind Polo. As discussed in point 1 of our response to reviewer #1, we now provide compelling evidence that Spd-2 interacts directly with the PBD when phosphorylated, and this interaction is abolished when the putative PBD-binding motifs are mutated (New Figure 1C). Second, the reviewer highlights that the Spd-2 mutant proteins still colocalise with Polo at centrioles, suggesting that the mutant proteins are still recruiting Polo to centrioles. We now clarify that we believe Spd-2 is only required to recruit Polo to the PCM, not to the centrioles. There is strong evidence that Polo can be recruited to centrioles independently of Spd-2 by phosphorylated S-S/T(P) motifs in centriole proteins such as Sas-4¹². Third, the reviewer asks whether the Spd-2 mutant proteins are still phosphorylated. Please see point 4 in our response to reviewer #1.

We have taken two approaches to address this point. First, we tested whether Spd-2 is a homodimer in fly embryos. This is crucial as, if so, a largely misfolded mutant protein might still localise to centrioles and centrosomes if it can still homodimerize with the WT protein. Using Fluorescence Correlation Spectroscopy (FCS), we show that cytoplasmic Spd-2 is monomeric (New Figure 5—figure supplement 1)—as was also recently shown for SPD-2 in worm embryos¹³. Second, with this information in hand, we examined how WT and mutant Spd-2-GFP fusion-proteins are recruited to centrioles in the presence of WT Spd-2-mCherry (which serves as a marker for the WT scaffold and also allows a scaffold to assemble even in the presence of the mutant proteins). The recruitment of the WT and mutant Spd-2-GFP fusions to the WT Spd-2-mCherry scaffold (scored blind) was essentially indistinguishable (New Figure 5). Thus, the mutant Spd-2 proteins can interact with all the proteins that recruit and maintain Spd-2 at centrioles and centrosomes, and this is unlikely to be because they are simply homodimerizing with the WT protein. Finally, we more clearly explain the significance of the data showing that the mutant Spd-2 proteins can rescue the defect in pronuclear fusion in embryos lacking Spd-2 (Figure 2C), but that these embryos still die of mitotic defects during early embryo development. This suggests that the mutant proteins allow centrioles to assemble sufficient PCM to promote the relatively slow process of pronuclear fusion, but not enough PCM to support the very rapid rounds of mitosis that follow. Taken together, these data suggest that the mutant proteins are not misfolded.

Additional concerns:

1) All experiments are performed in Drosophila embryos, whereas the authors claim that the mechanism is universal to the entire fly "in vivo". This should be toned down.

The reviewer points out that all our results are from fly embryos and may not apply to other cell types. We now make this clear throughout the manuscript.

2) The manuscript would benefit from more details of the mitotic defects in Figure 2A.

As requested, we now comment in more detail about the nature of the mitotic defects shown in Figure 2A (now Figure 3A).

3) Are Spd2ALL, Spd2^-/- heterozygotes viable or can reduced dosage explain the phenotypes in Figure 5—figure supplement 1?

The reviewer asks whether Spd2-ALL; Spd-2 ^-/- heterozygotes are viable, or whether reduced dosage can explain the phenotypes shown in Figure 5—figure supplement 1 (now Figure 6—figure supplement 1). These heterozygous flies are indeed viable, but this is to be expected as Spd-2^-/- mutants are also viable (but female sterile). We suspect the reviewer is suggesting that because there is only one copy of Spd-2-ALL in these flies, the phenotypes we observe may be due to lowering the genetic dosage of Spd-2 rather than to the mutations we introduce into the genes. We think this is unlikely for two reasons: (1) In all the experiments where we assess Spd-2-ALL- or Spd-2-CONS-fusion protein function the control is always the WT Spd-2-fusion tested at the same gene dosage; (2) The Spd-2-11A mutant protein is expressed at much lower levels than the WT protein (Figure 1—figure supplement 1A), yet it localises to centrosomes nearly as strongly as the WT protein (Figure 1—figure supplement 1B-D). Thus, the reduction in the centrosomal levels of Spd-2-ALL and Spd-2-CONS is unlikely to simply be due a general reduction in their protein levels. We now discuss this point.

4) The images of centrosomes in Figures 4-6 appear to be S-phase centrosomes with flares, not mitotic centrosomes. The authors reference mitosis and maturation when, in fact, they show images of interphase (S-phase) centrosomes. The authors should show low magnification images of mitotic spindles and then high magnification insets of centrosomes to demonstrate that they are, in fact, examining mitotic centrosomes.

The reviewer questions whether we are examining centrosomes in interphase or mitosis. We apologise for not explaining this properly. In these rapidly developing embryos the nuclei progress through successive S- and M-phases without intervening Gap-phases. As a result, the centrosomes are essentially always either in mitosis, or are preparing to enter mitosis—they are never truly in an “interphase” state where they organise only the tiny amounts of PCM one might observe in a fly somatic cell in interphase. Thus, as soon as mitosis finishes the embryos enter S-phase and the centrosomes start to mature in preparation for the next round of mitosis. We now clarify this important point.

Reviewer #3:

[…] 1) The authors should provide direct evidence about how much Spd-2-CONS and Spd-2-ALL are affects their binding with the Polo PBD compared to WT, which is critical for most of their conclusions. I'm concerned that mutating all 34 Serine sites may affect protein structure and this needs to be ruled out.

Please see point 1 of our response to reviewer #1: we now show that a recombinant WT Spd-2 MBP-fusion protein can bind to the PBD when phosphorylated by Plk1, but mutation of the S-S/T motifs to T-S/T prevents binding (New Figure 1C). The reviewer was also concerned that the multiple Ser-The substitutions we introduce in the mutant proteins may affect protein structure. Please see point 3 in our response to reviewer #2: we now present several lines of evidence that indicate that the mutant proteins are not simply misfolded.

2) Do the authors now if the phosphorylation status is altered during the cell cycle?

This is a good question that we have tried to address by raising phospho-specific antibodies to Spd-2—but so far without success.

3) Figure 4: authors conclude that the inability of mutant proteins to recruit Polo reduces the assembly and/or maintenance of this mixed scaffold, but the data shown can't exclude the possibility that mutant proteins are efficiently incorporated into the scaffold. The author should show a comparison of endogenous Spd-2 scaffold in all three conditions.

We apologise that we were unclear on this, as we do indeed believe that the mutant Spd-2 proteins can incorporate into the scaffold formed by the WT protein. This is entirely consistent with our model: as long as there is some WT Spd-2 in the scaffold, it should recruit Polo, which can then phosphorylate Cnn to support the expanding Spd-2 scaffold; thus, there is no requirement that every Spd-2 molecule in the scaffold be able to recruit Polo. As requested by the reviewer, we now directly address this issue by expressing either the WT or mutant Spd-2-GFP fusions in the presence of WT Spd-2-mCherry (but in the absence of any endogenous unlabelled Spd-2). This allows us to directly compare the ability of the WT and mutant GFP-fusions to incorporate into the expanding scaffold formed by the WT Spd-2-mCherry (New Figure 5). The incorporation of the WT and mutant Spd-2-GFP fusion proteins into the WT Spd-2-mCherry scaffold is essentially indistinguishable—consistent with our model.

At a first glance, this new data may appear at odds with the data shown in Figure 4B of our original submission (now superseded by New Figure 5). In this earlier experiment we showed that in the presence of WT unlabelled Spd-2, the scaffold formed by the mutant-GFP fusion proteins appears somewhat weaker than that formed by the WT-GFP fusion protein—suggesting that the presence of the mutant protein may partially inhibit the assembly of the WT scaffold (which is also consistent with our model, as presumably less Polo will be recruited to the PCM in the presence of the mutant proteins). These two experiments, however, are not comparable: in the old experiment, the scaffold is formed from 2 copies of the endogenous unlabelled Spd-2 gene and two copies of the Spd-2-GFP fusion (Old Figure 4B) while in the new experiment it is formed from one copy of WT-Spd-2-mCherry and one copy of the Spd-2-GFP fusions in the absence of any endogenous (unlabelled) protein (new Figure 5). Thus, there are several possible explanations for this minor discrepancy (for example, unlabelled Spd-2 might more efficiently compete with mutant Spd-2-GFP for incorporation into the scaffold than WT-Spd-2-mCherry, so the ratio of WT/mutant proteins in the scaffolds may be very different).

4) Figure 5: I'm not convinced by the authors' conclusion that the pericentriolar scaffold cannot expand because Spd-2-CON can't recruit Polo. First, they co-localize at the mother centriole, and second, Spd-2-CON itself is affecting scaffold formation in Spd-2 mutant background (Figure 4), so how can they be certain it's a Polo recruitment issue? I would potentially suggest a FRAP experiment to showcase the recruitment of Polo to the scaffold. In addition, apart from co-expressing with Polo-GFP, what's the difference between Figure 5B and Figure 4A (middle lane)? Why is there's a huge difference between the ratios (64/36 vs 9/91)? This was confusing to me.

The reviewer raises two specific reasons for this scepticism. First, the mutant proteins still co-localise with Polo at the centriole (suggesting that they can still recruit Polo). Please see point 2 in our response to reviewer #2where we discuss why we think that Spd-2 is not required to recruit Polo to centrioles. Second, although the Spd-2 mutants effect scaffold assembly we cannot be certain that this is due to a Polo recruitment defect. Please see point 2 of our response to reviewer #1where we describe new experiments that address this problem.

The reviewer suggests that we use FRAP to measure the dynamics of Polo recruitment to the scaffold. This is a good experiment, but we believe it would be hard to interpret. This is because Polo may have different turnover rates at the centriole and in the PCM and the ratio of Polo in these two structures is very different in embryos expressing WT Spd-2-mCherry (where it is mostly in the PCM) and mutant Spd-2-mCherry (where it is mostly in the centriole).

The reviewer asks why the PCM recruitment of the mutant Spd-2-CONS protein is so much worse in embryos expressing Polo-GFP (New Figure 6B) than in embryos not expressing Polo-GFP (New Figure 4). We apologise for not explaining this more clearly. The Polo-GFP line we use here is the healthiest available (an exon-trap insertion into the endogenous Polo gene), but the GFP tag partially disrupts Polo function and this line is only viable in the presence of an untagged copy of Polo. As we now explain more fully, there appears to be a strong genetic interaction between mutant Spd-2 proteins and the partially functional Polo-GFP. Embryos lacking endogenous Spd-2 and expressing Polo-GFP are viable in the presence of WT Spd-2-mCherry, but die very early in the presence of Spd-2-ALL-mCherry or Spd-2-CONS-mCherry (and much earlier than embryos expressing just Spd-2-CONS-mCherry or Spd-2-ALL-mCherry without the Polo-GFP). We do not understand the basis for this genetic interaction, but it suggests an intimate link between Spd-2 and Polo that is somehow perturbed when Spd-2 is mutated in this way and Polo is GFP-tagged.

5) Figure 7: what is the component of the scaffold? If the scaffold is made of Spd-2, when Spd-2 CONS is unable to form the scaffold, you won't be able to see other proteins there no matter how they are recruited. Is there any other marker for the scaffold?

This is correct. However, although we have long proposed that Spd-2, Polo and Cnn cooperate to form the PCM scaffold in flies, we feel that this is not yet an established fact. Hence, we sought to test whether other PCM components might still be recruited to an expanded PCM that might form independently of Spd-2 and Polo. Our data suggest that this is not the case, supporting the view that Spd-2, Polo and Cnn are the major components of the PCM scaffold. The reviewer asks if there is any other marker of the scaffold we can test, but these are the only three markers that we are aware of.

6) Figure 9: The C-terminus of Spd-2 contains most of the conserved Serine sites, but the authors show that mutation N-terminal sites have the biggest effect? What do the authors think the molecular basis for this is?

We apologise for being unclear on this point. The N- and C-terminal halves of Spd-2 that we test here each have 17 S-S/T motifs (see Figure 2A) and we mutated all 17 of these (not just the conserved ones) in the experiments comparing the effect of the mutations in each half of the protein (Figure 9). We have now clarified this point. We do not know why the N-terminal mutations have a larger effect, but suspect that the N-terminal half of Spd-2 normally plays the greater part in recruiting Polo to the PCM in vivo.

[Editors' note: the author responses to the re-review follow.]

Reviewer #1:

[…] Regarding the immunoprecipitation (IP) experiments between the MBP-phospho-Spd2 fragment and GST-PBD (new Figure 1C), the resolution of these immunoblots appear low and the brightness/contrast altered such that it is difficult to gauge the load between the samples. It appears that there is more GST-PBD in the MBP-Spd-2 WT immunoprecipitate, but how much more? Replacing these with higher resolution blots with accompanying quantitation would be appropriate. This is an important experiment/result because it demonstrates that the authors have in fact generated a Spd-2 Polo-binding mutant. Also, is the kinase being used human Plk1, whereas the Spd-2 and PBD proteins are fly? Whatever the species, the fact that the authors used proteins from different organisms should be stated in the Results section.

I also recommend a couple of text changes in this section. The authors state that binding of the 19T mutant was reduced to background levels. How do they know this is background? This could be real binding and, therefore, would remain at WT levels, not a reduced level. The authors end this section stating, "Thus, a fragment of Spd-2 can bind directly to the PBD when phosphorylated, and this binding is prevented when the S-S/T motifs are mutated to T-S/T". Based on the data, it looks to me that binding is not prevented. Instead, they should say that phosphorylation of Spd-2 increases PDB binding which, in my opinion, is what the data shows. The data also suggest that Polo activity primes it's binding to Spd-2 (as they state in the Discussion) but this important finding should also be emphasized here at the end of this Results section.

The reviewer made a number of points about our presentation and description of the immunoprecipitation experiment showing that phosphorylated Spd-2 fragment can bind recombinant GST-PBD in vitro (Figure 1C). These points were all valid and we have made the following changes:

1) We now show the quantification from three repeats of the experiment and state in the Results section that we use recombinant human Plk1 and a fragment of Drosophila Spd-2 in these experiments.

2) We agree that our original description of this new experiment was unclear in several respects. We have now rewritten this section to focus on the conclusion that the MBP-Spd-2 fragment binds more GST-PBD when phosphorylated by Plk1, and that this increase in binding is not seen when the S-S/T sites have been mutated to T-S/T sites.

3) We now emphasise the important finding that Polo activity may prime its own binding to Spd-2 in both the Results and Discussion sections.

Reviewer #2:

[…] The first concern was dealt with by providing in vitro biochemical interaction data (Figure 1C) and the second by performing a new type of quantification of the centrosomal Polo signal in cells expressing a mix of WT and mutant Spd-2, at a time point when Polo signal is just starting to become reduced (Figure 10). This approach allowed the authors to demonstrate reduced Polo signal even in cells where a Spd-2 PCM scaffold is still present. Although the data supports the authors' conclusion that Spd-2 mutants cause a specific loss of PCM associated Polo, the quantification would be more convincing if actual intensities would have been quantified for the Spd-2 scaffold and Polo signals rather than subjective classification into "weak" and "strong". This is because in contrast to a similar quantification in Figure 6, where an "all or nothing" effect is observed, here the relative intensities of PCM scaffold-associated Spd-2 vs Polo are crucial for the result. I would suggest to also adding quantification of intensities to the figure.

My other concerns were also addressed, not in all cases with experiments, but I recognize the technical challenges that have led to this decision.

I would now support publication.

The reviewer requested that we provide some additional quantification of fluorescent intensities for the experiment shown in Figure 10, rather than the more qualitative classification of images into “weak” or “strong” PCM signal. The reason we do not do this (for this experiment and for several others where we use 3D-SIM to assess the localisation of proteins at both the centriole and in the PCM) is that these 3D-SIM images should not really be used for intensity quantification as they are computer-based reconstructions of what the computer “thinks” the original image looks like. Our 3D-SIM experts in the Micron Oxford Advanced Bioimaging Facility tell us that this makes it unreliable to compare absolute intensity levels from such images, particularly if they have been collected from different samples. This is why in all figures like this we show several representative images to cover the range of phenotypes we observe and quantify in our classification of the reconstructed images. As all scoring is performed blind, we are confident that this “semi-quantitative” scoring system provides a reproducible and accurate reflection of the underlying data.

In addition we have made a few minor textual clarifications, corrected a few typos, and inserted an extra reference to a very recently published paper by Cabral et al., 2019, that is relevant to this work.

References:

1. Lee, K. S., Grenfell, T. Z., Yarm, F. R. & Erikson, R. L. Mutation of the polo-box disrupts localization and mitotic functions of the mammalian polo kinase Plk. Proceedings of the National Academy of Sciences 95, 9301–9306 (1998).

2. Song, S., Grenfell, T. Z., Garfield, S., Erikson, R. L. & Lee, K. S. Essential function of the polo box of Cdc5 in subcellular localization and induction of cytokinetic structures. Mol Cell Biol 20, 286–298 (2000).

3. Seong, Y.-S. et al. A spindle checkpoint arrest and a cytokinesis failure by the dominant-negative polo-box domain of Plk1 in U-2 OS cells. J Biol Chem 277, 32282–32293 (2002).

4. Reynolds, N. & Ohkura, H. Polo boxes form a single functional domain that mediates interactions with multiple proteins in fission yeast polo kinase. J Cell Sci 116, 1377–1387 (2003).

5. Elia, A. E. H., Cantley, L. C. & Yaffe, M. B. Proteomic screen finds pSer/pThr-binding domain localizing Plk1 to mitotic substrates. Science 299, 1228–1231 (2003).

6. Elia, A. E. H. et al. The molecular basis for phosphodependent substrate targeting and regulation of Plks by the Polo-box domain. Cell 115, 83–95 (2003).

7. Liu, J., Lewellyn, A. L., Chen, L. G. & Maller, J. L. The polo box is required for multiple functions of Plx1 in mitosis. J Biol Chem 279, 21367–21373 (2004).

8. Decker, M. et al. Limiting amounts of centrosome material set centrosome size in C. elegans embryos. Curr. Biol 21, 1259–1267 (2011).

9. Joukov, V., De Nicolo, A., Rodriguez, A., Walter, J. C. & Livingston, D. M. Centrosomal protein of 192 kDa (Cep192) promotes centrosome-driven spindle assembly by engaging in organelle-specific Aurora A activation. Proc Natl Acad Sci USA 107, 21022–21027 (2010).

10. Joukov, V., Walter, J. C. & De Nicolo, A. The Cep192-organized aurora A-Plk1 cascade is essential for centrosome cycle and bipolar spindle assembly. Mol Cell 55, 578–591 (2014).

11. Meng, L. et al. Bimodal Interaction of Mammalian Polo-Like Kinase 1 and a Centrosomal Scaffold, Cep192, in the Regulation of Bipolar Spindle Formation. Mol Cell Biol 35, 2626–2640 (2015).

12. Novak, Z. A., Wainman, A., Gartenmann, L. & Raff, J. W. Cdk1 Phosphorylates Drosophila Sas-4 to Recruit Polo to Daughter Centrioles and Convert Them to Centrosomes. Dev Cell 37, 545–557 (2016).

13. Wueseke, O. et al. The Caenorhabditis elegans pericentriolar material components SPD-2 and SPD-5 are monomeric in the cytoplasm before incorporation into the PCM matrix. Mol Biol Cell 25, 2984–2992 (2014).

Author details

1. Ines Alvarez-Rodrigo

   The Sir William Dunn School of Pathology, University of Oxford, Oxford, United Kingdom

   Contribution

   Conceptualization, Resources, Formal analysis, Validation, Investigation, Visualization, Writing—original draft, Writing—review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-2181-5535

2. Thomas L Steinacker

   The Sir William Dunn School of Pathology, University of Oxford, Oxford, United Kingdom

   Contribution

   Conceptualization, Resources, Formal analysis, Validation, Investigation, Visualization, Methodology

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-7244-5610

3. Saroj Saurya

   The Sir William Dunn School of Pathology, University of Oxford, Oxford, United Kingdom

   Contribution

   Resources, Investigation

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-4057-0123

4. Paul T Conduit

   The Sir William Dunn School of Pathology, University of Oxford, Oxford, United Kingdom

   Present address

   Department of Zoology, University of Cambridge, Cambridge, United Kingdom

   Contribution

   Formal analysis, Validation, Investigation, Visualization, Methodology

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-7822-1191

5. Janina Baumbach

   The Sir William Dunn School of Pathology, University of Oxford, Oxford, United Kingdom

   Present address

   Structural and Computational Biology Unit, European Molecular Biology Laboratory, Heidelberg, Germany

   Contribution

   Resources, Formal analysis, Validation, Investigation, Visualization

   Competing interests

   No competing interests declared

6. Zsofia A Novak

   The Sir William Dunn School of Pathology, University of Oxford, Oxford, United Kingdom

   Contribution

   Conceptualization, Resources, Formal analysis, Supervision, Methodology, Writing—review and editing

   Competing interests

   No competing interests declared

7. Mustafa G Aydogan

   The Sir William Dunn School of Pathology, University of Oxford, Oxford, United Kingdom

   Contribution

   Formal analysis, Methodology, Writing—review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0003-1673-0596

8. Alan Wainman

   The Sir William Dunn School of Pathology, University of Oxford, Oxford, United Kingdom

   Contribution

   Formal analysis, Supervision, Writing—review and editing

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-6292-4183

9. Jordan W Raff

   The Sir William Dunn School of Pathology, University of Oxford, Oxford, United Kingdom

   Contribution

   Conceptualization, Supervision, Funding acquisition, Visualization, Methodology, Writing—original draft, Project administration, Writing—review and editing

   For correspondence

   jordan.raff@path.ox.ac.uk

   Competing interests

   No competing interests declared

   "This ORCID iD identifies the author of this article:" 0000-0002-4689-1297

• 1,519

  Page views

• 222

  Downloads

• 21

  Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.