Microbial factors behind struvite enzymes and extracellular polymeric substances (EPS) as key players
Authors: Pelina Toprak, Anant Aishwarya Dubey, Abhijit Mukherjee, Allan Pring, Carlos Rodriguez-Navarro, Navdeep K. Dhami
Abstract
Abstract
Struvite is a phosphate-based mineral commonly formed in ammonia-rich settings such as soils, aquatic environments, wastewater systems, and the urinary tract. Its distinct morphology and mineral properties have drawn interest across disciplines, including microbiology, geology, medicine, and materials science, especially for its ammonia capturing ability. Microbial metabolic activities have been recently recorded to play a critical role in its formation. However, the influence of microbial characteristics, such as enzymatic activity, biofilm, and extracellular polymeric substances (EPS), on struvite morphology and precipitation mechanisms remains underexplored. This study investigates microbial struvite precipitation mediated by three distinct microbes, Sporosarcina pasteurii,* Bacillus subtilis*, and Pseudomonas fluorescens to investigate the role of different enzymes and EPS on struvite mineralisation. The bacterial cultures selected for the study have varying ureolytic and alkaline phosphatase activity, as well as EPS and biofilm production potential. Struvite precipitation kinetics under different microbial conditions were studied along with its morphology and biofilm-mineral interactions employing microstructural and mineralogical analyses (via field emission scanning electron microscopy, confocal laser scanning microscopy, X-ray diffraction, and Fourier transform infrared spectroscopy). This study reports that urease and phosphatase activities (between 7 and 7.5 U/mL) along with moderate EPS production (~6.5 g/L), lead to the generation of the highest quantities of struvite crystals with moderate size (103 ± 5 µm), contrary to the larger crystals (180 ± 5 µm) produced with a higher urease activity (8.4 U/mL) and lower EPS (2 g/L). The experimental findings suggest a complex interplay between enzymatic activity and EPS in shaping crystal structure and its features, which have a significant impact on mineral characteristics. Additionally, moderate biofilm formation (OD570 ~1.2) resulted in maximal struvite precipitation, indicating that structured biofilms support efficient mineral nucleation and accumulation. Overall, this study has unravelled key insights into microbial struvite precipitation mechanisms and can pave the way for tailored engineering and biotechnological applications utilising this phosphate mineral.
Key points
Biogenic struvite precipitation is structurally different from abiogenic struviteType of microbial species, surface features, enzymes and EPS impact crystallinity of struviteComposition of EPS effect phase and pattern of biogenic struvite crystal
Introduction
Microbes play a vital role in struvite (MgNH4PO4·6H2O) precipitation in diverse environmental conditions and industrial settings (Jaffer et al. 2002; Sinha et al. 2014). Struvite tends to form in environments rich in organic matter, where conditions support the decomposition of nitrogen-containing compounds (Žnidarčič et al. 2024). A few notable examples of such environments include guano deposits (accumulations of bird droppings), basaltic caves, and marshlands (Guan et al. 2023; Sinha et al. 2014). Another example of struvite formation in a contrastingly different environmental setting is sewage and wastewater treatment pipelines in form of the undesired persistent scale (Prywer & Torzewska 2010; Stratful et al. 2001), which is challenging to remove. Struvite is also commonly observed as infectious urinary stones in both humans and animals (Çiftçioglu et al. 1999).
Often, the formation of struvite in natural environments involves biomineralisation processes, where microorganisms play a crucial role in its precipitation. Specific microbes produce urease and phosphatase enzymes that break down organic nitrogen and phosphorous compounds, leading to the release of ammonium ions (NH4^+^) and phosphate ions (PO4^3−^), which results in an elevation of solution pH and creates favourable conditions for the formation of struvite, with magnesium commonly sourced from soil, water, or organic matter (Sadowski et al. 2014; Sinha et al. 2014). This highlights how microbial actions and the environment’s characteristics work together to form struvite. Chemical reactions involved in struvite formation include urea hydrolysis, phosphate release, and struvite precipitation as shown by Eqs. 1, 2 and 3, respectively.1\documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$
{\mathrm{CO}\left({\mathrm{NH}}{2}\right)}{2}+{2\mathrm{H}}{2}\mathrm{O}+{\mathrm{H}}^{+}\stackrel{\text{microbial urease}}{\to } 2{\mathrm{NH}}{4}^{+}+\mathrm{H}{{\mathrm{CO}}_{3}}^{-}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$
\text{Organic Phosphate}\stackrel{\text{microbial phosphatase}}{\to }\text{ P}{\mathrm{O}}_{4}^{3-}+\text{Organic residue}
$$\end{document}Organic Phosphate→microbial phosphatasePO43-+Organic residue3\documentclass[12pt]{minimal}
\usepackage{amsmath}
\usepackage{wasysym}
\usepackage{amsfonts}
\usepackage{amssymb}
\usepackage{amsbsy}
\usepackage{mathrsfs}
\usepackage{upgreek}
\setlength{\oddsidemargin}{-69pt}
\begin{document}$$
{\mathrm{Mg}}^{2+}+{\mathrm{NH}}_{4}^{+}+\text{ P}{\mathrm{O}}_{4}^{3-}+6{\mathrm{H}}_{2}\mathrm{O}\to \text{ Mg}{\mathrm{NH}}_{4}{\mathrm{PO}}_{4}.6{\mathrm{H}}_{2}\mathrm{O}
$$\end{document}Mg2++NH4++PO43-+6H2O→MgNH4PO4.6H2O
Microbial involvement in struvite precipitation not only influences the process itself but also affects the morphology of the resulting crystals (Zhao et al. 2021). Biogenic struvite often displays unique shapes, such as coffin-, X-like, dendritic, and prismatic forms, which differ from rod-, needle-like, and long tabular shapes formed under purely inorganic conditions (Zhao et al. 2021). These distinct morphologies suggest that bacteria play varying roles in both the mineralisation and morphology selection of the struvite crystals. The microbial cell surface, enzymatic activity, and EPS play crucial roles in the nucleation, growth, and morphology of struvite crystals (Leng & Soares 2021).
Moreover, in the context of engineering applications, the ability of microbially induced struvite precipitation (MISP) to utilise ammonia offers a significant advantage over the currently popular biomineralisation technique, microbially induced calcite precipitation (MICP), as the primary challenge in scaling up MICP is the significant environmental threat posed by ammonia/ammonium, the major byproduct of the urea hydrolysis process (Shi et al. 2022; van Paassen et al. 2010; Yu et al. 2022). MISP has the potential for application across a broad range of industries, including civil engineering, biomedicine, agriculture, waste stabilisation, and soil improvement due to its potential to use ammonium ions from biodegradation (Dong et al. 2020; He et al. 2023a, b; Yu et al. 2022). Interestingly, a variety of microorganisms that are capable of precipitating calcium carbonate, commonly used in MICP, have also been reported to induce struvite precipitation under suitable conditions (Simões et al., 2018). This overlap suggests that some carbonatogenic microorganisms possess metabolic versatility, enabling them to contribute to both MICP and MISP, depending on environmental factors and available nutrients. Therefore, a deeper knowledge of the mechanism(s) controlling struvite precipitation is critical for understanding the natural formation processes and advancing their engineering applications.
Previous studies on CaCO~3~ have shown how microbial interactions influence crystal structure and growth (Dhami et al. 2016; Li et al. 2015), suggesting a similar scenario for struvite. Biomineralisation predominantly occurs within microbial mats or biofilms in nature, which are complex, surface-attached clusters of microbial cells (Li et al. 2015). The environmental diversity within biofilms creates sharp gradients in oxygen, nutrients, and pH, resulting in varied microenvironments that affect microbial metabolism (Decho 2010; Decho and Gutierrez 2017). These conditions lead to differences in gene expression, respiration, and byproduct formation, such as organic acids that alter local pH. Such fluctuations create zones of supersaturation, promoting biomineralisation (Decho and Gutierrez 2017; Strathmann et al. 2002).
Biofilms develop when cells attach to a surface and secrete extracellular polymeric substances (EPS) (Lazarova and Manem, 1995, Ceyhan and Ozdemir 2008). EPS are high-molecular-weight polymers primarily composed of repeating sugar monomers, often accompanied by proteins, nucleic acids (extracellular DNA and RNA), lipids and various heteropolymers, that serve as the structural scaffold of biofilms, facilitating cell adhesion and cohesion (Chen et al. 2013; Di Martino 2018; Møller et al. 1997). The EPS scaffold concentrates ions and enzymes, facilitating the nucleation and growth of mineral crystals (Bains et al. 2015; Ercole et al. 2007; Flemming et al. 2007). The composition of EPS is essential for defining the mechanical and functional characteristics of biofilms (Chen et al. 2013; Di Martino 2018; Møller et al. 1997; Robles-Fernández et al. 2022), which must be investigated in-depth to understand the underlying mechanism of distinct morphology selection of microbial struvite in MISP processes.
This study aims to advance the existing knowledge in the domain of microbial struvite formation, with a focus on unpinning the impact of microbial enzymes (alkaline phosphatase, urease) and EPS on quality, quantity and morpho-mineralogical features of struvite precipitation. To achieve this, three bacterial strains with distinct enzymatic activities and EPS production profiles were selected, allowing a systematic comparison of microbial pathways involved in struvite biomineralisation. In summary, the goal is to address the The impact of biogenic activities on struvite crystal morphology compared to abiogenic precipitation.The influence of microbial species, their urease- and phosphatase-activity, and EPS formation capacity on reaction kinetics, precipitation rate, morphology, and crystallinity.The role of EPS in stabilising and guiding the nucleation and growth of distinct mineral phases in struvite precipitation.
## Materials and Methods
### Selection, cultivation and growth of bacterial strains
To understand the influence of the different enzymatic activities and EPS on struvite mineralisation, three distinct bacterial strains, namely *Sporosarcina pasteurii* (ATCC 11859), *Bacillus subtilis* (ATCC 6633), and *Pseudomonas fluorescens* (ATCC 13525), were cultivated in their respective growth media to initiate the experimental setup. *S. pasteurii*, a Gram-positive, rod-shaped soil bacterium measuring approximately 0.5-1.2 µm in width and 1.3-4.0 µm in length, is well known for its ability to precipitate CaCO₃ through high ureolytic activity with limited EPS production (Wormser and Stratton 2008; Dhami et al. 2017). They demonstrate biofilm formation capabilities in calcium-rich environments, adhering to and colonising concrete surfaces for applications such as biocementation (Achal & Pan 2011). EPS production in *S. pasteurii* includes polysaccharides and proteins that contribute to biofilm formation and mechanical stability, essential in biotechnological and environmental engineering contexts (Dejong et al. 2013). The bacterium was cultured in ammonium–yeast extract (YAT) medium at 34 °C and 180 rpm, following the method of Stocks-Fischer et al. (1999). *B. subtilis* is a Gram-positive, rod-shaped, biofilm and EPS producer bacterium measuring approximately 0.5-1.0 µm in width and 2.0-6.0 µm in length (Doyle & Koch 1987). It exhibits both urease and alkaline phosphatase activity at moderate levels, providing a balanced enzymatic profile that enables precipitation through both urea hydrolysis and phosphate solubilisation (Ren et al. 2004). They are known for structured biofilm formation, constructing tower-like biofilms reaching heights of up to 100 µm on agar surfaces, and synthesising poly-γ-glutamate (γ-PGA) as a major EPS component, with production levels varying between 10 and 100 µg/mL in biofilm cultures (Luo et al. 2016; Ren et al. 2004). For the present study, the bacterium was cultivated in nutrient broth medium under the same experimental conditions as *S. pasteurii*. *P. fluorescens* is a Gram-negative, rod-shaped bacterium measuring approximately 0.5 µm in width and 2.0-2.5 µm in length (Vesper 1987). They form biofilms up to 50 µm thick on surfaces such as glass and polystyrene within 24 h. (Angarano et al. 2020). Also, they are able to produce substantial amounts of alginate as EPS, with reported levels ranging from 50 to 200 µg/mL in biofilm cultures (Conti et al. 1994; Oliveira et al. 1994; Vidhyalakshmi et al. 2018). It represents a distinct strategy, characterised by low urease activity but high alkaline phosphatase activity, and secretes large amounts of EPS (Strathmann et al. 2002). It was cultured in Luria broth medium under the same experimental conditions as the other two bacteria. It is important to note that the strain-specific media (YAT, NB, and LB) were used exclusively during the cultivation phase to generate sufficient biomass. In the initial phase, assessing the growth and viability of the bacteria involved in the experiment is crucial. The growth curves of the microbes were recorded on a multimode microplate reader (Thermo Scientific™ Varioskan™ LUX) by measuring the optical density at 600 nm (OD~600~) of the cultivated media to monitor the progression and proliferation of microbes over time. The correlation between optical density (OD₆₀₀) and viable cell concentration was established by performing colony-forming unit (CFU) counts.
### Biogenic and abiogenic precipitation media
After reaching the exponential (log) growth phase at an optical density (OD₆₀₀) of 0.5 (corresponding to approximately 10⁷ CFU/mL, i.e., log-phase cells), the cells were harvested and washed with phosphate-buffered saline (PBS) to remove residual nutrients and ions from the growth medium. It is to be noted that the aim of the study is to determine how the three different microbial cells of different characteristics under the same environmental conditions affect struvite precipitation. To evaluate this, the standardised cell suspensions were transferred to the struvite precipitation medium. Although each bacterium was initially cultured in its growth medium for optimal growth, identical experimental conditions were maintained during the precipitation experiment. The precipitation medium was adapted from Rivadeneyra et al. (1992), with modifications inspired by the protocol of Zorzetto et al. (2022) to incorporate an organic phosphorus source. This medium consists of yeast extract (ATCC 1376) supplemented with 2 g/L magnesium sulphate heptahydrate (MgSO~4~·7H~2~O), 2 g/L β-glycerophosphate disodium salt (Sigma-Aldrich, Cat# 35,675) as the sole phosphorus source, and 0.5 M urea to supply ammonium via enzymatic hydrolysis.
Two control mediums were established to differentiate the microbial contributions from the plain chemical processes liable for struvite formation. The first control (medium A1) had the same composition as the precipitation medium described above but without bacterial inoculation and was used to verify whether struvite could form without microbial or enzymatic activity involving any impromptu phosphate or urea conversion. The second control (medium A2) represented an abiogenic inorganic system, adapted from Rivadeneyra et al. (1992), and consisted of yeast extract medium (ATCC 1376) containing equal concentrations of MgSO~4~ and KH~2~PO~4~ (inorganic phosphate), along with1 M NH~4~Cl, to evaluate the potential for chemical struvite precipitation. This abiotic setup served as a chemical control to evaluate the potential for non-biological struvite precipitation. The media sets were prepared in 250 mL Erlenmeyer flasks containing 100 mL of the specified media. All flasks were prepared in triplicate and incubated at 37 °C in an orbital shaker at 120 rpm for 5 days. Experiments were conducted in triplicate, and the data are presented as the mean value with standard deviations (mean ± SD).
### Quantification of enzymatic activities
The urease enzyme activity was measured in 96-well tissue culture plates by quantifying the release of ammonia from urea using the phenol-hypochlorite assay at various time points (Dhami et al. 2016). Alkaline phosphatase (ALP) activity was measured in 96-well tissue culture plates using p-nitrophenyl phosphate (p-NPP) as a phosphatase substrate. The substrate undergoes dephosphorylation by ALP, leading to a yellow colour change (Jones 1972).
### Spectrophotometric measurement of urea and orthophosphate concentrations in solution
Two millilitres of the experimental medium was sampled at different time intervals during the course of the precipitation experiments, sterile filtered through a 0.2 μm membrane, and stored at 4 °C until analysis. Urea concentrations were determined using the dimethyl-amino-benzaldehyde (DMAB) spectrophotometric method adapted from Murugan et al. (2021). In this method, 50 μL of each sample was mixed with 50 μL of a 12% trichloroacetic acid solution, followed by the addition of 100 μL of DMAB reagent containing 1.6% DMAB in concentrated hydrochloric acid with 10% (v/v) ethanol. The mixture was incubated at room *temperature* for 5 min, and the absorbance of the resulting yellow colour was measured at 425 nm using a plate reader. Urea concentration was quantified using a standard curve ranging from 0 to 125 mM.
Dissolved inorganic phosphate (*Pi*) concentrations were determined using the ascorbic acid spectrophotometry procedure adapted from Chen et al. (1956). In this method, orthophosphate reacts in an acidic medium with ammonium molybdate and potassium antimony tartrate to form phosphomolybdic acid, which is reduced by ascorbic acid to form a blue-coloured molybdenum complex. Absorbance was measured at 880 nm to determine *Pi* concentration.
The pH of each sample was measured prior to any filtration step to assess the chemical environment influencing precipitation processes.
### Crystal harvesting and quantification
Following a 5-day incubation period, the flask contents were filtered using 0.45 μm Whatman filter paper and rinsed with phosphate-buffered saline, and the filtrates were dried at 37 °C for 12 h. The dried filtrates were then weighed to determine the amount of precipitated crystals, which were subsequently prepared for morphological and chemical characterisation (Dhami et al. 2017).
### EPS isolation and composition analysis
The three selected bacteria were inoculated into 90 mL of their respective medium within 250 mL Erlenmeyer flasks, starting from 10 mL of an overnight culture. The flasks were incubated statically at 37 °C for 7 days to allow biofilm formation, following the protocol adapted from Ercole et al. (2007).
The biofilm quantities were assessed using the well-established crystal violet staining assay method (Borucki et al. 2003). The bacterial species were used for inoculation in their respective original growth cultures and were subsequently incubated for 24 h at 37 °C. Overnight culture of the bacteria was diluted to an OD~600~ of 0.5, and 1 µL was added to 99 µL culture medium inoculated into the microtiter plate. The edges of the plate were sealed with parafilm, and the plates were incubated at 37 °C. After 40 h, the liquid from each well was removed, and unattached cells were eliminated by rinsing three times with 150 μL of sterile water. The plates were then dried in an inverted position for 30 min. Biofilms were stained by adding 50 μL of a 0.1% crystal violet solution (in sterile water) to each well and incubating for 45 min at room *temperature.* Unbound dye was removed by rinsing three times with 150 μL of sterile water. The crystal violet was solubilised by adding 200 μL of 95% ethanol and incubated at 4 °C for 30 min. Subsequently, the contents of each well (100 μL) were transferred to a new sterile polystyrene microtiter plate, and the optical density at 595 nm (OD~595~) of each well was measured using a microplate reader.
Quantification of EPS was performed following the method outlined by Bains et al. (2015), with slight modifications. The bacterial cultures were inoculated into 100 mL of culture medium and incubated at 37 °C for 48 h. After incubation, culture broths containing EPS were centrifuged at 13,200 g for 25 min at 4 °C, ensuring that intact cells and cellular debris were removed. The cell-free supernatants were stored at -20 °C. EPS in 50 mL supernatant was precipitated by the addition of three volumes of chilled absolute ethanol. The mixture was held at 4 °C overnight and then subjected to centrifugation at 13,200 g for 15 min at 4 °C. The resulting pellet was dried at room *temperature* for 6 h, followed by drying at 100 °C till constant weight and the dry weight was measured.
Lyophilised EPS were analysed for total sugar content using the phenol-sulfuric acid method, as outlined by Masuko et al. (2005), and protein content was quantified by the Lowry method (Lowry et al. 1951), employing bovine serum albumin as the standard.
The functional groups present in the extracted EPS from the three different microbes were examined using a Fourier transform infrared (FTIR) spectrometer (Bruker IFS66). The analysis was performed in absorbance mode, with 20 consecutive scans in the spectral range 400 to 4000 cm^−1^, at a resolution of 4.0 cm^−1^.
### Tracking bacteria-biofilm-biomineral interactions through confocal laser scanning microscopy (CLSM)
To investigate the interactions between bacterial cells, biofilm and biominerals (struvite), the µ-Slide 4 well chambers were used. The glass slides were cleaned with ethanol and deionised water. They were then coated with a 0.01% solution of poly-L-lysine and allowed to air-dry in a sterile environment for 2 h (Colville et al. 2010). Following this, 20 µL of washed bacterial cells were introduced into each well, along with 100 µL of struvite precipitation media. The chambers were incubated at 37 °C without agitation overnight to facilitate biofilm formation.
To visualise biofilms under a confocal microscope, a combined staining technique was employed using Concanavalin A (ConA) and SYTO 9 dyes, adapted from Strathmann et al. (2002). Biofilms grown on a glass slide were first stained with ConA, which binds specifically to certain carbohydrate residues on the biofilm matrix. For this, stock solutions of ConA (Thermo Fisher) were prepared at a concentration of 1 mg/mL in 10 mM phosphate buffer (pH 7.5) and stored in 100 µL aliquots at -20 °C. Before application, a thawed aliquot was diluted with 10 mM phosphate buffer (pH 7.5) to a final lectin concentration of 10 µg/mL. The biofilms were incubated with the staining solution in the dark at room *temperature* for 30 min. Following incubation, excess staining solution was removed by washing the biofilms with phosphate buffer.
To visualise cell distribution in the biofilms, the fluorescent DNA-binding stain SYTO 9 (component A of the LIVE/DEAD® BacLight™ Bacterial Viability Kit, Molecular Probes, Eugene, OR, USA) was employed in conjunction with TRITC-lectin staining. Following the lectin staining procedure, the biofilms were treated with 100 µL of a freshly prepared solution containing 2 µL of SYTO 9 per ml of deionised water. The samples were incubated for 30 min at room *temperature* in the dark and then immediately examined using CLSM. The samples were imaged using a Nikon A1+ point scanning confocal microscope with NIS elements software (Nikon Instruments).
### Microstructural and mineralogical investigation
The morphology of the precipitated minerals was investigated by means of a Field Emission Scanning Electron Microscopy (FESEM) using a MIRA, TESCAN3 equipment. Prior to imaging, samples were affixed onto carbon-aluminium tape and coated with a thin layer of carbon using a carbon evaporation coater. The elemental composition of the biomineral was investigated using integrated Energy Dispersive X-ray spectroscopy (EDS).
Phase analysis of the crystals was carried out using X-ray powder diffraction (XRD). Samples were dispersed in ethanol and deposited onto low-background sample holders. Data collection was performed using a Bruker D8 Advance diffractometer with Ni-filtered Cu Kα radiation (40 kV, 40 mA) over the range 7-120° 2θ, with a step size of 0.015°. Phase identification utilised the Crystallography Open Database (COD) included in Bruker EVA 5.2 software, while phase quantification was achieved by the Rietveld method in Topas Academic 7, referencing crystal structures from the COD database.
## Results
### Bacterial growth characteristics and enzymatic activities
Time-series bacterial experiments were conducted employing *S. pasteurii*,* B. subtilis*, and *P. fluorescens* strains. Optical density (OD~600~) and bacterial concentration (colony-forming units, CFU/mL) were monitored over time, and the data were collected until the bacterial cultures reached the stationary phase. The production of urease and alkaline phosphatase was assessed at various stages of bacterial growth as illustrated in Fig. 1a-c.Fig. 1Production of urease and alkaline phosphatase across growth phases in (**a**) *S. pasteurii*, (**b**) *B. subtilis*, and (**c**) *P. fluorescens, shown alongside corresponding bacterial growth curves.* Error bars show standard deviation
*S. pasteurii* exhibited the highest urease production throughout its growth phases, when compared to *P. fluorescens* and *B. subtilis*, peaking during the exponential phase and maintaining significant activity into the stationary phase, as shown in Fig. 1a. In contrast, *P. fluorescens* demonstrated the highest alkaline phosphatase activity, particularly during the exponential growth phase, when compared to the other two strains, as illustrated in Fig. 1c. A notable decline in alkaline phosphatase activity was observed as the bacterial cultures transitioned into the stationary phase. *B. subtilis* exhibited intermediate levels of both urease and alkaline phosphatase activities compared to *S. pasteurii* and *P. fluorescens*, as illustrated in Fig. 1b. Urease production peaked during the exponential phase, while alkaline phosphatase activity remained relatively stable throughout growth.
### Bacterial impact on urea hydrolysis, phosphate utilisation, and struvite formation
Following the inoculation of the three different bacterial strains into the flasks containing struvite media, the environmental conditions within the flasks were determined over a 5-day period with a 1-day interval to monitor fluctuations in medium pH, urea and phosphate content during the mineralisation process. The precipitation rate (in g/L-day), urea utilisation and phosphate utilisation rate of *P. fluorescens*, *B. subtilis*, and S*. pasteurii* are shown in Fig. 2a, c and d, respectively.Fig. 2(**a**) Precipitation mass (dry weight), (**b**) pH, (**c**) soluble phosphate concentration, and (**d**) urea concentration during struvite precipitation by *P. fluorescens*,* B. subtilis*, and *S. pasteurii*. Error bars show standard deviation
After the 5-day incubation period, visible white precipitates from mineralisation experiments were filtered, and the quantity of struvite crystals produced by each bacterial strain was quantified, with *B. subtilis* demonstrating the highest production of crystals (3.19 g/L), followed by *S. pasteurii* (2.23 g/L), and *P. fluorescens* (1.11 g/L) showing the lowest production (Fig. 2a). No precipitate formed in the first control (medium A1), suggesting that struvite could not form under these conditions without any biological or enzymatic activity to release phosphate or hydrolyse urea. Therefore, this control was excluded from further investigation. In contrast, the second control (medium A2) produced a small amount of precipitate (0.8 g/L), suggesting that chemical precipitation occurs in the presence of inorganic phosphate and ammonium.
Soluble phosphate concentration profiles (Fig. 2c) showed an initial increase followed by a decline in all strains. *P. fluorescens* exhibited the fastest phosphate release, reaching its maximum soluble phosphate concentration earlier than the other strains. *B. subtilis* displayed moderate alkaline phosphatase activity, resulting in a slower phosphate release and a delayed peak compared to *P. fluorescens*. *S. pasteurii* exhibited the slowest phosphate mineralisation, consistent with its lower phosphatase activity. Consequently, the differences in soluble phosphate dynamics among the three strains primarily reflect variations in enzymatic hydrolysis efficiency, which in turn influenced the timing and extent of struvite formation.
As expected, the urea concentration also decreased gradually across all experimental flasks as illustrated in Fig. 2d. However, *S. pasteurii* exhibits notably higher urea hydrolysis rates compared to the other strains, achieving approximately 50% consumption within 24 h. This increased hydrolysis of urea by *S. pasteurii* was accompanied by a significant rise in pH, from 7 to 9.5 (Fig. 2b), indicating substantial ammonia production because of urease activity. Similarly, *B. subtilis* flasks showed a decrease in urea concentration and a corresponding pH increase from 7 to 9.2 (Fig. 2b), reflecting its urease activity and its impact on the flask environment. In contrast,* P. fluorescens* exhibited a more modest pH change, from 7 to 8.5 (Fig. 2b), which correlates with its lower level of urea hydrolysis and subsequent ammonia production.
### Biofilm and EPS characterisation
The results of the biofilm and EPS analyses from the three bacterial strains are illustrated in Fig. 3a. *P. fluorescens* exhibited the highest biofilm formation, with an OD₅₇₀ value of approximately 3.1 and EPS production reaching ~10 g/L. This was followed by *B. subtilis*, which demonstrated a moderate level of biofilm production (OD₅₇₀ ≈ 1.3) and EPS yield of ~5.5 g/L. In contrast, *S. pasteurii* displayed the lowest values, forming a biofilm with an OD₅₇₀ of only 0.39 and producing ~1.3 g/L of EPS. These findings highlight the differing capacities of *P. fluorescens*, *B. subtilis*, and *S. pasteurii* for biofilm and EPS production under the experimental conditions used here.Fig. 3(**a**) Biofilm and EPS secretion capacity (in g/L) and (**b**) carbohydrate and protein content in the EPS produced by *P. fluorescens*,* B. subtilis*, and *S. pasteurii.* Error bars show standard deviation
To further understand the impact of EPS on struvite precipitation, the sugar and protein content of the EPS produced by the three strains was also assessed (Fig. 3b). Among the three species, *P. fluorescens* EPS exhibited the highest sugar content (55%), followed by *B. subtilis* (50%), while *S. pasteurii* produced the EPS with the lowest sugar proportion (40%). Regarding total protein, *P. fluorescens* EPS had the highest protein content (~26%), followed by *S. pasteurii* (~21%), while *B. subtilis* produced the EPS with the lowest protein content (14%). Across all species, sugars were the dominant EPS component, though the sugar-to-protein ratio varied significantly, with *B. subtilis* displaying the highest ratio. These findings indicate species-specific differences in EPS composition, which may influence their roles in biofilm formation and mineralisation processes. The comparative analysis aimed to assess whether differences in EPS content and composition influenced struvite precipitation efficiency. This approach helps elucidate the role of EPS in biofilm formation and mineral precipitation processes.
The functional groups of EPS from *S. pasteurii*,* B. subtilis*, and *P. fluorescens* were identified by FTIR analysis. This analysis can provide valuable insights into the distinct biochemical signatures of EPS that influence struvite crystallisation. Figure 4 shows the FTIR spectra of EPS from the three different species.Fig. 4FTIR spectra of the EPS extracted from the three bacterial strains
The FTIR spectra of the extracted EPS from *S. pasteurii* and *B. subtilis* are nearly identical. While microbial activity clearly impacts the quantity of EPS produced, along with total carbohydrate and protein content, the FTIR analysis indicates that both microbes generate EPS with similar functional groups. The shaded regions in Fig. 3b highlight the key absorbance bands corresponding to these functional groups.
The absorption bands in the 3300-3500 cm^−1^ range are associated with hydroxyl (O-H) and amino (N-H) groups vibrational stretching, with the O-H stretching likely resulting from adsorbed water. A shoulder around 3100 cm^−1^ was evident, especially in *P. fluorescens*, and is assigned to Amide A, associated with N-H stretching in proteins or peptides. Distinct small bands between 2890 and 2950 cm^−1^ correspond to the C-H stretching of saturated carbohydrates. In the region between 1500 and 1700 cm^−1^, two distinct bands were a prominent peak near 1650 cm^−1^, attributed to Amide I and a shoulder near 1516 cm^−1^, corresponding to Amide II (N-H bending and C-N stretching). The broadening or splitting in this region, particularly in *B. subtilis* and *S. pasteurii*, may also include contributions from carboxylate (COO^−^) vibrations, which are linked to proteins within the EPS matrix. The vibrations of the carboxylate (C = O) group appear at ~1430 cm^−1^. The absorbance band near 1250 cm^−1^ is due to the asymmetric stretching vibrations of PO~2~^−^groups. This functional group is typically found in phosphorylated biomolecules, such as nucleic acids (DNA, RNA) or phospholipids, indicating that phosphate groups in the EPS may be biologically derived rather than from inorganic phosphate alone. The most prominent absorbance band, with peaks between 1060-1070 cm^−1^, corresponds to the complex vibrations of carbohydrates with a skeletal ring structure.
### Mineralogical and Morphological Analysis
XRD analysis confirms that the precipitates were crystalline, corresponding to struvite as illustrated in Fig. 5a–c.Fig. 5XRD patterns for precipitation by (**a**) *S. pasteurii*, **b**
*B. subtilis*, and (**c**) *P. fluorescens*
The experimental diffraction patterns (black lines, Fig. 5) were compared with the standard reference pattern for struvite (MgNH~4~PO~34~·6H~2~O, JCPDS file # 15–0762, similar to # 77–2303, with space group *P*mn2~1~, red bars, Fig. 5). In all samples, several major diffraction peaks, particularly those around 14°, 21°, 27°, and 31° 2θ, matched well with the reference peaks of struvite, confirming its formation. However, notable differences were also observed, including variations in peak intensities, slight shifts in peak positions, and the presence of additional minor peaks that are not assigned to pure struvite. The differences in peak intensities, in particular the very high intensity of the peaks at 15.82, 21.41, and 31.92°2θ corresponding to the (002), (012), and (004) planes, respectively, can be due to preferred orientation due to the overdevelopment of the corresponding (*hkl*) faces. In contrast, the presence of additional small Bragg peaks (not matching those of struvite) indicates that the precipitates are not composed exclusively of struvite but may include minor amounts of alternative magnesium ammonium phosphate phases with different hydration states or structural variations.
The morphology of the precipitated struvite crystals was investigated by means of FESEM. *S. pasteurii* produces large, elongated prismatic crystals ranging from 100 to 200 μm in size, with the micrographical images revealing cohesive and structured euhedral crystals that exhibit significant porosity on their surfaces, due to the presence of pervasive cracks/gaps displaying a triangular star shape, many of them arranged in an ordered pattern, as shown in Fig. 6a-c. The mean crystal size, determined by ImageJ analysis, was 180 ± 12 μm, based on measurements of 100 individual crystals.Fig. 6FESEM images of struvite precipitated by (**a–c**) *S. pasteurii*, (**d–f**) *B. subtilis*, and (**g–i**) *P. fluorescens*
In contrast, *B. subtilis* forms moderately sized crystals, with visible porosity and irregular, rounded shapes, many of which show penetration and contact twining (Fig. 6d-f). The mean crystal size, determined by ImageJ analysis, was 103 ± 5 μm, based on measurements of 100 individual crystals. *P. fluorescens* produces the smallest and most fragmented crystals, ranging from 1-40 μm. Most of them showed contact twining. The mean crystal size, determined by ImageJ analysis, was 37 ± 4 μm, based on measurements of 100 individual crystals. These crystals are characterised by a high degree of porosity, with predominantly prismatic shapes and a small proportion of X-like crystals (Fig. 6g–i).
Detailed examination with FESEM revealed that struvite crystals display a patterned surface structure made up of small crystallites from 1 to 3 μm in size (Fig. 7). The crystallites display well-defined facets limited by gaps showing the above-mentioned triangular star-shaped features. Although the patterns are assumed to be biogenic, previous studies have shown that these distinct patterns are likely to be dehydration artifacts on the struvite crystals (Hövelmann et al. 2019).Fig. 7FESEM micrograph showing small crystalline structures in *P. fluorescens*-induced struvite precipitation*.* Note the presence of bacterial cells on the surface of the patterned struvite crystals
To further disclose the role of bacteria in struvite formation, abiotic crystallisation experiments in the absence of bacteria were conducted by using sterile culture media (A2) containing KH~2~PO~4~ (inorganic phosphate source). In the absence of bacteria, our experiments demonstrated a significant reduction in total struvite crystal formation over a 5-day period as compared to bacterial-inoculated conditions. XRD analysis confirmed abiotic struvite precipitation, and subsequent FESEM analysis revealed the formation of larger crystals as shown in Fig. 8. The mean crystal size, determined by ImageJ analysis, was 287 ± 15 μm, based on measurements of 50 individual crystals. Abiogenic struvite crystals exhibited smooth, compact, and well-defined structures with prismatic geometry and sharp edges. Their surfaces were, however, less porous, with minimal surface irregularities and less well-developed surface patterns, in contrast to biogenic struvite crystals formed through bacterial processes. As indicated above, the latter displayed greater variability in shape and size, with rounded edges and more regular dehydration patterns. The abiotic control further confirmed that struvite can form without biological input when phosphate and ammonium are freely available, thereby isolating the chemical pathway from the biological one.Fig. 8FESEM analysis of abiogenic struvite precipitation at different scales. Note that cracks and surface patterns are less well developed than in bacterial struvite, but are still visible (right image)
FESEM analysis (Fig. 9) also reveals bacterial cells and EPS closely associated with the surface of struvite crystals, indicating microbial involvement in crystal formation.Fig. 9FESEM images showing (**a**) bacterial cells and EPS on the surface of struvite crystals, and (**b**) magnified view of the EPS matrix, whichappears as a fibrous, filamentous network on the crystal surface
### Investigation of bacteria-biofilm-biomineral interactions using confocal laser scanning microscopy (CLSM)
CLSM images highlight significant differences in struvite crystal morphology and biofilm interactions between *S. pasteurii*,* B. subtilis* and *P. fluorescens*. Figure 10 illustrates the spatial distribution of bacteria, EPS and struvite crystals for the three different strains. This reveals significant differences in crystal density, size, and spatial organisation across bacterial species, highlighting the role of bacterial activity in mineralisation. The green fluorescence indicates the presence of bacterial cells, stained using Syto9, which marks live cells. EPS was stained red using Concanavalin A (ConA), a lectin-based dye that specifically binds to polysaccharides present in EPS (Strathmann et al. 2002). The resulting struvite deposits were visualised using the laser reflectance signal of confocal microscopy. Biomineralised struvite crystals produced a distinctive reflection signal, revealing irregular crystal geometries that aligned with voids in the biofilm matrix. Larger mineral deposits displayed annular patterns due to reflection occurring at the crystal surfaces, allowing their outlines to be clearly distinguished in each focal plane. Struvite crystals, visualised using laser reflection, exhibit blue, purple, or green hues, highlighting their distinct structural variations within the biofilm matrix. In *S. pasteurii* (Fig. 10b), large, well-defined struvite crystals are embedded within a biofilm that is densely populated with bacterial cells, which likely enhance mineralisation by providing nucleation sites and localised supersaturation conditions. However, there is minimal EPS present in this thick cell structure, as indicated by the lack of red-stained areas (ConA), suggesting that in the absence of EPS, bacteria can provide fewer nucleation points for crystallisation, resulting in larger struvite crystals. This supports the ability of *S. pasteurii* to promote uninterrupted crystal growth, closely resembling abiogenic struvite. In contrast, *B. subtilis* and *P. fluorescens* produce smaller, irregular crystals surrounded by dense red fluorescence signals, indicative of substantial biofilm and EPS production (Fig. 10c, e). In *B. subtilis* (Fig. 10d), multiple smaller struvite crystals are distributed across the biofilm surface, with the biofilm showing a more heterogeneous structure with EPS and bacterial cells. In the case of *P. fluorescens*, crystals form at various depths, indicating a dynamic mineralisation process driven by widespread nucleation (Fig. 10f), although overall crystal formation remained minimal.Fig. 10Confocal laser reflection microscopy (CLSM) imaging of mineral deposits in (**a**,** b**) *S. pasteurii*, (**c**,** d**)*B. subtilis*, and (**e**,** f**) *P. flurosences* cultures. Struvite minerals, bacterial cells and biofilm morphology are imaged. Minerals imaged by laser reflection appear blue/purple, bacterial cells appear green, and the biofilm appears red. Scale bar is 50 μm
## Discussion
The trend observed in *S. pasteurii*’s enzyme activity aligns with previously reported roles of this species in nitrogen metabolism and environmental adaptation, where urease activity is crucial for urea hydrolysis under varying nitrogen conditions (Dhami et al. 2016). These two enzymes drive the biomineralisation of struvite and thereby also regulate phosphorus availability in nutrient-limited environments (Bhatti et al. 1976; Khan et al. 2009). A decrease in alkaline phosphatase activity during this phase indicates a shift in metabolic focus away from phosphorus acquisition towards maintenance processes. This reduction in enzyme activity aligns with the typical metabolic adjustments bacteria make in response to nutrient limitations and environmental stress during the stationary phase (Avigad 1967; Khan et al. 2009). This indicates that *B. subtilis* maintains a balanced approach to nitrogen and phosphorus acquisition across different growth conditions.
Microbial enzymes, i.e., reported urease and phosphatase activity in the current study, play a significant role in struvite precipitation by mediating key biochemical processes. A study involving 72 bacterial strains has shown that both urease-positive and urease-negative bacteria can induce struvite precipitation, suggesting that additional microbial factors beyond urease activity contribute to this process (Desmidt et al. 2013). This is further supported by recent findings showing that bacteria can promote struvite formation through mechanisms such as the production of microbial biopolymers, even in the absence of urease activity (He et al. 2023a, b). Furthermore, variations in urease activity have been reported to influence the efficiency, rate, and extent of struvite formation, highlighting the complexity of the underlying mechanisms (He et al. 2023a, b; Rivadeneyra et al. 1999). In addition, enzymatic hydrolysis of polyphosphates by phosphatases has been shown to facilitate the controlled precipitation of phosphate salts, including struvite (Kofina et al. 2009). Microbial phosphatase activity, through the metabolism of organic matter, generates phosphate ions essential (along with ammonium ions) for struvite formation, highlighting its significant role in the precipitation process (Rivadeneyra et al. 2014). Therefore, this study characterised three microbes with varying urease and phosphatase activity (as shown in Fig. 1a-c) to explore the mechanism of struvite precipitation in different scenarios.
Regarding phosphate dynamics, *P. fluorescens* reached maximum soluble phosphate concentration earlier than the other strains, consistent with its high alkaline phosphatase activity (Bhatti et al. 1976; Khan et al. 2009). This rapid mobilisation of phosphate favoured struvite precipitation. *B. subtilis* displayed moderate alkaline phosphatase activity, resulting in a slower phosphate release and a delayed peak compared to *P. fluorescens*. This trend aligns with previous findings indicating moderate phosphatase efficiency in *B. subtilis* (RGlenn, A. 1975; Ng et al. 2022). *S. pasteurii* exhibited the slowest phosphate mineralisation, consistent with its lower phosphatase activity. Although *S. pasteurii* is efficient in urea hydrolysis (Dhami et al. 2016), its capacity to mobilise phosphate from glycerophosphate was comparatively limited. Consequently, the differences in soluble phosphate dynamics among the three strains primarily reflect variations in enzymatic hydrolysis efficiency, which in turn influenced the timing and extent of struvite formation (Fig. 2c).
In terms of nitrogen metabolism, the increased hydrolysis of urea by *S. pasteurii* was accompanied by a significant rise in pH, indicating substantial ammonia production because of urease activity. The high urease activity observed in *S. pasteurii* is consistent with its known efficiency in urea hydrolysis and biocementation processes, as previously documented (Bachmeier et al. 2002; Dhami et al. 2016). Similarly, *B. subtilis* showed a pH increase reflecting its urease activity and its impact on the flask environment. In contrast, *P. fluorescens* exhibited a lower level of urea hydrolysis and subsequent ammonia production. These findings indicate that precipitation efficiency (weight of precipitates in g/L) is not directly correlated with EPS production (which was highest in the case of *P. fluorescens*) but rather depends on a well-balanced enzymatic activity, i.e., availability of ammonia ions, that facilitates struvite formation (Fig. 2b, d). While the differences in struvite precipitation were linked to an apparent balance between urease and phosphatase activity, other factors may also contribute.
It is well established that elevated pH levels (approximately 7.5-10) are essential for efficient struvite crystallisation (Münch and Barr 2001; Wu and Bishop 2004). Within this range, struvite solubility reaches a minimum at pH 9.0-10.0, favouring the supersaturation conditions necessary for nucleation and crystal growth. Additionally, pH significantly influences the growth rate and morphology of struvite crystals. Elevated pH levels enhance supersaturation conditions, thereby accelerating the crystal growth rate (Le Corre et al. 2009). However, several studies have also indicated that higher pH levels promote a greater density of nucleation events, leading to the formation of smaller and more heterogeneous crystal populations (Rahman et al. 2014).
In addition to enzymatic activity, the composition of EPS influences the formation of crystals during biomineralisation by regulating crystal nucleation, growth, morphology, and polymorphism (Al Disi et al. 2019; Tourney and Ngwenya 2014; Decho and Gutierrez 2017; Ercole et al. 2007). This is primarily due to the diverse functional groups present in EPS, which interact with ions and mineral surfaces (López-Moreno et al. 2014; Braissant et al. 2003). Negatively charged functional groups in EPS, such as carboxylate and phosphate, exhibit a high affinity for multivalent cations like Ca^2+^, accumulating these ions at specific sites and creating more nucleation sites (Dubey et al. 2025). Additionally, functional groups like hydroxyl and amine in EPS bind selectively to certain crystal faces, modifying their growth rates and resulting in distinct crystal shapes and sizes (Wang et al. 2022).
FTIR results showed that *S. pasteurii*, *B. subtilis*, and *P. fluorescens* have the same major functional groups, though variations in peak intensity reflect compositional and structural differences in their EPS. Although the EPS produced by *S. pasteurii* and *B. subtilis* showed no major differences in functional group profiles, *S. pasteurii* produced considerably less EPS with a relatively balanced sugar and protein composition, suggesting a more limited biochemical role in modulating nucleation and growth. *B. subtilis*, while producing moderate EPS (~5.5 g/L), had a relatively high sugar-to-protein ratio, indicating a polysaccharide-dominated matrix with enhanced ion-binding and nucleation potential (Marvasi et al. 2010). Although its protein content was lower (14%), proteinaceous components may still influence crystal morphology through interactions with mineral surfaces (Hu et al. 2024).
*P. fluorescens* exhibited the highest absorbance intensities across most FTIR peaks, indicating that its EPS matrix contains abundant hydroxyl, amide, and phosphate-related functional groups capable of binding cations such as Mg^2+^, Ca^2+^, and NH~4~^+^. This observation is consistent with its high sugar and protein content, as *P. fluorescens* produced the largest biofilm density and EPS yield (~10 g/L). Its EPS composition, comprising approximately 55% sugar and 25% protein, reflects a polysaccharide-rich matrix enriched in hydroxyl and carboxyl groups. Such a matrix has strong ion-binding capacity, favouring ion adsorption but potentially inhibiting their mobility in solution. In moderate amounts, EPS can promote nucleation by providing reactive sites; however, excessive EPS production often traps ions and forms a dense matrix that retards crystal growth or stabilises amorphous phases (Braissant et al. 2003). In this system, the EPS of *P. fluorescens* likely inhibited struvite mineralisation. Consequently, despite its chemically active and functionally diverse EPS, *P. fluorescens* exhibited the lowest struvite yield (≈1.1 g/L), indicating that its abundant, sugar-rich EPS hindered effective nucleation and crystal development through ion complexation and surface blocking.
These biochemical and biofilm differences are reflected in the XRD patterns of the precipitates. The variations in the XRD patterns among the three bacterial strains suggest that microbial activity significantly influences the crystallisation pathway. *S. pasteurii* produced sharp and intense peaks, indicating highly crystalline material, whereas *B. subtilis* and *P. fluorescens* showed broader peaks and minor mismatches, suggesting either lower crystallinity or the formation of mixed phases. Such diversity in mineral products is consistent with the known ability of microbial metabolism to locally alter pH, ionic strength, and supersaturation conditions, all of which can affect mineral nucleation and growth dynamics (Dhami et al. 2013). Nonetheless, the data obtained and the good match with struvite peaks in JCPDS 15–0762/77–2303 support the conclusion that the bacterial precipitates predominantly consist of struvite. Further advanced characterisation techniques could be applied in future studies to resolve the presence of mixed or less hydrated phases more precisely.
Following the mineralogical analysis, FESEM micrographs (Fig. 6) clearly illustrate the distinct morphology of struvite crystals produced by three bacterial strains. The crystals generated by *S. pasteurii* exhibit standard morphological features characteristic of both microbial and inorganic struvite, including well-defined euhedral shapes and surface porosity (Kofina and Koutsoukos 2005; Prywer and Torzewska 2010; Chen et al. 2010; Bazin et al., 2012; Prywer et al. 2012; Zhao et al. 2021). The ability of *S. pasteurii* to produce these highly crystalline, well-structured crystals is likely attributed to its high urease and phosphatase activity, combined with negligible EPS production, as shown in Fig. 3a. This enzymatic profile promotes localised supersaturation and controlled nucleation, enabling the formation of large, well-developed struvite crystals without significant interference from extracellular polymers (Li et al. 2021).
In comparison, *B. subtilis* formed moderately sized crystals with irregular, rounded shapes, visible porosity, and frequent penetration or contact twinning, features typically associated with crystallisation under relatively high supersaturation (Abbona and Boistelle 1979). *P. fluorescens* produced the smallest and most fragmented crystals, with high porosity, predominantly prismatic shapes, and some X-like morphologies, consistent with crystallisation at the highest supersaturation among the strains, likely influenced by its high EPS production.
Overall, the size and morphological features of struvite crystals produced by *P. fluorescens* point to their crystallisation at the highest supersaturation of all tested cases. Conversely, the euhedral shape and larger size of struvite crystals, lacking twinning, in the case of *S. pasteurii*, point to their formation at the lowest supersaturation. Interestingly, *S. pasteurii* was the one producing the lowest amount of EPS, whereas *P. fluorescens* was the one producing the largest amount. This suggests that EPS may play an important role in delaying or modulating struvite crystallisation, as suggested by Li et al. (2021), thereby enabling the system to reach a very high supersaturation before nucleation could take place.
At a finer structural level, FESEM images in Fig. 7 suggest that struvite crystals develop through the aggregation of 1-3 µm crystallites into mesocrystals, producing a porous structure with triangular star-shaped gaps between the building units. These could suggest that the struvite crystals form through the aggregation of these smaller units, resulting in a porous crystal structure with visible gaps between units. While dehydration during FESEM preparation may cause these patterns, their regular geometry and close association with bacterial cells may represent authentic microbial influences on crystal growth. It has been suggested that such structural features are due to the formation of struvite via a non-classical crystallisation mechanism involving the oriented aggregation of nanometric building units forming a mesocrystal (Cölfen and Antonietti 2005; Chen et al. 2010). The empty space between such building units self-assembled in a crystallographic register is apparently responsible for the reported high porosity and surface area (150-166 m^2^/g) of struvite crystals (Kofina and Koutsoukos 2005; Wei et al. 2018). According to Prywer et al. (2012), such an assumed oriented aggregation was likely induced by differences in the polarity of opposite faces along the *c*-axis due to the non-centrosymmetric structure of struvite.
Notably, bacterial cells were observed adhering to the surfaces of the struvite crystals, suggesting a direct association with the crystallisation process. This adhesion highlights that such porous features were observed across all precipitation experiments involving *S. pasteurii*,* B. subtilis*, and *P. fluorescens*, which might suggest that bacterial activity plays a critical role in guiding struvite crystal formation with these specific features. However, such mesostructural features are also observed in abiotic struvite crystals. Recent results by Hövelmann et al. (2019) clearly demonstrated that the surface texturing and development of porosity in struvite crystals are not necessarily related to a mesocrystallisation process. They also showed that struvite crystals display the thoroughly reported surface patterns and cracks when observed at high vacuum in an FESEM, but they do show smooth surfaces lacking any porosity, cracks, or surface texturing when observed using Cryo-SEM. The latter unambiguously demonstrates that previous observations of surface texturing and porosity on struvite surfaces are an artifact. The study also reported that the pressure in the FESEM chamber and the vacuum were enough to induce struvite dehydration, resulting in the observed surface patterning and porosity/cracks. Note that dehydration of hydrated salts typically results in the development of distinctive cracks and surface patterns (Beaugnon et al. 2022).
Although we cannot rule out that bacterial struvite formed non-classically by self-assembly of individual micrometric units forming a mesocrystal, our FESEM observations (as well as the SEM observations in other studies) cannot be used as proof for mesocrystal formation, as the observed textural features are due to dehydration. Similarly, the observation of these distinctive surface patterns cannot be directly associated with a microbial origin as previously suggested (Prywer and Torzewska 2010; Prywe et al. 2012).
Nevertheless, there are some key differences between the textural features of bacterial and abiotic Even if we consider that the surface patterns observed are an artifact, bio-struvite typically exhibits wider gaps and larger subunits (1.5-2 µm × 1–1.5 µm) compared to chemical struvite, which has smaller subunits (0.8 µm × 0.4 µm). These structural differences could be influenced by the degree of supersaturation during crystal formation, where lower supersaturation favours crystal growth over nucleation, resulting in larger and more perfect crystals (Leng and Soares 2021; Rahman et al. 2014), which, upon dehydration, show a better-structured surface pattern. These features might also be related to the presence of bacterial cells and organic byproducts of bacterial activity and their interaction with forming struvite crystals, which may get occluded in and/or adsorbed on the struvite crystals, modifying the dehydration dynamics and kinetics. The latter could lead to the well-structured cracks and surface pattern shown in Fig. 7. Microbial activity, particularly bacterial secretions and extracellular polymeric substances (EPS), creates a dynamic microenvironment that modulates crystallisation (Tansel et al 2018), facilitating crystal growth and stability (Leng and Soares 2021; Prywer et al. 2012).
A quantitative comparison reveals a significant difference in the efficiency of abiogenic and bacterial struvite precipitation. Under identical conditions and over the same period, abiogenic precipitation resulted in a yield of only 0.8 g/L of struvite. In contrast, bacterial precipitation was considerably more efficient, with *B. subtilis* producing the highest yield at 3.19 g/L, followed by *S. pasteurii* at 2.23 g/L and *P. fluorescens* at approximately 1.11 g/L (Fig. 2a). This confirms that under chemically favourable conditions-specifically, supersaturation of Mg^2+^, NH~4~^+^, and PO~4~^3−^ ions-abiotic struvite precipitation can occur. However, the rate of precipitation in the abiotic control was significantly lower compared to the bacterial cultures. This suggests that while abiotic mineralisation is possible, the presence of bacteria significantly accelerates struvite formation. Moreover, struvite morphology is greatly influenced by the presence of enzymes and EPS. Conversely, the abiogenic process relies solely on physicochemical factors, leading to lower precipitation rates.
Overall, the results indicate that while the fundamental lattice structure of struvite remains consistent across bacterial strains, the efficiency of crystal growth is heavily influenced by bacterial species and their metabolic byproducts*. S. pasteurii* demonstrates the most effective crystallisation process, producing euhedral crystals with limited porosity, more closely resembling those formed through abiogenic processes. Its efficient urease activity and low EPS production facilitate crystal growth at relatively low supersaturation, resulting in large, euhedral crystals. Previous studies have extensively explored the individual roles of urease and alkaline phosphatase enzymes in struvite precipitation separately (He et al. 2023a, b). Urease enzymes are crucial for hydrolysing urea into ammonia and carbon dioxide, which elevate pH levels and create conditions favourable for crystal formation (Castro-Alonso et al. 2019). On the other hand, alkaline phosphatase enzymes are involved in phosphate release from organic compounds, which is essential for providing phosphate ions necessary for a free phosphate source in struvite crystal growth (Leng & Soares 2021).
The presence of biofilm and EPS produced by bacterial strains such as *P. fluorescens* and *B. subtilis* appears to significantly influence struvite crystallisation, leading to smaller and less crystalline structures, reflecting crystallisation at a high supersaturation. While biofilms and EPS are known to facilitate nucleation when they act as a template (Arp et al. 2001), their overproduction may hinder nucleation and crystal growth (Li et al. 2021). This could result in irregular, twinned, or fragmented morphologies, which we observed for *P. fluorescens*. Although direct adsorption or inhibition assays were not performed, this interpretation is consistent with the relationship between high EPS production and reduced precipitation efficiency. This contrasts with *S. pasteurii*, which produces minimal EPS, thereby supporting the formation of larger, well-defined crystals resembling abiogenic struvite. However, the other reason behind the larger crystal produced with *S. pasteurii* could be credited to a higher urease activity, which resulted in high availability of ammonia for struvite precipitation. The observed variations in crystal morphology and sizes of precipitated struvite are the result of a complex interplay between the enzymatic activities and EPS (quantity, functional group and composition) of the different bacterial species.
The observed differences in crystal precipitation among bacterial strains suggest that microbial EPS plays a critical role in modulating biogenic crystal formation. EPS-producing bacteria exhibited more irregular and singular crystal morphologies, indicating that EPS affects nucleation and growth dynamics. While limited research has examined EPS effects on struvite, its role in calcium carbonate precipitation is well documented. Studies have shown that organic macromolecules, particularly EPS, can affect calcium carbonate polymorphism, morphology, and orientation (Addadi and Weinert 1985; Ercole et al. 2007; Perri et al. 2018; Dubey et al. 2025). In contrast, studies focusing on struvite crystals and the role of EPS are less extensive but provide useful insights. For instance, Cruz et al. (2023) demonstrated that increasing concentrations of sodium alginate, a model EPS, altered struvite crystal habit from prismatic to plate-like twinned forms, suggesting that EPS molecules can adsorb onto crystal faces and regulate their growth. This indicates that EPS can both promote heterogeneous nucleation and, when present in excess, may hinder crystal development by blocking active sites and disrupting lattice formation. In agreement with those results, the overdevelopment of struvite (001) faces, consistent with the high intensity of the 002 and 004 Bragg peaks (XRD results) observed here, likely resulted from the adsorption of EPS on such faces.
Following the morphological and microstructural analyses, the CLSM observations reveal clear strain-specific differences in crystal formation and biofilm structure. In the case of *P. fluorescens*, the lack of significant precipitation suggests lower enzyme activity related to struvite formation, potentially due to weaker urease activity or potentially combined with biofilm and EPS characteristics that sequester ions and modify surface reactivity. Overall, these observations demonstrate a clear correlation between bacterial activity, biofilm density, and mineralisation patterns.
These results imply that EPS plays a critical role in regulating the size, morphology, and nucleation density of struvite crystals. The lack of significant EPS in *S. pasteurii* seems to allow for larger crystals to form, whereas the abundant EPS secreted by *P. fluorescens* may adsorb ions or create barriers that slow crystal growth. Although direct adsorption or inhibition curve experiments were not performed in this study, such analyses would be valuable for quantitatively validating the proposed mechanism. Understanding the exact role of EPS and other biofilm characteristics in crystal formation will be essential for deeper insights into this biomineralisation process. The present study focused on EPS within living microbial systems, where enzymes, cells, and polymers interact simultaneously. However, a dose–response investigation applying the extracted EPS exogenously to sterile struvite precipitation media could yield valuable insights into its independent role. Furthermore, specific mechanisms involving functional groups, organic acids, and ion-chelating processes were not assessed in the current study. Future research that addresses these processes will be essential for more comprehensively understanding the mechanism of microbial struvite precipitation.
## Conclusion
The impact of three bacterial strains, *Sporosarcina pasteurii*,* Bacillus subtilis*,and* Pseudomonas fluorescens*, on the kinetics of struvite precipitation and its morphology was investigated. Clear differences in bacterial enzymatic activities, biofilm formation, and EPS production were observed, all of which appeared to influence struvite precipitation dynamics and crystal characteristics. The major findings from the current study are as This study identified a clear distinction between biogenic and abiogenic struvite in terms of crystal shape, morphology, size, quantity as well as dehydration features, highlighting the role of microbes in biogenic struvite precipitation.The enzymatic activities, EPS quantity and EPS composition appeared to influence the quantity and quality of the crystalised struvite. Availability of ammonia was associated with the formation of larger crystals. Compared to abiogenic crystals (with ammonium chloride) that reached an average size of 287 ± 15 µm, biogenic crystals by *S. pasteurii* were recorded to be 180 ± 12 µm in size. In contrast, *B. subtilis* and *P. fluorescens* produced smaller crystals of average size 103 ± 5 µm and 37 ± 4 µm. These size differences likely reflect interactions between EPS quantity/composition, urease activity, and ammonia availability.The quantity of biogenic struvite was influenced by the balance of urease and phosphatase activity, as urease is required to break down ammonia, whereas phosphatase is critical for the crystallisation of struvite. *B. subtilis* was observed to precipitate the highest quantity of struvite (3.19 g/L), followed by *S. pasteurii* (2.23 g/L) and then least in *P. fluorescens* (1.11 g/L), highlighting the significance and role of enzyme activity as well as EPS in struvite precipitation. Microbial cultures also influenced the reaction kinetics of struvite precipitation, and their mineralogical-morphological features.The effect of biofilm and EPS on crystal morphology was also demonstrated with confocal microscopy. Excessive EPS disrupted crystal growth, leading to irregular morphologies, especially in *P. fluorescens*. In contrast, the minimal EPS production by *S. pasteurii* was associated with a more cohesive crystal formation. The in-depth role of EPS in impacting morphological features of biogenic struvite was recorded under the specific experimental conditions of this study.Dehydration induced by the vacuum environment in the FESEM resulted in the formation of a distinctive surface pattern and extensive porosity generation due to the gaps developed between individual micrometric crystal sections leading to patterned surface. These structural features were observed in both the biogenic and abiogenic struvite crystals. However, biogenic struvite crystals developed better structured surface patterns during dehydration, composed of small crystallites, each displaying well-defined faces, with individual crystalline units averaging 1 to 3 μm in size, which were less evident in the abiogenic struvite.
Overall, the findings suggest that type of microbial culture, their metabolic activity (and byproducts) and EPS play a crucial role in struvite crystallisation. Amount and composition of EPS, biofilm and enzyme activity (both urease and phosphatase) influences the rate, reaction kinetics and morpho-mineralogical features of biogenic struvite under the conditions tested.
The findings of this study provide crucial insights into the microbial struvite biomineralisation process and hold significant potential for future engineering applications across diverse fields. In nutrient recovery, controlling enzyme activity and extracellular polymeric substance (EPS) levels can facilitate the formation of larger and more refined crystals, thereby improve phosphorus recovery efficiency and enhance product quality in wastewater treatment. In civil engineering contexts, these insights can be applied to capture ammonia during CaCO₃-based biocementation via ureolytic pathway. In agriculture, the release rate of N and P from struvite used as a fertilizer could be regulated by forming smaller (more reactive) or larger (less reactive) crystals using different bacterial strains, as shown here. Within bioremediation, the selection of appropriate microbial strains, alongside optimisation of their enzymatic activity and EPS production, can increase the effectiveness of contaminant removal. Additionally, in pathology, a deeper understanding of the microbial role in generating distinct struvite morphologies may contribute to improved diagnosis and treatment of infection-related urinary stones.
Further research can explore the impact of a broader range of microbial communities from different sites, environmental conditions, genetic features and substrates on biogenic struvite formation. Advanced tools to investigate the bio-mineral interactions at nanoscale can be utilised as Nano-SIMS (Nano Secondary Ion Mass Spectroscopy), ToF-SIMS (Time of Flight- Secondary Ion Mass Spectroscopy), AFM (Atomic Force Microscopy) to improve further our understanding of bacterial-induced struvite precipitation and EPS-mineral interactions.