|Home | About | Journals | Submit | Contact Us | Français|
Microtubule behavior changes during the cell cycle and during spindle assembly. However it remains unclear how these changes are regulated and coordinated. We describe a complex that targets the Protein Phosphatase 2A holoenzyme (PP2A) to centrosomes in C. elegans embryos. This complex includes RSA-1, a targeting subunit for PP2A, and RSA-2, a protein that binds and recruits RSA-1 to centrosomes. In contrast to the multiple functions of the PP2A catalytic subunit, RSA-1 and RSA-2 are specifically required for microtubule outgrowth from centrosomes and for spindle assembly. The centrosomally localized RSA-PP2A complex mediates these functions in part by regulating two critical mitotic effectors: the microtubule destabilizer KLP-7 and the C. elegans regulator of spindle assembly TPXL-1. Therefore, by recruiting the PP2A catalytic subunit to centrosomes, the RSA complex regulates a subset of PP2A functions in order to coordinate microtubule outgrowth from centrosomes and microtubule stability in the forming mitotic spindle.
A mitotic spindle is a bipolar, microtubule-based structure that ensures accurate inheritance of genetic material. As cells enter mitosis, the interphase microtubule cytoskeleton is reorganized to form the mitotic spindle. During spindle formation, microtubule dynamics are modulated globally but also locally at spindle poles and in the vicinity of chromatin (Desai and Mitchison, 1997; Karsenti and Vernos, 2001; Wittmann et al., 2001). Numerous studies have shown that the function of many organizers of the microtubule cytoskeleton such as microtubule motors and microtubule-associated proteins (MAPs) is modulated by their phosphorylation state (Verde et al., 1990; Nigg, 2001; Cassimeris, 1999). A key question that remains poorly understood is how protein phosphorylation is regulated temporally and spatially during spindle assembly. Phosphorylation states are determined by the balanced activities of kinases and phosphatases. Mitotic kinases (Nigg, 2001) rely largely on consensus sequence motifs for substrate recognition, however, recent work suggests that kinase activation and targeting can also occur through specific adaptor proteins. Examples include TPX2, which leads to activation of Aurora A kinase and its binding to spindle microtubules (Gruss and Vernos, 2004), and the chromosomal passenger proteins INCENP and Survivin, which localize and activate Aurora B kinase (Carmena and Earnshaw, 2003). Protein phosphatases, which counteract the activity of kinases, are also required for correct microtubule organization. Protein Phosphatase 2A (PP2A), for example, has been implicated in the regulation of microtubule dynamics during the cell cycle (Gliksman et al., 1992; Tournebize et al., 1997), and Protein Phosphatase 4 is required for centrosome maturation and function (Sumiyoshi et al., 2002). The de-phosphorylation of substrates during mitosis is presumably as tightly controlled as the corresponding phosphorylation reactions. However, there are far fewer phosphatases than kinases present in eukaryotic cells and unlike kinases, most phosphatases do not appear to target consensus sequence motifs. This implies a far more complex regulation of individual phosphatases, and regulatory subunits are crucial for the activity as well as the specificity of phosphatases (Faux and Scott, 1996; Dagda et al., 2003; Arnold and Sears, 2006). Protein Phosphatase 2A (PP2A), for instance, forms heterotrimeric complexes consisting of a catalytic C-subunit and a structural A-subunit (the core heterodimer) and a variable regulatory B-subunit. The association of the core heterodimer with different B-subunits can modulate phosphatase activity, localization or substrate specificity (Janssens and Goris, 2001). However, little is known about how regulatory subunits determine phosphatase function during mitosis and how they themselves are regulated.
In C. elegans embryos, centrosomes dominate the spindle assembly process (Hamill et al., 2002; Oegema and Hyman, 2006). Shortly before mitotic spindle assembly, microtubule levels at C. elegans centrosomes increase about five-fold (Hannak et al., 2002). A subset of these microtubules attach to the holocentric kinetochores to form kinetochore microtubules. Genome-wide RNA interference (RNAi) screens and forward genetics have identified a number of centrosome-localized effectors required for spindle assembly (reviewed in Oegema and Hyman, 2006). Microtubule nucleation requires centrosomal core components and a trimeric γ-tubulin complex (Hannak et al., 2002). Plus-end growth of microtubules away from the centrosome depends on free tubulin and the activity of a complex of ZYG-9 (the C. elegans orthologue of the microtubule stabilizer XMAP215) and its activator TAC-1 (Srayko et al., 2005). The number of microtubules that grow out from centrosomes is limited by the activity of the microtubule depolymerizing kinesin KLP-7 (Srayko et al., 2005). Proper length of kinetochore microtubules and therefore the stability of the assembling mitotic spindle depends on the activity of the Aurora A kinase AIR-1 and its activator TPXL-1 (Ozlu et al., 2005). However, it is still unclear how the process of spindle assembly is controlled temporally and spatially and how the different aspects of microtubule dynamics are coordinated to form this complex structure.
Here we identify a C. elegans protein complex that is required for two distinct processes in spindle formation: the outgrowth of microtubules from centrosomes and the stability of kinetochore microtubules. This complex consists of the new RSA-1 and RSA-2 proteins (for “Regulator of Spindle Assembly”) and the core centrosomal protein SPD-5 and constitutes a centrosome-targeting module for Protein Phosphatase 2A.
During a genome-wide RNA interference (RNAi) screen in C. elegans embryos, we identified C25A1.9, an uncharacterized gene whose silencing disrupted mitotic spindle assembly (Sonnichsen et al., 2005). Based on its RNAi phenotype and likely regulatory function (see below) we termed C25A1.9 rsa-1, for regulator of spindle assembly. In rsa-1(RNAi) embryos, centrosomal microtubules were reduced (figure 1A and supplemental movie S1) and centrosomes collapsed onto chromatin after nuclear envelope breakdown (NEBD), and separated again at anaphase (figure 1A, B). We used electron-microscope tomography to investigate the structure of the collapsed spindles more closely. Three-dimensional modeling of tomographic sections from rsa-1(RNAi) embryos with collapsed spindles, showed that microtubules still contacted the chromatin (figure 1D), suggesting that these spindles are otherwise intact and that kinetochore microtubules form but are less stable than in wild-type spindles.
We confirmed the rsa-1(RNAi) phenotype by isolating a mutant in rsa-1 using EMS mutagenesis and TILLING to identify mutants in the gene (supplemental figure 1; for a review on the TILLING method see Stemple, 2004). The mutant allele rsa-1(dd13), predicted to give rise to a C-terminally truncated protein (supplemental figure 1B), caused a phenotype indistinguishable from that of rsa-1(RNAi) embryos (figure 1A and supplemental movie S1). RSA-1 encodes a 404 amino acid protein with sequence similarity to B-type regulatory subunits of Protein Phosphatase 2A (PP2A) (supplemental figure 2). PP2A regulatory B subunits belong to one of at least three groups, the B-, B’- or B”-families (Janssens and Goris, 2001). Based on sequence conservation, the B” subunits can be grouped into two classes: the PR48/PR130 and the TON2 subfamilies (supplemental figure 2). RSA-1 is most closely related to B” subunits of the TON2 subfamily (supplemental figure 2). Consistent with this classification, the Arabidopsis thaliana B” PP2A subunit TON2 has been implicated in aspects of microtubule cytoskeleton organization (Camilleri et al., 2002).
To test biochemically whether RSA-1 associates with a PP2A core heterodimer in C. elegans, we generated a line expressing RSA-1 fused to GFP and immunoprecipitated the protein using anti-GFP antibodies. The gfprsa-1 transgene fully rescued rsa-1(dd13) mutants indicating that it is functional in vivo (see supplemental methods). GFPRSA-1 co-precipitated with the PP2A catalytic and structural subunits LET-92 and PAA-1, indicating that RSA-1 functions as part of a PP2A heterotrimeric complex. In addition to the PP2A core heterodimer, we consistently co-immunoprecipitated the core centrosomal protein SPD-5 and an uncharacterized protein, Y48A6B.11 (table 1). SPD-5 is a coiled-coil protein required for the recruitment of all known components of the pericentriolar material (PCM), and thus is essential for the formation of functional centrosomes (Hamill et al., 2002; Dammermann et al., 2004). Nevertheless, no direct interaction partners of SPD-5 have been described to date. Y48A6B.11 is a coiled-coil-containing 108kDa protein with no obvious homologues in other organisms. Y48A6B.11(RNAi) embryos displayed a phenotype indistinguishable from that of rsa-1(RNAi) embryos (figure 1A, C and supplemental movie S1). Therefore, we refer to this gene as rsa-2.
Consistent with the association of RSA-1 and RSA-2 with SPD-5, antibodies raised against these two proteins labeled centrosomes in wild-type embryos (figure 2A), as did GFP fusions of the proteins (data not shown). The centrosomal staining was strongly reduced after RNAi-knockdown of the respective proteins, confirming the specificity of the antibodies and the RNAi phenotypes (figure 2A). In spd-5(RNAi) embryos, we could not detect any localized intracellular staining for RSA-1 or RSA-2, indicating that both proteins depend on SPD-5 for binding to centrosomes (data not shown).
In order to investigate the assembly relationship between RSA-1 and RSA-2, we determined the location of each protein in the absence of the other. In both rsa-1(RNAi) embryos (figure 2A) and rsa-1(dd13) mutants (data not shown), RSA-2 localized correctly to centrosomes. However, in rsa-2(RNAi) embryos, RSA-1 was not detected on centrosomes (figure 2A). Western blotting showed that RSA-1 levels were decreased by about 50% upon rsa-2(RNAi) and RSA-2 levels were also reduced by about 50% in rsa-1(RNAi) (figure 2C). These results suggested that RSA-2 is specifically required for the centrosomal recruitment of RSA-1 while RSA-1 appears to be dispensable for RSA-2 binding to centrosomes.
A GFP-tagged version of the RSA-1-associated PP2A catalytic subunit LET-92 also localized to centrosomes. GFPPP2AcLET-92 fluorescence was additionally observed in the cytoplasm and around chromatin after NEBD. After depletion of RSA-1, the centrosomal GFPPP2AcLET-92 signal was no longer detectable while the cytoplasmic and nuclear pools persisted (figure 2B and supplemental movie S2). These localization results suggested an assembly hierarchy of SPD-5, RSA-2, RSA-1 and the PP2A catalytic subunit onto centrosomes.
In order to investigate how the RSA proteins facilitate PP2A binding to centrosomes, we performed a Yeast-Two-Hybrid analysis. These experiments suggested that RSA-2 could directly bind to both RSA-1 and the core PCM protein SPD-5 (figure 2D and supplemental table 1). Furthermore, we found that the amino-terminal part of RSA-2 interacted with SPD-5 and its carboxy-terminal half with RSA-1 (not shown), indicating that RSA-2 acts as a scaffold to link the phosphatase complex to centrosomes. These results support the view of a linear assembly pathway that is based on direct protein-protein interactions, whereby the centrosomal core protein SPD-5 links RSA-2 to centrosomes, which then recruits RSA-1, the localizing B regulatory subunit for the PP2A complex (figure 2E).
To more closely examine the functions of RSA-1 and RSA-2, we measured centrosomal GFPβ-tubulin fluorescence in rsa-1(RNAi) and rsa-2(RNAi) embryos. Centrosomal microtubules were reduced to 10–20% of wild-type levels during prophase, and a small nonuniform accumulation of tubulin at centrosomes appeared after NEBD (figure 3A–D, supplemental movie S1). Consistent with the microtubule reduction observed in rsa-1(RNAi) and rsa-2(RNAi) embryos, depletion of these proteins also caused a substantial decrease in outgrowth of microtubule plus ends from centrosomes (supplemental movie S3). The number of microtubules growing out from rsa-1(RNAi) centrosomes at metaphase was previously measured to be approximately 25% of the wild-type number (Srayko et al., 2005).
To determine whether RSA-1- and RSA-2-mediated stabilization of centrosomal microtubules requires the PP2A catalytic subunit LET-92, we examined let-92(RNAi) embryos. In contrast to RSA-1 or RSA-2 depletion, let-92(RNAi) resulted in a highly pleiotropic phenotype, including a failure in meiotic spindle disassembly, a failure in the formation of nuclear envelopes around the pronuclei and unseparated centrosomes. Nevertheless, a spindle-like bipolar structure eventually formed (figure 4A, supplemental movie S4), but with greatly reduced microtubule numbers at the poles (figure 4A, compare let-92(RNAi) panel IV with wild-type anaphase stage embryo in the right panel).
The pleiotropic defects observed in let-92(RNAi) embryos made it difficult to compare the function of this phosphatase with the functions of RSA-1 and RSA-2 in centrosomal microtubule stability. We therefore used the PP2A and Protein Phosphatase 1 (PP1) inhibitor Calyculin A to remove phosphatase activity from embryos that had progressed into mitosis normally. Embryos expressing YFPα-tubulin were mounted in the presence or absence of Calyculin A, which does not penetrate the eggshell. A UV laser was used to perforate the eggshell during mitosis, facilitating the entry of Calyculin A into the cytoplasm. In embryos exposed to Calyculin A, microtubules rapidly disappeared from centrosomes (figure 4B and supplemental movie S5). These observations indicate that phosphatase catalytic activity is required for microtubule stability at spindle poles.
Our results so far suggested that RSA-1 and RSA-2 mediate a specific subset of PP2A functions that are required for maintenance of normal microtubule levels at centrosomes. Microtubule amounts at C. elegans centrosomes are also modulated by KLP-7, the only identified Kinesin-13 in C. elegans (Desai et al., 1999; Siddiqui, 2002). However, in contrast to the RSA complex, which normally stabilizes microtubules, KLP-7 reduces the microtubule number at centrosomes. Specifically, the number of microtubule plus ends growing out from centrosomes is doubled in klp-7(RNAi) embryos (Srayko et al., 2005). Therefore, KLP-7 and the RSA complex have opposite effects on microtubule outgrowth. Indeed co-depletion of KLP-7 and either RSA-1 or RSA-2 restored centrosomal microtubules to wild-type levels. However, spindle poles still collapsed onto each other (figure 5A, B and supplemental movie S6 and supplemental movie S7, for controls see materials and methods). Similar results were obtained when KLP-7 was depleted from rsa-1(dd13) mutant embryos (supplemental figure 3 and supplemental movie S8). Thus, co-depletion of KLP-7 rescues the microtubule outgrowth defect in rsa-1(RNAi), rsa-1(dd13) and rsa-2(RNAi) embryos, but cannot rescue spindle instability. This result indicates that the RSA complex has at least two functions in the formation of a mitotic spindle: regulation of kinetochore microtubule length and regulation of microtubule outgrowth from centrosomes. The above result also suggested that the nucleation capacity of the centrosomes is intact in the absence of RSA-1 and RSA-2. Consistent with this notion, we found that γ-tubulin, the major microtubule nucleator in C. elegans embryos (Hannak et al., 2002) localizes normally in rsa-1(RNAi) and rsa-2(RNAi) embryos (figure 5C and supplemental figure 4A, B). We suggest that the RSA complex acts in a process downstream of microtubule nucleation per se and regulates the outgrowth of nucleated microtubules.
The RSA complex could control microtubule outgrowth from centrosomes directly by regulating KLP-7 catalytic activity, or indirectly, for instance by regulating KLP-7 levels at centrosomes. We tested whether the RSA-PP2A complex affects KLP-7 localization using a transgenic line expressing GFP-tagged KLP-7. KLP-7 levels at centrosomes were about 1.6 to 1.8 fold increased in rsa-1(RNAi) as compared to wild-type (figure 5D, E and supplemental movie S9). This observation indicated that the RSA-PP2A complex controls microtubule outgrowth from centrosomes, at least in part, by restricting the levels of the microtubule destabilizer KLP-7 at centrosomes.
The collapse of spindle poles at NEBD and their subsequent re-elongation observed in rsa-1(RNAi) (figure 1A, B and supplemental movie S1) as well as the ultrastructure of rsa-1(RNAi) spindles obtained by EM tomography (figure 1D) are very reminiscent of the depletion phenotype described earlier for the Aurora kinase activator TPXL-1 (Ozlu et al., 2005) suggesting a role of the RSA complex in regulating TPXL-1.
Consistent with this idea, TPXL-1 amounts at centrosomes were substantially reduced after depletion of RSA-1 (figure 6A). The decrease in TPXL-1 at the centrosome did not result from the reduced microtubule number as it was also observed when microtubule amounts were restored to wild type levels by co-depletion of KLP-7 and RSA-1 (figure 6A, B and supplemental movie S10). Similar results were obtained in rsa-2(RNAi) (data not shown). We next determined whether TPXL-1 and the RSA proteins also interact physically. Indeed, immunoprecipitation of TPXL-1 resulted in co-purification of the RSA complex as judged by Western blot and mass spectrometry (figure 6C and supplemental table 2). Depletion of the RSA complex did not affect the localization of other known centrosomal regulators, (see supplemental figure 4). Based on these results, we conclude that the RSA complex contributes to mitotic spindle assembly at least in part by targeting TPXL-1 to centrosomes, likely via a direct physical interaction.
One of the major questions regarding phosphatase regulation is how target specificity is achieved. Our data suggest that PP2AcLET-92 activity in the C. elegans embryo is regulated by the RSA centrosome-targeting complex, indicating that the RSA complex confers temporal and spatial specificity to the PP2A holoenzyme. This result further suggests that other regulatory subunits could mediate different functions of PP2AcLET-92 in the early embryo. Indeed, out of the potential PP2A subunits of the B, B’ and B” classes identified by BLAST searches (7 proteins in total), the B type subunit SUR-6 has a clear function in early embryonic mitoses (Kao et al., 2004; Sonnichsen et al., 2005; our unpublished observations). sur-6(RNAi) did not alter the targeting of LET-92 or TPXL-1 to centrosomes (our unpublished observations). SUR-6, therefore, appears to mediate LET-92 functions that are separate from those regulated by RSA-1. The Arabidopsis thaliana homologue of RSA-1, TON2, is also required for microtubule organization (Camilleri et al., 2001). This raises the possibility that PP2A-B” subunits of the RSA-1/TON2 subfamily have a conserved role in the regulation of the microtubule cytoskeleton.
Apart from the roles of TON2 and the RSA complex, there is little evidence of specific PP2A complexes regulating microtubule functions in other organisms. A PP2A complex of defined subunit composition was observed to bind microtubules and centrosomes in mammalian tissue culture cells (Sontag et al., 1995). However, the mode of interaction of the respective B-subunit with these structures was not determined and the regulatory significance of this binding remains unclear.
The RSA complex contains a linker protein, RSA-2, which connects the core PCM protein SPD-5 to the PP2A complex via the B” subunit RSA-1. RSA-2 depletion caused no additional defects compared to RSA-1 depletion and therefore, it probably has no additional functions. RSA-2 might modulate the number of phosphatase complexes that bind to centrosomes or facilitate the presentation of substrates to the phosphatase, and thereby adjust microtubule outgrowth to cellular needs. RSA-2 is the first identified target of the essential PCM recruiting protein SPD-5, and this interaction indicates that specific sub-complexes can be recruited by SPD-5, maybe via different domains of the protein. It will be interesting to determine how SPD-5 binds and recruits distinct proteins of the PCM.
Our data indicate that the RSA complex regulates microtubule growth from centrosomes through KLP-7. Recent work from our laboratory suggests that the Kinesin-13 proteins KLP-7 and MCAK limit the number of microtubules growing out from centrosomes, both in C. elegans embryos and in Xenopus egg extracts (Srayko et al., 2005; Kinoshita et al., 2005). As Kinesin-13 proteins are microtubule depolymerases that target microtubule ends, we presume that KLP-7 decreases the stability of nascent nucleated plus ends, although we cannot rule out a role of KLP-7 in nucleation. Possibly, the RSA phosphatase complex regulates the catalytic activity of KLP-7 itself. Experiments in tissue culture cells have shown that phosphorylation decreases the microtubule depolymerizing activity of MCAK, the human homologue of KLP-7 (Andrews et al., 2004; Lan et al., 2004). By analogy, dephosphorylation by PP2A should increase the activity of KLP-7, contradictory to the apparent increase of KLP-7 activity after RNAi knockdown of the RSA complex. We cannot exclude that the RSA complex controls KLP-7 through a different phosphorylation site with an opposite effect on its enzymatic activity. However, our data suggest that the primary role of the RSA complex is the regulation of KLP-7 targeting. Consistent with this idea, a non-phosphorylatable mutant of Xenopus laevis MCAK cannot target to centromeres (Ohi et al., 2004).
The most direct evidence we have found for how the RSA-PP2A complex regulates spindle assembly, is through physical interaction with, and targeting of TPXL-1 to centrosomes (figure 6A–C). Further support for the idea that TPXL-1 and the RSA complex work in the same pathway is that in both tpxl-1(RNAi) and rsa-1(RNAi) embryos, kinetochore microtubules form but are much shorter than in wild type spindles (Ozlu et al., 2005 and figure 1D).
C. elegans TPXL-1, like other microtubule-associated proteins, is highly basic and might thus bind to the acidic tails of tubulin. Phosphorylation is known to negatively regulate the binding of many MAPs to microtubules (Cassimeris and Spittle, 2001). It is conceivable that TPXL-1 is generally phosphorylated in C. elegans embryos and thereby prevented from binding to microtubules. Spatially restricted dephosphorylation at the centrosome by the RSA complex could enable TPXL-1 to bind microtubules and fulfill its microtubule-stabilizing function.
One of the least explored problems in the assembly of complex cytoskeletal structures, like the mitotic spindle, is how the cell coordinates the distinct processes that lead to the formation of this structure. In this context it is interesting that the RSA complex targets the appropriate amounts of two centrosome-effector proteins for correct spindle assembly. RSA-PP2A activity thereby potentially allows coordination of the different microtubule events required to form a spindle. By restricting the amount of KLP-7, the RSA complex ensures that enough microtubules grow from centrosomes to assemble a spindle. Concomitantly, recruitment of TPXL-1 by the RSA complex stabilizes some of these microtubules that are captured by kinetochores. The RSA complex could therefore integrate the requirements for microtubule stability at kinetochores and centrosomes, fine-tuning the number of microtubules contributing to the mitotic spindle accordingly.
All C. elegans strains were maintained as described (Brenner et al., 1974). Transgenic lines were created by microparticle bombardment as described (Praitis et al., 2001). The strains used are listed in supplemental table 3.
The isolation of mutant alleles by TILLING is outlined in the supplemental material section.
dsRNA synthesis was performed as previously described (Oegema et al., 2001) using N2 genomic DNA as template and the primer sequences documented at http://www.worm.mpi-cbg.de/phenobank2/cgi-bin/PrimersPage.py (rsa-1, rsa-2 and let-92 dsRNAs) and in Grill et al. (2001) (klp-7 dsRNA).
The efficiency of depletion of each single protein in the double RNAi experiments was confirmed by dilution of specific dsRNAs with a dsRNA targeting MIG-5, a protein with no function in the one cell stage embryo. Quantification of GFPβ-tubulin fluorescence revealed that rsa-1(RNAi) and rsa-1(RNAi); mig-5(RNAi) caused an identical reduction in centrosomal microtubule levels (not shown). Reduction of the two depleted proteins in the double RNAi experiment was also confirmed by immunostaining. Primers for mig-5 dsRNA production were: AATTAACCCTCACTAAAGGCAGTGGGCCTCAAGCAGT (forward) and TAATACGACTCACTATAGGCTGCCAGAGCATGTGGTG (reverse).
dsRNA was injected into the gonad of L4 stage hermaphrodites and the worms were incubated at 25°C for 22 to 26 hours before being examined.
GFPβ-tubulin, TPXL-1GFP, EBP-2GFP, GFPγ-tubulin and GFPAIR-1 movies were acquired as described in (Ozlu et al., 2005). For EBP-2GFP, imaging was performed as described in (Srayko et al., 2005). Quantification of centrosomal GFPβ-tubulin, GFPγ-tubulin, GFPKLP-7 and TPXLGFP was performed with Metamorph Software. A circular region was used to measure fluorescence intensity at the centrosome and at a nearby cytoplasmic area (“background”). The background subtracted intensity values were plotted over time using Excel.
For recordings of YFPα-tubulin embryos with Calyculin A treatment, microscope and laser setup were used as detailed in (Grill et al., 2003) but Metamorph software controlled the microscope. Worms were dissected in 10µl egg buffer (118mM NaCl, 48mM KCl, 2mM CaCl2, 2mM MgCl2, 25mM HEPES pH 7.3) on a poly-lysine (Sigma) coated microscope slide. Calyculin A (Sigma) from a 1mM stock in DMSO was added to a final concentration of 10µM; for controls, DMSO alone was added to 1% v/v. For each embryo, one control image was acquired at the start of metaphase as determined by DIC optics and entry of the drug was facilitated by targeting the embryonic eggshell with UV laser beam pulses.
Sample preparation for electron tomography was carried out essentially as published (Srayko et al., 2006). Briefly, isolated RNAi embryos were high-pressure frozen (Leica EMPACT2+RTS), freeze-substituted (Leica EM AFS), and thin-layer embedded in Epon/Araldite for serial sectioning. Electron tomography was performed on 300 nm plastic sections with a TECNAI F30 intermediate-voltage microscope (FEI) operated at 300 kV. Tomograms were computed and analyzed by using the IMOD software package as published (O’Toole et al., 2003).
RSA-1 antibodies were generated by immunizing rabbits with the RSA-1 COOH-terminal peptide H2N-AGFLSNSDDYMKYERREQ-COOH. This peptide with an additional N-terminal cysteine residue was used for coupling the peptide to SulfoLink resin (Pierce) for affinity purification of the antiserum. The antibody was directly labeled using Alexa Fluor 488 succinimidyl ester (Molecular Probes) according to the manufacturer’s instructions.
Two different antibodies were generated against RSA-2. αRSA-2(pep), was obtained by immunizing rabbits with the peptide H2N-CQMVLESEIDATVTDV-COOH (aa 933–947 plus aminoterminal cysteine) and affinity purification as described above. αRSA-2(M1522) was obtained by immunizing rabbits with a GST-RSA-2(aa 1–714) fragment and purification of the antiserum with MBP-RSA-2(aa 1–714). Purification of antisera was performed according to standard procedures (Harlow and Lane, 1988).
Immunoprecipitations were essentially performed as previously described (Desai et al., 2003). For RSA-1 IPs, whole worm extracts of transgenic worms expressing LAPRSA-1 were used. LAPRSA-1 was precipitated using an affinity purified antibody raised in goat against 6-His EGFP (Protein expression and purification facility, MPI-CBG). For controls, rabbit random IgG (dianova) was incubated with extracts from LAPRSA-1 worms or anti 6-His EGFP antibodies were incubated with wild-type extracts. Proteins were eluted in sample buffer and separated by SDS PAGE. Mass spectrometric analysis of these samples was performed by nanoLC-MS-MS as described in the supplemental material section.
For RSA-2 and TPXL-1 IPs, either αRSA-2(pep) or αTPXL-1(aa 1–210) (Ozlu et al., 2005) antibodies and 1ml of C. elegans embryo extract were used. Proteins were eluted in sample buffer for Western blot analysis or in 50 mM Tris (pH 8.5), 8 M urea for mass spectrometry. Mass spectrometry on these samples was conducted essentially as described in Cheeseman et al. (2001), except tandem mass spectra were searched against the most recent version of the predicted C. elegans proteins (Wormpep150).
Extract preparation and Western blotting was performed as described (Ozlu et al., 2005). 100% input corresponded to four worms. For the detection of RSA-2, the αRSA-2(M1522) antibody was used (at 0.3 µg/ml). Immunofluoresence experiments were performed as described (Oegema et al., 2001). Antibodies were used at 1µg/ml. The Alexa 488 conjugated αRSA-1 antibody was used for labelling RSA-1. The αRSA-2(pep) antibody was used for detection of RSA-2 and “converted” to a goat antibody by incubation with an anti-goat Fab fragment as described (Hannak, et al. 2002). DM1α (Sigma) was used at a dilution of 1:500 to visualize microtubules. The antibody against γ-tubulin has been described previously (Hannak et al., 2001). Z-stacks through entire embryos were acquired using a wide-field Delta Vision microscope (Applied Precision). The stacks were computationally projected and deconvolved using SoftWorx (Applied Precision).
For yeast two-hybrid assays, the Proquest Two-Hybrid System with Gateway Technology (Invitrogen) was used according to the manufacturer’s instructions. The host strain for the analysis was MAV203 (MATα, leu2–3,112, trp1–901, his3Δ200, ade2–101, gal4Δ, gal80Δ, SPAL10URA3, GAL1lacZ, HIS3UAS Gal1HIS3@LYS2, can1R, cyh2R).
Briefly, a vector encoding RSA-2 fused to the DNA binding domain (BD) of the GAL4 transcription factor was co-transformed with GAL4 activating domain (AD) fusions of either full-length RSA-1 or a SPD-5 fragment that corresponded to amino acids 280–1198. This SPD-5 fragment was obtained in a separate yeast-two-hybrid screen for interactors of RSA-2 (unpublished). Complex formation was detected by activation of the lacZ reporter gene in a β-galactosidase assay and growth on plates lacking either histidine or uracil (not shown). β-galactosidase units were quantified using ONPG as substrate according to the manufacturer’s instruction. Values for β-galactosidase units as well as the vectors used are described in supplemental table 1.
We thank W. Zachariae and all the members of the Hyman Lab, especially L. Pelletier, C. Cowan, N. Ozlu, M. van Breugel and C. Hoege, for stimulating discussions and experimental advice. We are grateful to E. Cuppen (Hubrecht Laboratory, Utrecht, The Netherlands) for advice on the TILLING procedure, to S. Winkler and the MPI-CBG TILLING facility as well as to the MPI-CBG sequencing facility headed by G. Wiebe for excellent technical support in isolating the rsa-1(dd13) allele, to Thomas Döbel for help with modeling of EM tomograms and to B. Habermann for help with bioinformatic analysis. We thank C. Cowan for C. elegans strains, M. Tipsword and M. Ruer for reagents and M. Boxem (Dana-Farber Cancer Institute, Boston, Massachusetts) for sharing data prior to publication. We are grateful to the Caenorhabditis elegans Genetics Center for providing strains. We also thank C. Cowan, N. Ozlu, C. Hoege, S. Schuck and W. Zachariae for comments on the manuscript.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.