LSY-2 is essential for maintaining the germ-soma distinction in C. elegans

The mechanisms that specify and maintain the characteristics of germ cells during animal development are poorly understood. In this study, we demonstrated that loss of function of the zinc-finger gene lsy-2 results in various somatic cells adopting germ cells characteristics, including expression of germline-specific P granules, enhanced RNAi activity and transgene silencing. The soma to germ transformation in lsy-2 mutants requires the activities of multiple chromatin remodeling complexes, including the MES-4 complex and the ISW-1 complex. The distinct germline-specific features in somatic cells and the gene expression profile indicate that LSY-2 acts in the Mec complex in this process. Our study demonstrated that lsy-2 functions in the maintenance of the soma-germ distinction.


INTRODUCTION
Germ cells are specified at an early embryonic stage in most sexually reproducing animals. In C. elegans, germline lineage is generated through each of four unequal divisions in which a blastomere generates one somatic founder cell and one germline blastomere (Strome, 2005). The primordial germ cell P4 then divides equally at the embryonic stage, giving rise to two germ precursor cells Z2 and Z3, which remain quiescent during embryogenesis and proliferate throughout larval development to produce germ cells (Strome, 2005). Germ P granules, which are aggregates of RNA and proteins, are maternally contributed and segregated exclusively into the germline lineages (Strome, 2005). P granules are synthesized in all of the descendants of P4 with the exception of mature sperm and are associated with the outer surface of the nuclear envelope (Hird et al., 1996). P granules contain constitutive components, which associate with P granules during all developmental stages, and also transient components, which interact with P granules in germline blastomeres and disappear in Z2 and Z3 (Strome, 2005). Constitutive components include the RGG box containing RNA binding proteins PGL-1 and PGL-3 and the RNA helicases GLH-1 to GLH-4 (Gruidl et al., 1996;Kawasaki et al., 1998;Kuznicki et al., 2000). During C. elegans embryogenesis, PGL-1 and PGL-3 are degraded by autophagy in somatic cells (Zhang et al., 2009). In autophagy mutants, PGL-1 and PGL-3 colocalize and accumulate into aggregates in somatic cells (Zhang et al., 2009).
Establishment of somatic and germ precursor cells during early embryogenesis in C. elegans requires maternally distributed proteins, including PIE-1, MEX-1, MEX-3, and POS-1, which are also associated with P granules during the first several cell divisions and disappear in the daughters of the germ precursor P4 cells Guedes and Priess, 1997;Tabara et al., 1999). PIE-1, MEX-1, and POS-1 prevent the germline blastomere from adopting somatic fates, while MEX-3 prevents certain somatic blastomeres from adopting a germ blastomere fate . Maintenance of germ cell fate during animal development involves multiple antagonistic chromatin remodeling complexes (Cui et al., 2006). Mutations in the SynMuv B genes, including the components of the C. elegans lin-35 Rb pathway and the MEP-1/LET-418/HAD-1 NURD complex, lead to the ectopic expression of germline traits by somatic cells (Unhavaithaya et al., 2002;Wang et al., 2005;Kunert et al., 2009). SynMuv B genes function redundantly with SynMuv A genes in preventing extra cells from adopting the vulval fate (Fay and Yochem, 2007). Animals defective in both SynMuv A and SynMuv B genes display a multi-vulva (Muv) phenotype. According to distinct germline-specific features in somatic cells and the profile of misexpressed small RNA and P granule genes, SynMuv B genes have been classified into three classes, representing three functional complexes: a LIN-35/RB-containing core complex (DRM), a heterochromatin complex and a SUMO-recruited Mec complex (Wu et al., 2012;Tables 1, 2). The soma to germ transformation in these mutants is antagonized by the ISW-1 complex and the MES-4 complex (Cui et al., 2006;Stielow et al., 2008), which appear to modulate the active chromatin structures. How these complexes are targeted to germlinespecific genes and faithfully maintain the repressive state of these genes in somatic cells is unknown.
lsy-2, encoding a C2H2 zinc-finger transcription factor, is involved in specifying the left-right asymmetry of the ASE neurons (Johnston and Hobert, 2005). In this study, we found that loss of function of lsy-2 causes ectopic expression of germ cell characteristics in various somatic cells, including the perinuclear localization of P granules, enhanced RNAi and transgene silencing. The phenotype resembles that in mutants for SUMO-recruited Mec complex factors MEP-1 and LET-418 (Stielow et al., 2008;Wu et al., 2012). The Mec complex also functions in specification of the ASE fate. Our study demonstrates that lsy-2 genetically interacts with the MEP-1/LET-418 complex and is involved in maintaining soma-germ distinction and ASE cell fate specification.

RESULTS
Loss of function of lsy-2 leads to transformation of somatic cell to germ cell fate To study how somatic-and germ cell-specific fates are specified, we screened a library of bacterial clones expressing dsRNAs designed to individually inactivate 16,749 genes (targeting about 87% of the predicted genes). We looked for gene inactivations that cause ectopic expression of the P granule-specific reporter, gfp::pgl-1, in somatic cells of larval animals. In addition to components of the lin-35 Rb pathway, including lin-35, lin-53, hpl-2, lin-9, dpl-1 and lin-52, and the Mec complex component mep-1, we identified that inactivation of the zinc-finger transcription factor lsy-2 and components of the sumoylation pathway, including smo-1, ubc-9, and uba-2, caused ectopic accumulation of GFP::PGL-1 granules in somatic cells.
We further examined the temporal expression pattern of GFP::PGL-1 in lsy-2 somatic cells. We found that ectopic GFP::PGL-1 or endogenous PGL-1 granules in lsy-2 mutants were not formed until the early L1 larval stage, when the germ precursor cells Z2 and Z3 start to proliferate ( Fig. 1J and 1K). Therefore, lsy-2 is not involved in de novo establishment of the soma-germ distinction in C. elegans, but appears to be required for its maintenance.

lsy-2 mutants show enhanced RNAi interference and display transgene silencing
We next determined whether lsy-2 mutants exhibit other characteristics associated with the germline. Germ cells exhibit elevated RNAi efficiency (Sijen and Plasterk, 2003;Robert et al., 2005). We determined whether mutations in lsy-2 enhanced the RNAi efficiency in somatic cells. Feeding of bacterial clones of his-44 and cel-1 had little effect on wildtype animals, but led to 95% (n = 106) and 33% (n = 124) larval arrest in lsy-2(ot90) mutants, respectively, indicating that the RNAi efficiency is enhanced in lsy-2 mutants ( Fig. 2A and 2B). Germ cells are more protected than somatic cells from foreign genetic elements, such as multicopy transgenes (Sijen and Plasterk, 2003;Robert et al., 2005). Therefore, transgenes introduced by microinjection into the germline are quickly silenced, a process mechanistically related to RNAi (Sijen and Plasterk, 2003;Robert et al., 2005). SCM:: GFP is strongly expressed in seam cells. We found that expression of SCM::GFP was significantly reduced in lsy-2 mutants ( Fig. 2C-E). The number of seam cells expressing SCM::GFP was decreased in lsy-2 mutants ( Fig. 2D and 2E). The SCM::GFP fluorescence intensity (SCM::GFP expressed in all seam cells of one worm) was also measured and we found that the intensity in wild-type worms was about twenty times higher than that in lsy-2 mutants (Fig. 2C). Simultaneous inactivation of mut-2 or mut-7, essential components in the RNAi pathway, partially restored the expression of SCM::GFP in lsy-2 mutants (Fig. 2C, 2F, and 2G). Therefore, somatic cells in lsy-2 mutants display characteristics of germ cells.

lsy-2 functions similar to Mec class genes in repressing expression of germline genes in somatic cells
Based on phenotypic differences and the different profile of misexpressed RNAi factors and P-granule components in somatic cells, SynMuv B genes can be divided into three distinct classes: a DRM core complex, a SUMO-recruited Mec complex, and a SynMuv B heterochromatin complex (Wu et al., 2012; Table 1). The target genes upregulated in different SynMuv B mutants have been categorized into germline-enriched common targets, ubiquitous common targets, germline-enriched DRM targets, and germline-enriched SynMuv B heterochromatin and Mec targets. We showed above that lsy-2 mutants exhibited transgene silencing ( Fig. 2C-G). The PGL-1 granules that were somatically misexpressed in intestine and hypodermal cells were small and densely distributed around the nuclei (Fig. 1E, 1F, and 1K). Mutations in lsy-2 also caused ectopic expression of LAG-2::GFP in the intestine, a phenotype associated with loss of function of Mec class genes ( Fig. 2H and 2I) (Poulin et al., 2005;Cui et al., 2006). Thus, lsy-2 mutants show similar phenotypes to those in Mec complex mutants.
We next tested lsy-2 mutants for the expression levels of the 13 genes that are differentially expressed in distinct SynMuv B classes (Wu et al., 2012). To avoid any possible contamination of somatic cells by the germline, we used glp-4(bn2) temperature sensitive mutants, which have no germline when cultured at 25°C. According to the real-time PCR results, germline-enriched common targets, ubiquitous common targets and, more significantly, germline-enriched SynMuv B heterochromatin and Mec-specific targets were dramatically upregulated upon inactivation of lsy-2, while germline-enriched DRM-specific targets showed little change ( Fig. 3A-D; Table 2), indicating that lsy-2 is related to either SynMuv B heterochromatin or Mec class genes. Taken together, lsy-2 acts similarly to Mec class genes in maintaining the germ-soma distinction.

Inactivation of lsy-2 in SynMuv mutants causes no SynMuv phenotype
The role of lsy-2 in repressing the expression of germline cell fates in somatic cells resembles that of SynMuv B genes. SynMuv B genes function redundantly with SynMuv A genes in repressing the vulva cell fate. Animals defective in both Role of lsy-2 in germ-soma specification
lsy-2 synergistically interacts with the LIN-35/Rbcontaining DRM complex in specifying somatic cell fate We next determined the relationship between lsy-2 and other SynMuv B genes in repressing the expression of germline cell fates. Neither lsy-2 nor lin-35 single mutations caused a larval arrest phenotype, while 100% of lin-35; lsy-2 double mutants (n = 93) arrested at the L1 stage. The synthetic lethal phenotype was also observed between lsy-2 and mutations in other DRM class genes, including lin-15B, dpl-1, lin-53, lin-37, and lin-9 (data not shown). This suggests that the DRM complex and lsy-2 may act in concert to regulate the expression of one or more critical targets during development. We next examined P granule formation in lin-35; lsy-2 double mutants. In lsy-2 and lin-35 single mutants, P granules were not ectopically expressed until the L1 larval stage (Figs. 1J, 1K, and 4A). In contrast, the onset of ectopic expression of P granules in lin-35; lsy-2 double mutants occurred much earlier. 20% (n = 15) and 74% (n = 19) of double mutant embryos contained multiple extra cells expressing perinuclear P granules at the 2-fold and 3-fold stage, respectively (Fig. 4B). Moreover, the number of somatic cells expressing P granules was dramatically increased in lin-35; lsy-2 mutant larvae ( Fig. 4C-E). We also examined P granule formation in lin-15B mutants after lsy-2 RNAi injection. The number of somatic cells expressing P granules was dramatically increased in lin-15B larval mutants after lsy-2 RNAi injection (Fig. 4F-H). Thus, lsy-2 and the LIN-35/Rb-containing DRM complex function in parallel to repress the expression of germ cell traits in somatic cells.

Formation of P granules in lsy-2 somatic cells requires the activities of multiple chromatin remodeling complexes
Formation of perinuclear P granules in somatic cells in the SynMuv B mutants requires the activity of chromatin remodeling complexes, including the MES-4 complex and the ISWI complex (Unhavaithaya et al., 2002;Cui et al., 2006). We investigated the role of several chromatin remodeling complexes in the ectopic expression of P granules in lsy-2 mutants. We found that inactivation of the NuA4 complex, the ISW1 complex, and the COMPASS complex dramatically decreased the formation of P granules in somatic cells in lsy-2(ot64), while mutations in components of the SWR1 complex and the SWI/SNF complex had no obvious effect ( Fig. 5A-C and Table 3). Thus, various chromatin-remodeling complexes can regulate the expression of different P-granule components or function non-redundantly in regulation of common P-granule components.

Mutations in lsy-2 and Mec class genes result in a defect in ASE asymmetry specification
LSY-2 is involved in specifying the asymmetry of the ASE neurons (Johnston and Hobert, 2005). Two taste receptor neurons, ASE left (ASEL) and ASE right (ASER) are morphologically bilaterally symmetric, but they display a left/right asymmetric function and express a distinct set of chemosensory receptors (Sarin et al., 2007;Ortiz et   ). In lsy-2 mutants, the ASEL neuron expresses the ASER-specific terminal fate markers and adopts the ASER cell fate (Fig. 6A and 6B). We showed above that LSY-2 genetically acts in the Mec complex to maintain the somagerm distinction. We determined whether inactivation of Mec complex components also results in ASE asymmetry defects. As neurons are refractory to RNAi, we used RNAisensitive rrf-3 animals. In rrf-3 animals, the ASER marker GCY-5::GFP is exclusively expressed in ASER (Fig. 6A). After RNAi inactivation of mep-1 and let-418, 25% (n = 32) and 12% (n = 40) animals showed ectopic expression of GCY-5::GFP in ASEL, respectively ( Fig. 6C and 6D). Inactivation of the DRM class genes lin-35 and lin-15B and the heterochromatin class genes hpl-2 and lin-65 failed to cause ectopic expression of GCY-5::GFP (data not shown). Misexpression of GCY-5::GFP in ASEL was also reported in smo-1 mutants (Poole et al., 2011). Therefore, Mec complex mutants also exhibit the ASE asymmetry specification defect, further supporting the idea that lsy-2 is genetically related to Mec class genes.

DISCUSSION
In this study, we demonstrated that loss of function of lsy-2 causes somatic cells to display some germline-specific features, including ectopic expression of P-granule components, enhanced RNAi efficiency, and transgene silencing. Consistent with these observations, mutations in lsy-2 results in the ectopic expression of a set of germ cell-specific genes in somatic cells (Fig. 6E). lsy-2 has previously been shown to be involved in the specification of the left-right asymmetry of ASE neurons by regulating the cell-type specific expression of lsy-6 miRNA (Johnston and Hobert, 2005). Loss of function of lsy-6 and other components involved in specification of ASEL fate, including lin-49, lim-6, and ceh-36, does not lead to the ectopic expression of P granules in somatic cells, indicating that lsy-2 functions in separate pathways in determining left-right asymmetry and soma-germ distinction.
How does lsy-2 function in these processes? lsy-2 encodes a zinc-finger protein that is distantly related to SP1/ KLF family transcription factors (Kaczynski et al., 2003;Johnston and Hobert, 2005). The repressed state of germ cell-specific genes is maintained over cell divisions in somatic cells during C. elegans development. During late embryogenesis and larval stages, this repressive state requires multiple functional antagonistic chromatin remodeling complexes. The LIN-35/Rb complex and the Mec complex are involved in maintaining the repressed state of these genes. The NuA4 complex, the COMPASS complex, and the ISW1 complex appear to promote the expression of germ cell fate by somatic cells in the absence of the LIN-35/ RB complex or the Mec complex. These complexes have been demonstrated to modify the chromatin structure that is associated with transcriptional activation. For example, the NuA4 complex contains histone acetylation activity (Cai et al., 2003;Doyon et al., 2004). Therefore, LSY-2 is involved in maintaining the repressive state of germ cell-specific genes in somatic cells. Three lines of evidence support the hypothesis that LSY-2 is likely to act in the same pathway as the Mec complex. First, loss of function of lsy-2 results in formation of small, densely clustered P granules, and transgene silencing in somatic cells, phenotypes also observed in mutants of SUMO-recruited Mec components (Wu et al., 2012). Second, the expression profile of misexpressed P-granule components and RNAi factors in lsy-2 mutants resembles that caused by Mec complex mutants (Wu et al., 2012). Finally, as in lsy-2 mutants, mutations in SUMO-recruited Mec components cause the ASE asymmetry specification defect.
lin-35 has been shown to function redundantly with other factors in specifying many developmental processes. For example, lin-35 and ubc-18, which encodes an E2 ubiquitinconjugation enzyme, function redundantly in controlling pharyngeal morphology (Fay et al., 2003;Fay et al., 2004). lin-35 cooperates with the SWI/SNF complex (Cui et al., 2004), including psa-1/swi3 and xnp-1/atr-x, and gon-14 (encoding a protein with similarity to LIN-15B) to control larval development (Cardoso et al., 2005;Chesney et al., 2006). Here we showed that lin-35 also functions redundantly with lsy-2 in promoting larval development and in maintaining the soma/germ distinction. lin-35; lsy-2 double mutants arrest at the early larval stage. The underlying causes of the larval arrest remain to be determined. The partially overlapping function between lsy-2 and lin-35 could be explained by the concerted activity of distinct complex on the same target genes or by additive effects due to misregulation of distinct genes.

RNAi screen
The RNAi feeding library was purchased from Geneservice. The library contains bacterial clones expressing dsRNA designed to individually inactivate 16,749 genes (targeting about 87% of the predicted genes). Synchronized L1 gfp::pgl-1; rrf-3 animals were fed on RNAi bacterial clones and the F1 progeny or arrested larvae were examined for reporter expression.

Indirect immunoflorescence
The monoclonal antibodies OIC1D4 and K76 for PGL-1 were obtained from the Developmental Studies Hybridoma Bank at the University of Iowa. Fragments of PGL-1 (95-551) and  were cloned into the pET-28a vector, expressed as His-tagged fusion proteins in E. coli BL21 and purified for use as an immunogen in rabbits (for PGL-1) or rat (for GLH-1).  For indirect immunofluorescence, the permeabilization of embryos and larvae was performed by freeze-cracking methods and was performed as previously described (Zhang et al., 2009).

RNAi feeding assay for enhanced RNAi
Gravid N2 and lsy-2(ot90) animals were grown with RNAi bacterial clones of his-44, cel-1, and vector L4440. After three hours of egglaying, adults were taken off the plates and embryos were allowed to hatch and grow at 20°C. Animals arrested at the L3 stage was scored when worms of the same genotype fed on vector RNAi grew into adults.

Real-time PCR assay
Synchronized L1 glp-4 mutant animals were fed on RNAi bacterial clones at 16°C until they reached adulthood. These adults were then transferred to a new plate with the same RNAi bacterial clone and allowed to lay eggs for three hours. Adults were then taken off the plate and embryos were allowed to hatch and grow at 25°C, L4 larvae were then collected for real-time PCR.  lin-59 N >40 *The percentage of lsy-2 animals in which the formation of ectopic P granules was suppressed by inactivation of each gene is indicated in parenthesis. N: no suppression was observed; Y: suppression.
Measure of total SCM::GFP fluorescence intensity Worms expressing the SCM::GFP reporter were photographed. The SCM::GFP fluorescence intensity of each seam cell in the photograph was measured by Zen 2011 and then added together to get the total SCM::GFP fluorescence intensity of the worm. Ten worms of each genotype were measured to calculate the average total SCM::GFP fluorescence intensity.