Discovery and identification of oxidative stress biomarkers in marine plankton using a new, open-source R package for lipidomics Jamie Collins 1,2 * , Bethanie Edwards 1,2,3 , Kevin Becker 1 , Helen Fredricks 1 , and Benjamin A.S. Van Mooy 1 1 Department of Marine Chemistry & Geochemistry, Woods Hole Oceanographic Institution 2 MIT/WHOI Joint Program in Oceanography; 3 School of Oceanography, University of Hawaii at Manoa Introduction: ROS and the algal lipidome Oxidative stress exerts a profound impact on the lives of marine microbes, including eukary- otic phytoplankton. Reactive oxygen species (ROS) can function as inter- or intracellular signals, alter various cellular metabolisms, and dramatically transform the individual mole- cules that make up the algal cell. As in terrestrial plants, the polar lipids of marine algae are a “front line” barrier in the battle against oxidative stress. Because these lipids are primary structural components of cell and or- ganelle membranes, they are particularly susceptible to chemical transformation by both inter- and extracellular ROS. The lipids transformed by ROS can be exploited as biomarkers for specific types of oxidative stress. We use lipidomics to analyze hundreds of these biomarkers simultaneously with their unoxi- dized parent compounds. Patterns in their distribution, abundance, and chemical structure can be used to: • Identify new biomarkers for vari- ous biological & abiotic stressors • Localize the impact of oxidative stress within an ecosystem or single cell • Examine the effect of ROS on bio- chemical pathways, microorgan- isms, and ecosystem-scale biogeo- chemistry xcms 3 and CAMERA 4 LOBSTAHS: An R package for lipidomics discovery We developed LOBSTAHS 1 (Lipid and Oxylipin Biomarker Screening through Adduct Hierarchy Sequences) to assist in the discovery and identification of new lipid biomarkers for oxidative stress. LOBSTAHS is an open-source pack- age for R that uses a unique screening criteria to make compound identifica- tions in HPLC-MS data. LOBSTAHS is available via Bioconductor or GitHub. The package can be downloaded via the QR code at right. Improvements, suggestions, and database additions are welcome – just submit a bug report. Increasing confidence in compound assignments LOBSTAHS 1 Screened & annotated “high confidence” dataset available for analysis Assign confidence scores Annotate possible iso- mers and isobaric masses Screen assignments using adduct ion hierarchy se- quences Screen assignments by re- tention time according to parent lipid class Initial compound assign- ments from database(s) Initial mass spectral feature discovery Retention time correction Peak grouping Eliminate 2° isotope peaks Data acquisition: HPLC-ESI-MS 2 ; high mass accuracy & resolution Complex dataset: Multiple samples across experi- mental treatments or environmental gradients OOH Lipoxygenase enzyme Lipoxygenase enzyme Abiotic Abiotic Biological Biological + O 2 + O 2 1 O 2 •OH H 2 O 2 1 O 2 •OH H 2 O 2 OOH Summary & significance Shifts in the distribution, abundance, and chemical structure of lipids in (a) P. tricornu- tum and (b) the microbial community at Sta- tion ALOHA can tell us about the effects and targets of oxidative stress within the algal cell and within ecosystems. • Extensive lipidome remodeling followed treatment with 150 μM H 2 O 2 in Phaeodacty- lum tricornutum Additionally, oxidative stress promoted reallocation of biomass to triacylglycerols (TAG) • At Station ALOHA, UVB radiation induced wholesale change in the oxidation state of the metalipidome Onboard, user-cus- tomizable databases: 14,068* unique IPL, lyso-IPL, ox-IPL, oxylipins, TAG, and photosynthetic pigments 1 Collins et al. 2016, Anal. Chem. 88:7154-7162. 2 Modified method of Hummel et al. 2011, Front. Plant Sci. 2:54. 3 Smith et al. 2006, Anal. Chem. 78:779-787; Tautenhahn et al. 2008, BMC Bioinformatics 9:504. 4 Kuhl et al. 2012, Anal. Chem. 84:283-289. 5 Graff van Creveld et al. 2015, ISME 9:285-395. In the study, cultures of P. tricornutum were treated with varying concentrations of hydrogen peroxide to induce an oxidative stress response. Experimental methods are described in the reference. 6 DGCC, diacylglyceryl carboxyhydroxymethylcholine; DGDG, digalactosyldiacylglycerol; DGTS & DGTA, diacylglyceryl trimethylhomoserine and diacylglyceryl hydroxymethyl-trimethyl-β-alanine; IPL, intact polar lipid; ox-IPL, oxidized intact polar lipid; MGDG, monogalactosyldiacylglycerol; PC, phosphatidylcholine; PE, phosphatidylethanolamine; PG, phosphatidylglycerol; SQDG, sulfoquinovosyldiacylglycerol; TAG, triacylglycerol 7 P≤ 0.05 (bold), P≤0.01 (*), P≤0.001 (**), and P≤0.0001 (***).. This research was supported by the Gordon and Betty Moore Foundation through Grant GBMF3301 to B.A.S.V.M. This research was also funded in part by a grant to B.A.S.V.M. from the Simons Foundation and is a contribution of the Simons Collaboration on Ocean Processes and Ecology. J.R.C. acknowledges support from a U.S. Environmental Protection Agency (EPA) STAR Graduate Fellowship (Fellowship Assistance Agreement No. FP-91744301-0). https://github.com/vanmooylipidomics/LOBSTAHS/ m z + double bond + acyl carbon 600 700 800 900 1000 1100 1200 5 10 15 20 +2O +3O +4O unoxidized +1O Oxylipin PE MGDG PC SQDG PG TAG DGTS & A DGDG DGCC Lipid class 6 ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● ● Corrected retention time (min) Within each lipid class 6 , certain molecules are associated only with oxidative stress TAG 26 11 145 26 11 145 MGDG 27 19 86 27 19 86 SQDG 23 23 96 23 23 96 PG 5 4 28 5 4 28 PC 49 45 106 49 45 106 PE 32 22 117 32 22 117 DGTS & DGTA 10 7 54 10 7 54 19 13 65 19 13 65 DGDG 191 145 703 191 145 703 All lipids DGCC 6 6 Healthy organism Lipids observed only in 0 μM H 2 O 2 control Stressed Lipids observed only under 150 μM H 2 O 2 Lipids observed in both treatments Results: Oxidative stress in the marine diatom Phaeodactylum tricornutum LOBSTAHS identifies 1039 unique oxidized and unoxidized lipids that serve as potential biomarkers Results in pure culture: Oxidative stress in the model marine diatom Phaeodactylum We applied LOBSTAHS to lipid data from an experiment 5 in which hydrogen peroxide (H 2 O 2 ) was used to induce oxidative stress in the marine diatom Phaeodactylum tricornutum. * ... and growing! * @jamesrco; future: Moore/Sloan & WRF Data Science Fellow at University of Washington eScience Institute & School of Oceanography New oxidized lipids are observed, but increase in oxidation state driven primarily by fall in concentration of unoxidized molecules Results at Station ALOHA: Effect of UVB exposure on a microbial metalipidome We applied different light screening treatments to large ( ~ 2 L ) volumes of unfiltered seawater collected at Station ALOHA, in the North Pacific Ocean, during the SCOPE/C-MORE HOE-Legacy III cruise. LOBSTAHS was used to identify patterns & potential biomarkers of natural, daily-recurring UV radiation stress in HPLC-MS lipid data from three treatments. 1,747 different oxidized and unoxidized compounds were identified & quantified. 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 25 20 15 10 5 Time (hours) 0 μM H 2 O 2 150 μM H 2 O 2 Fraction peak area ID’d as TAG Stress induces TAG production Initial + 9 h Mole fraction identified as oxidized lipid 0.0 0.1 0.2 0.3 0.4 Dark - UVB + UVB UVB exposure leads to increase in overall oxidation state of metalipidome Fold change from init. concentration DGTS_DGTA 27:1 +3O, RT 10.87 min. DGTS_DGTA 30:5 +4O, RT 14.01 min. DGTS_DGTA 45:1 +1O, RT 19.51 min. SQDG 39:1 +1O, RT 19.67 min. DGDG 30:0, RT 19.68 min. DGTS_DGTA 43:0 +1O, RT 19.51 min. DGTS_DGTA 43:2 +1O, RT 18.78 min. DGTS_DGTA 34:1, RT 17.59 min. DGTS_DGTA 41:1 +1O, RT 18.76 min. DGTS_DGTA 43:1 +1O, RT 19.12 min. DGTS_DGTA 41:0 +1O, RT 19.09 min. DGTS_DGTA 39:0 +1O, RT 18.7 min. TAG 40:4, RT 18.65 min. PE 41:1, RT 17.47 min. PG 43:4 +3O, RT 12.51 min. PG 31:3 +3O, RT 8.98 min. PE 33:4 +1O, RT 18.15 min. TAG 39:4 +3O, RT 18.57 min. SQDG 46:8 +4O, RT 10.53 min. PG 30:4, RT 7.55 min. DGDG 32:1 +4O, RT 8.73 min. PG 37:8 +3O, RT 7.76 min. SQDG 39:7 +3O, RT 9.98 min. SQDG 42:6 +2O, RT 8.73 min. PE 31:3 +2O, RT 7.55 min. Compounds exhibiting greatest relative increase in +UVB treatment 0 10 20 30 + UVB Dark control - UVB Not detected + UVB Dark control - UVB PG 41:5, RT 13.97 min. DGDG 36:9, RT 11.27 min. DGDG 28:1 +3O, RT 13.71 min. PC 38:6, RT 15.19 min. DGDG 34:7, RT 11.57 min. PG 41:5 +1O, RT 13.57 min. TAG 50:1, RT 21.84 min. TAG 50:2, RT 21.06 min. TAG 50:0, RT 22.77 min. TAG 52:1, RT 22.77 min. TAG 48:1, RT 21.06 min. TAG 48:0, RT 21.86 min. TAG 52:2, RT 21.83 min. SQDG 28:0, RT 12.07 min. PE 36:2, RT 17.2 min. DGDG 30:1, RT 14.07 min. SQDG 30:0, RT 13.32 min. MGDG 30:1, RT 14.92 min. SQDG 30:1, RT 12.3 min. PG 33:1, RT 14.81 min. PG 36:2, RT 15.24 min. SQDG 30:2, RT 11.36 min. Fold change from init. concentration -4 0 2 4 -2 Compounds with greatest absolute change in concentration from t = 0 (at t +9 h) (at t +9 h)