Introduction

Feeding a global population that has recently surpassed 8 billion represents a significant challenge, necessitating the continuous maintenance of agricultural production, a task heavily reliant on the use of fertilizers. Among these essential nutrients, phosphorus (P) stands out as the second most limiting factor. Indeed, tons of P, mainly in the form of orthophosphate ions (HPO42− and H2PO4−), have extensively been applied to crops. Looking ahead, global demand for P is projected to reach approximately 25·106 Mt P yr−1 by the year 2050 (Zou et al. 2022). However, the future availability of P fertilizers is decidedly uncertain. Currently, P is predominantly extracted from phosphate rock deposits, with nearly 50 Mt of P2O5 utilized annually for chemical fertilizers (Geissler et al. 2018). The main supplier of phosphorus fertilizers, rock phosphate mines, is expected to experience depletion in the coming decades, as highlighted by Brownlie et al. (2021), leading to what is commonly referred to as the "P crisis". However, recent discoveries of new mines have presumably extended the availability of P for several more decades. Despite this, embracing the reuse of waste and byproducts as fertilizers not only supports a circular economy but also helps mitigate concerns associated with contaminants linked to phosphate rock.

In this context, securing fertilizers rich in P, as well as in nitrogen (N), is essential to guarantee crop production. Meeting future P demands will hinge on efficiently recycling waste materials that contain ample P and converting them into effective and sustainable fertilizers capable of competing in the market against traditional chemical P fertilizers. One promising approach for addressing this challenge is the recycling of waste materials rich in P, such as sewage sludge obtained from wastewater treatment plants. Sewage sludge boasts a relatively high P content, both in the form of inorganic and organic P. When added to soil, sewage sludge not only aids in P recycling but also increases the content of organic carbon and nitrogen, microbial biomass and microbial activity in soil, thus improving soil fertility and functionality (Bastida et al. 2008). In the case of soil organic matter, this is particularly important in southern countries that use to have very low levels of organic matter in soil. It should be emphasized some of the detrimental effects of sewage sludge that necessitate special legislation to regulate its use. The Council Directive of 12 June 1986 on the protection of the environment, particularly the soil, when sewage sludge is used in agriculture (86/278/EEC), addresses these concerns. Sewage sludge generally contains high levels of heavy metals and other potentially hazardous compounds, such as bacteria, pathogens, and emerging contaminants, which pose risks to human, animal, and plant health. Therefore, it is essential for sludge to undergo processing and stabilization to ensure its application under the most optimal conditions. Treatments such as anaerobic digestion and thermal stabilization are commonly employed to reduce pathogen content and enhance safety.

However, it is worth noting that our understanding of the effects of organic amendments, including sludges, on soil P availability and related processes is still limited, with most studies focusing on carbon (C) and N dynamics (Johnston and Poulton 2019). In addition to sewage sludge, another potential source of P is meat-bone meal (MBM), which typically contains 3–6% total P in dry weight, along with notable amounts of total N (8%) and calcium (10%) (Römer and Steingrobe 2018). MBM is considered a slow-release fertilizer due to its mineral nature, with soil pH, texture, and cation content playing a significant role in P availability from mineral sources (Cabeza et al. 2011; Ruiz-Navarro et al. 2023). Nevertheless, there is evidence suggesting that MBM can provide available inorganic P over the long term, especially when used in combination with digested biosolids (Bøen ab and Knapp Haraldsen 2013). Therefore, further research efforts are needed to explore the potential applications of MBM as a soil fertilizer.

The availability of P from fertilizers is not solely determined by the chemical nature of the source but also by the diversity and activity of belowground soil microbes. The soil microbial community plays a pivotal role in maintaining soil functionality, and fertility by participating in various nutrient-provision processes (Delgado-Baquerizo et al. 2016; Fan et al. 2023). Microbes, for instance, are responsible for making P available (i.e. mobilizing) from both mineral and organic sources. When facing a scarcity of P, microorganisms upregulate the expression of genes involved in organic P mineralization and inorganic P solubilization, among others (Siles et al. 2022). Therefore, any fertilization approach should carefully consider evaluating the activity, biomass, and diversity of the soil microbial community. Rhizosphere is a hotspot for microbial activity and usually harbours high abundance of microbial biomass (Kuzyakov and Blagodatskaya 2015). Therefore, nutrient cycling is expected to occur more intensively in this soil compartment. Further, the community composition of the soil is expected to change with the developmental stage of plant (phenology), among other reasons, because the nutritional requirements of crops change with their development, which is expected to be concomitant with taxonomic changes in rhizosphere microbial community (De Zutter et al. 2022). Plant phenology can influence the diversity and composition of the soil microbial community through various mechanisms, such as altering the quantity and composition of root exudates and hormones (Chialva et al. 2020; Igwe et al. 2021; Molefe et al. 2023). In turn, belowground microbial communities can determine the fitness of aboveground plants and traits (Lau and Lennon 2011) and even plant phenology. For instance, in rice, the root microbiota is more dynamic during the vegetative phase of plant growth but becomes more stable during the rest of the cycle (Edwards et al. 2018). Similarly, the release of root-derived organic C (i.e. exudates) shows significant variations during different wheat growth phases, with a stronger correlation observed with the bacterial community than the fungal community (Chen et al. 2019). Other studies have highlighted that, while the abundance of some microbial taxa can change over time with plant development, core microbiota is consistently maintained (Trivedi et al. 2020). Yet, how plant-microbial associated dynamics are influenced by P-rich byproducts is not known.

In this context, we aim to: i) determine whether substituting conventional P fertilizer with alternative sources (MBM and/or sludges) can maintain wheat yield; ii) assess the impact of MBM, sludge, and their combination on the biomass, activity, and diversity of the soil microbial community; and iii) evaluate the impact of plant phenological stages on the soil microbial community’s responses to various fertilization strategies. We hypothesize that applying sludge and MBM can partially replace the excessive use of conventional NPK fertilizers and promote soil sustainability and ensure wheat yield. Specifically, we expect that sludge application will enhance the biomass and activity of the soil microbial community, leading to significant changes in the structure and diversity of the microbial population. In contrast, we expect that MBM will have a relatively minor impact on the overall microbial community, as this P source is rapidly immobilized in the soil and primarily affects a reduced subset of microorganisms, namely P solubilizers. Furthermore, we anticipate that the combination of sludge and MBM will synergistically boost microbial community development and P solubilization due to increased microbial activity and biomass resulting from sludge application. Finally, considering that plant P requirements vary during different phenological stages (Romer and Schilling 1986; Bojtor et al. 2021), we expect that shifts in community composition and P solubilizers will be influenced by plant phenological stages. These hypotheses will be tested in a wheat agroecosystem, given that wheat is one of the top three most-produced crops worldwide, with 737 million metric tons produced in 2020 (Food and Agriculture Organization 2022).

Material and methods

Experimental design and sampling

The agronomical study was conducted in the experimental field of ITAP (Barrax, Albacete (SE Spain) (39°02′57.0"N 2°04′54.1"W) with a Mediterranean semiarid climate. In this area, 16 plots of 15 m2 each were established. For each of them, a subplot of 7.5 m2 was left for crop harvesting, while the other 7.5 m2 subplot was used for soil samplings. Wheat (var. Califa Sur) sowing occurred in 5-February-2021. Plots were separated by 1 m corridor.

We established a completely randomized blocks design of 4 treatments with 4 replicates for each treatment. The treatments were applied in order to achieve the recommended dose for wheat, which was 190 UFN, 105 UFP and 105 UFK. Application rates were calculated according to the P content of each material. All P and K was applied in base fertilization, and then all treatments were supplemented with urea, ammonium nitrosulfate and calcium ammonium nitrate (Table S1) for top dressing fertilization when required to accommodate nutritional wheat demands.

The 4 treatments consisted in: i) traditional fertilization with NPK (Trad); ii) organic fertilization with thermo-stabilized sewage sludge from anaerobic digestion (2391 kg ha−1) (S); iii) mineral fertilization with meat-bone meal (342 kg ha−1) (MBM); and iv) organo-mineral fertilization consisting in a combination of sewage sludge plus MBM (1195 kg ha−1 of sludge plus 171.2 kg ha−1 of MBM) (SMBM). Base fertilization with the different treatments was applied at the beginning of February (2-Feb 2021). Sludge contained 31% of organic C, 5.4 of N and 1.92% of P. MBM contained 13% of P and 3.8% of N. The complete chemical composition of sludge and MBM appears in Table S2. No major differences exist in terms of the total amount of P, N and K applied in the different treatments, yet the distinct nature of fertilizers might affect plant and microbial responses. The water demand of the wheat used is 335 mm, and effective precipitation as 210 mm, 125 mm were added by sprinkler irrigation. In the subplot used for soil sampling, three different phenological stages were considered: germination (G, end February 2021), flag leaf (FL, mid March 2021) and maturation (M, end of July 2021). Rhizospheric soil samples were obtained from the soil attached to the root at a depth of 0–15 cm. For each subplot, rhizospheric soil from six plants was collected. Then, the six samples were combined to compose a representative sample per subplot. In total, the study consisted of 48 samples; 4 treatments × 3 sampling phenological stages × 4 replicates. Following field sampling, soils were sieved (< 2 mm). A portion of the samples was frozen immediately at -20ºC for fatty acid analysis and DNA extraction, other portion was conserved at 4ºC for enzyme analyses, and other portion was air-dried for chemical analyses.

Grain and plant analyses

Wheat was harvested at the end of July 2021, with a plot-sized combined harvester in the subplot dedicated to this end. The production of each subplot was calculated by weighing the produced grain. A part of this grain was stored 4 ºC for nutritional analyses. For plant biomass analysis, after harvesting, six plants per plot were sampled and merged to obtain a sample of vegetal material per plot. Plant weighing was used for assessment of plant biomass yield. This plant material was used to analyze the nutritional status of plants.

Leaf material was dried at 60 ºC for 48 h in an oven and subsequently milled. The nutrient content of plant material was then determined after HNO3–HClO4 digestion, using an ICP-OES spectrometer (ICAP 6500 DUO; Thermo-Scientific, Waltham, MA, USA) at the analytical services lab of the CEBAS-CSIC.

Soil physicochemical properties

Soil pH and EC were measured in a soil:distilled water extract (1:5, w:v). Briefly, 8 g of soils were placed in Falcon™ tubes. Then 40 ml of water were added and the tubes were placed in an orbital shaker for 2 h. After that EC was measured. The tubes were then centrifuged at 8500 rpm during 15 min, and filtered using a Whatman Ashless paper filter (110 mm diameter) in order to measure pH. Soil water-soluble C (WSC) and N (WSN) were measured in a C/N analyzer for liquid samples (Multi N/C 3100, Analytic Jena, Germany). Soil and leaf total inorganic carbon (TC), organic carbon (OC) and total nitrogen (TN) were analysed using an Elemental Analyzer (C/N Flash EA 112 Series-Leco Truspec). Total P was analyzed after wet acid digestion with hydrofluoric acid in an ICP-OES (ICAP 6500 DUO; Thermo-Scientific, Waltham, MA, USA). Soil and grain element contents were determined as the nutrient content of plant material. Olsen P was extracted using 0.5 M NaHCO3 and determined using a ICP-OES.

Soil enzyme activities

We selected enzymes related to C, N and P cycles. In semiarid soils, organic C is very scarce, and therefore β-glucosidase is widely used as soil quality indicator (Bastida et al. 2006), as it plays a key role in carbon cycle. It ultimately breaks down cellobiose into simple sugars which are accessible for microbes (Ferraz-Almeida et al. 2015). Alkaline phosphatase and urease are involved in P and N cycles. The β-glucosidase activity (EC 3.2.1.21) was assayed as described in Eivazi and Tabatabai (1988) by using 0.025 mM p-nitrophenyl β-D-glucopyranoside (Sigma-Aldrich) as substrate and 0.1 M modified universal buffer, pH 6.5. Alkaline (EC 3.1.3.1) phosphatase activity was assayed as described by Eivazi and Tabatabai (1977) using 0.1 M modified universal buffer at pH 11 and 0.115 M p-nitrophenyl phosphate (Sigma-Aldrich) as substrate. For both activities, after 1 h of incubation at 37 ºC, the p-nitrophenol released during the incubation was extracted and colorimetrically determined at 400 nm. Urease activity (E.C. 3.5.1.5) was analysed through the method developed by Kandeler and Gerber (1988), using an aqueous urea solution at 0.48% as substrate and 0.06 M borate buffer at pH 10. The NH4+ produced was extracted with 7.4% KCl and quantified by means of a modified indophenol reaction.

Analysis of fatty acid methyl esters (FAMEs)

Fatty acid methyl esters (FAMEs), hereafter fatty acids, were extracted from 3 g of soil according to (Schutter and Dick 2000). The fatty acids were analysed with an 8860 GC System gas chromatograph (Agilent Technologies, Santa Clara, USA) equipped with a flame ionization detector (FID), using a DB–FastFAME capillary column (30 m × 0.25 mm ID × 0.25 µm film) (Agilent Technologies), with helium as the carrier gas. The conditions were: an initial temperature of 80 °C for 1.5 min, then an increase to 160 °C with a ramp of 40 °C min−1, then to 167 °C at 0.5 °C min−1, then to 200 °C at 30 ºC min−1 and finally to 230 °C at 4 °C min−1. All the fatty acids mentioned in this article are described according to the standard nomenclature in Vestal and White (1989). The fatty acids i15:0, a15:0, i16:0, c16:1ω9c, 10Me16:0, i17:0, cy17:0, 10Me18:0 and cy19:0 were considered representative of the bacterial biomass and the fatty acids 18:2 ω6,9t and 18:2 ω6,9c indicators of the fungal biomass. The unspecific microbial fatty acids c14:0, c15:0, c16:0, c17:0, c18:0, c20:0, along with those specific for bacteria and fungi, were used to calculate the total microbial biomass (Joergensen 2021).

Soil DNA extraction

Total DNA from each sample (n = 64) was extracted from 250 mg of soil fresh soil using the Dneasy PowerSoil Pro Kit (QIAGEN) following the manufacturer’s instructions. The DNA quality was spectrophotometrically checked by NanoDrop (Thermo Fisher Scientific Inc.) based on the absorbance ratios A260/A280 and A260/A230 by quantification with the DeNovix dsDNA High Sensitivity Assay on a DS-11 DeNovix device.

Quantitative PCR analysis

Quantitative PCR (qPCR) analysis was conducted to determine the abundance of the phoD, gcd and pqqC genes in the soil DNA extracts. The pairs of primers ALPS-F730/ ALPS-R1101 (Sakurai et al. 2008), gcd-F/gcd-R and pqqC2-F/pqqC2-R (Zheng et al. 2018) for phoD, gcd and pqqC genes, respectively, were used. Single qPCR reactions were carried out containing 10 μL PerfeCTa SYBR® Green (Quantabio), 1 μL each primer (10 μM) (Genewiz), 2 μL DNA (3 ng μL−1) and 6 μL H2O, under the following thermal conditions: 95 °C for 3 min followed by 40 cycles of 95 °C for 10 s, 60 °C (phoD)/63 °C (gcd)/55 °C (pqqC) for 20 s, and 72 °C for 30 s. A QuantStudio 1 system (Applied Biosystems) was used to perform the quantitative PCR. In all the cases, after amplification reactions, melting curve and gel electrophoresis analyses were conducted to confirm that the amplified products had the appropriate size. Copy numbers for each gene were calculated using a regression equation for each assay relating the cycle threshold (Ct) value to the known number of copies in the standards of a plasmid standard curve containing the appropriate target gene. Construction of standard plasmid curves was done as previously reported (Siles and Margesin 2016).

Amplicon sequencing analysis and bioinformatics processing of the sequences

Prokaryotic (bacterial and archaeal) communities were characterized in terms of taxonomic diversity and composition by amplifying the V4 hypervariable region of the16S rRNA gene using the primers 515F (5’-GTGCCAGCMGCCGCGGTAA- 3’) and 806R (5’-GGACTACHVHHHTWTCTAAT- 3’). Fungal communities were investigated through the amplification of the internal transcribed spacer (ITS) 2 region with the primers gITS7 (5’-GTGARTCATCGARTCTTTG-3’) and ITS4 (5’-TCCTCCGCTTATTGATATGC- 3’). PCR mixtures and thermal conditions and library preparation for sequencing were as previously described (Martinović et al. 2022). Paired-end (PE) sequencing (2 × 300) was performed on an Illumina MiSeq sequencing platform (Illumina, USA).

Bioinformatic processing of the sequences was conducted using the USEARCH pipeline and UPARSE-out algorithm (Edgar 2013). Briefly, PE sequences were firstly merged with the command -fastq_mergepairs. Then, reads were quality-filtered allowing a maximum e-value of 1.0, trimmed (to 250 and 240 bp (base pair) for prokaryotic and fungal libraries, respectively), dereplicated, and sorted by abundance (removing singletons), prior chimera detection and determination of ASVs (amplicon sequence variants, zero-radius operational taxonomic units according to USEARCH pipeline) for prokaryotic community and OTUs (operational taxonomic units) at 97% sequence identity for fungi. Finally, original sequences were mapped to ASV or OTUs at the 97% identity threshold to obtain one ASV table for the prokaryotic community and OTU table for the fungal community. The taxonomic affiliation of each ASV/OTU was obtained using the -sintax algorithm against the RDP 16S rRNA training set 18 for 16 s rRNA gene sequences and utax_reference_dataset_25.07.2023 for ITS2 sequences with a 50% confidence threshold in both cases. The sequencing depth across samples was normalized to 4846 reads for prokaryotic community and to 2345 reads for fungal community using the -otutab_rare command and normalized ASV/OTU tables were used for downstream analyses.

The raw prokaryotic and fungal sequences associated with this study were deposited in the GenBank SRA database under BioProject accession number PRJNA1060287.

Statistical analyses

Normality and heteroscedasticity of data were tested by the Shapiro–Wilk (shapiro.test function, stats R package ver. 4.1.1) and Levene tests (leveneTest function, car R package ver. 3.1–1), respectively. In case that one of these conditions was not met, the values were transformed using the natural logarithm. Two-way ANOVA (analysis of variance) was applied to determine whether there were significant (p < 0.05) differences between treatments for the different variables measured (physicochemical, enzyme activity, soil content and plant responses) using the aov function of the R package stats ver. 4.1.1 (R Core Team). Tukey’s HSD (honest significance difference) test was used for pairwise comparisons of treatments within each phenological stage using the function TukeyHSD and the R packages multcomp ver. 1.4–23 (Bretz et al. 2016) and multcompView ver. 0.1–9 (Graves et al. 2015) were usedNMDS (non-metric multidimensional scaling) analysis based on Bray–Curtis dissimilarities was used as ordination method taking advantage of the metaMDS function in the R package vegan ver. 2.6–4 (Oksanen et al. 2013). The significance of the variations in the prokaryotic and fungal communities as a function of the treatment and phenological stage was tested using two-way PERMANOVA (permutational analysis of variance) with Bray–Curtis dissimilarities after 9,999 permutations using the R package vegan ver. 2.6–4 (Oksanen et al. 2013). SIMPER (similarity percentage) analysis was applied to assess the average dissimilarity between treatments and phenological stage and to identify the prokaryotic ASV and fungal OTUs responsible for the observed dissimilarities using Bray–Curtis dissimilarities with the software PAST ver. 4.10 (Hammer et al. 2008). Data visualizations were conducted using the R packages ggplot2 ver. 3.4.1 (Wickham 2009), pheatmap ver. 1.0.12 (Kolde 2015), and CorelDRAW ver. 2020.

Results

Plant responses

Replacement of traditional fertilization with MBM and SMBM did not have significant consequences on crop yield. However, sludge-fertilized soils showing the lowest yield (Table 1). This may be due to the increase of trace metals, which may cause a phytotoxic effect on plant development. Treatment significantly affected the content of chromium (Cr) in soil (p = 0.00468) but there was no effect of phenological stage. In the case of lead (Pb) phenological stage significantly influenced the content in soil (p = 1.48e -11), but there were no differences between treatments (Table S3). In the case of plant biomass, there were not significant differences between treatments. The contents of P and K in grain were greater in S and SMB than in the traditional fertilization and in the soil amended with MBM (p = 0.0048 and 0.010 respectively) (Table 1). The limits of trace metals in grain, according with the Commission Regulation (EU) 2023/915 of 25 April 2023 on maximum levels for certain contaminants in food Regulation (EC) No 1881/2006 is 0.2 mg/kg of cadmium. According with our results (Table 1), the presence of Cd in the grain in our experiment is below 0.01 mg/kg.

Table 1 Crop yield, biomass and macronutrients content in the grain at the maturation stage
Full size table

There were significant differences between treatments for the content of trace metals in grain (Table 1). For instance, Mn and Zn in grain resulted greater in S and SMBM than in the traditional and MBM treatments (p = 0.04 and 0.016 respectively).

Soil physicochemical parameters

EC was significantly influenced by treatment and phenological stage (p < 0.001), with the SMBM treatment showing the lowest values, especially at germination stage (Fig. 1a). Soil pH was significantly influenced by phenological stage (p = 0.0001), but not by treatment (Table S4). WSC and WSN contents were influenced by treatment (p < 0.001 and p = 0.031 respectively), phenological stage (p < 0.001) and their interaction (p = 0.026 and p =  < 0.001 respectively). The highest WSC content was observed in S and SMBM treatments which increased 46.30% and 25.67% compared to Trad at germination stage. Significant differences between treatments were particularly evidenced at the germination. WSN content decayed in the flag leaf and maturity stages in comparison with germination. The content of Olsen P was influenced by both treatment (P < 0.001) and phenological stage (p < 0.001). Soil amended with sludge had the highest Olsen P content in comparison with the other treatments and was followed by SMBM (Fig. 1d). In the S treatment, Olsen P content was lower in flag leaf than in germination and maturity stages and it increased 42.17% respect to Trad in germination and 24.901 in flag leaf. Treatment did not affect the content of total organic C and total N in the maturation stage (P = 0.38 and P = 0.64, respectively) (Table S5).

Fig. 1
figure 1

Box plots comparing (a) electrical conductivity (EC), (b) water soluble carbon (WSC), (c) water soluble nitrogen (WSN) and (d) Olsen phosphorus (P) between the soil amended with traditional NPK fertilization (Trad), sludge (S), meat bone (MBM) or the mixture sludge and meat bone (SMBM) at the germination (G), flag leaf (FG) and maturation (M) phenological stages of wheat. P-values of two-way ANOVA for the factors treatment (Tr) and phenological stage (PS), and their interaction are shown at the bottom of each figure panel. Different lowercase letters above each box denote significant differences among treatments within each phenological stage according to Tukey’s HSD test. The boxes represent the interquartile range (IQR) between the first and third quartiles (25th and 75th percentiles, respectively), and the vertical line inside the box defines the median. Whiskers represent the lowest and highest values within 1.5 times the IQR from the first and third quartiles, respectively. Dots represent outliers

Full size image

Enzyme activities, microbial biomass and copy numbers of functional genes

Alkaline phosphatase and β-glucosidase activities were influenced by treatment (p < 0.001) and phenological stage (p < 0.001) and their interaction (p < 0.001). In contrast, urease activity was only affected by the phenological stage (p < 0.001) (Fig. 2). Phosphatase activity was higher in S and SMBM at germination and flag leaf stages, which showed significant differences among treatments. However, phosphatase activity decreased in the maturation and there were no differences between treatments in this stage. Overall, β-glucosidase was also greater in S and SMBM. Significant differences for β-glucosidase were only observed in germination (S and SMBM increased 44.98 and 21.62 respectively compared to control) (Fig. 2b). Overall, the lowest values of this enzyme activity were recorded in the soil with traditional fertilization.

Fig. 2
figure 2

Box plots comparing activity of the (a) alkaline phosphatase, (b) β-glucosidase and (c) urease enzymes between the soil amended with traditional NPK fertilization (Trad), sludge (S), meat bone (MBM) or the mixture sludge and meat bone (SMBM) at the germination (G), flag leaf (FG) and maturation (M) phenological stages of wheat. P-values of two-way ANOVA for the factors treatment (Tr) and phenological stage (PS), and their interaction are shown at the bottom of each figure panel. Different lowercase letters above each box denote significant differences among treatments within each phenological stage according to Tukey’s HSD test. The boxes represent the interquartile range (IQR) between the first and third quartiles (25th and 75th percentiles, respectively), and the vertical line inside the box defines the median. Whiskers represent the lowest and highest values within 1.5 times the IQR from the first and third quartiles, respectively. Dots represent outliers

Full size image

The content of total, bacterial and fungal fatty acids was significantly influenced by treatment (P = 0.004, 0.004 and 0.003 respectively), phenological stage (p < 0.001) but not their interaction. The highest bacterial and fungal content was found in the maturation stage (Fig. 3). During the three phenological stages, the S treatment showed higher values of total, bacterial and fungal biomass, followed by the SMBM treatment. However, differences between treatments were only significant for S in the flag leaf stage (an increase of 29.18% compared to Trad treatment) (Fig. 3). The lowest fungal content was recorded in the traditional fertilization treatment (Trad). In comparison with the other treatments, the S showed the greatest fungal content, in particular in the germination and wheat flag leaf stage. Fungal-to-bacterial biomass was not influenced by treatment nor by phenological stage.

Fig. 3
figure 3

Box plots comparing (a) total, (b) bacterial and (c) fungal biomass and the (d) fungi to bacteria ratio (F/B) between the soil amended with traditional NPK fertilization (Trad), sludge (S), meat bone (MBM) or the mixture sludge and meat bone (SMBM) at the germination (G), flag leaf (FG) and maturation (M) phenological stages of wheat. P-values of two-way ANOVA for the factors treatment (Tr) and phenological stage (PS), and their interaction are shown at the bottom of each figure panel. Different lowercase letters above each box denote significant differences among treatments within each sampling time according to Tukey’s HSD test. The boxes represent the interquartile range (IQR) between the first and third quartiles (25th and 75th percentiles, respectively), and the vertical line inside the box defines the median. Whiskers represent the lowest and highest values within 1.5 times the IQR from the first and third quartiles, respectively. Dots represent outliers

Full size image

The gene copy numbers of phoD, pqqc and gcd were not affected by treatment. Yet phenological stage influenced these values in the case of phoD (p = 0.0046) and pqqC (p = 0.0002), with the greatest abundances for these genes being observed in the maturation stage (Table S7).

Diversity and composition prokaryotic and fungal communities

Prokaryotic richness was lower in the traditional fertilization in comparison with S, MBM and SMBM treatments at the maturation stage. However, these differences were not observed in germination and flag leaf stages (Fig. S1). Fungal richness was not different across treatments in any of the phenological stages.

Although the ASV-based structure of the prokaryotic community was significantly affected by the treatment factor, the effect of the phenological stage was much stronger (Fig. 4a). In fact, the main ordination factor of the samples in the NMDS was the phenological stage (Fig. 4a). Pairwise PERMANOVA comparisons demonstrated that the significance of the factor treatment was mainly due to differences between the traditional and S treatments (Table S8). SIMPER analysis quantified in 44.4% the differences across the 48 analyzed samples. At phylum level, Firmicutes, Verrucomicrobia and Chloroflexi was significantly affected by treatment (Fig. 5a; Table S9). The relative abundance of Firmicutes tended to increase in S and SMBM compared to the traditional and MBM treatments, especially in flag leaf and maturation stages. The relative abundance of Verrucomicrobia showed a temporal dynamic (Fig. 5, Table S9). In the germination stage, the abundance of this phylum was lower in the traditional treatment compared with the rest, while it was higher in S, MBM and SMBM compared to traditional in the maturation stage. At the class level, phenological stage influenced significantly the relative abundance of a wide number of populations, but the effect of treatment was only significant for Thermoleophilia, Planctomycetacia, Clostridia and Gp4 (Fig. 5, Table S10).

Fig. 4
figure 4

a Prokaryotic and (b) fungal NMDS ordination of soil samples amended with traditional NPK fertilization (Trad), sludge (S), meat bone (MBM) or the mixture sludge and meat bone (SMBM) at the germination (G), flag leaf (FG) and maturation (M) phenological stages of wheat. P-values of the two-way PERMANOVA for the factors treatment (Tr) and phenological stage (PS), and their interaction are shown for each community

Full size image
Fig. 5
figure 5

Taxonomic composition of the prokaryotic community at (a) phylum and (b) class levels of the soil amended with traditional NPK fertilization (Trad), sludge (S), meat bone (MBM) or the mixture sludge and meat bone (SMBM) at the germination (G), flag leaf (FG) and maturation (M) phenological stages of wheat according to 16S rRNA gene metabarcoding

Full size image

Further insights into the treatment- and phenological stage-mediated variations in prokaryotic taxonomy were obtained by investigating the relative abundance of the top-40 ASV contributing the most to the differences between treatments and phenological stages (Fig. 6). Among these, Micrococcaceae (Actinobacteria), Nitrososphaera, Gp16 and Blastococcus (Fig. 6) highlighted. Micrococcaceae and Blastococcus were particularly abundant in the flag leaf stage, and particularly in the SMB treatment. Nitrososphaera was abundant in germination and flag leaf stages, and was particularly abundant in SMB in the germination.

Fig. 6
figure 6

Heatmap showing the relative abundance of the top-40 prokaryotic ASV contributing the most to the differences between treatments according to SIMPER analysis. ASV are ordered in decreasing contribution to differences between treatments. Percentage to the right of each ASV indicates the percent dissimilarity contributed by each ASV. The best taxonomic hit for each ASV is provided. Color legend and scale (%) are provided in the figure. Trad = traditional fertilization, S = sludge, MBM = meat bone, SMBM = mixture of sludge and meat bone. G = germination, FL = flag leaf, M = maturation

Full size image

OTU-based structure of the soil fungal community significantly was significantly affected by treatment and plant phenological state (but not by their interaction), although the effect of the phenological stage was much stronger. This was clearly evidenced by NMDS ordination of samples (Fig. 4b). Differences between the 48 samples were shown to be of 52.3% by SIMPER analysis. Further analysis revealed that these differences were heavily weighted towards phenological stage and not treatment. In this way, treatment did not influence the relative abundance of fungal phyla, while phenological stage influenced that of Ascomycota, Mucoromycota and Mortierellomycota (Fig. 7a, Table S11). At the class level, treatment influenced the abundance of Eurotiomycetes, Dothideomycetes and Leotiomycetes (Fig. 7b, Table S12). As an example, the relative abundance of Eurotiomycetes was greater in S, MBM and SMBM than the traditional treatment in flag leaf and maturation stages. In the flag leaf stage, the abundance of Dothideomycetes was greater in MBM and SMBM compared to the other treatments. As for the prokaryotic community, the relative abundance of the top-40 fungal OTUs contributing the most to the differences between samples were investigated (Fig. 8). Lecarnoromycetes, Blumeria, Sordariomycetes and Dothideomycetes (Ascomycota), and Rhizopus (Mucoromycota) highlighted (Fig. 8). Blumeria was particularly abundant in the traditional treatment at the flag leaf stage, while Lecanoromycetes was it in the traditional and sludge treatment at the maturation.

Fig. 7
figure 7

Taxonomic composition of the fungal community at (a) phylum and (b) class levels of the soil amended with traditional NPK fertilization (Trad), sludge (S), meat bone (MBM) or the mixture sludge and meat bone (SMBM) at the germination (G), flag leaf (FG) and maturation (M) phenological stages of wheat according to ITS2 metabarcoding

Full size image
Fig. 8
figure 8

Heatmap showing the relative abundance of the top-40 fungal OTUs contributing the most to the differences between treatments according to SIMPER analysis. OTUs are ordered in decreasing contribution to differences between treatments. Percentage to the right of each OTU indicates the percent dissimilarity contributed by each OTU. The best taxonomic hit for each OTU is provided. Color legend and scale (%) are provided in the figure. Trad = traditional fertilization, S = sludge, MBM = meat bone, SMBM = mixture of sludge and meat bone. G = germination, FL = flag leaf, M = maturation

Full size image

Discussion

Meat-bone meal and organo-mineral fertilization maintain wheat yield

In comparison with the traditional NPK fertilization, wheat yield was sustained in mineral (MBM) and organo-mineral (SMBM) treatments. Previous studies have found that organo-mineral fertilizations can sustain grain production in wheat agroecosystems (Chagas et al. 2021). However, wheat yield decreased in the S application treatment in our assay. These results indicate that equaling the total P amounts added, based on plant demands, does not necessarily mean that production is maintained. The detrimental impact of S on crop production may be attributed to various mechanisms, including the presence of phytotoxic compounds, differential content of available P in organic versus mineral materials, and distinct biotic competition between plants and microbes for nutrients. Several studies have reported that S can diminish plant production due to phytotoxic compounds (Hechmi et al. 2020). However, the content of heavy metals in the soil did not significantly differ across treatments. Wheat plants may have absorbed heavy metals through the root system to varying degrees depending on the fertilization treatment. Indeed, we observed that the content of certain heavy metals in the grain, such as Mn and Zn, was higher in S treatment than in the Trad treatment, particularly in the combined organo-mineral treatment. These results indicate a translocation of heavy metals from soil to grain, emphasizing the importance of monitoring sludge amendments in crop fertilization to optimize food security (Seleiman et al. 2020). Nevertheless, we did not analyze other phytotoxic compounds that could potentially harm plant yield. Interestingly, despite lower wheat yield with sludge, the grain was more enriched in P and K, a positive outcome observed in both the S- and SMBM treatments. The increased phosphatase activity in S and SMB during earlier phenological stages (G and FL) might contribute to a later greater P content in the grain. Other studies have reported a greater accumulation of macronutrients in grain after sludge fertilization in wheat (Bose and Bhattacharyya 2008; Mañas and De Las Heras 2018).

The chemical composition of applied products differed. Sludge contains high portion of inorganic P, but also a high proportion of organic P compared to MB (Cabeza et al. 2011). Consequently, distinct microbial pathways, such as the mineralization of organic P versus solubilization, can be stimulated depending on the type of fertilization and thus contributing to different release of available P (Olsen P). Overall, S and SMBM demonstrated greater efficiency in generating available P compared to mineral (MBM) treatment, a phenomenon attributed to elevated phosphatase activity in those treatments. This enzyme, responsible for converting organic P sources into simpler P forms, is known to be stimulated by organic P substrates found in sludges (Bastida et al. 2008; Burns et al. 2013). This increased activity was particularly observed during the germination and flag leaf stages, coinciding with increased wheat P demand. In contrast, phosphatase activity diminished during the maturation stage, a period when soil Olsen P content was higher, suggesting a potential negative feedback effect regulating the Pho regulon responsible for enzyme production (Siles et al. 2022). Despite the greater availability of P resulting from sludge application compared to mineral bone application, this did not result in increased plant biomass or crop yield. Instead, it led to an enhanced P content in the grain in the sludge-treated soil. Other studies have indicated that sludge fertilization can enhance P availability in soil and uptake by wheat, but may also result in a reduced root system, providing a potential explanation for the decreased grain yield observed in our study (Wang et al. 2016).

Interestingly, sludge also contains organic C, leading to an increase in WSC, and provides N for microbial growth (Bastida et al. 2008). The microbial biomass, as estimated by fatty acids, was notably higher in soil treated with sludge at initial and medium stage of the assay. Various studies have reported an increase in microbial biomass with the application of organic amendments, including sludges, in semiarid agroecosystems (Bastida et al. 2008). Therefore, it cannot be ruled out that P might be sequestered into microbial biomass in sludge-fertilized soil, potentially limiting its availability for subsequent plant growth. The allocation of nutrients in microbial biomass has been identified as a mechanism that reduces availability for plants (Rousk et al. 2007). qPCR results indicate that the genetic potential of the soil to mobilize P remained unaffected by the type of fertilizers (organic vs mineral). Thus, the soil’s capacity to generate available P from mineral vs organic subproducts appears to be more closely linked to the current microbial activity and biomass rather than the genetic potential of the microbial community. Besides phosphorus, sludge incorporates many macro and micronutrients which can influence soil microbial responses. For instance, it is known that Fe, Mo, V and Cu can enhance the abundance of free living nitrogen fixating organisms (Dynarski and Houlton 2018). In this sense, among the dominant microbes, there are several microbial groups which have species with such functional capacity and increased their abundance in response to sludge application, such as Firmicutes (Table S5).

The high content of C and N and C/N ratio in sludge, but not in MBM, might explain the stimulatory effect of sludge in the microbial community. However, MBM alone sustained crop production, but did not increase microbial activity, microbial biomass or Olsen P. These findings, coupled with the absence of a stimulatory effect of MBM on the microbial community in the microbial fatty acids and copy number of pqqC and gcd genes, suggest that the increased P availability in MBM was not due the mechanisms studied (alkaline phosphatase and gluconic acid pathway) but by other mechanisms such as the secretion of different organic acids. MBM can increase Olsen P through either its high P content and the stimulation of the abundance and /or the activity of P solubilizers. It is likely that the proportion of P solubilizers is too small, and the differences between treatments are not detectable through fatty acid analysis or gene copy numbers. Sequencing results revealed that the relative abundance of microbial populations contributing most to differences in the MBM treatments was around 2% at best (e.g., some Micrococcaceae). Overall, these results highlight that some members of the rare microbial community play a key role in maintaining P cycle and crop nutrition, but that do not translate into a greater microbial biomass with the applied treatments.

In contrast to the sole application of sludge, the combined application of sludge and MBM not only sustained crop yield compared to traditional fertilization but also elevated microbial biomass and activity. Therefore, this organo-mineral fertilization approach presents itself as an alternative method for recycling waste in a sustainable and cost-effective manner. On the one hand, sludge promoted microbial growth and activity, including enhanced phosphatase activity and the generation of Olsen P. The feasibility of organo-mineral treatments in maintaining cereal production has been also highlighted in previous works (Bastida et al. 2019).

Crop phenology dominates plant-soil microbial interactions in response to fertilization treatments

Our results highlight that crop phenology has a dominant role in assembling soil microbial communities, as well as in modulating microbial biomass and functionality, and that the effect of fertilizers is secondary. Plant phenology can shape the soil microbial community through different mechanisms including the quantity and composition of root exudates, hormones, nutrient demand etc. (Chialva et al. 2020; Igwe et al. 2021). These phenological-associated changes create different niche environments that benefit some microbes over others, and affects microbial recruitment and establishment (Mamet et al. 2022), as well as functionality. In turn, belowground microbial communities can determine the fitness of aboveground plants and traits (Lau and Lennon 2011).

The activity of alkaline phosphatase was higher in flag leaf stage which is a stage, with higher P demand (Romer and Schilling 1986), followed by germination and maturation stages. In contrast, bacterial and fungal biomass –as estimated by fatty acids– was greater in the maturation stage. It is possible to argue that, at this stage, there is less competition between plants and soil microbes for nutrients and therefore a most effective use of nutrient by the soil microbial community translated into a greater microbial biomass. Olsen P content was greater at the maturation stage in comparison to germination and flag leaf stages, but just in the case of S and SMBM treatments. These results, together with the lower wheat production, can indicate that the increased phosphatase activity in earlier stages could generate more bioavailable Olsen P that was not completely utilized by plants at the end of the assay. Further, our results suggest that crop phenology has much more importance than the fertilization treatment in regulating the abundance of functional genes involved in P cycling, with a greater number of copies in maturation stage. These results might indicate that phenology may affect genetic potential involved in the provision of P for plants, while both phenology and fertilization regulate directly the activity of microbes (i.e. phosphatase activity) and P availability.

Moreover, our study provides evidences that, once satisfied nutritional demands with the different treatments, crop phenology more than the chemical form of fertilization, is fundamental to determine community composition, as observed in the NMDS. In the same vein, earlier results demonstrated that microbiome assembly is mainly determined by development stage of wheat rather than by fertilization (Xiong et al. 2021). Further, it is noteworthy to mention that the elevated prokaryotic richness in the maturation phase, as compared with germination and flag leaves stages, aligns with the high microbial biomass and WSC. As previously mentioned, it is conceivable that, since plant nutrient demands are stronger in flag leave stage, the competition between plants and microbes was less pronounced during maturation, thus enabling microbial richness and biomass to increase without biotic competition for nutrient resources (Bastida et al. 2021).

As in the case of community indicators, phenological stage influenced the relative abundance of most populations, while fertilization treatment only affected some of them. (Table S9). Indeed, only Thermoleophilia, Planctomycetacia, Clostridia and Gp4 were affected by the fertilization treatments. Thermoleophilia has been highlighted as a microbial group responding to fertilization and to be correlated with P limitation (Cui et al. 2018). In our study, Thermoleophilia has the greatest abundance in flag-leaf stage and under the MBM treatment, which is the treatment with less P availability, as regards of Olsen P. Further, previous study has shown that P fertilization affects positively the abundance of Planctomycetes (Randall et al. 2019), but here we found that such stimulatory effect only occurred at certain phenological stages. For instance, in the flag-leaf stage the relative abundance of Planctomycetes decreased in MBM and SMBM. Among the most important microbes contributing to explain the differences between treatments we can highlight Micrococcaceae (Actinobacteria). This group has been found to respond to mineral vs organic fertilization (Resendiz-Nava et al. 2023). This was also the case of Blastococcus that increased in the soils fertilized with MBM and SMBM at the flag leaf stage. Importantly, while treatments were set up to fulfill plant P demands, N is an important component of both S and MBM, and therefore microbial responses associate to N cycle can be also observed between treatments. The case of Nitrososphaera is noteworthy, as this is highly abundant archaea in our study explaining differences between treatments. This microbe plays a major role in ammonia oxidation and nitrification in terrestrial ecosystems (Tourna et al. 2011). Nitrososphaera reduced its abundance with sludge application in germination and flag leaf stages, likely given the high content of nitrates of sludge (Castro et al. 2009). These results are in agreement with previous findings that agricultural management may affect the abundance of this important microorganism in N cycling (Zhalnina et al. 2013).

In the case of fungal community, with the exception of some populations, the relative abundances of most of them were quite stable across treatment and phenological stages, which contrasts with changes in the composition of bacterial community. The abundance of Dothideomycetes changed across treatments, but these changes were highly dependent on phenological stage. At the flag leaf, soil treated with MBM and SMBM showed the highest abundance of Dothideomycetes, which is in agreement with the importance of this fungal group in fertilized soils. A notable example of phenological variations was observed in the abundance of Leotiomycetes, which was significantly higher during the flag leaf stage compared to other stages. In a wheat experiment, researchers found fluctuations in the abundance of Leotiomycetes across phenological stages (Sun et al. 2022). This fungal group comprises numerous endophytes, some of which possess pathogenic potential (Wang et al. 2006). Further, Mortierellomycetes, a group that significantly contributes to explaining differences between treatments, was more abundant in soils treated with alternative fertilization (S, MBM and SMBM) than with traditional NPK fertilization, particularly in the flag leaf stage, and notably in the S treatment. Mortierella, a member of this group, plays a role in soil carbon and nutrient transformation, functioning as a saprophytic fungus that converts complex organic matter into accessible components (Sun et al. 2022). Blumeria, a fungus with crucial importance in distinguishing treatments, also exhibited clear phenological patterns, with the flag leaf stage displaying the most intriguing differences between treatments. During this stage, the relative abundance of Blumeria decreased in all fertilization treatments compared to the traditional one. This fungus includes the well-recognized wheat pathogen Blumeria graminis f. sp. tritici, known to be influenced by the type of nitrogen fertilization (Maywald et al. 2022). Our results might suggest that these alternative fertilization methods have the potential to reduce the abundance of this pathogenic fungus.

Conclusions

Wheat yield was maintained by the partial substitution of traditional NPK mineral fertilization with meat-bone meal and the combination of meat-bone meal plus sludge. Overall, sludge and its combination with meat bone meal benefited the activity and biomass of the soil microbial community, and generated important amounts of Olsen P for plant growth, as well as available N at the initial stages. However, the benefits of MBM on plant yield and grain production are likely attributed to variations in the abundance of less dominant populations, potentially involved in P solubilization, rather than the whole stimulation of the soil microbial community.

Crop phenology, intricately linked with the time elapsed from treatment application, demonstrated a dominant role, even surpassing the impact of fertilization itself, in influencing plant-soil microbial interactions. Therefore, it is crucial to consider different phenologies when assessing the responses of the soil microbial community to fertilization strategies. Nevertheless, we should be aware that we cannot discriminate phenological effects themselves from pure seasonal effects, which might also influence our results. From a practical standpoint, the organo-mineral combination demonstrated positive effects on soil sustainability while ensuring wheat yield. As such, it is deemed a suitable approach for recycling these residues and applying them as fertilizer within the context of a circular economy perspective.