The low molecular weight (LMW) reduced sulfur substances (RSS) composition of dissolved organic matter (DOM) was examined along the GEOTRACES US-GP15 section in the Pacific Ocean. We demonstrate that LMW RSS constitutes a significant fraction of non-volatile dissolved organic sulfur (DOS). While thiols such as glutathione were below our detection limit (300 pM), RSS containing two carbon (C) sulfur (S) bonds were present at concentrations in the hundreds of nM range. RSS accumulation was observed in subtropical waters. The most likely source of these RSS is microbial alteration of sulfurized DOM with production of secondary thioamidated metabolites. RSS are initially produced by cyanobacteria to mitigate copper and oxidative stress induced by UV-B irradiance. A preferential remineralization of RSS over dissolved organic carbon (DOC) in the upper 350 meters suggests a partial lability of LMW DOS. Deeper, homogeneous concentrations and C:S ratio indicate increasing stability of this LMW DOS.

Samples were collected onboard the R/V Roger Revelle between September and November 2018). The seawater was collected using the Oceanographic Data Facility's (ODF, Scripps Institution of Oceanography) CTD rosette equipped with twelve 30 L Niskin bottles. Samples were filtered through 0.8/0.45 μm Acropak 500 filter cartridges and were stored in a -20°C freezer until analysis. 

The concentrations of DOC in samples were measured by size exclusion chromatography coupled to an organic carbon detector (SEC-OCD) (DOC-Labor®, Karlsruhe, Germany) following the protocol of Huber et al. (2011) adapted for seawater (Dulaquais et al., 2018; Fourrier at al., 2022a,b). The analysis of thiols were performed by Ad-CSV on a mercury drop electrode with a deposition potential of 0 V and a deposition time of 180 s following the recommendations of Pernet-Coudrier et al. (2013). Samples were thawed at 4°C and acidified at pH 2 (HCl, Suprapure, Merck©) immediately before analysis.

All the solutions as well as the washing of the material was performed with ultrapure water from a purification system, Milli-Q Element (resistivity > 18.2 MΩ) from Millipore®.

All the chemicals (mobile/acid phases) for SEC-OCD were prepared following the protocol from Huber et al. (2011). The data treatment was also done in the same way as these authors. The calibration of the OCD was performed as Dulaquais et al. (2018) which adapt DOC measurement by SEC for marine waters. The SEC device, equipped with two chromatographic columns (250 mm × 20 mm, TSK HW-50S, 3000 theoretical plates, Toso, Japan), permits the separation of DOM into five fractions of organic compounds with an optimal resolution.

Reduced sulfur species (RSS) such as thioacemamide (TA) form strong complexes with metals and are thus detectable by Ad-CSV at the surface of a mercury drop electrode. The TA concentration was determined by the standard addition method (example for a surface sample in Figure S1) using commercial TA (ACS reagent, Sigma-Aldrich, Switzerland, ≥ 99%) in ultrapure water (132 – 140 µM). Before analysis, the sample was introduced into a glass voltammetric cell, washed beforehand with acidified ultrapure water (pH 2, HCl), to limit any risk of contamination. 7.00 ± 0.01 g of water were weighed directly into the cell under laminar flow hood before lowering the pH to 2 by adding ~70 µL of a diluted HCl solution (~1 mol L-1). Measurements were performed using a three-electrode electrochemical device (Metrohm model 663 VA) connected to a potentiostat/galvanostat (µAutolab® Type III, Utrecht, Holland), controlled by GPES 4.9 software. The working electrode was a Static Mercury Drop Electrode (SMDE) with a drop size of ~ 0.52 mm². An Ag/AgCl electrode (salt bridge in KCl 3 mol L-1, Suprapur®, Merck) and a glassy carbon electrode were used as reference and counter (auxiliary) electrodes respectively. After a degassing step under stirring (60 s, N2 99.99%), the "Hg-(TA)2" complex adsorbed on the working electrode (Figure S2) under the application of a potential of 0 V for a duration depending on the TA concentration in the sample (60 – 120 s). The redissolution of the complexes deposited on the electrode occurs during the cathodic scan in differential pulse mode from 0 to -0.6 V.

To validate DOC measurements, an intercomparison was performed between the DOC concentrations measured by our SEC-OCD system and data acquired using the « classical » TOC-V method. We compared our measurements to the historic dataset from the P16 CLIVAR expedition at a station located at the same latitude but different longitude (20°S, 150°W) and it showed a good agreement between the two datasets in surface and deep waters.

All the DOC concentrations measured within each fraction lof each sample largely exceeded the limits of detection determined by Dulaquais et al. (2018) for marine waters. Deep seawater reference (DSR) samples used to validate the DOC measurements were provided by the Hansell research laboratory (DOC_DSR = 43.2 ± 1.7 µMC; n = 5; consensus value of lot #10 – 18: 43 – 45 µMC). For measurements by Ad-CSV, a reproducibility of 12% (n = 8) and a repeatability of 3.3% (n = 11) were calculated. For deposition times of 60 s and 300 s, limits of detection were estimated to be 81 nN and 22 nM, respectively.