Inhibition profile of three biological nitrification inhibitors and their response to soil pH modification in two contrasting soils

Abstract Up to 70% of the nitrogen (N) fertilizer applied to agricultural soils is lost through microbially mediated processes, such as nitrification. This can be counteracted by synthetic and biological compounds that inhibit nitrification. However, for many biological nitrification inhibitors (BNIs), the interaction with soil properties, nitrifier specificity, and effective concentrations are unclear. Here, we investigated three synthetic nitrification inhibitors (SNIs) (DCD, DMPP, and nitrapyrin) and three BNIs [methyl 3(4-hydroxyphenyl) propionate (MHPP), methyl 3(4-hydroxyphenyl) acrylate (MHPA), and limonene] in two agricultural soils differing in pH and nitrifier communities. The efficacies of SNIs and BNIs were resilient to short-term pH changes in the neutral pH soil, whereas the efficacy of some BNIs increased by neutralizing the alkaline soil. Among the BNIs, MHPA showed the highest inhibition and was, together with MHPP, identified as a putative AOB/comammox-selective inhibitor. Additionally, MHPA and limonene effectively inhibited nitrification at concentrations comparable to those used for DCD. Moreover, we identified the effective concentrations at which 50% and 80% of inhibition is observed (EC50 and EC80) for the BNIs, and similar EC80 values were observed in both soils. Overall, our results show that these BNIs could potentially serve as effective alternatives to SNIs currently used.


Introduction
In a gricultur al soils, nitr ogen (N) often originates fr om the a pplication of ammonium-based fertilizers, with ∼115 Mt yr −1 of fertilizer applied globally (FAO 2019 ).Ho w ever, modern fertilization pr actices ar e highl y inefficient and between 50% and 70% of the applied N is lost from agricultural systems (Subbarao et al. 2015, Coskun et al. 2017 ).It is estimated that about 50% of the global anthr opogenic nitr ous oxide (N 2 O) emissions originate fr om fertilized a gricultur al soils and that by 2030 a gricultur al N 2 O emissions are projected to further increase another 15% (Tian et al. 2020, IPCC 2022 ).Continuing to feed the world's growing population while reducing N losses from agriculture will, therefore only be possible with more efficient fertilization strategies.
A lar ge fr action of N fertilizer a pplied to soils is lost thr ough micr obiall y mediated nitrification, a process in which ammonia (NH 3 ) is oxidized to nitrite (NO 2 − ) and subsequently to nitrate (NO 3 − ).This process is mediated by three functional groups of micr oor ganisms: ammonia-oxidizing arc haea and bacteria (A O A and AOB), which oxidize NH 3 to NO 2 − , nitrite-oxidizing bacteria (NOB), which oxidize NO 2 − to NO 3 − , and complete ammonia-oxidizing (comammox) bacteria, which perform the complete oxidation of NH 3 to NO 3 − (Daims et al. 2015, van Kessel et al. 2015 ).As a result of nitrification, N is lost through leaching of highly mobile NO 2 − and NO 3 − , or is emitted as N 2 O due to nitrifier denitrification, incomplete hydroxylamine oxidation and abiotic reactions between nitrification pr oducts (Giguer e et al. 2017, Hink et al. 2018, Prosser et al. 2020, Hu et al. 2022 ).As ∼90% of N fertilizer in soils is nitrified (Subbarao et al. 2013 ), nitrification plays a centr al r ole in a gricultur al N use management.
One a ppr oac h to counter act N loss thr ough nitrification in a gricultural soils is the application of synthetic nitrification inhibitors (SNIs) together with N fertilizers (Huber et al. 1977, Slangen and Kerkhoff 1984, Ruser and Schulz 2015 ).Currently, the most widely used SNIs are dic y andiamide (DCD), 3,4-dimethylp yrazole phosphate (DMPP), and nitr a pyrin (2-c hlor o-6-(tric hlor omethyl)pyridine), whic h hav e been shown to effectiv el y incr ease soil N r etention while r educing fertilizer N losses (Zerulla et al. 2001, Subbarao et al. 2006, Zhou et al. 2020 ).Recentl y, plant-deriv ed compounds that suppress soil nitrification, known as biological nitrification inhibitors (BNIs), have increasingly been considered as natural and inexpensive alternatives to SNIs (Subbarao et al. 2012, 2015, Coskun et al. 2017 ).Compounds isolated from root exudates , plant tissues , or plant residues , ha v e pr e viousl y been shown to have nitrification inhibition activity, including methyl 3-(4-hydr oxyphen yl) pr opionate (MHPP) (Zakir et al. 2008 ), methyl 3-(4-hydr oxyphen yl) acrylate (MHPA), also known as methyl-pcoumarate (Gopalakrishnan et al. 2007 ), and limonene (White 1991 ).Ho w e v er, conflicting r esults hav e been obtained when testing the efficacy of these and other BNIs on a gricultur al soils due to the influence of widely varying soil properties, the lack of robust estimates of effective BNI concentrations, and the variable responses of the different naturally occurring soil nitrifier communities.
In soils, factors such as pH, temperature, organic matter content, and the complex interactions among them, influence not only nitrifying communities and nitrification rates, but also the efficacy of nitrification inhibitors (NIs; Keeney 1980, Subbarao et al. 2015 ).In particular, soil pH controls many soil biotic and abiotic pr ocesses, suc h as the depr otonation of ammonium (NH 4 + ) to NH 3 , whic h dir ectl y affects the activity of differ ent soil ammonia oxidizers due to their different affinities for NH 3 (Kits et al. 2017, Jung et al. 2022 ).In addition to this niche partitioning of ammonia oxidizers, soil pH also affects the sorption of NIs to soil particles, minerals, and soil organic matter; e.g. higher sorption of SNIs has been observed in alkaline soils (ASs) than in acidic soils (Guardia et al. 2018 ).
Inhibitor concentration is known to influence the efficacy of NIs (Nardi et al. 2020 ).In pur e cultur es of soil ammonia oxidizers, activity decreased linearly with BNI concentration (Kaur-Bhambra et al. 2022 ), suggesting that higher efficacies are achieved at higher BNI concentr ations.Pr e vious studies hav e shown the inhibitory effect of BNIs on soil nitrification (Zhang et al. 2015, Lu et al. 2019, Ma et al. 2021, Lan et al. 2022 ).Ho w e v er, differ ent nitrification inhibition r esults wer e observ ed when using the same BNI concentration acr oss differ ent soils (Wang et al. 2021 ).Ther efor e, elucidating the factors that influence the efficacy of BNIs is an important first step in determining their potential efficiency in a gricultur al soils and whether they are a suitable natural alternative to SNIs in counteracting N fertilizer losses.
The objective of this study was to e v aluate the efficacy of three SNIs (DCD, DMPP, and nitr a pyrin) and three BNIs (MHPP, MHPA, and limonene) in two contrasting agricultural soils differing in pH and nitrifier comm unities.Mor eov er, for the thr ee BNIs we determined the EC 50 (effective concentration at which 50% of inhibition is observed) and EC 80 in both soils.We hypothesized that soil pH has a strong direct influence on inhibitor efficac y b y influencing soil sorption and physicochemical properties, but also indirectly, by driving differences in nitrifier abundance and diversity.We further hypothesized that BNI efficacy is concentration dependent, and that the EC 50/80 of the tested BNIs differ between soils.While pr e vious studies have assessed the efficacy of NIs in different soil types, this study examines the direct effect of short-term soil pH perturbations and inhibitor concentration on the efficacy of NIs.

Site description and soil sampling
Samples were collected in spring 2022 from two agricultural soils located in Lo w er Austria, in the Marc hfeld r egion (48  (Lehtinen et al. 2014, Spiegel et al. 2018 ).The soil at the Marchfeld site is classified as Calcaric Phaeozem (sandy loam: 30.3% sand, 45.6% silt, and 24.2% clay), while the Alpenvorland site soil is a Gleyic Luvisol (loamy silt: 9.5% sand, 71.2% silt, and 19.4% cla y).T he two sites gr eatl y differ in their soil pH: Marchfeld is an AS with a pH in water of 8.50 ± 0.02, while Alpenvorland is a slightly acidic or circumneutral soil (CS) with a pH of 6.12 ± 0.09.Over the last 40 years, both soils r eceiv ed 120 kg N ha −1 yr −1 , 75 kg P 2 O 5 ha −1 yr −1 , and the crop rotation system consisted of 53%-55% cereals and 45%-47% root crops (Lehtinen et al. 2014, Spiegel et al. 2018 ).At each site, soil from the top 10 cm was collected from four field replicate plots, and each sample consisted of four soil cores.Samples were transported to the laboratory and sieved (2 mm mesh size).Soil for molecular analyses was stored immediately at −80 • C, while soil for microcosm incubations was stored at 4 • C.

Soil char acteriza tion
Soil water content was determined gr avimetricall y by drying 5 g of soil at 60 • C for 48 h.Dried and ground samples were analyzed for total carbon (C) and total N with an EA-IRMS (EA 1110, CE Instruments, Italy, coupled to a Finnigan MAT Delta Plus IRMS; Thermo Fisher Scientific, MA, USA).Soil pH w as measured in w ater using standard pH electrodes (1:5 ratio w/v).The content of CaCO 3 , EDTA-extr actable ir on, manganese, copper, zinc, and CEC were determined by the Austrian Agency for Health and Food Safety (AGES) follo wing the standar d protocols ÖNORM L1084, L1098, and L1086-1.Microbial biomass C and N (C mic and N mic ) were determined by c hlor oform fumigation extr action and corr ected for extraction efficiency using a factor of 0.45 (Vance et al. 1987, Jenkinson et al. 2004 ).Fumigated and nonfumigated soil samples wer e extr acted with 1 M KCl (2 g soil in 15 ml KCl) and anal yzed for dissolved organic carbon (DOC) and total dissolved N (TDN) content on a TOC/TN Analyzer (TOC-V CPH E200V/TNM-122 V; Shimadzu, Austria).A modified photometric indophenol reaction method was used to determine NH 4 + in the KCl extracts (Kandeler and Gerber 1988 ), and the acidic VCl 3 /Griess reaction was used to determine NO 3 − and NO 2 − (Griess-Romijn v an Ec k 1966, Mir anda et al. 2001 ).  1 .

Microbial and nitrifier community char acteriza tion
The amoA and 16S rRNA gene amplification and sequencing was carried out at the Joint Microbiome Facility of the Medical University of Vienna and the University of Vienna (JMF) under project IDs JMF-2110-10 and JMF-2206-04.For amoA gene sequencing, the primers described above for qPCR were used.The 515F/806R primers were used for 16S rRNA gene sequencing (Apprill et al. 2015(Apprill et al. , P ar ada et al. 2016 ) ). Target gene amplification and sample bar coding w ere performed with a two-step PCR protocol using the aforementioned primer pairs modified with linker sequences, as described in Pje v ac et al. ( 2021 ).Sequencing was performed on the Illumina MiSeq platform, using the 600-cycle v3 chemistry (2 × 300 bp pair ed-end r eads).Raw sequencing data wer e pr ocessed using FASTQ w orkflo w (Basespace , Illumina) with default settings , and amplicon sequence variants (ASV) were inferred using the D AD A2 pipeline (Callahan et al. 2016 )

Soil net nitrification potential and A O A contribution
Soil slurry incubations were performed to e v aluate the contributions of A O A and A OB/comammox to the net nitrification potential of both soils.For the assays, fresh soil (4 g) and 30 ml of MilliQ w ater w ere w eighed in 125 ml Wheaton bottles, and NH 4 Cl w as added to a final concentration of 1 mM.All bottles were sealed gas tight with 33 mm interlock butyl septa.To differentiate between the contribution of A O A and A OB/comammox to soil nitrification the AOB and comammox specific inhibitors 1-octyne (4 μM) and all yl-2-thiour ea (ATU, 100 μM) (Sigma-Aldrich, St. Louis, MO, USA) tr eatments wer e included (Taylor et al. 2013(Taylor et al. , 2015 ) ).The negativ e contr ol, in whic h nitrification activity was completel y inhibited, r eceiv ed 1 mM phenylacetylene (McCarty and Bremner 1986 ). Sealed bottles were shaken at 180 r/m at 23 • C for 72 h.Each treatment included four replicate soil slurries and 1 ml subsamples were collected with a 1-ml syringe through the butyl septa daily during the incubation period to measure NO 3 − accumulation as described abo ve .NO 2 − accumulation was also assessed but was below the limit of quantification of 15 μM in all samples at all timepoints.Soil net nitrification potential ( μg N g −1 soil day −1 ) was calculated by subtracting the initial NO 3 − concentr ation fr om the accum ulated NO 3 − concentr ation, normalized to soil weight and incubation time.

Short-term soil pH perturbations
To e v aluate the effect of soil pH on the efficacy of SNIs and BNIs further soil slurry incubations were conducted as described abo ve .
The treatments consisted of modifying the pH of the soil slurries before NH 4 Cl and inhibitor addition.Soil slurries with the AS (nativ e pH 8.50) wer e adjusted to pH 7.18 and 6.15 using 0.3 ml and 2.5 ml of 0.5 M HCl, r espectiv el y.Similarl y, the pH of the CS slurries (native pH 6.12) was adjusted to pH 7.60 and 8.50 by adding 80 μl and 200 μl of 0.5 M NaOH, r espectiv el y.After the initial pH adjustment, slurries were left to stabilize for 24 h at 23 • C, shaking at 180 r/m, before the addition of 1 mM NH 4 Cl and each of the inhibitors .T he pH was measur ed periodicall y to confirm that the adjusted pH was stable throughout the experiment.
Nitr a pyrin, DCD (Sigma-Aldric h), and DMPP (Cayman Chemical, Mic higan, USA) wer e a pplied at r ates pr e viousl y used in a gricultural soils (Lu et al. 2019, Dawar et al. 2021, Lan et al. 2022 ).Specificall y, 3.5 μM nitr a pyrin (equiv alent to 24 μg g −1 soil) dissolved in 99.5% DMSO (Lactan Chemikalien & Labor ger äte GmbH, Graz, Austria) (0.035% v/v DMSO final concentration in the soil slurries), 100 μM DCD (equivalent to 252 μg g −1 soil) and 10 μM DMPP (equivalent to 58 μg g −1 soil) were added to the soil slurries.A DMSO control (0.035% v/v) was also included.As 1000 μg g −1 soil is the highest concentration of BNIs pr e viousl y tested in different soil types (Wang et al. 2021 ), MHPP, MHPA, and limonene were tested at a concentration of 200 μM (equivalent to 1081.12, 1069.08, and 817.44 μg g −1 soil, r espectiv el y).In addition to the SNIs and BNIs , octyne , ATU, and phen ylacetylene contr ols wer e included as described abo ve .All treatments were performed in triplicates and incubated under the conditions described abo ve .Pr e vious experiments with the AS sho w ed higher nitrification activity than the CS, so the incubation period for the AS was set to 48 h, while the incubations with the CS were carried out for 72 h.Subsamples of 1 ml were taken daily for NO 3 − accum ulation measur ements to determine net nitrification potential as described abo ve .T he percentage of nitrification inhibition caused by each inhibitor was calculated using the following equation: where T is the NO 3 − concentration ( μg N g −1 dry soil) produced in the NI treatment, C is the NO 3 − per gr am of soil pr oduced in the positiv e contr ol (i.e.no inhibitor added) and N is the NO 3 − consumption that occurred in the negative control (i.e.phenylacetylene addition).

Net carbon dioxide production
Additionally, to assess the effect of the BNIs on the heterotrophic micr obial comm unity r espir ation, net carbon dioxide (CO 2 ) pr oduction was quantified.Ambient and headspace gas samples were taken to determine initial (0 h) and final (72 h) CO 2 concentrations in the treatments with and without BNI addition.CO 2 concentr ations wer e determined using an infr ar ed gas analyzer (EGM-4 Environmental gas analyzer for CO 2 , PP Systems; Hertfordshir e, UK).Net CO 2 pr oduction in eac h tr eatment was calculated by subtracting the initial from the final CO 2 concentration.

EC 50 and EC 80 determination
To determine the EC 50 and EC 80 values of the three BNIs used in this study (MHPP, MHPA, and limonene) (Sigma-Aldrich), similar net nitrification potential incubations were performed as described abo ve .For eac h BNI, up to six concentr ations r anging fr om 50 μM to 1.5 mM were tested in each soil.All BNIs were dissolved in DMSO at concentr ations r anging fr om 0.004% to 0.2% v/v, and a DMSO control (0.25% v/v) without inhibitors was also included.
Mor eov er, octyne and phenylacetylene controls, as well as a positiv e contr ol without inhibitors were included.The percentage of nitrification inhibition for each BNI concentration was calculated as described abo ve .T he EC 50 and EC 80 values for MHPP, MHPA, and limonene, w ere calculated b y fitting a dose-response model (Motulsky and Christopoulos 2003 ) to the percentage of nitrification inhibition at each concentration using the equation: where y is the percentage of nitrification inhibition, x is the concentration of inhibitor ( μM), h is the Hill coefficient, F is the target percentage of inhibition (i.e.50% or 80%), and EC F is the effective concentration at which 50% or 80% of inhibition is observed.The model was constrained to two parameters, setting the minimum ( Min ) and maximum ( Max ) effect of inhibition to 0% and 100%, respectiv el y.Additionall y, the tar get percenta ge of inhibition ( F ) was set to either 50 for EC 50 or to 80 for EC 80 calculations .T hese constraints simplified the equation to: Nonlinear least squares regression (Kemmer and Keller 2010 ) was used to estimate the EC 50 and EC 80 values as well as the Hill coefficient, along with their r espectiv e standard errors.

Sta tistical anal ysis
All statistical analyses were performed in R version 4.3.1 (R Core Team 2023 ) using Rstudio version 2023.6.0.421 (Posit Team 2023 ).Differences in soil parameters between sites were tested using a t -test (for Mn, Cu, Zn, DOC, soil C:N, C mic , N mic , microbial C:N, ammonium, and nitrate) or a Wilcoxon signed-rank test (for pH, % CaCO 3, Fe , CEC , total C , and total N).For the differences in ammonia oxidizer abundances, a generalized least squares model was run using the gls function of the nlme 3.1-162 pac ka ge (Pinheir o et al. 2023 ).Her e, 'site' and 'type of ammonia oxidizer' were used as fixed factors and the abundance of each gene as the response variable.A tw o-w ay ANOVA w as used to test for differences in the total net nitrification potential and A O A activity between soils.T -tests were conducted to compare the efficacy of each NI among sites and a one-w ay ANOVA w as used to test whether the inhibitor efficacy depended on adjusted soil pHs.For all tests, a P-value < .05 was considered to mark significant differences between sites or inhibitor treatments, unless otherwise noted.
Sequence data analyses were performed using the phyloseq 1.44.0 (McMurdie and Holmes 2013 ), and vegan 2.6-4 pac ka ges (Oksanen et al. 2022 ).For all genes, a canonical correspondence analysis on a Bray-Curtis dissimilarity matrix was performed using the ordinate function from the phyloseq package .T he significance of the environmental variables was assessed with the envifit function of the vegan pac ka ge and only the variables with a P < .01wer e c hosen for the canonical correspondence analysis ( Supplementary Table 2 ).Alpha-diversity was determined on a dataset r ar efied to 6126, 3189, 905, and 2688 reads sequencing depth for the 16S rRNA, A O A, A OB, and comammox clade B amoA gene datasets, r espectiv el y.To test for the differ ences in the 16S rRNA and amoA alpha diversity between sites, a t -test or a Wilcoxon signed-rank test were performed.To test for significant differences between the taxa with mean relative abundance ≥ 2% for the 16S rRNA, and ≥ 5% for the amoA gene amplicon datasets, the ANCOMBC 2.2.0 (Lin and Peddada 2020 ) pac ka ge was used.

Soil physicochemical properties of the studied sites
Two distinct a gricultur al soils with differing pH were selected for this study, an AS (AS, pH 8.50 ± 0.02) and a CS (pH 6.12 ± 0.09) (Table 1 ).While many soil properties significantly differed between the soils, the TDN and EDTA-extractable zinc did not.Besides a high pH, the AS is c har acterized by a significantly higher CaCO 3 content, CEC, EDTA-extractable (i.e .plant a vailable) copper concentr ation, and nitr ate concentr ation.In addition, the AS also has a significantly higher total C, total N, soil C:N r atio, DOC, micr obial C, and microbial N. The CS, in contr ast, has significantl y higher ammonium, and EDTA-extr actable ir on and manganese concentrations (Table 1 ).

Total microbial community composition
Micr obial comm unity structur e, based on 16S rRNA gene amplicon sequencing, significantly differed between the AS and CS in the most abundant orders (only orders with ≥ 2% mean r elativ e abundance at each site were compared; Fig. 1 A).In both soils, the A O A order Nitr ososphaer ales was among the most abundant microbial orders but accounted for a significantly higher r elativ e abundance in the AS than in the CS ( P < .0001).In the AS, the Nitr ososphaer ales comprised 22 ± 3.05% of the 16S rRNA genebased community, while in the CS they represented only 7.2 ± 0.84% (Fig. 1 A).Other putative nitrifiers within the genera Nitrospira (1.4 ± 0.25% in the AS and 0.99 ± 0.11% in the CS) and Nitrosospira (0.27 ± 0.03% in the AS and 0.28 ± 0.02% in the CS) were also detected in both soils, but at low r elativ e abundances ( < 2%).
The diversity of the total microbial community was significantly higher in the CS, while the richness and observed diversity did not differ significantly between the two sites ( Supplementary Fig. 1 ).Micr obial comm unity composition dissimilarities between the two investigated soils corresponded to differences in pH, CEC, and DOC, which were higher in the AS, as well as Fe and Mn, which were higher in the CS.Together, these factors explain over 62.1% of the community variance along the CCA1 ( Supplementary Fig. 2 ).

Ammonia oxidizer community and net nitrification potential
Analysis of the ammonia oxidizer community with amoA gene amplicon sequencing also sho w ed significant differences between the AS and the CS.The A O A comm unities wer e dominated by ASVs affiliated with the genus TA-21 from the NS-δ-2.1 clade (Alv es et al. 2018 ), r epr esenting 99.2 ± 0.2% of the A O A ASVs in the AS and 49.9 ± 7.8% in the CS (Fig. 1 B, Supplementary Fig. 3 ).Additional A O A amoA ASVs, belonging to the NS-ε clade (11.79 ± 1.7%) wer e onl y pr esent in the CS.Members of the w ell-kno wn terrestrial AOB genus Nitrosospira were dominant in both soils, but distinct ASVs were detected between each soil.Similarly, distinct comammox clade B ASVs w ere detected betw een soils (Fig. 1 B).Both richness and evenness of the AOB and comammox clade B comm unities significantl y differ ed between the AS and CS, while not significant differences were observed in the A O A community ( Supplementary Fig. 4 ).Much like with the 16S rRNA total microbial community, pH and DOC were the strongest drivers of amoA community dissimilarities ( Supplementary Fig. 5 ).
The quantification of the amoA genes sho w ed significant differences in the total amoA gene abundance between sites.Ov er all, the AS harboured a higher number of amoA genes than the CS (6.7 × 10 7 ± 2.5 × 10 6 amoA gene copies g −1 soil and 2.5 × 10 7 ±  1.8 × 10 6 amoA gene copies g −1 soil, r espectiv el y; P = < .01).In fact, the AS sho w ed significantly higher abundances of both A O A and AOB; ho w e v er, both sites contained similar comammox clade B abundances (Fig. 2 A).No comammox clade A amoA genes were detected in either soil.The AS soil was dominated by A O A, containing 4.3 × 10 7 ± 5.2 × 10 6 amoA A O A gene copies g −1 soil.In addition, 2.2 × 10 7 ± 1.2 × 10 6 amoA gene copies g −1 soil and 1.0 × 10 6 ± 7.7 × 10 4 amoA gene copies g −1 soil of the AOB, and comammox clade B communities were detected, respectively (Fig. 2 A).In contrast, the CS had similar A O A and A OB abundances (1.3 × 10 7 ± 2.1 × 10 6 amoA gene copies g −1 soil and 1.1 × 10 7 ± 1.7 × 10 6 amoA gene copies g −1 soil, r espectiv el y) and less comammox clade B bacteria (1.3 × 10 6 ± 2.8 × 10 5 amoA gene copies g −1 soil) (Fig. 2 A).The total net nitrification potential and octyne resistant (i.e.A O A contribution) activity fraction were determined in both soils.Total net nitrification potential was 2-fold and 10-fold higher than octyne resistant activity in AS and CS, respectively (Fig. 2 B).In the AS the octyne resistant fraction contributed about ∼35% of the total activity, compared to less than 7% in the CS (Fig. 2 B).Overall, the AS had higher total net nitrification potential and octyne resistant activity, along with a larger total ammonia oxidizer and A O A abundances.

Differences in NI efficacy between soils
Due to the low solubility of nitr a pyrin, MHPP, and MHPA in water, these NIs were dissolved in DMSO for application and therefore the effect of DMSO (0.035% v/v) alone on the net nitrification potential was also assessed in both soils.DMSO had no significant effect on the net nitrification potential ( Supplementary Fig. 6 , P = .106in the AS and P = 0.747 in the CS).In addition, phenylacetylene (1 mM) was used as a total net NI and caused 95.4 ± 1.8% and 93.2 ± 1.4% inhibition in the AS and CS, r espectiv el y (Fig. 3 ).The AOB and comammo x-selecti ve inhibitors, octyne and ATU, were used to differentiate between bacterial and archaeal ammonia oxidizer contribution to net nitrification activity in both soils.As expected, octyne and ATU sho w ed similar nitrification inhibition patterns across both soils.In the AS, octyne (4 μM), and ATU (100 μM) inhibited 65 ± 2.3% and 51 ± 2.9% of the net nitrification potential, r espectiv el y, indicating that the r emaining activity was A O A driv en.In contr ast, both compounds inhibited ∼94% of the net nitrification potential in the CS, highlighting that the ammonia oxidation activity in this soil was almost exclusiv el y AOB/comammo x dri ven.
The efficacy of all tested SNIs and BNIs was significantly higher in the CS than in the AS (Fig. 3 ).Ov er all, the efficacy of SNIs sho w ed greater variability in the AS with inhibition efficacies between 20% and 75%, while in the CS all SNIs were highly effective, inhibiting 80%-100% of nitrification.Among the three SNIs tested, nitr a p yrin w as the most effective inhibitor in both soils, with an av er a ge of 75 ± 5.2% net nitrification inhibition in the AS and 100 ± 0.7% in the CS.DMPP inhibited net nitrification potential by 47 ± 1.6% in the AS and 90 ± 1.1% in the CS.DCD was the least effective SNI tested in both soils, with 20 ± 5.4% and 80 ± 9% of inhibition in the AS and the CS, r espectiv el y (Fig. 3 ).
Among the three BNIs, MHPA had the highest inhibition efficacy in both soils with 54 ± 1.8% and 72 ± 1.9% of inhibition in the AS and CS, r espectiv el y.That w as follo w ed b y limonene with 40 ± 4.3% of inhibition efficacy in the AS and 68 ± 2.9% in the CS.MHPP was the least effective BNI tested in both soils, inhibiting 35 ± 6.4% in the CS, and having almost no effect on the net nitrification potential in the AS with 2.7 ± 1.3% of inhibition (Fig. 3 ).Inter estingl y, all inhibitors (selective inhibitors , SNIs , and BNIs) had a significantly higher efficacy in the CS at pH 6.12 ± 0.09 than in the AS at pH 8.50 ± 0.02.

NI efficacy is not exclusi v ely dri v en by soil pH
As soil pH was significantl y differ ent betw een the tw o analysed soils and has pr e viousl y been r eported to affect NI efficacy (K eeney 1980 , Bac htse v ani et al. 2021 , Yin et al. 2023 ), we hypothesized that the observed differences in the efficacy of the BNIs and SNIs are predominantly driven by soil pH.T herefore , short-term soil pH manipulations were performed for each soil and the efficacy of all NIs at the modified pHs was determined.Inter estingl y, pH modifications by almost one pH unit (from 8.5 to pH 7.18) significantl y incr eased the net nitrification potential in the AS, while a decrease by a second pH unit (from 8.5 to pH 6.15) did not significantly impact net nitrification potential ( Supplementary Fig. 7 A).In contrast, pH modifications by one pH unit (from 6.15 to pH 7.6) did not significantly impact the overall net nitrification potential in the CS but pH modifications by a second pH unit (from 6.12 to pH 8.5) resulted in a significant decrease in the net nitrification potential ( Supplementary Fig. 7 B).Importantly, net nitrification was still observed in all pH modified soils ( Supplementary Fig. 7 ).In each soil, the pH modifications did not have a significant effect on the percentage of inhibition in the phenylacetylene or DMSO controls (Fig. 4 ).For all the tested NIs, the percentage of inhibition at a given soil pH was determined relative to the respective net nitrification potential at that pH.
In the AS, when the pH was reduced from 8.5 to 6.15, the A O Adriven net nitrification potential decreased, as observed by a significant increase in the efficacy of octyne (24%, P = < .01,Fig. 4 A).Similarly, the efficacy of MHPP and MHPA also significantly increased (34% and 42%, respectively) when the soil pH was reduced from 8.5 to 6.15 (Fig. 4 A).Interestingly, the efficacy of octyne and MHPP ( r = 0.80 P = < .01)as well as octyne and MHPA ( r = 0.92, P = < .001)were positively correlated.While limonene follo w ed a similar pattern where the highest efficac y w as observed at pH 6.15, its efficacy across soil pHs did not significantly correlate with the efficacy of octyne ( r = 0.32, P = .4,Supplementary Fig. 8 ).In contrast, the efficacy of the three SNIs tested was not significantly affected  by the soil pH modifications in the AS.In the CS, raising the soil pH from 6.12 to 8.5 did not have a significant effect on the efficacies of any of the selective inhibitors , BNIs , or SNIs (Fig. 4 B).

BNI effecti v e concentr a tions
An important c har acterization of NIs is the concentr ation or field a pplication r ate necessary to effectiv el y inhibit soil nitrification.
Due to the differences in soil physicochemical properties, ammonia oxidizer community, and net nitrification activity between the AS and CS, we hypothesized that the EC 50 and EC 80 values of the three tested BNIs would differ significantly between the soils .T his was the case for the EC 50 values of MHPP (460.3 ± 31.6 μM in the AS and 359.4 ± 18.6 μM in the CS) and MHPA (103.5 ± 10.1 μM in the AS and 34.7 ± 10.7 μM in the CS), which both had signif- icantly lo w er EC 50 values (i.e. higher efficac y) in the CS (Fig. 5 , Supplementary Table 3 ).Ho w e v er, the EC 50 v alues of limonene were not significantly different between the AS and the CS.Among the three BNIs tested, MHPP again had the lo w est efficac y (highest EC 50 and EC 80 values) in both soils.Notably, the EC 80 values of all three tested BNIs were not significantly different between soils (Fig. 5 , Supplementary Table 3 ).

Efficacy of NIs in two contrasting agricultural soils
While variations in NI efficacy between soils have previously been reported (Lu et al. 2019, Zhou et al. 2020, Bachtsevani et al. 2021, Ma et al. 2021, Wang et al. 2021, Yin et al. 2023 ), the extent to which soil pH, soil physicochemical properties, and microbial community affect the efficacy of SNIs and BNIs r emains lar gel y unexplor ed.Consequentl y, in this study, the efficacy of three BNIs: MHPP, MHPA, and limonene, and three SNIs: nitrapyrin, DCD, and DMPP, in two contrasting agricultural soils were in vestigated.T he efficacy of all the NIs tested here was higher in the CS (pH 6.12 ± 0.09) than in the AS (pH 8.50 ± 0.02) (Fig. 3 ).While the CS had similar abundances of A O A and A OB/comammox, the net nitrification potential was AOB/comammox driven with A O A activity contributing < 7% (Fig. 2 ).In contrast, the abundance of A O A in the AS soil was higher than the abundances of AOB/comammox, and contributed ∼35% of the total net nitrification potential.Although the incubation conditions used in this study might have altered the proportion of AOB/comammox and AOA activities compared to the more competitive interactions observed in whole soil incubations (Prosser andNicol 2012 , Rütting et al. 2021 ).This trend of higher NI efficacy in more neutral soils, where net nitrification potential is AOB driv en, compar ed to AOA dominated or more ASs has been observ ed pr e viousl y (Guardia et al. 2018, Lu et al. 2019, Lei et al. 2022 ).
In our study, the SNI nitr a p yrin w as the most effectiv e NI r egardless of soil type e v en when a pplied at a low field a pplication rate (3.5 μM, 0.35% of the N applied) (Fig. 3 ).This is in line with other studies comparing the efficacy of SNIs, wher e nitr a pyrin is the most effective inhibitor regardless of soil type (Hayden et al. 2021, Lan et al. 2023 ), as long as it is applied at a sufficient rate (Zhou et al. 2020 ).In a gr eement with pr e vious r eports (Guardia et al. 2018, Zhou et al. 2020, Lan et al. 2023 ), DMPP was more effective than DCD, even though the applied DCD concentration was 10 times higher (10 μM and 100 μM, r espectiv el y).
Similar to the tr end observ ed for the SNIs, the efficacy of all BNIs tested here was also higher in the CS than in the AS, which is in a gr eement with pr e vious studies, whic h observ ed gr eater BNI inhibition efficacy in acidic soils compared to ASs (Nardi et al. 2013, Lu et al. 2019, Subbarao et al. 2021 ).Among the BNIs tested, MHPA had the highest efficacy in both soils follo w ed b y limonene (Fig. 3 ).Notably, this is the first study assessing the efficacy of MHPA and limonene in a gricultur al soils.MHPP, a commonly tested BNI, sho w ed the lo w est inhibition efficac y of all the BNIs and SNIs tested in both soils (Fig. 3 ).Although MHPP has sho wn higher efficac y compared to DCD in a soil of similar pH as the CS (Nardi et al. 2013 ), a lower efficacy of MHPP has also been r eported pr e viousl y in ASs (Lu et al. 2019, He et al. 2023 ).Together, our results support the majority of previous studies, which also report higher BNI efficacy in lower-pH soils (Lu et al. 2019, Subbarao et al. 2021 ).
Ov er all, the differ ences in NIs efficacy acr oss soils observ ed here and in several other studies (Shi et al. 2016, Zhou et al. 2020, Lu et al. 2022, Lan et al. 2023 ), suggest that the efficacy of SNIs and BNIs in soils is affected by soil pH (Shi et al. 2016, Lu et al. 2022 ), the v arying physicoc hemical pr operties of soils (Marsden et al. 2016, McGeough et al. 2016, Guardia et al. 2018 ) and the micr obial comm unity abundance and composition (Zhou et al. 2020, Bac htse v ani et al. 2021 ).

The role of pH on NI efficacy
Significant differences in NI efficacies between soils of differing pH wer e observ ed her e and hav e been pr e viousl y r eported (Shi et al. 2016, Lu et al. 2019, Zhou et al. 2020, Bac htse v ani et al. 2021 ).Ho w e v er, a r ecent study concluded that pH is not the main driving factor of the efficacy of the SNI, DMPP in acidic a gricultur al soils (Oliv eir a et al. 2022 ).Ther efor e, we conducted short-term soil pH manipulation experiments to determine if soil pH dir ectl y acts as the main driver of the differences observed in NI efficacy.Contrary to our hypothesis, the efficacy of most SNIs and BNIs was not significantly affected by short-term soil pH manipulations (Fig. 4 ).This suggests that soil pH alone is not the main driver of the differ ences observ ed her e in NI efficac y betw een the AS and CS.Instead, other soil physicochemical properties such as organic matter, clay content, and the concentration of specific micronutrients ma y ha v e a mor e dir ect short-term effect on the efficacy of NIs in soils.

The role of soil physicochemical properties on NI efficacy
Pr e vious studies have identified that sorption of NIs to organic matter can affect NI efficacy in soils (Marsden et al. 2016, Guardia et al. 2018 ).Under the incubation conditions used here, it is likely that the interactions of NIs with the dissolved organic matter are facilitated and ther efor e, a lo w er NI efficac y is expected in the soil with high organic content.The AS did not only have higher pH, but also higher total N, total C and CEC (Table 1 ), suggesting that the lo w er efficac y of all NIs tested may be due to higher sorption of the NIs to the soil matrix (Fig. 3 ).Pr e viousl y, a str ong corr elation between NI half-life time and total N across nine different soil types was observ ed, wher e for instance, the half-life time of DCD was reduced in a soil with similar total N and CEC as the AS in our study (McGeough et al. 2016 ).Aside fr om or ganic matter , clay particles can also reduce the bioavailability of NIs through sorption.In our study, the clay contents of the AS and CS wer e compar able (24.2% in the AS and 19.4% in the CS).Among the NIs tested here, MHPP had the lo w est efficac y in both soils (Fig. 3 ).This is in a gr eement with the low efficacy of MHPP, observed in a soil with similar clay content as that of the two soils tested here (Lu et al. 2019 ), and in contrast to the higher efficacy of MHPP in a soil with a very low clay content (2.1%) (Nardi et al. 2013 ).Altogether, these observations suggest that high organic matter, CEC, and clay content dir ectl y affects the efficacy of NIs by reducing their mobility, half-life time or bioavailability.
In addition, the availability of micronutrients, such as iron and copper, also influences ammonia oxidation (Reyes et al. 2020, Shafiee et al. 2021 ).While the exact mode of action of many NIs is unclear, it has been proposed that se v er al NIs (e.g.DCD and DMPP) can act as copper chelators ( Supplementary Table 4 ) suggesting their efficacy is affected by soil copper content.The significantly higher EDTA-extractable copper content in the AS (Table 1 ) is one possible explanation for the high net nitrification potential activity (Fig. 2 B), as well as for the gener all y lo w er NI efficac y observed in the AS (Fig. 3 ).In fact, a negativ e corr elation between copper content and nitrification inhibition by DCD has been pr e viousl y reported (McGeough et al. 2016 ).Likewise, the low organic matter content, CEC, and copper content in the CS ma y ha ve facilitated the higher inhibition efficacy observed for all NIs tested.Howe v er, the individual effects of soil organic matter, clay, and copper content on NI efficacy can not be determined from this study as all thr ee soil pr operties cov aried, wer e significantl y higher in the AS soil, and likely have ad diti ve effects (Table 1 ).

Role of the soil microbial community on NI efficacy
Both SNIs and BNIs sho w ed higher efficac y in the CS, which had fewer ammonia oxidizers in total, a lower net nitrification rate, and the AOB/comammox communities contributed to > 90% of the net nitrification potential (Figs 2 and 3 ).Regarding the role of ammonia oxidizer community abundance and diversity on NI efficacy in soils, a selective inhibition of AOB over AOA has been proposed (Zhou et al. 2020, Bachtsevani et al. 2021, Nair et al. 2021 ).Notably, despite the exceptionally high relative abundance of A O A ( ∼22%) and the 1.9-fold higher copy number of A O A amoA genes than AOB amoA genes in the AS, only 35% of total activity was octyne resistant (i.e.A O A driven).A recent meta-analysis on the effects of SNIs on the microbial community concluded that the application of DMPP, DCD, and nitr a pyrin significantl y r educed the gene and transcript abundances of AOB amoA , rather than A O A amoA (Yin et al. 2023 ).This selective effect of DMPP has also been observed when tested with ammonia-oxidizing pure cultur es (P a padopoulou et al. 2020 ) and in soils where nitrification activity was pr edominantl y AOB driv en (Bac htse v ani et al. 2021 ).The higher NI efficacy in the CS observed here (Fig. 3 ) may further indicate that AOB/comammox are more susceptible to SNIs and BNIs than A O A. Ho w e v er, the AS and CS harbour distinct AOB/comammox communities (Fig. 1 B).These observed differences in nitrifying communities may be driven by long-term differ ences in pH, whic h taken together may also contribute to the differences in NI efficacy observed.
Inter estingl y, lo w ering the pH in the AS caused a significant increase in the efficacy of the selective inhibitor, octyne, suggesting that A O A activity decreased with decreasing pH (Fig. 4 ).Decreasing A O A activity at lo w er pH is in contrast to other studies whic h hav e observ ed that nitrification is most often A O A-driven in low-pH soils (Prosser andNicol 2012 , Zhang et al. 2012 ).The A O A community in the AS was dominated by the genus TA-21 from the NS-δ-2.1 clade (Fig. 1 and Supplementary Fig. 3 ).This A O A clade has been pr e viousl y found in soils (Lee et al. 2023 ), but significant contributions to nitrification have not been reported pr e viousl y and little is known about the pH pr efer ences or physiology of A O A in the NS-δ-2.1 clade.Ho w e v er, under the incubation conditions used here, the genus TA-21 (NS-δ-2.1 clade) was seemingl y r esponsible for the octyne-r esistant fr action observ ed in the AS as it makes up 99.2 ± 0.2% of the A O A community in this soil (Fig. 1 and Supplementary Fig. 3 ).
Notably, in addition to octyne, MHPP and MHPA also displayed a seemingly AOB/comammox selective inhibition beha viour.T his is illustrated by the strong positive correlation between their efficacy and the efficacy of octyne in the AS across the pHs tested ( Supplementary Fig. 8 ).A similar indication that MHPP is a potential selective inhibitor of AOB/comammox was r ecentl y observed with ammonia-oxidizing pure cultures (Kaur-Bhambra et al. 2022 ), where double the inhibitor concentration was required to inhibit soil A O A cultures compared to soil AOB cultures.In contr ast, dir ect e vidence fr om ammonia-oxidizing pur e cultur es is r equir ed to confirm the selectiv e inhibition potential of MHPA.
Aside from the different ammonia oxidizer communities, other members of the soil micr obial comm unity may also influence NI efficacy.Higher soil microbial biomass has been previously suggested as an indicator for higher microbial NI degradation (McGeough et al. 2016 ).Additionall y, high CO 2 r espir ation r ates have been observed for the commonly used SNIs DMPP and DCD (Guardia et al. 2018 ).Ther efor e, it is expected that organic compounds such as BNIs will also be susceptible to microbial uptake, tr ansformation, and degr adation whic h may affect NI efficacy (Marsden et al. 2016 ).While the CS and the AS harbour distinct microbial communities, the AS has significantly higher microbial N and C biomass than the CS (Fig. 1 and Table 1 ), which is an additional explanation of the ov er all lo w er NI efficac y in the AS.In support of this hypothesis, significantly higher CO 2 production was observed in the MHPP and MHPA treatments in the AS and the CS, compared to the control without BNI addition ( Supplementary Fig. 9 ).Although the C added by the application of BNIs (200 μM equivalent to 0.18 mg C g −1 soil) r epr esented onl y 0.8% and 2% of the total C content in the AS and the CS (Table 1 ), r espectiv el y, mor e C as CO 2 was produced than what can be attributed to the BNIs alone .T hese observations suggest that the addition of BNIs might indir ectl y stim ulate other soil microbial pr ocesses, whic h could subsequentl y affect the micr obial uptake, tr ansformation, degr adation, and ultimatel y the efficacy of NIs in soils.

Effecti v e concentr a tions of BNIs in tw o contr asting agricultur al soils
While there has been an increasing interest in the applicability of BNIs in a gr oecosystems (Lu et al. 2019, Wang et al. 2021, Lan et al. 2022, 2023 ), this study is the first to report the efficacy (EC 50 and EC 80 values) of three BNIs: MHPP, MHPA, and limonene in agricultural soils.All three BNIs tested here follo w ed a log-logistic inhibition model (Fig. 5 ).Inter estingl y, while the EC 50 v alues of MHPP and MHPA differed between soils, the EC 80 values of the three BNIs were similar between the AS and the CS.
Pr e vious r eports of EC 50 v alues for BNIs ar e scar ce, y et the EC 50 values for the three BNIs tested here are within similar ranges to the values previously reported.With an EC 50 34.7 ± 10.7 μM in the CS, the EC 50 of MHPA is comparable to the value previously determined with the AOB, Nitrosomonas europaea (19.5 μM) (Gopalakrishnan et al. 2007 ).Similarly, the EC 50 of limonene in the CS (96.8 ± 30.5 μM), is only 2.5-fold higher than the value previously reported for N. europaea (38 μM) (Ward et al. 1997 ).A more recent study reported EC 80 values of MHPP in several pure A O A and A OB cultures, and ranged between 78.8 and 647.3 μM for soil A O A, and between 46.8 and 341.3 μM for soil AOBs (Kaur-Bhambra et al. 2022 ).While the EC 80 values of MHPP in the CS (711.9 ± 48 μM) and the AS (870.5 ± 81.7 μM) are higher than the values reported for soil AOB cultur es, our EC 80 v alues ar e v ery similar to the highest effectiv e concentration of MHPP for soil A O A cultur es.Ov er all, higher EC 50 and EC 80 values for MHPA, MHPP, and limonene were observed in the AS and in the CS than in pr e vious pur e cultur e studies, whic h is likely a result of a more complex soil microbial community and interactions with soil properties.
Notabl y, a selectiv e or mor e pr onounced inhibition of soil AOB was r ecentl y determined for MHPP with A O A and A OB pure culture isolates (Kaur-Bhambra et al. 2022 ).Similarly, MHPP and MHPA had significantly higher efficacy (lower EC 50 values) in the soil where the net nitrification potential was AOB/comammox driven (CS).This higher AOB/comammox-sensitivity to MHPP and MHPA, together with the significant positive correlation with the efficacy of the selective inhibitor octyne ( Supplementary Fig. 8 ), suggest that these two BNIs are potentially selective AOB/comammox inhibitors.Ne v ertheless, further r esearc h on the inhibition patterns of pure ammonia oxidizer cultures by MHPA, and additional evidence of A O A-driven nitrification activity in the presence of MHPA and MHPP is r equir ed to confirm the selective inhibition of these BNIs (Taylor et al. 2013 ).
In addition to the degree of inhibition, chemical structure can also play a role in the mode of action or selectivity of NIs.It has been proposed that specific chemical structures favour inhibition of nitrification, and thus compounds with similar structures to known NIs are often expected to result in similar degrees of inhibition (White 1988, Subbarao et al. 2013 ).MHPP and MHPA are very similar phenylpropanoids, yet large differences in their EC 50 v alues wer e observ ed (Fig. 5 , Supplementary Table 4 ).In fact, similar differences in NI efficacies due to specific chemical structures hav e been observ ed pr e viousl y.For example, a c hange in the configuration of one of the isomers of the NI br ac hialactone r educed its nitrification inhibition activity (Egenolf et al. 2020 ).Despite being an inhibitor of all known ammonia oxidizers, phenylacetylene inhibits A O A and A OB thr ough a differ ent mode of action (Wright et al. 2020 ).Pr e vious studies hav e also shown that the degr ee of inhibition of n -alkynes, with different chain lengths, varies between A O A and A OB (Taylor et al. 2013, 2015, Wright et al. 2020 ).These observations indicate that besides the effect of soil pH, soil physicoc hemical pr operties, and the micr obial comm unity, the r ole of the c hemical structur e in NI efficacy is of r ele v ance particularl y amidst the complexity of natural environments such as soils.

Conclusions
In this study w e sho w ed that se v er al SNIs and BNIs are resilient to short-term soil pH c hanges, whic h indicates that besides soil pH, other soil physicochemical properties, the abundance and composition of ammonia oxidizers, and NI chemical structure play a role in the efficacy of NIs.Additionally, we showed the first evidence of two BNIs (MHPP and MHPA) as putative selective inhibitors and that MHPA and limonene inhibition was equal and, in some cases, superior to the inhibition caused by commonly used SNIs.Notably, the EC 80 values of MHPA were similar to currently used DCD concentr ations in a gricultur al soils (10% N a pplied), highlighting that some BNIs could be an effective alternative to SNIs.In summation, if BNIs are to be mor e widel y adopted as a natur al alternativ e to reduce N losses, it is k e y to understand the extent to which soil biotic and abiotic factors affect their efficacy to establish an optimal a pplication r ate in a gricultur al systems.

Figure 1 .
Figure 1.Soil microbial community composition in the AS and the CS.(A) 16S rRNA gene-based microbial community composition depicting the top taxa with ≥2% mean r elativ e abundance at each site (B) amoA microbial community composition depicting the top amoA taxa with ≥5% mean r elativ e abundance.Taxa with mean r elativ e abundance below 2% (for 16S rRNA gene) or 5% (for amoA ) are clustered as 'other'.Samples collected in spring 2022 ( n = 4, error bars represent standard error) ( * P < .001).Absolute abundances of the amoA genes are shown in Supplementary Fig.3.

Figure 2 .
Figure 2. Abundance of ammonia oxidizers, total net nitrification potential activity, and A O A contribution to the net nitrification potential.(A) Absolute abundance of ammonia oxidizers in the AS and the CS.(B) Total net nitrification potential activity and A O A contribution to the net nitrification potential.Lo w er case letters depict significant differences within and betw een soils.Samples collected in spring 2022.Err or bars r epr esent the standard error in each case ( n = 4, P < .05).

Figure 4 .
Figure 4. Effect of soil pH manipulation on the efficacy of SNIs and BNIs in the (A) AS (pH 8.50), (B) CS (pH 6.12) PA: 1 mM Phen ylacetel yne, 4 μM Octyne, 100 μM ATU, 3.5 μM Nitr a pyrin, 100 μM DCD, 10 μM DMPP, 200 μM MHPP, 200 μM MHPA, and 200 μM limonene.Samples collected in spring 2022.The median is depicted as the middle hinge in the boxplots.Upper and lo w er hinges r epr esent the first and third quartile .T he length of the whispers is determined by the largest and the smallest value in the dataset that are within 1.5 times the interquartile range.( n = 3, * P < .05,* * P < .01).

Figure 5 .
Figure 5. Log-logistic model fitting to estimate the EC 50 and EC 80 values of three BNIs in the AS and CS.(A) and (D) log-logistic model for MHPP, (B) and (E) for MHPA, and (C) and (F) for limonene.(A), (B), and (C) correspond to the inhibitors assessed in the AS, and (D), (E), and (F) in the CS.The corr esponding EC 50 , EC 80 v alues (dashed lines), and hill coeficient, with their r espectiv e standard err ors, as well as the R 2 ar e displayed for eac h BNI in each soil.

Table 1 .
Soil parameters of the AS and the CS.CEC: cation exchange capacity; soil C:N: molar based soil C:N ratio; DOC: dissolved organic carbon; TDN: total dissolv ed nitr ogen; C mic : micr obial biomass C; N mic : microbial biomass N; and microbial C:N: molar based microbial C:N ratio.Samples collected in spring 2022 ( n = 4, SE: standard error).