Evaluation of an untargeted nano-liquid chromatography-mass spectrometry approach to expand coverage of low molecular weight dissolved organic matter in Arctic soil

Abstract

Characterizing low molecular weight (LMW) dissolved organic matter (DOM) in soils and evaluating the availability of this labile pool is critical to understanding the underlying mechanisms that control carbon storage or release across terrestrial systems. However, due to wide-ranging physicochemical diversity, characterizing this complex mixture of small molecules and how it varies across space remains an analytical challenge. Here, we evaluate an untargeted approach to detect qualitative and relative-quantitative variations in LMW DOM with depth using water extracts from a soil core from the Alaskan Arctic, a unique system that contains nearly half the Earth's terrestrial carbon and is rapidly warming due to climate change. We combined reversed-phase and hydrophilic interaction liquid chromatography, and nano-electrospray ionization coupled with high-resolution tandem mass spectrometry in positive- and negative-ionization mode. The optimized conditions were sensitive, robust, highly complementary, and enabled detection and putative annotations of a wide range of compounds (e.g. amino acids, plant/microbial metabolites, sugars, lipids, peptides). Furthermore, multivariate statistical analyses revealed subtle but consistent and significant variations with depth. Thus, this platform is useful not only for characterizing LMW DOM, but also for quantifying relative variations in LMW DOM availability across space, revealing hotspots of biogeochemical activity for further evaluation.

Figure 1

Scatter plot of the features detected (intensity >1.0E4, +/−0.005 m/z) in a single soil water extract and the elution profiles for HILIC (top) and RP (mirrored bottom) in positive-ion mode demonstrating different separation profiles of LMW DOM on each LC phase. Each marker matches to a m/z and retention time (RT). The corresponding normalized base peak chromatograms are overlaid on top to show a typical elution profile for each LC condition and display trends between m/z and RT. Additional examples in negative-ion mode are shown in Fig. S4.

Figure 2

Schematic of the untargeted metabolomics approach established and applied in the present study for the analysis of LMW DOM from Arctic soil water extracts. After the filtration step, triplicate extracts for each section of the core (n = 9) were split and handled separately. The resulting concentrated aliquots (18 samples) were run on two LC phases and in two MS polarities, resulting in four analytical conditions per sample. RT: retention time. ESI (+/−): electrospray ionization positive or negative mode. DDA: data-dependent acquisition (unbiased precursor selection for MS² fragmentation).

Figure 3

Overlap of HQFs detected by HILIC and RP in positive- and negative-ion MS polarities (based on the MS¹ neutral mass for the corresponding [M + H]⁺ or [M − H]⁻ ion, +/−0.001 Da).

Figure 4

High-quality features ranked by abundance (1 = most abundant, 1705 = least abundant) and the relative contribution of each to the cumulative abundance demonstrating the depth of measurement for each LC-MS condition evaluated. The number of LMW DOM features detected by each LC-MS condition accounting for half and the total cumulative abundance are reported.

Figure 5

Principal component analyses of high-quality features detected in soil water extracts analyzed by (a) HILIC (+) and (b) HILIC (−) demonstrating the sensitivity of the untargeted technique to detect subtle variations in LMW DOM across space in these organic-rich soils.

Figure 6

(a) Heatmap of the unique IDs and normalized log₂ peak areas for each abundant HQF detected by HILIC (+). The dendrogram to the left of the heatmap clusters LMW DOM metabolites that varied similarly between the top and bottom of the soil core. To the right, we call out two clusters that starkly increased or decreased with depth after normalizing to per gram dry weight soil. One feature that decreased with depth, highlighted in red, is further analyzed in (b). (b) Cross-sectional diagram of the soil core showing three depth increments and corresponding extraction replicate sample numbers. Stacked extracted ion chromatograms and mass spectra (MS¹ and MS², insets) for a feature (116.0703 m/z) detected by HILIC (+) at RT ~6.1 min. The feature was detected (intensity >1.0E5) in all six soil extracts but not in the blanks or controls. Extraction triplicates yielded similar amounts (CV < 3%) despite some peak splitting, and there was a 4-fold difference between the log₂ peak areas of the top and bottom sections of the core (p-value < 0.05 by t-test). The feature was putatively identified as proline by matching the MS¹ spectrum in MMCD and confirming with the MS² spectrum in MassBank.