Control of Apoptosis by Asymmetric Cell Division

Asymmetric cell division and apoptosis (programmed cell death) are two fundamental processes that are important for the development and function of multicellular organisms. We have found that the processes of asymmetric cell division and apoptosis can be functionally linked. Specifically, we show that asymmetric cell division in the nematode Caenorhabditis elegans is mediated by a pathway involving three genes, dnj-11 MIDA1, ces-2 HLF, and ces-1 Snail, that directly control the enzymatic machinery responsible for apoptosis. Interestingly, the MIDA1-like protein GlsA of the alga Volvox carteri, as well as the Snail-related proteins Snail, Escargot, and Worniu of Drosophila melanogaster, have previously been implicated in asymmetric cell division. Therefore, C. elegans dnj-11 MIDA1, ces-2 HLF, and ces-1 Snail may be components of a pathway involved in asymmetric cell division that is conserved throughout the plant and animal kingdoms. Furthermore, based on our results, we propose that this pathway directly controls the apoptotic fate in C. elegans, and possibly other animals as well.


Introduction
Asymmetric cell division is essential for the generation of cellular diversity during animal development [1]. In certain cases, one of the cells derived from an asymmetric division is specified to undergo apoptosis (programmed cell death) [2][3][4][5][6][7][8][9][10][11]. However, the genetic and cell biological mechanisms that permit the coupling of asymmetric cell division and the adoption of the apoptotic fate are not well understood.
Previous studies have implicated members of the Snail family of transcriptional repressors in both asymmetric cell division and apoptosis, but there has been no demonstration that these processes are integrated in a given developmental context. In mammals, Snail-related proteins have been shown to have pro-survival as well as anti-apoptotic activities, and they have been causally linked to tumorigenesis and tumor progression in mammals [12,13]. In Drosophila melanogaster, the Snail-related genes snail, escargot, and worniu have been shown to function redundantly in asymmetric cell division [14,15]. Specifically, rather than dividing asymmetrically and producing two daughter cells of different sizes and fates, neuroblasts in the central nervous system of D. melanogaster mutants lacking snail, escargot, and worniu function divide symmetrically to produce two daughter cells of similar sizes and fates. The effect of snail, escargot, and worniu on asymmetric neuroblast division is mediated in part by their ability to promote the expression of the gene inscuteable, which encodes an adaptor protein that is required for the establishment and maintenance of neuroblast polarity [16]. The protein Inscuteable is localized in a polar fashion through its interaction with a complex composed of the PDZ domaincontaining proteins Bazooka (also referred to as Par-3) and Par-6, which are found on the apical cell cortex of the neuroblasts [17]. Inscuteable in turn recruits the adaptor protein Pins and the alpha subunit of the heterotrimeric G protein Gi (Gai), which initiates the displacement of the mitotic spindle along the cell division axis of the neuroblasts, resulting in their asymmetric division. Inscuteable is also at least partially required for the enrichment of cell-fate determinants such as Prospero and Staufen on the basal cell cortex of the neuroblasts and their asymmetric segregation into the basal daughter cell [17].
In Caenorhabditis elegans, the Snail-related protein CES-1 (cell-death specification) has been implicated in the suppression of a specific apoptotic event, the death of the neurosecretory motoneuron (NSM) sister cell [18,19]. During embryonic development, the NSM neuroblast divides to give rise to two daughter cells: the NSM, which differentiates into a serotonergic motoneuron, and the NSM sister cell, which undergoes apoptosis [3]. A gain-of-function mutation of the ces-1 gene, which most likely results in the mis-or overexpression of ces-1 in the NSM lineage, prevents the death of the NSM sister cell [18]. The death of the NSM sister cell is dependent on the transcriptional upregulation of the proapoptotic BH3-only gene egl-1 (egg-laying abnormal) in the NSM sister cell, a process that is at least partially dependent on a heterodimer composed of the bHLH transcription factors HLH-2 and HLH-3 (HLH-2/HLH-3) [20] (Figure 1). HLH-2/HLH-3 can bind to E-boxes/Snail-binding sites in a cisregulatory region of the egl-1 locus referred to as Region B, which is required for the expression of egl-1 in the NSM sister cell in vivo. Therefore, it has been proposed that HLH-2/ HLH-3 is a direct activator of egl-1 transcription in the NSM

Results bc212 Prevents the Death of the NSM Sister Cells
The NSM sister cells are generated about 410 min after the first cleavage of the C. elegans zygote and undergo apoptosis at about 430 min [3]. We screened for mutations that cause the NSM sister cells to survive, and we isolated the recessive mutation bc212 (J. Hatzold, B. Conradt, unpublished data). bc212 causes 12% and 50% NSM sister cell survival when raised at 25 8C or 15 8C, respectively, and hence causes a coldsensitive NSM sister cell survival phenotype (Table 1, dnj-11(bc212)). In addition, bc212 is maternally rescued: in homozygous bc212 progeny (dnj-11(bc212)) of heterozygous bc212 hermaphrodites (dnj-11(bc212)/þ), only 1% of the NSM sister cells survive (Table S1). To determine whether bc212 prevents the death of cells other than the NSM sister cells, we analyzed additional cells that undergo apoptosis during C. elegans development. We found that bc212 had no effect on their deaths, which demonstrates that bc212 does not block apoptosis in general (Table S2). However, we found that bc212 causes a variety of other defects such as morphological defects, lethality, slow growth, and reduced brood size ( Figure  S1 and Tables S3-S5). Therefore, the gene defined by bc212 is required for the death of the NSM sister cells and for additional processes that are important for C. elegans development and fertility.
bc212 Is a Putative Null Allele of the Gene dnj-11, Which Is Widely Expressed and Encodes a Protein That Primarily Localizes to the Cytosol We cloned the gene defined by bc212 and found that it is identical to the previously uncharacterized gene dnj-11 (F38A5.13) (DNJ-11 accession number NP_501006) on linkage group (LG) IV ( Figure 2A). dnj-11 encodes a 589-amino acid (aa) protein that is most similar to members of the family of MIDA1-like proteins, which are found in plants and animals (see below). bc212 is a C-to-T transition at position 21 of the nucleotide sequence of the dnj-11 gene and transforms codon 7 into a stop codon, which is predicted to truncate the DNJ-11 protein after aa 6 ( Figure 2A and 2B). tm2859, a 641base pair (bp) deletion of the dnj-11 gene, removes base pairs 402-1042 of the dnj-11 coding region. This deletion is predicted to result in a frameshift leading to the generation of a truncated protein composed of the first 116 aa of the wild-type protein and a 67-aa, C-terminal extension in a different reading frame (Figure 2A and 2B). Like bc212, tm2859 causes a cold-sensitive NSM sister cell survival phenotype and 50% NSM sister cell survival at 15 8C (Table  1, dnj-11(tm2859)). In addition, bc212 and tm2859 cause 24% and 26% lethality, respectively, at 15 8C (Table S3). Based on these results, we conclude that bc212 represents a strong lossof-function mutation and putative null allele of the dnj-11 gene. Figure 1. The Genetic Pathway of the NSM Sister Cell Death egl-1 is required for the death of the NSM sister cell. egl-1 activity in the NSM sister cell is positively regulated by the genes hlh-2 and hlh-3 and negatively regulated by the gene ces-1.

Author Summary
Asymmetric cell division and apoptosis (programmed cell death) are two fundamental processes that are important for the development and function of multicellular organisms. Asymmetric cell division creates daughter cells of different fates, and this is critical for the generation of cellular diversity. Apoptosis eliminates superfluous cells from the organism, which is critical for cellular homeostasis. We found that the processes of asymmetric cell division and apoptosis can be functionally linked. Specifically, we show that asymmetric cell division in the nematode Caenorhabditis elegans is mediated by a pathway involving three genes, dnj-11 MIDA1, ces-2 HLF, and ces-1 Snail, that directly control the enzymatic machinery responsible for apoptosis. Interestingly, the role of this pathway in asymmetric cell division and the control of apoptosis might be evolutionarily conserved. Furthermore, it might have an unexpected role in stem cell biology: the process of asymmetric cell division plays an essential role in the ability of stem cells to self-renew, and the mammalian counterparts of two components of the dnj-11 MIDA1, ces-2 HLF, ces-1 Snail pathway have recently been implicated in stem cell function. For this reason, we speculate that a dnj-11 MIDA1, ces-2 HLF, ces-1 Snail-like pathway might function in stem cells to coordinate self-renewal and apoptosis and, hence, the number of stem cells.
Using a functional transgene that drives the expression of a DNJ-11::green fluorescent protein (GFP) fusion protein under the control of the dnj-11 promoter (P dnj-11 dnj-11::gfp), we determined the expression pattern of the dnj-11 gene and the sub-cellular localization of the DNJ-11 protein. P dnj-11 dnj-11::gfp expression was observed in most if not all somatic cells of embryos, larvae, and adult animals, including the cells of the NSM lineage ( Figure 2C and unpublished data). DNJ-11::GFP protein primarily localized to the cytoplasm in a punctate pattern and could not be detected in nuclei. In addition, the dnj-11 gene is most likely also expressed in the adult germ line, because dnj-11(bc212) is maternally rescued (Table S1).

dnj-11 Is Required for Asymmetric NSM Neuroblast Division
The V. carteri ortholog of DNJ-11, GlsA, has been implicated in asymmetric cell division [35]. Specifically, during V. carteri development, an asymmetric cell division occurs that results in the generation of two daughter cells of different sizes and fates, namely a large reproductive cell and a small somatic cell. In glsA mutants, this cell division occurs symmetrically, resulting in two cells of equal size, both of which differentiate into somatic cells [26].
Most of the 131 cells that undergo apoptosis during C. elegans development are thought to be the result of an asymmetric cell division that gives rise to a large cell, which is programmed to survive, and a small cell, which is programmed to die [2,3]. Therefore, we determined whether the NSM neuroblast, which gives rise to the NSM and the NSM sister cell, divides asymmetrically with respect to size as well.
To that end, we identified the NSM neuroblast based on its position in the embryo, observed its division at about 410 min, and determined the sizes of its daughter cells immediately after its division had been completed ( Figure 3A). Using a transgene that expresses a plasma membrane-targeted GFP fusion protein (P pie-1 gfp::ph(PLC1d1)) [36], we found that in wild-type animals, the size of the NSM sister cell on average is 0.46 times the size of the NSM ( Figure 3B, þ/þ). Therefore, the NSM neuroblast divides asymmetrically to give rise to a large cell (the NSM) and a small cell (the NSM sister cell).   The alignment of the DNJ-11 sequence with the sequence of mouse MIDA1, human MPP11, V. cateri GlsA, and the S. cerevisiae Zuotin was performed using the DIALIGN algorithm. Identical amino acids and similar amino acids are shaded in black and gray, respectively. The orange box marks the J domain and the conserved tripeptide is shaded in orange. The blue box indicates the M domain. The maroon bars above the sequences represent the Myb domains. Arrows point to conserved aromatic residues in the Myb domains. bc212 results in the truncation of the DNJ-11 protein after aa 6 (asterisk). tm2859 results in a frame shift after aa 116 (red asterisk) and a stop after an additional 67 aas in a different frame. (C) Embryos transgenic for the functional P dnj-11 dnj-11::gfp transgene were fixed as described in Materials and Methods and stained with an anti-GFP antibody (aGFP) to detect DNJ-11::GFP and with DAPI (DAPI) to detect nuclei. doi:10.1371/journal.pbio.0060084.g002 (B) The ratio of the sizes of the NSM sister cell and the NSM was determined as described in Materials and Methods. The NSM sister cell was defined as the lateral cell of the two daughter cells. Each diamond represents the ratio of a single, independent NSM neuroblast division. The numbers above the diamonds represent the average ratio obtained for a given genotype. The apoptotic fate of the NSM sister cells was determined in a subset of dnj-Furthermore, we found that in dnj-11(bc212) animals, the difference in size between the NSM sister cell and the NSM is highly variable, ranging from a 2-fold difference as observed in wild-type animals to no difference ( Figure 3B, dnj-11(bc212)). To determine whether the range observed reflects the incomplete penetrance of the NSM sister cell survival phenotype caused by dnj-11(bc212) (see Table 1), we followed the fate of NSM sister cells after the division of the NSM neuroblasts. We found that in five out of five embryos in which the division of the NSM neuroblast had occurred asymmetrically (NSM sister cell size on average 0.51 times the size of the NSM), the NSM sister cells died ( Figure 3B, red diamonds). By contrast, four out of four NSM sister cells that were similar in size to the NSMs (NSM sister cell size on average 0.83 times the size of the NSM) survived (green diamonds). To rule out the possibility that the increase in NSM sister cell size observed in dnj-11(bc212) animals is a result of inappropriate NSM sister cell survival rather than a defect in asymmetric cell division, we analyzed egl-1(lf) mutants, in which many apoptotic events, including the death of the NSM sister cells, are blocked. We found that, like in wild-type animals, the NSM neuroblast divided asymmetrically in egl-1(lf) mutants (NSM sister cell size on average 0.52 times the size of the NSM) ( Figure 3B, egl-1(n1084n3082)). Thus, the increase in NSM sister cell size observed in dnj-11(bc212) animals is a result of a defect in asymmetric cell division. Therefore, we conclude that dnj-11 is required for asymmetric NSM neuroblast division and, by inference, is required for the displacement of the mitotic spindle along the cell division axis. Furthermore, because in dnj-11(bc212) animals the defect in asymmetric NSM neuroblast division correlates with the defect in NSM sister cell death, asymmetric NSM neuroblast division might be critical for the specification of the apoptotic fate of the NSM sister cell.
To determine whether dnj-11 is required for asymmetric cell division in lineages other than the NSM lineage, we analyzed a subset of the other asymmetric cell divisions (including the first division of the zygote) that take place during C. elegans development and that give rise to daughter cells of different sizes and fates. We found that none of these divisions was affected by dnj-11(bc212), suggesting that dnj-11 is not required for asymmetric cell division in general (Table  S7 and unpublished data).
To determine whether ces-1 also acts downstream of dnj-11 and ces-2 during asymmetric NSM neuroblast division, we determined whether the loss of ces-1 function can suppress the defects in asymmetric NSM neuroblast division observed in dnj-11(bc212) and ces-2(bc213) animals. We found that ces-1(n703n1434) partially suppressed the defects in asymmetric NSM neuroblast division that are observed in dnj-11(bc212) and ces-2(bc213) animals ( Figure 3B). This finding suggests that ces-1 acts downstream of dnj-11 and ces-2 during asymmetric NSM neuroblast division as well.
dnj-11, ces-2, and ces-1 Also Play a Role in Establishing the Correct Orientation of the Cell Division Axis During Asymmetric NSM Neuroblast Division Shortly after the division of the NSM neuroblast has been completed, the NSM is located at a position that is medial and ventral to the NSM sister cell ( Figure 3A). The position of the NSM relative to the NSM sister cell implies that in wild-type animals, the cleavage plane of the NSM neuroblast is along the ventral/lateral to dorsal/medial axis. While determining the size of the NSM and NSM sister cell using the plasma membrane-targeted GFP fusion protein, we observed that in two out of nine ces-2(bc213) animals, the medial daughter cell was located dorsally rather than ventrally to the lateral daughter cell ( Figure 3A). This observation suggests that the loss of ces-2 function not only affects the displacement of the mitotic spindle along the cell division axis in the NSM neuroblast, but also the orientation of the cell division axis and, hence, the orientation of the cleavage plane.
dnj-11 and ces-2 Reduce ces-1 Transcription in the NSM Lineage Thereby Restricting the Presence of CES-1 Protein to the NSM To investigate how dnj-11 and ces-2 antagonize ces-1 function in the NSM lineage, we constructed a functional transgene that drives the expression of a CES-1::yellow fluorescent protein (YFP) fusion protein under the control of the ces-1 promoter (P ces-1 ces-1::yfp). In wild-type animals, we failed to detect CES-1::YFP in NSM neuroblasts or NSM sister cells. However, we observed CES-1::YFP in two out of 17 NSMs analyzed ( Figure 5A and 5B, þ/þ). (CES-1::YFP was observed in lineages other than the NSM lineage in all animals examined.) This observation suggests that the level of CES-1::YFP and hence, most probably of endogenous CES-1 protein, is relatively low in the NSM lineage and that it is higher in NSMs than in NSM neuroblasts or NSM sister cells. In dnj-11(bc212) and ces-2(bc213) animals, we observed CES-1::YFP not only in NSMs but also in NSM sister cells and NSM neuroblasts ( Figure 5A and 5B, dnj-11(bc212), ces-2(bc213)). These results demonstrate that dnj-11 and ces-2 antagonize ces-1 function in the NSM lineage by reducing the level of CES-1 protein in the NSM neuroblast, thereby restricting the presence of CES-1 protein to the NSM. To determine whether dnj-11 and ces-2 act at the transcriptional or posttranscriptional level to affect the level of CES-1 protein, we constructed a transgene that drives the expression of the GFP protein under the control of the ces-1 promoter (P ces-1 gfp) and analyzed GFP expression in the NSM lineage. In wild-type animals, GFP was not detected in NSM neuroblasts, NSMs, or NSM sister cells ( Figure 5B, þ/þ). However, in dnj-11(bc212) and ces-2(bc213) animals, GFP was detected in NSM neuroblasts, NSMs, and NSM sister cells ( Figure 5B, dnj-11(bc212), ces-2(bc213)). Based on these results, we conclude that dnj-11 and ces-2 restrict the presence of CES-1 protein to the NSM by reducing ces-1 transcription in the NSM lineage.

Discussion
It is well established that asymmetric cell division can be coupled to the apoptotic fate [2][3][4][5][6][7][8][9][10][11]. However, the genetic and mechanistic basis for this relationship has not been well defined. Here we present evidence that the regulation of asymmetric cell division and apoptosis are functionally linked by a pathway that involves three evolutionarily conserved genes: dnj-11 MIDA1, ces-2 HLF and ces-1 Snail.

CES-1 Snail Levels Determine the Apoptotic Fate
Snail family transcription factors have previously been implicated in the regulation of apoptosis in mammalian cells. In the case of C. elegans, it has been suggested that the Snail protein, CES-1, might normally function within the NSM lineage to repress transcription of the pro-apoptotic gene, egl-1, thereby allowing NSM survival. Our data support this idea, because we find that CES-1 protein is present at higher levels in the NSMs than in the NSM sister cells. Therefore, we propose that CES-1 acts as a cell-fate determinant in the NSMs to ensure the survival of the (B) Summary of the data on the presence of CES-1 protein in the NSM lineage using the P ces-1 ces-1::yfp transgene and ces-1 expression using the P ces-1 gfp transgene. Green circles represent cells in which CES-1::YFP or GFP were detected. White circles represent cells in which CES-1::YFP or GFP were not detected. The complete genotypes of the embryos analyzed from top to bottom were: bcIs58 (P ces-1 ces-1::yfp), dnj-11(bc212) bcIs25; bcIs58, ces-2(bc213); bcIs25; bcIs58, bcEx619 (P ces-1 gfp), dnj-11(bc212) bcIs25; bcEx619, and ces-2(bc213); bcIs25; bcEx619. doi:10.1371/journal.pbio.0060084.g005 NSMs and their differentiation into serotonergic neurons. Conversely, we propose that the absence of CES-1 protein in the NSM sister cells determines the death of the NSM sister cells ( Figure 6A).

CES-1 Snail Forges a Link between Asymmetric Cell Division and Apoptosis
Our results also indicate that the CES-1 protein acts in the NSM neuroblasts to affect the orientation of the cleavage plane and, hence, asymmetric cell division ( Figure 6). Therefore, at least in the NSM lineage, the CES-1 protein represents a functional link between the cellular machinery that causes asymmetric cell division and the cellular machinery that causes the apoptotic death of specific cells during C. elegans development. At least to our knowledge, this is the first demonstration that apoptosis can directly be controlled by asymmetric cell division. Like the function of ces-1 in NSM fate determination, the function of ces-1 in Figure 6. Regulation of Asymmetric NSM Neuroblast Division and the Apoptotic Death of the NSM sister Cell (A) Molecular models. (Wild type) By repressing the transcription of the Snail-related gene ces-1, the proteins CES-2 HLF and DNJ-11 MIDA1 ensure that the CES-1 Snail protein is present at an appropriate, low level in the unpolarized, early NSM neuroblast. Blue oval indicates polarity factor that is required to establish polarity in the late NSM neuroblast and the synthesis of which is promoted by CES-1. Indicated in red is a complex that restricts the polarity factor to the ventral/medial side of the NSM neuroblast. Once localized to the ventral/medial side, the polarity factor is involved in restricting CES-1 to the ventral/medial side of the late NSM neuroblast. See text for details. (dnj-11(lf), ces-2(lf), ces-1(gf)) Loss-of-function mutations of dnj-11 or ces-2, or a gain-of-function mutation of ces-1 result in an increased level of CES-1 Snail protein in the early NSM neuroblast. This increased level of CES-1 protein results in a level of polarity factor too high to establish asymmetry and restrict CES-1 protein to the ventral/medial side. See text for details. (B) Genetic pathway. dnj-11 acts upstream of or in parallel to ces-2 to negatively regulate the function of ces-1, thereby causing asymmetric NSM neuroblast division and NSM sister cell death. ces-1 function can affect asymmetric NSM neuroblast division by regulating the function of a gene or genes required for the process. By negatively regulating egl-1, ces-1 function can also prevent the hlh-2, hlh-3-dependent death of the NSM sister cells. ces-1 function in NSM sister cell death is affected by the process of asymmetric NSM neuroblast division, which causes the asymmetric distribution of the product of the ces-1 gene. See text for details. doi:10.1371/journal.pbio.0060084.g006 asymmetric NSM neuroblast division is redundant to that of one or more unidentified genes. Similarly, the Snail-related proteins Snail, Escargot, and Worniu of D. melanogaster act redundantly to cause asymmetric cell division [14,15,17].

DNJ-11 and CES-2 Regulate ces-1 Transcription in the NSM Lineage
The HLF-like bZIP transcription factor, CES-2, is thought to act as a direct repressor of CES-1 Snail transcription [19,21]. We have found that the dnj-11 MIDA1-like gene acts in concert with CES-2 as a negative regulator of CES-1 expression. Reducing either dnj-11 or ces-2 function results in an increased level of ces-1 transcription within the NSM lineage, disrupts asymmetric NSM neuroblast division, and prevents the death of the NSM sister cells. Based on our data, dnj-11 and ces-2 could either act in parallel or in a single linear pathway to antagonize ces-1 function (Figure 6). MIDA1-like proteins have been implicated in transcriptional regulation [28][29][30] and are components of a ribosome-associated chaperone referred to as RAC, which co-translationally interacts with nascent polypeptides thereby affecting translational accuracy and termination as well as protein folding [31,32]. The DNJ-11 protein predominantly localizes to the cytosol in a punctate pattern, which suggests that rather than regulating ces-1 transcription directly, DNJ-11 might affect the translation and/or folding of a regulator of ces-1 transcription.

Snail Pathway in Asymmetric NSM Neuroblast Division
Based on our results, we propose that by repressing the transcription of the snail-related ces-1 gene, the HLF-like bZIP transcription factor CES-2 and the MIDA1-like chaperone DNJ-11 ensure that the CES-1 protein is present at an appropriate, low level in the early NSM neuroblast ( Figure  6A). A low level of CES-1 protein in the early NSM neuroblast would allow the expression at a certain level of a ''polarity factor'' that is required for the asymmetric division of the late NSM neuroblast. Furthermore, a complex that localizes to the ventral/medial side of the late NSM neuroblast would restrict the polarity factor to the ventral/medial side, thereby promoting the displacement of the mitotic spindle along the cell division axis and resulting in a shift of the cleavage plane and the asymmetric division of the cell. Finally, the asymmetrically localized polarity factor would also restrict CES-1 to the ventral/medial side and thereby cause its segregation predominantly into the NSM, thus resulting in the repression of egl-1/BH3-only transcription in the NSM and the survival of the NSM ( Figure 6A). In this model, the CES-1 protein not only is a component of the cellular machinery that causes asymmetric NSM neuroblast division, but also one of its targets. The identity of the polarity factor as well as the signals and mechanisms that cause its asymmetric localization or activation remain to be determined. The Snail-related genes snail, escargot, and worniu of D. melanogaster function in asymmetric neuroblast division by promoting the expression of the gene inscuteable, which encodes an adaptor protein required for asymmetric neuroblast division. It will be of interest to determine whether the C. elegans ortholog of inscuteable, the gene insc-1 (F43E2.3) (INSC-1 accession number AAC71125) [37], plays a role in asymmetric cell division and in asymmetric NSM neuroblast division, in particular.
The Role of the dnj-11 MIDA1, ces-2 HLF, ces-1 Snail Pathway in Asymmetric Cell Division and the Control of Apoptosis Might be Conserved At least to our knowledge, this is the first evidence that a HLF-like bZIP transcription factor plays a role in asymmetric cell division. Furthermore, our studies provide a functional link between the known roles in asymmetric cell division of MIDA1-like chaperones and Snail-related transcription factors, and hence suggest the existence of a pathway involved in asymmetric cell division that is conserved throughout the plant and animal kingdoms [14][15][16]26]. Therefore, it will be of interest to determine whether HLF-like and Snail-related proteins also contribute to asymmetric cell division in V. carteri and whether MIDA1-like and HLF-like proteins also participate in asymmetric neuroblast division in D. melanogaster. Furthermore, our results hint at the possibility that MIDA1-like, HLF-like, and Snail-related proteins might play a role in asymmetric cell division in vertebrates. Specifically, it has recently been reported that HLF and the Snail-related protein Slug of mammals have functions in stem cells [12,42]. In addition, there is increasing evidence that the process of asymmetric cell division plays a crucial role in the ability of stem cells to self renew [43]. Therefore, we hypothesize that a dnj-11 MIDA1, ces-2 HLF, ces-1 Snail-like pathway might be important for stem cell renewal by allowing asymmetric stem cell division.
At least to our knowledge, this is also the first evidence that a MIDA1-like protein plays a role in the regulation of apoptosis. Furthermore, our studies have identified a new component of a conserved cell-death specification pathway composed of C. elegans ces-2 HLF and ces-1 Snail. Like CES-2 and CES-1, HLF and the Snail-related protein Slug of mammals have previously been implicated in the regulation of apoptosis [19,21,[44][45][46]. Therefore, it will be of interest to determine whether MIDA1-like proteins also have an apoptotic role in mammals and act in a HLF-, and Slugdependent pathway. Finally, based on our work, we consider it a possibility that the roles in vertebrates of a dnj-11 MIDA1, ces-2 HLF, ces-1 Snail-like pathway in asymmetric cell division and apoptosis might be functionally linked as well. Specifically, we speculate that such a pathway could cause stem cell renewal through asymmetric cell division and control stem cell numbers through apoptosis [47].

Implications for the Functions of the Proto-Oncoprotein HLF and the Snail-Related Protein Slug of Mammals
The oncogenic form of human HLF, the E2A-HLF fusion protein, found in patients carrying the t(17;19) (q22;p13) translocation, gives rise to pro-B cell acute lymphoblastic leukemia (ALL) in adolescents [48]. The E2A-HLF fusion protein is composed of the trans-activation domain of the bHLH protein E2A and the DNA-binding domain of HLF [49]. It has been proposed that E2A-HLF causes leukemic transformation of pro-B cells by blocking their apoptotic death. Specifically, it has been proposed that E2A-HLF inappropriately activates the transcription of the snail-related gene, slug, which encodes a direct repressor of the egl-1-like, pro-apoptotic BH3-only gene puma, thereby causing the survival of pro-B cells that are normally programmed to undergo apoptosis [44][45][46]. Based on our results, we speculate that like C. elegans CES-2 and CES-1, the proteins HLF and Slug might not only function to control the expression of a pro-apoptotic BH3-only gene but to cause asymmetric cell division. Hence, in patients with the t(17;19) (q22;p13) translocation, the presence of the E2A-HLF fusion protein and elevated levels of Slug protein might affect aspects of the pro-B cell fate other than their apoptotic fate and/or might alter the division of the lymphoid progenitors that produce pro-B cells. Finally, it will be of interest to determine whether the human MIDA1-like gene MPP11 [27], which is expressed in hematopoietic lineages as well as other tissues, plays a role in E2A-HLF-and Slug-mediated tumorigenesis.
ces-1 plasmids: The plasmid pBC510 (P ces-1 ces-1::yfp) was generated by cloning an AflII-SpeI fragment from cosmid F43G9 containing the ces-1 rescuing fragment [19] into the EcoRV site of pBluescript II KS þ (pBC482A), and inserting yfp amplified from pvdB#3 [59] into the SwaI site to generate a C terminal in frame fusion. The P ces-1 ces-1::yfp transgene was able to block the ability of ces-1(n703n1434) to suppress the NSM sister cell survival phenotype of dnj-11(bc212) and ces-2(n732) animals, demonstrating that it is functional (unpublished data). To generate plasmid pBC664 (P ces-1 gfp), first ces-1 upstream and downstream regulatory regions were amplified by PCR, a PmeI site was introduced 39 of upstream and 59 of downstream regulatory regions, and PCR products were cloned into XmaI/NcoI digested pBC482A to obtain the ces-1 locus without coding region, 59 and 39 untranslated regions (UTRs), and introns (pBC656). Because intron 4 of ces-1 is highly conserved between C. elegans and C. briggsae, intron 1 of the gfp sequence of plasmid pPD95.77 (gift of A. Fire) was replaced with intron 4 of ces-1 by PCR fusion. The SmaI/SpeI fragment of pPD95.77_ces-1 intron containing the gfp sequence and the unc-54 39 UTR was inserted into the PmeI site of pBC656 to generate pBC664.
Phenotypic analysis and microscopy. The NSM sister cell survival was scored as previously described [20]. Microscopy of living embryos was performed by mounting embryos on 2%-5% agar pads in M9 buffer, sealing them with petroleum jelly, and using a Zeiss Axioskop2 equipped with epifluorescence, a Micromax CCD camera (Princeton Instruments), and Metamorph software. NSM neuroblasts, NSMs, and NSM sister cells were identified based on the position of their nuclei using Nomarski optics. Z-series were taken with a Z-distance of 0.5 lm (analysis of P ces-1 ces-1::yfp, P ces-1 gfp, and P his-24 his-24::gfp expression) and 0.25 lm (determination of cell size). Epifluorescence Z-series were deconvolved using the AutoDeblur Gold WF AutoVisualize software (Media Cybernetics). The cell size of NSMs and NSM sister cells was determined 10 to 15 min after the NSM neuroblast had started to divide, as indicated by the breakdown of the nuclear envelop, which was observed by Nomarski optics. To visualize the outline of a cell, a plasma membrane-targeted GFP fusion protein (P pie-1 gfp::ph(PLC1d1)) was used. The area of the cross section of a cell was measured in each section of a Z-series using Metamorph software. To estimate the difference in volume between NSM and NSM sister cell, the values of each section were added, and the sum obtained for the NSM sister cell was divided by the sum obtained for the NSM. To determine the orientation of the cleavage plane of the NSM neuroblast, chromatids were visualized using a His24-GFP fusion protein (P his-24 his-24::gfp, dtIs372) (gift of M. Dunn and G. Seydoux, Johns Hopkins University, Baltimore, Maryland; and J. Waddle, Southern Methodist University, Dallas, TX). The NSM neuroblast was identified by Nomarski optics before the start of the division and consecutive Z series were taken in 1 min time intervals until the completion of the NSM neuroblast division.