Germ-line stem cells are unique because they either self-renew through mitosis or, at a certain frequency, switch to meiosis and produce gametes. The switch from proliferation to meiosis is tightly regulated, and aberrations in switching result in either too little or too much proliferation. To understand the genetic basis of this regulation, we characterized loss-of-function mutations and a novel tumorous allele of Caenorhabditis elegans mett-10, which encodes a conserved putative methyltransferase. We show that METT-10 is a nuclear protein that acts in the germ line to inhibit the specification of germ-cell proliferative fate. METT-10 also promotes vulva, somatic gonad, and embryo development and ensures meiotic development of those germ cells that do differentiate. In addition, phenotypic analysis of a mett-10 null allele reveals that METT-10 enables mitotic cell cycle progression. The finding that METT-10 functions to inhibit germ-cell proliferative fate, despite promoting mitotic cell cycle progression of those germ cells that do proliferate, separates the specification of proliferative fate from its execution.
FROM early development to later life, tissues are formed from and maintained by stem cell populations. Stem cells have the unique ability to give rise to both differentiated cell types and self-renewing daughters and must regulate the choice between the two. A balance between proliferation and differentiation is critical for normal tissue development and the avoidance of disease. For example, uncontrolled self-renewal is a hallmark of cancer (Krivtsov et al. 2006), while a failure to maintain proliferation is often equally detrimental (Kauffman et al. 2003).
The Caenorhabditis elegans germ line provides a model system to study factors regulating this balance. C. elegans germ cells arise from germ-line precursors set aside early in life (Sulston et al. 1983). Following initiation of post-embryonic development, germ-line stem cells either maintain a proliferative fate and undergo mitotic cell division or differentiate into gametes, whereby they enter a meiotic developmental program (Hansen and Schedl 2006; Kimble and Crittenden 2007). This decision is spatially regulated within the gonad. Proliferating germ cells reside in the distal end and enter meiosis more proximally at a region of the germ line called the “transition zone,” which corresponds to leptotene/zygotene of meiosis I (Figure 1A). There is no evidence for asymmetric division within the proliferative zone, and it is generally thought that differentiation is a consequence of progressive displacement away from the niche (Morrison and Kimble 2006).
The signal for continued germ-cell proliferation is provided by activation of Notch signaling in distal germ cells (Austin and Kimble 1987; Berry et al. 1997). The ligands LAG-2 and APX-1, members of the conserved Delta/Serrate family, are expressed by a somatic cell called the distal tip cell, creating a “niche” (Henderson et al. 1994; Fitzgerald and Greenwald 1995; Nadarajan et al. 2009). Ligand binding to GLP-1 (one of two C. elegans Notch receptors) likely results in GLP-1 cleavage, generating GLP-1(Intra), which translocates to the nucleus and complexes with factors, such as the LAG-1 DNA-binding protein, to activate transcription of target genes that promote the proliferative fate (Mumm and Kopan 2000). Thus, loss of glp-1 function results in the premature differentiation and meiotic entry of all germ cells (Austin and Kimble 1987), while constitutive activation of glp-1 causes germ-line tumors (Berry et al. 1997).
Although glp-1 signaling is a central component controlling the specification of germ-cell proliferative fate, regulation of this decision is not fully understood. A complete picture of the components that transduce or inhibit the glp-1 signal, govern the maintenance of proliferative cells, and ensure their proper differentiation and entry into meiosis will enhance our understanding of cell fate decisions, including aberrant cell fate decisions, leading to human cancers (Weng et al. 2004; Koch and Radtke 2007).
To further characterize the specification of germ cell proliferative fate, we characterized a novel tumorous allele of the mett-10 gene isolated in a forward genetic screen for factors that disrupt the balance between germ-cell proliferation and differentiation (Francis et al. 1995a). We demonstrate that METT-10, a conserved putative methyltransferase, inhibits germ cell proliferative fate in C. elegans, possibly through inhibition of the pro-proliferative functions of the glp-1/Notch signaling pathway.
MATERIALS AND METHODS
Worm culture and genetics:
Standard procedures for culture and genetic manipulation of C. elegans were followed (Brenner 1974). To isolate tumorous mutants, N2 Bristol animals were mutagenized with 50 mm ethyl methanesulfonate, and a dissecting microscope was used to screen an aged F2 generation for tumorous animals (Francis et al. 1995a). For temperature-sensitive phenotypes, hermaphrodites laid eggs overnight at 20° and were removed, plates were shifted to 25°, and the F1 was scored at 1 day past the L4 stage. All experiments analyzed hermaphrodite germ lines with the exception of the enhancement of gld-1(q485), in which male germ lines were analyzed. However, germ-line tumors were observed in rare, spontaneous smg-2(e2008); mett-10(oz36) males, indicating that the effect of the oz36 mutation is not entirely sex-specific. Complementation between mett-10(oz36) and other mett-10 alleles was tested by crossing potential mett-10 alleles to the hT2[qIs48]gfp balancer, and the resulting gfp(+) males to an unc-32(e189) mett-10(oz36) strain. Non-green, non-unc hermaprodite progeny were picked and shifted to 25°. To examine interaction between mett-10 and lin-12, we built double mutants with the lin-12(n379) gain-of-function allele and mett-10(g38) or mett-10(oz36). No enhancement of the multi-vulval phenotype of lin-12(n379) was observed, although mett-10(-) lin-12(n379) animals lacking maternal gene product were sick, grew slowly, and had few progeny. To remove unlinked and linked mutations, mett-10 alleles obtained from the knockout consortium and National Bioresource Project-Japan were outcrossed to N2 Bristol three times, or unc-32 and dpy-18 were recombined on and off the third chromosome. Strain constructions were verified by PCR and DNA sequencing.
The following alleles were used in this study:
Chromosome III: lin-12(n379), glp-1(oz264), glp-1(bn18), glp-1(q175), unc-32(e189), unc-36(e251), mett-10(oz36) [nucleotide change: III:10115924 C>T], mett-10(g38) [III:10115898 G>A], mett-10(tm2697), mett-10(ok2204), mett-10(oj32)[III: 10115334 G>A], unc-119(ed3), dpy-18(e364), nDf40
Mapping of S-adenosyl-methionine-binding residues in METT-10:
Because of relatively low sequence conservation, mapping of characterized methyltransferase motifs (reviewed in Malone et al. 1995) onto C. elegans METT-10 was done by aligning the C. elegans protein to human METT10D, a portion of which was crystallized complexed with S-adenosyl methionine (SAM) by the Structural Genomics Consortium [Structure (PDB 2H00), which can be found at http://www.rcsb.org/pdb/explore.do?structureId=2H00]. This was aided by structure prediction using Jpred3 (http://www.compbio.dundee.ac.uk/∼www-jpred). Many methyltransferase motifs are defined by interaction with SAM, so we examined the METT-10-SAM interaction interface using RSCB ligand Explorer (available through the above website), which also predicted residues required for SAM-binding.
Immunohistochemistry and image capture:
Germ-line staining was carried out as previously described (Lee et al. 2007). Antibodies used were the following: monoclonal mouse anti-GFP (mAb3E6; Invitrogen), rabbit anti-GLD-1 (Jones et al. 1996), rabbit anti-phosphohistone-3 (Upstate), guinea pig anti-Ce-lamin (gift from J. Liu, Cornell University), donkey anti-mouse Alexa 594, goat anti-rabbit Alexa 488, goat anti-rabbit Alexa 594, and goat anti-guinea pig-Alexa 488 obtained from Molecular Probes. For 5-ethynyl-2′-deoxyuridine (EdU) labeling and costaining with antibodies, we modified a protocol from Sarah Crittenden (University of Wisconsin-Madison, personal communication). To make EdU plates, MG1693 Escherichia coli (thymidine deficient, E. coli Stock Center) were grown overnight at 37° and diluted 1/50 in 1% glucose, 1.25 μg/ml thiamine, 0.5 μm thymidine, 1 mm MgSO4, and 20 μm EdU in M9 minimal media. This culture was grown at 37° for 24 hr in the dark, pelleted, and resuspended in a small volume of M9, and plated on 60-mm M9 plates. To label and stain proliferative cells, worms were picked onto EdU plates, washed off in PBS, dissected, and fixed for <5 min in 3% paraformaldehyde/0.1 m K2HPO4 (pH 7.2) at room temperature. Germ lines were taken through the antibody staining protocol, including DAPI-staining, followed by the Click-IT EdU reaction, performed according to the manufacturer's instructions (Molecular Probes).
Fluoresence micrographs were taken on a Zeiss compound microscope using AxioPlan 2.0 imaging software and a Hamamatsu camera and processed as described in Arur et al. (2009). Confocal images were captured on a Perkin-Elmer Ultraview microscope using 1 μm z-steps. Where stated, multiple slices were assembled into an extended focus image using Volocity software.
Construction of transgenes and generation of transgenics:
To create a METT-10∷GFP construct, the METT-10 genomic region from 1.6 kb upstream to 1.9 kb downstream of the coding region was amplified as two PCR products meeting at the translational stop with KOD polymerase (Novagen), using the following primers: 5′ region—ATTA-NotI- GCGTCGTAGCCTGTGTTCAATTCC (F) and ATTA-BamHI-TCGGCATATTAAATTTTTCAAATACTGCACTAG (R) and 3′ region—ATTA-BamHI-TAAATAGCATAATTTTATTTATTCAATATTTATACC (F) and ATTA-NotI- GGATACTGTATGTACAACTGATCTCTGAGC (R). This 5.5-kb region includes parts of adjacent genes. The products were digested with BamHI and NotI and ligated to NotI-digested pMM016 (Praitis et al. 2001), a plasmid containing the unc-119+ transformation marker, to create pMM016(mett-10+). GFP(ivs):FLAG was amplified from pPD115.70 (gift of Andrew Fire) and cloned into pMM016(mett-10+) using the BamHI site. Orientation was determined by PCR, and the mett-10 region was sequenced. Microparticle bombardment was used to create low-copy integrated transgenic lines (Praitis et al. 2001).
To generate arrays for mosaic analysis, pMM016 [mett-10+] was co-injected with pTG96 [sur-5∷gfp] (gift of Min Han) at concentrations of 20 ng/μl and 80 ng/μl, respectively. These arrays were crossed into smg-2(e2008); mett-10(oz36) worms. GFP+ animals were cloned individually at the L4 stage. After 2.5 days, plates were screened for animals with GFP+ embryos or progeny to determine if the array had passed through the germ line. Animals were segregated on the basis of germ-line transmission, germ lines were dissected, fixed briefly for 10 min in 3% paraformaldehyde to preserve GFP fluorescence, DAPI stained, and mounted for scoring of germ-line phenotype and GFP distribution.
A mett-10 allele with a tumorous phenotype:
To identify genes regulating the transition between proliferative and meiotic fates, we carried out a screen for recessive tumorous mutants exhibiting excessive germ-cell proliferation (Francis et al. 1995a). We isolated an allele of mett-10, the worm ortholog of the METT10D putative methyltransferase of vertebrates (see below, this section); mett-10(oz36) exhibits a cold-sensitive and partially penetrant tumorous phenotype (Table 1). The tumorous phenotype is variable and includes both regions of ectopic proliferation near the “loop” region of the germ line normally occupied by cells in meiotic pachytene, as well as what is typically called the “late-onset tumorous” phenotype, in which the distal proliferative zone progressively increases in length as the animals age. These phenotypes indicate an imbalance of proliferation over differentiation similar to glp-1/Notch gain-of-function mutations (Berry et al. 1997; Pepper et al. 2003) and are distinct from tumors that arise from a return to proliferation from meiotic development (Francis et al. 1995a; Subramaniam and Seydoux 2003) or that result from disrupting the function of multiple factors acting downstream of glp-1 to promote meiotic entry (Kadyk and Kimble 1998; Eckmann et al. 2004; Hansen et al. 2004a).
We confirmed that the suspected ectopic proliferating cells are cycling with the nucleotide analog, EdU, which labels replicating DNA in S-phase (Salic and Mitchison 2008). In wild-type germ lines, a 3-hr EdU pulse labels most cells in the proliferative zone, but not in the adjacent meiotic region (Figure 1B). The suspected ectopic proliferating cells in mett-10(oz36) mutants label with EdU within a 3-hr period, showing that they are cycling (Figure 1C). Because EdU labels cells in both mitotic and meiotic S-phase, we also stained for additional mitotic and meiotic markers. Ectopic phospho-histone-3 (pH 3, a metaphase marker) nuclei were seen in mett-10(oz36) mutants, indicative of ectopic mitosis (Figure 1, D and E). These cells were negative for the meiotic marker GLD-1 (Hansen et al. 2004b). Thus, these data argue that the mett-10(oz36) mutation causes proliferation of germ cells that should have entered meiosis. Because C. elegans germ-line stem cells have not been unequivocally identified, we cannot definitively attribute the phenotype to overproliferation of the germ-line stem cells, rather than a potential transient amplifying population. However, mett-10(oz36) shares the unique late-onset tumorous phenotype with glp-1/Notch gain-of-function mutants (Berry et al. 1997), suggesting that it may affect stem cell fate specification.
We cloned mett-10 by genetic mapping, cosmid rescue, and sequencing of candidate genes. Several lines of evidence show that the gene affected by the oz36 mutation is ZK1128.2, the ortholog of the human Methyltransferase-10 domain-containing protein (METT10D) (Figure 2A). First, the mutation responsible for the tumorous phenotype maps to a genomic region containing ZK1128.2 (data not shown). Second, a transgene containing the ZK1128.2 coding region rescues the mett-10(oz36) mutant phenotype (Table 1). Finally, sequencing of the oz36 allele, as well as other mutations in the same complementation group (see below), all reveal molecular lesions within the ZK1128.2 coding region (Figure 2B).
The N-terminal domain of human METT10D was crystallized in complex with SAM, the substrate used by SAM-dependent methyltransferases to catalyze transfer of a methyl group to various substrates, including proteins, DNA, RNA, small molecules, and lipids (Martin and McMillan 2002; Wu et al. 2009). Significantly, on the basis of sequence conservation and secondary structure prediction, C. elegans METT-10 is predicted to bind SAM (Figure 2A). Moreover, both human and C. elegans METT-10 contain a SAM-dependent methyltransferase fold, including conserved protein motifs responsible for catalyzing methyl transfer (Figure 2A; Malone et al. 1995). With only two known exceptions, proteins containing the SAM-dependent methyltransferase fold act as methyltransferases (see discussion). Thus, it is likely that METT-10 is a methyltransferase, although the substrate is unknown.
Loss of mett-10 function causes multiple temperature-sensitive phenotypes:
The oz36 mutation creates a premature stop codon that truncates the protein shortly after the methyltransferase domain. While mett-10(oz36) behaves as a complex allele at lower temperatures to cause a tumorous phenotype, including both loss-of-function and poisoning properties (see below), it behaves as a true loss-of-function allele at 25°, failing to complement a deficiency (nDf40). At this nonpermissive temperature, mett-10(oz36) animals are sterile (Ste) or maternal-effect lethal (Mel), with somatic phenotypes including slow growth (Gro), protruding vulva (Pvl), and distension of the intestinal lumen (data not shown; Table 2; Figure 3, B and E).
We found two more mett-10 alleles by testing uncloned Ste/Pvl mutants mapping to the same genetic region for complementation of mett-10(oz36) at 25° (Figure 2B). mett-10(g38) was first identified as let-42 in screens for maternal-effect-lethal mutations (Cassada et al. 1981). mett-10(oj32) was isolated in a screen for temperature-sensitive sterile and uncoordinated mutants and was named stu-18 (O'Connell et al. 1998). On the basis of sequence conservation between C. elegans METT-10 and Human METT10D, the oj32 mutation (G110R) alters a conserved glycine that binds to S-adenosyl methionine, the substrate used by SAM-dependent methyltransferases, and thus should disrupt enzymatic activity (Figure 2; Malone et al. 1995).
Two mett-10 deletions were obtained from the Japanese National Bioresource Project (tm2697) and the C. elegans gene knockout consortium (ok2204) (Figure 2B). tm2697 is an internal in-frame deletion C-terminal to the core methyltransferase domain. The ok2204 deletion removes the entire methyltransferase domain and is predicted to be a null allele.
Like mett-10(oz36), all other mett-10 alleles exhibit temperature-sensitive Ste/Mel/Pvl phenotypes (Table 2; Figure 3, C and F). Germ-line sterility may stem from abnormalities in gonad morphology and defects in germ cell meiotic development. At 25°, 30% of mett-10(ok2204) germ lines have at least one extra protrusion in this distal gonad (Figure 4, A and B). Moreover, many germ lines contain pachytene cells in the proximal-most region of the germ line, which is normally occupied by diakinetic oocytes (data not shown). In situations where proximal mett-10(ok2204) germ cells progress from pachytene to diplotene, there is often disruption of normal linear oocyte organization (Figure 4, E and F).
In addition to abnormalities in meiotic development, and unlike mett-10(oz36) germ line tumors, mett-10(ok2204) germ cells have difficulty in progressing through mitotic cell cycles. At all temperatures, the proliferative zone of mett-10(ok2204) germ lines contains enlarged, diffuse nuclei (Figure 4, B and D). Since this nuclear morphology is associated with cell cycle arrest (Gartner et al. 2000; Moser et al. 2009), we tested whether these cells were cycling by EdU labeling and staining for phospho-histone-3. We found that the enlarged nuclei are not pH 3 positive and do not incorporate EdU during a time sufficient to label essentially all proliferative zone nuclei in wild-type germ lines (Figure 4, C and D). These data indicate that the enlarged nuclei arrest at some point other than M phase and that METT-10 promotes cell cycle progression. Alternatively, cell cycle arrest may arise from activation of a checkpoint. Overall, mett-10 loss-of-function germ lines are quite small at 25°, suggesting that METT-10 contributes to proliferation per se.
The somatic and germ line mett-10 loss-of-function defects are rescued by a wild-type mett-10 transgene (Table 2; data not shown) and, with the exception of maternal-effect embryonic lethality, they are rescued maternally (Table 2). Significantly, while all alleles exhibit the Ste/Mel/Pvl phenotype with variable penetrance at 25°, in contrast to mett-10(oz36), none are tumorous at any temperature tested (Table 3).
mett-10(oz36) produces a “poisonous” product:
Genetic experiments indicate that mett-10(oz36) likely encodes a product that interferes with processes in which mett-10(+) participates. If mett-10(oz36) encodes a poisonous product, the severity of the tumorous phenotype should depend on mett-10(oz36) dosage. The tumorous phenotype is recessive and completely suppressed by a wild-type transgene (Table 1). To increase mett-10(oz36) dosage, we took advantage of the fact that the mutation is a premature stop codon and that the transcript should be degraded by nonsense-mediated RNA decay (NMD). Disruption of NMD should therefore increase production of the truncated “poisonous product” (Hodgkin et al. 1989; Page et al. 1999). Indeed, disruption of NMD by the smg-2(e2008) mutation significantly increases penetrance of the mett-10(oz36) overproliferation phenotype to 98% (Table 1). Furthermore, tumors in smg-2(e2008); mett-10(oz36) animals are larger than in mett-10(oz36) animals (Figure 5, B and C). Interestingly, even in an NMD-defective background, mett-10(oz36) is recessive, indicating that wild-type METT-10 fully suppresses the tumorous phenotype (Table 1). Thus the tumorous phenotype of mett-10(oz36) mutants may be caused by simultaneous reduction of METT-10 function and poisoning of another protein by the truncated METT-10(oz36) product. Supporting this idea, heteroallelic mett-10(oz36)/mett-10(g38) mutants exhibit a partially penetrant tumorous phenotype that is less severe than mett-10(oz36) homozygotes in an NMD-defective background (Table 1).
NMD disruption also enhances maternal-effect embryonic lethality in the mett-10(oz36) mutant, as the few progeny laid by smg-2(e2008); mett-10(oz36) mothers at 15° or 20° die late in embryogenesis or shortly after hatching with various abnormalities (Figure 3, G–I). Thus, METT-10(oz36) expression may also disrupt embryonic development.
METT-10 inhibits germ cell proliferative fate:
A simple interpretation of the mett-10(oz36) tumorous phenotype is that wild-type METT-10 inhibits specification of germ cell proliferative fate. However, we also considered the possibility that the METT-10(oz36) poison indirectly affects proliferative fate specification and thus tested whether mett-10 loss-of-function alleles also affect the proliferative/meiotic fate decision. Because none of these alleles are tumorous on their own, we examined their ability to enhance tumor formation in animals with elevated glp-1/Notch activity. Here we employed glp-1(oz264) (Kerins 2006), which is close to wild type at 20°, but shows late-onset and proximal proliferation tumors at 25°, analogous to previously described weak glp-1(gf) alleles (Pepper et al. 2003). We found that mett-10 alleles enhance the glp-1(oz264) tumorous phenotype at 20° (Table 3), showing that METT-10 either inhibits germ cell proliferative fate and/or promotes meiotic entry.
The conclusion that METT-10 normally inhibits germ cell proliferation and/or promotes meiotic entry is supported by the finding that mett-10 loss-of-function alleles partially suppress the premature meiotic entry phenotype of a temperature-sensitive glp-1 loss-of-function mutant [glp-1(bn18)] (Kodoyianni et al. 1992; Qiao et al. 1995). Adult glp-1(bn18) animals raised at 20° and shifted to 25° for 6 hr have minimal proliferative zones and an absence of nuclei in M-phase (as shown by a lack of pH 3-positive nuclei) because removal of glp-1/Notch signaling induces all germ cells to enter meiosis (Kodoyianni et al. 1992; Qiao et al. 1995; P. M. Fox and T. Schedl, unpublished data; Figure 6, E, F, and J). However, in contrast to glp-1(bn18) mutants, we detected pH 3-positive nuclei, albeit reduced in number, in glp-1(bn18) mett-10(g38) double mutants after a 6-hr shift, despite equivalent numbers of pH 3-positive nuclei in glp-1(bn18) and glp-1(bn18) mett-10(g38) unshifted controls. (Figure 6, G, H, and J). The reduction in proliferative zone size and number of pH 3-positive nuclei in glp-1(bn18) mett-10(g38) germ lines upon temperature shift shows that germ cells are still induced to enter meiosis upon removal of glp-1 signaling in the presence of the mett-10 mutation, although with different temporal dynamics. Indeed, longer shifts of glp-1(bn18) mett-10(g38) animals result in meiotic entry of all germ cells (Figure 6I).
Given the central role of glp-1/Notch signaling in the specification of germ cell proliferative fate, we sought to determine if METT-10 normally acts downstream of, upstream of, or in parallel to glp-1 to inhibit germ cell proliferative fate. Several lines of evidence indicate that it is unlikely that GLP-1 normally acts to inhibit METT-10. The first makes use of the fact that METT-10(oz36) acts as an antimorphic protein, in that increased expression of the mutant protein drives the overproliferation phenotype toward the mett-10 loss-of-function direction. Thus, if GLP-1 acts upstream of METT-10, we would not expect the mett-10(oz36) tumorous phenotype to depend on glp-1 activity. However, tumor formation in mett-10(oz36) mutants is exquisitely sensitive to glp-1 signaling, and glp-1(bn18) suppresses the smg-2(e2008); mett-10(oz36) tumorous phenotype (but not the embryonic lethality) at the permissive temperature of 20° (Table 1, Figure 5, C–E). This argues that METT-10 may negatively regulate glp-1 signaling to inhibit specification of proliferative fate.
The idea that METT-10 acts upstream of or in parallel to glp-1 signaling to inhibit germ cell proliferation is further supported by our analysis of genetic interactions between mett-10 and the gld-1 and gld-2 pathways that function downstream of glp-1 to redundantly promote meiotic entry. In animals that lack either gld-1 or gld-2, germ cells enter meiosis but do not undergo normal meiotic progression (Francis et al. 1995b; Kadyk and Kimble 1998). Germ cells that have lost both gld-1 and gld-2 are defective in meiotic entry, resulting in tumors that are independent of glp-1 activity (Kadyk and Kimble 1998; Hansen et al. 2004a). We made double mutants between mett-10(oz36) and null alleles of gld-1, gld-2, and gld-3, which act in the gld-2 pathway (Eckmann et al. 2004). Because gld-1(q485) hermaphrodites have germ-line tumors caused by a return to mitosis, while males are unaffected (Francis et al. 1995a), we analyzed gld-1(q485); mett-10(oz36) males and found that mett-10(oz36) does not act redundantly with gld-1 to promote meiotic entry (Table 4). Although gld-2(q497) enhanced the mett-10(oz36) tumorous phenotype, this was completely suppressible by a reduction in glp-1 activity caused by the glp-1 null allele, glp-1(q175), or by the temperature-sensitive glp-1 hypomorph, glp-1(bn18), raised at the 20° permissive temperature (Table 4). Results similar to those for gld-2 were observed for gld-3 (Table 4). Thus, in contrast to gld-1 gld-2 synthetic tumors, mett-10(oz36) synthetic tumors are dependent on glp-1(+) activity, supporting a model in which mett-10 acts upstream of or in parallel to glp-1 to inhibit proliferative fate.
METT-10 is a nuclear protein expressed in many cell types:
To localize METT-10 protein in vivo, we expressed a GFP-tagged version under control of the endogenous promoter as an extrachromosomal array and as an integrated transgene that rescues all mett-10 mutant phenotypes (Tables 1–3⇑⇑). We found that METT-10∷GFP is a nuclear protein expressed in many of the tissues where mett-10 disruption causes phenotypic abnormalities, such as intestine, vulva, and cells of the somatic gonad including the distal tip cell, gonadal sheath cells, and spermatheca (Figure 7, A–E). The germ-line localization of METT-10∷GFP is consistent with its role in inhibition of germ cell proliferative fate: the levels are lower in the distal proliferative zone and increase as cells enter meiosis, although some proliferative nuclei are positive (Figure 8A). METT-10∷GFP persists throughout meiotic prophase as a predominantly nuclear protein that does not colocalize with DNA (Figure 7, F and G).
The germ-line expression pattern of METT-10 is altered in both glp-1 loss-of-function and gain-of-function mutant backgrounds. In line with the finding that METT-10 levels increase as germ cells enter meiosis, nuclear METT-10 is uniformly detected in distally located meiotic nuclei in shifted glp-1(bn18) animals in which almost all germ cells have entered meiosis (Figure 8B). However, there is also an increase in distal METT-10 levels when the glp-1(ar202) gain-of-function mutant is shifted to 25° to induce tumors, although groups of proliferating cells expressing lower levels of METT-10 can be found throughout the germ line (Figure 8C). It is possible that this reflects an expansion of the normally smaller proliferative zone population that accumulates METT-10 and that may require METT-10 for cell cycle progression. The finding that METT-10 expression does not inversely correlate with GLP-1 activity is consistent with the idea that METT-10 functions upstream of or in parallel to GLP-1 to antagonize its pro-proliferative functions.
METT-10 acts in the germ line to inhibit germ cell proliferative fate:
Since METT-10 is expressed in both the germ line and the soma, including cells of the somatic gonad that can affect germ cell fate (Kimble and White 1981; McCarter et al. 1997; Killian and Hubbard 2004; Voutev et al. 2006), we used mosaic analysis to determine whether METT-10 acts in the soma or the germ line to inhibit germ cell proliferative fate. We generated mosaics by taking advantage of the fact that extrachromosomal arrays are imperfectly transmitted during cell division (Yochem and Herman 2005). To form marked extrachromosomal arrays carrying wild-type mett-10, we co-injected a rescuing mett-10(+) construct with a plasmid carrying the sur-5∷GFP cell-autonomous marker that is expressed in almost all somatic cell nuclei (Yochem et al. 1998). Even with loss in the somatic gonad, the presence of the sur-5∷GFP; mett-10(+) array in the germ line (detected by transmission of the array to progeny) fully suppresses the tumorous phenotype of the smg-2(e2008); mett-10(oz36) double mutant (Figure 9). Conversely, expression of mett-10(+) in cells of the somatic gonad did not suppress the smg-2(e2008); mett-10(oz36) tumorous phenotype in the event of germ-line loss (Figure 9). We conclude that METT-10 acts in the germ line to inhibit the specification of proliferative fate.
Unlike the post-mitotic soma, the adult C. elegans germ line is dynamic; germ cells proliferate through mitosis to maintain their population and, at some frequency, switch to meiosis to produce gametes (Crittenden et al. 2003; Hansen and Schedl 2006; Kimble and Crittenden 2007). The temporal and spatial dynamics of this decision, as well as the conservation of the Notch pathway that governs germ cell proliferation, makes it an excellent system for studying the regulation of stem cell proliferation and fate specification. Here we identify a role for the highly conserved and previously uncharacterized METT-10 putative methyltransferase in inhibiting the choice of proliferative fate in the C. elegans germ line. In addition to a role in proliferative fate specification, we find that METT-10 is essential for the normal development of the vulva, somatic gonad, and embryo and contributes to progression of germ cells through both mitosis and meiosis. The role of METT-10 in promoting mitotic cell cycle progression, while inhibiting the specification of proliferative fate, delineates a genetic difference between the developmental decision of proliferative fate specification and the process of cell proliferation per se.
METT-10 acts as an inhibitor of germ cell proliferation in C. elegans:
Germ-line stem cell proliferation is promoted by the conserved glp-1/Notch signaling pathway (Crittenden et al. 2003; Hansen and Schedl 2006; Kimble and Crittenden 2007). One approach to identifying regulators of germ cell proliferation is to study mutations that phenocopy the germ cell overproliferation phenotype observed in Notch gain-of-function mutants (Berry et al. 1997; Pepper et al. 2003). mett-10(oz36) causes late-onset tumors, similar to glp-1 hyperactivation. Interestingly, while we can demonstrate that METT-10 normally plays a role in regulating germ cell proliferative fate (see below), mett-10(oz36) is unique in that it is the only mett-10 allele that is tumorous in an otherwise wild-type background.
Because mett-10(oz36) has, with respect to germ-line proliferation, a phenotype more severe than the null, it has properties beyond simple loss of METT-10 function that contribute to the tumorous phenotype. Our results argue that mett-10(oz36) not only reduces METT-10 function, but also produces a protein that poisons or interferes with unknown factor(s) to cause germ-line tumors; thus, the overproliferation observed in mett-10(oz36) mutants is likely a synergistic phenotype caused by simultaneous reduction of mett-10 function and the function of at least one other factor. Our model makes predictions that will aid in the identification of the factor(s) targeted by the METT-10(oz36) mutant protein, including not only that disrupting the function of the target should enhance overproliferation of mett-10 simple loss-of-function alleles, but also that it may physically interact with METT-10(oz36).
The fact that it takes a special mett-10 allele to recapitulate the glp-1/Notch gain-of-function phenotype highlights the redundancy governing glp-1 signal inhibition and germ cell fate specification and suggests that glp-1 occupies a unique “constriction point” in this genetic network. This is further supported by the identification of many negative regulators of Notch signaling that act at multiple steps in the signal transduction process, as well as by the redundancy between the downstream pathways controlling meiotic entry (Sundaram and Greenwald 1993; Hubbard et al. 1997; Crittenden et al. 2003; Hansen and Schedl 2006; Macdonald et al. 2008).
Although the unique mett-10(oz36) allele is needed to see the effects of mett-10 on the proliferative fate decision in otherwise wild-type animals, mett-10 loss-of-function alleles enhance the tumorous phenotype of glp-1 gain-of-function mutations while suppressing premature meiotic entry upon reduction of glp-1 signaling. Thus, METT-10 normally participates in this fate decision.
How does METT-10 inhibit germ cell proliferation? The accumulation of METT-10 in nuclei as germ cells enter meiosis suggests that it may act in the nucleus to switch from proliferative to meiotic fates. One possibility is that it inhibits the pro-proliferative functions of glp-1/Notch signaling because the mett-10-associated overproliferation phenotypes resemble glp-1 hyperactivation and require glp-1 activity. Moreover, genetic analysis suggests that mett-10 does not function in the known meiotic entry pathways that act downstream of GLP-1. While these data suggest that METT-10 may negatively regulate glp-1/Notch signaling, it is still possible that it acts in parallel to GLP-1. Moreover, we cannot say whether METT-10 inhibits Notch signaling in general, as we were unable to detect clear genetic interactions with alleles of lin-12, the other C. elegans Notch receptor (materials and methods).
It is likely that METT-10 is a methyltransferase and that this enzymatic activity is important for inhibiting proliferative fate. METT-10 contains a well-conserved SAM-dependent methyltransferase fold, which binds SAM in the human protein (Wu et al. 2009). The mett-10(oj32) mutation disrupts a key SAM-binding residue and the balance between germ cell proliferation and meiotic entry in a weak glp-1 gain-of-function genetic background. Nearly all of the many proteins that contain a SAM-dependent methyltransferase fold are methyltransferases (Martin and McMillan 2002). The two notable exceptions are putrescine aminopropyltransferase, which binds a decarboxylated form of SAM (and not SAM itself) to catalyze the transfer of an amino-propyl group, rather than a methyl group, during spermidine synthesis (Korolev et al. 2002), and Dnmt2, which contains a full methyltransferase fold and binds SAM, but lacks detectable DNA methyltransferase activity (Dong et al. 2001). Although methyltransferases can methylate protein, nucleic acid, small molecules, and lipids, in the absence of close homologs with known functions and target specificity, the biochemical nature of their target substrates cannot be easily predicted from protein or sequence structure (Martin and McMillan 2002). Thus, it is currently difficult to predict METT-10 substrates.
Loss of METT-10 function causes diverse developmental defects:
In addition to inhibiting proliferative fate in the distal germ line, mett-10 has multiple developmental functions. The phenotypes of mett-10 loss-of-function mutants, including abnormal vulva and gonad morphology and multiple defects in germ cell and embryonic development, testify to the potential diversity of mett-10 functions and targets. Intriguingly, many defects observed in these mutants, specifically in vulval morphogenesis and germ cell mitotic progression, indicate defects in cell division. Multiple screens for cell division mutants have been based on these phenotypes, including the one that identified the mett-10(oj32) allele (O'Connell et al. 1998; Fay and Han 2000). Indeed, we have shown that loss of mett-10 function leads to a cell cycle progression defect in germ cells, as evidenced by enlarged nuclei that fail to cycle. The role for mett-10 in enabling cell division, coupled with its developmental role as an inhibitor of proliferative fate, highlights the importance of distinguishing the specification of developmental cell fates from the cellular mechanisms that ensure their proper execution.
We thank the Caenorhabditis Genetics Center, the Japanese National Bioresource Project, the C. elegans Gene Knockout Consortium, and Randall Cassada for strains and the Structural Genomics Consortium for making the crystal structure of human METT10D freely accessible. We thank Momoyo Hanazawa for constructing the glp-1(oz264) mett-10(g38) strain and Jim Collins for help with injections for mosaic analysis. We extend our sincerest thanks to Swathi Arur, Dale Dorsett, Kevin O'Connell, Jim Skeath, Paul Fox, Justin Fay, Mike Nonet, and three reviewers for experimental suggestions, provisions of reagents, or comments on the manuscript. This work was supported by GM63310 (T.S.) from the National Institute of General Medical Sciences.
↵1 Present address: Cambria Pharmaceuticals, Inc., 8A Henshaw St., Woburn, MA 01801.
Communicating editor: D. I. Greenstein
- Received May 20, 2009.
- Accepted July 11, 2009.
- Copyright © 2009 by the Genetics Society of America