Previously, it was shown that san-1(RNAi) embryos are viable, yet sensitive to conditions that depolymerize microtubules (anoxia or nocodazole treatment), providing evidence that the san-1 gene product has a role in the spindle checkpoint pathway 2122. To investigate the role san-1 has in C. elegans we obtained a san-1 deletion strain from the C. elegans Knockout Consortium. We verified this deletion by DNA sequence analysis and found that the san-1(ok1580) allele has a 992 kb deletion, which is consistent with what is documented in wormbase. The san-1 gene contains 8 predicted exons, of which the deletion in san-1(ok1580) removes all of exons 3–7 and the first 13 bp of intron 7. Thus, the san-1(ok1580) allele is minimally a catalytic null as the predicted kinase domain and Mad3/Bub1 homology region predicted to be essential for binding to CDC20p is absent (data not shown).

Others have shown that san-1(ok1580) hermaphrodites grown at 24°C have several phenotypes including a higher incidence of males in the population, egg laying defects, and a higher death rate in adult hermaphrodites due to vulva bursting or formation of a bag of worms 19. As prior work with san-1(RNAi) was done at 20°C, we characterized the phenotype of san-1(ok1580) animals grown at 20°C. Similar to animals maintained at 24°C, we found that the san-1(ok1580) embryos were viable, yet had some low penetrant phenotypes (Table 1). In comparison to wild-type animals, the san-1(ok1580) animals have a larvae arrest and slow growth phenotype (Table 1). The san-1(ok1580) adults have a higher incidence of death resulting from defects in vulva morphogenesis (Table 1). We further investigated these vulva defects using DIC microscopy to examine the vulva region of san-1(ok1580) animals and determined that the san-1(ok1580) animals contained tissue damage, abnormal vulva anatomy, and an increase in the protruding vulva phenotype (Figure 1B–D, Table 1). In comparison to wild-type adults, the average number of eggs held within the uterus of san-1(ok1580) animals is not significantly different (P = .176) (Table 1). However, the number of eggs held in the uterus was significantly different, ranging from 8 to 18 for wild-type animals, and 5 to 27 for san-1(ok1580) animals, indicating that some of the san-1(ok1580) animals were not laying embryos. Consistent with this idea, we observed that 18% (n = 16) of the san-1(ok1580) 1-day old adults had larvae hatching out in the uterus. We analyzed progeny production in san-1(ok1580) hermaphrodites and determined that the average number of progeny produced was significantly decreased in comparison to wild-type hermaphrodites (Table 1). Some of the san-1(ok1580) hermaphrodites died as young adults and produced less than 60 progeny, whereas others did not die, yet they still produced significantly less progeny than wild-type hermaphrodites (P = .001). The san-1(ok1580) animals also have a higher incidence of males (Table 1). Additionally, san-1(ok1580) males have various defects in tail morphology including an increase in tissue damage, and abnormal ray morphology or tissue structure (Figure 1F–H). In our studies we used the san-1(ok1580) males for mating indicating that the san-1(ok1580) males are not sterile. In summary, these results are consistent with defects in cell cycle progression leading to abnormal vulva development in the hermaphrodite and somatic cell defects in the adult male 28.

To further analyze the san-1 function in C. elegans we used RNAi to identify additional genes that genetically interact with san-1(ok1580). We were interested in determining if RNAi of specific gene products resulted in a more severe phenotype in the san-1(ok1580) background. This approach is similar in concept to a synthetic lethal screen, an established technique to identify genetic interactions in which the double mutant has a phenotype distinct or more severe than the phenotype(s) of the individual single mutants. The genes that we evaluated included those that are involved with kinetochore assembly or spindle checkpoint function and do not have severe embryonic lethality phenotypes. Thus, we analyzed hcp-1, hcp-2, the putative bub-3 homologue Y54G9A.6, the known spindle checkpoint genes, mdf-1 and mdf-2, and the predicted kinetochore protein kbp-5 18. The hcp-1/2 genes encode kinetochore-localized proteins required for kinetochore assembly, and animals only have a significant defect in viability when both HCP-1/2 are simultaneously depleted by RNAi 2325. The gene product of Y54G9A.6 has homology to the spindle checkpoint gene BUB-3 (see Additional file 1). BLASTp analysis of C. elegans Y54G9A.6 predicted gene product suggests there is homology with yeast BUB3 (E value = 4e-28), Homo sapiens BUB3 (E value = 3e-27), and D. malanogaster BUB3 (E value = 2e-74). The amino acid identity between the C. elegans Y54G9A.6 predicted gene product and BUB3 proteins is 44% for H. sapiens, 43% for D. melanogastor and 19% for Saccharomyces cerevisiae, as determined using the ClustalW software. Thus, the Y54G9A.6 gene will be referred to as bub-3 in this report. Other gene candidates likely to interact with san-1, such as other spindle checkpoint genes (bub-1), genes that encode products essential for kinetochore assembly and function (hcp-3, knl-1, knl-3, cls-2) and homologues to zw10 and rod (czw-1, rod-1), were not analyzed because RNAi of these gene products result in severe embryonic lethality, thus testing a synthetic lethal interaction with san-1(ok1580) is not possible. We chose to focus our analysis instead on genes bub-3, mdf-1, mdf-2, hcp-1, hcp-2 and kbp-5.

A reduction in the spindle checkpoint genes (san-1, mdf-1, mdf-2, bub-3) or the kinetochore assembly genes (hcp-1, hcp-2) individually did not result in severe embryonic lethality and the majority of animals developed into adults (Figure 2A). The mdf-1(RNAi) and mdf-2(RNAi) animals had several phenotypes as adults including death due to vulva bursting or formation of a bag of worms (data not shown), which is consistent with what others have observed 18. The mdf-1(gk2) deletion mutant has a significant embryo lethal phenotype 18, suggesting that the mdf-1(RNAi) acts as a null. In comparison to san-1(ok1580) animals, the san-1(ok1580);bub-3(RNAi) (P < .04), san-1(ok1580);mdf-1(RNAi) (P < .01) and san-1(ok1580);hcp-1(RNAi) (P < .0003) animals displayed a reduction in the ability for embryos to develop to adulthood (Figure 2B). Nearly all of the san-1(ok1580);hcp-1(RNAi) animals died as embryos or young larvae. The san-1(ok1580);hcp-2(RNAi) animals had a decrease in the ability to develop to adulthood, in comparison to san-1(ok1580) animals (P < .05), however the phenotype was not as a severe as the san-1(ok1580);hcp-1(RNAi) animals (Figure 2b). The san-1(ok1580);mdf-2(RNAi) had a reduced but not significant difference in viability (P = .073). We did not observe a viability defect in san-1(ok1580);kbp-5(RNAi) animals thus kbp-5 was not further analyzed (data not shown).

Because RNAi does not always completely remove all of a genes function, we used a genetic deletion of hcp-2 to gain a greater understanding of the genetic interaction between hcp-2 and san-1. We crossed the hcp-2(ok1757) and san-1(ok1580) animals to produce a san-1(ok1580);hcp-2(ok1757) animal. The hcp-2(ok1757) allele is a 1.281 kb deletion, which removes most of exon 5 resulting in a frame-shift; hcp-2 has 8 predicted exons (data not shown). We verified this deletion by DNA sequence analysis. We determined that 26.25% of the F2 offspring (n = 80) from a san-1(ok1580) and hcp-2(ok1757) cross had phenotypes including embryo lethality, larvae lethality, odd morphology and lethargy. The number of F2 progeny with abnormal phenotypes was higher than the expected 1/16 (6.25%) for a homozygous double mutant animal (san-1(ok1580);hcp-2(ok1757)), suggesting that some of the F2 progeny that are a combination of homozygous and heterozygous mutations for either san-1(ok1580) or hcp-2(ok1757) may have an abnormal phenotype. We genotyped 39 F2 animals, using single worm PCR to detect the san-1 and hcp-2 deletions, and determined that the homozygous double mutant (san-1(ok1580);hcp-2(ok1757)) and animals homozygous for the san-1(ok1580) deletion and heterozygous for the hcp-2(ok1757) deletion displayed embryonic and larval lethality phenotypes. The hcp-2(ok1757) animals appeared to be lethargic as adults, as noted by their reduction in motility, yet did not exhibit embryonic lethality and the majority reached adulthood three days after egg laying (Figure 2A). We isolated a san-1(ok1580);hcp-2(ok1757) double mutant and determined that the majority of the embryos hatched, yet only 7.3 % ± 2.4 of the larvae reached adulthood within three days at 20°C. However, if given four days to develop the majority of san-1(ok1580);hcp-2(ok1757) animals developed to adulthood (Figure 2b). Several of the san-1(ok1580);hcp-2(ok1757) animals had morphological defects suggesting that the double mutant had a more severe phenotype then the single mutants (data not shown). Our results indicate that the phenotype observed in san-1(ok1580);hcp-2(ok1757) is not identical to that observed in san-1(ok1580);hcp-1(RNAi) animals.

Nearly all of the hcp-1(RNAi);hcp-2(RNAi) embryos die during embryogenesis; however, over half of the san-1(ok1580);hcp-1(RNAi) or mdf-2(RNAi);hcp-1(RNAi) animals survive embryogenesis but die as larvae (Figure 2B, 2C). Thus, the embryo lethality of san-1(ok1580);hcp-1(RNAi) or mdf-2(RNAi);hcp-1(RNAi) animals is not as severe in comparison to the hcp-1(RNAi);hcp-2(RNAi) animals 23. In comparison to mdf-2(RNAi) animals, the mdf-2(RNAi);hcp-1(RNAi) animals had a significant decrease in the ability to develop and survive to adulthood (P < .011). The mdf-2(RNAi);hcp-2(ok1757) animals also had a decrease in the ability to survive to adulthood (P < .025), however the viability defect was not as severe as that observed in mdf-2(RNAi);hcp-1(RNAi) animals (Figure 2C). Similar to bub-3(RNAi) or hcp-1(RNAi) animals, the bub-3(RNAi);hcp-1(RNAi) animals did not have a viability defect (Figure 2C). Furthermore, the bub-3(RNAi);hcp-2(ok1757) (Figure 2C) and bub-3(RNAi);hcp-2(RNAi) (data not shown) animals did not have viability defects. Using RT-PCR, in comparison to wild-type, a greater than 5 fold reduction in bub-3 transcript level (P < .029) was observed in bub-3(RNAi) animals, thus demonstrating RNAi efficiency. Together, these results suggest that san-1 genetically interacts with bub-3, mdf-2, mdf-1, hcp-1 and hcp-2 and that C. elegans requires both hcp-1 and either of the checkpoint genes, san-1 or mdf-2 but not bub-3. Furthermore, these results suggest that hcp-1 and hcp-2 have redundant but not identical functions.

To further characterize the phenotype of the san-1(ok1580);mdf-2(RNAi), san-1(ok1580);bub-3(RNAi) and san-1(ok1580);hcp-1(RNAi) animals, we analyzed the animals for developmental defects. The majority of animals, that did not die as embryos, arrested as larvae or had an abnormal gonad as adults (Figure 3, Table 2). First, the san-1(ok1580);mdf-2(RNAi), san-1(ok1580);bub-3(RNAi) and san-1(ok1580);hcp-1(RNAi) animals had larvae arrest or slow growth phenotypes (Figure 3A). Additional defects observed in the san-1(ok1580);mdf-2(RNAi) or san-1(ok1580);bub-3(RNAi) animals included abnormal cell morphology in the mitotic region of the gonad, abnormal cell morphology in the meiotic region of the gonad, meiotic cells that had button like cell morphology reminiscent of apoptotic cells, and abnormal oocyte structure (Figure 3B). Some of the animals produced sperm but the spermatheca appeared to be abnormal in structure as determined by sperm localization (Figure 3B, black arrows). We also observed gonad migration defects (data not shown). The majority of san-1(ok1580);hcp-1(RNAi) animals died or arrested as larvae. Abnormal gonad morphology was not observed in the bub-3(RNAi) or hcp-1(RNAi) animals. Some of the san-1(ok1580) and mdf-2(RNAi) had abnormal gonad morphology (Table 2).

The hcp-1(RNAi);hcp-2(RNAi) animals die as embryos whereas many of the san-1(ok1580);hcp-1(RNAi) embryos hatch and later die as young larvae (Figure 2). To determine if the san-1(ok1580);hcp-1(RNAi) embryos have developmental defects we used the pharynx as a marker for developmental progression, by analyzing the expression of myo-2::GFP. Briefly, myo-2 encodes a muscle-type specific myosin heavy chain isoform and is expressed during embryogenesis in the primordial pharyngeal muscle cells. Similar to the control embryos (Figure 4A–D), the hcp-1(RNAi) (Figure 4E–H) and san-1(ok1580) (Figure 4I–L) animals had normal pharynx development. However, the san-1(ok1580);hcp-1(RNAi) embryos had either a reduction in myo-2::GFP (Figure 4M–N, Q–R) or abnormal pharynx structure (Figure 4O–P). Together, these results indicate that the san-1(ok1580);hcp-1(RNAi) embryos have developmental defects.

Our data suggest that the development of san-1(ok1580) animals depends upon bub-3 and hcp-1, however it is not clear if, like the san-1,mdf-1 and mdf-2 genes, the bub-3 and hcp-1 gene products function in the spindle checkpoint pathway. It is known that blastomeres of wild-type embryos exposed to oxygen deprivation (anoxia) reversibly arrest cell cycle progression 2129. Unlike wild-type embryos, san-1(RNAi) and mdf-2(RNAi) embryos are sensitive to the microtubule disrupting environments of anoxia or nocodazole treatment. Specifically, san-1(RNAi) and mdf-2(RNAi) embryos exposed to anoxia have a significantly reduced number of metaphase blastomeres and an increase in anaphase bridging and chromosomal abnormalities; supporting the idea that san-1 and mdf-2 function as spindle checkpoint genes 2122. Similar to what is known regarding the san-1(RNAi) embryos, the san-1(ok1580) embryos exposed to anoxia had a reduced survival rate (Table 3). The bub-3(RNAi) embryos exposed to anoxia also had a reduced survival rate in comparison to wild-type animals suggesting that bub-3 functions as a spindle checkpoint protein in developing embryos. The hcp-1(RNAi) embryos were not sensitive to anoxia (Table 3). Additionally, we did not observe a sensitivity to anoxia in hcp-2(RNAi) or hcp-2(ok1757) embryos (data not shown). These results suggest that hcp-1 and hcp-2 are not required for anoxia induced metaphase arrest and do not function as a spindle checkpoint protein.

To gain a greater understanding of the functional relationship between the hcp-1 and san-1 gene products in developing embryos we determined if chromosome loss was occurring in the san-1(ok1580);hcp-1(RNAi) embryos. We analyzed chromosome structure in mitotic blastomeres from wild-type, hcp-1(RNAi), san-1(ok1580), and san-1(ok1580);hcp-1(RNAi) embryos collected, fixed and stained as previously described 22. We used mAb414, which recognizes the nuclear pore complexes and anti-PhosH3, which recognizes the phosphorlyated form of histone H3 to distinguish mitotic blastomeres 22293031. We found that unlike wild-type, san-1(ok1580), and hcp-1(RNAi) embryos (Figure 5A–L), the san-1(ok1580);hcp-1(RNAi) embryos had a significant increase in abnormal nuclei (44%, n = 50) (Figure 5M–T). These abnormalities included abnormal mitotic structure, lagging chromosomes, and anaphase bridging (Figure 5M, 5Q, arrows). These defects were also observed in young embryos (2–4 cell embryos; 15.4% (n = 13)), indicating that the chromosome segregation defects occur early in embryo development.

In young wild-type embryos (less than 30 blastomeres) the nuclear pore complexes detected by mAb 414 do not diminish during mitosis 31. In contrast, embryos with greater than 30 blastomeres, the nuclear pore complexes detected by mAb 414 diminish from prometaphase to anaphase. Similar to wild-type embryos (Figure 5A–D), with greater than 30 blastomeres, the nuclear pore complexes, detected by mAb414, were diminished in metaphase blastomeres of hcp-1(RNAi) (Figure 5E–H), and san-1(ok1580) embryos (Figure 5I–L). However, in san-1(ok1580);hcp-1(RNAi) embryos with greater than 30 blastomeres, the nuclear pore complex, detected by mAb414, were in small aggregates surrounding the metaphase plates (72%, n = 11) (Figure 5Q–T, 5S arrow head). This phenotype can only be observed in the older embryos, since the younger embryos do not have diminished nuclear pore complexes. The accumulation of abnormal nuclei and aggregates of the nuclear pore complex suggests that progression through mitosis was aberrant in the san-1(ok1580);hcp-1(RNAi) embryos.

To gain a greater understanding of mitotic progression in the san-1(ok1580);hcp-1(RNAi) embryos we analyzed chromosome, kinetochore and spindle microtubule structure using indirect immunofluorescence. We first stained embryos with MPM-2, an antibody that recognizes mitotic proteins localized to the kinetochore and centrosomes 32. The kinetochore structure in san-1(ok1580) and hcp-1(RNAi) embryos was similar to that of wild-type embryos (Figure 6C,G,K). However, the san-1(ok1580);hcp-1(RNAi) embryos contained both normal and abnormal kinetochore structure, as determined by the presence or absence of the two-line formation along the sister chromatids (Figure 6O, 6S). The abnormal kinetochore structure correlated with abnormal mitotic structure.

We used anti-HCP-3 and YL1/2 to analyze the localization of the centromeric histone HCP-3 (CENP-A) and spindle microtubules, respectively 2326. HCP-3 is at the top of the kinetochore assembly pathway 2325. The centromere structure and spindle microtubules in san-1(ok1580) and hcp-1(RNAi) embryos were similar to that of wild-type embryos (Figure 7A–L). The san-1(ok1580);hcp-1(RNAi) blastomeres with normally formed metaphase plates had normal centromere and microtubule structure (Figure 7M–P). The san-1(ok1580);hcp-1(RNAi) blastomeres with abnormal chromosome structure contained HCP-3 localized with chromosomes (Figure 7U–X); yet in some blastomeres HCP-3 was not always associated with the chromosomes (Figure 7Q–T). In these abnormal blastomeres the microtubule structure was abnormal (Figure 7Q–X). Embryos analyzed for these experiments contained ~20–30 blastomeres so it is possible that these defects are due to altered gene expression resulting from chromosome segregation defects.

Given the presence of anaphase bridging, lagging chromosomes, and abnormal chromosome structure observed in the san-1(ok1580);hcp-1(RNAi) animals, it is likely that the blastomeres were transitioning through mitosis abnormally. To determine if this occurred in vivo we produced a san-1(ok1580);tbg-1::GFP;pie-1::GFP::H2B strain (PM115 strain, histone and gamma-tubulin GFP-tagged proteins) for live cell imaging using a spinning disc confocal microscope. The san-1(ok1580);hcp-1(RNAi);tbg-1::GFP;pie-1::GFP::H2B embryos contained blastomeres with abnormal mitotic progression. Analysis of a typical san-1(ok1580);hcp-1(RNAi) embryo showed one blastomere which progressed to anaphase with normal chromosome segregation, and another blastomere with abnormal chromosome segregation leading to anaphase bridging (see Additional file 2). These results further demonstrated that abnormal chromosome segregation occurred in the san-1(ok1580);hcp-1(RNAi) embryos.

Thus far, we have shown that the san-1(ok1580);hcp-1(RNAi) embryos abnormally segregate chromosomes which likely leads to developmental defects and lethality. The hcp-1(RNAi) embryos have not been shown to have chromosome segregation issues that lead to increased lethality. One possible reason for the synthetic lethal interaction between san-1 and hcp-1 is that loss of san-1 amplifies a subtle hcp-1 defect. Given the defect in progression through mitosis observed with mAb 414, timing of mitotic progression is possibly perturbed when hcp-1 is absent and loss of san-1 increases this defect. Therefore, we conducted live cell imaging to compare mitotic progression in hcp-1(RNAi);tbg-1::GFP;pie-1::GFP::H2B and control animals. We analyzed mitosis in the AB and P cell blastomeres of a 2-cell embryo. We determined the time it took to transition from nuclear envelope break down (NEB) to formation of the metaphase plate, metaphase to anaphase onset and NEB to anaphase onset in the P cell of a wild-type and hcp-1(RNAi) 2-cell embryo. We also determined the time between NEB of the AB cell and NEB of the P cell. We found that the mitotic progression of hcp-1(RNAi) embryos analyzed at 21°C did not have any chromosomal defects and the time it took to progress through mitosis was not significantly different in comparison to control embryos (Figure 8). Together, our data indicates that hcp-1(RNAi) animals require a functional spindle checkpoint for normal development.

The use of synthetic lethal screens is a genetic tool to elucidate genes that interact in similar cellular functions or pathways. Thus, when a genetic mutant does not have a severe phenotype the use of a synthetic lethal screen to identify double mutants that have a phenotype distinct or more severe then the phenotype(s) observed for the individual single mutants often will aid in further understanding the function of the gene products. To analyze san-1 function in C. elegans embryos we used RNAi to identify genes that could genetically interact with san-1(ok1580). We determined that san-1(ok1580);hcp-1(RNAi), san-1(ok1580);bub-3(RNAi), san-1(ok1580);mdf-1(RNAi), and san-1(ok1580);mdf-2(RNAi) had a reduced ability to survive. The genetic interaction between san-1, mdf-1, and mdf-2 is expected, since these genes are thought to function in the spindle checkpoint pathway in mitotic and meiotic cells 18192021. The predicted C. elegans bub-3 gene product has high homology to BUB-3 from other systems and bub-3(RNAi) embryos are more sensitive to the microtubule disrupting process of anoxia exposure, supporting the idea that the Y54G9A.6 gene encodes the bub-3 spindle checkpoint gene.

The genetic interaction between hcp-1/2 and the spindle checkpoint genes is likely to be more complex. The viability of san-1(ok1580);hcp-1(RNAi) and mdf-2(RNAi);hcp-1(RNAi) animals was severely compromised, suggesting that HCP-1 interacts with MDF-2 and SAN-1. However, the bub-3(RNAi);hcp-1(RNAi) animals did not have a significant decrease in survival rate. This could be due to BUB-3 and HCP-1 not involved together in a specific function. The san-1(ok1580);hcp-2(ok1757) animals have embryonic and larvae lethal phenotypes. We isolated a san-1(ok1580);hcp-2(ok1757) double mutant and determined that it had more severe phenotypes than that observed in the single mutants; this suggests that hcp-2 and san-1 genetically interact. However, the phenotype of the san-1(ok1580);hcp-2(ok1757) double mutants was not as severe as that observed san-1(ok1580);hcp-1(RNAi) animals, suggesting that HCP-1 and HCP-2 are not be completely redundant. HCP-1 and HCP-2 are 54% similar mostly at their N- and C-termini 23. Although HCP-1 shares some similarity to human CENP-F, primarily in a tandem repeat present in both HCP-1 and CENP-F, HCP-2 lacks significant similarity to CENP-F 23. Thus, it is possible that HCP-1 and HCP-2 have overlapping and distinct functions. It will be of interest to determine the specific molecular interactions between HCP-1, HCP-2, SAN-1 and MDF-2.

The majority of san-1(ok1580);hcp-1(RNAi) animals had severe developmental defects including embryo and larvae lethality. The viability defects observed are likely due to abnormal chromosome segregation, which will in turn compromise cellular structure and function. The san-1(ok1580);bub-3(RNAi) and san-1(ok1580);mdf-2(RNAi) animals had gonad defects. We and others have shown that san-1(ok1580) and mdf-2(RNAi) animals have low level gonad defects 1819; yet these defects are much more severe if two spindle checkpoint genes are reduced. We observed a reduction in brood size for san-1(ok1580) hermaphrodites and others have shown that mdf-2(av14) mutants have a reduced brood size, further supporting the role the checkpoint genes have in germline function 19. Others have shown that hcp-1(RNAi);hcp-2(RNAi) animals also have meiotic defects, supporting the idea that hcp-1 has a role in meiosis 27. The exact molecular function the spindle checkpoint proteins and HCP-1/2 have in meiosis needs to be further investigated.

The san-1(ok1580);hcp-1(RNAi) embryos have severe chromosome segregation defects, and on average 55.3% of the embryos die and approximately 0.9% of the animals reach adulthood. Thus, in san-1(ok1580);hcp-1(RNAi) embryos, embryogenesis can progress, but the ability to produce viable adults is severely compromised. We used myo-2::GFP to analyze the structure of the developing pharynx in the san-1(ok1580);hcp-1(RNAi) embryos and found that there was not only abnormal pharynx morphology but we often observed a reduction in the myo-2::GFP in the embryos. It is not known if the reduction in myo-2::GFP in the san-1(ok1580);hcp-1(RNAi);myo-2::GFP animals is due to abnormal differentiation, the loss of myo-2::GFP DNA or abnormal regulation of the myo-2::GFP. An increase in genome loss due to chromosome segregation issues could result in either of these possibilities.

Closer examination of the chromosomes, nuclear pore proteins, centromere, kinetochore, and microtubules indicates that the san-1(ok1580);hcp-1(RNAi) embryos have chromosome segregation defects leading to abnormal nuclei, anaphase bridging and lagging chromosomes. We determined that the nuclear pore complexes, detected by mAb414, formed small aggregates surrounding the metaphase plates in san-1(ok1580);hcp-1(RNAi) embryos, indicating that the nuclear membrane pore complexes are not completely breaking down. This may indicate that SAN-1 and HCP-1 influences additional mitotic events in addition to chromosome segregation. The centromeric protein HCP-3 was observed to have a normal or an abnormal localization pattern in the san-1(ok1580);hcp-1(RNAi) blastomeres. The abnormal localization of HCP-3 can be interpreted in several ways. First, it could be due to SAN-1 and HCP-1 directly regulating the localization of HCP-3 to the region that marks the centromere, however this seems unlikely since the kinetochore assembly pathway clearly places HCP-3 upstream of HCP-1 and SAN-1. Alternatively, the abnormal HCP-3 localization could be due to chromosome segregation defects causing loss of chromatin that marks the centromere. Finally, it is possible that the segregation defects result in chromatin fragments that HCP-3 associates with. Further studies would be required to determine the mechanism leading to abnormal HCP-3 localization.

In san-1(ok1580);hcp-1(RNAi) embryos, the kinetochore proteins recognized by the MPM-2 antibody were not associated with the chromosomes when the chromosome structure was aberrant. Furthermore, the microtubule structure was altered in blastomeres with abnormal chromosome structure. Together, these data suggests that in the san-1(ok1580);hcp-1(RNAi) blastomeres the kinetochore and microtubule interaction is sometimes, but not always, compromised. Given the increase in anaphase bridges, lagging chromosomes and abnormal mitotic nuclei, the microtubule and kinetochore interaction in san-1(ok1580);hcp-1(RNAi) animals is likely not sufficient for proper chromosome segregation.

Others have shown using mammalian cell culture (HeLa cells) that there is a prolonged mitosis in CENP-F (RNAi) cells and this prolonged mitosis is dependent upon the spindle checkpoint gene BUBR1 33. In C. elegans hcp-1(RNAi);hcp-2(RNAi) blastomeres take longer to progress through mitosis 20. Our data suggest that hcp-1(RNAi) blastomeres of embryos analyzed at 21°C did not take longer to progress from prometaphase to metaphase, in comparison to control blastomeres, suggesting that mitotic timing of hcp-1(RNAi) blastomeres is normal as long as hcp-2 gene product is present.

CENP-F is required for chromosome alignment 34 and silencing of CENP-F in HeLa cells activates the spindle checkpoint 33. The misalignment defect is thought to be due to kinetochores that do not correctly biorient rather than a defect in microtubule and kinetochore attachment 33. Our results support the idea that in hcp-1(RNAi) animals the spindle checkpoint is being activated to ensure proper chromosome segregation and genome defects will occur without normal spindle checkpoint function; specifically SAN-1 and MDF-2 function are required. Given that the involvement of the spindle checkpoint in tumor progression, it is of interest to fully understand the genes products that interact with the spindle checkpoint pathway to ensure genome fidelity and maintenance.

The wild-type Bristol strain (N2) and mutant strains were cultured on NGM plates seeded with E. coli (OP50) and raised at 20°C as described (Sulston and Hodgkin, 1988). The following strains were obtained from the Caenorhabditis elegans Genetics Center: RB1391(san-1(ok1580)), PD4790 (myo-2::GFP), TH32 (tbg-1::GFP, pie-1::GFP::H2B), and RB1492 (hcp-2(ok1757)). The Caenorhabditis elegans Gene Knockout Consortium (Oklahoma Medical Research Foundation) produced the san-1(ok1580) and hcp-2(ok1757) deletion alleles. The hcp-2(ok1757) mutant (strain RB1492) was back-crossed to wild-type animals three times (strain PM117). The san-1(ok1580) allele was crossed, using standard genetic techniques, to produce the following strains: PM107 (san-1(ok1580);myo-2::GFP), PM115 (san-1(ok1580);tbg-1::GFP;pie-1::GFP::H2B), and PM116 (san-1(ok1580);hcp-2(ok1757)). The san-1(ok1580) allele was backcrossed into wild-type background once. The san-1(ok1580) and hcp-2(ok1757) alleles were genotyped by single worm PCR with the following primers: san-1 forward primer (CGC TTA AAG CTT GAT CAA CTT CTCG), san-1 reverse primer (GCT AGT GAT TTC TCC TCC GTT TTC TCA), hcp-2 forward primer (ACT CTG AAG TCG GAA CAT GAA ATT), and hcp-2 reverse primer (TGA AGA GCC TTC TGT GCA AA). DNA sequence analysis to verify the deleted region of the san-1(ok1580) and hcp-2(ok1757) strains was conducted using standard molecular techniques.

The BUB-3 protein from the yeast Saccharomyces cerevisiae was used to conduct BLASTp against the C. elegans, Homo sapiens and D. melanogastor genome to identify the most identical protein for each species. The primary amino acid sequences were aligned using ClustalW and processed using the LaserGene sequence analysis software package (DNASTAR). Clustal analysis between the C. elegans Y54G9A.6 gene product (referred to as BUB-3) and the S. cerevisiae, H. sapiens and D. melanogastor BUB-3 proteins was conducted to determine percent identity between the proteins.

For all experiments animals were kept at 20°C and grown on E. coli (OP50). To assay for the high incidence of males (him) phenotype several hermaphrodites, grown in the absence of males, were allowed to lay eggs for several hours and the percent of male progeny was determined (for wild-type and san-1(ok1580), n = 244, n = 325, respectively). The san-1(ok1580) animals were assayed for larval arrest and slow growth phenotypes (for wild-type and san-1(ok1580), n = 210, n = 218, respectively). To determine if larval arrest or slow growth was observed we collected a synchronized population of L1 larvae and placed on OP50 NGM plates at 20°C for three to five days. We quantified the number of animals that developed to adulthood within three days, animals that developed to adulthood within three to five days (slow growth) and those that remained arrested as larvae (larval arrest). To assay for adult death phenotype, animals were grown for the first 5 days of adulthood and analyzed every day for death or survivorship (for wild-type and san-1(ok1580), n = 210, n = 207, respectively). Hermaphrodites were moved daily to fresh plates and lost worms were eliminated from analysis. The average brood size was determined by quantifying the number of offspring individual hermaphrodites produced (for wild-type and san-1(ok1580), n = 5, n = 10, respectively). To quantify the number of eggs within the uterus we collected 1-day old adults, placed on an agar pad and visualized using a Zeiss compound microscope; the number of eggs within the uterus was determined (for wild-type and san-1(ok1580), n = 16). To assay for protruding vulva phenotype we collected 1-day old adults and counted the number of adult animals that had a protruding vulva (for wild-type and san-1(ok1580), n = 210, n = 207, respectively). The p values (P) were determined using standard student t test.

RNA interference (RNAi) was used to decrease the expression of specific genes. For all experiments, the methodology was similar to that previously described (Hajeri et al). Briefly, a synchronous population of L1 larvae (N2, san-1(ok1580)) were grown to adulthood on NGM plates supplemented with 200 μg/ml ampicillin, 12.5 μg/ml tetracycline and 1 mM IPTG, seeded with a bacterial strain expressing dsRNA specific for mdf-2, mdf-1, bub-3, hcp-1, hcp-2 or control food 35. Several adults were then placed on a secondary RNAi plate and allowed to lay eggs for 3–4 hours. The adults were removed and the embryos were grown at 20°C and assayed for embryonic lethality, larvae lethality, and morphological defects as seen by Nomarski microscopy using a motorized Zeiss Axioskop and imaged using Openlab 3.17. For each viability experiment at least three independent assays were conducted. The P-value (P) was determined using a standard student t-test. In addition to using the RNAi food from the MRC, 35 we also produced RNAi food specific for hcp-1 and hcp-2. Briefly, plasmid vectors for feeding RNAi were constructed by cloning either a 1.1 kb Spe I – Dra I fragment from the hcp-1 cDNA LM46-3 or a 1.6 kb Sal I fragment from the hcp-2 cDNA, yk19h10 into pPD129.36, which allows IPTG induction of double stranded RNA. Respective feeding vectors were transformed into the HTH115(BL23) E. coli strain and fed to animals, consistent with previously described RNAi methodology 35. To conduct double RNAi experiments, both RNAi foods were plated, in equal amounts, onto the NGM IPTG plates. As a control we grew the E. coli strain, which has a plasmid with no insert, on identical NGM IPTG plates to account for any differences attributable to the vector. We conducted RT-PCR experiments to analyze the efficiency of RNAi by quantifying the level of bub-3 transcript in wild-type and bub-3(RNAi) animals. Methodology is as previously described 36; we conducted at least three independent experiments and each experiment was performed in triplicate. Primer sequences used for RT-PCR reactions of bub-3 are as follow: 5'TGGGATCCTTTCAATCGGAAGC3' and 5'TTATTTCGGTCTGCTCCGGA3'. The P value was determined using the Mann-Whitney U Test.

To assay whether specific genes are required for spindle checkpoint activity we exposed embryos to anoxia as previously described 2129. Briefly, the wild-type and san-1(ok1580) animals were grown, from the L1 larval stage to the 1 day old adult stage, on RNAi food (vector with no insert as a control or the specific strain to reduce expression of bub-3, hcp-1, or hcp-2). The adults were placed on a fresh RNAi plate and allowed to lay eggs for 1–2 hours. The adults were removed and the embryos were placed into anoxia, for 1 day, using the BioBag type A environmental chamber. The embryos were allowed to recover in air and assayed 24 hours later for the ability to hatch and three days later for the ability to reach adulthood. Four independent experiments with at least a total of 200 embryos were analyzed.

The L1 larvae (N2, san-1(ok1580), or san-1(ok1580);myo-2::GFP) were grown to adulthood on appropriate RNAi plates (control, bub-3, or mdf-2 food). To analyze the developing pharynx in embryos, both the san-1(ok1580);myo-2::GFP and the san-1(ok1580);hcp-1(RNAi);myo-2::GFP adult animals were dissected and the embryos were placed onto 2.5% agarose pads for microscopy analysis. To analyze the morphology of control, mdf-2(RNAi), hcp-1(RNAi), san-1(ok1580), san-1(ok1580);hcp-1(RNAi), san-1(ok1580);mdf-2(RNAi), and san-1(ok1580);bub-3(RNAi) animals, RNAi was conducted as stated above and the adult animals were placed on fresh RNAi plates and allowed to lay eggs for several hours. The embryos developed for three days at 20°C before DIC microscopy analysis. For all experiments the animals were placed on a 2.5% agarose pad and visualized using a motorized Zeiss Axioskop Fluorescent microscope and imaged using Openlab 3.17. At least three independent experiments were conducted.

Wild-type and san-1(ok1580) animals were grown on appropriate RNAi food from L1 stage to adulthood. The adults were collected and dissected to release embryos. Embryos were collected, fixed and stained as previously described 2229. Briefly, the Phos H3 antibody recognizes the phosphorylated (Ser10) form of Histone H3 specific for mitotic cells (Upstate Biotechnology, Lake Placid NY)30; the mAb414 recognizes the nuclear pore complex (Babco, Berkeley, CA) 37; mAb MPM-2 recognizes mitotic proteins located at the kinetochore, centrioles and P granules (DAKO, Carpinteria, CA) 32; anti-HCP-3 detects HCP-3 26; and YL1/2 detects the spindle microtubules (Amersham Life Science, Little Chalfont, Buckinghamshire, England). Microscopy was done using either a Zeiss Axioskop or spinning disc confocal microscope (McBain, CA). Images were collected using the spinning disc confocal microscope and processed using Image J and Adobe Photoshop version 8.0.

To analyze mitotic progression the chromosomes and centriole were used as markers, using the TH32 (tbg-1::GFP, pie-1::GFP::H2B) or PM115 (san-1(ok1580);tbg-1::GFP;pie-1::GFP::H2B) strain. L1 larvae were grown on RNAi control or hcp-1 RNAi food to adulthood at 20°C. 1-day old gravid adults were dissected and embryos of the appropriate stage were collected by mouth pipet and placed on a 2.5% agarose pad. The temperature of the room in which the embryos were analyzed was at 21°C. Mitotic progression was quantified for 2-cell embryos using a spinning disc confocal microscope. To examine mitotic progression, the time it took, in seconds, to transition from the onset of prometaphase (as seen by nuclear envelope breakdown and condensed chromosomes) to the formation of the metaphase plate, and from the formation of the metaphase plate to the onset of anaphase was determined. Seven embryos were analyzed, using the Simple PCI Version 6 software program. The p-value (P) was determined using a standard student t-test. Movies, to demonstrate mitotic progression in the san-1(ok1580);tbg-1::GFP;pie-1::GFP::H2B embryos, were obtained by analyzing live, young embryos using a spinning disc confocal microscope. Images were collected, processed using NIH Image, and imported into Quick Time for display.