Response of bacterial communities to the application of sewage sludge biochar and Penicillium aculeatum in rhizosphere and bulk soil of wheat

The application of sewage sludge biochar (SSB) in agriculture is a promising solution for detoxifying and recycling the nutrient-rich sludge. However, knowledge is required on how SSB alone or in combination with biofertilizers influences the soil microbial communities. This work used a wheat pot experiment with a root free soil compartment to study effects of SSB and the phosphate-solubilizing Penicillium aculeatum on bacterial communities in wheat rhizosphere and bulk soil. Treatments were applied only to the root free soil compartment. Analysis by 16S rRNA gene amplicon sequencing showed that SSB increased the alpha diversity and induced taxon-specific shifts. These shifts occurred in both rhizosphere and bulk soil


Introduction
Sewage sludge is being generated in vast amounts and needs to be recycled, e.g., as a nutrient rich soil amendment, but metals, toxins and pathogens in the sludge constitute an environmental threat. Pretreatment of the raw sludge is therefore required and its conversion into biochar by pyrolysis is considered an acceptable method for immobilizing the metals and removal of pathogens and hazardous compounds (Chagas et al., 2021;Goldan et al., 2022). The application of biochar to soil most often leads to enhanced fertility through effects on nutrient availability and physiochemical characteristics (e.g., pH, soil structure, release of soluble C and availability of micronutrients) (Joseph et al., 2021) and sewage sludge biochar (SSB) in particular appears to be a valuable fertilizer due to its high nutrient content (Efthymiou et al., 2018;Faria et al., 2018;de Figueiredo et al., 2021).
Biochar forms biogeochemical interfaces, due to its high porosity and presence of functional groups, which can support the growth of microorganisms or adsorb toxic compounds (Ali et al., 2019). Biochar may also increase the sorption of minerals, organic compounds, carbon or other nutrients (Zhu et al., 2017) and the resulting charosphere; i.e., the soil surrounding the biochar particle (Quilliam et al., 2013) may hence affect the soil-plant-microbe interactions. Since biochar properties such as surface charge, density and pore size distribution vary with both the feedstock (e.g., wood, straw, wastes) and the production technology (e. g., pyrolysis temperature), biochars from different feedstocks will differ in effects on the soil microbial community (Latini et al., 2019). The biochar-induced changes in the soil microbiota may benefit plant growth by enhancing the abundance of taxa that modulate e.g., soil physiochemical properties (Cheng et al., 2019) or nutrient cycling . The few studies that have addressed effects of SSB on soil microorganisms reported significant effects on the microbial communities (Luo et al., 2018;Ahmad et al., 2022;J. Chen et al., 2022).
Studies of soil microbiomes typically target the rhizosphere communities since they play a key role in plant nutrient acquisition, abiotic stress tolerance and protection against soil borne pathogens (Mendes et al., 2013). The rhizosphere is indeed the primary space for plant-rootmicrobe interactions and the rhizosphere communities are also influenced by root physiology and genetics. Still, the rhizosphere communities are predominately shaped by microorganisms recruited from the bulk soil (Bakker et al., 2015;Liu et al., 2019) and bulk soil samples should therefore be included in studies of soil amendment effects on microbial communities.
The bulk and rhizosphere microbiomes may also respond to the introduction of microbial inoculants which are being increasingly used in sustainable cultivation systems (Elnahal et al., 2022). Such inoculants include the widely studied phosphate solubilizing microorganisms which may improve plant P nutrition (Alori et al., 2017) but it is still unclear whether and how they affect the native rhizosphere and bulk soil communities. Inoculation is usually associated with the addition of the introduced microorganisms at high densities to ensure rapid colonization, which may change, at least temporarily, the equilibrium of the native communities (Trabelsi and Mhamdi, 2013). Furthermore, studies report that inocula of different beneficial microorganisms (e.g., symbiotic and free-living N-fixing bacteria, PGPR, biocontrol agents, AM fungi) did alter the soil microbial communities (de Salamone et al., 2010;Ravnskov et al., 2006;Yin et al., 2017). Penicillium spp. is predominant among soil fungi being able to solubilize various P compounds (Richardson, 2001) and P. aculeatum strain ATCC 10409 was selected for its ability to enhance the P fertilizer value of SSB by solubilizing biochar-P through the production of citrate (Efthymiou et al., 2018). Still, it is unknown how the soil microbiota will respond to such combined amendment.
The aim of this study was to test the following hypotheses: (1) the bacterial community responses to the combined application of SSB and P. aculeatum will differ from responses to the single-factor applications, (2) treatment effects on the bacterial communities will be related to selected soil characteristics and (3) community changes will differ between rhizosphere and bulk soils. To explore these hypotheses, bacterial communities and soil parameters were analysed in a wheat pot experiment where the pots had a mesh-separated root-free compartment that permitted a well-defined sampling of rhizosphere and bulk soil. The experimental treatments were applied to soil in the root-free compartment only. Wheat was chosen as model plant due to its relatively long root hairs (Wang et al., 2016) that would grow through the mesh and enable a rhizosphere-like soil layer at the proximal end of the root-free compartment.

Experimental design and model system
A two-compartment pot model system was constructed from 50 mm PVC tubing and comprised a 300 mm vertical plant root compartment (RC) that was separated from a 70 mm horizontal lateral root-free compartment (RFC) by means of a 25 μm nylon mesh (Ravnskov and Jakobsen, 1995). The RFC contained a 25 mm soil core and was closed by a capped vial fitting inside the tube (Fig. 1). Spring wheat plants (Triticum aestivum L. cv. Dacke) were grown in the model system and exposed to four treatments: (1) no amendment (Control), (2) amendment of the RFC soil with sewage sludge biochar (SSB), (3) inoculation of the RFC soil with P. aculeatum (Pa) and (4) combined application of SSB and P. aculeatum to the RFC soil (SSB + Pa) (Fig. 1). The pot design allowed for analysis of the microbial community structure at three levels: zone Arhizosphere soil within the RC without direct exposure to the experimental treatments; zone Ba 1.5 mm layer of rhizospherelike soil adjacent to the mesh in the RFC, with direct exposure to the treatments and to root hairs and exudates from the RC roots at the mesh; and zone C -bulk soil (15-20 mm from the mesh) with direct exposure to treatments but without direct exposure to root exudates (Fig. 1). Each treatment had eight replicates and sets of four were harvested and sampled after three and five weeks.

Growth substrate and sewage sludge biochar
The potting medium was a 3:1 (w/w) mixture of soil (sieved to 4 mm) and quartz sand. The soil, collected from the long-term nutrient depletion trial at the experimental research farm of the University of Copenhagen in Taastrup, Denmark (55 o 40 ′ N, 12 o 17 ′ E), was a sandy loam (17 % clay, 17 % silt, 66 % sand, 1.7 % soil organic matter, 312 mg total P kg − 1 , EC 0.30 dS m − 1 and pH 5.5 (1:5, soil:water)) or Luvisol in the FAO classification system. The soil-sand mixture, hereafter referred to as soil, contained 0.96 mg kg − 1 soil of water-extractable P (1:60, soil: water). The RC and the RFC contained 720 g and 60 g soil, respectively. In order to support plant growth in the nutrient-depleted soil, nutrients were mixed into the soil prior to potting (mg kg − 1 soil): NH 4 NO 3 , 100 N; KH 2 PO 4 , 10 P; K 2 SO 4 , 166 K; The SSB powder (particle size <1 mm) was uniformly mixed into the RFC soil for the SSB treatments at 1 % w/w which is a typical application rate in pot and field studies of plant responses to SSB (Junior and Guo, 2023;Faria et al., 2018). The SSB was produced at the Technical University of Denmark (see Thomsen et al., 2017 for more details) by slow pyrolysis (600 • C for 2 h) of sewage sludge originating from the Bjergmarken waste water treatment plant, Roskilde, Denmark (55 o 39 ′ N, 12 o 03 ′ E). The characteristics of the SSB were: pH 11.6, 23 % C, 2.2 % N, 7 % total P, 1.8 % bicarbonate-extractable P, nil water-extractable P, EC 0.36 dS m − 1 , 75 % ash content and a BET surface area of 25 m 2 g − 1 (see Efthymiou et al., 2018 andRaymond et al., 2018 for concentrations of further elements and FTIR-PAS spectrum). The addition of 1 % SSB corresponded to 0.7 mg P g − 1 soil.

Plant and fungal inoculum
Spring wheat (Triticum aestivum L.) cv. Dacke (Moerdrupgaard, Lynge, Denmark) seeds of uniform size were selected and surfacesterilized twice by gently shaking in a 3 % sodium hypochlorite solution for 3 min and rinsing six times for 1 min with sterile MilliQ water. Seeds were germinated on moist sterile filter paper at 25 • C for 2 days. The Penicillium aculeatum strain ATCC 10409 was obtained from the American Type Culture Collection and cultivated from a glycerol stock on potato dextrose agar (Difco) plates for 7 days at 25 • C. Spores were scraped off the plate, using a loop and Milli-Q water, and a 10 8 spores mL − 1 solution was achieved by filtration, centrifugation and resuspension in Milli-Q water. The P. aculeatum suspension was then mixed with the RFC soil of the two P. aculeatum treatments at a rate of 5 × 10 5 spores g − 1 soil. One week later, two germinated seeds were planted at 2 cm depth in each pot (RC) and were thinned to one per pot after emergence.

Growth conditions
The experiment was conducted in a growth chamber with a day/ night cycle of 16/8 h, temperatures of 20/15 • C, a relative air humidity of 65/72 % and a photosynthetically active radiation of 500 μmol m − 2 s − 1 . Prior to sowing, the soil moisture content was adjusted to 60 % of water-holding capacity (WHC) and the soil surface was covered by 10 g of plastic beads to minimize evaporation. The pots were arranged in a randomized design and watered to weight with deionized water every second day for the first three weeks, and daily during the final two weeks of the experiment. Water supply increased to 70 % and then 80 % of WHC during the last three weeks of growth.

Soil sampling and analyses
Soil was sampled from zones A, B and C of each pot at 3 and 5 weeks after sowing (Fig. 1). Rhizosphere soil from zone A was carefully sampled by shaking off the soil adhered to the roots from that zone. The soil attached to the RFC side of the mesh (1.5 mm layer in zone B) was considered as the rhizosphere soil of zone B. Soil sampled approximately 1.5-2 cm away from the mesh was considered as the bulk soil of zone C. Soil pH in H 2 O (1:5, w/w) and water-extractable P (WEP) were measured on duplicate samples obtained by pooling of fresh samples from replicate pots two-by-two. The WEP was extracted by shaking samples for 1 h in MilliQ water (1:60) followed by filtration (0.45 μm) (Van der Paauw, 1971) and analysis of ortho-P by flow injection (FIA star 5000, Foss Analytical, Denmark).

DNA extraction
Soil subsamples were freeze-dried for 24 h and homogenized using a pestle. DNA was extracted from 0.5 g of homogenized soil using the ® 96 Soil kit (Macherey-Nagel, Düren, Germany) adapted to a Biomek® FCP Laboratory Automation Workstation (Beckman Coulter™, CA, USA).
Negative controls (500 μL of PCR graded water) were included in the extraction and used for 16S rRNA amplicon library preparation together with the soil samples. All extracts were quantified using a Qubit 3.0 fluorometer (Invitrogen, by Life Technologies, Naerum, Denmark) with a Qubit® dsDNA HS Assay Kit (range 0.2-100 ng; Invitrogen, by Life Technologies, Naerum, Denmark) and stored at − 20 • C. DNA concentrations of soil extracts ranged from 1 to 32 ng μL − 1 .

Quantification of P. aculeatum by digital droplet PCR (ddPCR)
Absolute quantification of P. aculeatum by ddPCR was performed on DNA samples from the three sampling zones of the control and the P. aculeatum-inoculated treatments (Pa and SSB + Pa). The following primers and probe were used: (Efthymiou et al., 2018) and 5 ′ -FAM CAGCGGCGACGTCCGTCTGG BHQ-3 ′ . Droplet production and analysis was carried out using the QX200™ droplet digital PCR system and the QuantaSoft analysis software (Bio-Rad Technologies, Inc., Mississauga, ON, Canada), following the manufacturer's instructions. Reactions were prepared with 15 μL 2× ddPCR Supermix for probes (No dUTP) (Bio-Rad Technologies), 0.9 μM of each of the previously described forward and reverse primers, 0.25 μM of the probe, 1.5 μL of 10 x diluted DNA template and DNase/RNase-free water (Sigma-Aldrich) to reach a final volume of 30 μL. The thermal cycling program was as follows: 95 • C for 3 min followed by 5 cycles touchdown starting at 64.6 • C and decreasing by 0.4 • C per cycle, after which 25 cycles of 30 s at 95 • C, 30 s at 63 • C, 30 s at 72 • C and a final extension for 7 min at 72 • C.

Sequencing and trimming of 16S rRNA amplicon libraries
Extracted DNA was used for 16S rRNA gene sequencing, using the primers 341F (5 ′ CCTACGGGNGGCWGCAG-3 ′ ) and 805R (5 ′ GACTACHVGGGTATCTAATCC-3 ′ ) with adapters to amplify a fragment of the V3-V4 region of the 16S rRNA gene. Library preparation was performed as described in van der Bom et al. (2018) and paired-end sequencing of the 16S rRNA amplicon library was done using the MiSeq reagent kit v3 (600 cycles) and a MiSeq sequencer (Illumina Inc., San Diego, California, USA).

Fig. 1.
Pot design, treatment description and sampling zone description. Zone A represents rhizosphere soil from the root compartment (RC) which connects to the root free compartment (RFC) via a nylon mesh. In the RFC, zone B represents rhizosphere soil exposed to root hairs and exudates from RC roots at the mesh, and zone C represents bulk soil.
using the MiDAS database v.1.20 (Mcllroy et al., 2015). The raw sequences were uploaded to NCBI Sequence Read Archive (SRA) under the bioproject number PRJNA949852. The number of sequences before and after cleanup, the final number of OTUs, and the individual SRA sample accession numbers are presented in Table S1 (Supporting Information).

Plant analyses
Shoot dry weights were determined after drying at 70 • C for 48 h. Ground shoot samples (~100 mg) were oxidized in a mixture of 70 % nitric acid and 15 % hydrogen peroxide (2.5:1, v/v) in a microwave digestion system (UltraWAVE, Milestone Inc., USA) and the total shoot P content was determined by flow injection (FIA star 5000, Foss Analytical, Denmark).

Data analyses and statistics
All statistical analyses were performed in Rstudio (Version 0.98.1091). For the plant and soil data, one-or two-way analysis of variance (ANOVA) were performed to test the effects of the treatment at each time point and/or soil zone, and the effects of the treatment * zone interaction at each time point. P-values were further adjusted using the Tukey's honest significance difference (HSD) test and considered significant when <0.05. All assumptions were tested for normality and variances.
For sequencing data, the R packages vegan (Oksanen et al., 2017), MASS (Venables and Ripley, 2002) and mvabund (Wang et al., 2012) were used. Rarefaction curves at OTU level were computed indicating that the number of detected OTUs increased with the number of sequences but did not reach a true plateau (Table S1) The rrarefy function was used to rarefy the generated annotation tables to 5000 reads per sample, and alpha diversity indices (Richness (S = number of different OTUs) and Shannon's diversity (H = − ∑ p i * ln(p i ) where p i is the proportion of total OTUs made up of species OTU i )) at OTU level were then calculated using the PAST software ver.2.17 (Hammer et al., 2001). Differences in diversity indices were determined using one-or two-way ANOVA followed by HSD test and considered significant when P < 0.05. Beta-diversity at OTU level was analysed on non-rarefied data and represented by Non-metric multidimensional scaling (NMDS) ordination with Bray-Curtis dissimilarity index. Effects of zone and treatment factors at each time point were evaluated by PERMANOVA (10.000 permutations; Bray-Curtis dissimilarity index). Relative abundance of high taxonomic ranks (Phylum) was plotted in box-plots and effects of treatment within each zone were determined using one-way ANOVA followed by HSD test and considered significant when P < 0.05. Selection of OTUs having their abundance significantly affected by the treatments was done by fitting the non-rarefied data to a negative binomial model (nbGLM) and subsequent analysis of deviance. Selected OTUs were plotted in a heatmap of centered and scaled counts, as previously described (Nunes et al., 2016).

Results
This experiment was sampled three and five weeks after sowing but, for the sake of clarity, only results from the five weeks sampling are reported here. Results from the three weeks sampling were similar to the five weeks results, but with larger between-replicate variation. All data are available upon request to the authors.

Treatment effects on water-extractable P, pH, abundance of P. aculeatum and plant growth
Water-extractable P (WEP) was affected by SSB-treatments showing values 4.5-fold higher (P < 0.001) in zones B (RFC rhizosphere soil) and C (bulk soil) than in either control treatments or zone A (RC rhizosphere soil). In contrast, WEP was not or was only marginally affected by inoculation with P. aculeatum (Fig. 2a).
The alkaline biochar treatments (SSB and SSB + P. aculeatum) also increased pH in zones B and C by 1.1 unit (avg.) above pH in the control (Fig. 2b). The pH of the individual P. aculeatum treatment was similar to pH of the controls in all zones. In zone A, pH was stable across treatments. However, pH gradients across zones differed markedly between treatments: in the control, pH decreased by approximately one unit (P < 0.001) from zone A to zone C, whereas in the SSB treatments, pH increased by 0.5 units from zone A to zone B (P < 0.01). The across-zone pH gradient in the P. aculeatum treatment increased in a similar manner as the control.
Penicillium aculeatum, as quantified by ddPCR was not detected or was detected at very low abundance (< 60 DNA copies g ¡1 soil) in uninoculated controls (Fig. S2). In contrast, inoculation of RFC soil with P. aculeatum led to high abundances in zones B and C (6 × 10 5 -10 6 DNA copies g ¡1 soil). The small responses to P. aculeatum inoculation in zone A were non-significant.
Shoot dry weight and shoot P content were in the range of 0.9-1.2 g dry wt. and 1.1-1.3 mg P, respectively, and were not significantly influenced by the SSB or P. aculeatum treatments.

Alpha diversity of bacteria in rhizosphere and bulk soil
Both sampling zone and treatment had significant effects on alphadiversity of the bacterial communities estimated as Richness and Shannon indices (Fig. 3). Richness responded to both treatment and sampling zone and the two factors interacted significantly (Fig. 3a). This interaction resulted from a zone-dependent increase of Richness by SSB, which mitigated the marked decrease in Richness of the control from zone A to zone C. However, this effect of SSB was only significant in zone C (Fig. 3a). P. aculeatum did not influence Richness in any zone and accordingly, the individual application of SSB and its combination with P. aculeatum had similar effects in zone C. The Shannon diversity index also responded significantly to treatment and zones and again, the two factors interacted (Fig. 3b). In this case, both SSB treatments fully mitigated the decrease in Shannon index of the control from zone A to zone B, while in zone C only the SSB + P. aculeatum treatment increased the Shannon index significantly above that of the control. As observed for Richness, P. aculeatum inoculation did not affect the Shannon index in any zone.

NMDS ordination of bacterial communities in rhizosphere and bulk soil
The analysis of NMDS ordination data by PERMANOVA showed that the composition of the bacterial communities depended on both treatment (P ≤ 0.002) and sampling zone (P ≤ 0.004), and that the two factors interacted (P ≤ 0.002) (Fig. 4a-c). These interactions were mainly due to the absence of treatment effects in zone A, the rhizosphere soil from the root compartment.
The community composition in zones B and C responded mostly to SSB and the largest changes occurred in zone C (Fig. 4a). Penicillium aculeatum also influenced the communities, but less than SSB, and the fungus appeared to only have significant effects in zone B which was exposed to root hairs and root exudates (Fig. 4b). The dominant effect of SSB on the bacterial communities is also visible when comparing the effects of the combined application of SSB and P. aculeatum with the effects of the individual treatments (Fig. 4c). Hence, samples of the combined treatment (SSB + Pa) clustered tightly with SSB samples in both zone B and zone C, but not with P. aculeatum samples.

Relative abundance of high taxonomic ranks of bacteria in rhizosphere and bulk soil
The dominant phyla across treatments were Actinobacteria, Firmicutes and Proteobacteria and significant treatment effects were observed for Alphaproteobacteria, Gemmatimonadetes, Acidobacteria, Chloroflexi, Betaproteobacteria and Bacteroidetes when analysed across sampling zones (Fig. S3). However, within each individual zone, four additional taxa were also identified as being significantly affected by treatment: Firmicutes (P = 0.014), Deltaproteobacteria (P < 0.001), Gammaproteobacteria (P = 0.015) and Verrucomicrobia (P = 0.017). Furthermore, these zone dependent effects were mainly observed in the root free compartment zones B (rhizosphere) and C (bulk soil) where treatments had been applied (Fig. 5).
The relative abundance of phylum Gemmatimonadetes increased significantly in response to the two SSB treatments in zones B and C (P < 0.001; Fig. 5) and similar responses to SSB treatments were seen for Chloroflexi (P < 0.001). In contrast, the individual P. aculeatum treatment resulted in lower relative abundances of four phyla as compared to   3. Diversity indices calculated from 16S rRNA gene sequence data at OTU level for the control (green), sewage sludge biochar (SSB, red), P. aculeatum (Pa, blue) and biochar+P. aculeatum (SSB + Pa, black) treatments, at five weeks after sowing. a) Richness and b) Shannon Index. Samples were rarefied to 5000 reads prior to analysis. Bars represent means of four replicates ± SE. P-values of ANOVAs are shown and different letters indicate significant differences between the treatments (control, SSB, Pa and SSB + Pa) and across zones.
abundances for one or more of the other treatments: Bacteriodetes, zone C (P < 0.001), Betaproteobacteria, zones A and C (P < 0.01), Gammaproteobacteria, zone C (P = 0.015) and Deltaproteobacteria, zones A and B (P = 0.05 and P < 0.001, respectively) (Fig. 5). However, in two cases P. aculeatum increased the relative abundance of phyla. Hence, the abundance of Firmicutes in zone B became higher than in the control treatment (P = 0.014) and the abundance of Verrucomicrobia, zone C became higher than in the combination treatment (P = 0.017) (Fig. 5). Finally, some phyla were relatively more abundant in the control treatment than in one or more of the experimental treatments: Acidobacteria, zones B (P = 0.015) and C (P = 0.04), Chloroflexi, zone A (P < 0.001) and Alphaproteobacteria, zone B (P < 0.001) and zone C (P < 0.01).

Potential mechanistic relationships between treatment effects on bacteria phyla and soil characteristics
The positive effect of SSB on Gemmatimonadetes in zones B and C cooccurred with large increases in levels of water extractable P, but SSB also increased pH levels in the two zones (Figs. 2 and 5). In contrast, the suppressive effect of P. aculeatum on Bacteroides, Betaproteobacteria and Deltaproteobacteria showed no relationship to either WEP or pH, both of which were similar in the P. aculeatum and in the control treatment. Accordingly, the P. aculeatum-induced stimulation of Firmicutes and Verrucomicrobia showed no relationship with WEP or pH.

Discussion
This work provides novel insight into (i) specific effects of P rich sewage sludge biochar (SSB) on soil bacterial communities in both rhizosphere and bulk soil of wheat plants and (ii) possible interactions between SSB and inoculation with a P-solubilizing Penicillium aculeatum strain. So far, studies of biocharsoil microbiome interactions have primarily used biochar from feedstocks such as wood chips and straw (Latini et al., 2019;Yin et al., 2021;Joseph et al., 2021;Tan et al., 2022) which are less nutrient rich than biochar derived from sewage sludge and manure.

Biochar-induced increase in water extractable soil P and pH
The four-fold increased P availability in SSB-treated soil was probably related to a high abundance of Ca-associated P forms in the SSB. Such P forms are extractable with HCl (Hedley et al., 1982) and represented 69 % of total P in a very similar SSB based on sludge from the same treatment plant as in this work (Mackay et al., 2017). Since this Ca -P rich and alkaline SSB was added to a relatively acid soil, some P release is expected (Schneider and Haderlein, 2016). Furthermore, the increased pH observed in soil treated with alkaline SSB probably contributed to the higher WEP levels. Hence, the soil mineral surfaces would have become more negatively charged with the subsequent release of some of the adsorbed P (Glaser and Lehr, 2019). The direct alkalinization effects of SSB in zones B and C probably occurred concurrently with indirect effects related to plant mineral uptake. Such effects were observed as pH increases from bulk soil zone C to true rhizosphere soil A in control and P. aculeatum treatments. These pH gradients might result from the alkalinization of the rhizosphere soil due to the common predominant uptake of NO 3 − over NH 4 + (Custos et al., 2020).

Bacterial community responses to sewage sludge biochar and P. aculeatum
Overall, SSB effects on the bacterial communities were stronger than the effects of P. aculeatum. The combined application of SSB and P. aculeatum induced similar changes in the bacterial communities, as when SSB was applied alone, leading to rejection of the first hypothesis.
The impact of the nutrient-rich SSB on bacterial communities occurred at all levels of analysis, i.e. alpha diversity, beta diversity and relative abundance of high taxonomic ranks. Although this work does not allow for a conclusive identification of all the underlying mechanisms, the SSB-induced increases in water extractable P and pH will likely have played a role. Indeed, higher richness and diversity are wellknown biochar effects (Li et al., 2022) that have been linked to nutrient enrichment, higher pH, and more niche space (Singh et al., 2022;Zhu et al., 2017). More specifically, the suggested roles of the SSB-induced P and pH increases are in accordance with other studies reporting that abundance and diversity of soil bacterial communities are positively related to pH (Rousk et al., 2010;Zhang et al., 2017) and to soil P availability (Tan et al., 2013;Siegenthaler et al., 2022). Besides, possible factors behind the SSB-induced changes in the beta diversity would include the soil P enrichment as well as biochar-soil aggregates providing protected and nutrient-rich micropore habitats (Palansooriya et al., 2019).
The SSB-induced increase in the relative abundance of Gemmatimonadetes and Chloroflexi aligns with the results of a recent study where these phyla, together with Cyanobacteria and Nitrospirae, comprised biochar-enhanced taxa (Li et al., 2023). Indeed, the frequently reported enrichment of the Gemmatimonadetes phylum in biochar-amended soils has been associated with common properties of biochar such as alkalinity, high nutrient content and complex structure (Cheng et al., 2019;Han et al., 2017;Khodadad et al., 2011;Sheng and Zhu, 2018;Xu et al., 2014;Yao et al., 2017;Yu et al., 2018). In the present work, the co-occurrence of an increased relative abundance of Gemmatimonadetes and an enhanced P availability in SSB-amended soil may suggest that this phylum participates in microbial P cycling by utilizing pools of poorly available P in the SSB such as Ca-phosphates. This is also supported by a recent finding that the genome of the Gemmatimonadetes contains the gcd gene encoding glucose dehydrogenase (GCD) (Liang et al., 2020). This enzyme is responsible for the production of gluconic acid and taxa from Gemmatimonadetes might therefore be involved in the microbial solubilization of recalcitrant inorganic P in biochar. This agrees with the observed ability of Gemmatimonas to decompose pyrogenic C (Whitman et al., 2016) and thereby get access to P. The potential importance of Gemmatimonadetes in the cycling of biochar P is corroborated by the present work. Here, the OTUs from this phylum constituted <28 % of the SSB-enriched OTUs. In addition to this possible specific role of Gemmatimonadetes, soil pH may have also played a role since the increasing abundance of Gemmatimonadetes cooccurred with increased soil pH. This role of pH is supported by observations that Gemmatimonadetes not only prefers neutral over acidic pH, but dominates in alkaline soils (see refs in Mujakić et al., 2022). The

Fig. 5.
Relative abundance of the taxa that responded significantly to treatments (control, green; sewage sludge biochar, SSB, red; P. aculeatum, Pa, blue; SSB + Pa, black) within each zone (A, B or C) at five weeks after sowing. Data are presented at the phylum level, except for Proteobacteria, which are classified to class level. Pvalues of ANOVAs (P < 0.05) are shown at each soil zone. Letters indicate significant differences between the treatments (control, SSB, Pa and SSB + Pa) at each soil zone.
increased abundance of Chloroflexi in response to SSB is supported by other biochar studies (Cheng et al., 2019;Li et al., 2023) and might be explained by the associated pH increase Sheng and Zhu, 2018). In contrast to pH, the increased P availability in the SSB treatments may have exerted less influence on Chloroflexi which did not become more abundant with increasing soil P in previous studies (Y. Chen et al., 2022;Cheng et al., 2020).
The relatively smaller and less frequent effects of P. aculeatum, as compared to effects of SSB, suggest that the individual application of this fungus may not induce major shifts in the bacterial community composition. The observed zone-specific increases (Firmicutes and Verrucomicrobia) or decreases (Bacteriodetes, Betaproteobacteria, Deltaproteobacteria) in response to the inoculation with P. aculeatum occurred without any changes in soil pH and WEP. Therefore, the possible underlying mechanisms would need to be identified among an array of complex interactions between bacteria and fungi as reviewed by Deveau et al. (2018) and would need a different experimental design. Consequently, potential mechanisms behind the observed effects were only indicated for the SSB treatment leading to a partial acceptance of our second hypothesis.

Zone-specific treatment effects: rhizosphere versus bulk soil bacterial communities
The considerably higher effect of SSB and P. aculeatum on alpha and beta diversity and on the relative abundance of specific taxa in root-free zones B and C than in rhizosphere zone A was expected since only zones B and C were directly exposed to the SSB and P. aculeatum treatments. Yet, the bacterial communities in bulk soil C presented several differences as compared to zone B. This narrow zone is in the root-free compartment, but it still represents the rhizosphere soil that is directly exposed to root exudates and physio-chemical changes known to significantly influence the microbial communities (Bais et al., 2006). Such root-derived modifications can therefore explain the generally higher alpha diversity indices observed in zone B. The higher similarity between the rhizosphere and bulk soil bacterial communities in the SSB Fig. 6. Heatmap of the OTUs that significantly differed in relative abundance between the control (green), the sewage sludge biochar (BC, red), the P. aculeatum (Pa, blue) and the biochar+P. aculeatum (SSB + Pa, black) treatments in zone C. Data are centered and scaled to the average of each taxon's abundance and each vertical column corresponds to one sample (treatments are separated by white lines). Sample clustering (constrained average agglomeration method) is based on Euclidean distance. The left vertical bar in the heatmap corresponds to the phylum level classification of each OTU, with the exception of Proteobacteria, which are classified to class level. treatment than in the control suggests that the SSB effect was stronger than the rhizosphere effect. Looking at specific phyla, rhizosphere specific effects were only observed in a few cases such as the P. aculeatuminduced stimulation of Firmicutes or suppression of Deltaproteobacteria in zone B but not C. Since the abundance of P. aculeatum was similar in zone B and zone C, such results would seem to involve rhizosphere-and phylum specific interactions with the fungal inoculant. On the other hand, the relative abundance of Gemmatimonadetes and Chloroflexi was enhanced by SSB in both zone B and zone C, i.e., irrespective of rhizosphere-specific conditions. Overall, the rather complex picture leads to partial acceptance of our third hypothesis. Nevertheless, the need to target not only rhizosphere but also bulk soil in studies of the effects of resource management on soil microbial communities is supported by the observation that legacy effects of altered bulk soil communities persisted into the rhizosphere (Bakker et al., 2015).

Conclusion
This work reports that bacterial communities in rhizosphere and bulk soil responded strongly to sewage sludge biochar and this response was not modified by a P solubilizing P. aculeatum strain. The consistent higher abundance of Gemmatimonadetes and Chloroflexi observed in response to SSB was associated with increased P availability and increased pH, respectively. The mechanisms behind the less consistent effects of P. aculeatum were associated with unidentified rhizosphere processes. It is also shown that the effects of treatments to the overall bacterial communities or even to the relative abundance of specific phyla in bulk soil are not always extended to the rhizosphere soil. This leads to the suggestion that bulk soil should be included in future work evaluating bacterial community responses to resource management. The absence of significant interactions between SSB and P. aculeatum in this study might have been related to the selected model system and, in future studies, such interactions should be studied in the direct presence of roots and exploring different timings, fungal strains and application methods. Future studies should also pay attention to the persistence of effects and should address not only bacteria but also fungi and other soil biota components.

Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability
Data will be made available on request.