Progression of Meiosis Is Coordinated by the Level and Location of MAPK Activation Via OGR-2 in Caenorhabditis elegans

Abstract

During meiosis, a series of evolutionarily conserved events allow for reductional chromosome division, which is required for sexual reproduction. Although individual meiotic processes have been extensively studied, we currently know far less about how meiosis is regulated and coordinated. In the Caenorhabditis elegans gonad, mitogen-activated protein kinase (MAPK) signaling drives oogenesis while undergoing spatial activation and deactivation waves. However, it is currently unclear how MAPK activation is governed and how it facilitates the progression of oogenesis. Here, we show that the oocyte and germline-related 2 (ogr-2) gene affects proper progression of oogenesis. Complete deletion of ogr-2 results in delayed meiotic entry and late spatial onset of double-strand break repair. Elevated levels of apoptosis are observed in this mutant, independent of the meiotic canonical checkpoints; however, they are dependent on the MAPK terminal member MPK-1/ERK. MPK-1 activation is elevated in diplotene in ogr-2 mutants and its aberrant spatial activation correlates with stages where meiotic progression defects are evident. Deletion of ogr-2 significantly reduces the expression of lip-1, a phosphatase reported to repress MPK-1, which is consistent with OGR-2 localization at chromatin in germ cells. We suggest that OGR-2 modulates the expression of lip-1 to promote the timely progression of meiosis through MPK-1 spatial deactivation.

MEIOSIS is a specialized, coordinated program of cell division where one cycle of chromosome duplication is followed by two consecutive rounds of segregation. This process reduces the chromosome number by half, thereby generating haploid gametes from diploid cells. This reduction is achieved through unique and highly coordinated chromosomal interactions where homologous chromosomes pair, align, and form physical connections prior to the first division [reviewed in Jasin and Rothstein (2013), Lui and Colaiacovo (2013), Mézard et al. (2015), and Gray and Cohen (2016)]. A challenging question in the study of meiosis is: how are these intricate interactions controlled and phased?

The nematode Caenorhabditis elegans is an excellent model organism that is especially suited for investigating how meiosis is governed and orchestrated. In the hermaphrodite gonad arm, nuclei are arranged in a well-defined spatial temporal order from the stem cell niche to the fertilized embryo, enabling cytological observations of all meiotic stages at the same time and in the same organ. The distal region of the C. elegans gonad, termed the mitotic zone or the proliferative zone, contains proliferating germ cell nuclei (Kimble and White 1981). This proliferation is essential for maintaining the progenitor pool that repeatedly produces meiocytes. As cells move proximally, they enter meiosis in the leptotene/zygotene (LZ or transition) zone, where recognition and pairing of homologous chromosomes are initiated. At this time, double-strand break (DSB) formation commences, catalyzed by SPO-11, a highly conserved topoisomerase-like endonuclease (Keeney et al. 1997; Dernburg et al. 1998), and repair begins via homologous recombination (Dernburg et al. 1998). As nuclei enter the pachytene stage, a tripartite proteinaceous scaffold, known as the synaptonemal complex (SC), holds pairs of homologous chromosomes together (MacQueen et al. 2002; Schild-Prüfert et al. 2011). Throughout this stage, most meiotic crossover (CO) events mature within the context of fully synapsed chromosomes. During diplotene, the SC starts to disassemble and homologs remain attached mostly by a single chiasma, the cytologically detectable connection, resulting from a crossing-over event that occurred in each of the homologous chromosomes. At the diakinesis stage, six discrete bivalents (pairs of homologs) can be detected in mature oocytes. Timely progression of each stage during meiosis is a prerequisite for the formation of functional euploid gametes [reviewed in Lui and Colaiacovo (2013) and Hillers et al. (2017)]. However, our current understanding of the regulatory mechanisms that dictate meiotic progression is highly limited and further research is required.

Mitogen-activated protein kinase (MAPK) signaling is a master regulator of key meiotic events in C. elegans (Gumienny et al. 1999; Yin et al. 2016). This pathway includes the signaling molecules LET-60 (Ras), LIN-45 (Raf), and MEK-2 (Mek), which relay extracellular signals that result in phosphorylation and activation of its terminal member MPK-1 (Erk) (Chen et al. 2001). Active MPK-1 phosphorylates substrates that control and coordinate multiple essential aspects of C. elegans germline development [reviewed in Arur (2017)]. MPK-1 activation is temporally regulated; it is activated in pachytene, where its signal regulates pachytene progression and exit (Lee et al. 2007b). Indeed, mutations in any of the core genes in the MAPK signaling pathway—namely, let-60, lin-45, mek-2, or mpk-1—result in a pachytene-arrest phenotype, i.e., the failure of germ cells to successfully progress from pachytene into diplotene (Church et al. 1995; Gumienny et al. 1999). After pachytene exit, MPK-1 is rapidly inactivated. This inactivation was previously attributed to LIP-1, a nematode homolog of the vertebrate’s MAPK phosphatases (Hajnal and Berset 2002; Rutkowski et al. 2011). As the oocytes approach the spermatheca, the secreted sperm signal, MSP, reactivates MPK-1, which consequently induces oocyte maturation and allows meiosis to progress (Miller et al. 2001; Kosinski et al. 2005). When either synapsis or DSB repair (DSBR) is not completed, owing to various environmental or genetic causes, most of the affected meiocytes are eliminated by germ cell apoptosis (Gartner et al. 2000, 2008). However, this quality control mechanism is prevented if MAPK signaling is blocked and, conversely, it occurs at excessive levels in null lip-1 mutants, where MPK-1 activity was suggested to be increased (Hajnal and Berset 2002; Lee et al. 2006; Rutkowski et al. 2011). MPK-1 also has a nonessential function in promoting the proliferative germ cell fate. Null mpk-1 mutants have a small germline, which can be explained by reduced germline stem cell proliferation (Lee et al. 2007a,b; Narbonne et al. 2017). MPK-1 controls and coordinates other essential aspects of cell biology during oogenesis, including cellular organization, SC disassembly, oocyte organization, and differentiation, as well as oocyte maturation and ovulation (Lee et al. 2007b; Nadarajan et al. 2016). Therefore, MAPK signaling acts at nearly all stages of oogenesis; it has been suggested to be a major contributor to meiotic coordination (Arur 2017). Even though the different roles of MPK-1 in meiosis have been extensively studied, our knowledge of the underlying mechanisms controlling its spatial activation, allowing meiotic progression, remains incomplete.

Here, we describe the meiotic roles of the previously uncharacterized SPK [SET (Su(var)3-9, Enhancer-of-zeste and Trithorax) and PHD (plant homeodomain) domain-containing proteins, and protein kinases] domain-containing OGR-2 protein. Loss of ogr-2 affects several processes, ranging from entering into meiosis to mature oocytes’ fates, as well as the correct formation of bivalent structures. Deletion of ogr-2 leads to high levels of germ cell apoptosis through a noncanonical MPK-1-dependent pathway. In agreement with these results, we found that ogr-2 negatively influences the spatial activation of MPK-1 in the gonad. Importantly, we show that ogr-2 and lip-1 mutants exhibit similar, but nonadditive, defects in germline development and in the regions of MPK-1 activation, indicating that they act through the same genetic pathway. OGR-2 is localized to the chromatin of germ cells in foci and patches, which can explain the ogr-2 requirement for normal levels of lip-1 mRNA. These findings suggest that ogr-2 may act via lip-1 to control the level and location of MAPK signaling activation, thereby coordinating the timely progression of key meiotic processes.

Materials and Methods

Strains and alleles

All strains were cultured under standard conditions at 20° unless specified otherwise (Brenner 1974). The N2 Bristol strain was utilized as the wild-type background. Worms were grown on NGM plates with Escherichia coli OP50 (Brenner 1974). Unless otherwise stated, all experiments were conducted using adult hermaphrodites 20–24-hr post-L4 stage. The following mutations and chromosome rearrangements were used: LGI: hus-1(op241), cep-1(gk138), LGII: meIs8 [pie-1p::GFP::cosa-1 + unc-119(+)] II, LGIII: mpk-1(ga111), LGIV: lip-1(zh15), and LGV: syp-1(me17) V/nT1 [unc-?(n754) let-? qIs50] (IV;V). The following strains have been previously described: UV7: unc-119(ed3) III; jfIs2[piepromoter::GFP::zhp-3] (Bhalla et al. 2008), CV64: rjEx02[-Ppie-1::GFP::lab-1::HA; unc-119(+)] (de Carv alho et al. 2008), and RW10226: stIs10226[his-72p::HIS-24::mCherry::let-858 39 UTR + unc-119(+)] (Murray et al. 2012).

The YBT5: ogr-2(huj1) II, YBT38: ogr-2(huj18) II, and YBT44: ogr-2(huj20[ogr-2::3XFLAG]) II strains were generated by clustered regularly interspaced short palindromic repeat (CRISPR)-Cas9 engineering (see below). Homozygous ogr-2(huj1) worms were outcrossed five times with the N2 wild-type strain. Unless otherwise stated, all experiments and downstream crosses reported here were carried out using the outcrossed strain.

Generation of ogr-2-engineered strains by CRISPR-Cas9 genome editing

To generate the ogr-2(huj1) deletion, we used the procedure described in Friedland et al. (2013), but with several modifications. We used the following plasmids: pU6::unc-119 sgRNA (single-guide RNA) (Friedland et al. 2013) and peft-3::cas9-SV40_NLS::tbb-2 3UTR (Friedland et al. 2013), together with the co-injection markers pCFJ90 (pmyo-2::mCherry::unc-54utr) (plasmid # 19327; Addgene) and pCFJ104 (pmyo-3::mCherry::unc-54) (plasmid # 19328; Addgene), which were a kind gift from Erik Jorgensen. We generated four new sgRNA plasmids by site-directed mutagenesis PCR (Weiner et al. 1994) with the vector pU6::unc-119 sgRNA to replace unc-119 with the sgRNA of interest [see below and Friedland et al. (2013)]. We designed ogr-2 sgRNAs to target two DNA sites before the ATG codon of ogr-2 ORF and two sites after the stop codon. The sequences of the sgRNAs were as follows:

• ogr-2-sgRNA-1: 5′-ACC GCA AAC TCA CGA GAT GC-3′.

• ogr-2-sgRNA-2: 5′-AAA ATA TAT TGT TTT CTT TC-3′.

• ogr-2-sgRNA-3: 5′-ACG ATA CTT GAC AGC ACT GG-3′.

• ogr-2-sgRNA-4: 5′-GAA ATT TTA AAA AAC TCA AG-3′.

Transgenic worms were created via microinjection of a mixture of DNA plasmids into the syncytial gonad of young adult N2 worms as described previously (Friedland et al. 2013; Tzur et al. 2013). The injection mix contained the following plasmids: four ogr-2 sgRNAs at 50 ng/μl each, peft-3::cas9-SV40_NLS::tbb-2 3UTR at 100 ng/μl, pCFJ90 at 5 ng/μl, and pCF104 at 100 ng/μl. Transformants were selected by isolating marker-positive (mCherry+) worms.

A second ogr-2 deletion strain, YBT38, was generated by CRISPR-Cas9/crRNA-tracrRNA ribonucleprotein (RNP) injection using the protocol described in Paix et al. (2015). Two ogr-2-gRNAs (analogous to ogr-2-sgRNA-2 and 3) were used.

We tagged the ogr-2 gene by inserting 3XFLAG before the stop codon using CRISPR-Cas9/RNP complexes containing an ogr-2-FLAG-sgRNA and an ssODN repair template that included a 66-nt insert coding for a 3XFLAG tag: Ogr-2-flag-sgRNA: 5′-ATATCTTCCGACATGTATTA-3′ and 3XFLAG ssODN sequence: 5′-GACTACAAAGACCATGACGGTGATTATAAAGATCATGATATCGATTACAAGGATGACGATGACAAG-3′.

Fecundity assays

To determine the fecundity of the ogr-2 mutants, at least 19 individual L4 worms were placed on seeded NGM plates, transferred to new plates every 24 hr, and their embryos and hatched progeny were counted during a 3-day period.

To determinie the mpk-1(ga111) effects on brood size, mutant lines were maintained at 15° before initiating the assays, being isolated at the L4 stage, and were maintained at 25° throughout the analyses.

Cytological analysis and immunostaining

DAPI and immunostaining of dissected gonads were carried out as described in Colaiácovo et al. (2003) and Saito et al. (2009). Worms were permeabilized on Superfrost+ slides for 2 min with methanol at −20° and fixed for 30 min in 4% paraformaldehyde/phosphate‐buffered saline (PBS). Staining with DAPI was carried out for 10 min at 500 ng/ml, followed by destaining in PBS containing 0.1% Tween 20 (PBST); then, the slides were mounted with Vectashield antifading medium (Vector Laboratories, Burlingame, CA). Primary antibodies were used at the following dilutions: rabbit α-LAB-1 (1:200) (de Carvalho et al. 2008), rabbit α-RAD-51 (1:10,000, SDIX), mouse α-MAPK-YT (1:500, M8159; Sigma [Sigma Chemical], St. Louis, MO), rabbit α-SYP-2 (1:200, a kind gift from S. Smolikove), rabbit α-pH3 (D5692, 1:1000; Sigma), mouse anti-FLAG M2 (F1804, 1:500; Sigma), rabbit anti-HIM-8 (1:200; Novus Biological), and guinea pig α-HTP-3 (1:200) (Goodyer et al. 2008). The secondary antibodies used were Cy2-goat anti-rabbit, Cy3-goat anti-rabbit, Cy3-goat anti-mouse, and Cy2-donkey anti-guinea pig (all used at 1:500 dilution; Jackson ImmunoResearch Laboratories, West Grove, PA).

Imaging and microscopy

For GFP::ZHP3, α-pH3, mouse anti-FLAG M2, double-phosphorylated MPK-1 (dpMPK-1) immunostaining, and DIC visualization, images were acquired using the Olympus IX83 fluorescence microscope system (Olympus). Optical Z-sections were collected at 0.30/0.60-µm increments with a Hamamatsu Orca Flash 4.0 v3 and CellSens Dimension imaging software (Olympus). Pictures were deconvolved using AutoQuant X3 (Media Cybernetics).

For bivalent analysis and LAB-1, SYP-2, HTP-3, HIM-8, and RAD-51 immunostaining, Z-stack three-dimensional (3D) images were taken at 0.3-µm increments using an Olympus FV1000 Inverted Confocal IX81 microscope and FV10-ASW 3.1 Software (Olympus).

Immunofluorescence images of the RNA interference (RNAi) experiment were collected at 0.3-µm increments with an IX-70 microscope (Olympus) and a cooled CCD camera (model CH350; Roper Scientific) controlled by the DeltaVision system (Applied Precision). Images were subjected to deconvolution analysis using the SoftWorx 3.0 program (Applied Precision) as in Nabeshima et al. (2005).

For germ cell apoptosis, worms were transferred onto a drop of M9 on 1.5% agarose pads on slides, and assayed using a Hamamatsu Orca II ER/AG camera mounted on a Zeiss ([Carl Zeiss], Thornwood, NY) Axioplan II microscope equipped for fluorescence and DIC.

Small oocyte analyses were performed on 3-day post-L4 adult hermaphrodite whole mounts. Worms were mounted in M9 and extracted gonads were subjected to 3D DIC microscopy (Sulston and Horvitz 1977). Oocytes images were captured and midoocytes plane areas were measured with ImageJ software.

Prophase I-stage nuclei quantification

The number of nuclei at each meiotic stage, from the distal tip to the end of pachytene, was quantified manually on DAPI-stained gonads. The mitotic region was defined from the distal tip of the gonad to the first row of nuclei with clustered chromosome (crescent) morphology. The LZ zone was defined as extending from the first to the last rows, with at least two nuclei having a distinct clustered chromosome morphology. Only nuclei with clustered chromatin were counted to avoid confusion with early pachytene. The pachytene zone was defined as starting at the first row after the LZ zone and extending until diplotene.

Quantitative analysis of germ cell apoptosis

Germ cell corpses were scored in 20-hr post-L4 adult hermaphrodites using acridine orange (AO), as described in Kelly et al. (2000). A minimum of 16 gonads were scored for each genotype. For the temperature-sensitivity assays, mpk-1(ga111) and wild-type worms were shifted to 25° 4-hr prior to AO incubation. Statistical analyses were performed using the two-tailed Mann–Whitney U-test (95% C.I.).

Quantification of immunofluorescence signals

Activated MPK-1 fluorescence intensity was quantified on raw images taken from whole-mounted gonads of wild-type, ogr-2(huj1), lip-1(zh15), and ogr-2(huj1); lip-1(zh15) worms stained with an anti-dpMPK-1 antibody, using the same experimental conditions and identical acquisition parameters. ImageJ software was utilized to measure the fluorescence intensity level in a rectangular area at the middle of each meiotic stage.

Time-course analysis for RAD-51 foci

RAD-51 foci were quantified for all seven zones of the gonad, starting with the premeiotic tip to late pachytene regions of the germline, as in Colaiácovo et al. (2003), McClendon et al. (2016). The total number of nuclei was scored per zone from three gonads, each for wild-type and ogr-2 mutant worms. The data were also analyzed according to meiotic stages. Statistical comparisons between genotypes were performed using the two-tailed Mann–Whitney U-test, 95% C.I.

Quantification of ZHP-3 foci

For quantification of GFP::ZHP3 foci, nuclei that were in the last four-to-five rows of late pachytene and were completely contained within the image stack were analyzed (see Imaging and microscopy). Foci were quantified manually from deconvolved 3D stacks.

RNAi

Feeding RNAi experiments were performed at 20°, as described in Govindan et al. (2006, 2009). Control RNAi was performed by feeding HT115 bacteria carrying the empty pL4440 vector. A feeding vector from the C. elegans RNAi collection (Source Biosciences, Nottingham, UK) was used to deplete ogr-2 (F52H3.4).

HIM-8 pairing analysis

The gonads were divided into five developmental stages corresponding to the mitotic, LZ, early-, mid-, and late-pachytene stages. Nuclei were scored for HIM-8 foci pairing. Scoring was done similarly to FISH analysis in MacQueen et al. (2002), with a 0.7-µm distance between HIM-8 foci used as the cutoff for paired homologous chromosomes.

Quantitative real-time PCR

For quantitative real-time PCR (qRT-PCR), total RNA was isolated from whole worms using Direct-zol Miniprep Plus (Zymo Research), according to the manufacturer’s instructions. Worms were subjected to nine rapid freeze–thaw cycles in Trizol (Invitrogen, Carlsbad, CA) prior to RNA isolation. Next, 2 μg of total RNA were reverse-transcribed using SuperScript III Reverse Transcriptase (Invitrogen). Quantitative PCR was performed using Power SYBR Green Mastermix (Applied Biosystems, Foster City, CA) using an Applied Biosystems Step One Plus apparatus (Applied Biosystems). Experiments were performed in triplicate and all data were normalized to glyceraldehyde 3-phosphate dehydrogenase-1 (gpd-1). The primers used were as follows:

• lip-1 Forward: 5′-GTG ATT CAG TGA AAC GAG CCA AT-3′.

• lip-1 Reverse: 5′-TTG AAC CGC CCA TTA AAC G-3′.

• gpd-1 Forward: 5′-ACT CGT CCA TTTT CGA TGC T-3′.

• gpd-1 Reverse: 5′-TCG ACA ACA CGG TTC GAG TA-3′.

Data availability

Strains and plasmids are available upon request. Supplemental Material, File S1 contains descriptions of all supplemental figures. File S2 contains detailed descriptions of all primers used for genotyping. Figure S1 contains the sequences for the engineered mutations. Supplemental material available at https://doi.org/10.25386/genetics.7825712.

Results

SPK is a domain, 45 residues in length, which is present in protein kinases of undetermined specificity (STYKc), caspases, SET methyltransferases, and other proteins (Doerks et al. 2002). Currently, little is known about this domain beyond its presence in this group of proteins. The domain is present in the N-terminus of the 349-aa protein encoded by F52H3.4, a previously uncharacterized gene expressed in the germline (Reinke et al. 2000; Ortiz et al. 2014; Diag et al. 2018; Tzur et al. 2018) (Figure S1A). Below, we describe multiple roles for F52H3.4 in the hermaphrodite adult germline. Accordingly, we renamed this gene oocyte and germline-related 2 (ogr-2).

CRISPR/Cas9-engineered deletions reveal roles for ogr-2 in fertility and prophase I progression

To uncover the genetic roles of ogr-2 in the germline, we used the CRISPR/Cas9 system (Friedland et al. 2013; Chen et al. 2014; Paix et al. 2015) and isolated two independent null alleles, ogr-2(huj1) and ogr-2(huj18), with deletions encompassing the entire ogr-2 ORF (see Materials and Methods and Figure S1, B–D). Both alleles showed similar phenotypes (Figure S2). Unless otherwise specified, all further experiments were conducted using the five-time outcrossed homozygous ogr-2(huj1) mutant (hence forward ogr-2Δ).

We first determined whether ogr-2 affects fertility. Worms carrying the ogr-2(huj1) allele exhibit a mild, albeit statistically significant, reduction in brood size compared to wild-type (251 ± 25 vs. 305 ± 38, n = 19 and 20, respectively, see Figure S2). We further monitored embryonic lethality (Emb), a phenotype associated with meiotic defects (as well as other developmental perturbations), and found that it is very slightly increased in ogr-2Δ compared to wild-type worms (0.5% ± 0.1 vs. 0.2% ± 0.1, n = 19 and 20 for ogr-2Δ and wild-type, respectively, Figure S2). These results suggest that ogr-2 may have a role in fertility or development, as well as potential roles in meiosis.

To determine whether ogr-2 deletion leads to meiotic defects, we examined DAPI-stained mutant gonads. In wild-type young adult hermaphrodite worms, the U-shaped gonads display a highly ordered progression of nuclei [Figure 1A, reviewed in Schvarzstein et al. (2010), Lui and Colaiacovo (2013), Pazdernik and Schedl (2013), and Hillers et al. (2017)]. We found that the chromatin morphology in most regions of the ogr-2Δ gonads is normal; however, we noticed that the relative size of the stages was altered (Figure 1, B–F). More specifically, in ogr-2Δ gonads, we found a significant elongation of the mitotic zone. We counted the number of nuclei between the distal tip until the first crescent-like nucleus, which marks the beginning of the first meiotic stage, the LZ stage. We found that the mitotic population is significantly increased in the mutant compared with wild-type worms (260 ± 15 vs. 226 ± 16, n = 8 and 5, respectively, Figure 1C), indicating that there are more premeiotic nuclei in the ogr-2 mutant gonads than in the wild-type. To determine whether this increase results from the proliferative nuclei population taking up a larger part of the total number of germ cells, we compared the number of proliferative nuclei relative to the total number of nuclei from the distal tip to the end of pachytene. Indeed, the proliferative nuclei in ogr-2Δ compose 0.35 ± 0.03 of this population vs. 0.27 ± 0.02 in the wild-type (Figure 1D). We also tested whether this increase is due to a change in the proliferation rate. To this end, we monitored germ cell proliferation by staining the gonads with an antibody against phospho-histone H3 (pH3), a marker of nuclei in M-phase (Hendzel et al. 1997; Hans and Dimitrov 2001). We found that the number of pH3-positive nuclei was not significantly higher in ogr-2Δ vs. wild-type gonads (Figure S3). These results suggest that the lack of OGR-2 increases the size of the mitotic population and that ogr-2 plays a role in the proper timing of the mitotic-to-meiotic switch.

Figure 1.

The roles of ogr-2 and lip-1 in meiotic progression. (A) An illustration of the C. elegans gonad. (B) DAPI staining of WT, lip-1(zh15), ogr-2Δ (ogr-2(huj1), a complete deletion of ogr-2), and ogr-2; lip-1(zh15) gonads. The mitotic zone is denoted by yellow lines, the LZ region by green lines, and pachytene by blue lines. (C–F) Box plots depict the average number of nuclei (C and D) in the mitotic and (E and F) LZ zones. (C and E) The absolute number of nuclei (D and F), the number of nuclei within the indicated zones relative to the total number of nuclei present from the distal tip to the end of pachytene [the red lines in (B)]. Bar, 10 µM. * P < 0.05, by two-tailed Mann–Whitney U-test. Bars represent the SD. “X” indicates the mean, which is also indicated above each plot. The lengths of the mitotic and LZ zones are significantly altered in the ogr-2Δ gonads. Note that the lip-1(zh15) mutation reduces the total number of germ cells; however, similar changes in the relative size of the mitotic and LZ zones are observed in both mutants. LZ, leptotene/zygotene; NS, not significant; WT, wild-type.

Our finding that the lack of OGR-2 leads to elongation of the mitotic zone suggests that some regions of prophase I may be shorter. Indeed, although in ogr-2Δ the nuclei enter later into the LZ stage and the chromosomes acquire the polarized spatial organization characteristic of this stage, the chromosomes redisperse and adopt a characteristic pachytene morphology earlier, resulting in a significantly shorter LZ zone in ogr-2Δ (66 ± 9 n = 8) vs. wild-type worms (136 ± 20, n = 6, Figure 1, E and F). These results suggest that ogr-2 is involved in leading to timely mitotic-to-meiotic entry and may also play a role in ensuring the correct length of the LZ stage.

Deletion of ogr-2 has a subtle effect on CO designation

Meiotic prophase progression was previously linked to CO control and the formation of bivalent structures at diakinesis (Carlton et al. 2006). During the transition from pachytene to diakinesis, several proteins relocalize to specific parts of the chromosome axes (Martinez-Perez and Villeneuve 2005; Nabeshima et al. 2005; de Carvalho et al. 2008; Martinez-Perez et al. 2008; Severson et al. 2009; Tzur et al. 2012; Severson and Meyer 2014). The axial regions extending from the single chiasma to the farthest end are termed “the long arms” of the bivalents, whereas the regions from the chiasma to the closer end are termed “the short arms” of the bivalents (Figure 2A). During chromosome remodelling, LAB-1 (Long Arms of the Bivalent-1) becomes restricted only to the long arms of the bivalent, thus ensuring that the sister chromatids segregate together at metaphase I (de Carvalho et al. 2008). Both ogr-2Δ and ogr-2 RNAi animals exhibited incompletely penetrant chromosome remodeling defects (Figure 2, A and B). RNAi knockdown of ogr-2 in a strain stably expressing LAB-1, fused to GFP, and mCherry, fused to histone H2B, led to LAB-1 mislocalization in 15% of the oocytes (n = 20), where it was present in both the long and short arms of the bivalents (Figure 2A). Similarly, immunostaining using LAB-1 antibodies showed its mislocalization to the short arms of the bivalent in 23% of ogr-2Δ oocytes (Figure 2B, n = 48–1 oocytes). However, in wild-type worms, LAB-1 was detected only on the long arms (n = 50). During this analysis, we also observed that a subpopulation of ogr-2Δ −1 oocytes presented one bivalent with a ring shape (Figure 2C). This type of bivalent structure is indicative of double chiasmata created by two COs, which is extremely rare in wild-type worms (Meneely et al. 2002; Nabeshima et al. 2004; Carlton et al. 2006; Lim et al. 2008; Martinez-Perez et al. 2008; Libuda et al. 2013; Gabdank and Fire 2014). In ogr-2 mutants, we detected bivalents with a ring shape in 6.2% (n = 48) of diakinesis nuclei, which we never observed in the wild type (n = 50). To verify this result and to determine whether CO designation is perturbed in ogr-2Δ, we quantified CO designation levels cytologically. Since ogr-2 is genomically located near the integration site of the tagged cosa-1/CNTD1, an early CO designation marker (Yokoo et al. 2012), we were unable to cross the transgene into the ogr-2 deletion background. Therefore, we used a GFP-tagged ZHP-3/Zip3 strain (Bhalla et al. 2008). ZHP-3, a conserved CO-promoting protein, initially localizes in stretches along the full length of the SCs, and then begins to concentrate at one side of each presumptive CO site during late pachytene (Jantsch et al. 2004; Bhalla et al. 2008). ZHP-3::GFP foci were previously used to quantify CO precursor sites (Bhalla et al. 2008; Youds et al. 2010; Yokoo et al. 2012; Nadarajan et al. 2016). In wild-type gonads, we detected six GFP::ZHP3 foci in 93.5% (n = 139) of nuclei reflecting a single-CO designation site on each homolog pair (Figure 2, D and E). In contrast, in ogr-2 mutants, we observed an increase in the number of foci per nucleus (9% of the nuclei had seven GFP::ZHP3 foci vs. 0.72% in the wild type, P-value < 0.0005 Figure 2, D and E), suggesting a role for OGR-2 in limiting ZHP3-marked CO precursors to one per chromosome pair. Taken together, these results indicate that ogr-2 plays a role in chromosome remodeling and may also be involved in maintaining tight CO interference.

Figure 2.

Deletion of ogr-2 leads to the formation of aberrant bivalent structures and increased ZHP-3 foci with normal SC formation. (A) Images of bivalents in diakinesis oocytes with mCherry-tagged histone H2B and GFP-tagged LAB-1 of control and ogr-2 RNAi worms. (B) Immunostaining for LAB-1 in a bivalent of a diakinesis oocyte in WT and ogr-2Δ gonads. Red: LAB-1. Blue: DAPI. Arrows: LAB-1 mislocalization to the short arms. Bar, 1 µM. (C) DAPI staining of bivalents in diakinesis nuclei. In total, 6.2% of ogr-2Δ nuclei contain a bivalent with a ring shape (depicted). Arrowheads point toward chiasmata. Bar, 1 µM. (D) Representative images of gonads with ZHP-3 foci (green) stained with DAPI (blue) from late-pachytene nuclei in WT and ogr-2Δ mutants. Arrows indicate a nucleus with seven foci. Bar, 5 µM. (E) Distribution of the percentage of ZHP-3 foci per late-pachytene nucleus. Note the increase in seven foci in ogr-2Δ. (F) Pachytene nuclei immunostained for SYP-2 (red) and HTP-3 (green). Bar, 4 µM. RNAi, RNA interference; SC, synaptonemal complex; WT, wild-type.

The SC is properly formed in ogr-2 mutants

During the LZ stage, the SC begins to assemble along the length of the paired chromosomes to keep them closely associated and aligned (Colaiácovo et al. 2003; Couteau and Zetka 2005; Hayashi et al. 2010; Schild-Prüfert et al. 2011). This zipper-like structure is composed of lateral element proteins that are recruited to the chromosome axes and to central region proteins that localize between them [reviewed in Zetka (2009)]. We suspected that in ogr-2 mutants, the premature redispersion of the nuclei, which denotes the end of the LZ zone, may allow less time for synapsis, resulting in defects in SC assembly. To determine whether SC morphology was altered in ogr-2Δ, we performed immunostaining using antibodies against SYP-2, a central region protein (Colaiácovo et al. 2003), and HTP-3, an axial component of the SC (MacQueen et al. 2002; Goodyer et al. 2008; Severson et al. 2009). We did not observe any change in the staining of these components between wild-type and ogr-2Δ. In wild-type as well as in ogr-2Δ pachytene nuclei, both SYP-2 and HTP-3 were colocalized in long contiguous linear stretches between DAPI-stained parallel chromosome tracks, corresponding to fully synapsed homologs (Figure 2F). Similar to SC formation, its disassembly was also normal in ogr-2Δ. SYP-2 staining became restricted to the short arms at diplotene and was no longer visible on the −3 oocyte in 80 (n = 12) and 83% (n = 15) of the oocytes scored for wild-type and ogr-2Δ, respectively (Figure 1A). We never found SYP-2 staining on the −2 oocyte. This suggests that ogr-2 is not required for normal SC assembly and disassembly.

We also evaluated the effect on homologous pairing at the X chromosome pairing centers using HIM-8 antibodies. Pairing at this region was almost identical between ogr-2Δ and the wild type. We found 94 ± 6% and 93 ± 5% paired chromosomes in LZ, and 99.8 ± 0.7% and 99 ± 3% in midpachytene, for ogr-2Δ and wild-type, respectively (Figure S4). These results indicate that, at least for this chromosomal region, ogr-2 plays no role in meiotic pairing.

Taken together, these results show that although in ogr-2 mutants the LZ stage is considerably shorter, at least at this level of cytological observation, ogr-2 is not required for X chromosome pairing, or for either the formation or the disassembly of the SC.

Deletion of ogr-2 does not lead to a change in the meiotic staging of DSBR

Programmed meiotic DSB formation is initiated at the LZ stage in C. elegans (Dernburg et al. 1998; Colaiácovo et al. 2003). To determine whether this process was affected by the shorter LZ stage, we assessed DSBR progression using an antibody against RAD-51, a protein involved in strand invasion/exchange during DSBR (Rinaldo et al. 2002; Alpi et al. 2003; Colaiácovo et al. 2003). We quantified the number of RAD-51 foci in seven regions from the distal premeiotic tip to the end of pachytene. In young adult wild-type germlines, zones one and two contain mostly premeiotic nuclei, zone three contains mostly LZ nuclei, and zones four-to-seven correspond to progression from early to late pachytene [Figure 3A; Colaiácovo et al. 2003; McClendon et al. 2016)]. In wild-type worms, the number of RAD-51 foci rose at the end of zone three, peaked in zone four, and dropped in zones five and six, reflecting the progression of DSBR (Figure 3B). In ogr-2 mutants, we noticed a relative shift in the gonad position regarding the appearance of RAD-51 foci. The number of foci peaked in zone five and only decreased in zone seven (Figure 3B). This shift could be explained by the delay in meiotic onset detected in ogr-2 mutants. To test this possibility, we reanalyzed our data and binned the nuclei according to their meiotic stage. We found that RAD-51 dynamics during oogenesis were practically identical in ogr-2Δ and wild-type worms (Figure 3C). We conclude that ogr-2 deletion leads to a change in oogenesis progression, which may also lead to a shift in the gonadal position of meiotic stages and thus the position at which DSBR occurs, but it is executed at the correct meiotic stage.

Figure 3.

Double-strand break repair position in ogr-2Δ. Double-strand break repair quantification by immunostaining of RAD-51. (A) Illustration of the C. elegans gonad. The zones of RAD-51 analysis are depicted. (B) Mean foci/nuclei in the seven equally divided zones from the premeiotic tip to late pachytene. (C) The same analysis binned according to developmental stages. Bars represent the SD. LZ, leptotene/zygotene.

Apoptosis is increased in the ogr-2-deleted worms

Our observation that the number of progeny laid by ogr-2Δ mutants is significantly smaller than in wild-type, yet the embryonic lethality is almost indistinguishable from wild-type levels, suggests that many meiocytes are removed by apoptosis. We explored this possibility by scoring the number of apoptotic nuclei by AO staining, taking advantage of its preferential uptake by apoptotic corpses (Mpoke and Wolfe 1997; Gartner et al. 2000). In the bend region of wild-type gonads, we counted an average of 3.6 ± 1.6 apoptotic nuclei (Figure 4), which is in line with previous reports [e.g., Lettre et al. (2004)]. We found more than a twofold increase in the number of apoptotic corpses in ogr-2 mutant gonads: 8.1 ± 2.5 (Figure 4).

Figure 4.

Elevated apoptosis levels in ogr-2 mutants are dependent on MPK-1, but not on DSBR and synapsis. Quantification of germline apoptosis by acridine orange staining. * P < 0.05, *** P < 0.0001 by the two-tailed Mann–Whitney U-test. Bars represent the SD. Deletion of ogr-2 leads to an increase in apoptosis, which is partially additive when combined with syp-1 mutation but is blocked by mutation in mpk-1. DSBR, double-strand break repair; NS, not significant; WT, wild-type.

During oogenesis, stringent quality control systems ensure the production of normal haploid gametes. Checkpoint mechanisms prevent meiotic segregation errors by eliminating defective cells prior to the completion of meiosis (Gumienny et al. 1999; Gartner et al. 2000; Bhalla and Dernburg 2005). In C. elegans, two canonical meiotic checkpoints exist: the synapsis checkpoint and the recombination/DNA damage checkpoint. These checkpoints trigger apoptosis of the affected meiocytes in response to incomplete synapsis and/or unrepaired recombination intermediates in late meiotic prophase (Gartner et al. 2000; Bhalla and Dernburg 2005; Woglar et al. 2013). In ogr-2Δ we detected normal SC formation (Figure 2F) and normal meiotic staging of DSBR (Figure 3C), making this level of increase in apoptosis unexpected. To determine whether the increase in apoptosis is related to DSBR or synapsis, we utilized the syp-1(me17) mutation, where both synapsis and DSBR checkpoints are activated, leading to a large number of apoptotic nuclei (MacQueen et al. 2002; Ye et al. 2014). If the increase in apoptosis observed in ogr-2Δ is due to the canonical checkpoints (either DSBR or synapsis), we would expect no additive effect of these two pathways in the double mutant of ogr-2 and syp-1 worms. On the other hand, if the increase in apoptosis in ogr-2Δ is not due to activation of these checkpoints, but rather is due to a different mechanism, then the double mutant should display an additive effect.

In syp-1 mutants, we scored an average of 15.8 ± 2.6 apoptotic nuclei per gonad arm, which is in line with previous reports (Ye et al. 2014). In the ogr-2; syp1 double mutant, we found a significant increase in the number of apoptotic nuclei (18.8 ± 3.7, Figure 4). We note that the increase in apoptotic nuclei in the double mutant is less than the complete additive values of the single mutants. We hypothesize that this incomplete additive value stems from the high apoptotic levels, which may be close to the limit of the neighboring cells’ ability to process all the apoptotic nuclei with the same dynamics. We can conclude that this increase in germline apoptosis is consistent with a possible role that is independent of synapsis and DSBR.

To verify that the increase in apoptosis is not due to unrepaired DSBs, we quantified the number of apoptotic nuclei in a hus-1-mutant background. It was previously shown that HUS-1 is a DNA damage checkpoint protein and that DNA damage-induced germ cell death is abrogated in hus-1 mutants. Highly reduced levels of germ cell apoptosis are detected in hus-1 mutants following treatment with ionizing radiation (Hofmann et al. 2002; Quevedo et al. 2007). If the increased apoptosis in ogr-2Δ is due to unrepaired DSBs, then the hus-1; ogr-2Δ double mutant is expected to exhibit a lower number of apoptotic nuclei. Alternatively, if the increased apoptosis in ogr-2Δ is unrelated to DSB repair, then the hus-1 mutation in ogr-2Δ should not alter the level of apoptosis. We found 7.6 ± 1.7 apoptotic nuclei in gonads of ogr-2Δ hus-1, which is not significantly different from the levels found in ogr-2Δ (see above; Figure 4). We can conclude that the increased apoptosis in ogr-2Δ is not due to unrepaired DSBs.

MAPK, through MPK-1/ERK, was shown to increase germline apoptosis independently of the synapsis and recombination/DNA damage checkpoints (Schumacher et al. 2001; Rutkowski et al. 2011) by directly interacting with CEP-1, the C. elegans p53 homolog (Derry et al. 2001; Rutkowski et al. 2011). We investigated whether the ogr-2Δ increased apoptosis is MPK-1-dependent. To this end, we eliminated MAPK signaling in ogr-2 mutants by combining it with the temperature-sensitive mutant mpk-1(ga111ts). This mutation is a weak temperature-sensitive loss-of-function allele that reduces the rate of activation and phosphorylation of MPK-1 (Lee et al. 2007b). At a restrictive temperature (25°), the number of apoptotic nuclei in the bend region of the germline in the ogr-2; mpk-1 double mutant was significantly lower than in ogr-2Δ (3.4 ± 1.5, Figure 4) and was indistinguishable from mpk-1(ga111ts) (3.7 ± 1.6). Thus, the mpk-1 mutation suppresses the increase in germ cell apoptosis observed in ogr-2Δ. These results are consistent with our hypothesis that the elevated level of apoptosis in ogr-2 mutants is dependent on MPK-1/ERK activation, but not on the synapsis and recombination checkpoints.

It has been previously reported that activation of MPK-1 in diplotene increases apoptosis through CEP-1 (Lee et al. 2007b; Rutkowski et al. 2011). Therefore, we investigated whether the aberrant rise in apoptosis found in ogr-2 mutants is CEP-1-dependent. In cep-1; ogr-2 double mutants, the enhanced apoptosis was suppressed (2.5 ± 1.9 for cep-1; ogr-2Δ and 1.6 ± 1.6 for cep-1, Figure 4). The dependence of ogr-2 apoptotic control on cep-1 once again supports our conclusion that OGR-2 apoptotic control is dependent on MPK-1.

Inactivation of LIP-1 leads to an increase in germline apoptosis by relieving the negative regulation of MAPK signaling (Hajnal and Berset 2002; Rutkowski et al. 2011). Therefore, we investigated whether ogr-2 apoptotic control operates via the same pathway as lip-1. We found that ogr-2; lip-1(zh15) double mutants did not display an additive increase in germline apoptosis over lip-1(zh15) (11 ± 3 vs. 10 ± 2 for ogr-2Δ; lip-1(zh15) and lip-1(zh15), respectively, Figure 4). We can conclude that ogr-2 restricts germline apoptosis through mpk-1 and possibly lip-1.

Increased MPK-1 activation in ogr-2Δ

Previous studies indicated that LIP-1-mediated dephosphorylation deactivates MPK-1 in diplotene and, consequently, worms carrying a lip-1 loss-of-function mutation had elevated MPK-1 activity in this region, resulting in increased apoptosis (Hajnal and Berset 2002; Rutkowski et al. 2011). We tested the hypothesis that OGR-2 affects the proper spatial activation of MPK-1, similar to LIP-1, by assessing the localization pattern of the MPK-1 active form, i.e., dpMPK-1 in ogr-2 mutant germlines. Staining of wild-type gonads using an antibody directed against dpMPK-1 recapitulated the previous results. Two main activation regions of MPK-1 were found: the region that spans mid-to-late pachytene and a proximal region of activation in diakinesis oocytes [reviewed in Arur (2017); Figure 5A]. At other stages of the gonad, the staining was very weak (Lee et al. 2007a,b; Yin et al. 2016). Unlike in wild-type worms, dpMPK-1 staining persists in the diplotene region of ogr-2 gonads, similar to lip-1(zh15) mutants (Figure 5A; Hajnal and Berset 2002; Rutkowski et al. 2011). These results suggest that ogr-2 plays a role in limiting MPK-1 activation in the diplotene stage; thus, it regulates apoptosis levels.

Figure 5.

Aberrant MPK-1 activation regions in ogr-2Δ. (A) Immunostaining of activated MPK-1 (dpMPK-1) in WT, ogr-2Δ, lip-1(zh15), and ogr-2Δ; lip-1(zh15) gonads. Note the aberrant staining at early stages and at diplotene present in ogr-2Δ, lip-1, and ogr-2Δ; lip-1. Nonrelated gut nuclei are denoted by an asterisk. (B) Quantification of the dpMPK-1 intensity. Bars represent SEM. Asterisks indicate P < 0.05 by the two-tailed Mann–Whitney U-test for all mutants compared to WT. dpMPK-1, double-phosphorylated MPK-1; WT, wild-type.

To better characterize the effects of ogr-2 on MPK-1 activation, we quantified the intensity of dpMPK-1 staining along the different gonad regions. We detected a significant ectopic activation of MPK-1, not only in diplotene, but also in the mitotic and LZ zones of ogr-2Δ gonads (Figure 5B). Thus, ogr-2 is involved in reducing MPK-1 activation along most of the adult gonad. The correlation between the regions in ogr-2Δ gonads where MPK-1 was aberrantly activated and where meiotic progression was abnormal suggests that OGR-2 meiotic control acts via restriction of MPK activation (see the Discussion).

Mutations in lip-1 and ogr-2 result in similar defects in oocyte morphology and MPK-1 activation

In both lip-1 and ogr-2Δ, aberrant apoptotic control exists and dpMPK-1 persists in diplotene. This raises the hypothesis that they may also act at other meiotic stages via the same pathway. To test that possibility, we quantified the relative dpMPK-1 staining in lip-1 gonads. Interestingly, higher activation levels were observed not only in diplotene but also in the mitotic and LZ stages, almost identically to ogr-2Δ (Figure 5B). However, this aberrant activation was not additive. Indeed, in the double mutant ogr-2; lip-1, we found no statistically significant differences in dpMPK-1 staining compared with the single mutants; however, in all three strains we measured significantly higher dpMPK-1 levels than in the wild type (Figure 5B).

The activation of MPK-1 in lip-1 and ogr-2 was accompanied by almost identical changes in meiotic progression in the two mutants. In the early stages of lip-1(zh15) gonads, we found relative elongated mitotic zones and shorter LZ zones, very similar to the ogr-2Δ gonads. These changes were not additive, since in the double mutant we found a similar relative nuclei number compared to the single mutants (Figure 1, D and F), indicating that they work through the same pathway.

Similarities also exist between the fates of mature oocytes in ogr-2Δ and lip-1(zh15) mutants (Figure 6, A–I). Three‐day-old lip-1(zh15) mutants exhibit a defective G2/M-phase arrest, with endomitotic germ cells and smaller oocytes aligned in multiple rows owing to aberrant oocyte maturation (Hajnal and Berset 2002). Using DIC optics, we measured the size of mature oocytes and found that, in both ogr-2Δ and lip-1(zh15), they were significantly smaller than in wild-type worms (Figure 6, F–I). In line with these results, DAPI staining indicated that oocytes in both ogr-2Δ and lip-1(zh15) were often aligned in multiple rows (41.2%, n = 17 and 100%, n = 7, respectively, Figure 6, B and D). Moreover, in the gonads of both genotypes we found endomitotic nuclei strongly stained with DAPI (41.2%, n = 17 in ogr-2Δ vs. 28.5%, n = 7 in lip-1(zh15), Figure 6, C and E). These two phenotypes were never observed in the gonads of wild-type worms of the same age (n = 15, Figure 6A). The changes we found in oocyte morphology were not due to the lack of sperm, since we found a similar number of sperm in ogr-2Δ and wild-type gonads (131 ± 25 in ogr-2Δ vs. 127 ± 22 in wild-type in day 1 adult worms, and 3 ± 3 in day 3 adult worms for both).

Figure 6.

Interactions and similarities between ogr-2Δ and lip-1. DAPI-stained gonads of (A) wild-type (WT), (B and C) ogr-2Δ, and (D and E) lip-1(zh15). Bar, 50 µM. Arrows indicate oocytes stacked in pairs. Arrowheads indicate endomitotic nuclei in which the acquisition was oversaturated to facilitate the observation of both endomitotic and meiotic nuclei. (F–H) Images using DIC optics of dissected gonads of (F) N2, (G), ogr-2Δ, and (H) lip-1(zh15). (I) Quantification of oocyte size. *** P < 0.001 using the two-tailed Mann–Whitney U-test. (J) Real-time PCR analysis of lip-1 expression. * P < 0.05 using the two-tailed Mann–Whitney U-test.

The elevated level of activated MPK-1 in lip-1(zh15) was previously shown to relieve some of the phenotypes observed in temperature-sensitive mutations in mpk-1 (Hajnal and Berset 2002). We investigated whether the deletion of ogr-2 can result in the same outcome. When mpk-1(ga111ts) worms were shifted to the restrictive temperature, very few progeny were observed (6 ± 2 per worm, n = 49). Similar to previous results (Hajnal and Berset 2002), we found that a mutation in lip-1 partially relieved this effect (22 ± 5 per worm, n = 30). In ogr-2; mpk-1 double mutants, we also found a significantly greater number of progeny (47 ± 9, n = 30, P-value < 0.00001 by the two-tailed Mann–Whitney U-test, compared with mpk-1). Thus, mutations in lip-1 and ogr-2 lead to similar effects on apoptosis, MPK-1 activation, and oocyte maturation, and both partially suppress the effects of mpk-1(ga111ts) mutation.

OGR-2 localizes to chromatin and may act to enhance lip-1 transcription

To gain insight into the mechanism by which ogr-2 acts to restrict the activation of MPK-1 via lip-1, we used CRISPR engineering to insert a triple FLAG-tag before the STOP codon of the endogenous ogr-2 gene to create the OGR-2-tagged version strain, YBT44. After staining gonads of this strain with anti-FLAG antibodies, we observed an OGR-2-FLAG signal on the chromatin of all nuclei in the gonad (Figure 7). This signal was not detected in wild-type gonads (Figure S5). Interestingly, at higher magnification, stronger OGR-2-FLAG signals were observed as foci and patches on chromatin, and became stronger during mid/late pachytene (Figure 7).

Figure 7.

OGR-2 is present on germ cell chromatin. Gonad of ogr-2::3XFLAG stained with DAPI and anti-FLAG antibody. Bar for whole mount, 20 µM. Bar for mitotic zone and midlate pachytene, 10 µM.

It was previously reported that LIP-1 localizes as distinct cytoplasmic puncta during pachytene (Hajnal and Berset 2002). The increased OGR-2 chromatin-associated staining during mid/late pachytene, at the same region where dpMPK-1 staining becomes stronger, raised the hypothesis that OGR-2 acts in the nucleus to enhance lip-1 expression. To test this hypothesis, we quantified lip-1 mRNA levels by qRT-PCR. Importantly, we found that the levels were reduced by 40% in ogr-2Δ vs. wild-type (Figure 6J). We conclude that OGR-2 is present in germline nuclei, where it acts to maintain normal levels of lip-1 mRNA.

Discussion

Successful gametogenesis requires the concerted execution of complex cellular processes within the context of a viable organ. We show here that oogenesis coordination, ranging from proliferation to the mature oocyte, is skewed in a strain with complete deletion of ogr-2. Without ogr-2, entry into meiosis is delayed, the LZ stage is shorter, and the apoptosis rates are elevated (Figure 1 and Figure 4). One caveat that must be considered is that due to the similarities between LZ and early pachytene morphologies, the purity of the scored LZ population may be incomplete. In ogr-2Δ gonads, proper chromosome remodeling fails and low levels of bivalents with double chiasmata are present (Figure 2). We also found age-dependent aberrations in oocyte morphology in ogr-2Δ and stacked oocytes entering endomitosis without fertilization (Figure 6). The aberrant developmental progression and high apoptosis levels probably lead to the lower numbers of progeny observed in these worms. Concurrent with the developmental changes observed in this strain, we found that MPK-1 activation is elevated in ogr-2Δ gonads in the proliferative, LZ, and diplotene stages; we propose that this increase is the source of many of the observed meiotic alterations (discussed next).

The roles of ogr-2 in attenuating MPK-1 activation

Oogenesis, like many other developmental and differentiation processes, is orchestrated by executing a genetic program that is activated by intra- and internuclei signaling. MAPK signaling has been found to be involved in many processes throughout oogenesis, ranging from yeast to mouse (Zhang et al. 2010; Sukegawa et al. 2011; Nabti et al. 2014; Arur 2017). In C. elegans, where oogenesis is continuous, MAPK signaling is required for meiotic coordination and, when it is blocked, the program fails and the nuclei are arrested at the end of pachytene (Church et al. 1995). The model in which MAPK signaling facilitates oogenesis progression is supported by the activation and deactivation of MPK-1 at different parts of the gonad, a switch that was shown to regulate sets of genes required for meiotic coordination (Leacock and Reinke 2006; Lin and Reinke 2008; Arur 2017). Mutations that affect MPK-1 activation result in aberrations during nearly every stage of this developmental program (Gumienny et al. 1999).

The deletion of ogr-2 leads to higher levels of MPK-1 activation in many regions including the mitotic and the LZ regions, as well as during the diplotene stage (Figure 5). Comparable to other mutations that change MAPK signaling, this deletion also results in meiotic progression phenotypes at many stages. These include delayed meiotic entry, elevated apoptosis levels, altered chromosome remodelling, and endomitosis (Church et al. 1995; Hajnal and Berset 2002; Lee et al. 2006, 2007a; Yin et al. 2016). Surprisingly, the meiotic changes in ogr-2Δ only slightly affect embryonic lethality. We speculate that this is due to the high level of apoptosis, which removes most of the affected nuclei. The elevated apoptosis levels were MPK-1-dependent, supporting our conclusion that the ogr-2Δ germline phenotypes at least partly result from the elevated activation levels of MPK-1.

The model that emerges from these results is that OGR-2 attenuates MPK-1 activation levels, keeping them low before midlate pachytene and during diplotene, to allow the timely execution of meiosis. When OGR-2 is not present, entry into meiosis is delayed, which probably allows less time for recombination. It is possible that the aberrations we observed at the late stages in ogr-2Δ gonads, which include altered chromosome remodelling and chiasmata levels, are the secondary outcome of earlier events. Alternatively, these defects could directly result from the elevated MAPK signaling at these later stages, since numerous studies showed that correct MPK-1 activation is required for the last steps of prophase I (Church et al. 1995; Hajnal and Berset 2002; Lee et al. 2007a,b; Arur et al. 2009).

Does OGR-2 affect germline development through LIP-1?

Other proteins were previously shown to control meiotic progression by restricting the extent of MPK-1 activation. For example, a mutation in kin-18 leads to “flickering” MPK-1 activation from the early-to-late oogenesis stages. These regions of activation encompass different numbers of nuclei (Yin et al. 2016). These worms have a mixture of pachytene and diakinesis nuclei, and nuclei with late prophase I chromosome morphology are present at distal parts of the gonad. In kin-18 mutant gonads, RAD-51 and COSA-1 foci appear much earlier (Yin et al. 2016). Mutations in the germinal center kinase gene gck-1 lead to increased staining of activated MPK-1 in pachytene. This results in fewer germ cells, higher apoptotic levels, small oocytes, and nuclei with an undefined chromatin morphology (Schouest et al. 2009). These mutations indicate that correct MPK-1 activation along the gonad is a prerequisite for proper oogenesis progression.

One of the best-characterized regulators of MPK-1 is LIP-1. Similar to other genes that affect MPK-1 activation, lip-1 is also required for normal oogenesis progression, germline proliferation, and oocyte maturation (Hajnal and Berset 2002; Lee et al. 2006; Rutkowski et al. 2011; Cha et al. 2012). Several lines of evidence support a common pathway for lip-1 and ogr-2 in controlling meiotic progression through MPK-1 deactivation: (1) in both mutants there is a similar change in the spatial activation of MPK-1, and it is present in the mitotic and diplotene regions (Figure 5); (2) both are involved in maintaining the relative size of the mitotic and LZ populations (Figure 1); (3) in mutants of both genes, there is increased apoptosis in the bend region, which is not dependent on recombination or synapsis checkpoints (Rutkowski et al. 2011; Figure 4); (4) both play a role in oocyte maturation and G2/M transition block prior to fertilization (Hajnal and Berset 2002; Figure 6) in older hermaphrodites; and (5) mutations in both genes can relieve the sterility of mpk-1 temperature-sensitive alleles [this work and Hajnal and Berset (2002)]. Interestingly, lip-1 and ogr-2 transcripts were found to be associated with both FBF-1 and FBF-2 (Lee et al. 2006; Prasad et al. 2016), two partially redundant proteins expressed at the distal part of the gonad following GLP-1/Notch signaling. The FBF proteins reduce the stability and translation efficiency of hundreds of transcripts to maintain the proliferative population (Prasad et al. 2016). Thus, the same mechanism probably prevents the expression of lip-1 and ogr-2 distally to maintain the proliferative population. The common phenotypes of lip-1 and ogr-2Δ suggest that ogr-2 is part of the same pathway that utilizes the phosphatase LIP-1 to restrict the activation of MPK-1; thus, ogr-2 coordinates oogenesis progression. We note that some phenotypes are not shared between the mutants. For example, LAB-1 is not mislocalized in lip-1 mutants (Nadarajan et al. 2016), as it is in the minority of ogr-2 diakinesis bivalents (Figure 3, A and B). On the other hand, in lip-1 mutants, defects exist in SC disassembly (Nadarajan et al. 2016); these defects are not observed in ogr-2 mutants (this study). In addition, lip-1 has ∼50% embryonic lethality (Hajnal and Berset 2002), compared with only 0.5% in ogr-2Δ. We hypothesize that this could result from the following: (1) the MPK-1-independent roles of LIP-1; (2) the roles of OGR-2, which are mediated through other pathways; or (3) the partial effect of OGR-2 on LIP-1 (Figure 6J).

How can OGR-2 promote LIP-1 activity? To the best of our knowledge, the function of the SPK domain has not yet been resolved, but since it is also present in proteins that are part of chromatin-modifying complexes (Doerks et al. 2002), one attractive possibility is that OGR-2 is part of such a complex. This possibility is supported by the presence of OGR-2 on the chromatin in foci and patches (Figure 7), and it can explain how ogr-2 promotes lip-1 expression through transcriptional regulation. This regulation could either affect lip-1 directly or it could act through genes upstream of lip-1, such as dpl-1 or efl-1 (Lin and Reinke 2008). This possibility can also explain why we failed to find additive effects in the lip-1 and ogr-2 double mutant for most of the phenotypes tested (Figure 1, Figure 4, and Figure 5), and why sometimes the ogr-2Δ phenotype is milder than the lip-1 phenotype. If a lack of ogr-2 affects a chromatin modifier complex, it is expected to influence the expression of many genes, which would explain why some of the ogr-2 phenotypes are not shared with lip-1 (as mentioned before). In the future, it would be interesting to determine which proteins interact with OGR-2. Note that we measured only a 40% reduction of lip-1 in ogr-2Δ worms, while to the best of our knowledge no haploinsufficiency was reported for lip-1 mutants, yet there are still similar phenotypes in lip-1 and ogr-2Δ. We believe that this can be explained by the fact that we measured lip-1 levels in mixed populations of whole worms, whereas OGR-2 may play specific roles in the gonad.

Most of the data presented here suggest that OGR-2 functions by reducing MPK-1 activation. Nevertheless, it is also possible that OGR-2 plays MPK-1-independent meiotic roles that remain to be defined. Genetic interactions suggest that OGR-2 attenuates MAPK signaling via LIP-1. We note that in both mutants the increased MPK-1 activation (as observed by dpMPK-1 staining) is less than an order of magnitude greater than in wild-type worms. Nevertheless, the effects of this mild difference are evident in several oogenesis stages, highlighting the importance of accurate MAPK signaling. It is possible that small changes in MAPK activation are amplified to other local and global signaling molecules and effectors, and, if so, the outcome is detrimental. Spatial proteomics analyses in the gonad will advance our understanding of how oogenesis is coordinated in the syncytial environment of the worm’s gonad. Applying these future tools in lip-1 and ogr-2Δ gonads would help reveal the effects that small changes in MAPK signaling induce, as well as how this controls meiotic progression at each step.