HEK293T cell preparationHEK293T cells (ATCC CRL-3216) were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (Gibco, 11995-065) with 1% penicillin/streptomycin (Thermo Fisher Scientific, 15-140-122) and 10% fetal bovine serum (FBS) (HyClone, 89133-098) at 37 °C and 5% CO2. Cell line authentication was performed by the commercial distributor. Culture media was replaced every 24 h. Cells were expanded to appropriate cell number, detached from tissue culture plate with 0.05% trypsin-EDTA (Gibco, 25300062), washed once with phosphate buffered saline (PBS) (Gibco, 02-0119-0500), cell number determined, and 20 × 107 cells pelleted. Cell pellet was immediately stored at −80 °C until use. The cells were low passage number and tested negative for mycoplasma contamination. Frozen cell pellets were resuspended in 5.4 M guanidine hydrocholoride (from Sigma Life Science, 8 M, pH 8.5, G7294-100mL) in 100 mM Tris, pH 8 (Invitrogen, 1 M Tris pH 8.0, 0.2 µm filtered, AM9856) via vortexing, followed by heating in a sand bath for 5 min at 105 °C prior to brief (10–15 s) sonication with a probe sonicator. The sample was diluted with the guanidine buffer above to give a ~1.5 mg/mL estimated protein concentration via NanoDrop (Thermo Scientific) prior to beginning digestion.EGF-stimulated HeLa cell preparationHeLa cells (ATCC CCL-2) were cultured in Dulbecco’s Modified Essential Medium (DMEM) (Gibco, 11995-065) with 10% FBS (HyClone, 89133-098) and 1% penicillin/streptomycin (Thermo Fisher Scientific, 15-140-122) in 37 °C/5% CO2 incubator. Cell line authentication was performed by the commercial distributor. Cells were resuspended in PBS (Gibco, 02-0119-0500) with or without 100 ng/mL human Endothelial Growth Factor (Thermo, AF-100-15) for 15 min at room temperature. Cells were gently resuspended every 5 min during incubation. After incubation, cells were washed twice with PBS and stored in −80 °C until use. Protein extraction was performed as described above for the HEK293T samples. Protein concentrations were estimated via protein BCA (Pierce, 23235). Three biological replicates each were prepared for the control and EGF treatment groups.Yeast cell preparationSaccharomyces cerevisiae S288C-derivative strain, BY4741 (SRD GmbH, Y00000) was cultured in triplicate for ~5 generations into log phase in rich YPD medium at 30 °C. Authentication was performed by the commercial distributor. Cells were centrifuged at 2000 g, 3 min, rinsed in water, transferred to smaller tubes and centrifuged at 1600 g, 3 min, prior to snap-freezing in liquid nitrogen. Frozen cell pellets were resuspended in 8 M urea Sigma-Aldrich, U5378) in 100 mM Tris, pH 8 (Invitrogen, 1 M Tris pH 8.0, 0.2 µm filtered, AM9856) and vortexed with glass beads (425-600 µm, Sigma-Aldrich, G8772-500G) to lyse (2 min of total vortexing with 30 s vortexing followed by 30 s on ice). Protein concentrations were estimated via protein BCA (Pierce, 23235).Mouse tissue preparationAll experiments were performed in accordance with the National Institute of Health Guide for the Care and Use of Laboratory Animals and were approved by the Animal Care and Use Committee at the University of Wisconsin-Madison. Mice were kept on a 12-h light–dark cycle at 23 °C and within a humidity range between 30% and 50%. Mice were fed standard chow diet (Teklad #2018). Six-week-old male C57BL6/J mice (n = 3) were euthanized by cervical dislocation and tissues were immediately collected, and flash frozen in liquid nitrogen. A total of twelve tissues (pancreas, small intestine, spleen, liver, kidney, testes, heart, lung, subcutaneous white adipose tissue (WAT), brown adipose tissue (BAT), gastrocnemius, and brain) were collected. All tissues were stored at −80 °C prior to cryo-pulverization. For each tissue, samples from three mice were pulverized together into a fine power under liquid nitrogen (i.e., one biological replicate per tissue). For each pulverized tissue, ~60 mg frozen wet weight was resuspended in 4 mL of 5.4 M guanidine hydrocholoride (Sigma Life Science, 8 M, pH 8.5, G7294-100 mL) in 100 mM Tris, pH 8 (Invitrogen, 1 M Tris pH 8.0, 0.2 µm filtered, AM9856) with Pierce™ Phosphatase Inhibitor Mini Tablets (A32957) with one tablet/10 mL. Tissue samples were vortexed and sonicated for 20 min in a bath sonicator (chilled to 4 °C) to homogenize. A probe sonicator was used briefly to sonicate samples on ice as needed. The protein concentration was estimated via protein BCA (Pierce, 23235) and samples were diluted with the guanidine hydrochloride buffer above to give a protein concentration of ~2 mg/mL prior to digestion.Protein digestionAfter extracting proteins from human cells, yeast cells, or mouse tissues, methanol (Optima LC/MS grade, Fisher Scientific) was added to 90% (v/v) to precipitate protein and samples were vortexed prior to centrifugation at 4000 g for 15 min. The supernatant was removed, and the pellet was resuspended in 8 M urea (Sigma-Aldrich, U5378), 100 mM Tris (Invitrogen, 1 M Tris pH 8.0, 0.2 µm filtered, AM9856), 10 mM TCEP (Sigma-Aldrich, C4706-2G), 40 mM 2-chloroacetamide (Sigma-Aldrich, ≥98%, C0267-100G) pH 8 at ~1.5 mg protein/mL. Lysyl Endopeptidase (LysC, 100369-826, VWR) was added at a ratio of 1:50 enzyme:protein and gently rocked at ambient temperature for 4 h, followed by the dilution of the solution to 2 M urea with 100 mM Tris, pH 8 (Invitrogen, 1 M Tris pH 8.0, 0.2 µm filtered, AM9856). Promega Sequencing Grade Modified Trypsin (V5113) was added at a ratio of 1:50 enzyme:protein and incubated overnight. Following overnight digestion, the solution was acidified to <pH 2 with 10% trifluoroacetic acid (Sigma-Aldrich, HPLC grade, >99.9%) to quench the digestion. The sample was then centrifuged at 4000 g for 10 min to remove particulate matter prior to desalting with a Strata-X 33 µm polymeric reversed phase SPE cartridge. Peptides were dried via a SpeedVac (Thermo Scientific) and stored at −80 °C until phosphopeptide enrichment or, in the case of the mouse proteomics experiments, until fractionation.Phosphopeptide enrichmentPhosphopeptides were enriched from digested peptides using MagReSyn Ti-IMAC HP beads (ReSyn Biosciences, MR-THP005). A volume of 100 µL beads were used per 1 mg of peptides. Input peptide masses of 2–3 mg were utilized for human and yeast enrichments, and input peptide masses ranging from ~0.3 mg to 1.2 mg were utilized for the different mouse tissues. Beads were washed three times with 1 mL 80% acetonitrile (Optima LC-MS grade, Fisher Scientific)/6% trifluoracetic acid (Sigma-Aldrich, HPLC grade, >99.9%) prior to resuspending the sample in 1 mL 80% acetonitrile (Optima LC-MS grade, Fisher Scientific)/6% trifluoracetic acid (Sigma-Aldrich, HPLC grade, >99.9%) and vortexing the sample with the beads for 1 h. After the 1-h of vortexing, the beads were washed three times with 1 mL 80% acetonitrile (Optima LC-MS grade, Fisher Scientific) /6% trifluoracetic acid (Sigma-Aldrich, HPLC grade, >99.9%), once with 1 mL 80% acetonitrile, once with 1 mL 80% acetonitrile (Optima LC-MS grade, Fisher Scientific)/0.5 M glycolic acid (Sigma-Aldrich, 99%, 124737-500 G), and three times with 1 mL 80% acetonitrile (Optima LC-MS grade, Fisher Scientific). The phosphopeptides were eluted from the beads with the addition of 300 µL 50% acetonitrile (Optima LC-MS grade, Fisher Scientific)/1% ammonium hydroxide (28% in H2O, ≥99.99% trace metals basis, Sigma-Aldrich), followed by a second elution with another 300 µL 50% acetonitrile/1% ammonium hydroxide. The samples were acidified via addition of 15 µL 10% trifluoroacetic acid (Sigma-Aldrich, HPLC grade, >99.9%). The samples were then dried down in a SpeedVac (Thermo Scientific) prior to being resuspended in 0.2% trifluoroacetic acid (Sigma-Aldrich, HPLC grade, >99.9%) and desalted as described above. Desalted phosphopeptide samples were dried and resuspended in 0.1% formic acid (Fisher Scientific, LC-MS grade). The phosphopeptide concentration was estimated via NanoDrop (Thermo Scientific). The HEK293T phosphopeptides were pooled to generate a sample for method evaluation. Each mouse phosphopeptide sample was fractioned as described below.High-pH peptide fractionationHigh-pH fractionation of peptides was performed on an Agilent 1260 Infinity BioInert LC with an automated fraction collector. A 20-min method was performed on a Waters XBridge, Peptide BEH C18, 3.5 µm, 130 Å, 4.6 mm × 150 mm column with a flow rate of 0.8 mL/min. Mobile phase A and B were 10 mM ammonium formate (Sigma-Aldrich, >99/0%, LC-MS grade, 70221-100GF), pH 10 and 20% 10 mM ammonium formate (pH 10)/80% methanol (Optima LC/MS grade, Fisher Scientific), respectively. The gradient went from 0 to 35%B from 0 to 2 min, 35 to 75%B from 2-8 min, 75% to 100%B from 8 to 13 min, followed by washing at 100%B from 13 to 15 min and equilibration at 0%B from 15 to 20 min. UV absorbance at 210 and 280 nm was recorded. For HEK293T spectral library generation and mouse proteomics samples, 16 fractions were collected from 5 to 18 min and concatenated into the final fractions by combining fraction 1 and 9, fraction 2 and 10, etc., resulting in 8 final fractions. For mouse phosphoproteomics samples, 8 fractions were collected from 5 to 18 min and concatenated into the final fractions by combining fraction 1 and 5, fraction 2 and 6, etc. Samples were dried down in a SpeedVac (Thermo Scientific) prior to being resuspended in 0.1% formic acid (Fisher Scientific, LC-MS grade) for LC-MS analysis.Synthetic phosphopeptide dilution series preparationFive sets of phosphopeptide standards were acquired: SpikeMix PTM-Kit 52 (JPT, SPT-PTM-POOL-Yphospho-1), SpikeMix PTM-Kit 54 (JPT, SPT-PTM-POOL-STphospho-1), MS PhosphoMix 1 (Sigma, MSP1L-1VL), MS PhosphoMix 2 (Sigma, MSP2L-1VL), and MS PhosphoMix 3 (Sigma, MSP3L-1VL). The standards were reconstituted in 0.2% formic acid/20% acetonitrile/80% water via vortexing. The standards were pooled into an equimolar mixture of the 225 total phosphopeptide standards. The pooled equimolar mixture was then diluted and mixed with the yeast phosphopeptide sample to construct a dilution series comprised of five points of four-fold dilutions starting at 10,000 amol, resulting in total quantities loaded onto the column of 10000, 2500, 625, 156.25, and 39.0625 amol per phosphopeptide standard along with a constant yeast phosphopeptide load of 250 ng. A summary of all the phosphopeptide standards used in the analysis is provided in Supplementary Data 1.LC-MS operation for phosphoproteomics with Orbitrap Astral analysisNanoflow capillary columns (75 µm I.D., 360 µm O.D.) with pulled nanoESI emitters were packed to 40 cm at high pressures with C18 1.7 µm diameter, 130 Å pore size BEH C18 particles (Waters) as previously described39. A slurry of C18 particles dissolved in chloroform was loaded into a custom-built packing setup with a high pressure pneumatic pump (Haskel) and ultrahigh-pressure capillary fittings (HiP) and packed into the column while slowly ramping up to 30,000 psi, holding at 30,000 psi for 4 h, and then allowed to depressurize slowly. Samples were analyzed with a Vanquish Neo UHPLC (Thermo Scientific) coupled to an Orbitrap Astral mass spectrometer (Thermo Scientific) using a NanoSpray Flex source (Thermo Scientific). A source voltage of 2000 V was used for all experiments. Mobile phase A and B were 0.1% formic in water (Fisher Scientific, Optima LC-MS grade) and 0.1% formic acid/80% acetonitrile (Fisher Scientific, Optima LC-MS grade), respectively. The column was heated to 50 °C with the Column Oven PRSO-V2 (Sonation Lab Solutions) and the flow rate was set to 400 nL/min at the start of the method to decrease delay time and turned to 300 nL/min at the start of the active gradient. Initial conditions of 2%B were ramped to 14% from 0 to 5 min. The active gradient was generally set to 14% to 54%B with curve type 6 beginning at 5.2 min, with the exact %B settings adjusted for each active gradient length to evenly distribute peptide signal across the gradient. The column was washed for 5 min at 100%B and 400 nL/min at the end of the gradient, followed by fast equilibration on the Vanquish Neo LC with an upper-pressure limit of 1100 bar.For initial DDA experiments on the Orbitrap Astral MS, MS1 spectra were collected in the Orbitrap every 0.6 s at a resolving power of 240,000 at m/z 200 over m/z 350–1350 with a normalized AGC target of 300% (3e6 charges) and a maximum injection time of 10 ms. The MIPS filter was applied with Peptide mode and “Relax Restrictions when too few Precursors are Found” set to True. Precursors were filtered to charges states 2-6. A Dynamic Exclusion filter was applied with 10 s duration and 10ppm low and high mass tolerance and exclude isotopes set to True. An intensity filter was applied with a minimum precursor intensity of 5000 required for selection. MS2 scans were collected in the Astral mass analyzer with an isolation window of 0.7 m/z, normalized collision energy of 27, a scan range of 150–2000 m/z, an AGC target of 100% (1e4 charges), and a maximum injection time of 10 ms.For DIA experiments on the Orbitrap Astral MS, MS1 spectra were collected in the Orbitrap every 0.6 s at a resolving power of 240,000 at m/z 200 over m/z 380–980. The MS1 normalized AGC target was set to 300% (3e6 charges) with a maximum injection time of 10 ms. DIA MS2 scans were acquired in the Astral analyzer over a 380–980 m/z range with a normalized AGC target of 500% (5e4 charges) and a maximum injection time of 3.5 ms and an HCD collision energy setting of 27% and a default charge state of +2. Window placement optimization was turned on. Isolation widths of 2 Th and active gradient lengths of 30 min were used with HEK293T fraction analysis for HEK293T spectral library generation. Isolation bin widths of 2 Th and 4 Th with 1 Th overlap were compared for HEK293T phosphoproteomics method evaluation. Isolation widths of 2 Th were used for mouse phosphoproteomics experiments. Comparisons of different DIA m/z ranges and AGC targets are shown in Supplementary Fig. 2D–G.Eight HEK293T phosphopeptide fractions were injected once for library generation. ForHEK293T shotgun experiments, one injection replicate per loading mass was performed for the loading mass experiment. Three injection replicates per method were performed for the gradient length and isolation width experiments. For the phosphopeptide standard analysis, each concentration point was analyzed with three injection replicates. For the EGF treatment vs control HeLa phosphopeptide and proteomics experiments, each of the three biological replicates per condition was injected once. For the HeLa phosphopeptide method evaluation experiments, two injection duplicates were performed per method. For the mouse phosphopeptide analysis, each tissue sample was fractioned into 4 fractions and injected once.LC-MS operation for benchmarking on the Orbitrap AscendNanoflow capillary columns (75 µm I.D., 360 µm O.D.) with pulled nanoESI emitters were packed to 40 cm at high pressures with C18 1.7 µm diameter, 130 Å pore size BEH C18 particles (Waters) as previously described39. See above for column packing description. Samples were analyzed with a Vanquish Neo UHPLC (Thermo Scientific) coupled to an Orbitrap Ascend mass spectrometer (Thermo Scientific) using a NanoSpray Flex source (Thermo Scientific) incorporating a homebuilt column heating compartment. The column was heated to 50 °C. The flow rate was set to 400 nL/min at the start of the method to decrease delay time and turned to 300 nL/min at the start of the active gradient. Initial conditions of 2%B were ramped to 14% from 0 to 5 min. The active gradient was generally set to 14–54%B with curve type 6 beginning at 5.2 min, with the exact %B settings adjusted for each active gradient length to evenly distribute peptide signal across the gradient. The column was washed for 5 min at 100%B and 400 nL/min at the end of the gradient, followed by fast equilibration on the Vanquish Neo LC with an upper pressure limit of 1100 bar.DIA experiments on the Orbitrap Ascend utilized similar method settings to those previously described by Bekker-Jensen et al.41, but with the same m/z range as utilized on the Orbitrap Astral to enable direct comparisons. Briefly, MS1 spectra were collected in the Orbitrap at a resolving power of 60,000 at m/z 200 over m/z 380–980 with a AGC target of 250% and maximum injection time of 123 ms. DIA MS2 scans were collected in the Orbitrap with 15,000 at 200 m/z resolving power, a scan range of 150–2000 m/z, an AGC target of 200%, a maximum injection time of 27 ms, and a collision energy of 30%. A m/z range from 380–980 m/z was iterated through with 14 Th isolation windows with 1 Th overlap.For the phosphopeptide standard analysis, each concentration point was analyzed with three injection replicates. For the EGF treatment vs control HeLa phosphopeptide and proteomics experiments, each of the three biological replicates per condition was injected once.LC-MS operation for proteomics with Orbitrap Eclipse analysisThe chromatography setup described above for the Orbitrap Ascend analysis was utilized for the Orbitrap Eclipse experiments. A source voltage of 2000 V was used for all experiments. Mobile phase A and B were 0.2% formic acid (Optima LC-MS grade, Fisher Scientific) in water (Optima LC-MS grade, Fisher Scientific) and 0.2% formic acid/80% acetonitrile (Optima LC-MS grade, Fisher Scientific), respectively. The column was heated to 50 °C and a flow rate of 300 nL/min was used. Initial conditions of 0%B were ramped to 6%B from 0 to 1 min. The active gradient was set to 6% to 52%B with curve type 6 from 1 to 73 min. The column was washed for 5 min at 100%B, followed by fast equilibration on the Vanquish Neo LC with an upper pressure limit of 1150 bar.For DDA experiments on the Orbitrap Eclipse MS, MS1 spectra were collected in the Orbitrap analyzer every 0.6 s at a resolving power of 240,000 at m/z 200 over m/z 300–1350 with a normalized AGC target of 250% and a maximum injection time of 50 ms. The MIPS filter was applied with Peptide mode and “Relax Restrictions when too few Precursors are Found” set to True. Precursors were filtered to charges states 2-5. A Dynamic Exclusion filter was applied with 10 s duration and 25ppm low and high mass tolerance and exclude isotopes set to True. MS2 scans were collected in the ion trap with an isolation window of 0.5 m/z, normalized collision energy of 25, a scan range of 150–1350 m/z, an AGC target of 300%, and a maximum injection time of 14 ms.Each mouse tissue was fractioned into 8 fractions, each of which was injected once for DDA proteomics.DDA data processingProteomics DDA data was processed in MaxQuant 2.4.9.0 with default parameters40. Phosphoproteomics DDA data was processed with the same version of MaxQuant with default parameters and Phospho (STY) enabled as a variable modification. The “Phospho (STY)Site.txt” was used for analysis with a filter for localization probabilities greater than or equal to 0.75.DIA data processingPhosphoproteomics DIA data was processed in Spectronaut version 17.6.230428.55965 or 18.6.231227. A HEK293T spectral library was generated in Spectronaut (searching against the human proteome database (Swiss-Prot and TrEMBL) downloaded from UniProt on 15-Jan-2023) from the eight HEK293T phosphopeptide fractions analyzed with a 30-min active gradient 2 Th DIA method and the library was used to search the HEK293T phosphoproteomics method evaluation experiments shown in Fig. 2. Note that the entrapment search in Fig. 2G and the phosphoproline variable modification search in Fig. 2H were performed as a library-free search. Mouse phosphoproteomics data was searched with the directDIA mode. Default Spectronaut search parameters were used with a variable Phospho (STY) modification included. The PTMSiteReport file was used for analysis of phosphorylation sites. To count the number of unique phosphorylation sites detected within an experiment, the PTMSiteReport was filtered for rows with “Phospho (STY)” in the “PTM.ModificationTitle” column and values greater than or equal to 0.75 in the “PTM.SiteProbability” column. The data table was then sorted according to “PTM.CollapseKey” and filtered for unique values in the “PTM.Group”. At this stage, there can be rows with the same values of “PTM.CollapseKey” that are not filtered by “PTM.Group” due to PTM grouping differing across multiple files, so the dataset was filtered for unique values of “PTM.CollapseKey” to determine the total number of unique phosphorylation sites measured in a single file. Mouse phosphoproteomics data was searched against the mouse proteome downloaded from UniProt from Swiss-Prot and TrEMBL (downloaded on 25-Aug-2023). For the mitochondrial biology investigation, a separate search was performed with just the Swiss-Prot database to facilitate the analysis.HEK293T phosphoproteomics DIA data was also processed in Proteome Discoverer 3.1.0.638 with CHIMERYS. Numbers of localized phosphosites were reported from the Modification Sites table after filtering for rows with “Phospho” in the Modification Name column. These searches were performed in a developmental version of Proteome Discoverer 3.1 to show that the phosphoproteomic depth achieved with our methods is not restricted to Spectronaut searches. However, this developmental version does not yet incorporate quantitative values for phosphorylation sites detected with DIA methods. Furthermore, from internal conversations, we know the results of this development build should be treated as preliminary and that the algorithms being utilized are still in active development. Consequently, we chose to perform our analyses using Spectronaut, which provides identification, localization, and quantification for DIA phosphoproteomics.Localization error rate calculationTo calculate the localization error rate, a reference sheet of all the possible precursors that could be detected from the phosphopeptide standards was constructed with the phosphorylation sites specified. Precursors with sequences that could also arise from the yeast proteome (based on an in-silico digest) were removed from consideration. Then, the error rate was calculated as \({{{\rm{Error}}}}{\mbox{Rate}}(\%)=\frac{{{{\rm{False\; Sites}}}}}{{{{\rm{False\; Sites}}}}+{{{\rm{True\; Sites}}}}}\times 100\), where “False Sites” is the number of phosphopeptide standard precursors detected with phosphorylation states not indicated in the reference sheet, and “True Sites” is the number of phosphopeptide standard precursors detected with phosphorylation states indicated in the reference sheet. This calculation was performed for each phosphopeptide standard raw data file, and the average error rates are reported as a function of the localization probability cutoff.Phosphopeptide standard quantification calculationsThe phosphopeptide standard reference sheet described above was modified so that each row corresponds to a phosphorylation site, so that the “PTM.Quantity” value in the PTM Site Report could be utilized for quantification assessment. As described above, phosphopeptides with sequences that could arise from the yeast proteome were removed from consideration. Linear regression and R2 calculations were performed using the “pearsonr” and “linregress” functions within the “scipy.stats” module for phosphosites detected across at least three concentration points76.EGF-stimulated HeLa differential expression and pathway analysisDifferential expression analysis for the EGF treatment experiment was performed on all phosphosites that were detected across triplicates in either the EGF-treatment or control group. Missing value imputation was performed in Perseus77 assuming a normal distribution with “width = 0.3” and “down shift = 1.8”. Enriched phosphosites were defined as those with an absolute fold change of more than 2 and a p value less than 0.05 using a two-sided t-test without a multiple testing correction.Pathway enrichment analysis was performed using Enrichr78. The enriched subset contains genes corresponding to phosphosites that are enriched as defined above, and the background was performed for the rest of genes with at least one detected phosphosite.Sequence-based phosphosite clusteringPhosphosite sequences, along with their 5-amino acid flanking regions, were compared using the Blosum62 substitution matrix. The top 4 scores, representing a typical number of amino acids in kinase recognition motifs, were selected. These scores were subsequently averaged and transformed to a range between 0 (indicating identical sequences) and 1 (representing maximally different sequences). Subsequently, a similarity matrix between all sequence pairs was constructed and utilized to generate a 2D embedding space through t-SNE, implemented using the scikit-learn library79. Default parameters were used, except for metric = “precomputed”, init = “random”, n_iter=500, n_iter_without_progress=150, and random_state=42. Additionally, Perseus was employed for visualization purposes77.Kinase motif enrichmentInitially, each phosphorylation site and its flanking region in our dataset was used to search the kinase library website (https://kinase-library.phosphosite.org/site, accessed on February 22nd, 2024) for predictions of all 303 serine/threonine kinases for each site. For kinase motif enrichment analysis, only the top 15 kinases with regards to percentile score for each site were retained. To nominate tissue-enriched phosphorylation sites, we proceeded as follows: for each tissue-tissue comparison the log-transformed abundances of all shared phosphorylation sites were subtracted from each other. For selection of top phosphorylation sites per tissue two different approaches were applied: a z-score threshold of 2 was used, or the top and bottom 200 sites were selected. For the final analysis, the top and bottom 200 ranked sites were used. To calculate the frequency factor for each kinase the proportion of how often a kinase was predicted within the top phosphorylation sites versus the total of top phosphorylation sites was computed, divided by the same ratio for the unchanged sites, i.e., not ranked among the top/bottom 200. A chi-squared test was calculated after applying Haldane’s correction to the contingency table of the two proportions and the p value was extracted.Human variant informationAnnotation of human variants in CPS1 and OPA1 were identified using The Human Gene Mutation Database at the Institute of Medical Genetics in Cardiff (HGMD) using the public site entries (retrieval date 2023-11-1880).Multiple sequence alignmentClustal Omega81 hosted at the EMBL-EBI webserver was used for protein sequence alignment using the following UniProt identifiers for CPS1: P08686, F7EZ, P07756, Q8C196, F1ML89, A0A8C0NKH0; for OPA1: P58281, P58281-2, O60313, P32266; and other dynamin family members: Q811U4, Q80U63, Q8K1M6 (retrieved from UniProt 2023-11-18).Structural modeling predictionThe PyMOL Molecular Graphics System, Version 2.5.7 Schrödinger, LLC. was used for predicting structural models for CPS1 (RCSB PDB: 5DOT (Apo), 5DOU (NAG)66) and OPA1 (RCSB PDB structure 6JTG74). For modeling purposes, the human crystal structures (highly homologous to mouse) were used.Reporting summaryFurther information on research design is available in the Nature Portfolio Reporting Summary linked to this article.
