From May 27 - June 18, 2021, marine microbial samples were collected in East Sound, WA, USA every 4 hours. This high temporal resolution sampling effort captured the initiation of two separate Chaetoceros socialis blooms. A metaproteomic data independent analysis was performed on the 6 days of the time series preceding the second bloom (June 12 – 18). A subsequent targeted analysis using parallel reaction monitoring was performed on the time period preceding the first bloom (May 28 – June 2). Prior to the PRM analysis, a matrix-matched calibration curve was analyzed to assess if peptides from a representative microbiome sample could be accurately quantified.

Sample collection and processing Samples were collected near Orcas Island in East Sound, WA, USA from May 27 - June 18, 2021, by pumping water from 100 m offshore (Lat: N 48o .40.590', Long: W 122o 52.994') at a constant depth of 2 m and flow rate of 5 L min-1. Whole water was filtered through a 200 µm nylon mesh followed by three subsequent filters at 100 µm, 10 µm, 1 µm before final collection of the bacterial fraction of the water column onto a 0.22 µm polyethersulfone (PES) 47 mm filter. Final metaproteomic samples were stored long term in -80°C on 0.22 µm polyethersulfone (PES) filters. Sample filters (0.22 µm PES) containing the bacterial fraction of the microbiome were manually agitated within a 3x4 cm Ziplock baggie with 100 μl S-Trap solubilization buffer (5% SDS, 50 mM TEAB, 2 mM MgCl2, 1X HALT protease and phosphatase inhibitors in nanopure water) to ensure equal spreading of the buffer across the filter surface. The cell lysate was collected via pipette in a 2 mL microfuge tube. This was followed by additional rinse of the filter with 100 μl nanopure water followed by manual agitation and immediate transfer of solution to pool fractions in the 2 mL tube. Two additional rinses of the filter with 100 μl nanopure water without manual agitation were performed and the solution added to the pool. The resulting 400 μl whole cell solution was then sonicated (Branson 250 Sonifier; 20 kHz, 30 × 10 s on ice). Samples were then evaporated in a SpeedVac (ThermoFisher Scientific SRF110) at room temperature to reduce the volume to 100 µL (5% SDS). Protein concentrations for each sample were quantified in triplicate using a Bicinchoninic Acid protein assay kit (Pierce Thermo Scientific) using a microplate reader. Enolase protein standard (0.16 μL 100 ng μl−1 enolase per 1 μg protein) was added to each sample to track digestion efficiency. Each sample (20 µg protein) was purified according to manufacturer’s recommendations (Protifi S-Trap micro protocol). Proteins were digested with Promega modified trypsin (2 μg for 1:10 ratio, 1 hour 37°C). Resulting peptides were evaporated to dryness and resuspended in 2% acetonitrile (ACN), 0.1% formic acid with final concentration of 0.5 μg protein μl−1. MS methods All samples were analyzed on an Orbitrap Lumos Fusion Tribrid instrument with an inline EASY-nLC 1200 (ThermoFisher Scientific). Reverse phase chromatography was achieved using a manufactured PicoTip fused silica capillary column (75 μm i.d., 40 cm long) packed in-house with C18 beads (Dr. Maisch ReproSil-Pur; C18-Aq, 120 Å, 3 μm) and maintained at 50°C. The analytical column was fitted with a 4 cm long, 150 μm i.d. in-house packed kasil-frit precolumn (Dr. Maisch ReproSil-Pur; C18-Aq, 120 Å, 3 μm). Peptides were eluted using an acidified (formic acid, 0.1% v/v) water-acetonitrile gradient (2-45% acetonitrile). Sample analyses on the MS were randomized to reduce batch effects and quality control (QC) peptide mixtures were analyzed every sixth injection to monitor chromatography and MS sensitivity. Skyline was used to evaluate the consistency of the resulting peak areas and retention times of QC standards through all analyses. Data independent acquisition Samples from June 12, 2021 (17:00) to June 18, 2021 (13:00) were analyzed using DIA mode. For each DIA MS experiment, 4 µg digested peptides with 7 fmol PRTC were analyzed. Centroid full MS resolution data was collected at 120,000 and MS2 data was collected at a resolution of 15,000 with an AGC target of 4 ×105 across isolation windows of 8 m/z. Prosit (Gessulat et. al., 2019) was used to process a DDA-refined metagenomic database created via whole genome sequencing of a pooled sample of DNA from every sample in the time series (DDA data: PRIDE identifier PDX052873) using an HCD intensity and ion retention time prediction model according to methods from Searle et. al., 2020. Six gas phase fractionations (GPFs) spanning 100 m/z for each analysis were completed with narrow isolation window (4 m/z) to generate a chromatogram library using Encyclopedia v. 1.4.01 following Searle et. al., 2018 and Pino et. al., 2020. The 36 individual time series MS experiments were searched against the resulting chromatogram library using Encyclopedia v. 1.12.31. Encyclopedia-exported results were visualized using Skyline-daily version 23.0.9.187 (MacLean et. al., 2010), including quality assurance that every peptide contained at least three quantifiable transitions and external standards (PRTC peptides) had a coefficient of variation below 20%. Quantitative values for each peptide were generated using Skyline which calculated the area under the curve (AUC) for each peptide after total ion current (TIC) normalization using Encyclopedia-inferred peak boundaries. Parallel reaction monitoring Samples from May 28, 2021 (01:00) to June 03, 2021 (01:00) were analyzed using PRM mode. The targeted peptide list was generated using the Skyline analysis of the DIA data from June 12 – 18. Peptides were trapped and separated on a near identical chromatography setup detailed earlier with the exception of the analytical column length (30 cm). Mass spectrometry experiments were completed with 2 µg digested peptides with 150 fmol PRTC. Peptides were eluted using the gradient above followed by a 30 minute wash with 80% ACN and 10 minute re-equilibration with 3% ACN. Targeted ions were selected for MS2 acquisition from precursor ion scans of 400–1000 m/z with an isolation window of 1.6 m/z, flow rate of 300 nL/min, and resolution of 30,000. Centroid full MS and MS2 data was collected with an AGC target of 4 ×105. Data was analyzed with Skyline software to identify peptide peaks and chromatographic time boundaries. Matrix-matched calibration curve A pooled sample consisting of 2 µg aliquots of digested peptides from the time series samples was used as a representative bacterial microbiome for constructing a matrix-matched calibration curve. This microbiome sample was serially diluted with a matched matrix of digested peptides from the marine bacterium Colwellia psychrerythraea 34H. The resulting dilution series consisted of six microbiome-to-matrix ratios that spanned the expected dynamic range of targeted peptides: 0%, 0.3%, 3%, 50%, 70%, 100% microbiome. These samples were analyzed using PRM as per the methods above.

Time series of the bacterial fraction of an ocean microbiome across two separate six day periods (PRM on 36 samples from May 28 – June 2; DIA on 36 samples from June 12 - 18). Additional matrix-matched calibration curve analysis by PRM using a representative pooled sample of the time series.