Iron Chelation in Soil: Scalable Biotechnology for Accelerating Carbon Dioxide Removal by Enhanced Rock Weathering

Enhanced rock weathering (EW) is an emerging atmospheric carbon dioxide removal (CDR) strategy being scaled up by the commercial sector. Here, we combine multiomics analyses of belowground microbiomes, laboratory-based dissolution studies, and incubation investigations of soils from field EW trials to build the case for manipulating iron chelators in soil to increase EW efficiency and lower costs. Microbial siderophores are high-affinity, highly selective iron (Fe) chelators that enhance the uptake of Fe from soil minerals into cells. Applying RNA-seq metatranscriptomics and shotgun metagenomics to soils and basalt grains from EW field trials revealed that microbial communities on basalt grains significantly upregulate siderophore biosynthesis gene expression relative to microbiomes of the surrounding soil. Separate in vitro laboratory incubation studies showed that micromolar solutions of siderophores and high-affinity synthetic chelator (ethylenediamine-N,N′-bis-2-hydroxyphenylacetic acid, EDDHA) accelerate EW to increase CDR rates. Building on these findings, we develop a potential biotechnology pathway for accelerating EW using the synthetic Fe-chelator EDDHA that is commonly used in agronomy to alleviate the Fe deficiency in high pH soils. Incubation of EW field trial soils with potassium-EDDHA solutions increased potential CDR rates by up to 2.5-fold by promoting the abiotic dissolution of basalt and upregulating microbial siderophore production to further accelerate weathering reactions. Moreover, EDDHA may alleviate potential Fe limitation of crops due to rising soil pH with EW over time. Initial cost-benefit analysis suggests potassium-EDDHA could lower EW-CDR costs by up to U.S. $77 t CO2 ha–1 to improve EW’s competitiveness relative to other CDR strategies.

-S2, and 1 x Supporting Information Note.Details are as listed below: Figure S1.Field experimental set-up for rock grain bags at the Energy Farm, US. Figure S2.Genus-specific expression of the desferrioxamine and arthrobactin siderophore biosynthesis gene desB.Figure S3.Elemental release from basalt in the absence or presence of high-affinity (siderophores) and low-affinity (citrate) iron chelators at different concentrations.Figure S4.Dissolution of phosphorus from basalt in response to in vitro chelator-driven weathering.Figure S5.Mobilization of titanium (Ti) at exchange sites in response to EDDHA-driven weathering.Figure S6.Sequential extraction procedure for incubated soil and soil+basalt samples in this study.Figure S7.Charge balance in the chelator-free control, K-DF siderophore, and synthetic chelator K-EDDHA weathering solutions considering K adsorption onto basalt mineral surfaces from initial solutions.Table S1.Hillhouse basalt mineralogy as outlined in Lewis et al, 2021.Table S2.Correlation table: change in soil exchangeable concentration of different elements between end (20 days) and start (0 days) of experimental incubation in soil and soil+basalt substrates in response to EDDHA concentrations.Supplementary Note 1. Modified Arthrobacter sp.JG-9 bioassay to measure hydroxamates in soil.

Bags of basalt weathered in soil
Figure S1.Field experimental set-up for rock grain bags at the Energy Farm, US.Heat-sealed polyethylene mesh bags (30 µm pore diameter) containing 4 grams of 53-90 µm fresh basalt grains were placed in the soil between plant rows.This bag deposition happened in control block 5 previously untreated with basalt.Bags were placed in a plastic washing basket pre-filled with soil with holes drilled on the side and holes on the bottom to allow for water flow and root interaction.This apparatus was covered with soil (not shown) and bags were left to weather for 1.5-years before they and surrounding soil were collected for omics analyses.The metatranscriptomic expression data indicate that genera in weathered basalt microbiomes express desB to greater levels relative to their core metabolic activity (here exemplified by the housekeeping gene rpoA) than the same genera in soil.These data support the view that in addition to positive selective pressure recruiting siderophore-producing bacteria in basalt microbiomes relative to soil (Figure 1E), siderophore-producing bacteria are also actively utilizing the ability to produce siderophores during their life on basalt grains more so than members of the same genus in soil microbiomes.Error bars show SEM.
fold over ligand-free control fold over ligand-free control fold over ligand-free control  All the soil samples were extracted with 3 ml of ultrapure water as shown in Figure S6.After shaking and centrifuging, an aliquot of the cold-water extraction in the supernatant (i.e.room-temperature) was filtered and used for further assays.The remaining contents of the tube were reconstituted with the leftover water, and the sample heated up in a water bath as specified in the figure above.Note that no further water was added.After incubation and centrifuging, all remaining hot-water extract in the supernatant was taken out, filtered and used for further assays.Following this, the soil pellet was resuspended and extracted with 12.5 ml 1M ammonium acetate buffered at pH 7.0.Hydroxamate siderophore concentration was determined using the hydroxamate auxotrophic bacterium Arthrobacter sp.JG-9 (now Microbacterium flavescens JG-9).The bacterium was grown in the terregens factor assay medium (TAM) medium 1 using soil extract of specified volume.Standard curve was based on Arthrobacter sp.JG-9 growth (measured by OD600) in TAM medium with added desferrioxamine B mesylate at known concentrations.The hot water extracts and exchangeable extracts were diluted and acidified to 2% nitric acid and send for ICP-MS together with appropriate blanks.
Dimitar Z. Epihov * , Steven A. Banwart, Steve P. McGrath, David P. Martin, Isabella L. Steeley, Vicky Cobbold, Ilsa B. Kantola, Michael D. Masters, Evan H. DeLucia & David J. Beerling *Corresponding author: d.z.epihov@sheffield.ac.ukOverall contents: 7 x Supporting Information Figures labelled Figure S1-S7, 2 x Supporting Information Tables labelled Table Figure S2.Genus-specific expression of the desferrioxamine and arthrobactin siderophore biosynthesis gene desB across abundant Actinobacteriota genera in field-weathered basalt and soil microbiomes.The metatranscriptomic expression data indicate that genera in weathered basalt microbiomes express desB to greater levels relative to their core metabolic activity (here exemplified by the housekeeping gene rpoA) than the same genera in soil.These data support the view that in addition to positive selective pressure recruiting siderophore-producing bacteria in basalt microbiomes relative to soil (Figure1E), siderophore-producing bacteria are also actively utilizing the ability to produce siderophores during their life on basalt grains more so than members of the same genus in soil microbiomes.Error bars show SEM.

Figure S3 .Figure S4 .Figure S5 .
Figure S3.Elemental release from basalt in the absence or presence of high-affinity (siderophores) and low-affinity (citrate) iron chelators at different concentrations.Note that major divalent cations mainly Ca 2+ and to a lesser extent Mg 2+ , important for CDR, show patterns of increase only in response to siderophores but not citrate.The release of Fe, Al, and Ti from basalt is generally promoted by the presence of all tested chelators.There are no significant differences in the release of Mn from any of the chelators relative to control.Error bars show SEM.

Figure S6 .
Figure S6.Sequential extraction procedure for incubated soil and soil+basalt samples in this study.All the soil samples were extracted with 3 ml of ultrapure water as shown in FigureS6.After shaking and centrifuging, an aliquot of the cold-water extraction in the supernatant (i.e.room-temperature) was filtered and used for further assays.The remaining contents of the tube were reconstituted with the leftover water, and the sample heated up in a water bath as specified in the figure above.Note that no further water was added.After incubation and centrifuging, all remaining hot-water extract in the supernatant was taken out, filtered and used for further assays.Following this, the soil pellet was resuspended and extracted with 12.5 ml 1M ammonium acetate buffered at pH 7.0.Hydroxamate siderophore concentration was determined using the hydroxamate auxotrophic bacterium Arthrobacter sp.JG-9 (now Microbacterium flavescens JG-9).The bacterium was grown in the terregens factor assay medium (TAM) medium 1 using soil extract of specified volume.Standard curve was based on Arthrobacter sp.JG-9 growth (measured by OD600) in TAM medium with added desferrioxamine B mesylate at known concentrations.The hot water extracts and exchangeable extracts were diluted and acidified to 2% nitric acid and send for ICP-MS together with appropriate blanks.

Figure S7 .
Figure S7.Charge balance in the chelate-free control, K-DF siderophore, and synthetic chelate K-EDDHA weathering solutions considering K adsorption onto basalt mineral surfaces from initial solutions.A,B -Charge equivalents Ca 2+ + Mg 2+ plotted against HCO3 -.The plots show a relationship more closely matching the 1:2 mixing line than the 1:1 line.These data are suggestive of other negatively charged ions different than bicarbonate.C,D.Proportion of K + adsorbed onto basalt surfaces and unbalanced negative charge.Since weathering reactions took place in 0.001M KCl solution (equivalent to 1000 µM K + and 1000 µM Cl -), adsorption of K + onto basalt mineral surfaces (negative charge of reactive surfaces or basalt-derived clays) can leave unbalanced Cl -in the weathering solution.Moreover, added chelators (particularly K3 + :EDDHA 3-) contain a large amount of K + which upon the

Table S2 .
Correlation table: change in soil exchangeable concentration of different elements between end (20 days) and start (0 days) of experimental incubation in soil and soil+basalt substrates in response to EDDHA concentrations.Based on soil hot-water extraction as exchangeable Al was below detection limits.Pearson correlation test, *** P < 0.001, ** < 0.01, * < 0.05, ^ < 0.10, ns > 0.10.