A Genetic Cascade of let-7-ncl-1-fib-1 Modulates Nucleolar Size and rRNA Pool in Caenorhabditis elegans

Ribosome biogenesis takes place in the nucleolus, the size of which is often coordinated with cell growth and development. However, how metazoans control nucleolar size remains largely unknown. Caenorhabditis elegans provides a good model to address this question owing to distinct tissue distribution of nucleolar sizes and a mutant, ncl-1, which exhibits larger nucleoli than wild-type worms. Here, through a series of loss-of-function analyses, we report that the nucleolar size is regulated by a circuitry composed of microRNA let-7, translation repressor NCL-1, and a major nucleolar pre-rRNA processing protein FIB-1/fibrillarin. In cooperation with RNA binding proteins PUF and NOS, NCL-1 suppressed the translation of FIB-1/fibrillarin, while let-7 targeted the 3’UTR of ncl-1 and inhibited its expression. Consequently, the abundance of FIB-1 is tightly controlled and correlated with the nucleolar size. Together, our findings highlight a novel genetic cascade by which post-transcriptional regulators interplay in developmental control of nucleolar size and function.


Introduction
Among the RNA/protein bodies within the nucleus, nucleoli bear the essential function of being the factories for ribosome subunit production and assembly, a stress sensor for cell cycle control, as well as a site for hepatitis D virus (HDV) replication and adenovirus-associated virus (AAV) assembly [1][2][3]. The size and morphology of the nucleolus is a cytological manifestation of ribosome biogenesis and therefore protein biosynthesis and is closely coordinated with cell growth and development [4]. Accordingly, these attributes sometimes are also physiological indicators of cell cycle, cancer growth and malignancy as well as stem cells differentiation and pluripotency [5,6]. However, without membrane delimitation, the principles that define nucleoli size and shape are poorly understood. Furthermore, spatiotemporal regulation of nucleolar size and output, particularly in coordination with development and in nondividing cells, are not fully characterized.
Caenorhabditis elegans represents an exploitable model for further interrogating nucleolus biology owing to distinct distribution of nucleolar sizes in different cell types. A C. elegans mutant, ncl-1, described as a recessive mutation with enlarged nucleoli in nearly all cells of the worm [7,8], has this phenotype consistent with its role as a suppressor of rRNA biosynthesis. C. elegans ncl-1 phenotypes can be rescued by its Drosophila homolog, brat [9,10]. Mutations in the fly brat gene have a similar phenotype to the defect of ncl-1 mutants in C. elegans, affecting nucleolar size. In addition, brat mutants induce brain tumor formation [11]. These homologous proteins belong to a TRIM/RBCC/NHL (NCL-1, HT2A, and LIN-41] family characterized by the presence of a RING domain, a B-box zinc finger, and a coil-coiled domain [12,13]. Because its lack of an RNA binding motif, Brat protein was previously shown to associate with the 3'UTR of the hunchback transcript in partnership with two RNA-binding proteins Pumillio (PUF) and Nanos (NOS), and suppress expression of Hunchback protein at the translational level [14].
In this study, we dissected the molecular mechanism through which NCL-1 controls nucleolar size and function, and pinpointed fibrillarin-the rRNA 2'-O-methyltransferase and pre-rRNA processing factor [1,[15][16][17]-as a downstream effector. Further, this regulation is dynamically coordinated with development as part of a functional axis driven by let-7, a critical developmental regulator of heterochronic development in worms and flies [18][19][20] and of cancer formation and stem cell maintenance in the mammals [21].

Results
Suppression of nucleolar size and rRNA expression by NCL-1 is associated with nucleolar protein FIB-1 Although abundantly expressed in the gonads of C. elegans [9], the effect of ncl-1 on the nucleoli of germ cells was not characterized. In intact gonads of wild-type (N2) young adult worms, nucleolar structure is nearly absent in the -1 oocyte, which is immediately adjacent to the spermatheca (Fig 1A, upper panel). In contrast, the nucleolus was readily detectable in the -1 oocyte of ncl-1 (e1942) mutant (Fig 1A, lower panel). While nucleoli were evident in the germ cells and the -3 and -2 oocytes of both worms, ncl-1 worms exhibited considerably larger average nucleoli size ranging from 119% to 176% of wild-type diameter (Fig 1A and 1B). Profiling of the ncl-1 mRNA expression by RT-qPCR revealed a progressive decline in mRNA abundance from the embryo to and throughout the four larva stages, followed by subsequent up-regulation in the adult (S1 Fig). This developmental stage-specific expression is consistent with previous in situ immuno-staining of NCL-1 that demonstrated its expression in the proximal gonad and early embryos and the subsequent gradual disappearance in the late stages of embryos [9]. Further, this expression is in line with the non-detectable to small sizes of nucleoli in the -1 oocyte and early embryos (Fig 1C, left  panel), supporting the notion that NCL-1 is a negative regulator of nucleolar size.
We also examined nucleolar morphology in worms devoid of functional fib-1. Consistent with its significance, fib-1 mutation led to lethality [22]; we thus characterized fib-1 mutant larvae (L1 stage) and found that nucleoli therein displayed size reduction (S2 Fig). To next determine if FIB-1/fibrillarin is involved in the nucleolar appearance and size, we depleted fib-1 in ncl-1(e1942) worms by RNAi feeding and measured the nucleolar size. The increase in , were quantitatively determined. Asterisks signify differences between the two worms: ***P < 0.0001; n = 8-14 gonad arms. (C) DIC microscopy of the blastomeres of wild-type (wt) and the ncl-1 embryos, with or without fib-1 RNAi. Each blastomere is indicated as ABa, ABp, EMS and P. Insets represent enlarged versions of the boxed regions of ABp cells to highlight the nucleoli. Scale bar, 20 μm. (D) Quantitative representation of the results shown in (C), which illustrates the distribution of nucleolar areas in the four blastomeres. Asterisk signifies difference between the indicated strains: *P < 0.05; n 31 embryos. (E) Knockdown of fib-1 was done in the indicated worms. The expression of Actin (lower panel) and the endogenous FIB-1 (upper panel) were examined by Western blot analysis. (F) RT-qPCR analysis of 26S rRNA expression in the indicated strains of worms as shown in (E) *P < 0.05; ***P < 0.001; ns, no significant; n = 3.
doi:10.1371/journal.pgen.1005580.g001 nucleolar size in the blastomeres of ncl-1(e1942) worms ( Fig 1C, middle panel) was significantly reversed by fib-1 abrogation as shown by image analysis (Fig 1C, right panel, and 1D). This observation supports a notion that the amount of FIB-1 expression is directly associated with the control of nucleolar size by NCL-1. Moreover, Western blot and RT-qPCR analyses showed that worms expressing a greater amount of FIB-1 generally had a higher level of rRNA abundance (Fig 1D and 1F). Conversely, knockdown of FIB-1 led to an overall reduction in the rRNA levels, further indicating a positive role of FIB-1 in this functional regard.

NCL-1 is a suppressor of FIB-1 expression
To examine whether NCL-1-mediated nucleolar size alternations is through the regulation of FIB-1 expression, we generated a pair of transgenic worms that express FIB-1::GFP chimeric protein in both the N2 and ncl-1 backgrounds [respectively designated as cguIs1 (strain SJL1) and ncl-1(e1942); cguIs1 (strain SJL14), see S1 Table]. Time-lapse fluorescence microscopy of embryos was performed to trace the level of GFP expression during early stages, and showed progressively higher GFP signals (Fig 2A and S1-S3 Movies). Dynamic up-regulation of GFP levels was more prominent in the ncl-1(e1942); cguIs1 embryos (62.8%) than in cguIs1 (26.9%) (Fig 2B). Random collections of embryos from both transgenic worms were further examined to quantify the GFP intensity of each embryo in the same field ( Fig 2C) and subsequently revealed that the embryos in the absence of NCL-1 exhibited higher levels of FIB-1::GFP (about 2 fold) ( Fig 2D). Further expression analyses consistently showed elevated levels of FIB-1 in ncl-1(e1942) embryos (5.2 fold) and adult worms (1.7 fold) ( Fig 2E). Unexpectedly, RT-qPCR analysis revealed comparable levels of fib-1 mRNA in wild type and ncl-1(e1942) in embryo and adult stages (Fig 2F). Taken together, these findings indicate that ncl-1 is an upstream negative regulator of fib-1 expression at the post-transcriptional/translational stage.

NCL-1 cooperates with PUF and NANOS to modulate fib-1 mRNA translation
We next aimed to test whether NCL-1 acts as its fly homologue Brat, which suppresses its target gene at the translational level by binding to the 3' UTR of transcripts [14]. Towards this end, we created two more pairs of transgenic worms [cguIs2 and ncl-1(e1942); cguIs2 (strain SJL2/strain SJL15), and cguIs19 and ncl-1(e1942); cguIs19 (strain SJL34/strain SJL38), see S1 Table]; SJL2 and SJL15 harbored a plasmid similar to cguIs1 worms that contains the fulllength fib-1 3' UTR, while in SJL34 and SJL38 the fib-1 3' UTR was replaced with unc-54 3' UTR sequence ( Fig 3A). In agreement with the above observations, enlarged nucleoli and a significantly increased levels of FIB-1::GFP expression were both evident in the tail hypodermis of ncl-1(e1942); cguIs1 and ncl-1(e1942); cguIs2 worms ( Fig 3B, top two panels at right, and S3 Fig). In contrast, for the transgene harboring the unc-54 3' UTR, ncl-1 inactivation did not lead to discernable difference in GFP intensity, despite the occurrence of enlarged nucleoli of cells in cguIs19; ncl-1 transgenic worms ( Fig 3C). These observations and the quantitative data for the whole worms ( Fig 3D and 3E) strongly support the notion that, rather than being the consequence of altered nucleolus, the suppression of FIB-1 may arise from direct targeting of its 3' UTR by NCL-1.
Since Brat mediates its repressive role through other RNA-binding factors, we further tested the roles of C. elegans pumillio and nanos in the translational suppression of fib-1. A potentially direct involvement of these RNA-binding proteins was first supported by the sequence analysis of the fib-1 3'UTR, which revealed a consensus PUF binding motif ( Fig 4A). To demonstrate the link between this 3'UTR element and NCL-1-dependent control, we then generated worms with 3'UTR reporter carrying mutations in the PUF binding sequence (cguEx18; Figs 3A and 4A) [23,24]. Fluorescence microscopy showed that, in comparison to the wild-type reporter ( Fig 3B and 3D), this particular transgene exhibited considerably diminished responsiveness to the loss of ncl-1 (Fig 4B and 4C), giving rise to a lower level of fluorescence intensity. In further support to the roles of the PUF proteins, RNAi knockdown puf-5, puf-8 and puf-9 and nos-2 in cguIs1 worms resulted in the appearance of brighter GFP signals (Fig 4D and 4E). However, such effect of nos/puf knockdown (puf-8 and puf-9 in particular) on the GFP reporter was reduced in the cguEx18 worms, in which the PUF binding sequence was altered ( Fig 4F). Consistently with the ncl-1 knockdown and mutant worms, immunoblotting showed a rise in FIB-1::GFP abundance in these knockdown worms ( Fig 4G). Collectively, these data imply that ncl-1 may coordinate with puf-5, -8, -9 and nos-2 to act directly on the 3' UTR element of fib-1, likely through a similar regulatory mechanism exhibited by brat, pumillio and nanos in the fly   [14]. This demonstration of a response element in the fib-1 3'UTR and its regulatory relevance would certainly strengthen a specific and direct control mechanism.
We next interrogated the significance of let-7 in the regulation of nucleolar size by assessing vulva cells in the temperature-sensitive, loss-of-function let-7(n2853) mutants. Mutants grown at non-permissive temperature (25°C) displayed a significant reduction in nucleolar size in these cells, by 25% as compared to those at permissive condition (15°C) (Fig 5F and 5G). However, such temperature-sensitive nucleolar size alteration was not observed in a double mutant let-7; ncl-1 (strain SJL39) (Fig 5F and 5G), implying that let-7 acts upstream of ncl-1 transcript to directly suppress NCL-1 translation and regulate nucleolar sizes of the vulva cells. We further verified the link of let-7 to NCL-1-mediated regulation by assessing downstream FIB-1 expression and rRNA abundance in let-7(n2853) and let-7(n2853); ncl-1(e1942) worms. To this end, expression profiling revealed higher amounts of both FIB-1 ( Fig 5H) and ribosomal RNA species (Fig 5H and 5I and S5 Fig) in let-7(n2853) worms grown at 15°C vs. 25°C, in contrast to a lack of discernable differences in the let-7(n2853); ncl-1(e1942) worms between these rearing temperatures (Fig 5H and 5I, and S5 Fig). Such loss of phenotypes in the ncl-1(e1942) background is in agreement with let-7-ncl-1 interaction and functional antagonism. Based on these findings, we hypothesize that the genetic circuit of let-7-ncl-1-fib-1 constitutes a critical determinant in the regulation of nucleolar size and rRNA pool (Fig 6).
Discussion let-7 is known as a critical regulator of heterochronic development in worms and flies [18,29]. Our studies outlined for the first time a genetic cascade through which the coordinated actions of let-7 and NCL-1 modulate the expression of a major nucleolar protein FIB-1, thereby fine-10 μm (right panels). (C) In Fig 4C (the cguEx18 and ncl-1(e1942); cguEx18 worm pair containing the mutated fib-1 3' UTR) were quantitatively determined, with ratios between the indicated strains being shown in the bar graph. Asterisk signifies the difference: ***P < 0.001; n = 136-156 animals. (D) cguIs1 worms with RNAi targeting ncl-1, nos-2, puf-5, puf-8, or puf-9 were assessed for FIB-1::GFP expression as in (B). Scale bar: 100 μm (left panels) and 10 μm (right panels). (E) Quantitative image analysis for the results shown in (D), showing the relative ratios of average FIB-1::GFP signals between the indicated worm pairs. The bar graph depicts means ± S.E.M.; ***P < 0.001; n = 30-198 animals. (F) Quantitative image analysis for the FIB-1::GFP reporter expression in worm strains derived from cguEx18, showing the relative ratios of average FIB-1::GFP signals between the indicated worm pairs. The bar graph depicts means ± S.E.M.; *P < 0.05; **P < 0.01;***P < 0.001; ns, no significance; n = 30-198 animals. (G) Expression of the FIB-1::GFP reporter in worms with RNAi targeting the indicated genes was monitored by anti-GFP immunoblotting. tuning the size and function of the nucleolus (Fig 6). This circuit of let-7-ncl-1-fib-1 and nucleolus size may represent an adaptive mechanism that couple cellular protein production capacity to the metabolic state of individual cell types. Interestingly, in a recent genome-wide RNAibased screening for molecular networks underlying nucleolus size regulation in Drosophila, both brat and fib were identified [4], substantiating the possibility that these factors constitute a conserved core of regulatory network. Moreover, Vogt et al. has demonstrated that nucleolus maturation during early embryonic development in mice is dependent on the pluripotency factor LIN28 [30], which is known as an essential regulator of let-7 biogenesis [19,20,31]. Intriguingly, Chan and Slack have also shown that ribosomal protein RPS-14 is able to modulate let-7 function [32], which hints at the possibility for a feedback regulation between let-7 and nucleolar dynamics. Our work thus contributes to these findings by reinforcing the relevance of hierarchical organization of post-transcriptional regulators in the fundamental process of nucleogenesis. As FIB-1 expression in C. elegans is also regulated by the die-1 and let363/TOR pathways [33,34], our findings further support the notion that intricate integration of multiple mechanisms underpins nucleolus integrity.
NCL-1 is a member of TRIM/RBCC-NHL protein family, which has been implicated in the regulation of tumor suppression, cell growth, and cell differentiation. In Drosophila larval neuroblasts (stem cell-like precursors), the Brat homologue is distributed to only one daughter cell through asymmetric cell division and acts as an inhibitor of its self-renewal through posttranscriptional suppression of Myc expression. In Brat mutant, both daughter cells grow and lead to the formation of larval brain tumor [35]. Similarly, the mammalian homologue TRIM3 has been reported as a tumor suppressor in human glioblastoma (GBM), a highly malignant human brain tumor, through its suppression on Myc [36]. Our study complements these findings on the NCL-1 homologues and further provides significant insight into understanding how microRNA cooperates with TRIM/RBCC-NHL proteins to suppress tumor formation. A schematic model of the let-7-ncl-1-fib-1 circuit and its regulation of nucleolus size and function. Since let-7 is a heterochronic gene linked to the control of vulva formation in the L4 larva stage, this model depicts a novel let-7-driven regulatory cascade-the let-7-ncl-1-fib-1 pathway-that regulates the nucleolus size and rRNA expression in the vulva cells. In this context, let-7 increases in the L4 larva and targets the 3' UTR of ncl-1 transcript to suppress NCL-1 translation. In other types of cells with low levels of let-7, such as hypodermis for example, NCL-1 may be accumulated and cooperates with two other RNA binding proteins, PUF and NOS, to suppress translation of a nucleolar protein FIB-1 and consequently the size of the nucleolus (see Fig 4B and 4C). However, in the vulva cells in which NCL-1 is down-regulated, a higher abundance of FIB-1 enters the nucleolus to facilitate rRNA processing and likely contributes to enlarged nucleolus exhibited by this particular cell type (see Fig 5F and 5G). Possible FIB-1 action on Pol I activity is not resolved in this study (the question mark in the scheme), although one recent study (Tessarz et al., Nature 505, 564-568, 2014) [47] has shown that FIB-1 impacts Pol I transcription through an epigenetic control. Despite the prevalent requirement for proper maintenance of nucleolus size, our data did not exclude the possibility that the NCL-1-dependent control mechanism may have tissue-and developmental stage-specific relevance. First, while elevated FIB-1::GFP expression was robustly observed in the ncl-1 mutant, the extent to which it was up-regulated was varied between cells/tissues. A strong evidence for this phenotype is shown in Fig 3B, in which we observed variation in nucleolar size changes between hyp 9 and hyp 10 cells. Second, and perhaps more intriguingly, even in the absence of putative PUF binding site, loss of ncl-1 led to a prominently up-regulated GFP reporter expression in the head region of the cguEx18 worms (Fig 4B). This observation of differential regulation thus implies that 1) there is additional cisacting element(s) in the fib-1 3' UTR, through which a yet unknown protein mediates brainspecific expression suppression, and/or 2) NCL-1 may functionally cooperates with other neuronal RNA-binding protein(s) to exert a context-dependent regulation of fib-1. This possibility of a modular organization of NCL-1-based regulatory network, as well as its developmental implications, may be further resolved by genetic screens and/or biochemical characterization of NCL-1-interacting factors.

Worm transformation (microinjection and bombardment)
Germ line transformation by microinjection was performed as described by Mello and Fire [40]. Plasmids at the concentration of 100 ng/μl were injected into young adult N2 worms. An integrated line containing the plasmid of P fib-1 ::fib-1::gfp::3' UTR fib-1 in about a hundred copies (determined by RT-qPCR) was first obtained in the wild-type background (designated as SJL1 cguIs1). A male of cguIs1 was then crossed with ncl-1(e1942) hermaphrodites, and GFP positive worms were selected. This was followed by hermaphrodite selfing to generate a homozygote worms [SJL14 ncl-1(e1942); cguIs1]. The same method was used to generate the other integration lines (see S1 Table), whereas strains of SJL6 to SJL12 (S1 Table) were obtained by the bombardment method [41].

RNAi treatment
The RNAi library was obtained from Julie Ahringer's group [42][43][44]. Bacteria clones producing double-stranded RNA to each target gene were grown in LB broth containing ampicillin and tetracycline for 7 to 8 hrs, and subsequently induced to produce double-stranded RNA by 1 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) for 2 hrs. Concentrated bacteria were then seeded on RNAi plates (NGM agar, 1 mM IPTG, 100 mg/ml ampicillin, and 5 mg/ml tetracycline), onto which synchronized L1-L2 stage worms were placed and cultured for 36 hrs at 25°C. Young adult worms were collected for microscopy, RT-qPCR, and/or Western blot analyses.

Western blot
Protein extracts from embryos or worms at L4 or young adult stage were prepared by sonication and separated on 10% or 15% SDS-PAGE and transferred onto polyvinylidene fluoride (PVDF) membranes. Blocked membranes were then incubated with anti-FIB-1 (1:2,000 dilution, Santa Cruz) or anti-Actin (1:200,000 dilution, Millipore) antibody overnight at 4°C, and subsequently probed with secondary antibody-horseradish peroxidase conjugate (1:5,000 dilution, Sigma). Signals were detected with the ECL Western blot detection system (Thermo Scientific Inc., Waltham, MA).

RT-qPCR
Synchronized worms were collected by washing with M9 buffer and then subjected to sucrose density centrifugation to remove OP50 (E. coli) contamination. Total RNA was isolated from a frozen 1 ml aliquot (100 μl worm pellet dissolved in 1 ml TRIzol) by thawing and vigorous mixing according to the manufacturer's instructions. The genomic DNA was digested by DNase I (Promega). Reverse transcription reactions were performed with iScript Reverse Transcription Supermix for RT-qPCR (Bio-Rad) with 1 μg of RNA. Fifty ng of cDNA was used for each realtime PCR reaction, which was performed with the iCycler IQ real-time PCR detection system (Bio-Rad). For the quantitative detection of ncl-1, fib-1, act-1, gfp and 26S rRNA transcripts, the following primer pairs were respectively used (the act-1 transcript was simultaneously quantified as an internal control): Qncl-

Northern blot analysis
Synchronized late L4 worms grown at 15°C or 25°C as indicated were homogenized by a beadbeating homogenizer (FastPrep-24, MP Biomedicals) and total RNA was isolated by acid guanidinium thiocyanate-phenol-chloroform extraction [45]. Total RNA was subjected to 1.2% agarose-formaldehyde gel electrophoresis (5 μg/lane) and transferred to a Hybond-N+ membrane (GE Healthcare). DNA probes were generated from PCR products amplified from C. elegans genomic DNA and labeled with 32 P-dCTP (Perkin Elmer, PK-BLU513H) by hexamer priming. Primers for generating the ribosomal RNA species and actin probes were performed as described by Voutev et al. [46]. Hybridization was carried out at 55°C in 0.36 M Na 2 HPO 4 , 0.14 M NaH 2 PO 4 , 1 mM EDTA, 10% SDS, 25% formamide and 0.1 mg/ml salmon sperm DNA. Washes were done at 55°C sequentially in 4× SSPE, 4% SDS and 0.1× SSC, 0.1% SDS. Membranes were exposed to Kodak BioMax MS film.

Light microscopy and quantitative image analysis
To observe the FIB-1::GFP expression, embryos or young adult worms of cguIs1 were prestained with WGA 555 (50 μg/ml) (Alexa Fluor 555 conjugate of wheat germ agglutinin, Invitrogen) at room temperature for 30 mins (embryos) or 4 h (worms) and collected by washing 3 times with M9 buffer. They were then mixed with embryos or worms of ncl-1(e1942); cguIs1 in an equal ratio and mounted onto 5% agar pad (worms) or a chamber coverglass (embryo) (Thermo) for image acquisition. Bright field and fluorescence images were captured on an inverted or upright microscopy (Leica DMIRE2 and DM2500) using a 10×/NA 0.3 air immersion objective lens and a cool CCD (CoolSNAP K4, Roper Scientific). In order to distinguish the levels of GFP in the experimental and control embryos or worms under a same fluorescence microscope field, the average fluorescence intensity of different strains in the same images was measured using Metamorph 7.7.10.0 offline (Molecular Devices) and quantitatively determined by using Microsoft Excel software. For visualization of FIB-1::GFP expression and nucleolus size in worms, a upright microscope (Leica DM2500) with high-magnification, differential interference contrast (DIC) and fluorescence channels was used; images (shown in enlarged insets) were captured using a 63×/NA 1.4 oil immersion objective lens and a cool CCD (CoolSNAP K4). Metamorph 7.7.10.0 and Microsoft Excel software were used to measure the nucleolus size.

Deconvolution microscopy
For visualization of GFP signals in the vulva and seam cells, transgenic worms at the L4 stage were paralyzed and mounted onto 5% agar pad for z-series image recording. The DIC and fluorescence signals were collected on a Deltavision deconvolution microscope (PersonalDV, Applied Precision) using a 60×/NA 1.4 oil immersion objective lens and a cool CCD (Cool-SNAP HQ2, Roper Scientific). The Metamorph software version 7.7.10.0 offline was used in image analysis.

Time-lapse images recording
Embryos of cguIs1 or ncl-1(e1942); cguIs1 as described above were plated onto a chamber coverglass for image acquisition. Phase contrast and fluorescence images were captured on an inverted microscope (Leica DMIRE2) using a 25×/NA 0.95 water immersion objective lens and an electron multiplying (EM) CCD (iXon ultra 897, Andor Technology). Images were recorded at 30s intervals and converted to pseudo-color using Metamorph software.

Statistical analysis
Statistical analyses were performed with a two-tailed Student's t-test for independent samples by using GraphPad Prism 5 software. P<0.05 was considered statistically significant.
Supporting Information (WMV) S1 Table. Strains of worms generated in this study.