Soil and seed both influence bacterial diversity in the microbiome of the Cannabis sativa seedling endosphere

2.2. Sources of soils

“Innotech” topsoil was collected from a field at InnoTech Alberta (Vegreville, AB) (53.504377, -112.088324) in which the hemp cultivar Silesia was been grown. “Okanagan” soil was harvested from Westpoint Vineyard in Kelowna, BC (49.821298, -119.470140) growing Chardonnay’ (Vitis vinifera L.) wine grape Clone 99 (121) on S04 (V. vinifera) rootstock. These soils were selected to represent both a soil in which cannabis has already been grown, and soil in which cannabis has not been previously grown. To sterilize the soils, a small volume of each soil type was sealed in a 2 L glass mason jar and sterilized by two successive cycles of autoclave sterilization for 55 minutes at 121°C, and the start of each cycle of autoclaving was separated by 24 hours (Carter et al., 2007). Absence of culturable microorganisms in autoclaved soil was confirmed by spreading approximately 1 g of soil from each jar on Petri dishes containing R2A agar in a laminar flow hood and inspecting plates for visible colonies after 3 days incubation at 25°C.

2.3. Seedling growth and experimental design

Each of four types of soil (Innotech non-sterilized, Innotech sterilized, Okanagan non-sterilized, Okanagan sterilized) were distributed into sterilized 25 mm x 200 mm borosilicate test tubes. Working in a laminar flow hood, soil was added to fill tubes to 5 cm height (approximately 25 mL of soil). One seed of each of the three genotypes (BC Big Bud, Katani, X59) was placed in the soil in each tube, watered with 1 mL sterile RO water, and the tubes were sealed with cotton balls. The tubes were placed under fluorescent light at 8250 lux, 7 days prior to removal of the seedlings. There were four replicates for each combination of the four soil types and three cannabis genotype, thus 48 seedlings per replicate. Two independent replicates were conducted for a total of 96 seedlings analyzed.

2.4. Seedling surface sterilization and DNA extraction

After removal from soil, seedlings were first washed with filter-sterilized 0.1% Triton-X for 10 min, and then 3% sodium hypochlorite for 20 min. Samples were drained and rinsed with autoclaved, deionized water, followed by treatment with 70% ethanol for 10 minutes. Samples were rinsed 5 times to remove ethanol before being frozen for DNA isolation. An aliquot of the final rinse was plated on R2A medium to test sterility. DNA from whole 7-day old seedlings (including both roots and shoots) was isolated using the Qiagen DNeasy Plant Pro kit according to the manufacturer’s instructions and using a Powerlyser24 homogenizer at 1500 rpm for 1 min, then 2000 rpm for 1 min.

2.5. 16S rRNA amplicon sequencing

The V4 hypervariable region of the 16S rRNA gene was amplified using barcoded primers with Illumina adapters: 515FB: GTGYCAGCMGCCGCGGTAA, 806RB: GGACTACNVGGGTWTCTAAT. 1 μM pPNA and 1 μM mPNA (PNA Bio Inc, Newbury Park, CA) were included in the PCR to inhibit amplification of plant chloroplast and mitochondria sequences. The PCR cycle was as follows: 95°C for 5 min, followed by 35 cycles of 95°C for 30 s, 55°C for 30 s and 72°C for 50 s, and then extended at 72°C for 10 min. Amplicons were visualized on a 1% TAE agarose gel. Amplicons were cleaned using Mag-Bind^® TotalPure NGS magnetic beads (Omega), followed by indexing PCR with i7 and i5 primers (95°C for 3 minutes followed by 8 cycles of 95°C for 30 seconds, 55°C for 30 seconds, 72°C for 30 seconds, followed by 72°C for 5 minutes). Indexed amplicon libraries were bead-cleaned a second time before quantification using a Nanodrop spectrophotometer. Equimolar samples were then pooled and sent to Genome Quebec for DNA sequencing using the Illumina MiSeq Reagent Kit v3 (2×300 cycles).

4.1. Seed-dependent effects on seedling endosphere

In the present study, the variety (genotype) of seed was an experimental factor comprising both the functional properties of the plant and the microorganisms it vectors from its parent. An effect of genotype on beta-diversity in rhizosphere, endorhizosphere, and phylosphere has previously been demonstrated in three types of vegetatively propagated drug-type cannabis (Comeau et al., 2020). This is consistent with our observation of a significant effect of cannabis genotype in beta-diversity (Figure 6, Supplementary Table S1). Interestingly, within a given sterilized soil treatment, PERMANOVA based on Bray-Curtis distance matrix found genotypes to be significantly different from each other (Supplementary Table S1), suggesting that differences in the genotypes influenced the microbial communities of their endospheres.

Our results also showed that a core microbiome of bacterial endophytes existed across samples: Gammaproteobacteria and Bacilli, the most abundant taxa, were found in all treatments (Figures 2, 3). Betaproteobacteria, Flavobacteriia, and Alphaproteobacteria were also found across all soil treatments but typically at lower abundance. This suggests that some portion of seed-borne endophytes inhertited from the seed-bearing parent were maintained whether seedlings were germinated in the presence or absence of soil microorganisms.

4.2. Soil-dependent effects on seedling endosphere

We observed that the Shannon alpha-diversity of autoclave sterilized soils was significantly lower than non-sterilized soils, when all genotypes and soil sources were grouped together (Figure 4A). This is unsurprising as non-sterilized soil contains abundant microorganisms that can colonize seedlings and alter the endosphere microbiota. This leads to speculation about the impact of growing drug-type cannabis in soilless medium, as is common practice for licenced cultivators in Canada to limit pathogens. However, when genotypes and soil sources were considered individually, only Katani plants grown in Okanagan soil showed a significant reduction in alpha-diversity of sterilized soil compared to non-sterilized soil (Figure 5). This is indicative of genotype-specific factors and the presence of particular soil microganisms in the Okanagan soil source. Other genotypes grown in soil from either Okanagan or Innotech showed non-significant responses to soil sterilization, in terms of alpha-diversity. Interestingly, when all soil treatments (i.e. both source and sterilization) were considered together, and the effect of genotype was tested, drug-type genotype Katani and X59 had significantly lower alpha-diversity as compared to the two hemp genotypes: (Figure 4B). Further experiments would need to be conducted with additional examples of drug-type and hemp-type cannabis to see whether this difference between these two types of cannabis is generalizable.

The relative abundance of many taxa were significantly different between endospheres of seedlings grown in sterilized soil as compared to non-sterilized (Figures 2, 3). Genotype had clear and significant effect on Bray-Curtis distances, however, cannabis genotypes as a whole clustered into their soil treatment groups on the PCoA (Figure 6). Genotypes planted in untreated Okanagan and Innotech soils were dramatically different from one another, while the genotypes grown in the sterilized soils clustered far closer together, and so it is unsurprising that they would have an effect on endosphere communities. Meanwhile, the closer clustering of cannabis genotypes grown in both sterilized Okanagan and Innotech soil is likely a result of seed-borne endophytes being the sole colonizers of niche (Figure 6).

There were clear differences in relative abundance of some taxa in the endosphere when cannabis seedlings were grown on sterilized versus non-sterilized soil. In the absence of viable soil inocula, cannabis seedlings were significantly enriched in Betaproteobacteria (1.42x) (Supplementary File 2). Betaproteobacteria are a class of gram-negative bacteria with important roles in cycling nitrogen, sulfur, and other nutrients (Kowalchuk et al., 1997; Schmalenberger and Kertesz, 2007; Falca et al., 2010). Some also have plant growth-promiting properties, and their abundance in the potato rhizoshere is cultivar dependent and plant growth-stage dependent (Falca et al., 2010). In high-starch potato cultivars, abundance of Betaproteobacteria in the rhizosphere increased towards the flowering stage of the plant. Perhaps cannabis-cultivars show a similar effect which leads to an enrichment of these bacteria in flowering tissue, allowing them to be vectored to seedlings, as was inferred from our results.

Seedlings grown in sterilized Okanagan soil were significantly enriched in Bacilli, as compared to seedlings grown in non-sterilized Okanagan soil. This enrichment was as large as 49X when all genotypes were considered as a group, and was also significant in the two hemp genotypes individually (in X59 when rare taxa (<50 frequency) were filtered), although the enrichment in BC Big Bud (1.8X) was insignifcant. The increase in Bacilli between sterilized and non-sterilized Innotech soil was generally not as large or significant as Okanagan soil, perhaps because of abundant Bacilli in the non-sterilized Innotech soil. Bacilli are a class of economically and agriculturally relevant bacteria, some of which promote plant growth and create a variety of antimicrobial secondary metabolites (Ongena and Jacques, 2008; Cochrane and Verderas, 2014; Hashem et al., 2019). Additionally, they form robust endospores capable of surviving harsh environental conditions for long periods of time (Nicholson, 2002; Hashem et al., 2019). Interestingly, Bacilli spores have been revived from 25-40 million year old amber (Raul and Monica, 1995). It is therefore unsurprising that this class of bacteria would persist in cannabis seeds (which can be dormant for several years) and be vectored across plant generations. Previous research demostrated cannabis plants vector culturable Bacillus species to seedlings and that these bacteria antagonize seed-borne pathogens (Dumigan and Deyholos, 2022).

Gammaproteobacteria showed the most dramatic differences in abundance in endospheres of cannabis seedlings grown in non-sterilized soils as compared to sterilized soils. This taxon was enriched 329x in non-sterilized soils, when all genotypes and soil sources were considered collectively. Gammaproteobacteria is a class of bacteria commonly associated with plants, and has even been identified as an indicator of a healthy microbiome and suppressive of wilt caused by Fusarium spp. in banana plants (Köberl et al., 2017). Pseudomonas, a genus within this class, is the source of bacteria with biocontrol activity against cannabis fungal pathogens (Balthazar et al., 2022). Interestlingly, in an previous report, Pseusdomonas was found to have plant growth promotion activity, but only when inoculated with a consortia of Bacillus and Pseudomonas species (Comeau et al., 2021). It is difficult to study the activity of these-non cultured Gammaproteobacteria (Dumigan and Deyholos, 2022), however, this previous research leads to questions about potential evolution of plant growth promoting and antifungal consortia that may be vectored on seed or obtained from soil to protect emerging cannabis seedlings.

Flavobacteriia and Sphingobacteriia were also differentially abundant and higher in endospheres of cannabis seedlings grown in non-sterilized Okanagan as compared to sterilized Okanagan soil (Supplementary File S2). Interestingly, this was not seen in Innotech soil, where Flavobacteriia were significantly enriched only in the genotype X59 when compared sterilized Innotech soil. This is likely a result of an abundance of the classes of bacteria in Okanagan soil. Flavobacteriia are often found in root-associated microbiomes, but the factors influencing their abundance are unclear. A recent study found that abundance of Flavobacterium (a genus of Flavobacteriia) is correlated with pectin availability in plant-associated communities (Kraut-cohen et al., 2021). Cannabis fiber is composed of 1-17% pectin (Liu et al., 2017), and perhaps soil Flavobacteriia species have evolved to associate with these plants for their pectin content. The genus Sphingobium from the class Sphingobacteriia has been well-studied for its biodegradative properties and for bioremediation (Kunihiro et al., 2013; Boss et al., 2022). They are also known antagonize pathogenic Pseudomonas (Takeuchi et al., 2001).

Abiotic soil characteristics could also be inferred to affect microbiome composition (Figure 6; Supplementary Table S1). This is demonstrated by the differences endosphere beta-diversity observed when seedlings were grown on sterilized soils, each from two different sources. Physical properties such as the sand, silt, and clay content may influence this, as well as the differences in organic matter, pH, and nutrient content in the Okanagan vineyard soil versus the Innotech hemp farm soil.

The authors thank Dr. Jan Slaski for providing hemp germplasm for this study.