PARPs and DNA Damage Response

Poly(ADP-ribose) polymerases (PARPs) mediate poly(ADP-ribosyl)ation (PARylation), a critical post-translational modification involved in DNA repair, chromatin remodeling, and cell death. PARP-1 and PARP-2 are activated by DNA strand breaks and catalyze the transfer of ADP-ribose units from NAD⁺ to nuclear proteins, facilitating the recruitment of DNA repair factors. Poly(ADP-ribose) glycohydrolase (PARG) reverses PARylation by degrading PAR chains, thereby restoring protein function and terminating the DNA damage response in mammalian cells [1,2]. These enzymes play central roles in base excision repair (BER) and also contribute to homologous recombination (HR) and non-homologous end joining (NHEJ).

Functional Overlap and Divergence Between PARP1 and PARP2

While PARP1 is the most abundant and catalytically active family member, PARP2 shares overlapping functions in DNA repair and cell cycle regulation. Both are involved in maintaining genomic stability, repairing single-strand breaks, restarting stalled replication forks, and suppressing tumorigenesis [3–6]. These redundancies are likely due to structural similarities in their catalytic domains [7]. However, PARP2 is preferentially activated by DNA gaps, flaps, and recombination intermediates rather than the nicks and DSBs that activate PARP1 [8–11]. These distinctions suggest that PARP-1 and PARP-2 have complementary but non-redundant roles in DNA repair.

PARP-2-Specific Roles in DNA damage repair

Accumulating evidence highlights its unique and non-redundant roles in DNA damage repair. For instance, PARP2 expression is induced by DNA-damaging agents such as mitomycin C, radiation, and doxorubicin, with elevated levels correlating with more aggressive tumor phenotypes [12,13]. Importantly, depletion of PARP2—but not PARP1—results in heightened sensitivity to neocarzinostatin, a double-strand break (DSB)-inducing agent [14]. Moreover, PARP2 plays a protective role against illegitimate IgH/c-myc recombination during class switch recombination, highlighting its specificity in maintaining genomic stability [15]. Lastly, PARP2 limits the accumulation of 53 BP1 at DSB sites, thereby facilitating CtIP-mediated end resection regardless of PAR synthesis activity in mammalian cells.

However, most of these insights rely on transformed cell lines or somatic tissues, leaving the spatiotemporal dynamics of PARP2-mediated DNA-damage responses in an intact organism largely uncharted. The C. elegans germline offers a unique opportunity to close this gap: its stereotyped mitotic, transition, and pachytene zones can be visualized at single-cell resolution with CRISPR knock-in reporters, while genome-wide RNA-seq and ChIP-seq can be performed in an isogenic background.

Evolutionary Conservation and C. elegans PARP Homologs

The functional conservation of PARPs is evident across species, with homologs identified in humans, mice, plants, and Caenorhabditis elegans [3,16,17]. In C. elegans, five genes are involved in PAR metabolism: parp-1 (pme-1), parp-2 (pme-2), parg-1 (pme-3), parg-2 (pme-4), and tank-1 (pme-5). PARP-1/PME-1 is a 108 kDa protein with approximately 31% overall structural similarity and 76% catalytic domain similarity to human PARP-1 (Fig 1A). PARP-2/PME-2 likewise shares structural similarity with its human counterpart, particularly in the conserved catalytic PARP signature motif. PARG-1 and PARG-2 are homologous to mammalian PARG and are responsible for the degradation of poly(ADP-ribose) chains [18]. TANK-1/PME-5, the C. elegans ortholog of human tankyrases (PARP5), is also implicated in PAR signaling and telomere regulation [19].

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Characterization of C. elegans PARP proteins and the parp-2 mutants.

(A) Schematic representation of domain structures of C. elegans PARP-1, PARP-2, and PARG-1. Functional domains, including the WGR (tryptophan–glycine–arginine-rich) domain, zinc finger domains, and the catalytic domain, are indicated. Positions of the mutations used in this study (ok344, ok988, gk120, and E509K) are marked on the corresponding proteins. Numbers denote amino acid positions. The parp-2(ok344) mutation carries a 1576 bp deletion that eliminates 20% of Exon 1 and all four remaining exons and beyond, resulting in a predicted null allele. This deletion removes the entire PARP superfamily domain responsible for polymerase catalysis and NAD+ contact residues, as well as part of the DNA-binding domain located in exon 1 suggesting a null mutant [23]. The parp-2(E509K) mutation corresponds to a glutamate-to-lysine substitution at position 988 in human PARP1, a change known to abolish PARylation activity. The parp-1(ok988) mutation is a complex allele involving a 1007 bp deletion and a 40 bp insertion. The deleted region is replaced by a 40 bp insertion. The parg-1(gk120) mutation results in a deletion of the first four exons and approximately half of the fifth exon. Due to alternative splicing, this may vary slightly among the nine known parg-1 isoforms. Notably, the deletion does not affect the catalytic domain, which resides near the C-terminus. (B) Alignment of the C-terminal amino acid sequences of C. elegans PARP-1 (NP_491072.1), PARP-2 (NP_001022057.1), and PARG-1 (NP_001023135.1) with their mammalian counterparts: human PARP1 (NP_001609.2), PARP2 (NP_005475.2), and mouse PARP2 (NP_033762.1). Arrows indicate conserved glutamate residues known to be critical for PARylation activity [24,25]. Dots and asterisks indicate the degree of conservation among the aligned sequences. (C) Quantitative PCR analysis shows approximately 66% reduction in parp-2 transcript levels in ok344 mutants compared to wild type (P = 0.0021, two-tailed Mann-Whitney test). (D) Western blot analysis revealed a strong ~61 kDa band in wild-type animals, corresponding to PARP-2, and a faint band in ok344 mutants. Animals expressing FLAG::3xHA::PARP-2 exhibited a slightly heavier band, as expected due to the presence of the epitope tags. (E) Western blot analysis of PARylation in WT, parp-2(ok344), and E509K animals generated via CRISPR-Cas9 at the endogenous parp-2 locus. WT shows strong PARylation signals at ~180 kDa and higher–molecular-weight regions, with additional lower–molecular-weight bands. The ok344 mutant displays reduced PARylation (0.68-fold overall; 0.64-fold at ~180 kDa, normalized to GAPDH). E509K shows a modest decrease at ~180 kDa (0.78-fold) but exhibits smear in the 120–180 kDa range, consistent with loss of PARylation but retention of MARylation activity. (F) PARylation in parp-2(ok344) mutants and parp-2(ok344) + parp-1 RNAi double depletion. The ok344 single mutant shows ~29% reduction of the ~ 180 kDa band relative to WT, whereas double depletion results in ~52% reduction (additional ~32% decrease compared to ok344 alone). Non-specific bands detected between 130–55 kDa were excluded from quantification.

https://doi.org/10.1371/journal.pgen.1012062.g001

Functional Roles of PAR Genes in DNA damage and meiotic repair

These five C. elegans PAR genes, which mediate or reverse PARylation, contribute to various biological processes, including cell development, apoptosis, and DNA damage signaling and repair [19,20]. Notably, PARP-1, PARP-2, and PARG-1 are essential for the cellular response to ionizing radiation (IR). Following γ-irradiation, wild-type worms show a sharp decline in NAD⁺ levels and a marked accumulation of PAR polymers, indicating the activation of PARP-1 and PARP-2 in response to DNA damage [21]. Also, RNAi-mediated knockdown of tank-1/pme-5 leads to enhanced germ cell apoptosis in a radiation time-dependent manner, suggesting its involvement in IR-induced DNA damage repair [22]. Moreover, PARG-1 has been shown to facilitate the formation and repair of meiotic double-strand breaks (DSBs) via homologous recombination during meiosis [23].

PARP-2 as a Unique regulator of replication stress and DNA Repair

Here, using C. elegans parp-2 null (ok344) and catalytic-dead (E509K) mutants, we dissected the functional significance of PARP-2 through genetic, molecular, and transcriptomic analyses. We show that PARP-2 displays dynamic subcellular localization during germline development—associating with both chromatin and the nucleolus—and is essential for fertility, embryogenesis, and germline architecture. Loss of PARP-2 confers hypersensitivity to replication stress and DNA cross-linking agents, accompanied by checkpoint activation, RAD-51 and FCD-2 accumulation, and defects in homologous recombination and interstrand crosslink repair.

Transcriptomic profiling revealed stronger DNA-damage pathway induction in E509K than in null alleles, including the FA, homologous recombination, and base-excision repair networks. By contrast, nucleolar/RNA-synthesis genes split into two classes: four loci responded equally in both mutants, whereas five loci were up-regulated exclusively in the null, indicating gene-class-specific requirements for PARP-2 catalysis. ChIP-seq further identified PARP-2 binding at promoter-proximal regions of stress-response genes, supporting both transcriptional and scaffolding functions beyond catalysis. Collectively, our data demonstrate that PARP-2 fulfills non-redundant roles in germline genome maintenance, cooperates with the FA pathway under replication stress, and differentially regulates DNA-repair and ribosomal gene expression through distinct catalytic and structural modes.

Conserved Architecture and Catalytic Features of PARylation-Associated Enzymes in C. elegans and Mammals

The domain structures of C. elegans PARP-1, PARP-2, and PARG-1 are well conserved when compared to human counterparts [24]. Key functional regions—including the DNA-binding domain and the catalytic domain (also known as the PARP signature)—are also well preserved.

At the amino acid level, a glutamate residue at position 988, known to be critical for the polymerization activity of human PARP-1 (hsPARP1), is also conserved in C. elegans PARP-1, PARP-2, and PARG-1, as well as in hsPARP2 and mouse PARP2 (mmPARP2) (Fig 1A and 1B, [24,25]).

We validated the expression of PARP-2, the main focus of our research, using qPCR and Western blot analysis. The ok344 mutant is a deletion that includes the end of exon 1 through the last exon and beyond (Fig 1A), and is thus considered a predicted null mutant. Indeed, the mRNA expression level in the ok344 a decrease by a 66% compared to control wild type (Fig 1C, 1.0 vs 0.34 in qPCR, P = 0.0021).

Consistently, Western blot analysis showed that, while wild-type animals expressed a strong band at the predicted size of 61 kDa using a PARP-2 antibody, the signal intensity was reduced to approximately 53% in parp-2(ok344) mutants (Fig 1D). Notably, a residual signal remained detectable, which may reflect cross-reactivity of the antibody—raised against human PARP-2—with conserved epitopes in C. elegans. Alternatively, this residual signal could arise from low levels of PARylation activity independent of PARP-2 and PARP-1, or from incomplete depletion in RNAi-treated backgrounds. The CRISPR-generated reporter expressing FLAG::3xHA::PARP-2 at its endogenous locus exhibited a slightly higher molecular weight band, as expected.

While C. elegans PARP-2 activity has been assessed in vitro [24], its in vivo function had not been characterized. In wild-type (wt) animals, PARylation signals were detected at ~180 kDa and in the higher–molecular-weight region (indicated by the arrow), and multiple bands as well as smear-like patterns were also observed below 50 kDa (Fig 1E). Compared to WT, the parp-2(ok344) mutant showed an overall reduction of PARylation signals from ~35–180 kDa and higher to 0.68-fold relative to wt after normalization to GAPDH (32% decrease). The prominent high–molecular-weight PARylation band around ~180 kDa was also reduced to 0.64-fold relative to wt (GAPDH-normalized), corresponding to a 36% decrease.

The CRISPR-generated parp-2 (E509K) mutants—corresponding to a glutamate-to-lysine substitution at position 988 in human PARP1 [25] - exhibited a similar reduction but a distinct pattern. The ~ 180 kDa band was reduced to 0.78-fold relative to GAPDH-normalized wt, corresponding to a 22% decrease. In addition, smear-like or multiple bands appeared in the 180–120 kDa range, which were not clearly detected in wt or ok344. This pattern is consistent with the interpretation that E509K fails to form long PAR chains, resulting in reduced high–molecular-weight PARylation and the accumulation of lower–molecular-weight species likely corresponding to mono-ADP-ribosylated proteins.

These observations support the idea that ok344 is a mutant that lacks PARylation as well as mono-ADP-ribosylation (MARylation) activity, whereas E509K retains mono-ADP-ribosylation activity but is defective in PARylation.

To evaluate the functional redundancy between PARP-1 and PARP-2 in PARylation, we compared the parp-2(ok344) single mutant and the parp-2(ok344) + parp-1 RNAi double depletion (Fig 1F). Consistent with Fig 1E, the parp-2(ok344) mutant showed an approximately 29% reduction in the ~ 180 kDa PARylation band relative to wt after GAPDH normalization. In contrast, the double depletion resulted in a 52% reduction relative to wt, which corresponds to an additional ~32% decrease compared to the normalized ok344 single mutant. Non-specific signals were occasionally detected across all samples in the 130–55 kDa region; therefore, this region was excluded from quantification.

Taken together, these results indicate that PARP-1 partially compensates for the loss of PARP-2, supporting the interpretation that the two enzymes exhibit redundant PARylation functions.

PARP-2 Dynamically Distributes in the Germline during Meiotic Prophase and Shows Increased Signal upon UV or HU Treatment

1. a. Dynamic Localization of PARP-2 During Meiotic Prophase

To gain insight into the function of PARP-2, we examined the localization of the PARP-2 reporter worm (FLAG::3xHA::PARP-2) at its endogenous locus by immunostaining dissected gonads from wild-type hermaphrodites. PARP-2 signal is observed in mitotic nuclei at the distal tip (premeiotic tip, Fig 2A and 2B). This signal is then reduced upon entrance into meiosis (leptotene/zygotene stages = transition zone) and remains weak through the mid-pachytene stage while control groups bearing no reporter genes did not exhibit distinct signals. The signal increases once again in late pachytene nuclei and persists through late diakinesis oocytes. This dynamic pattern of expression suggests regulation of PARP-2 during meiotic prophase. PARP-2 localization is also observed in embryonic nuclei and this signal is specific since it is absent in control groups bearing no reporter genes (Fig 2B).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Stage-specific nuclear localization of PARP-2 in the germline, its colocalization with nucleolar structures, and induction by DNA damage.

(A) Immunolocalization of PARP-2 expression using CRISPR-generated FLAG::3xHA::PARP-2 worms. Anti-HA antibody was used to highlight the PARP-2 signal at different stages in the gonads of wild-type control and FLAG::3xHA::PARP-2 worms. Strong HA-PARP-2 signals are observed in the nuclei of pre-meiotic tips, late pachytene to diplotene, and embryos. Left: Wild-type control. Right: FLAG::3xHA::PARP-2 worm. Bar = 10 μm. (B) Dynamic localization of PARP-2 in various nuclear stages from mitotic to meiotic prophase I and embryos. Bar = 10 μm. (C) Pre-meiotic tips and late-pachytene germline nuclei of PARP-2 co-stained with nucleolar protein 1 (NOP-1). Bar = 5 μm. (D) Line scan of relative fluorescence intensity of PARP-2, NOP-1, and DAPI in FLAG::3xHA::PARP-2 pre-meiotic tips and late pachytene nuclei. (E) PARP-2 signal induction upon treatment with 300 J/m² UV and 30 mM HU. Bar = 5 μm. (F) Quantification of signal intensity of panel E. Asterisks indicate statistical significance compared to the control group. P = 0.0004 for 0 and 300 J/m² UV at pre-meiotic tip (PMT), P < 0.0001 for 0 and 300 J/m² UV at pachytene, P < 0.0001 for 0 and 30 mM HU at PMT, and P < 0.0001 for 0 and 30 mM HU at pachytene, all with two-tailed Mann-Whitney test. (G) Quantification of mRNA expression for parp-2 following HU and UV treatments. Asterisks represent statistical significance with P = 0.0063,0.0446 from left to right, using a two-tailed t-test.

https://doi.org/10.1371/journal.pgen.1012062.g002

At late pachytene, a portion of the PARP-2 signal is centered, rather than localized to the peripheral chromatin area stained by DAPI. In contrast, the centered signal co-localizes with NOP-1, which marks the nucleolar region (Fig 2C). Consistent with this observation, a linescan analysis of the PARP-2 signal revealed a similar signal pattern across the DAPI-stained chromatin in the pre-meiotic tip (Fig 2D). However, in late pachytene nuclei, the PARP-2 signal pattern becomes synchronized with the nucleolus/NOP-1, rather than with the DAPI-stained chromatin, suggesting a partial re-localization of PARP-2 from chromatin to the nucleolus during pachytene.

Consistently, the PARP-2 signal is observed at diplotene and disappears at diakinesis, following the same localization pattern as the nucleolus (Fig 2B), which further supports the idea of PARP-2 co-localizing with the nucleolus.

To assess the functional relevance of this association, we quantified NOP-1 intensity across the -1 to -3 oocyte transition and monitor the intensity changes along the oocyte progression. NOP-1 fluorescence was markedly elevated in parp-2 mutants across −1 to −3 oocytes compared to wild type (S1 Fig, 2.14 vs 1.26 in -2; 2.00 vs 1.34 in -3 position, P = 0.0021 in wt and parp-2 at -2 position, two-tailed Mann–Whitney test). In wild-type animals, NOP-1 levels declined gradually from −3 to −1, whereas parp-2 mutants exhibited a sharp drop over the same interval. These data indicate that PARP-2 is essential for maintaining proper nucleolar homeostasis; its absence leads to aberrant accumulation of the nucleolar protein NOP-1 without a corresponding change in nucleolar size (parp-2/wt = 1.04, P = 0.39).

1. b. PARP-2 Signal upon UV or HU Treatment

We further analyzed whether PARP-2 localization is altered in response to DNA damage. In response to HU or UV, PARP-2 signal intensity was elevated compared to untreated worms, supporting its role in the replication stress response (Figs 2E and S2). Specifically, UVC exposure increased the average intensity of PARP-2 localization by 1.35-fold at the premeiotic tip (PMT) compared to untreated worms (Fig 2F, P = 0.0004). Similarly, UVC exposure increased the intensity by 1.23-fold at pachytene (P < 0.0001). HU exposure also increased PARP-2 intensity by 1.36-fold at PMT (P < 0.0001) and by 1.61-fold at pachytene (P < 0.0001), indicating that PARP-2 responds to HU-induced stress across multiple germline stages, including stages that are not undergoing bulk DNA replication.

Taken together, PARP-2 is initially enriched in mitotic nuclei at the premeiotic tip, diminishes during early meiotic stages, and reappears in late pachytene through diplotene, where it co-localizes with the nucleolar marker NOP-1. This dynamic localization pattern suggests regulated PARP-2 distribution during meiotic prophase. Upon exogenous replication stress, PARP-2 signal intensity increases significantly in both the premeiotic and pachytene regions. Additionally, PARP-2 transcript levels are upregulated under replication stress, further supporting its role in the germline replication stress response (Fig 2G).

Functional Analysis of PAR genes: Catalytic PARP-2 Is Essential for Development While PARG-1 Uniquely Regulates Meiotic X Chromosome Segregation

To determine whether PARylation-associated genes contribute to developmental or morphological processes, we assessed growth parameters, including brood size (the total number of fertilized eggs), embryonic survival, larval development, and the High Incidence of Males (HIM) phenotype, in the absence of exogenous stress. parp-2(ok344) mutants exhibited a 27.4% reduction in brood size compared to wild type (Fig 3A, 217 vs. 299, P < 0.0001).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. PARP-1, PARP-2, and PARG-1 differentially regulate fertility, embryogenesis, larval development, and X chromosome segregation in C. elegans.

(A) Brood-size measurements. Total eggs laid per hermaphrodite (n ≥ 15/genotype) revealed a marked reduction in parp-2(ok344) and parp-2(E509K) mutants versus WT, indicating that PARP-2 catalytic activity is essential for fertility. A comparable decrease was observed in parp-1(ok988) (P < 0.0001), whereas parg-1(gk120) showed no significant change (P = 0.1488). (B) Quantification of embryonic lethality. All mutant strains exhibited a significant rise in embryonic lethality relative to WT (P < 0.001), except the FLAG::3xHA::PARP-2 reporter line and E509K/WT heterozygotes. These data indicate that three PAR genes are essential for normal embryonic development. E509K mutants exhibited a 95-fold increase in embryonic lethality, highlighting the essential role of PARP-2’s catalytic activity in embryonic development. Embryonic lethality was scored based on the percentage of unhatched embryos among progeny (n ≥ 15 per genotype). (C) Larval arrest frequency was markedly elevated in parp-1, parg-1, and parp-2 mutants. Homozygous E509K mutants are sterile with 57.4% embryonic lethality, and the remaining larvae (~43%) arrest at L1-L3 and do not survive to adulthood. In contrast, the FLAG::3xHA::PARP-2 line exhibited normal development, supporting that the tagged protein retains functional activity. (n ≥ 15 per genotype). Right, nearly all arrested E509K larvae (~99.8%) halted at the L1 stage, with only 0.2% reaching L3. The arrested larvae were smaller but did not display dauer-, dumpy-, or skinny-like traits. Plate-tapping response assays revealed severely reduced motility in E509K larvae (99.6%, 99.3%, and 24.8% responsive in wild type, ok344, and E509K, respectively; P < 0.0001 in WT compared to E509K and also between ok344 and E509K). (D) High Incidence of Males (HIM) phenotype was significantly increased only in parg-1 mutants, suggesting a distinct role in X chromosome segregation during meiosis. Male frequency was calculated as the percentage of male offspring produced by self-fertilizing hermaphrodites (n ≥ 15 per genotype). All assays were conducted using hermaphrodites 24 hours post-L4 stage under non-stress conditions. Statistical significance was determined using two-tailed Mann–Whitney tests with a 95% confidence interval. Asterisks indicate significant differences relative to wild type (P < 0.05). (E) Germline length measurements in wild-type and mutants. Both parp-2 (ok344) and parg-1(gk120) mutants exhibited shorter pachytene stages and overall germline lengths. Data analyzed using two-tailed unpaired t-test.

https://doi.org/10.1371/journal.pgen.1012062.g003

The catalytically inactive parp-2 (E509K) mutant was used for further analyses. Homozygous F2 E509K/E509K worms derived from E509K/mIn1 heterozygous parents developed normally and reached adulthood, but their F3 progeny exhibited high embryonic lethality and larval arrest. Fertility and immunofluorescence measurements were performed using fertile second-generation F2 progeny, whereas developmental abnormalities of embryo lethality and larval arrest were measured for F3 progeny. The parp-2 (E509K) mutant also showed a significant reduction in brood size compared with wild type (233 vs. 299, P = 0.0002). Similarly, parp-1 mutants displayed a 22% reduction in brood size (Fig 3A, 232 vs. 299, P = 0.0371), while parg-1 mutants did not show a significant difference (262 vs. 299, P = 0.1488).

All mutants exhibited increased embryonic lethality relative to wildtype. Specifically, parp-1 mutants showed a 9.7-fold increase (Fig 3B, 5.8% vs. 0.6%, P < 0.0001), parg-1 a 5.2-fold increase (3.0% vs. 0.6%, P = 0.0017), and parp-2(ok344) an 8.0-fold increase (4.8% vs. 0.6%, P = 0.0002). In contrast, the FLAG::3xHA::PARP-2 reporter, knocked in at the endogenous parp-2 locus, showed no significant difference in embryo lethality (0.3% vs. 0.6%), indicating that the tagged protein retains full functionality.

Surprisingly, E509K catalytic-site mutation caused a 95-fold increase in embryonic lethality (57.4% vs. 0.6% in WT, P < 0.0001), establishing that PARP-2 enzymatic activity is indispensable for embryogenesis. This effect was ~ 10-fold stronger than the parp-2(ok344) null allele (P < 0.0001), suggesting that loss of catalytic activity in E509K may lead to aberrant accumulation of inactive protein, exacerbating the phenotype. Heterozygous E509K/ + embryos generated by crossing E509K homozygous hermaphrodites with wild-type males survived at near wild-type levels, indicating that E509K behaves as a recessive, catalytically inactive mutation (0.6% vs. 0.6%, P = 0.3389).

A similar pattern emerged in larval development. The ok344 null allele produced a modest 2.4-fold rise in larval arrest (Fig 3C, 1.7% vs. 0.7%, P = 0.0134), whereas E509K homozygotes exhibited complete arrest (100%, P < 0.0001). Introducing one WT allele (E509K/+) reduced arrest to 13%, reaffirming haplosufficiency and implicating the same catalytic-loss-plus-toxicity mechanism. Also, both parp-1 and parg-1 animals showed a 4.6-fold increase in larval arrest (Fig 3C, 3.1% vs. 0.7%, P < 0.0001). In contrast, FLAG::3xHA::PARP-2 animals exhibited no significant difference (1.4% vs. 0.7%, P = 0.0962).

To further define the nature of the larval arrest observed in parp-2 (E509K) mutants, we examined synchronized embryos and characterized the arrested larvae. Nearly all arrested progeny (~99.8%) halted development at the L1 stage, with only 0.2% reaching L3. The L1-arrested larvae were smaller but did not display dauer-, dumpy-, or skinny-like features (S3 Fig). Their motility was severely reduced, with only 24.8% responding to plate tapping compared with 99.6% and 99.3% in wild type and ok344, respectively (P < 0.0001 in WT compared to E509K and also between ok344 and E509K; Fig 3C). Measurement of body length at ~48 hours after egg laying confirmed that all parp-2 mutants were smaller than wild type (S3 Fig). These findings confirm that E509K homozygotes arrest predominantly at the L1 stage with impaired responsiveness and vitality.

Mis-segregation in X chromosome during meiosis causes an increased frequency of male offspring referred to as the Him (High Incidence of Males) phenotype, since a self-fertilizing XX hermaphrodite lays less than 0.2% XO males in normal conditions To assess meiotic chromosome segregation, we evaluated the Him phenotype, which results from X chromosome missegregation during meiosis. Under normal conditions, self-fertilizing XX hermaphrodites produce <0.2% XO males (Fig 3D, [26]). Neither parp-1 (0.13% vs. 0.03%, P = 0.3357) nor parp-2(ok344) (0.09% vs. 0.03%, P = 0.8626) mutants showed a significant increase in Him frequency, indicating that PARP-1 and PARP-2 are not essential for sex chromosome segregation. In contrast, parg-1 mutants exhibited a 100-fold increase in Him (3.0% vs. 0.03%, P < 0.0001), suggesting that loss of PARG-1–mediated dePARylation may disrupt faithful segregation of the X chromosome during meiosis, consistent with a previous report [23]. Consistent with ok344, E509K, E509K/wt and FLAG::3xHA::PARP-2 strains did not exhibit increased HIM. As noted above, the FLAG::3xHA::PARP-2 allele shows no obvious larval arrest or embryonic lethality, suggesting that the tagged protein largely retains its functionality.

Since PARP genes are required for proper fertility, we further investigated their roles during meiotic development. In C. elegans, germline nuclei are organized in a temporal and spatial manner, moving from mitosis through various stages of meiotic prophase I in a distal-to-proximal direction. Leveraging this well-defined organization, we quantified germline length and nuclear shapes in both mitotic (premeiotic tip) and meiotic nuclei (transition zone and pachytene) and compared the results between wild-type and mutant strains [27,28].

While deficiencies in these genes do not impact the mitotic or transition zones—where mitotic and meiotic progression overlap—they result in a shortened pachytene stage, where chromosomes exchange genetic material, as well as a reduced overall germline length (Fig 3E). Specifically, during pachytene, the lengths were 154.4 µm in parp-2 (ok344) vs. 200.7 µm in wild-type (P < 0.0001), and 181.4 µm in parg-1(gk120) vs. 200.7 µm in wild-type (P = 0.0318). For overall germline length, which includes the region from the distal tip to the late pachytene stage, measurements were 272.9 µm in parp-2 (ok344) vs. 321.4 µm in wild-type (P < 0.0001) and 297.7 µm in parg-1 (gk120) vs. 321.4 µm in wild-type (P = 0.0238). These findings suggest that PARP-2 and PARG-1 play significant roles specifically during the pachytene stage of meiosis, which is critical for genetic recombination and maintaining proper germline length. In contrast, parp-1 mutants did not show a significant difference in overall germline length compared to wild-type animals (333.2 µm vs. 321.4 µm, P = 0.1388), further highlighting the non-redundant roles of PARP-1 and PARP-2.

Taken together, our results establish that PARP-2’s catalytic activity is indispensable for normal development. The catalytically dead E509K allele causes far more severe embryonic lethality and complete larval arrest than the null mutant, revealing that loss of catalysis—augmented by the intrinsic toxicity of the mutant protein—underlies these extreme phenotypes. PARG-1, while dispensable for overall brood size, is required for faithful X-chromosome segregation and maintenance of pachytene length, pointing to a meiosis-specific function.

PARP-2 Is Required for Survival under Replication-Associated DNA Damage and Plays a Broad Role in Replication-Stress Response

PARP-2 levels are elevated in response to replication stress. To dissect its contribution to DNA-damage repair, we monitored sensitivity to various DNA damaging agents. To do this, adult C. elegans hermaphrodites were exposed to DNA damaging agents such as UV light, hydroxyurea (HU), HN₂, and cisplatin to examine how PARP-2 and other DNA repair components respond

UV light induces DNA damage primarily through the formation of pyrimidine dimers, which distort the DNA structure and interfere with replication and transcription. HU is a ribonucleotide reductase inhibitor that blocks DNA synthesis by preventing the expansion of the dNTP pool. This exposure led to replication-related DNA damage. HN₂ is a type of alkylating agent which leads to cross-linking of the DNA strands, which prevents the DNA from being properly replicated. Cisplatin is another chemotherapeutic drug and it works similarly to HN₂ in that it also forms cross-links in DNA. Cisplatin binds to purine bases in DNA and forms intrastrand and interstrand cross-links, with the intrastrand link between adjacent guanines being the most common.

Exposure to UV light significantly reduced hatching levels in all three mutants compared to the wild type. Specifically, at a dose of 300 J/m², the parp-2, parp-1, and parg-1 mutants displayed reductions in hatching of 56.2%, 48.9%, and 58.2%, respectively, compared to the wild type (Fig 4A). These changes are statistically significant, with P = 0.0211 for parp-2, 0.0073 for parp-1, and 0.0257 for parg-1. These results suggest that they are required for UVC-induced DNA damage repair, which is consistent with reports on other organisms [29–31].

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. PARP-2 is critical for embryonic viability under replication stress.

Synchronized embryos from indicated genotypes were exposed to (A) UV, (B) hydroxyurea (HU), (C) nitrogen mustard (HN₂), or (D) cisplatin, and hatching rates were scored. (A) UV hypersensitivity was evident in parp-1(ok988), parp-2(ok344), and parg-1(gk120) mutants. (B) HU treatment caused a severe hatching defect in parp-2(ok344) (34.3%) compared with WT (91.7%), whereas parp-1(ok988) showed a milder reduction (78.4%) and parg-1(gk120) was unaffected. (C, D) Both parp-2(ok344) and catalytically inactive parp-2(E509K) displayed pronounced sensitivity to HN₂ (C; ok344 vs WT P = 0.0004, E509K vs WT P = 0.0003) and cisplatin (D; ok344 vs WT P < 0.0001, E509K vs WT P < 0.0060). For E509K, E509K/mIn1 (F1) animals were treated with the indicated drugs, and embryonic viability was quantified in their F2 E509K/E509K progeny. Because F3 progeny of E509K mutants exhibit high baseline embryonic lethality and larval arrest, drug-induced survival was not assessed in the F3 generation. Data are mean ± SEM from ≥3 independent experiments; significance by two-tailed Mann-Whitney test.

https://doi.org/10.1371/journal.pgen.1012062.g004

Exposure to HU, resulted in significant alterations in hatching rates among parp-2 and parp-1 mutants relative to wild-type counterparts. When treated with 30 mM HU, parp-2 mutants showed a dramatically reduced hatching rate of 34.3% (P < 0.0001), while parp-1 mutants exhibited a moderate decrease to 78.4% (P = 0.023, Fig 4B). In contrast, hatching rates in parg-1 mutants did not significantly differ from the wild type across three HU doses, maintaining a rate of 86.8% (P = 0.3115 at 30 mM). Taken together, the hatching (%) in parp-2 mutants decreased by 57.4% (from 91.7 to 34.3%) compared to the wild type — a substantial drop. Meanwhile, parp-1 mutants experienced a milder decline of 13.3% (from 91.7 to 78.4%), underscoring a more significant role for PARP-2 in the management of replication fork stress than PARP-1.

In addition to the replication stress induced by HU, we next examined whether PARP-2 mutants respond to DNA cross-linking agents, such as nitrogen mustard (HN₂) and cisplatin, which also induce DNA replication stress. Similar to HU-induced stress, parp-2 (ok344) mutants exhibited increased sensitivity compared to wild type at 100 and 150 µM of nitrogen mustard (Fig 4C, 22.4 vs. 43.8 at 150 µM, P = 0.0004). Likewise, E509K mutants also showed significantly reduced survival compared to wild type, as assessed in F2 homozygous progeny following drug treatment of heterozygous parents (23.9 vs. 43.8, P = 0.0003). Consistently, upon Cisplatin exposure, ok344 mutants displayed increased sensitivity at 200 µg/mL (Fig 4D, 46.2 vs. 58.4, P < 0.0001), and E509K mutants showed a similar response (38.9 vs. 58.4, P < 0.0060). Thus, both the ok344 and E509K alleles confer comparable sensitivity to DNA damage induced by HU and Cisplatin. In contrast, parp-1 and parg-1 mutants did not exhibit significant differences compared with wild type. Under HN₂ treatment (Fig 4C), parp-1 mutants showed 41.34% survival (P = 0.606), and parg-1 mutants showed 44.5% survival (P = 0.842). Under cisplatin treatment (Fig 4D), parp-1 mutants exhibited 55.75% survival (P = 0.56), and parg-1 mutants exhibited 58.13% survival (P = 0.95), none of which differ significantly from wild type.

In summary, all three PARylation-associated genes exhibited UVC sensitivity. parp-2 mutants were sensitive to HU-, HN₂-, and Cisplatin-induced DNA damage, suggesting that parp-2 is required for responding to replication stress.

PARP-2 attenuates spontaneous CHK1/ATR activation and basal replication-stress signaling

To elucidate the downstream checkpoint response triggered by these lesions, we examined CHK1 activation. When DNA is damaged or replication stress is sensed by the cell, CHK1 is activated through phosphorylation by the upstream kinases ATM and ATR. Phosphorylated CHK1 (pCHK1) is the active form of the enzyme and plays a central role in the DNA damage response by initiating a cascade of events that halt the cell cycle.

We observed activation of DNA damage checkpoints in parp-1, parp-2, and parg-1 mutants even in the absence of exogenous DNA damage, as evidenced by significantly higher levels of pCHK1 foci compared to wild type (Fig 5A and 5B; 1.45 vs. 0.42 in parp-2, P = 0.0007; 1.32 vs. 0.42 in parp-1, P = 0.0040; 1.34 vs. 0.42 in parg-1, P = 0.0453).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. parp-2 Mutants Exhibit Elevated S-Phase Checkpoint Activation Under Replication Stress, While PARP Inhibitors Mimic the Absence of PARP Expression in C. elegans.

(A) Representative immunofluorescence images showing phosphorylated CHK-1 (pCHK1) foci in the germline of wild-type, parp-1, parp-2, and parg-1 mutants under untreated conditions. ATL-1 foci are also accumulated in parp-2 (ok344) mutants compared to wild type. Scale bar = 2 μm. (B) Quantification of pCHK1 foci reveals significantly increased DNA damage checkpoint activation in parp-1 (1.32 ± 0.13), parp-2 (1.45 ± 0.11), and parg-1 (1.34 ± 0.16) mutants compared to wild type (0.42 ± 0.09) (P = 0.0040, 0.0007, and 0.0453, respectively). Right, expression levels of atl-1 and atm-1 are also significantly upregulated in parp-2 (ok344) mutants relative to the control group (P = 0.0020 for atl-1, P = 0.0037 for atm-1). (C) Quantification of the number of germ cell nuclei in the premeiotic tip after 8 mM hydroxyurea (HU) treatment. Both wild-type and parp-2(ok344) mutants show a significant reduction in nuclei number upon HU exposure (P = 0.0020 and 0.0238, respectively), indicating proper activation of the S-phase checkpoint. E509K mutants also show reduced nuclei upon HU treatment (P < 0.0160), but consistently exhibit fewer nuclei than ok344 mutants under both treated and untreated conditions (P < 0.0001), suggesting a distinct difference in checkpoint activation dynamics. All data represent mean ± SEM. Scale bar = 10 μm. Statistical significance was determined using unpaired t-tests (D) qPCR analysis of parp-2 mRNA expression following treatment with human PARP inhibitors PJ34 and Olaparib. Both inhibitors significantly reduced parp-2 expression, with PJ34 causing an 81% reduction (P = 0.0067) and Olaparib resulting in a 78% reduction (P = 0.0125) compared to untreated controls. Data represent mean ± SEM. (E) Immunofluorescence analysis of CRISPR-generated FLAG::3xHA::PARP-2 worms at the pachytene stage of germline development. Treatment with 100 μM PJ34 or Olaparib caused a significant reduction in HA-PARP-2 signal intensity. The intensity was reduced by 51% (P = 0.0003) and 57% (P < 0.0001) for PJ34 and Olaparib, respectively, compared to untreated controls. No significant change was detected in the premeiotic tip (PMT), consistent with stage-specific regulation of PARP-2 nuclear localization. (F) pCHK-1 foci count as a measure of checkpoint activation in response to PARP inhibition. Treatment with 100 μM PJ34 or Olaparib significantly increased the number of pCHK-1 foci per nucleus. The mean number of foci was 2.60 and 2.41 for 100 μM PJ34 and Olaparib, respectively, compared to 0.81 in the control (1% DMSO). For 1 mM PJ34 and Olaparib, the mean number of foci increased to 3.08 and 2.77, respectively, compared to 1.88 in the control (10% DMSO).

https://doi.org/10.1371/journal.pgen.1012062.g005

Consistent with these findings, the upstream kinase ATL-1 (the C. elegans ATR ortholog)—which is recruited to stalled replication forks in the mitotic zone of the gonad—accumulates in parp-2(ok344) animals (Fig 5A). Moreover, transcript levels of both atl-1 and atm-1 are significantly elevated in parp-2(ok344) mutants (Fig 5B), supporting the idea that PARP-2 is required to resolve endogenous replication stress.

The replication-dependent S-phase checkpoint, activated in response to replication stress, induces S-phase arrest. This arrest is characterized by an enlarged nuclear diameter and a reduced number of nuclei in the premeiotic tip region [32,33]. Upon exposure to hydroxyurea (HU), both wild-type and ok344 mutants exhibited a significant reduction in the number of nuclei in the premeiotic tip, indicating that the absence of PARP-2 does not impair checkpoint activation (Fig 5C, P = 0.0020 for WT vs. WT + HU; P = 0.0238 for ok344 vs. ok344 + HU). This suggests that PARP-2 is not essential for S-phase arrest. Similarly, the catalytically inactive mutant E509K showed a comparable reduction following HU treatment (P < 0.0160), suggesting that enzymatic activity may not be required for checkpoint engagement. Interestingly, however, E509K mutants displayed significantly fewer nuclei than ok344 mutants under both untreated and HU-treated conditions (P < 0.0001), pointing to a distinct difference in checkpoint regulation between the two strains.

Human PARP Inhibitors Revealed Functional Conservation of PARP-1 and PARP-2 for DNA Damage Response in C. elegans

To further explore the Functional Conservation and the potential of C. elegans as a model for screening PARP inhibitors, we investigated the effects of existing human PARP inhibitors, namely PJ34 and Olaparib and, on wild type worms. Our qPCR analysis revealed that both inhibitors efficiently reduced the expression of parp-2 mRNA: PJ34 led to an 81% reduction (1.0 vs 0.19, P = 0.0067) and Olaparib led to a 78% reduction in parp-2 mRNA expression (1.0 vs 0.22, P = 0.0125, Fig 5D).

To further validate these findings, we conducted immunofluorescence analysis using reporter worms expressing FLAG::3xHA::PARP-2. We observed a significant reduction in PARP-2 signal intensity in pachytene-stage germline nuclei when worms were treated with 100 μM PJ34 or olaparib. Specifically, PARP-2 signal intensity was reduced by 51% and 57% for PJ34 (Fig 5E, P = 0.0003) and Olaparib(P < 0.0001), respectively, compared to untreated controls, indicating effective inhibition of PARP-2 in pachytene. Notably, no difference was detected in the premeiotic tip (PMT), suggesting stage-specific regulation of PARP-2 nuclear localization.

Moreover, in line with previous findings showing parp-2 null mutants exhibit more active cell cycle checkpoints, we assessed the functional impact of PJ34 and Olaparib by quantifying pCHK-1 foci as a measure of checkpoint activation. Both inhibitors significantly increased the number of pCHK-1 foci, with treatments at 100 μM and 1 mM showing similar results (Fig 5F). Specifically, the mean number of pCHK-1 foci per nucleus was 2.60 and 2.41 for 100 μM PJ34 and Olaparib and 3.08 and 2.77 for 1mM PJ34 and Olaparib, respectively, compared to 0.81 in the control and 1.88 in the control indicating robust activation of the DNA damage checkpoint in response to human PARP inhibitor treatment.

These findings mirror the decrease in PARP-2 mRNA and the loss of PARP-2 signal observed upon mammalian PARP-inhibitor treatment, indicating that the DNA-damage response is regulated similarly in C. elegans and mammals. Thus, the mechanisms governing PARP activity are conserved, validating C. elegans as a robust model for PARP-inhibitor screening.

PARP-2 Deficiency Disrupts Double-Strand Break Repair and Activates the Fanconi Anemia Pathway, Leading to Elevated RAD-51 and FCD-2 Foci and Increased Apoptosis in Germline Cells

RAD-51 plays a critical role in double-strand break repair (DSBR), as it protects stalled replication forks and is a central factor in strand exchange during homologous recombination (HR) [34] Given that parp-2 mutants exhibit sensitivity to replication-associated stress, including HU, HN2, and Cisplatin, we aimed to determine whether PARP-2 expression is essential for DSBR by assessing the levels of RAD-51 foci. For this, we observed and compared the whole germline nuclei of wild-type and parp-2 (ok344) mutant worms.

In wild-type animals, RAD-51 foci began to appear at the premeiotic tips and reached their peak levels in mid-pachytene (Fig 6A). Subsequently, the foci gradually declined from mid- to late pachytene and disappeared in diplotene [32]. In contrast, both E509K and ok344 mutants exhibited elevated levels of RAD-51 expression compared with wild-type. Specifically, the mean number of RAD-51 foci per nucleus in the premeiotic tips was higher in the E509K (0.15) and ok344 (0.29) mutants than in wild-type (0.04). This increase persisted during meiotic prophase, with foci numbers of 1.71 and 2.04 (vs. 1.4 in wild-type) in early pachytene, and 4.8 and 5.21 (vs. 4.13 in wild-type) in mid-pachytene.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. PARP-2 functionally intersects with FCD-2 during replication stress, potentially contributing to ICL repair. (A) RAD-51 foci analysis across different meiotic stages in wild-type and parp-2 mutants. RAD-51 foci, marking double-strand break (DSB) sites, appear at premeiotic tips and reach their peak in mid-pachytene in wild-type germline nuclei. In contrast, both E509K and ok344 mutants exhibit a marked increase in RAD-51 foci from premeiotic to late pachytene stages, indicating persistent DSBs. The temporal appearance and heightened accumulation of RAD-51 foci in mutants suggest a defect in the resolution of DSBs in the germline. (B) Quantification of apoptotic nuclei by acridine orange staining of age-matched adults (20 h post-L4). Apoptotic nuclei count in the pachytene region of the germline are elevated in parp-2 (ok344), E509K and parg-1 mutants, consistent with enhanced checkpoint activation and unresolved replication-associated DNA damage. (C) FCD-2, a key component of the Fanconi Anemia (FA) pathway, accumulates significantly in parp-2 mutants compared to wild type. This supports the notion that the FA pathway is activated in response to replication stress, particularly in the absence of PARP-2, leading to increased recruitment of FCD-2 foci during meiosis. Bar = 10 μm. (D) Representative images showing immunofluorescence signals for FCD-2 and PARP-2 in the pre-meiotic tip and pachytene region. Arrows indicate co-localized foci. Right, quantitative co-localization of FCD-2 and PARP-2 along the germline. (n = 172, 252, and 4673 FCD-2 foci quantified, respectively). Error bars indicate SDs.

https://doi.org/10.1371/journal.pgen.1012062.g006

RAD-51 foci levels also remained elevated in late pachytene (0.85 and 1.11 vs. 0.67 in wild-type). However, by diplotene, the differences were no longer evident. These data indicate that both the E509K and ok344 alleles of parp-2 result in increased accumulation of RAD-51 foci in mitotic and meiotic germline nuclei, suggesting impaired resolution of double-strand breaks (DSBs) when PARP-2 catalytic activity is lost.

We speculate that defective replication repair may result in increased apoptosis when DNA lesions are not effectively resolved. Indeed, parp-2 and parg-1 mutants exhibited significantly elevated levels of apoptosis compared to the wild-type control, supporting the notion that impaired replication repair activates the DNA damage checkpoint and can ultimately lead to cell death (Fig 6B). The average number of apoptotic nuclei was 3.87 in wild type, 6.09 in parp-2 (P = 0.0014), 4.48 in parp-1 (P = 0.6207), 6.47 in E509K (P < 0.0001), and 7.71 in parg-1 (P < 0.0001).

Since parp-2 mutants are sensitive to interstrand crosslink (ICL) damage (HN₂ and Cisplatin) and accumulate DSB intermediates, we next investigated the Fanconi anemia (FA) pathway, which is upstream of DSB repair [35]. FCD-2, a homolog of mammalian FANCD2, is a key component of the FA pathway. In PMT, we observed significantly higher levels of FCD-2 foci in parp-2(ok344) mutants compared to wild-type animals (2.36 vs. 18.39 foci/nucleus, P = 0.0001, Fig 6C), with similar induction also seen in the E509K mutant (2.36 vs. 10.71 foci/nucleus, P < 0.0001). Consistently, FCD-2 foci levels were elevated in the pachytene stage of both mutants (6.82 vs. 28.68 foci/nucleus, P < 0.0001 in wt and ok344; 6.82 vs. 17.72 foci/nucleus, P < 0.0001 in wt and E509K).

The concomitant rise of FCD-2 and RAD-51 foci in parp-2 mutants indicates that PARP-2 loss stalls resolution of replication-induced damage. To test whether PARP-2 and FCD-2 act jointly during replication stress, we examined their co-localization. Partial co-localization was evident from PMT to pachytene stage (Fig 6D): 63% of FCD-2 foci co-localized with PARP-2 in the PMT, 51% in the transition zone, and 43% in the pachytene region.

Altogether, PARP-2 is indispensable for the efficient resolution of replication-associated DNA lesions in both mitotic and meiotic germline nuclei. Its loss stalls repair, triggering compensatory recruitment of the Fanconi anemia core component FCD-2 and the strand-exchange protein RAD-51; the resulting accumulated foci peak at pachytene and elicit a DNA-damage-checkpoint-mediated apoptotic response. Spatiotemporal co-localization of PARP-2 and FCD-2 throughout the germline suggests that PARP-2 may act in concert with components of the FA pathway during replication stress. While our data indicate partial overlap between these factors, further studies will be needed to determine whether this cooperation is specific to ICL repair or reflects a broader interaction with multiple DNA repair pathways.

Catalytic Inactivity of PARP-2 Triggers Broad DNA Damage Responses Not Seen in Null Mutants

In mammalian systems, PARP-1 and PARP-2 have been shown to associate with promoter-proximal chromatin and to regulate transcription primarily through chromatin remodeling or recruitment of transcriptional machinery, rather than via well-defined sequence-specific DNA recognition. However, it remains unclear whether PARP-1/2 possess any intrinsic sequence preference, as this aspect has not been systematically investigated [36–38]. To better understand the molecular basis of the observed phenotypes, we performed chromatin immunoprecipitation followed by ChIP-seq to map genome-wide PARP-2 binding sites. We found that PARP-2 was enriched at transcription start sites (TSSs) and enhancer-proximal regions, suggesting a potential role in transcriptional regulation. A metagene profile centered around TSSs revealed a sharp enrichment within ±100 bp, indicating promoter-proximal binding (Fig 7A). Furthermore, the broad distribution of peaks extending up to ±1–2 kb from the TSS suggests that PARP-2 may play a role as a chromatin modifier involved in processes such as nucleosome remodeling, chromatin loosening, or the formation of enhancer-promoter loops. Notably, this enrichment was specific to PARP-2 IP samples and absent in input controls, ruling out general chromatin accessibility as the cause.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 7. PARP-2 chromatin binding, motif preference, and transcriptomic changes in parp-2 mutants.

(A) Metagene profile of PARP-2 ChIP-seq signal relative to transcription start sites (TSS) across the genome. The signal shows modest enrichment near TSS (±100 bp), suggesting PARP-2 association with promoter-proximal regions. Bottom: Motif enrichment analysis of ChIP-seq peaks identifies significant overrepresentation of GAGA-repeat and Trl(Zf)/S2-GAGA factor motifs, both associated with chromatin architecture in Drosophila. (B) Differential gene expression profiles of parp-2(ok344) and parp-2(E509K) mutants. RNA-seq analysis revealed that both mutants altered the expression of genes involved in DNA damage response, reproduction, metabolic processes, and cellular processes. Both mutants affected similar functional pathways, the ratio of upregulated to downregulated genes within these pathways differed substantially between the two genotypes. (C) Differentially expressed gene (DEG) clustering reveals downregulation in parp-2 (ok344) mutants and upregulation in parp-2(E509K) compared to wild type. These patterns are consistent with parp-2 expression levels confirmed by qPCR analysis, which showed reduced parp-2 expression in ok344 and increased expression in E509K. parp-1 is also upregulated in both mutants, with stronger induction in E509K. Right, RNA-seq analysis reveals partial transcriptional overlap between E509K mutation and ok344 treatment. Venn diagrams show numbers of downregulated and upregulated genes unique to each condition and shared. (D) KEGG pathway enrichment reveals shared pathways in both mutants, but DNA damage repair pathways are significantly enriched only in E509K. (E) Expression analysis of selected DNA repair pathways shows stronger upregulation in E509K than in ok344, suggesting a more robust DNA damage response in the catalytic dead mutant. The y-axis represents the average expression of genes within each indicated DNA damage repair pathway. Numbers correspond to the average expression level for genes comprising indicated pathways.

https://doi.org/10.1371/journal.pgen.1012062.g007

To explore whether PARP-2 targets specific DNA sequences, we analyzed known and de novo motif enrichment within ChIP-seq peaks. Surprisingly, Among the top enriched motifs were GAGA repeats and Trl(Zf)/GAGA factor binding sites, both of which are associated with chromatin architecture and remodeling in Drosophila (Fig 7A).

Although the fold enrichment of GAGA motifs within PARP-2 binding sites was moderate (~1.4-fold), the association was statistically robust (q < 0.001), suggesting a non-random and potentially functional involvement of GAGA-related elements in PARP-2 recruitment. This unexpected enrichment suggests that PARP-2 recruitment may involve recognition of GAGA-associated chromatin features.

Also, interestingly, some peaks overlapped with promoters of key DNA repair genes such as rad-51 and pif-1, indicating that PARP-2 may directly regulate the expression of genes involved in genome maintenance. These findings, together with PARP-2’s partial nucleolar localization, support a model in which PARP-2 modulates stress-responsive gene expression programs.

Building on these findings, we conducted transcriptome profiling using RNA sequencing (RNA-seq) to determine whether PARP-2 binding correlates with gene expression changes in the DNA damage response pathways. Gene Ontology (GO) analysis categorizes genes based on their biological processes, molecular functions, and cellular components. Consistent with our results, both mutants affected the expression of genes associated with the DNA damage response, reproduction, metabolic processes, and cellular processes. These findings suggest that both mutants are involved in similar pathways (Fig 7B); however, they exhibited a strong difference in the ratio of upregulated to downregulated genes within these pathways.

Overlap analysis using Venn diagrams (S4 Fig) comparing wild-type and parp-2(ok344) mutants revealed that genes down-regulated in ok344 are more numerous and significantly enriched for PARP-2 promoter binding than up-regulated genes. This indicates that PARP-2 functions primarily as a transcriptional activator, although a subset of targets shows upregulation, suggesting potential repressor roles at specific loci.

To further validate these observations, we performed a Differentially Expressed Genes (DEG) cluster analysis comparing both ok344 and E509K mutants to wild type (Fig 7C). This analysis shows that the ok344 mutant exhibits more downregulated genes (blue bars), while the E509K mutant shows more upregulated genes (red bars). We hypothesize that the observed downregulation in the parp-2(ok344) mutant is due to the absence of parp-2 expression, leading to suppression of associated pathways. In contrast, the E509K mutant, despite lacking catalytic enzymatic activity, still expresses PARP-2, which does not result in similar global suppression.

Consistent with this idea, parp-2(ok344) mutants exhibit a reduced level of parp-2 expression by 0.34-fold (Fig 7C, FDR = 3.29 × 10^-41), while E509K shows a 2.4-fold increase (FDR = 7.71 × 10^-139) in its expression. Furthermore, the ok344 mutant also showed 2.4- and 1.2-fold upregulation of PARP-1 and PARG-1, respectively, while the E509K mutant exhibited even stronger 4.8- and 2.3-fold upregulation. This suggests that the mere absence of PARP-2 catalytic activity (as observed in E509K), even in the presence of the protein, triggers the induction of certain genes, potentially as a compensatory mechanism or in response to un-PARylated targets.

RNA-seq analysis revealed substantial but incomplete overlap between transcriptional changes induced by ok344 and the E509K mutation. Among downregulated genes, 3,687 and 3,328 genes were downregulated in E509K and ok344, respectively, with 2,489 genes shared between the two conditions. For upregulated genes, 2,930 genes were upregulated in E509K and 1,689 in ok344, with 441 genes commonly upregulated.

We further analyzed the pathways using KEGG mapping to validate the roles of PARP-2. Both ok344 and E509K mutants highlighted common pathways, including neuroactive ligand-receptor interaction, calcium signaling, and purine metabolism, all with p-values < 0.05 (Fig 7D). However, while no DNA damage repair pathways were enriched in the ok344 mutants, the E509K mutants showed significant enrichment of multiple DNA damage repair pathways, including the Fanconi anemia pathway, homologous recombination, DNA replication, nucleotide excision repair, mismatch repair, base excision repair, and non-homologous end-joining.

Upon closer inspection of individual gene expression, we observed a significant upregulation of genes in seven DNA damage repair pathways in both ok344 and E509K mutants compared to wild-type controls. For example, the average gene expression of the Fanconi anemia (FA) pathway was 1.5-fold higher in ok344 and 2.9-fold higher in E509K compared to controls (Fig 7E). Similarly, the average gene expression of Base excision repair (BER) was 1.7-fold higher in ok344 and 2.9-fold higher in E509K compared to controls.

These results suggest a similar activation of DNA damage repair pathways in both mutants. However, E509K mutants exhibited stronger induction across all pathways compared to ok344 mutants. For instance, the FA pathway showed the highest relative increase (ratio) in E509K mutants (1.93-fold increase, Table 1). Likewise, nucleotide excision repair (NER) and homologous recombination (HR) showed a 1.86-fold and 1.80-fold increase, respectively, in E509K mutants. These findings suggest that the E509K mutant triggers a broadly enhanced DNA damage response, particularly through pathways involved in interstrand crosslink repair and excision-based repair mechanisms. This was further validated by qPCR analysis (S5 Fig).

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. DNA Repair Pathway Induction in E509K vs ok344 Mutants in transcriptomic analysis. Average fold change(FC) in the DNA damage repair pathway is indicated.

https://doi.org/10.1371/journal.pgen.1012062.t001

By contrast, nucleolar and RNA-synthesis genes followed two distinct patterns of PARP-2 dependency, underscoring a fundamentally different regulatory logic from DNA repair pathways. Genes such as cpb-3, iff-2 and rpl-11.1 were regulated independently of catalytic activity, pointing to a scaffolding or structural function of PARP-2 protein in ribosomal networks (Fig 7E and S1 Table). Conversely, other genes including c43e11.9, nol-16, and iff-1 required PARP-2 catalysis for repression, as they were up-regulated only in the null mutant but not in E509K. Notably, none of the nucleolar genes showed the strong up-regulation observed in DNA-damage-responsive genes, suggesting that catalytically inactive PARP-2 has a limited impact on ribosomal gene expression. However, these findings are preliminary and do not provide direct evidence for a role of PARP-2 in ribosome biogenesis.

Taken together, our data demonstrate that both ok344 and E509K mutants exhibit significant upregulation of multiple DNA damage response and repair pathways, including Fanconi anemia, homologous recombination, and nucleotide excision repair, highlighting a broad role for PARP-2 in genome maintenance. Notably, the magnitude of induction was consistently greater in E509K than in ok344, indicating that catalytic inactivation has a more profound impact on DNA damage signaling than complete protein loss.

In sharp contrast, while DNA damage repair pathways are strongly dependent on the catalytic state of PARP-2, nucleolar and ribosomal gene regulation can proceed through either catalytic or scaffolding functions, underscoring a context-dependent mode of regulation.

PARP-2 links germline genome maintenance to replication-coupled DNA repair

Our findings identify PARP-2 as a pivotal regulator of genome integrity during germline development. PARP-2 localizes dynamically across mitotic and meiotic stages, re-localizing from chromatin to the nucleolus during meiotic prophase.

This re-localization is further enhanced by replication stress and likely reflects a role in nucleolar signaling rather than direct rDNA transcriptional regulation, as previous mammalian studies distinguish PARP-2 from the nucleolar transcriptional functions ascribed to mammalian PARP-1 [37,39].

Given that the nucleolus functions as a genome surveillance hub [40], our results suggest that PARP-2 may participate in nucleolar responses to replication stress, potentially contributing to the stability of rDNA and other fragile genomic loci. However, these observations are preliminary and do not provide direct evidence for a role of PARP-2 in ribosome biogenesis.

Loss of PARP-2 leads to embryonic lethality, larval arrest, and hypersensitivity to replication-blocking or crosslinking agents. These phenotypes are accompanied by RAD-51 and FCD-2 foci accumulation, implicating PARP-2 in replication-coupled DNA repair. The observed co-localization with FCD-2 is consistent with, but not exclusive to, potential engagement with the FA pathway. Together, these data support a model in which PARP-2 safeguards replication fork stability and orchestrates DNA repair under stress conditions.

Catalytic activity as a regulatory brake on DNA damage responses

The catalytic-dead E509K mutant, analogous to the E988K mutation in human PARP1, reveals an unexpected regulatory dimension of PARP-2. Absence of catalytic activity in the E509K mutant results in phenotypes that are comparable to or even more severe than those observed in the null allele, including reduced brood size, high embryonic lethality, and complete larval arrest. This suggests that the enzymatic function of PARP-2 is essential for meiosis and early development. In addition, E509K animals show hypersensitivity to interstrand crosslinking agents and increased germline apoptosis, further underscoring the indispensable role of catalytic activity in DNA repair and stress response.

Transcriptomic profiling (Fig 7) revealed that E509K mutants show altered activation patterns of DNA damage repair pathways, including FA, HR, NER, and BER. Although some DDR genes appear upregulated, this likely reflects compensatory activation due to impaired repair efficiency rather than a more effective DNA damage response. Thus, the data suggest that loss of catalytic activity disrupts normal DDR regulation, leading to uncoordinated or excessive pathway activation rather than enhanced repair capability. Importantly, mammalian studies indicate that PARP2 can promote HR and alternative end-joining by restraining 53 BP1 accumulation independent of PAR synthesis [41], suggesting that loss of catalytic activity may disrupt DNA repair pathway choice in addition to impairing repair efficiency. Such disruption could explain why uncontrolled upregulation of multiple DNA damage pathways is observed in the E509K catalytic-dead mutation.

Alternatively, the observation that the E509K mutation leads to more pronounced defects than the null allele suggests that catalytically inactive PARP-2 may interfere with normal repair processes—possibly by sequestering DNA substrates or protein partners—thereby intensifying the cellular response to replication stress. Catalytic-dead mutants may disrupt the balance of PARylation signaling or compete with catalytically active PARPs, such as PARP-1, for chromatin binding, ultimately leading to aberrant transcriptional and repair outcomes. This highlights a critical distinction between loss of protein function and loss of catalytic function, with implications for interpreting cancer-associated mutations of PARP family members [42]. Future work should also investigate the complex interplay between PARP-2 and PARG-1, particularly by assessing genetic interactions and the potential suppressive effects of combinatorial drug treatments.

This highlights a critical distinction between loss of protein function and loss of catalytic activity, underscoring that PARP-2 possesses unique regulatory roles beyond PAR synthesis. Notably, dysregulation of PARP-2 targets has been linked to diverse cancers [42], suggesting that cancer-associated mutations should not be viewed simply as generic loss-of-function events but rather in the broader context of PARP-2–specific contributions to genome stability, chromatin regulation, and tumorigenesis—insights that carry direct implications for isoform-selective therapeutic strategies.

This model contrasts sharply with the behavior of PARG-1, whose catalytic-dead mutants maintain meiotic DNA repair through scaffolding roles [23]. Thus, PARP-2 represents a unique case in which catalytic activity is not redundant but actively restrains overactivation of DNA damage repair pathways. To more precisely resolve the sequestration and hyperactivation models, future studies are warranted to test the genetic interaction with parg-1 deletion and to assess the suppressive effects of clinical PARP inhibitors on the E509K phenotype. Consistent with checkpoint engagement, PARP inhibitor treatment reduced parp-2 mRNA and PARP-2 signals and increased pCHK-1 foci. However, embryonic lethality and larval arrest were not assessed under these conditions, and thus the developmental consequences of pharmacological inhibition cannot be directly equated with those observed in the catalytic-dead mutant.

Chromatin surveillance and functional specialization of PARP-2

ChIP-seq analyses reveal that PARP-2 preferentially occupies transcription start sites (±1–2 kb), enriched at GAGA motif-containing loci. These regions correspond to transcriptionally active and structurally vulnerable chromatin domains, suggesting that PARP-2 functions as a chromatin-associated co-regulator. The enrichment of PARP-2 at these loci even under basal conditions—absence of exogenous DNA damage—implies a constitutive role in maintaining chromatin architecture, possibly through general modulation of chromatin structure. In light of recent findings that mammalian PARP2 contributes to telomere stability and fragile site protection under replication stress [43], our data suggest a broader chromatin surveillance function that extends beyond canonical DNA repair.

Importantly, PARP-2 performs functionally non-redundant roles within the PARP network. First, although PARP-1 is significantly up-regulated in parp-2 null mutants (Fig 7C) and PARP-2 signals are reciprocally elevated in parp-1 mutants (S6 Fig), this localized, partial compensation is restricted to specific germline stages and fails to rescue the respective phenotypes, demonstrating that the two genes are not interchangeable. Second, E509K catalytic-dead mutants, which over-express both PARP-1 and PARG-1, display more severe defects than other PAR-related null alleles, underscoring the essential, non-redundant contribution of PARP-2 catalytic activity. Interestingly, this dependency is not uniform across biological processes: DNA damage repair pathways are highly sensitive to the catalytic state of PARP-2, whereas nucleolar and ribosomal gene expression may be supported through non-catalytic scaffolding functions. This dichotomy highlights a pathway-specific specialization of PARP-2 that differentiates genome repair from transcriptional homeostasis.

Collectively, these findings demonstrate that PARP-1 and PARP-2 share overlapping yet distinct, spatially restricted functions, establishing PARP-2 as a unique, non-redundant guardian of germline genome integrity.

While mammalian systems feature partially redundant functions among PARP1 and PARP2, C. elegans appears to allocate distinct roles to each component. PARP1 primarily responds to single-strand breaks and nucleolar transcriptional regulation, whereas PARP2 engages with replication-associated DNA gaps, recombination intermediates, and fragile chromatin domains [8–11,37]. The phenotypic severity observed in the E509K mutant, not recapitulated by loss of PARP-1 or PARG-1 alone, further supports the idea of specialized, non-redundant functions within the PAR network in nematodes. Although non-catalytic roles are plausible, particularly in light of PARG-1’s partial activity in meiotic repair [23], our data indicate that for PARP-2, catalysis remains the dominant determinant of function.

A unified model for PARP-2 in genome stability

We propose a model in which PARP-2 integrates catalytic and structural functions to preserve genome stability. Under replication stress, catalytic PARylation promotes fork stabilization and DNA repair while simultaneously constraining excessive DNA damage repair pathways (Fig 8). In parallel, PARP-2’s chromatin association enables surveillance of transcriptionally active and structurally fragile loci, thereby preventing their destabilization. Our results further suggest that this dual role operates in a context-dependent manner: DNA repair requires PARP-2 enzymatic activity, whereas nucleolar and ribosomal gene expression may be supported by non-catalytic scaffolding functions. This division of labor underscores the multifunctional nature of PARP-2 in safeguarding both genome integrity and transcriptional equilibrium.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 8. PARP-2 integrates DNA repair, transcriptional control and nucleolar surveillance to preserve germline genome integrity. (A) PARP-2 is recruited to stalled replication forks, inter-strand cross-links (ICLs) and meiotic DSBs, where its catalytic activity promotes FCD-2 loading and efficient inter-strand cross-link repair. The same catalysis simultaneously operates as a regulatory rheostat that prevents over-amplification of DNA-damage signalling. (B) Independent of exogenous damage, PARP-2 binds transcription start sites (TSS) and GAGA-motif-rich regions, activating programs that drive meiotic progression and replication-stress responses while fine-tuning chromatin accessibility. (C) PARP-2 localizes within the nucleolus and controls expression of ribosome-biogenesis genes.

https://doi.org/10.1371/journal.pgen.1012062.g008

Given that clinical PARP inhibitors target both PARP-1 and PARP-2, our findings suggest that therapeutic strategies should consider the specialized contribution of PARP-2 to replication stress responses and pathway choice regulation. Consistent with this, PARP inhibitor treatment induces cellular responses characteristic of replication stress, including reduced nuclear PARP-2 signal and activation of DNA damage checkpoints. Notably, PARP inhibition itself may perturb transcriptional homeostasis. Clinical PARP inhibitors principally act via catalytic inhibition and PARP–DNA trapping rather than by reducing total PARP protein levels. To date, the literature does not provide consistent evidence that PARP inhibitor treatment generally decreases PARP1 or PARP2 abundance, although a few reports in specific cellular contexts have observed partial reductions [44,45]. Therefore, the reduced nuclear PARP-2 fluorescence observed here is interpreted as likely reflecting altered nuclear/chromatin association or detection rather than definitive protein depletion; direct biochemical assays would be required to resolve this.

We observed a reduction in parp-2 transcript levels following olaparib treatment, which likely reflects an indirect effect rather than direct suppression of gene expression. Previous studies indicate that PARP1/2 participate in transcriptional regulation, and PARP inhibition can disrupt transcription–replication balance, thereby altering global transcriptional output [46,47]. Thus, reduced parp-2 mRNA levels may represent feedback regulation of DNA damage response genes or general transcriptional stress rather than a direct consequence of catalytic inhibition. Cancer-associated PARP mutations may deregulate not only repair capacity but also transcriptional homeostasis, emphasizing the need for approaches that can differentially modulate PARP-1 versus PARP-2 activity [42].

Strains and alleles

All C. elegans strains were maintained at 20°C under standard laboratory conditions, as previously described [48]. The N2 Bristol strain was used as the wild-type reference.

The parp-2(ok344) allele, carrying a 1,576 bp deletion that removes exons 1–4, including the catalytic domain, was obtained from the Caenorhabditis Genetics Center (CGC). The deletion was confirmed by PCR genotyping. Total RNA was extracted from synchronized young adults, and parp-2 transcript levels were measured by quantitative PCR using gene-specific primers. Protein expression and PARylation status were analyzed by Western blot using an anti–PARP-2 antibody.

The parp-2 (E509K) allele was generated by CRISPR-Cas9–mediated genome editing to substitute the catalytic glutamate residue (E509) with lysine, analogous to the E988K mutation in human PARP1 using CRISPR-Cas9 (Suny Biotech, Fuzhou, China). The single-nucleotide substitution was verified by Sanger sequencing, and catalytic activity was assessed by in vitro PARylation assay using worm lysates.

A PARP-2 reporter strain (FLAG::3xHA::PARP-2) was generated by inserting an N-terminal FLAG–3 × HA tag at the endogenous parp-2 locus using CRISPR-Cas9 (Suny Biotech, Fuzhou, China). Correct insertion was confirmed by PCR and Sanger sequencing, and protein expression and localization were examined by Western blot and immunostaining.

All deletion alleles used in this study were verified by PCR and/or Sanger sequencing. Deletion mutant strains were outcrossed at least three times prior to phenotypic analysis.

RNA sequencing Analysis

Synchronized C. elegans populations were grown to 1 day post-L4 stage at 20 °C, washed three times in M9 buffer, and pelleted. Total RNA was extracted using TRIzol reagent, quantified with Nanodrop, and integrity was verified. After DNase I treatment, samples were stored at –80 °C and shipped on dry ice to the sequencing facility.

Chip Sequencing Analysis

For ChIP-seq, staged C. elegans (∼10,000–20,000 embryos or adults per replicate) were quenched with 125 mM glycine and snap-frozen in liquid nitrogen. After lysis, chromatin was immunoprecipitated with the HA-Tag antibody (6E2, Cell Signaling), washed, and eluted. Eluted samples were shipped on dry ice to the sequencing core.

Sample staging for RNA-seq and ChIP-seq

Worms were synchronized and collected at 1 day post-L4 (young adult stage), which is after the larval arrest period. At this stage, for ok344 deletion mutants, no larval arrest was observed in the populations used for RNA extraction. For E509K homozygotes, first-generation animals were used because second-generation animals are completely arrested at the larval stage. Whole animals were harvested, and total RNA was extracted using TRNzol reagent (Tiangen, Cat. No. GDP424) according to the manufacturer’s instructions. Because young adults contain developing embryos, a small number of dead embryos may have been included in the RNA preparations.

DNA damage sensitivity assay

For HN₂ sensitivity, animals were treated with 0, 50, 100 or 150 µM of HN₂ (mechlorethamine hydrochloride) in M9 buffer containing E. coli OP50 with slow shaking in the dark for 20 hours. Cisplatin treatment was similar but with doses of 0, 100, or 200 µg/ml. After treatment with either HN₂ or Cisplatin, animals were washed twice with M9 containing TritonX100 (100 ml/L) and plated to allow recovery for 3 hours [49,50]. UV irradiation treatment was performed utilizing the XL-100 Spectrolinker UVC. Worms were exposed to 0, 100, 200 or 300 J/m² of UVC and plated to allow recovery for 3 hours. HU sensitivity was assessed by placing animals on seeded NGM plates containing either 0, 10, 20 or 30 mM HU for 20–24 hours. Hatching sensitivity was examined in at least 24 animals 4 hours after HU treatment. Each damage condition was replicated at least twice in independent experiments.

Immunofluorescence staining

Immunofluorescence staining of whole-mount gonads was performed as described in [32,49,51]. Primary antibodies used included rabbit anti-pCHK-1 (1:250, Cell Signaling, Ser345), rabbit anti-FCD-2 (1:250, custom antibody), rabbit anti-RAD-51 (1:2000; SDIX), mouse anti-HA (1:250, Jackson Immunochemicals), rabbit anti-HA (1:250, Jackson Immunochemicals), mouse anti-NOP-1 (1:100, Santa Cruz, 28F2), rabbit anti-ATL-1(1:500, [33]) and secondary antibodies were Cy3 anti-rabbit (1:300, Jackson Immunochemicals), FITC anti-rabbit (1:300, Jackson Immunochemicals), Cy3 anti-mouse (1:250, Abcam), FITC anti-mouse (1:250, Abcam).

Fluorescence images were captured at 0.2 μm intervals using an Eclipse Ti2-E inverted microscope equipped with a DSQi2 camera (Nikon) or CSU-X1. Images were obtained with a 60x objective, combined with 1.5x auxiliary magnification, and subjected to deconvolution using NIS Elements software (Nikon) or VisiView. Partial projections of half-nuclei are shown.

Western blotting

Proteins were separated by SDS–PAGE and transferred onto PVDF membranes (Millipore). The membranes were blocked with 5% non-fat dry milk in TBST (Tris-buffered saline, 0.1% Tween-20) for 1 h at room temperature and then incubated overnight at 4 °C with the following primary antibodies: PARP2 antibody (PA5–116682, Thermo Fisher) and anti-poly(ADP-ribose) polymer antibody (ab14459, Abcam) both diluted at 1:1000, as well as GAPDH antibody (Adamas Life) diluted at 1:2000. After washing, the membranes were incubated with HRP-conjugated secondary antibodies (Jackson ImmunoResearch) for 30–60 min at room temperature. Signals were detected using enhanced chemiluminescence (ECL, Thermo Fisher) and imaged with a ChemiDoc imaging system (Bio-Rad).

Quantitation of germline apoptosis

Germline apoptosis was assessed by acridine orange staining in age-matched (20 hours post-L4) animals, following the method described in [52]. Briefly, live worms were incubated with acridine orange in M9 buffer for ~2 hours in the dark, washed three times with M9 buffer, and transferred to fresh NGM plates for ~2 hours to clear residual dye. A Nikon Ti2-E fluorescence microscope was used to score between 20 and 30 gonads per treatment. Statistical significance was determined using the two-tailed Mann–Whitney test, with a 95% confidence interval.

Quantitative real-time PCR

cDNA was synthesized from RNA extracted from young hermaphrodite worms using the ABscript II First Strand synthesis kit (ABclonal, RK20400). Real-time PCR was performed using ABclonal 2X SYBR Green Fast Mix (RK21200) in a LineGene 4800 system (BIOER, FQD48A). The initial denaturation step was carried out at 95°C for 2 minutes, followed by 40 cycles of 95°C for 15 seconds, 60°C for 20 seconds, and elongation. A melting curve analysis (60°C to 95°C) was conducted to verify the specificity of the PCR products. Tubulin encoding gene tba-1 was used as a reference gene based on microarray data for C. elegans. Each PCR experiment was repeated at least twice.