Mixing regimes shape microbial community composition, nutrient regimes, and plant growth attributes in Jeevamrit: metagenomics and culturomics-based insights

Abstract

Jeevamrit, a microbial inoculant widely used in zero-budget natural farming (ZBNF) that relies on local farm-based resources to enhance overall biological health of soil, is reported for inconsistent crop yield enhancements. This is mainly due to variability in its preparation methods, e.g., mixing intensity, incubation regimes, and quality of ingredients used. Hence, the current study aimed to decipher the effect of mixing intensity (extent of oxygenation) on microbial community composition, nutrient transformation, and plant growth attributes of Jeevamrit, using a combined metagenomics-culturomics approach. Frequent mixing (Constant/Intermediate) enhanced nutrient solubilization (Fe, Zn, Cu, Mn) with higher total N and dissolved organic carbon, while less mixing (Anoxic/No-mix) led to accumulation of soluble Fe and NH₄⁺-N with higher microbial diversity. Mixing-driven differential enrichment of taxa were noted, i.e., constant mixing (CM) dominated by Acinetobacter (~ 40%), Comamonas, Pseudomonas, and Lysinibacillus, linked to oxidative C/N cycling and metal dissolution. Whereas, anoxic (AO) favored Clostridium sensu stricto, Lactobacillales, Enterococcus, and Enterobacterales (> 60%), correlating to fermentative metabolism-driven reductive elemental cycling. Co-occurrence network analysis identified Acinetobacter, Pseudomonas, Comamonas, Trichococcus, and Stenotrophomonas as hubs, indicating keystone functions in structuring metabolic interactions. The metagenome-recovered MAGs belonged to Acinetobacter sp., Clostridium saccharobutylicum, Trichococcus flocculiformis, and Enterococcus gallinarum with potential to participate in multiple nutrient cycling. Cultivable members of Shigella, Rhodococcus, and Bacillus spp. showed high IAA production (135–145 µg mL⁻¹), NH₃ release (~ 0.12 µg mL⁻¹), and K and P solubilization (~ 55.2 µg mL⁻¹). We hypothesize that oxygenation drives the Jeevamrit’s microbial guild assembly, where mixing intensity modulates oxido-reductive metabolism and nutrient mobilization efficiency, indicating the requirement for standardization of formulation aligned to soil-specific conditions.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-36414-4.

Introduction

Soil degradation mediated by accelerated erosion, depletion of the soil organic carbon (SOC) pool, loss in biodiversity, nutrient imbalance caused by excessive input of chemical fertilizers, is leading to major catastrophes in the soil ecosystem, e.g., fertility issues, acidification, drought, and salinization¹. It further leads to a decline in soil health and ecosystem services, thus imposing pressure on agricultural production costs and sustainability. In the context of the Indian agricultural system, nearly 30% of the agricultural land, approximately 146.8 million hectares, is classified as degraded, majorly driven by decades of chemical overuse, i.e., input of N-P-K fertilizers, pesticides, growth regulators, along with intensive tillage, monocropping, imbalanced irrigation regimes, etc.². The skewed NPK ratio in Indian soils, i.e., 7.7:3.1:1 vs the ideal 4:2:1 (10.9:4.9:1 for the recently concluded kharif season), exemplifies the extent of chemical dependence and nutrient imbalance, leading to stagnation, soil sickness, deficiency of secondary (micro)nutrients, alkalinity, and salinity, and hence threatening long-term food security, water quality, and climate resilience³. Low-input and sustainable practices, which boost soil health, are needed for restoring/soil health with optimum yields. Such an approach can improve overall soil nutrient budget, biotic functions (micro, meso, and macro) and diversity, structural/physico-chemical stability, which can ultimately produce “more from less” by increasing soil, water, and nutrient use efficiency¹.

In this context, Zero-Budget Natural Farming (ZBNF) has emerged as a promising soil-conserving approach tailored to small- and marginal landholders in Indian agro-ecological settings. It advocates the elimination of synthetic inputs/agrochemicals and promotes on-farm resource use, such as cow-based microbial inoculants like Jeevamrit, Beejamrit, Ghanjeevamrit, etc., combined with low-till practices, mulching, and diversified cropping⁴. Owing to its nature of relieving farmers from externally high-cost inputs (synthetic fertilizers, pesticides, and nutrient supplements), with high reliance on locally available farm-based biomass resources (on-farm livestock dung-urine used formulations as concoctions), ZBNF has been recognized for its multiple benefits in the mainstream Indian agricultural methods. Indian states like Andhra Pradesh, Maharashtra, and Karnataka, with reports of higher yield and gross/net income in the range of 14.2–50%, and a decrease in cost of cultivation by 23.7% under irrigated and rainfed conditions⁵. However, the systematic investigations providing microbial and molecular nexus towards its formulation-level as well as field-level translation warrant microbiological and chemical-biology studies.

“Jeevamrit”, a fermented microbial inoculant made up of cow dung (10 kg), cow urine (10 L), jaggery (2 kg), pulse flour (2 kg), and pristine soil (in 200 L of water, as per NITI Aayog), ascribed to be a plant and soil immunity booster, is a key input in ZBNF⁵. It is hypothesized that Jeevamrit’s effectiveness is attributed to two primary mechanisms: microbial enhancement and chemical modulation (nutrients and metabolites), along with a certain scope for improvement. Also, repeated booster application of Jeevamrit has been reported to increase total microbial populations in soil, i.e. 528-fold (from 3.05 to 1610 million CFU g⁻¹) dominated by Pseudomonadota (40%) and Actinobacteria (21%) as the core/stable taxa with enhanced enzymatic activities (77% increase in dehydrogenase, β-glucosidase, and urease), thus displaying better nutrient cycling and disease suppression^6,7. Consistent with these observations, several meta-analyses and field experiments have demonstrated that organic amendments similar to Jeevamrit substantially enhance soil microbial diversity and functional capacity. A global synthesis of 219 studies reported significant increases in Shannon diversity, richness, and phylogenetic diversity under organic inputs compared to mineral-only fertilization, accompanied by clear shifts in bacterial and fungal community structure⁸. Likewise, a meta-analysis of 94 studies showed that organic amendments markedly elevated microbial biomass (both bacterial and fungal fractions) and increased activities of key C-, N-, and P-cycling enzymes, including β-glucosidase, urease, and phosphatases⁹. Long-term amendment trials further indicate sustained increases in microbial biomass carbon and nitrogen, higher basal respiration, and enhanced functional diversity linked to extracellular hydrolases responsible for carbon decomposition and nutrient mineralization¹⁰. These findings collectively support the notion that organic formulations like Jeevamrit act primarily as microbiome stimulants, reinforcing microbial growth, metabolic activity, and enzymatic turnover in soil ecosystems. Jeevamrit has been reported for its higher organic carbon content (up to 46%: 0.61% to 0.92% OC), available P (by 270%), micronutrient (Zn by 98%, Fe by 23%), and pH stability, which in turn enhances soil structure, thus better root development⁶. These attributes could be linked to better plant-soil interactions like P solubilization, IAA production, and biological N-fixation, with biocontrol against pathogens, with over 90% mycelial growth inhibition against Alternaria spp.¹¹. In addition to increasing organic carbon and available nutrients, application of such amendments significantly enhances soil physical structure and biological functioning. The added organic matter stimulates microbial decomposition, resulting in expansion of fungal hyphae and microbial polysaccharides that bind soil particles into macro-aggregates, thereby improving soil porosity, aeration, water-holding capacity, aggregate stability, and facilitating better root penetration¹². Such amendments also substantially increase microbial biomass (bacteria and fungi) and stimulate soil enzyme activities (e.g., dehydrogenase, phosphatase, and urease), which are critical for decomposition of organic matter and mineralization of nutrients, thereby sustaining nutrient cycling (C, N, P) and maintaining nutrient availability over time⁹. Cumulatively, these effects manifest in superior nutrient use efficiency, with treated crops showing enhanced uptake of N (60.58 kg ha⁻¹), P (7.25 kg ha⁻¹), and K (37.88 kg ha⁻¹), with improved microbial biomass carbon (204.5 mg kg⁻¹) in the soil, representing sustained carbon sequestration and soil resilience¹³. Consistent application of Jeevamrit has been found to improve SOC levels ranging from 1.2 to 1.5%, and promotes aggregate stability in Indo-Gangetic alluvial soils. Further, when nitrogen-fixing and phosphate-solubilizing microorganisms are present in the amendment, Jeevamrit may contribute to biological nitrogen fixation and enhanced phosphorus and micronutrient mobilization, improving N and P availability to plants beyond what is provided by the amendment directly⁶. The crop-specific benefits of jeevamrit have also been reported across diverse agroecosystems. In rice–wheat cropping systems, its application under water-stressed conditions enhances rhizosphere microbial diversity, resulting in an 11–14% increase in yield⁷. In the horticultural crop, Jeevamrit has been shown to elevate flower yield by up to 20% while enhancing carotenoid content¹⁴. Similarly, a soil of cropping system integrating vermicompost, ghanjeevamrit/dravajeevamrit, crop residue, and biofertilizers under the rice–mustard–green gram exhibited improved physical properties (e.g., reduced bulk density), enhanced biological activity, better nutrient availability and yield, indicating that natural-input based farming can match productivity while improving long-term soil fertility¹⁵.

Such multidimensional effects demonstrate Jeevamrit’s capacity to create self-reinforcing cycles of soil improvement through integrated biological, chemical, and physical mechanisms. On the contrary, though proven to be beneficial, Jeevamrit has several key limitations. It often provides a limited usable nitrogen source, leading to 20–48% yield reductions in high-input systems¹⁶. Furthermore, cow dung base used can harbour harmful enteric pathogens like E. coli, Salmonella, and Cryptosporidium, posing serious health risks. However, the optimization of such formulation preparation and effective treatments could nullify such limitations and improve its efficacy. It has also been shown that a lack of standardized preparation leads to inconsistent quality in terms of microbial, nutrient, and metabolite signatures, and even poor storage could create contamination risk. Despite the increasing adoption of Jeevamrit with wider benefits, significant variability in its preparation methods (mixing effects, incubation duration, temperature effect, light–dark periods), ingredient quality, and application practices contribute to inconsistent results, emphasizing the need for formulation standardization and adoption of optimized protocols. Knowledge gaps also persist in understanding how Jeevamrit performs across varying agro-ecological zones, soil types, cropping systems, and climatic stress conditions such as salinity and drought. In the context of agricultural relevance, the identification of key microbial signatures and their metabolic interplay that could be linked to elemental cycling and plant-growth/ yield outcomes need to be evaluated across defined modes of Jeevamrit preparations. The comprehensive understanding of the effects of preparation method (mixing, as a major driver) on change in community composition/dynamics and their functional role (genes, metabolites) relevant for plant growth promotion and soil nutrient cycling is not yet fully elucidated. Their factor-specific understanding will be critical for developing soil–plant-edaphic factor-specific application strategies to unlock its full potential in sustainable agriculture.

Considering the research gap highlighted, current study was rationalized to answer a) change in Jeevamrit microbiome composition when incubated under different mixing regimes: constant mixing (CM), intermediate mixing (IM), no mixing (NM), and anoxic condition (AO), b) involvement of key microbial taxa (keystone/indicators) and their metabolic interactions which could influence community structure and be associated with key physico-chemical parameters, c) effect of such mixing on the Jeevamrit metabolite pool, and d) plant growth beneficial traits imparted by the representative microbial members of such formulation.

Materials and methods

Preparation of Jeevamrit and assessment of physico-chemical attributes

Jeevamrit was formulated under four distinct mixing conditions: (i) Constant mixing (“CM”: four time mixing by hand in a day, at a periodic interval of 2–3 h), (ii) Intermediate mixing (“IM”: one time mixing in a day, at every 24 h), iii) No mixing (“NM”: without mixing) and iv) Anoxic (“AO”: without mixing and kept under anoxic condition, air-tight) (Fig. S1). Here the mixing condition is based on the intensity of mixing (as an experimental factor) and the level of oxygenation by measuring redox parameters (mentioned below). Briefly, Jeevamrit was prepared (g L⁻¹) by adding cow dung (50 g, from a native breed “Gir” of Gujarat, India), cow urine (50 mL, from a healthy cow without any diseases in the past few years), jaggery (5 g), and chickpea flour (5 g), and pristine soil (approx. 200 g, without any anthropogenic impact: from a forest area “Aranya Uddhyaan”, Gujarat, India), as per the recommendations of NITI Aayog, Govt. of India⁵. The mixture was incubated for 7 days at room temperature under shaded conditions with the respective mixing regimes and subsequently analyzed for geo-microbiological attributes. Various physico-chemical parameters of the aqueous phase of the formulation, viz., oxidation–reduction potential (ORP), conductivity (EC), pH, dissolved oxygen (DO), total dissolved solids (TDS), and salinity were measured using a multiparameter instrument (Orion Star A329, Thermo Scientific, NC, USA). The readings were noted after the probe had stabilized. Further, ions [i.e., total nitrogen (TN), nitrate (NO₃⁻), ammonium (NH₄⁺), total iron (Fe²⁺/Fe³⁺), and alkalinity (CO₃²⁻/HCO₃⁻)] were estimated spectrophotometrically using commercial kits (DR6000, HACH, CO, USA). Total organic carbon (TOC) was estimated by the wet-oxidation method using a TOC analyzer (Aurora 1030W TOC Analyzer, OI Analytica, Texas, USA) that converted organic matter into CO₂, which was then measured by a solid-state, non-dispersive infrared detector (SS-NDIR). Elemental analysis of the solid phases was performed by acid digestion with aqua regia (1:5, v/v), subsequently diluted, filtered (0.22 µm), and analyzed using an inductively coupled plasma-optical emission spectrophotometer (ICP-OES: Optima 7300 DV, Perkin Elmer, CT, USA) following the recommended procedure (SIST ISO 11466:1996).

Microbial community analyses

16S rRNA gene amplicon sequencing and data analysis

Metagenomic DNA was extracted using a DNA isolation kit (DNeasy Powersoil Pro kit, QIAGEN) with appropriate modifications (30 min incubation after adding C2 and C3 solutions each). For each sample, DNA eluted from replicate extraction vials was pooled, concentrated, and quantified using a multimode reader (Cytation 5, BioTek Instruments, Inc., VT, USA). The V3-V4 region of the 16S rRNA gene was amplified with barcoded primers (515F/806R), libraries were purified, processed, and sequenced through NovaSeq 6000 (Illumina) using 500 Cycles (250 × 2) reagent kit on SP flow cell. Raw 16S rRNA gene amplicon reads were quality-filtered [removal of homopolymer runs (> 6 bp), sequences outside the 200–300 bp range, ≥ 3 primer mismatches] and subsequently analyzed using the QIIME v2 pipeline¹⁷. De novo-based clustering (99% identity) of filtered reads to operational taxonomic units (OTUs) was performed using VSEARCH under the QIIME workflow. Taxonomic assignment of representative reads from each OTU was performed using the SILVA v138 database (https://www.arb-silva.de/documentation/release-138). Diversity metrics (observed features, Chao1 richness estimator, Shannon, Simpson, and ACE) were derived using the core-diversity plugin within QIIME2. The core, unique, and accessory OTUs amongst the samples were estimated by using InteractiVenn (https://www.interactivenn.net/)¹⁸. The top 30 OTUs (core, unique, and rare) from each sample were aligned against quality-controlled 16S rRNA gene sequences of type and non-type strains in the BLAST-NR database, and phylogenetic trees were constructed in MEGA v11 using the maximum-likelihood method with 1000 bootstrap replicates¹⁹. iTOL v6.9 (https://itol.embl.de/) was used to draw, optimize, and visualize the phylogenetic tree²⁰. Metabolic inventories of the community were predicted using Functional Annotation of Prokaryotic Taxa (FAPROTAX) based on annotation information of the cultured prokaryotes²¹. It was performed using the collapse_table.py script on the OTU BIOM table generated through QIIME2. Microbial metabolic pathways were estimated based on the 16S rRNA gene data from the closed OTU picking method using the PICRUSt2 software package on the web-based Galaxy server using the Greengenes database²².

Shotgun metagenome sequencing, assembly, binning, and annotation

For shotgun sequencing, metagenomic DNA was extracted from 5 mL samples using the cetyltrimethylammonium bromide (CTAB) method, quantified with the Qubit ™ dsDNA BR Assay Kit using a Qubit ™ 4 Fluorometer (Thermo Fisher Scientific, MA, USA). Library preparation (50 ng) was done using the Nextera DNA Flex library preparation kit (Illumina) according to the manufacturer’s guidelines. An equimolar ratio of the libraries (0.7 nM) was pooled together and sequenced in a paired-end manner for 300 cycles on the NovaSeq 6000 (Illumina). The amplicon and metagenomic shotgun sequencing were performed at Gujarat Biotechnology Research Centre, Gandhinagar, Gujarat, India. The shotgun metagenomic raw reads were demultiplexed, and both primers and poor-quality sequences (Phred score < 25, chimeras, and adapters) were removed using Trimmomatic (v0.36) within the OmicsBox software package²³. FragGeneScan within OmicsBox was used for predicting open reading frames (ORFs)²⁴. Taxonomic classification was performed using Kraken 2 software²⁵. Orthologous gene families were screened using EggNOG through the GO classifier²⁶. Assembly of the reads was done using SqueezeMeta v1.3.0 in sequential mode with the MegaHIT assembler (default parameters)²⁷. Metagenomic binning was performed using MaxBin2 v2.2.5²⁸ and MetaBAT2 v2.12.1²⁹. A combination of binning results was performed using the DAS Tool³⁰ and was assessed in CheckM³¹. Bins with > 90% completeness and < 20% contamination were selected for annotation using the RAST-tk server (https://rast.nmpdr.org/rast.cgi) and visualized using the SEED server³². Metagenome-assembled genomes annotated with the DRAM tool, and distilled annotations were used to generate interactive functional summaries per genome, formatted as KBase-compatible assembly objects using the KBase platform³³.

Analysis of untargeted metabolites using liquid chromatography-mass spectrometry (LC–MS)

Metabolites were extracted by mixing 20 mL Jeevamrit with 50 mL LC–MS grade ethyl acetate (SRL, India), followed by vortexing (1 h) and overnight incubation at 200 rpm (Orbitrek shaker (Model LE), Scigenics biotech Pvt. Ltd., TN, India). This solvent-extraction workflow selectively captures the dissolved, low-molecular-weight compounds and their metabolite fractions representing the dissolved organic matter (DOM) pool of the samples. The extract was transferred to a separating funnel, and the organic phase was collected following liquid–liquid extraction (LLE). The solvent was evaporated to dryness at 40 °C using a N₂-evaporator (Turbo Athena Technology, Model: AT-EV-50, India). The dried extract was reconstituted in LC–MS grade ethyl acetate, centrifuged at 5000 × g for 5 min to pellet insoluble salts, filtered through a 0.22 µm nylon syringe filter (Axiva, India), and injected into LC–MS using an autosampler. Liquid chromatography run was performed using a UHPLC (Thermo Scientific Vanquish, MA, USA) coupled to an Orbitrap (Exploris 240 HRMS, Thermo Fisher Scientific, MA, USA) equipped with a heated electrospray ion source (H-ESI). The separation and detection of metabolites was achieved using a reversed-phase C18 analytical column of 100 mm × 2.1 mm × 1.9 μm, maintained at 35 °C. The sample injection volume was 10 μl. The mobile phase consisted of 0.1% formic acid in water (solvent A) and LC–MS grade 0.1% formic acid in acetonitrile (solvent B), at a flow rate of 0.3 mL/min. LC separation was performed with the following gradient: 100% solvent A held for 8 min, ramped to 100% B by 9.5 min, held until 10.5 min, returned to 100% A by 12 min, then shifted to 100% B at 29 min and held until 30 min, totaling the run to be 50 min. ESI ± mode with capillary voltages of 4.0 kV (+) and 2.5 kV (−) was set. The data analysis was performed using Compound Discoverer software v3.3 (Thermo Fisher Scientific) with an m/z range of 70–1050.

Evaluation of plant growth-promoting traits of cultivable members of Jeevamrit

Enumeration of cultivable microbial groups

To assess the plant growth-promoting traits of cultivable isolates of Jeevamrit, enumeration of total heterotrophs, Actinobacteriales, Azotobacter spp., Beijerinckia spp., and Rhizobial members was performed using selective agar media (Nutrient agar, Actinomycetes isolation agar, Ashby’s mannitol agar, Beijerinckia medium, and Rhizobium medium) following serial dilution series (until 10^–8) in phosphate buffer (20 mM, v/v), and incubated at 30 °C for 24–72 h. Axenic cultures (morphotypes) were purified by repeated sub-culturing on the same plates and subjected to various plant-growth beneficial traits (detailed in the subsequent sections).

Screening for mineral solubilization (P and K)

The extent of phosphate solubilization (solubilization index “SI”) was quantified by measuring available PO₄³⁻ in the spent medium (after bacterial growth) following the molybdenum blue method with a standard curve of KH₂PO₄ (1–20 µg mL⁻¹, w/v)³⁴. The concomitant decrease in pH during incubation (up to 5 days) was periodically monitored using a pH meter (A4220 Orion Star, Thermo, NC, USA). Solubilization of mineral K (feldspar, as potassium aluminosilicates) was assessed by using modified Aleksandrov agar medium (HiMedia Labs, India)³⁵. The medium was inoculated with a bacterial suspension (10⁸ CFU mL⁻¹) and incubated at 30 °C for 3–5 days. Observation of the halo zone around the colonies confirmed the solubilization trait, and SI was calculated by using the formula: SI = [(Diameter of zone of clearance/diameter of growth) × 100].

Phytohormone (IAA) production assay

Phytohormone (IAA) production by the isolates was assayed in the presence and absence of L-tryptophan (5 mM, precursor) using the Salkowski colorimetric method (500 mL reagent: 150 mL conc. H₂SO₄, 250 mL distilled water, 7.5 mL 0.5 M FeCl₃, v/v)³⁶. The standard curve was prepared by using the IAA standard (SRL, concentration range of 0–25 µg mL⁻¹, v/v, from 100 µg mL⁻¹ stock solution).

Ammonia (NH₃) production

NH₃ production was assessed by inoculating actively growing cultures (18 h) into peptone water and incubating at 30 °C for 3 days at 120 rpm. Nessler’s reagent was used (1:1, reagent: supernatant, prepared with NH₃-free water) for detection using a spectrophotometer at 450 nm with a standard curve of (NH₄)₂SO₄ (Conc. of 0.2-1 µmol).

Hydrogen cyanide (HCN) production

Production of HCN was assessed by cultivating cells of the isolates in nutrient broth amended with L-glycine (0.44%, w/v). A solution of Na₂CO₃ (2%, w/v) and picric acid (0.5%, w/v) was used for detection by impregnating it onto a Whatman cellulose filter paper (Grade-1) and placing it on the top of the plate (stuck to the lid). The development of an orange to red color indicated a positive reaction, thus denoting HCN production.

Lytic enzyme production (amylase, protease, cellulase)

All the isolates were screened for production of lytic enzymes (i.e., amylase, protease, and cellulase) following the methods described³⁷. For both amylase and cellulase production, actively grown bacterial cultures were spotted onto agar-supplemented minimal salt medium [gL^-1: MgSO₄·7H₂O, 0.20; NaCl, 2.00; (NH₄)₂SO₄, 1.00; FeSO₄·7H₂O, 0.01; yeast extract, 1.00; bacteriological agar, 15.00] supplemented with either starch (1%,w/v) or carboxy-methyl cellulose (CMC, 0.3%,w/v) as sole source of hydrolytic substrates. Observation of the zone of clearance surrounding the growth, followed by Gram’s iodine treatment (33% Iodine), was used for assessing starch hydrolysis. Plates were flooded with Congo red solution (1%, w/v) and washed with NaCl (1%, w/v) to visualize the zone of cellulose hydrolysis. For proteolytic activity, isolates were spotted onto skim milk agar, and the clear zone surrounding the bacterial growth indicated positive activity. Based on overall performance (i.e., the number and extent of PGP traits expressed), isolates were categorized from highly beneficial to low beneficial. Accordingly, the best-performing strains were selected for molecular identification.

Molecular identification of cultivable bacterial isolates and phylogenetic analyses

A total of ten isolates were selected for molecular identification using full-length 16S rRNA gene sequencing and phylogenetic reconstruction. High-quality genomic DNA was extracted using the Genomic DNA kit (HiPurA^®, Himedia) according to the manufacturer’s instructions. The full-length 16S rRNA gene was PCR amplified using the universal eubacterial primers (27F/1492R) (reaction mix: TaKaRa Taq™ master mix, 50 µl; 29.5 µl of nuclease-free water, 3 µl of each primer, and 4 µl template DNA; thermal cycling steps: 5 min at 95 °C, 30 cycles of 95 °C for 30 s, 55 °C for 30 s, 72 °C for 60 s, and a final extension at 72 °C for 10 min), gel purified using a QIAquick Gel Extraction Kit (QIAGEN), and sequenced using capillary electrophoresis by 3500 XL Genetic Analyzer (Applied Biosystems). Raw sequences were processed using BioEdit v3.7.22, and a contig was formed using the Contig assembly program (CAP)³⁸. Homology search of the 16S rRNA gene sequences with reference strains (type and non-type) was carried out using the EzTaxon Biocloud server (https://www.ezbiocloud.net/). Best hits (> 97% similarity) were selected for multiple alignments using the CLUSTAL-W package of MEGA v11¹⁹. Phylogenetic reconstruction, validation, and taxonomic assignment were performed using the neighbour-joining (NJ) method based on bootstrap analysis with 1000 replications with the Jukes-Cantor distance model. Both maximum-likelihood (ML) and minimum-evolution (ME) methods were employed to test the robustness of the trees. The tree was optimized using iTOL with default parameters.

Statistical analysis

Canonical correspondence analysis (CCA) was performed to assess relationships between environmental variables and OTUs (Top 50), while principal component analysis (PCA) was applied to explore overall variation in the dataset. Both analyses were conducted using PAST v4.16c software³⁹. Multivariate data analysis (Pearson correlation, screening bar plots, and compositional stacked bar plots) was performed using OriginPro 8.0 (OriginLab Corporation) and visualized using Chiplot (https://www.chiplot.online/). Microbial co-occurrence network analysis was conducted using the integrated network analysis pipeline (iNAP)⁴⁰, employing SparCC methods to construct networks from compositional data based on Spearman’s rank correlation. The network matrix file generated by iNAP was visualized using Cytoscape (v3.7.1). Co-occurrence networks were constructed using SparCC⁴¹ to identify significant correlations (|r|≥ 0.5, p < 0.05, 100 bootstraps) between the top 50 core OTUs and differentially enriched OTUs. The resulting adjacency matrix was visualized and analyzed in R using igraph with hubs defined as nodes with degree > mean + 1 SD and visualized using the Fruchterman–Reingold layout⁴². The heat tree technique utilizes the hierarchical organization of taxonomic classifications quantitatively (via median abundance) and statistically (using the non-parametric Wilcoxon Rank Sum test) to illustrate the disparities in taxa (greater abundance within each experimental condition) among various communities and represented using a heat tree generated using MicrobiomeAnalyst 2.0 platform (https://www.microbiomeanalyst.ca)⁴³. A metabolic reconstruction of the microbial functions (metagenomic dataset) was carried out using the gene annotation table parsed from EggNOG annotation and visualized using Biorender (https://www.biorender.com/). The Voronoi heatmap was generated to depict common taxa and their distribution for functional gene categories using SR Plot (https://www.bioinformatics.com.cn/en). Metabolite heatmaps and metabolic pathway analyses were conducted using MetaboAnalyst 6.0 (http://www.metaboanalyst.ca). Data were normalized via log transformation and Pareto scaling. Metabolites with variable importance in projection (VIP) scores > 1.5 and statistically significant differences (p < 0.01) were identified as major contributors to sample variability. Pathway enrichment analysis was performed using the KEGG database⁴⁴.

Results

Effects of mixing conditions on the physico-chemical changes of the Jeevamrit formulation

Elemental and physico-chemical variations

To understand the mixing-driven compositional differences of Jeevamrit formulations, elemental/ ionic and other physico-chemical properties were analyzed and presented in Table 1. Both hardness and total alkalinity (CO₃²⁻ + HCO₃⁻) varied markedly among the samples. Elevated soluble and total Fe, along with higher Zn, Cu, Mn, Ca, and Mg in CM (1.6–2.9 times), suggest enhanced metal solubilization/mineral weathering under higher mixing rates. Similarly, the higher total N in CM (1.5–1.8 times majorly organic N) may result from the aerobic hydrolysis of polymeric N present in the added substrates, and elevated levels of NH₄⁺-N, soluble Fe (Fe²⁺), and moderate DOC indicate reductive pathways in AO.

Table 1.

Physico-chemical properties and microbial diversity indices of Jeevamrit samples prepared under different mixing conditions.

	Parameters	CM	IM	NM	AO
Physico-chemical properties	pH (initial/final)*	7.1/5.53	7.1/5.43	7.2/5.36	6.8/5.13
	DO (initial/final)*	8.1/0.4	7.5/0.26	7.5/1.0	7.2/1.5
	ORP (RmV)*	551.4/255.2	493.2/234.1	472.3/260.9	470.5/261.8
	EC (mS)*	3.14/5.84	3.07/5.53	3.04/5.26	3.0/5.04
	TDS (ppt)*	1.57/2.91	1.53/2.76	1.52/2.62	1.50/2.51
	Hardness (CaCO3)	17.4	8.1	19.6	0.06
	Alkalinity	867	730	650	861
	Nitrate (NO3−)	0.4	0.2	0.4	0.4
	Nitrogen, Ammonia (Nessler’s)*	29.6	20.4	24.4	34.4
	Total Nitrogen (N)	138	77.1	90.8	93.6
	Available P (Olsen)*	732	610	620	730
	Total Soluble Iron (Ferrozine)	4.6	3.3	3	2.9
	Total Iron*	9.3	5.8	3.2	3.4
	Total Organic Carbon (TOC)	161.2	116.7	132.7	146.8
	Tannin & Lignin	98	76	78	108
	Ca	11.57	11.04	10.37	11
	Cu	0.33	0.20	0.20	0.18
	Co	0.032	0.032	0.026	0.028
	Cd	0.039	0.038	0.039	0.038
	Cr	0.057	0.040	0.038	0.034
	Mg	7.94	7.79	6.87	6.48
	Mn	0.69	0.62	0.45	0.50
	Mo	0.05	0.03	0.04	0.04
	Zn	0.51	0.29	0.20	0.17
Community Analysis	Number of reads for analysis	296,612	207,334	444,007	595,954
	Total No. of OTUs	21,150	15,753	14,236	31,597
	Percent of Core OTUs	8.66	11.63	12.87	5.80
	Percent of Unique OTUs	21.6	15	10.6	36.1
Alpha Diversity	Shannon Index	8.36	8.77	7.68	9.18
	Simpson’s Diversity Index	0.95	0.96	0.92	0.98
	ACE index	71,280	45,068	110,044	91,119
	Chao1	66,957	41,260	101,550	85,585

*Denotes the analysis was performed in triplicate and the average value was quoted with < 1% of standard deviation. All units are expressed as mg L-1, except pH, ORP, EC, TDS for which the appropriate unit is mentioned in parenthesis.

CM: Constant mixing, IM: intermediate mixing, NM: No mixing, AO: Anoxic.

Microbial diversity, community patterns, and biogeochemical transformations

A higher prokaryotic diversity and richness (OTUs as species) in AO, followed by NM (Table 1), was observed, which corroborated the total bacterial counts (CFU mL⁻¹), i.e., maximum heterotrophic growth observed for AO (lawn growth in the highest dilution, approximating to 10^12–13 on nutrient Agar), followed by NM (1.3 × 10⁹). Higher counts of actinobacteria (7 × 10⁷, 6.9 × 10⁷), Rhizobium and Azotobacter (2 × 10⁸) were noted (Table S1). A clear separation between AO and CM, and a closer association between NM and IM (low-mixing samples) was noted (Fig. S2). Furthermore, corroborating the physico-chemical data, it was observed that both CM and AO were more constrained with total N, TOC, alkalinity, tannin-lignin, and Ca/metals, whereas NM and IM were constrained with hardness and pH. A negative correlation between (NO₃⁻)/ammonium nitrogen (NH₄⁺–N) with total organic carbon (TOC) and pH, and a positive correlation with micro elements, including Fe (TFe/Fe²⁺), and OTUs was noted (Fig. 1a, b). The occurrence of oxidation of complex organic matter to yield simpler organic carbon, transformation of nitrogen species, and dissolution of metals is thus evident in CM, whereas reductive N and Fe metabolism with higher microbial diversity was noted in anoxic solution.

Fig. 1.

Microbial community composition and beta diversity analyses w.r.t physico-chemical changes of Jeevamrit samples: (a) Canonical Correspondence Analysis (CCA) biplot illustrating the relationships between environmental variables (green arrows) and microbial community composition based on the top 50 most abundant operational taxonomic units (OTUs) in Jeevamrit samples. OTUs are represented by numbered blue dots, while treatment groups are indicated by colored asterisks: CM (constant mixing), IM (intermediate mixing), NM (no mixing), and AO (anoxic). Arrows indicate the direction and magnitude of the influence of individual environmental parameters on microbial distribution. Axis 1 and Axis 2 explain 46.9% and 34.37% of the total constrained variance, respectively; (b) Heatmap depicting the Pearson correlation of physico-chemical properties, where correlation coefficients are color-coded from positive (yellow) to negative (purple) and the extent of similarities is represented by a cladogram made by hierarchical clustering; (c) Bar plots of microbial community composition (relative abundance) at the phylum (left) and genus level (right) across different samples. The relative abundance (%) of dominant phyla (Pseudomonadota, Bacillota, Bacteroidota, and others) is shown, along with hierarchical clustering based on Bray–Curtis dissimilarity; (d) Comparison of relative abundance (%) of dominant genera in CM and NM samples determined using whole-genome sequencing (WGS, blue) and 16S rRNA gene sequencing (orange), where coherence/differences highlight sequencing biases in detecting such microbial taxa.

Microbial community composition with respect to mixing conditions

Taxonomic composition

An average of 3.85 × 10⁵ 16S rRNA gene sequence reads per sample were assigned to a total of 65,534 OTUs (with an average of 20,684 OTUs per sample). Bacteria predominated (average 99.94%) over archaea (0.03%). The dominant phyla included Pseudomonadota, Bacillota, Bacteroidota, Verrucomicrobiota, Campylobacteriota, and Actinobacteriota, constituting 90–95% of the community (Fig. 1c). Members of Gammaproteobacteria, Bacilli, Bacteroidia, and Clostridia collectively contributed 90–94% of communities across the samples and were thus regarded as the dominant members. Correspondingly, Pseudomonadales [with ~ 44% in both CM and NM], Lactobacillales, Bacteroidales, Enterobacterales, Bacillales, Oscillospirales, and Burkholderiales were the major taxa (Fig. S3a). At the genus level, Acinetobacter exhibited the highest prevalence across all samples (11.42–38.59%). A marked decrease of Trichococcus and Clostridium sensu stricto 12 and an increase in Comamonas and UCG-005 (a Ruminococcaceae member) relative abundance were noted in CM (Fig. 1c). Whereas an increase in Trichococcus, Dickeya, Streptococcus, and Enterobacterales in AO and Lysinibacillus and Pseudomonas was noted in IM and NM.

Functional potential and metabolic profiling with diversity shifts

Metabolic prediction showed a preponderance of Chemoheterotrophy, aromatic compound degradation, and fermentation across all the samples (AO, CM, IM, NM), indicating active organic matter degradation by the community (Fig. S3b). CM/NM was noted with oxidative carbon, nitrogen, and sulfur metabolism, reflecting intensive nutrient cycling. Whereas, enrichment of terpenoid and polyketide metabolism, reductive nitrogen (nitrogenase, ferredoxin-nitrate reductase), and fermentative carbon metabolism (lactate dehydrogenase, propanediol dehydratase), in AO reflected the prevalence of anaerobic metabolic routes, validating the observed physico-chemical data (Fig. S3c). To assess the impact of mixing on the diversity, a heat tree (Wilcoxon signed-rank test) was generated to identify the differential assemblage of community members (Fig. S4). A marked shift in Bacillota, e.g., Bacilli (Lactobacillales, Carnobacteriaceae, and Enterococcaceae) and Bacteroidia was evident under anoxic conditions, whereas Aerococcaceae and Streptococcaceae remained unaffected (p < 0.05). Amongst Clostridia, Oscillospirales members were strongly perturbed, while Lachnospiraceae, Ruminococcaceae, and Clostridiaceae were enriched but remained relatively stable. In Gammaproteobacteria, members of Comamonadaceae, Pseudomonadaceae, and Moraxellaceae were markedly shifted under anoxic set-up, while Enterobacterales and Burkholderiales remained unaltered. Rank abundance plot of the top 20 OTUs revealed that taxa from one sample contributed moderately to high abundances in other samples, indicating core and dominant members to be the major drivers of the community. It is worth highlighting that Acinetobacter OTU was abundant in CM (32%) compared to the other samples, while AO exhibited the highest contribution for most of the remaining OTUs (Fig. S5).

Core members comprised of [1832 OTUs (2.8%)] Acinetobacter, Trichococcus, Pseudomonas, Clostridium sensu-stricto, Bacteroides, and Lysinibacillus, constituting 5.8–12.7% of the respective samples, and were dominant in all the samples (Fig. S6 a, b). Prediction of the observed metabolic trait of these members revealed that Acinetobacter members (OTUs) differed in NH₄⁺–N and NO₃⁻–N assimilation, N-intermediate respiration (NO₂⁻, NO, N₂O), phosphate/polyphosphate (PO₄³⁻) solubilization, IAA and cellulase production, and ethanolamine/acetate metabolism. Similarly, Trichococcus varied in fermentative traits; Pseudomonas in terms of denitrification, pollutant degradation, heavy metal biosorption, antifungal activity, and secondary metabolite production; and Clostridium sensu stricto for H₂ metabolism, Fe³⁺–Fe²⁺ reduction, acetone–butanol–ethanol (ABE) fermentation, and cellulolytic activity (Fig. S6 b). Pearson correlation of dominant OTUs with environmental factors revealed a positive correlation of N, NH₄⁺-N, and total alkalinity [TA; (CO₃²⁻ + HCO₃⁻)] with Streptococcus to Clostridium sensu stricto 12. Tannin and lignin, NO₃⁻, TFe, TN, and TOC were positively correlated with Acinetobacter (OTU-13 and OTU-18) (Fig. S7). The relative abundance of Comamonas_uncultured compost, Pseudomonas, Lysinibacillus xylanilyticus, and Stenotrophomonas showed a positive correlation with most environmental parameters. The CCA plot further supported these results, indicating closer affiliation of CM (eigen-vector) to Mn, Co, TN, and TFe, indicating a strong positive relationship (Fig. 1a). A positive correlation between TN, Mo, and Cu and OTUs 11 (Comamonas_uncultured compost), 30 (Lysinibacillus xylanilyticus), and 38 (Clostridiaceae_unidentified), indicated the probable involvement of these taxa in mediating such specific transformation of elements. Co-occurrence analysis of core and dominant OTUs suggested 86 associations among 44 nodes with 32 positive (co-occurrence) and 39 negative (co-exclusion) interactions. Acinetobacter (n = 11), Dickeya (n = 5), Streptococcus (n = 8), Trichococcus (n = 7), Comamonas (n = 6), and Macelibacteroides (n = 6) emerged as the hub taxa, where Comamonas was noted to be a key member of the hub, having strong affiliation with aerobic organic matter metabolizing genera Stenotrophomonas, Acinetobacter, Advenella, and negative connections with fermentative Clostridiaceae, Enterococcaceae, and Trichococcus (Fig. S8a, b). Unique taxa (~ 9–12%, exclusively present in one sample but absent in others) w.r.t incubation conditions (oxygenic/anoxygenic) were also analyzed (Fig. S9a). In CM, members of Moraxellaceae (50%), followed by Pseudomonadaceae and Comamonadaceae, constituted ~ 65% of the community, as top unique taxa. These members are non-fermenting, catalase-positive, aerobic/facultative anaerobes. In contrast, in AO, we observed equal representations of anaerobic and fermentative members of Carnobacteriaceae, Clostridiaceae, Streptococcaceae, and Enterobacteriaceae, together constituting 63% of the community, thus justifying their role w.r.t the condition provided. Members with mixed metabolic repertoires (aerobic, anaerobic metabolism, and fermentative abilities), i.e., Planococcaceae, Clostridiaceae, and Lachnospiraceae, were abundant in IM (85% relative abundance). Predictive metabolic profiling of these unique OTUs revealed that the gondoate biosynthesis pathway was enriched under anaerobic conditions, specifically in AO (4.2%) and NM (4.4%). Whereas aerobic respiration-1 was up-regulated in CM (5.19%) and IM (4.90%), consistent with aerobic conditions provided (Fig. S9b). The isobutanol biosynthetic pathway, responsible for converting pyruvate to isobutanol via fermentation, exhibited a decreasing trend from CM (4.0%) to AO (3.5%). Overall, the mixing condition had a profound effect on the community’s composition, enrichment of the taxa, and metabolic traits of these members, which could ultimately affect the physicochemical changes of the Jeevamrit formulation.

Shotgun metagenomics-based taxonomic and gene inventories of the communities

Community composition and taxonomic profiles

The community and genetic repertoire of the microbiome of Jeevamrit formulations (CM and NM) are inferred, and key parameters are summarized in Table S2. The prokaryotic members (bacterial diversity and gene abundance) were found to be considerably higher, i.e., 97.9% in CM and 96.5% in NM, over viruses (1.7–3.2%), archaea (0.18–0.02%), and eukaryota (0.15–0.22%). Pseudomonadota, Bacillota, and Bacteroidota were noted to be abundant (Fig. S10a). Coherent taxonomic distribution (between 16S rRNA sequencing and shotgun metagenomics) was observed, denoting Acinetobacter as the predominant member. However, members of Klebsiella and Bacteroides were well represented in WGS, whereas Dickeya and Pseudomonas were variably detected (Fig. 1d).

Functional gene repertoires and nutrient transformation potential

COG distribution of genes revealed overall higher gene allocation to amino acid transport (E), energy conversion (C), and carbohydrate metabolism (G) categories (Fig. S10b). CM showed a higher proportion of genes related to Information Storage & Processing (19.66% vs. 18.03%), particularly replication and repair (L) and translation (J), whereas Cellular Processes & Signaling were comparable across samples. Considering organic matter-driven nutrient transformation (N, Fe, K, P, S) as key processes, genes about such pathways were mined (Fig. 2i, ii). Abundance of denitrification (narG, nirK/S, norB, nosZ), nitrogen fixation (nifH), and nitrate reduction (nrfA/H, nasAB) was primarily attributed to members of Pseudomonas (Pseu), Gammaproteobacteria (Gamm), Clostridiaceae (Clos), and Rhodocyclales (Rhod). The absence of genes for nitrification (amo, hao, nxrA) and anammox (hzo) suggested limited aerobic or oxidative nitrogen-processing capacity (Fig. 2i, a). P solubilization appeared to occur through acidolysis (pqq, gcd) and enzymolysis (phoR, pstS, appA, phnP, ppx, ppk), releasing orthophosphate from mineral and organic pools. Solubilization of K-bearing minerals (e.g., feldspar, mica) was linked to organic acid–mediated weathering, driven by gltA and ppc genes. Moraxellaceae (Mora) and Comamonadaceae (Coma) emerged as key mobilizers of these nutrients (Fig. 2i, b, c). Siderophore-mediated Fe-acquisition highlighted the major involvement of Comamonadaceae (Coma) and Gammaproteobacteria (Gamm), underscoring the dominance of Gram-negative Fe uptake mechanisms (Fe³⁺–catecholate complexes via FepA–TonB–ABC systems (FepC/D/G) over Gram-positive (Substrate binding protein, permease) (Fig. 2i, d). In the sulfur cycle, sulfate (via cysPUWA) and alkanesulfonate (via ssuABCD) are found to be reduced to sulfide via cysND, cysC, cysH, cysJI, and ssuE/D. Sulfide was subsequently incorporated into L-cysteine (cysK) and further into L-homocysteine via metE; however, the cth gene, which encodes cystathionine γ-lyase, was absent. Gammaproteobacteria and Pseudomonas spp. remained consistently dominant across treatments (Fig. 2i, e). Metabolism of glucose, glycerol, benzene, toluene, and phenol proceeded through intermediates such as pyruvate, acetyl-CoA, and succinate, primarily driven by Moraxellaceae (Mora) and Clostridiaceae (Clos). Although benD was absent, the catechol (ortho/para) pathway represented the principal route for aromatic metabolism in these samples. In addition, anaplerotic reactions leading to fermentative products (alcohols and organic acids) were observed in abundance (Fig. 2ii, a). Pseudomonadaceae (Pseu), Moraxellaceae (Mora), and Gammaproteobacteria (Gamm) are attributed for phytohormone modulation, specifically auxin (IAA) biosynthesis via the IAM, IPA, and partial TAM pathways (iaaM, iaaH, ipdC, tynA). The absence of oxr, iad1, nthA, tdc, aldh, aatA, and YUCCA suggested that auxin biosynthesis was routed through indole-derived aldehyde, amide, and pyruvate intermediates, but not through aldoxime or acetonitrile pathways (Fig. 2ii, b). Hydrolysis of complex polymers/polysaccharides, mediated by amyA, celA, xynA, and chiA, was implicated in carbon turnover and the formation of a reserve/available organic matter pool. These activities were primarily associated with Bacteroidota (Bact), Lysinibacillus (Lysi), and Marinilabiliaceae (Mari) in CM and Porphyromonadaceae (Porp) and Alphaproteobacteria (Alph) in NM, highlighting treatment-specific metabolic repertoires (Fig. 2ii, c). Differential gene abundance patterns revealed that CM exhibited higher frequencies of genes related to auxin (nit, iaaM, tryA), carbohydrate (pyk, catA, dhaB, adh, ppc), and potassium metabolism (glt, ppc), suggesting enhanced auxin production, microbial organic carbon assimilation, and nutrient solubilization. In contrast, NM showed relatively higher abundance in phosphorus (phnP, phoR), iron (fepA, tonB, entS), and sulphur metabolism (cysP, cysD, ssuA), indicating greater potential for elemental redox transformations and uptake. Nitrogen metabolism displayed a mixed pattern: narG (nitrate reductase) was more abundant in NM, whereas nifH (nitrogen fixation) and nirK (nitrite reduction) were more enriched in CM. Hydrolytic enzymes also followed this treatment-specific trend, with celA and amyA more abundant in NM, and xynA enriched in CM.

Fig. 2.

(i, ii) Shotgun metagenome-based mining of genes related to nutrient cycling and microbial taxa harboring such genes for constant mixing (CM) and no mixing (NM) jeevamrit samples: (i,a) Nitrogen transformation pathways e.g., nitrogen fixation, nitrification, denitrification, ammonification, and nitrate reduction; (i,b) Phosphorus solubilization e.g., acidification and hydrolytic enzymatic mobilization; (i,c) Potassium solubilization e.g., acidolysis-mediated K⁺ release from feldspar and mica; (i,d) Iron acquisition e.g., siderophore biosynthesis, secretion, and transport in Gram +ve and −ve bacteria; (i,e) Sulfur metabolism e.g., sulfonate and sulfate assimilation, including transport and reduction to sulfide and cysteine; (ii,a) Organic matter degradation i.e., aromatic and carbohydrate metabolism e.g., degradation of benzoate, toluene, and phenolics via catechol and β-ketoadipate pathways, central carbon and glycerol metabolism; (ii,b) Indole-3-acetic acid (IAA) biosynthesis e.g., tryptophan-dependent pathways suggesting microbial contributions to plant growth promotion; (ii,c) Hydrolysis of complex polymers for polysaccharide and amino acid derivative degradation e.g., cellulase, amylase, xylanase, and chitinase. Bar graphs represent functional gene abundance, and Voronoi heatmaps show dominant bacterial taxa associated with each pathway.

Metagenome-assembled genomes (MAGs) and core functional guilds

Metagenomic binning yielded 20 high-quality (completeness > 90%, contamination < 20%) bins as metagenome-assembled genomes (MAGs) (Table 2), which belonged to Pseudomonadota, Bacillota, and Bacteroidota. Overall, nucleotide metabolism (purine and pyrimidine), amino acid and cell wall biosynthesis (lysine, peptidoglycan, lipopolysaccharide), and pathways related to membrane transport (ABC transporters), signal transduction (two-component systems), and translation (ribosomal genes) were consistently enriched across all bins (both CM and NM). Phylogenetic analysis of MAGs with corresponding 16S rRNA-based OTUs indicated coherent clustering of Clostridium saccharobutylicum (Concoct 133-CM), Lysinibacillus sp. (Concoct 67-CM), and Veillonella dispar (Concoct 139-CM) with Clostridiaceae OTU3, Lysinibacillus sp. OTU2, and Veillonella sp. OTU2, respectively, confirming their representation in the CM (Fig. S11a). In contrast, Concoct 9-CM, Concoct 44-CM, and Concoct 2.60-CM lacked direct alignment with known OTUs. Similarly, in the NM sample, Veillonella atypica (Concoct 101-NM) clustered with Veillonella sp. OTU2, supporting its classification, while Acinetobacter sp. NCu2D-2 (Concoct 99-NM) showed no OTU-level correspondence (Fig. S11b). Comparative gene annotation using RAST showed that MAGs of CM [Klebsiella pneumoniae (Concoct 44-CM, Gammaproteobacteria-GAMM) and Pseudaeromonas aegiceratis (Concoct 9-CM, Pseudomonas-Pseu)] shared similar metabolic profiles and thus formed a conserved metabolic guild, notably contributing to elemental nutrient cycling e.g. denitrification and nitrate ammonification (narG, nrfA, nosZ, nirK) and siderophore-mediated iron acquisition (tonB, fepA, fepC, entS) (Fig. 3). In parallel, P. aegiceratis supports nitrosative stress responses and nitrogen fixation (nosZ, nirK, limited nifH), with Pseu-specific enrichment. Both taxa are active in sulfur assimilation (cysP in GAMM, cysD in both) and phosphate metabolism (pstS, phnP); however, the phoR gene, which regulates phosphate uptake, is noted to be absent. For carbohydrate metabolism, K. pneumoniae MAG showed a higher frequency of genes pta, alsS, and dhaB, whereas P. aegiceratis showed restricted activity, limited to pta-driven mixed acid fermentation. This observation corroborated the principal role of Gammaproteobacteria (GAMM) in CM in mediating redox transformation coupled to organic carbon metabolism (Fig. 3). It is notable that MAGs sharing similar taxonomy but are assigned to different samples, such as Gemmobacter fulvus (Concoct120-NM and Concoct38-CM), Lysinibacillus spp. (Concoct67-CM and Concoct18-NM), and Acinetobacter spp. (Concoct99-NM and Metabet2.60-CM) exhibited comparable metabolic capabilities, indicating their common role in driving similar changes across both samples and suggesting these taxa as core community members. Gemmobacter fulvus MAGs displayed high pathway potential for butanol biosynthesis and acetyl-CoA to butyrate fermentation, yet lacked key genes (pflB, dhaB, adhE, pta) in the Voronoi gene distribution, whereas Acinetobacter MAG contained a complete set of genes for fermentative and alternative carbon utilization pathways. This comparative analysis highlights that MAGs belonging to different taxa, i.e., K. pneumoniae and P. aegiceratis, but originating from the same CM sample, demonstrated distinct, condition-specific functional roles shaped by the local environment. In contrast, MAGs with identical taxonomic affiliations (Gemmobacter, Lysinibacillus, and Acinetobacter) across different samples (CM and NM) retained conserved functional traits. DRAM-based annotation of CM-derived MAGs (Comamonadaceae, Acinetobacter sp., Trichococcus sp.) revealed broader metabolic capabilities, including complex polysaccharide degradation (α/β-glucans, pectin, xylan, chitin), aerobic carbon metabolism (acetate, lactate, formate utilization), and key nitrogen cycling functions such as ammonia oxidation, nitrate/nitrite reduction, etc. These traits suggest enhanced aerobic respiration, nutrient solubilization, and higher energy flux. In contrast, NM-derived MAGs (e.g., Clostridium sp., Enterococcus sp., Lactobacillus sp.) encoded pathways characteristic of anaerobic and fermentative metabolism, including lactate and acetate production, partial nitrate reduction, arsenate and selenate respiration and dissimilatory dissolution. Methanogenesis-associated pathways were also more abundant, a metabolic shift toward anaerobic energy conservation (Fig. S12).To provide gene-level confirmation of these aerobic and fermentative shifts, the representative oxidative (e.g., sdhA, gcd, catA) and fermentative/anaerobic (e.g., ldhA, adhE, pflB, frdA) marker genes enriched across CM and NM are presented in Table S3. Overall, CM MAGs promote nutrient turnover and bioavailability through aerobic processes, whereas NM MAGs are adapted to low-oxygen environments, favoring fermentation and redox buffering. This highlights that mixing-driven conditions modulate the functional potential of Jeevamrit microbiomes.

Table 2.

Comprehensive genomic and functional profiling of metagenome-assembled genomes (MAGs) from Jeevamrit samples (Constant mixing (CM) and No mixing (NM)).

Sample	Bin ID	Contigs	Length	N50	Completeness (%)	Contamination (%)	Disparity	Strain heterogeneity (%)	GC (%)	CDS	RNAs	Subsystems	Phylum	Top hit (Genomic blast)
CM	Concoct.9	517	3,307,077	8597	96.08	5.89	0.016	14.29	61.89	3550	60	284	Pseudomonadota	Pseudaeromonas aegiceratis strain ZJS20 (PQ168984)
	Concoct.38	802	2,677,557	4118	91.06	4.39	0.036	66.67	58.71	3313	40	255		Gemmobacter fulvus strain con5 (NR_181766.1)
	Concoct.44	325	6,151,882	30,755	98.94	5.03	0.031	65.43	56.61	6417	85	395		Klebsiella pneumoniae strain Kp_03 (JAPIVU000000000.1)
	Metabet 2.60	116	2,390,829	29,339	77.23	5.17	0	62	41.49	2409	65	234		Acinetobacter sp. NCu2D-2 (CP015594.1)
	Concoct.36	361	2,693,805	12,211	94.06	2.93	0.009	14.29	46.7	2894	26	208	Bacillota	Trichococcus flocculiformis (Y17301)
	Concoct.67	550	3,514,077	11,065	93.86	19.7	0.012	16.98	37.22	3913	40	243		Lysinibacillus sp. 3P01SB (CP154840.1)
	Concoct.133	221	3,086,410	25,100	99.19	1.69	0.04	0	28.94	3404	77	203		Clostridium saccharobutylicum strain BAS/B3/SW/136 (NZ_CP016089.1)
	Concoct.139	64	1,880,027	60,806	98.86	0.1	0.039	0	39.99	1807	45	205		Veillonella dispar strain NCTC11831 (NZ_LR134375.1)
	Concoct.75	79	3,673,511	84,368	99.25	1.28	0.063	14.29	41.51	3273	62	230	Bacteroidota	Bacteroides graminisolvens: MAG (NZ_OY762556.1)
	Metabet 2.10	393	3,570,439	11,741	87.78	2.19	0.04	43.33	37.58	2509	32	202		Dysgonomonas capnocytophagoides strain SMU001 (NZ_AP028867.1)
NM	Concoct.99	93	2,710,978	55,590	95.34	2.3	0	39.13	40.97	2620	88	260	Pseudomonadota	Acinetobacter sp. NCu2D-2 (CP015594.1)
	Concoct.120	517	2,611,235	7648	89.54	3.87	0.03	46.67	58.61	3041	41	236		Gemmobacter fulvus strain con5 (NR_181766.1)
	Concoct.18	76	3,018,196	140,470	97.35	1.21	0.179	0	36.73	2953	36	247	Bacillota	Lysinibacillus parviboronicapiens strain VT1065 (NZ_CP073713.1)
	Concoct.101	49	1,932,039	76,500	100	0	0.051	7.14	40.01	1862	45	208		Veillonella atypica strain G765s (CP103397.1)
	Metabet 2.21	65	2,904,910	74,557	98.44	1.12	0.032	50	40.84	2875	50	228		Enterococcus gallinarum strain K205-4a (CP116510.1)
	Concoct.4	197	3,349,813	49,061	100	13.46	0.01	10	36.02	3194	37	195	Bacteroidota	Crocinitomicaceae bacterium: MAG (OZ252050.1)
	Concoct.52	529	3,264,722	9111	92.69	5.79	0.012	100	34.33	3529	91	191		Flavipsychrobacter sp. Isolate: MAG (OZ246011.1)
	Concoct.71	879	8,193,507	12,223	94.8	51.07	0.005	12.15	40.95	7848	95	252		Macellibacteroides fermentans: MAG (OY761811.2)
	Metabet 2.31	375	3,566,855	12,644	95.36	3.42	0.022	66.67	37.48	3392	37	205		Dysgonomonas capnocytophagoides strain SMU001 (NZ_AP028867.1)
	Metabet 2.35	356	3,054,732	10,957	93.65	2	0.003	40	42.15	2800	60	218		Macellibacteroides fermentans: MAG (OY761811.2)

Fig. 3.

Details of metagenome-assembled genomes (MAGs) recovered from constant mixing (CM) and no mixing (NM) samples: (a) Mega BLAST (Basic Local Alignment Search Tool) based top hit-based homology of MAGs showing affiliation to bacterial taxa and relative abundance of such taxa in the respective samples performed using OmicsBox; (b) bubble plot showing gene counts (color gradient, purple to yellow) assigned to SEED (Subsystems-based annotation) functional categories e.g., auxin biosynthesis, and metabolism of aromatic compounds, carbohydrates, K, P, S, Fe, and N using the RAST-tk (Rapid Annotations using Subsystems Technology toolkit) server. MAGs are clustered using Bray–Curtis dissimilarity values using the k-mean clustering method and represented by a cladogram.

Metabolite profiling of Jeevamrit formulation

PLS-DA analysis showed clear metabolomic divergence in samples. Under high O₂ (> 1.5 VIP score), indole-3-propionic acid, cannabichromene, 6-deoxoteasterone, 4′-hydroxyacetophenone, and 19-oxoandrost-4-ene-3,17-dione showed a strong influence on the PLS-DA model. Whereas, upregulation of 4,4-dimethylcholesta-8,14,24-trienol, gibberellin A12 aldehyde, and 9-hydroxylarone was observed. Members of Acinetobacter, Comamonas, and Dickeya were strongly associated (r > 0.8, p < 0.05) with fatty acid, steroid, and linoleic acid biosynthesis (Fig. S13a). KEGG functional allocation further supported these associations, showing enhanced activity in β-alanine metabolism, CoA biosynthesis, tyrosine metabolism, fatty acid degradation, and steroid hormone biosynthesis under oxic conditions (Fig. S13b). Conversely, under low O₂ saturation, elevated levels of dimethyl benzimidazole, obtusifoliol, and quinoline-4,8-diol indicated involvement in the biosynthesis of sterols, vitamin precursors, and redox-active compounds (Fig. S13c). KEGG enrichment study corroborated such observations, denoting steroid biosynthesis, metabolism of bile acid, tryptophan, and CoA biosynthesis to be upregulated. Thus, distinct metabolic patterns were mediated by a specific set of microbes like Clostridium sensu stricto, Dickeya, Acinetobacter, and Trichococcus (r > 0.6–0.7, p < 0.05) (Fig. S13d), thus highlighting condition-specific microbe-metabolite networks. These results demonstrate that microbial communities dynamically configure their metabolic outputs in response to oxygen availability, thus needing optimization for yielding plant-growth-specific metabolites for better agricultural application.

Characterization of cultivable members of Jeevamrit and PGP traits

In vitro seed germination assay revealed 100% germination of mung bean (Vigna radiata (L.) when applied with IM on day 2 (vs. 20% in control), while NM and AO achieved 100% germination, compared to 70% in CM and control soil. Differential effects on higher root lengths (6.56 ± 0.5 cm in IM; 6.43 ± 1.1 cm in NM) and shoot lengths (5.95 ± 0.9 cm in IM; 6.35 ± 0.1 cm in NM) were observed (Fig. S14; Table S3). A total of 25 phenotypically distinct microbial isolates were obtained from such Jeevamrit formulations. These isolates were further screened for key plant growth-promoting (PGP) traits and identified using 16S rRNA gene sequence similarity. Indole-3-acetic acid (IAA) production varied considerably among isolates, with Shigella spp. RMJ(O)B and NMJ(O)D producing the highest levels (145.4 and 135.66 µg mL^–1, respectively) in LB supplemented medium with L-Trp (1%) as precursor, representing 11- and fivefold increases compared to unamended LB. Similarly, Sphingobium sp. AIJ(O)6 and Shigella sp. JNM(O)6 also exhibited up to 11-fold higher IAA production (Fig. 4a). Furthermore, Shigella sp. JNM(O)6 and Bacillus sp. AIJ(O) displayed elevated NH₃ production (0.12 and 0.11 µg mL⁻¹), followed by Shigella sp. JNM(O)4 (Fig. 4b). Potassium (K) solubilization was most pronounced in Shigella sp. JNM(O)4 and NMJ(O)D, with additional activity in RMJ(O)B and JNM(O)6, while Rhodococcus sp. NMJ(O)C exhibited the highest phosphorus (P) solubilization (55.2 µg mL^–1), accompanied by a pH reduction from 7.0 to 4.5 (Fig. 4c). Hydrogen cyanide (HCN) production was detected only in Enterococcus sp. AIJ(O)B and Bacillus sp. AIJ(O). Enzymatic assays revealed that Shigella sp. RMJ(O)B and Bacillus sp. AIJ(O) showed both protease and cellulase activity, whereas Bacillus sp. RMJ(O)E exhibited amylase and cellulase activity. Based on cumulative PGP potential, the top-ten-performing isolates were identified and belong to Bacillus sp. AIJ(O) and RMJ(O)E, Shigella sp. RMJ(O)B, JNM(O)4, NMJ(O)D, JNM(O)6, NMJ(O)A, Rhodococcus sp. NMJ(O)C, Sphingobium sp. AIJ(O)6, and Enterococcus sp. AIJ(O)B. These isolates were further confirmed by 16S rRNA gene-based phylogenetic analysis, which showed close alignment with OTUs identified from metagenomic datasets, supporting the ecological relevance of cultured strains (Fig. 4d). Specifically, Rhodococcus sp. NMJ(O)C (OTU_58529), enriched in IM treatment, exhibited NH₃ production and P-solubilization but lacked K-solubilization; Sphingobium sp. AIJ(O)6, affiliated with OTUs 48,757 and 41,849 from CM treatment, demonstrated IAA production, P-solubilization, and cellulase activity. Several Shigella isolates and Bacillus sp. AIJ(O) (OTU_32368) that were primarily recovered from AO and NM consistently showed IAA, P- and K-solubilization, and NH₃ production, though their hydrolytic enzyme activity was limited. In contrast, Enterococcus sp. AIJ(O)G (OTU_624), dominant in AO and NM, exhibited IAA synthesis, P-solubilization, and HCN production, highlighting its potential role in nutrient cycling and plant defense promotion (Table S4). BLAST-based homology of isolates the closest affiliation of members Bacillus altitudinis, Bacillus safensis, reported for enhancing plant growth via IAA production, phosphate solubilization, and stress tolerance^45,46. Enterococcus sp. AIJOG showed similarity to E. gallinarum, isolated from the rhizosphere of Lathyrus sativus exhibiting multiple PGPR traits, i.e., IAA production, phosphate solubilization, and antifungal activity⁴⁷. Sphingobium sp. AIJ06 clustered with S. xenophagum, known for IAA production, siderophore synthesis, and stress tolerance, promoting plant growth⁴⁸. The Shigella isolates (JNMO4, JNMO6) are related to S. sonnei and S. flexneri, but there is no strong evidence of PGPR activity for these species (Fig. 4e). Together, these findings demonstrate the functional richness and phylogenetic coherence of Jeevamrit-derived microbes and underscore their potential as plant growth enhancers. Collectively, these findings confirm that cultured isolates are integral constituents of the Jeevamrit microbiome and highlight their role in plant growth promotion and nutrient cycling. The concordance between phylogenetic identity, PGP traits, and treatment-specific abundance further supports their candidacy as bioinoculants for agricultural application.

Fig. 4.

In-vitro assessment of plant growth-promoting (PGP) traits and phylogenetic characterization of cultivable bacterial isolates of Jeevamrit; (a) Quantitative estimation (µg mL⁻¹) of Indole-3-acetic acid (IAA) production by bacterial strains cultured in Luria–Bertani (LB) broth, both in the absence (LB; red bars) and presence of L-tryptophan precursor (LB + Trp; blue bars); (b) Ammonia production (µg mL⁻¹); (c), Phosphate solubilization, where blue bars indicate the concentration of available phosphate (µg mL⁻¹; left y-axis), corresponding changes in the pH of the culture medium were recorded on Day 2 (red line) and Day 5 (blue line) of incubation (right y-axis). A non-inoculated medium was used as a control; (d) Maximum-likelihood-based phylogenetic tree was constructed from the 16S rRNA gene sequences of PGP isolates (indicated by a yellow symbol preceding their names) and similar OTUs present in the OTU data (16S rRNA gene amplicons), illustrating their taxonomic relationships. The key PGP traits of these isolates are depicted with colored circles. Relative abundance of the affiliate OTUs in Jeevamrit samples is presented next to the tree. The scale bar in the tree represents a 0.01 substitution (1%) rate per nucleotide position; (e) Neighbor-joining phylogenetic tree of 16S rRNA gene sequences showing the relationship of the top PGPR isolates (bold) with their closest BLAST hits (regular font). Bootstrap values (> 50%) based on 1000 replications are indicated at branch nodes. Scale bar represents 0.07 substitutions per nucleotide position. Data are mean ± standard deviation (SD) of three replicates.

Discussion

Influence of mixing conditions on physico-chemical properties of Jeevamrit

Jeevamrit is a low-cost, on-farm resource-based formulation, prepared via mixing of cow dung (from an indigenous breed), cow urine, jaggery (molasses), pulse flour (typically chickpea or pigeon pea), and pristine bund soil under aerobic to anoxic conditions, and recommended to be used as a liquid inoculant at a dose of 200 L/acre¹³. Upon combination of these components under different mixing regimes, a plethora of microbial activity/networking and physiological cross-talk is expected, which could be associated with changes in the physicochemical properties of such formulation. This formulation has been reported to exert certain plant and soil beneficial traits, but the details of the microbial, metabolic, and molecular nexus of such properties are explained eventually remain unexplained, thus needing a detailed investigation.

Owing to the on-farm resources-based preparation method of Jeevamrit, it is subject to different mixing conditions/intensity, which contributes to compositional heterogeneity and displays inconsistent effects on crop growth⁴⁹. To address this issue, we investigated the influence of mixing regimes on microbial, metabolic, and nutrient dynamics of the Jeevamritformulation. We observed that mixing conditions resulted in significantly higher levels of TN, Fe²⁺/Fe³⁺, TOC, and several trace elements (Mg, Zn, Mn, Ca), thus highlighting oxidative microbial metabolism and mineral weathering processes favored under higher oxygenation. Elevated levels of ammonium (NH₄⁺), Fe²⁺ (soluble), slightly lower pH, moderate TOC with higher microbial diversity were noted in anoxic set-up, implying the occurrence of respiratory/reductive metabolic processes-mediated nutrient cycling. Significant correlation between soluble organic matter (OM) and physico-chemical properties denoted OM to be a key factor associated with complex interactions with edaphic factors, i.e., change in pH, OC-driven mineral weathering, or metal bioavailability through complexation, chelation, and precipitation, mediated through microbial metabolic processes⁵⁰. Such OM is possibly derived from cow dung, jaggery, and pulse flour (as key ingredients), supplying partially decomposed/fermentable organics (a pool of labile organic C/N) that could stimulate microbial activity and nutrient solubilization⁵¹. Additionally, pristine soil and cow dung can contribute to a metabolically active microbial population, thereby accelerating the overall metabolic channeling of such organics⁵². Negative correlations between NO₃⁻ and NH₄⁺ with TOC, pH, and hardness suggested that elemental transformations are tightly regulated by carbon availability and redox conditions. Microbe-OM interactions are reported to create a positive feedback loop, where OM degradation releases organic ligands that further bind or mobilize metals for redox cycling and microbial assimilation. Oxygen profoundly influences DOM turnover both by directly accelerating its oxidative breakdown and by indirectly modulating redox reactions of associated elements that shape DOM dynamics⁵³. Consistent with these trends, higher solubilization of Fe, Zn, Cu, and Mn occurred under oxygen-rich conditions via oxidative dissolution and enhanced dispersion of metal–organic complexes⁵⁴, whereas elevated soluble Fe and NH₄⁺–N in anoxic/no-mix conditions corroborate earlier studies that attribute such increases to reductive dissolution of Fe(III) oxides and anaerobic ammonification^55,56. Together, these observations indicate that oxidative (high-O₂) and reductive (low-O₂) pathways both promote solubilization, albeit through distinct oxygen-structured biogeochemical mechanisms. Under oxic conditions, microbial metabolism can be strongly influenced by DOM–metal complexation rather than solely by direct metal uptake. Similarly, Wang et al.⁵⁷ showed that humic acid addition increased Cu and Pb mobility while enhancing microbial diversity. These findings align with Reuter et al.⁵⁰, which demonstrated that aerobic degradation of protein-rich substrates supports elevated microbial growth. In contrast, higher microbial richness and diversity (OTUs, Shannon indices) in less-mixed samples (AO, NM) likely result from multipartite interactions among anaerobic and facultative microbes, facilitated by micro-niche formation. The absence of agitation is attributed to preserving the fragile microbial aggregates, enabling the establishment of structured microbial interactions⁵⁹. Under such conditions, OM of soil is reported to form organo-metal/mineral complexes which stabilize aggregates, and in turn regulate moisture retention, modulate pore structure, and oxygen diffusion⁶⁰. Within macroaggregates, limited diffusion can create anaerobic microsites, providing a specialized niche for activity of key anaerobic microbes, thus implying higher diversity in anoxic conditions⁶¹. Availability of OM is directly related to macroaggregate formation, thus influences oxygen diffusion, and thereby alters OM mineralization and DOM dynamics under oxygenic conditions.

Microbial community dynamics of Jeevamrit variants

Mixing-induced shifts in microbes, nutrients, and organic matter–metal interactions highlight the need for metagenomics to reveal microbial guilds associated with DOM turnover and redox activity. Within the community, Acinetobacter, Comamonas, Pseudomonas, and Stenotrophomonas emerged as dominant taxa in oxygenic set-up, exhibited strong positive correlations (Pearson’s r > 0.5) with TN, total Fe, alkalinity, and pH, and are characterized by traits such as heavy-metal tolerance, degradation of complex organic matter, and nitrogen transformation (Fig. S7). Metabolic mapping of such members to their nearest relatives suggested that Acinetobacter spp. (e.g., A. baumannii) utilize siderophores such as acinetobactin and FeO systems for efficient iron uptake, while also contributing to heterotrophic nitrification under low-temperature conditions⁶². Comamonas (C. terrigena) oxidizes Fe²⁺ under nitrate-reducing conditions to Fe³⁺ (as oxides) that co-precipitate metals such as As and Ni, thereby coupling iron and nitrogen cycling⁶³. Pseudomonas spp. demonstrates robust siderophore-mediated acquisition of Fe and Cu, coupled with complex organics and hydrocarbon degradation, coupled to denitrification (narG–nosZ) under micro-oxic conditions^64,65. S.maltophilia reported for high metal tolerance, aromatic hydrocarbon metabolism, and N and S redox-cycling through Fe–S enzymes^66,67. Collectively, these taxa mediate redox-driven mineral dissolution, organic carbon turnover, and nitrogen cycling, highlighting their relevance in aerated Jeevamrit systems. On the other hand, the prevalence of Clostridium sensu stricto, Lactobacillales, Enterococcus, and Enterobacterales was noted in an anoxic set-up. Clostridium spp. are well-documented for performing organic carbon-coupled reduction of nitrate to ammonium in oxygen-depleted environments⁶⁸. Such members are well-described for dissimilatory Fe (III) reduction coupled to oxidation of organic acids as electron donors, e.g., lactate and acetate. Members of Enterobacterales, including Enterobacter and Citrobacter, harbor enzymatic machinery enabling both DNRA and iron respiration, thereby contributing to N-retention and Fe solubilization under anaerobic conditions⁶⁹. This fermentation-driven redox niche is consistent with the observed increase in NH₄⁺ and Fe²⁺ in AO, thus confirming the preponderance of reductive N and Fe pathways. Further, PCA projections along with community composition revealed clear separation between CM and AO, illustrating differential change in physicochemical gradients w.r.t oxygenation, as similarly observed in suspended growth systems with aerobic and anoxic reactors⁷⁰. The study emphasized a higher and adaptive community diversity under fully anoxic conditions with a variety of metabolically diverse groups (denitrifier and phosphate-accumulating), and the community assembly was most stochastic in response to environmental variables compared to its oxic counterpart. Collectively, these findings emphasize the pivotal role of oxygen availability, modulated by mixing regimes, in steering differential microbial assembly and nutrient transformations in Jeevamrit. These shifts in microbial guilds, redox processes, and resulting nutrient signatures under oxic versus anoxic regimes are summarized in (Fig. S15).

Microbial community of Jeevamrit was noted to be densely interconnected with co-occurring networks (|r|> 0.5) [83 nodes, 1320 edges, average node degree = 31.8, network density = 0.388, global clustering coefficient = 0.698] with 708 positive (co-presence) and 612 negative (mutual-exclusion) interactions (Fig. 5a). Among the regulatory hubs (n = 10, as keystone members) formed by Pseudomonadota (n = 37) and Bacillota (n = 34), like Cellulolysiticum, Lactobacillales, Enterococcus, Stenotrophomonas, Comamonas, Trichococcus, which were differentially enriched and responsive to mixing regimes denoting their specific metabolic role, while Acinetobacter, Lactobacillus, and Pseudomonas were the core members (Fig. 5b). Such dynamic recruitment of taxa, i.e., some respond specifically to environmental cues (differential), and a core community (generalist) that persists across conditions, emphasized the synergistic microbial and metabolic dependency in the community linked to specific physico-chemical change (Fig. 5c).

Fig. 5.

Microbial interaction network and differential enrichment of microbial members in samples subjected to differential mixing conditions (oxygenation); (a) Co-occurrence network based on SparCC (Sparse Correlations for Compositional data) analysis displaying robust and significant correlations (|r|≥ 0.5, p < 0.05) considering the top 50 abundant OTUs and differentially enriched abundant OTUs. Node size represents degree centrality, color indicates phylum affiliation, and hub OTUs (key taxa) are shown as square nodes. Blue and red edges denote positive and negative correlations, respectively; (b) Grouped divergence analysis showing differentially abundant taxa in samples. Blue bars indicate taxa significantly enriched under aerobic conditions, while red bars represent those enriched under anaerobic conditions; (c) Z-score normalized heatmap of statistically significant hub taxa across treatments (constant mixing (CM), intermediate mixing (IM), no mixing (NM), and anoxic (AO)), highlighting treatment-specific abundance patterns. Data were scaled by row (taxon) and clustered using hierarchical clustering to reveal co-varying microbial signatures associated with enrichment conditions.

Metabolic versatility and functional roles of community members

A key characteristic of the community is the versatile metabolic repertoire exhibited by its hub taxa. For e.g., Pseudomonas and other Gammaproteobacteria (in frequent mixing) harbored near-complete denitrification genes (narG–nirK–nosZ), reflecting N-responsive redox plasticity under fluctuating oxygen regimes⁷¹. Whereas, Clostridium spp. (both CM and NM) encoded nifH and glnA, supporting their diazotrophic and N assimilatory roles in anaerobic soils/sediment systems⁷². Similarly, Gammaproteobacteria members, Comamonas, Acinetobacter, Dickeya, and Stenotrophomonas have high-affinity phosphate uptake (pstS, phnP) and organophosphonate breakdown, justifying their role in P-acquisition. It is well known that Comamonas phosphati and similar members solubilize phosphate by releasing glycolic acid, rhamnose, and d-melibiose⁷³, and Stenotrophomonas maltophilia can solubilize both inorganic and organic P by acid-based hydrolysis and chelation⁷⁴. The predominance of K-acquisition genes (glt, ppc) in such members further substantiates their role in mineral solubilization, as soil Acinetobacter soli and Comamonas spp. have been reported to efficiently solubilize K-feldspar^75,76. It is interesting to note that Pseudomonas spp. displayed diversity in terms of carbon metabolism, i.e., oxidative central pathways (pyk, pta, pcaF, catA) under mixing and reductive pathways (ppc, ldh) under no-mixing conditions, reflecting adaptive energy generation under fluctuating O₂ concentration. Whereas Comamonas possesses central and aromatic compound metabolic pathways (pyk, ppc, pflB, catA, and protocatechuate catabolic genes), which justify lignin/complex organic matter oxidation in the system⁷⁷. On the contrary, Acinetobacter with lactate dehydrogenases, Dickeya for adhE and pflB, and Clostridium with fermentative dehydrogenases (adhE, dhaB, pflB, ldh) are ascribed for the production of mixed acids and alcohol (ethanol, butanol, lactate, etc.)^78,79. Abundance of hydrolytic enzymes (cellulases, xylanases, and other polysaccharide-degrading enzymes (chiA, amyA, celA, xynA) in Clostridial members (C. thermocellum, C. stercorarium) under cellulose-rich, oxygen-limited conditions further justified their role in mediating fermentative metabolism of such substrates and production of organic acids⁸⁰. The release of these organic acids not only reflects active carbon metabolism but also creates a favorable biochemical environment for metal solubilization, which was conserved across taxa, with Pseudomonas (frpA–entS–tonB), Gammaproteobacteria (tonB–fepA–entS–fepC), and Clostridium (fepC–entS) actively scavenging Fe. Pseudomonas uses TonB-dependent transporters (TBDTs) such as frpA and fpvA for pyoverdine and pyochelin uptake⁸¹, Acinetobacter produces acinetobactin with entA essential for siderophore function⁸², and Stenotrophomonas maltophilia synthesizes ferric enterobactin via entC⁸³. Such observation was well correlated to enhanced metal solubility in the setup. We noted that Auxin (IAA) biosynthesis varied with mixing, majorly mediated by Pseudomonas (P. syringae, P. savastanoi) via iaaM and iaaH⁸⁴, while Comamonas (C. testosteroni) likely employs analogous or alternative pathways to modulate root architecture and plant growth⁸⁵. Collectively, these patterns indicate that Jeevamrit’s core microbiome is endowed with varied N redox-transformation and assimilation pathways in combination with polyphosphate metabolism, organic carbon oxidation, polymer hydrolysis, and acid production, correlating to the observed physico-chemical changes of the formulation. In addition, some members were identified as key network components (degree ≥ 71) that showed functional dependencies with hubs, forming cooperative metabolic guilds to bridge aerobic–anaerobic niches and support nutrient cycling. For example, OTU_28_DEO (Stenotrophomonas), predominantly S. maltophilia, contributed to aerobic oxidation of complex organic matter and nutrient assimilation, while also exhibiting phosphate solubilization, siderophore/IAA production, and suppression of soil-borne pathogens. Enterococcus and Lactobacillales collectively represented fermentative, lactic acid-producing lineages that adapt to oxygen-limited micro-niches and contribute to organic matter turnover⁸⁶. Similarly, Paludibacteraceae and Ruminococcaceae OTUs likely contributed to anaerobic saccharolysis, fermenting plant-derived polysaccharides into short-chain fatty acids (propionate and acetate) and thereby driving fermentative carbon cycling under anoxic conditions⁸⁷. Together with Trichococcus and Lactobacillales, these members are attributed to building the anaerobic metabolic guild under oxygen-depleting conditions. Recovery of MAGs for Acinetobacter sp. NCu2D-2, Klebsiella pneumoniae Kp_03, Trichococcus flocculiformis, Clostridium saccharobutylicum BAS/B3/SW/136, and Enterococcus gallinarum K205-4a revalidated their roles as core functional members of the Jeevamrit consortium. Comparative genomic analyses of Acinetobacter strains revealed high genomic plasticity, including diverse mechanisms for nutrient acquisition and the presence of biosynthetic gene clusters encoding secondary metabolites such as aryl polyenes, β-lactones, and siderophores. C. saccharobutylicum and T. flocculiformis cooperatively mediate polysaccharide breakdown, enhancing short-chain fatty acid flux, and ABE fermentation⁸⁸. E. gallinarum K205-4a encodes bacteriocins and stress response genes, aiding OM turnover and antagonism⁸⁹, which could help in stabilizing the community. This alignment of community profiling and MAG recovery identifies these taxa as key drivers of decomposition, nutrient cycling, and microbial interaction in Jeevamrit. The metabolite–microbe associations observed further substantiate the fact that Jeevamrit undergoes a structured metabolic succession shaped by oxygen availability through key microbial players pathways, which are hallmarks of aerobic lipid and steroid catabolism. As oxygen levels decline, the microbial community shifts toward fermentative specialists such as Clostridium sensu stricto and Trichococcus, which are well‐documented for their roles in saccharolytic fermentation and syntrophic interactions^90,91. Collectively, these findings suggest the occurrence of synergistic microbial-metabolic interactions to stabilize the community’s core functions, reinforcing the structural stability of the microbial network in Jeevamrit. Whereas specific responsive taxa (differentially recruited) are necessary for rendering metabolic/functional specificity under contrasting mixing regimes.

Cultivable microbes and their plant growth-promoting traits

25 bacterial isolates, Bacillus sp. AIJ(O) and RMJ(O)E, Shigella sp. RMJ(O)B, JNM(O)4, NMJ(O)D, JNM(O)6, NMJ(O)A, Rhodococcus sp. NMJ(O)C, Sphingobium sp. AIJ(O)6, and Enterococcus sp. AIJ(O)B-exhibited the most potent plant growth-promoting (PGP) properties. Bacillus sp. AIJ(O) exhibited broad PGPR activity with strong nutrient solubilization, NH₃ production, HCN generation, and protease activity, while Bacillus sp. RMJ(O)E was notable for high extracellular enzyme activity, especially protease and cellulase. Bacillus spp., dominant in CM and IM formulations, are core rhizosphere inhabitants whose ecological adaptability and endospore formation ensure survival under variable soil conditions, supporting their consistency as bioinoculants. Many Bacillus spp. (e.g., B. pumilus, B. megaterium, B. subtilis) harbor nifH and other N-responsive genes, enabling biological N₂ fixation. Members of the B. subtilis and B. cereus lineages produce IAA via Trp-dependent pathways, stimulating root growth⁹², solubilize phosphate through organic acids, secrete hydrolytic enzymes for organic matter turnover and pathogen suppression⁹³, and release HCN and siderophores to enhance Fe acquisition to limit pathogen growth⁹⁴. The genomic plasticity of Bacillus, mediated by global regulators, σ-factors, and BGCs, enables coordinated PGP trait expression, making it a versatile and dominant bioinoculant for agricultural applications. Shigella isolates, though less studied as PGPR, demonstrated IAA synthesis in the presence of L-Trp, alongside efficient NH₃ release and P and K solubilization, underscoring their potential functional relevance in the rhizosphere. Rhodococcus sp. NMJ(O)C solubilizes phosphate and produces NH₃ via secretion of organic acids (gluconic, citric, acetic, formic) that chelate cations and acidify the microenvironment, while its nitrogen-metabolizing enzymes assimilate and deaminate amino acids, amides, and nitriles to release ammonia, enhancing rhizospheric N availability⁹⁵. Enterococcus sp. AIJ(O)B exhibits a multifunctional PGP profile IAA synthesis via Trp-dependent pathways (as in E. faecium LKE12), producing 4.5 µg mL⁻¹IAA and promoting root elongation in Arabidopsis⁹⁶. Phosphate solubilization through organic acid–mediated cation chelation and pH reduction, and HCN production, while its facultative anaerobic metabolism, broad stress tolerance, and robust catabolic systems support adaptation to diverse rhizosphere conditions, reinforcing its role as a versatile PGP microbe. Sphingobium sp. AIJ(O)6 exhibits IAA production, phosphate solubilization, and cellulose degradation, underpinned by its enriched genomic and metabolic repertoire. Comparative genomics of Sphingobium sp. AEW4 revealed genes for Trp-dependent IAA biosynthesis and an expanded set of carbohydrate-active enzymes, including endoglucanases and xylanases, enabling cellulose and hemicellulose breakdown. Phosphate solubilization is mediated via organic acid biosynthesis and acid phosphatases, mobilizing mineral-bound phosphorus under low-pH conditions⁹⁷, collectively supporting nutrient cycling and carbon turnover in the rhizosphere. Metagenomic and culture-based studies validated such observed results, showing dominance of Pseudomonadota and Bacillota (Bacillus, Priestia, Pseudomonas, Rhizobium/Bradyrhizobium, Paenibacillus, Streptomyces, Sphingomonas, Azotobacter, and Rhodococcus) as cultivable representatives harboring multiple plant growth-promoting traits. Culture-based work frequently recovers Bacillus, Priestia, and Rhodococcus as dominant cultivable taxa, while 16S and metagenomic analyses highlight Pseudomonas, Rhizobium, and actinobacterial groups associated with nitrogen fixation, phosphate solubilization, and pathogen suppression. Although exact proportions may vary with feedstock inputs, preparation method, and local seedbank, convergence across independent studies supports the view that Jeevamrit contains a consistent consortium of beneficial bacteria underpinning its crop growth benefits^98,99. To contextualize these assignments, certain lineages particularly Shigella-like Enterobacteriaceae and Enterococcus require clarification. It is important to note that 16S rRNA-based taxonomy often overestimates Shigella due to its > 99% sequence identity with Escherichia coli, resulting in frequent misclassification of environmental E. coli strains as “Shigella”¹⁰⁰. Also, non-pathogenic Enterococcus lineages are commonly being detected in soil, compost, and organic amendments without posing clinical risk⁸⁶. However, considering the established biofertilizer guidelines, safe field application should include biosafety screening or simple mitigation steps like thermophilic or alkaline treatment which effective could reduce such populations. Overall, our findings indicate that Jeevamrit harbors metabolically versatile microbial communities whose functions are strongly shaped by mixing regimes, which in turn facilitates the diversity of plant growth-promoting activities, thus implicated for better nutrient cycling, and crop productivity in natural farming settings.

Conclusion

This study demonstrates that the aeration/mixing regime profoundly influences the microbial composition and metabolic repertoire, which in turn affects the overall physico-chemistry of the Jeevamrit formulation. Metagenomic analyses highlighted aerobic, organic matter oxidizing, mineral-solubilizing taxa (e.g., Pseudomonas, Acinetobacter) enriched with siderophore, IAA, and phosphate-solubilization genes for continuous mixing (CM) preparation. Whereas static or anoxic conditions (NM, AO) favored obligatory and facultative fermenters (e.g., Clostridia, Enterococcus) that were linked to diverse fermentation metabolites and ammonium formation, reflecting anaerobic N‐mineralization pathway. Intermediate mixing (IM) showed a balanced trend, combining nutrient release with microbial diversity, and showed a positive effect on seed germination of the mung bean. The culture-based study showed dominance of Pseudomonadota and Bacillota, having multiple plant beneficial traits that were also corroborated by metagenomic data. It is important to note that oxygen supply/mixing modulates the Jeevamrit functional guilds: intense aeration is associated with oxidative solubilization, whereas static fermentation enhances reductive metabolism. The experimental results suggested that via tailoring oxygenation extent (mixing intensity) of Jeevamrit preparation, the properties could be fine-tuned, i.e., moderate agitation yields a balanced metabolites and biofertilizer potential, while vigorous mixing maximizes nutrient solubilization. Ultimately findings of this work provide a framework for further optimizing Jeevamrit formulation to improve key microbial and metabolic signatures that could enhance soil fertility and crop growth.

While this study provides mechanistic insights into how mixing-driven oxygenation shapes the microbial composition, metabolic potential, and nutrient dynamics of Jeevamrit, several limitations should be acknowledged. The experiments were conducted under controlled laboratory-scale conditions using a single preparation cycle per mixing regime; thus, the consistency of these microbial and functional patterns across repeated preparations, seasons, and farm-scale settings remains to be established. Although metagenomic and culturomic analyses revealed functional potential and representative plant growth–promoting traits, direct assessment of in-situ nutrient fluxes and long-term soil or crop responses under field conditions was beyond the scope of this study. In addition, metabolomic profiling was restricted to water-extractable fractions and may not capture transient or volatile metabolites relevant after soil application. Furthermore, the study employed a single ingredient composition and incubation duration; variations in substrate quality, climatic conditions, and application rates in farmer-managed systems may further influence Jeevamrit performance. Future studies should therefore focus on multi-batch validation, field-scale evaluations across diverse agro-ecological zones, and integration of process-based nutrient flux measurements to improve reproducibility and agronomic relevance.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

AO

Anoxic

CM

Constant mixing

IM

Intermediate mixing

NM

No mixing

ZBNF

Zero-budget natural farming

SOC

Soil organic carbon

TOC

Total organic carbon

DOC

Dissolved organic carbon

DOM

Dissolved organic matter

TN

Total nitrogen

ORP

Oxidation–reduction potential

EC

Electrical conductivity

DO

Dissolved oxygen

TDS

Total dissolved solids

OTU

Operational taxonomic unit

MAG

Metagenome-assembled genome

PGP

Plant growth-promoting

PGPR

Plant growth-promoting rhizobacteria

IAA

Indole-3-acetic acid

LC–MS

Liquid chromatography–mass spectrometry

ESI

Heated electrospray ionization

LLE

Liquid–liquid extraction

PCA

Principal component analysis

CCA

Canonical correspondence analysis

PLS-DA

Partial least squares-discriminant analysis

VIP

Variable importance in projection

COG

Clusters of orthologous groups

KEGG

Kyoto Encyclopedia of Genes and Genomes

ORF

Open reading frame

WGS

Whole-genome sequencing

CFU

Colony-forming unit

HCN

Hydrogen cyanide

SI

Solubilization index

LB

Luria–Bertani

CTAB

Cetyltrimethylammonium bromide

SS-NDIR

Solid-state non-dispersive infrared

Author contributions

Ayush G. Jain: Investigation, data curation, formal analysis, visualization, writing-original draft, revision. Draksha Agwan: Investigation, methodology, data curation, writing-original draft, revision. Ashutosh Kumar: Bioinformatics analysis and validation. Imran Pancha: Conceptualization, review, and editing. Jagat Rathod: Conceptualization, supervision, project administration, funding acquisition, and writing-review & editing. Balaram Mohapatra: Conceptualization, supervision, project administration, funding acquisition, and writing-review & editing.

Funding

We gratefully acknowledge the financial support provided by Gujarat State Biotechnology Mission (GSBTM), Department of Science and Technology, Government of Gujarat, India (Grant ID: GSBTM/JD(R&D)/661/2022-23/00173054) under the “Network Research Program on Natural Farming through Biotechnology Interventions”.

Data availability

The 16S rRNA amplicon sequencing and whole genome shotgun (WGS) sequencing of Jeevamrit samples were performed using the Illumina NovaSeq X platform. 16S amplicon sequence data are available in the NCBI SRA under BioProject: PRJNA1156891 with (run accessions: SRR30573539–SRR3057354). Shotgun metagenomic data have been deposited under BioProject PRJNA1269043 (run accessions: SRR33746293 and SRR33746294). Additionally, bacterial isolates were identified by almost complete-length 16S rRNA gene sequencing, and the sequences are available in GenBank under the accession numbers: PV342468–PV342477.

Declarations

Competing interests

The authors declare no competing interests.