Sample collection for the isolation of endophytic microorganisms was done at an altitude ranging from 1500 to 2200 meters above mean sea level. The roots, stems and leaves of disease-free and healthy rice plants (10 from each variety) in disease-affected fields were randomly selected at the panicle initiation stage (active phase) from four varieties of rice (C1039, SR3, SR4 and SR5). Paper bags were used to transport samples to the laboratory.

To isolate endophytic microorganisms, plant tissues were washed under running tap water to remove soil residues and then surface sterilized with 70% alcohol and 5% sodium hypochlorite for 5 min each, followed by three washes with sterile distilled water. For purification, well-established endophytic fungal and bacterial colonies were removed from plates and sub-cultured in test tubes [14].

Dual culture [15] was used to test the antagonism of endophytic bacterial and fungal isolates with diffusible compound(s) against Pyricularia oryzae. In the case of endophytic fungi, a 5 mm mycelial disc of pathogenic fungi was put on one side of the potato dextrose agar (PDA) plate, and an endophytic fungal disc was put on the other side. In the case of bacterial endophytes, one loopful of 48-hour-old culture was streaked on one side of the plate, and a 5 mm mycelial disc of pathogenic fungi was placed on the other side of the plate with PDA and nutrient agar (NA) in a 2:1 ratio [8]. The plates were put in a biological oxygen demand (BOD) incubator for 20 days at 251°C for fungi and 281°C for bacteria. The experiment was done using a complete randomised design with three replications. The percentage of inhibition of mycelial growth over control was calculated using the formula below. The isolates showing potent antifungal activity were selected for further studies.

Percent inhibition = (C − T/C) 100, where C represents the radial growth of the test pathogen in the control plates (measured in mm), and T is the radial growth of the test pathogen in the test plates (mm) [16].

The volatile compound(s) (VOCs) production of the superior antagonists was screened according to Dennis and Webster (16). A 5 mm disc of P. oryzae and the fungal endophyte were placed at the centre of two different Petri plates containing PDA. In the case of bacterial endophytes, on one plate, a 5 mm disc of P. oryzae was placed at the centre of the Petri plate, and the bacterial endophytes were streaked on another Petri plate containing NA. The lids were then removed and replaced by the bottom of the Petri plate containing the testing endophytic isolates. The two dishes were then taped with parafilm. On the control plate, no bioagent was inoculated but taped in the same manner. The plates were incubated at 251°C in a BOD incubator for 20 days. The percent inhibition was measured by the formula mentioned above.

A loopful of endophytic bacteria and a mycelial disc of endophytic fungus were added to 3.9 mL of Pikovskaya’s broth medium [17]. The slants were kept in an incubator shaker and incubated at 28°C with shaking at 200 rpm for 3 days for bacterial endophytes and 5 days for fungal endophytes. After that, 1 mL of culture was centrifuged at 14,000 rpm for 10 min, and the soluble P in the supernatant was determined with the method of Bray et al. [18]. The amount of soluble P was calculated from the standard curve of P concentration using KH₂PO₄.

The production of IAA was demonstrated by the method of Bano et al. [19]. The fungal endophytes were inoculated in the PDA amended with 2 mg mL⁻¹ of L-tryptophan (Trp) and incubated at 120 rpm, 25°C, for 5 to 7 days, whereas the bacterial endophytes were grown in Luria broth amended with 2 mg mL⁻¹ of L-tryptophan (Trp) and incubated at 120 rpm, 25°C, for 3 days. The broth was then centrifugated at 10,000 rpm for 10 min, and 1 mL of the supernatant was mixed with 2 mL of Salkowski reagent. Pink colour production indicated IAA production. For the quantitative assay, the above mixture was incubated for 20 min, and a reading was taken at 530 nm.

In a 250-mL Erlenmeyer flask, 100 mL of nutrient broth (NB) and potato dextrose broth (PDB) were sterilised, and 1 mL of standard inoculum (10⁹ cells mL⁻¹) of endophytic isolates was added, which was shaken at 30°C. The experiment was conducted according to standard techniques [20]. The absorbance at 254 nm was measured using a spectrophotometer.

The fungal and bacterial endophytes were grown in 10 mL of peptone broth and kept in an incubator shaker (120 rpm, 25°C ± 1°C) for 4–5 days. A volume of 0.5 mL of Nessler’s reagent was added to each tube. The development of a brown-to-yellow color indicated ammonia production. The liberated ammonia was determined spectrophotometrically at 450 nm [21].

Colloidal chitin was prepared by the method adopted by Prabavathy et al. [22]. The bacterial and fungal endophytes were grown in the slants containing colloidal chitin media for 3 days and 5 days, respectively. A colorimetric chitinase enzymatic assay was conducted according to Rojas-Avelizapain triplicates [23]. The absorbance was measured at 535 nm. One unit (U) of chitinase activity was expressed as the amount of enzyme required to produce 1 mol of N-acetyl glucosamine (Sigma assay kit) per hour under the above conditions [23].

Siderophore production was assayed by the method given by Schwyn et al. [24]. In this assay, CAS media was prepared by dissolving 60.5 mg CAS (Sigma Aldrich, USA) in 50 mL of water and mixing it with 10 mL of iron (III) solution (1 mM FeCl₃•6H₂O and 10 mM HCl). This solution was slowly added to 72.9 mg of HDTMA dissolved in 40 mL of water. Also, minimal media were prepared, and both were autoclaved separately. The colonies of bacterial and fungal endophytes were then added to the broth. The reference was also made to the reference, which had exactly the same components but lacked the endophyte. The cultures were incubated in an incubator shaker at 28°C for 3–5 days. After that, the absorbance was measured at 630 nm. The experiment was carried out in three replications.

The bacterial endophytic cultures were streaked on pre-poured plates of King’s medium B amended with 4.4 gL⁻¹ glycine, while the fungal endophytic cultures were grown on PDA amended with 4.4 gL⁻¹ glycine. The filter paper (Whatman No. 1) strips were saturated in 0.5% picric acid and 0.2% sodium carbonate. Now the strips were placed in the center of the Petri plates. The parafilm-sealed Petri plates were transferred into an incubator at 28°C for 1–5 days. The control (uninoculated plate) was used for comparison. The experimental plates were analyzed through observation of the change in color of the filter paper from yellow to orange-brown or dark brown [25].

The cultures of endophytic fungi were morphologically identified by observing their morphology, both macro- and micro- characteristics. The macro morphological characteristics tested included the colour of the colony, the texture, and the surface of the colony. The microscopic morphological observations were made by forming a smear of the different isolates of endophytic fungi on the glass slide cover and staining it with lactophenol cotton blue. The isolates were examined under a light microscope. The result of morphological identification (macro and micromorphological characteristics) was compared with the reference available [26]. The colony morphology of bacterial endophytes was studied on plates after streaking a loopful of the isolated colony, and different morphological characteristics like shape, size, texture, margin, and colour were observed. The shape of a single colony was observed under a light microscope. The isolates of Bacteria were gramme-stained. The bacterial smear was prepared using isolated colonies and stained with Gramme’s stain.

The isolates of endophytes that showed antifungal properties of the tested pathogens were cultured and selected at 27°C for 8 days using PDB (Himedia). The mycelial mat was separated from the PDB and freeze-dried. The mats were quenched in liquid nitrogen, and the genomic DNA was isolated by the Cetyl Trimethyl Ammonium Bromide (CTAB) procedure with some modifications following the protocol of Sambrook et al. [27]. Similarly, for bacterial endophytes, selected endophytic bacterial isolates were cultured for 24 h at 28°C in 10 mL of Luria broth medium, and DNA was extracted following the standard protocol [28]. The DNA extracted was separated on a 0.8 percent (w/v) agarose gel in 1 TAE buffer for 30 min and visualised under UV light by staining with ethidium bromide. The extracted DNA was amplified using ITS1 (5′-TCCGTAGGTGAACCTGCGG-3′) and ITS4 (5′-TCCTCCGCTTATTGATATGC-3′) primers for fungi and 27F (5′-AGAGTTTGATCTTGGCTCAG-3′) and 1492R (5′-TACGGTTACCTTGTTACGACTT-3′) for bacteria [29,30] in a thermocycler. The PCR products were run on electrophoresis along with a molecular ladder in a 1.5% agarose gel.

The PCR products of partial 16S rRNA and ITS regions of the gene were custom sequenced by Biokart, Bangalore, India. Consensus sequences were constructed in DNA Baser version 5.15. Before, manual editing was conducted, and then pairwise alignment was accomplished by BioEdit software version 7.2. The similarity of these sequences was checked with the already present sequences in the National Centre for Biotechnology Information (NCBI) gene bank database using the Basic Local Alignment Search Tool (BLASTn) program. The isolates were identified through the observation of a percentage of sequence identity (90%) with that of a prototype strain sequence in the GenBank [31]. The consensus sequences were deposited in GenBank using the BankIt tool of the NCBI.

The retrieved sequences were subjected to BLAST analysis after sequencing to check for the best hits. The consensus sequences were submitted to GenBank. A neighbour-joining phylogenetic study was performed by MEGA v. 7 using generated sequences [32]. Bootstrap analysis was done with 10,000 replicate run data sets, and the sequence identity matrix was done using BioEdit version 7 software.

A greenhouse experiment was performed to investigate the efficacy of endophytes against P. oryzae using a completely randomised design. The experiment was performed in pots containing sterile soil. The soil and vermicompost were mixed in a ratio of 2:1 (w/w) and sterilised by autoclaving at 121°C at 15 psi for two hours for three days. The susceptible variety of P. oryzae Mushkbudji was used. The seeds were soaked in the 30 mL spore suspension of potential bacterial (10⁸ cfu mL⁻¹) and fungal (10⁶ spores mL⁻¹) endophytes for 24 h. At the planting stage, additional 5 mL inoculums were added to the inoculated seeds in the pots. The pathogen was added 7 days after the inoculation of endophytes. Untreated rice seeds were planted in control pots. After germination, each container had five plants. The containers were watered with sterile water to keep the soil wet. Samples from rice plants were obtained 30 days after sowing (DAS) for assessing various plant biometric characteristics. Disease incidence and disease intensity (%) were evaluated using the Standard Evaluation System (1–9 scale) (IRRI, 1980). The lengths of the shoot and root, as well as the fresh and dry weights, were measured [33].

The data from the experiments were subjected to statistical analysis as per the methods detailed by Gomez et al. [34]. A one-way analysis of variance (ANOVA) was carried out using OPSTAT online software. Molecular data analysis was carried out by BioEdit software version 7.2 and MEGA software version 11.0 [32].

A total of 125 endophytic microorganisms were isolated from the leaves, stems, and roots of the four different varieties of rice grown in the Kashmir Himalayas, out of which 100 were bacterial endophytes and 25 were fungal endophytes. It was found that the highest number of endophytes were isolated from roots, followed by stems and leaves (Table 1). In particular, 48% of bacterial endophytes were isolated from the root. Whereas the endophytes isolated from stem and leaf were 35% and 20% of the total number of isolated endophytes, respectively. The highest number of 14 endophytes was isolated from SR4 (36) followed by C1039 (32) and SR5 (31) while the fewest endophytes were isolated from the SR3 variety of rice.

The isolates were evaluated for antagonistic activity against P. oryzae using the dual culture method. Five isolates among bacterial endophytes and five isolates among fungal endophytes showed antagonistic activity against pathogens with percent inhibition greater than 50%. It was found that in the case of bacterial endophytes, isolates b18 and b7 had the highest percent inhibition of 71.56% and 70.43% respectively, which were statistically significant (p < 0.05), followed by b17 (67.83%), b55 (62.93%), and b9 (61.23%). However, the lowest inhibition was found by the isolate b45 (8.46%) (Table 2; Figs. 1A, 1B). In the case of fungal endophytes, the highest inhibition was found by f18 with a percent inhibition of 90.46% which was significant at p < 0.05 followed by f17 (87.39%). However, the lowest inhibition was found by f3 (5.77%) (Table 3; Figs. 1A, 1C). The endophytes that showed greater inhibition in dual culture tests were used for further studies.

All the isolates produced volatile compounds that inhibit the growth of the pathogen. Among the bacterial isolates, b18 showed the highest inhibition of 65.6% followed by b7 and b17 with percent inhibitions of 61.91% and 58.41%, respectively (Table 4). Among the fungal isolates, the highest inhibition was shown by f18 (77.74%) followed by f17 (68.58%), and f11 (67.58%) (Table 4).

The phosphorus solubilization test revealed that in the case of fungal endophytes, phosphorous solubilization varied from 34.28 (f12) to 88.44 g mL⁻¹ (f18). On the other hand, in the case of bacterial isolates, it varied from 48.24 (b7) to 70.73 g mL⁻¹ (b55) (Tables 5, 6).

All the bacterial and fungal isolates were able to produce auxin as well as gibberellic acid. The auxin production varied from 7.14 (g mL⁻¹) to 97.77 (g mL⁻¹). Among the bacterial endophytes, the highest auxin production was shown by isolate b55 (97.77 µgml⁻¹) (significant at p < 0.05), followed by isolate b17 (84.56 µgml⁻¹). The lowest auxin production was shown by isolate b18 (14.46 µgml⁻¹). Among the fungal isolates, the highest production was shown by f18 (significant at p < 0.05), which showed auxin production of 35.83 µgml⁻¹ followed by f12 (17.37 µgml⁻¹), whereas the lowest auxin production was exhibited by f13 (7.14 µgml⁻¹). In the case of gibberellic acid (GA) production, isolate b7 showed the highest GA production of 42.19 µgml⁻¹ (significant at p < 0.05), and the lowest production was shown by isolate b9 (6.32 µgml⁻¹) among bacterial isolates. Among fungal isolates, the highest production was shown by isolate f18 (48.98 µgml⁻¹), followed by isolate f13 (44.37 µgml⁻¹), whereas the lowest production was shown by isolate f12 (1.26 µgml⁻¹) (Tables 5, 6).

Among the five bacterial isolates, four were confirmed to produce siderophores by CAS assay, while, only three were able to produce siderophores among the fungal isolates. The isolates that produce siderophores change the colour of the media from blue to orange. In the case of bacteria, the isolate b18 showed the highest siderophore production of 52.5 M mL⁻¹, whereas in the case of fungi, the isolate f18 produced 56.64 M mL⁻¹ of siderophore (Tables 5, 6). Among all the ten isolates tested for HCN production, only two bacterial isolates were able to produce HCN. The isolate b9 produced HCN of 36.15 µgml⁻¹, while b17 produced HCN of 14.4 µgml⁻¹ (Tables 5, 6).

All the bacterial isolates were able to produce ammonia, varying from 83.42 µgml⁻¹ (b55) to 12.78 µgml⁻¹ (b18). However, only three fungal isolates were able to produce ammonia, where the highest production was found by isolate f12 (37.99 µgml⁻¹) followed by f18 (35.39 µgml⁻¹) and f13 (27.57 µgml⁻¹) (Tables 5, 6). Also, some isolates were able to produce chitinase. The isolate b18 had the highest chitinase activity of 9.54 units/ml, followed by b7 (8.76 units/ml) and b17 (3.97 units/ml). In the case of fungi, only two fungi were able to produce the chitinase enzyme. The isolate f18 showed chitinase production of 13.76 units/ml, and f13 showed a production of 2.34 units/ml (Tables 5, 6). The isolates were also tested for nitrogen fixation by using Jensen’s media, and it was confirmed that all the bacterial isolates were able to fix the atmospheric nitrogen by forming glistering colonies, while none of the fungal isolates were able to fix the atmospheric nitrogen.

The morphological characteristics of bacterial endophytes revealed that b7 and b9 were gram-negative, while b18, b55, and b17 were gramme-positive in gramme staining. The colony morphology of the bacterial isolates varied from smooth and round in the case of b7 to irregular and rough in b18, while b17 has granular and wavey margins. The colony’s colour varied from white to greyish white. All the bacterial endophytes were rod-shaped (Supplementary Table 1). In the case of fungal isolates, the colour of the colony was initially white, which later turned dark olivaceous green in the cases of f11 and f17 (Supplementary Table 2). The mycelium is abundant, cottony, dark olive green in colour, and aerial. The conidia are globose to subglobose with an average size of 5.21-3.14.13-2.17 m (av. 4.033.11 m) and 4.27-2.783.87-2.17 m (av. 3.783.18 m), respectively. The average width of mycelium was found to be 5.524.3 m (av. 4.96 m) and 5.633.66 m (av. 4.85 m), respectively. In f12, the mycelium is floccose, tufted green, fast-growing, and sporulates abundantly. The edge of the colony was smooth and spreading. The conidia were light green in colour and globose to subglobose in shape, with an average size of 3.88–2.7–2.97–1.61 (av. 2.97–2.51) m and an average mycelial width of 4.94–3.64 (av. 4.59) m. In the case of f18, the colony was a powdery mass with a yellow-green colour and abundant sporulation. Usually, the margins of the colony were white. The conidia were globose in shape, with an average size of 4.67-3.85.61-4.00 (av. 4.873.92) m and a mycelial width of 8.644.74 (av. 6.53) m. In the case of f13, the colonies were bluish-green in colour with a white periphery and were radially sulcate. The globose to subglobose with an average size of 4.89–3.95–4.45–3.18 (av. 3.84–3.78) m and mycelial width of about 4.97–3.26 (av. 3.74) m. The phialides were flask-shaped, and their size was also measured, which ranged from 4.3-9.53.1-4.5 m (av. 5.83.2) m in the case of f12 to 5.2-7.84.6-8.2 m (av. 6.87.5 m) in f13 (Supplementary Table 2).

To confirm the identity of potential bacterial and fungal endophytes, the 16S rRNA gene in the case of bacteria and the ITS region in the case of fungal endophytes were amplified using PCR. The 16S rRNA gene was amplified from all five bacterial endophytes, and an amplicon of 1200 bp was obtained after PCR amplification. Similarly, the ITS region was amplified from all five fungal endophytes, and an amplicon of 750 bp was obtained.

Both the fragments from bacteria and fungi were sequenced and analysed. All the sequences generated were submitted to NCBI GenBank, and accession numbers were received. Detailed illustrations of these bacterial and fungal endophytes and their accession numbers, percentage similarity to the closest species, and genetic identity based on BLASTn searches are given in Supplementary Table 3. The results revealed that the endophytes were identified as: Trichoderma afroharzianum (f11 and f17), Trichoderma harzianum (f12), Penicillium rubens (f13), Aspergillus flavus (f18), Stenotrophomonas rhizophila (b7), Stenotrophomonas maltophilia (b9), Bacillus cereus (b17), Enterobacter cloacae (b55), and Bacillus licheniformis (b18). The percent similarity of endophytes varied from 93 percent in the case of Stenotrophomonas maltophilia to 100 percent in the cases of Trichoderma harzianum and Aspergillus flavus. The sequences of isolated bacterial and fungal endophytes were compared with sequences of other bacterial and fungal species, respectively, using a similarity of above 99% retrieved from GenBank. MEGA 11 software was used to analyse and construct phylogenetic relationships between the endophytes and their related fungi and bacteria. The evolutionary history was inferred using the neighbour-joining method. In the bacterial phylogenetic tree, it was evident that two broad clusters were formed, with Bacillus licheniformis (B. licheniformis), Bacillus cereus (B. cereus), and Stenotrophomonas rhizophila (S. rhizophila), S. maltophilia, and E. coaleae in another cluster (Supplementary Fig. 1). All the isolates were grouped into two phyla, viz., Firmicutes and Proteobacteria. In the case of a fungal phylogenetic tree, the tree divides into two broad clusters. T. afroharzianum and T. harzianum were in one cluster, and P. crysogenum and A. flavus were in a separate cluster. Both clusters further subdivide into two more clusters. All the fungal endophytes fall under the division Ascomycota.

The sequence identity matrix of our endophytes with other endophytes in the NCBI gene bank revealed that similarity ranged between 23% and 100%. Most of the bacterial endophytes were similar to endophytes from India and China. However, Stentrophomonas rhizopila was 93.5% similar to isolates from Iran and China (MZ895036 and EU135790). Similarly, Bacillus licheniformis was 96% similar to isolates from Denmark and South Korea (CP045814 and CP042252). S. maltophila was 93% similar to Indian isolates (MW282918 and HE963840). Furthermore, Enterobacter cloacae were 96% similar to isolates obtained from China and France (KX431213 and KJ49995). Bacillus cereus was 99% similar to isolates from India (KC68396 and HQ388817) (Supplementary Fig. 1). In the case of fungal endophytes, the sequence identity matrix revealed that T. harzianum showed 100% similarity with the isolates from China (MT102390) and Hungary (MN516457), and T. afroharzianum showed 100% similarity with isolates from Brazil (KC576738) and the USA (KC007211). P. chrysogenum showed 99% similarity with isolates from the USA (MN644620) and South Africa (MN413181). Similarly, A. flavus was showing 100% similarity with the isolates of Cassava (KC994648) and France (KC964101) (Supplementary Fig. 2).

The highest disease control was observed by A. flavus with a percent inhibition of 79.43%, followed by T. afroharzianum with an inhibition of 71.63% (Table 7). Among bacterial endophytes, the highest inhibition was found by B. licheniformis with an inhibition of 78.89% and was statistically significant at p < 0.05, followed by S. rhizophila with an inhibition of 77% (Table 7). The plants were also evaluated for growth parameters, and the results indicate that all the isolates had a positive impact on considered plant growth parameters like elongation of the shoot (38.64–31.13 cm) and root (32.35–21.42 cm) when compared with the control seedling treated with only distilled water. Among bacterial isolates, the isolate E. cloacae showed a significant increase in root length (28.94 cm), shoot length (38.64 cm), fresh root weight (4.95 g), fresh shoot weight (5.60 g), dry root weight (0.6 g), and dry shoot weight (1.29 g) over the control at p < 0.05 (Table 8). In the case of fungal isolates, the f18 showed the highest root growth (32.35 cm), shoot growth (37.23 g), dry root weight (0.67 g), dry shoot weight (1.01 g), fresh root weight (5.91 g), and fresh shoot weight (4.87 g) at p < 0.05 (Table 8).