Nucleotide depletion signals early-stage soil stable carbon collapse in anthropogenically disturbed alpine ecosystems

Nucleotide depletion signals early-stage soil stable carbon collapse in anthropogenically disturbed alpine ecosystems | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Nucleotide depletion signals early-stage soil stable carbon collapse in anthropogenically disturbed alpine ecosystems Mengdi Xie, Qiang Lin, Lingling Feng, Huan Zhao, Can Tang, Ke Tan, and 6 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-9373312/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract Anthropogenic disturbances are a primary cause of soil organic carbon (SOC) destabilization in alpine ecosystems. However, effective management is hindered by a "response delay" of macro metrics used to characterize stable carbon pools (e.g., microbial necromass carbon and aggregate associated organic carbon), which often fail to capture the immediate impairment of the soil carbon sequestration. To address this, we conducted a multi-omics study on 204 soil samples along a 621-km transect. We observed a "temporal lag": while physical disturbance caused rapid depletion of nucleotide metabolites, stable carbon fractions showed limited responsiveness, masking carbon depletion onset. Through metabolomics and 100-fold stratified subsampling, we identified depletion of specific nucleotides—notably thymidine and guanosine—as early-warning signatures (mean AUROC > 0.90). Metagenomic profiling revealed this depletion is driven by a disturbance-induced taxonomic shift triggering a synchronized suppression: the simultaneous inhibition of genetic capacity for de novo synthesis (mediated by pyrD ) and salvage pathways (mediated by deoA ). Furthermore, the concurrent downregulation of korA indicates the disruption of the "Glutamine Bridge," effectively severing the metabolic link between nucleotide turnover and central carbon/energy metabolism. Our findings identify molecular "early-warning biomarkers" that precede observable carbon loss, providing a sensitive tool for monitoring incipient soil degradation. Earth and environmental sciences/Environmental sciences/Environmental chemistry/Geochemistry Earth and environmental sciences/Ecology/Biogeochemistry/Carbon cycle alpine soil soil stable carbon multi-omics integration biomarkers nucleotide Figures Figure 1 Figure 2 Figure 3 Figure 4 Figure 5 Introduction Global soil organic carbon (SOC) stocks, estimated at 1500–2400 Pg, constitute the largest active terrestrial carbon reservoir [1] . In alpine ecosystems such as the Qinghai-Tibet Plateau, SOC pools are highly vulnerable to recent infrastructure expansion [1,2] . The construction of tunnels, highways, and pipelines imposed severe physical disturbances—such as vegetation removal and soil aggregate disruption [3] . Consequently, with estimated carbon losses of 7–12 tons per kilometer of road construction [4,5] , large-scale projects in China could drive massive SOC depletion on the order of thousands of tons. While these direct physical losses are calculable, monitoring the stability of the remaining soil carbon pool is hindered by a temporal lag. Current monitoring relies largely on "macro-indicators" such as aggregate-associated organic carbon (AAOC) and microbial necromass carbon (MNC) [6-8] . While these metrics define the cumulatively stable carbon status, they exhibit significant temporal lag [9] . MNC represents a slow-turnover pool, often revealing detectable losses only years after the initial environmental changes (such as ecological restoration and physical disturbance) [10] . Because AAOC is stabilized within organo-mineral complexes through adsorption onto soil colloids, its response to environmental disturbance is significantly delayed [11] . Effective ecological restoration requires early warning systems to track carbon dynamics in real time. This diagnostic delay limits the capacity for real-time monitoring: by the time statistical decreases in soil stable carbon pool, such as MNC and AAOC are detected, the microbial processes governing carbon sequestration may have already been compromised [12] , ultimately jeopardizing the long-term stability of the soil carbon pool. To overcome this diagnostic delay, indicators that reflect the instantaneous physiological state of the soil microbiome are required. Unlike macro-carbon pools, microbial communities are highly sensitive to environmental disturbances (e.g., habitat destruction and vegetation removal), altering their turnover rates immediately upon stress [13,14] . Intracellular metabolites—specifically purine and pyrimidine nucleotides—are tightly coupled to these microbial dynamics. As the fundamental building blocks for cell proliferation and the primary currency for cellular energy transfer (e.g., ATP/GTP) [15] , the soil nucleotide pool scales directly with active microbial biomass. Consequently, physical excavation that induces a sudden disruption of nutrient inputs causes a rapid depletion in nucleotide concentrations. Because microbial biomass turnover is the primary engine for stable soil carbon formation, the collapse of nucleotide metabolism may serve as an immediate precursor to the impairment of the soil carbon sink. To evaluate the physiological link between physical disturbance and early carbon dynamics, we conducted a short-term trial profiling metabolite responses to soil excavation. We found that within one month of physical exposure, thymidine and adenosine were depleted, whereas MNC and lignin remained unchanged. These results support that nucleotide pools collapse prior to observable carbon loss. Therefore, we hypothesize that the depletion of specific purines and pyrimidines modulated the efficiency of the microbial carbon cycle by disrupting central carbon and energy metabolism and cell proliferation, allowing these nucleotides to serve as quantitative early predictors for the decline of stable carbon pools in response to physical disturbance. To test our hypothesis, we integrated metabolomic, metagenomic, and geochemical profiling of paired natural and disturbed soils across a 621 km alpine transect to identify early-warning biomarkers of soil carbon impairment. Our study was designed to: (1) detect a conserved metabolic signature of disturbance that transcends spatial heterogeneity. (2) elucidate the underlying genetic mechanism driving these shifts via metagenomic analysis; and (3) construct a multi-omics pathway map to mechanistically link metabolic biomarkers to their genomic and taxonomic regulators. Consequently, we provide the first evidence that nucleotide depletion serves as a precursor to soil carbon collapse, offering a new metric for early diagnosis in soil carbon loss. Results 2.1 Evidence of Carbon Sink Impairment: Hysteresis in Macro-Carbon Indicators To investigate the early-stage impairment of the soil carbon sink driven by anthropogenic disturbance, we established a paired-group study design across a 621-km alpine transect (Fig. 1 a). Based on this framework, we first quantified traditional macro-carbon pools in a representative subset (n = 78) to validate their diagnostic responsiveness ( Supplementary Table 3 ). Consistent with our "diagnostic lag" hypothesis, the soil stable carbon pools exhibited a delayed response to physical disturbance. Specifically, while SOC, easily oxidizable carbon (EOC) and micro-AAOC were significantly depleted in disturbed soils ( P < 0.001, Fig. 1 b), stable fractions—such as lignin phenols, MNC and macro-AAOC—did not change significantly. This divergence provides direct evidence for "temporal lag". Because these stable fractions are physically protected or chemically recalcitrant, this protection delays the detection of reduced microbial carbon sequestration [ 16 , 17 ] . For instance, the physical disturbance disrupted soil architecture, resulting in a 79.5% reduction in micro AAOC (< 0.053 mm). Despite this significant change, the contribution ratio of MNC to total SOC (characterizing the size of soil stable organic carbon pool) remained statistically indistinguishable between the disturbed and natural groups (20.19% vs. 20.47%, Supplementary Table 3 ). These results confirm that while the carbon stock is already compromised, traditional stable fractions fail to capture the immediate decline of the microbial carbon pump. This temporal diagnostic gap necessitates the identification of more sensitive, instantaneous molecular indicators to monitor early-stage carbon degradation. 2.2 Metabolomic Profiling Identifies Nucleotide Depletion as a Sensitive Early-warning Signature To identify instantaneous molecular indicators capable of bridging the diagnostic gap identified above, we performed a targeted metabolomic analysis on the aggregate dataset (n = 204). By spanning a 621-km transect, we aimed to isolate metabolic signatures that are highly sensitive to physical disturbance and conserved across diverse alpine regional gradients. Multivariate analysis (PCA and OPLS-DA) demonstrated a clear metabolic separation between natural and disturbed soils across the entire elevational gradient ( R 2 Y = 0.721, Q 2 = 0.673, Fig. 2 b), indicating an immediate and consistent metabolic suppression driven by physical disturbance. Differential abundance analysis identified 47 significantly differential metabolites (SDMs) that were universally downregulated in disturbed soils (Fig. 2 c, d; Supplementary Table 4 ). KEGG pathway analysis reveals that SDMs were heavily enriched in nucleotide metabolism (purines and pyrimidines) rather than central carbohydrate pathways (Fig. 2 d). Specifically, key nucleotides—such as adenine, adenosine, thymine, thymidine, guanosine, and uridine—were strongly depleted in disturbed soil ( P < 0.0001, fold change < 0.5, Supplementary Table 4 ). Concurrently, we observed a severe suppression of L-glutamine, a critical metabolic hub bridging carbon fixation, nitrogen assimilation, and organic synthesis [ 18 ] . This suggests a disruption in the fundamental carbon-nitrogen flux that fuels microbial growth [ 19 , 20 ] . To identify the primary metabolic signature of soil disturbance, we distilled the 47 SDMs down to 10 core markers (Fig. 2 f; Supplementary Fig. 1a ). These 10 SDMs belong entirely to purine and pyrimidine nucleotide families and exhibited high sensitivity to infrastructural disturbance (Cohen’s d ranging from 0.87 to 4.60, Supplementary Table 5 ). Furthermore, pattern analysis via K-means clustering resolved the abundance trends of SDMs into 7 distinct patterns (Fig. 2 e), revealing that these nucleotide features followed a synchronized decline. This implies they act as a cohesively regulated functional module responding instantaneously to environmental disturbance. The specific suppression of these nucleotides—as opposed to generalized stress markers like proline [ 21 ] —indicates a disruption of DNA replication and high-energy substrate turnover [ 22 ] , signifying an immediate arrest of microbial proliferation upon disturbance. Finally, ROC analysis confirmed the diagnostic robustness of this module, yielding high AUC values ranging from 0.91 to 0.97 (Fig. 2 g). Furthermore, we reconstructed the metabolic network using KEGG pathway mapping to elucidate the mechanistic link between nucleotide depletion and the failure of carbon sequestration. L-glutamine serves as the mechanistic bridge connecting these distinct pathways: it serves as the precursor for de novo biosynthesis of both purines and pyrimidines (ko 00230 and ko 00240), while simultaneously acting as the core node feeding into central carbon and energy metabolism (via the TCA cycle), methane metabolism, and pyruvate metabolism (Fig. 2 h). This reveals that the early-stage suppression of soil carbon cycling is coupled to nucleotide depletion via the collapse of the L-glutamine hub. Finally, to confirm the relationship between the metabolome and stable carbon pools, Mantel tests demonstrated a significant correlation between the nucleotide matrix and the stable carbon indices (MNC, lignin, and macro AAOC) ( R > 0.6, P < 0.001; Supplementary Fig. 1b ). 2.3 Loss of the Carbon-Metabolizing Genome Drives Nucleotide Depletion and Disrupts Central Carbon Fixation To elucidate the specific genetic machinery driving the observed nucleotide depletion and carbon sink impairment, we conducted metagenomic sequencing on the matched soil samples (n = 204, Fig. 3 a). Functional profiling based on KEGG revealed clear shifts in central carbon and energy metabolism, notably within the Glycolysis/Gluconeogenesis pathways, Purine/Pyrimidine Metabolism, and Carbon Metabolism (Fig. 3 b). To pinpoint the exact drivers of this metabolic shift, we mapped the metagenomic reads against the curated carbon-cycle database (CCycDB). From an initial alignment of 3,744 C-genes, we isolated 48 altered candidates ( P < 0.05, Supplementary Fig. 2a-c ). We then intersected these candidates with genes governing the upstream biosynthesis and downstream degradation of our feature metabolites (thymidine, guanosine, adenosine, and L-glutamine). This locked down 5 core regulatory C-genes— pdp , deoA , gsk , pyrD and yrfG —that constitute the enzymatic backbone controlling nucleotide turnover (Fig. 3 c). Metabolic pathway analysis revealed that the suppression of nucleotides is achieved through a synchronized genomic downregulation. First, pyrD (dihydroorotate dehydrogenase), which governs the conversion of glutamine-derived precursors into the pyrimidine metabolism, Fig. 3 c, was significantly downregulated (FC = 0.14). This suppression hinders the flux of carbon and nitrogen from L-glutamine, cutting off the primary supply for UMP and Thymidine, Fig. 3 c. Simultaneously, the microbial capacity to recycle existing nucleotides (salvage pathways) [ 22 ] was also impaired. Specifically, we observed a severe depletion of deoA (thymidine phosphorylase, FC = 0.03) alongside an increase in pdp (pyrimidine-nucleoside phosphorylase, FC = 2.02), restricting the phosphorolysis of thymidine and preventing the recycling of deoxyribose-derived carbon back into glycolysis. Similarly, the imbalance between yrfG (downregulated, FC = 0.51)) and gsk (upregulated, FC = 3.54) promoted the net consumption of guanosine. Furthermore, we substantiated the link between these nucleotides-regulatory genes and soil carbon metabolism by identifying concurrent alterations in specific carbon fixation and degradation genes, notably korA (FC = 0.32), oorA (FC = 3.58), and cbbM (FC = 2.19). The downregulation of korA is significant, as this enzyme controls the entry of glutamine-derived carbon (via 2-oxoglutarate) into the TCA cycle. The simultaneous suppression of pyrD (nucleotide synthesis) and korA (TCA entry) functionally severs the "glutamine bridge," depriving the microbiome of both the genetic materials for replication and the carbon and energy substrates required for carbon stabilization, Fig. 3 c. To verify the spatial robustness of these core genetic signatures and ensure they are not artifacts of local sampling bias, we stratified the 204 samples into 5 distinct groups based on geographical location and altitude ( Supplementary Table 1,2 ) for validation. First, to assess local robustness, we performed 10-fold cross-validation and Leave-One-Out Cross-Validation (LOOCV) within each individual group (Fig. 3 d-h). Second, to test spatial transferability, we conducted 5-fold cross-group validation and Leave-One-Cohort-Out (LOCO) validation (Fig. 3 i). This testing yielded AUROC values ranging from 0.86 to 0.99. These results substantiate that the genomic alteration in this study is a highly conserved biological response to physical disturbance across the alpine transect, rather than a localized anomaly. Consistent with these functional and genetic disruptions, PCoA and ANOSIM of the microbiome revealed a clear separation between natural and disturbed soils (* P < 0.05, FDR 3.0) revealed a shift from “carbon-building” to “carbon-consuming” microbial groups ( Supplementary Fig. 2d ). Specifically, we observed a consistent depletion of taxa associated with AAOC synthesis, including Acidobacteriota , Alphaproteobacteria , Verrucomicrobiota , and Deltaproteobacteria [ 23 – 25 ] . Conversely, taxa harboring oligotrophic lifestyles specialize in degrading recalcitrant soil organic matter (e.g., Actinomycetota and Chloroflexota [ 26 – 28 ] ) were significantly enriched. These taxonomic shifts suggest that the microbial community, deprived of fresh inputs, pivoted toward consuming the stable carbon pool for survival. 2.4 Robustness Verification via Stratified Random Subsampling To rigorously validate the spatial robustness of these nucleotide markers—and to ensure they are not artifacts of local sampling bias, we implemented a 100-fold stratified random subsampling strategy. In each iteration, a subset of 100 samples was randomly drawn from the aggregate dataset (n = 204). This was constrained to maintain a balanced distribution between natural and disturbed groups (47–53 samples per group), effectively testing the markers against varying subsets of environmental heterogeneity. For each randomized subset, the full analytical workflow—including PCA, OPLS-DA modeling, differential analysis, and ROC evaluation—was executed independently. This iterative testing revealed stability. Specifically, the feature selection process showed convergence: Thymidine, Guanosine, and Adenosine were repeatedly identified as significantly differential metabolites (VIP > 1, P < 0.05, FC < 0.4) in 100%, 99%, and 98% of the trials, respectively (Fig. 4 d). Furthermore, ROC analysis across the 100 independent subsets yielded a stable mean ROC of 0.95, 0.97, and 0.92 for thymidine, guanosine, and adenosine, respectively. As visualized in Fig. 4 a-c, the narrow 95% confidence intervals (CI) surrounding the mean ROC curve confirm that the diagnostic performance of these nucleotide markers is not driven by specific regional outliers. Instead, their depletion represents a consistent biological response to physical disturbance across the alpine transect. 2.5 Multi-Omics Pathway Connects Nucleotide Depletion to Central Carbon Metabolism To highlight how nucleotide depletion connects to carbon fixation and the overall soil carbon stability, we integrated the identified taxonomic, genomic, and metabolic signatures into a unified biological correlation map. The resulting network reveals three connected components (Fig. 5 ), with L-Glutamine acting as the central metabolic hub, bridging purine/pyrimidine metabolism with central carbon and energy pathways. Specifically, the nucleotide markers exhibited strong positive correlations with the genes governing their biosynthesis ( pyrD , deoA , yrfG ) and with korA , the critical enzymatic bridge linking glutamine to the TCA cycle. These positive links extended to key carbon-fixing taxa (e.g., Acidobacteriota ). In contrast, we observed significant negative correlations between these nucleotides and the genes driving their degradation ( gsk , pdp ). This mirrored the enrichment of mineralization-associated taxa (e.g., Actinomycetota ). This pathway map provides confirmation of our central hypothesis: the impairment of soil carbon sequestration is a direct downstream consequence of the genetic blockade in nucleotide turnover. Discussion Nucleotide Depletion Suppresses Soil Stable Carbon Accumulation by Constraining the Ex Vivo Modification and In Vivo Turnover of Microbial Carbon Pump Detecting early soil carbon loss following anthropogenic disturbance is challenging due to the relative stability of macro-carbon pools (e.g., AAOC and MNC) [ 31 ] . Our integrated analysis of diverse soil carbon fractions and metabolomic profiles corroborates this, revealing that the impairment of microbial nucleotide turnover significantly precedes detectable loss in stable carbon pools. Nucleotide metabolites reside in the active intracellular pool and turn over rapidly, making them sensitive to environmental changes. However, why their depletion predicts soil carbon dynamics has remained unclear. By integrating targeted metabolomics and metagenomics across a 621-km alpine transect, we uncovered how physical disturbance impairs the process of soil microbial carbon sequestration by reducing the synthesis of key nucleotides. Physical disturbance reduces the microbial nucleotide supply primarily through the following two processes. First, the removal of vegetation directly severs the continuous supply of plant root exudates, depriving the rhizosphere microbiome of bioavailable precursors (e.g., L-glutamine) and exogenous nucleotides [ 32 ] . Second, the mechanical disruption of soil aggregates destroys microbial microhabitats, leading to a decrease in copiotrophic, "carbon-building" taxa (such as Acidobacteriota and Alphaproteobacteria responsible for synthesizing soil aggregate carbon and Betaproteobacteria which is associated with labile carbon turnover) [ 23 – 25 ] . Because these specific taxa are the primary carriers of the nucleotide-processing machinery, their decline manifests at the metagenomic level as a synchronized downregulation of both the de novo biosynthesis ( pyrD ) and salvage ( deoA ) pathways of nucleotides (Fig. 3 c, 5 a). This suggests that the observed genetic shift reflects a change in community composition rather than a transient stress response. Furthermore, the concurrent downregulation of korA blocks the entry of remaining L-glutamine into the TCA cycle, effectively severing the metabolic bridge between nucleotide synthesis and central energy metabolism. Consequently, the intracellular concentrations of key nucleotides, such as thymidine and L-guanosine, decrease significantly (Fig. 3 c,). Because nucleotides, including ATP, serve as the primary energy currency [ 15 ] and essential coenzymes, their depletion reduces the available energy and metabolic resources required for the two sequential stages of the microbial carbon pump (MCP) [ 33 ] : ex vivo modification and in vivo turnover of microorganisms. First, this energy limitation affects the ex vivo modification pathway, which is the initial step for microbial acquisition of environmental carbon [ 34 ] . This process relies on the synthesis and secretion of extracellular enzymes to depolymerize complex organic matter. However, the transcription, translation, and export of these enzymes are highly energy-dependent and require nucleotide-derived cofactors [ 35 , 36 ] . The reduction of the nucleotide pool restricts the ATP available to drive this machinery, decreasing the microbial capacity to process and acquire external carbon sources. Second, the reduced nucleotide pool compromises the in vivo turnover pathway. Even when simple carbon substrates are assimilated, the shortage of nucleotides—the fundamental building blocks for DNA and RNA [ 37 ] —limits microbial anabolism. This limitation directly suppresses microbial carbon use efficiency (CUE), a physiological trait that governs the partitioning of assimilated carbon [ 38 ] . Under energy constraints, microbes allocate a larger proportion of available carbon to respiratory maintenance rather than to biosynthesis [ 39 ] . This shift lowers CUE and limits cellular proliferation. Since the continuous generation of microbial biomass carbon (MBC) is the required precursor for the formation of stable microbial necromass carbon (MNC) [ 40 ] , the reduction in cell growth directly leads to a decrease in MNC accumulation, halting the transformation of labile organic matter into the long-term stable carbon pool. Ultimately, the combined constraints on ex vivo modification and in vivo turnover slow down the overall MCP process. Among the nucleotide markers, thymidine and guanosine showed the strongest and most consistent depletion (AUROC > 0.90). Thymidine is primarily derived from DNA turnover and its salvage pathway depends on deoA , which was severely downregulated. Guanosine, linked to purine metabolism, was similarly affected by the upregulation of gsk . The sensitivity of these specific nucleotides may reflect their roles in DNA replication and energy transfer. Unlike stable carbon content, intracellular nucleotides like thymidine and guanosine are not physically protected and turn over rapidly [ 41 ] . Therefore, their depletion reflects immediate soil stable carbon loss upon disturbance [ 42 , 43 ] . A Consistent Response to Environmental Heterogeneity A major challenge in soil omics is site-specificity, where biomarkers identified in one region fail to generalize due to environmental heterogeneity [ 44 , 45 ] . To rule out the influence of local environmental noise, we spanned a 621 km transect with great altitudinal (700–4300 m) and spatial gradients. The conserved performance of Thymidine in our 100-fold stratified random subsampling confirms that its rapid depletion is not an artifact of localized environmental noise. While the roles of central carbon genes in ecosystem functioning are well-established, the nucleotide-processing machinery identified here offers a novel diagnostic method. To determine if the suppression of this machinery represents a conserved response beyond physical excavation, we cross-referenced our findings with independent metagenomic datasets. Notably, we confirmed that the suppression of these nucleotide-processing genes represents a consistent response to diverse environmental stress (e.g., land-use change and chemical contamination). For example, the specific suppression of deoA was consistently observed in soil under pesticide and acid stress [ 46 ] . Similarly, the dysregulation of gsk and pdp mirror their well-established role in plant biology, where they serve as key modulators of growth metabolism and adaptation to diverse abiotic stresses (e.g., salt, hormone signaling, drought, and light stress) [ 47 – 50 ] . This external validation confirms that nucleotide depletion is a physiological response of the soil microbiome to environmental stress, supporting its use as an early molecular indicator of soil carbon loss. Practically, our findings offer targets for the rapid assessment of soil carbon sinks during ecological restoration. Future interventions should therefore transition from conventional physical stabilization to targeted metabolic or microbial resuscitation. Whether the soil carbon sequestration can be reactivated through targeted chemical amendments—such as exogenous nucleotide or L-glutamine addition—needs to be further investigated. Materials and Methods Study Design and Sample Collection To investigate soil carbon dynamics under anthropogenic disturbance, a 621-km transect was established spanning the transition from the Baiyu in Ganzi Tibetan Autonomous Prefecture to the Tianquan in Sichuan (elevations 700–4300 m), Fig. 1 a. The study employed a paired-group design comprising 204 soil samples (102 disturbed vs. 102 natural pairs). To account for spatial heterogeneity, the sampling transect was stratified into five distinct altitudinal cohorts: Baiyu (4000–4300 m), Litang (3600–4000 m), Yajiang (3200–3600 m), Kangding (2500–2800 m), and Tianquan (700–1000 m), Supplementary Table 1,2 . Disturbed samples were collected from tunnel construction sites exposed for less than three months, while matched natural controls were obtained from adjacent undisturbed forest ecosystems within a 1 km radius to minimize climatic and geological confounding. At each sampling site, 3 ~ 6 circular plots (radius 0.2 m) were randomly established ( Supplementary Fig. 3 ). Surface soil (5–20 cm) was collected from three points within each plot and pooled to form a composite sample. Samples were divided into two aliquots: one portion was air-dried for geochemical characterization (subset n = 78), and the remainder was sieved (2 mm) and flash-frozen at − 80°C for multi-omics analysis (total dataset n = 204). Quantification of soil stable carbon pools To quantify the 'temporal lag' of carbon fractions, we analyzed five key carbon pools in a representative subset (n = 78, Supplementary Table 3 ). This subset covered 13 sampling sites stratified into five regional cohorts based on altitude: Sites 1–2 (Baiyu), Sites 3–5 (Litang), Sites 6–9 (Yajiang), Sites 10–11 (Kangding), and Sites 12–13 (Tianquan). Soil aggregates were fractionated into macro- (> 0.25 mm), medium- (0.053–0.25 mm), and micro-aggregates (< 0.053 mm) via dry-sieving. Total SOC and AAOC was determined using the potassium dichromate-sulfuric acid oxidation method with external heating (digestion at 135°C for 30 min), and measured spectrophotometrically at 585 nm against a glucose standard. The organic carbon content within AAOC was quantified by titrimetric dichromate oxidation, followed by titration with 0.2 M FeSO4 using o-phenanthroline. EOC was assessed via the 333 mM KMnO4 oxidation method [ 50 ] . Briefly, samples were reacted with KMnO4 (25°C, 2 h, 120 rpm), and the consumed permanganate was measured spectrophotometrically at 565 nm (Shimadzu UV-1800, Kyoto, Japan). Lignin phenols and amino sugars were quantified using an Agilent 6890 GC-MS system (Agilent Technologies, CA, USA). Lignin phenols were released via alkaline CuO oxidation, extracted, purified, and derivatized to volatile trimethylsilyl (TMS) ethers. Quantification was performed using external calibration, and total lignin was calculated as the sum of five monomers: vanillic acid, acetosyringone, syringic acid, p-hydroxycinnamic acid, and ferulic acid. For microbial necromass carbon (MNC), soil samples (> 0.3 mg N) were hydrolyzed in 6 M HCl at 105°C for 8 h with myo-inositol as an internal standard. Following filtration, neutralization (pH 6.6–6.8), and desalting, amino sugars were derivatized using a mixture of 32 M hydrochloric carboxymethoxylamine and 40 M 4-dimethylaminopyridine (4:1, v/v; 80°C, 30 min), followed by acetylation with acetic anhydride. The final derivatives were dissolved in ethyl acetate-hexane (1:1) and injected (1.0 µL, split 10:1) using high-purity N 2 carrier gas (0.8 mL min − 1 ). The GC oven temperature was programmed as follows: 120°C (4 min hold), ramped at 10°C min − 1 to 230°C, 5°C min − 1 to 250°C (4 min hold), and finally 40°C min − 1 to 300°C (5 min hold). Bacterial and fungal necromass carbon were calculated from muramic acid and glucosamine concentrations, respectively, using established conversion coefficients. Targeted metabolomics and biomarker identification Metabolite Extraction and LC-MS/MS Analysis Soil metabolites were extracted and profiled using a targeted 600-MRM platform (Biotree, Shanghai). Briefly, 25 mg of soil was extracted using an acetonitrile-methanol-water system (2:2:1, v/v) spiked with isotopically labeled internal standards. Following homogenization with zirconia beads, low-temperature sonication, and centrifugation, the supernatant was analyzed. Chromatographic separation was performed on an Agilent 1290 UHPLC system equipped with a Waters Atlantis Premier BEH Z-HILIC column (1.7 µm, 2.1 × 150 mm). The mobile phase consisted of 10 mmol/L ammonium formate in water/acetonitrile (9:1, v/v) (Phase A) and 1:9 (v/v) (Phase B). Mass spectrometry was conducted on an AB Sciex QTrap 6500 + system with the following source parameters: IonSpray Voltage + 5500V/-4500V, Curtain Gas 35 psi, Temperature 400°C, and Ion Source Gas 1/2 at 50 psi. To monitor instrumental stability, quality control (QC) samples (pooled standard mixture) were injected every 10 samples ( Supplementary Fig. 4 ; Supplementary Table 6 ); the relative standard deviation (RSD) of internal standard retention times was maintained within ± 10 seconds (RSD ≤ 20%), confirming high-quality data acquisition [ 51 ] . Statistical Analysis and Biomarker Screening Raw data were processed using SCIEX Analyst Workstation (v1.7.3) and BIOTREE BioBud (v2.0.3) for peak integration and quantification. Preprocessing included filtration, missing value recoding, and internal standard normalization. Multivariate pattern recognition, including Principal Component Analysis (PCA) and Orthogonal Partial Least Squares-Discriminant Analysis (OPLS-DA), was performed using SIMCA (v18.0.1). Significantly Differential Metabolites (SDMs) were identified based on a strict combinatorial threshold [ 52 ] : (1) Variable Importance in Projection (VIP) score > 1; (2) Fold Change (FC) > 2 or < 0.5; and (3) False Discovery Rate (FDR)-adjusted P-value < 0.05 (Wilcoxon rank-sum test). Feature Selection and Diagnostic Validation To isolate the most robust signatures, SDMs were standardized via Z-score normalization and grouped into co-regulated functional modules using K-means clustering (R package cluster). The top 10 metabolites with the highest Z-score variance were selected as candidate biomarkers. For these candidates, biological effect sizes were quantified using Cohen’s d (difference in means divided by pooled SD). Subsequently, the diagnostic performance of this 10-metabolite signature was evaluated using Receiver Operating Characteristic (ROC) analysis (R package pROC or plotROC). The Area Under the Curve (AUROC) was calculated to quantify sensitivity and specificity, identifying markers with AUROC > 0.90 as robust indicators of soil disturbance. Finally, SDMs were mapped onto the KEGG database to elucidate perturbed metabolic networks, with a specific focus on carbon metabolism pathways. Robustness verification via stratified random subsampling To distinguish universal biological signatures from artifacts of local environmental heterogeneity (e.g., elevation or vegetation differences), a rigorous 100-fold stratified random subsampling strategy was implemented. In each of the 100 iterations, a subset of 100 samples was randomly drawn from the total dataset (n = 204), balanced between natural and disturbed groups. For each subset, differential abundance analysis and ROC evaluation were independently executed. A biomarker was considered robust only if it maintained statistical significance (VIP > 1, P < 0.05) in more than 95% of the iterations. This approach ensured that the identified markers represented a physiological response rather than site-specific noise. Validation of Metabolic Biomarkers via Mantel test To validate the quantitative predictive power of the identified biomarkers for soil carbon sequestration, the matrix of signature metabolites was correlated with soil stable carbon indices, including SOC, lignin phenols, MNC, AAOC (macro-, medium-, micro-), and EOC fractions. Global associations between the metabolic matrix and carbon factors were assessed via Mantel tests (999 permutations, R package vegan), which identified five metabolites significantly associated with the majority of carbon indicators. Metagenomic sequencing and functional profiling DNA Extraction, Sequencing, and Assembly Total genomic DNA was extracted from 204 soil samples using the Mo Bio PowerSoil DNA Isolation Kit (Mo Bio Laboratories, CA, USA). Paired-end sequencing (2 × 150 bp) was performed on an Illumina NovaSeq 6000 platform. Raw reads were processed using Fastp for quality control (low-quality trimming and adapter removal). High-quality reads were assembled into contigs using MEGAHIT with a multi-k-mer strategy. Contigs < 500 bp were discarded. Open Reading Frames (ORFs) were predicted using MetaGeneMark, and a non-redundant gene catalogue was constructed using MMseqs2 (95% identity, 90% coverage). Gene abundance was calculated based on read mapping (Bowtie2) and normalized to Transcripts Per Million (TPM) or relative abundance. Taxonomic Profiling and Biomarker Identification Taxonomic annotation was performed by aligning non-redundant genes against the NR database using DIAMOND (e-value < 1 × 10 − 5 ). Species-level abundance tables were rarefied to the minimum sequencing depth to standardize sampling effort. Alpha diversity (Chao1, Shannon) and Beta diversity (PCoA based on Bray-Curtis distance) were calculated using QIIME. Significant taxonomic differences between natural and disturbed soils were assessed via PERMANOVA and ANOSIM. To identify robust biomarkers, Linear Discriminant Analysis Effect Size (LEfSe) was applied (LDA score > 3.0, FDR-corrected P < 0.001). Functional Annotation and Key C-Gene Screening Functional profiling was conducted using the KEGG database for metabolic pathway reconstruction and CCycDB for specific carbon-cycle gene annotation. From an initial pool of 48 differentially abundant C-genes ( P < 0.05), we applied a mechanistic filtering strategy to isolate core drivers of metabolic disruption. Genes were retained only if they: (1) exhibited significant differential abundance between groups; and (2) encoded enzymes acting as direct upstream/downstream regulators of the identified nucleotide biomarkers (e.g., connecting Glutamine to UMP/TMP). This targeted screening narrowed the candidate list to 5 core regulatory genes ( pyrD , deoA , gsk , korA / oorA , cbbM ) ( https://ccycdb.github.io/ ). Robustness Validation of Genomic Signatures To verify the generalizability of the identified core genomic signatures, a dual-level validation framework was implemented using Random Forest classifiers (R package randomForest): (1) Local robustness (internal validation): Within each of the five regional cohorts, performance was evaluated using 10-fold cross-validation and Leave-One-Out Cross-Validation (LOOCV). (2) Spatial transferability (external validation): To test stability across environmental gradients, we performed 5-fold cross-group validation and Leave-One-Cohort-Out (LOCO) validation, where the model was trained on four regions and tested on the independent fifth region. Construction of Multi-Omics Pathway Map To elucidate the mechanistic links between the microbiome and the metabolome, multi-omics pathway map was constructed by integrating taxonomic source tracking and correlation networks. The taxonomic origins of signature metabolites and the carriers of differential C-genes were identified using the BIOTREE internal database (Biotree, Shanghai, China), which integrates curated associations from public repositories including KEGG, BIOML, CGR, HBC, and GMrepo. Intersection analysis was subsequently performed to isolate consistent species-gene-metabolite trios that were both differentially abundant and linked. Based on these validated interactions, an integrated framework was visualized to encapsulate the flow of carbon from specific microbial taxa through regulatory enzymes (C-genes) to metabolic end-products (signature metabolites). Declarations Competing interests The authors declare no competing interests. Author Contributions The study was conceived and designed by M.X., L.Q. and X.T. Sampling, and soil pretreatment were carried out by X.W. Y.C. and Y.L. The sequencing data from metagenomics and metabolomics were analyzed by C.T., K.T., and H.Z. The reliability analysis was assisted by H.T. The original draft was prepared by M.X., and all authors did the revisions. All authors read, discussed, and approved the final version of the manuscript. Acknowledgments This work was supported by various funding sources. We are grateful for the funding provided by the National Natural Science Foundation of China (grant no. 42307028 to M.X.), Sichuan Provincial Natural Science Foundation (grant no. 25QNJJ4171 to M.X.), the special fund of ‌National Environmental Protection Key Laboratory of Synergetic Control and Joint Remediation of Water & Soil Pollution (grant no. GHBK-2024-15 to M.X.), the Science and Technology Research Project of the Tianfu Yongxing Laboratory (grant numbers 2023KJGG06 to X.T. and X.P.), and the China Giant Panda Conservation and Research Center Foundation Project (grant no. CCRCGPBY202409 to M.X.). The authors would like to express their gratitude to the China State Railway Group Co., Ltd. for their assistance on soil sample collection at ongoing tunnel construction sites (Sichuan—Xizang section). We also thank Biotree Biotech Co., Ltd. (Shanghai, China) for the assistant with determination of metabolome and metagenome References Blakemore, R. J. Biomass Refined: 99% of Organic Carbon in Soils. Biomass 4(4), 1257–1300 (2024). Ma, W. et al. Carbon budgets and environmental controls in alpine ecosystems on the Qinghai-Tibet Plateau. CATENA 229, 107224(2023). Li, Q. et al. The link between landscape characteristics and soil losses rates over a range of spatiotemporal scales: hubei province, China. Inter. J. Environ. Res. Public Health 18(21), 11044 (2021). Chen, L. et al. Effect of organic material addition on active soil organic carbon and microbial diversity: A meta–analysis. Soil Till. Res. 241, 106128 (2024). Al–Shammary, A. A. G., Al–Shihmani, L. S. S., Fernandez–Galvez, J. & Caballero–Calvo, A. Optimizing sustainable agriculture: A comprehensive review of agronomic practices and their impacts on soil attributes. J Environ. Manage. 364, 121487 (2024). Xie, M.D. et al. Estuarine wetland tidal organic carbon activates microbial carbon pump and increases long–term soil carbon stability. CATENA 247, 108559 (2024). Yang, Y. et al. Unlocking mechanisms for soil organic matter accumulation: carbon use efficiency and microbial necromass as the keys. Glob. Change Biol . 31, (2025). Zhao, Y. et al. A global meta–analysis of land use change on soil mineral–associated and particulate organic carbon. Glob. Change Biol . 31, (2025). Slessarev, Eric W et al. Initial soil organic carbon stocks govern changes in soil carbon: Reality or artifact?. Glob.Change Biol . 29(5), 1239–1247 (2023). Zhang, T. et al. Temporal thresholds and depth–specific mechanisms of soil organic carbon stabilization during 65 years of revegetation in the Tengger Desert. J Environ. Manage. 385, 125633 (2025). Lavallee, J.M., Soong, J.L., Cotrufo, M.F. Conceptualizing soil organic matter into particulate and mineral-associated forms to address global change in the 21st century. Glob. Change Biol . 26(1), 261–273 (2020). Han, X. et al. Understanding soil carbon sequestration following the afforestation of former arable land by physical fractionation. CATENA 150, 317–327 (2017). Li, L. et al. Assessment of transcriptional reprogramming of lettuce roots in response to chitin soil amendment. Front. Plant Sci. 14, 1158068 (2023). Mazzoleni, S. et al. Metabolomic changes in Arabidopsis thaliana exposed to extracellular self–and nonself–DNA: A reversible effect. Environ. Exp. Bot. 234, 106149 (2025). Heinke, L. Good neighbours transfer nucleotides. Nat. Rev. Mol. Cell Biol. 26, 582 (2025). Han, C. et al. Joint regulation of the soil organic carbon accumulation by mineral protection and microbial properties following conservation practices. Catena 245, 108298 (2024). Hu, P. et al. Lithological controls on soil aggregates and minerals regulate microbial carbon use efficiency and necromass stability. Environ. Sci. Technol. 48, 58, (2024). Georgiou, K. et al. Emergent temperature sensitivity of soil organic carbon driven by mineral associations. Nat. Geosci. 17(3), p205 (2024). Bogati, K. & Walczak, M. The impact of drought stress on soil microbial community, enzyme activities and plants. Agronomy Basel 12, (2022). Bhattacharjya, S. et al. Utilizing soil metabolomics to investigate the untapped metabolic potential of soil microbial communities and their role in driving soil ecosystem processes: A review. Appl. Soil. Ecol. 195, 105238 (2024). Zhang, J. et al. Enhancing rice salt tolerance: mechanisms of compound functional liquid in alleviating salt stress during the seedling stage. Plant Physiol. Biochem. 229, 110273 (2025). Heinke, L. Good neighbours transfer nucleotides. Nat. Rev. Mol. Cell Biol. 26, 582 (2025). Wirbel, J. et al. Meta–analysis of fecal metagenomes reveals global microbial signatures that are specific for colorectal cancer. Nat. Med. 25, 679–689 (2019). Yang, X. et al. Rice Ferredoxin–Dependent Glutamate Synthase Regulates Nitrogen–Carbon Metabolomes and Is Genetically Differentiated between japonica and indica Subspecies. Mol. Plant 9, 1520–1534 (2016). Li, Y. et al. Soil organic carbon loss decreases biodiversity but stimulates multitrophic interactions that promote belowground metabolism. Glob. Change Biol . 30, e17101 (2024). Gralka, M., Pollak, S. & Cordero, O.X. Genome content predicts the carbon catabolic preferences of heterotrophic bacteria. Nat. Microbiol. 8, 1799 (2023). He, C. et al. Decoupled fungal and bacterial functional responses to biochar amendment drive rhizosphere priming effect on soil organic carbon mineralization. Biochar 6, 84(2024). Goldfarb, K.C. et al. Differential growth responses of soil bacterial taxa to carbon substrates of varying chemical recalcitrance. Front. Microbiol. 2, 94 (2011). Mogens, K. et al, Nucleotide metabolism and its control in lactic acid bacteria. FEMS Microbiol. Rev. 29(3), 555–590 (2005). Keenor, S.G., Lee, R.; Reid, B.J. Physical protection of soil carbon stocks under regenerative agriculture. Soil 11(2), 957–973 (2025). Amundson, R. The Pandora's box of soil carbon. PNAS 119(11), e2201077119 (2022). Deng L. et al. Autotoxic ginseno-side stress induces changes in root exudates to recruit the beneficial Burkholderia strain B36 as revealed by transcriptommic and metabolomic approaches. J. Agric. Food Chem . 71(11), 4536–49 (2023). Jiao, N. et al. The microbial carbon pump and climate change. Nat. Rev. Microbiol. 22(7), 408–419 (2024). Michael, L. & Georgi, K.M. The bioenergetic costs of a gene. PNAS . 112(51), 15690–15695 (2015). Wang, C.Q. & Kuzyakov, Y. Energy use efficiency of soil microorganisms: driven by carbon recycling and reduction. Glob. Change biol. 29(22), 6170–6187 (2023). Witte, C., & Herde, M. Nucleotide metabolism in plants. Plant Physiol . 182(1), 63–78 (2020). Allison, S.D. Rethinking microbial carbon use efficiency in soil models. Nat. Climate Change 5(1), 56–60 (2025). Liu, X.J.A. et al. Soil aggregate-mediated microbial responses to long-term warming. Soil Biol. Biochem . 152, 108055 (2021). Peng, Z. et al. Plant detritus carbon dominates over microbial necromass carbon in topsoil of alpine ecosystems. Commun. Earth Environ . 6(1), 912 (2025). Gunina, A., & Kuzyakov, Y. From energy to (soil organic) matter. Glob.Change Biol . 28(7), 2169–2182 (2022). Zheng, Y. et al. Purines enrich rootassociated Pseudomonas and improve wild soybean growth under salt stress. Nat Commun. 15(1), 3520 (2024). Liu, H. et al. Nucleotides enriched under heat stress recruit beneficial rhizomicrobes to protect plants from heat and root-rot stresses. Microbiome 13, 160 (2025). Hansen, P.M. et al. Microbial community regulation of extracellular enzyme production can mediate patterns of particulate and mineral-associated organic matter accumulation in undersaturated soils. Soil Biol. Biochem. 214, 110056 (2026). Zhou, H.B. et al. Soil heterogeneity influences the biotoxicity of photoaged tire wear particles in Eisenia fetida: A comparative assessment. J. Hazard. Mater. 496, 12 (2010). Yun, Y.F., Gai, J.Y., Zhao, T.J. Identification of regulated genes conferring resistance to high concentrations of glyphosate in a new strain of enterobacter. FEMS Microbiol. Letter. 349(2), 135–143 (2013). Huang, S.H. et al. Genome-wide identification and expression analysis of the GSK gene family in Solanum tuberosum L. under abiotic stress and phytohormone treatments and functional characterization of StSK21 involvement in salt stress. Gene 766, 145156 (2021). Zhang, Peipei, et al. "Genome-wide identification and expression analysis of the GSK gene family in wheat (Triticum aestivum L.). Molec. Biol. Rep. 49(4), 2899–2913 (2022). Wang, L.L. et al. Genome-wide characterization and phylogenetic analysis of GSK gene family in three species of cotton: Evidence for a role of some GSKs in fiber development and responses to stress. BMC Plant Biol. 18, 1 (2018). Xu, X.F. et al. GmGGDR gene confers abiotic stress tolerance and enhances vitamin E accumulation in arabidopsis and soybeans. Agronomy 15(2), 351 (2025). Blair, G.J., Lefroy, R.D.B. & Lise, L. Soil carbon fractions based on their degree of oxidation, and the development of a carbon mangement index for agricultural systems. Austral. J Agricul. Res. 46(7), 1459–1466 (1995). Yuan, M., Breitkopf, S., Yang, X. & Asara, J.M. A positive/negative ion–switching, targeted mass spectrometry–based metabolomics platform for bodily fluids, cells, and fresh and fixed tissue. Nat. Protoc. 7, 872–881 (2012). Trivedi, P. et al. Microbial regulation of the soil carbon cycle: evidence from gene–enzyme relationships. ISME J. 10, 2593–2604 (2016). Additional Declarations There is NO Competing Interest. Supplementary Files SupplementaryfilesCEE.docx supplementary file floatimage1.png Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-9373312","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":620816187,"identity":"f794f73b-25bf-45b5-8670-4d74585ee050","order_by":0,"name":"Mengdi Xie","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Mengdi","middleName":"","lastName":"Xie","suffix":""},{"id":620816188,"identity":"9828e7f6-d434-413b-8204-493b07259f51","order_by":1,"name":"Qiang Lin","email":"","orcid":"https://orcid.org/0000-0001-8315-3464","institution":"University of Antwerp","correspondingAuthor":false,"prefix":"","firstName":"Qiang","middleName":"","lastName":"Lin","suffix":""},{"id":620816189,"identity":"ee0deece-2ca7-4111-91df-5800884a1886","order_by":2,"name":"Lingling Feng","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Lingling","middleName":"","lastName":"Feng","suffix":""},{"id":620816190,"identity":"a58396de-f0e1-4004-baaa-03d050a848b4","order_by":3,"name":"Huan Zhao","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Huan","middleName":"","lastName":"Zhao","suffix":""},{"id":620816191,"identity":"cea58106-7bec-4ffc-9e52-cda6cc7cb254","order_by":4,"name":"Can Tang","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Can","middleName":"","lastName":"Tang","suffix":""},{"id":620816192,"identity":"85e261aa-6a6c-4253-bbb3-439cfec4f456","order_by":5,"name":"Ke Tan","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Ke","middleName":"","lastName":"Tan","suffix":""},{"id":620816193,"identity":"d5f0709f-3fe0-4373-82b3-57fbda801bcd","order_by":6,"name":"Tristram Hales","email":"","orcid":"https://orcid.org/0000-0002-3330-3302","institution":"Cardiff University","correspondingAuthor":false,"prefix":"","firstName":"Tristram","middleName":"","lastName":"Hales","suffix":""},{"id":620816194,"identity":"c6fc6f10-9d46-4c52-a34b-c7627be6182f","order_by":7,"name":"Xiaoyi Wu","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Xiaoyi","middleName":"","lastName":"Wu","suffix":""},{"id":620816195,"identity":"661361e4-f825-445e-bee8-f76b8f4b78e1","order_by":8,"name":"Yixuan Cheng","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Yixuan","middleName":"","lastName":"Cheng","suffix":""},{"id":620816196,"identity":"9f2f1f87-873f-4445-9f85-878f4d073fbd","order_by":9,"name":"Yishi Lin","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Yishi","middleName":"","lastName":"Lin","suffix":""},{"id":620816197,"identity":"6c670ad7-7b19-4a90-968b-8e75cbf60704","order_by":10,"name":"Xiangjun Pei","email":"","orcid":"","institution":"","correspondingAuthor":false,"prefix":"","firstName":"Xiangjun","middleName":"","lastName":"Pei","suffix":""},{"id":620816186,"identity":"967ed494-bcdc-46f7-9715-6c147fcfd1c4","order_by":11,"name":"Xiaolu Tang","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAAyklEQVRIiWNgGAWjYBACfvbGxgcfeGzk+NkbiNQi2XO42XCGTJqxZM8BIrUY3Ehvk+axOZS44UYCsS67kdgmOSPnQOLMmY833mCosYkmqIOx52GzxYczd4z7pdOKLRiOpeU2ENLCzJ7YeHNmzzPZmbNzzCQYGw4T1sLGkNggzfvvMOOGm2eI1MLDkdgkzcNzWHHDDR4itUjwHAQGMg8okIF+SSDGL/bH2x9Co/LwxhsfamwIa0EGBhIJpCiHaCFVxygYBaNgFIwMAACs5Ebu8RwmAwAAAABJRU5ErkJggg==","orcid":"https://orcid.org/0000-0001-5079-7650","institution":"Chengdu University of Technology","correspondingAuthor":true,"prefix":"","firstName":"Xiaolu","middleName":"","lastName":"Tang","suffix":""}],"badges":[],"createdAt":"2026-04-10 01:45:53","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-9373312/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-9373312/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":107900090,"identity":"f3c6c1da-ada5-4a1c-9c2e-0ea6a4d8aa3a","added_by":"auto","created_at":"2026-04-27 11:27:09","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":764633,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eStudy sampling framework and the variations of soil carbon pools across different regions.\u003c/strong\u003e \u003cstrong\u003ea\u003c/strong\u003e, study area and sampling scheme. A map illustrating the 621 km transect spanning varying elevations and geographic regions. A total of 204 soil samples were collected and stratified into natural (n = 102) and disturbed (n = 102) groups. \u003cstrong\u003eb\u003c/strong\u003e, Quantitative profiling of soil organic carbon fractions in the representative subset of soil samples (n = 78). Samples were collected from 13 representative sites stratified into five regional cohorts: Sites 1–2 (Baiyu), Sites 3–5 (Litang), Sites 6–9 (Yajiang), Sites 10–11 (Kangding), and Sites 12–13 (Tianquan). While SOC and easily oxidizable carbon (EOC) exhibits significant depletion in disturbed soils compared to natural baselines (\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001), stable fractions (lignin phenols, MNC, macro-AAOC) show statistically limited reduction (\u003cem\u003eP\u003c/em\u003e \u0026gt; 0.001), illustrating the \"masking effect\" of physical disturbance.\u003c/p\u003e","description":"","filename":"floatimage2.png","url":"https://assets-eu.researchsquare.com/files/rs-9373312/v1/cc42c5f11a584271a5f5e8a7.png"},{"id":107900113,"identity":"748f7f35-631c-4cbf-9516-babbd1c259e1","added_by":"auto","created_at":"2026-04-27 11:27:12","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":1169612,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cbr\u003e\n\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eUniversal metabolomic signatures and the central role of glutamine in carbon flux disruption. a\u003c/strong\u003e, Overall study design. The unified metabolomics dataset comprises samples collected across diverse environmental gradients (elevation and geography), analyzed as an aggregate to identify conserved metabolic responses to physical disturbance.\u003cstrong\u003e b\u003c/strong\u003e, Multivariate statistical analysis. PCA and OPLS-DA score plots demonstrate a robust separation between natural and disturbed soils across the aggregate dataset (R\u003csup\u003e2\u003c/sup\u003eY = 0.721, Q\u003csup\u003e2\u003c/sup\u003e = 0.673, \u003cem\u003eP\u003c/em\u003e \u0026lt; 0.05), confirming a consistent metabolic suppression regardless of regional heterogeneity.\u003cstrong\u003e c\u003c/strong\u003e, Differential abundance profiling. Volcano plot visualizing the distribution of metabolites. Blue dots represent the 47 Significantly Differential Metabolites (SDMs), all of which are universally downregulated in disturbed soils. Point size corresponds to OPLS-DA VIP values.\u003cstrong\u003e d\u003c/strong\u003e, Chemical taxonomy of SDMs. Classification reveals a specific enrichment of downregulated markers in nucleotide metabolism (purines and pyrimidines) rather than central carbohydrate pathways (\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.0001).\u003cstrong\u003e e\u003c/strong\u003e, Functional clustering. K-Means analysis resolves the evolutionary trends of SDMs into 7 distinct patterns based on Z-score standardized content. Core nucleotide signatures (e.g., Adenine, Thymidine) cluster together (Cluster 5), indicating a coordinated functional response to disturbance.\u003cstrong\u003e f\u003c/strong\u003e, Top predictive markers. Z-score ranking highlight the top 10 core nucleotide metabolites as the primary signature of soil disturbance. \u003cstrong\u003eg\u003c/strong\u003e, Diagnostic performance. Receiver Operating Characteristic (ROC) analysis of the signature metabolites. High area under the ROC curve (AUROC) values (\u0026gt; 0.9) confirm their robustness as universal biomarkers for detecting soil carbon impairment. \u003cstrong\u003eh\u003c/strong\u003e, Mechanistic reconstruction of the metabolic network. The KEGG pathway map illustrates the synergistic collapse of nucleotide and central carbon metabolism. Glutamine is highlighted as the mechanistic bridge: it acts as the obligate precursor linking the TCA cycle (Ko01200) and de novo nucleotide biosynthesis (Ko00230, Ko00240), indicating that the observed nucleotide depletion stems from the suppression of glutamine-mediated carbon flux.\u003c/p\u003e","description":"","filename":"floatimage3.png","url":"https://assets-eu.researchsquare.com/files/rs-9373312/v1/ab98e0d942641acbe0915ef1.png"},{"id":107900088,"identity":"43b12ff4-394e-4617-9fb4-0391a00ed936","added_by":"auto","created_at":"2026-04-27 11:27:08","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":954010,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMechanistic suppression of the carbon-metabolizing microbiome as a validation of metabolomic changes. a\u003c/strong\u003e, Overall composition of metagenomic datasets across the disturbed sol and natural soil (n = 204). \u003cstrong\u003eb\u003c/strong\u003e, Functional genomic profiling. Heatmap of KEGG genomic functional annotations showing the broad suppression of genes in Central Carbon Metabolism (Glycolysis, Pyruvate metabolism) and Nucleotide Metabolism pathways in disturbed soils. \u003cstrong\u003ec\u003c/strong\u003e, Mechanistic blockade of nucleotide turnover. A reconstructed metabolic pathway map integrating the 8 core regulatory C-genes (\u003cem\u003epyrD, deoA, yrfG, gsk, korA, oorA, and cbbM\u003c/em\u003e) identified via CCycDB screening. The diagram illustrates the systemic collapse across three interconnected modules: The downregulation of \u003cem\u003epyrD\u003c/em\u003e blocks the utilization of L-Glutamine for pyrimidine synthesis; The depletion of \u003cem\u003edeoA\u003c/em\u003e and \u003cem\u003eyrfG\u003c/em\u003eprevents the recycling of nucleosides; ;The suppression of \u003cem\u003ekorA\u003c/em\u003e marks the disconnection of the Glutamine-to-TCA bridge, hindering the entry of nitrogenous carbon into the central metabolic cycle. Blue spots indicate downregulated metabolites or genes, and red spots indicate upregulated ones.\u003cstrong\u003ed\u003c/strong\u003e-\u003cstrong\u003eh\u003c/strong\u003e, Local robustness validation within stratified cohorts. Receiver Operating Characteristic (ROC) curves evaluating the diagnostic performance of the 14-species Random Forest classifier within each of the five geographically distinct groups (stratified by altitude and region). For each group, model stability was assessed using both 10-fold cross-validation (blue lines) and Leave-One-Out Cross-Validation (LOOCV) (dark red lines). The high AUROC values (0.86~0.99) confirm the model's accuracy within specific local environments. i, Spatial transferability and generalization assessment. ROC curves representing the model's performance in cross-group validation scenarios. The curves display the results of 5-fold cross-group validation and Leave-One-Cohort-Out (LOCO) validation (training on n-1 groups and testing on the excluded group). The consistent performance (AUROC = 0.94 and 0.83) demonstrates that the identified microbial signature is universally applicable and robust against spatial and environmental heterogeneity. \u003cstrong\u003ej\u003c/strong\u003e,\u003cstrong\u003ek\u003c/strong\u003e,\u003cstrong\u003e \u003c/strong\u003eCommunity-level divergence. Principal Coordinate Analysis (PCoA) based on Bray-Curtis distances reveals a significant separation of microbial communities between natural and disturbed soils across the aggregate dataset (\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.001), further validated by Analysis of Similarity (ANOSIM, R \u0026gt; 0.5). The inset violin plots show the distribution of Alpha-diversity indices (Chao1 and Shannon), indicating that structural re-organization, rather than species loss, drives the observed differences.\u003c/p\u003e","description":"","filename":"floatimage4.png","url":"https://assets-eu.researchsquare.com/files/rs-9373312/v1/f91d84a6d784ee2641e92dac.png"},{"id":107900114,"identity":"61ce9023-39f8-4a13-ae21-5dcb9e184ecb","added_by":"auto","created_at":"2026-04-27 11:27:12","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":364545,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cbr\u003e\n\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eMulti-omics synthesis and predictive modeling of soil carbon sequestration. a\u003c/strong\u003e, The Multi-Omics Biological Correlation (MOBC) map. A network visualizing the correlations between the signature microbial taxa, the 4 core regulatory genes, and the 10 nucleotide markers. Blue lines indicate positive correlations (e.g., between Thymidine, \u003cem\u003edeoA\u003c/em\u003e, and \u003cem\u003eAcidobacteriota\u003c/em\u003e), while red lines indicate negative correlations (e.g., between Guanosine and the \u003cem\u003egsk\u003c/em\u003e gene). The topology shows that the decrease in nucleotides is statistically coupled with the loss of specific synthesis genes and bacterial taxa. \u003cstrong\u003eb\u003c/strong\u003e, Cross-scale statistical linkage. Mantel test results demonstrating the significant correlations between the nucleotide matrix and macro-scale stable carbon fractions (MNC, lignin, AC). Line width corresponds to the Mantel R statistic (\u003cem\u003eP\u003c/em\u003e \u0026lt; 0.05). \u003cstrong\u003ec\u003c/strong\u003e, Predictive regression models. Linear regression analysis establishing the quantitative relationships between the two genetically-anchored markers (Thymidine and Guanosine) and soil stable carbon indices (SOC, lignin and MNC). The high correlation coefficients (\u003cem\u003eR\u003c/em\u003e = 0.8) validate the utility of Thymidine as sensitive indicators for assessing soil carbon health, while the embedded values indicate the robustness verified via 5-fold cross-validation (mean \u003cem\u003eR\u003c/em\u003e\u003csup\u003e2\u003c/sup\u003e = 0.75, max \u003cem\u003eR\u003c/em\u003e\u003csup\u003e2\u003c/sup\u003e = 0.97).\u003c/p\u003e","description":"","filename":"floatimage5.png","url":"https://assets-eu.researchsquare.com/files/rs-9373312/v1/567d48c71c00a2dfc16a8913.png"},{"id":107900057,"identity":"77374b92-5862-4266-95fa-3bb720b03770","added_by":"auto","created_at":"2026-04-27 11:27:05","extension":"png","order_by":5,"title":"Figure 5","display":"","copyAsset":false,"role":"figure","size":288076,"visible":true,"origin":"","legend":"\u003cp\u003e\u003cstrong\u003eMulti-omics pathway network depicting the response of the carbon cycle-related metagenome and metabolome to physical disturbance. \u003c/strong\u003eThis map reveals the correlations among 10 nucleotide signatures, 8 core regulatory genes, and signature microbial taxa. Notably, L-glutamine serves as a core node bridging central carbon, purine, and pyrimidine metabolism. Guanosine and thymidine emerge as the most predictive markers for alterations in the carbon-related microgenome. Edge colors denote correlation types: blue lines represent positive associations (e.g., among thymidine, \u003cem\u003edeoA\u003c/em\u003e, and \u003cem\u003eAcidobacteriota\u003c/em\u003e), whereas red lines signify negative associations (e.g., between guanosine and \u003cem\u003egsk\u003c/em\u003e). The overall network topology demonstrates that nucleotide depletion is statistically coupled with the loss of specific synthesis genes and associated bacterial taxa.\u003c/p\u003e","description":"","filename":"floatimage6.png","url":"https://assets-eu.researchsquare.com/files/rs-9373312/v1/75558318de2864c3bca2e7d9.png"},{"id":109067720,"identity":"5de5876d-1bdf-4c7a-a164-129eed3e5ad7","added_by":"auto","created_at":"2026-05-12 10:00:15","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3927957,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-9373312/v1/e7b000e9-eb89-4e5a-9acb-e581f70a0579.pdf"},{"id":107900048,"identity":"701f936b-b221-4ee0-a572-76b8c0563390","added_by":"auto","created_at":"2026-04-27 11:26:57","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":1255908,"visible":true,"origin":"","legend":"supplementary file","description":"","filename":"SupplementaryfilesCEE.docx","url":"https://assets-eu.researchsquare.com/files/rs-9373312/v1/769f5972788dc415cfdcf040.docx"},{"id":107900112,"identity":"155d572d-9830-43f3-9976-d4779b1aceab","added_by":"auto","created_at":"2026-04-27 11:27:12","extension":"png","order_by":2,"title":"","display":"","copyAsset":false,"role":"supplement","size":1116797,"visible":true,"origin":"","legend":"","description":"","filename":"floatimage1.png","url":"https://assets-eu.researchsquare.com/files/rs-9373312/v1/d521a9585231d1c90825acf6.png"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Nucleotide depletion signals early-stage soil stable carbon collapse in anthropogenically disturbed alpine ecosystems","fulltext":[{"header":"Introduction","content":"\u003cp\u003eGlobal soil organic carbon (SOC) stocks, estimated at 1500\u0026ndash;2400 Pg, constitute the largest active terrestrial carbon reservoir\u003csup\u003e[1]\u003c/sup\u003e. In alpine ecosystems such as the Qinghai-Tibet Plateau, SOC pools are highly vulnerable to recent infrastructure expansion\u003csup\u003e[1,2]\u003c/sup\u003e. The construction of tunnels, highways, and pipelines imposed severe physical disturbances\u0026mdash;such as vegetation removal and soil aggregate disruption\u003csup\u003e[3]\u003c/sup\u003e. Consequently, with estimated carbon losses of 7\u0026ndash;12 tons per kilometer of road construction\u003csup\u003e[4,5]\u003c/sup\u003e, large-scale projects in China could drive massive SOC depletion on the order of thousands of tons.\u003c/p\u003e\n\u003cp\u003eWhile these direct physical losses are calculable, monitoring the stability of the remaining soil carbon pool is hindered by a temporal lag. Current monitoring relies largely on \u0026quot;macro-indicators\u0026quot; such as aggregate-associated organic carbon (AAOC) and microbial necromass carbon (MNC)\u003csup\u003e[6-8]\u003c/sup\u003e. While these metrics define the cumulatively stable carbon status, they exhibit significant temporal lag\u003csup\u003e[9]\u003c/sup\u003e. MNC represents a slow-turnover pool, often revealing detectable losses only years after the initial environmental changes (such as ecological restoration and physical disturbance)\u003csup\u003e[10]\u003c/sup\u003e. Because AAOC is stabilized within organo-mineral complexes through adsorption onto soil colloids, its response to environmental disturbance is significantly delayed\u003csup\u003e[11]\u003c/sup\u003e. Effective ecological restoration requires early warning systems to track carbon dynamics in real time. This diagnostic delay limits the capacity for real-time monitoring: by the time statistical decreases in soil stable carbon pool, such as MNC and AAOC are detected, the microbial processes governing carbon sequestration may have already been compromised\u003csup\u003e[12]\u003c/sup\u003e, ultimately jeopardizing the long-term stability of the soil carbon pool.\u003c/p\u003e\n\u003cp\u003eTo overcome this diagnostic delay, indicators that reflect the instantaneous physiological state of the soil microbiome are required. Unlike macro-carbon pools, microbial communities are highly sensitive to environmental disturbances (e.g., habitat destruction and vegetation removal), altering their turnover rates immediately upon stress\u003csup\u003e[13,14]\u003c/sup\u003e. Intracellular metabolites\u0026mdash;specifically purine and pyrimidine nucleotides\u0026mdash;are tightly coupled to these microbial dynamics. As the fundamental building blocks for cell proliferation and the primary currency for cellular energy transfer (e.g., ATP/GTP)\u003csup\u003e[15]\u003c/sup\u003e, the soil nucleotide pool scales directly with active microbial biomass. Consequently, physical excavation that induces a sudden disruption of nutrient inputs causes a rapid depletion in nucleotide concentrations. Because microbial biomass turnover is the primary engine for stable soil carbon formation, the collapse of nucleotide metabolism may serve as an immediate precursor to the impairment of the soil carbon sink.\u003c/p\u003e\n\u003cp\u003eTo evaluate the physiological link between physical disturbance and early carbon dynamics, we conducted a short-term trial profiling metabolite responses to soil excavation. We found that within one month of physical exposure, thymidine and adenosine were depleted, whereas MNC and lignin remained unchanged. These results support that nucleotide pools collapse prior to observable carbon loss. Therefore, we hypothesize that the depletion of specific purines and pyrimidines modulated the efficiency of the microbial carbon cycle by disrupting central carbon and energy metabolism and cell proliferation, allowing these nucleotides to serve as quantitative early predictors for the decline of stable carbon pools in response to physical disturbance.\u003c/p\u003e\n\u003cp\u003eTo test our hypothesis, we integrated metabolomic, metagenomic, and geochemical profiling of paired natural and disturbed soils across a 621 km alpine transect to identify early-warning biomarkers of soil carbon impairment. Our study was designed to: (1) detect a conserved metabolic signature of disturbance that transcends spatial heterogeneity. (2) elucidate the underlying genetic mechanism driving these shifts via metagenomic analysis; and (3) construct a multi-omics pathway map to mechanistically link metabolic biomarkers to their genomic and taxonomic regulators. Consequently, we provide the first evidence that nucleotide depletion serves as a precursor to soil carbon collapse, offering a new metric for early diagnosis in soil carbon loss.\u003c/p\u003e"},{"header":"Results","content":"\u003cp\u003e \u003cb\u003e2.1 Evidence of Carbon Sink Impairment: Hysteresis in Macro-Carbon Indicators\u003c/b\u003e \u003c/p\u003e \u003cp\u003eTo investigate the early-stage impairment of the soil carbon sink driven by anthropogenic disturbance, we established a paired-group study design across a 621-km alpine transect (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea). Based on this framework, we first quantified traditional macro-carbon pools in a representative subset (n\u0026thinsp;=\u0026thinsp;78) to validate their diagnostic responsiveness (\u003cb\u003eSupplementary Table\u0026nbsp;3\u003c/b\u003e). Consistent with our \"diagnostic lag\" hypothesis, the soil stable carbon pools exhibited a delayed response to physical disturbance. Specifically, while SOC, easily oxidizable carbon (EOC) and micro-AAOC were significantly depleted in disturbed soils (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001, Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb), stable fractions\u0026mdash;such as lignin phenols, MNC and macro-AAOC\u0026mdash;did not change significantly.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eThis divergence provides direct evidence for \"temporal lag\". Because these stable fractions are physically protected or chemically recalcitrant, this protection delays the detection of reduced microbial carbon sequestration\u003csup\u003e[\u003cspan citationid=\"CR16\" class=\"CitationRef\"\u003e16\u003c/span\u003e, \u003cspan citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e]\u003c/sup\u003e. For instance, the physical disturbance disrupted soil architecture, resulting in a 79.5% reduction in micro AAOC (\u0026lt;\u0026thinsp;0.053 mm). Despite this significant change, the contribution ratio of MNC to total SOC (characterizing the size of soil stable organic carbon pool) remained statistically indistinguishable between the disturbed and natural groups (20.19% vs. 20.47%, \u003cb\u003eSupplementary Table\u0026nbsp;3\u003c/b\u003e). These results confirm that while the carbon stock is already compromised, traditional stable fractions fail to capture the immediate decline of the microbial carbon pump. This temporal diagnostic gap necessitates the identification of more sensitive, instantaneous molecular indicators to monitor early-stage carbon degradation.\u003c/p\u003e \u003cp\u003e \u003cb\u003e2.2 Metabolomic Profiling Identifies Nucleotide Depletion as a Sensitive Early-warning Signature\u003c/b\u003e \u003c/p\u003e \u003cp\u003eTo identify instantaneous molecular indicators capable of bridging the diagnostic gap identified above, we performed a targeted metabolomic analysis on the aggregate dataset (n\u0026thinsp;=\u0026thinsp;204). By spanning a 621-km transect, we aimed to isolate metabolic signatures that are highly sensitive to physical disturbance and conserved across diverse alpine regional gradients. Multivariate analysis (PCA and OPLS-DA) demonstrated a clear metabolic separation between natural and disturbed soils across the entire elevational gradient (\u003cem\u003eR\u003c/em\u003e\u003csup\u003e2\u003c/sup\u003e\u003cem\u003eY\u003c/em\u003e\u0026thinsp;=\u0026thinsp;0.721, \u003cem\u003eQ\u003c/em\u003e\u003csup\u003e2\u003c/sup\u003e\u0026thinsp;=\u0026thinsp;0.673, Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb), indicating an immediate and consistent metabolic suppression driven by physical disturbance. Differential abundance analysis identified 47 significantly differential metabolites (SDMs) that were universally downregulated in disturbed soils (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ec, d; \u003cb\u003eSupplementary Table\u0026nbsp;4\u003c/b\u003e). KEGG pathway analysis reveals that SDMs were heavily enriched in nucleotide metabolism (purines and pyrimidines) rather than central carbohydrate pathways (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ed).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eSpecifically, key nucleotides\u0026mdash;such as adenine, adenosine, thymine, thymidine, guanosine, and uridine\u0026mdash;were strongly depleted in disturbed soil (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.0001, fold change\u0026thinsp;\u0026lt;\u0026thinsp;0.5, \u003cb\u003eSupplementary Table\u0026nbsp;4\u003c/b\u003e). Concurrently, we observed a severe suppression of L-glutamine, a critical metabolic hub bridging carbon fixation, nitrogen assimilation, and organic synthesis\u003csup\u003e[\u003cspan citationid=\"CR18\" class=\"CitationRef\"\u003e18\u003c/span\u003e]\u003c/sup\u003e. This suggests a disruption in the fundamental carbon-nitrogen flux that fuels microbial growth\u003csup\u003e[\u003cspan citationid=\"CR19\" class=\"CitationRef\"\u003e19\u003c/span\u003e, \u003cspan citationid=\"CR20\" class=\"CitationRef\"\u003e20\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e \u003cp\u003eTo identify the primary metabolic signature of soil disturbance, we distilled the 47 SDMs down to 10 core markers (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ef; \u003cb\u003eSupplementary Fig.\u0026nbsp;1a\u003c/b\u003e). These 10 SDMs belong entirely to purine and pyrimidine nucleotide families and exhibited high sensitivity to infrastructural disturbance (Cohen\u0026rsquo;s d ranging from 0.87 to 4.60, \u003cb\u003eSupplementary Table\u0026nbsp;5\u003c/b\u003e). Furthermore, pattern analysis via K-means clustering resolved the abundance trends of SDMs into 7 distinct patterns (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ee), revealing that these nucleotide features followed a synchronized decline. This implies they act as a cohesively regulated functional module responding instantaneously to environmental disturbance. The specific suppression of these nucleotides\u0026mdash;as opposed to generalized stress markers like proline\u003csup\u003e[\u003cspan citationid=\"CR21\" class=\"CitationRef\"\u003e21\u003c/span\u003e]\u003c/sup\u003e\u0026mdash;indicates a disruption of DNA replication and high-energy substrate turnover\u003csup\u003e[\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]\u003c/sup\u003e, signifying an immediate arrest of microbial proliferation upon disturbance. Finally, ROC analysis confirmed the diagnostic robustness of this module, yielding high AUC values ranging from 0.91 to 0.97 (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eg).\u003c/p\u003e \u003cp\u003eFurthermore, we reconstructed the metabolic network using KEGG pathway mapping to elucidate the mechanistic link between nucleotide depletion and the failure of carbon sequestration. L-glutamine serves as the mechanistic bridge connecting these distinct pathways: it serves as the precursor for \u003cem\u003ede novo\u003c/em\u003e biosynthesis of both purines and pyrimidines (ko 00230 and ko 00240), while simultaneously acting as the core node feeding into central carbon and energy metabolism (via the TCA cycle), methane metabolism, and pyruvate metabolism (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eh). This reveals that the early-stage suppression of soil carbon cycling is coupled to nucleotide depletion via the collapse of the L-glutamine hub.\u003c/p\u003e \u003cp\u003eFinally, to confirm the relationship between the metabolome and stable carbon pools, Mantel tests demonstrated a significant correlation between the nucleotide matrix and the stable carbon indices (MNC, lignin, and macro AAOC) (\u003cem\u003eR\u003c/em\u003e\u0026thinsp;\u0026gt;\u0026thinsp;0.6, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001; \u003cb\u003eSupplementary Fig.\u0026nbsp;1b\u003c/b\u003e).\u003c/p\u003e \u003cp\u003e \u003cb\u003e2.3 Loss of the Carbon-Metabolizing Genome Drives Nucleotide Depletion and Disrupts Central Carbon Fixation\u003c/b\u003e \u003c/p\u003e \u003cp\u003eTo elucidate the specific genetic machinery driving the observed nucleotide depletion and carbon sink impairment, we conducted metagenomic sequencing on the matched soil samples (n\u0026thinsp;=\u0026thinsp;204, Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ea). Functional profiling based on KEGG revealed clear shifts in central carbon and energy metabolism, notably within the Glycolysis/Gluconeogenesis pathways, Purine/Pyrimidine Metabolism, and Carbon Metabolism (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eb). To pinpoint the exact drivers of this metabolic shift, we mapped the metagenomic reads against the curated carbon-cycle database (CCycDB). From an initial alignment of 3,744 C-genes, we isolated 48 altered candidates (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05, \u003cb\u003eSupplementary Fig.\u0026nbsp;2a-c\u003c/b\u003e). We then intersected these candidates with genes governing the upstream biosynthesis and downstream degradation of our feature metabolites (thymidine, guanosine, adenosine, and L-glutamine). This locked down 5 core regulatory C-genes\u0026mdash;\u003cem\u003epdp\u003c/em\u003e, \u003cem\u003edeoA\u003c/em\u003e, \u003cem\u003egsk\u003c/em\u003e, \u003cem\u003epyrD\u003c/em\u003e and \u003cem\u003eyrfG\u003c/em\u003e\u0026mdash;that constitute the enzymatic backbone controlling nucleotide turnover (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ec).\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eMetabolic pathway analysis revealed that the suppression of nucleotides is achieved through a synchronized genomic downregulation. First, \u003cem\u003epyrD\u003c/em\u003e (dihydroorotate dehydrogenase), which governs the conversion of glutamine-derived precursors into the pyrimidine metabolism, Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ec, was significantly downregulated (FC\u0026thinsp;=\u0026thinsp;0.14). This suppression hinders the flux of carbon and nitrogen from L-glutamine, cutting off the primary supply for UMP and Thymidine, Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ec. Simultaneously, the microbial capacity to recycle existing nucleotides (salvage pathways)\u003csup\u003e[\u003cspan citationid=\"CR22\" class=\"CitationRef\"\u003e22\u003c/span\u003e]\u003c/sup\u003e was also impaired. Specifically, we observed a severe depletion of \u003cem\u003edeoA\u003c/em\u003e (thymidine phosphorylase, FC\u0026thinsp;=\u0026thinsp;0.03) alongside an increase in \u003cem\u003epdp\u003c/em\u003e (pyrimidine-nucleoside phosphorylase, FC\u0026thinsp;=\u0026thinsp;2.02), restricting the phosphorolysis of thymidine and preventing the recycling of deoxyribose-derived carbon back into glycolysis. Similarly, the imbalance between \u003cem\u003eyrfG\u003c/em\u003e (downregulated, FC\u0026thinsp;=\u0026thinsp;0.51)) and \u003cem\u003egsk\u003c/em\u003e (upregulated, FC\u0026thinsp;=\u0026thinsp;3.54) promoted the net consumption of guanosine. Furthermore, we substantiated the link between these nucleotides-regulatory genes and soil carbon metabolism by identifying concurrent alterations in specific carbon fixation and degradation genes, notably \u003cem\u003ekorA\u003c/em\u003e (FC\u0026thinsp;=\u0026thinsp;0.32), \u003cem\u003eoorA\u003c/em\u003e (FC\u0026thinsp;=\u0026thinsp;3.58), and \u003cem\u003ecbbM\u003c/em\u003e (FC\u0026thinsp;=\u0026thinsp;2.19). The downregulation of \u003cem\u003ekorA\u003c/em\u003e is significant, as this enzyme controls the entry of glutamine-derived carbon (via 2-oxoglutarate) into the TCA cycle. The simultaneous suppression of \u003cem\u003epyrD\u003c/em\u003e (nucleotide synthesis) and \u003cem\u003ekorA\u003c/em\u003e (TCA entry) functionally severs the \"glutamine bridge,\" depriving the microbiome of both the genetic materials for replication and the carbon and energy substrates required for carbon stabilization, Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ec.\u003c/p\u003e \u003cp\u003eTo verify the spatial robustness of these core genetic signatures and ensure they are not artifacts of local sampling bias, we stratified the 204 samples into 5 distinct groups based on geographical location and altitude (\u003cb\u003eSupplementary Table\u0026nbsp;1,2\u003c/b\u003e) for validation. First, to assess local robustness, we performed 10-fold cross-validation and Leave-One-Out Cross-Validation (LOOCV) within each individual group (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ed-h). Second, to test spatial transferability, we conducted 5-fold cross-group validation and Leave-One-Cohort-Out (LOCO) validation (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ei). This testing yielded AUROC values ranging from 0.86 to 0.99. These results substantiate that the genomic alteration in this study is a highly conserved biological response to physical disturbance across the alpine transect, rather than a localized anomaly.\u003c/p\u003e \u003cp\u003eConsistent with these functional and genetic disruptions, PCoA and ANOSIM of the microbiome revealed a clear separation between natural and disturbed soils (*\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05, FDR\u0026thinsp;\u0026lt;\u0026thinsp;0.05), whereas alpha diversity (Chao1 and Shannon) showed no significant differences (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ej,k). Linear Discriminant Analysis Effect Size (LEfSe, LDA\u0026thinsp;\u0026gt;\u0026thinsp;3.0) revealed a shift from \u0026ldquo;carbon-building\u0026rdquo; to \u0026ldquo;carbon-consuming\u0026rdquo; microbial groups (\u003cb\u003eSupplementary Fig.\u0026nbsp;2d\u003c/b\u003e). Specifically, we observed a consistent depletion of taxa associated with AAOC synthesis, including \u003cem\u003eAcidobacteriota\u003c/em\u003e, \u003cem\u003eAlphaproteobacteria\u003c/em\u003e, \u003cem\u003eVerrucomicrobiota\u003c/em\u003e, and \u003cem\u003eDeltaproteobacteria\u003c/em\u003e\u003csup\u003e[\u003cspan additionalcitationids=\"CR24\" citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]\u003c/sup\u003e. Conversely, taxa harboring oligotrophic lifestyles specialize in degrading recalcitrant soil organic matter (e.g., \u003cem\u003eActinomycetota\u003c/em\u003e and \u003cem\u003eChloroflexota\u003c/em\u003e\u003csup\u003e[\u003cspan additionalcitationids=\"CR27\" citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e]\u003c/sup\u003e) were significantly enriched. These taxonomic shifts suggest that the microbial community, deprived of fresh inputs, pivoted toward consuming the stable carbon pool for survival.\u003c/p\u003e \u003cp\u003e \u003cb\u003e2.4 Robustness Verification via Stratified Random Subsampling\u003c/b\u003e \u003c/p\u003e \u003cp\u003eTo rigorously validate the spatial robustness of these nucleotide markers\u0026mdash;and to ensure they are not artifacts of local sampling bias, we implemented a 100-fold stratified random subsampling strategy. In each iteration, a subset of 100 samples was randomly drawn from the aggregate dataset (n\u0026thinsp;=\u0026thinsp;204). This was constrained to maintain a balanced distribution between natural and disturbed groups (47\u0026ndash;53 samples per group), effectively testing the markers against varying subsets of environmental heterogeneity.\u003c/p\u003e \u003cp\u003eFor each randomized subset, the full analytical workflow\u0026mdash;including PCA, OPLS-DA modeling, differential analysis, and ROC evaluation\u0026mdash;was executed independently. This iterative testing revealed stability. Specifically, the feature selection process showed convergence: Thymidine, Guanosine, and Adenosine were repeatedly identified as significantly differential metabolites (VIP\u0026thinsp;\u0026gt;\u0026thinsp;1, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05, FC\u0026thinsp;\u0026lt;\u0026thinsp;0.4) in 100%, 99%, and 98% of the trials, respectively (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ed). Furthermore, ROC analysis across the 100 independent subsets yielded a stable mean ROC of 0.95, 0.97, and 0.92 for thymidine, guanosine, and adenosine, respectively. As visualized in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea-c, the narrow 95% confidence intervals (CI) surrounding the mean ROC curve confirm that the diagnostic performance of these nucleotide markers is not driven by specific regional outliers. Instead, their depletion represents a consistent biological response to physical disturbance across the alpine transect.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003e2.5 Multi-Omics Pathway Connects Nucleotide Depletion to Central Carbon Metabolism\u003c/b\u003e \u003c/p\u003e \u003cp\u003eTo highlight how nucleotide depletion connects to carbon fixation and the overall soil carbon stability, we integrated the identified taxonomic, genomic, and metabolic signatures into a unified biological correlation map. The resulting network reveals three connected components (Fig.\u0026nbsp;\u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003e), with L-Glutamine acting as the central metabolic hub, bridging purine/pyrimidine metabolism with central carbon and energy pathways.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003eSpecifically, the nucleotide markers exhibited strong positive correlations with the genes governing their biosynthesis (\u003cem\u003epyrD\u003c/em\u003e, \u003cem\u003edeoA\u003c/em\u003e, \u003cem\u003eyrfG\u003c/em\u003e) and with \u003cem\u003ekorA\u003c/em\u003e, the critical enzymatic bridge linking glutamine to the TCA cycle. These positive links extended to key carbon-fixing taxa (e.g., \u003cem\u003eAcidobacteriota\u003c/em\u003e). In contrast, we observed significant negative correlations between these nucleotides and the genes driving their degradation (\u003cem\u003egsk\u003c/em\u003e, \u003cem\u003epdp\u003c/em\u003e). This mirrored the enrichment of mineralization-associated taxa (e.g., \u003cem\u003eActinomycetota\u003c/em\u003e). This pathway map provides confirmation of our central hypothesis: the impairment of soil carbon sequestration is a direct downstream consequence of the genetic blockade in nucleotide turnover.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003e \u003cb\u003eNucleotide Depletion Suppresses Soil Stable Carbon Accumulation by Constraining the Ex Vivo Modification and In Vivo Turnover of Microbial Carbon Pump\u003c/b\u003e \u003c/p\u003e \u003cp\u003eDetecting early soil carbon loss following anthropogenic disturbance is challenging due to the relative stability of macro-carbon pools (e.g., AAOC and MNC)\u003csup\u003e[\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e]\u003c/sup\u003e. Our integrated analysis of diverse soil carbon fractions and metabolomic profiles corroborates this, revealing that the impairment of microbial nucleotide turnover significantly precedes detectable loss in stable carbon pools. Nucleotide metabolites reside in the active intracellular pool and turn over rapidly, making them sensitive to environmental changes. However, why their depletion predicts soil carbon dynamics has remained unclear. By integrating targeted metabolomics and metagenomics across a 621-km alpine transect, we uncovered how physical disturbance impairs the process of soil microbial carbon sequestration by reducing the synthesis of key nucleotides.\u003c/p\u003e \u003cp\u003ePhysical disturbance reduces the microbial nucleotide supply primarily through the following two processes. First, the removal of vegetation directly severs the continuous supply of plant root exudates, depriving the rhizosphere microbiome of bioavailable precursors (e.g., L-glutamine) and exogenous nucleotides \u003csup\u003e[\u003cspan citationid=\"CR32\" class=\"CitationRef\"\u003e32\u003c/span\u003e]\u003c/sup\u003e. Second, the mechanical disruption of soil aggregates destroys microbial microhabitats, leading to a decrease in copiotrophic, \"carbon-building\" taxa (such as \u003cem\u003eAcidobacteriota\u003c/em\u003e and \u003cem\u003eAlphaproteobacteria\u003c/em\u003e responsible for synthesizing soil aggregate carbon and \u003cem\u003eBetaproteobacteria\u003c/em\u003e which is associated with labile carbon turnover)\u003csup\u003e[\u003cspan additionalcitationids=\"CR24\" citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e]\u003c/sup\u003e. Because these specific taxa are the primary carriers of the nucleotide-processing machinery, their decline manifests at the metagenomic level as a synchronized downregulation of both the \u003cem\u003ede novo\u003c/em\u003e biosynthesis (\u003cem\u003epyrD\u003c/em\u003e) and salvage (\u003cem\u003edeoA\u003c/em\u003e) pathways of nucleotides (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ec, \u003cspan refid=\"Fig5\" class=\"InternalRef\"\u003e5\u003c/span\u003ea). This suggests that the observed genetic shift reflects a change in community composition rather than a transient stress response. Furthermore, the concurrent downregulation of \u003cem\u003ekorA\u003c/em\u003e blocks the entry of remaining L-glutamine into the TCA cycle, effectively severing the metabolic bridge between nucleotide synthesis and central energy metabolism.\u003c/p\u003e \u003cp\u003eConsequently, the intracellular concentrations of key nucleotides, such as thymidine and L-guanosine, decrease significantly (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ec,). Because nucleotides, including ATP, serve as the primary energy currency\u003csup\u003e[\u003cspan citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e]\u003c/sup\u003e and essential coenzymes, their depletion reduces the available energy and metabolic resources required for the two sequential stages of the microbial carbon pump (MCP)\u003csup\u003e[\u003cspan citationid=\"CR33\" class=\"CitationRef\"\u003e33\u003c/span\u003e]\u003c/sup\u003e: ex vivo modification and in vivo turnover of microorganisms.\u003c/p\u003e \u003cp\u003eFirst, this energy limitation affects the ex vivo modification pathway, which is the initial step for microbial acquisition of environmental carbon\u003csup\u003e[\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e]\u003c/sup\u003e. This process relies on the synthesis and secretion of extracellular enzymes to depolymerize complex organic matter. However, the transcription, translation, and export of these enzymes are highly energy-dependent and require nucleotide-derived cofactors\u003csup\u003e[\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e, \u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e]\u003c/sup\u003e. The reduction of the nucleotide pool restricts the ATP available to drive this machinery, decreasing the microbial capacity to process and acquire external carbon sources.\u003c/p\u003e \u003cp\u003eSecond, the reduced nucleotide pool compromises the in vivo turnover pathway. Even when simple carbon substrates are assimilated, the shortage of nucleotides\u0026mdash;the fundamental building blocks for DNA and RNA\u003csup\u003e[\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e]\u003c/sup\u003e\u0026mdash;limits microbial anabolism. This limitation directly suppresses microbial carbon use efficiency (CUE), a physiological trait that governs the partitioning of assimilated carbon\u003csup\u003e[\u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e38\u003c/span\u003e]\u003c/sup\u003e. Under energy constraints, microbes allocate a larger proportion of available carbon to respiratory maintenance rather than to biosynthesis\u003csup\u003e[\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e]\u003c/sup\u003e. This shift lowers CUE and limits cellular proliferation. Since the continuous generation of microbial biomass carbon (MBC) is the required precursor for the formation of stable microbial necromass carbon (MNC) \u003csup\u003e[\u003cspan citationid=\"CR40\" class=\"CitationRef\"\u003e40\u003c/span\u003e]\u003c/sup\u003e, the reduction in cell growth directly leads to a decrease in MNC accumulation, halting the transformation of labile organic matter into the long-term stable carbon pool.\u003c/p\u003e \u003cp\u003eUltimately, the combined constraints on ex vivo modification and in vivo turnover slow down the overall MCP process. Among the nucleotide markers, thymidine and guanosine showed the strongest and most consistent depletion (AUROC\u0026thinsp;\u0026gt;\u0026thinsp;0.90). Thymidine is primarily derived from DNA turnover and its salvage pathway depends on \u003cem\u003edeoA\u003c/em\u003e, which was severely downregulated. Guanosine, linked to purine metabolism, was similarly affected by the upregulation of \u003cem\u003egsk\u003c/em\u003e. The sensitivity of these specific nucleotides may reflect their roles in DNA replication and energy transfer. Unlike stable carbon content, intracellular nucleotides like thymidine and guanosine are not physically protected and turn over rapidly\u003csup\u003e[\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e41\u003c/span\u003e]\u003c/sup\u003e. Therefore, their depletion reflects immediate soil stable carbon loss upon disturbance\u003csup\u003e[\u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e42\u003c/span\u003e, \u003cspan citationid=\"CR43\" class=\"CitationRef\"\u003e43\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eA Consistent Response to Environmental Heterogeneity\u003c/h2\u003e \u003cp\u003eA major challenge in soil omics is site-specificity, where biomarkers identified in one region fail to generalize due to environmental heterogeneity\u003csup\u003e[\u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e44\u003c/span\u003e, \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e45\u003c/span\u003e]\u003c/sup\u003e. To rule out the influence of local environmental noise, we spanned a 621 km transect with great altitudinal (700\u0026ndash;4300 m) and spatial gradients. The conserved performance of Thymidine in our 100-fold stratified random subsampling confirms that its rapid depletion is not an artifact of localized environmental noise.\u003c/p\u003e \u003cp\u003eWhile the roles of central carbon genes in ecosystem functioning are well-established, the nucleotide-processing machinery identified here offers a novel diagnostic method. To determine if the suppression of this machinery represents a conserved response beyond physical excavation, we cross-referenced our findings with independent metagenomic datasets. Notably, we confirmed that the suppression of these nucleotide-processing genes represents a consistent response to diverse environmental stress (e.g., land-use change and chemical contamination). For example, the specific suppression of \u003cem\u003edeoA\u003c/em\u003e was consistently observed in soil under pesticide and acid stress\u003csup\u003e[\u003cspan citationid=\"CR46\" class=\"CitationRef\"\u003e46\u003c/span\u003e]\u003c/sup\u003e. Similarly, the dysregulation of \u003cem\u003egsk\u003c/em\u003e and \u003cem\u003epdp\u003c/em\u003e mirror their well-established role in plant biology, where they serve as key modulators of growth metabolism and adaptation to diverse abiotic stresses (e.g., salt, hormone signaling, drought, and light stress)\u003csup\u003e[\u003cspan additionalcitationids=\"CR48 CR49\" citationid=\"CR47\" class=\"CitationRef\"\u003e47\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e]\u003c/sup\u003e. This external validation confirms that nucleotide depletion is a physiological response of the soil microbiome to environmental stress, supporting its use as an early molecular indicator of soil carbon loss.\u003c/p\u003e \u003cp\u003ePractically, our findings offer targets for the rapid assessment of soil carbon sinks during ecological restoration. Future interventions should therefore transition from conventional physical stabilization to targeted metabolic or microbial resuscitation. Whether the soil carbon sequestration can be reactivated through targeted chemical amendments\u0026mdash;such as exogenous nucleotide or L-glutamine addition\u0026mdash;needs to be further investigated.\u003c/p\u003e \u003c/div\u003e"},{"header":"Materials and Methods","content":"\u003cdiv id=\"Sec5\" class=\"Section2\"\u003e \u003ch2\u003eStudy Design and Sample Collection\u003c/h2\u003e \u003cp\u003eTo investigate soil carbon dynamics under anthropogenic disturbance, a 621-km transect was established spanning the transition from the Baiyu in Ganzi Tibetan Autonomous Prefecture to the Tianquan in Sichuan (elevations 700\u0026ndash;4300 m), Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea. The study employed a paired-group design comprising 204 soil samples (102 disturbed vs. 102 natural pairs). To account for spatial heterogeneity, the sampling transect was stratified into five distinct altitudinal cohorts: Baiyu (4000\u0026ndash;4300 m), Litang (3600\u0026ndash;4000 m), Yajiang (3200\u0026ndash;3600 m), Kangding (2500\u0026ndash;2800 m), and Tianquan (700\u0026ndash;1000 m), \u003cb\u003eSupplementary Table\u0026nbsp;1,2\u003c/b\u003e. Disturbed samples were collected from tunnel construction sites exposed for less than three months, while matched natural controls were obtained from adjacent undisturbed forest ecosystems within a 1 km radius to minimize climatic and geological confounding. At each sampling site, 3\u0026thinsp;~\u0026thinsp;6 circular plots (radius 0.2 m) were randomly established (\u003cb\u003eSupplementary Fig.\u0026nbsp;3\u003c/b\u003e). Surface soil (5\u0026ndash;20 cm) was collected from three points within each plot and pooled to form a composite sample. Samples were divided into two aliquots: one portion was air-dried for geochemical characterization (subset n\u0026thinsp;=\u0026thinsp;78), and the remainder was sieved (2 mm) and flash-frozen at \u0026minus;\u0026thinsp;80\u0026deg;C for multi-omics analysis (total dataset n\u0026thinsp;=\u0026thinsp;204).\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eQuantification of soil stable carbon pools\u003c/h3\u003e\n\u003cp\u003eTo quantify the 'temporal lag' of carbon fractions, we analyzed five key carbon pools in a representative subset (n\u0026thinsp;=\u0026thinsp;78, \u003cb\u003eSupplementary Table\u0026nbsp;3\u003c/b\u003e). This subset covered 13 sampling sites stratified into five regional cohorts based on altitude: Sites 1\u0026ndash;2 (Baiyu), Sites 3\u0026ndash;5 (Litang), Sites 6\u0026ndash;9 (Yajiang), Sites 10\u0026ndash;11 (Kangding), and Sites 12\u0026ndash;13 (Tianquan).\u003c/p\u003e \u003cp\u003eSoil aggregates were fractionated into macro- (\u0026gt;\u0026thinsp;0.25 mm), medium- (0.053\u0026ndash;0.25 mm), and micro-aggregates (\u0026lt;\u0026thinsp;0.053 mm) via dry-sieving. Total SOC and AAOC was determined using the potassium dichromate-sulfuric acid oxidation method with external heating (digestion at 135\u0026deg;C for 30 min), and measured spectrophotometrically at 585 nm against a glucose standard. The organic carbon content within AAOC was quantified by titrimetric dichromate oxidation, followed by titration with 0.2 M FeSO4 using o-phenanthroline. EOC was assessed via the 333 mM KMnO4 oxidation method\u003csup\u003e[\u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e50\u003c/span\u003e]\u003c/sup\u003e. Briefly, samples were reacted with KMnO4 (25\u0026deg;C, 2 h, 120 rpm), and the consumed permanganate was measured spectrophotometrically at 565 nm (Shimadzu UV-1800, Kyoto, Japan).\u003c/p\u003e \u003cp\u003eLignin phenols and amino sugars were quantified using an Agilent 6890 GC-MS system (Agilent Technologies, CA, USA). Lignin phenols were released via alkaline CuO oxidation, extracted, purified, and derivatized to volatile trimethylsilyl (TMS) ethers. Quantification was performed using external calibration, and total lignin was calculated as the sum of five monomers: vanillic acid, acetosyringone, syringic acid, p-hydroxycinnamic acid, and ferulic acid. For microbial necromass carbon (MNC), soil samples (\u0026gt;\u0026thinsp;0.3 mg N) were hydrolyzed in 6 M HCl at 105\u0026deg;C for 8 h with myo-inositol as an internal standard. Following filtration, neutralization (pH 6.6\u0026ndash;6.8), and desalting, amino sugars were derivatized using a mixture of 32 M hydrochloric carboxymethoxylamine and 40 M 4-dimethylaminopyridine (4:1, v/v; 80\u0026deg;C, 30 min), followed by acetylation with acetic anhydride. The final derivatives were dissolved in ethyl acetate-hexane (1:1) and injected (1.0 \u0026micro;L, split 10:1) using high-purity N\u003csub\u003e2\u003c/sub\u003e carrier gas (0.8 mL min\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e). The GC oven temperature was programmed as follows: 120\u0026deg;C (4 min hold), ramped at 10\u0026deg;C min\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e to 230\u0026deg;C, 5\u0026deg;C min\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e to 250\u0026deg;C (4 min hold), and finally 40\u0026deg;C min\u003csup\u003e\u0026minus;\u0026thinsp;1\u003c/sup\u003e to 300\u0026deg;C (5 min hold). Bacterial and fungal necromass carbon were calculated from muramic acid and glucosamine concentrations, respectively, using established conversion coefficients.\u003c/p\u003e\n\u003ch3\u003eTargeted metabolomics and biomarker identification\u003c/h3\u003e\n\u003cdiv id=\"Sec8\" class=\"Section2\"\u003e \u003ch2\u003eMetabolite Extraction and LC-MS/MS Analysis\u003c/h2\u003e \u003cp\u003eSoil metabolites were extracted and profiled using a targeted 600-MRM platform (Biotree, Shanghai). Briefly, 25 mg of soil was extracted using an acetonitrile-methanol-water system (2:2:1, v/v) spiked with isotopically labeled internal standards. Following homogenization with zirconia beads, low-temperature sonication, and centrifugation, the supernatant was analyzed. Chromatographic separation was performed on an Agilent 1290 UHPLC system equipped with a Waters Atlantis Premier BEH Z-HILIC column (1.7 \u0026micro;m, 2.1 \u0026times; 150 mm). The mobile phase consisted of 10 mmol/L ammonium formate in water/acetonitrile (9:1, v/v) (Phase A) and 1:9 (v/v) (Phase B). Mass spectrometry was conducted on an AB Sciex QTrap 6500\u0026thinsp;+\u0026thinsp;system with the following source parameters: IonSpray Voltage +\u0026thinsp;5500V/-4500V, Curtain Gas 35 psi, Temperature 400\u0026deg;C, and Ion Source Gas 1/2 at 50 psi. To monitor instrumental stability, quality control (QC) samples (pooled standard mixture) were injected every 10 samples (\u003cb\u003eSupplementary Fig.\u0026nbsp;4\u003c/b\u003e; \u003cb\u003eSupplementary Table\u0026nbsp;6\u003c/b\u003e); the relative standard deviation (RSD) of internal standard retention times was maintained within \u0026plusmn;\u0026thinsp;10 seconds (RSD\u0026thinsp;\u0026le;\u0026thinsp;20%), confirming high-quality data acquisition\u003csup\u003e[\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e51\u003c/span\u003e]\u003c/sup\u003e.\u003c/p\u003e \u003c/div\u003e\n\u003ch3\u003eStatistical Analysis and Biomarker Screening\u003c/h3\u003e\n\u003cp\u003eRaw data were processed using SCIEX Analyst Workstation (v1.7.3) and BIOTREE BioBud (v2.0.3) for peak integration and quantification. Preprocessing included filtration, missing value recoding, and internal standard normalization. Multivariate pattern recognition, including Principal Component Analysis (PCA) and Orthogonal Partial Least Squares-Discriminant Analysis (OPLS-DA), was performed using SIMCA (v18.0.1). Significantly Differential Metabolites (SDMs) were identified based on a strict combinatorial threshold\u003csup\u003e[\u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e52\u003c/span\u003e]\u003c/sup\u003e: (1) Variable Importance in Projection (VIP) score\u0026thinsp;\u0026gt;\u0026thinsp;1; (2) Fold Change (FC)\u0026thinsp;\u0026gt;\u0026thinsp;2 or \u0026lt;\u0026thinsp;0.5; and (3) False Discovery Rate (FDR)-adjusted P-value\u0026thinsp;\u0026lt;\u0026thinsp;0.05 (Wilcoxon rank-sum test).\u003c/p\u003e\n\u003ch3\u003eFeature Selection and Diagnostic Validation\u003c/h3\u003e\n\u003cp\u003eTo isolate the most robust signatures, SDMs were standardized via Z-score normalization and grouped into co-regulated functional modules using K-means clustering (R package cluster). The top 10 metabolites with the highest Z-score variance were selected as candidate biomarkers. For these candidates, biological effect sizes were quantified using Cohen\u0026rsquo;s d (difference in means divided by pooled SD). Subsequently, the diagnostic performance of this 10-metabolite signature was evaluated using Receiver Operating Characteristic (ROC) analysis (R package pROC or plotROC). The Area Under the Curve (AUROC) was calculated to quantify sensitivity and specificity, identifying markers with AUROC\u0026thinsp;\u0026gt;\u0026thinsp;0.90 as robust indicators of soil disturbance. Finally, SDMs were mapped onto the KEGG database to elucidate perturbed metabolic networks, with a specific focus on carbon metabolism pathways.\u003c/p\u003e \u003cdiv id=\"Sec11\" class=\"Section2\"\u003e \u003ch2\u003eRobustness verification via stratified random subsampling\u003c/h2\u003e \u003cp\u003eTo distinguish universal biological signatures from artifacts of local environmental heterogeneity (e.g., elevation or vegetation differences), a rigorous 100-fold stratified random subsampling strategy was implemented. In each of the 100 iterations, a subset of 100 samples was randomly drawn from the total dataset (n\u0026thinsp;=\u0026thinsp;204), balanced between natural and disturbed groups. For each subset, differential abundance analysis and ROC evaluation were independently executed. A biomarker was considered robust only if it maintained statistical significance (VIP\u0026thinsp;\u0026gt;\u0026thinsp;1, \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05) in more than 95% of the iterations. This approach ensured that the identified markers represented a physiological response rather than site-specific noise.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec12\" class=\"Section2\"\u003e \u003ch2\u003eValidation of Metabolic Biomarkers via Mantel test\u003c/h2\u003e \u003cp\u003eTo validate the quantitative predictive power of the identified biomarkers for soil carbon sequestration, the matrix of signature metabolites was correlated with soil stable carbon indices, including SOC, lignin phenols, MNC, AAOC (macro-, medium-, micro-), and EOC fractions. Global associations between the metabolic matrix and carbon factors were assessed via Mantel tests (999 permutations, R package vegan), which identified five metabolites significantly associated with the majority of carbon indicators.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec13\" class=\"Section2\"\u003e \u003ch2\u003eMetagenomic sequencing and functional profiling\u003c/h2\u003e \u003cdiv id=\"Sec14\" class=\"Section3\"\u003e \u003ch2\u003eDNA Extraction, Sequencing, and Assembly\u003c/h2\u003e \u003cp\u003eTotal genomic DNA was extracted from 204 soil samples using the Mo Bio PowerSoil DNA Isolation Kit (Mo Bio Laboratories, CA, USA). Paired-end sequencing (2 \u0026times; 150 bp) was performed on an Illumina NovaSeq 6000 platform. Raw reads were processed using Fastp for quality control (low-quality trimming and adapter removal). High-quality reads were assembled into contigs using MEGAHIT with a multi-k-mer strategy. Contigs\u0026thinsp;\u0026lt;\u0026thinsp;500 bp were discarded. Open Reading Frames (ORFs) were predicted using MetaGeneMark, and a non-redundant gene catalogue was constructed using MMseqs2 (95% identity, 90% coverage). Gene abundance was calculated based on read mapping (Bowtie2) and normalized to Transcripts Per Million (TPM) or relative abundance.\u003c/p\u003e \u003c/div\u003e \u003c/div\u003e \u003cdiv id=\"Sec15\" class=\"Section2\"\u003e \u003ch2\u003eTaxonomic Profiling and Biomarker Identification\u003c/h2\u003e \u003cp\u003eTaxonomic annotation was performed by aligning non-redundant genes against the NR database using DIAMOND (e-value\u0026thinsp;\u0026lt;\u0026thinsp;1 \u0026times; 10\u003csup\u003e\u0026minus;\u0026thinsp;5\u003c/sup\u003e). Species-level abundance tables were rarefied to the minimum sequencing depth to standardize sampling effort. Alpha diversity (Chao1, Shannon) and Beta diversity (PCoA based on Bray-Curtis distance) were calculated using QIIME. Significant taxonomic differences between natural and disturbed soils were assessed via PERMANOVA and ANOSIM. To identify robust biomarkers, Linear Discriminant Analysis Effect Size (LEfSe) was applied (LDA score\u0026thinsp;\u0026gt;\u0026thinsp;3.0, FDR-corrected \u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.001).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec16\" class=\"Section2\"\u003e \u003ch2\u003eFunctional Annotation and Key C-Gene Screening\u003c/h2\u003e \u003cp\u003eFunctional profiling was conducted using the KEGG database for metabolic pathway reconstruction and CCycDB for specific carbon-cycle gene annotation. From an initial pool of 48 differentially abundant C-genes (\u003cem\u003eP\u003c/em\u003e\u0026thinsp;\u0026lt;\u0026thinsp;0.05), we applied a mechanistic filtering strategy to isolate core drivers of metabolic disruption. Genes were retained only if they: (1) exhibited significant differential abundance between groups; and (2) encoded enzymes acting as direct upstream/downstream regulators of the identified nucleotide biomarkers (e.g., connecting Glutamine to UMP/TMP). This targeted screening narrowed the candidate list to 5 core regulatory genes (\u003cem\u003epyrD\u003c/em\u003e, \u003cem\u003edeoA\u003c/em\u003e, \u003cem\u003egsk\u003c/em\u003e, \u003cem\u003ekorA\u003c/em\u003e/\u003cem\u003eoorA\u003c/em\u003e, \u003cem\u003ecbbM\u003c/em\u003e) (\u003cspan class=\"ExternalRef\"\u003e\u003cspan class=\"RefSource\"\u003ehttps://ccycdb.github.io/\u003c/span\u003e\u003cspan address=\"https://ccycdb.github.io/\" targettype=\"URL\" class=\"RefTarget\"\u003e\u003c/span\u003e\u003c/span\u003e).\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec17\" class=\"Section2\"\u003e \u003ch2\u003eRobustness Validation of Genomic Signatures\u003c/h2\u003e \u003cp\u003eTo verify the generalizability of the identified core genomic signatures, a dual-level validation framework was implemented using Random Forest classifiers (R package randomForest): (1) Local robustness (internal validation): Within each of the five regional cohorts, performance was evaluated using 10-fold cross-validation and Leave-One-Out Cross-Validation (LOOCV). (2) Spatial transferability (external validation): To test stability across environmental gradients, we performed 5-fold cross-group validation and Leave-One-Cohort-Out (LOCO) validation, where the model was trained on four regions and tested on the independent fifth region.\u003c/p\u003e \u003c/div\u003e \u003cdiv id=\"Sec18\" class=\"Section2\"\u003e \u003ch2\u003eConstruction of Multi-Omics Pathway Map\u003c/h2\u003e \u003cp\u003eTo elucidate the mechanistic links between the microbiome and the metabolome, multi-omics pathway map was constructed by integrating taxonomic source tracking and correlation networks. The taxonomic origins of signature metabolites and the carriers of differential C-genes were identified using the BIOTREE internal database (Biotree, Shanghai, China), which integrates curated associations from public repositories including KEGG, BIOML, CGR, HBC, and GMrepo. Intersection analysis was subsequently performed to isolate consistent species-gene-metabolite trios that were both differentially abundant and linked. Based on these validated interactions, an integrated framework was visualized to encapsulate the flow of carbon from specific microbial taxa through regulatory enzymes (C-genes) to metabolic end-products (signature metabolites).\u003c/p\u003e \u003c/div\u003e"},{"header":"Declarations","content":"\u003cp\u003e \u003ch2\u003eCompeting interests\u003c/h2\u003e \u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e \u003c/p\u003e\u003ch2\u003eAuthor Contributions\u003c/h2\u003e \u003cp\u003eThe study was conceived and designed by M.X., L.Q. and X.T. Sampling, and soil pretreatment were carried out by X.W. Y.C. and Y.L. The sequencing data from metagenomics and metabolomics were analyzed by C.T., K.T., and H.Z. The reliability analysis was assisted by H.T. The original draft was prepared by M.X., and all authors did the revisions. All authors read, discussed, and approved the final version of the manuscript.\u003c/p\u003e\u003ch2\u003eAcknowledgments\u003c/h2\u003e \u003cp\u003eThis work was supported by various funding sources. We are grateful for the funding provided by the National Natural Science Foundation of China (grant no. 42307028 to M.X.), Sichuan Provincial Natural Science Foundation (grant no. 25QNJJ4171 to M.X.), the special fund of \u0026zwnj;National Environmental Protection Key Laboratory of Synergetic Control and Joint Remediation of Water \u0026amp; Soil Pollution (grant no. GHBK-2024-15 to M.X.), the Science and Technology Research Project of the Tianfu Yongxing Laboratory (grant numbers 2023KJGG06 to X.T. and X.P.), and the China Giant Panda Conservation and Research Center Foundation Project (grant no. CCRCGPBY202409 to M.X.). The authors would like to express their gratitude to the China State Railway Group Co., Ltd. for their assistance on soil sample collection at ongoing tunnel construction sites (Sichuan\u0026mdash;Xizang section). We also thank Biotree Biotech Co., Ltd. (Shanghai, China) for the assistant with determination of metabolome and metagenome\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\u003cli\u003e\u003cspan\u003eBlakemore, R. J. Biomass Refined: 99% of Organic Carbon in Soils. \u003cem\u003eBiomass\u003c/em\u003e 4(4), 1257\u0026ndash;1300 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMa, W. et al. Carbon budgets and environmental controls in alpine ecosystems on the Qinghai-Tibet Plateau. \u003cem\u003eCATENA\u003c/em\u003e 229, 107224(2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi, Q. et al. The link between landscape characteristics and soil losses rates over a range of spatiotemporal scales: hubei province, China. \u003cem\u003eInter. J. Environ. Res. Public Health\u003c/em\u003e 18(21), 11044 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eChen, L. \u003cem\u003eet al.\u003c/em\u003e Effect of organic material addition on active soil organic carbon and microbial diversity: A meta\u0026ndash;analysis. \u003cem\u003eSoil Till. Res.\u003c/em\u003e 241, 106128 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAl\u0026ndash;Shammary, A. A. G., Al\u0026ndash;Shihmani, L. S. S., Fernandez\u0026ndash;Galvez, J. \u0026amp; Caballero\u0026ndash;Calvo, A. Optimizing sustainable agriculture: A comprehensive review of agronomic practices and their impacts on soil attributes. \u003cem\u003eJ Environ. Manage.\u003c/em\u003e 364, 121487 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXie, M.D. et al. Estuarine wetland tidal organic carbon activates microbial carbon pump and increases long\u0026ndash;term soil carbon stability. \u003cem\u003eCATENA\u003c/em\u003e 247, 108559 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang, Y. et al. Unlocking mechanisms for soil organic matter accumulation: carbon use efficiency and microbial necromass as the keys. \u003cem\u003eGlob. Change Biol\u003c/em\u003e. 31, (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhao, Y. et al. A global meta\u0026ndash;analysis of land use change on soil mineral\u0026ndash;associated and particulate organic carbon. \u003cem\u003eGlob. Change Biol\u003c/em\u003e. 31, (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eSlessarev, Eric W et al. Initial soil organic carbon stocks govern changes in soil carbon: Reality or artifact?. \u003cem\u003eGlob.Change Biol\u003c/em\u003e. 29(5), 1239\u0026ndash;1247 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang, T. et al. Temporal thresholds and depth\u0026ndash;specific mechanisms of soil organic carbon stabilization during 65 years of revegetation in the Tengger Desert. \u003cem\u003eJ Environ. Manage.\u003c/em\u003e 385, 125633 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLavallee, J.M., Soong, J.L., Cotrufo, M.F. Conceptualizing soil organic matter into particulate and mineral-associated forms to address global change in the 21st century. \u003cem\u003eGlob. Change Biol\u003c/em\u003e. 26(1), 261\u0026ndash;273 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHan, X. et al. Understanding soil carbon sequestration following the afforestation of former arable land by physical fractionation. \u003cem\u003eCATENA\u003c/em\u003e 150, 317\u0026ndash;327 (2017).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi, L. et al. Assessment of transcriptional reprogramming of lettuce roots in response to chitin soil amendment. \u003cem\u003eFront. Plant Sci.\u003c/em\u003e 14, 1158068 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMazzoleni, S. et al. Metabolomic changes in Arabidopsis thaliana exposed to extracellular self\u0026ndash;and nonself\u0026ndash;DNA: A reversible effect. \u003cem\u003eEnviron. Exp. Bot.\u003c/em\u003e 234, 106149 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHeinke, L. Good neighbours transfer nucleotides. \u003cem\u003eNat. Rev. Mol. Cell Biol.\u003c/em\u003e 26, 582 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHan, C. et al. Joint regulation of the soil organic carbon accumulation by mineral protection and microbial properties following conservation practices. \u003cem\u003eCatena\u003c/em\u003e 245, 108298 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHu, P. et al. Lithological controls on soil aggregates and minerals regulate microbial carbon use efficiency and necromass stability. \u003cem\u003eEnviron. Sci. Technol.\u003c/em\u003e 48, 58, (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGeorgiou, K. \u003cem\u003eet al.\u003c/em\u003e Emergent temperature sensitivity of soil organic carbon driven by mineral associations. \u003cem\u003eNat. Geosci.\u003c/em\u003e 17(3), p205 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBogati, K. \u0026amp; Walczak, M. The impact of drought stress on soil microbial community, enzyme activities and plants. \u003cem\u003eAgronomy Basel\u003c/em\u003e 12, (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBhattacharjya, S. et al. Utilizing soil metabolomics to investigate the untapped metabolic potential of soil microbial communities and their role in driving soil ecosystem processes: A review. \u003cem\u003eAppl. Soil. Ecol.\u003c/em\u003e 195, 105238 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang, J. et al. Enhancing rice salt tolerance: mechanisms of compound functional liquid in alleviating salt stress during the seedling stage. \u003cem\u003ePlant Physiol. Biochem.\u003c/em\u003e 229, 110273 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHeinke, L. Good neighbours transfer nucleotides. \u003cem\u003eNat. Rev. Mol. Cell Biol.\u003c/em\u003e 26, 582 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWirbel, J. et al. Meta\u0026ndash;analysis of fecal metagenomes reveals global microbial signatures that are specific for colorectal cancer. \u003cem\u003eNat. Med.\u003c/em\u003e 25, 679\u0026ndash;689 (2019).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYang, X. et al. Rice Ferredoxin\u0026ndash;Dependent Glutamate Synthase Regulates Nitrogen\u0026ndash;Carbon Metabolomes and Is Genetically Differentiated between japonica and indica Subspecies. \u003cem\u003eMol. Plant\u003c/em\u003e 9, 1520\u0026ndash;1534 (2016).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLi, Y. et al. Soil organic carbon loss decreases biodiversity but stimulates multitrophic interactions that promote belowground metabolism. \u003cem\u003eGlob. Change Biol\u003c/em\u003e. 30, e17101 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGralka, M., Pollak, S. \u0026amp; Cordero, O.X. Genome content predicts the carbon catabolic preferences of heterotrophic bacteria. \u003cem\u003eNat. Microbiol.\u003c/em\u003e 8, 1799 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHe, C. et al. Decoupled fungal and bacterial functional responses to biochar amendment drive rhizosphere priming effect on soil organic carbon mineralization. \u003cem\u003eBiochar\u003c/em\u003e 6, 84(2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGoldfarb, K.C. et al. Differential growth responses of soil bacterial taxa to carbon substrates of varying chemical recalcitrance. \u003cem\u003eFront. Microbiol.\u003c/em\u003e 2, 94 (2011).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMogens, K. et al, Nucleotide metabolism and its control in lactic acid bacteria. \u003cem\u003eFEMS Microbiol. Rev.\u003c/em\u003e 29(3), 555\u0026ndash;590 (2005).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eKeenor, S.G., Lee, R.; Reid, B.J. Physical protection of soil carbon stocks under regenerative agriculture. \u003cem\u003eSoil\u003c/em\u003e 11(2), 957\u0026ndash;973 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAmundson, R. The Pandora's box of soil carbon. \u003cem\u003ePNAS\u003c/em\u003e 119(11), e2201077119 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eDeng L. et al. Autotoxic ginseno-side stress induces changes in root exudates to recruit the beneficial Burkholderia strain B36 as revealed by transcriptommic and metabolomic approaches. \u003cem\u003eJ. Agric. Food Chem\u003c/em\u003e. 71(11), 4536\u0026ndash;49 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eJiao, N. et al. The microbial carbon pump and climate change. \u003cem\u003eNat. Rev. Microbiol.\u003c/em\u003e 22(7), 408\u0026ndash;419 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eMichael, L. \u0026amp; Georgi, K.M. The bioenergetic costs of a gene. \u003cem\u003ePNAS\u003c/em\u003e. 112(51), 15690\u0026ndash;15695 (2015).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, C.Q. \u0026amp; Kuzyakov, Y. Energy use efficiency of soil microorganisms: driven by carbon recycling and reduction. \u003cem\u003eGlob. Change biol.\u003c/em\u003e 29(22), 6170\u0026ndash;6187 (2023).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWitte, C., \u0026amp; Herde, M. Nucleotide metabolism in plants. \u003cem\u003ePlant Physiol\u003c/em\u003e. 182(1), 63\u0026ndash;78 (2020).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eAllison, S.D. Rethinking microbial carbon use efficiency in soil models. \u003cem\u003eNat. Climate Change\u003c/em\u003e 5(1), 56\u0026ndash;60 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu, X.J.A. et al. Soil aggregate-mediated microbial responses to long-term warming. \u003cem\u003eSoil Biol. Biochem\u003c/em\u003e. 152, 108055 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003ePeng, Z. et al. Plant detritus carbon dominates over microbial necromass carbon in topsoil of alpine ecosystems. \u003cem\u003eCommun. Earth Environ\u003c/em\u003e. 6(1), 912 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eGunina, A., \u0026amp; Kuzyakov, Y. From energy to (soil organic) matter. \u003cem\u003eGlob.Change Biol\u003c/em\u003e. 28(7), 2169\u0026ndash;2182 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZheng, Y. et al. Purines enrich rootassociated Pseudomonas and improve wild soybean growth under salt stress. \u003cem\u003eNat Commun.\u003c/em\u003e 15(1), 3520 (2024).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eLiu, H. et al. Nucleotides enriched under heat stress recruit beneficial rhizomicrobes to protect plants from heat and root-rot stresses. \u003cem\u003eMicrobiome\u003c/em\u003e 13, 160 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHansen, P.M. et al. Microbial community regulation of extracellular enzyme production can mediate patterns of particulate and mineral-associated organic matter accumulation in undersaturated soils. \u003cem\u003eSoil Biol. Biochem.\u003c/em\u003e 214, 110056 (2026).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhou, H.B. et al. Soil heterogeneity influences the biotoxicity of photoaged tire wear particles in Eisenia fetida: A comparative assessment. \u003cem\u003eJ. Hazard. Mater.\u003c/em\u003e 496, 12 (2010).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYun, Y.F., Gai, J.Y., Zhao, T.J. Identification of regulated genes conferring resistance to high concentrations of glyphosate in a new strain of enterobacter. \u003cem\u003eFEMS Microbiol. Letter.\u003c/em\u003e 349(2), 135\u0026ndash;143 (2013).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eHuang, S.H. et al. Genome-wide identification and expression analysis of the GSK gene family in Solanum tuberosum L. under abiotic stress and phytohormone treatments and functional characterization of StSK21 involvement in salt stress. \u003cem\u003eGene\u003c/em\u003e 766, 145156 (2021).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eZhang, Peipei, et al. \"Genome-wide identification and expression analysis of the GSK gene family in wheat (Triticum aestivum L.). \u003cem\u003eMolec. Biol. Rep.\u003c/em\u003e 49(4), 2899\u0026ndash;2913 (2022).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eWang, L.L. et al. Genome-wide characterization and phylogenetic analysis of GSK gene family in three species of cotton: Evidence for a role of some GSKs in fiber development and responses to stress. \u003cem\u003eBMC Plant Biol.\u003c/em\u003e 18, 1 (2018).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eXu, X.F. et al. GmGGDR gene confers abiotic stress tolerance and enhances vitamin E accumulation in arabidopsis and soybeans. \u003cem\u003eAgronomy\u003c/em\u003e 15(2), 351 (2025).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eBlair, G.J., Lefroy, R.D.B. \u0026amp; Lise, L. Soil carbon fractions based on their degree of oxidation, and the development of a carbon mangement index for agricultural systems. \u003cem\u003eAustral. J Agricul. Res.\u003c/em\u003e 46(7), 1459\u0026ndash;1466 (1995).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eYuan, M., Breitkopf, S., Yang, X. \u0026amp; Asara, J.M. A positive/negative ion\u0026ndash;switching, targeted mass spectrometry\u0026ndash;based metabolomics platform for bodily fluids, cells, and fresh and fixed tissue. \u003cem\u003eNat. Protoc.\u003c/em\u003e 7, 872\u0026ndash;881 (2012).\u003c/span\u003e\u003c/li\u003e \u003cli\u003e\u003cspan\u003eTrivedi, P. et al. Microbial regulation of the soil carbon cycle: evidence from gene\u0026ndash;enzyme relationships. \u003cem\u003eISME J.\u003c/em\u003e 10, 2593\u0026ndash;2604 (2016).\u003c/span\u003e\u003c/li\u003e\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":true,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"alpine soil, soil stable carbon, multi-omics integration, biomarkers, nucleotide","lastPublishedDoi":"10.21203/rs.3.rs-9373312/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-9373312/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eAnthropogenic disturbances are a primary cause of soil organic carbon (SOC) destabilization in alpine ecosystems. However, effective management is hindered by a \"response delay\" of macro metrics used to characterize stable carbon pools (e.g., microbial necromass carbon and aggregate associated organic carbon), which often fail to capture the immediate impairment of the soil carbon sequestration. To address this, we conducted a multi-omics study on 204 soil samples along a 621-km transect. We observed a \"temporal lag\": while physical disturbance caused rapid depletion of nucleotide metabolites, stable carbon fractions showed limited responsiveness, masking carbon depletion onset. Through metabolomics and 100-fold stratified subsampling, we identified depletion of specific nucleotides\u0026mdash;notably thymidine and guanosine\u0026mdash;as early-warning signatures (mean AUROC\u0026thinsp;\u0026gt;\u0026thinsp;0.90). Metagenomic profiling revealed this depletion is driven by a disturbance-induced taxonomic shift triggering a synchronized suppression: the simultaneous inhibition of genetic capacity for \u003cem\u003ede novo\u003c/em\u003e synthesis (mediated by \u003cem\u003epyrD\u003c/em\u003e) and salvage pathways (mediated by \u003cem\u003edeoA\u003c/em\u003e). Furthermore, the concurrent downregulation of \u003cem\u003ekorA\u003c/em\u003e indicates the disruption of the \"Glutamine Bridge,\" effectively severing the metabolic link between nucleotide turnover and central carbon/energy metabolism. Our findings identify molecular \"early-warning biomarkers\" that precede observable carbon loss, providing a sensitive tool for monitoring incipient soil degradation.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e","manuscriptTitle":"Nucleotide depletion signals early-stage soil stable carbon collapse in anthropogenically disturbed alpine ecosystems","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-04-27 11:26:08","doi":"10.21203/rs.3.rs-9373312/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"9f659382-ad21-4879-9819-0076b5056689","owner":[],"postedDate":"April 27th, 2026","published":true,"recentEditorialEvents":[{"type":"decision","content":"Reject after peer review","date":"2026-05-07T14:58:40+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"This content is not available.","date":"2026-05-06T07:03:10+00:00","index":1,"fulltext":"This content is not available."}],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":66068591,"name":"Earth and environmental sciences/Environmental sciences/Environmental chemistry/Geochemistry"},{"id":66068592,"name":"Earth and environmental sciences/Ecology/Biogeochemistry/Carbon cycle"}],"tags":[],"updatedAt":"2026-05-07T15:01:51+00:00","versionOfRecord":[],"versionCreatedAt":"2026-04-27 11:26:08","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-9373312","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-9373312","identity":"rs-9373312","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}