Suspended Animation in C. elegans Requires the Spindle Checkpoint

See allHide authors and affiliations

Science  07 Nov 2003:
Vol. 302, Issue 5647, pp. 1038-1041
DOI: 10.1126/science.1089705


In response to environmental signals such as anoxia, many organisms enter a state of suspended animation, an extreme form of quiescence in which microscopically visible movement ceases. We have identified a gene, san-1, that is required for suspended animation in Caenorhabditis elegans embryos. We show that san-1 functions as a spindle checkpoint component in C. elegans. During anoxia-induced suspended animation, embryos lacking functional SAN-1 or a second spindle checkpoint component, MDF-2, failed to arrest the cell cycle, exhibited chromosome missegregation, and showed reduced viability. These data provide a model for how a dynamic biological process is arrested in suspended animation.

Adverse environmental conditions such as extreme temperature, decreased nutrient availability, or anoxia cause some organisms, including mammals (1), fish (2, 3), and invertebrates (4, 5), to enter a reversible state of suspended animation. For instance, nearly 100 different mammals enter into diapause, in which maternal cues induce an arrest of embryogenesis that can last for several months (1). Other examples of quiescence include hibernation (6) and estivation (7), which are programs that allow organisms to survive harsh seasonal conditions. In the zebrafish (Danio rerio), exposure to anoxia (operationally defined as <0.001 kPa O2) rapidly leads to a complete arrest of cell division, developmental progression, movement, and heartbeat. Upon reoxygenation, these biological functions are restored. Adult mammals can enter into suspended animation under conditions in which exsanguination restricts oxygen availability in the tissues (8, 9). The nematode C. elegans can enter into anoxia-induced suspended animation from any stage in the life cycle and remain suspended for several days with high viability (10).

Severe oxygen deprivation is likely to have broad physiological effects (11), and the mechanism of suspension probably involves a complex interaction of several different factors. One contributing factor may be the passive consequences of energy depletion that result from anoxia-induced inhibition of aerobic metabolism. Another possibility is that there is an active program in which adaptation to decreased oxygen tension is coordinated by gene products that direct the cessation of individual biological processes (12). This second possibility is especially attractive, considering the enormous degree of cellular organization that must be maintained in order for the organism to survive while suspended. If there is an active program, it should be possible to find gene products that are required for coordination of the suspended animation process. One possible candidate would be the transcription factor HIF-1 (hypoxia-inducible factor 1) (13, 14), which is activated at low oxygen tensions and promotes adaptation to hypoxia. However, whereas HIF-1 is required for survival of hypoxia in C. elegans (15), HIF-1 is not required for survival in anoxia (10). Here, we present an anoxia-specific mechanism by which a component of suspended animation is activated.

To identify genes required for suspended animation in C. elegans, we performed an RNA interference (RNAi)–based screen of 2445 open reading frames (ORFs) on chromosome I (16). Scoring for anoxia-specific lethality revealed an ORF (ZC328.4) that we have named san-1 (for suspended animation–1). Upon exposure to anoxia for 24 hours, san-1(RNAi) embryos exhibited a significantly lower viability compared to control embryos (Fig. 1). We observed high viability in san-1(RNAi) and control embryos that were either left in normoxia or exposed to hypoxia (0.5 kPa O2) for 24 hours, indicating that the lethality in the san-1(RNAi) strain is specific to anoxic exposure (Fig. 1).

Fig. 1.

san-1(RNAi) and mdf-2(RNAi) embryos exhibit anoxia-specific lethality. Viabilities to adulthood were assayed after exposure to 24 hours of normoxia (21 kPa O2, 21% O2), hypoxia (0.5 kPa O2, 0.5% O2), or anoxia (<0.001 kPa O2, <0.001% O2) in san-1(RNAi), mdf-2(RNAi), or control (RNAi) worms. Viability to adulthood was assayed after reexposure to normoxia; worms that could not be accounted for were dropped from the total. All data points are the result of at least three independent experiments, and significant differences between viabilities in anoxia were determined by pairwise comparisons of wild type and san-1(RNAi) and wild type and mdf-2(RNAi) using the Student's t test. In both cases, P < 0.05.

san-1 encodes a protein that shares 27% sequence identity with the yeast Saccharomyces cerevisiae spindle checkpoint component Mad3p, raising the possibility that SAN-1 is a component of the spindle checkpoint in C. elegans. SAN-1 was detected in the nucleus during prophase (17) and localized to the poleward faces of chromosomes in metaphase, overlapping with the kinetochore marker HCP-3 (18) (Fig. 2). This staining pattern is consistent with that of other spindle checkpoint components, including MDF-2 in C. elegans (19), Mad3p in S. cerevisiae (20), and the Mad3p homolog hBubR1 in human cells (21). Staining with antibody against SAN-1 was absent in san-1(RNAi) embryos, indicating that staining was specific for the SAN-1 protein (17). Treatment with nocodazole, a spindle depolymerizing drug, also induced spindle checkpoint activity in early embryos, and both san-1 and the previously identified spindle checkpoint component mdf-2 were required for this process (22). To test the requirement for an additional spindle checkpoint component in suspended animation, we examined the viability of mdf-2(RNAi) embryos in anoxia. As with san-1(RNAi), mdf-2(RNAi) embryos exhibited significantly reduced survival to adulthood in anoxia but not in normoxia or hypoxia (Fig. 1). The requirement for san-1 and mdf-2 in anoxia indicates that spindle checkpoint components are required for suspended animation.

Fig. 2.

SAN-1 protein localizes to the poleward faces of the metaphase plate. A wild-type embryo (A to D) and a high-resolution image of a single metaphase plate (E to H) stained with 4′,6′-diamidino-2-phenylindole (DAPI) (A and E), antibody against HCP-3 (B and F), and antibody against SAN-1 (C and G). (D) and (H) show the merged image of all three channels. White scale bar represents 5 μm.

This requirement of spindle checkpoint components for survival in suspended animation led us to examine whether the spindle checkpoint is active in wild-type embryos in anoxia. In wild-type embryos, the percent of mitotic blastomeres in metaphase increased from 18.2% in normoxia to 42.9% in anoxia (Table 1; also see Fig. 3 for an example of a typical wild-type anoxic embryo). In contrast, the percentages of mitotic blastomeres in metaphase decreased to 0.7% and 0.4% in san-1(RNAi) and mdf-2(RNAi) embryos, respectively (Table 1). Furthermore, whereas anoxic wild-type embryos contained no blastomeres in anaphase, anoxic san-1(RNAi) and mdf-2(RNAi) embryos did contain blastomeres in anaphase, many of which appeared abnormal. This result demonstrates that san-1 and mdf-2 are required for the increase in frequency of blastomeres in metaphase normally seen in anoxic wild-type embryos and, therefore, that the spindle checkpoint is activated in suspended animation.

Fig. 3.

Anoxic wild-type embryos have abundant metaphases, whereas anoxic san-1(RNAi) embryos have abundant anaphases. Anoxic wild-type embryos have a high percentage of blastomeres in metaphase and no blastomeres in anaphase. In contrast, anoxic san-1(RNAi) embryos have a high percentage of blastomeres in anaphase and few blastomeres in metaphase. A representative example of an anoxic wild-type embryo (A) and an anoxic san-1(RNAi) embryo (B) is shown. White arrows indicate three examples of blastomeres in metaphase in (A) and three examples of blastomeres in anaphase in (B), though in both cases other examples can be found. White scale bar represents 5 μm.

Table 1.

Mitotic stage of san-1(RNAi) and mdf-2(RNAi) embryos exposed to anoxia. Three independent experiments with a total of at least 280 mitotic blastomeres were evaluated. The average number of blastomeres in early embryos that are in mitosis is 27.4% (n = 100 embryos). Abnormal mitotic nuclei include anaphase bridging and abnormal telophase nuclei.

Normoxia Anoxia
Control san-1(RNAi)mdf-2(RNAi) Control san-1(RNAi)mdf-2(RNAi)
Prophase 60.8% 50.6% 53.1% 55.9% 66.9% 72.9%
Prometaphase 4.3% 4.1% 3.2% 1.7% 0.8% 0%
Metaphase 18.2% 20.3% 16.2% 42.9% 0.7% 0.4%
(P > 0.05) (P > 0.05) (P < 0.001) (P < 0.001)
Anaphase 10.1% 11.4% 11.7% 0% 0.2% 0%
Telophase 6.7% 11.4% 17.7% 0% 0.6% 0%
Abnormal 0% 0% 0% 0.2% 30.7% 26.7%
(P < 0.001) (P < 0.001)

Because the spindle checkpoint is normally activated in anoxia, we expected that progression through metaphase during the anoxic treatment in blastomeres of embryos without a functional checkpoint would result in missegregation of chromosomes in anaphase. We observed defects at anaphase in san-1(RNAi) and mdf-2(RNAi) anoxic embryos (Fig. 4, A and B). We did not find these defects in normoxic embryos, indicating that the missegregation phenotype is specific to anoxic exposure. To quantify the degree to which aneuploidy was occurring in san-1(RNAi) and mdf-2(RNAi) embryos, we used a strain (jm93) that contains a series of LacI promoters repeated in an array on one autosome (23). In this assay, two dots of staining in the nucleus indicates that the normal complement of two copies of the marked autosome are present, whereas more than two dots in a single nucleus demonstrates that a missegregation event has occurred. In anoxic san-1(RNAi) and mdf-2(RNAi) jm93 embryos, 9% (n = 100) and 11.6% (n = 103) of the embryos, respectively, had at least one blastomere with more than two discrete dots of staining (Fig. 4, E and F). Taken together, these data demonstrate that lack of spindle checkpoint function in anoxia results in missegregation of chromosomes and aneuploidy and therefore provide a compelling explanation for the anoxia-specific lethality in san-1(RNAi) and mdf-2(RNAi) embryos.

Fig. 4.

The san-1(RNAi) and mdf-2(RNAi) embryos exhibit evidence of chromosome missegregation in anoxia. Anaphase bridging was observed in mdf-2(RNAi) (A) and san-1(RNAi) (B) embryos after exposure to 24 hours of anoxia. To quantitate the extent of this missegregation, we exposed mdf-2(RNAi) and san-1(RNAi) jm93 embryos to 24 hours of anoxia and scored them for evidence of aneuploidy. Embryos with at least one nucleus containing more than two copies of the marked autosome occurred 11.6% of the time (n = 103) in mdf-2(RNAi) and 9% of the time (n = 100) in san-1(RNAi). (C) to (F) show an example of an anoxic mdf-2(RNAi) jm93 embryo (C and D) and an anoxic san-1(RNAi) jm93 embryo (E and F) exhibiting evidence of aneuploidy. (C) and (E) are DAPI, and (D) and (F) are LacI stained. The LacI insertions are visualized as individual dots in the nucleus. Nuclei with the normal complement of two copies of the tagged chromosome contain two dots of staining. The white arrows identify a blastomere in each embryo that contains more than two dots and is therefore aneuploid. White scale bar represents 5 μm.

We have demonstrated that san-1 is a component of the spindle checkpoint in C. elegans and that activation of the spindle checkpoint is essential for successful execution of the suspended animation program. Anoxic san-1(RNAi) and mdf-2(RNAi) embryos do not contain blastomeres arrested in metaphase and instead accumulate abnormal anaphases and telophases. That is, removal of san-1 or mdf-2 function in anoxia results in the continuation of the cell cycle beyond the metaphase-to-anaphase transition. It therefore seems that the function of the spindle checkpoint in anoxia is to prevent the initiation of anaphase in blastomeres that are energetically able to engage in the metaphase-to-anaphase transition but that are incapable of completing anaphase correctly. Thus, this work shows that cessation of this important biological process in response to oxygen deprivation is not just the inevitable result of energy depletion; there exists an active mechanism within the cell that coordinates the suspension process.

The responses of organisms to various oxygen tensions are highly conserved and very specific to particular oxygen tensions (24). HIF-1 is an important factor that mediates the response to hypoxia but not to anoxia (10). We show here that spindle checkpoint components are necessary for the response to anoxia but not hypoxia. Both HIF-1 and spindle checkpoint components are highly conserved throughout the animal kingdom, suggesting that the cellular response to low oxygen tensions is also conserved. This idea is corroborated by the similarities in the physiological aspects of suspended animation in different organisms. For example, both dogs (8) and pigs (9) can withstand exsanguination-induced oxygen deprivation in their tissues by entering into suspended animation in which, like zebrafish, these animals are immobile and not breathing and have no heartbeat. Upon reperfusion, they can be resuscitated and demonstrate normal cognitive function. The similarities between this phenomenon and those in other model systems leave open the possibility that the program of suspended animation may be conserved from worms to mammals.

Supporting Online Material

Materials and Methods

Table S1

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article