Siderophores and siderophore-associated metabolites from Antarctic Pseudomonas sp. ANT_H12B contribute to iron and phosphorus mobilization in soils with contrasting properties

Iron (Fe) oxy(hydr)oxides regulate soil Fe and phosphorus (P) cycling, but how bacterial siderophores and co-produced metabolites mobilize these nutrients across pH remains unclear. We hypothesized that siderophore-mediated Fe complexation and P mobilization depend on pH and mineral crystallinity and that siderophore-associated metabolites (SAM) are more effective than purified siderophores. A purified pyoverdine-like siderophore fraction from Antarctic Pseudomonas strain ANT_H12B was chemically and spectroscopically characterized and evaluated for Fe³ + complexation from FeCl 3 between pH 5 and 9. Iron mobilization from ferrihydrite (Fh), goethite (Gt), and hematite (Hm) was examined at pH 5.5 and 8.5 across mineral concentrations of 0.5–2 g L⁻¹; purified siderophores and SAM were compared in four P-deficient soils of contrasting pH. Iron³ + complexation increased from 6% at pH 5 to 10% at pH 9, reflecting ligand deprotonation. Ferrihydrite showed the greatest dissolution, although Fe release at pH 8.5 was 30% and 37% lower than at pH 5.5 at 1 and 1.5 g L⁻¹, respectively. Conversely, Gt dissolution at 2 g L⁻¹ was 40% higher at pH 8.5, whereas Hm showed the lowest dissolution, reaching 3 and 1.9 μM at pH 5.5 and 8.5, respectively. Purified siderophores moderately increased Olsen-P in neutral and high-Fe alkaline soils, whereas SAM increased Olsen-P approximately 1- to 26-fold across all soils, with the greatest effect in high-Fe alkaline soil. SAM also increased soluble P 11- to 40-fold and NaOH-EDTA-extractable P by 25–104% relative to controls. The results support our hypothesis, showing that siderophores contribute to Fe and P mobilization but act more effectively within SAM, where organic acids provide complementary dissolution and sorption-competition mechanisms. This finding agrees with previous ANT_H12B studies reporting improved nutrient availability and plant performance and supports using unpurified metabolite mixtures for sustainable P management. Indeed, retaining the co-produced mixture could reduce purification requirements and preserve complementary functions relevant to sustainable P management.

Phosphorus (P) is an essential macronutrient for plant growth, acting as a key component of fundamental biological molecules, participating in phosphorylation reactions, and playing a central role in energy metabolism (1, 2). However, due to its extremely low bioavailability in soil, P is one of the major limiting factors for crop productivity on more than 40% of the world’s arable land (3). The slow diffusion coefficient of inorganic P (Pi), typically present as orthophosphate ions, combined with its strong interactions with soil constituents, often results in soil solution concentrations ranging from 0.1 to 10 μM, insufficient to sustain optimal plant growth and productivity (4, 5). Phosphorus availability is largely constrained by adsorption and precipitation processes. In acidic soils, Pi is strongly adsorbed onto iron (Fe) and aluminum (Al) oxy(hydr)oxides and forms poorly soluble complexes with Fe 3+ and Al 3+. In contrast, in alkaline soils, it precipitates as calcium (Ca 2+) phosphates (6, 7).

Soils contain various forms of Fe oxy(hydr)oxides, differing in distribution and crystallinity, due to the complex interplay among pedogenic factors such as time, lithology, and climate (8, 9). Soil pH strongly influences the solubility of Fe oxy(hydr)oxides, which decreases with increasing pH, reaching a minimum under neutral to alkaline conditions. The high affinity of Fe oxy(hydr)oxides for phosphate plays a central role in regulating P availability in soils (10). Phosphate ions are strongly adsorbed onto these mineral surfaces through inner-sphere complexation with surface hydroxyl groups, forming monodentate, bidentate, or binuclear complexes depending on mineral structure and pH (11, 12).

Poorly crystalline Fe oxy(hydr)oxides, such as ferrihydrite (Fh), exhibit the highest P sorption capacity due to their large specific surface area and high reactivity. In contrast, more crystalline phases, such as goethite (Gt) and hematite (Hm), generally retain less P but with greater binding strength and stability (13, 14). Over time, phosphate becomes progressively stabilized through a process known as “P aging”, which reduces its desorption and, consequently, its bioavailability in soils. Furthermore, the transformation of poorly crystalline Fe phases into more crystalline forms during pedogenesis further decreases the reversibility of P adsorption, thereby exacerbating P limitation in older and highly weathered soils (15).

Siderophores are low-molecular-weight organic ligands produced by bacteria, fungi, and graminaceous plant roots under Fe-limiting conditions because of their high affinity for Fe³ +, which enables efficient Fe chelation and uptake (16–19). Bacterial siderophores are classified into catecholates, hydroxamates, and carboxylates based on their Fe³ + -coordinating functional groups (20). Among them, pyoverdines, primarily associated with hydroxamate and mixed-type coordination, are the main siderophores produced by fluorescent Pseudomonas species.

More than 60 structurally distinct pyoverdines synthesized by different Pseudomonas strains have been identified, sharing a conserved dihydroquinoline-type chromophore linked to a strain-specific peptide chain that determines binding properties and specificity (21). The ability to produce and utilize high-affinity Fe³ + -siderophore complexes gives microorganisms a competitive advantage under Fe-deficient conditions (22, 23). Siderophores also form stable complexes with metals such as Al, Ca, Zn, Mo, and Mn, influencing nutrient solubility, mobility, and soil nutrient dynamics (24). These interactions may indirectly influence the solubility, mobility, and availability of other nutrients, thereby affecting overall soil nutrient dynamics.

Siderophores can influence the solubility and dissolution kinetics of Fe minerals by promoting Fe oxy(hydr)oxide dissolution through surface-controlled, proton-, and/or ligand-promoted mechanisms. By disrupting Fe-P associations, they may also mobilize mineral-bound P and potentially improve Fe and P acquisition by microorganisms and plants (25). Nevertheless, despite the recognized role of Fe oxy(hydr)oxides as major phosphate sinks, previous research has focused primarily on siderophore-mediated Fe acquisition, whereas their contribution to P mobilization from Fe-associated pools remains comparatively poorly understood. In particular, limited comparative information is available on siderophore-mediated dissolution of Fe oxy(hydr)oxides differing in crystallinity and mineralogy, and the combined influence of mineral structure and pH on dissolution efficiency remains unclear. Moreover, most mechanistic studies have considered purified siderophores, leaving uncertain whether naturally siderophore-associated metabolites (SAM) can produce comparable or greater effects through complementary mechanisms. Resolving this question is important because metabolite mixtures requiring limited downstream purification could provide a more practical and sustainable basis for agricultural applications.

The central question of this study was therefore: how do Fe mineralogy, environmental pH, and co-produced metabolites interact to regulate siderophore-mediated Fe dissolution and the associated mobilization of P? We hypothesized that (i) siderophore-mediated Fe complexation and mineral dissolution would vary with pH and the crystallinity and mineralogy of Fe oxy(hydr)oxides; (ii) these mineral-scale interactions would translate into soil-specific changes in Fe extractability and P availability governed by soil pH and dithionite-citrate-bicarbonate-(DCB)-extractable Fe; and (iii) the complete pool of SAM would mobilize more Fe and P than purified siderophores because of complementary activities among its constituents. These hypotheses were tested using purified pyoverdine-like siderophores and SAM produced by the Antarctic Pseudomonas strain ANT_H12B. Specifically, we compared siderophore-mediated Fe complexation and dissolution among Fe oxy(hydr)oxides of contrasting crystallinity under acidic and alkaline conditions, quantified changes in Fe and P pools in four low-P soils with contrasting properties with soil pot assays, and compared the effects of purified siderophores and SAM. The contribution of this study consists in combining controlled experiments with soil incubations to examine how Fe mineralogy, pH, soil Fe status, and co-produced metabolites may influence siderophore-mediated nutrient mobilization. It also provides further mechanistic insight into the conditions under which siderophores may contribute to the mobilization of Fe-associated P and assesses whether retaining the naturally co-produced metabolite mixture could offer functional and technological advantages over the use of purified siderophores.

Production of SAM and purification of siderophores

Pseudomonas sp. ANT_H12B, a psychrotolerant strain isolated from Antarctic soil on King George Island (62°09.6010′ S, 58°28.4640′ W), was used for SAM and siderophore production. The strain was previously characterized for its genomic and metabolic traits and shown to produce more than 500 μM of a pyoverdine-like siderophore fraction, quantified by HPLC using commercial pyoverdine as an external standard (26).

SAM production followed a two-step procedure. The strain was first grown overnight in LB medium (10 g L -1 tryptone, 5 g L -1 yeast extract, 5 g L -1 NaCl) at 20 °C with shaking (150 rpm) to obtain the inoculum. Bacterial cultures (50 mL) were then centrifuged (8,000 rpm, 5 min), washed twice with 0.85% (w/v) NaCl solution and the resulting biomass was resuspended in 50 mL of sterile 0.85% NaCl and used to inoculate GCl medium (5.25 g L -1 glucose, 1.35 g L -1 NH 4 Cl, 0.96 g L -1 Na 2 HPO 4, 0.44 g L -1 KH 2 PO 4, and 0.2 g L -1 MgSO 4 ·7H 2 O) to an initial OD 600 of 0.06, specifically formulated to promote siderophore production. Cultures were incubated for 48 h at 10 °C with shaking (150 rpm). Bacterial growth parameters were not monitored during the incubation period, as growth-normalized siderophore production was not an endpoint of the present study. Quality control was performed by measuring siderophore activity, as described in Section 2.2.3, and comparing the results with those reported in our previous study (26). After cultivation, cultures were centrifuged (8,000 rpm, 8 min) and the supernatants filtered (0.22 µm) to obtain sterile supernatants (SAM), which were stored at -20 °C.

Siderophores were purified from SAM using a modified solid-phase extraction (SPE) protocol (27). Strata-X polymeric reversed-phase cartridges (500 mg, 6 mL-1; Phenomenex, Torrance, CA, USA) were conditioned with HPLC-grade methanol (VWR Chemicals, Pennsylvania, USA), and equilibrated with ultrapure water. Acidified SAM aliquots (100 mL, 1% (v/v) formic acid) (VWR Chemicals, Pennsylvania, USA) were loaded onto the cartridges, washed three times with ultrapure water, and siderophores were eluted using a mixture of deionized water and methanol (7:3, v/v) containing 0.1% (v/v) formic acid. Because siderophores produced by Pseudomonas sp. ANT_H12B were expected to be pyoverdine-like, their concentration in SAM was determined by HPLC analysis according to Bultreys et al. (28), using commercial pyoverdine from Pseudomonas fluorescens (Sigma-Aldrich, USA) as the external standard.

Siderophore and SAM chemical characterization

FT-IR and ¹H NMR spectroscopy of purified siderophores

The Fourier-transform infrared (FT-IR) spectrum of purified siderophores was recorded in transmission mode using a Bruker Vector 22 spectrophotometer equipped with a Globar source and a DTGS detector (Billerica, MA, USA). Spectra were acquired over the 4000–400 cm -1 range, with 128 scans at a resolution of 4 cm -1. Background spectra were collected against air before each measurement, and data were processed with OPUS 6 software. Functional group assignments are reported in Supplementary Table S1. 1 H nuclear magnetic resonance (1 H NMR) spectroscopy was acquired on a JEOL ECZR 600 MHz instrument (1H nLarmor = 600.23 MHz), using an 8 µs pulse (90°), 5.9 s delay, 2.9 s acquisition time, and 8 transients. The spectrum was processed with 0.3 Hz line broadening. Peak integration was performed using Delta 5.3.1.

Elemental composition of purified siderophores and SAM

The concentrations of total carbon (C), nitrogen (N), hydrogen (H), and sulfur (S) were determined by dry combustion using a CHNS macro-analyzer (UNICUBE, Elementar Analyzensysteme GmbH, Langenselbold, Germany). Samples were analyzed in triplicates. Oxygen (O) concentration was calculated by difference. Alfalfa OAS (B2273, Elemental Microanalysis Limited, Okehampton, UK) was used as a reference material. Quality assurance and quality control (QA/QC) recoveries were 98.8% for C, 100.2% for N, 97.7% for H, and 102.9% for S. The concentrations of total C, N, and S in SAM were also determined using a CHNS elemental analyzer. Nutrient content (P, potassium - K, and magnesium - Mg) was measured with inductively coupled plasma-optical emission spectroscopy (ICP-OES) iCAP 7400 (Thermo Fisher, USA).

Metabolic profile of SAM

Siderophore activity in SAM was assessed using the Chrome Azurol S (CAS) assay according to Schwyn and Neilands, as previously applied to Pseudomonas sp. ANT_H12B (26). Cell-free supernatants were mixed with CAS reagent at a 1:1 (v/v) ratio and incubated for 1 h in darkness. Absorbance was measured at 630 nm using a microplate reader. A calibration curve was prepared with deferoxamine mesylate (Sigma-Aldrich); 0.025 mM deferoxamine and sterile GCl medium were used as positive and negative controls, respectively. Measurements were performed in triplicate. The CAS assay was used to assess overall siderophore-mediated Fe 3+ -chelating activity, whereas the concentration of the pyoverdine-like fraction was estimated separately by HPLC using commercial pyoverdine from Pseudomonas fluorescens as an external standard.

The qualitative and semi-quantitative composition of extractable organic compounds was analyzed by gas chromatography-mass spectrometry (GC-MS) as described by Musialowski et al. (26). Briefly, bacterial cultures were centrifuged (8,000 rpm, 5 min) to obtain cell-free supernatants. Organic compounds were extracted from 100 mL of cell-free culture supernatant using 25 mL of chloroform in a separatory funnel for 3 min, and the extraction was repeated three times. The combined chloroform extracts were dried over anhydrous Na 2 SO 4 and evaporated under a stream of N 2. Samples were subsequently derivatized with 0.5 mL BSTFA (99:1, v/v) for 30 min at 70 °C. A blank sample was prepared using the same procedure. Compound identification was performed using Agilent Enhanced ChemStation (G1701EA, version E.02.00.493) and the Wiley Registry of Mass Spectral Data (version 3.2, 8th Edition with Structures). Compounds showing ≥80% spectral similarity were considered identified, whereas compounds with lower similarity were assigned to major chemical classes based on characteristic fragmentation patterns. Only reproducible peaks representing >0.1% of the total chromatogram area were considered. All experiments were performed in triplicate.

Assay of the pH-dependent Fe³ + chelation capacity of siderophores using FeCl 3

To assess the Fe 3+ chelation under different pH, purified siderophore (60 μM final concentration) solution were mixed with FeCl 3 ·6H 2 O solutions (1:1, v/v) to obtain final Fe³ + concentrations of 15, 30, 45, and 60 μM. This resulted in final concentrations of 60 μM siderophores and 15, 30, 45, and 60 μM Fe 3+ in the mixtures. The pH was adjusted to 5.0, 6.0, 7.0, 8.0, and 9.0 (± 0.1), and samples were incubated at room temperature for 30 min. The pH was readjusted, and the solutions were further incubated for 24 h. After incubation, samples were centrifuged (3,000 rpm, 10 min) and filtered (0.45 μm). Fe³ + concentrations in solution were determined spectrophotometrically using the o - phenanthroline method after reduction of Fe³ + to Fe² + with hydroxylamine hydrochloride (29). The amount of siderophore-chelated Fe³ + was calculated as the difference between initial and final Fe³ + concentrations in solution.

Synthesis and characterization of Fe oxy(hydr)oxides and siderophore-Fe complexation assays

Three Fe oxy(hydr)oxides were used. Ferrihydrite (Fh) was synthesized according to Schwertmann and Cornell (30). Briefly, FeCl 3 ·6H 2 O was dissolved in deionized water and the pH was adjusted to 7–8 with 1M KOH under vigorous stirring. The resulting suspension was centrifuged, dialyzed, and freeze dried. Goethite (Gt) and hematite (Hm) were synthesized and characterized as described by Celi et al. (31). Specifically, Gt was prepared by dissolving Fe(NO 3) 3 ·9H 2 O in deionized water, followed by the rapid addition of 5 M KOH under stirring. The resulting suspension was immediately diluted with deionized water and aged at 70 °C for 60 h. The product was then collected by centrifugation, washed with deionized water, and freeze-dried. Hematite was prepared by dissolving Fe(NO 3) 3 ·9H 2 O in HNO 3 preheated to 98 °C. The resulting suspension was let at 98 °C for 7 days, after which the product was collected by centrifugation, dialyzed, and freeze-dried. The specific surface area determined by N 2 adsorption was 232 m 2 g -1 for Fh (this study), whereas values of 40.2 and 46.8 m 2 g -1 were reported for Gt and Hm, respectively, by Celi et al. (31).

Zeta (ζ) potential was measured at pH 5.5 and 8.5 to characterize mineral surface charge. Briefly, oxides (0.02 g) were dispersed in 0.01 M KCl, shaken for 24 h, and adjusted to the target pHs. KCl (0.01 M) was added to obtain a final volume of 10 mL. Electrophoretic mobility was determined by Laser Doppler Velocimetry coupled with Photon Correlation Spectroscopy using a DELSA 400 LVD-PCS spectrometer (Beckman Coulter Inc., Canada), equipped with a 5 mW He-Ne laser operating at 632.8 nm. Data were converted into ζ values using the Smoluchowski equation (32). Results are reported in Supplementary Table S3.

For testing siderophore effects, Fe oxy(hydr)oxide suspensions were prepared at concentrations of 0.5, 1.0, 1.5, and 2.0 g L⁻¹, equilibrated to target pHs (5.5 or 8.5 ± 0.1) using 0.01 M NaOH or HCl to simulate contrasting soil conditions (acidic vs. alkaline), and mixed with a 60 µM purified siderophore solution. After 24 h of incubation at room temperature (22 ± 1 °C) under gentle shaking (100 rpm), samples were centrifuged (3,000 rpm, 10 min), filtered (0.45 µm), and Fe 3+ concentration in the filtrates was determined spectrophotometrically using the o -phenanthroline method after Fe 3+ reduction to Fe 2+ (29). A siderophore-free control was conducted with ferrihydrite at 2.0 g L⁻¹ and pH 5.5. This condition was selected as a screening condition for abiotic Fe release. Ferrihydrite is poorly crystalline and generally more reactive and soluble than the more crystalline goethite and hematite. Moreover, acidic conditions favor Fe³ + oxy(hydr)oxide dissolution, and the highest mineral loading provided the greatest amount of solid potentially available for dissolution, thus maximizing the possibility of detecting any background Fe release from the solid phase. No measurable Fe release was detected under this condition. Therefore, abiotic Fe release was expected to be negligible from the less reactive, more crystalline minerals and at pH 8.5. This conclusion represents a conservative inference rather than a direct measurement of background Fe release under each individual condition. Accordingly, Fe detected in siderophore-containing treatments is interpreted as predominantly associated with siderophore-mediated mobilization. After 24 h of incubation, the samples were centrifuged (3,000 rpm, 10 min) and filtered through a 0.45 µm membrane filter. Iron concentrations in the filtrates were determined by ICP-OES iCAP 7400 (Thermo Fisher, USA).

Experimental design for assessing Fe and P mobilization by purified siderophores and SAM in soils with contrasting properties

Soil sampling and physicochemical characterization

Four representative soils (Soils 1-4) were collected from different locations in the Piedmont region (Italy) and characterized for their physicochemical properties (Table 1). Soil samples were collected from the 0–20 cm depth. At each site, soil samples collected from three sampling points were thoroughly mixed to obtain one representative composite sample. The soils differed markedly in pH, Fe oxy(hydr)oxide, carbonate contents, and P forms. Two soils were subacidic to acidic, with high free Fe content and contrasting available P, where P immobilization was likely mostly attributable to sorption onto Fe minerals. The other two soils were alkaline, with different carbonate content, where P immobilization was expected to be dominated by Ca phosphate precipitation. Soil samples were air-dried and passed through a 2 mm sieve for texture, pH, and Olsen-P determination. A subsample was then further homogenized and passed through a 0.5 mm sieve for the determination of total C and N, CaCO 3, total P, and DCB-extractable Fe and P. All processed soil samples were stored at room temperature until analysis. Soil pH was determined potentiometrically in a 1:2.5 (w/v) soil-to- water suspension (33). Total carbon (TC) and total nitrogen (TN) were determined by dry combustion using a CHNS analyzer (UNICUBE, Elementar Analyzensysteme GmbH, Langenselbold, Germany). The CaCO 3 content was determined volumetrically using a Scheibler calcimeter (34), based on the volume of CO 2 released after reaction of dried soil with HCl. Total P (TP) was determined following sulfuric-perchloric acid digestion and quantified spectrophotometrically using the malachite green method at 600 nm (35), while bicarbonate-extractable P (Olsen-P) was measured on air-dried soils according to Olsen using molybdate colorimetry at 650 nm (36). Iron oxides were extracted using the DCB procedure (37). Briefly, 0.500 g of <0.5 mm soil was extracted with 25 mL of citrate-bicarbonate solution (8:1, v/v; 0.3 M sodium citrate:1 M NaHCO 3) and 0.5 g of sodium dithionite for 16 h. After centrifugation at 3000 rpm for 10 min and filtration through Whatman No. 4 filter paper, Fe was quantified by AAS (PerkinElmer AAnalyst 400, Norwalk, CT, USA), while DCB-extractable P was quantified by the molybdenum blue method at 650 nm using a KH 2 PO 4 calibration curve ranging from 0 to 0.8 mg P L⁻¹ (Table 1).

Pot incubation experiment with purified siderophores and SAM and determination of Fe and P pools

For the pot incubation experiment, each of the four soils received one of three treatments in triplicate: control (no siderophore addition), purified siderophores (SP), or siderophore-associated metabolites (SAM). To ensure identical P inputs across treatments, the control and SP treatments were supplemented with the same amount of P introduced with the SAM preparation. This design enabled a direct comparison of the effects of purified siderophores and SAM on soil P mobilization, such that any treatment differences reflected changes in soil P release rather than unequal P additions.

Before treatment application, soils were pre-incubated with deionized water for 24 h to equilibrate soil moisture. The incubation was conducted in pots (8 × 8 × 8 cm) containing 230 g of soil adjusted to 70% of field capacity. The SP treatment received purified siderophores (pH 6.0), whereas the SAM treatment received siderophore-associated metabolites (pH 5.21), with the amount applied adjusted to deliver an equivalent siderophore dose (500 µM) to the soil. Control pots received the same volume (20 mL) of sterile deionized water supplemented with the same amount of P as the SAM and SP treatments. After 24 h of incubation, soil samples were collected and stored at −20 °C until analysis of Fe and P pools (Supplementary Figure S1).

Two Fe and four P pools were determined using different extractants. Citrate and NaOH-EDTA extractions allowed direct comparison between metal mobilization and P release. Citrate-extractable fraction was obtained by shaking 1 g of soil with 40 mL of 0.20 M citrate/citric acid buffer (pH 6.0; 1:40, w/v) for 16 h at room temperature (38). The buffer was prepared by dissolving citric acid and sodium (Na) citrate in deionized water and adjusting the pH to 6.0 with HCl or NaOH. After centrifugation, supernatants were collected for both Fe and P determination.

NaOH-EDTA extractable fraction was obtained by extracting 1 g of soil with 20 mL of 0.25 M NaOH and 0.05 M EDTA (1:20, w/v) for 16 h at room temperature (39). Following centrifugation, supernatants were collected and analyzed for Fe and P. In addition, soluble P was extracted using 0.01 M CaCl 2 (1:40, w/v) for 16 h at room temperature, and Olsen-P was extracted with 0.5 M NaHCO 3 (pH 8.5) at a soil-to-solution ratio of 1:20 (w/v) for 30 min. Phosphorus was quantified using the malachite green method for CaCl 2 extracts (40) at 600 nm with KH 2 PO 4 standards ranging from 0 to 0.5 mg P L⁻¹ and the molybdenum blue method for citrate (41), NaOH-EDTA (42), and Olsen extracts (36) at 650 nm using KH 2 PO 4 standards over the range of 0–1 mg P L⁻¹. Iron concentrations in citrate and NaOH-EDTA extracts were determined by atomic absorption spectrometry (AAS; Thermo Fisher Scientific, USA). All analytical determinations were performed in triplicate.

Statistical analysis

Data are presented as mean ± standard error (SE) of three replicates (n = 3). One-way analysis of variance (ANOVA) was performed to assess the effects of siderophore-mediated Fe chelation in FeCl 3 solutions and in the presence of Fe oxy(hydr)oxides under different pH, as well as the effects of purified siderophores and SAM on Fe and P pool releases from soils with contrasting properties. Significant differences (p ≤ 0.05) were determined using the least significant difference (LSD) post hoc test. Differences in Fe 3+ chelation by purified siderophores between pH conditions within the same FeCl 3 or oxy(hydr)oxide concentration were evaluated using Student’s t -tests. Two-way ANOVA followed by the LSD post hoc test was performed to assess the effects of pH, FeCl 3 concentration, and their interaction on Fe³ + chelation. Statistical analysis was performed using RStudio (version 4.2.2). The results were presented on graphs obtained with ggplot2 (version 4.0.0).

Siderophore chemical and spectroscopic characterization

Elemental analysis (CHNSO) of the purified siderophore fraction confirmed its organic composition. Among the elements analyzed, O was the most abundant element (43.85%), followed by C (40.23%), N (9.28%), H (5.92%), and S (0.73%). This elemental composition is consistent with previously reported values for pyoverdines produced by Pseudomonas strains, supporting the successful purification of siderophore molecules belonging to this structural class. Notably, compared with canonical pyoverdines, S was detected in the siderophore produced by Pseudomonas sp. ANT_H12B, although at a relatively low level (0.73%).

The FT-IR spectrum of the purified siderophores is reported in Figure 1. A broad band observed in the 3200–3500 cm⁻¹ region was attributed to O–H stretching, suggesting the presence of hydroxyl groups and/or hydrogen-bonded water. Weak bands near 2900 cm⁻¹ corresponded to aliphatic C–H stretching. A prominent and sharp absorption peak at approximately 1650–1700 cm⁻¹ was assigned to C=O stretching vibrations, typically indicative of carbonyl functionalities such as amides or carboxyl groups. Additional bands in the 1550–1600 cm⁻¹ region likely arose from N–H bending and C–N stretching (amide II) and/or aromatic C=C vibrations. Signals detected between 1400 and 1500 cm⁻¹ were consistent with symmetric stretching of carboxylate groups (COO⁻) and C–H bending modes. In the fingerprint region, strong and multiple peaks were evident in the range of 1200–1300 cm⁻¹ and assigned to C–N and C–O stretching, while peaks between 1000 and 1200 cm⁻¹ were attributed to C–O stretching, supporting the presence of alcohol, ether, or carboxylate groups. The combination of these characteristic peaks suggests that the purified siderophores contained hydroxyl, carboxyl, and amide functional groups, which were consistent with the molecular features commonly found in pyoverdine-type siderophores.

The ¹H NMR spectrum (interval 0–9 ppm) was consistent with a predominantly aliphatic compound incorporating O-containing functional groups and a minor aromatic component (Figure 2). More specifically, a set of intense signals could be observed in the 0.81–5 ppm region, and assigned to terminal methyl (CH 3) and methylene (CH 2) groups of saturated alkyl chains, indicating that the molecule contained a structurally diverse set of aliphatic units. Additional peaks in the 1.5-3.0 ppm range were consistent with methylene and methine protons in β- or α-positions relative to electron-withdrawing groups, such as carbonyls or aromatic systems. Distinct resonances in the 3.0-4.2 ppm region were indicative of protons attached to carbons bearing O atoms (C–O), supporting the presence of oxygenated functionalities, such as alcohols, ethers, or esters. The chemical shift and relative intensity of these signals suggest that such functionalities were significant structural features. Weak downfield signals between 7.0 and 8.0 ppm could be attributed to aromatic or unsaturated protons, representing a limited proportion relative to the aliphatic framework. No resonances were observed in the 9–10 ppm or 10–12 ppm regions, thereby excluding the presence of aldehydic or carboxylic acid protons, respectively.

Chemical characterization and metabolic profile of SAM

The elemental composition of SAM was as follows: 1210 mg L⁻¹ for C, 296 mg L⁻¹ for N, 23.0 mg L⁻¹ for S, 26 mg L⁻¹ for P, 141 mg L⁻¹ for K, and 18.1 mg L⁻¹ for Mg.

The qualitative and semi-quantitative characteristics of SAM characterized by a strong predominance of fatty acids (total concentration of 77.85 mg L⁻¹), with comparatively minor contributions from other metabolite classes, alongside a substantial concentration of siderophores (Table 2). Among the quantified extractable organic metabolites reported in mg L⁻¹, fatty acids accounted for approximately 83% of the total detected mass, confirming their predominance within the GC–MS-detectable fraction of SAM. Among fatty acids, hexadecanoic acid (28.69 mg L⁻¹) and cis-9-hexadecenoic acid (26.64 mg L⁻¹) were the most abundant compounds, followed by trans-13-octadecenoic acid (7.86 mg L⁻¹) and tetradecanoic acid (4.49 mg L⁻¹), while other fatty acids such as octadecanoic, n-pentadecanoic, dodecanoic, and nonanoic acids were present at lower concentrations. Minor components included alcohols (7.7 mg L⁻¹) and esters (7.6 mg L⁻¹), whereas sugars were detected only in trace amounts (0.17 mg L⁻¹). An indolic compound was also identified at low concentration (2.5 μg L⁻¹). In addition to these metabolites, pyoverdine-like siderophores were quantified at a concentration of 500 μM.

Siderophore complexation capacity

Capacity for siderophore-mediated Fe³ + chelation using FeCl 3

Iron 3+ chelation capacity of the purified siderophore fraction in a FeCl 3 solution at different concentrations and pH is reported in Figure 3. Overall, Fe chelation increased progressively with increasing FeCl 3 concentrations. At lower FeCl 3 concentration (15 µM), the amount of complexed Fe was relatively limited across all pH conditions, with values of 0.87, 0.48, 1.26, and 1.52 µM at pH 5, 6, 8, and 9, respectively. As the FeCl 3 concentration increased to 30 and 45 µM, a concomitant increase in chelated Fe 3+ was observed, though significant only at pH 6, 8 and 9 (p ≤ 0.01). At 30 µM, the concentrations of chelated Fe³ + reached 2.23, 1.46, 2.38, and 2.44 µM at pH 5, 6, 8, and 9, respectively, compared to 15 µM. A more pronounced increase was observed at 45 µM, with values of 2.57, 2.38, 4.27, and 4.34 µM at pH 5, 6, 8, and 9, respectively. At the highest concentration (60 µM), the chelation capacity further raised, with chelated Fe³ + concentrations of 3.82, 5.26, 5.19, and 6.24 µM at pH 5, 6, 8, and 9, respectively.

The effect of pH on siderophore Fe³ + binding capacity by siderophores was also evident, by varying significantly across the tested pHs. The aliquot of chelated Fe compared to the added amount was influenced by the pH, passing from about 6-9% at pH 5 and 6, to 9% at pH 8 and 10% at pH 9. Conversely, at each pH the chelated aliquot remained nearly constant at increasing Fe additions, indicating that the binding sites were not yet saturated. In particular, Fe complexation increased progressively with increasing pH, reaching its highest values at pH 9.

Two-way ANOVA confirmed significant effects of both pH and FeCl 3 concentration on Fe³ + chelation, as well as a significant interaction between the two factors (Supplementary Table S2). FeCl 3 concentration showed the strongest effect (F = 308.484), followed by pH (F = 39.511), while the pH × Fe³ + interaction was also significant (F = 7.628; p ≤ 0.01 for all effects).

Capacity for siderophore-mediated Fe³ + mobilization from Fe oxy(hydr)oxides

The capacity of siderophores to chelate Fe³ + in the presence of three Fe oxy(hydr)oxides (i.e., Fh, Gt, Hm) at two different pH conditions (5.5 and 8.5) and across four increasing mineral concentrations (0.5, 1, 1.5, 2 g L⁻¹) is shown in Figure 4.

Poorly ordered Fh exhibited the highest overall Fe 3+ dissolution among the tested oxy(hydr)oxides (Figure 4A). In the presence of Fh, dissolved Fe 3+ concentration extracted by complexing agents increased with mineral concentration at both pHs. At pH 8.5 the dissolved concentration increased nearly linearly at increasing Fh concentrations, while at pH 5.5 the increments progressively declined. However, dissolution efficiency was consistently higher at pH 5.5 than at pH 8.5, especially at intermediate concentrations, where Fe³ + dissolution at pH 8.5 was 30% and 37% lower compared to pH 5.5 at Fh concentrations of 1 and 1.5 g L⁻¹, respectively. Even in the presence of Gt, Fe 3+ dissolution increased progressively with increasing Gt concentration under both pHs (Figure 4B). However, only with this Fe mineral the extent of solubilization was consistently higher at the alkaline pH. At low Gt concentrations (< 1.5 g L⁻¹), differences between pH conditions were not statistically significant, whereas they became more pronounced at the highest concentration (2 g L⁻¹), with Fe³ + dissolution at pH 8.5 being 40% higher than at pH 5.5. In contrast, Hm showed the lowest Fe 3+ dissolution overall, with maximum values of 3 μM at pH 5.5 and 1.91 μM at pH 8.5 at the highest mineral concentration (Figure 4C).

As observed in the chelation experiment with FeCl 3, the aliquot of dissolved Fe remains nearly constant at increasing Fe oxy(hydr)oxide dose. For Fh, averagely 5.8 and 4.8 μmol of Fe were extracted from each g of solid at pH 5.5 and 8.5 respectively; for Gt, the extracted amounts were averagely 1.6 and 1.7 μmol g -1 at pH 5.5 and 8.5, respectively; finally, for Hm, the averages were 1.5 and 1.0 μmol g -1 at pH 5.5 and 8.5, respectively (data not shown). Under alkaline conditions, Hm consistently exhibited lower Fe dissolution than Fh and Gt across all concentrations, with reductions of 83% and 42%, respectively, at 2 g L⁻¹. At acidic pH, Fe dissolution from Hm was always lower than with Fh, while the comparisons with Gt depended on mineral concentration. Specifically, Hm showed slightly lower values than Gt at 0.5 and 1 g L⁻¹, by 31% and 12%, respectively, but similar chelation at 1.5 and 2 g L⁻¹, increasing by 10% and 27%.

Figure 5 provides a comparative heatmap of Fe mobilization from the three Fe oxy(hydr)oxides across increasing mineral concentrations under acidic (pH 5.5; Figure 5A) and alkaline conditions (pH 8.5; Figure 5B). At pH 5.5, Fh showed significantly greater Fe mobilization than Gt and Hm at all tested mineral concentrations (p ≤ 0.05), whereas no significant differences were observed between Gt and Hm (Figure 5A). At pH 8.5, Fh also showed significantly greater Fe mobilization than both Gt and Hm across the entire concentration range (Figure 5B). However, under alkaline conditions, Gt and Hm showed a concentration-dependent differentiation: Gt mobilized significantly more Fe than Hm at 1.0 and 1.5 g L⁻¹ (p ≤ 0.01), whereas the two minerals did not differ significantly at 0.5 and 2.0 g L⁻¹. The color intensity also shows the general increase in Fe mobilization with increasing mineral concentration, which was particularly pronounced for Fh at both pH values.

The ζ-potential of the Fe oxy(hydr)oxides showed marked mineral- and pH-dependent differences (Supplementary Table S3). At pH 5.5, Fh and Gt exhibited positive surface charges, with ζ-potentials of 19.0 ± 0.53 and 44.0 ± 0.39 mV, respectively, whereas Hm was slightly negatively charged (−2.1 ± 1.44 mV). Increasing the pH to 8.5 markedly decreased the ζ-potential of all three minerals. Ferrihydrite became strongly negatively charged (−34.6 ± 0.42 mV), Gt approached electrical neutrality (−0.1 ± 0.82 mV), and Hm became more negatively charged (−25.8 ± 0.56 mV). Overall, the shift in surface charge with increasing pH was greatest for Fh and Gt, whereas Hm remained negatively charged under both pH conditions.

No Fe release from any of the three Fe oxy(hydr)oxides was observed at each tested pH in absence of siderophore (data not shown).

Soil properties and effect of siderophores and SAM on Fe and P mobilization from soils

Physicochemical characteristics of surveyed soils

The physicochemical properties of the four surveyed soils, showing substantial variability in pH, nutrient concentrations, and texture, are summarized in Table 1. Soil texture ranged from coarse (sandy loam) to finer textures (loam and silty clay loam), with sand content varying from 9% to 59%. The OC (0.37-4.36%) and TN (0.09-0.22%) contents were also variable, reflecting differences in organic matter inputs and soil fertility. Soil pH ranged from acidic (4.66 in soil 1) to strongly alkaline conditions (up to 8.48 in 4); two of the soils (soil 3 and 4) developed on carbonatic parent material and exhibited different contents of CaCO 3. The other two soils (soil 1 and 2) were free of carbonates, but richer in free Fe oxy(hydr)oxides. Indeed, DCB-extractable Fe ranged from 5.54 g kg -1 in soil 4 to 21.05 g kg -1 in soil 1, indicating differences in the dominant mechanisms of P immobilization. Total P and DCB-extractable P exhibited indeed marked differences, reflecting variability in both total P reserves and Fe-associated P fractions. The two calcareous soils had greater total P contents, however only a small fraction was associated with the free Fe-minerals, DCB-extractable P representing only 12 to 23% of the total P, while in the other soils this fraction varied from 66% in soil 2 to 86% in soil 1. Available P (Olsen-P) was consistently low to medium-low across all soils (3.62-11.99 mg kg⁻¹), providing suitable conditions to investigate P and Fe release induced by siderophore and SAM application.

Effect of siderophores and SAM on soil P and Fe pools across contrasting soil types and Fe levels

Changes in Fe and P pools were evaluated under control conditions (C; H 2 0 addition) and two treatments (SP and SAM) (Figures 6, 7). The extraction with 0.01M CaCl 2 allowed to evaluate the readily soluble pool; 0.5 M bicarbonate at pH 8.5 the plant-available pool; a citric acid/citrate buffer at pH 6.0 the fraction exchangeable with organic anions and/or associated with loosely bound Fe forms (43). The extraction with NaOH-EDTA is commonly used for characterizing organic P forms (39), but it may also extract some loosely adsorbed Pi. Potential changes in Fe extractability induced by siderophores and SAM were evaluated by measuring Fe, beside P, in the latter two extracts.

The effects of siderophores and SAM on Fe extractability varied markedly across soils differing in pH and Fe status, as reflected in both NaOH-EDTA and citrate-extractable Fe (Figure 6). Without siderophores or SAM addition, NaOH-EDTA-extractable Fe was highest in the acidic soil (soil 1) and lower in the circumneutral (soil 2) and alkaline soils (soil 3 and 4), consistently with their free Fe content (Figure 6A). Following siderophore addition (SP treatment), NaOH-EDTA-extractable Fe decreased significantly in the acidic soil by 45% (p ≤ 0.05), slightly increased in the circumneutral soil, and remained relatively similar in both alkaline soils. When soils were supplemented with microbial metabolites (SAM treatment), NaOH-EDTA-extractable Fe further decreased in the acidic soil by 77% compared to its corresponding control, while the circumneutral and alkaline soils showed relatively stable and comparable values to each other and to previous treatments, with only minor variation (Figure 6A). Regarding citrate-extractable Fe, in the absence of complexing agents, the circumneutral soil showed the highest values, followed by the acidic soil, while both alkaline soils had substantially lower concentrations (Figure 6B). Under SP treatment, citrate-extractable Fe increased in the neutral soil by 8% relative to the C treatment, remained similar in the acidic soil, and showed slight increases in alkaline soil 4. Under SAM treatment, citrate-extractable Fe further increased in the neutral soil by 22% relative to the C treatment, remained relatively unchanged in the acidic soil, and increased slightly in the alkaline soils 3 and 4 (Figure 6B).

Phosphorus availability and distribution in different pools were strongly influenced by both treatment (SP and SAM) and soil type (Figure 7). The promptly soluble P extracted with CaCl 2 was negligible in the most acidic soil followed by the most alkaline one and was higher in the two soils with intermediate pH. SP treatment slightly increased P solubility in the soils with extreme pH values, while it did not change or even decreased P release in the soils with intermediate pH. SAM treatment markedly increased CaCl 2 -extractable P, particularly in the two alkaline soils, with increases of ~11–40- fold relative to the C treatment, and with ~10- and 15-fold increases compared with SP treatment. Both SP and SAM treatment had negligible or only minor effects in the acidic and circumneutral soils compared to the C treatment (Figure 7A). SAM treatment also produced the highest increases in NaHCO 3 -extractable P (Olsen-P), representing the bioavailable P, across all soils, with increases of ~1–26- fold relative to the C treatment, with the greatest effect observed in alkaline soil with higher Fe (soil 3), followed by alkaline soil with lower Fe (soil 4) and circumneutral soil (soil 2) (Figure 7B). The acidic soil showed the lowest bioavailable P across treatments. SP treatment moderately increased bioavailable P compared to the C treatment, particularly in the circumneutral soil and in the alkaline one with higher Fe (soils 2 and 3, respectively), but the effect was smaller than that of SAM treatment. The addition of SAM significantly enhanced citrate-extractable P, with increases of ~0.3–5- fold (p ≤ 0.01), especially in the alkaline soil with higher Fe (Figure 7C). SP treatment significantly decreased the citrate-extractable P fraction in acid and alkaline soils by 10-76% (p ≤ 0.05), while no significant change was observed in the circumneutral soil (Figure 7C). NaOH-EDTA extractable P was highest in the circumneutral soil across all treatments, followed by the alkaline soil with higher Fe (soil 3) (Figure 7D). SAM treatment further increased this P pool by 25-104% relative to the C treatment in all soils, while the SP treatment showed no effect (Figure 7D).

The elemental analysis showed C, H, N, and O contents consistent with those reported for pyoverdine-like non-ribosomal peptide siderophores produced by Pseudomonas spp. (44, 45). In addition, a minor S fraction was detected, which may reflect the presence of S-containing amino acids, possibly cysteine, and suggest additional structural variability in ligand composition (46). The spectroscopic analyses provided further information on the functional groups present in the siderophore purified fraction and identified structural features that could influence its reactivity in soils. The FT-IR spectrum showed signals attributable to hydroxyl, carboxylate, and amide functional groups which may participate in Fe³ + coordination. In particular, the strong C=O stretching signal and the presence of amide II bands were consistent with peptide-like structures typical of pyoverdines, while abundant C–O and C–N bands suggested the presence of a high density of O- and N-donor ligands. These groups could provide multiple coordination sites for Fe³ + and facilitate complex formation. Signals potentially attributable to S-containing moieties were also identified by FT-IR, in agreement with the detection of S by elemental analysis, However, additional structural analyses would be required to establish the identity and coordination role of these moieties. The ¹H NMR data further support our interpretation, revealing a structure dominated by aliphatic components together with protons associated with O- and N-containing environments, consistent with peptide-based siderophores, such as pyoverdines. Minor aromatic resonances may be associated with a pyoverdine-like chromophore (47).

To evaluate the functional behavior of these purified pyoverdine-like siderophores, Fe³ + -complexation experiments were conducted. The results showed a clear concentration-dependent binding behavior, with increasing FeCl 3 levels resulting in progressively higher Fe complexation. However, the relatively low Fe binding efficiency suggested that only a fraction of the added Fe was complexed under the experimental conditions. The linear increase in complexed Fe with increasing FeCl 3 addition (Supplementary Figure S2), together with the nearly constant fraction of complexed Fe across the tested concentrations, indicated that an apparent saturation plateau was not reached. Several factors could account for this behavior, including the presence of binding sites with different affinities or accessibility, the formation of complexes with different stoichiometries, steric constraints, partial occupation of functional groups, and competing reactions such as Fe³ + hydrolysis, but the present data do not allow these possible explanations to be distinguished.

The presence of S and the weak, non-stoichiometric, Fe binding behavior could indicate some functional similarities with other mixed-donor siderophores, such as pyochelin (48). However, this comparison represents a hypothesis, because neither the identity of the S-containing moieties nor their involvement in Fe coordination was directly determined. Similarly, the absence of a saturation plateau could reflect the progressive contribution of additional binding sites or the formation of complexes with different stoichiometries, as previously observed for some bacterial siderophores (49), but verification would require speciation modeling and more detailed structural and thermodynamic analyses. The co-occurrence of other factors, including partial occupation of functional groups, steric constraints within the molecular structure, or competition with other processes co-occurring in solution (e.g., Fe³ + hydrolysis), cannot be ruled out. The observed influence of pH on Fe³ + complexation was consistent with a possible role of protonation-deprotonation equilibria in controlling pyoverdine-like siderophore activity. The greater Fe³ + binding observed under alkaline conditions may reflect increased deprotonation of key functional groups (e.g., hydroxyl and carboxylate moieties), which enhanced their electron-donating capacity and promoted Fe³ + coordination (50). Nevertheless, the protonation state of the ligands and the chemical speciation of Fe were not directly determined; therefore, this explanation should be regarded as a mechanistic hypothesis.

Compared with the ideal 1:1 stoichiometry reported for highly efficient catecholate siderophores such as bacillibactin from Bacillus subtilis (51), the pyoverdine-like siderophores produced by the Pseudomonas strains investigated in this study exhibited a comparatively low apparent complexation capacity. This behavior may be related to differences in ligand composition, coordination geometry, accessibility of donor groups, or experimental conditions. Because Fe³ + affinity varies substantially among ligand classes, with catecholate siderophores generally exhibiting higher stability constants than hydroxamate or mixed-ligand systems (18, 52, 53), the structural characteristics of Pseudomonas -derived siderophores may result in lower or more variable binding efficiency. Although a 1:1 Fe³ + complexation stoichiometry is commonly reported for pyoverdine-like siderophores, structural variability among bacterial strains and the possible presence of more than one compound in the purified fraction may account for deviations from ideal complexation behaviour (54–56). Together, these results show that the Fe³ + complexation capacity by the siderophore fraction was strongly governed by the interplay between intrinsic molecular features and external environmental conditions (e.g., pH and Fe concentration).

Beyond complexation in solution, the interaction between siderophores and Fe oxy(hydr)oxides may also depend on mineralogy and pH (Supplementary Figure S3). Ferrihydrite has poor crystallinity, abundant structural defects and high specific surface area, properties generally associated with relatively high reactivity and susceptibility to ligand-promoted dissolution. At pH 5.5, Fh showed a positive ζ-potential (+19.0 mV) suggesting that under these conditions the observed Fe mobilization may have involved ligand interactions with accessible surface sites, followed by complexation of released Fe. At pH 8.5, Fh exhibited a negative ζ-potential (-34.6 mV), which could restrict the approach of negatively charged ligand groups through electrostatic repulsion. Nevertheless, its high intrinsic reactivity may have allowed interactions at accessible reactive surface sites. In contrast, Gt showed a stronger response to alkaline conditions. This pattern could be consistent with increased ligand deprotonation and ligand-exchange reactions at higher pH, but adsorption and surface-complexation measurements would be required to confirm this interpretation. Hematite was the least soluble and most stable phase, and displayed limited responsiveness, particularly at alkaline pH where the siderophore fraction had greater complexing capability. However, at pH 8.5, Gt was close to its point of zero charge (-0.1 mV), whereas Hm exhibited a clearly negative surface charge (-25.8 mV). Therefore, the lower Fe mobilization from hematite could reflect electrostatic constraints on ligand–surface interactions, together with its lower intrinsic solubility and reactivity. These findings collectively indicate that the capacity of siderophores in mobilizing Fe is not solely controlled by ligand properties and environmental pH, but also by mineral phase characteristics that govern dissolution kinetics and surface reactivity (18, 57–59).

In the soil experiments, the Fe extraction patterns provided further evidence that siderophores and SAM affected Fe and P dynamics, although the underlying processes could not be identified directly. In acidic soils, the decrease in NaOH-EDTA-extractable Fe following siderophore and SAM addition may indicate that the adsorption of these organic compounds on Fe-mineral surfaces hampered the direct contact between the extractant (citrate or EDTA) and surface Fe, thus limiting Fe extraction, as already observed for other adsorbed anions (31). Alternative explanations, including changes in Fe partitioning or extractability, cannot be ruled out.

Besides siderophores, the organic/fatty acids detected in SAM by GC-MS contain carboxyl groups that could participate in Fe interactions. Although their chelating capacity is expected to be weaker than that of pyoverdine-like siderophores, they may further contribute to interactions with Fe mineral surfaces. Minor contributions from indolic compounds, sugars, alcohols, and esters cannot be excluded, as these molecules contain O- or N-bearing groups capable of weak metal interactions. We hypothesize that competitive interactions of these organic ligands with binding sites on soil Fe oxy(hydr)oxides contributed to the observed P mobilization. In addition, Fe mobilization through the formation of soluble siderophore–Fe complexes may have contributed to weakening Fe–P associations and increasing P availability, particularly in neutral to alkaline soils. However, the limited complexation efficiency observed in solution suggests that only a fraction of Fe is effectively mobilized, constraining the overall impact on P release.

Soil pH is a major factor controlling Fe and P availability, and differences in the pH of the applied solutions could potentially influence their mobilization. In the present study, purified siderophores and SAM were applied at pH 6.0 and 5.21, respectively. However, considering the small amount of solution applied relative to the amount of soil and the buffering capacity of the soil matrix, these differences in solution pH did not substantially alter soil pH. Therefore, although a slight contribution of local and transient pH changes at the soil-solution interface cannot be completely excluded, the observed differences in Fe and P mobilization between SP and SAM are not primarily driven by changes in bulk soil pH. Rather, they reflect differences in the composition and activity of the compounds present in the two treatments.

The enhancement of P release following SAM addition was observed with all P extraction methods employed. With the mild CaCl 2 extractant, the enhancement in P solubility was particularly evident in soil 4 (nearly 40 times), which was characterized by the most alkaline pH, the greatest organic matter content and the lowest Fe Fe oxide concentration. These properties could be associated with a comparatively low availability of anion-binding sites on oxide surface. Under these conditions, organic anions supplied by SAM may have competed for the binding sites with phosphate and the relatively high concentration of native soil organic matter, thus favoring competitive phosphate displacement and subsequent desorption. This interpretation is consistent with the observed response. Alternatively, complexation of Ca² + by siderophores and other SAM components may have contributed to P dissolution by decreasing Ca² + activity in solution. Although this mechanism was not directly assessed in the present study, its interpretation is aligned with recent evidence indicating that siderophores may contribute to P mobilization in calcareous soils not only indirectly through Fe complexation, but also through interactions with Ca-bound P pools. In alkaline, carbonate-rich environments, where phosphate availability is primarily controlled by precipitation and association with Ca phases, siderophore ligands may decrease free Ca² + activity and shift dissolution equilibria toward phosphate release. Cui et al. (25) experimentally demonstrated that microbial siderophores can act as phosphate-solubilizing agents and improve soil P availability, although their study primarily addressed Fe-associated P in acidic soil. More recently, Tabbasum et al. (60) found that siderophore production was particularly pronounced among phosphate-solubilizing bacteria isolated from P-deficient alkaline calcareous soils and associated this trait with the chelation of Ca² + and the release of Ca-bound phosphate. However, siderophore production co-occurred with organic-acid secretion, indicating that these mechanisms are likely complementary rather than mutually exclusive. Accordingly, the large response observed in soil 4 may have resulted from the combined displacement of surface-bound phosphate and ligand-induced perturbation of Ca–P equilibria. Because Ca complexation and individual Ca–P fractions were not directly measured here, this mechanism should nevertheless be regarded as plausible rather than conclusively demonstrated.

Concerning plant-available P (Olsen-P), the most acidic soil (soil 1) exhibited the greatest increase following SAM addition (nearly 27 times), despite having the highest Fe oxide content and more than 80% of its total P being DCB-extractable, compared with only 12% of soil 4 (not shown). The response observed during Olsen extraction at pH 8.5 may be consistent with enhanced interaction between deprotonated organic ligands and mineral-associated cations, shifting the available P from the negligible concentration of the control (less than 1 mg kg -1) up to the satisfactory value of 18.3 mg kg -1, thus demonstrating that these compounds can substantially contribute to plant P nutrition even under the most unfavourable scenario. Similarly, SAM addition increased P availability in all other soils, shifting them from P-deficient to P-sufficient conditions.

In all soils, the P pools extracted with EDTA or citrate, also increased by SAM addition, although the differences from the controls were less pronounced. In the carbonate soils, which had lower Fe oxy(hydr)oxide content, the increments in P release with siderophore or SAM addition within each soil tended to coincide with increases in Fe release, a pattern consistent with a possible contribution from Fe mobilization. In the non-calcareous soils, which contained more Fe, P release became more relevant with siderophore and SAM addition, despite Fe release declined with both extractants in soil 1 and with EDTA in soil 2. This decoupling between P and Fe release suggests that competitive ligand adsorption or other processes may have contributed more strongly than Fe dissolution under these conditions.

Overall, these results indicate that the tested siderophores exhibited moderate Fe chelation ability in solution and produced a comparatively limited effect on P mobilization in soils. Siderophore-mediated Fe dissolution may have contributed to P release, but the results suggest that this was not the only controlling process. In contrast, SAM produced a stronger increase in P availability, even in alkaline soils, where P is typically immobilized by Ca. This suggests that the stronger effect of SAM may involve multiple, complementary processes, potentially including competition for sorption sites and metal complexation. The molecules produced by the Antarctic Pseudomonas strain ANT_H12B may represent a remarkable resource for enhancing P availability and the exploitation of soil legacy-P in a wide range of soil types and P immobilization mechanisms. The important, unexpected contribution of the associated metabolites roles out the need for siderophore purification, thus simplifying the potential productive process for agricultural applications.

From an applied perspective, the consistently stronger effects of SAM compared with the purifiedsiderophore fraction suggest that retaining the naturally co-produced metabolite mixture may offer both functional and technological advantages. Avoiding extensive downstream purification could reduce processing requirements while preserving potentially complementary activities of different microbial metabolites. This interpretation is consistent with our recent studies using cell-free siderophore-associated metabolites from ANT_H12B, in which SAM enhanced soil Fe, P and K availability and maintained or improved radish productivity under reduced P and K fertilization (61), while a SAM-based formulation with molasses enhanced basil growth and modified rhizosphere functioning (62). Thus, SAM efficacy is increasingly supported experimentally, but the mechanisms underlying its activity, including interactions between siderophores and co-produced metabolites, remain in part unresolved. The present study moves beyond demonstration of efficacy toward mechanistic resolution of SAM activity by examining nutrient mobilization in a controlled system. Understanding these mechanisms is essential for rational optimization of SAM composition, dose and application for soil biostimulation and plant biofortification. Future work should address dose–response relationships, persistence, plant nutrient uptake and yield, formulation stability, repeated application, microbiome interactions, and performance across contrasting soils and field conditions.

This study demonstrates that siderophore-mediated Fe chelation contributes to P mobilization, although its effectiveness depends strongly on pH, Fe oxy(hydr)oxide crystallinity, and competing soil processes. Spectroscopic analyses identified peptide-like structures and O- and N-donor groups consistent with pyoverdine-type siderophores. However, the purified fraction showed limited Fe³ + -binding efficiency and only moderate effects on soil P availability, whereas SAM produced more pronounced responses, particularly in alkaline soils. This contrast suggests that interactions between siderophores and co-produced metabolites are more effective than the purified fraction alone and that preserving this metabolic complexity may be essential for developing practical microbial-derived products for nutrient management. By linking the chemical properties of the siderophore fraction to the functional behavior of SAM, our findings provide a mechanistic basis for the previously observed capacity of ANT_H12B-derived SAM to enhance nutrient availability and support plant productivity under reduced fertilization. The study thus advances the evidence from demonstrating SAM efficacy toward identifying the processes underlying its activity. A clearer understanding of these mechanisms will support the rational optimization of SAM-based formulations, including their composition, dosage, and application strategy, for soil biostimulation and plant biofortification. Future research should evaluate metabolite persistence, plant nutrient uptake, microbiome responses, the effects of repeated applications, and performance across contrasting soils under plant-based and field conditions to translate these mechanistic insights into effective agricultural applications.