A Targeted RNAi Screen Identifies nanos Family Genes as meg-1 Interactors
In order to gain new insight into the mechanisms of meg-1 function, we conducted a targeted RNAi screen in a meg-1 mutant background to identify factors that either enhanced or suppressed the meg-1 phenotype. We selected 90 germline-expressed genes (Reinke et al.,2004) from the Ahringer RNAi feeding library (Kamath and Ahringer,2003), giving preference to genes with predicted roles in RNA metabolism, as this process is likely to be affected by MEG-1, given its localization to P granules and interactions with PGL-1 and GLH-1. Additionally, we restricted our selection of genes to those not previously reported to induce sterility when depleted by RNAi. The full list of genes tested can be found in Supporting Information File 1.
Loss of meg-1 function results in a partially penetrant maternal-effect sterility, or “grandchildless” phenotype, in that homozygous mutants born from heterozygous mothers are fertile, but ∼60% of their offspring are sterile at 25°C (Leacock and Reinke,2008). To identify genes that influence meg-1-induced sterility, we placed meg-1 mutant animals at 25°C as L4 larvae, and fed them bacteria expressing dsRNA corresponding to individual candidate genes to induce an RNAi response. Small numbers of F1s were transferred to a new plate, and we assessed the amount of F2 progeny as an indicator of F1 sterility (Fig. 2a). Genes that enhanced the penetrance of the meg-1 phenotype decreased the number of F2 worms produced, while genes that suppressed the phenotype increased the size of the F2 generation. The negative control included a mock RNAi treatment (L4440), while the positive control consisted of meg-2(RNAi) treatment, which increases the penetrance of the meg-1 phenotype to 100% (Leacock and Reinke,2008). We screened each candidate gene in triplicate and concluded that a gene was affecting the penetrance of the meg-1 phenotype if it consistently increased or decreased the number of progeny in at least two out of the three wells.
Figure 2. New meg-1 genetic interactions identified through a targeted RNAi screen. (a) RNAi screen design. Synchronized meg-1(vr10) L4 worms were added to wells containing bacteria from the Ahringer RNAi feeding library and grown in liquid culture at 25°C for 36 h. F1 progeny were transferred to new wells and grown for another 54 h. The number of F2 worms in each well was then compared to the number produced by meg-1 animals given a mock RNAi treatment. (b) RNAi depletion of five genes affected the penetrance of the meg-1 phenotype. The results from a secondary screen conducted on plates are plotted. The percentage of sterile progeny was determined by scoring for the presence or absence of eggs in the uterus. Error bars, standard deviation.
Download figure to PowerPoint
Genes we identified as affecting the meg-1 phenotype in the first round of screening were then subjected to secondary screening on individual plates rather than in culture wells, and in wild type animals as well as in meg-1 mutants (see Supporting Information File 1). We then isolated progeny and determined the sterility of individual worms. Of the 90 genes tested, four significantly increased sterility in meg-1 mutants (pgl-2, C37A2.8, nos-2, and T02G5.11; Fig. 2b). RNAi-mediated depletion of C37A2.8, a gene of unknown function, produced a modest increase in the penetrance of sterility in the F1 generation, from 60 to 72%, in meg-1 mutants, but did not induce any sterility in wild-type animals. pgl-2(RNAi) increased the penetrance of sterility to 80% in the progeny of meg-1 mutants but also caused 15% of wild type worms to become sterile, suggesting that the enhanced sterility in the meg-1 mutant background is merely additive. Moreover, pgl-2(RNAi) did not enhance meg-1 sterility at 20°C (data not shown), and thus was not pursued further.
The other two genes whose reduction enhanced meg-1(vr10) sterility are two closely related sequences, nos-2 and T02G5.11. nos-2 is a member of the nanos family of germ cell regulatory genes, is expressed in embryonic germ cells, and is required redundantly with nos-1 for germ cell maintenance (Subramaniam and Seydoux,1999). nos-2(RNAi) produced the most dramatic increase in the penetrance of sterility observed in the screen, as greater than 99% of F1 meg-1(vr10); nos-2(RNAi) animals were sterile. Similarly, T02G5.11(RNAi) resulted in 90% sterility in the F1 generation of meg-1 mutants. However, the sequence used to target T02G5.11 for RNAi corresponds to the most highly conserved region between T02G5.11 and nos-2 (>300 bp of uninterrupted identical sequence), and T02G5.11 lacks the characteristic zinc binding motifs that true members of the nanos family possess. It is therefore likely that T02G5.11(RNAi) is affecting nos-2 levels as well. The sequence targeted in nos-2(RNAi), on the other hand, corresponds to a non-conserved region that should not cross-react with T02G5.11 or any other nanos family members.
In addition to the four genes identified that increased the penetrance of the meg-1 phenotype, RNAi of a single gene, nos-3, decreased sterility in the F1 generation of meg-1 mutants to 35% from 60% (Fig. 2b). nos-3 is also a member of the nanos family and is present continually in the germ line, where it interacts with the FBF proteins to promote the switch from sperm to oocyte production in hermaphrodites, in a functionally redundant manner with nos-1 and nos-2 (Kraemer et al.,1999). Additionally, nos-3 promotes the translation of GLD-1, a key pro-meiotic regulator (Hansen et al.,2004).
In sum, our screen identified two of the three characterized C. elegans nanos family members as meg-1 interactors, although these genes displayed opposing effects on the meg-1 phenotype. These opposing effects are not surprising, as nos-3 appears to function quite differently from nos-1 and nos-2. nos-1 and nos-2 are required for germ cell survival, while nos-3 is not. Moreover, nos-3(RNAi) can suppress the lethality of par-2 embryonic polarity mutants (Labbé et al.,2006). Despite their different phenotypes, both nos-2 and nos-3 are expressed in embryonic germ cells during the time of MEG-1 expression (only maternal NOS-2 contributes to its expression from the 28-cell to 500-cell stage, while maternal and zygotic NOS-3 contribution overlaps for continuous expression), making both genes promising targets to shed light on meg-1 function.
Loss of nos-3 and gld-1 Restore Proliferation in meg-1 Mutants
Because the suppression of the meg-1 phenotype by nos-3(RNAi) was so striking, we decided to focus on this interaction first. We first repeated the RNAi experiment and assessed individual animals on plates rather than in liquid culture, and found that suppression of the meg-1(vr10) phenotype by nos-3(RNAi) was reproducible (Fig. 3a). Importantly, nos-3(RNAi) also suppressed the essentially complete sterility of meg-1(vr10); meg-2(RNAi) animals. The progeny of nos-3(RNAi); meg-1(vr10) meg-2(RNAi) animals exhibited a significant reduction of sterility compared to meg-1(vr10); meg-2(RNAi) animals, both at 20°C (from 93 to 67%; P < 0.001) and 25°C (from 99 to 79%). Loss of nos-3 can thus suppress the Meg phenotype in even the most compromised germ cells. To further confirm the RNAi results, we generated a nos-3(q650); meg-1(vr10) mutant strain. At 25°C, nos-3(q650); meg-1(vr10) mutants have an incidence of sterile progeny similar to nos-3(RNAi); meg-1(vr10) animals (31% compared to 35%; Table 1), indicating that loss of nos-3 is indeed capable of suppressing the meg-1 mutant phenotype (Fig. 3b,c).
Figure 3. Loss of nos-3 suppresses the meg-1(vr10) phenotype. (a) meg-1(vr10) animals were fed bacteria expressing dsRNA targeting L4440 (mock RNAi), nos-3, a 50:50 mixture of meg-2 and L4440, or a 50:50 mixture of meg-2 and nos-3 at 20 and 25°C. The percentage of sterile progeny was determined by scoring for the presence or absence of eggs in the uterus. Error bars, standard deviation (*P < 0.05), **P < 0.005). (b–e) DIC images from F1 animals raised at 25°C. Percentages are for the incidence of the phenotype of the total worms. (b) Fertile nos-3(q650); meg-1(vr10) double mutant. (c) Sterile meg-1(vr10) mutant. (d) Sterile nos-3(q650); meg-1(vr10) double mutant with under-proliferated gonad. (e) Sterile nos-3(q650); meg-1(vr10) double mutant with an intermediate phenotype. Scale bar, 50 μM. Asterisks indicate the distal end of the gonad.
Download figure to PowerPoint
Table 1. meg-1 Interactions With the nanos Family
|Genotype||% sterile F1 worms at 25°C||n|
|nos-2(RNAi); nos-3(q650); meg-1(vr10)||99||227|
Despite this suppression, a large fraction of nos-3(q650); meg-1(vr10) mutant animals are still sterile. Observation of nos-3(q650); meg-1(vr10) mutant gonads at elevated temperatures by DIC revealed that most of the sterile worms possessed severely under-proliferated gonads similar to meg-1 mutants (Fig. 3d), but a small fraction of the total (5%) had fully proliferated gonads but lacked differentiated oocytes (Fig. 3e), a phenotype never seen in the meg-1 mutant. This phenotype might represent a rescue of the meg-1 proliferation defect but not other aspects of germ cell differentiation.
NOS-3 positively regulates translation of GLD-1 in the adult germ line (Hansen et al.,2004). GLD-1 is a central regulator of the switch from mitosis to meiosis in the adult gonad, where it represses the translation of mitotic transcripts such as glp-1, allowing germ cells to enter meiosis (Crittenden et al.,2003; Hansen et al.,2004; Marin and Evans,2003). Thus, loss of nos-3 might suppress the meg-1 phenotype by decreasing GLD-1 levels, promoting germ cell proliferation and rescuing sterility.
Therefore, to test whether gld-1 is involved in suppression of the meg-1 phenotype, we generated gld-1(q485); meg-1(vr10) mutants. gld-1(q485) mutant germ cells are unable to progress through meiosis and instead re-enter mitosis, producing over-proliferated, tumorous gonads (Francis et al.,1995 ; Subramaniam and Seydoux, 1999; Kraemer et al., 1999). gld-1; meg-1 mutants at 25°C have fewer under-proliferated gonads when compared to meg-1 mutants (from 60% to less than 20%; Fig. 4a,b), while at the same time also have fewer over-proliferated gonads compared to gld-1 mutants (from 100% in gld-1 mutants to 80%; Fig. 4c). We observed similar results in the progeny of gld-1(q485); meg-1(vr10) meg-2(RNAi) animals, with the proportion of under-proliferated gonads decreasing from 98% in the meg-1(vr10); meg-2(RNAi) animals to 23% in gld-1(q485); meg-1(vr10) meg-2(RNAi) worms (Fig. 4a).
Figure 4. Loss of gld-1 rescues proliferation in meg-1 mutants. (a) gld-1(q485/+); meg-1(vr10) double mutants were fed bacteria expressing dsRNA targeting L4440 or meg-2 on RNAi feeding plates at 25°C. Progeny homozygous for gld-1(q485) were compared by DIC to the gld-1 over-proliferation phenotype (*P < 0.05), **P < 0.005) (b). Example of an over-proliferated (c) and under-proliferated (d) gld-1(q485); meg-1(vr10) gonad. Asterisks indicate the distal end of the gonad. Scale bar, 50 μM. (e) WT and meg-1(vr10) animals were fed bacteria expressing dsRNA targeting L4440 or meg-2 at 25°C. Progeny were moved to plates with bacteria targeting gld-1 and the sterility of these animals assessed. Error bars, standard deviation. (f) The progeny of meg-1(vr10) meg-2(RNAi) animals were subjected to gld-1, nos-3, or L4440 RNAi beginning at L1, and assessed for germ cell number as adults. Gonads were classified as either wild type, moderately under-proliferated (with undifferentiated germ cells), or severely under-proliferated (<50 germ cells).
Download figure to PowerPoint
GLD-1 is expressed embryonically as well as in larval and adult gonads (Jones et al.,1996). To test whether the critical period of gld-1 function was during embryogenesis or larval development, we shifted either meg-1(vr10) or meg-1(vr10) meg-2(RNAi) animals to 25°C at the L4 stage, and allowed them to produce progeny. We then performed gld-1(RNAi) in early F1 larva. Reducing the larval expression of gld-1 produced a significant, though mild, reduction in sterility compared to meg-1(vr10) mutants (from 58 to 40%; P < 0.05; Fig. 4d). Additionally, larval gld-1(RNAi) reduced sterility in meg-1(vr10) meg-2(RNAi) animals (from 98 to 78%; P < 0.01). Therefore, the suppression of the Meg phenotype by loss of gld-1 activity can occur post-embryonically. However, these effects were less complete than when the gld-1(q485) allele was used, either because RNAi is less effective, or because gld-1 function in the embryo also affects meg-1 activity.
Notably, larval gld-1(RNAi); meg-1(vr10) meg-2(RNAi) animals had dramatically fewer gonads with severe under-proliferation and more with moderate under-proliferation compared to meg-1(vr10) meg-2(RNAi) animals (Fig. 4f). Additionally, the over-proliferation phenotype seen in gld-1(q485); meg-1(vr10) meg-2(RNAi) animals was not observed in the gld-1(RNAi); meg-1(vr10) meg-2(RNAi) animals, indicating that RNAi was not completely effective in abolishing gld-1 activity. Together, these results suggest that the degree to which germ cells proliferate is determined by a balance between meg-1 and gld-1 activity.
Hyperactivation of Notch Signaling Rescues Proliferation Defects in meg-1Mutants
One of the main forces driving proliferation in the C. elegans gonad is Notch signaling from the somatic distal tip cell to adjacent undifferentiated germ cells via the Notch receptor GLP-1 (Crittenden et al.,2003; Kimble and Simpson,1997). Notch signaling also results in the translational repression of gld-1 mRNA in the distal region of the gonad (Crittenden et al.,2002). However, as germ cells move away from the distal tip cell, GLD-1 levels increase and GLD-1 then prevents the translation of GLP-1/Notch (Hansen et al.,2004). This switch from GLP-1 to GLD-1 activity contributes to the transition of germ cells from proliferation to meiosis and differentiation (Marin and Evans,2003). We therefore investigated whether loss of MEG activity impairs the ability of germ cells to respond to Notch signaling. We therefore performed RNAi of meg-1 and/or meg-2 in a heterozygous strain carrying a glp-1 gain-of-function allele balanced by a glp-1 loss-of-function allele, and analyzed the offspring bearing different glp-1 alleles. Offspring homozygous for the glp-1(gf) allele have two copies of the constitutively active Notch receptor and consequently have over-proliferated tumorous gonads (Berry et al.,1997). Performing meg-1(RNAi) in glp-1(gf) mutants did not diminish the glp-1(gf) tumorous phenotype (data not shown). However, meg-1(RNAi) meg-2(RNAi) in glp-1(gf) mutants did result in under-proliferation of 24% of gonads (see Fig. 5d), indicating that extensive loss of MEG activity could suppress the glp-1(gf) phenotype. Additionally, the slight reduction in GLP-1 activity of glp-1(gf)/glp-1(lf) animals, which have only one copy of the constitutively active Notch receptor, led to even stronger effects by meg-1(RNAi) meg-2(RNAi). The percentage of under-proliferated gonads in glp-1(gf)/glp-1(lf); meg-1(RNAi) meg-2(RNAi) animals at 20°C increased to 77% from the 0% seen in controls. We conclude that MEG-deficient germ cells are capable of undergoing proliferation when Notch signaling is elevated, but that proliferation is partially impaired.
Figure 5. Hyper-activation of Notch signaling rescues proliferation in meg-1 mutants. (a) WT and glp-1(gf/lf) animals were subjected to RNAi for L4440 or a mixture of meg-1 and meg-2 at 20°C. Progeny were then DAPI stained and gonads were either classified as (b) wild-type, (c) over-proliferated, (d) under-proliferated. Scale bar, 50 μM (*P < 0.05).
Download figure to PowerPoint
Combined Depletion of meg-1 and nos-2 Can Result in the Loss of Primordial Germ Cells Z2/Z3
Because of the opposing effects on meg-1 function by nos-3 and nos-2, we decided to further investigate the enhancement of the meg-1 phenotype by nos-2 as well. In C. elegans, the phenotypes of many mutants that disrupt P granule function are only evident when subjected to the stress of elevated temperatures (Updike and Strome,2010), likely due to some functional redundancy among family members, which provide a buffering effect at low temperatures (Kawasaki et al.,2004; Leacock and Reinke,2008). However, even at the low temperature of 20°C, RNAi depletion of nos-2 in meg-1 mutants increased the frequency of sterility in progeny to greater than 90% (P < 0.001) while causing little sterility in the wild type background (Fig. 6a), suggesting a severe defect. The nos-2(RNAi); meg-1(vr10) phenotype at 20°C is similar to that of meg-1(vr10) animals raised at 25°C, with a small number of aberrant, undifferentiated germ cells present in severely under-proliferated gonads.
Figure 6. nos-2 and meg-1 are required for germ cell maintenance. (a) WT and meg-1(vr10) animals were subjected to nos-2 or L4440 RNAi at 20 or 25°C. The percentage of sterile progeny was determined by scoring for the presence or absence of eggs in the uterus. Error bars, standard deviation (*P < 0.05), **P < 0.005). (b–f) RFP and DIC images of L1 (b) PGL-1::RFP; L4440(RNAi) (c) PGL-1::RFP; nos-2(RNAi), (d) PGL-1::RFP; meg-1(vr10), (e, f) PGL-1::RFP; nos-2(RNAi); meg-1(vr10) animals. Brackets indicate location of primordial germ cells, Z2 and Z3. Scale bar, 10 μM.
Download figure to PowerPoint
Interestingly, we found that raising the temperature to 26°C caused nos-2(RNAi); meg-1(vr10) animals to produce some progeny that developed into adults with gonads entirely devoid of germ cells. To determine more precisely when germ cells in the nos-2(RNAi); meg-1(vr10) animals were first being affected, we utilized a PGL-1::RFP reporter transgene that marks P granules to follow germ cells during embryogenesis and the first larval stage. In the progeny of PGL-1::RFP; nos-2(RNAi); meg-1(vr10) animals, we found that Z2 and Z3 are initially formed, but in 9% of embryos after the ∼500-cell stage, P granules lost their perinuclear localization, dispersing into the cytoplasm (Table 2). The percentage of animals without perinuclear P granules increased to 71% by 6-h post-hatching, with 57% of gonads failing to proliferate beyond the first two primordial germ cells, Z2 and Z3. DIC microscopy revealed that the germ cells are degenerating, and had grossly altered morphology, often lacking recognizable nuclei (Fig. 6b–f). By contrast, in PGL-1::RFP; nos-2(RNAi) animals at 26°C, Z2 and Z3 are maintained and proliferate through the first two larval stages, after which point they too begin to degenerate and die. The combined loss of meg-1 and nos-2 thus seems to increase the severity of the nos-2 phenotype, as well as advance its onset from the end of the second larval stage to, in some cases, late embryogenesis, resulting in the loss of Z2 and Z3.
Table 2. meg-1 and nos-2 Are Required for Germ Cell Maintenance at 26°C
|Genotype||Embryonic||Larval||% of L1s with more than two germ cells|
|% of late stage embryos with perinuclear P granules (n)||% of L1s with perinuclear P granules|
|PGL-1::RFP; nos-2(RNAi)||100 (52)||100||100|
|PGL-1::RFP; meg-1(vr10); L4440||99 (27)||88||100|
|PGL-1::RFP; meg-1(vr10); nos-2(RNAi)||91 (20)||29||43|
We noticed that occasionally the germ cells of meg-1(vr10); nos-2(RNAi) L1 animals adopted the “button” morphology characteristic of programmed cell death. To test whether germ cells were undergoing apoptosis, we performed nos-2(RNAi) in ced-4(n1162); meg-1(vr10) mutants, which are unable to undergo apoptosis (Ellis and Horvitz,1986). At 26°C, the progeny of these animals still underwent germ cell degeneration, and 60% of gonads failed to proliferate beyond Z2 and Z3 (Table 3). We also tested whether the dying germ cells are engulfed by crossing meg-1(vr10) with ced-5(n1812) mutants, which are defective for engulfment of cell corpses (Ellis et al.,1991). RNAi depletion of nos-2 in ced-5(n1812); meg-1(vr10) mutants did not prevent germ cell degeneration, and 63% of gonads still failed to proliferate beyond more than two germ cells. The loss of germ cells observed in meg-1(vr10); nos-2(RNAi) animals therefore appear to be independent of the apoptotic pathway as well as the ced-5 corpse engulfment pathway.
Table 3. Apoptosis and Engulfment Do Not Mediate nos-2(RNAi); meg-1(vr10) Germ Cell Degeneration
|Genotype||% of L1 with more than two germ cells||n|