Relationships between nitrogen cycling microbial community abundance and composition reveal the indirect effect of soil pH on oak decline

The study addresses increasing decline and mortality of oak trees in temperate regions, where multiple biotic and abiotic stressors (e.g., temperature rise, pollution, invasive pests and pathogens) interact in complex ways, complicating diagnosis and management. Across the UK and Europe, increased nitrogen deposition and acidifying compounds may destabilize nutrient cycling and reduce tree stress tolerance. In N-limited systems, elevated N can promote vegetative growth and increase susceptibility to pests and pathogens, while N deficiency can inhibit growth and potentially increase vulnerability. Microorganisms regulate soil N availability through nitrification and denitrification; however, links between soil N-cycling microbial communities and tree health are unclear. Ammonia-oxidising bacteria (AOB) and archaea (AOA) exhibit niche differentiation, with AOA often dominant in acidic soils and under low ammonium, and plants potentially modulate their abundance via litter inputs. The study aims to determine whether destabilization of soil N-cycling is associated with oak stress and decline by characterizing abundance and diversity of N-cycling microbes (AOB, AOA, denitrifiers) in soils under symptomatic versus asymptomatic oaks across seven UK woodlands. The authors hypothesised that N-cycling microbial community structure and abundance would differ between symptomatic and asymptomatic trees due to inherent differences in soil chemical properties affecting microbial composition and function.

Prior work indicates niche differentiation between AOB and AOA: AOA dominance in acidic soils and preference for low ammonium conditions; AOB often associated with higher N availability, including under plant canopies due to increased litter and N inputs. Changes in microbial utilization of organic C and N can shift overall soil microbial community structure and influence nitrification. Increased N deposition and acidifying compounds have been proposed to destabilize nutrient cycles and reduce tree resilience, potentially increasing susceptibility to insect herbivory, pathogens, and frost damage. These background studies motivate examining how soil abiotic factors modulate N-cycling microbial communities in relation to oak health.

Study design and sites: Seven oak-dominated broadleaf woodlands in England were sampled (Attingham, Bigwood, Langdale, Winding Wood, Speculation, Chestnuts Wood, Great Monks Wood) during 2016–2017. At each site, 10 symptomatic (declining) and 10 asymptomatic (apparently healthy) oaks were selected based on crown condition and presence of stem lesions associated with acute oak decline or root decay fungi (Armillaria sp., Gymnopus fusipes). Major soil groups were identified (surface water gleys at Attingham, Bigwood, Great Monks Wood, Langdale, Speculation; brown soils at Chestnuts Wood, Winding Wood). Oak density and basal area varied by site. Sampling: For each tree, three soil cores (0–20 cm depth) were taken 2 m from the base, yielding 420 cores total. Samples were kept on ice (~5 h) then stored at -20 °C. Rainfall occurred either one day before or during sampling across sites. Soil chemistry and nitrification potential: Soil moisture (105 °C oven-dry to constant weight), total C and N (Carlo Erba Flash EA 1112), pH in water (1:2.5), extractable NO3- and NH4+ (1 M KCl; colorimetric analysis). Nitrification potentials measured via shaken soil slurry assay supplemented with 0.3 mM NH4Cl; NH4+ quantified by Dionex ICS-3000; NH4+ removal rates by linear regression over time. DNA extraction and qPCR: DNA extracted from 0.25 g wet soil (Norgen Soil DNA Isolation Plus). qPCR quantified gene abundances using SensiFAST SYBR NO-ROX on BioRad CFX96 for: amoA (AOA: CrenamoA-23F/CrenamoA-616R; AOB: amoA-1F/amoA-2R), nirS (Cd3aF/nirS-R3cd), nirK (nirK-1F/nirK-5R), nosZ (nosZ2F/nosZ2R). Thermocycling: 95 °C 3 min; 40 cycles of 95 °C 5 s, 60 °C 30 s; melt curve 52–95 °C. Absolute quantification with standard curves (R2>0.99; slopes -3.2 to -3.4; efficiency ~95–104%). Standards, samples, and NTCs in duplicate; samples diluted 10×. Amplicon sequencing: Illumina Nextera-adapted primers; 28-cycle PCR with MyTaq Red DNA polymerase for 16S rRNA (bacteria: Bakt 341F/805R; archaea: 344F/915R) and functional genes (amoA AOB/AOA, nirS, nosZ). Amplicons purified with AMPure XP, indexed via 8 additional PCR cycles, quantified (Picogreen/Nanodrop 3300), pooled equimolarly, quality checked (Agilent 2100 Bioanalyzer DNA 1000), and sequenced on Illumina MiSeq V3 (2×300 bp) at University of Essex. Bioinformatics: Reads demultiplexed, quality-trimmed (Sickle q20), error-corrected (SPAdes BayesHammer), dereplicated and clustered into OTUs at 97% similarity (VSEARCH). Singleton OTUs removed; chimeras filtered de novo and reference-based (UCHIME). Taxonomy for 16S with RDP Classifier. Functional genes filtered for locus specificity via FrameBot. Dominant OTU centroid sequences (>99% reads per gene) codon-aligned (MUSCLE, MEGA6) with references (FunGene, BLAST). Phylogenetic trees for top 50 OTUs per gene by neighbor-joining with 1,000 bootstraps (Geneious 9.0.2). Statistics: Conducted in R. Piecewise structural equation modelling (piecewise SEM) assessed direct and indirect relationships among soil abiotic variables (total C, pH, NH4+, NO2-+NO3-, C:N ratio, soil moisture, nitrification potential), N-cycle gene abundances (amoA AOA, amoA AOB, nirS, nirK, nosZ), and tree health (binary: 1 asymptomatic, 0 symptomatic). Gene abundances log-transformed to improve normality; random effects included site. Model fit assessed with Shipley’s test of directed separation (non-significant Fisher’s C indicates adequate fit). A GLM with quasibinomial errors assessed AOB:AOA ratios vs tree health per site. For community composition, OTU tables rarefied (AOB 900, AOA 800, nirS 500, nosZ 700 sequences/sample), multivariate abundance analyzed with mvabund; visualized via NMDS (Bray-Curtis; metaMDS). Raw reads deposited to ENA (PRJEB35364).

- Soil heterogeneity: pH ranged broadly by site (e.g., Winding Wood ~3.6–7.6; Langdale ~4.7–8.3). Sites grouped by NH4+: higher (Attingham, Great Monks Wood, Langdale, Winding Wood) vs lower (Bigwood, Chestnuts, Speculation). Nitrification potentials generally low but highest at Langdale corresponding to high NH4+. - Gene abundance ranges: AOA amoA: 2.6×10^1–1.4×10^5 copies g−1; highest at Chestnuts, Bigwood, Speculation. AOB amoA: 3.1×10^1–4.1×10^5 copies g−1; highest at Winding Wood, Great Monks Wood. Denitrification genes varied significantly across sites (P<0.001): nirS 1.4×10^3–8.4×10^4; nirK 2.7×10^2–8.7×10^5; nosZ 1.1×10^3–2.2×10^7 copies g−1. - SEM model fit: Fisher’s C non-significant (P=0.977; df=8), indicating adequate model fit. - Tree health associations: • Asymptomatic trees had higher AOB amoA abundance than symptomatic trees (path coefficient=0.23; R^2=0.07; P=0.008). • No association between AOA amoA abundance and tree health. - Environmental drivers of nitrifiers: • AOB amoA abundance positively driven by soil pH (path coefficient=0.26; R^2=0.77; P<0.001). • AOA amoA abundance primarily driven by lower NH4+ concentrations (consistent with AOA preference for low ammonium). - Denitrifiers: • C:N ratio was the primary direct driver of nitrite reductase gene abundance (nirS path coefficient=0.22; P=0.01; nirK path coefficient=0.26; P=0.02). • Strong positive covariation among denitrification genes: nirS–nirK (path coefficient=0.73; P<0.001); both with nosZ (0.69; 0.91; P<0.001). • AOB amoA positively covaried with nirS (0.21; P<0.001), nirK (0.21; P=0.007), and nosZ (0.26; P<0.001). - Soil parameter interrelationships: • C:N ratio influenced by NO2-+NO3- (negative; path coefficient=−0.14; P=0.04), NH4+ (negative; −0.20; P=0.003), and total C (positive; 0.69; P<0.001); C:N in turn drove pH (negative; −0.21; P<0.05). • Moisture positively affected NO2-+NO3- (0.39; P=0.03), NH4+ (0.20; P<0.001), and total C (0.16; P=0.04); total C covaried positively with NH4+ (0.34; P<0.001). • Moisture positively influenced nitrification potential (0.08; P=0.04) and pH (0.11; P<0.001). Overall, results indicate an indirect effect of soil pH on oak health mediated through AOB abundance, AOA linked to ammonium availability, and denitrifier abundance governed largely by soil C:N ratio, with strong inter-gene covariation among denitrifiers and with AOB.

The findings support the hypothesis that soil abiotic conditions, particularly pH, indirectly influence oak health via effects on ammonia-oxidising bacteria. Asymptomatic trees exhibited higher AOB amoA abundance, and AOB abundance increased with higher soil pH, indicating that more neutral conditions may favor bacterial ammonia oxidation, potentially stabilizing N transformations and reducing plant stress. In contrast, AOA abundance did not vary with tree health and was instead associated with lower NH4+, consistent with niche differentiation where AOA prefer low-ammonium conditions. Denitrifier gene abundances were not directly linked to tree health but were strongly governed by the soil C:N ratio and showed tight covariation among nirS, nirK, and nosZ, as well as positive covariance with AOB, suggesting coordinated N-cycling responses to organic matter quality and N availability. The complex network of soil parameters—moisture, inorganic N species, total C, C:N ratio, and pH—highlights multiple interacting pathways that shape microbial communities and N-process rates. Management strategies that ameliorate soil acidification and adjust C:N balance could enhance AOB abundance and function, potentially mitigating stress in declining oaks by influencing nitrification-denitrification dynamics and N availability. The site-level heterogeneity and grouping by ammonium status, alongside variable nitrification potentials, suggest local edaphic context is important for predicting microbial community responses and associated tree health outcomes.

By integrating soil chemistry, qPCR quantification, amplicon sequencing, and piecewise SEM across seven UK oak woodlands, the study demonstrates that soil pH exerts an indirect effect on oak decline via modulation of AOB abundance: asymptomatic trees are associated with higher AOB amoA gene copies, and AOB abundance increases with pH. AOA abundance is decoupled from tree health and relates primarily to ammonium levels. Denitrifier abundance is principally driven by soil C:N ratio and shows strong inter-gene covariation and positive associations with AOB. These results suggest that managing soil acidity and balancing C:N may influence N-cycling microbial communities and N transformations in ways that reduce stress on declining oaks. Future work should include mechanistic and longitudinal studies to establish causality, assess microbial activity (not just gene abundance), and test management interventions (e.g., liming or organic matter amendments) aimed at adjusting pH and C:N to support oak health.

The study is observational and cross-sectional across heterogeneous sites; while SEM identifies direct and indirect associations, it does not establish causality. Gene abundances (qPCR) serve as proxies for microbial potential rather than in situ activity or process rates, and amplicon-based approaches can introduce primer and amplification biases. Tree health status was categorized from crown condition and lesion presence, which may reflect multiple concurrent stressors (e.g., pathogens, insects) not fully disentangled from soil effects. Nitrification potentials were measured under laboratory conditions and may not fully capture field rates.