Functional Redundancy of Paralogs of an Anaphase Promoting Complex/Cyclosome Subunit in Caenorhabditis elegans Meiosis
Kathryn K. Stein, Jessica E. Nesmith, Benjamin D. Ross, Andy Golden


The anaphase promoting complex/cyclosome (APC/C) mediates the metaphase-to-anaphase transition by instructing the ubiquitination and turnover of key proteins at this stage of the cell cycle. We have recovered a gain-of-function allele in an APC5 subunit of the anaphase promoting complex/cyclosome. This finding led us to investigate further the role of APC5 in Caenorhabditis elegans, which contains two APC5 paralogs. We have shown that these two paralogs, such-1 and gfi-3, are coexpressed in the germline but have nonoverlapping expression patterns in other tissues. Depletion of such-1 or gfi-3 alone does not have a notable effect on the meiotic divisions; however, codepletion of these two factors results in meiotic arrest. In sum, the two C. elegans APC5 paralogs have a redundant function during the meiotic divisions.

THE anaphase promoting complex or cyclosome (APC/C) is an E3 ubiquitin ligase composed of >10 subunits, which is required for progression through the cell cycle (Peters 2006). An activating subunit associates transiently with the APC/C in a cell cycle-dependent manner to positively regulate APC/C activity: CDC20/fzy-1 at M phase and CDH1/fzr-1 during the G1 phase of the cell cycle (Pesin and Orr-Weaver 2008). During meiosis and mitosis, the ubiquitination of substrates by the APC/C results in protein turnover, which drives these processes. One such APC/C target is securin. Upon degradation of securin, separase, an enzyme required for sister chromatid separation, is activated and promotes the metaphase-to-anaphase transition.

The timing of APC/C activation during M phase must be carefully regulated because precocious or aberrant segregation of chromosomes can result in aneuploidy of daughter cells, leading to embryonic death or uncontrolled cell proliferation. At metaphase, the APC/C is quiescent while the chromosomes are captured by microtubules at their kinetochores and aligned on the metaphase plate. At this time, the APC/C is held in check by the components of the spindle assembly checkpoint (SAC). The SAC regulates the activity of the APC/C, depending on the state of chromosome attachments (Musacchio and Salmon 2007). When the kinetochores are unattached, the SAC holds the APC/C in the inactive state by the inhibition of the positive APC/C regulator, CDC20/FZY-1. Once the chromosomes are properly aligned and oriented, the SAC releases CDC20 so that the APC/C is activated to ubiquitinate the appropriate substrates. The components of the SAC were first discovered in Saccharomyces cerevisiae and the core proteins are functionally conserved in multicellular organisms (Musacchio and Salmon 2007).

In Caenorhabditis elegans, most of the APC/C subunits are required for cell division. In homozygous null mutants segregating from a heterozygous hermaphrodite, maternal stores perdure and rescue embryonic and postembryonic development; cell divisions in the late larvae fail and the terminal phenotype of null mutants is sterility due to the absence of germline proliferation (Furuta et al. 2000). Temperature-sensitive mutants of five APC/C subunits have been isolated (Furuta et al. 2000; Golden et al. 2000). When these hypomorphic mutants are shifted to restrictive temperature following the completion of the mitotic divisions required for germline establishment, a brood of one-cell embryos arrested at metaphase of meiosis I is produced. These results indicate that there is a requirement for the APC/C during the meiotic divisions (Furuta et al. 2000; Golden et al. 2000; Shakes et al. 2003). RNAi-mediated knockdown of individual APC/C subunits in late larval hermaphrodites also leads to one-cell meiotic arrest with the exception of two subunits, APC5 and APC10 (Davis et al. 2002), which exhibit arrest at a multicellular stage when depleted. It recently has been established that there are two genes encoding each of these subunits in C. elegans (Tarailo et al. 2007). Therefore, APC5 and APC10 may not exhibit an RNAi meiotic arrest phenotype because the two paralogs of each subunit might function redundantly at the meiotic divisions. Alternatively, it is possible that these subunits are not required at the meiotic divisions.

A suppressor screen was performed using the meiosis-specific temperature-sensitive loss-of-function allele of mat-3/APC8/CDC23, mat-3(or180ts), to discover meiotic regulators of the APC/C in C. elegans (Stein et al. 2007). Most of the suppressors cloned from that study function in the SAC pathway. The SAC mutants probably suppress mat-3(or180ts) by increasing the level of APC/C activity via a reduction of the strength of the spindle checkpoint. The interdependence of the regulation of the SAC and the APC/C is demonstrated by reciprocal studies in which a loss-of-function in the SAC can be suppressed by reduction-of-function APC/C mutants (Furuta et al. 2000; Tarailo et al. 2007). Further, Bezler and Gönczy (2010, accompanying article in this issue) assert that proper mitotic timing is achieved through the negative regulation of the SAC by the APC/C. In this study, we identify the mat-3 suppressor av9gf as a gain-of-function allele of one of the APC5 paralogs, such-1. In addition, our studies with the such-1(h1960ts) loss-of-function allele and APC5 RNAi have led us to conclude that the APC5 paralogs such-1 and gfi-3 function redundantly during the meiotic divisions.


Genetic mapping of av9:

av9 was isolated in an EMS-based genetic screen as a suppressor of mat-3(or180ts) embryonic lethality at the nonpermissive temperature of 24°. Genome-wide snip-SNP mapping was performed to localize this mutant to chromosome III as described previously (Stein et al. 2007). Classical three-factor genetic mapping was performed in the mat-3(or180ts) dpy-18(e364) unc-25(e156) background by mating with mat-3(or180ts) av9; him-8(e1489) males. In the dpy-18 unc-25 interval 15/17 Unc non-Dpys were suppressed and 3/27 Dpy non-Uncs were suppressed; these data predict a map position of ∼10.01 ± 1.45 on LG III. The such-1 gene is located at 10.99 on LG III (


Two such-1 RNAi feeding constructs were made. They contained genomic DNA from (1) a 275-bp fragment spanning parts of exon and intron 10 and (2) exons 6–8 with intervening introns. RNAi feeding constructs were created using the Gateway system (Invitrogen, Carlsbad, CA) and a customized L4440 vector (Timmons and Fire 1998), pCR88, then transformed into HT115(DE3) competent cells for use. The gfi-3 clone from the Ahringer RNAi library (Kamath et al. 2003) was used for gfi-3 RNAi experiments (Geneservice, Cambridge, UK). This RNAi clone included 907 bp spanning a genomic region from exons 14–16 and intervening introns. At 24°, L4 hermaphrodites were fed bacteria containing the RNAi construct for 24–28 hr and then moved to a new RNAi plate for another 20–24 hr. Hermaphrodites were then removed. After another 24 hr the second RNAi plate was scored. Embryos were assessed microscopically for one-cell arrest and larvae were counted. To control for RNAi treatment, animals were grown on smd-1 bacteria, which does not produce a phenotype. Animals of the following genotypes were tested: wild type, mat-3(or180ts), mat-3(or180ts) such-1(av9), and such-1(h1960ts). such-1(h1960ts) was isolated from the such-1(h1960ts); mdf-1(gk2) line generated in the Rose lab and was backcrossed five times with dpy-18(e364) unc-25(e156).


Genomic DNA of such-1 was amplified by PCR from the mat-3(or180ts) av9 strain and sequenced by Macrogen (Rockville, MD). A mutation was discovered in exon 10 (G to A 8260 nucleotides downstream of the start site), which created a Snip-SNP not present in wild type. The mutation was confirmed by digestion of a PCR fragment with AvaI. Alignments were done with the CLUSTAL 2.0.11 multiple sequence alignment program (Larkin et al. 2007).

Brood size:

For each experiment, 5 L4 hermaphrodites of each genotype were placed on OP50 plates at 16° or 24° and moved each day to a new plate until embryos were no longer observed. Embryos were counted 24 hr after hermaphrodites were removed from each plate and larvae were counted 24 hr later. Brood size was calculated by totaling all larvae and unhatched embryos and percentage of hatching was equal to the number of hatched embryos divided by the brood size and multiplied by 100.

Double mutant analysis:

Genetic enhancement was evaluated between the loss-of-function mutant, such-1(h1960ts), and temperature-sensitive alleles of APC/C subunit mutants: mat-3(or180ts) dpy-1(e1), emb-27(g48ts) unc-4(e120), and unc-74(x19) mat-1(ax144ts), by mating such-1(h1960ts) males with hermaphrodites of the marked lines. In the F2 generation, marked hermaphrodites were picked to single plates at 16° and scored for sterility, maternal-effect lethality (Mel), or production of a normal brood. In the case of emb-27(g48ts) and mat-1(ax144ts), it was anticipated that 25% of the F2 would also be homozygous for the unmarked such-1(h1960ts) mutation because the two genes reside on different chromosomes. Hermaphrodites that were sterile or Mel were genotyped for the presence of such-1(h1960ts), utilizing the Snip-SNP created by the such-1(h1960ts) mutation and the SgrA1 enzyme. Because mat-3 and such-1 are both on LG III, F2 Dpy animals were picked to single plates and 10 F3 progeny were picked from each plate to detect an F2 recombinant animal by the presence of F3 sterile animals or F4 embryonic lethality at the permissive temperature. Animals were genotyped to confirm the presence of such-1(h1960ts).

Most of the double mutants were generated with marked or unmarked alleles of APC/C subunits and dpy-18(e364) such-1(av9). The following strains marked with a morphological marker were combined with an unmarked such-1(av9gf) allele: unc-74(x19) mat-1(ax144ts), mat-1(ax212ts) unc-13(e51), and mat-3(ax148ts) dpy-1(e1). In all cases, control strains contain the morphological marker but not the test allele. In each experiment, at least 400 embryos were evaluated from the experimental and control double mutants. Double mutants were shifted to the restrictive temperature at the L4 stage for 24 h, and then were removed. Embryos and larvae were counted the next day and embryos were closely examined with a Nikon SMZ-2B dissecting microscope to assess the stage of arrest, one-cell or multicellular. Percentage of hatching was calculated as described above (n = 3). Double mutants between unc-17(e245) mdf-2(av14) or fzy-1(av15gf) unc-4(e120) and dpy-18(e364) such-1(av9gf) were generated and confirmed by molecular genotyping. Five animals were shifted to 24° and brood size and hatching was determined for each experiment. The Dpy Unc phenotype of these animals sometimes led to a small brood and premature death of the hermaphrodites, as a result of egg-laying defects. To account for the early death of hermaphrodites in the brood size calculations, total progeny for each day was calculated by adding the number of embryos and larvae on the plate and dividing by the number of adults that were alive on that plate at the beginning of the day to obtain a value of progeny per animal for that day. The progeny per animal per day were then summed for the course of the experiment to produce the value of total brood per animal.

Transgenic animals:

For both such-1 and gfi-3, the 5′ promoter region (750 bp upstream of the translational start site for such-1, 885 bp for gfi-3) and 3′-UTR (739 bp downstream of stop codon for such-1, 758 bp for gfi-3) were amplified by PCR and entry clones were generated using Gateway technology (Invitrogen). All entry clones were sequence confirmed. These regulatory regions were then recombined with mCherry∷H2B coding sequence in a destination vector containing a wild-type copy of unc-119. unc-119(ed3) animals were bombarded by biolistic transformation (Wilm et al. 1999) using a Model PDS-1000/He biolistic particle delivery system (BioRad, Hercules, CA) and plated on fifteen 100-mm plates. Postbombardment, animals were allowed to grow for 7 days at 24° until the plate was starved. A pedestal of rescue (a small piece of agar seeded with bacteria) was then introduced to each plate and after 48 hr pedestals were examined for wild-type (non-unc-119) animals. Transformed animals contain the unc-119 rescuing construct and are normally motile. Therefore, transformed animals, but not Unc animals, are able to crawl up the piece of agar and access the food, which allows them to reproduce and also facilitates detection. Wild-type animals were picked to single plates to determine which were stable trangenics and to distinguish between integrated and extrachromosomal lines. Finally, animals were examined microscopically to assess if any were expressing mCherry∷H2B. Positive lines were maintained and used for further studies. For such-1, four lines were generated with an identical expression pattern. One was integrated and the remainder expressed extrachromosomal arrays. For gfi-3, three lines were generated with an identical expression pattern. Two were integrated and the other expressed an extrachromosomal array. Live animals and embryos were examined using a Nikon Eclipse E800 microscope equipped with a PerkinElmer Ultraview LCI CSU10 scanning unit (PerkinElmer, Fremont, CA) and an ORCA ER cooled CCD camera (Hamamatsu, Japan). The objective was a ×40 Nikon Plan Apo oil with a numerical aperature of 1.0.

such-1(h1960ts)/Df RNAi experiments:

To generate such-1(h1960ts)/ctDf3 animals, such-1(h1960ts) males were mated into ctDf3/qC1 dpy-19(e1259) glp-1(q339) hermaphrodites and cross progeny were individually moved to RNAi or control OP50 plates at 25°. Hermaphrodites were moved to new RNAi plates after 24 hr and left on these plates for an additional 24 hr. Unhatched embryos and larvae were counted for the 24- to 48-hr plate and percentage of hatching was calculated. Plates segregating sterile Dpy animals were judged to contain qC1 dpy-19(e1259) glp-1(q339) and were used as a control. All other plates were such-1(h1960ts)/ctDf3. A minimum of 25 single plates were assessed for such-1(exon 10) RNAi, gfi-3 RNAi, and smd-1 RNAi. Embryos were visually examined to ascertain whether they were arrested at a one-cell or multicellular stage.


av9 is a gain-of-function allele of the APC5 homolog, such-1:

av9 was isolated in a large-scale genetic screen as a semidominant suppressor of the embryonic lethality of the temperature-sensitive APC8 subunit mutant, mat-3(or180ts) (Stein et al. 2007). mat-3(or180ts) is a maternal-effect lethal mutation that produces a brood of one-cell embryos arrested at metaphase of meiosis I when shifted to restrictive temperature. In the presence of the suppressor av9, ∼60% of embryos hatch and become fertile adults (Stein et al. 2007). av9 was mapped to a small interval on LG III between dpy-18 and unc-25 using snip-SNP and classical genetics methods. This region of LG III was examined for candidate genes involved in cell cycle regulation or the spindle assembly checkpoint. One candidate, Y66D12A.17, encodes such-1 (suppressor of spindle checkpoint defect), a homolog of the APC5 subunit. RNAi was used to test whether av9 was likely to be an allele of such-1. If av9 was a loss-of-function allele of such-1, RNAi of such-1 would be expected to restore viability to mat-3(or180ts) embryos at the restrictive temperature. If av9 was a gain-of-function allele of such-1, RNAi of such-1 would eliminate the viability of mat-3(or180ts) av9 embryos at the restrictive temperature. Wild-type animals are viable on such-1 RNAi (Figure 1). We found that knockdown of such-1 by RNAi did not rescue the lethality of mat-3(or180ts) at 24° but depletion of such-1 in a mat-3(or180ts) av9 background eliminated the suppression provided by av9 (Figure 1). This result indicates that the semidominant suppressor, av9, may contain a gain-of-function mutation in the such-1 gene and will hereafter be referred to as such-1(av9gf).

Figure 1.—

such-1 RNAi eliminates the suppression of mat-3(or180ts) by such-1(av9). Wild-type, mat-3(or180ts), and mat-3(or180ts) such-1(av9) animals were grown at 24° on smd-1 control or such-1(exons 6-8) RNAi and embryonic hatching was assessed. 1C indicates the presence of arrested one-cell embryos. Statistics, Student's T-test; asterisk indicates a P-value <0.001. Error bars indicate standard deviation.

The such-1 exons were sequenced from the av9gf strain and found to contain a point mutation that results in a glutamic acid (E)-to-lysine (K)-amino-acid change at residue 693. The APC5 homologs have few regions of extended conservation between species and only a single putative TPR domain has been reported (Bentley et al. 2002). Thus, there does not appear to be a conserved functional role for this amino acid across species. C. elegans contains an additional ortholog of APC5, GFI-3 [GEI-4 (Four)-Interacting protein], which is 38% identical and 71% similar to SUCH-1 at the amino acid level, including identity at the site that is mutant in av9gf (for an amino acid alignment, see Tarailo et al. 2007). All other model organisms contain only a single copy of APC5. Identity between each C. elegans paralog and the ortholog in human, Drosophila, and S. cerevisiae is comparable, indicating that both C. elegans paralogs are equally similar to APC5 (Tarailo et al. 2007).

To investigate whether the such-1(av9gf) allele has a phenotype on its own, we tested the viability of this mutant in an otherwise wild-type background. When compared to the wild-type strain, there were no significant differences between such-1(av9gf) alone and wild-type animals with regard to brood size or embryonic hatching (Figure 2, A and B). Therefore, while the av9gf mutation is capable of significant suppression of mat-3(or180ts), it has no deleterious effects on the viability or fecundity of the animal.

Figure 2.—

such-1(av9gf) animals are fertile and healthy. Wild-type and such-1(av9gf) animals were grown at 16° and 24°. Brood size (A) and percentage of embryonic hatching (B) were evaluated and found to be not significantly different. Statistics, Student's T-test; error bars indicate standard deviation.

The such-1 gain-of-function allele may not reflect the true function of this gene. Therefore, we examined the such-1 loss-of-function phenotype. Two loss-of-function alleles exist for such-1, a penetrant paternal-effect lethal mutation, such-1(t1668) (Gönczy et al. 1999, Bezler and Gönczy 2010) and such-1(h1960ts), a point mutant that exhibits near-normal hatching at 20° but reduced hatching and brood size at 25° (Tarailo et al. 2007). such-1(h1960ts) embryos that fail to hatch arrest at a range of embryonic stages, from rare one-cell embryos to multicellular embryos. The absence of a complete meiotic one-cell arrest in such-1(h1960ts) embryos at the restrictive temperature raised the possibility that APC5 does not serve as a canonical APC/C subunit in C. elegans. Characteristically, the temperature-sensitive APC/C mutants isolated to date enhance each other when combined at the permissive temperature and double mutants are sterile or produce broods of unhatched embryos (Shakes et al. 2003; our unpublished observations). To address the possibility that such-1(h1960ts) encodes a functional APC/C subunit, the loss-of-function allele was combined with three representative temperature-sensitive APC/C loss-of-function mutants at permissive temperature and assayed for viability. Double mutants between such-1(h1960ts) and the APC/C mutants mat-3(or180ts), emb-27(g48ts), or mat-1(ax144ts) exhibited a sterile or maternal-effect lethal phenotype (Table 1). This finding is consistent with a role for such-1 in APC/C function. such-1(h1960ts) may not exhibit a one-cell arrest because it is a weak hypomorphic allele or alternatively, it has overlapping function with the APC5 paralog, gfi-3.

View this table:

The loss-of-function mutant, such-1(h1960ts), enhances temperature-sensitive APC/C mutants

such-1(av9gf) suppresses APC/C subunit mutations but does not interact genetically with SAC component mutations:

We have shown previously that many of the suppressors isolated in the mat-3(or180ts) suppressor screen are capable of suppressing mutations in other APC/C subunits (Stein et al. 2007). The majority of these suppressors were spindle assembly checkpoint genes, so it is thought that they suppress through a common mechanism. Because SUCH-1 is a component of the APC/C itself, we wondered whether this suppressor would be as promiscuous or whether suppression might be limited to mat-3(or180ts). To address this hypothesis, we made double mutants of such-1(av9gf) with a strong and weak allele of each of the five APC/C subunits for which temperature-sensitive alleles are available (Table 2) (Golden et al. 2000). In addition, we generated a double mutant with emb-1(hc62ts), an allele that also exhibits a one-cell–arrest phenotype (Miwa et al. 1980; Schierenberg et al. 1980; our unpublished observations). We found that such-1(av9gf) was a suppressor of three weak APC/C alleles in addition to mat-3(or180ts): emb-30(or420ts), mat-1(ax212ts), and emb-1(hc62ts) (Table 2). Notably, suppression of the meiotic one-cell arrest was observed: most of the embryos were multicellular. Despite this meiotic suppression, only a small percentage of embryos were capable of hatching, probably due to mitotic defects arising later in development that were not suppressed by such-1(av9gf). These results indicate that such-1(av9gf) suppression is not limited to mat-3(or180ts) and, in addition, is not specific for the MAT-3 subunit. Thus, such-1(av9gf) appears to increase the activity of various mutant forms of the APC/C.

View this table:

such-1(av9gf) suppression is not specific to mat-3(or180ts)

We have shown that such-1(av9gf) does not confer an apparent mutant phenotype in an otherwise wild-type background (Figure 2). This observation indicates that a gain-of-function mutation, which likely increases the activity of the APC/C, on the basis of its ability to suppress mat-3(or180ts), does not have detrimental effects on the normal development and function of the organism. However, if APC/C activity was elevated even further, what would be the consequences for the cell and organism? We tested this question in two ways. First, we reduced the spindle checkpoint activity by using a loss-of-function mutation in mdf-2. mdf-2 is the ortholog of S. cerevisiae MAD2 (Kitagawa and Rose 1999), which inhibits the APC/C by sequestering FZY-1. In the presence of mdf-2(av14), a greater amount of FZY-1, the positive regulator of the APC/C and ortholog of CDC20, should be available to drive APC/C activity. Previous studies have shown that mdf-2(av14) is a strong suppressor of APC/C loss-of-function mutants and also that its embryonic lethality is enhanced by another SAC loss-of-function mutation in san-1/mdf-3, the MAD3 ortholog (Nystul et al. 2003). We combined mdf-2(av14) with such-1(av9gf) and examined these double mutants for the presence of any enhanced phenotype. We found that such-1 and mdf-2 do not appear to have any genetic interaction (Figure 3, A and B). We then tested whether increasing fzy-1 activity directly through a fzy-1 gain-of-function mutation would lead to overactivated APC/C when combined with such-1(av9gf). However, the fzy-1(av15gf) allele, which is healthy on its own (like av9gf), does not exhibit a synthetic phenotype when in combination with av9gf (Figure 3, C and D). Thus, the change in APC/C activity provided by these two mutants either is not detrimental to the cell or is insufficient to alter the bulk of APC/C activity in an appreciable way.

Figure 3.—

such-1(av9gf) does not genetically interact with the spindle checkpoint mutants mdf-2(av14) or fzy-1(av15gf). Double mutants were made with dpy-18(e364) such-1(av9gf) and two spindle checkpoint mutants, unc-17(e245) mdf-2(av14) (A and B) and fzy-1(av15gf) unc-4(e120) (C and D). Brood size per animal (A and C) and embryonic hatching (B and D) are not significantly different between the control and experimental strain. Statistics, Student's T-test; error bars indicate standard deviation.

C. elegans contains two APC5 paralogs:

To date, C. elegans is unique among model organisms, as it contains two genes, such-1 and gfi-3, which are equally similar to APC5 orthologs. What is the nature of the two APC5 genes in C. elegans? It is possible that these two paralogs of APC5 have a redundant function in the same tissue or cell type, serve different functions in the same cell, or have the same function in different tissues, such as the soma and the germline. To understand the role of APC5 in C. elegans, we first determined the expression pattern of such-1 and gfi-3 and then further investigated the loss-of-function phenotype of each gene to establish the processes for which each gene is required.

We examined the expression pattern of such-1 and gfi-3 by constructing transcriptional fusions and generating transgenic animals. The fusion protein mCherry∷histone H2B was driven by the promoter and 3′-UTR of either such-1 or gfi-3. such-1p∷mCh∷H2B∷such-1 3′-UTR is expressed throughout the developing germline (Figure 4A) as well as in the meiotic embryo (Figure 4B) and throughout embryogenesis (Figure 4C). Limited expression is observed in the soma of the adult, including some head neurons and vulval precursor cells (data not shown). gfi-3p∷mCh:H2B∷gfi-3 3′-UTR was similarly expressed throughout the germline, meiotically, and in all embryonic stages (Figure 4, E–G). Only the gfi-3 transgene is observed in the soma of the L1–L4 larval stages and is also expressed in the gut cells of the adult animal (Figure 4E, data not shown). In hermaphrodites, both such-1– and gfi-3–driven transgene expression is prominent in mature sperm stored in the spermatheca. We generated male transgenics and found that these transgenes are expressed throughout the male germline (Figure 4, D and H). such-1 and gfi-3 exhibit overlapping expression patterns in the germline, but are expressed specifically in other tissues and stages. These data indicate that such-1 and gfi-3 have the opportunity to act redundantly during germline development and meiosis.

Figure 4.—

such-1 and gfi-3 transcriptional fusions are expressed predominantly in the adult germline and embryo. such-1∷mCherry∷H2B∷such-1 and gfi-3∷mCherry∷H2B∷gfi-3 stable transgenes were imaged by fluorescence microscopy (A–H) and DIC (B, C, F, and G). such-1 (A) and gfi-3 (E) adult hermaphrodites exhibit expression in the germline extending from the distal tip (inset) to the −1 oocyte (arrowhead, A and E). Sperm expression is observed in the spermathecae (arrow, A and E). Somatic gut nuclei (asterisk, E) also express gfi-3. such-1 is expressed in meiotic (arrow, B) and mitotic embryos (C). gfi-3 is expressed at all embryonic stages, including meiosis (arrow, F) and mitosis (G). Male germlines also express the such-1 (D) and gfi-3 (H) transgenes; mature sperm are indicated with an arrow. Bars, 10 μm.

such-1 and gfi-3 function redundantly during meiosis:

The expression data suggest that such-1 and gfi-3 could be redundant during meiosis, explaining the lack of meiotic arrest phenotypes observed with the such-1(h1960ts) loss-of-function mutant. To test this, two such-1 RNAi constructs were used, derived from genomic DNA containing exons 6–8 or exon 10, to knock down the message levels in wild-type animals (Figure 5A). Wild-type animals fed bacteria expressing either such-1 RNAi construct exhibited no embryonic lethality (Figure 5B and data not shown). While the such-1 RNAi was sufficiently effective to block the suppression of the sensitized mat-3(or180ts) such-1(av9gf) strain (see Figure 1), such-1 RNAi of wild-type animals does not recapitulate the phenotype of the such-1(h1960ts) loss-of-function allele (Figure 5B). To further reduce SUCH-1 levels, such-1(h1960ts) loss-of-function animals were grown on such-1 RNAi at the restrictive temperature of 25°. Reduction of such-1 by RNAi reduces the viability of such-1(h1960ts) mutants (Figure 5B), indicating that such-1(h1960ts) is not null, consistent with its weak phenotype. such-1(h1960ts) hermaphrodites grown on either such-1 RNAi produce embryos with a significantly reduced hatching rate, and one-cell–arrested embryos are only rarely observed (Figures 5, B and C). To further deplete SUCH-1, we created animals hemizygous for such-1(h1960ts) [such-1(h1960ts)/ctDf3] and grew them on such-1 RNAi. At 25°, the hatching rate of untreated such-1(h1960ts)/ctDf3 is 33%. When animals of this genotype are grown on such-1(exon 10) RNAi, a 2.4% hatching rate is observed; however, all of the such-1 RNAi-treated embryos are multicellular (data not shown). These results indicate that a severe depletion of the levels of such-1 is not sufficient to yield a meiotic one-cell arrest typically observed with an APC/C mutant or RNAi of an APC/C subunit (Davis et al. 2002).

Figure 5.—

such-1 and gfi-3 act redundantly in meiosis. (A) Schematic of such-1 and gfi-3 gene structure and position of genomic fragments used in RNAi constructs (solid bars below gene structure). The scale bar to the bottom right of each gene structure indicates 100 base pairs (bp). Total genomic length is to the right of the gene structure. (B) Percentage of embryonic hatching of wild-type and such-1(h1960ts) animals grown on several different RNAi conditions at 25°. N > 750 for each sample. (C) Percentage of dead embryos exhibiting a one-cell–arrest phenotype at 25°. N > 435 for each sample. Statistics, Student's T-test; asterisk indicates a significant difference between sample and smd-1 control, with a P-value <0.05. Error bars indicate standard deviation.

Because such-1 depletion does not produce a one-cell arrest, we sought to deplete gfi-3 and such-1 in combination to directly address the possible redundancy of these two genes. There are no mutant alleles available for gfi-3 and wild-type animals fed bacteria expressing gfi-3 RNAi do not exhibit lethality (Figure 5B). When such-1 and gfi-3 RNAi is introduced as a mix, wild-type animals show a slight increase in embryonic lethality (Figure 5B). such-1(h1960ts) hermaphrodites grown on gfi-3 RNAi produce embryos with a significantly reduced hatching rate and a significant population of one-cell embryos (Figure 5, B and C). Under conditions where such-1 and gfi-3 are simultaneously depleted by RNAi in combination with such-1(h1960ts) loss-of-function, a brood composed almost exclusively of one-cell embryos is produced (Figure 5C). The “escapers” that hatch before RNAi becomes fully effective become sterile F1 adults. This finding is strong evidence that such-1 and gfi-3 are redundant at the metaphase-to-anaphase transition of meiosis. To confirm this redundancy, we grew animals hemizygous for such-1(h1960ts) [such-1(h1960ts)/ctDf3] on gfi-3 RNAi at 25° and found that hatching is reduced to 0.7% (n = 1659). Importantly, the embryos laid on gfi-3 RNAi are almost all meiotic one-cell embryos, in contrast to the such-1 RNAi-treated embryos, which are multicellular. In sum, gfi-3 and such-1 are likely redundantly required for the first meiotic division.


APC5 is required for viability in all organisms where it has been studied to date. In the yeasts Schizosaccharomyces pombe and S. cerevisiae, null mutations in APC5 exhibit the “cut” phenotype or arrest at metaphase, typical of an APC/C subunit mutant in each species (Yu et al. 1998; Zachariae et al. 1998; Ors et al. 2009). In Drosophila, null alleles of the APC5 ortholog, ida, result in larval lethality following the depletion of wild-type maternal stores during embryogenesis. When ida is manipulated so that expression is eliminated in the germline, flies produce an extremely small brood, which is 100% embryonic lethal (Bentley et al. 2002). However, ida mutant cells are not arrested in metaphase, but rather seem to enter anaphase precociously, which generates aneuploid cells. The disparate phenotypes observed between organisms suggest that APC5 may have divergent functions within the APC/C complex. In this study, we demonstrate that C. elegans has two APC5 subunits that are redundantly required for meiosis. Although we have not been able to generate a null for both APC5 subunits, a severe loss of the function of both of these genes together leads to meiotic metaphase arrest, suggesting that APC5 in C. elegans operates as a classical APC/C subunit. Our findings and the previous studies in fungi suggest that the function of Drosophila ida is divergent and that the canonical role for APC5 is in the metaphase-to-anaphase transition.

The two C. elegans APC5 paralogs, such-1 and gfi-3, both function during the meiotic divisions. Depletion of such-1 alone does not result in a penetrant one-cell–arrest phenotype, although we occasionally observe a few one-cell embryos from such-1(h1960ts) mothers at the restrictive temperature. gfi-3 RNAi in combination with a reduction of such-1 yields a brood almost exclusively composed of meiotic one-cell embryos, which further supports redundancy of these genes during the meiotic divisions. It is formally possible that such-1 is, in fact, the sole APC5 subunit required for the meiotic divisions and that such-1 is not completely depleted by the combination of the such-1(h1960ts) allele, the deficiency chromosome and such-1 RNAi; however, we consider this very unlikely.

The APC5 paralogs are likely not redundant at all cell divisions. such-1(h1960ts) exhibits some embryonic lethality, suggesting a nonredundant role for embryonic viability. Additionally, the such-1(h1960ts) loss-of-function allele was originally recovered as a suppressor of the deletion allele of the spindle checkpoint gene, MAD1/mdf-1 (Tarailo et al. 2007). If such-1 and gfi-3 were completely redundant, this allele would not be anticipated to suppress the deletion mutant phenotype on its own. It is likely that the two APC5 genes are partially redundant and that such-1(h1960ts) reduces the activity of the APC/C in a subset of somatic cells to allow survival of the mdf-1 deletion strain. The recovery of such-1(h1960ts) and the novel gain-of-function allele, such-1(av9gf), underscores the utility of genetic screens to isolate unique genetic alleles.

Why might there be two redundant APC5 genes in Caenorhabditis? Although there is an overlap in function at meiosis, our studies with transgenic animals indicate that larval and adult somatic expression is not equivalent. For example, GFI-3 may play a more prominent role during larval development, where it is the primary APC5 paralog expressed. Alternatively, it is possible that the activity of the APC/C containing either subunit is appreciably different. For this reason, one form of the APC/C may be preferable in certain cell types, for example, if it required a quicker-acting APC/C. However, it remains to be determined whether there are distinct APC5-containing complexes. Interestingly, C. elegans also has two paralogs of APC10. No comprehensive study has been done on APC10, so the redundancy of these subunits is not known. It is possible that the two APC10 subunits are also redundant at meiosis because RNAi of one of the subunits does not produce a one-cell meiotic-arrest phenotype (Davis et al. 2002). RNAi of the second paralog has not yet been reported.

What is the possible mechanism of suppression in mat-3(or180ts) such-1(av9gf) animals? Structural studies have recently produced models of how the APC/C is assembled, and these studies cast some light on the mechanism of suppression. APC5 sits at the junction of the two arms of the APC/C, one containing the target-binding TPR domain-containing proteins and the other the catalytic subunits (Thornton et al. 2006). In that study, APC5 and CDC23/APC8/MAT-3 were found to directly interact. The ability of such-1(av9gf) to strongly suppress mat-3(or180ts) to viability may be attributable to a compensatory modification of this interaction in the two mutants. Mutations in two other APC/C subunits, CDC27/APC3/mat-1 and APC4/emb-30, and emb-1, are suppressed at the meiotic stage by such-1(av9gf), but most embryos do not hatch. This may be due to subtle differences in strength and requirements of these alleles. mat-1(ax212ts) is defective in postembryonic divisions; these animals are sterile if shifted to the restrictive temperature at the L1 stage. Perhaps such-1(av9gf) is not sufficient to suppress the APC/C in this context. Additionally, while such-1(av9gf) may suppress the meiotic divisions, the fidelity of these divisions may be compromised and result in multicellular embryonic lethality due to aneuploidy. It is likely that the overall APC/C structure or assembly is disrupted in the mat-3(or180ts) mutants and the av9gf allele is able to compensate for the mat-3 mutant and hold the complex together more effectively, which might influence the ubiquitin ligase activity of APC/C. Exploration of the mechanism of suppression awaits future biochemical analysis of APC/C complexes bearing this APC5 mutation.


We thank Paula Fearon for critical reading of the manuscript and Risa Kitagawa for the such-1(h1960) allele. We also thank Kevin O'Connell for the “pedestal of rescue” technique for recovering transgenes from ballistic transformations as well as his useful suggestions for the manuscript. This research was supported by the Intramural Research Program of the National Institutes of Health (NIH), National Institute of Diabetes and Digestive and Kidney Diseases. Some nematode strains used in this work were provided by the Caenorhabditis Genetics Center, which is funded by the NIH National Center for Research Resources.


  • 1 Present address: Biological and Biomedical Sciences Program, University of North Carolina, Chapel Hill, NC 27599.

  • 2 Present address: Department of Molecular and Cellular Biology, University of Washington, Basic Sciences Division, Fred Hutchinson Cancer Research Center, Seattle, WA 98109.

  • Communicating editor: D. I. Greenstein

  • Received September 22, 2010.
  • Accepted October 11, 2010.


View Abstract