Abundance and Distribution of GDGTs in Incubated Artificial Soils with No Fossil Pool

Significance: Glycerol dialkyl glycerol tetraethers (GDGTs) are a series of microbial membrane lipids produced by archaea and bacteria. They are widely distributed in a range of natural environments including soils, lakes, oceans, peat bogs, stalagmites, and fossil bones^[1-15], relatively resistent to degradation, and sensitive to environmental variables. Therefore, they have increasingly been used as a popular tool in the study of past cliamte changes^[16-17]. In soils, the distribution of brGDGTs are significantly correlated with soil moisture condition and temperature, therefore, many GDGT-based paleotemperature and paleohydrological proxies have been developed. Particularly, for soil moisture, modern investigations suggest that, the branched and isoprenoid tetraether index (BIT) which represents the ratio of bacterial brGDGTs to crenarchaeol, can track soil water content (SWC) changes and thus can be used as a soil moisture proxy^[24]. On the other hand, soil mosture condition can also affect the applicablity of brGDGT-based paleotemperautre proxies^[2,21,23]. The mechanisms for the effect of soil moisture on the BIT soil moisture proxy and other paleotemperautre proxies, however, remain inadequately understood.

　　Incubation experiments are helpful to clarify the mechanisms for the responses of GDGTs to environmental variables, and to test the reliability of related climatic proxies. Two recent studies incubated Candidatus Solibacter usitatus and observed abundant brGDGTs^[31-32], but the results of this single-strian incubation cannot fully represent those for all brGDGT-producing microrganisms in the soil matrix. Some other works tested the correlation between incubated parameters and brGDGTs by in situ experiments in the field^[38-40] or laboratory incubation experiments^[33,41-45], but the interference of fossil signals cannot be exluded.

　　In this paper, we performed a 4-year incubation experiment using artificial soils without fossil GDGTs. By incubating these artificial soils under the same temperature but different soil water content (SWC), the response of newly produced GDGTs to incubated parameters was explored, and the impact of brGDGT distribution on the applicablity of the brGDGT paleothermometer was assessed. While the abundances of incubated GDGTs were positively correlated with SWC, the BIT soil moisture proxy was not significantly correlated with SWC, suggesting that this index may indirectly (rather than directly) respond to SWC changes. Additionally, 6-methyl brGDGTs were more abundant than 5-methyl brGDGTs, resulting in extremely high values of MBT'5ME and low MBT', confirming that high relative abundance of 6-methyl brGDGTs can affect the applicability of the MBT'_5ME paleothermometer in soils. To our knowledge, this is the first work that incubates GDGTs using artificial materials to imitate soil matrix conditions, but excludes the interference of fossil GDGT signals. The results will provide important implications of the study of GDGTs and their related paleoclimatic proxies.

Methods: Wheat bran, combused sand, and red clay were mixed in a ratio of 5:93:2 to create an artificial soil. Sand and red clay can provide a soil-like condition while wheat bran can provide organic carbon for the growth of microrganisms. No GDGTs were detected in the artificial soil and therefore the GDGTs obtained after incubation should all be produced in situ. A total of 8 samples (each 53g in a 250mL beaker) were used in this experiment. These samples were divided into two groups, and phosphate buffer solution was added to one group (A₁−A₄) to adjust pH during incubation but not added to the other group (B₁−B₄). Each group was incubated under relatively constant temperature (21.3±2.3℃) with a maximum SWC gradient from 10% to 40% (Table 1). The incubated samples were harvested after 4 years.

　　The incubated soils were ground after freeze-drying, and ultersonicly extracted 3 times with MeOH:dichloromethane (1∶9, V/V). The total lipid extracts containing GDGTs were dried under nitrogen gas, re-dissolved in hexane and filtered through a PTFE filter. GDGTs were analyzed on two high performance liquid chromatography/atmospheric pressure chemical ionization-mass spectrometry (HPLC-APCI-MS) systems (Shimadzu HPLC-MS 8030 and ThermoFisher Vanquish Core HPLC-Orbitrip Exploris 120) in Institute of Earth Environment, Chinese Academy of Sciences. The separation of lipids was obtained with coupled silica columns following the chromatography methods capable of separating 5-methyl and 6-methyl brGDGTs^[25,46]. MS scanning was performed in selected ion monitoring (SIM) mode that targeted specific [M+H]⁺ ions for GDGTs. The chromatograms of GDGTs on HPLC-Orbitrap Exploris 120 are shown in Fig.1. The uncertainty for peak area is less than 3% for continuous measurements. C₄₆ GTGT was used for quantification of GDGTs and archaeol on the two HPLC-APCI-MS, assuming that their response factors are identical.

Data and Results: There are impurity peaks which cannot be fully resolved with GDGTs in the chromatograms of both HPLC-MS 8030 and HPLC-Orbitrap Exploris 120 (at 100ppm resolution). However, when analyzing the chromatograms of HPLC-Orbitrap Exploris 120 at 20ppm resolution, the impurity peaks no longer exist. Therfore, high-resolution mode for GDGT quantification was used in this work.

　　The concentrations of GDGTs range from 5.14 to 39.89ng/g for incubated soils with phosphate buffer solution, are systematically higher than those for incubated soils without phosphate buffer solution (21.79−73.95ng/g) (Table 1, Fig.4). For both groups, the concentrations of GDGTs increase with increasing SWC (Fig.4), in agreement with those observed from natural soils^[23-24,26]. The BIT index varies between 0.11 and 0.91, but exhibits no obvious relationship with SWC (Table 1). This confirms the field observation of a positive relationship between BIT and SWC or precipitation^[1,24,57]. Moreover, the Archaeol and Caldarchaeol Ecometric (ACE) salinity proxy ranges from 19.6 to 47.4 and from 3.7 to 8.5 for incubated soils with and without phosphate buffer solution, respectively (Table 1, Fig.5), indicating that the addition of phosphate buffer solution can result in higher salinity.

　　For most of the incubated soil, relatively pure 6-methyl brGDGTs can be obtained (Fig.1), and the relative proportion of 6-methyl brGDGTs versus 5-methyl brGDGTs, expressed in the isomer ration (IR_6ME) index, is quite high (≥0.99) (Table 1). The MBT'_5ME values are extremely high while MBT' values are extremely low when IR_6ME approaches 1 (Table 1). For samples with IR_6ME values ≥0.99, the average MBT' value is 0.07±0.04 (N=7), representing the MBT'₆ endmember of 6-methyl brGDGTs at this incubation temperature (21.3±2.3℃) (Fig.6).

　　The weak correlation between BIT and SWC indicates that the BIT soil moisture proxy may indirectly, rather than directly, respond to SWC changes. The relatively lower concentration of GDGTs in incubated soils with phosphate buffer solution suggests that this procedure will inhibit the production of GDGTs, and therefore caution should be used when adding phosphate buffer solution in incubation experiments. Moreover, the significantly biased MBT'_5ME and MBT' values confirm that high relative abundance of 6-methyl brGDGTs can affect the applicability of the brGDGT paleothermometer in soils.