The centrosome duplicates during the cell cycle but functions as a single microtubule-organising centre until shortly before mitosis. This raises the question of how centrosome cohesion is maintained throughout interphase. One dynamic model proposes that parental centrioles are held together through centriole-associated, entangling filaments. Central to this model are C-Nap1, a putative centriolar docking protein and rootletin, a fibrous component. Here we identify two novel proteins, Cep68 and Cep215, as required for centrosome cohesion. Similar to rootletin, Cep68 decorates fibres emanating from the proximal ends of centrioles and dissociates from centrosomes during mitosis. Furthermore, Cep68 and rootletin depend both on each other and on C-Nap1 for centriole association. Unlike rootletin, overexpression of Cep68 does not induce extensive fibre formation, but Cep68 is readily recruited to ectopic rootletin fibres. These data suggest that Cep68 cooperates with rootletin and C-Nap1 in centrosome cohesion. By contrast, Cep215 associates with centrosomes throughout the cell cycle and does not appear to interact with Cep68, rootletin or C-Nap1. Instead, our data suggest that Cep215 functionally interacts with pericentrin, suggesting that both proteins influence centrosome cohesion through an indirect mechanism related to cytoskeletal dynamics.
The centrosome is the major microtubule-organising centre (MTOC) of animal cells (Bornens, 2002; Doxsey, 2001; Luders and Stearns, 2007; Nigg, 2004). The single centrosome present in a mammalian G1-phase cell comprises two centrioles surrounded by a fibrous protein matrix, the so-called pericentriolar material (PCM). The PCM is responsible for the nucleation and anchoring of microtubules (MTs) (Bornens, 2002; Job et al., 2003). It comprises several large coiled-coil proteins (Doxsey et al., 2005; Ou et al., 2002) and provides docking sites for γ-tubulin ring complexes (Luders and Stearns, 2007) as well as cell-cycle-regulatory proteins (Doxsey, 2001; Fry and Hames, 2004). The two centrioles present within a G1 centrosome are structurally and functionally distinct. Only the mature centriole carries appendages at its distal end and is competent to function as a basal body in ciliogenesis (Sorokin, 1968). It is also this mature centriole that anchors the bulk of the MTs, whereas the other centriole is relatively free to move (Piel et al., 2000). The average distance between these two centrioles is variable, cell-type dependent and influenced by cytoskeletal dynamics and physiological conditions (Buendia et al., 1990; Euteneuer and Schliwa, 1985; Holy et al., 1997; Jean et al., 1999; Rodionov and Borisy, 1997; Schliwa et al., 1983; Schliwa et al., 1982; Sherline and Mascardo, 1982).
During S phase, a single new centriole forms at an orthogonal angle close to the proximal end of each parental centriole, suggesting that the wall of the parental centriole favours the formation of exactly one progeny centriole (Bettencourt-Dias and Glover, 2007; Nigg, 2007). As a result of this duplication, a G2 cell harbours two centrosomes, each comprising two closely associated centrioles. Remarkably, the attachment between parent and progeny centriole within each centrosome – termed `engagement'– persists throughout G2 and early M phase, until the two centrioles separate at the end of mitosis. This `disengagement' is then thought to constitute a necessary licensing step for a further round of duplication (Tsou and Stearns, 2006a; Tsou and Stearns, 2006b).
At the onset of mitosis, the activation of several protein kinases triggers centrosome separation, leading to the formation of a bipolar spindle (Berdnik and Knoblich, 2002; Blangy et al., 1995; Fry et al., 1998a; Giet et al., 1999; Glover et al., 1995; Hannak et al., 2001; Lane and Nigg, 1996; Sawin and Mitchison, 1995). However, before late G2 phase, the two centrosomes function as a single MTOC. This raises the question of how this `centrosome cohesion' is maintained throughout the major part of interphase. Two prominent models for centrosome cohesion have been proposed. The first model emphasises the role of cytoskeletal dynamics and proposes that the close proximity between parental centrioles is primarily a consequence of forces exerted by the cytoskeleton (Euteneuer and Schliwa, 1985; Jean et al., 1999; Thompson et al., 2004). The second model postulates the existence of dedicated `linker proteins' that function to provide a dynamic connection between parental centrioles. Clearly, the two models are not mutually exclusive. Importantly, the influence of cytoskeletal alterations on centrosome cohesion is not necessarily independent of linker structures, because interference with the cytoskeleton may well affect the transport of structural or regulatory components that are important for the functionality of the putative linker (Meraldi and Nigg, 2001). One major attraction of the `dynamic linker model' is that it readily explains how cell-cycle-dependent phosphorylation can regulate centrosome cohesion (reviewed by Meraldi and Nigg, 2001). The linker model is further supported by electron microscopy (Bornens et al., 1987; Paintrand et al., 1992), and the identification of two specific proteins, C-Nap1 (Fry et al., 1998a; Mayor et al., 2000) and rootletin (Bahe et al., 2005; Yang et al., 2006), that are required for the maintenance of centrosome cohesion. Remarkably, both proteins dissociate from the centrosome when the activity of the protein kinase Nek2 exceeds that of a counteracting type 1 phosphatase, concomitant with centrosome separation at the G2-to-M transition (Bahe et al., 2005; Fry et al., 1998b; Helps et al., 2000; Yang et al., 2006). Additional proteins, notably pericentrin (Jurczyk et al., 2004) and dynamin-2 (Thompson et al., 2004), have also been implicated in centrosome cohesion, but these latter proteins are known to play multiple roles. Depletion of all the above proteins causes centrosome splitting but whether they function through a common mechanism is not presently clear.
We emphasise that the term `centrosome splitting' is used here to describe the separation of parental centrioles. From a mechanistic perspective, this event is expected to be intimately related to centrosome separation, which occurs under physiological conditions at the G2-to-M transition. It is important to distinguish centrosome splitting from centriole disengagement (which is occasionally also called `centriole splitting'), a term that describes the separation of parent and progeny centrioles (within the same centrosome) that normally occurs at the end of mitosis (see Nigg, 2006).
In this study, we used a siRNA screen to search for proteins involved in centrosome cohesion. In addition to confirming the roles of C-Nap1, rootletin and pericentrin, this screen identified only two additional proteins, Cep68 and Cep215, whose depletion triggered substantial centrosome splitting. This indicates that centrosome splitting is not a common consequence of interference with centrosome structure but rather reflects the impairment of specific protein functions. Both Cep68 and Cep215 localised to centrosomes, but differed in their precise localisations during both interphase and mitosis. Our functional and cytological studies indicate that Cep68 contributes to a dynamic linker structure, whereas the role of Cep215 in centrosome cohesion is likely to be more indirect, possibly related to centrosome-MT interactions.
siRNA screen to identify proteins involved in centrosome cohesion
To search for proteins that might play a role in centrosome cohesion, we performed a siRNA screen focused on centrosomal proteins (supplementary material Table S1) and monitored centrosome splitting in U2OS cells. Centrosomes were counted as split whenever the distance between the two parental centrioles was >2 μm. Of 38 proteins tested, depletion of only five proteins resulted in substantial centrosome splitting (supplementary material Fig. S1, Table S1). These include two new proteins Cep68 and Cep215, as well as three proteins previously implicated in centrosome cohesion, namely C-Nap1, rootletin and pericentrin (Bahe et al., 2005; Jurczyk et al., 2004; Mayor et al., 2000; Yang et al., 2006). Centrosome splitting in response to depletion of the above proteins was also observed in two other cell lines, A549 and hTERT-RPE1 (RPE1), although effects were quantitatively reduced in RPE1, presumably because of lower transfection efficiency (data not shown). The above results indicate that splitting is not an inevitable consequence of interference with overall centrosome structure. Yet, the available evidence indicates that centrosome cohesion is affected through multiple mechanisms and its regulation is likely to be complex (Meraldi and Nigg, 2001). In this regard, two observations made during the above screen are noteworthy. First, depletion of ninein caused centrosome splitting only in RPE1 cells, and depletion of Cep164 caused splitting only in serum-deprived RPE1 cells (data not shown). This confirms that the strength of centrosome cohesion differs between cell types and growth conditions. Second, a quantitative analysis of centrosome splitting revealed that, of all proteins examined so far, depletion of C-Nap1 and rootletin produced the most drastic phenotype, both with regard to penetrance – the percentage of cells showing split centrosomes (supplementary material Fig. S1) – and the extent of separation between centrioles (see below).
Characterisation of Cep68 and Cep215
The two proteins newly identified here as being involved in centrosome cohesion, Cep68 and Cep215, were originally described in the course of a proteomic characterisation of the human centrosome (Andersen et al., 2003). Orthologues of both proteins are readily detectable in other mammalian genomes. Cep68 is encoded on human chromosome 2p14 and alternative splicing is thought to give rise to two isoforms comprising 757 and 620 amino acids, respectively (Fig. 1A). Here, these isoforms are distinguished by using the suffix L or S (referring to long and short isoforms, respectively). Both isoforms are predicted to share a globular domain at the C-terminus (Fig. 1A). Cep215 is also known as Cdk5rap2 (Cdk5 regulatory subunit associated protein 2) and has attracted considerable interest because of its implication in the control of brain size. In fact, Cep215/Cdk5rap2 is encoded on human chromosome 9q33.2 and homozygous mutations in this gene have been linked to primary microcephaly (Bond et al., 2005). As in the case of Cep68, alternative splicing is known to produce two isoforms, but these differ only in one relatively small region (Fig. 1B). Sequence analysis using the Pfam database (Finn et al., 2006) predicts several coiled-coil domains as well as an N-terminal `microtubule association region' in both isoforms. The latter region is conserved in myomegalin and Drosophila centrosomin, but its functional significance is not presently known.
Rabbit antibodies were raised against both Cep68 (residues 1-497) and Cep215 (residues 503-1010) expressed in E. coli. In western blots performed on total lysates of HeLaS3, U2OS and 293T cells anti-Cep68 antibodies recognised two bands of ∼90 and 67 kDa, respectively, in good agreement with the predicted sizes of the two Cep68 isoforms (81 and 67 kDa for Cep68L and -S, respectively), whereas pre-immune serum showed no specific bands (Fig. 1C, left panels). Furthermore, ectopic expression of Cep68L in 293T cells enhanced the expected signal, giving confidence that the 90 kDa band represents Cep68L (data not shown). Similarly, anti-Cep215 antibodies recognised a band of the expected size, ∼210 kDa, on western blots of HeLaS3, U2OS and 293T cells, whereas pre-immune serum showed no specific bands (Fig. 1D, left panels). Because the two isoforms of Cep215 are expected to display very similar migration behaviour, these data do not provide information about isoform expression. Anti-Cep68 and anti-Cep215 antibodies also detected their antigens in isolated centrosomes, purified from human KE37 T-lymphoblastoid cells (Fig. 1C,D, central panels). In the case of Cep68, the larger version was predominant, indicating that KE37 cells predominantly express the Cep68L isoform. Finally, antibody specificity was confirmed by siRNA. Transfection of U2OS cells with siRNA oligonucleotides targeting either Cep68 (for 48 hours) or Cep215 (for 72 hours), using two different duplexes to target each protein, resulted in the (near-)complete loss of Cep68 or Ce215, respectively (Fig. 1C,D, right panels).
Endogenous Cep68 and Cep215 could readily be visualised at centrosomes by immunofluorescence microscopy (Fig. 2), confirming and extending previous data (Andersen et al., 2003; Bond et al., 2005). Attesting to staining specificity, the corresponding pre-immune sera did not produce any specific signals and siRNA-mediated depletion of the antigens essentially abolished the staining (Fig. 2). Furthermore, siRNA-mediated depletion of either Cep68 or Cep215 caused centrosome splitting (Fig. 2A,B, compare γ-tubulin staining in bottom panels with upper panels), confirming the results of the siRNA screen and the role of these proteins in centrosome cohesion. Although Cep68 and Cep215 both associate with centrosomes, they clearly display distinct localisations. Cep68 localised to thin fibres protruding away from the two centrioles (Fig. 2A), highly reminiscent of rootletin (Bahe et al., 2005; Yang et al., 2006; Yang et al., 2002). By contrast, Cep215 displayed a rather compact localisation at the centrosome (Fig. 2B). The two proteins also displayed distinct behaviours during cell cycle progression. Whereas Cep68 staining was progressively reduced during prophase, resulting in Cep68 being undetectable on mitotic spindle poles (Fig. 3A), Cep215 clearly persisted on centrosomes throughout mitosis, in agreement with previous data (Bond et al., 2005). We cannot rigorously exclude the fact that the disappearance of Cep68 from mitotic spindle poles might reflect epitope masking, but emphasise that a similar cell-cycle-regulated displacement from mitotic centrosomes has previously been documented for C-Nap1 and rootletin (Bahe et al., 2005; Fry et al., 1998a; Mayor et al., 2002; Mayor et al., 2000; Yang et al., 2006; Yang et al., 2002), two proteins with which Cep68 interacts functionally (see below).
Subcellular localisation of Cep68 and Cep215 at high resolution
To corroborate the above results, we carried out both high-resolution fluorescence microscopy, using a Deltavision deconvolution instrument, and pre-embedding immuno-electron microscopy (immuno-EM). Both approaches confirmed that Cep68 localises to striking fibres originating from centrioles (Fig. 4A; Fig. 4C, upper panel). These fibres were clearly associated with the proximal ends of centrioles, as inferred from EM images revealing the position of appendages (Fig. 4C; upper panel; arrowhead) and/or the presence of nascent pro-centrioles next to Cep68 fibres (arrow). In general, several (most often two to four) Cep68-positive fibres emanated from individual centrioles and the length of some fibres exceeded 0.5 μm. Control experiments (without primary antibody) showed background labelling but no staining of fibres (Fig. 4C, right panels). Clearly, the localisation of Cep68 is remarkably similar to that reported previously for rootletin (Bahe et al., 2005; Yang et al., 2006; Yang et al., 2002). By contrast, Cep215 was associated with the centriolar cylinders, often embedding them and occasionally joining them (Fig. 4B,C, lower panels).
Interdependency of proteins implicated in centrosome cohesion
To explore the functions of Cep68 and Cep215 in centrosome cohesion, we used siRNA to determine to what extent these proteins were required for centrosomal localisation of other proteins implicated in centrosome cohesion. When compared with GL2-treated control cells, where all proteins showed the expected centrosomal localisation (supplementary material Fig. S2), depletion of Cep68 mostly abolished rootletin staining, but did not produce any significant effects on the localisation of Cep215, C-Nap1 or pericentrin (Fig. 5A and Table 1). Depletion of Cep215, on the other hand, exerted at most a marginal influence on any of the other proteins examined (Fig. 5B and Table 1). In reciprocal experiments, we examined the fate of Cep68 and Cep215 in response to depletion of rootletin, C-Nap1 or pericentrin. Depletion of rootletin caused a loss of Cep68 from centrosomes but did not detectably influence either Cep215 or any of the other cohesion proteins [Fig. 6A and Table 1; for C-Nap1 see Bahe et al. (Bahe et al., 2005)]. Following C-Nap1 depletion, Cep68 staining at centrosomes was mostly abolished, whereas the localisation of the other proteins was largely unaltered [Fig. 6B and Table 1; for rootletin see Bahe et al. (Bahe et al., 2005)]. Finally, depletion of pericentrin caused an almost complete loss of Cep215 from centrosomes, a detectable reduction in centrosomal levels of Cep68 and rootletin, but no significant effect on C-Nap1 (Fig. 6C and Table 1). Taken together, these results point to functional (and perhaps molecular) interactions between (1) Cep68 and rootletin and (2) Cep215 and pericentrin.
Next, we performed overexpression experiments to assess the ability of the above proteins to colocalise in vivo. In particular, we exploited the previous observation that overexpression of GFP-rootletin causes the extensive formation of both centrosome-associated and cytoplasmic filaments (Bahe et al., 2005; Yang et al., 2002). When Myc-Cep68 was coexpressed with GFP-rootletin, extensive colocalisation could be seen, whereas Myc-Cep215 showed no significant association with filamentous structures (Fig. 7A). When expressed alone, Myc-Cep68 associated with centrosomes, but did not produce any filaments, even though excess protein could be seen throughout the cytoplasm (Fig. 7B). Interestingly, endogenous Cep68, but not Cep215, was also recruited to conspicuous filaments that formed when Myc-rootletin was expressed ectopically (Fig. 7C). These results demonstrate that although Cep68 is not able to form fibres on its own, it is readily recruited to rootletin fibres. Thus, Cep68, but not Cep215, is able to interact with rootletin in vivo. This suggests that Cep68 cooperates with C-Nap1 and rootletin in forming a dynamic linker structure, whereas Cep215 is likely to affect centrosome cohesion through a distinct, more indirect mechanism.
Here, we have performed a siRNA screen (encompassing 38 proteins) to search for centrosomal proteins that play a role in centrosome cohesion. Under the conditions of this assay, the depletion of only five proteins (including three previously known ones) caused robust centrosome splitting. This is an important finding because it indicates that centrosome splitting is not a common consequence of interfering with centrosome integrity. Instead, our data indicate that only few proteins are critically required for centrosome cohesion, consistent with the existence of specific cohesion structures. Because the extent of centrosome cohesion depends on cell type and physiological conditions (Euteneuer and Schliwa, 1985; Jean et al., 1999; Meraldi and Nigg, 2001), we would expect that additional proteins will be found to affect centrosome cohesion under appropriate circumstances.
The two proteins newly identified here as being required for centrosome cohesion, Cep68 and Cep215, were studied further. Like rootletin, endogenous Cep68 localises to fibres emanating from the proximal ends of parental centrioles. Depletion of Cep68 caused a loss of rootletin from centrioles and vice versa, and depletion of C-Nap1 often caused a loss of both Cep68 and rootletin. Moreover, as described previously for C-Nap1 (Fry et al., 1998a; Mayor et al., 2002) and rootletin (Bahe et al., 2005; Yang et al., 2006; Yang et al., 2002), Cep68 was displaced from centrosomes at the onset of mitosis and absent from mitotic spindle poles, consistent with the notion that a linker structure needs to be dismantled for centrosome separation at the onset of mitosis. In future studies, it will thus be interesting to explore whether, similarly to C-Nap1 and rootletin, Cep68 is also regulated by an antagonism between Nek2 kinase and a type 1 phosphatase (Bahe et al., 2005; Fry et al., 1998b; Helps et al., 2000; Yang et al., 2006).
Overexpression of Cep68 in U2OS cells did not induce the formation of extended fibres, suggesting that Cep68 is not able to form polymers on its own. However, Cep68 was readily recruited to fibrous structures formed by overexpressed rootletin. It is possible that the two proteins co-polymerise, but considering that only rootletin (but not Cep68) comprises coiled-coil domains, it is perhaps more likely that Cep68 stabilises rootletin homopolymers. Taken together, our data strongly indicate that Cep68 cooperates with C-Nap1 and rootletin in the formation of a dynamic linker structure connecting parental centrioles. So far, no conclusive biochemical evidence could be obtained for a molecular interaction between Cep68 and either rootletin or C-Nap1, presumably because of the small amounts of the endogenous proteins and antibody-related limitations. Thus, it will be interesting to examine the expression and localisation of Cep68 in ciliated cells that express higher levels of rootletin and harbour prominent striated fibrous networks known as ciliary rootlets. Moreover, studies involving recombinant proteins will be required to explore whether Cep68 interacts with rootletin directly or indirectly.
In contrast to Cep68, we obtained no evidence for a direct functional interaction between Cep215 and any of the proteins implicated in forming a dedicated linker structure. Cep215 showed no obvious fibre localisation and no mutual dependency with C-Nap1, rootletin or Cep68. Furthermore, Cep215 persisted at the centrosome throughout mitosis, making it unlikely that it is part of a cell-cycle-regulated linker structure. In all these aspects, Cep215 more closely resembles pericentrin, the one other protein implicated in centrosome cohesion in this study as well as previously (Jurczyk et al., 2004). Interestingly, pericentrin depletion caused not only centrosome splitting but also a strong reduction of Cep215 levels at the centrosome, suggesting that displacement of Cep215 contributes to explain this phenotype. How exactly Cep215 and pericentrin contribute to centrosome cohesion remains to be clarified.
Cep215 is of considerable medical interest because mutations in the corresponding gene are linked to autosomal recessive primary microcephaly (MCPH) (Bond et al., 2005). Whether the pathology resulting from mutations in Cep215 relates to the function of this protein in centrosome cohesion will require further study. However, it is intriguing that Cep68 might also potentially be related to human disease. The corresponding gene in fact maps to a locus on chromosome 2p14 that is frequently mutated in retinitis pigmentosa (Kumar et al., 2004). Specifically, haplotype analysis of an Indian family identified Cep68 (KIAA0582) as one of only 14 genes expressed in the eye or retina that map to a critical 1.06 cM region (Kumar et al., 2004). Thus, it may be rewarding to further explore the possibility that Cep68, much like Cep215, is implicated in human disease.
In conclusion, the data reported here support the view that centrosome cohesion is determined through both direct and indirect mechanisms. A prominent direct mechanism is likely to involve entangling filaments and requires the proteins C-Nap1 (Fry et al., 1998a; Mayor et al., 2000), rootletin (Bahe et al., 2005; Yang et al., 2006) and, as shown here, Cep68. All three proteins show not only interdependencies in their localisations but also a common cell-cycle-regulated association with the centrosome (Bahe et al., 2005; Mayor et al., 2000; Yang et al., 2006; Yang et al., 2002) (this study). By contrast, pericentrin and Cep215 appear to function primarily in a distinct context. Both proteins remain associated with spindle poles throughout mitosis and they do not appear to form part of a linker structure. Our observation that Cep215 may function downstream of pericentrin suggests that the two proteins affect centrosome cohesion through a common mechanism. This mechanism remains to be elucidated, but in view of the reported functions of pericentrin (Dictenberg et al., 1998; Jurczyk et al., 2004; Miyoshi et al., 2006; Takahashi et al., 2002), it seems likely to relate to centrosome architecture and positioning, microtubule organisation and/or intracellular transport.
Materials and Methods
Plasmid preparation and recombinant proteins
Polymerase chain reaction was used to amplify full-length human Cep68 from KIAA0582 clone (from Kazusa DNA Research Institute, Kisarazu, Japan). The cDNA was then subcloned into a mammalian expression vector providing a C-terminal Myc-tag. The Cep215 cDNA sequence was obtained from KIAA1633 clone (from Kazusa DNA Research Institute). Using PCR, a frameshift within the sequence was corrected and missing 5′ coding information was obtained by PCR amplification from Marathon cDNA library (Clontech). Constructs were fused to yield the complete coding sequence of Cep215, and subcloned into a mammalian expression vector providing a C-terminal Myc-tag. All constructs were confirmed by sequencing. For expression of recombinant protein fragments, bp 1-1491 of Cep68 and bp 1508-3030 of Cep215 were PCR amplified, inserted into the expression vector pGEX-5X-2 (Stratagene) and confirmed by sequencing. GST-tagged fragments were expressed in E. coli strain BL21(DE3) and purified under denaturing conditions using standard protocols (QIAexpressionist system, Qiagen).
Rabbit anti-Cep68 antibodies (R169 and R170) were raised against an N-terminal fragment (aa 1-497) and anti-Cep215 antibodies (R173 and R174) against an internal fragment (aa 503-1010) (both Charles River Laboratories, Chatillon-sur-Chalaronne, France). Similar results were obtained with both antibodies against the respective proteins. Antibodies R170 (anti-Cep68) and R174 (anti-Cep215) were affinity purified in a two-step process using Affigel (Bio-Rad) according to standard protocols: first GST-protein bound to Affigel-10 was used to remove any anti-GST antibodies, followed by purification of specific antibodies via antigen bound to Affigel-15. R170 and R174 were used for all experiments shown in this study.
Cell culture and transfections
Cells were grown at 37°C under 5% CO2. For U2OS, HeLaS3, A549 and 293T cells, DME medium was used, supplemented with 10% FCS and penicillin-streptomycin (100 IU/ml and 100 μg/ml, respectively). hTERT-RPE1 cells were grown in DMEM nutrient mixture F-12 Ham supplemented with 10% FCS, penicillin/streptomycin, 2 mM glutamine and 0.348% sodium bicarbonate. 293T cells were transfected using the calcium phosphate precipitation method (Krek and Nigg, 1991). U2OS cells were transfected using FUGENE6 reagent (Roche) according to the manufacturer's protocol. Cells were analysed 24–36 hours post transfection.
For pre-embedding immuno-EM of whole cells, U2OS cells, grown on coverslips, were fixed with 4% formaldehyde for 10 minutes, permeabilised with PBS + 0.5 % Triton X-100 for 2 minutes. Blocking and primary antibody incubations were performed as described for immunofluorescence microscopy, followed by incubation with goat anti-rabbit IgG-Nanogold (1:50; Nanoprobes). Nanogold was silver-enhanced with HQ Silver (Nanoprobes). For controls, the primary antibody was omitted. Cells were further processed as described (Fry et al., 1998a).
Immunofluorescence microscopy and immunoblotting
Cells were grown on coverslips, washed once in PBS and fixed in –20°C methanol for 10 minutes. Then, coverslips were washed in PBS and blocked in 1% bovine serum albumin (BSA) in PBS for 30 minutes, before incubation with primary antibodies (diluted in 3% BSA-PBS) for 1 hour at room temperature. After three washes in PBS for 10 minutes each, incubations with secondary antibodies and subsequent washes were done the same way. Coverslips were mounted on slides using glycerol-based mounting medium containing p-phenylenediamine as anti-fading agent. Primary rabbit antibodies were anti-Cep68 affinity-purified IgG (R170, 1 μg/ml) or corresponding pre-immune serum (1:1000), anti-Cep215 affinity purified IgG (R174, 1 μg/ml) or corresponding pre-immune serum (1:1000), anti-rootletin serum (R145, 1:1000), anti-C-Nap1 affinity purified IgG [R63 (Mayor et al., 2000)] and anti-pericentrin (1 μg/ml, ab4448, Abcam). Primary mouse monoclonal antibodies were anti-γ-tubulin (1:1000, GTU-88, Sigma), anti-α-tubulin (1:5000, DM1A, Sigma) and anti-Myc [1:3, hybridoma tissue culture supernatant, 9E10 (Evan et al., 1985)]. Secondary antibodies were Alexa Fluor 488 or 555-conjugated donkey IgGs (1:1000, Molecular Probes) and Cy2-/Cy3-conjugated donkey IgGs (1:1000, Dianova). DNA was stained with 4,6-diamidino-2-phenylindole (DAPI; 2 μg/ml).
Immunofluorescence microscopy was performed using a Zeiss Axioplan II microscope (Carl Zeiss, Jena, Germany) equipped with an Apochromat 63× oil immersion objective, and images were acquired using a Micromax charge coupled device (CCD) camera (model CCD-1300-Y; Princeton Instruments) and MetaView software (Visitron Systems). Alternatively (Fig. 4A,B), a Deltavision microscope on a Olympus IX71 base (Applied Precision) equipped with an Apo 100×/1.35 oil immersion objective and a CoolSnap HQ camera (Photometrics) was used for collecting 0.2 μm distanced optical sections in the z-axis. Images at single focal planes were processed with a deconvolution algorithm (100×: Olympus_100X_140_10103.otf). Settings were `enhanced ratio', with noise filtering set to medium and ten deconvolution cycles. The number of z-stacks collected was variable (between 6 and 14), depending on the height of the individual cell. Images were projected into one picture using softWoRx 3.5.0 (Applied Precision). Exposure times and settings for image processing (deconvolution) were constant for all samples to be compared within any given experiment. Images were opened in Adobe Photoshop CS and then sized and placed in figures using Adobe Illustrator CS2 (Adobe Systems).
For immunoblotting experiments, performed as described previously (Fry et al., 1998a), primary antibodies were used at the following concentrations: rabbit anti-Cep68 affinity-purified IgG (R170, 1 μg/ml), rabbit anti-Cep215 affinity-purified IgG (R174, 1 μg/ml) or corresponding pre-immune sera (1:1000), monoclonal mouse anti-α-tubulin (1:5000, DM1A, Sigma). Secondary antibodies were HRP-conjugated goat anti-rabbit (1:7000, Bio-Rad) or anti-mouse (1:7000, Bio-Rad) IgGs.
Proteins tested in the siRNA screen were depleted using siRNA duplex oligonucleotides targeting the sequences described in supplementary material Table S1. [Most target sequences were chosen by Qiagen based on protein accession numbers: sequences targeting BBS4 (Kim et al., 2004), CAP350 and FOP (Yan et al., 2006), centrin-2 (Salisbury et al., 2002), Cep170 (Guarguaglini et al., 2005), chTOG (Cassimeris and Morabito, 2004), C-Nap1 and rootletin (Bahe et al., 2005), hSAS6 (Leidel et al., 2005), PCM1 and pericentrin (Dammermann and Merdes, 2002) and Plk4 (Habedanck et al., 2005) have been described previously.] For detailed analyses, Cep68 was depleted using siRNA duplex oligonucleotides 262 targeting the coding region 1035-1055, and Cep215 was depleted using siRNA duplex oligonucleotide 283 targeting the coding region 466-486. Qualitatively similar results were obtained using oligonucleotide 263 targeting the coding region 451-471 in the Cep68 gene and oligonucleotide 282 targeting coding region 1306-1326 in the Cep215 gene. An oligoduplex targeting luciferase [GL2 (Elbashir et al., 2001)] was used as a control. RNA oligonucleotides were diluted to 20 μM in 1× oligonucleotide buffer provided by the manufacturer (Qiagen). Cells were methanol-fixed 48 and 72 hours later.
Quantitative real-time PCR (qRT-PCR)
For the analysis of siRNA efficiency, total RNA was extracted from HeLa S3 cells treated for 72 hours with siRNA oligonucleotide duplexes targeting individual centrosomal proteins (supplementary material Table S1). To analyse expression levels of transcripts, total RNA was extracted from HeLa S3 cells at various cell cycle stages using an RNeasy Mini Kit (Qiagen). Then, cDNAs were synthesised from RNA samples using random hexamers and Superscript II reverse transcriptase (Invitrogen) following the manufacturer's instructions. PCR reactions contained cDNA, Power SYBR Green Master Mix (Applied Biosystems) and 300 nM of forward and reverse primers. Primers were designed with Primer Express Software (Applied Biosystems) and amplified fragment corresponded to an exon-exon junction. All primer sequences are available on request. qRT-PCR was carried out in optical 384-well plates and fluorescence was quantified with a Prism 7900 HT sequence detection system (Applied Biosystems). Samples were analysed in triplicate and the raw data consisted of PCR cycle numbers required to reach a fluorescence threshold (Ct). Raw Ct values were obtained using SDS 2.0 (Applied Biosystems). The relative expression level of target genes was normalised according to geNorm (Vandesompele et al., 2002) using EEF1A1 (eukaryotic translation elongation factor a-1) and GusB (β glucuronidase) genes as references to determine the normalisation factor. The thermal profile recommended by Applied Biosystems was used for amplification (50°C for 2 minutes, 95°C for 10 minutes, 40 cycles of 95°C for 15 seconds and 60°C for 1 minute). To verify the specificity of amplification, a melting curve analysis was included according to the thermal profile suggested by the manufacturer (95°C for 15 seconds, 60°C for 15 seconds, and 95°C for 15 seconds). The generated data were analysed with SDS 2.2 software (Applied Biosystems).
We thank T. Mayor for the C-Nap1 antibody, C. Wilkinson for cloning myc-Cep68 and myc-Cep215, R. Malik for help with bioinformatics and E. Nigg for expert technical assistance. This work was supported by the Max-Planck-Society and the `Deutsche Forschungsgemeinschaft' (SFB413).
Supplementary material available online at http://jcs.biologists.org/cgi/content/full/120/24/4321/DC1
- Accepted September 22, 2007.
- © The Company of Biologists Limited 2007