In this work we have examined for the first time in vivo spatiotemporal expression profiles of eight spindle-checkpoint genes in C. elegans. Among these eight genes, five are conserved from yeast to human (mdf-1, mdf-2, san-1, bub-1 and bub-3) [9–13], while three are conserved in higher eukaryotes (hcp-1, hcp-2 and rod-1), including C. elegans[12, 16, 45]. Our study focused on analysis of the expression patterns by using extra-chromosomal arrays. To maximally reduce the effect of mosaicism, the known caveat of this approach, we analyzed a large number of animals for each developmental stage, and recorded the tissues and cells where GFP expression was consistently observed. On the other hand, we found the mosaicism to be beneficial for a better identification of tissues where GFP is expressed. When promoters drive GFP expression in more than one tissue types, then expression restricted to only small groups of cells, due to loss of the array, offers more confident identification of these tissues. Also, GFP expression is a sensitive technique which is important for SAC gene expression analysis because generally SAC genes do not produce an abundant number of transcripts. Concatamer arrays were previously suggested to be a sensitive tool for detecting gene expression for genes with low levels of transcription . We confirmed the sensitivity of this approach when we generated a pmdf-2::GFP stable line using MosSCI . This stable line had very low GFP signal intensity and required long exposure times for the expression to be observed.
The 5' DNA sequences selected as containing putative promoters of the SAC genes displayed common early embryonic activities in the majority, if not all, of the rapidly dividing embryonic cells. This finding is consistent with the known roles of the checkpoint genes in cell division. We expected this result because of the fact that 556 of the 959 somatic cells present in adult hermaphrodite are generated during embryogenesis . Furthermore, our observations of early embryonic expression is supported by published analyses which used antibodies against some of the SAC gene products [9, 11, 15, 27, 45]. Thus, it is likely that these transcriptional fusions recapitulate endogenous SAC gene promoter activities. Importantly, this common "ubiquitous" expression of SAC genes (including mdf-1) during early embryogenesis, suggests that expression of mdf-1, the only one located within an operon, has to be driven by the internal promoter (Additional file 1, Figure S1). Thus, the mdf-1 containing operon is likely a "hybrid operon" .
czw-1 (known as ZW10 in other organisms) was also included in our study; however, analysis of two different constructs did not reveal any detectable GFP expression. It is possible that expression of the analyzed transgenes was either too low for visible detection, germline specific, conditional, or that regulatory elements of this gene are located in regions not included by our putative promoter selection criteria.
In contrast to expression in embryos, postembryonic expression of SAC genes in C. elegans is more localized. During the four larval stages in a hermaphrodite, the 53 undifferentiated somatic blast cells generate an additional 403 somatic nuclei . The somatic blast cell divisions generate somatic gonad, muscle, coelomocytes, nerves, hypodermis and intestine [30, 46, 47]. If all of the checkpoint genes played the same role in postembryonic development, one would expect to observe the same expression patterns for the SAC genes. However, our analysis revealed that checkpoint promoters generally dictate differential postembryonic expression patterns. For example, it is very interesting that mdf-1
and the rod-1 promoters drive GFP expression exclusively in intestine after embryogenesis, while the hcp-1 promoter drives GFP expression in multiple tissues (Table 1). These findings suggest distinct, yet overlapping, roles of the checkpoint genes in postembryonic development and provide an excellent resource for further research. Recently, staining of newly hatched L1 larva with anti-MDF-1 antibody revealed specific localization of MDF-1 to intestinal cells and germ cell precursors , which further supports our findings from using the transcriptional reporter system. We did not observe expression in germ cell precursors or any other germ cells possibly due to silencing of concatamer transgenes in the germinal gonad.
An unexpected finding from our analysis was tissue-specific expression of SAC genes in late L4 and adults that contain no somatic cells destined to divide. Considering that tissue-specificity observed in these stages was similar to the tissue-specificity observed in larval stages, it is possible that the observed patterns reflect longer turnover times for the GFP carried over from earlier larval stages . On the other hand, it is possible that 5' upstream sequences used in our analysis do not include important "repressor" elements that are required for proper expression of SAC genes. Alternatively, it may be that SAC genes have roles in these adult tissues that remain to be uncovered.
We have found that spindle-checkpoint genes reveal an intriguing co-expression in hypodermal seam cells. This finding prompted us to use the seam cell lineage to test the functional importance of the checkpoint for proper postembryonic cell proliferation. Here, we demonstrated that the knockout allele, tm2190, of mdf-2 results in defective seam cell development that is mainly attributed to seam cell proliferation failure at L2. In the absence of MDF-2, on average 14 seam cell nuclei were observed instead of expected 16. The number of SCM::GFP nuclei per side of an animal ranges from 8 to 19 in the absence of MDF-2 (Table 3). While the majority of the Δmdf-2 homozygotes contains less than expected 16 seam cell nuclei per side in young adults, we also observed animals that had more than 16 seam cell nuclei (Figure 5), which could be attributed to defective cell division. The results presented in this paper provide the first evidence that embryonic cell divisions are more tolerant to the loss of SAC, in particular MDF-2, than postembryonic cell divisions, as determined using the seam cell lineage. Furthermore, we show that the importance of MDF-2 for proper seam cell proliferation depends on its regulation of APC/CCDC20. The seam cell defect in Δmdf-2 homozygotes cannot be explained by cell damage followed by caspase-dependent apoptotic cell death, since ced-3 mutant had no effect on seam cell defect in Δmdf-2 worms. Furthermore, fzy-1(h1983) rescued all of the Δmdf-2 phenotypes, except for the brood size. On the other hand, G1 phase regulators, LIN-35 and FZR-1, when defective affect only brood size in the absence of MDF-2. The analysis presented here, using the Δmdf-2, serves as an excellent model for further studies on effects of a defective SAC on development of different tissues in a multicellular organism.
A striking emerging pattern is that essentially all SAC genes are expressed in intestine and hypodermis. SAC components MDF-2  and MDF-1  have previously been observed to be localized to gut cells by using antibody staining. Endoreduplication, also known as endoreplication, is a process in which S phases are not followed by mitosis. This process gives rise to cells with extra copies of chromosomes, permitting amplification of the genome in specialized cells. In humans, these include hepatocytes, cardiomyocytes and megakaryocytes . In C. elegans, two tissues are polyploid: the hypodermis and the intestine . Our finding of co-expression of SAC genes in these tissues may suggest a possible role of these genes in the process of endoreduplication in C. elegans. Furthermore, our findings clearly suggest that SAC genes are differentially regulated at the transcription level at different developmental stages.