Wildtype zebrafish (AB strain) were bred and kept in the University of Ottawa Aquatics core facility at 28°C on a 14:10h light/dark cycle. All the fish were fed twice a day. Sexually mature zebrafish were anaesthetized using buffered 4.5% tricaine, then sacrificed by spinal transection. Brain, pituitary and gonads from 42 females and 12 males were dissected and immediately transferred into 1.5 mL centrifuge tubes containing 200 µL of homogenization solution on ice. To homogenize the tissues, a probe homogenizer was utilized at a power output of 10 Watts for 20 seconds. The homogenization solution dissolves the peptide and steroid hormones, while precipitating the proteins to facilitate analysis. After centrifugation, the resulting homogenate was separated into pellet and supernatant fractions, which were then evaporated using a speed-vac.

Dried pellets from the tissue homogenates were solubilized in 200 µL digestion buffer. Dithiothreitol was added to a final concentration of 10 mM and incubated at room temperature for 1hr for reduction. Iodoacetamide was added to a final concentration of 20 mM and incubated at room temperature for 40 min in dark for alkylation. In the end, the sample was diluted 5-fold by 50 mM ammonium bicarbonate and digested by 1 µg trypsin/20 µg sample protein overnight at room temperature.

The solid phase extraction (SPE) plates were prepared for both peptide-steroid co-extraction and protein hydrolysate extraction by packing 10 mg of extraction sorbent into filtered 96-well plates. Prior to extraction, the pre-packed extraction plates were activated by adding 200 µL of activation solution for 5 times, followed by 5 times 200 µL of equilibrium solution. The samples were dissolved and loaded in the equilibrium solution, and the loaded SPE plates were washed another 5 times with 200 µL of the same solution. The elution was performed by adding 20 µL of elution solution for 6 times. The activation elution solutions were flushed through the packed SPE bed by centrifugation at 500 rcf, while the equilibrium solution was flushed at 1000 rcf. The stepwise procedure is presented in Supplementary File 2 .

For both untargeted proteomics and peptidomics, the system consisted of an Agilent 1100 micro-HPLC system coupled with a Thermofisher LTQ-Orbitrap mass spectrometer equipped with a nano-electrospray interface. Peptide separation was performed on an in-house packed 50 μm × 100 mm analytical column. For proteomics, the gradient was from 5% to 35% mobile phase B in 120 min. The mass spectrometry (MS) method consisted of one full MS scan, followed by data-dependent MS/MS scan of the 5 most intense ions using the LTQ-Orbitrap mass spectrometer. The full MS scan range was 150 to 1500 m/z for peptides and steroids, and 300 to 1700 m/z for digested samples. To improve mass accuracy, all measurements in the Orbitrap mass analyzer were performed with internal recalibration (“Lock Mass”). On the Orbitrap, the charge state rejection function was enabled, with single and “unassigned” charge states rejected for untargeted experiments. The raw files generated by the LTQ-Orbitrap were processed and analyzed using MaxQuant, Version 1.2.2.5 with the Uniprot protein FASTA database, including commonly observed contaminants. The following parameters were used: cysteine carbamidomethylation was selected as fixed modification; methionine oxidation, protein N-terminal acetylation, and enzyme specificity was set to trypsin. Up to two missing cleavages of trypsin were allowed to compensate insufficient enzyme digestion. Precursor ion mass tolerances were 7 ppm, and fragment ion mass tolerance was 0.8 Da for MS/MS spectra. Identified peptide sequences from one protein that were equal to or contained within another protein’s peptide set were grouped together and reported as one protein group. The false discovery rate (FDR) for peptide and protein was set at 1%, and a minimum length of six amino acids was used for peptide identification. Quantification was performed using normalized label-free quantitation (LFQ) intensity.

Untargeted peptidomics followed a similar analysis pipeline. However, we selected methionine oxidation, protein N-terminal acetylation, and no enzyme digestion as modifications, and cysteine carbamidomethylation was not fixed.

For targeted peptide and steroid quantification, the HPLC was run in micro mode with a pre-column splitter on load and elute modes, using an in-house-prepared 50 µm × 150 mm analytical column with 3 µm C18-AQ sorbent from Dr. Maisch. The sample was loaded on the column using 95% mobile phase A (A1 for positive mode, A2 for negative mode) at a flow rate of 6 μL/min for 3 min. Then, we used a gradient of 20-90% B1 for positive mode analysis and a gradient of 50-90% B2 for negative mode analysis. We set the mass spectrometer to run on data-dependent acquisition, with the orbitrap analyzer performing the master scan at resolution 60,000 (FWHM), and the linear ion-trap analyzer performing the fragmentation and secondary MS scan according to the mass list and master scan. The mass list collects all primary ions of analytes of interest. We used Thermo Xcalibur software, version 2.2, for data acquisition and analysis. We confirmed peptide structures by three MS2 ions and steroid structures by one MS2 ion. We plotted chromatograms by accurate MS1 with 10 ppm mass tolerance and 120 sec identification windows, using Genesis peak detection when there were over five data points to form a chromatographic peak. We determined the level by comparing the peak area to the standard curve generated by standard compounds. Steroid standards were commercially available. Peptide standards were prepared by solid phase peptide synthesis and purified by HPLC.

Standard peptides were synthesized in-house using Fmoc peptide synthesis chemistry. Solid-phase synthesis was carried out on Tentagel resin. For deprotection of the Fmoc group on the N-terminus of each amino acid, 2 times 5 min 20% piperidine treatment was employed. Standard coupling reactions were performed with 4-fold HBTU (0.5 M), 4.4-fold amino acids (0.6 M), and 4-fold NMM (4 M) for 40 mins. Double coupling reactions were used to attach each amino acid for sequences with a length of 15 or below, while triple coupling reactions were used for sequences with a length greater than 15.

At the end of synthesis, the resin was treated with a cleavage cocktail for two hours to harvest crude peptides. To introduce cysteine into the peptide sequences of zebrafish oxytocin and vasopressin, Fmoc-Cys(Acm)-OH was used. The Acm-protected cysteine was oxidized on-resin by suspending the synthesis resin in a 4 M Iodine/DMF solution with mild mixing for 2.5 h, followed by washing with DMF 5 times and with DCM 3 times before cleavage. To introduce pyroglutamic acid (Pyr), Boc-Pyr-OH was used with a standard coupling protocol. The crude peptides were precipitated with ethyl ether on ice, dissolved in 50% ACN water, and lyophilized. Lyophilized crude peptide samples were stored at -20°C for purification.

Peptide purification was performed on an Agilent 1100 HPLC system with Luna Omega 5 µm 100 Å PS C18 RP-HPLC column (250×10 mm) at a flow rate of 4 ml/min. The column oven was set at 50°C. To purify each peptide, the HPLC gradients were adjusted to have the target peptide elute at least 7.5 min after the gradient started. The fraction collector was set to start collection 7.5 min after the gradient started on 400 mAU threshold, 500 mAU/s upslope, and 200 mAU/s downslope according to the 214 nm chromatogram. The collected fractions were chilled in liquid nitrogen for 5 min and lyophilized accordingly. Lyophilized purified peptide powders were stored at -20°C. The molecular weight of each peptide was confirmed on LC-MS.

A total of 8 peptide and 5 steroid hormones were selected to evaluate the possibility of targeted analysis in unpooled, single tissues harvested from multiple females and males. The supernatants were resuspended and subjected to SPE using in-house prepared cartridges, as described in the methods and Table 1 . Peptide standards were prepared in-house with Fmoc/tBu chemistry, purified, and confirmed by MS as described above. The level of each analyte was determined by comparing the chromatogram peak areas to standard curves. Chromatograms of 3 representative analytes (vasotocin, also known as fish arginine vasopressin (AVP), gonadotropin releasing hormone 3 (GNRH3) and testosterone) are shown in Figure 1C , and chosen for their diversity in size, structure, and hydrophilicity/lipophilicity. The limits of detection and maximum amounts of these and other analytes measured from tissues of both females and males are listed in Table 2 . This new analysis strategy provides the possibility to measure diverse peptide and steroid hormones in tissue samples as small as a zebrafish pituitary. The performance of the in-house packed SPE cartridge was also tested with the LC-MS/MS system, and the overall recovery rates for the 3 representative analytes were over 85% on 0.01-10 nmol loading.

To perform untargeted peptidomics, we applied the peptide-steroid extracts from pituitary to LC-MS/MS analysis using a shallow gradient and data dependent MS acquisition. We used the acquired MS/MS spectra to search against the zebrafish International Protein Index protein sequence database (Version 3.85) with Maxquant and the LFQ option. Our analysis successfully identified 260 different peptides derived from 168 different proteins ( Supplementary File 1 , Figure 1D ).

The protein pellets were subjected to reduction, alkylation, and digestion, and the extracts were further processed using the same SPE procedure as described in the Supplementary. The LC-MS/MS analysis and data processing followed a similar protocol, with trypsinization and alkylation being used as the default methods. Analysis of the resulting data revealed the identification of 1758 proteins, of which 93 were found to be shared with the peptidomics results ( Supplementary File 1 , Figure 1D ).

We conducted gene set enrichment analysis (GSEA) using precursor proteins and associated peptides from both proteomics and the peptidomics results ( Supplementary File 1 ). This analysis led to the identification of 393 pathways from the pituitary protein lysate (P<0.05) and 202 pathways from the peptide (P<0.05) datasets ( Supplementary File 1 ). This approach allowed us to investigate biological processes and molecular pathways more comprehensively, providing valuable information for further exploration and interpretation of our results. Interestingly, 123 pathways were discovered in both peptidomics and proteomics experiments, suggesting their importance in the same pituitary tissue. The details of all pathways are in Supplementary File 1 , and specific examples of hormonal pathways enriched in both datasets are presented in Table 3 . Given the importance of the pituitary hormones in homeostasis and physiological regulation, detection of some of these were expected.

For example, pathways related to proopiomelanocortin (POMC) production, secretion and action are well represented in the data ( Figure 2 , Table 3 ). A lesser studied aspect of pituitary function is that of leptin action via STAT, ELK/SRF expression targets ( Table 3 ). In fish, leptin can regulate prolactin (PRL) (11) and the gonadotropins (12). We also noted that enzymatic processing of classical hormones such as PRL can give rise to angiogenic peptides (13, 14) that were also found in the pituitary gland.

A key example of the utility of the methodology is the detection and quantification of the emerging reproductive hormone secretoneurin (SN). It is derived from prohormone convertase-mediated processing of the larger secretogranin-2 (SCG2) precursor which was detected in the proteomic samples ( Figure 2 ). In zebrafish and other teleosts there are 2 precursor forms, SCG2a and SCG2b, respectively giving rise to SNa1-34 and SNb1-31 which we synthesized and then quantified in brain, pituitary and gonads of both females and males ( Table 2 ). TALEN-mediated frameshift mutation of the zebrafish SCG2a and SCG2b genes led to reduced levels of sexual activity, suboptimal ovulation and egg laying, fertility, and embryonic survival 24-hr postfertilization (10). The SN peptides stimulate LH production and release (15). Gene set enrichment analysis (GSEA) was conducted in Pathway Studio 12.0 (Elsevier). A total number of 2,095 proteins (proteomics) and 202 proteins (based on peptidomics) were mapped into the program for zebrafish using the official gene name. GeneCards (www.genecards.org/) was used to manually retrieve mammalian homologs for mapping. Gene set enrichment analysis proceeded with 1000 permutations to generate the distributions for statistical testing (permutation test) using a Kolmogorov-Smirnov Distribution test. Parameters were as follows: p-value ≤ 0.05, maximum number of networks shown in result = 1000 with a minimum network overlap of 5 entities. The enrichment identifies cell processes and signaling pathways most represented in the data sets.

We also analyzed pituitary protein lysates using targeted proteomics to investigate the presence of specific proteins, namely the subunits of luteinizing hormone (LH). Critically, LH is a dimeric pituitary hormone responsible for driving ovulation in all vertebrate animals (16) and inappropriate LH secretion can lead to various fertility problems (7). It is traditionally measured by radiometric or enzyme-liked immunoassays, with associated challenges (3, 17, 18). There is currently no specific immunoassay for zebrafish LH. As an alternative, we optimized proteomic digestion steps and then identified two fragments from the glycoprotein hormones alpha polypeptide (CGA) (CGA 48-62 and CGA 27-38) and two fragments from the luteinizing hormone subunit beta (LHB) (LHB 9-29 and LHB 30-43) for the detection of LH ( Figure 1B ).