Morpholinos for splice modificatio

Morpholinos for splice modification

Advertisement

zyg-11 and cul-2 regulate progression through meiosis II and polarity establishment in C. elegans
Rémi Sonneville, Pierre Gönczy

Summary

The mechanisms that ensure coupling between meiotic cell cycle progression and subsequent developmental events, including specification of embryonic axes, are poorly understood. Here, we establish that zyg-11 and the cullin cul-2 promote the metaphase-to-anaphase transition and M phase exit at meiosis II in Caenorhabditis elegans. Our results indicate that ZYG-11 acts with a CUL-2-based E3 ligase that is essential at meiosis II and that functions redundantly with the anaphase-promoting complex/cyclosome at meiosis I. Our data also indicate that delayed M phase exit in zyg-11(RNAi) embryos is due to accumulation of the B type cyclin CYB-3. We demonstrate that PAR proteins and P granules become polarized in an inverted manner during the meiosis II delay resulting from zyg-11 or cul-2 inactivation, and that zyg-11 and cul-2 can regulate polarity establishment independently of a role in cell cycle progression. Furthermore, we find that microtubules appear dispensable for ectopic polarity during the meiosis II delay in zyg-11(RNAi) embryos, as well as for AP polarity during the first mitotic cell cycle in wild-type embryos. Our findings suggest a model in which a CUL-2-based E3 ligase promotes cell cycle progression and prevents polarity establishment during meiosis II, and in which the centrosome acts as a cue to polarize the embryo along the AP axis after exit from the meiotic cell cycle.

Introduction

The meiotic cell cycle differs from the mitotic cell cycle in several respects. One key difference lies in the fact that the two meiotic divisions occur without an intervening S phase, thus ensuring the generation of haploid gametes. Another distinctive feature is that exit from the meiotic cell cycle is followed by important developmental events such as embryonic axis specification. The mechanisms allowing progression through the meiotic cell cycle and their coupling to subsequent development are incompletely understood.

Progression through mitosis is governed by cyclin-dependent kinases (Cdks), which drive cells into metaphase, and by the anaphase-promoting complex/cyclosome (APC), a multi-subunit E3 ligase that poly-ubiquitinates substrate proteins to target them for destruction by the proteasome (reviewed by Nurse, 2002; Peters, 2002). APC substrates include securins, whose destruction is essential for the metaphase to anaphase transition, and B-type cyclins, whose destruction is essential for mitotic exit. Although the APC is a universal regulator of mitosis, the situation appears to be different during meiosis. In Xenopus laevis, the APC is required for the metaphase to anaphase transition of meiosis II, but not for that of meiosis I (Peter et al., 2001; Taieb et al., 2001). Conversely, in C. elegans, the APC is required at meiosis I but seemingly not at meiosis II (Golden et al., 2000; Shakes et al., 2003; Wallenfang and Seydoux, 2000). It is likely that components other than the APC ensure progression through meiosis I in X. laevis and meiosis II in C. elegans, although their identity is not known.

Progression through the mitotic cell cycle also requires cullin-based E3 ligases, including Skp1-Cul1-F-box-protein complexes (SCF) and Elongin C-Cul2-SOCS box complexes (ECS) (reviewed by Deshaies, 1999). In Saccharomyces cerevisiae, SCF activity is required at the G1 to S transition to target Cdk inhibitors for degradation (Schwob et al., 1994), and plays a role at the G2 to M transition as well (Goh and Surana, 1999). ECS activity is thought to be required for the G1 to S transition, as germ cells homozygous for a cul-2 deletion in C. elegans accumulate in G1 (Feng et al., 1999). Although embryos lacking cul-2 function often fail to extrude a second polar body (Feng et al., 1999), the exact requirement of cullin-based E3 ligases in meiotic cell cycle progression remains to be elucidated.

Establishment of polarity along the anteroposterior (AP) embryonic axis in C. elegans occurs shortly after exit from the meiotic cell cycle. The two female meiotic divisions take place shortly after fertilization, which occurs when the oocyte enters the spermatheca (McCarter et al., 1999). Because the nucleus is located distally in the mature oocyte, oocyte-derived chromosomes and the meiosis I and II spindles are on the distal side of the newly fertilized embryo, whereas sperm-derived chromosomes and centrosomes are on the proximal side. AP polarity is initiated by a sperm component that confers posterior character to the proximal side and becomes apparent during S phase of the first mitotic cell cycle (Cuenca, 2003; Goldstein and Hird, 1996). This results in a polarized distribution of maternally contributed PAR and MEX proteins in the one-cell stage embryo (reviewed by Lyczak et al., 2002; Pellettieri and Seydoux, 2002; Rose and Kemphues, 1998). PAR-3 and PAR-6 localize to the anterior cortex, PAR-2 and PAR-1 localize to the posterior cortex, whereas MEX-5 and MEX-6 are restricted to the anterior cytoplasm. Polarity established by PAR and MEX proteins translates into differential segregation of developmental factors. For instance, P granules and PIE-1, which are both destined to the germ lineage, are segregated to the posterior of the one-cell stage embryo. Polarity also translates into asymmetric elongation of the mitotic spindle, resulting in an unequal first division that generates a larger anterior blastomere and a smaller posterior one.

The prevailing view is that the sperm component acting as polarity cue corresponds to astral microtubules nucleated by paternally contributed centrosomes. Relevant to this view are experiments conducted with conditional mutants in APC components. At the restrictive temperature, such mutants arrest at the metaphase to anaphase transition of meiosis I (Golden et al., 2000; Wallenfang and Seydoux, 2000). Approximately 40% of these embryos establish inverted polarity, with the posterior markers PAR-1, PAR-2 and PIE-1 enriched in the vicinity of the persisting meiotic spindle, and the anterior marker PAR-3 enriched on the opposite side (Wallenfang and Seydoux, 2000). The fraction of embryos with inverted polarity is halved when microtubules are depolymerized with nocodazole, and PIE-1 is no longer segregated when the meiotic spindle is lacking following inactivation of the Cdc2 homologue ncc-1 (Wallenfang and Seydoux, 2000). These observations suggest that microtubules from the persisting meiosis I spindle can act as a surrogate cue sufficient for polarity establishment. It has also been proposed that astral microtubules may be necessary, because polarity is not established in embryos with aberrant sperm asters that nucleate few astral microtubules following the inactivation of air-1, spd-2 or spd-5 (Hamill et al., 2002; O'Connell et al., 2000; Schumacher et al., 1998; Wallenfang and Seydoux, 2000). The APC itself has also been implicated in polarity establishment. Hypomorphic APC mutants that allow meiotic cell cycle completion result in polarity defects, which are also observed following mild inactivation of separase (Rappleye et al., 2002). However, the view that the APC plays a direct role in polarity establishment is controversial, because there is a strict correlation between defects in meiotic cell cycle progression and defects in polarity, suggesting that the latter may be a consequence of the former (Shakes et al., 2003).

zyg-11 is another component that appears to play a role in meiotic cell cycle progression and AP polarity. Analysis of live embryos from the temperature-sensitive allele zyg-11(b2), and of fixed specimens from the non-conditional allele zyg-11(mn40), indicate that zyg-11 is required for progression through meiosis II (Kemphues et al., 1986). The division of zyg-11(mn40) one-cell stage embryos can be asymmetric as in wild type, yielding a larger anterior blastomere and a smaller posterior one, asymmetric but inverted, or symmetric (Kemphues et al., 1986). Moreover, P granules are mislocalized in zyg-11(mn40) embryos, further indicating that zyg-11 is somehow required for proper AP polarity (Kemphues et al., 1986). Although defects are observed at later stages of embryogenesis, they probably reflect an earlier requirement for zyg-11 function, as the temperature-sensitive period of zyg-11(b2) spans the time from fertilization until the end of meiosis II (Kemphues et al., 1986). zyg-11 encodes a 799 amino acid protein that bears at least two leucine-rich repeats and one Armadillo repeat (Carter et al., 1990). There are ZYG-11 homologues in other metazoans, including Drosophila and human, suggesting that this family of proteins performs an evolutionarily conserved function.

Here, we establish that embryos with impaired zyg-11 function have a delayed metaphase to anaphase transition and M phase exit at meiosis II. We find an identical phenotype after inactivating cul-2 and suggest that ZYG-11 acts with a CUL-2-based E3 ligase during meiosis. We provide evidence that the B-type cyclin CYB-3 is a target of this E3 ligase at meiosis II. We uncover that PAR proteins and P granules become localized in an inverted manner during the meiosis II delay following zyg-11 or cul-2 inactivation, and establish that these two components can regulate polarity establishment independently of their role in cell cycle progression. Unexpectedly, we find that microtubules appear dispensable for ectopic polarity in embryos lacking zyg-11 function, as well as for polarity establishment in wild-type embryos. We propose a model in which the centrosome acts as a cue to polarize the embryo along the AP axis following exit from the meiotic cell cycle.

Materials and methods

Nematodes

C. elegans strains of the following genotypes were maintained according to standard procedures (Brenner, 1974): him-1(e879) I; mnC1 dpy-10(e128) unc-52(e444)/zyg-11(mn40) unc-4(e120) II (Kemphues et al., 1986); cul-2(ek1)/unc-64(e246) III (Feng et al., 1999), as well as unc-32(e189) cul-2(t1644) and unc-32(e189) cul-2(t1664), both balanced by qC1 dpy-19(e1259ts) glp-1(q339) (Gönczy et al., 1999); mat-1(ax161) I; him-8(e1489) IV and mat-1(ax144) I; him-8(e1489) IV (raised at 16°C and shifted 8-10 hours to 25°C for restrictive conditions) (Wallenfang and Seydoux, 2000); nDf29/unc-13(e1091) spd-2(oj29) I (O'Connell et al., 2000). Transgenic animals expressing GFP-histone2B (GFP-HIS), GFP-β-tubulin (GFP-TUB), GFP-PAR-2 and GFP-ZIF-1 have been described (DeRenzo et al., 2003; Strome et al., 2001; Wallenfang and Seydoux, 2000).

For generation of GFP-ZYG-11 and GFP-CYB-3 transgenic animals, PCR-derived genomic fragments (sequences of these and other primer pairs available upon request) were subcloned into pSU25, a modified version of the pie-1-gfp vector containing the unc-119 cDNA (a gift from Michael Glotzer). Sequence-verified constructs were bombarded essentially as described (Praitis et al., 2001). For GFP-ZYG-11, two integrated and one non-integrated lines were retained. For GFP-CYB-3, one non-integrated line was recovered; GFP-CYB-3 was present within nuclei of oocytes and embryonic blastomeres throughout the cell cycle, and was released in the cytoplasm at NEBD (data not shown).

zyg-11 and cul-2 mutant alleles

Most of the zyg-11 coding sequence was sequenced following PCR reactions; a C to T alteration at position 483 of the zyg-11 gene was found in zyg-11(mn40) in two independent PCR reactions, predicted to truncate ZYG-11 after amino acid 161.

Two alleles (t1664 and t1644) of a maternal-effect locus mapping close to cul-2 give rise to a phenotype similar to that of cul-2(RNAi) (Gönczy et al., 1999). Sequencing of RT-PCR products and parts of the cul-2 genomic sequence following PCR reactions revealed mutations in the donor site of the third intron (G to A alteration at position 1200 of the gene in t1664) or the fourth intron (G to A alteration at position 1472 of the gene in t1644). Both mutations result in the use of a new donor site (G at position 1152 for t1664, G at position 1496 for t1644) predicted to yield a protein truncation and an internal deletion, respectively.

RNAi

RNAi feeding strains for zyg-11, cul-2 (a gift from Lionel Pintard), cyb-3, elc-1, ncc-1 and tba-2 (a gift from Michael Glotzer) were generated essentially as described (Timmons et al., 2001). The mat-1 feeding strain was from the chromosome I feeding library (Fraser et al., 2000). RNAi was performed by feeding L4 larvae for 24-36 hours essentially as described (Kamath et al., 2001).

Slightly milder conditions were used for zyg-11(RNAi) embryos expressing GFP-HIS, GFP-TUB and GFP-PAR-2 to improve recovery. Although this reduced the frequency of chromosome segregation defects, inverted polarity was nevertheless always established during the meiosis II delay under these conditions (see Fig. 7). Of 25 embryos analyzed in this manner, 6 failed to develop until the two-cell stage and were not considered for analysis.

Fig. 7.

Evolution of polarity in zyg-11(RNAi) embryos. (A-C) Images from time-lapse sequences of zyg-11(RNAi) embryos expressing GFPPAR-2 and GFP-HIS, as well as GFP-TUB, shown during the meoisis II delay (left panel), the first mitotic division (middle panel) and the twoto four-cell stage transition (right panel) (see also movies 8-10 at http://dev.biologists.org/supplemental/). Recordings were started during the meiosis II delay. Arrowheads point to first polar body, arrows indicate condensed sperm DNA; stippled arrowhead and arrows indicate that these positions can be determined only from the movies. Scale bar: 10 μm. (D) Schematic representation of two-cell stage embryos (n=19) corresponding to recordings A (6 embryos), B (10 embryos) or C (3 embryos). See text for details.

The penetrance of the tba-2(RNAi) phenotype was somewhat variable. Those embryos with microtubule remnants were not retained for analysis; occasional gigantic polyploid embryos, which probably resulted from oocyte fusion events, were also not retained. Most reliable phenotypes were obtained after feeding for 28 hours for wild-type and 34 hours for zyg-11(mn40). Shorter feeding resulted in more embryos with partial inactivation, longer feeding in more gigantic embryos. In the absence of a first polar body, zyg-11(mn40) tba-2(RNAi) embryos at meiosis II were distinguished from those at meiosis I by the fact that the former are slightly smaller, being separated from the eggshell, and have a more intact cortex after fixation and staining than meiosis I embryos.

The following 26 genes were inactivated using RNAi by feeding to search for components required for meiosis II. Ten were genes whose reported phenotype was potentially similar to that of zyg-11(RNAi), C47B2.4 (pbs-2), F25H2.9 (pas-5), F33D11.10, F39H11.5 (pbs-7), F48E8.5 (paa-1), K07C11.2 (air-1), T09A5.9, Y45F10A.2 (puf-3), ZK1127.5, ZK520.4 (cul-2); 10 were putative cell cycle regulators, C09G4.3 (cks-1), H31G24.4 (cyb-2.2), K06A5.7a (cdc-25.1), T05A6.1 (cki-1), T05A6.2a (cki-2), T06E6.2 (cyb-3), W02A2.6 (rec-8), Y60A3A.12 (chk-2), ZC168.4 (cyb-1), ZK1307.6 (fzr-1); and six were genes with homologies to zyg-11, C33A12.12, Y9C9A.13, Y39G10AR.5, F44E5.2, F47D12.5, T24C4.6.

Microscopy

Pressure-free and isotonic conditions (Shelton and Bowerman, 1996) were used because embryos shortly after fertilization are fragile. To achieve precise timing of events, embryos were imaged essentially as described (Brauchle et al., 2003), collecting 1 frame every 30 seconds either in utero, starting at fertilization, or ex utero, starting before anaphase I. Duration of events obtained with either method was comparable. For in utero recordings, 4-10 worms were soaked for ∼45 minutes in M9 buffer, supplemented with 0.01% tetramisole and 0.1% Tricaine (Kirby et al., 1990), and mounted under a cover-slip on a 2% agarose pad surrounded by vaseline. For visualizing chromosomes in live embryos expressing GFP-PAR-2 and that are osmosensitive, samples were bathed in Hoechst 33342 for four minutes and then washed for two minutes prior to imaging.

Fixation and indirect immunofluorescence were essentially as described (Gönczy et al., 1999). Primary antibodies were mouse anti-tubulin (DM1A; 1:400; Sigma) and the following antibodies raised in rabbits: anti-PAR-1 (1:2000) (Gönczy et al., 2001), anti-PAR-2 (1:400) (Pichler et al., 2000), anti-PAR-3 (1:60) (Pichler et al., 2000), anti-PGL-1 (1:2000) (Kawasaki et al., 1998), anti-SAS-4 (1:1600) (Leidel and Gönczy, 2003), anti-phosphorylated histone H3 (1:1200; Upstate) and anti-GFP (1:600, a gift from Viesturs Simanis). Slides were counterstained with Hoechst 33258 (Sigma) to reveal DNA. Indirect immunofluorescence was imaged on an LSM510 Zeiss confocal microscope; quantification of GFP-CYB-3, GFP-ZYG-11 and GFP-ZIF-1 signals was performed using a 12-bit CCD Camera and Metamorph software (Universal Imaging). Images were processed with Adobe Photoshop.

Results

zyg-11 is required for timely metaphase to anaphase transition and M phase exit at meiosis II

To analyze the consequences of zyg-11 inactivation with high temporal and spatial resolution, we performed dual differential interference contrast (DIC) and fluorescence time-lapse microscopy using transgenic animals expressing GFP-HIS (histone 2B). In wild type (Fig. 1, Fig. 2A-C, Table 1; see Table S1 and Movies 1 and 2 at http://dev.biologists.org/supplemental/), the meiosis I spindle assembles shortly after fertilization. Following the metaphase to anaphase transition of meiosis I, homologous chromosomes segregate during anaphase I and extrusion of the first polar body ensues. After a brief time interval that we refer to as meiotic interphase, the meiosis II spindle assembles. Following the metaphase to anaphase transition of meiosis II, sister chromatids segregate during anaphase II, after which the second polar body is extruded. Embryos then exit the meiotic cell cycle, oocyte- and sperm-derived chromosomes decondense as pronuclei form and the first mitotic cell cycle ensues.

Fig. 1.

zyg-11, cul-2 and cyb-3 are required for progression through meiosis II. (A) Images from time-lapse sequences of embryos expressing GFP-HIS (see also Movies 1-7 at http://dev.biologists.org/supplemental/). Metaphase views are: polar (P), the spindle is viewed from the pole and five homologues (meiosis I) or sister chromatids (meiosis II) are observed in a circle around the sixth; lateral (L), the spindle is viewed from the side and up to three pairs of homologues (meiosis I) or sister chromatids (meiosis II) are observed; or intermediate (I), between the polar and lateral views. In this and other figures, embryos are oriented with polar bodies to the left and all panels of a kind are at the same magnification. Scale bar: 5 μm. (B) Timing of events in embryos from fertilization (t=0) until the end of meiosis II. Average duration of each stage, defined in Table 1, is indicated. Letters in the time-lines indicate the time at which the corresponding images in (A) are taken. Vertical white lines indicate the time at which GFP-PAR-2 first becomes enriched at the cell cortex in the vicinity of oocyte-derived chromosomes in zyg-11(RNAi) (n=4) or zyg-11(mn40) (n=3) embryos imaged starting at the end of meiosis I.

Fig. 2.

zyg-11 is required for timely metaphase to anaphase transition and M phase exit at meiosis II. (A-N) Lateral views of wild-type or zyg-11(RNAi) embryos at the indicated stages, stained with antibodies against α-tubulin (A-C,G-J), or α-tubulin and phosphorylated histone H3 (D-F,K-N). Insets below panels A-C and G-J, as well as the entire panels D-F and K-N are magnified views of a meiotic spindle and have a width of ∼7 μm. Insets below panels A-C and G-J showα -tubulin (green) and DNA (blue) on the left, and DNA alone on the right. Panels D-F and K-N show α-tubulin (green), phosphorylated histone H3 (red) and DNA (blue) on the left, and phosphorylated histone H3 alone on the right. To view polar bodies and sperm chromosomes, the DNA signal in the low magnification images is a projection of several 1-μm confocal optical sections. Arrowheads point to the first polar body, arrows indicate condensed sperm DNA. Scale bar: 10 μm. Note that the focus of phosphorylated H3 lies between homologues at metaphase I and sister chromatids at metaphase II in both wild-type and zyg-11(RNAi) embryos. (O,P) Wild-type (O) and zyg-11(RNAi) (P) anaphase II embryos expressing GFP-CYB-3 stained with antibodies against α-tubulin (not shown) and GFP (red); DNA is shown in blue. (Q) Quantification of signal intensities in wild-type (n=6), zyg-11(RNAi) (n=14) and cul-2(RNAi) (n=19) embryos in metaphase II or anaphase II stained as in (O); embryos were staged using the anti-α-tubulin and DNA signals. The difference between wild type and zyg-11(RNAi) or cul-2(RNAi) is statistically significant (Student's t-test: P=6×10-5 and P=4×10-6, respectively).

View this table:
Table 1.

Duration of events during meiosis in wild type, and following zyg-11, cul-2 or cyb-3 inactivation

We analyzed zyg-11(RNAi) embryos and zyg-11(mn40) mutant embryos, which are likely to be null for zyg-11 function (see Materials and methods). In both cases (Fig. 1, Fig. 2G-J, Table 1; Table S1 and Movie 3 at http://dev.biologists.org/supplemental/), the timing of events is indistinguishable from wild type prior to the metaphase of meiosis II. However, the metaphase to anaphase transition of meiosis II is significantly delayed, with embryos exhibiting metaphase-like figures for ∼30 minutes, as compared with ∼6 minutes in wild type. During this delay, chromosomes lose their tight arrangement on the metaphase plate and the spindle becomes less organized (Fig. 2I). Thereafter, anaphase-like figures are observed, although chromosome bridges are often apparent and sister chromatids are not segregated properly (Fig. 2J). Moreover, anaphase II is lengthened. Chromosomes display the M phase-specific phosphorylated histone H3 (Hsu et al., 2000) throughout the meiosis II delay in zyg-11(RNAi) embryos, including when anaphase-like configurations are apparent, indicating that it corresponds to a prolonged M phase (Fig. 2L-N). Eventually, oocyte and sperm-derived chromosomes decondense, pronuclei form and the first mitotic cell cycle ensues. We conclude that zyg-11 is required for timely metaphase to anaphase transition and M phase exit at meiosis II.

cyb-3 and cul-2 are also required for meiosis II progression

As the sequence of ZYG-11 does not offer obvious clues as to its biochemical activity, we sought to identify other genes whose inactivation results in delayed progression through meiosis II to gain insight into the mechanism of action of ZYG-11. We examined the results of RNAi-based functional genomic screens available as DIC movies through www.wormbase.org. Out of ∼5500 genes considered, we found 10 whose reported phenotype resembles that of zyg-11(RNAi). We also selected 10 putative cell cycle regulators and six ORFs that bear homology with zyg-11 for further analysis (see list of genes in Materials and methods). We inactivated these 26 genes using RNAi and analyzed meiotic cell cycle progression using time-lapse DIC microscopy. We found two genes whose inactivation does not affect meiosis I but results in delayed progression through metaphase of meiosis II: the B type cyclin cyb-3 and the cullin cul-2.

In cyb-3(RNAi) embryos (Fig. 1, Table 1; Table S1 and Movie 4 at http://dev.biologists.org/supplemental/), the delay is already apparent prior to metaphase, earlier than in the absence of zyg-11 function. Another difference is that anaphase II is not delayed in cyb-3(RNAi) embryos. By contrast, we found that cul-2(RNAi) embryos (Fig. 1, Table 1; Table S1 and Movie 5 at http://dev.biologists.org/supplemental/) have a meiotic phenotype indistinguishable from that of zyg-11(RNAi) or zyg-11(mn40) embryos. In particular, the metaphase to anaphase transition of meiosis II is significantly delayed, with embryos exhibiting metaphase-like figures for∼ 33 minutes. Moreover, anaphase II is lengthened and often aberrant. We found a similar phenotype in the deletion allele cul-2(ek1), and in two maternal-effect embryonic lethal alleles of cul-2 (Table 1; see Materials and methods). Importantly, inactivation of cul-2 by RNAi in zyg-11(mn40) mutant embryos does not result in a more prolonged meiosis II delay (Fig. 1, Table 1; see Movie 6 at http://dev.biologists.org/supplemental/), compatible with the notion that zyg-11 and cul-2 act in a common process. A phenotype similar to that observed by inactivating zyg-11 or cul-2 has also been obtained by inactivating Elongin C (elc-1) and Rbx (rbx-1), two other core components of ECS E3 ligases (Liu et al., 2004). Taken together, these results establish that a CUL-2-containing E3 ligase is essential for timely metaphase to anaphase transition and M phase exit at meiosis II, and suggest that ZYG-11 somehow acts in concert with this ECS.

CYB-3 accumulates following zyg-11 inactivation, delaying M phase exit

Given that cyb-3 acts at meiosis II, and that degradation of B type cyclins is generally required for M phase exit, we investigated whether persistence of CYB-3 may be causing delayed M phase exit when the CUL-2-based E3 ligase is inactivated. We generated a fusion protein between GFP and CYB-3; in wild type, we found levels of GFP-CYB-3 to be high until metaphase I, and significantly lower after the metaphase to anaphase transition of meiosis I (data not shown). Similar low levels were observed during meiosis II (Fig. 2O), although these low levels precluded assessing whether levels drop further after the metaphase to anaphase transition of meiosis II. Importantly, we found that GFP-CYB-3 levels during the meoisis II delay in zyg-11(RNAi) and cul-2(RNAi) embryos were high compared with wild-type meiosis II embryos (Fig. 2P,Q). To test whether CYB-3 accumulation was causing delayed M phase exit in the absence of zyg-11 function, we examined zyg-11(mn40) cyb-3(RNAi) embryos. Strikingly, whereas the metaphase to anaphase transition of meiosis II is still delayed in such embryos, the anaphase II delay is abolished (Fig. 1, Table 1; see Movie 7 at http://dev.biologists.org/supplemental/). Therefore, the persistence of CYB-3 appears to be responsible for delayed M phase exit following zyg-11 inactivation.

ZYG-11 may be a substrate recruitment subunit of a meiosis-specific CUL-2-containing ECS

ECS E3 ligases contain a SOCS-box-containing protein that bridges Elongin C and the substrate (Kile et al., 2002). We considered whether ZYG-11 could be such a substrate recruitment subunit, even though it does not contain an obvious SOCS-box. Substrate recruitment subunits of the ECS and other cullin-based E3 ligases are unstable and are targeted for degradation by the very same complex of which they are a part (DeRenzo et al., 2003; Galan and Peter, 1999; Geyer et al., 2003; Pintard et al., 2003b; Wirbelauer et al., 2000; Zhou and Howley, 1998). To test whether ZYG-11 behaves in this manner, we generated a transgenic line expressing GFP-ZYG-11 and compared GFP levels in wild-type and cul-2(RNAi) animals. Although wild-type animals express the fusion protein at very low levels in the germ-line and early embryos (Fig. 3A,D), GFP-ZYG-11 levels are dramatically increased when cul-2 is inactivated (Fig. 3B,E,G). As anticipated, analogous results were obtained after elc-1 inactivation (Fig. 3G). Elevated levels of GFP-ZYG-11 do not result merely from slower progression through meiosis II, because GFP-ZYG-11 is not increased in cyb-3(RNAi) animals (Fig. 3C,F,G). These observations are compatible with ZYG-11 being a substrate recruitment subunit of a CUL-2 containing ECS ligase.

Fig. 3.

cul-2 destabilizes GFP-ZYG-11; zyg-11 does not regulate GFP-ZIF-1. Images of paralyzed animals (A-C,H-J) or of fixed embryos stained with anti-GFP antibodies (D-F,K-M) from strains expressing GFP-ZYG-11 (A-F) or GFPZIF-1 (H-M). Plain outlines indicate oocytes, stippled outlines embryos. Note also the gut autofluorescence outside of embryos in live specimens. Scale bars: 50 μm. (G,N) Quantification of GFPZYG-11 (G) and GFP-ZIF-1 (N) signal intensities in fixed early embryos (i.e. <10 cell stage) of the indicated genotypes stained as in (D-F,K-M). Increase of GFP-ZYG-11 in cul-2(RNAi) is 16.4±3.0-fold, of GFP-ZYG-11 in elc-1(RNAi) is 17.3±3.2-fold, and of GFPZIF-1 in cul-2(RNAi) is 5.8±3.3-fold.

We next addressed whether zyg-11 is required for the activity of all ECS complexes. ZIF-1 is the SOCS-box protein of an ECS that targets germ plasm proteins for degradation in somatic lineages in early embryos (DeRenzo et al., 2003). In cul-2(RNAi) animals, levels of GFP-ZIF-1 are increased compared with otherwise wild-type embryos (Fig. 3H,J,K,M,N) (DeRenzo et al., 2003). By contrast, we found that levels of GFP-ZIF-1 are not altered in zyg-11(RNAi) animals (Fig. 3I,L,N). We conclude that zyg-11 is not required for the function of all CUL-2-containing E3 ligases but plays a more restricted role during meiosis.

zyg-11 and cul-2 function redundantly with APC during meiosis I

In addition to an essential requirement at meiosis II, we uncovered a non-essential function for zyg-11 and cul-2 at meiosis I. The mat-1 locus encodes the APC component Cdc27/APC3 and embryos from the conditional allele mat-1(ax161) raised at 25°C arrest at the metaphase to anaphase transition of meiosis I (Fig. 4A-C, Table 2) (Golden et al., 2000; Wallenfang and Seydoux, 2000). Such arrested embryos fall into three categories on the basis of their position with respect to the spermatheca in the animal (data not shown), and the appearance of chromosomes and microtubules (Fig. 4). First, embryos closest to the spermatheca and thus the youngest, which have a canonical metaphase I arrest configuration (Fig. 4A, metaphase I). Second, embryos located further away from the spermatheca, which have chromosomes with a looser arrangement and a less organized spindle (Fig. 4B, metaphase I late). Third, embryos located further still from the spermatheca and thus the oldest, which have elongated chromosomes located in the embryo center and a completely disassembled spindle (Fig. 4C, metaphase I very late).

Fig. 4.

zyg-11 and cul-2 function redundantly with APC at meiosis I. Meiotic figures from embryos of the indicated genotypes stained with antibodies against α-tubulin (green in merge) are shown; DNA is shown in blue in merged image and alone on the right. Width of panel represents ∼7μ m. NA, not available.

View this table:
Table 2.

zyg-11 and cul-2 act redundantly with APC at meiosis I

Although 15°C mat-1(ax161) embryos are indistinguishable from wild type (Fig. 4G,K,L; Table 2) (Golden et al., 2000; Wallenfang and Seydoux, 2000), we found that they exhibit meiosis I defects when, in addition, either zyg-11 or cul-2 are inactivated, accumulating with chromosomes and spindle configurations resembling those of 25°C mat-1(ax161) embryos (Fig. 4H-J, compare with Fig. 4A-C; Table 2). Although some 15°C mat-1(ax161) zyg-11(RNAi) and 15°C mat-1(ax161) cul-2(RNAi) embryos bypass the meiosis I block, the first polar body is small or absent (data not shown) and metaphase II figures invariably exhibit supernumerary chromosomes, which is indicative of defective anaphase I (Fig. 4M,N). Similar results were obtained using a distinct conditional mat-1 allele, mat-1(ax144) (Table 2). These findings indicate that zyg-11 and cul-2 also act at meiosis I, although this role is revealed only when APC function is slightly compromised. Because zyg-11 behaves like cul-2 in enhancing the mat-1 phenotype, these findings also reinforce the notion that ZYG-11 and CUL-2 function in a common process.

Interestingly, we noted that a sizeable fraction of 15°C mat-1(ax161) zyg-11(RNAi) and 15°C mat-1(ax161) cul-2(RNAi) metaphase II embryos have elongated chromosomes located in the center of the embryo and a completely disassembled spindle (Fig. 4O, Table 2), resembling the last category of 25°C mat-1(ax161) embryos (see Fig. 4C). Such configurations are never observed when zyg-11 and cul-2 are inactivated on their own (Fig. 2). Therefore, it appears that some embryos lacking zyg-11 or cul-2 function cannot exit meiosis II if APC function is compromised. Although this may be a consequence of the meiosis I defects, it could also reflect a requirement for APC at meiosis II.

Inverted polarity is established during the meiosis II delay in the absence of zyg-11 or cul-2 function

Next, we investigated the cause of the alterations in cleavage pattern and P granule distribution that have been reported in zyg-11(mn40) embryos (Kemphues et al., 1986) by analyzing the distribution of polarity markers. In wild-type meiosis II, PAR-1, PAR-2 and PAR-3 proteins are not polarized, and P-granules are present throughout the cytoplasm (Fig. 5A-D, Table 3). By contrast, we found that ∼50% of zyg-11(RNAi) meiosis II embryos exhibit polarized distribution of PAR proteins and P-granules (Fig. 5E-H, Table 3), although enrichment of all four markers is typically less pronounced than for wild-type embryos during the first mitotic cell cycle (Fig. 5Q-T). Strikingly, in all zyg-11(RNAi) meiosis II embryos with polarized distribution, PAR-1, PAR-2 and P-granules are enriched in the vicinity of the meiotic spindle, whereas PAR-3 is enriched on the opposite side. These distributions are inverted compared with wild-type embryos during the first mitotic cell cycle. We next examined cul-2(RNAi) meiosis II embryos and found similar alterations in the distribution of PAR-1, PAR-2, PAR-3 and P granules (Fig. 5I-L, Table 3). We conclude that inverted polarity is established during meiosis II in the absence of zyg-11 or cul-2 function.

Fig. 5.

Inverted AP polarity during meiosis II delay in the absence of zyg-11 or cul-2 function. Wild-type (A-D), zyg-11(RNAi) (E-H), cul-2(RNAi) (I-L), and cyb-3(RNAi) (M-P) embryos during meiosis II, and wild-type embryos during the first mitotic cell cycle (Q-T), stained with antibodies against α-tubulin (green in merged image) and a polarity marker as indicated (PAR-1, PAR-2, PAR-3 or PAR-1+PGL-1 to detect P granules; red in merged image and viewed alone in the top panel); DNA is shown in blue in merged image. Scale bar: 10 μm. Note that enrichment of PAR proteins and P granules in the absence of zyg-11 or cul-2 function is somewhat variable during meiosis II, and is less pronounced, both in intensity and in lateral extent, than in wild-type embryos during the first mitotic cell cycle.

View this table:
Table 3.

Distribution of polarity markers following zyg-11, cul-2 or cyb-3 inactivation

zyg-11 and cul-2 prevent polarity establishment independently of promoting meiosis II cell cycle progression

Inverted polarity could be a consequence of prolonged meiosis II or reflect a more direct requirement of zyg-11 and cul-2 to negatively regulate polarity establishment. To distinguish between these possibilities, we first examined the distribution of polarity markers in cyb-3(RNAi) embryos, which also exhibit delayed progression through meiosis II (Fig. 1 and Table S1). Importantly, we found that PAR-1, PAR-2, PAR-3 and P granules are not polarized in cyb-3(RNAi) meiosis II embryos (Fig. 5M-5P), indicating that the inverted polarity established in zyg-11(RNAi) or cul-2(RNAi) embryos is unlikely to result merely from prolonged meiosis II.

Because the meiosis II delay in cyb-3(RNAi) embryos is slightly less pronounced and of a different nature than that observed in zyg-11(RNAi) or cul-2(RNAi) embryos, we conducted two experiments to test whether zyg-11 and cul-2 can regulate polarity establishment independently of their requirement for cell cycle progression during meiosis II. First, we made use of 25°C mat-1(ax161) embryos, which are arrested in meiosis I with evenly distributed P granules (Fig. 6A,B, Table 4) (Wallenfang and Seydoux, 2000). Because P granules are polarized in an inverted manner in∼ 50% of embryos lacking zyg-11 or cul-2 function (Table 3), we performed a molecular epistasis experiment using 25°C mat-1(ax161) embryos subjected to zyg-11(RNAi) or cul-2(RNAi). Interestingly, we found that ∼30% of 25°C mat-1(ax161) embryos lacking, in addition, either zyg-11 or cul-2 function, exhibit an enrichment of P granules in the vicinity of the meiotic spindle (Fig. 6D, Table 4). Because the CUL-2-based E3 ligase plays a non-essential role at meiosis I, we considered whether this may reflect a more robust cell cycle arrest in such embryos. However, we found that the fraction of embryos in each of the three meiosis I arrest categories (Fig. 4) is the same in 25°C mat-1(ax161) embryos irrespective of whether they have been subjected to zyg-11(RNAi) or cul-2(RNAi) (Table 2), indicating that cell cycle progression is similarly affected.

Fig. 6.

zyg-11 can prevent polarity establishment independently of promoting meiosis II cell cycle progression. (A-D) 25°C mat-1(ax161) or 25°C mat-1(ax161) zyg-11(RNAi) embryos stained with antibodies againstα -tubulin (green) and PGL-1 to detect P granules (red); DNA is shown in blue. Depicted are embryos from the metaphase I (A,C) and the metaphase I (late) (B,D) categories (see Fig. 4). Insets show magnified views of meiotic chromosomes. Scale bar: 10 μm; width of inset represents ∼7 μm. (E,F) Live ncc-1(RNAi) [heterozygous for zyg-11(mn40)] or ncc-1(RNAi) zyg-11(mn40) embryos during meiosis expressing GFP-PAR-2 (green) and stained with Hoechst 33342 to view DNA (blue in merged image). Insets show magnified views of meiotic chromosomes; width of inset represents∼ 8 μm. Left panels show a DIC image of the same embryo. (G-J) Live mat-1(RNAi) [heterozygous for zyg-11(mn40)] or mat-1(RNAi) zyg-11(mn40) embryos expressing GFP-PAR-2 (green) and stained with Hoechst 33342 to view DNA (blue in merged image). Insets show magnified views of meiotic chromosomes; width of inset represents ∼8μ m. Scale bar: 10 μm.

View this table:
Table 4.

zyg-11 and cul-2 can prevent polarity establishment independently of promoting meiosis II cell cycle progression

In a second experiment, we made use of ncc-1(RNAi) embryos. ncc-1 encodes a Cdc2 kinase acting during meiosis; in its absence, chromosomes do not congress to form a metaphase plate and both meiotic divisions fail (Boxem et al., 1999). Nevertheless, the time interval between fertilization and the appearance of pronuclei is unchanged as the meiotic divisions are bypassed (Boxem et al., 1999). We found the average number of embryos in meiosis, as judged by examining chromosomes with GFP-HIS, to be 1.5/gonad for ncc-1(RNAi) (n=17 gonads), 3.0/gonad for zyg-11(mn40) (n=23 gonads) and 1.6/gonad for ncc-1(RNAi) zyg-11(mn40) (n=25 gonads). This indicates that the time separating fertilization from pronuclear appearance is not changed in zyg-11(mn40) ncc-1(RNAi) embryos compared with ncc-1(RNAi) embryos, enabling us to test the requirement of zyg-11 in polarity establishment independently of that in meiotic cell cycle progression. As shown in Fig. 6F, we observed cortical enrichment of GFP-PAR-2 in the vicinity of oocyte chromosomes in ∼60% of zyg-11(mn40) ncc-1(RNAi) embryos prior to pronuclear appearance (n=29), although to lower levels than in zyg-11(mn40) embryos (data not shown). Taken together, our findings indicate that zyg-11 and cul-2 can regulate polarity establishment independently of their requirement for meiosis II cell cycle progression.

Timing of inverted polarity establishment

We noted that only 25°C mat-1(ax161) zyg-11(RNAi) or 25°C mat-1(ax161) cul-2(RNAi) embryos of the `metaphase I late' and `metaphase I very late' categories show polarized distribution of P granules (Fig. 6D; Table 4). Compatible with this view, inverted distribution of GFP-PAR-2 in mat-1(RNAi) or mat-1(RNAi) zyg-11(mn40) embryos is found predominantly in embryos of these two categories (Fig. 6H,J; Table 4). Therefore, inverted polarity is established some time after the initial metaphase I arrest in embryos lacking both mat-1 and zyg-11 function.

By analogy, we considered whether inverted polarity might be established some time after the initial metaphase II block in embryos lacking zyg-11 or cul-2 function. We imaged GFPPAR-2 in zyg-11(RNAi) and zyg-11(mn40) embryos, and found that cortical enrichment in the vicinity of the meiotic spindle always becomes apparent during the second half of the delay (Fig. 1B, vertical white lines; n=7). This is likely to explain why polarity inversion is observed in∼ 50% of fixed zyg-11(RNAi) or cul-2(RNAi) meiosis II embryos, and indicates that inverted polarity is invariably established by the end of the meiotic cell cycle in the absence of zyg-11 or cul-2 function.

Plasticity of polarity after the meiotic cell cycle in zyg-11(RNAi) embryos

We next investigated how polarity inversion by the end of meiosis II generates the diverse cleavage patterns that occur in the absence of zyg-11 function (Kemphues et al., 1986). By live imaging of zyg-11(RNAi) embryos expressing GFP-HIS and GFP-PAR-2, we found that a second GFP-PAR-2 domain, which is variable in size, is always established on the opposite side of the embryo at the onset of the first mitotic cell cycle, as in wild type (see Movies 8-10 at http://dev.biologists.org/supplemental/; n=19). Therefore, zyg-11(RNAi) embryos experience two polarizing signals, one during meiosis II and a second one at the onset of the first mitotic cell cycle.

We found that these embryos can be placed into three classes according to the evolution of the two GFP-PAR-2 domains. In six embryos, the first GFP-PAR-2 domain diminishes in intensity, whereas the second one becomes more robust; embryos in this first class divide asymmetrically as in wild type, yielding a smaller blastomere on the side of the prevailing GFP-PAR-2 domain (Fig. 7A,D; see Movie 8 at http://dev.biologists.org/supplemental/). In 10 embryos, the first GFP-PAR-2 domain remains robust, whereas the second one diminishes in intensity; embryos in this second class divide asymmetrically, but in an inverted manner compared with wild type (Fig. 7B,D; Movie 9 at http://dev.biologists.org/supplemental/). In the remaining three embryos, both domains of GFP-PAR-2 become undetectable by the end of the first mitotic cell cycle; embryos in this third class divide symmetrically (Fig. 7C,D; Movie 10 at http://dev.biologists.org/supplemental/). These findings suggest that one PAR-2-containing domain at most can be present by the end of the first mitotic cell cycle and that this domain is predictive of spindle positioning during the first cleavage division.

We used GFP-HIS to follow sperm-derived chromosomes in the three classes of zyg-11(RNAi) embryos (n=14). In the first class, where the second GFP-PAR-2 domain prevails, sperm chromosomes remain in the position they occupy in wild type, opposite the first GFP-PAR-2 domain (Movie 8 at http://dev.biologists.org/supplemental/; 6/6 embryos). In most embryos of the second class, where the first GFP-PAR-2 domain prevails, sperm chromosomes are initially in their normal position, but then move towards the inverted GFP-PAR-2 domain (Movie 9 at http://dev.biologists.org/supplemental/; 5/6 embryos). In the third class, where both GFP-PAR-2 domains become undetectable, sperm chromosomes are also initially in their normal position, but then move towards the cell center (Movie 10 at http://dev.biologists.org/supplemental/; 2/2 embryos). These observations suggest that the cleavage pattern of zyg-11(RNAi) embryos is dictated primarily by the position of a sperm-derived component during the course of the first mitotic cell cycle.

To test this hypothesis further, we analyzed polarity in embryos lacking both zyg-11 and spd-2, which is required for establishing polarity induced by the sperm-derived component (O'Connell et al., 2000). We found that spd-2 is dispensable for establishing inverted polarity in the absence of zyg-11 at meiosis II, as PAR-1 is still inverted in∼ 50% of fixed spd-2(oj29) zyg-11(RNAi) meiosis II embryos (Table 3). In addition, the fraction of embryos with an inverted first division is increased from 36% in zyg-11(RNAi) embryos to 60% in zyg-11(RNAi) spd-2(oj29) (Table 3). These findings indicate that a sperm-derived component that is in part spd-2 dependent competes with the inverted polarity established during the meiosis II delay to result in a single GFP-PAR-2 domain by the end of the first mitotic cell cycle.

Microtubules and polarity establishment during meiosis II and the first mitotic cell cycle

We next investigated whether establishment of inverted polarity during the meiotic II delay is microtubule dependent. We eliminated microtubules from zyg-11(mn40) mutant embryos using RNAi against the alpha tubulin gene tba-2, thus likely inactivating all alpha tubulin genes owing to cross-RNAi between highly related sequences (Wright and Hunter, 2003). In such embryos, all microtubule processes, including the two meiotic divisions, are defective, and all oocyte chromosomes are found approximately two-thirds of the way down the length of the egg (Yang et al., 2003) (Fig. 8B and data not shown). Importantly, we found that ∼65% of zyg-11(mn40) tba-2(RNAi) meiosis II embryos have a patch of cortical GFP-PAR-2 (Fig. 8B; n=18), which is comparable to the fraction of zyg-11(RNAi) meiosis II embryos exhibiting inverted PAR-2 cortical localization (Table 3). Similar findings were made examining PAR-1 distribution (data not shown). Moreover, P granules tend to be enriched in the vicinity of cortical GFPPAR-2 (Fig. 8B). Interestingly, we noted also that the patch of GFP-PAR-2 is located invariably on the cortex closest to oocyte chromosomes (Fig. 8C), raising the possibility that they may somehow influence positioning of cortical GFP-PAR-2. Together, these observations indicate that microtubules are not essential for imparting ectopic polarity in the absence of zyg-11 function.

Fig. 8.

Establishment of AP polarity occurs independently of astral microtubules. zyg-11(mn40) (A) or zyg-11(mn40) tba-2(RNAi) (B) meiosis II embryos, wild-type (E,G,I,K) or tba-2(RNAi) embryos (F,H,J,L) during the first mitotic cell cycle, and tba-2(RNAi) embryos during the second mitotic cell cycle (D) are shown. All embryos are stained with antibodies againstα -tubulin (green in merged image) and a centriolar or polarity marker as indicated (SAS-4, PAR-1, GFP-PAR-2, GFP-PAR-6, PGL-1; red in merged image); DNA is shown in blue in the merged image. A dozen ∼1 μm confocal optical sections were imaged. The top panels show a projection of all slices for the α-tubulin channel. Insets show magnified views of centrosomes; width of inset represents ∼5 μm. The merged images show a single section for α-tubulin and the centriolar or polarity marker, along with a projection of all slices for the DNA signal. Scale bar: 10 μm. (C) Plots of position along the circumference in zyg-11(mn40) tba-2(RNAi) embryos (x-axis, degrees, with 0 degrees being posterior-most) as a function of the distance separating oocyte chromosomes from the cortex (y-axis, μm). Red dots indicate cortical locations where GFP-PAR-2 levels are at least five times higher than in the cytoplasm (determined with Metamorph software). Top plot, ten embryos; bottom plot, embryo shown in B. Note that GFP-PAR-2 is present in the cortical region closest to oocyte chromosomes. (E-L) Pairs of wild-type and tba-2(RNAi) embryos, from approximately the same stage, during prophase of the first cell cycle. In all tba-2(RNAi) embryos lacking detectable astral microtubules, including those with only two tiny dots of α-tubulin presumably corresponding to paternally contributed centrioles (F,L, see also D), polarity markers were distributed as in wild type [number of embryos examined for each polarity marker (the number of embryos that only have two tiny dots is given in parentheses): PAR-1, 10 (5); GFP-PAR-2, 6 (1); GFP-PAR-6, 6 (0); PGL-1, 10 (5)].

These results raise the possibility that microtubules may also not be essential for establishing AP polarity during the first mitotic cell cycle. To test whether this is the case, we examined markers of polarity in tba-2(RNAi) embryos. Cortical, astral and spindle microtubules are not detected in one-cell stage tba-2(RNAi) embryos, and the sole remnant of the microtubule cytoskeleton is the centrosome, presumably because it is the favoured site of microtubule nucleation for any tubulin dimers remaining after RNAi treatment (Fig. 8F,H,J,L). In severely affected embryos, only two tiny dots are observed, which co-localize with the centriolar protein SAS-4 (Fig. 8D), indicating that they correspond to paternally contributed centrioles. Strikingly, we found that PAR-1, GFP-PAR-2, GFP-PAR-6 and P granules always localize correctly in tba-2(RNAi) embryos during the first mitotic cell cycle (Fig. 8E-L). Furthermore, the onset of polarized distribution is comparable to that of wild-type embryos; for instance, PAR-1 and P granules enrichments begin during early prophase. The same conclusion is reached with live imaging of tba-2(RNAi) embryos expressing GFP-PAR-2 and GFP-PAR-6 (data not shown). Therefore, microtubules appear not to be necessary for establishing AP polarity in one-cell stage C. elegans embryos.

Discussion

A novel CUL-2-based E3 ligase acting in early C. elegans embryos

The molecular nature of ZYG-11 does not offer immediate clues as to its mode of action. Our analysis, together with the accompanying work by Liu et al. (Liu et al., 2004), reveals that zyg-11 is likely to act with a CUL-2-containing E3 ligase at meiosis. Could ZYG-11 be a substrate of this ECS, given that GFP-ZYG-11 levels are dramatically increased in the absence of cul-2? This possibility seems unlikely, given that inactivation of the ECS and of its substrate are expected to yield opposite, not identical, phenotypes. Although other scenarios can be envisaged, a plausible hypothesis is that ZYG-11 is the substrate recruitment component of this ECS. Although this seems at odds with the fact that ZYG-11 does not contain a recognizable SOCS box, this motif can be very divergent, as is the case for ZIF-1 (DeRenzo et al., 2003), and Drosophila Zyg11 contains a potential SOCS box (R.S. and P.G., unpublished). Whether C. elegans ZYG-11 truly acts as a substrate recruitment subunit remains to be ascertained biochemically.

Regulated protein degradation by way of E3 ligases plays a crucial role in other aspects of early C. elegans development. Another ECS that uses ZIF-1 as a substrate recruitment subunit is essential for removing several CCCH finger proteins from somatic lineages (DeRenzo et al., 2003). Moreover, a CUL-3-based complex is essential for degradation of the meiosis-specific microtubule-severing protein MEI-1 during the first mitotic cell cycle (Kurz et al., 2002; Pintard et al., 2003a). The activity of these two E3 ligases is dependent on the DYRK kinase MBK-2 (Pellettieri et al., 2003). This does not seem to be the case for the ECS described in this work, as embryos lacking mbk-2 function complete the meiotic divisions normally and have no apparent polarity defects (Pellettieri et al., 2003).

The APC and an ECS together ensure progression through the two meiotic divisions

Previous work established that the APC is essential for the metaphase to anaphase transition of meiosis I in C. elegans (Golden et al., 2000; Wallenfang and Seydoux, 2000). Hypomorphic APC mutants that affect meiosis I without arresting cell cycle progression exhibit defective sister chromatid segregation at meiosis II (Shakes et al., 2003). However, this may result from the aberrant meiosis I, as no semi-permissive conditions were found that yield a meiosis II arrest, raising the possibility that APC is not required at meiosis II (Shakes et al., 2003). Compatible with this view, the APC component FZY-1 localizes to chromosomes during meiosis I but not meiosis II (Kitagawa et al., 2002). Here, we establish that a CUL-2-based E3 ligase is required for progression through meiosis II. Though unlikely given the distinct phenotypes of APC hypomorphic mutants and of zyg-11 or cul-2 inactivation, the possibility that this ECS is an essential positive regulator of the APC cannot be excluded for meiosis II. By contrast, this cannot be the case at meiosis I, because inactivation of zyg-11 or cul-2 does not affect meiosis I, whereas that of the APC results in metaphase I arrest. Although embryos lacking zyg-11 or cul-2 do not arrest at meiosis I, compromising APC activity slightly in these embryos results in meiosis I defects, indicating that the ECS plays a non-essential role at meiosis I.

Interestingly, progression through the meiotic cell cycle in other organisms also rests on distinct E3 ligases at meiosis I and meiosis II. In Saccharomyces pombe, the Fizzy/Cdc20-related APC activator mfr1 is required specifically at meiosis II for degradation of the B-type cyclin cdc13 (Blanco et al., 2001). Similarly, in Drosophila, the Cdc20/Fizzy protein Cortex is required for timely metaphase to anaphase transition at meiosis II (Chu et al., 2001; Page and Orr-Weaver, 1996). Thus, whereas meiosis II in S. pombe and Drosophila relies on specific APC variants, it requires the activity of an ECS in C. elegans.

What substrates of the CUL-2-based E3 ligase must be targeted for degradation to ensure progression through meiosis II in C. elegans? The metaphase to anaphase transition is generally triggered when securin is targeted for degradation. IFY-1 is a C. elegans destruction-box protein that has properties expected from a securin (Kitagawa et al., 2002), and it will be interesting to test whether the CUL-2-based E3 ligase mediates IFY-1 degradation at meiosis II. Furthermore, M phase exit is generally triggered when B type cyclins are targeted for degradation. Our findings that cyb-3 inactivation prevents the anaphase delay of embryos lacking zyg-11 function strongly suggests that CYB-3 is a target of the CUL-2-based E3 ligase at meiosis II.

zyg-11 and cul-2 prevent polarity establishment

In wild type, the first signs of polarized GFP-PAR-2 distribution occur shortly after meiosis II. Whether polarity is established when a set time is reached (e.g. after fertilization), or when a given cell cycle stage is encountered, has not been addressed prior to this work. We found that asymmetric distribution of polarity markers in cyb-3(RNAi) embryos is not established during the meiosis II delay, but instead during the first mitotic cell cycle (R.S. and P.G., unpublished). Therefore, polarity establishment does not occur at a fixed time after fertilization or meiosis I, but is coupled instead to exit from the meiotic cell cycle and the onset of the first mitotic cell cycle.

In contrast to the situation in cyb-3(RNAi) embryos, inverted polarity is established during the meiosis II delay in embryos lacking zyg-11 or cul-2 function. We show that zyg-11 and cul-2 can regulate polarity establishment independently of promoting progression through meiosis II, suggesting that there are distinct polarity substrates that must be degraded by the CUL-2-based E3 ligase to prevent polarity establishment during meiosis II. As inverted polarity is also observed in metaphase I-arrested embryos lacking APC function (Wallenfang and Seydoux, 2000), it is tempting to speculate that at least some of these polarity substrates are shared between the two E3 ligases.

Interestingly, cortex mutant embryos in Drosophila have impaired polyadenylation of bicoid and Toll mRNAs, resulting in inefficient translation and ensuing defective embryonic polarity (Lieberfarb et al., 1996). Although the mechanisms by which APC promotes polyadenylation of these mRNAs remains to be clarified, it is remarkable that E3 ligases that ensure progression through meiosis II also serve to promote correct establishment of embryonic axes in diverse metazoan organisms.

Mechanisms of polarity establishment in C. elegans

It has been proposed that astral microtubules nucleated by the sperm aster are essential for establishing AP polarity in C. elegans embryos (O'Connell et al., 2000; Wallenfang and Seydoux, 2000). Our findings challenge this view. We establish that PAR-1, GFP-PAR-2, GFP-PAR-6, P granules and GFPPIE-1 (R.S. and P.G., unpublished) all localize correctly in the first mitotic cell cycle in tba-2(RNAi) embryos. These results raise the possibility that microtubules are dispensable for AP polarity.

We recognize that we cannot exclude that minute microtubules undetectable by immunofluorescence or damaged during fixation may play a role. Nevertheless, we note that embryos lacking spd-2, air-1 or spd-5 function, while exhibiting delayed microtubule nucleation compared with wild type, have a more extensive microtubule network than tba-2(RNAi) embryos do (Hamill et al., 2002; O'Connell et al., 2000; Schumacher et al., 1998; Wallenfang and Seydoux, 2000). How can these apparently discrepant observations be reconciled? In wild type, the onset of GFP-PAR-2 accumulation at the cortex coincides with that of GFP-TUB on asters, early in the cell cycle (Cuenca et al., 2003). It may be that, at that early stage, minute microtubules are present in tba-2(RNAi) embryos, but not in embryos lacking spd-2, air-1 or spd-5 function. If this were the case, our findings would merely demonstrate that microtubules are dispensable for the expansion phase of the posterior PAR-2 cortical domain and not for the preceding initiation phase (Cuenca et al., 2003). However, we favour an alternative explanation in which microtubules are dispensable throughout the process, and in which spd-2, air-1 and spd-5 have a requirement for polarity establishment that is independent from their role in microtubule nucleation.

As sperm chromosomes are not essential for AP polarity (Sadler and Shakes, 2000), our findings lead us to suggest that, rather than astral microtubules, the sperm component acting as the polarity cue is the centrosome. Compatible with centrosomes being key, spd-2, spd-5 and air-1 are all required for centrosome maturation (Hamill et al., 2002; Hannak et al., 2001; O'Connell et al., 2000), and the corresponding proteins localize to the centrosome (Hamill et al., 2002; Schumacher et al., 1998; Hannak et al., 2001; Kemp et al., 2004). It will be interesting to investigate whether such a hypothetical polarity cue resides in centrioles or the surrounding pericentriolar material. Equally interesting will be to elucidate how a component somehow correlated with the position of oocyte chromosomes, which do not have associated centrosomes, can act as a surrogate polarity cue during meiosis II.

Our findings taken together suggest a working model (Fig. 9) in which a CUL-2-based E3 ligase targets polarity substrates for degradation during meiosis II; as a result, polarity cannot be established. After exit from the meiotic cell cycle, polarity substrates accumulate, perhaps due to diminished E3 ligase activity, and thus the centrosome can induce polarity. We postulate that in the absence of zyg-11 or cul-2, substrate accumulation occurs during the meiotic cell cycle, resulting in polarity establishment in response to a surrogate polarity cue.

Fig. 9.

Working model of polarity establishment in C. elegans embryos. (A) In wild-type meiosis II, a CUL-2-based E3 ligase that also requires zyg-11 function ubiquitinates a substrate (X) that can induce polarity when triggered by a surrogate polarity cue, thus targeting the substrate for degradation by the proteasome. After exit from the meiotic cell cycle, E3 ligase activity is downregulated, allowing substrate accumulation and polarity establishment in response to the bona fide centrosome polarity cue. Green oval, oocyte chromosomes; green disk, centrosomes; blue crescent, cortical PAR-2; orange disks, P granules. Ubiquitin (Ub, blue circles) moieties are also shown. (B) In the absence of zyg-11 or cul-2, substrate accumulation occurs during the meiotic cell cycle, resulting in polarity establishment in response to a surrogate polarity cue that correlates with the position of oocyte chromosomes. A related outcome may ensue when the APC is inactivated at meiosis I (not depicted here). After exit from the meiotic cell cycle in the absence of zyg-11 or cul-2, a second focus of polarity is established in response to the centrosome polarity cue.

Acknowledgments

For help in improving the manuscript, we thank Jean-Michel Bellanger, Ken Kemphues, Lionel Pintard and Viesturs Simanis. We are grateful to Ken Kemphues also for advice, encouragement, and zyg-11 mutant alleles; to Michael Glotzer, Lionel Pintard, Geraldine Seydoux and Viesturs Simanis for reagents; to Edward Kipreos for communicating results prior to publication; and to Karine Baumer for generating transgenic lines. Some strains were obtained from the Caenorhabditis Genetics Center, which is funded by the NIH National Center for Research Resources (NCRR). This work was supported by grant 31-62102.00 from the Swiss National Science Foundation.

Footnotes

  • Movies and supplemental data available online

    • Accepted April 30, 2004.

References

View Abstract