Uranium is a naturally occurring element that is ubiquitous in the Earth’s crust (e.g., soil and water). However, anthropogenic activities related to uranium mining, the nuclear energy industry, or weapon manufacturing may involve the redistribution of this element in the environment, sometimes resulting in locally high U concentrations (Kolhe et al., 2018). Due to U radiological and chemical toxicity, remediation strategies are highly recommended to prevent the potentially harmful effects for the environment and health.

Several remediation approaches have been proposed to remove inorganic contaminants from polluted sites leading to a decrease in their associated risks (Gavrilescu et al., 2009). Traditional strategies based on physico-chemical methods, such as precipitation, evaporation, or extraction amongst others, have been used to treat heavy-metal contaminated sites (Selvakumar et al., 2018). However, these approaches are known to be costly at environmental level in comparison with biologically based remediation. Both types of techniques are often additionally used for more efficient and economical rehabilitation of contaminated areas (Khalid et al., 2017). In recent years, a wide range of microorganisms isolated from contaminated sites (such as mining areas) have been shown to possess the ability to immobilize U (Banala et al., 2021). The main strategies which have been adopted by microorganisms to avoid U cytotoxic effects are based on different mechanisms such as U biosorption at the cell surface (Merroun et al., 2006), U biomineralization (Macaskie et al., 2000; Beazley et al., 2007; Lopez-Fernandez et al., 2021), U intracellular accumulation (Merroun et al., 2003; Gerber et al., 2016) and U biotransformation, either by reduction (Lovley et al., 1991; Williams et al., 2013) or by oxidation (DiSpirito and Tuovinen, 1982). In particular, one of the most promising approaches consists of U-biomineralization under aerobic conditions (Beazley et al., 2011; Tu et al., 2019; Sánchez-Castro et al., 2020).

The uranium biomineralization process probably occurs in the presence of ligands such as carbonates, phosphates and hydroxides, which act as nucleation sites generated by microbial activity (Lloyd and Macaskie, 2000; Wufuer et al., 2017). Amongst the different mineralization processes, U-phosphate precipitation has been well documented in recent years by numerous studies (Zhang et al., 2020). They are generally bi-phasic processes based on passive U binding at the cell surface through phosphate and/or carboxyl groups and a metabolism-dependent release of orthophosphates from organic phosphate substrates [e.g., glycerol-2-phosphates (G2P), glycerol-3-phosphates (G3P)] mediated by phosphatase activity (Lopez-Fernandez et al., 2021). The resulting orthophosphates interact with uranyl ion (UO₂²⁺), the most common soluble form of environmental U, making it immobile and consequently less toxic. Specifically, meta-autunite-like U phosphate forms have been widely studied due to their long-term stability under different physico-chemical conditions (Beazley et al., 2009). These U-phosphate compounds are attractive for bioremediation purposes rather than the uraninite forms produced during bioreduction processes which are known to be less stable and easily oxidized (Wang et al., 2013). Previous studies have reported the formation of U phosphates in presence of determined microbial strains and organic phosphate sources such as G3P (Beazley et al., 2011) or G2P (Macaskie et al., 2000; Newsome et al., 2015; Sánchez-Castro et al., 2020), as well as inorganic phosphates (Pi). The microbial-mediated formation of U phosphates is known to be affected by different physico-chemical factors such as pH (Zheng et al., 2018), redox conditions (Salome et al., 2013), co-existing cations/anions (Wei et al., 2019) and humic substances (Boiteau et al., 2018), which may decrease the bioprecipitation efficiency. However, to the best of our knowledge, no studies on the impact of different forms of phosphates (organic/inorganic) on U biomineralization have been conducted so far. The presence of different phosphate forms should be considered, especially when phosphate solubilizing bacteria (PSB) are used as bioremediation agents or are present in the medium. PSB are known to play a major role in enhancing phosphate-induced immobilization of heavy metals such as U for bioremediation purposes (Sowmya et al., 2014). However, eutrophication issues should be contemplated since there is a limit on the amount of phosphate which can be added to the environment (Park et al., 2011). Therefore, the use of PSB (e.g., the species of the genus Microbacterium) in the bioremediation of heavy metals could contribute to the circular economy by using cheap and readily available material which is rich in insoluble phosphates.

This study addresses the impact of phosphate forms (organic/inorganic) in U biomineralization by Microbacterium sp. Be9, a bacterial strain isolated from U mill tailings. In previous studies, this bacterium showed high tolerance to different heavy metals (Sánchez-Castro et al., 2017). Moreover, members of the Microbacterium genus have been described for their potential role in uranium-contaminated water remediation. They have been shown to accumulate uranium both at cell surface level and also extracellularly as uranium phosphates at pH 4.5 (Nedelkova et al., 2007) displaying multiple detoxification mechanisms involving phosphates (Theodorakopoulos et al., 2015). According to the latter works, Microbacterium sp. strain Be9 would display similar U interaction mechanisms mediated by different enzymatic activities upon phosphate source (phosphatase in presence of organic phosphates and polyphosphatase in a free phosphate system) leading to the biomineralization of U as U phosphate mineral phases. Thus, Be9 could precipitate U under the presence of organic phosphates. Bacterial strains with phosphatase activity are capable of releasing orthophosphates from organophosphate sources in order to avoid the stress caused by U. In presence of inorganic phosphates, most of the U should remain in the insoluble form. In our study, systems with three different sources of phosphate were assayed (no-phosphate, organic and inorganic phosphates) to investigate the mechanisms displayed by the Be9 strain under U presence. We have used a multidisciplinary approach combining colorimetric methods for measuring residual uranium and inorganic-phosphate release, as well as flow cytometry to assess bacterial activity and viability. A characterization of Be9 membrane surface was determined by potentiometric titrations and X-ray photoelectron spectroscopy (XPS) analysis. Moreover, we also used microscopic and spectroscopic techniques [Scanning Transmission Electron Microscopy-High Angle Annular Dark-Field (STEM-HAADF) and X-ray Diffraction (XRD)] to localize at cellular level the U precipitates obtained and to analyze their composition.

The main objective of this study is to investigate the effect of phosphates in the biomineralization of U (VI) by Microbacterium sp. Be9, a uranium mill tailings porewater isolate. To the best of our knowledge, this the first study determining the effect of organic and inorganic phosphates in the U phosphate biomineralization process. Previously, environmental factors such as pH (Beazley et al., 2007; Chandwadkar et al., 2018; Zhang et al., 2018), temperature (Li et al., 2013), the presence of carbonates (Wei et al., 2019), NH₄⁺ incorporation (Yong and Macaskie, 1995) and organic matter (Medina et al., 2017; Boiteau et al., 2018) have been described to affect the U phosphate biomineralization. Our experiments have shown that the U biomineralization ability by the Be9 strain is drastically reduced by both organic and inorganic phosphates. However, in an exogenous phosphate-free system, the cells removed up to 88% U within 24 h (Figure 3) through intracellular U phosphate biomineralization. In addition, as a result of the kinetic studies, we can confirm that in the presence of previously abiotically precipitated U phosphates, the Be9 cells were able to solubilize this radionuclide (Figure 7) through a metabolism-dependent process.

The first step in this microbial U biomineralization process by the Be9 strain appears to be surface biosorption. Bacterial surfaces act as nucleation sites for the formation of biogenic U(VI)-phosphates through combined surface sorption of this radionuclide and phosphatase-mediated U biomineralization (Pan et al., 2015; Theodorakopoulos et al., 2015). Membranes play a major role in sorption of metals where electrostatic interactions between positively charged metal species and negatively charged functional groups (e.g., phosphates and carboxyles) occur. Gram-positive bacteria (such as the Be9 strain) exhibit greater biosorption capacity than Gram-negative bacteria due to their cell envelope components (Hufton et al., 2021). Liu et al. (2015) demonstrated the role of carboxyl, phosphoryl, and amino functional groups of Synechococcus sp. PCC 7002 cells as metal surface ligands by means of potentiometric titrations. Therefore, in our study, the chemical properties of the cell surface of Be9 strain were characterized using potentiometric titration and XPS analysis. The pH zero proton charge (pH_zpc) around 6.61 ± 0.07 (Supplementary Table 2) indicated that Be9 cells developed a negative net charge at the pH value we studied. These data show that electrostatic attraction of negatively ionizing groups (e.g., carboxyl, phosphoryl and hydroxyl groups) with positively charged U species [e.g., (UO₂)₃(OH)⁵⁺ and (UO₂)₄(OH)⁷⁺] is favorable. Thermodynamic calculation showed that under phosphate free system and in presence of G2P, these favorable compounds as (UO₂)₃(OH)⁵⁺ and (UO₂)₄(OH)⁷⁺ represented 78 and 20% species of U (Supplementary Table 3). In the case of phosphate-free system, the negatively ionizing groups of the cell wall could sorb 70% of positively charged U species in the first 10 min as was shown by the U removal kinetics studies (Figure 3). XPS analysis exhibited a high proportion of polysaccharides in the Be9 cell surface. Polysaccharides have been described as playing an important role in enhancing the tolerance of microorganisms under metal stress (Sun et al., 2020). In the cell surfaces of Pseudomonas putida similar percentages of peptides were obtained and were involved in the interaction with nZVI/Pd particles (Lv et al., 2017).

Under the absence of phosphates (MC1 system), a metabolism dependent precipitation of U intracellular phosphates seemed to take place after the surface sorption. The U removal capacity increased slowly throughout the time, without any significant differences between 24 and 48 h. These results show that U removal by this strain is a biphasic process with a fast first phase mediated by a passive process (e.g., sorption at cell surfaces) and a slow second phase driven by metabolically active processes (e.g., intracellular accumulation, biomineralization, etc.). However, in comparison to other bacterial strains, Be9 showed low concentrations of phosphate or carboxyl groups at the cell surface (Supplementary Table 2) which could probably explain why U is mainly accumulated intracellularly as was demonstrated by electron microscope analysis. Previous studies showed a strong correlation between the concentration of phosphates and carboxyl groups of the cell wall and cellular localization of the metal of some bacterial strains such S. bentonitica BII-R7 (Sánchez-Castro et al., 2020). The latter strain exhibited 12 and 10 orders of magnitude of phosphate and carboxyl groups (Ruiz-Fresneda et al., 2020) higher than the Be9 strain, and high capability to bind and biomineralize U at the cell surface (Sánchez-Castro et al., 2020). However, in spite of the low amount of phosphate and carboxyl groups, seems to be sufficient to act as a first passive sorption step followed by intracellular precipitation of uranium.

Biphasic U removal process has also been reported by other authors in different bacteria such as Acidovorax facilis (Gerber et al., 2016), Stenotrophomonas bentonitica BII-R7 (Pinel-Cabello et al., 2021) and Microbacterium sp. A9 (Theodorakopoulos et al., 2015). Flow cytometry studies revealed that 100% of the cells were viable and metabolically active during the first 24 h, supporting the involvement of a metabolic-dependent process in the interaction with U besides biosorption. A decrease in cell viability at 48 h seems to be clearly related to the high U-bioprecipitation inside the cells. In the case of uranium, numerous studies have reported the immobilization of this radionuclide by intracellular polyphosphate bodies through different analyses as TEM and EDX (Merroun et al., 2003, 2006; Suzuki and Banfield, 2004; Li et al., 2016). The low concentrations of phosphate and carboxyl groups at the Be9 cell surface indicated by potentiometric titration studies (Supplementary Table 2) could explain the low U amount bound to the surface. However, in spite of the low amount of phosphate and carboxyl groups, seems to be sufficient to act as a first passive sorption. In our study, the presence of U in the cells is speculated to induce the activity of enzymes such as exopolyphosphatase to degrade polyphosphate, thus releasing orthophosphates for the biomineralization of U. Theodorakopoulos et al. (2015) proposed an induction of enzymatic pathways from a Microbacterium strain which causes bound breakage of polyphosphate granules in the cytoplasm for the biomineralization of U intracellularly. However, the intracellular biomineralization of U associated with degradation of polyphosphates require the U uptake by the cells. For metal uptake in microorganisms, two main mechanisms were described: first, an active process through metal transporters for essential metals, and then, a passive process by an increase of cell membrane permeability due to damage for toxic metals (Suzuki and Banfield, 1999). In the case of Microbacterium, it has been demonstrated that iron transportation proteins are involved in the uranium uptake by Microbacterium oleivorans A9 (Gallois et al., 2018). A recent study about different Microbacterium strains revealed a transmembrane protein with binding affinity for UO₂²⁺ and Fe³⁺ specific to U-tolerant strains (Gallois et al., 2022). Regarding to the Be9 strain, in a previous study, its genome was annotated and ABC-type Fe³⁺ siderophore transport proteins, which may be involved in U uptake, were identified (Martínez-Rodríguez et al., 2020). Transcriptomics and proteomics studies are needed to further address the nature of transporters used for U intracellular accumulation in the strain Be9.

Scanning transmission electron microscopy (STEM) micrographs of Be9 cells showed a gradual formation of needle-like structures composed of U and P in the cytoplasm over time. The U phosphate nature of similar needle-like structures was confirmed in the literature by the use of different spectroscopic techniques such as EXAFS, TRLFS, etc. (Lopez-Fernandez et al., 2018), also in Microbacterium genus (Theodorakopoulos et al., 2015). Microbes are able to precipitate U phosphate phases with different structures (e.g., autunite, chernikovite, etc.) through phosphatase activity which degrade organic phosphates (e.g., glycerol phosphates) and release orthophosphates (Merroun et al., 2011; Skouri-Panet et al., 2018; Sánchez-Castro et al., 2020). In our study, although phosphatase activity was not detected, U phosphates were evidenced in the cells treated with U. Thus, the U phosphate biomineralization induced by this strain is probably not associated with phosphatase activity. Similarly, other U-tolerant Microbacterium strains showed a very low abundance of acid and alkaline phosphatase activity in their proteome (Gallois et al., 2022). The needle-like structures identified in STEM micrographs could be raised from the interaction of U species with existent orthophosphates in intracellular polyphosphates since no exogenous phosphate substrate was added. Previous studies have reported some bacterial strains capable of precipitating U into uramphite without additional phosphorous sources (Pan et al., 2015; Zhang et al., 2018). Polyphosphates (polyPs) granules are ubiquitous polymers occurring in many microorganisms (Seufferheld et al., 2008). They contain a high number of orthophosphate residues that are linked to phosphoanhydride bonds to form linear polymers (Mandala et al., 2020). However, a recent study described the existence of small metaphosphates, cyclic oligomers of [PO₃]⁽⁻⁾ made up of 3 to 8 phosphate groups (Mandala et al., 2020) as a candidate for precipitation of positively charged U species. Polyphosphate bodies possess various biological functions including phosphate removal, metal chelation, stress response, etc. (Achbergerová and Nahálka, 2011). Previous studies have demonstrated the role of bacterial polyphosphates in the accumulation of heavy metals and radionuclides (Brim et al., 2000; Acharya and Apte, 2013; Villagrasa et al., 2020). Be9 whole genome sequence analysis indicated the presence of genes codifying for enzymes involved in the synthesis (e.g., polyphosphate kinase) and degradation (e.g., exopolyphosphatase) of polyphosphates (Martínez-Rodríguez et al., 2020). In addition, Microbacterium spp. are well documented as potential polyphosphate accumulating microorganisms (Theodorakopoulos et al., 2015; Suzina et al., 2022). Therefore, the results obtained are in agreement with the expected hypothesis, where Be9 cells are able to release orthophosphates for the U biomineralization through polyphosphatase activity on intracellular polyphosphate granules. No phosphatase activity was involved in this metal/bacteria interaction process.

Surprisingly, in MC2 system (G2P presence), Be9 cells exhibited low U removal capacity contrary to expected, as compared to free phosphate system. In addition, no phosphatase activity was detected. Microbial phosphatase enzymes (acid and alkaline) catalyze the hydrolysis of G2P into glycerol and orthophosphates which lead to the precipitation of U phosphates with different structures such as efficient strategy for U bioremediation (Beazley et al., 2011; Merroun et al., 2011; Liang et al., 2016; Sánchez-Castro et al., 2020, 2021). However, although no phosphatase activity was measured, a certain concentration of orthophosphates was detected under these conditions. Thus, since G2P is not degraded extracellularly, it is probably transported intracellularly to be used as an organic carbon source for the metabolism of carbohydrates and lipids. This hypothesis is supported by the fact that the whole genome sequence of Be9 exhibited genes codifying for proteins for the uptake and utilization of G3P and glycerol (glycerol-3-phosphate ABC transporter, ATP-binding protein UgpC, glycerol-3-phosphate dehydrogenase, glycerophosphoryl diester phosphodiesterase). Gallois et al. (2018) reported the upregulation of proteins involving transport of glycerol-3-phosphate in the cells of Microbacterium oleivorans A9 treated with uranium. In contrast with the previous conditions (phosphate-free system), cell viability under the presence of U and G2P increased from 24 to 48 h of incubation, which supports the assumption that G2P is used by Be9 cells as a carbon or phosphorus source. Tu et al. (2019) described that strong organic-ligands such as oxalate and citrate competed with biotic PO₄³⁻ for uranyl, inhibiting the U-phosphate biomineralization by the cells of Bacillus sp. dw-2. Speciation of U in MC2 system (Supplementary Table 3) was dominated by hydroxo-uranyl complexes as (UO₂)₃(OH)⁵⁺ (78.51%) and (UO₂)₄(OH)⁷⁺ (19.22%) and no G2P-U complexes were expected. However, the mechanisms responsible for the decrease of U removal and biomineralization rate remains unclear.

Finally, we used the culture medium LPM in different treatments (MC3 system) to induce abiotic precipitation of U phosphates. As expected, the chemical speciation of U (at 0.5 and 1 mM) in this medium and in its diluted form (0.2 mg/l peptone concentration) resulted in a high positive saturation index (Supplementary Table 3), predicting the precipitation of various uranyl compounds including U phosphates. These results were confirmed by abiotic precipitation experiments where U precipitation reached high values after 48 h (Figure 6). Numerous amorphous structures such as U-phosphates and U-carbonates were identified in HR-TEM micrographs by EDX analysis (Supplementary Figure 7). However, in presence of Be9 cells, 94.5% of U was re-solubilized from abiotically precipitated U phosphates. Under this U concentration, the Be9 cells remained viable until 48 h, reacting with the precipitated U species, mainly U inorganic phosphates. Thus, contrary to the expected hypothesis, the obtained results demonstrated the Microbacterium sp. strain Be9 capacity of U phosphate solubilization, due to its belonging to PSB group. The cells seemed to display active mechanisms to solubilize this radionuclide. Microorganisms known as PSBs are able to solubilize both organic and inorganic phosphorus from insoluble compounds. Recently, numerous U-tolerant PSB strains have been isolated from mining sites of this radionuclide (Sowmya et al., 2014, 2020; Lv et al., 2022). Mechanisms of phosphate solubilization by PSB strains involve hydroxyl and carboxyl groups present in the cells or in released organic acids (Chen et al., 2006). In addition, a number of genes encoding enzymes responsible for P solubilization (e.g., encoding quinoprotein glucose dehydrogenase (gcd), phnP C-P lyase subunit (phnP), phoA alkaline phosphatase (phoA), phoD alkaline phosphatase (phoD), and phoN acid phosphatase (phoN)) have been reported (Rawat et al., 2021). In our study, we detected the U-mobilization activity by the Be9 strain reacting with inorganic phosphates. This fact indicates that this behavior may involve the solubilization of U from precipitated U-phosphates. In this way, bonded heavy metals (such as U) could be released in a soluble and toxic form to the environment. However, some bioremediation strategies require interaction with an available form of the metal to facilitate its removal. To this aim, the Be9 solubilization behavior could be applied in combined U remediation technologies.

In this study we have demonstrated that the Microbacterium sp. strain Be9 possesses a dual behavior toward U and is capable of precipitating and solubilizing this radionuclide under different conditions. In addition, the presence or absence of different phosphate sources had an influence on the U biomineralization ability of the Be9 strain. Therefore, this strain could be used as U bioremediation agent in a free exogenous phosphate system. And due to its potential in the solubilization of phosphates from organic and inorganic P sources, this strain could contribute indirectly to providing this inorganic anion for the precipitation of this radionuclide. Thus, our results provide new insights into the impact of microbes on the biogeochemical cycle of U in the presence of different forms of phosphates. This information leads to the understanding of the conditions regarding biogenic U (VI) phosphate mineral formation and the physico-chemical parameters which hinder biomineralization and remediation strategies in oxidizing conditions. The correct selection of the supplemented organic phosphate donor could be crucial depending on the microorganism used. Therefore, the importance of gaining knowledge about microbial diversity and its role with the target metal is crucial to achieve the correct application of selected bioremediation technologies.

The original contributions presented in the study are included in the article/Supplementary material, further inquiries can be directed to the corresponding author.

PM-R, IS-C, and MM designed the experimental setup and methodology, and major contributors in writing the manuscript. PM-R and IS-C performed all laboratory works, analyzed the data, and created the graphs and figures. MA carried out microscopy analyses presented here. JO designed and supervised the cell surface analysis reported here. JO and MD revised the manuscript. MD and MM conceived the concept, led this project and obtained funds for carrying out these studies. All authors contributed to the article and approved the submitted version.

This work was supported by ORANO Mining (France; collaborative research contract no 3022 OTRI-UGR). It results from a Joint Research Project between Orano Mining R&D Department and the Department of Microbiology of the University of Granada.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.