(C) PLOS One
This story was originally published by PLOS One and is unaltered.
 . . . . . . . . . .



SUMO-mediated regulation of H3K4me3 reader SET-26 controls germline development in C. elegans [1]

['Cátia A. Carvalho', 'Department Of Cell', 'Developmental Biology', 'Faculty Of Medical', 'Health Sciences', 'Tel Aviv University', 'Tel Aviv', 'Ulrike Bening Abu-Shach', 'Asha Raju', 'Zlata Vershinin']

Date: 2025-01 

Sumoylation is a posttranslational modification essential for multiple cellular functions in eukaryotes. ULP-2 is a conserved SUMO protease required for embryonic development in Caenorhabditis elegans. Here, we revealed that ULP-2 controls germline development by regulating the PHD-SET domain protein, SET-26. Specifically, loss of ULP-2 results in sterility and a progressive elevation of global protein sumoylation. In the germline of ulp-2 null mutant, meiosis is arrested at the diplotene stage and the cells in the proximal germline acquire a somatic fate. Germline RNAseq analysis revealed the down-regulation of numerous germline genes in ulp-2 mutants, whereas somatic gene expression is up-regulated. To determine the key factors that are regulated by ULP-2, we performed a yeast two-hybrid screen and identified the histone methylation reader, SET-26 as a ULP-2 interacting protein. Loss of SET-26 enhanced the sterility of ulp-2 mutant animals. Consistently, SET-26 is sumoylated and its sumoylation levels are regulated by ULP-2. Moreover, we detected a reduction in H3K4 tri-methylation (H3K4me3) histone levels bound to SET-26 in the ulp-2 mutant background suggesting a dependence of this histone reader on balanced sumoylation. Finally, a comparative proteomics screen between WT and ulp-2 loss of activity identified the predicted methyltransferase SET-27 as a ULP-2-dependent SET-26-associated protein. SET-27 knockout genetically interacts with ULP-2 in the germline, but not with SET-26. Taken together, we revealed a SUMO protease/H3K4me3 histone reader axis which is required for the maintenance and regulation of germline development.

C. elegans SET-9 and SET-26 are 2 histone readers with a PHD-SET domain that binds to H3K4me3 marks in germline and somatic genes [ 31 ]. Even though they are 97% identical in their sequence, they perform different functions. SET-9 and SET-26 synergize to maintain germline function across generations; however, SET-26 plays an additional role in regulating lifespan in a germline-independent manner and in regulating resistance to heat stress [ 31 , 35 , 36 ]. The C. elegans ULP-2 protein occupies the Ulp2-like branch of the SUMO proteases and shares 17% similarity with the catalytic domains of the mammalian SUMO proteases SENP6 and SENP7 [ 37 ]. Proteomics studies revealed that SENP6 is associated with chromatin organization complexes [ 38 ] and is required to stabilize the centromeric H3 variant CENP-A [ 39 , 40 ]. Here, we revealed a role for the ULP-2 SUMO protease in regulating a reader of the active H3K4me3 histone methylation mark in the germline. We demonstrated that ULP-2 plays a key role in germline development and maintenance of germ cell fate and identified the SET-26 histone reader as a central target for this activity.

SUMO, a small ubiquitin-like modifier, is covalently attached to target proteins [ 1 , 2 ] and is essential for development and viability in eukaryotes [ 3 ]. SUMO modification is highly dynamic and reversible; its deconjugation is mediated by specific SUMO proteases that cleave the isopeptide bond between the SUMO moiety and substrates [ 4 , 5 ]. The important role of SUMO in the regulation of gene expression has been extensively demonstrated [ 6 ], and many transcription factors and chromatin regulators have been identified as SUMO target proteins [ 7 , 8 ]. Sumoylation is associated with both transcriptional repression and activation [ 9 ]. Posttranslational histone modifications, including methylation, are epigenetic marks that generate binding sites for histone reader proteins, which transduce these signals to downstream effectors [ 10 , 11 ]. Methylation of histone H3 on lysine residues K4, K36, and K79 is often linked to transcriptional activation, whereas methylation events on lysine residues K9 and K27 are associated with gene repression [ 11 , 12 ]. It has been demonstrated that the SUMO machinery contributes to the regulation of repressive histone methylation; for example, sumoylation has been shown to be required for Polycomb group protein (PcG) Pc2 recruitment to repressive H3K27me3 marks in mouse embryonic fibroblast cells [ 13 ] and for the activity of the PcG-like protein SOP-2 in Caenorhabditis elegans [ 14 ]. Interestingly, Pc2 itself acts as a SUMO E3 ligase in mammalian cell culture [ 15 ]. Down-regulation of sumoylation by knockdown of Ubc9 in embryonic stem cells leads to genome-wide loss of H3K9me3-dependent heterochromatin [ 16 ]. In Drosophila, deposition of H3K9me3 marks depends on SUMO and the PIAS SUMO E3 ligase Su(var)2–10, which recruits the SetDB1/Wde complex [ 17 ]. Many protein domains mediating reading of histone methylation states have been identified [ 18 ], several of which are linked to SUMO modifications. One of these is the chromodomain found in PcG proteins and the heterochromatin protein 1 (HP1). Chromodomain proteins play an important role in maintaining repressed chromatin states [ 19 , 20 ]. In C. elegans, knockdown of SUMO alters the chromatin-binding pattern of the chromodomain protein MRG-1 [ 21 ]. An additional histone methylation reader domain family is the plant homeodomain (PHD) zinc fingers which includes approximately 100 distinct PHD fingers with histone-binding activity [ 22 ]. Structural studies revealed that a modified histone H3 tail is extensively associated with the PHD finger domain, providing a high degree of specificity of this domain to the active mark H3K4me3 [ 23 , 24 ]. PHD domains were suggested to function as SUMO E3-ligases [ 25 ]. Several histone readers contain more than one domain and are therefore potentially able to receive and translate complex signals. One of them is the mixed lineage leukemia 5 protein (MLL5), which contains both a Su(var)3–9, Enhancer-of-zeste, Trithorax (SET) domain, and a PHD domain [ 26 ]. MLL5 binds to H3K4me3 marks through the PHD domain and associates with chromatin downstream of the transcriptional start sites of active genes [ 27 – 29 ]. Based on a similar sequence of their SET domains, MLL5 appears to be the mammalian homolog of yeast SET3/SET4 paralogs, Drosophila UpSET, and the C. elegans SET-9/SET-26 paralogs [ 30 , 31 ]. The unique property of the SET domain in these proteins is the lack of histone methyl transferase activity [ 26 , 32 , 33 ] (and this study). The SET domains of SET3 and UpSET were shown to be associated with histone deacetylases [ 32 , 33 ] and to regulate H3K9me2 levels [ 34 ].

Results

Increased sumoylation levels in ulp-2 mutant inversely shapes fertility output and lifespan We have previously reported that the SUMO protease ULP-2 is required for embryonic epidermal morphogenesis. Closer examination revealed that in addition, loss of ULP-2 leads to a developmental delay and germline sterility. The first generation of homozygous ulp-2 mutants is viable but produces 63% lethal embryos [37]. The remaining 37% complete embryonic development and develop into sterile adults. The development of this escaper group was highly delayed compared to both wild-type (WT) controls or F1 ulp-2 homozygous mutant animals (8 to 9 days, Figs 1A and S1). We measured the brood size (unhatched embryos and larvae) of first (F1) and second (F2) generation ulp-2 mutant animals and found that F1 ulp-2 mutants already displayed reduced fertility compared with WT animals (approximately 55% reduction: 153 ± 53 versus 277 ± 30 progeny). The second generation of homozygous mutants was almost completely sterile (approximately 99% sterility, 0.2 ± 2 progeny) (Fig 1B). This suggests that there is a progressive reproductive fitness decline in the homozygous generations of ulp-2 mutant animals. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 1. Loss of function of ULP-2 induces fertility decline and increases longevity accompanied by the accumulation of SUMO conjugates. (A) Schematic representation of the developmental progress of ulp-2(tv380) animals. C. elegans schemes were generated with BioRender.com. (B) Quantification of the number of progeny laid by WT (30 animals), F1 ulp-2(tv380) (53 animals), and F2 ulp-2(tv380) (94 animals) in 2 biological replicates; the Shapiro–Wilk and one-way ANOVA tests on ranks (Kruskal–Wallis), followed by Dunn’s post hoc test, were used; ns = p > 0.05. (C) Survival curves for WT (252 death events + 100 censored), F1 ulp-2(tv380) (220 death events + 162 censored), and F2 ulp-2(tv380) (213 death events + 228 censored) animals in 2 biological replicates; the log-rank (Mantel–Cox) test was used; ns = p > 0.05. (D) Representative immunoblot of the total sumoylation levels of WT, F1 ulp-2(tv380), and F2 ulp-2(tv380) animals; 3 biological replicates; the upper panel was probed with anti-sumo and the bottom panel was probed with anti-α-tubulin. (E) Quantification of the total levels of sumoylated proteins in D; the Shapiro–Wilk and one-way ANOVA tests, followed by Tukey’s post hoc statistical test, were used; ns = p > 0.05. The numerical data presented in this figure can be found in S1 Data. https://doi.org/10.1371/journal.pbio.3002980.g001 Since the reproductive system plays a key role in regulating the lifespan of the organism [41,42], we determined whether the lifespan of the ulp-2 mutants was affected. We found that the median lifespan of WT animals is 16 days, compared with 19 days in F1 ulp-2(tv380) and 24 days in F2 ulp-2(tv380) animals (Fig 1C and S1 Data). These results reflect a progressive increase in the lifespan from F1 to F2 animals lacking ULP-2 activity which correlates inversely with the progressive loss of reproductive fitness. Since both generations of homozygous animals are genetically identical but differ in the levels of the maternal ulp-2 that they have received, we reasoned that the phenotypic differences could be associated with an accumulation of SUMO conjugates over time. Indeed, whereas F1 ulp-2(tv380) mutant animals contained double the amount of sumoylated substrates compared with WT (mean = 1.6 ± 0.4 versus 0.7 ± 0.3); the F2 ulp-2(tv380) mutant animals contained triple the amount of SUMO conjugates (mean = 2.4 ± 0.7) (Fig 1D and 1E). All together, these results suggest that the accumulation of SUMO conjugates in subsequent generations of ulp-2 mutant animals contributes to the decline in fertility and concomitant increase in lifespan.

Loss of ULP-2 activity disrupts the transcriptional program in the germline We hypothesized that abnormal desumoylation of germline proteins caused by ULP-2 loss impairs the overall developmental program of the germline leading to sterility. To reveal the key affected pathways, we performed transcriptomics analysis of isolated germlines of ulp-2 sterile animals and WT control animals. Morphologically, the germlines of F2 ulp-2(tv380) animals have a reduction of approximately 50% in length (269 ± 58 μm) compared with WT germlines (533 ± 73 μm) (S2A Fig). The mutant germlines also harbor 45% fewer nuclei, indicating that ULP-2 loss of function negatively affects the cellular content of the germline (S2B Fig). The transcriptomics analysis revealed that a total of 7,202 genes are differentially expressed in the germlines of F2 ulp-2(tv380) relative to WT, with 3,577 genes down-regulated and 3,625 genes up-regulated (Figs 2A and S2C and S1 and S2 Tables). Gene Ontology (GO) analysis revealed that the down-regulated genes in F2 ulp-2(tv380) germlines are associated with biological processes essential for germline maintenance—such as cell cycle regulation, P-granule assembly, and oocyte development and specification (S3A Fig). On the other hand, the GO analysis of the genes up-regulated in the F2 ulp-2(tv380) germlines is associated with somatic processes such as immune response, cuticle development, and neurogenesis (S3B Fig). The intersection of our transcriptomics data set with 2 previously defined data sets [43] substantiates the GO analyses; whereas the down-regulated genes in the F2 ulp-2(tv380) germlines largely overlap with the “germline-enriched genes” data set (R.F. = 2.6), the up-regulated genes significantly intersect with the “somatic-specific genes” data set (R.F. = 2.4) (Fig 2B). Taken together, these observations suggest that ULP-2 promotes the transcription of genes required to maintain the germline and negatively regulates somatic transcription in the germline. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 2. Loss of ULP-2 disrupts the germline gene expression program, leading to impaired oocyte formation and loss of germ cell identity. (A) A volcano plot displaying the distribution of the differentially expressed genes of the F2 ulp-2(tv380) germlines in comparison to WT. The cut-off values used were FDR > 2 (-log10) and fold change>|1.2| (log2). (B) A Venn diagram representing the intersection between down-regulated and up-regulated genes in the F2 ulp-2(tv380) germlines and the established data sets of germline-enriched and somatic-specific genes [43]; the representation factor (R.F.) of each intersection and the associated p-value are shown; Fisher’s exact test was used. (C) Representative images of the isolated germlines of WT and F2 ulp-2(tv380) animals; the germline zones and sperm are indicated. A single nucleus in diplotene is labeled with dashed circle, a single nucleus in diakinesis is labeled with a rectangle, and a large nucleus is denoted by a white arrow in the ulp-2(tv380) germline. Scale bar = 30 μm. Insets below the images show magnification of the proximal gonad and sperm. (D) Quantification of the number of nuclei per region in the germlines of WT (7 germlines, 3,834 nuclei) and F2 ulp-2(tv380) (6 germlines, 1,809 nuclei) animals; the Shapiro–Wilk and the two-tailed Mann–Whitney U tests were used; ns = p > 0.05. Lines on the Y axis indicate a scale change to allow observation of the quantification. (E) Representative images of UNC-119::GFP localization in WT and F2 ulp-2(tv380) animals. White line marks the proximal gonad, most proximal oocyte (oo) and spermatheca (sp). Scale bar = 10 μm. The numerical data presented in this figure can be found in S1 Table (A) and S1 Data. https://doi.org/10.1371/journal.pbio.3002980.g002

Impaired desumoylation prevents gamete formation and the maintenance of germ cell identity We next investigated the cytological defects in the germline that underlie the observed sterility. Germ cells are produced by mitotic divisions in the distal region of the C. elegans gonad with meiosis initiating and progressing in the distal to proximal direction. At the proximal end, germ cells mature into oocytes, which are then fertilized in the spermatheca [44]. The gonads of F2 ulp-2(tv380) animals are relatively smaller with fewer germ cells and almost no mature oocytes (Fig 2C). The mutant germlines appear to be relatively healthy at the L4 stage but deteriorate when the animals continue development to adulthood. The number of germ cells undergoing mitosis was reduced by 41% in F2 ulp-2(tv380) animals. Accordingly, we observed a sharp reduction in the number of nuclei in all meiotic stages. The most dramatic reduction was in diakinesis, with 81% fewer nuclei, resulting in an almost complete absence of oocytes formed in F2 ulp-2(tv380) germlines (Fig 2C and 2D). Interestingly, we identified apparently undifferentiated larger misshapen nuclei appearing in variable sizes and numbers in the proximal gonad (Fig 2C, arrow in inset), which resembles the abnormal differentiation of germ cells into somatic cells observed in the lin-41 mutants [45]. To determine whether these larger cells lost their germline fate, we crossed the ulp-2 mutant animals with the pan-neuronal marker UNC-119::GFP [46], which has been used to detect germ cell reprogramming to a neuron-like somatic fate [47–49]. We observed germline expression of UNC-119::GFP in cell clusters in the proximal gonad with axonal-like cellular extensions (85%; n = 13) that are not detected in WT animals (n = 8) (Fig 2E). Analysis of an additional neuronal/hypodermal UNC-13::GFP reporter tagged endogenously with CRISPR/Cas9 also revealed ectopic germline expression in the proximal gonad in F2 ulp-2(tv380) mutants (59%; n = 32) while no such expression was detected in WT control (n = 30) (S4 Fig). This analysis suggests that ULP-2 is essential for protecting germ cell fate in the proximal germline, in line with the results of the transcriptomic analyses.

ULP-2 interacts with the SET-26 H3K4me3 reader to maintain germline function To identify targets for the SUMO protease activity of ULP-2 in the germline, we performed a yeast two-hybrid (Y2H) screen of a C. elegans mixed-stage cDNA library. The N-terminal region of SENPs/ULPs has been identified as the substrate recognition domain [50,51], and therefore, we used the N-terminal domain of ULP-2 as bait (Fig 3A). We identified the SET-26 protein and a homologous protein, Y73B3A.1, as putative interactors of ULP-2. SET-26 is a PHD-SET protein while Y73B3A.1 has a high sequence homology to SET-26 but lacks the PHD-SET domain (S5A Fig). SET-9 and SET-26 are paralogs that have been established as H3K4me3 readers required for a functional germline [31]; therefore, SET-9, SET-26, and Y73B3A.1 are potential targets for ULP-2 activity in the germline. To examine if ulp-2 genetically interacts with set-9, set-26, or Y73B3A.1, we first simultaneously down-regulated all 3 genes (“Set3” RNAi) in homozygous F1 ulp-2(tv380) animals by RNAi. We observed a decrease in the average number of progeny produced when compared with F1 ulp-2(tv380) and WT (S5B Fig). Next, we down-regulated ulp-2 by RNAi in individual mutants of set-9, set-26, and Y73B3A.1. We observed increased embryonic lethality and near sterility when inactivating ulp-2 and set-26, but not when combining ulp-2 inactivation with set-9 or Y73B3A.1 mutants (S5C and S5D Fig). We therefore decided to focus on the interaction of ULP-2 with SET-26. PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 3. SET-26 is required for maintenance of the germline function of ULP-2. (A) Exon intron structure of ULP-2 with the region used as bait labeled in blue. (B) A yeast two-hybrid analysis of the interaction between the N- and C-terminal domains of SET-26 and the N-terminal domain of ULP-2. The top row is a non-selective growth plate. The middle 3 rows are selective plates that only allow yeast growth in the presence of an interaction. The bottom row is a control plate where growth indicates the auto-activation of reporter genes by the DNA-binding domain protein fusion. Controls are the protein pairs of a known reporter activation strength (S3 Table). (C) Quantification of the number of progeny produced by WT (30 animals, 277 ± 30 offspring per mother), ulp-2(tv380) (32 animals, 153 ± 63 offspring per mother), F2 ulp-2(tv380) (94 animals, 0.2 ± 2 offspring per mother), set-26(tm2467) (25 animals, 207 ± 27 offspring per mother), ulp-2(tv380);set-26(tm2467) (39 animals, 0.3 ± 2 offspring per mother), set-26(tm3505) (25 animals, 218 ± 25 offspring per mother) and ulp-2(tv380);set-26(tm3505) (37 animals, 2 ± 4 offspring per mother); 2 biological replicates; the Shapiro–Wilk and one-way ANOVA tests on ranks (Kruskal–Wallis), followed by Dunn’s post hoc test, were used; ns = p > 0.05. (D) Survival curves for WT (252 death events + 100 censored animals), ulp-2(tv380) (220 death events + 162 censored animals), F2 ulp-2(tv380) (213 death events + 228 censored animals), set-26(tm2467) (254 death events + 92 censored animals), F1 ulp-2(tv380);set-26(tm2467) (157 death events + 210 censored animals), set-26(tm3505) (245 death events + 81 censored animals), and F1 ulp-2(tv380);set-26(tm3505) (205 death events + 188 censored animals); the log-rank (Mantel–Cox) test was used; ns = p > 0.05. The numerical data presented in this figure can be found in S1 Data. https://doi.org/10.1371/journal.pbio.3002980.g003 To confirm the interaction and map the binding region, we tested the N-terminal (exons 1–4) and the C-terminal (exons 6–9) regions of SET-26 for interaction with ULP-2 by Y2H (Fig 3B). The results indicated that the N-terminal region of SET-26 mediates its interaction with the N-terminal domain of ULP-2 (Fig 3B). To further examine the genetic interaction between ULP-2 and SET-26, we generated double mutants of ulp-2(tv380) with 2 loss of function alleles of set-26. In these 2 alleles, the N-terminal region is intact and the deletion causes early termination of the reading frame before or inside the PHD domain, leading to a shorter protein with no PHD and SET domains (S5E Fig). Double ulp-2; set-26 mutants rendered the first generation of homozygous ulp-2(tv380) near sterile, resembling the F2 ulp-2(tv380) phenotype (Fig 3C). Thus, F1 homozygous ulp-2(tv380) animals express sufficient maternal ULP-2 for germline development, but this process is impaired in the background of the set-26 loss of function alleles. These results suggest that SET-26 is a key target of ULP-2 activity in the germline. In addition to its function in the germline, SET-26 has been shown to regulate somatic-driven maintenance of a normal lifespan [35,36]. To determine whether ULP-2 and SET-26 interaction extends to somatic tissues, we measured the lifespan of the double mutants. In double mutants, we observed a shorter extension of lifespan than observed in either single mutant (Fig 3D and S1 Data). This suggests that while in the germline both factors act positively to maintain the germline, in the soma their mode of interaction is possibly indirect and involves additional mechanisms.

ULP-2 controls the sumoylation levels of SET-26 SUMO proteases mediate the enzymatic removal of SUMO moieties from sumoylated proteins [5]. The physical and genetic interaction between SET-26 and ULP-2 suggests that SET-26 is a target of ULP-2 protease activity. To investigate this hypothesis, we first assessed whether SET-26 is a SUMO-acceptor protein. Using an in vitro sumoylation assay, the SET domain with or without the adjacent PHD domain of SET-26, accumulated multiple higher molecular weight bands detected by anti-SUMO antibody, suggesting that SET-26 can be either sumoylated on several lysine residues or it can be polysumoylated (Fig 4A). The reaction proceeded more rapidly when the region including the PHD-SET domain of SET-26 was included (Fig 4A), suggesting that the PHD domain in SET-26 may act as an intramolecular SUMO E3 ligase, analogous to the enhancement of sumoylation of the bromodomain of the KAP1 corepressor by its SET domain [25]. Next, we performed in vitro desumoylation reactions and found that incubation of sumoylated PHD-SET domains with ULP-2 resulted in the removal of the SUMO peptides, as manifested by a sharp decrease in signal, compared with the sumoylated input (Fig 4B). Overall, these results indicate that ULP-2 can regulate the sumoylation levels of SET-26 in vitro. If SET-26 is a target of ULP-2 in vivo, increased levels of sumoylated SET-26 should be observed in the ulp-2 loss of function background. Immunoprecipitation of SET-26::GFP did not exhibit strongly reacting bands with anti-SUMO antibody in a WT background (Fig 4C). However, SET-26::GFP immunoprecipitated from F2 ulp-2(tv380) mutant animals exhibited a strong reactivity with the anti-SUMO antibody, corresponding to a 3-fold increase in sumoylation levels (Fig 4C and 4D). These results suggest that ULP-2 regulates the sumoylation levels of SET-26 in C. elegans. SET-26 is the ortholog of human MLL5, fly UpSET, as well as yeast SET3 and SET4 that lack histone methyltransferase activity [26,32,33]. We tested whether the sumoylation of the PHD-SET domain of SET-26 can activate the SET methyltransferase in vitro, but no such activity was detected (S6 Fig). PPT PowerPoint slide

PNG larger image

TIFF original image Download: Fig 4. SET-26 is a SUMO target and is desumoylated by ULP-2. (A) In vitro sumoylation of the SET and PHD-SET domains of SET-26. Bacterially expressed GST-SET and GST-PHD-SET were incubated with E1 (SAE1/SAE2), E2 (UBC9), SUMO1, and ATP for 0, 15, 30, and 60 min at 37°C. Control reactions were incubated for 60 min without the substrate. The anti-SUMO and anti-GST antibodies were used to detect sumoylation and the GST proteins, respectively. (B) The ULP-2 catalytic domain desumoylates the PHD-SET domain in vitro. The GST-PHD-SET was sumoylated for 2 h before adding ULP-2 for the indicated time points. Reactions were carried out at 37°C. The anti-SUMO and anti-GST antibodies were used similarly to A. (C) SET-26 in vivo sumoylation in WT and F2 ulp-2(tv380) backgrounds; anti-GFP and anti-SUMO antibodies were used to probe SET-26::GFP and the sumoylation levels of SET-26, respectively. (D) Quantification of the normalized levels of the sumoylation of SET-26 in C for 3 biological replicates; one sample t test was used; ns = p > 0.05. The numerical data presented in this figure can be found in S1 Data. https://doi.org/10.1371/journal.pbio.3002980.g004

[END]
---
[1] Url: https://journals.plos.org/plosbiology/article?id=10.1371/journal.pbio.3002980

Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.

via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/