Microbiome diversity and variations in industrial hemp genotypes

Industrial hemp and marijuana belong to a single biological species, Cannabis sativa L., but they differ in their uses, cultivation, and chemical composition1,2. The identity of two different genotypes is mainly based on a threshold of the psychoactive compound delta-9-tetrahydrocannabinol (THC) and cannabidiol (CBD)3,4,5. Recently, the United States Farm Bills have reduced restrictions to produce modern industrial hemp with less than 0.3% THC content, which led to a renewed interest in hemp cultivation6,7. The hemp generated a significant market size in North America, mainly fiber, oilseed, and pharmaceuticals7. Selective breeding and cultivation practices have also played a significant role in the diversity of hemp genotypes8. Industrial hemp has been selectively bred for desirable traits such as fiber quality, seed yield, oil content, and psychoactive properties, leading to different phenotypes and chemotypes9,10,11,12. Despite technological advances in Cannabis breeding, the proportion of (CBD) and THC fluctuates greatly depending on the gender (male or female), genotypes, cultivation practices, and biotic or abiotic stresses. More recently, it has been shown that microorganisms such as fungi or bacteria play a crucial role in defining phenotype and chemotype traits13. Specific genotypes can host specific microbes in the phyllosphere and rhizosphere14.

Microbes living with host plants, either as endophytes (inside) or epiphytes (outside), provide access to essential nutrients, produce beneficial metabolites and enzymes, and protect plants from biotic and abiotic stress factors14,15. The microbial communities associated with host plants can perform specific functions, and their diversity can be based on the (i) type of host’s genetic makeup, (ii) physio-photosynthetic responses to the environment, (iii) habitat of growth, and (iv) exposure to stress and their responses14,15,16. In addition, the beneficial and core-microbiome (a microbial taxonomy abundantly available in a particular habitat17) species show a significant variation across bulk soil to rhizosphere (below-ground) and phyllosphere (above-ground) parts of the plant18. Microbiome variation could be defined as changes in the diversity of specific genera across different parts and segments of the host plant14. This variation significantly impacts several functions of microbes during the growth stages of the host14,19. Hence, it is essential to understand these variations to help improve the growth promotion of the host for obtaining specific economically important traits, for example, CBD and fiber in hemp. Several studies have recently shown that host genotypes could impact specific phenotypic traits, including the production of specialized metabolite and the symbiosis of specific microbes. For example, the yews plant produces paclitaxel, whereas the symbiotic Pestalotiopsis microspore Ne32 adopted the same trait. Additionally, the same species also improve plant growth20. A similar conclusion was drawn on core-microbiome diversity when C. sativa “TJ’s CBD� was used across six field locations5.

Our initial work involved the cultivation of over 100 accessions sourced from various regions across the United States. Following a thorough evaluation, we identified and selected cultivars that demonstrated optimal performance within the unique climatic conditions of Texas21. Our focus has been refined to center on the most high-performing cultivars concerning CBD (Sweet Sensi and Cherry Wine) and fiber (American Victory and Unknown cultivar). The unknown cultivar used in this study is still in the process of breeding and trial stages. Hemp cultivars currently need standardized botanical cultivar nomenclature. The industry still relies on various non-scientific common names. In this context, the term ‘unknown cultivar/genotype’ signifies that we do not even know the common name of this genotype. We have utilized these cultivars as focal points for comprehensive field-level trials. Going beyond conventional assessments of yield and quality, our study extended its scope to unravel the intricate microbiome diversity associated with these strains/genotypes.

Through this exploration, we aimed to decipher the intriguing relationship between Hemp cultivars and their microbiota, thereby contributing valuable insights to the broader understanding of sustainable and resilient agricultural practices for industrial hemp in Texas (USA). However, how different hemp cultivars recruit and colonize the microbial communities has yet to be fully understood. Hence, in this study, we evaluated (i) the microbiome structure and diversity of two CBD (Sweet Sensi and Cherry Wine) and two fiber (American Victory and Unknown cultivar) producing industrial hemp genotypes and (ii) assess the core microbiome structure across the rhizosphere (soil and root) and phyllosphere (stem and leaf). The unknown cultivar is in the process of characterization and taxonomic process. We will elucidate two critical factors that can influence the microbiome composition, viz. (i) genotypes, either its CBD or Fiber trait-oriented hemp cultivar and (ii) variation across rhizosphere and phyllosphere microbiome diversity. We demonstrate that these factors can impact microbiome diversity using next-generation sequencing and analysis. The results from this study will help to understand the microbiome function of hemp plants, which can have the potential to develop plant growth-promoting inoculants for commercially cultivating important hemp strains.

Plant material and growth conditions

Sweet Sensi, Cherry Wine, American Victory-1, and one unknown genotype of industrial hemp were grown in field conditions, and at the mature stages, the samples were collected from the field of Prairie View A&M University (PVAMU; Latitude:30.0919 Longitude: − 95.9894). Initially, the Tetra Hemp Company (Texas, USA) provided Sweet Sensi (Fig. 1A), Cherry Wine (Fig. 1B), and Unknown Germplasms (Fig. 1D). The American Victory germplasm (Fig. 1C) was provided by Aqui Flow LLC, Texas, to PVAMU. The PVAMU possesses a license from the Texas State for hemp growth and research, and all the plants were grown per suggested guidelines and legislation. All the samples were collected from 9 different plants and their rhizosphere soil. Plant and soil sampling was performed in compartments of phyllosphere (leaf and stem) and rhizosphere (root and soil) and brought to the lab in sterilized conditions at 4 °C. Extraction of DNA was performed from 3 replicates of each sample (leaf, stem, root, and soil). The soil samples were collected from rhizospheric regions (near roots, ~ 6 inches) and processed through selective sieves (2–4 mm) for DNA extraction. The root, stem, and leaf parts were surface sterilized to remove the epiphytic microbes. For surface sterilization, the samples were washed with sodium hypochlorite (5%), ethanol (70%), and autoclaved distilled water22. After preparing samples, all four genotypes were immediately transferred to liquid nitrogen and kept at − 80 °C until further analysis.

Microbiome DNA extraction and analysis

The soil samples were processed using the protocol Khan et al.23. The soil DNA was extracted using the ZymoBIOMICS Miniprep Kit according to manufacturing protocol. The MagMAX™ Plant DNA Kit (Thermo Scientific, California, USA) was used to extract the plant’s leaf and stem DNA for the endophytic microbiome. The manufacturer’s instructions were used with a few modifications to extract high molecular weight DNA. Similarly, the sterilized root DNA was extracted using a manual protocol that was recently published by Khan et al.23. The sterilized samples (root, stem, and leaf) were ground into a fine powder using a sterile mortar and pestle using liquid nitrogen. We used about 500 mg of plant sample (a pool of nine samples) and homogenized in extraction buffer to yield one replicate. The DNA from three replicates of each plant compartment from each genotype was passed through a DNA Cleanup kit (ThermoFisher, California, USA) to remove impurities. The quantity and purity of all DNA samples were assessed using a NanoDrop Lite Spectrophotometer and Qubit™ 4.0 Fluorometer (ThermoFisher, California USA) using the manufacturer protocols and High Sensitivity DNA kit (ThermoFisher, California USA). Gel electrophoresis was performed to verify the integrity of DNA fragments. Samples with low-quality DNA fragments were repeated for DNA extractions. Thus, a minimum of 100–150 ng/uL of DNA with a purity ratio of up to ~ 1.8 was processed for microbiome sequencing24.

Microbiome sequencing

The genomic DNA (~ 100–150 ng per microliter) was processed for amplicon library preparation and sequencing. PCR-free libraries of each DNA sample were generated by amplifying the internal transcribed spacer (Fungal ITS1-1F for leaf, root, and stem endophytes and ITS-5F for soil samples) and 16S rRNA (V5–V7 for leaf, root, and stem endophytes and V4-V5 for soil) for fungal and bacterial communities, respectively. The primers used for the 16S V4–V5 region were 515F (GTGCCAGCMGCCGCGGTAA) and 907R (CCGTCAATTCCTTTGAGTTT), and for the V5–V7 endophytic primers were 799F (AACMGGATTAGATACCCKG) and 1193R (ACGTCATCCCCACCTTCC). PNA (peptide nucleic acid) clamps were utilized to minimize the chloroplast and mitochondrial contaminations. Illumina MiSeq instrument (Illumina Inc., San Diego, CA, USA) was used to perform the paired-end sequencing approach of 300 bp operating with v2 chemistry (User Guide Part # 15,027,617 Rev. L). The sequencing facilities of the Novogene (San Diego, CA, USA) were used for this purpose, and the clean reads (Tables S1, S2, and S3) were further assessed and analyzed.

Microbiome analysis

QIIME2.0 v2022.225 was used to analyze the sequencing reads for ease of use and interpretable results. Quality assessment of the reads was performed with FastQC26. Sequences were filtered (by overall quality), and chimeric sequences were removed in denoising. The widely used DADA2 algorithm was utilized for denoising and generating the ASVs (amplicon sequence variants, Tables S4, S5)27. Amplicon sequence variants were opted for instead of operational taxonomic units (OTUs) due to the overestimation by OTUs compared to ASVs28. The SILVA database (v138) was used for the taxonomical classification of 16S sequences, whereas the ITS sequences UNITE database (v9.0) was used29. Chloroplast and mitochondrial sequences were filtered from the classified sequences by continuing the QIIME2.0 workflow. Shannon diversity index and Bray–Curtis distances were calculated for alpha and beta diversities, respectively30,31,32,33. All data sets generated by QIIME2.0 were exported to R for further statistical analysis and figure creation25,34. The R packages PhyloSeq35, R package ggplot236, and MicrobiotaProcess37 were used to analyze output files from QIIME2.025. The microbiome38 R package was used for core microbiome analysis.

Statistical analysis

To test the significant effect of factors (plant compartment, genotype, and CBD-Fiber), beta diversity distance matrixes were analyzed via permutative multivariate analysis of variance (PERMANOVA, 999 permutations)39. Their interaction on bacterial and fungal community composition using the “adonis2 function� from the vegan R package40 was also used to perform sums of squares of a multivariate data set. This test is analogous to multivariate analysis of variance41. We also performed the Linear discriminant analysis effect size (LEfSe)37,42. LEfSe was used to explain the differences between classes by coupling standard tests for statistical significance (p < 0.05) with additional tests encoding biological consistency and effect relevance43. The MicrobiotaProcess37 package was used to test the effects of the factors on the fungal and bacterial communities using R. An alpha value of 0.05 was used unless otherwise stated.

Microbiome structure of hemp

Microbiome diversity is defined as the number and abundance distribution of specific species of microbes44,45. Thus, the current research evaluated the microbial diversity and abundance of hemp species rich in CBD (Sweet Sensi and Cherry Wine) and Fiber (American Victory and Unknown). In this study, we evaluated both the rhizosphere (root and soil) and the phyllosphere (stem and leaf). The results from all samples showed that over 1,900,000 bacterial genera raw ASVs and over 4,200,000 fungal raw ASVs comprised 469 different genera. We found the fungal ASVs to be more abundant than bacterial ASVs. We also found that bacterial ASVs were significantly higher in rhizosphere soil (70,000) and root (47,000) when compared to phyllosphere stem and leaf (28,000–36,000) (Table S6). Interestingly, the overall fungal ASV abundance showed significantly higher counts across the phyllosphere than the rhizosphere. We found 118,000–131,000 ASVs in leaf and stem compartments and 74,000–81,000 counts in soil and root compartments. In the case of genotypes, a varying abundance of ASV count was observed across different compartments. Overall, the Cherry Wine showed significantly higher ASVs, followed by American Victory-1. We observed relatively high ASVs in the endospheric microbiome compared to the soil microbiome (Table S6).

We assessed the shared and unique ASVs across different genotypes and their compartments. The results showed that the bacterial ASVs were significantly shared in soil between the different genotypes, most likely due to their being grown in similar soil conditions. However, in the case of fungi, 14.6% more highly unique ASVs were found in fiber than in CBD-rich hemp species. Contrarily, a more varying distribution of ASVs was observed across different genotype compartments. The uniqueness and sharing of bacterial and fungal ASV recovered reduced from soil > root > stem > leaf compartments. More unique ASVs were retrieved from the CBD leaf (53.6%) than the fiber genotype (Fig. 2). The beta diversity data was subjected to PERMANOVA analysis, and interestingly, the stem compartment was significantly different (p < 0.05). This could be because the stem composition is different in CBD when compared to fiber-producing hemp varieties. Fiber-producing varieties tend to recruit different bacterial and fungal microbiomes than CBD-producing varieties. By digging into the CBD vs. fiber-producing varieties, the root compartment had significant (p < 0.05) variation in bacterial microbiomes. Likewise, the fungal microbiome was significantly (p < 0.05) different in the rhizosphere and stem compartment.

Rhizospheric and phyllosphere microbiome of hemp

In the soil, the most abundant (≥ 1% relative abundance) bacteria phyla were Firmicutes, Proteobacteria, Myxococcota, Actinobacteriota, Bacteroidota, Plantomycetota, Acidobacteriota, and Chloroflexi. Of the eight phyla in the data, the top 3 phyla were Proteobacteria, Acidobacteria, and Actinobacteria. Except for root parts, Firmicutes are significantly abundant compared to soil samples. Proteobacteria and Actinobacteria remain highly abundant in the root system. The top genera (≥ 1% relative abundance) are Streptomyces, Bacillus, and Pseudomonas across all genotypes. Among the keystone genera of American Victory-1 were from Family Kineosporiceae, which is roughly 15% relatively abundant, or genus Xylella, present at 16% relative abundance (Fig. 3).

In the stems and leaves of the plants, we found that Proteobacteria remain the top phylum, accounting for almost 97.5% of the Cherry Wine genotype. At the same time, Proteobacteria remains the top phylum for all other genotypes, between 60 and 98%. Firmicutes (0.5–21%) and Actinobacteria (1–14%) continue to play an important role in the stem. At the same time, Actinobacteria’s presence appears to be reduced in the leaf of hemp, with the percent change being between 34 and 89% decrease. The top genera (≥ 1% relative abundance) for hemp leaves were Bacillus, Pseudomonas, Pantoea, Xylella, Sphingomonas, and Methylobacterium-Methylorubrum (Fig. 3).

The top fungal phylum abundant (≥ 1% relative abundance) in the soil were Ascomycota, Mortierellomycota, and unidentified organisms. In the root system, we found that Ascomycota remains highly abundant across all hemp genotypes and their organs, while Basidiomycota was abundant in genotype American Victory-1. For both the stem and leaves, the fungal phyla Ascomycota and Basidiomycota were the most prevalent phyla, with unidentified phyla still being present at very low abundance for all plant organs. Ascomycota is the most abundant phylum, with relative abundance being around 77–90% in the stem and leaves.

Many fungal genera were found throughout the plant genotypes. In the root fungal diversity, we found the six most abundant genera (≥ 1% presence): Gibberella, Fusarium, unidentified (Class: Sordariomycetes), Alternaria, Neocosmospora, and Septoria. These belong to Ascomycota. As expected, fewer genera were found from the phylum Basidiomycota. These genera were Athelia, and Rhizoctonia. We also found that unidentified organisms were abundant here. The top genera found in the stem and leaves of hemp genotypes appear to differ from that of the root system: Septoria, Cladosporium, Alternaria, Gibberella, and Moesziomyces (Fig. 4). The only genus Moesziomyces belonged to Basidiomycota, whereas most genera were Ascomycota.

Microbiome variation in genotype

When analyzing the Principal Coordinate Analysis (PCoA), the Bray–Curtis distance method was utilized alongside the Hellinger method for all ß-diversity testing. All soil communities are tightly grouped around the center of the PCoA graph, showing high relatedness in all samples. Sweet Sensi has the most significant distance from the rest in bacterial communities, while Cherry Wine has the largest distance in the fungal communities. Using PERMANOVA, both communities and all organs had significant variation. The leaf bacterial communities are roughly equivalent, with American Victory-1 and Sweet Sensi having the closest relatedness. Fungal communities in plant leaves, closely related to American Victory-1, seem to vary from the other genotypes. The bacterial stem communities are more tightly clustered than those in the fungal stem communities. However, there was still significant community variation with P-values of about 0.04 and 0.01, respectively (Table S8) on genotype. However, the bacterial stem communities are far from the Cherry Wine and other genotypes. For all other organs, genotypes showed high statistically significant differences in both bacterial and fungal community structures (p values, ≤ 1.00E−04), as shown in Table S8.

The LEfSe identifies discriminative fungal and bacterial taxa from the different Hemp genotypes (Fig. 5). When analyzing the LDA effect size (LEfSe, α = 0.05), we focused on discriminative taxa with a p value of < 0.025 due to the smaller sample size. There were no bacterial or fungal taxa that were determined to be discriminative. While there were discriminative bacterial taxa (p < 0.05, none below 0.025), meaning the variation might not be significant. We identified discriminatory bacterial taxa within the roots of all genotypes; all P-values are 0.02 unless noted otherwise. The unknown genotype had 13 identified taxa. Eight of the taxa belonged to the Actinobacteria phylum, those being Lechevalieria, Mycobacterium, Nocardioides, Actinophytocola, Actinoplanes, Kibdelosporangium, and Amycolatopsis. The other taxa are composed of Bacteroidota (genus Niastella, LDA score = 3.55), Proteobacteria (genera Shinella [3.46], SM2D12 [3.09], and Microvirga [3.07]), Bdellovibrionota (genus Bdellovibrio, 3.13), Myxococcota (genus Sandaracinus, 3.08). Genotype Sweet Sensi only had three relevant taxa: two from the phylum Firmicutes, genera Paenibacillus (4.16) and Shimazuella (2.97), and one from the phylum Proteobacteria, genus Pseudomonas (4.78). The Cherry Wine genotype had nine relevant taxa from phyla Proteobacteria, Actinobacteria, Bacteroidota, Chloroflexi, and Myxococcota. The genera of these phyla are Rhizobium (4.26), Methylophilus (3.97), Sphingobacterium (3.72), Nonomuraea (3.40), Cellbivrio (3.28), and Stenotrophomonas (3.26). Lastly, genotype American Victory-1 had four taxa identified family Kineosporiaceae of the phylum Actinobacteria (4.84), order Armatimonadales (3.05), genus Pir4_lineage of phylum Plantomycetota (3.60), and genus Subgroup_10 of family Thermonanaerobaculaceae (3.02).

However, a significant taxon was present in the other organs (leaf, root, and soil), including fungal taxa identified within these parameters. Soil fungal communities based on genotype had the least variation with only five differential taxa, two belonging to Unknown, another to Sweet Sensi, and one for Cherry Wine. For the Unknown genotype genera, Mariannaea (3.81) and Calvatia (3.57) were represented, and Sweet Sensi Neocosmospora and Thanatephorus were represented (4.91, 3.68, respectively). Lastly, the genotype Cherry Wine had only one discriminatory fungal genus, Gibberella (4.80). Moving up the plant into the root system, nine significant taxa were found overall. The unknown genotype had genera Botryotrichum and Xylaria, and unidentified from order Branch06 of class Sordariomycetes along with family Cerotavasidiaceae (5.06, 4.86, 4.35, and 3.91, respectively). Cherry Wine genotype had genera Neocosmospora (5.20), Athelia (5.16), and Pseudallescheria (4.01) shown as discriminatory taxa. American Victory-1 and Sweet Sensi genotypes had only one distinct taxon found: Candida (4.71) and Alternaria (5.04), respectively. Finally, we will describe the taxa of relevance within the leaves of different genotypes. Sweet Sensi was the most distinct, with four genera of distinction: Zymoseptoria, Meira, Dimorphiseta, and Pseudozyma. Cherry Wine had only one taxon, the family Bionectriaceae, with an LDA score of 3.67. American Victory-1 had two distinct taxa: one of the genera Aspergillus (3.77) and one of the families Chaetomiaceae (3.29). The unknown hemp genotype had no distinct taxa in their leaf fungal communities.

Microbiome diversity based on genotype traits: CBD and fiber-richness

Grouping the genotypes by their production use of either Fiber or CBD, we discovered that only the roots and soil had significantly different bacterial communities (p < 0.002). The fungal communities differed significantly across the soil, root, and stem but not the leaf communities (p < 0.003, 0.05, 0.03, and 0.30, respectively; Fig. 6). Running LEfSe analysis (α = 0.01) on the samples when grouped by production types, some taxa were identified as differentially abundant between the two groups (Fig. 7). Only the rhizosphere had differentially abundant taxa with the strict alpha values (0.01) set due to the small sample size and to reduce false positive findings. Soil bacterial diversity was very similar. The bacterial root diversity was found to have multiple bacterial taxa and was significantly upregulated in Fiber production plants, with the genera Asticcacaulis (3.05) and Shinella (3.37), Saccharimonadales (3.25), Nonomuraea (3.28), Nocardioides (3.57), Niastella (3.35), Mycobacterium (3.95), Blrii41 (3.13), Aeromicrobium (3.77), and Actinoplanes (3.63).

In contrast, fungal diversity had a more significant variation in soil between the plants, with 14 differentially abundant taxa. The genera Neocosmospora (4.77) and Myrothecium (3.33) are of note, with the most significant differentiation in CBD-producing plants. The fungal taxa in the soil near Fiber-producing plants had four distinct genera: Fusicolla (3.88), Cladosporium (3.17), and Triangularia (3.44). The fungal taxa along the root system differed significantly from the genus Pseudallescheria (3.52). The fungal root community for CBD production plants had relevant taxa of genera Alternaria (4.78), Stemphylium (3.17), Candida (4.46), Condenascus (3.64), and Achroiostachys (3.22).

Core-microbiome player of hemp genotypes

Investigating the core microbiome, we found significant variation across CBD vs Fiber genotypes. We found 16 and 11 core microbiome bacterial and fungal species across genotypes (Table S7). The results showed that Sphingomonas, Pseudomonas, and Bacillus were the core microbiome species of fiber genotypes with a highly abundant presence compared to CBD genotypes. Microbacterium and Rhizobium were significantly higher in CBD than fiber. In the case of the fungal microbiome, the Alternaria and Gibberella formed a core microbiome of fiber genotype rather than CBD. Contrarily, Penicillium and Nigrospora were significantly more abundant in CBD than in fiber genotypes (Table S9).

Elucidating the microbial community structures associated with the industrially important crop is critical because it’s the pivotal stride toward understanding the crosstalk between a plant and its microbiome. It is important to emphasize that the microbiome influences crop health, yield, and development. The current study evaluated the bacterial and fungal microbiome in the phyllosphere (leaf and stem) and rhizosphere (root and soil) of 4 different hemp cultivars, including two CBD and two fiber-producing varieties. We evaluated and identified the dominant taxa across four compartments, characterized the core microbiome of all four varieties, and performed comparative analysis using CBD vs fiber varieties. We found that the plant compartments contain an abundance of fungal and bacterial taxa. By comparing the rhizosphere and phyllosphere ASVs, the rhizosphere (root and soil), AVSs were significantly higher in bacterial taxa compared to the phyllosphere with high fungal taxa ASVs. Microbiome variations observed in each compartment were evaluated in all four genotypes. In the bacterial microbiome, the leaf compartment showed higher variation in bacterial ASVs, where CBD and fiber-producing varieties had 53 and 33 unique AVSs, respectively, with 108 shared ASVs.

Similarly, in root and stem compartments, the bacterial ASVs variations were higher, although the soil compartment of both CBD and fiber-producing genotypes had shared ASVs. Furthermore, the leaf compartment also had the highest variation in fungal ASVs, with CBD having 16 and fiber producing 36 unique ASVs with 230 shared ASVs. The fungal ASVs showed the most negligible variation in stem and soil compartments. The plant compartment and field conditions based on bacterial and fungal ASVs variation in hemp genotypes have also been reported previously46,47,48.

The beta diversity analysis was performed via PCoA using the Bray–Curtis distance matrix and Hellinger method. The CBD-producing variety Sweet Sensi and fiber-producing American Victory showed higher distances in bacterial and fungal biomes, respectively. This result can be attributed to the plants being grown in the same field, meaning similar soil conditions, and having less influence on soil microbes the further they are from the root system49,50,51. The hemp microbiome originated from soil and is mainly filtered and localized into various compartments. A similar pattern of low-distance matrix was also reported in other hemp varieties52. In the case of diversity indices, it was shown that recruiting different microbiomes in different compartments by trait-specific varieties is highly ideal, which conforms with our study53.

The Cherry wine and Sweet sensi had Methylobacterium-methylorubrum as the dominant genera, followed by Pseudomonas, Xylella, and Sphingomonas in the leaf compartment. The American victory and Unknown genotypes had Pseudomonas as a highly abundant genus, followed by Methylobacterium and Xylella occupying the leaf part. The American victory and Cherry wine stem compartment showed a high abundance of Xylella, whereas the unknown genotype had Bacillus as the top abundant genera in the stem. The root compartment of the genotypes showed the highest abundance of the Bacillus genus, followed by Pseudomonas and Streptomyces. However, the soil compartment showed a similar abundance pattern, and the statistical differences were not variable for the soil part of the hemp genotypes.

Interestingly, a fiber-producing hemp microbial diversity study reported Methobacterium, Pseudomonas, Bacillus, and Sphingomonas as abundant taxa, which agrees with our results54. These genera potentially facilitate the separation of fibers from other stem tissues by producing hydrolyzing enzymes and improving the retting of industrial hemp54. Moreover, plant microbiomes are potential players in protecting the host plants from biotic and abiotic stresses55, acquiring immunity56, suppressing diseases57, and supplying nutrients58. The microbiota in the rhizosphere and phyllosphere significantly influence the plant metabolome by synthesizing bioactive compounds and signaling molecules that influence the metabolic pathways59,60. Bacteria and fungi in the rhizosphere produce enzymes, vitamins, and phytohormones that facilitate plant growth, development, and nutrient acquisition61,62 and produce antibiotics and siderophores that help suppress pathogens. Similarly, in the phyllosphere, microbiomes produce antibiotics, enzymes, volatile organic compounds, and phytohormones that help and promote plant growth and development63. Interestingly, endophytes promote the plant’s secondary metabolite production by supplying the plant with essential nutrients and phytohormones; this symbiotic relationship plays a pivotal role in the accumulation of cannabinoids in CBD-producing hemp varieties64.

Similarly, Pseudomonas is reported to be highly abundant and a core microbiome genus in the cannabis microbiome, and it correlates with secondary metabolite production, including CBD production65. Furthermore, Scott et al.66Pseudomonas and Bacillus are highly abundant genera in industrial hemp. Similarly, Sphingomonas and Methylobacterium are highly associated with hemp cultivars and have been shown to promote plant growth and pathogen resistance48. Sphingomonas is highly abundant in hemp plants2 and, specifically, the phyllosphere52. Recently, a potentially novel strain of Sphingomonas was isolated from Cannabis sativa67. This highlights the potential for these dominant genera to play a role in CBD-producing genotypes, but further research is required. The Methylobacterium is reported to play a key role in modulating the production of flavor-related phytometabolites in plants68.

In the current study, phyllosphere fungal genera Septoria, Alternaria, Gibberella, Moesziomyces, and Cladosporium were abundant, whereas, in the rhizosphere, Neocosmospora, Alteraria, Septoria, and Fusarium were abundant genera across all four genotypes. Contrarily across genotypes, only Cherry wine had Athelia (in root) and candida (in soil) genera, which were also highly abundant48,66,69,70. Interestingly, we found Alternaria, Fusarium, and Cladosporium, genera that are pathogenic to plants. Alternaria and Cladosporium are also reported as human pathogens in cannabis-associated microbe studies64,71, but the involvement of cannabis and its products in these diseases is not reported. Moreover, Gibberella, Septoria, Moesziomyces, Candida, and Alternaria are endogenous plant symbionts that can help plants in secondary metabolite production and promote plant health, growth, and development72.

In the current study, we detected eleven core bacterial species of Bacillus, Pseudomonas, and Sphingomonas genera widely distributed across different hemp genotypes. Previously, these genera have been reported multiple times as core bacterial communities48,65,66. Recently, these species have been manipulated and applied to various crops in rhizospheric soil, enhancing yields and sustained agroecosystems for different crops, such as corn, wheat, and soybeans53. Thus, these species could be pivotal in hemp secondary metabolites production, such as CBD, growth, and development. Similarly, Alternaria and Gibberella were abundant in fiber-producing varieties, whereas Penicillium and Nigrospora were in CBD genotypes.

Moreover, the Penicillium species are reported to enhance plant growth properties and help plants in drought conditions73. Surprisingly, American Victory (fiber) and Sweet Sensi (CBD) were similar in bacterial leaf communities and fungal rhizosphere communities, with the most variation between CBD and fiber-producing plants found in the bacterial soil and root biomes. Fungal microbes varied in all plant compartments except in the leaf compartment. This shows a potential for fungal communities to play a role in secondary metabolite production.

In this study, we evaluated and characterized four genotypes’ rhizosphere and phyllosphere microbiomes and compared the microbiome variation in fiber vs CBD-producing hemp varieties. The microbiome variation across the genotypes, plant site (phyllosphere vs. rhizosphere), and nature (fiber vs. CBD) advances our knowledge in understanding the modulating role of all these factors in the hemp microbiome. In a natural field, we have tried to establish microbial diversity, community structure, and interkingdom interactions between bacteria and fungi in hemp phyllo and rhizosphere. The results showed that phyla Ascomycota, Basidiomycota, and core microbiome genera Alternaria, Penicillium, and Gibberella were significantly more abundant in CBD-rich genotypes than fiber-rich genotypes. Similarly, the concentric phyla Proteobacteria, Acidobacteria, Actinobacteria, and the core-bacterial genera were Sphingomonas, Pseudomonas, and Bacillus. This suggests that specific genotypes recruit specialized microbial communities across different spheres of association. Our results emphasize the variation in the native microbiome pool in the local soil environment and plant genotype74. This study has several prospective new approaches, such as:

(1)

Investigate the specific endophytes that positively correlate with increased metabolite synthesis in industrial Hemp. Employ advanced genetic and microbial engineering techniques to enhance the abundance or activity of these beneficial endophytes. This could lead to developing symbiotic relationships that boost the production of desirable metabolites such as CBD (cannabidiol) or other valuable compounds (terpenes).

(2)

Explore the endophytes that positively impact fiber quality in Hemp plants. Investigate how these endophytes influence the plant’s structural components and work towards optimizing these interactions. This could result in the development of Hemp cultivars with improved fiber properties.

(3)

Delve deeper into how specific endophytes contribute to the plant’s stress resistance. Investigate their role in mitigating environmental stresses like drought, pests, or diseases. Understanding these mechanisms can inform the development of resilient Hemp varieties capable of thriving in varying agroecological conditions in Texas.

4)

By understanding how these endophytes facilitate nutrient absorption and utilization in industrial hemp plants, we can develop strategies to optimize nutrient management in the crop. This has the potential to increase overall plant health and productivity.

(5)

Validate the efficacy of endophyte-mediated enhancements in real-world agricultural settings. Develop endophyte-inoculated industrial hemp seeds for commercial production. This step will be crucial in translating laboratory findings into practical applications, ensuring the scalability and viability of endophyte-assisted cultivation practices on a larger scale.