Introduction

Every large multicellular organism studied so far, including plants, interacts with a plethora of microorganisms in its environment. The plant microbiota, which includes bacteria, fungi, protists, nematodes, and viruses, can form complex co-associations with its host. Plant microbial communities have been shown to promote plant growth, nutrient uptake, as well as pathogen resistance, thus promoting plant health in its natural environment (Richardson and Simpson, 2011, Pieterse et al., 2014, Trivedi et al., 2016, Backer et al., 2018, Gouda et al., 2018, Trivedi et al., 2020). However, our understanding of plant-microbial alliances’ functional mechanisms and broad ecological implications is still in its early stages (Martin et al., 2017).

Among the bacterial communities interacting with plants, a subset known as the root-associated bacteria (RAB) colonizes root compartments, such as the rhizosphere (i.e. the narrow zone surrounding and influenced by plant roots growth and metabolic activity) and intraradical regions and has the ability to modulate plant functions (Mendes et al., 2013, Qu et al., 2020, Dobbelaere et al., 2003, Walker et al., 2003). It has been suggested that the holobiont, defined as the host and its associated microbiome, forms a discrete symbiotic ecological unit that has co-evolved to maintain host function and fitness (Hassani et al., 2018, Uroz et al., 2019, Vandenkoornhuyse et al., 2015, Trivedi et al., 2020, Liu et al., 2019). Plants selectively recruit RAB depending on different factors such as host genotype, metabolic profiles, root exudates, and soil physicochemical properties (Bulgarelli et al., 2013, Hassani et al., 2018, Liu et al., 2019, Uroz et al., 2019). Although the plant and its exudates mostly influence this microenvironment, environmental influences also significantly shape its composition (Vives-Peris et al., 2020, Najera et al., 2020). Furthermore, diverse endophytic bacteria possess the ability to produce bioactive compounds of usually low molecular mass, with a wide range of reported functions, as part of their secondary metabolism. These compounds, which are not crucial for the fundamental cellular processes of bacteria in controlled laboratory conditions, often confer selective advantages within natural ecosystems. Such secondary metabolites (SMs) include antibiotics, pigments, toxins, enzyme inhibitors, immunomodulators, effectors of ecological competition or symbiosis, as well as compounds with hormonal activity or particular effects on lipids or carbohydrate metabolism (reviewed in (Fouillaud and Dufosse, 2022)). Many RABs have been reported to enhance plant growth through processes such as nutrient acquisition, hormonal regulation, protection against pathogens, and mitigation of abiotic stresses (reviewed and discussed in (Narayanan and Glick, 2022, Glick, 1995, Olanrewaju et al., 2017, Glick, 2012, Hassani et al., 2018, Mendes et al., 2013, Qu et al., 2020, Trivedi et al., 2020, Uroz et al., 2019, Vandenkoornhuyse et al., 2015)). These RAB, conferring growth benefits to plant hosts, are collectively referred to as plant growth-promoting rhizobacteria (PGPR), which typically belong to the Proteobacteria, Actinobacteria, Firmicutes, and Bacteroidetes phyla (Santoyo et al., 2016). Within these phyla, genera such as Rhizobium, Bacillus, Pseudomonas, and Burkholderia are prominently recognized as PGPR and have demonstrated functional contributions in mediating seed germination, plant growth and biocontrol of plant disease, notably by conveying secondary metabolites inclusive of antibiotics, volatile compounds, phytohormones and siderophores (Glick, 2012, Kang et al., 2012, Walia et al., 2013, Bloemberg and Lugtenberg, 2001). Notably, the publication of whole-genome sequences of rhizosphere-inhabiting bacteria facilitates the identification of genes involved in the regulation and production of SMs by comparative and functional genomics. However, despite the established knowledge that symbiotic associations between roots and PGPR provide multiple growth benefits to host plants, little is known regarding the role of PGPR and other RAB in shaping plant ecology through their secondary metabolism (Siefert et al., 2019, Kaur and Sharma, 2021).

Orchids, which account for approximately 10% of all seed plants, exhibit intricate associations with diverse microorganisms in their natural habitat (Givnish et al., 2016, Givnish et al., 2015). One example is mycoheterotrophy, an alternative nutritional strategy observed in certain plant species, particularly orchids, which derive sugars and essential nutrients from symbiotic associations with mycorrhizal fungi (Timilsena et al., 2023). The marsh orchid Dactylorhiza traunsteineri is an allotetraploid (2n = 80), mainly found in the eastern hemisphere (Paun et al., 2010). Dactylorhiza traunsteineri is part of the D. majalis complex, which has evolved polytopically from unidirectional hybridization between the diploid species Dactylorhiza fuchsii and D. incarnata (Pillon et al., 2007, Brandrud et al., 2020). It is slender, usually low-grown with narrow leaves and a lax few-flowered inflorescence, has narrow soil moisture and pH tolerance, and grows in calcareous fens, typically on naturally open sites associated with seepage zones (Kull and Hutchings, 2006, Soó, 1980, Baumann and Künkele, 1988, Delforge, 2001). The species occurs in low-nutrient soils, characterized in particular by very low levels of available nitrate and other macro and micronutrients (Wolfe et al., 2023). A recent study documented, at two different localities, a high diversity of fungi associated with the roots of D. traunsteineri in its natural environment, in comparison to its sibling D. majalis (Emelianova et al., 2024). Although hybrids and allopolyploids are known to be more resistant to environmental changes due to their larger genomic potential, D. traunsteineri is classified as endangered by several organizations in Europe and has proven challenging to grow and maintain under artificial environmental conditions, which is a requirement for ex-situ conservation efforts (Paun et al., 2007, Blinova and Uotila, 2012, Bartók et al., 2018, Paun et al., 2011).

This study presents a comprehensive analysis of the rhizosphere-associated bacterial microbiome of Dactylorhiza traunsteineri. Our community analysis, based on 16S ribosomal DNA (rDNA)-targeted sequencing, identified Proteobacteria, Actinobacteria, Myxococcota, Bacteroidota, and Acidobacteria as the predominant phyla within the D. traunsteineri rhizosphere. Leveraging deep shotgun metagenomics and de novo metagenome assembly, we extracted 47 MAGs, unveiling significant metabolic and biosynthetic capabilities of the orchid’s RAB. Notably, this study represents the first comprehensive investigation into the microbial community of the D. traunsteineri rhizosphere, highlighting a potential novel source of critical bioactive substances. Our findings illuminate previously uncharacterized PGPR SM pathways and the discovery of “dark matter” biosynthetic gene clusters, underscoring the rhizosphere’s untapped potential for biotechnological and ecological applications. This research advances our understanding of D. traunsteineri’s microbial interactions and sets the stage for future studies to exploit these microbial consortia for conservation and bioprospecting efforts.

Materials & methods

Sampling and sample processing

Dactylorhiza traunsteineri individuals (from the turfosa subspecies) were isolated from the Kitzbühel alpine town region, in the western Austrian province of Tyrol on the 25th of May 2018. The average temperatures recorded in this region during summer are between 13°C and 26°C, and between 0°C and 6°C during winter, with an annual total rainfall estimated to 1,443 mm, the majority occurring primarily during summer in June (average 188 mm). The individuals were potted and maintained at the Botanical Garden of the University of Vienna under conditions mimicking those of the sampling region. The sampling of the rhizosphere soil was performed on 03.02.2020 under sterile laboratory conditions. The plantlets were carefully uprooted, the soil discarded by shaking the roots, and the adherent soil directly in contact with the root system collected. Additionally, the roots of the plantlets were washed for 5 min with sterile DNA-free laboratory-grade Sartorius avium® filtered water. Samples were collected in sterile 50 ml Cellstar tubes. Both samples were shock-frozen in liquid nitrogen and kept at -80°C until the metagenomic DNA extraction was performed.

Metagenomic DNA extraction

Isolation of the metagenomic DNA was performed from the rhizosphere soil and the plantlet wash lyophilized samples in a sterile environment. The root wash samples were lyophilized using the FreeZone Freeze Dryer system [LabConco; Cat. 700201000]. 500 mg of lyophilized mass from each sample was transferred to a 2 mL screw cap reaction tube with sterile glass beads. Samples were first disrupted using the FastPrep-24 Classic homogenizer [MPBio; Cat. 116004500] at setting level 6 for 30 seconds. 1 mL of 2xCTAB buffer (pH 8.0) [100 mM Tris (pH 8.0); 20 mM EDTA (pH 8.0); 1.4 M NaCl; 2% (w/v) CTAB] with 4 mL β-mercaptoethanol, was preheated at 55°C and added to the samples, before disrupting the biomass a second time using the FastPrep-24 Classic homogenizer, twice for 30 seconds at setting level 5. The samples were then incubated for 20 minutes at 65°C and centrifuged for 1 minute at 10,000 rpm before transferring the supernatant and foam to fresh, sterilized 2 mL reaction tubes. Next, 400 µL Phenol and 400 µL SEVAG [4% (v/v) isoamyl alcohol; 96% (v/v) chloroform] were added to the samples before vigorous mixing and incubation for 1 hour at room temperature. The aqueous phases were transferred to fresh, sterilized 2 mL reaction tubes, and 800 µL chloroform was added to the samples. The samples were mixed vigorously and centrifuged for 1 minute at 10 000 rpm, before transferring the aqueous phase to fresh sterilized 2 mL reaction tubes. 2 µL RNase A [10 mg/ml [ThermoFisher; Cat: EN0531]] were then added, and the samples mixed by inversion and incubated for 15 minutes at 60°C.

Next, 2 volumes of ice-cold 96% ethanol were added to the samples and incubated overnight at -20°C for RNase A inactivation and DNA precipitation. Following overnight incubation, the samples were centrifuged for 20 minutes at 10,000 rpm at 4°C. The supernatants were discarded, and the DNA pellets were washed once in 500 µL 96% ethanol. The samples were then centrifuged for 10 minutes at 10,000 rpm at 4°C, and the DNA pellets washed twice with 500 µL 70% ethanol. After centrifuging for an additional 10 minutes at 10,000 rpm at 4°C, the DNA pellets were dried and resuspended in 100 µL ultra-pure water and incubated overnight at 4°C, before quality assessment. Samples were stored at 4°C.

16S rDNA amplicon sequencing and analysis

For bacterial community composition analysis, the V3-V4 fragment of the 16S rDNA gene was amplified from the genomic DNA extracted from the soil and root wash samples. 16S rRNA fragment concentrations in the DNA extracts were quantified using domain-specific quantitative PCR (Fontaine et al., 2023). DNA extracts were normalized based on the 16S rRNA gene concentrations to ensure 16S rRNA gene template molarity for amplification and a two-step barcoding procedure. A first amplification reaction was performed using the 16S rDNA 341F and 805R primer set, containing adapters for the introduction of Illumina adapters (341F + adapter: ACACTCTTTCCCTACACGACGCTCTTCCGATCTNNNNCCTACGGGNGGCWGCAG; 805R + adapter: AGACGTGTGCTCTTCCGATCTGACTACHVGGGTATCTAATCC), using Q5 high-fidelity DNA polymerase [New England Biolabs; Cat: M0491] and the following cycling conditions: 98°C for 1 minute, followed by 20 cycles of 98°C for 10 seconds, 62°C for 30 seconds, 72°C for 30 seconds, and a final extension at 72°C for 2 minutes. The 16S rRNA gene copy concentrations in DNA extracts were determined through quantitative PCR and normalized to equal 16S rRNA fragment copy number to enhance comparability and reduce PCR bias, and purified using the Agencourt AMPure XP [Beckman; Cat: B23319] purification system. Sequencing indexes were introduced in a second amplification reaction with sample-specific barcodes (FWD: AATGATACGGCGACCACCGAGATCTACAC-[index]-ACACTCTTTCCCTACACGACG; REV: CAAGCAGAAGACGGCATACGAGAT-[index]-GTGACTGGAGTTCAGACGTGTGCTCTTCCGATCT) using Q5 high-fidelity DNA polymerase and the following cycling conditions: 98°C for 1 minute, followed by 15 cycles of 98°C for 10 seconds, 66°C for 30 seconds, 72°C for 30 seconds, and a final extension at 72°C for 2 minutes. Amplicon libraries were purified using the Agencourt AMPure XP purification system and quantified using PicoGreen kit [ThermoFisher; Cat: P7589]. The libraries were pooled and diluted to 4 nM before being sequenced for 300 cycles pair-end on the Illumina MiSeq system using the MiSeq® Reagent Kit v2 [Illumina; Cat: MS-102-2002].

Following sequencing, the quality of the 16S amplicon sequences was verified using FastQC (v.0.11.9) (Andrews, 2010) and sequencing adapter sequences, as well as low-quality reads, were removed from the raw read data using Trimmomatic v.0.40 (Bolger et al., 2014) with the following parameters: ILLUMINACLIP:TruSeq3-PE-2.fa:2:30:10:2:keepBothReads LEADING:3 TRAILING:3 SLIDINGWINDOW:4:15 MINLEN:36. 16S amplicon sequencing data analysis has been performed using the R package dada2 (v.1.26.0) (Callahan et al., 2016), which includes quality filtering, dereplication, dataset-specific error model determination, ASV interference, chimeral removal, as well as taxonomic assignment, as part of its standard workflow.

Shotgun Sequencing

Shotgun metagenomics sequencing was conducted on the genomic DNA extracted from the soil and root samples. Evaluation of the initial DNA concentrations in the prepared samples was carried out using the Qubit dsDNA HS Assay Kit [ThermoFisher; Cat: Q32851]. The libraries were prepared using 1000 ng of genomic DNA, which was fragmented using the Bioruptor sonication device [Diagenode; Cat: B01020001]. Fragments were cleaned up using the GeneJet PCR Purification Kit [ThermoFisher; Cat. #K0701] and size-selected for a 300 bp using Agencourt AMPure XP beads. Sequencing libraries were prepared following the standard NEBNext Ultra II DNA Library Prep with Sample Purification Beads [New England Biolabs; Cat: E7103] procedure. The final library concentrations were determined using the Qubit dsDNA HS Assay Kit and the average size of the library was determined using the Agilent 5200 Fragment Analyzer System [Agilent; Cat: M5310AA]. The libraries were pooled and diluted to 4 nM before being sequenced for 600 cycles pair-end on the Illumina MiSeq system using the MiSeq® Reagent Kit v3 [Illumina; Cat: MS-102-3303].

Metagenome assembly, genome binning and quality assessment

The quality of the metagenomic reads was verified using FastQC (v.0.11.9) (Andrews, 2010) and sequencing adapter sequences, as well as low-quality reads, were removed from the sequenced data using Trimmomatic v.0.40 (Bolger et al., 2014) with the following parameters: ILLUMINACLIP:TruSeq3-PE-2.fa:2:30:10:2:keepBothReads LEADING:3 TRAILING:3 SLIDINGWINDOW:4:15 MINLEN:36. The quality trimmed reads from both samples were combined into one large dataset, containing the reads from both samples. The combined dataset was assembled using MEGAHIT (v.1.2.9) (Li et al., 2015) (parameters --k-min 27 --k-max 247 --k-step 10). Next, the paired-end metagenomic trimmed reads were mapped back to the generated contigs using bowtie2 (v.2.3.5.1) (Langmead and Salzberg, 2012) to generate coverage information, and sorted using SAMtools (v.1.10). To generate MAGs, contigs were binned using Autometa (v2.0) (Miller et al., 2019). Unclustered contigs were succinctly binned using MetaBAT2 (v.17) (Kang et al., 2015). For further improvement, the corresponding reads from each MAGs’ contigs were extracted and reassembled using SPAdes (v.3.14.1) (Nurk et al., 2017, Bankevich et al., 2012) using the meta option. Finally, the quality of each MAG was evaluated using CheckM (v.1.1.3) (Parks et al., 2015) and QUAST (v.5.0.2) (Mikheenko et al., 2018), to determine their completeness, contamination and heterogeneity.

MAG analysis and taxonomical classification

For each MAG, ribosomal sequences were extracted based on the KEGG (Kanehisa, 2019, Kanehisa et al., 2023, Kanehisa and Goto, 2000) and PANNZER2 (Toronen et al., 2018) annotations, and quantified by type to identify the most abundant representatives in each MAG. For each binned MAG, taxonomic and lowest common ancestor predictions were generated using Autometa (Miller et al., 2019), based on the non-redundant NCBI database (O’Leary et al., 2016). Furthermore, a library containing genomic sequences of 199 reference organisms (Supp. Table 1), obtained from the NCBI database, was generated to calculate the average nucleotide identify (ANI) of each MAG (Supp. Figure 1) using FastANI (v.1.33) (Jain et al., 2018).

Sequencing statistics (16S profiling)

Gene prediction and annotation

For each generated MAG, genes and proteins were predicted using Prodigal (v.2.6.3) (Hyatt et al., 2010), and the sequences annotated using the KEGG (Kanehisa, 2019, Kanehisa et al., 2023, Kanehisa and Goto, 2000) and PANNZER2 (Toronen et al., 2018) databases. The KEGG annotation was further used to investigate metabolic pathways, whereas PANNZER2 was used to annotate gene ontology and functional domains. Additionally, biosynthetic gene clusters (BGCs) for each MAG were predicted using antiSMASH (v.6.1.1) (Blin et al., 2017, Blin et al., 2021), DeepBGC (v.0.1.31) (Hannigan et al., 2019), and GECCO (v.0.9.10) (Carroll et al., 2021), annotated using the KEGG and PANNZER2 databases and evaluated for completeness.

Results

16S rRNA gene amplicon sequencing/taxonomic identification

To gain insights into the rhizosphere of D. traunsteineri, a potted orchid (Individuum 1 from Kitzbühel, sampled on 25.05.2018), maintained in the botanical garden of the University of Vienna (Universität Wien), has been transferred to our laboratory environment (Figure 1). To taxonomically characterize the predominant bacterial and archaeal community inhabiting the rhizosphere of D. traunsteineri, we conducted PCR amplification and sequencing of the V3-V4 variable region of the 16S rRNA gene from the soil and root wash samples in technical duplicates. Following paired-end sequencing, a total of 240,935 reads were generated, each with a read length of 300 bp (Table 1). After rigorous processing steps, including trimming, filtering of low-quality reads, merging of forward and reverse reads, and removal of chimeric sequences, 106,648 sequences were clustered into 2350 amplicon sequence variants (ASV) with taxonomic assignments spanning 31 phyla and over 84 classes (Figure 2). In total, we identified 2349 bacterial and 1 archaeal ASVs. The prevalence of the five major bacterial phyla ranged from 38.43% to 41.33% for Proteobacteria, 11.09% to 13.55% for Actinobacteriota, 9.40% to 11.90% for Myxococcota, 9.72% to 11.37% for Bacteriodota, and from 8.72% to 9.26% for Acidobacteriota (Table 2). Comparisons of ASV abundances across the four samples using permutational multivariate analysis of variance (PERMANOVA) with the R package vegan revealed no significant differences (similarity p-value < 0.09), indicating a consistent sampling and sequencing approach. Encouraged by the stability of the bacterial composition in the soil surrounding the roots of D. traunsteineri and the bacteria adhering to the roots themselves (root wash samples), we opted to analyze and process the shotgun metagenomic sequences as a unified sample.

Dactylorhiza traunsteineri sampling.

D. traunsteineri individuals from the Kitzbühel region (Tyrol, Austria) (A). D. traunsteineri individuum 1 from Kitzbühel in the laboratory environment (B). Washing of the roots with laboratory-grade water (C).

Bacterial Community Composition Analysis based on 16S rDNA sequencing.

Alpha-diversity (Shannon and Simpson indexes) (A). Relative abundance of identified ASVs at Phylum (B) and Class (C) levels.

ASVs abundance (by Phylum) (%)

Metagenome assembled genomes (MAGs) generation

To investigate the functional properties of the D. traunsteineri rhizobiome, we proceeded to shotgun metagenomic sequencing using short-read technology. We generated 33,043,213 paired-end metagenomic reads from the soil and root wash samples, totalling over 9 Gbp, with each sample ranging from over 1.8 to over 4.9 Gbp (Table 3). After in silico pruning of adapter sequences and low-quality reads, 31,541,795 reads were retained, representing over 95% of the initial dataset. De novo metagenomic assembly on the combined datasets using MEGAHIT (Li et al., 2015, Li et al., 2016) resulted in a total of 859,083 contigs, totalling over 7.5 Gbp (Table 4).

Sequencing statistics (Shotgun sequencing)

Metagenomic assembly statistics

The obtained contigs were binned using Autometa (Miller et al., 2019), resulting in 37 predicted MAGs, as well as a group of unclustered contigs, which was further binned using MetaBAT2 (Kang et al., 2015), thus resulting in 10 additional MAGs (Figure 3). Overall, from the 248,298 contigs that were binned as Bacteria, 44,138 contigs could be grouped into 47 MAGs, from which 36 MAGs could successfully be reassembled using SPAdes (Bankevich et al., 2012, Nurk et al., 2017). These MAGs represent 359 Mbp of assembled reads, while 14.86% of all reads mapped back to the sequences contained in the combined bacterial MAGs (Table 5). MAG size ranges from 361 kbp to 45.81 Gbp, with N50 ranging from 0.9 kbp to 221 kbp (Table 6). The average predicted completeness of the MAGs stands at 33.64%, with a mean coverage spanning from 2.11 to 25.00x. Following a quality assessment, 13 initially generated MAGs were excluded due to either a contamination score exceeding 50% or completeness falling below 10%. The taxonomic distribution of the top-represented groups revealed that 28 MAGs were classified as Proteobacteria, eight as Acidobacteria, and eight as Bacteroidetes (Table 7). Notably, our assembly efforts yielded one high-quality draft genome (MAG_Dt_26), satisfying the stringent Minimum Information about a Metagenome-Assembled Genome (MIMAG) criteria, boasting less than 5% contamination and over 90% completeness, as well as three medium-quality draft genomes (MAG_Dt_01, MAG_Dt_03, MAG_Dt_23), with estimated completeness above 50% with less than 10% estimated contamination (Bowers et al., 2017).

Barnes-Hut t-SNE representation of MAG contigs.

BH-t-SNE representation of MAG contigs coloured by MAG (A) and taxonomic assignment (B).

Metagenomic binning statistics

Gene prediction and secondary metabolite potential

Gene prediction was performed using PRODIGAL (Hyatt et al., 2010) for each MAG, and the resulting predicted proteins were annotated using the KEGG (Kanehisa and Goto, 2000, Kanehisa et al., 2023) and PANNZER2 (Toronen et al., 2018) databases for metabolic pathway enrichment (Table 8). Furthermore, each MAG was mined for SM BSGs using the nf-core funcscan pipeline (Ewels et al., 2020). A total of 1,741 BGCs were predicted from the D. traunsteineri rhizosphere bacterial MAGs (Figure 4; Table 9).

Overview of number of predicted SM BGCs per MAG and their product classes.

In MAG_Dt_26 (Proteobacteria; completeness of 92.03%; contamination of 2.53%), a total of 181 SM BGCs were predicted. These included six nonribosomal peptide (NRP) clusters, 24 polyketide clusters, 22 ribosomally synthesized and post-translationally modified peptide (RiPP) clusters, two RiPP-like clusters, five saccharide clusters, as well as four terpene clusters (Supp. Table 2). Additionally, 101 gene clusters were categorized as “Other”, potentially representing multiple product types and rarer SM classes, while 17 clusters were classified as “Unknown” (Supp. Table 2). From MAG_Dt_01 (Actinobacteria; completeness of 80.24%; contamination of 4.58%), a total of nine SM BGCs were predicted, representing one NRP cluster, one polyketide cluster, two RiPP clusters, two terpene clusters, as well as three “Other” product class clusters (Supp. Table 2). From MAG_Dt_03 (Acidobacteria; completeness of 65.94%; contamination of 3.99%), four SM BGCs were predicted, representing one NRP cluster, one terpene cluster, and two “Other” product class clusters (Supp. Table 2). From MAG_Dt_23 (Proteobacteria; completeness of 61.01%; contamination of 2.68%), four SM BGCs were predicted, representing two RiPP-like and two terpene clusters (Supp. Table 2).

Discussion

Through comprehensive 16S rDNA profiling, a total of 2,349 bacterial and one archaeal ASVs were identified within the rhizosphere of D. traunsteineri. The predominant bacterial phyla detected included Proteobacteria, Actinobacteria, Myxococcota, Bacteroidota, and Acidobacteria, which collectively accounted for 95% of the sequences identified. Importantly, there were no significant differences in the bacterial community composition between the soil and the root-adhering bacteria associated with D. traunsteineri. This consistency underscores the robustness and reproducibility of our sample preparation and sequencing methodologies. Moreover, in our review of existing literature on orchid RAB ecology, we validated that our approach successfully recovered representatives from all phyla previously reported in orchid rhizospheres (Kaur and Sharma, 2021). Notably, the scientific documentation of orchid RAB through culture-independent methods is relatively scarce. Previous studies have predominantly utilized conventional Sanger sequencing methods, while others employing high-throughput sequencing technologies have primarily focused on 16S rDNA profiling of bacterial communities directly extracted from root samples. Nevertheless, the field of orchid RAB research is undergoing a transformative shift with the advent of innovative methodologies. In this context, we introduce a pioneering approach that leverages shotgun metagenomics, followed by the reconstruction of MAGs. This cutting-edge technique promises to offer deeper insights into the complex ecosystem of the orchid rhizobiome, providing unparalleled understanding of its microbial diversity and functional potential. Utilizing shotgun sequencing and de novo metagenomic assembly, we successfully obtained 47 bacterial MAGs from the rhizobiome of D. traunsteineri. Our detailed analysis focused on four high-quality MAGs: MAG_Dt_01, MAG_Dt_03, MAG_Dt_23, and MAG_Dt_26. Specifically, MAG_Dt_01 was taxonomically assigned to Actinobacteria bacterium (taxid 1883427), MAG_Dt_03 to Acidobacteria bacterium (taxid 1978231), and MAG_Dt_23 and MAG_Dt_26 to Dongia sp. URHE0060 (taxid 1380391) and Steroidobacter (taxid 469322), respectively, both belonging to the phylum Proteobacteria.

The phyla Proteobacteria and Actinobacteria, generally enriched in the rhizosphere, alongside Acidobacteria, typically enriched in bulk soil, play pivotal roles in soil carbon and nitrogen cycles as well as the decomposition of organic matter (Eichorst et al., 2018, Song et al., 2021, Ling et al., 2022). The Actinobacteria phylum represents one of the largest and most diverse bacterial group, inhabiting soil, water and plant tissue (Diab et al., 2024, Boubekri et al., 2022), and are prolific producers of SMs including well-known antibiotics such as streptomycin, tetracycline, erythromycin, and vancomycin (Berdy, 2005). Beyond antibiotics, they produce a wide array of SMs with potential applications as biopesticides (Barka et al., 2016). For example, polyketides, such as macrolides, exhibit antifungal activity, while non-ribosomal peptides (NRPs), like cyclopeptides, demonstrate insecticidal properties (Gaynor and Mankin, 2003, Quiroz-Carreno et al., 2022). The Acidobacteria phylum, although highly prevalent within soil ecosystems, is characterized by its challenging cultivation (Kalam et al., 2020). This phylum fulfills crucial ecological functions through its significant involvement in carbon, nitrogen, and sulphur biogeochemical processes (Ward et al., 2009, Eichorst et al., 2018, Hausmann et al., 2018). Furthermore, Acidobacteria possess genes conducive to survival and competitive colonization within the rhizosphere, thereby facilitating the formation of symbiotic relationships with plants (Kalam et al., 2020, Liu et al., 2024). Members of the Proteobacteria phylum, such as the Dongia and Steroidobacter genera, have been frequently isolated from rhizosphere soil and display the potential to degrade complex organic compounds, including various polysaccharides (Ikenaga et al., 2021, Huang et al., 2019, Liu et al., 2024, Baik et al., 2013, Kim et al., 2016, Liu et al., 2010, Palla et al., 2022). The high-quality MAGs obtained in this study present a promising opportunity to elucidate in greater detail the roles of these phyla in plant microbiome interactions, particularly regarding their secondary metabolite potential and their impact on plant health and development.

Given that our sampling and sequencing efforts were confined to the rhizobiome of a solitary specimen of D. traunsteineri, caution must be exercised in generalizing the presence and prevalence of the identified MAGs across all populations of D. traunsteineri. However, it is important to note that the sampled individual originated from an extant population and has been maintained under conditions that closely simulate its natural habitat at the Botanical Garden of the University of Vienna. These conditions include soil composition, humidity, and temperature, which are critical for mimicking the plant’s natural environment. The controlled yet naturalistic setting in which the specimen was studied provides a valuable context for our findings. Despite the inherent limitations of a single-sample study, our research offers a foundational framework for future interaction studies between RABs and D. traunsteineri. This study lays the groundwork for more extensive investigations into the microbiome of this orchid species across different populations and environmental conditions. Furthermore, our findings open new avenues for exploring the complex processes involved in the adaptation, ecology, and evolution of this allopolyploid orchid. Understanding the interactions between D. traunsteineri and its associated microbiome can shed light on the mutualistic relationships that facilitate plant health, nutrient acquisition, and stress resilience. Such insights are crucial for conservation strategies, especially for orchids that are often sensitive to environmental changes and habitat disturbances. Additionally, the use of advanced metagenomic techniques in our study underscores the potential for discovering novel microbial taxa and functional genes that contribute to the unique ecology of orchid rhizospheres. This approach not only enhances our understanding of plant-microbe interactions but also provides a deeper comprehension of the evolutionary dynamics within the orchid family. Our study, therefore, serves as a stepping stone for future research aimed at unraveling the intricate web of interactions that sustain and influence the biodiversity and ecological success of D. traunsteineri.

Conclusion

This study provided a comprehensive overview of the microbiome inhabiting the rhizosphere of the endangered orchid Dactylorhiza traunsteineri. By conducting an in-depth bacterial community composition analysis through 16S rDNA-targeted sequencing, we identified Proteobacteria, Actinobacteria, Myxococcota, Bacteroidota, and Acidobacteria as the most abundant phyla within the D. traunsteineri rhizosphere. This foundational data highlights the complexity and diversity of the microbial communities associated with this orchid species. Our investigation extended beyond community profiling to include deep shotgun metagenomics sequencing and subsequent de novo metagenome assembly. This approach enabled us to extract high-quality metagenome-assembled genomes (MAGs), providing a rich resource for further functional analysis. Each MAG was meticulously analyzed for metabolic pathway enrichment and secondary metabolite (SM) biosynthetic gene clusters (BGCs), revealing a previously unexplored niche of the rhizobiome. These analyses uncovered potential metabolic capabilities and biosynthetic potentials that are integral to the rhizosphere’s ecological dynamics.

The insights gained from this study not only enhance our understanding of the microbial diversity within the D. traunsteineri rhizosphere but also lay the groundwork for future research aimed at unravelling the intricate interconnections between the rhizobiome and its host. The high-quality MAGs we have made available serve as a valuable resource for further exploration of microbial functions and their contributions to the health, adaptation, and survival of D. traunsteineri. Moreover, our findings underscore the importance of employing advanced metagenomic techniques to study plant-associated microbiomes, particularly for endangered species. By elucidating the functional roles of key microbial players, we can develop more effective conservation strategies and foster a deeper understanding of the symbiotic relationships that underpin plant resilience and ecological success. In conclusion, our study represents a significant step forward in orchid microbiome research, providing novel insights and a robust platform for future investigations into the dynamic interactions within the rhizosphere of D. traunsteineri. These efforts are crucial for advancing our knowledge of plant-microbe interactions and supporting the conservation of this and other endangered orchid species.

Software

Supplementary Information

ANI of MAGs and reference organisms.

Genomes used for comparative analysis

Availability of data and materials

The datasets supporting the conclusions of this article are publicly available in the National Center for Biotechnology Information (NCBI) repository under BioProject PRJNA1124010.

Acknowledgements

We would like to thank David Pressler and the Botanical Garden of the University of Vienna for maintaining the plant in cultivation, and the Tyrol County administration for issuing necessary permits.

Additional information

Authors’ contribution

GAV, CZ, DFS and OV conceptualized and designed the study. GAV and LZ generated the datasets. GAV and JC performed the data analysis. GAV and JC generated the figures and drafted the manuscript. ARMA and RLM revised and contributed to the manuscript. All authors read and approved the final manuscript.

Funding

This study was supported by a grant from the Austrian Science Fund (FWF): P29556 given to RM, and the PhD program TU Wien bioactive.

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Abbreviations

  • ANI: average nucleotide identity

  • ASV: amplicon sequence variant

  • BGC: biosynthetic gene cluster

  • MAG: metagenome assembled genome

  • MIMAG: Minimum Information about a Metagenome-Assembled Genome

  • OTU: operational taxonomic unit

  • PERMANOVA: permutational multivariate analysis of variance

  • PGPR: plant growth promoting Rhizobacteria

  • RAB: root associated bacteria

  • SM: secondary metabolite

Additional files

Supplementary Table 2

Table 6

Table 7

Table 8

Table 9