Abiotic and biotic drivers of struvite solubilization in contrasting soils

Antonio RUIZ-NAVARRO,Manuel DELGADO-BAQUERIZO,Concha CANO-DÍAZ,Carlos GARCÍA and Felipe BASTIDA,

1Centro de Edafología y Biología Aplicada del Segura,Consejo Superior de Investigaciones Científicas(CEBAS-CSIC),Department of Soil and Water Conservation,Espinardo,Murcia E-30100(Spain)

2ICL SpecialtyFertilizers Iberia,Totana,Murcia E-30850(Spain)

3Laboratorio de Biodiversidad yFuncionamiento Ecosistemico,Instituto de Recursos Naturales yAgrobiología de Sevilla(IRNAS),CSIC,Av.Reina Mercedes 10,Sevilla E-41012(Spain)

4IRNAS-CSIC,Unidad Asociada Universidad Pablo de Olavide,Universidad Pablo de Olavide(UPO),Sevilla 41013(Spain)

5Centre for Researchand Development in Agrifood Systems and Sustainability(CISAS),Instituto Politécnico de Viana do Castelo(IPVC),Rua Escola Industrial e Comercial Nun’Álvares 34,Viana do Castelo 4900-347(Portugal)

ABSTRACT Phosphorus(P)limitation in the coming decades calls for the utilization of alternative fertilizers in agriculture.Struvite is a promising P source,but its potential role as a fertilizer is dependent on different physical,chemical,and biological properties,which are very heterogeneous in soil,complicating the prediction of the best soil conditions for its application.Here,we evaluated the solubility of struvite in soil,its redistribution into P fractions,and its potential abiotic and biotic drivers in 62 globally distributed soils with contrasting properties through an incubation assay.We found that after 40 d,about 35%of struvite P was redistributed into soil fractions more accessible to plants and microbes.Phosphorus redistribution from struvite was driven by a complex suite of soil physical,chemical,and microbial properties as well as environmental factors that varied across soils.Soil texture played a critical role in determining the redistribution of P in struvite-amended soils in soluble(H2O extraction),labile(NaHCO3 extraction),and moderately labile(NaOH extraction)fractions.In addition,the soil solution cation concentration was one of the most important drivers of available struvite-derived P fractions.The great importance of texture and cations in determining struvite-derived P fractions in soil was contrasted with the relatively minor role of pH.At the microbial level,the number of bacterial operational taxonomic units(OTUs)from the unfertilized soils that correlated with struvite-derived P fractions was higher than that of fungi.The number of OTUs that correlated with the struvite-derived soluble P fraction was dominated by fungi,whereas the number of OTUs that correlated with the struvite-derived labile P fraction was dominated by bacteria.Overall,this study provided a predictive framework for the potential use of struvite as a P fertilizer in contrasting soils.

KeyWords: bacteria,fungi,environmental factor,operational taxonomic unit,phosphorus fertilizer,phosphorus fraction,phosphorus solubilization,soil fertility,soil property

INTRODUCTION

Unlike soil carbon (C) and nitrogen (N), which are available through atmospheric biological fixation,soil phosphorus (P) is mostly available from the solubilization of the bedrock.Mineral P fertilizers are widely used to promote crop productivity.However,the main source of these P fertilizers—the phosphate rock—is a finite and nonrenewable resource mined only in a few regions, mainly in Western Sahara mines(Gilbert,2009;Filippelli,2018).Phosphate rock is expected to be exhausted in 70—100 years,which may compromise the maintenance of food security along with geopolitical disequilibration(Cordell and White,2015),leading to the so-called“P crisis”(Abelson,1999;Heckenmlleret al.,2014).New regulations of the European Union are restricting the use of phosphate rock because it usually contains certain heavy metals(Huygens and Saveyn,2018).In this scenario,there is a strong need to find alternative sources of P fertilizers that can satisfy global demands and maintain food security(Schrderet al.,2011;Daneshgaret al.,2018;Menezes-Blackburnet al.,2018).

The recycling and recovery of P from different wastes are widely regarded as essential for the sustainable maintenance of soil fertility.Struvite(MgNH4PO4·6H2O)is a salt of magnesium ammonium phosphate recovered by precipitation and crystallization from wastewater and sewage sludge.It has a greater fertilizer potential than other recovered P products,such as biosolid and municipal solid waste compost, due to its low hazardous substance concentration and/or high P concentration(Schneideret al.,2019).Therefore,several initiatives have been developed for fostering the biotechnological production of struvite(Gntheret al.,2018).Some countries such as the USA and Canada have already allowed the use of struvite as a fertilizer, while the regulations in Europe have been revised to permit its use in the coming years(Huygens and Saveyn,2018).Despite its low solubility,struvite has shown a similar agronomic performance to that of water-soluble mineral P fertilizers;however,its fertilizing effects are highly variable.The effects of soil properties on struvite P dissolution remain poorly understood although struvite P dissolution was found to be correlated with soil pH and texture in field and pot experiments(Vaneeckhauteet al., 2016; Degryseet al., 2017; Nongqwengaet al.,2017; do Nascimentoet al., 2018; Robles-Aguilaret al.,2020).In general, struvite P availability in soil decreases as pH increases, reaching a minimum at pH 9—11.A higher dissociation rate of struvite has been observed in soils with high sorption capacity, such as those with high clay and silt concentrations,because sorption sites act as sinks for phosphates and cations.How other soil and environmental properties influence struvite-derived P availability remains unknown.Certain soil microbial populations are important for P solubilization through the production of organic acids.For example,phosphate-solubilizing bacteria and mycorrhizal fungi are key in driving P solubilization(Di Tomassiet al., 2021; Sileset al., 2022).However,the composition of microbial communities in soil is extremely variable(Delgado-Baquerizoet al.,2018)and can determine soil capacity for solubilizing struvite P.A better understanding of the global soil drivers of P solubilization is essential for designing efficient P fertilizers based on new mineral sources,which are fundamental for the maintenance of soil fertility,agricultural productivity,and food security.

Given the forecasted exhaustion of phosphate rock,the supply of alternative P sources is fundamental for ensuring crop production in the coming decades and,therefore,for achieving the Sustainable Development Goal 2(zero hunger)of the 2030 Agenda of the United Nations.In this study,we investigated the environmental factors controlling the solubilization and fractionation of struvite P.Soils from contrasting biomes (from deserts to tropical ecosystems)were sampled to capture a wide range of soil properties such as pH,nutrient concentration,organic C(OC)concentration,texture, salinity, and microbial biomass, activity, and diversity.The selected soils were not subjected to any previous agricultural practices.We hypothesized that P solubilization from struvite worldwide would be dependent on multiple soil factors rather than on a single one.However,considering the mineral characters of struvite and the fact that different local studies have highlighted pH as a major driver of mineral P solubilization(Cabezaet al.,2011;Bastidaet al.,2019), we expected that soil pH would play an important role in controlling the solubilization and distribution of P from struvite into different P fractions and that other factors,also varying on broad scales, may weaken the role of pH in soil P solubility.A recent meta-analysis demonstrated that several environmental factors, including climate and vegetation cover,also influence soil P supply to plants(Houet al., 2020).Therefore, we also hypothesized that these environmental factors would play a key role in struvite P supply to globally distributed soils.Our study will help to elucidate the behaviour of struvite P in soil and reveal soil conditions relevant for fertilization with struvite.

MATERIALS AND METHODS

Field surveyand soil characterization

Sampling was designed to obtain a wide gradient of edaphic and environmental properties that allow us to study the abiotic and biotic factors governing solubilization and distribution of struvite-derived P in soil.For this purpose,selecting sites from different climates maximized the likelihood for obtaining a gradient of soils with contrasting properties(Tables SI and SII,see Supplementary Material for Tables SI and SII).Field data were collected from 62 globally distributed soils throughout the six continents with contrasting climates(cold,Mediterranean,oceanic,semiarid,and subtropical)and vegetation cover(forest,shrubland,and grassland)(Fig.S1,see Supplementary Material for Fig.S1).These sites were selected because they have never been fertilized,which eliminated potential effects from artificial fertilizers.Field surveys were conducted according to a standardized sampling protocol(Maestreet al.,2012).At each site, we surveyed a 50 m× 50 m plot.Three parallel transects of the same length, spaced 25 m apart, were established.

One composite topsoil(0—10 cm)sample was composed of five soil cores collected under the dominant ecosystem vegetation across each plot.Following field sampling,soils were sieved(<2 mm)and frozen at-20◦C for later analyses.Electrical conductivity(EC),pH,texture,and concentrations of OC, N, P in different fractions, and water-soluble Ca,Na, Mg, K,Al, and Fe were measured in all soil samples after sieving.Electrical conductivity and pH were measured in a soil and water suspension(1:2.5,weight:volume).Soil texture was determined as percentage of fine fractions(clay and silt)according to Kettleret al.(2001).Organic C concentration was determined by colorimetry after oxidation with a mixture of potassium dichromate and sulfuric acid.Nitrogen was analyzed using a CN analyzer(LECO CHN628 Series,LECO Corporation,St Joseph,USA).Phosphorus concentration was analyzed after wet acid digestion with hydrofluoric acid in an inductively coupled plasma-optical emission spectrometry(ICP-OES,Elemental Iris Intrepid II XDL,Thermo Scientific,Franklin,USA).Soluble P,Ca,Na, Mg, K, Al, and Fe concentrations were measured in water extracts by ICP-OES,and the sum of soil base cation concentrations in the water extract was calculated as the sum of Ca, Na, Mg, and K concentrations.Soil OC concentration ranged between 1 and 270 g kg-1,the C:N ratio ranged between 1.4 and 170.6,total P ranged between 64.5 and 3 101 mg kg-1,the sum of base cation concentrations ranged between 53.3 and 1 153.3 mg kg-1,Fe ranged between 0.11 and 100.2 mg kg-1, Al ranged between 0.31 and 214.7 mg kg-1,pH ranged between 3.9 to 8.8,and the percentage of clay and silt varied between 0.5%and 86%(Table SI).

Soil incubation withstruvite

Frozen soil samples were pre-incubated in laboratory conditions to reactivate soil and microbial communities.For this purpose, 15 g of each soil sample was placed into 125 mL microcosms,with moisture at 60% maximum water-holding capacity,for 7 d at 28◦C.Struvite from the Murcia-Este Wastewater Treatment Plant(Spain)was used in the microcosm assays.The mineralogical purity of struvite was high(98%).The concentrations of total K,P,Mg,Ca,Fe, Cu, Zn, and Mn (Table I) in struvite were analyzed by ICP-OES.Electrical conductivity and pH of struvite were measured in a struvite and water suspension (1:2.5,weight:volume).The concentrations of hazardous materials,such as heavy metals,were low or non-detectable,as in case of microbial pathogens(Table I).

After the pre-incubation of soils,two parallel soil sample series were set up,one without struvite(unfertilized,control)and one with struvite(fertilized)at a dose of 8.7 mg P g-1soil.Unfertilized and struvite-fertilized soils were incubated at 28◦C and 60% maximum water-holding capacity for 40 d.Because this assay was conducted without plants,the struvite dose was not selected according to agronomical criteria.This allowed us to search for the factors that determine struvite P solubilization and distribution in a wide range of soils.

Phosphorus fractionation

After 40 d of incubation,P fractionation was carried out on struvite, unfertilized soils, and struvite-fertilized soils according to the Hedley fractionation method(Hedleyet al.,1982)modified by Tiessen and Moir(1993).Soil samples were sequentially extracted by 30 mL of the respective reaction reagents:H2O(water-soluble P,H2O-P),0.5 mol L-1NaHCO3(exchangeable P,NaHCO3-P),0.1 mol L-1NaOH(secondary mineral P,organic matter-,Fe-,and Al-associated P,NaOH-P),and 1 mol L-1HCl(carbonate-associated P as apatite,HCl-P)for 16 h on a shaker.Subsequently,the samples were centrifuged at 4 100×gfor 15 min,and P concentrations in the extracts were determined using ICP-OES.Soil residue was digested in a microwave oven (ETHOS-1600, Milestone, Sorisole, Italy) with HNO3:H2O2(3:1,volume:volume),and P concentration was measured;however, the P concentration in this fraction was negligible,showing that all the P was extracted in the previous stages.These fractions represent a gradient of P availability for plants and soil microbes: soluble or completely available(H2O-P),labile or potentially available(NaHCO3-P),moderately labile or partially available(NaOH-P),and immobile or hardly available (HCl-P) (Weihrauch and Opp, 2018;Leiet al., 2020).The difference in P concentration between struvite-fertilized and unfertilized soils represented the struvite-derived P concentration retained in each fraction.

TABLE I Mineralogical,chemical,and microbial properties of the struvite used in this study

Microbiological analysis in the unfertilized soil

Different microbial and biochemical parameters were analyzed in the unfertilized soil after incubation to ascertain the biotic factors that could drive the solubilization of P from struvite in soil.Soil microbial biomass was estimated by phospholipid-derived fatty acids(PLFAs)extracted from a 0.5 g freeze-dried soil sample using the method described by Bligh and Dyer(1959)and modified by Buyer and Sasser(2012).The extracted fatty acid methyl esters were analyzed using an Agilent Technologies 7890B gas chromatograph with an Agilent DB-5 column(Agilent Technologies,Santa Clara,USA).The fatty acids selected to represent bacterial biomass are the PLFAs i15:0,a15:0,15:0,i16:0,16:1ω7,17:0,i17:0,a17:0,cy17:0,18:1ω7,and cy19:0,and the fatty acid representative of fungal biomass is 18:2ω6(Frostegård and Bååth,1996;Rinnan and Bååth,2009).The fatty acid 16:1ω5 was selected as representative of mycorrhizal fungi(Olssonet al.,1995).Total biomass was calculated as the sum of bacteria and fungi,and the fungal-to-bacterial biomass(F:B)ratio was estimated as an indicator of microbial community structure.Microbial biomass ranged between 5.7 and 6 213.8 nmol g-1,while the F:B ratio was between 0.01 and 0.33(Table SI).Soil basal respiration was analyzed through the quantification of CO2by gas chromatography and used as an indicator of the general microbial activity.Soil phosphatase activity was measured following the method of Tabatabai and Bremner(1969).

The composition of bacterial and fungal communities was analyzed through amplicon sequencing using the Illumina MiSeq platform.Ten grams of frozen soil(per sample)were ground under liquid nitrogen using a mortar to homogenize soil and obtain a representative soil sample.Soil DNA was extracted using the Powersoil®DNA isolation kit(MoBio Laboratories,Carlsbad,USA)according to the manufacturer’s instructions.A portion of the bacterial 16S(V3—V4 region)and eukaryotic 18S(V9 region)rRNA genes was sequenced using the 341F/805R and Euk1391f/EukBr primer sets,respectively.Bioinformatic processing was performed using a combination of QIIME76,USEARCH77 and UNOISE378.The relative abundance of microbial phyla was obtained from these analyses.In total,52 samples of fungi and 62 samples of bacteria were successfully sequenced and used for statistical analyses.

He took the precaution of surrounding the palace with a dense32 cloud, and then hastened to his Court, where his prolonged absence was causing much anxiety

Statistical analysis

The Kruskal-Wallis one-way non-parametric analysis of variance was used to elucidate the differences in struvite P fractions between the raw material(without soil)and fertilized soils after incubation.Spearman correlation analysis was used to explore relationships between struvite-derived P fractions in soil with environmental,chemical,biochemical,and microbial variables,and random forest(RF)analysis was performed to identify the main predictors of P concentration in different fractions as described by Delgado-Baquerizoet al.(2016).Furthermore, to explore the relationship between microbial populations and struvite-derived P in different fractions,we developed bacterial and fungal phylogenetic trees with coloured rings to relate P fractions to biodiversity from a cladistic perspective.For this,sequences were aligned using the online tool of Silva Incremental Aligner(Pruesseet al., 2012) enabling the options “Search and Classify”and“Compute Tree”to search and obtain the most similar sequences(closest neighbours)from the Silva SSU database(Silva SSU database 138 release)and then reconstruct the phylogenetic trees using the tree of neighbor sequences as a reference.This method is recommended for obtaining more robust and consistent tree hypotheses when the sequences are short.The visualization and annotation of the trees were conducted with the online tool iTol(Letunic and Bork,2019).

RESULTS

Struvite P fractions

The analysis of raw struvite indicates that a great part of P was present in the non-available(HCl-P,69.4%)and moderately labile(NaOH-P,28.6%)fractions(Fig.1).Minor proportions of struvite P were found in the potentially(NaHCO3-P,1.7%)and completely(H2O-P,0.2%)available fractions.The more available struvite-derived P fractions noticeably increased when struvite was added to soil.Considering the average of all struvite-fertilized soils, 23.8% of struvite-derived P was found in the HCl-P fraction,40.7%in the NaOH-P fraction,26.3%in the NaHCO3-P fraction and 9.2%in the H2O-P fraction.On average,our results indicate that 35.5% of struvite P added to the soil was potentially available for plants after 40-d incubation.

Fig.1 Distribution of different P fractions of raw struvite and struvite in struvite-fertilized soils.Values are means with standard deviations shown by vertical bars(n=62).Differences(D)in each P fraction between raw struvite and struvite in struvite-fertilized soils are considered significant at P <0.05 according to one-way non-parametric analysis of variance.H2O-P,NaHCO3-P,NaOH-P,and HCl-P=H2O-,NaHCO3-,NaOH-,and HCl-extractable P,respectively.

Relationships of environmental factors and soil properties withstruvite-derived P fractions in soil

The random forest analysis provided a general overview of the importance of several drivers of struvite-derived P fractions in soil(Fig.2).Soil physical and chemical parameters,including soil texture,concentrations of soil solution cations (sum of Ca, Na, Mg, K, Fe, and Al), total P, and OC,contributed to the P concentrations of soluble(H2O-P)and moderately labile(NaOH-P)fractions.In comparison,microbial and environmental attributes were more relevant for labile P fraction(NaHCO3-P).Soil texture was a significant factor for the soluble,labile,and moderately labile P fractions,showing the overall importance of texture as a predictor of struvite-derived P availability in soil,particularly in the moderately labile fraction.

Fig.2 Random forest mean importance of selected predictor variables for different P fractions extracted by H2O(H2O-P,a),NaHCO3 (NaHCO3-P, b), NaOH (NaOH-P, c), and HCl (HCl-P, d) using the random forest analysis.Model predictable capacity was higher for soluble(H2O-P,R2 =0.356)and labile(NaHCO3-P,R2 =0.553)P fractions than for moderately labile (NaOH-P, R2 = 0.189) and immobile (HCl-P, R2 = 0.001) P fractions.Asterisk*indicates the variable is a significant factor at P <0.05.MAT=mean annual temperature;AI=aridity index;PC=plant cover; OC = organic C; EC = electrical conductivity; TP = total P;ΣCationH2O = sum of soil base cations in water extract; FeH2O and AlH2O =Fe and Al in water extract,respectively;BR=basal respiration;MB=microbial biomass;F:B ratio=fungal-to-bacterial biomass ratio;EB=ectomycorrhizal biomass.

Soil solution cation concentrations were one of the most important drivers of available struvite-derived P fractions(Fig.2).Soil phosphatase activity was also a relevant predictor for the completely available fraction(H2O-P),while soil total P concentration,OC concentration,and pH significantly affected the potentially available P fraction (NaHCO3-P).The concentrations of Fe and Al in the soil solution were important parameters for struvite-derived NaHCO3-P and NaOH-P,which was also significantly affected by soil total P concentration and EC.Moreover,environmental factors,such as aridity index(AI)and mean annual temperature(MAT),and microbial variables, such as biomass and respiration,were also important predictors for the NaHCO3-P fraction.

Correlations of struvite-derived P fractions in soil with environmental factors and soil physical,chemical,and microbial properties weekend with the recalcitrance of the P fractions(Fig.3).

Fig.3 Spearman correlations of struvite-derived P fractions in soil with selected soil properties and environmental factors.See Fig.2 for detailed description of each P fraction,soil property,and environmental factor.

Indeed,the most recalcitrant fraction(HCl-P)was correlated only with soil C:N ratio,while the moderately labile fraction(NaOH-P)was positively associated with OC concentration,texture,microbial biomass,and total P concentration(Fig.3).A greater number of significant correlations was observed for the most labile fractions(H2O-P and NaHCO3-P); both fractions were significantly positively correlated with soil solution cation concentrations.Positive relationships were observed between H2O-P and EC,a parameter associated with ion concentration in soil solution,and between NaHCO3-P and pH,a parameter linked to the saturation of bases in the soil cation exchange complex.Phosphatase activity correlated positively with the H2O-P fraction.Soil basal respiration, microbial biomass, and OC correlated positively with the H2O-P and NaOH-P fractions.Among environmental factors,only plant cover correlated positively with the H2O-P fraction.

Struvite-derived P fractionation and its association withsoil microbial community

In order to further investigate the potential involvement of native soil microbial populations in struvite-derived P solubilization,we analyzed the correlation between the relative abundance of microbial groups in the unfertilized soils and P solubilized from struvite in different fractions after incubation(P<0.01,Fig.4,Table SIII,see Supplementary Material for Table SIII).

In total, the relative abundance of 104 operational taxonomic units (OTUs) correlated with struvite-derived P concentrations in all fractions (Tables SIII).The number of bacterial OTUs that correlated with struvite-derived P fractions (67 OTUs) exceeded that of fungal OTUs (37 OTUs).Overall,the number of OTUs that correlated with struvite-derived P solubilization was greater for the labile fraction (NaHCO3-P) (47 OTUs) than the water-soluble(H2O-P)(27 OTUs)and moderately labile(NaOH-P)(29 OTUs)fractions.Only one OTU correlated significantly with the immobile fraction(HCl-P).

Fig.4 Phylogenetic trees of soil bacteria(a)and fungi(b)in the unfertilized soils after incubation.Phylogenetic trees were constructed using the maximum likelihood model(RAxML)of the online tool of Silva Incremental Aligner.The tree of the closest neighbours(full length sequences)from Silva SSU database was used as a reference tree.H2O-P,NaHCO3-P,NaOH-P,and HCl-P=H2O-,NaHCO3-,NaOH-,and HCl-extractable P,respectively.

In the case of the struvite-derived moderately labile P fraction(NaOH-P)(Fig.4,Table SIII),the correlated OTUs preferentially originated from bacteria rather than fungi.Among bacteria,several Acidobacteria(i.e.,Solibacterales)and Actinobacteria (especially Gaiellales) dominated, followed by different Proteobacteria:Alphaproteobacteria such as Rhizobiales(BradyrhizobiumandRhodoplanes),Betaproteobacteria (Variovorax), Deltaproteobacteria (Myxococcales), and Verrucomicrobia (Pedosphaerales and Chthoniobacterales).The number of fungal OTUs related to this fraction was much lower than that of bacterial OTUs and was dominated by Ascomycota(Pezizomycotina).Only one OTU,an Actinobacteria of the Acidimicrobiales order,correlated significantly with the struvite-derived immobile P fraction.

DISCUSSION

Soil properties driving the redistribution of P in struviteamended soils

The limitation of P fertilizers in the coming decades calls for the utilization of alternative P sources in agriculture and restoration efforts.However,the large variability of soils requires the creation of a framework that allows identification of soils suitable for the application of novel mineral P sources such as struvite.The effects of soil properties on the solubilization of struvite-derived P remain unclear,although the effects of pH and texture have been reported by field and pot experiments(Vaneeckhauteet al.,2016;Degryseet al.,2017;Nongqwengaet al.,2017;do Nascimentoet al.,2018;Robles-Aguilaret al.,2020).Through a broad screening of soil types,our results suggest that within 40 d of incubation,an average of 35%of struvite-derived P was redistributed into soil fractions that are accessible to plants and microbes,such as soluble(H2O-P)and labile(NaHCO3-P)fractions(Fig.1).These results were in agreement with previous studies suggesting the potential of struvite to promote P availability and fertility in soil (Bastidaet al., 2019; Andersonet al., 2021).It is important to highlight that this average value(35%)was obtained across a wide variety of soils representing contrasting biomes and soil properties.

Our results highlight the critical role of soil texture in determining the redistribution of struvite-derived P in soluble (H2O-P), labile (NaHCO3-P), and moderately labile(NaOH-P)fractions,as observed by the random forest analysis(Fig.2).Moreover,this was particularly evident in the case of texture and the NaOH-P fraction shown by the Spearman correlation analysis(Fig.3).This result suggests that the concentration of clay plus silt is more important in retaining P in the less available fraction than in the soluble and labile fractions.There are strong electrostatic affinites between dissolved P ions in the soil solution and free charges within the soil matrix(exchangeable soil cations and clayhumic complex); therefore,P ions can be easily bound to soil particles(Turanet al.,2012;Weihrauch and Opp,2018).Indeed, the important role of texture in determining the solubility of phosphate in soil has been described by several local (Löpez-Hernández and Niño, 1993; Huffmanet al.,1996) and meta-analysis (Yuet al., 2021) studies, which usually found a greater capacity of clay-rich soils to adsorb P(Gérard,2016).Likewise,Achatet al.(2016)found that the capacity of fine-textured soils to release available P was higher than that of coarse-textured soils.The sum of soil base cation concentrations was also a critical factor regulating the distribution of P into different fractions, in particular in the most accessible fractions (H2O-P and NaHCO3-P)for plant and microbe nutrition (Figs.2 and 3).Thus, in general,soils with a greater base cation concentration have a greater capacity to redistribute P from struvite into the most available fraction(H2O-P).In acidic soils,trivalent Fe and Al can be found at high concentrations in the soil solution,while in more alkaline soils, the dominant cations are Ca and Mg(Hinsinger,2001).The Fe and Al phosphates tend to have a greater solubility with increasing pH values,while the solubility of Ca phosphates decrease with increasing pH(Hinsinger,2001).Further,since large parts of the soluble P is in the form of Mg-P and Ca-P complexes in more alkaline soils (Ruiz, 1992), it is logical that the concentration of soil cations can be an important determinant of the struvitederived P fractions.The random forest analysis demonstrated the important role of Fe and Al cations in several P fractions(Fig.2).

As depicted by the random forest analysis, compared to the significant importance of texture and cation concentrations in determining struvite-derived P fractions in soil,pH only showed a significant correlation with NaHCO3-P(Fig.2),even though pH was suggested as a critical factor determining P solubility.The solubility of P-precipitated salts such as struvite has been suggested to be dependent on soil pH,with a minimum solubility at pH 9—11, however,some studies have sparked controversy on this topic(Hiltet al.,2016;Huygens and Saveyn,2018).Our results,based on a global survey of soils,did not support the finding that more alkaline soils increased the potential solubilization of struvite P.This is a valuable finding considering that soils with contrasting pH values were included.To some extent, this observation agrees with results from Cabezaet al.(2011)who indicated that struvite-like products can be useful in soils with a wide range of pH values.Our study clearly pointed out that other soil properties, such as soil cation concentrations and texture,might play a more significant role in determining the solubility of struvite-derived P in soil at least in the short-term,and that pH was related only to the NaHCO3-P fraction(Fig.3).Recent studies have suggested that the disparity of struvite dissolution in soils with different pH was caused by adsorption of dissolved P into clay minerals,thus highlighting the critical role of pHvstexture in struvite solubilization(Guet al.,2021).We found positive correlations between soil pH and the NaHCO3-P fraction(Fig.3).Indeed,NaHCO3-P is usually utilized in calcareous and alkaline soils(Hinsinger,2001).

Relationships between microbes and struvite-P redistribution into different fractions

Soil microbial community is diverse,and the number of microbes potentially involved in P solubilization must also be very large.Furthermore,there is a lack of model approaches to understand how microbes affect P mobilization, which is fundamental not only for plant productivity but also for C and N cycles.The relationship between either microbial biomass or basal respiration and P in different fractions could not be unequivocally established and was dependent on each fraction (Fig.3).Soil microbial biomass is more affected by soil C and N availability than by P availability at continental and global scales(Bastidaet al.,2021;Smithet al.,2021).Phosphatase activity,one of the most widely studied extracellular enzyme activities (Nannipieriet al.,2011)was a critical driver of the most accessible P fraction(H2O-P)in struvite-amended soils(Fig.3).Yet,the relative importance of this enzyme activity was lower than that of other edaphic properties such as the sum of base cation concentrations.Soils with greater phosphatase activity had a higher capacity to mobilize P from struvite into the soluble fraction.The P starvation regulon in bacteria has been linked to the production of phosphatase,indicating that phosphatase activity is induced by the low availability of PO-4in soil(Schneideret al.,2019).Our results indicate that soils with a naturally greater capacity to produce phosphatase also have a greater potential to mineralize P from struvite.

The negative correlations between P in the most accessible fractions(H2O-P and NaHCO3-P)and the biomass of ectomycorrhizal fungi indicate that this type of mycorrhiza,which are also known for their capacity to take up/dissolve P,are outcompeted by other microbes that take advantage of available P(Fig.3).Recent studies have shown that despite the recognized role of arbuscular mycorrhizal fungi in the use and mobilization of P(Di Tomassiet al.,2021),the inoculation with these microbes did not contribute to plant P uptake from struvite in a chamber assay with rye(Schwalbet al.,2021).Nevertheless,we should stress that the incubations in the present study were conducted with sieved soil,which may have altered the ecological function of fungi(including mycorrhiza)in P solubilization.

Furthermore, correlation between the relative abundance of microbial populations and different P fractions in struvite-amended soil was evaluated to generate information about the potential involvement of autochthonous microbial populations in P solubilization in a large variety of soils.For example,the unique population associated with P solubilization from the most recalcitrant fraction(HCl-P)was Acidimicrobiales,which highlights the limited access to P in the most recalcitrant sources such as minerals.In contrast,several populations were associated with P solubilization in the NaHCO3-P fraction,such asSphingomonasandMycobacterium,which have been previously proposed as P solubilizers(Vassilevet al.,2012;Alemnehet al.,2021).Similarly,several actinobacterial populations(i.e.,Pseudonocardia,Rubrobacter,andSolirubrobacter)were positively correlated with these P fractions (Fig.4).The role of some of these actinobacterial populations in P solubilization has been also previously described(Vassilevet al., 2012; Borah and Thakur, 2020).Correlations were observed for different fungal populations(Fig.4).For example,Aureobasidium(Ascomycota) has been suggested to possess a certain capacity for P solubilization (Mestreet al., 2016), and our study suggests that this population may play an important role in P solubilization from struvite in different soils.Regarding the NaOH-P fraction,several microbial populations showed a positive correlation with the NaOH-P fraction.For example,Bradyrhizobium,which is a recognized bacterium with plant-growth promotion capacity,has also been identified as a critical organism in P solubilization in the rhizosphere(Alemnehet al.,2021;Huanget al., 2021).Our work also highlights the capabilities of certain populations,such asRhodoplanes,to solubilize P in all soil fractions(except HCl-P).

CONCLUSIONS

Our results indicate that struvite-derived P was quickly redistributed into soil fractions and became more accessible to plants and microbes.Thus,struvite can be used as a P-rich fertilizer.This P redistribution in soil was driven by a complex suite of chemical, physical, microbial, and environmental properties that varied across soils.This study evidenced the important roles of soil texture and base cation concentrations in determining the distribution of struvite-derived P into accessible fractions.We cannot dismiss the findings that these properties overcome the previously recognized role of soil pH in struvite-derived P solubilization.Among biological properties,soil phosphatase was a significant driver of struvite-derived P redistributed into the water-soluble fraction.This study statistically linked autochthonous microbes from a variety of soils to struvite-derived P solubilized into different fractions and highlighted their potential role as P solubilizer.Overall, this study provides predictive hints for the most appropriate conditions under which struvite application could provide efficient P fertilization.

ACKNOWLEDGMENTS

Antonio Ruiz-Navarro is thankful for the financial support by the Fundación General CSIC, Spain (Programa ComFuturo).This research is part of the project PID2020-114942RB-I00 funded by MCIN/AEI//10.13039/5011000 11033.Manuel Delgado-Baquerizo was supported by a project from the Spanish Ministry of Science and Innovation (No.PID2020-115813RA-I00) and a project of the Fondo Europeo de Desarrollo Regional(FEDER)and the Consejería de Transformación Económica, Industria,Conocimiento y Universidades of the Junta de Andalucía(FEDER Andalucía 2014-2020 Objetivo temático “01—Refuerzo de la investigación, el desarrollo tecnológico y la innovación”,ANDABIOMA,No.P20_00879).Concha Cano-Díaz was supported by a postdoctoral scholarship as part of the FCT-funded project“Soil Ecosystems in the XXI Century:Drivers,Conservation and Future Scenarios”(No.FCT-PTDC/BIACBI/2340/2020)led by IPVC,Portugal.

SUPPLEMENTARY MATERIAL

Supplementary material for this article can be found in the online version.