Phosphoproteomics of ATR Signaling in Prophase I of Mouse Meiosis

During mammalian meiosis, the ATR kinase plays crucial roles in the coordination of DNA repair, meiotic sex chromosome inactivation and checkpoint signaling. Despite the importance of ATR in meiosis, the meiotic ATR signaling network remains largely unknown. Here we defined ATR signaling during prophase I in mice. Quantitative analysis of phosphoproteomes obtained after genetic ablation of the ATR-activating 9-1-1 complex or chemical inhibition of ATR revealed over 12,000 phosphorylation sites, of which 863 phosphorylation sites were dependent on both 9-1-1 and ATR. ATR and 9-1-1-dependent signaling was enriched for S/T-Q and S/T-X-X-K motifs and included proteins involved in DNA damage signaling, DNA repair, and piRNA and mRNA metabolism. We find that ATR targets the RNA processing factors SETX and RANBP3 and regulate their localization to the sex body. Overall, our analysis establishes a comprehensive map of ATR signaling in spermatocytes and highlights potential meiotic-specific actions of ATR during prophase I.

dependent sites (112 sites in Q2) at the S/T-P-X-K remains unknown. Interestingly, we noticed that several components of the piRNA network were enriched for 2 4 2 RAD1-and ATR-dependent phosphorylation within the S/T-P-X-K motif (Fig. 3E). piRNAs are suggesting that TDRD9 may be regulated by kinases in a RAD1-and ATR-dependent regulating the piRNA pathway. Further work will be important to dissect the role of ATR in 2 5 5 regulating piRNA proteins.  Given that ATR preferentially phosphorylates S/T-Q motifs 54,55 , we reasoned that most 2 6 0 phosphorylation sites at S/T-Q in Q2 are more likely to reflect direct ATR substrates in 2 6 1 meiosis. As expected, proteins involved in DNA damage signaling and repair were found to 2 6 2 contain Q2 S/T-Q phosphorylation, including ATR itself, TOPBP1, RAP80, and components have additional phosphorylation sites that did not change with RAD1 CKO or ATR inhibition 2 6 8 ( Fig. 4A-B), indicating that the overall abundance of these proteins was not likely changing. Furthermore, many Q2 proteins contained both an S/T-Q motif and another non-S/T-Q 2 7 0 phosphorylation site that was also ATR-dependent ( Fig. 4A-B). Taken together, these results 2 7 1 reveal a set of proteins that may potentially be direct substrates of ATR in meiosis and define 2 7 2 a set of RNA regulatory proteins subjected to 9-1-1 and ATR-dependent phosphoregulation. Although ATR localizes to the sex body to promote MSCI, it is not known if ATR directly 2 7 6 regulates RNA metabolic proteins to promote silencing or processing of RNAs. To investigate 2 7 7 how meiotic ATR may regulate RNA metabolism we focused on RNA metabolic proteins with 2 7 8 S/T-Q phosphosites in Q2. We found serine 353 in Senataxin (SETX), an RNA:DNA helicase 2 7 9 with established roles in transcriptional regulation and genome maintenance 56 , to be 2 8 0 downregulated upon RAD1 loss and ATR inhibition ( Fig. 4B  Senataxin in meiotic spreads derived from both ATR inhibitor treated and RAD1-CKO mice.

8 9
In accordance with previous work, we found that Senataxin localizes to the sex body at 2 9 0 pachynema control spreads 58,59 . Strikingly, Senataxin accumulation at the sex body was 2 9 1 significantly reduced in pachytene spreads derived from ATRi treated mice ( Fig. 5A-B, S2A-2 9 2 B). While it is difficult to morphologically distinguish the X and Y chromosomes in the Rad 2 9 3 CKO spreads, we observed no enrichment of Senataxin around any selection of signaling at the sex body, implicating Senataxin in promoting ATR signaling 58,59 . Our results 2 9 7 further suggest that ATR-dependent phosphorylation promotes the recruitment or retention of 2 9 8 Senataxin at the sex body, consistent with a model in which Senataxin and ATR act in a feed-2 9 9 forward loop to cooperatively promote their recruitment and efficient sex body formation and Another protein with a S/T-Q phosphorylation site identified in Q2 was RANBP3 (serine 283). RANBP3 is a relatively unknown protein with connections to miRNA and protein export in although one study has found an association with decreased RANBP3 expression and 3 0 6 human infertility 62 . We investigated the localization of RANBP3 in meiotic spreads and found 3 0 7 that in cells derived from wild type or vehicle treated mice, RANBP3 localizes to the sex body proper localization of SETX and RANBP3 to the sex body in pachynema. in mouse testis using phosphoproteomics, these datasets lack experimentally established 3 2 5 kinase-substrate relationships [70][71][72][73][74] . Given the utmost importance of defining the ATR-3 2 6 mediated signaling events in mammalian meiosis to mechanistically dissect its function and 3 2 7 mode of action, here we performed an in-depth phosphoproteomic analysis of ATR signaling 3 2 8 in meiosis. The success of our work mostly relied on a two-part approach for identifying high- confidence ATR-dependent phosphorylation events. By combining the datasets from the 3 3 0 Rad1 CKO genetic mouse model and ATR inhibitor treated mice, we enhanced our 3 3 1 confidence in identifying meiosis-specific ATR functions. As validation of our dataset, we 3 3 2 detected known ATR targets such as MDC1 and TOPBP1 in the set of ATR and RAD1-3 3 3 dependent signaling events. Additionally, we observed the expected enrichment for DNA 3 3 4 metabolism, DNA repair and cell cycle gene ontology categories. We anticipate that this 3 3 5 database will be a useful resource for the meiosis community for further study into the 3 3 6 mechanisms of meiotic ATR activity. The set of ATR-mediated signaling events detected included phosphorylation in proteins 3 3 9 involved in RNA metabolism, such as Senataxin and RANBP3. It is tempting to speculate that 3 4 0 these are direct targets since they were phosphorylated at the S/T-Q motif and their 3 4 1 localization to the XY body was compromised during ATR inhibition. We propose that these 3 4 2 data highlight a need for ATR in regulating distinct aspects of RNA metabolism such as R- piRNA network were also found to be regulated in an ATR-dependent manner. The the XY is inextricably linked to prophase I progression, it is likely that the connection of 3 5 0 meiotic ATR signaling to RNA metabolism is even more relevant compared to its mitotic signaling. An interesting model to be explored in future work is that SETX and RANBP3 may 3 5 2 both play a key coordinated role in removing RNA from XY DNA, and exporting it, to establish 3 5 3 MSCI ( Fig. S3F). Further work will be needed to establish these as direct ATR substrates and 3 5 4 to dissect the mechanism by which ATR promotes the localization and action of Senataxin, 3 5 5 RANBP3 and other RNA metabolic proteins identified in this study. Since our phosphoproteomic is unbiased and not only directed at the preferred S/T-Q motif, 3 5 8 we were able to capture a range of phosphorylation events in other motifs suggesting that ATR regulates multiple downstream kinases during meiosis. Strikingly, we observed a strong 3 6 0 enrichment for ATR-dependent phosphorylation sites at the S/T-P-X-K motif, suggesting that predict that these kinases are activated by ATR during prophase I, which could be tested by 3 6 6 future phosphoproteomic analysis of testes from mice treated with inhibitors for these 3 6 7 kinases. It is worth mentioning that in mitosis, the canonical action of ATR in promoting DNA 3 6 8 damage checkpoint, and consequent cell cycle arrest, is mediated via inhibition CDK activity, 3 6 9 and consequent reduction in S/T-P phosphorylation sites 76 . In this sense, the observed 3 7 0 dependency of S/T-P-X-K motif for ATR in meiosis is the opposite to what would be predicted 3 7 1 from mitotic cells. Since the high activity of ATR in meiosis does not result in meiotic arrest, 3 7 2 but is actually required for meiotic progression, it is possible that our data is revealing a of ATR-and RAD1-dependent sites. We cannot exclude that the presence of other motifs in 3 7 6 Q2 is due to the function of ATR regulation of phosphatases. Importantly, several 3 7 7 phosphatases, including PPM1G and PP1R7, were also identified in Q2. Overall, our work represents an initial attempt to reveal the scope of targets and processes ATR may be activated in a 9-1-1-independent manner, via ETAA1. Therefore, additional 3 8 5 phosphoproteomic analyses from mouse mutants/CKOs of ETAA1 are likely to identify 3 8 6 different subsets of ATR targets that may represent different modes of ATR signaling in 3 8 7 meiosis. Another key outstanding question is to understand how the ATR kinase, which 3 8 8 imposes cell cycle checkpoints in most other cell types, is so highly activated in 3 8 9 spermatocytes without inducing cell cycle arrest. A potential explanation may lay at the 3 9 0 specificity of ATR's action at the sex body, which may be devoted to the regulation of 3 9 1 checkpoint-independent processes such as the control of RNA processing during meiotic 3 9 2 prophase I, as supported by our data. Finally, there are medical implications of understanding 3 9 3 ATR signaling in meiosis, since many ATR inhibitors are currently in phase 2 clinical trials for 3 9 4 cancer treatment and determining the impact of these inhibitors in meiotic cells will be 3 9 5 relevant to define the effects of these treatments in patient fertility. and 50% water. Wild-type C57BL/6 male mice aged to 8 weeks-old were gavaged with 4 1 2 50mg/kg of AZ20 (Selleckchem) and euthanized at indicated time points. Specific timepoints 4 1 3 examined in this study include collection after 3 days of 50mg/kg, 2.5 days of 50mg/kg per 4 1 4 day or 4 hours after one dose of 50mg/kg AZ20. All mice used for this study were handled  Whole, decapsulated testes were collected and frozen at -80 o C from 8 week-old AZ20 and 4 2 0 vehicle-treated C57BL/6 mice 4 hours after treatment as indicated. Whole, decapsulated 4 2 1 testes from Rad1 CKO and littermate control mice were collected at 8 or 12 weeks of age. aliquots (10%, 30%, 30% and 30%) and dried in silanized glass tubes. For each experiment, 4 3 8 the three 30% aliquots from each control and AZ20-treated or control and RAD1-KO samples  Dried TMT-labeled phosphopeptides were ressuspended in 16.5uL water, 10uL formic acid Bioscience), using a three-solvent system: buffer A (90% acetonitrile), buffer B (75% 4 5 2 acetonitrile and 0.005% trifluoroacetic acid), and buffer C (0.025% trifluoroacetic acid). The 4 5 3 chromatographic runs were carried out at 150uL/min and gradient used was: 100% buffer A 4 5 4 at time = 0 min; 94% buffer B and 6% buffer C at t = 3 min; 65.6% buffer B and 34.4% buffer 4 5 5 C at t = 30 min with a curve factor of 7; 5% buffer B and 95% buffer C at t = 32 min; isocratic UltiMate™ 3000 RSLC nano chromatographic system coupled to a Q-Exactive HF mass    Meiotic spreads were produced as previously described 78 . Briefly, decapsulated testes Triton-X for 10 minutes and blocked with 10% antibody dilution buffer. Secondary antibodies 5 0 7 were diluted as indicated and incubated on slides with a parafim strip at 37 o C for 1 hour.

0 8
Slides were then washed in 0.4% photoflo/PBS for 10 minutes twice followed by 0.4% 5 0 9 photoflo/H 2 O for 10 minutes twice and allowed to dry before mounting with DAPI/antifade. Slides were imaged on a Leica DMi8 Microscope with a Leica DFC9000 GTC camera using the sex body and mean intensity of the underlying pixels was recorded. Additionally, two ROI 5 1 7 lines of equal length were placed over two autosome cores and the mean pixel intensity was 5 1 8 also recorded to serve as an internal control for background florescence. The sex body ROI 5 1 9 intensity was then normalized to the average of the two autosomal ROI intensities for each

Enrichment in Q2 Depletion in Q2
Center Q2 3.3% A) Immunofluorescence of meiotic chromosome spreads with SETX (green) and SYCP3

Q2 vs Center
(red) from mice collected 4 hours after 50 mg/kg treatment with AZ20 or vehicle. B) Quantification of pachytene spreads from 5A (4 vehicle mice; n=237 cells; 4 ATRi mice; n=283 cells p=0.00435 measured by student's t-test

Supplemental Figure 1. yH2AX is reduced at the sex body after 4 hours of 50
mg/kg treatment with AZ20 and gene ontology of quadrants.
a) quantification mean intensity of the ratio of yH2AX signal as depicted in Figure 1C separated by individual animal replicates. yH2AX intensity is measured as described in methods. Datapoints indicate the ratio of signal intensity across sex body to average of intensity across two autosomes for an individual pachytene-stage meiotic spread. b) example of pachytene spreads showing variation in signal intensity and pattern from quantification in S1A with yH2AX (green) and SYCP3 (red). c) scatterplot identifying peptides in each quadrant with phosphopeptide count (bar graph, lower left) and corresponding top five STRING analysis categories in d) bar graph.

Supplemental Figure 2. pMDC1 signal is lost at the sex body upon ATR inhibition.
a) Schematic of detected phosphorylation sites in MDC1 b) quantified pMDC1 intensity from 2G separated by individual mice after treatment with vehicle or ATR inhibitor for 4 hours as indicated. Quantification was done as described in methods. c) example meiotic spreads depicting variation in signal intensity and pattern for pMDC1 (green) and SYCP3 (red) for ATRi and vehicle-treated mice. a) quantification of SETX signal at the sex body separated by individual mice with c) example images from vehicle or ATR inhibitor treated mice collected 4 hours after 50 mg/kg treatment with AZ20 or vehicle. c) quantification of SETX at sex body or sex chromosomes of control or Rad1 CKO mice, respectively. d) example spreads of SETX (green) and SYCP3 (red) staining. Rad1 CKO 'pachytene-like' stage was defined as having 3 or more fully synapsed autosomes. See methods for more details on quantification.

Supplemental Figure 5. RANBP3 signal is lost at the sex body upon ATR inhibition.
a) Quantification RANBP3 intensity from meiotic spreads separated by animal with b) example spreads from mice collected 4 hours after 50 mg/kg treatment with AZ20 or vehicle. c) quantification of RANBP3 at the sex body or sex chromosomes body of control Based on the model, ATR directly phosphorylates proteins involved in RNA metabolism, such as SETX and RANBP3, to promote their robust localization to the sex body. In this scenario, sex body-localized SETX and RANBP3 would contribute to ATR-mediated silencing by favoring the disengagement of mRNA from XY chromatin and mRNA export from the sex body. ATR also regulates the activity of additional kinases that in turn phosphorylate piRNA biogenesis-related proteins that may also contribute to sex body morphology and function. Collectively, ATR signaling promotes a range of events contributing to sex body formation.