Sample Descriptions:
See Supplemental File "Field Sampling Info" (PDF) for the dates and locations of collection of the four field samples. In short:
Cyano = a cyanobacteria sample provided by Matt McCarthy lab (UCSC, δ15N-Phe and δ15N-Glu are 7.41±0.83‰ and 7.64±0.62‰, respectively).
Zoop = a zooplankton sample collected near Bermuda.
LIS = a sinking particle sample collected by a sediment trap in Long Island Sound, NY.
GoCA = a surface sediment sample collected from the Gulf of California under the oxygen-deficient zone.
Std1-4 = mixtures of 16 amino acids with different δ15N-Phe and δ15N-Glu values (see Supplemental File "Std1-4 Info" (PDF))
M-std = a standard mixture of 15 amino acids with known δ15N values from the Matt McCarthy lab (UCSC, δ15N-Phe and δ15N-Glu are 9.17‰ and -4.13‰, respectively).

The methodology for nitrogen isotopic compositions (d15N) analysis in amino acids (AA) employed here is comprised of four main steps: (1) sample preparation and amino acid extraction, (2) amino acid separation and collection, (3) conversion of amino acid to nitrite, and (4) conversion of nitrite to nitrous oxide (N2O) for nitrogen (N) isotope analysis using purge-and-trap continuous-flow isotope ratio mass spectrometry (PT/CF/IRMS).

Four mixtures of 16 proteinaceous AAs were created to identify AA in natural samples, as well as to quantify their concentrations and d15N values. USGS L-glutamic acid (Glu): USGS40 (-4.52 ‰) and 41a (+47.55‰) were mixed by weight with proportions of 62% vs 38% and 34% vs 66%, respectively, to create four different d15N values for Glu (United States Geological Survey, Reston, VA, USA). A series of phenylalanine (Phe) standards were created to cover a broad natural-abundance range in d15N adapted from a previously published procedure (Zhang and Altabet 2008), which were made by mixing isotopically labeled L-Phe (15N>98%, Cambridge Isotope, Cambridge, MA, USA) with unenriched Phe (Alfa Aesar, Haverhill, MA, USA). The d15N of these Phe at four different isotopic levels were determined using an adapted persulfate oxidation method in the Altabet Lab (Bronk et al., 2001, Knapp et al., 2005). These Glu and Phe were mixed with L-alanine (Ala), L-arginine (Arg), D/L-aspartic acid (Asp), L-cysteine (Cys), glycine (Gly), L-histidine (His), L-isoleucine (Ile), L-leucine (Leu), L-lysine (Lys), D/L-methionine (Met), D/L-serine (Ser), D/L-threonine (Thr), L-tyrosine (Tyr), and D/L-valine (Val). The AA in the mixtures were made with a concentration of 2.5mM except for Glu and Asp, which were at 5mM, respectively. The detailed isotopic composition of Phe and Glu in each standard mixture can be found in the Supplemental File "Std1-4 Info" (PDF).

Sample analyses took place between January and August 2020.

(1) Sample preparation and amino acid extraction:
The extraction of AA from proteinaceous samples followed previously established procedures (e.g., Zhang et al., 2021). Various amounts (Table SI-2 of Zhang et al. (2021)) of cyanobacteria, zooplankton, sinking particles, and sediment samples were placed in 20 milliliter (mL) glass vials with 10mL 6 M hydrochloric acid (HCl) (Thermo Fischer Scientific, Waltham, MA, USA), were flushed with nitrogen gas, sealed, and hydrolyzed at 110 °Celsius for 20 hours.

(2) Amino acid separation and collection:
Separations and collection of AAs were conducted using an ICS-5000+ Ion-change chromatography system (Thermo Fischer Scientific, Waltham, MA, USA). Baseline separation of 9 AAs (Arg, Lys, Met, Ser, Phe, His, Glu, Asp, and Tyr) was achieved on a CarboPac PA 10 semi-preparative scale column (9×250 mm, 10 micrometer (μm) particle size, pore size of < 10 angstroms (Å); Thermo Fischer Scientific, Waltham, MA, USA). Each sample injection was set at 25 microliters (μL) and the column compartment was set at a constant 30°C. The mobile phases used were (A) MilliQ water, (B)1M NaOH in MilliQ water, and (C)100mM NaOH in MilliQ water, and programmed as follows for each run: 0 to 30 min (A: 80%, C: 20%), 35 to 52 min (A: 50%, C: 50%), 55 to 62 min (A: 85%, B: 15%), 65 to 75 min (A: 75%, B: 25%), 80 to 140 min (A: 70%, B: 30%).

At the end of each run, 1M sodium acetate (NaOAc) was used to flush the column for 20 minutes to clean out any residue. Before the next run, the column was conditioned with 70% A and 30% B for 15 minutes, and then 80% A and 20% C for another 15 minutes. See Figure 2 in Zhang et al. (2021) for chromatograms and sample collection-window time intervals. 10% of the flow coming out of the ion chromatography (IC) column was directed to a Pulse Amperometric Detector (PAD; Thermo Fischer Scientific, Waltham, MA, USA) by a flow splitter for peak detection and quantitation. The rest of the flow was sent to an automated fraction collector (AFC-3000; Thermo Fischer Scientific, Waltham, MA, USA), which was placed in a glove bag filled with N2 gas. Purified AA fractions were collected into 40mL flat-bottom glass vials (Thermo Fischer Scientific, Waltham, MA, USA) using retention-time based collection method. IC procedural blanks were collected close to each AA (Phe and Glu) fractions and had the same concentration of NaOH as the corresponding AA fractions.

(3) Conversion of amino acid to nitrite:
A previously developed protocol (Zhang and Altabet 2008) was employed for the conversion of collected AA to nitrite. This adapted protocol involves three types of reagents:
1) a catalytical reagent that was prepared by mixing 0.665 grams (g) of sodium bromide (NaBr; Thermo Fischer Scientific, Waltham, MA, USA) with 52.5mLof 50% w/w sodium hydroxide (NaOH; Thermo Fischer Scientific, Waltham, MA, USA), and then diluted to 100mL with MilliQ water;
2) an oxidizing reagent that was prepared daily by diluting 4.25mL of commercial bleach (5.25% NaClO; Thermo Fischer Scientific, Waltham, MA, USA) to 100 mL using MilliQ water;
and 3) a quenching reagent that was made by adding 5.1g sodium arsenite (Na2AsO2) into 100ml MilliQ water. 0.1mL of the catalytical reagent was added into the 10mL AA fractions and mixed well, followed by the addition of 0.1mL oxidizing reagent.

(4) Conversion of nitrite to N2O:
The analytical procedure of converting nitrite to nitrous oxide (N2O), and subsequent N isotope analysis of N2O, have been documented in previous studies (McIlvin and Altabet, 2005, Zhang et al., 2007). A slightly modified procedure was performed at UMD-SMAST. About 15 nanomoles (nmol) of nitrite converted from the AA samples or reference materials were pipetted into 20mL crimp-top sealed headspace glass vials and diluted to a consistent volume (i.e. 4mL) using MilliQ water. Then, pre-conditioned sodium azide solution with acetate buffer (0.2 M NaN3 in 85.5% acetic acid (aq), prepared by adding 10ml of a solution of 2M NaN3 in 45% acetic acid (aq) to 90ml of a 90% acetic acid (aq) solution) was pipetted into the same vials at a 0.14/1 (v/v) reagent‐to‐sample ratio and immediately capped with a butyl rubber stopper and crimped with aluminum seal followed by gentle swirling to fully mix the reagent and sample. It took an hour for this reduction reaction to complete at room temperature. The produced N2O in these crimped vials were purged with Helium using a Trace Gas interface autosampler (PAL), which was subsequently cryogenically focused and separated from N2, O2, CO2, and H2O using a gas chromatography column before sending into the IRMS for isotope analysis.