The differential assimilation of nitrogen fertilizer compounds by soil microorganisms

Abstract The differential soil microbial assimilation of common nitrogen (N) fertilizer compounds into the soil organic N pool is revealed using novel compound-specific amino acid (AA) 15N-stable isotope probing. The incorporation of fertilizer 15N into individual AAs reflected the known biochemistry of N assimilation—e.g. 15N-labelled ammonium (15NH4+) was assimilated most quickly and to the greatest extent into glutamate. A maximum of 12.9% of applied 15NH4+, or 11.7% of ‘retained’ 15NH4+ (remaining in the soil) was assimilated into the total hydrolysable AA pool in the Rowden Moor soil. Incorporation was lowest in the Rowden Moor 15N-labelled nitrate (15NO3−) treatment, at 1.7% of applied 15N or 1.6% of retained 15N. Incorporation in the 15NH4+ and 15NO3− treatments in the Winterbourne Abbas soil, and the 15N-urea treatment in both soils was between 4.4% and 6.5% of applied 15N or 5.2% and 6.4% of retained 15N. This represents a key step in greater comprehension of the microbially mediated transformations of fertilizer N to organic N and contributes to a more complete picture of soil N-cycling. The approach also mechanistically links theoretical/pure culture derived biochemical expectations and bulk level fertilizer immobilization studies, bridging these different scales of understanding.


Introduction
Nitrogen (N) fertilizers are essential to modern food production and 105 Tg N fertilizers were used in 2016 (FAO 2019 ).It is estimated, ho w e v er, onl y 17% of N applied to crops ultimately supports human nutrition, with the remainder being lost to the environment during food production and processing (Leach et al. 2012, Fowler et al. 2013 ).This brings the low nutrient use efficiency of the human food-chain into critical focus .T he interaction of applied fertilizer N with the soil N-cycle, and influence on soil organic N, r epr esents an important determinant of the fate of fertilizer N, the N balance of soil and e v entual efficiency of production systems.Major gaps exist regarding the biological processing of N fertilizers in soils, particularly the routes and proportions of conversion into soil organic N.
Processing of N fertilizers has traditionally been quantified using isotope pool dilution to determine rates of N mineralization, immobilization, and nitrification in soils.Ho w e v er, e v en in agricultur al soils, N stor ed in or ganic forms dominates inor ganic N (Dungait et al. 2012 ).This large and heterogeneous soil N pool still underpins soil N dynamics and the supply of N to micr oor ganisms , plants , and loss pathways (in some cases providing 30%-50% of the inorganic N for crop uptake; Macdonald et al. 1997, Murphy et al. 2000, Dungait et al. 2012 ).In order to provide a new perspective on the biomolecular fate and partitioning of different common N fertilizer compounds into soil organic N, we herein describe the application of compound-specific amino acid (AA) 15 N-stable isotope probing (SIP) to investigate N-cycling into the soil protein pool (Charteris et al. 2016 ).The a ppr oac h combines compound-specific gas c hr omatogr a phy-combustion-isotope r atio mass spectrometry (GC-C-IRMS) and 15 N-SIP in the metametabolome of the whole soil system (Knowles et al. 2010 ; or other complex media, e.g.river water; Mena-Riv er a et al. 2022 ) and is essentially a targeted 15 N fluxomics a ppr oac h (Cascante and Marin 2008 ).The soil protein pool is the largest (20%-50% of total soil N), and ar guabl y most important, identifiable class of soil organic N (Stevenson 1982 ).Microbially mediated N transformations through the AA glutamate (Glu; Santero et al . 2012 ) r epr esent the gatew ay betw een the inorganic and organic soil N pools ( Supplementary Fig. S1 ).The extent to which fertilizer N is incor por ated into soil pr otein has implications for its tempor al av ailability to plants and loss pathwa ys [e .g. nitrate (NO 3 − ) leaching, ammonia (NH 3 ) volatilization, and nitrous oxide (N 2 O) emissions] and we can now r e v eal distinct differ ences between thr ee differ ent fertilizer N compounds in two differ ent gr assland soils.

Materials and methods
We explore whether differences exist in the processing of three different 10 atom % 15 N-labelled fertilizer N compounds-potassium nitrate (K 15 NO 3 ), ammonium chloride ( 15 NH 4 Cl), and urea (CO( 15 NH 2 ) 2 ); henceforth r eferr ed to as the 15 NO 3 − treatment; the 15 NH 4 + treatment, and the 15 N-U treatment, r espectiv el y-in two differ ent soils, identified by site-Rowden Moor (RM) and Winterbourne Abbas (WA)-using soil microcosms (Table 1 ).

Sites and soil sampling
Soil was sampled to a depth of 15 cm along a random W transect from plot six of RM experimental site at Rothamsted Research North W yke, Devon, UK (50 The RM site was a long-term grassland ( > 40 years) dominated by Lolium spp.interspersed with Cynosurus , Festuca , Agrostis , Holcus , and Dactylis spp.It had been grazed by cattle for around 25 years and had r eceiv ed ∼200-250 kg N ha −1 year −1 as cattle slurry.The WA site, on the other hand, had been used for spring cropping before being converted to a grass ley ( Lolium perenne and Trifolium repens ) and used for dairying with a mobile milking parlour for 2 years prior to sampling.The ley was fertilized with 40 kg N ha −1 (pr e viousl y as ammonium sulfate [(NH 4 ) 2 SO 4 ] and then as sulfurcoated urea [CH 4 N 2 O]) every 40 days from spring until the start of the 'closed period' on 15th September.whic h pr ohibits N fertilizer a pplication on gr asslands in nitr ate vulner able zones (Defr a 2013 ).The samples of each soil were combined in equal weights and homogenized to produce a pooled soil sample for each site .P ooled samples were air-dried to allow sieving to < 2 mm and then double distilled water (DDW) added to attain 50% water holding capacity (WHC).

Incubations
Each experimental unit consisted of 10 g soil at 50% WHC contained in a 10-cm high by 2-cm diameter glass tube.Maintenance of the soil at 50% WHC was selected to pr e v ent leac hing and the tubes were fitted with furnaced and pierced aluminium foil lids to minimize volatile and e v a por ativ e losses.All incubations wer e carried out in triplicate so ther e wer e thr ee tubes for eac h time point of each treatment.Incubation treatments and periods are summarized in Table 1 .Treatments were injected into the soil and distributed over the full core depth.Incubations were halted at the r equir ed time by immersion in liquid nitrogen (N 2 ) and stored at −20 • C prior to fr eeze-drying.Whole fr eeze-dried soil cor es wer e finel y gr ound and homogenized using a pestle and mortar and stored in sealed 28 ml vials at −20 • C.

Extr action, isola tion, and deriv a tiza tion of hydrolysable AAs
Freeze-dried and ground incubation soil samples (100 mg) with an added internal standard of 100 μl norleucine in hydr oc hloric acid (400 μg ml −1 Nle in 0.1 M HCl) were hydrolyzed with 5 ml 6 M HCl at 100 • C for 24 hours under an atmosphere of N 2 (Fountoulakis andLahm 1998 , Roberts andJones 2008 ).Acid hydr ol ysis extracts both free and proteinaceous AAs as well as catalyzing the breakdown of living microbial biomass (Roberts and Jones 2008 ).The r elativ el y harsh conditions ar e necessary for the cleav a ge of peptide bonds between hydrophobic residues [e.g.isoleucine (Ile), leucine (Leu), and valine (Val)], but also result in the deamination of aspar a gine (Asn) to Asp and glutamine (Gln) to Glu and Table 1.b Not all time-points analysed for AAs, only 3 and 6 hours and 2, 4, 16, and 32 da ys .
c Not all time-points analysed for AAs, only 1.5, 3, and 12 hours and 2, 8, and 32 da ys .
the complete destruction of cysteine (Cys) and tryptophan (Trp; Fountoulakis andLahm 1998 , Roberts andJ ones 2008 ).T he technique may also partiall y destr oy serine (Ser; ca.10% loss), threonine (Thr; ca. 5% loss), and tyrosine (Tyr; loss depends on le v el of trace impurities in hydrolysis agent; Fountoulakis and Lahm 1998 ) and has the potential to hydr ol yse AA c hains fr om nonpr oteinaceous sources, such as peptidoglycan, resulting in an overestimation of some AAs, mostly alanine (Ala), Glu, glycine (Gly), and lysine (Lys; Roberts and Jones 2008 ).The technique is, howe v er, consider ed the most reliable method for determining the total protein content of soils (Roberts and Jones 2008 ) and as such, it is reasonable to equate total hydr ol ysable AA concentr ations to the size of the soil protein pool.The hydr ol ysis is performed under N 2 as the presence of O 2 can induce the thermal breakdown of hydroxyl-and sulfur-containing AAs [e .g. Ser, T hr, Tyr, and methionine (Met); Roberts and Jones 2008 ].Hydr ol ysates wer e collected by centrifugation, dried at 60 • C under a stream of N 2 , and stored at −20 • C under 1 ml 0.1 M HCl.Cation-exchange column chromatogr a phy with acidified Do w ex 50WX8 200-400 mesh ion-exchange resin was used to isolate AAs from the hydrolysates (Metges and Petzke 1997 ).Finally, the hydrolysed soil AA mixtures were converted to their N -acetyl, O -isopropyl deri vati ves for analysis (Corr et al. 2007 ).

Instrumental analyses
Bulk soil percentage total N (% TN) and δ 15 N analyses were carried out by elemental analysis-isotope ratio mass spectrometry (EA-IRMS) at the Lancaster node of the Natural Environment Research Council Life Sciences Mass Spectrometry Facility (NERC LSMSF).AAs as their N -acetyl, O -isopropyl deri vati ves were quantified by comparison with the Nle internal standard using gas c hr omatogr a phy-flame ionization detection (GC-FID).The Nacetyl, O -isoprop yl AAs w ere identified by their known elution order and by comparison with N -acetyl, O -isopropyl derivatized-AA standar ds.Data w er e acquir ed and anal ysed using Clarity c hr omatogr a phic station for Windows by DataApex.The δ 15 N values of individual AAs as their N -acetyl, O -isopr opyl deriv ativ es wer e determined using GC-C-IRMS.Data were acquired and analysed using Isodat NT 3.0 (Thermo Electr on Cor por ation).Bulk soil percentage total C (% TC) analyses were carried out on a Eurovector EA3000 elemental analyser.

Sta tistical informa tion and calcula tions
AA plateau 15 N values and % 15 N R incorporations were determined by curve fitting with a simple exponential equation using Genstat ® statistical software for biosciences (19th edition, VSNI): where α is the plateau AA 15 N value or % 15 N incorporation, α + β is the AA 15 N value or % 15 N incorporation at t = 0 (which is 0 by definition for these parameters) and θ is the rate at which AA 15 N values or % 15 N incorporations increase.In addition, due to the temporal trend of Glx 15 N values in the 15 NH 4 + and 15 N-U tr eatments, these r esponses wer e also fitted with a critical exponential r egr ession: where α is again the plateau AA 15 N value or % 15 N incor por ation and α + β is again the AA 15 N value or % 15 N incor por ation at t = 0 (again 0 by definition).The balance between γ (increase) and θ (decay) controls the height and positioning ( x value) of the peak in the critical exponential function, where γ can be used to assess the rate of increase in AA 15 N values or % 15 N incorpor ated (lar ger γ = faster, although comparison between γ values becomes less clear where θ values differ).Lack of error bar overlap between mean 15 N values at t = 32 days was used as an indicator of significant statistical difference between final AA 15 N values .T his approach was used because formal statistical testing would confirm a significant statistical difference between means with separated error bars, and would, r ather, onl y be useful to determine whether ther e wer e an y statisticall y significant differ ences between means with some error bar ov erla p.This further le v el of inspection was not deemed to add sufficient value to the inter pr etation of this work as the complex statistical modelling r equir ed to rigor ousl y determine the statistical difference between plateau 15 N values (using constrained curve fitting) would not be proportionate for the additional information obtained.Simple t -tests or analysis of variance using final t = 32-day values would be based on very small datasets and would ther efor e onl y pr ovide confirmation wher e err ors bars ar e separ ated, whic h can already be observed.
The percentage of the applied 15 N incorporated into each AA is as follows: where E is the 15 N enrichment of the AA following application of a 15 N-labelled substrate (taking into account the moles of N in the AA per gram of sample and the excess atom fraction of the AA after incubation, compared with the control).The percenta ge of r etained 15 N [based on n E ( 15 N) P/C , the excess moles of 15 N pr esent/r etained per gram bulk sample at time, t ] incorporated into each AA at time, t is as follows: Finall y, the percenta ge of a pplied/r etained 15 N incor por ated into ne wl y synthesized soil pr otein w as determined b y summing the results of Equations ( 3) or (4) , respectively, for individual AAs.

Results and discussion
Ancillary data for the incubation experiments is given in Supplementary Note 1 and Supplementary Tables S1 -S8 .AA 15 N-SIP exposes patterns in the biochemical assimilation pathways of applied 15 N-labelled substrates via changes in the measured isotopic compositions ( δ 15 N values) of each hydrolyzable AA over time (Charteris et al. 2016 ).AA δ 15 N v alues r eflect the r elativ e 15 N content in the AA pool at that time, with any additional 15 N ( cf. t = 0 AA δ 15 N values , i.e . 15N v alues; Fig. 1 A-F) being deriv ed fr om the applied 15 N-labelled substrate.
Individual AAs demonstrated different levels and patterns of 15 N incor por ation in eac h tr eatment, but in both 15   Since AA concentrations (and thus the balance of AA degradation/biosynthesis/turnover) did not change markedly during the incubation experiments ( Supplementary Note 1 ; Supplementary Tables S3 -S8 ), 15 N may be expected to be distributed (after initial uptake) in proportion to the quantity of N in each AA pool.Ho w e v er, 15 N can onl y be incor por ated into activ el y cycling pools, so a large, but stable AA pool would incorporate less 15 N than expected based on the amount of N in that AA pool.Deviations fr om a pr oportional distribution, ther efor e, r esulted fr om activity differences between AA pools and from the different biochemical routing of 15 N.These deviations are reflected in differing fitted (Equations 1 and 2 ) or 'plateau' AA 15 N values (if 15 N is distributed in proportion with AA concentration, AA 15 N values would be a ppr oximatel y equal for all AAs in a given experiment; Supplementary Note 2 ; Supplementary Tables S9 and S10 ).
AA δ 15 N (and 15 N) v alues indicate the pr oportion of N deriv ed fr om the a pplied 15 N but not the total flux of that 15 N into in each hydr ol yzable AA [or, ther efor e, the distribution of applied 15 N or 15 N still present in the soil ( retained 15 N ) amongst the AAs].AAs present in higher concentrations require larger amounts of 15 N to raise the N isotopic composition of the whole pool.It is, therefore, useful to consider the excess moles of 15 N in each AA and, to provide some context, in comparison with the excess moles 15 N applied (Equation 3), or alternatively, the excess moles 15 N retained in the soil at that time (Equation 4; Supplementary Fig. S2 ).P er centa ge a pplied 15 N incor por ations (% 15 N A incor por ation) ar e useful in providing an indication of the ov er all fate of applied 15 N (affected by heterogenous treatment applications and any losses of 15 N fr om the system, whic h w ould occur in a field).P er centage r etained 15 N incor por ations (% 15 N R incor por ation) r eflect the partitioning of 15 N present (or retained) in the system at the time, but as these data are calculated based on bulk soil δ 15 N values, could be affected by volatile losses of lighter 14 N raising values.
Temporal patterns in the % 15 N R incorporation into each AA under eac h tr eatment (Fig. 2 A-F) wer e similar to those of incr easing AA 15 N values (Fig. 1 A-F) but were dependent on the quantity of AA N in each pool ( Supplementary Tables S3 -S8 ; to reflect the routing/partitioning of 15 N) and smoothed by the availability of 15 N in the bulk soil.As for AA plateau 15 N values, AA plateau % 15 N R incor por ations wer e determined by fitting simple exponential r egr essions (as well as critical exponential r egr essions for Glx in the 15 NH 4 + and 15 N-U treatments; Equations 1 and 2 ; Supplementary Tables S11 and S12 ).The largest plateau hydr ol yzable AA % 15 N R incor por ations wer e found in Glx in fiv e out of the six tr eatments, r anging fr om 2.65 ± 0.15% of retained 15 N in RM-15 NH 4 + to 1.0 ± 0.21% in WA-15 NO 3 − (Fig. 2 A-F).Using an analogous experimental a ppr oac h (kinetic flux profiling) on an Esc heric hia coli culture, Yuan et al. ( 2006 ) similarly found largest fluxes of 15 N into Glu and Gln and surmized that Glu N was quic kl y tr ansferr ed into other AAs (Reitzer 2003 ).The exception to this was the RM-15 NO 3 − treatment, in which the highest % 15 N R was observed in Ala (0.4% r etained 15 N).In gener al, and particularly in the 15 NO 3 − treatments, AAs present at higher concentrations ( Supplementary Tables S3 -S8 ) demonstrated larger % 15 N R incor por ations (Fig. 2 A-F), as might be expected to maintain the AA concentr ation pr ofile of the soil, whic h did not v ary.As highlighted by differences in AA 15 N values, ho w ever, applied 15 N was not homogeneously distributed across the AA pools due to differ entl y r esponding subpools of AAs and/or the differential bioc hemical r outing of 15 N (Fig. 2 A-F; Supplementary Note 3 ).That the plateau 15 N le v els (as depicted in the pie charts in Fig. 2 C-F) for the 15 NH 4 + and 15 N-U tr eatments ar e v ery similar, but those of the 15 NO 3 − are different both from these four and one another, suggests that the two soils responded differently to nitrate, but similarly to the other two substrates.
A summation of the results of Equations ( 3) and ( 4) for eac h hydr ol yzable AA giv es the % 15 N A incor por ation and % 15 N R incor por ation into the total hydr ol yzable AA pool, r espectiv el y (Fig. 3 ).Ther e wer e onl y minor differ ences between the % 15 N A incor por ation and % 15 N R incor por ation into the total hydr ol yzable AA pool, whic h wer e due to bulk soil 15 N contents ( Supplementary Table S1 ).As before, plateau % 15 N incor por ations into the total hydr ol yzable AA pool were determined by fitting simple exponential r egr essions (Equation 1 ; Supplementary Table S16 ).Differences between the three N sources and two soils are clear-the three substrates are assimilated to significantly different extents ( 15 NH 4 + > 15 N-U > 15 NO 3 − ) in the RM soil, but not in the WA soil (based on error bar overlap).
Although the two soils in these experiments were sampled fr om cattle-gr azed gr asslands in southwest England, they had differ ent mana gement histories and contr asting compositions (noncalcar eous v ersus calcar eous), whic h affected the biotic and abiotic processing of applied N (Müller et al. 2011 ).The RM soil receiv ed onl y cattle slurry for the 25 years prior to soil sampling while the WA soil also r eceiv ed r egular additions of ammonium sulfate or urea (since 2011 when it was conv erted fr om spring crops to grass le y).Man uring, and higher soil percentage total organic carbon (% TOC) and percentage total N (% TN) contents have been related to greater soil microbial biomass activity (RM > WA; t = 0% TOC 6.80% cf.4.17% and % TN 0.63% cf.0.45; Söderström et al. 1983, Černý et al. 2003, Edmeades 2003, Booth et al. 2005, Müller et al. 2011 ).
Substrate assimilation in the RM soil matched expectations based on N assimilation biochemistry and previous studies assessing fertilizer N immobilization with bulk measurements (e.g.Wic kr amasinghe et al. 1985, Jackson et al. 1989, Recous et al. 1990, Christie and Wasson 2001 ).NO 3 − -15 N was not used extensiv el y as an anabolic N source.Both NO 3 − uptake and incor por ation into cell material (via reduction to NH 4 + ) require more energy (and thus C) than NH 4 + assimilation and NO 3 − uptake can be inhibited by only low concentrations of NH 4 + (Rice and Tiedje 1989, Recous et al. 1990, Magasanik 1993, Geisseler et al. 2010 ).Urea-15 N incor por ation w as slo w er and less extensive than 15 NH 4 + incorpor ation as ur ea m ust first be hydr ol yzed.Ur ease is ubiquitous in soils, ho w e v er, and ur ea hydr ol ysis can occur extra-or intracellularly (Mobley et al. 1995, Geisseler et al. 2010 ), at a lo w er metabolic cost than 15 NO 3 − reduction.
The operation of a more active (or larger) soil microbial biomass in the RM soil is supported by the significantly higher (almost double) plateau le v el of incor por ation of 15   (Robertson and Groffman 2007 ).Indeed, attunement to urea fertilization of this soil could also be responsible for the faster (initial and ov er all) assimilation of urea-15 N compared with the RM soil through increased endogenous urease concentrations.Further, differences in the active microbial community, such as r elativ e bacterial and fungal r atios, arising fr om differing management, may also influence dynamics of uptake for differing N amendments.Other work at the RM site using amino sugar (AS) 15 N-SIP allo w ed quantification of 15 N assimilation in this smaller, but more specific soil organic N pool (Reay et al. 2019a , Joergensen 2018 ).Assimilation into bacterial AS pools reflected dynamics observ ed her ein for AAs (Reay et al. 2019b ), while fungal AS exhibited slo w er uptake, and a lo w er pr efer ence for NH 4 + ov er NO 3 − , likel y r eflecting uptake of secondary N sources (Marzluf 1997, He et al. 2011 ).Hence the differing soil types, and management at the RM and WA sites her ein likel y r esulted in differing micr obial communities (Malik et al. 2018, Romdhane et al. 2022 ), and thus attunement to N amendments.Ov er all, a maxim um of 12.9% of a pplied 15 N (as 15 NH 4 + ), or 11.7% of 'retained' 15 N was assimilated into the total hydrolyzable AA pool (in RM-15 NH 4 + ; Fig. 3 ; Supplementary Table S16 ).Incorporation was lowest in RM-15 NO 3 − , at 1.7% of applied 15 N, or 1.6% of retained 15 N.These maximal plateau % 15 N incor por ations ar e unlikel y to hav e been caused by 15 N-substrate limitation during the incubations since 15 N remained in the soil (based on bulk soil δ 15 N values) and other processes: are either considered poor competitors for NH 4 + (e.g.nitrification); would not reduce 15 N availability (e.g.denitrification or other gaseous losses, whic h wer e not observed to occur extensively, and would likely increase, rather than decrease, bulk soil δ 15 N values); or were not observed to occur (e.g. 15 N loss via leaching).Maintenance of the soil at 50% WHC prev ented leac hing losses and made anaer obic micr osites suitable for denitrification and dissimilatory nitr ate r eduction to ammonium (DNRA) less likely to develop (Tiedje et al. 1984, Sexstone et al. 1985 ). Rather, maximal 15 N assimilations pr obabl y r esulted fr om regulation of N uptake/assimilation as limitation by another essential nutrient (e.g.C or P) arose in the soil.Physical and chemical protection of soil organic C reduces microbial availability, resulting in C-limitation, which is consistent with lo w er NO 3 − assimilation observed in the WA soil, which had lo w er C content compared to the RM soil (Soong et al. 2019 ).The application of our new 15 N-AA SIP approach provides new insights into inorganic and organic N assimilation biochemistry by soil micr obes.Criticall y, it pr ovides vital mechanistic links between theor etical/pur e cultur e deriv ed bioc hemical expectations and bulk le v el fertilizer immobilization studies, bridging these different scales of understanding.Moreover, the work demonstrates that simple biochemical processes (N assimilation in this case) oper ating in physiologicall y r ele v ant complex matrices are subject to additional biotic and abiotic environmental influences.This includes substr ate suppl y by similarl y influenced upstr eam pr ocesses and can ov er all r esult in quite differ ent a ppar ent pr ocess efficiencies in different settings (here , soils).Hence , the work constitutes a k e y ste p to w ar d gr eater a ppr eciation of the micr obiall y mediated transformations of fertilizer N to organic N and contributes to a more complete picture of soil N-cycling in response to fertilizer N a pplications.Finall y, the quantitativ e estimates r egarding these transformations generated through time-course incubation experiments are vital parameters for the next generation of soil N-cycling models.
fertilization r ate calculated based on a 0.3-m soil depth and an av er a ge of fiv e to six treatments between February and October.The rates are generally within the range recommended for grasslands for dairy grazing (140-340 kg N ha −1 year −1 ; Defra 2010 ).
NO 3 − treatments, 15 N v alues initiall y dipped befor e rising (Fig.1 A and B).All AAs exhibited a similar temporal pattern, but a range of responses (AA 15 N values) was observed at all time points.In the 15 NH 4 + and 15 N-U tr eatments (Fig.1C-F), glutamate [abbr e viated to 'Glx' since acid hydr ol ysis deaminates glutamine to glutamate, so the measured glutamate pool includes contributions from glutamic acid (Glu) and glutamine (Gln)] had a differ ent tr end fr om the two-phase rise of other AAs rising more quickly to an early peak (at ca.t = 2 days).

Figure 1 .
Figure 1.Time-course plots of AA 15 N values revealing 15 N assimilation into individual AAs in the six treatments.(A) RM-15 NO 3 − , (B) WA-15 NO 3 − (error bars at t = 16 and 32 days are coloured to aid differentiation), (C) RM-15 NH 4 + , (D) WA-15 NH 4 + , (E) RM-15 N-U, and (F) WA-15 N-U.RM and WA refer to the two different soils from the two sites, RM and WA and the three amendments were potassium nitrate (K 15 NO 3 ), ammonium chloride ( 15 NH 4 Cl), and urea (CO( 15 NH 2 ) 2 ).Error bars are ± SE ( n = 3).See Supplementary Fig. S3 .For individual figures for each AA in each treatment for additional clarity.

Figure 2 .
Figure 2. Time-course plots of AA % 15 N R incor por ations r e v ealing 15 N assimilation into individual AAs in the six tr eatments, alongside pie c harts of the r elativ e percenta ge of r etained 15 N in eac h AA pool this r epr esented, based on the plateau partitioning of 15 N in eac h total hydr ol yzable AA pool (deriv ed fr om simple exponential r egr essions of the % 15 N R incor por ated into AAs ov er time; Equation 1 ).(A) RM-15 NO 3 − (err or bars for Ala and Gl y ar e coloured to aid differentiation), (B) WA-15 NO 3 − (error bars for Glu, Asp, and Ala are coloured to aid differentiation), (C) RM-15 NH 4 + , (D) WA-15 NH 4 + , (E) RM-15 N-U, and (F) WA-15 N-U.Error bars are ± SE ( n = 3).Adapted from Charteris ( 2019 ).

F
igure 3. P er centage of 15 N incorporated into the total hydrolyzable AA pool for all treatments, labelled with the plateau % 15 N R incorporations determined by simple exponential r egr essions.(A) Percenta ge of applied 15 N and (B) Percentage of the 15 N still present in the soil or 'retained' at that time.Error bars are ± SE ( n = 3), the error bars of the WA-15 NO 3 − treatment are highlighted in red as the bar at t = 32 days is large and otherwise difficult to distinguish.Adapted from Charteris ( 2019 ).

Soil Substr a te Key Labelling Substr a te applied 10 −1 g soil Mass N 10 −1 g soil Equivalent a fertilization r a te kg −1 N ha −1 year −1 Incubation periods
Table summarizing the laboratory incubation experiments conducted.