Phosphorylation of HORMA-domain protein HTP-3 at Serine 285 is dispensable for crossover formation

Abstract Generation of functional gametes is accomplished through a multilayered and finely orchestrated succession of events during meiotic progression. In the Caenorhabditis elegans germline, the HORMA-domain-containing protein HTP-3 plays pivotal roles for the establishment of chromosome axes and the efficient induction of programmed DNA double-strand breaks, both of which are crucial for crossover formation. Double-strand breaks allow for accurate chromosome segregation during the first meiotic division and therefore are an essential requirement for the production of healthy gametes. Phosphorylation-dependent regulation of HORMAD protein plays important roles in controlling meiotic chromosome behavior. Here, we document a phospho-site in HTP-3 at Serine 285 that is constitutively phosphorylated during meiotic prophase I. pHTP-3S285 localization overlaps with panHTP-3 except in nuclei undergoing physiological apoptosis, in which pHTP-3 is absent. Surprisingly, we observed that phosphorylation of HTP-3 at S285 is independent of the canonical kinases that control meiotic progression in nematodes. During meiosis, the htp-3(S285A) mutant displays accelerated RAD-51 turnover, but no other meiotic abnormalities. Altogether, these data indicate that the Ser285 phosphorylation is independent of canonical meiotic protein kinases and does not regulate HTP-3-dependent meiotic processes. We propose a model wherein phosphorylation of HTP-3 occurs through noncanonical or redundant meiotic kinases and/or is likely redundant with additional phospho-sites for function in vivo.


Introduction
Sexual reproduction relies on the formation of haploid gametes through meiosis, a specialized cell division mechanism that ensures equal distribution of the genetic material in the daughter cells Kleckner 1999, 2015). Faithful chromosome segregation depends on the recognition of the homologous chromosomes (pairing), stabilization of their association through the synaptonemal complex (SC; synapsis), and establishment of chiasmata (recombination; Kleckner 1999, 2015). The latter arises from crossover (CO)-dependent repair of programmed double-strand breaks (DSBs), which are generated during meiotic prophase I by the topoisomerase-like enzyme Spo11 (Keeney et al. 1997). The SC is a proteinaceous tripartite structure composed of lateral and central elements that assembles in a "zipper"-like fashion to keep the homologous chromosomes tightly juxtaposed, thus allowing for the physical exchange of DNA molecules during homology-mediated DSB repair.
A family of HORMA-domain-containing proteins, composed of HTP-1/2, HTP-3, and HIM-3 in Caenorhabditis elegans, localizes along the axial elements of the SC and exerts essential functions for axes morphogenesis and the licensing of the SC loading between the homologs (Couteau et al. 2004;Martinez-Perez and Villeneuve 2005;Goodyer et al. 2008;Severson et al. 2009). HTP-3 forms the base of the scaffold, and lack of HTP-3 prevents (1) SC polymerization by abolishing proper formation of chromosome axes and (2) a dramatic reduction of recombination intermediates, indicating an important role for this protein in the efficient induction of meiotic DSBs (Goodyer et al. 2008). Recent evidence has shown that HTP-3 can be directly regulated by kinases during meiotic progression. For example, Das et al. (2020) observed that ERK/MPK-1 phosphorylates HTP-3 in vitro, however, the actual phosphorylation site remains to be determined.
We have previously shown that parg-1, the nematode ortholog of mammalian poly(ADP-ribose) glycohydrolase PARG, is an important factor required to coordinate DSB induction and repair and identified physical interactors of PARG-1 in worms (Janisiw et al. 2020). In a mass spectrometry analysis performed on PARG-1::GFP pull downs, we identified a putative phosphorylated form of HTP-3 at Serine 285, which we further investigated. A phospho-specific HTP-3 S285 antibody detected phosphorylation of HTP-3 throughout meiotic prophase I, confirming that this site is phosphorylated in vivo. Phosphorylated-HTP-3 S285 localization overlaps with that of total HTP-3 and is independent of meiotic DSBs, synapsis, or COs. Individual removal of canonical meiotic kinases did not alter HTP-3 S285 phosphorylation pattern, suggesting that HTP-3 may be phosphorylated at multiple sites and these phospho-sites may function redundantly. Alternatively, S285 may be phosphorylated by multiple kinases in a redundant manner. Preventing HTP-3 S285 phosphorylation in vivo does not impair axes morphogenesis and recombination, indicating that this site is not essential to successfully achieve chiasmata formation. Altogether, our data show that phosphorylation of HTP-3 at Ser285 likely functions in a redundant manner with other phospho-residues that are yet to be identified.

Caenorhabditis elegans genetics and viability assays
The Bristol N2 C. elegans strain (Brenner 1974) was used as the wild-type control. The htp-3(S285A) was generated by CRISPR/ Cas9 genome editing by SUNY Biotech. Silent mutations encoding for an HhaI restriction site were included for screening purposes. The strains generated by CRISPR/Cas9 were outcrossed to wildtype N2 worms at least twice before use. All strains were maintained at 20 C under standard conditions for all experiments unless otherwise indicated. Viability and male progeny assessment were performed on single animals plated as L4 and then moved onto fresh NG plates every 24 h for 3 days. Dead embryos/total embryos were scored 24 h after the mother had been moved and the presence of males was evaluated 3 days later. Strains used for this study were: CA1199: unc-119(ed3) III; ieSi38 [sun-1p::TIR1::mRuby::sun-1 3'UTR þ Cbr-unc-119(þ)] IV.

RNAi experiments and auxin treatment
RNAi experiments were conducted by feeding L4 staged animals on HT115(DE3) bacteria expressing dsRNA of the relevant target gene (chk-1) for 48 h until dissection.
NGM plates containing 1-mM 5-Hydroxyindole-3-acetic acid (Sigma) dissolved in absolute ethanol (þauxin) were used for auxin-induced degradation of cdk-1, cdk-2, and chk-2. The control plates (-auxin) were poured in the same way with the addition of an identical volume of ethanol without auxin. Given that presence of auxin inhibits bacterial growth, a saturated OP50 culture was concentrated 5Â before being spotted onto NGM plates. Plates were left to dry overnight at room temperature before L4 animals were plated and then dissected 24 h later.

Immunostaining and images acquisition
About 20-24 h post-L4 stage animals were dissected in 15 ml of 1xPBS and fixed with an equal amount of 2% PFA (diluted in 1xPBS from a 16% stock) for 5 min at room temperature. A 24 Â 24 coverslip was gently applied and slides were submerged in liquid nitrogen for freeze-crack. Samples were placed in methanol at À20 C for 5 min and then washed thrice for 5 min at room temperature in 1xPBS with 0.1% Tween.
Blocking was performed by leaving the slides for 1 h at room temperature in 1% BSA (dissolved in 1xPBS with 0.1% Tween), followed by primary antibody treatment overnight at 4 C in a humid chamber. The following day, slides were washed in 1xPBS with 0.1% Tween thrice for 10 min each and secondary antibodies were left in incubation on the slides for 2 h at room temperature in the dark.
After 3 washes in the dark at room temperature for 10 min each, a 60 ml drop of DAPI (2 mg/ml) was placed onto the samples, allowed to stain for 1 min in the dark, and then washed for at least 20 min in 1xPBS with 0.1% Tween. A drop of 12 ml of Vectashield was placed onto the samples and a 22 Â 22 coverslip was then sealed using nail polish.
Images were acquired with an upright fluorescence microscope Zeiss AxioImager.Z2 equipped with a Hamamatsu ORCA Flash 4.0, sCMOS sensor camera, using UPlanSApo 100Â/1.4 Oil objective with Z-stacks at 0.24 mm thickness. Images were deconvolved with ZEN 3.0 Blue software (Zeiss), using "constrained iterative" algorithm at maximum strength.

Generation of anti-phospho-HTP-3 S285 and anti-RAD-51 antibodies
Phospho-specific antibodies against S285 in HTP-3 were produced by immunizing rabbits with the synthetic peptide CNPELDEIYFpSPGR (Genscript) in which the cysteine was added for conjugation to KLH (keyhole limpet hemocyanin). Polyclonal phosphoHTP-3 S285 antibodies were purified and then further absorbed against htp-3(S285A) mutant worms to reduce background. The specificity of the antibody was assessed by immunofluorescence where lack of staining was observed in htp-3(S285A) mutants but not in WT animals.
The synthetic peptide CSAQASRQKKSDQEQRAADQA corresponding to the amino acids 40-59 of C. elegans RAD-51 isoform a (including the underlined cysteine required for conjugation to KLH) was used to perform 4 immunization rounds in 2 rabbits (Genscript). RAD-51 polyclonal antibodies were separately affinity purified from raw sera of both animals against the same synthetic peptide employed as the immunogen. The specificity of the antibody was assessed by immunostaining of spo-11 mutants in which, unlike WT control animals, detection of RAD-51 foci was abrogated due to the absence of physiological DSBs.

Immunoprecipitation and sample preparation for mass spectrometry
Nuclear extracts from parg-1::GFP and untagged WT animals were produced as detailed in Silva et al. (2014). One milligram of pooled nuclear-soluble and chromatin-bound fractions were used for GFP immunoprecipitations by employing agarose GFPtraps (Chromotek), which had been pre-equilibrated in Buffer D [20% glycerol, 0.2 mM EDTA pH 8, 150 mM KCl, 20 mM Hepes-KOH (pH 7.9), and 0.2% Triton X-100, supplemented with protease inhibitor cocktail (Roche)]. Incubation of beads with the extracts was carried out over night at 4 C on a rotating shaker. The following day, the beads were spun down and washed extensively in Buffer D (without Triton X-100) before an equal amount of 2Â Laemmli buffer was added. Samples were boiled for 10 min and then the immunoprecipitated complexes were separated on a 4-12% acrylamide gel.
For mass spectrometry analysis, agarose beads were resuspended in 30 ml elution buffer (2 M urea, 50 mM ammonium bicarbonate), disulfide bonds reduced with 10 mM dithiothreitol for 30 min at room temperature, and then alkylated with 25 mM iodoacetamide for 15 min in the dark. After quenching with another 5 mM dithiothreitol, 150 ng of trypsin (Trypsin Gold, Promega) was added followed by 90-min incubation at room temperature in the dark. The supernatant without beads was transferred to a new tube and another 30 ml of elution buffer was added to the beads. The supernatants without beads were combined, diluted to 1 M urea concentration, and another 150 ng of trypsin was added before incubation at 37 C in the dark overnight. The digest was stopped by the addition of trifluoroacetic acid to a final concentration of 1%, and the peptides were desalted using C18 Stagetips (Rappsilber et al. 2007).

Mass spectrometry data acquisition and analysis
The mass spectrometer was operated in data-dependent acquisition mode, survey scans were obtained in a mass range of 380-1,650 m/z with lock mass activated, at a resolution of 120k at 200 m/z and an AGC target value of 3E6. The 10 most intense ions were selected with an isolation width of 2 m/z, fragmented in the HCD cell at 27% collision energy and the spectra recorded for max. 250 ms at a target value of 1E5 and a resolution of 30k. Peptides with a charge of þ1 or >þ6 were excluded from fragmentation, the peptide match feature was set to preferred, the exclude isotopes feature enabled, and selected precursors were dynamically excluded from repeated sampling for 30 s.
Raw data were processed using the MaxQuant software package (version 1.5.5.1; Tyanova et al. 2016) and the Uniprot C. elegans reference proteome (www.uniprot.org), as well as a database of most common lab contaminants. The search was performed with full trypsin specificity and a maximum of 2 missed cleavages at a protein and peptide spectrum match false discovery rate of 1%. Carbamidomethylation of cysteine residues was set as fixed, phosphorylation (serine, threonine, or tyrosine), oxidation (methionine), and N-terminal acetylation as variable modifications. Label-free quantification and the "match between runs" feature were activated-all other parameters were left at default. The spectrum supporting phosphorylation of pHTP-3 S285 was validated manually.

Protein expression and in vitro kinase assay
Recombinant HTP-3 protein was generated by cloning full-length cDNAs of htp-3 into pTrcHis Topo (Invitrogen, catalog no. K4410-01) bacterial expression vectors to generate N-terminally 6Â Histagged proteins (Das et al. 2020). Mutant HTP-3 S285A protein was generated through site-directed mutagenesis as described (Arur et al. 2009). Positive clones were then sequence verified. Recombinant proteins were expressed in BL21(DE3) (Sigma-Aldrich) cells at 37 C by using 1-mM isopropyl b-d-1-thiogalactopyranoside (dioxane free) for 3 h. Proteins were then purified by using Ni-nitrilotriacetic acid resin (Thermo scientific, no. 88221). The expression of proteins was confirmed by Western blot analysis with anti-His (Sigma-Aldrich, no. H1029). In vitro kinase assay was performed using purified ERK2 kinase (New England Biolabs, catalog no. P6080S), and the purified recombinant proteins according to the methods previously described (Arur et al. 2009). After phosphotransfer, the proteins were resolved onto a 10% SDS-polyacrylamide gel electrophoresis (Bio-Rad, catalog no. 4561033). The gel was then dried at 60 C under vacuum for 1 h and exposed to the autoradiographic film (Sigma-Aldrich, catalog no. 864 6770) for 4 h at À80 C followed by the development of the film using the Kodak X-OMAT 2000A processor machine.

Results and discussion
HORMA-domain-containing protein HTP-3 is phosphorylated at S285 in vivo We have recently shown that in C. elegans, the PARG-1/PARG protein, ortholog to mammalian poly(ADP-ribose)glycohydrolase, localizes along both the central and lateral elements of the SC, and its function is important for the induction of optimal levels of SPO-11-dependent DSBs and to promote their repair through HR (Janisiw et al. 2020). Furthermore, PARG-1::GFP requires HTP-3 to be properly loaded onto the chromosomes and coimmunoprecipitates with factors localizing along the SC (Janisiw et al. 2020).
To determine whether HTP-3 is phosphorylated at S285 in vivo, we generated a specific phospho-HTP-3 S285 antibody and performed immunofluorescence analysis.
Phospho-specific HTP-3 S285 was localized to nuclei in meiotic prophase I (Fig. 1c) and the signal was abolished in the htp-3(S285A) mutant (Fig. 1d, Supplementary Fig. 1), confirming that the antibody is specific to the phosphorylated form of HTP-3 S285 .
The expression of phospho-HTP-3 S285 overlaps with the pan-HTP-3 expression across the distal-proximal axis of the gonad (Fig. 1, c and e). Careful examination of the two localization patterns revealed that at meiosis onset the pHTP-3 S285 staining was rather uneven, relative to total HTP-3; most nuclei displayed a patchy localization of pHTP-3 S285 , with only a subset of them showing the phospho-staining along the whole length of the chromosome (Fig. 1e, transition zone/Early pachytene). Chromosome axes are established upstream to synapsis and therefore lateral elements such as HTP-3 are loaded earlier than the SYP proteins (Colaiá covo et al. 2003;Couteau et al. 2004;Martinez-Perez and Villeneuve 2005;Goodyer et al. 2008). At this stage, the establishment of the SC occurs gradually (Colaiá covo et al. 2003), thus, we hypothesize that phosphorylation of HTP-3 S285 might rely upon the completion of synapsis, which would be consistent with the gradual increase in its loading along the chromosomes (Fig. 2). By the early pachytene stage, the pHTP-3 S285 staining was no longer distinguishable from the panHTP-3 expression except in the cells undergoing physiological apoptotic cell death, where pHTP-3 S285 was absent (Fig. 1e, mid/latepachytene, dotted circle). Thus, we conclude that HTP-3 is phosphorylated at S285 in vivo during meiotic prophase I.
Given the gradual recruitment of p-HTP-3 S285 in the nuclei at meiosis entry, we wondered whether HTP-3 S285 phosphorylation was dependent on early events that occur during meiotic progression, such as the establishment of synapsis and formation of physiological DSBs, respectively. To assess whether phosphorylation of HTP-3 at Ser285 required the establishment of synapsis, we performed immunostaining analysis of pHTP-3 S285 in syp-2 mutants. We observed that pHTP-3 S285 was detectable along chromosomes upon loss of syp-2 indicating that its localization is independent of SC establishment (Fig. 2).
In C. elegans, SC formation takes place independently of induction of SPO-11-mediated DSBs (Dernburg et al. 1998). Programmed DSB induction relies on the catalytic activity of the topoisomerase-like SPO-11, which exerts its function in combination with several cofactors in worms (Reddy and Villeneuve 2004;Wagner et al. 2010;Meneely et al. 2012;Rosu et al. 2013;Stamper et al. 2013;Hinman et al. 2021). Importantly, deletion of htp-3 impairs meiotic break formation, most likely due to its activity in recruiting the MRN/X complex or perhaps a direct role in promoting loading of SPO-11 itself (Goodyer et al. 2008). Previous studies analyzed RAD-51 foci and indirectly posited that induction of DSBs takes place at early meiotic entry (Alpi et al. 2003;Colaiá covo et al. 2003). To determine whether HTP-3 Fig. 1. Phospho-HTP-3 S285 accumulation follows total HTP-3 staining pattern. a) Table showing several of the putative interactors identified by mass spectrometry analysis on the PARG-1::GFP pull downs. b) Annotated fragmentation mass spectrum of the HTP-3 peptide carrying pS285. Red and blue lines indicate b-and y-ions, respectively, y ions with phosphate losses are marked in light red. The illustration was generated using PeptideShaker ver. 2.2.0 (Vaudel et al. 2015). c) Whole-mount gonad stained with panHTP-3 and phosphoHTP-3 S285 antibodies. Scale bar 20 mm. d) htp-3(S285A) mutants stained with panHTP-3 and phosphoHTP-3 S285 antibodies showing specificity of the phospho-antibody. Scale bar 20 mm. e) Insets showing nuclei at different stages of meiotic prophase I. Note that pHTP-3 S285 antibody staining is absent in apoptotic cells (dotted circles). Scale bar 5 mm.
phosphorylation is triggered by CO formation, we analyzed the localization of pHTP-3 S285 in spo-11 mutants. However, removal of spo-11 did not affect pHTP-3 S285 localization, suggesting that phosphorylation and localization of HTP-3 are independent of physiological DNA damage (Fig. 2).
The cyclin homolog COSA-1/CNTD1 is essential to convert recombination intermediates into mature COs, and its depletion leads to nearly complete lack of chiasmata in both worms and mice (Yokoo et al. 2012;Holloway et al. 2014). To determine whether HTP-3 phosphorylation is triggered by CO formation, we analyzed its localization in the cosa-1 mutants. We did not observe any obvious aberrations in pHTP-3 285 localization pattern in the cosa-1 mutants (Fig. 2).
Taken together, these results indicate that phosphorylation of HTP-3 at Ser285 is independent of synapsis and COs, and it does not require SPO-11-mediated DSBs, suggesting that this occurs as an early event at meiosis onset, which most likely takes place either contemporaneously or immediately after axes morphogenesis.

Phosphorylation of HTP-3 S285 is highly redundant
Having confirmed the phosphorylation of HTP-3 at Ser285 in vivo, we then wanted to identify the kinase/s driving this modification. Fig. 2. Phosphorylation of HTP-3 S285 is independent of synapsis and recombination. Mid-pachytene nuclei stained with pHTP-3 S285 and panHTP-3 antibodies in the indicated mutant backgrounds. Scale bar 5 mm.
ERK/MPK-1 phosphorylates HTP-3 in vitro (Das et al. 2020), however, the site has not yet been determined. Therefore, we investigated whether HTP-3 S285 is phosphorylated by ERK/MPK-1. To test this, we performed an in vitro kinase assay using bacterially expressed HTP-3 WT and HTP-3 S285A proteins, employing active ERK2 enzyme (Fig. 3). We observed that the HTP-3 Ser285Ala mutant protein is phosphorylated by active ERK in vitro (Fig. 3a), suggesting that S285 is a phosphor-acceptor for ERK2.
Recent work has shown that the 2 cyclin-dependent kinases CDK-1/2 perform crucial functions in the C. elegans germline, by promoting phosphorylation of the SC component SYP-1 and by exerting a direct regulatory function on pro-CO factors, thus promoting chiasmata formation (Brandt et al. 2020;Haversat et al. 2021;Zhang et al. 2021). We tested whether CDK1/2 mediates pHTP-3 285 phosphorylation using a cdk-1/2-AID-tagged lines, in which degradation of CDK-1/2 was elicited by exposure to auxin. We observed that depletion of either cdk-1 (Fig. 4a) or cdk-2 (Fig. 4b), as well as contemporaneous removal of both kinases (Fig. 4c) did not affect the phosphorylation and localization of HTP-3 S285 suggesting that S285 is likely not a substrate of these 2 kinases.
Altogether, these results indicate that none of the major kinases that have been shown to exert important roles during meiotic prophase I in the C. elegans gonad regulate phosphorylation of HTP-3 S285 , suggesting that this residue may be recognized by a different kinase/s or subjected to a highly redundant phospho-regulation. Fig. 3. HTP-3 S285 phosphorylation is independent of ERK, plk-2, chk-1/chk-2, and the DNA damage kinases atm-1 and atl-1. a) In vitro kinase assay of HTP-3 wild-type and HTP-3 S285A mutants with active recombinant ERK2. b) Mid-pachytene nuclei stained for pHTP-3 S285 and panHTP-3 in plk-2(ok1936). Scale bar 10 mm. c) HTP-3 S285 phosphorylation in ATM and ATR mutants. Dotted circles indicate nuclei with impaired axes morphogenesis. Scale bar 10 mm. d) HTP-3 S285 phosphorylation in chk-1 (RNAi) . Scale bar 10 mm. e) HTP-3 S285 phosphorylation upon auxin-induced degradation of CHK-2. Scale bar 10 mm.

Phosphorylation of HTP-3 S285 is dispensable for chiasmata formation
To determine the biological significance of HTP-3 S285 phosphorylation, we generated an unphosphorylatable HTP-3 by changing the Serine 285 to Alanine (Supplementary Fig. 3) using CRISPR/ Cas9 method. We assessed the hatching rates in the htp-3(S285A) unphosphorylatable mutant worms, as well as monitored the establishment of synapsis and induction/resolution of the recombination intermediates by analyzing RAD-51 dynamics. We found that compared to wild type, there are no defects in either embryonic viability or generation of male progeny (Him phenotype) in the htp-3(S285A) mutant (Fig. 5a), indicating that HTP-3 Fig. 4. Phosphorylation of HTP-3 S285 does not require cdk-1 and cdk-2. HTP-3 S285 phosphorylation in mutant germlines upon loss of cdk-1 (a), cdk-2 (b), or both (c). Scale bar 10 mm. phosphorylation at this site is not essential to preserve fertility. We also performed viability analysis in worms grown at 25 C, known to induce destabilization of the SC and increased lethality in DNA repair-defective mutants. However, we did not observe any significant differences between WT and htp-3(S285A) mutants (Fig. 5a).
Additionally, we did not observe any defects in the establishment of synapsis (Fig. 5b) or in the formation of recombination intermediates (Fig. 5c), as indicated by colocalization between panHTP-3 and SYP-1 and comparable number of RAD-51 foci in the htp-3(S285A) mutants as in WT controls. However, we found representative examples of mid-pachytene nuclei from the indicated genotypes labeled with HTP-3 and SYP-1. Scale bar 10 mm. c) Time-course analysis of RAD-51 foci formation and removal in htp-3(S285A) mutants and wild-type controls. Gonads were divided into 7 equal regions from the mitotic tip to diplotene entry and the number of RAD-51 foci in each nucleus was counted. Bars in the charts indicate mean and standard deviation (ns, nonsignificant, ****P < 0.0001, **P ¼ 0.0072, *P ¼ 0.047 as calculated by T-test). Insets show representative images of early-pachytene nuclei stained with anti-RAD-51 antibodies in the indicated genotypes. Scale bar 10 mm. a small, although statistically significant acceleration in the RAD-51 foci turnover. The RAD-51 foci peaked and disappeared earlier in the htp-3(S285A) mutants compared to the WT animals (Fig. 5c, zones 3-6). This could indicate a role for HTP-3 S285 phosphorylation in regulating the timing of DSB induction or alternatively in influencing the DSB repair kinetics.
Altogether, these data demonstrate that HTP-3 is phosphorylated at Ser285 during meiotic prophase I. It is interesting to note that HTP-3 S285 is not conserved in the Caenorhabditis species, suggesting that its regulation may be specific to C. elegans. Functionally, HTP-3 phosphorylation at Ser285 is dispensable for the execution of the HTP-3-dependent functions during axes morphogenesis and DSB induction. We believe that phosphoregulation of HTP-3 might proceed in a highly redundant fashion, since lack of different kinases did not abrogate phosphorylation at Ser285 and further, the functional requirements imposed by HTP-3 phosphorylation could be shared through multiple accessory sites yet to be identified.

Data availability
The data underlying this article are available in the article and in its online supplementary material. All reagents are available upon request.
Supplemental material is available at G3 online.