Particulate intact polar diacylglycerol (IP-DAG) lipids measured in water column samples from the Drake Passage and Bellingshausen Sea

Seawater samples were retrieved from depth using standard oceanographic sampling equipment and then collected by vacuum filtration onto 0.2 µm pore size Durapore membrane filters; these were frozen immediately at -80°C. Extraction was performed using a modified Bligh and Dyer (Bligh and Dyer, 1959; https://doi.org/10.1139/o59-099) method described in Popendorf et al. (2013; https://doi.org/10.1007/s11745-012-3748-0). Lipid extracts were analyzed by HPLC-ESI-MS with data dependent-MS2 acquisition on a high-resolution, accurate mass Thermo Q Exactive Hybrid Quadrupole-Orbitrap mass spectrometer (ThermoFisher Scientific, Waltham, MA, USA) coupled to an Agilent 1200 HPLC system (Agilent, Santa Clara, CA, USA). The HPLC-ESI-MS method is described in Collins et al., 2016 (https://doi.org/10.1021/acs.analchem.6b01260).

The LOBSTAHS lipidomics discovery software (Collins et al., 2016; https://doi.org/10.1021/acs.analchem.6b01260) was used to putatively identify HPLC-MS features in the data. We confirmed each LOBSTAHS identification using two additional means: (1) via comparison of data-dependent MS2 spectra with those from authentic standards or published reference spectra and (2) by requiring the presence of the same compound identity in data acquired in the opposite HPLC-ESI-MS ionization mode. We confirmed all LOBSTAHS identities at the lipid class level (e.g., PC versus PE, or MGDG versus TAG) using a new, experimental LOBSTAHS feature which automatically detects diagnostic product ion fragments and constant neutral losses (as given in Popendorf et al., 2013; https://doi.org/10.1007/s11745-012-3748-0) in the available data-dependent MS2 spectra for each sample. After identification, quantification of analytes was performed using a series of standard curves, followed by normalization to concentration of an internal standard.

Lipid identities are resolved only to the level of bulk fatty acid composition (i.e., the sum of the properties of the substituents at both the sn-1 and sn-2 positions). As a result, in some cases (e.g., DGTS_DGTA 32:1), several different molecules having the same bulk fatty acid composition have been identified. These can be distinguished by retention time, which is given in the "Comment" field for each entry.

采用标准海洋学采样设备从指定深度采集海水样本，随后通过真空过滤将样本截留于孔径0.2 µm的杜拉孔（Duropore）膜滤器上，并立即将滤膜置于-80℃环境中冷冻保存。

脂质提取采用经改良的布莱根-戴尔（Bligh and Dyer）法，该改良方案源自Popendorf等于2013年的研究，原始经典方法参见Bligh和Dyer于1959年的工作（Bligh and Dyer, 1959; https://doi.org/10.1139/o59-099；Popendorf et al., 2013; https://doi.org/10.1007/s11745-012-3748-0）。脂质提取物通过高效液相色谱-电喷雾电离质谱（HPLC-ESI-MS）联用系统进行分析，采用数据依赖性二级质谱（data-dependent MS2）采集模式，所用仪器为赛默飞世尔科技（ThermoFisher Scientific，美国马萨诸塞州沃尔瑟姆）的Thermo Q Exactive混合四极杆-轨道阱高分辨精准质谱仪，搭配安捷伦1200高效液相色谱系统（Agilent，美国加利福尼亚州圣克拉拉）。该HPLC-ESI-MS分析方法详见Collins等（2016）的研究（https://doi.org/10.1021/acs.analchem.6b01260）。

本研究使用LOBSTAHS脂质组学发现软件（Collins et al., 2016; https://doi.org/10.1021/acs.analchem.6b01260）对数据中的高效液相色谱-质谱特征峰进行推定注释。随后通过两种额外方式验证每一条LOBSTAHS注释结果：(1) 将数据依赖性二级质谱谱图与标准品谱图或已发表的参考谱图进行比对；(2) 要求在相反电离模式下获取的质谱数据中检测到相同的化合物注释。我们通过LOBSTAHS新增的实验性功能完成所有脂质类别层面的注释验证（例如磷脂酰胆碱（PC）与磷脂酰乙醇胺（PE）、单半乳糖基二酰基甘油（MGDG）与三酰甘油（TAG）），该功能可自动在每个样本的可用数据依赖性二级质谱谱图中识别特征性产物离子片段与恒定中性丢失碎片（Popendorf et al., 2013; https://doi.org/10.1007/s11745-012-3748-0）。

注释完成后，通过一系列标准曲线对分析物进行定量，并以内标物浓度作为参照完成数据归一化处理。

本研究仅能将脂质注释至总脂肪酸组成层面（即sn-1与sn-2位取代基的总属性之和）。因此在部分案例中（例如DGTS_DGTA 32:1），已注释得到多个具有相同总脂肪酸组成的不同分子。此类分子可通过保留时间进行区分，相关保留时间信息已在每条记录的"Comment"字段中提供。