Genetic and Chemical Controls of Sperm Fate and Spermatocyte Dedifferentiation via PUF-8 and MPK-1 in Caenorhabditis elegans

Using the nematode C. elegans germline as a model system, we previously reported that PUF-8 (a PUF RNA-binding protein) and LIP-1 (a dual-specificity phosphatase) repress sperm fate at 20 °C and the dedifferentiation of spermatocytes into mitotic cells (termed “spermatocyte dedifferentiation”) at 25 °C. Thus, double mutants lacking both PUF-8 and LIP-1 produce excess sperm at 20 °C, and their spermatocytes return to mitotically dividing cells via dedifferentiation at 25 °C, resulting in germline tumors. To gain insight into the molecular competence for spermatocyte dedifferentiation, we compared the germline phenotypes of three mutant strains that produce excess sperm—fem-3(q20gf), puf-8(q725); fem-3(q20gf), and puf-8(q725); lip-1(zh15). Spermatocyte dedifferentiation was not observed in fem-3(q20gf) mutants, but it was more severe in puf-8(q725); lip-1(zh15) than in puf-8(q725); fem-3(q20gf) mutants. These results suggest that MPK-1 (the C. elegans ERK1/2 MAPK ortholog) activation in the absence of PUF-8 is required to promote spermatocyte dedifferentiation. This idea was confirmed using Resveratrol (RSV), a potential activator of MPK-1 and ERK1/2 in C. elegans and human cells, respectively. Notably, spermatocyte dedifferentiation was significantly enhanced by RSV treatment in the absence of PUF-8, and its effect was blocked by mpk-1 RNAi. We, therefore, conclude that PUF-8 and MPK-1 are essential regulators for spermatocyte dedifferentiation and tumorigenesis. Since these regulators are broadly conserved, we suggest that similar regulatory circuitry may control cellular dedifferentiation and tumorigenesis in other organisms, including humans.


Introduction
The nematode C. elegans is a multicellular organism that has become a popular model for biological and basic medical research. It has also been widely used as a model system to explore fundamental questions in multiple aspects of biology, including development, stem cell regulation, cell fate decision, tumorigenesis, and aging [1][2][3][4][5]. C. elegans has two sexes: male and hermaphrodite. In males and hermaphrodites, spermatogenesis begins in the L4 larval stage [6]. Spermatogenesis continues throughout the lifetime of a male, whereas it ceases and switches to oogenesis in the late L4 stage in a hermaphrodite [6] ( Figure 1A). Development of the C. elegans germline progresses by many of the same steps typical of other animal germlines [8]. The C. elegans germline is organized in a simple linear fashion that progresses from germline stem cells (GSCs) at one end to maturing gametes at the other ( Figure 1A). Germ cells progress from GSCs at the distal end through meiotic prophase as they move proximally to become differentiated gametes at the proximal end [6]. C. elegans germline development is tightly regulated by conserved external signaling pathways, including GLP-1/Notch signaling, and intrinsic regulators, including RNA binding proteins and cell cycle regulators [9] ( Figure 1B). The Notch signaling pathway and its core components in C. elegans are highly conserved. C. elegans has two Notch receptors, GLP-1 and LIN-12, which mediate cell-cell interaction during the development [10]. Specifically, GLP-1/Notch signaling in the C. elegans germline is critical for GSC maintenance and continued mitotic division through its direct target genes-sygl-1 and lst-1 [9,11] (Figure 1B). In addition to GLP-1/Notch signal pathways, a battery of RNA regulators, including PUF (Pumilio/FBF) RNA-binding proteins, play critical roles in GSC maintenance, differentiation, and cell fate specification in the C. elegans germline [7] ( Figure 1B). In vertebrates, PUF proteins control various physiological processes such as stem cell proliferation [12,13], tumorigenesis [13], neurogenesis [14,15], germline development [16,17], mesenchymal cell fate decision [12], and mitochondrial dynamics/mitophagy [18] by interacting with the 3 untranslated regions (UTRs) of specific mRNAs to repress the mRNA translation or stability ( Figure 1C).
C. elegans has 11 PUF proteins that recognize a family of related sequence motifs in the target mRNAs ( Figure 1D), yet individual PUF proteins have distinct biological functions [19]. Among them, the PUF-8 (mainly like the Drosophila and human PUFs) protein controls multiple cellular processes, including GSC proliferation, differentiation, dedifferentiation, and sperm-oocyte decision, depending on the genetic context [20] ( Figure 1E). Most puf-8(q725 or ok302) single mutants make both sperm and oocytes, and they are self-fertile at the permissive temperature (20 • C) [21][22][23][24]. However, MPK-1 activation promotes sperm fate, resulting in masculinization of the germline (Mog) phenotype at 20 • C [25,26] and spermatocyte dedifferentiation, resulting in germline tumors at 25 • C in the absence of PUF-8 [21,23] (Figure 1E,F). Dedifferentiation is a cellular process by which cells from partially or terminally differentiated stages revert to a less differentiated stage. This cellular phenomenon has been implicated in regenerative medicine and tumorigenesis [27]. Although this cellular process is observed in vivo in many eukaryotes, its cellular mechanism remains poorly understood.

Worm Maintenance and Strains
C. elegans strains were maintained at 20 • C or 25 • C as previously described [1]. C. elegans strains were provided by Caenorhabditis Genetics Center (CGC) and Dr. Kimble's lab (University of Wisconsin-Madison, Madison, WI, USA) or generated by us using a standard genetic method. Supplemental Table S1 lists strains used in this study.

RNA Interference (RNAi)
RNAi experiments were performed by feeding bacteria expressing double-strand RNAs (dsRNAs) corresponding to the gene of interest [33]. Briefly, synchronized L1 staged worms were plated onto RNAi plates and incubated at 25 • C. Germline phenotypes were determined by staining dissected gonads with specific markers and DAPI. For mpk-1b RNAi, the unique region (exon 1; 1-240 nt) of the mpk-1b gene was amplified by PCR from C. elegans genomic DNA and cloned into the pPD129.36 (L4440) vector containing two convergent T7 polymerase promoters in opposite orientations separated by a multi-cloning site [25,34]. Other RNAi bacteria were from the C. elegans RNAi feeding library (Source Bioscience LifeSciences) and C. elegans ORF-RNAi library (Open Biosystems).

5-Ethynyl-2 -deoxyuridine (EdU) Labeling
To label mitotically cycling cells, worms were incubated with rocking in 0.2 mL M9 buffer (3 g KH 2 PO 4 , 6 g Na 2 HPO 4 , 5 g NaCl, 1 mL 1M MgSO 4 , H 2 O to 1 L) containing 0.1% Tween 20 and 1 mM EdU for 30 min at 20 • C. Gonads were dissected and fixed in 3% paraformaldehyde/0.1 M K 2 HPO 4 (pH 7.2) solution for 20 min, followed by −20 • C methanol fixation for 10 min. Fixed gonads were blocked in 1 × PBST/0.5% BSA solution for 30 min at 20 • C. EdU labeling was performed using the Click-iT EdU Alexa Fluor 488 Imaging Kit (Thermo Fisher Scientific, Waltham, MA, USA, #C10337), according to the manufacturer's instructions. For co-staining with antibodies, EdU-labeled gonads were incubated in the primary antibodies after washing three times and subsequently in the secondary antibodies as described above.

Resveratrol (RSV) Treatment
RSV (Sigma-Aldrich, St. Louis, MO, USA; Cat# R5010) was dissolved in ethanol (EtOH) to stock concentrations of 100 mM. RSV was directly added to the NGM media before pouring the solution into Petri dishes. The worms were transferred to the EtOH-or RSV-containing NGM agar plates. All worms tested were transferred to fresh plates every 2 days, and their germline phenotypes were determined by staining dissected gonads with cell type-specific antibodies or an EdU-labeling kit.

Western Blot Analysis
Cells were lysed as previously described [36]. Proteins were subjected to 10% SDS PAGE. Gels were transferred to the iblot transfer stack (Invitrogen, MA, USA, #IB4010 01) using a transfer apparatus (Invitrogen iBlot 2). Primary antibody incubations were performed in a blocking solution (5% BSA, 1 × TBS, 0.1% Tween20) overnight at 4 • C after blocking for 1 h in the blocking solution. Secondary antibody incubations were performed for 1 h at room temperature in a blocking solution. After washing three times, bands were visualized using Clarity Western ECL substrate (Bio-Rad, Hercules, CA, USA, #1705061) and calibrated by the Chemidoc Imaging System (Bio-Rad, Hercules, CA, USA). Supplemental Table S3 lists antibodies used in this study.

Resveratrol Induces Germline Tumors by Activating MPK-1 in the Absence of PUF-8
Our previous study found that Resveratrol (RSV) maintains MPK-1 activity throughout the lifespan of C. elegans [49] (Figure 5B). To better understand the effect of RSV on ERK1/2, two human cell lines, MDA MB231 and WPMY-1, were treated with the different concentrations (0, 10, 20, 50 µM) of RSV for 24 h after 80% confluence. The expression levels of total ERK1/2 and phospho-ERK1/2 (pERK1/2) proteins were examined by Western blot using anti-ERK1/2 and anti-pERK1/2 antibodies ( Figure 5C). RSV significantly increased the levels of pERK1/2 proteins dose-dependently up to 20 µM ( Figure 5C). However, their levels significantly decreased at 50 µM RSV in both cell lines ( Figure 5C). Similarly, C. elegans pMPK-1 levels were increased up to 3.3-fold by 100 µM RSV treatment, but their increasing levels were decreased up to 2.3-fold by 200 µM RSV treatment [49]. We also found that 50 µM RSV treatment significantly decreased cell viability in a dose-dependent manner (data not shown), as previously reported [50][51][52]. Thus, the reduction in pERK1/2 and pMPK levels in cells and worms exposed to high RSV concentration might be due to an increased cell death. This result indicates that the effects of RSV on the activation of MPK-1 and ERK1/2 are conserved in human cell lines and C. elegans.
Similarly, C. elegans pMPK-1 levels were increased up to 3.3-fold by 100 µ M RSV treatment, but their increasing levels were decreased up to 2.3-fold by 200 µ M RSV treatment [49]. We also found that 50 µ M RSV treatment significantly decreased cell viability in a dose-dependent manner (data not shown), as previously reported [50][51][52]. Thus, the reduction in pERK1/2 and pMPK levels in cells and worms exposed to high RSV concentration might be due to an increased cell death. This result indicates that the effects of RSV on the activation of MPK-1 and ERK1/2 are conserved in human cell lines and C. elegans.  Based on these findings, we tested whether 100 µM RSV could induce the formation of germline tumors via spermatocyte dedifferentiation by activating MPK-1 signaling in puf-8(q725); fem-3(q20gf) mutant germlines. Synchronized L1 staged puf-8(q725); fem-3(q20gf) mutant worms were cultured on NGM agar plates containing 100 µM RSV or 0.1% ethanol (EtOH) control at 25 • C. Their germline phenotypes were determined daily by staining dissected gonads with an EdU-labeling kit and DAPI. Notably, RSV significantly induced the formation of germline tumors via spermatocyte dedifferentiation from day 3 in puf-8(q725); fem-3(q20gf) mutant germlines ( Figure 5D,E). This result suggests that RSV is a potential inducer of spermatocyte dedifferentiation in vivo in the absence of PUF-8. Next, to test whether RSV-induced spermatocyte dedifferentiation relies on MPK-1 activity, we depleted the expression of mpk-1 by RNAi from L1 staged puf-8(q725); fem-3(q20gf) mutants in the presence of 100 µM RSV. While vector RNAi control did not suppress the formation of germline tumors via spermatocyte dedifferentiation, mpk-1 RNAi significantly suppressed the formation of puf-8(q725); fem-3(q20gf) germline tumors even in the presence of 100 µM RSV ( Figure 5F). This result was confirmed by staining dissected gonads with anti-HIM-3 antibodies which recognize meiotic differentiating cells ( Figure 5G). mpk-1 RNAi inhibited the formation of germline tumors via spermatocyte dedifferentiation and induced HIM-3-postive meiotic cells (non-proliferative cells) in the puf-8(q725); fem-3(q20gf) mutant germline ( Figure 5G). Therefore, we suggest that MPK-1 activation could chemically induce the formation of germline tumors via spermatocyte dedifferentiation in the absence of PUF-8 in vivo ( Figure 6).

Discussion
RNA-binding proteins (RBPs) bind to either single-stranded or double-stranded RNA and play a role in the post-transcriptional control of RNAs, such as mRNA stabilization, localization, splicing, polyadenylation, and translation [53]. PUF family RBPs are highly conserved among most eukaryotic organisms [7]. PUF proteins are conserved RBPs that maintain GSCs in worms and flies and have also been implicated in this role in mammals [12,15,17,[54][55][56][57]. PUF proteins bind specifically to PUF binding elements (PBE: UGUAnAUA) within the 3′ untranslated region (3′UTR) of their direct target mRNAs to

Discussion
RNA-binding proteins (RBPs) bind to either single-stranded or double-stranded RNA and play a role in the post-transcriptional control of RNAs, such as mRNA stabilization, localization, splicing, polyadenylation, and translation [53]. PUF family RBPs are highly conserved among most eukaryotic organisms [7]. PUF proteins are conserved RBPs that maintain GSCs in worms and flies and have also been implicated in this role in mammals [12,15,17,[54][55][56][57]. PUF proteins bind specifically to PUF binding elements (PBE: UGUAnAUA) within the 3 untranslated region (3 UTR) of their direct target mRNAs to repress their translation [7,19,20] (Figure 1C). PUF proteins also have diverse roles depending on the organism. For example, in Drosophila melanogaster, Pumilio is required for embryonic development through the regulation of Hunchback (necessary for the establishment of an anterior-posterior gradient) [58] and GSC maintenance [59]. The yeast PUF protein Mpt5, a broad RNA regulator in Saccharomyces cerevisiae, binds to more than 1000 RNA targets [60]. The Mpt5 is required to promote G2/M cell cycle progress [61] and cell wall integrity [62]. Humans have two PUF proteins, PUM1 and PUM2. The PUM1 and PUM2 have high structural similarity and recognize the same RNA binding motif. Despite their similarities, PUM1 is critical for stem cell proliferation, and PUM2 is more important for stem cell differentiation and cell lineage specification [12]. Notably, PUF proteins repress mRNAs encoding MAPK enzymes in worms, flies, yeast, and humans [34,63,64]. However, the role of PUF proteins in limiting dedifferentiation remain poorly understood.
Cellular dedifferentiation counteracts the decline of stem cells during aging but has also been implicated in the formation of tumor-initiating cells [65]. Thus, a comprehensive examination of what causes stem cells to differentiate into desired cell types and how committed cells return to undifferentiated cells is a central question in stem cell biology, regenerative medicine, and tumorigenesis [66]. This cellular dedifferentiation can take many forms depending on the specific organism and tissue type. In zebrafish (Danio rerio), cellular dedifferentiation occurs in a controlled environment where cardiomyocytes partially dedifferentiate to repopulate lost ventricular tissue [67]. In fruit flies (Drosophila melanogaster), it has been shown that differentiating germ cells can revert into functional stem cells both in second instar larval ovaries and in adult fruit flies [68]. Drosophila has also been shown to induce dedifferentiation in spermatogonia cells as there is considerable plasticity due to Jak-STAT signaling [69]. Reduction in stem cell division in Drosophila has been shown due to an accumulation of GSCs with misoriented centrosomes that increases as the flies age [70]. In Mus musculus, dedifferentiated basal-like cells originating from luminal airway cells can function as stem cells in the repopulation of damaged airway epithelia [71]. Dedifferentiation in the mouse model has also been shown in the intestine, where tumorigenesis is initiated due to increased Wnt-activation allowing for polyp formation [72]. In these different examples of dedifferentiation, other signaling pathways are reverting differentiated cells into either stem cell-like or tumor-initiating cells.
During the testing of germline tumors with mpk-1 RNAi, we found that puf-8(q725); lip-1(zh15) mutants fed OP50 had a lower percentage of tumor formation than that fed HT115, an RNase III-deficient E. coli strain used for feeding RNAi in C. elegans [81,82] (see Figure 5D,F; Supplementary Figure S2). The HT115 E. coli provided a greater metabolic energy source due to the recycling of excess nucleotides by the bacteria being RNase IIIdeficient [83]. Due to this increase in energy boosting the formation of germline tumors, we decided to look at the levels of active MPK-1. Notably, the worms fed HT115 E. coli had increased pMPK-1 than those fed OP50 E. coli (unpublished result). These unpublished results suggest that the HT115 diet may increase spermatocyte dedifferentiation via activating MPK-1 signaling. Notably, increasing evidence suggests that metabolic changes alter cell fates by changing multiple signaling pathways. For example, starvation or starvationinduced quiescence maintains GSCs, independent of GLP-1/Notch signaling [84,85]. Furthermore, short-term starvation stress enhances the meiotic activity of germ cells to prevent age-related declines in sperm production [86]. Similarly, the dedifferentiation of primary hepatocytes is accompanied by the reorganization of lipid metabolism [87]. Therefore, non-genetic factors may play a vital role in cell fate reprogramming and tumorigenesis. Our findings reveal fundamental mechanisms of the differentiation/dedifferentiation decision in vivo and may provide a future platform for identifying therapeutic targets for dedifferentiation-mediated tumorigenesis.