Ammonia-oxidizing archaea are integral to nitrogen cycling in a highly fertile agricultural soil

Human activity has substantially altered the global nitrogen cycle through fertilizer use, atmospheric deposition, and crop N fixation, accelerating nitrogen cycling and contributing to eutrophication and increased N₂O emissions. Nitrification, the oxidation of ammonia to nitrite and nitrate, is central to the N cycle and supplies substrates for denitrification and anammox. Ammonia oxidation—the rate-limiting step of nitrification—is performed by three distinct groups: AOB, AOA, and comammox Nitrospira. AOA often dominate in pristine neutral or acidic soils, while fertilized agricultural soils are typically thought to be dominated by AOB and associated with higher N losses and N₂O emissions. Comammox organisms are widespread but their roles in agricultural soils remain poorly resolved; the only isolated comammox strain shows low ammonia Kₘ and low N₂O yield similar to AOA. Recent inhibitor-based studies using 1-octyne (AOB-selective) indicate niche separation between AOA and AOB, with AOA favored at low NH₄⁺ supply and lower N₂O yields than AOB. AOA activity in soils has been linked to organic matter mineralization with low NH₄⁺ fluxes. The authors hypothesized that (i) in highly fertile soils with high organic matter mineralization, AOA-linked nitrification with low N₂O yields would be detectable in situ without selective inhibitors, and (ii) competition between plants and nitrifiers for reduced N would reduce nitrifier abundance in cropped soils, whereas fallow treatments would enrich nitrifiers. They selected highly organic Everglades Agricultural Area (EAA) soils (60–85% OM) that sustain sugarcane without N fertilizer to test these hypotheses.

Prior work shows AOA frequently outnumber AOB in neutral/acidic, unfertilized soils, whereas AOB often dominate fertilized agricultural soils and are implicated in higher N losses and N₂O emissions. Comammox Nitrospira are widespread in terrestrial systems, but their ecological significance in agricultural soils is unclear; the isolated comammox strain Nitrospira inopinata exhibits low ammonia Kₘ and low N₂O yield, but potentially lower specific affinity than some AOA. Studies using 1-octyne as an AOB-selective inhibitor demonstrated that AOA dominate ammonia oxidation at low NH₄⁺ inputs, while AOB are stimulated by ammonium fertilization; AOA show lower N₂O yields than AOB. PTIO can inhibit AOA and comammox but is unsuitable for soil applications. AOA activity has been linked to organic matter mineralization and low NH₄⁺ fluxes. Environmental controls commonly linked to AOA abundance/diversity include pH, temperature, NH₄⁺, moisture, and organic matter. These findings frame the current study’s focus on highly fertile, organic soils to assess in situ group contributions and N₂O yields without inhibitors.

Study site and sampling: In December 2017, soils were sampled in the Everglades Agricultural Area (EAA) near Belle Glade, Florida. Five plots were selected: one unmanaged native plot (P1) and four agriculturally managed plots (P2–P5) in close proximity. Management histories: P2 (sugarcane 3 years, then spinach Jan–May 2017, 28-week fallow to sampling), P3 (sweet corn Jan–May 2017, then flooded rice May–Oct 2017, 8-week fallow), P4 (sweet corn Jan–May 2017, 28-week fallow), P5 (sugarcane third consecutive year, harvested one week before sampling). Native plot P1 had mixed plant cover. For each plot, three composite samples were collected (nine 10-cm topsoil cores per composite within 10×10 m²; composites spaced 50–200 m apart). Samples were transported on ice; subsamples for molecular analyses were frozen at −80°C. Net nitrification and N₂O production: Microcosms with 10 g field-moist soil incubated at 28°C for 144 h following Hink et al. (2017, 2018). Rates were calculated from linear accumulation of NO₂⁻+NO₃⁻ (NOx-N) and headspace N₂O over time. 0.1% acetylene controls confirmed negligible nitrification/N₂O in preliminary tests (samples from Oct 2017). Nitrification potential: Modified Hart et al. method using synthetic freshwater Crenarchaeota medium instead of phosphate buffer to better capture AOA/AOB/comammox activity. Slurries: 1.0 g field-moist soil in 20 mL medium, incubated at 28°C; potential rates from linear NOx-N accumulation over first 48 h. Carbon mineralization and soil chemistry: Dissolved organic carbon (DOC), inorganic N (NH₄⁺, NO₂⁻, NO₃⁻), available P, moisture, pH, and C mineralization were measured; Pearson correlations among edaphic factors and microbial activity metrics were computed. DNA extraction and molecular analyses: DNA extracted from 0.25 g soil (DNeasy PowerSoil). Clone libraries for amoA were prepared for archaeal, bacterial, and comammox clade A/B (Pjevac et al. primers). Comammox clade B PCRs were negative for all plots and a wetland reference site (SRS3); thus, not sequenced. Sanger sequencing performed; sequences trimmed and aligned (ARB), and clustered into OTUs (QIIME 1.9.1) using similarity thresholds approximating strain-level resolution (archaeal 96%, bacterial 97%, comammox 94%). 16S rRNA gene amplicons: EMP protocols with primers 515F/806R; Illumina MiSeq (v3, 150 cycles). Data processed with DADA2 in QIIME2; taxonomy via sklearn classifier against SILVA v132; detailed phylogenies in ARB; diversity metrics in QIIME2. PCoA, CCA, and dbRDA performed in R (vegan 2.5-6). Quantitative PCR: amoA genes quantified via qPCR for archaeal, betaproteobacterial (AOB), and comammox Nitrospira (clade A/B). Standards: Nitrososphaera viennensis (AOA), Nitrosospira briensis (AOB), and a comammox amoA clone (OTU11). PCR efficiencies: 86% (archaeal amoA), 90% (bacterial amoA), 96% (comammox amoA). Gene copy numbers normalized per g dry soil. Statistics and visualization: Pearson correlations, ANOVA with post hoc Tukey HSD at α=0.05 unless noted (R 3.6.3). Community ordinations and constrained analyses (dbRDA, CCA) tested edaphic drivers (notably pH, moisture, and fallow period). Visualization in R (ggplot2).

• Net nitrification rates were high across plots, ranging from 2.11 ± 1.12 to 5.10 ± 1.68 µg NOx-N g⁻¹ dry soil day⁻¹ (lowest P3; highest P5). Rates did not differ significantly between plant-covered and fallow plots.
• Nitrification potentials tended to be higher in fallow plots (e.g., P2: 10.95 ± 1.29 µg NOx-N g⁻¹ dw day⁻¹) than plant-covered plots (P1: 4.76 ± 2.10; P5: 5.70 ± 0.58), with significant differences over net rates only in fallow plots. In P1 and P5, net and potential rates were similar, suggesting organic matter mineralization supported nitrification near potential capacity.
• N₂O production ranged from 0.40 ± 0.09 to 1.83 ± 0.30 ng N₂O-N g⁻¹ dw day⁻¹, with slightly lower rates in fallow plots.
• N₂O yield was low and narrowly distributed: 0.18 ± 0.04 to 0.41 ± 0.22 ng N₂O-N per µg NOx-N (average 0.31 ± 0.09), aligning with AOA-associated yields from inhibitor studies and lower than typical AOB yields (>0.9 ng N₂O-N per µg NOx-N). This indicates AOA-dominated nitrification in situ.
• Strong positive correlations among DOC, carbon mineralization, N₂O production, and net nitrification suggest tight coupling of microbial C and N mineralization. Fallow period negatively correlated with DOC, C-mineralization, and N₂O, and positively with NO₂⁻/NO₃⁻ and nitrification potential.
• AOA numerical dominance: Archaeal amoA gene copies were 1.10×10⁶ to 2.97×10⁷ g⁻¹ dw; AOB were 27–120× less abundant; comammox clade A were 65–430× less abundant; comammox clade B was undetected by endpoint PCR and qPCR. Assuming 1 amoA per AOA/comammox genome and 2–3 per AOB genome, AOA outnumbered AOB and comammox by 54–430×.
• 16S rRNA data corroborated AOA dominance: Thaumarchaeota constituted 2.42–6.37% of reads; Nitrosomonadaceae (AOB) were rare (0.006–0.055%); Nitrospiraceae (NOB) were 0.80–1.33%. Putative comammox-affiliated ASV accounted for only 0.025–0.116% of total 16S reads.
• Community diversity: AOA showed high lineage diversity (45 archaeal amoA OTUs across 10 lineages), dominated by NS-8 lineage in managed plots; AOB amoA diversity was low (10 OTUs, Nitrosospira cluster 3) with plot-specific dominant OTUs; comammox amoA diversity was low (14 OTUs; clade A only), with a predominant EAA-specific cluster.
• Community structure drivers: dbRDA and CCA indicated pH and soil moisture as primary correlates of both whole-community and thaumarchaeal community structure; models including fallow period provided only slightly lower fit.
• Plant N source inference: Similar nitrifier abundances and net nitrification across cropped vs fallow plots, alongside NO₃⁻ accumulation in fallow but not cropped plots, indicate plants relied heavily on NO₃⁻ as primary N source. Estimated nitrification in P5 (~1.59 t NO₃⁻-N ha⁻¹ over 10 months) could meet sugarcane N demand (~1.03 t N ha⁻¹) over a growing season.
• Comammox contribution likely minor: Given low abundances, comammox cells would require near/above known maximum per-cell rates to contribute ≥10% of observed nitrification, making substantial comammox contribution unlikely.

The study set out to test whether highly fertile, organic-rich agricultural soils would exhibit AOA-linked nitrification with low N₂O yields in the absence of selective inhibitors, and whether plant presence would suppress nitrifiers via N competition. Results support these hypotheses: (i) nitrification exhibited consistently low N₂O yields within AOA-associated ranges, indicating AOA-dominated activity in situ; (ii) nitrifier abundances and net nitrification rates were similar in plant-covered and fallow plots, arguing against strong direct competition with plants for reduced N. Strong correlations among DOC, carbon mineralization, and nitrification indicate that in these soils, AOA activity is closely coupled to microbial decomposition of organic matter producing low, steady NH₄⁺ fluxes. Community ordinations show pH and moisture structure both overall and AOA communities, suggesting that edaphic conditions regulate community composition, while mineralization dynamics regulate activity intensity. The extremely low abundance and diversity of comammox, combined with rate calculations, suggest negligible comammox contributions under the studied conditions. The low N₂O yields observed without inhibitors corroborate prior inhibitor-based findings and reinforce the potential to mitigate N₂O emissions by management practices that constrain AOB. The plant N mass balance implies that AOA-driven nitrification supplies NO₃⁻ used by crops and native plants, reframing nitrification as an integral, potentially beneficial process in such high-OM systems rather than solely a loss pathway. Overall, the findings underscore the ecological importance of AOA in agricultural nitrogen cycling, their link to carbon mineralization, and their favorable N₂O emission profile relative to AOB.

Highly fertile, organic agricultural soils in the EAA harbor AOA-dominated nitrifier communities that drive most nitrification irrespective of plant cover and produce consistently low N₂O yields. Molecular and rate measurements show AOA outnumber AOB and comammox by orders of magnitude and largely control nitrification activity. Nitrification rates appear tightly coupled to DOC-driven carbon mineralization, and plants in these systems rely predominantly on nitrification-derived NO₃⁻. These outcomes validate previous inhibitor-based inferences under natural conditions and point to management strategies that favor AOA over AOB to reduce N₂O emissions. Future work should disentangle seasonal moisture/pH effects, directly quantify in situ nitrifier activity and plant N uptake dynamics, and further resolve the ecological roles and substrate spectra (e.g., ammonia, urea, cyanate) sustaining AOA in diverse soils.

• Activity partitioning: No single approach definitively separates AOA, AOB, and comammox activity in situ; interpretations rely on N₂O yield ranges, abundance data, and indirect calculations.
• Inhibitor constraints: PTIO is unsuitable for soils; 1-octyne was not used here to avoid perturbation, so attributions are inferential.
• In situ fluxes: Net nitrification and N₂O production were measured in microcosms; true in situ fluxes and plant N uptake forms were not directly measured.
• Sequencing depth and bias: Clone libraries had limited depth; potential taxonomic biases exist between amoA and 16S datasets. Comammox amoA primer set showed off-target amplification, potentially overestimating comammox presence despite low abundance.
• Temporal scope: Single time point (early dry season) limits assessment of seasonal dynamics; moisture and pH effects inferred from between-plot variation.
• Comammox physiology: Conclusions about minor comammox contribution assume uniform activity and known kinetic constraints; uncharacterized comammox variants could differ.
• Statistical power: Community–environment relationships were evaluated with limited plots (n=5), reducing power to detect weaker drivers beyond pH and moisture.