Linking meta-omics to the kinetics of denitrification intermediates reveals pH-dependent causes of N₂O emissions and nitrite accumulation in soil

The study addresses how soil pH shapes denitrification phenotypes and emissions of intermediates (NO₂⁻, NO, N₂O, HONO) by linking community genetic potential and transcription to kinetic measurements in soil microcosms. Denitrification reduces nitrate to N₂ via nitrite, nitric oxide, and nitrous oxide, using modular reductases (Nar/Nap, NirK/NirS, qNor/cNor, NosZ clade I/II). Prior work shows pH strongly affects intermediate accumulation: NO₂⁻ tends to accumulate at higher pH, whereas N₂O emissions are negatively correlated with pH. Despite transcription of nosZ at low pH, acidic soils often show negligible or delayed N₂O reduction, suggesting post-transcriptional interference (e.g., impaired NosZ maturation or electron transfer). Low NO₂⁻ in acidic soils is commonly attributed to abiotic decomposition, but recent work suggests rapid biological nitrite reduction may dominate. The authors hypothesize that pH-dependent regulation and enzymology of denitrification steps, particularly NosZ maturation and Nir/Nor activities, underlie observed kinetic differences. They aim to resolve which genes and taxa are active under anoxia across pH, whether DNRA contributes, whether nos accessory genes are present/transcribed at low pH, and how these factors explain NO₂⁻/NO/N₂O/HONO patterns.

- Denitrification pathway and modular gene distribution are well established; presence/absence and regulation influence intermediate accumulation and N₂O/N₂ ratios. - Soil pH is a major control: higher pH favors NO₂⁻ accumulation; lower pH correlates with higher N₂O/(N₂O+N₂) ratios. Evidence shows nosZ can be transcribed under acidic conditions but N₂O reduction is delayed, implying a post-transcriptional block (likely NosZ maturation or electron delivery issues). Prior quantifications often targeted only nosZ clade I. - The assumption that low NO₂⁻ in acidic soils is mainly abiotic has been challenged by kinetic modeling indicating biological NO₂⁻ reduction is a dominant sink at low pH. - The relationship between nosZ abundance and N₂O emissions is inconsistent across studies, questioning reliance on gene counts as predictors. - Primer biases in amplicon-based surveys can misrepresent nirk vs nirs and nor gene pools, potentially skewing interpretations of community function.

Study system: Two peat soils from a long-term liming field experiment in western Norway: SoilA (unlimed, pHCaCl2 3.8) and SoilN (limed, pHCaCl2 6.8). Both have 40–45% organic C and ~2% organic N. Microcosm setup and incubation: To stimulate detectable transcription, soils were amended with 5 mg dried, powdered clover per g soil wet weight and pre-incubated at 15°C for 72 h. Aliquots corresponding to 1.5 g soil organic C (5–8 g ww) were placed in sealed glass vials. KNO₃ was added to reach 80% water holding capacity and 6.2–7.1 mM NO₃⁻ in soil moisture (initial total NO₃⁻: 37 µmol/vial in SoilA; 26 µmol/vial in SoilN). Vials were rendered anoxic via six evacuation/He-fill cycles and incubated at 15°C. Headspace gases (CO₂, O₂, NO, N₂O, N₂) were measured every 3 h using an autosampler with GC and NO analyzer. At each gas sampling, a replicate vial was sacrificed for NO₂⁻ extraction and quantification. Portions of the same vials were snap-frozen for nucleic acid extraction. Enzymatic step rates V_NAR (NO₃⁻→NO₂⁻), V_NIR (NO₂⁻→NO), V_NOR (NO→N₂O), V_NOS (N₂O→N₂) were calculated from measured NO₂⁻, NO, N₂O, N₂ kinetics and corrected for abiotic NO₂⁻ decomposition (significant in SoilA) per Lim et al. (2018). Nucleic acid extraction: DNA and RNA were co-extracted from frozen soils using a CTAB/phenol-chloroform bead-beating protocol optimized for inhibitor-rich soils. Post-precipitation cleanup used Zymo kits. RNA was DNase-treated; absence of gDNA was verified by qPCR (16S rRNA primers; Ct ≤35 defined contamination). RNA was reverse-transcribed with random hexamers. DNA/cDNA quality was verified by qPCR (16S rRNA and nosZ clade I primers). Sequencing: Metagenomes (MG; triplicate DNA) and metatranscriptomes (MT; duplicate RNA per time point) were sequenced on Illumina HiSeq 2500 (CBC/Keck Center). RNA integrity and gDNA absence were independently confirmed prior to sequencing. Community composition was profiled via 16S rRNA gene amplicons (V4 region; primers 515f/806rB) on Illumina MiSeq (2×300 bp) by StarSEQ. Sampling for MT: SoilA at 0.5 h and 3 h after anoxia; SoilN at 0.5, 3, 9, 12, and 27 h (aligned with observed kinetic transitions). Bioinformatics and quantification: Quality control of reads with BBTools BBDuk. Functional annotation: DIAMOND alignments (e-value ≤1e−3) against a curated nitrogen metabolism database. Matches required >30 aa aligned region and >60% identity. Read counts were normalized to reads per million (RPM). Taxonomic assignment of extracted gene/transcript reads via Kaiju on KBase. 16S reads processed with GHAP using USEARCH clustering at 97% identity and RDP-based classification; rarefaction to 38,287 reads/sample for richness and Pielou’s evenness. Statistical analysis: In-house R scripts; normalization to RPM; ratios computed among gene families (NAR=NarG+NapA; NIR=NirK+NirS; NOR=qNor+cNor; NOS=nosZ clade I+II). HNO₂ concentrations estimated from measured total nitrite-N using Henderson–Hasselbalch with pKa=3.398 to infer potential HONO emissions. qPCR: Gene and transcript quantification for 16S rRNA, nirk, nirs, nosZ clade I using standard primer sets (27F/518R; 517F/1055R; cd3aF/R3cd; Z-F/1622R). SYBR chemistry, optimized cycling, detection limit ~5 copies/µL reaction (~4×10² copies g⁻¹ soil ww).

- Kinetic phenotypes differ by pH: - SoilA (pH 3.8): Very low NO₂⁻ in soil moisture (mostly 20–50 µM; brief ~100 µM at 36–40 h), strong N₂O accumulation with almost no N₂ for ~35 h; V_NAR, V_NIR, V_NOR similar; V_NOS ~0 initially. NO transient peaks ~1.5–2 µM similar to SoilN. - SoilN (pH 6.8): Pronounced NO₂⁻ accumulation (2–3 mM at ~20 h), concurrent N₂O and N₂ production from start; V_NAR initially exceeded V_NIR, V_NOR, V_NOS then declined with NO₃⁻ depletion; V_NOS high from the start. - HONO potential: Despite high total nitrite in SoilN, calculated undissociated HNO₂ ≤1.4 µM. SoilA had nearly two orders of magnitude lower HNO₂, yet overall estimated HONO production was ~10× higher in SoilA than SoilN. - Metagenomes (MG) and metatranscriptomes (MT): - NAR genes more abundant than NIR, NOR, NOS in both soils. narG dominated over napA (narG/napA ~6.1 in SoilA; ~2.1 in SoilN MG). - NIR genes similar total abundance across soils, but nirk >> nirs (MG nirk/nirs ≈40 in SoilA; ≈7 in SoilN). MT showed nirk transcripts dominate; nirs transcripts undetected in SoilA. - NOR genes more abundant in SoilA (MG NOR RPM ~74.9 vs 48.1 in SoilN), with strong dominance of qnor over cnor (qnor/cnor ~11 in SoilA; ~4 in SoilN MG). MT NIR/NOR ratios were <1 in SoilA (avg ~0.7) and >1 in SoilN (avg ~2.2), indicating strong NO control in SoilA. - NOS genes (nosZ clade I+II) somewhat higher in SoilN MG (25.8 RPM) than SoilA (18.2 RPM). Clade distribution: SoilA enriched in clade I; SoilN enriched in clade II. MG nosZ clade I/clade II ratio ~28.1 in SoilA vs ~0.25 in SoilN. - MT: In SoilN, nosZ transcripts increased ~7-fold from 0.5 h to 3 h (from ~56 to ~380 RPM). In SoilA, nosZ transcripts roughly doubled between 0.5 h and 3 h, yet N₂O reduction remained negligible, indicating post-transcriptional impairment. - Nos accessory genes: Genes and transcripts for nosR, nosL, nosD, nosF, nosY detected in both soils and upregulated 0.5→3 h, indicating presence of maturation/electron transfer machinery; their absence does not explain failure of N₂O reduction at low pH. - DNRA potential vs activity: nrfA and nirB genes and transcripts were detected (sometimes at levels comparable to NIR/NOR), yet mass balances showed ~100% recovery of NO₃⁻-N as N₂ in SoilN and ~94% accounted (N-gases + nitrosylation) in SoilA, indicating minimal DNRA activity. - Taxonomy: Denitrification genes/transcripts primarily from Proteobacteria, Actinobacteria, Bacteroidetes; qnor and nirk widely distributed across phyla (including Acidobacteria and Planctomycetes for qnor). nosZ clade I transcripts mainly Proteobacteria; clade II transcripts diverse, dominated by Bacteroidetes in both soils. - Primer bias evidence: qPCR (standard primers) showed nirs >> nirk and lower nosZ clade I, contradicting -omics (nirk dominance; comparable nosZ). Demonstrates that standard primers dramatically underestimate nirk and bias nor assessments. - Interpretation of NO₂⁻ control: Low NO₂⁻ in SoilA primarily due to high biological NO₂⁻ reduction rates (and potentially higher NirK activity at low pH) rather than abiotic decomposition alone; NO kept low via high qnor expression/activity. - Post-transcriptional NosZ impairment at low pH appears widespread across taxa and both nosZ clades.

The integrated kinetics and meta-omics show that soil pH reshapes denitrification by altering both regulation and enzymatic performance. In acidic soil, despite rapid and broad nosZ transcription (including accessory nos genes), N₂O reduction was delayed for >25–35 h, strongly supporting a generic, post-transcriptional impairment of NosZ maturation or function at low pH, across diverse taxa and affecting both clade I and II. Consequently, N₂O accumulates, increasing emission risk under natural conditions. Conversely, neutral pH soils exhibited immediate V_NOS, aligning with effective N₂O reduction. Nitrite dynamics were pH-dependent: high NAR relative to NIR led to NO₂⁻ accumulation in SoilN, while in SoilA, low NO₂⁻ concentrations resulted mainly from high biological NO₂⁻ reduction rates and effective NO detoxification via qNor, not merely from abiotic nitrite decomposition. The MT NIR/NOR ratios (<<1 in SoilA; >1 in SoilN) match observed NO control, highlighting qNor’s central role in acidic soils and suggesting pH selects for qnor over cnor. The study reveals that gene or transcript abundances alone (and especially their ratios) do not reliably predict phenotypic rates in situ: NAR/NIR transcript ratios were similar between soils despite markedly different NO₂⁻ accumulation; MG NAR/NIR ratios were nearly identical while kinetics diverged. Functional outcomes are influenced by enzyme pH optima, post-transcriptional processes, and community regulatory phenotypes. DNRA genes/transcripts were relatively abundant, yet kinetic budgets indicated negligible DNRA activity, implying that the presence and expression of DNRA genes do not equate to measurable pathway flux under the tested conditions. Estimated HONO formation was substantially higher from the acidic soil, contributing to atmospheric chemistry, though most NO₂⁻ was funneled through denitrification rather than emitted as HONO. Finally, strong discrepancies between qPCR and -omics underscore substantial primer bias in common functional gene assays, particularly underestimating nirk and misrepresenting nor composition, which can lead to erroneous ecological inferences.

This multi-omics and kinetic integration clarifies pH-dependent controls on denitrification in soils. Key contributions include: (1) demonstration that delayed N₂O reduction at low pH is a community-wide, post-transcriptional limitation of NosZ maturation/function affecting both clades across diverse taxa; (2) evidence that low NO₂⁻ in acidic soils is primarily due to high biological NO₂⁻ reduction and strong qNor-mediated NO control; (3) identification of widespread qnor dominance and functional importance in acidic soils; (4) revelation that DNRA contributes little to nitrate reduction under the studied conditions despite gene/transcript presence; (5) quantification suggesting HONO production potential is much higher from acidic soil; and (6) clear documentation that standard qPCR primers can severely bias interpretation of denitrifier community structure and activity. Future work should: directly measure HONO emissions across pH and moisture regimes; dissect biochemical steps of NosZ maturation under acidic conditions in situ; refine and validate primer sets or adopt primer-free methods for functional gene surveillance; and link meta-omics with isotope tracing/enzymatic assays to better map transcripts to process rates.

- Metatranscriptomic sampling for the acidic soil covered only the first 3 h of anoxia; later transcriptional dynamics during the period when N₂O reduction eventually commenced (>25–35 h) were not captured. - Accessory nos gene databases and annotations (e.g., nosF) are incomplete/uncertain, limiting confident interpretation of maturation machinery abundance; protein function/phylogeny remain partly unresolved. - HONO was inferred from calculated HNO₂ based on equilibrium assumptions rather than directly measured emissions; actual fluxes may differ under soil–atmosphere nonequilibrium conditions. - Findings are based on two peat soils from a single long-term liming site and microcosm conditions; generalizability to other soil types and field conditions may be constrained. - qPCR comparisons highlight primer bias; however, alternative primer sets or primer-free validation across broader gene families were not exhaustively tested in this study.