Brassica oleracea var. acephala (kale) improvement by biological activity of root endophytic fungi

Brassica oleracea var. acephala (kale) is a cruciferous vegetable widely cultivated for its leaves and flower buds in Atlantic Europe and the Mediterranean area, being a food of great interest as a "superfood" today. Little has been studied about the diversity of endophytic fungi in the Brassica genus, and there are no studies regarding kale. In this study, we made a survey of the diversity of endophytic fungi present in the roots of six different Galician kale local populations. In addition, we investigated whether the presence of endophytes in the roots was beneficial to the plants in terms of growth, cold tolerance, or resistance to bacteria and insects. The fungal isolates obtained belonged to 33 different taxa. Among those, a Fusarium sp. and Pleosporales sp. A between Setophoma and Edenia (called as Setophoma/Edenia) were present in many plants of all five local populations, being possible components of a core kale microbiome. For the first time, several interactions between endophytic fungus and Brassica plants are described and is proved how different interactions are beneficial for the plant. Fusarium sp. and Pleosporales sp. B close to Pyrenophora (called as Pyrenophora) promoted plant growth and increased cold tolerance. On the other hand, isolates of Trichoderma sp., Pleosporales sp. C close to Phialocephala (called as Phialocephala), Fusarium sp., Curvularia sp., Setophoma/Edenia and Acrocalymma sp. were able to activate plant systemic resistance against the bacterial pathogen Xanthomonas campestris. We also observed that Fusarium sp., Curvularia sp. and Setophoma/Edenia confered resistance against Mamestra brassicae larvae.


Results
Fungal diversity. From the 900 root-fragments (180 per B. oleracea accession), 376 fungal isolates were obtained, at a rate of 54-98 fungal isolates per accession. Isolates were obtained from 41.77% of the root fragments plated. All sampled plants harbored fungi in their roots, and on average, 13 isolates were obtained from the roots of each plant.
The isolates were grouped into morphotypes, obtaining 27 morphotypes for MBG-BRS0106, 27 for MBG-BRS0292, 45 for MBG-BRS0426, 33 for MBG-BRS0446 and 21 for MBG-BRS0468. After sequencing one or more isolates of each morphotype, 179 ITS1-5.8S-ITS2 nucleotide sequences were obtained. After those sequences were clustered, considering that those being 97% or more identical belonged to the same taxon, 33 different fungal taxa were identified (Table 1). A phylogenetic tree made with the sequences of these 33 taxa helped to accommodate into taxonomic orders isolates whose sequences were less than 95% identical to those of a type strain (Supplementary Fig S1).
The most prevalent taxa were a Fusarium sp. found in 63.3% of the plants collected, and Pleosporales sp. A, found in 56.7%, both were present in all B. oleracea accessions (Table 1). Thanks to the creation of the phylogenetic tree and the comparison of the representative sequence of the proposed taxon Pleosporales sp. A, we have determined how this sequence represents a species not yet described, between the genus Edenia and Setophoma (called from now Setophoma/Edenia). The distribution of the fungal taxa showed few commonalities among the kale accessions (Fig. 1). MBG-BRS0426 and MBG-BRS0468 have 5 fungal taxa in common, 2 of which were not shared with any other accession. Several taxa occurred in only one accession: 10 taxa in MBG-BRS0426, 5 taxa in MBG-BRS0106, 4 taxa in MBG-BRS0292, 4 taxa in MBG-BRS0446, and 3 taxa in MBG-BRS0446. The similarity analysis based on the Jaccard index (JI) ( Table 2), shows that only the accessions MBG-BRS0292 and MBG-BRS0446 present a similarity between their fungal taxa greater than 0.3, while above 0.25 we find MBG-BRS0292 and MBG-BRS0468, MBG-BRS0426 and MBG-BRS0468, and MBG-BRS0446 and MBG-BRS0106.
Pleosporales is the most representative order in terms of the number of taxa (32.3%) ( Table 1). Most taxa are present only in one of the accessions (58.8%), with only 5.9% of the taxa present in all (Table 1). Taking into account the incidence of each order in each variety (Fig. 2), it can be seen that the most abundant order is Pleosporales (27-68%), followed by Hypocreales (7-34%), both present in all varieties. Helotiales is present in all accessions, except in MBG-BRS0106, where we find Diaporthales as an specific order. On the other hand, the distribution of taxa according to their incidence can be visualized in the rank-abundance curve shown in Fig. 3 Increased plant growth, tolerance to cold and resistance to biotic stresses. Nine fungal isolates were used to study beneficial effects on kale. Significant differences in the growth of the aerial part of kale were observed with the application of Fusarium sp. (5.66 g) and Acrocalymma sp. (4.83 g), being their weight almost double that observed for the control plants without inoculation (2.92 g). Growth of plants inoculated with Pleosporales sp. B next to Pyrenophora (called from now Pyrenophora) and Curvularia sp. was also significantly higher than that of the control plants. Likewise, although there was no significant difference respect to the control, we observed a reduction in the weight of the plants inoculated with Chaetomium sp. (1.71 g) (Fig. 4a); as well as dry weight (Fig. 4b).
Under cold conditions (constant 12 °C) is observed how inoculation with Pyrenophora, Chaetomium sp., Helotiales sp.next to Phialocephala (called from now Phialocephala), Diaporthe sp., Fusarium sp. and Acrocalymma sp. caused a significant increase in growth, and therefore, of tolerance to abiotic stress. Inoculation with Phialocephala, Fusarium sp. and Acrocalymma sp. increased the plant weight almost twice (2.72 g, 2.82 g, and 2.78, respectively) respect to uninoculated control plants (1.47 g) (Fig. 4c); as in dry weight (Fig. 4d). www.nature.com/scientificreports/ Regarding the activation of systemic resistance against biotic stressors, leaves of B. oleracea were inoculated with the pathogenic Xcc bacteria (Fig. 5). We observed a significant decrease in the damage caused by the bacteria at 8 d.p.i (Fig. 5a) in the plants inoculated with Trichoderma sp., Phialocephala, Fusarium sp., Curvularia sp., Setophoma/Edenia and Acrocalymma sp., being highly significant (P < 0.01) with the inoculation of the first four fungi indicated. On the other hand, at 15 d.p.i. (Fig. 5b) we only obtained a significant reduction in the incidence of the disease in the plants with the inoculation with the Setophoma/Edenia and Acrocalymma sp. strains.
Regarding the possible activation of systemic resistance in the inoculated plants against the insect pest Mb, the inoculation of kale with the fungi Setophoma/Edenia (DI 2.3) and Fusarium sp. (DI 2.2) strains supposed a significant decrease in the damage index, compared to the control plants without inoculation (DI 3.1). Furthermore, although no significant differences were observed with respect to the leaf area consumed by the larvae, both Setophoma/Edenia and Fusarium sp. had the lowest average leaf area consumed (Fig. 6).

Discussion
Few studies have investigated fungal diversity in varieties of the same plant species within a study site. In other words, whether genotypic variation affects the structure of the root microbiome. Regarding the B. oleracea accessions of our study, we have reported a possible specificity of accession. Most of the determined fungal taxa are only present in one accession and the similarity indices give great differences despite the fact that all the plants Table 1. Core and abundant fungal species isolated from surface sterilized roots of B. oleracea (kale) from 5 different accessions. A core microbiome is defined as the group of microbes commonly found within a host's microbiome, using persistence of the association as the criterion to select microbes potentially providing critical function within the habitat in which they are found 67 . www.nature.com/scientificreports/ are in the same geographical plot. These differences could be due to differences in the root exudates produced by each kale accession, which could affect the ability of endophytic fungi to colonize the roots, just as it has been observed in different tomato genotypes and their rhizospheric microbiota 19 . Despite this, it is very difficult to consider the existence of such specificity in a study like this, since the possibility of detecting those fungi with low prevalence is more difficult than those that are present to a greater extent, also the great differences existing at the level of alpha (Table 1, Fig. 2) and beta ( Table 2) microbial diversity. On the other hand, the presence in all accessions of Fusarium sp. and Setophoma/Edenia, suggests that those two taxa could be components of a core microbiome of kale. This could be confirmed by surveying kale root endophytes at different locations. The diversity of endophytic fungi present in the roots of plants of the Brassica genus has been scarcely studied so far, as has the diversity of fungi between different varieties of the same species. In our study, most isolated fungi belonged to the Pleosporales order. To our knowledge, only one species of the Pleosporales order, Leptosphaeria biglobosa, has been found in B. napus 13 . Besides, order Hypocreales was found in all accessions. This order is more common in Brassicaceae, and taxa such as Fusarium sp. 20 , C. rosea 13 and M. anisopliae 15 have been reported in B. napus. Mucorales order was found in 3 of the 5 accessions, having previously being reported Mucor sp. in B. campestris 21 and B. chinensis 22 . Sordariales order was only found in MBG-BRS0292 and MBG-BRS0468 accessions, having previously being described the fungus Chaetomium globosum in B. napus roots 13 . And Eurotiales order was only found in the MBG-BRS0446 accession, having previously reported the fungus Penicillium sp. 20 , A. flavipes 13 and A. capensis 14 in B. napus roots. Therefore, our study reports the first description of the Helotiales, Capnodiales, Diaporthales, Chaetosphaeriales and Minutisphaerales orders in roots of a plant of the Brassica genus; Heliotales order was widely present in 4 of the 5 accessions.
In our study carried out with different accessions of B. oleracea (kale), we have been able to determine the presence of two dominant taxa in all the accessions examined (Fusarium sp. and Setophoma/Edenia). Species of the genus Fusarium, such as F. oxysporum can be neutral, beneficial, or detrimental for host plants. Fusarium wilt is one of the most devastating diseases in agriculture, since some strains can also cause vascular wilts resulting    www.nature.com/scientificreports/ in B. oleracea 27,28 . So, this is the first study that describes Fusarium as an endophyte of brassica roots, although Fusarium oxysporum and other species of the genus have been described as dominant components of the microbiome of several plant species 29 .
As regards the taxa Setophoma/Edenia, there is information about pathogenic and endophytic species in both genus. Setophoma includes disease-causing pathogenic fungal species such as leaf spot in Camellia sinensis    31 , being the only report of S. terrestris-Brassica interaction described so far. On the other hand, S. terrestris has been described as an endophyte of other plant species such as Gloriosa superba 32 and horseradish (Armoracia rusticana), where is able to decompose various glucosinolates 33 . In turn, Edenia is a well-known genus of endophytic fungi that produce compounds of high medical interest such as anti-inflammatories 34 or antileshmanials 35 . Furthermore, the E. gomezpompae species is capable of producing highly toxic compounds for plants, which are used as natural herbicides 36 . Both within the Setophoma genus and the Edenia genus, our isolate represents the first endophyte identified in roots of Brassica plants.
An Acrocalymma sp. was present in 3 of the 5 accessions studied. A. medicaginis has been described as a causal agent of crown and root rot in Medicago plants 37 . However, A. vagum is a dark septate endophyte able to decrease heavy metal content in tobacco 38 and to promote plant growth in Glycyrrhiza uralensis 39 , but never before reported in Brassica plants.
Therefore, in the present study we have been able to identify several taxa of endophytic fungi in kale roots, which had not been described so far as endophytic fungi (although in some cases as pathogens) within the Brassica genus.
Once the different endophytic fungi in the kale roots were identified, we made a selection of them to determine their possible biological role in the plant. To do this, we inoculated kale plants and tried to determine if the endophytes were capable of promoting plant growth, increasing cold tolerance, or inducing systemic plant resistance against pathogens and/or pests.
The ability of different species of endophytes to promote plant growth as well as tolerance to abiotic stresses such as cold, or resistance to pathogens has been described 40 . In our study, we have observed that the Pyrenophora strain was capable of promoting the growth of the plant and its cold tolerance. This would be the first time that a benefit of the Pyrenophora-plant interaction has been described, this genus is best known for pathogens, such as Pyrenophora teres, causal agent of net blotch of barley 41 . The beneficial role observed in B. oleracea by Pyrenophora could be due to the defensive capacity of cruciferous plants through secondary metabolites such as glucosinolates, not present in other plant groups, which is why a possible pathogen in a species may be a harmless symbiont in Brassica. Moreover, inoculation with Fusarium sp. in B. oleracea also increased its growth and cold tolerance, more significantly than with Pyrenophora. Although there are numerous studies that demonstrate www.nature.com/scientificreports/ the ability of non-pathogenic strains of Fusarium to increase the resistance of plants against various pathogens, such as F. oxysporum 23 , their role as a possible promoter of plant growth or cold tolerance has not been studied. The interaction B. oleracea-Acrocalymma sp. resulted in a significant increase in plant growth, something previously observed in other plants such as Glycyrrhiza uralensis 39 , Medicago sativa and Ammopiptanthus mongolicus 42 in interaction with A. vagum. Similar results have been reported in B. oleracea-Curvularia sp. interaction, fungal genus that, although it includes important plant pathogens 43 , also has species classified as endophytes that promote plant growth, such as C. geniculata in Parthenium hysterophorus through phosphate solubilization and phytohormone production 44 .
Regarding the increase in cold tolerance, we have observed that the B. oleracea-Chaetomium sp. interaction results in a significant benefit for the plant. In this sense, several authors have described different mechanisms such as heat shock proteins or antioxidant compounds that, for example, the endophyte C. globosum has to tolerate cold, some of which may also be beneficial for the plant thanks to the interaction [45][46][47] . The interaction of B. oleracea with Phialocephala also produced an increase in plant tolerance to cold, as has been verified in different subarctic herbaceous plants by Phialocephala fortinii 48 . Also, we have observed the capacity of Diaporthe sp. and Acrocalymma sp. to increase the tolerance of B. oleracea to cold, being the first time, even within their fungal genera, that these beneficial capacities for plants have been described.
After infection with Xcc, we quantified a decrease in the harmful effect of the bacteria on B. oleracea plants preinoculated with Trichoderma sp. The ability of different Trichoderma species to activate plant systemic resistance against different pathogens has been widely proven in various plant species 49 , including Brassica plants as B. napus with T. harzianum 50,51 or T. viride 52 , B. rapa with T. pseudokoningii 53 , and even B. oleracea var. capitate with T. harzianum against Rhizoctonia solani 54 .
As far as the Fusarium sp. strain is concerned, the ability of F. oxysporum endophytic strains to induce systemic plant resistance has been described in zucchini plants (Cucurbita pepo) against the insect Aphis gossypii 55 58 . While in the case of inoculation with Phialocephala, Curvularia sp., Setophoma/Edenia and Acrocalymma sp., the data obtained represents the first report of activation of plant systemic resistance by these fungi as root endophytes.
As conclusions, the diversity of endophytic fungi found in the kale roots of various accessions is lesser compared to non-brassicaceae species, where up to 49 fungal taxa can be found per plant 59 . In addition, several of the isolated fungal taxa have been widely described as pathogens of different crops, but never as endophytes, as species of the genus Pyrenophora, an aspect that occurs in kale probably due to their powerful defensive capacity through the use of glucosinolates against fungi 60 . On the other hand, we find taxa present in several of the accessions and even in all of them, as is the case of Fusarium sp. and Setophoma/Edenia, representing a possible core microbiome for kale. It is precisely these taxa that have shown in our preliminary studies a greater capacity to promote the productivity of kale, by promoting its growth, increasing its tolerance to cold and increasing its defensive capacity against pathogens and pests. The study was conducted in 2016 at MBG-CSIC (Galicia, NW Spain). Two hundred seeds of each accession were sown in trays in May 2016, and kept in a greenhouse under environmental conditions. After 50 days, when most seedlings were at the 4-5 leaf stage, one hundred and fifty plants of each accession were transplanted into an experimental plot randomly distributed. The distance was 0.8 m between rows and 0.6 m between plants. The experimental plot was established in a slightly acidic soil (pH close to 5.5), organic matter content of 6.3%, 100 ppm available P, 335 ppm assimilable K, 133 ppm changeable Mg, and 7.83 cmol/kg Ca 2+ . As management practices, no fungicide was applied, or any other pesticide, or any form of fertilization. To prevent the development of weeds, the soil was covered with a mesh that prevents their presence in the plot.

Materials and methods
Sampling was carried out in 30-week-old plants, randomly choosing 6 plants per accession. After digging out the entire root system, samples of the roots were collected, and stored in cold for a few hours until processing in the laboratory.

Isolation of fungi.
A subsample of 30 root fragments of 4-5 cm per plant was collected. Disinfection of roots and fungal isolation was done following the procedure described by Ref. 29 . Each root sample was washed with tap water and then surface-disinfected with a solution of 20% commercial bleach (1% active chlorine) containing 0.02% Tween 80 (v:v) for 6 min, followed by treatment with an aqueous solution of 70% ethanol for 30 s. Finally, the roots were rinsed with sterile water and cut into pieces about 5 cm long. Thirty root pieces of each sample were plated in three Petri plates (10 pieces/plate) with potato dextrose agar (PDA) (Sigma-Aldrich, St. Louis, USA) containing 200 mg/L of chloramphenicol (the antibiotic was used to avoid the isolation of endophytic bacteria), and kept in the dark at room temperature. As mycelium emerged from a root fragment into the agar, a small piece of the mycelium from the leading edge of the colony was transferred to a new PDA plate and maintained at room temperature. The root fragment and remaining mycelium were taken out of the original plate to avoid overgrowth. The plates with root samples were checked daily for the presence of fungi for about 4 weeks. All the sequences obtained were grouped into operational taxonomic units (OTU), considering that groups of sequences with a similarity greater than 97% belonged to the same OTU. This clustering operation was done using CD-Hit-Est software 62 . For taxonomic identification, a representative sequence of each OTU was used to search the GenBank nucleotide database using BLAST. This search was limited to sequences belonging to the ITS region from fungi type and reference material. Taxonomic identification was limited to genus rank because often occurs that ITS sequences are not accurate for species rank identification (i.e. Ref. 30 ). When the identity between our sequences and that of a type strain was lower than 95%, the genus of our sequence was considered to be unknown. Further taxonomic information was obtained by means of a phylogenetic tree made with representative sequences from each taxon. This tree was made with MEGA6 software using the maximum likelihood method with distances calculated according to the Tamura 3 parameter model 63 . Tree branch confidence values were estimated by bootstrapping with 1000 replications.
For each kale accession, the incidence of each fungal taxon in plants was calculated, relating the number of plants in which it was found against the total of plants sampled (30 root fragments per plant, 6 plants per accession, 30 plants in total), as alpha diversity analysis. Additionally, the distribution of the relative abundance of each taxon was observed with a rank-abundance curve. Furthermore, beta diversity was analyzed by comparing between accessions, using the Jaccard index.
In planta assays. Plant-fungus interaction. One of the local varieties previously used in the diversity study (MBG-BRS0106, from here on referred to as kale) was used for the reinoculation tests and the study of the biological activity of the selected fungi.
Different fungal isolates (Table 3) were chosen based on their presence in all accessions, or their proven biological activity in other plant species (bibliographic search). To inoculate plants, one part of beet pulp inoculum was mixed with seven parts (w/w) of a substrate consisting of peat moss (Profi-Substract, Gramoflor, Valencia, Spain) previously treated at 80 °C for 24 h. The fungal inoculum was a 4-week-old culture of each fungus grown in sugar beet pulp medium 64 . Uninoculated controls were transplanted to soil mixed with uninoculated beet pulp medium. By qPCR, the ability of fungi to effectively colonize kale roots could be verified (data not shown).
Growth promotion. For the analysis of plant growth promotion kale plants were inoculated in 20 L pots. The plants were kept under controlled greenhouse conditions until their aerial part was harvested in 8-week-oldplants to record fresh weight. The plants were watered 2-3 times a week, according to the observed needs, always with the same amount of water in all the plants. Exogenous fertilization was not used. Greenhouse conditions were 14 h photoperiod, environmental temperature (12-30 °C) and relative humidity above 80%. A total of 25 plants were used for each fungal isolate inoculated.
Cold tolerance. For the cold tolerance assay, we stablished the same conditions reported in Ref. 65 . Seeds of kale were planted in 1 L pots and grown under fluorescent light (228 μmol m −2 s −1 ) in a 14 h light/10 h dark photoperiod regime and watered as needed. The temperature in the cold-exposure treatment was set at 12 ± 1 °C, since www.nature.com/scientificreports/ lower temperatures reduced seed germination and seedling survival dramatically. A total of 25 plants were used for each inoculation.
Biotic resistance. Inoculation of kale leaves with the bacterial plant pathogen Xanthomonas campestris pv. campestris (Xcc) was made according to Ref. 66 . Bacterial cultures were grown in PDA at 30 °C for 48 h. A bacterial suspension was diluted with sterile water to reach a turbidity value of 0.5, which corresponds to a concentration of 5 × 10 8 cfu/mL. Turbidity of the suspension was measured with a microplate spectrophotometer (Spectra MR; DynexTechnologies, Chantilly, VA) at a wavelength of 600 nm. The second leaf counting from the apex was inoculated when plants were at a 5-6 leaf stage. Inoculation was made by biting on three veins located on the edge of the leaf with a mouse-tooth forceps wrapped in cotton submerged in the inoculum. Xcc race 1 type strain HRI3811, provided by Joana Vicente (University of Warwick, UK), was used to inoculate the plants. After the inoculation, plants were kept under controlled greenhouse conditions (20 °C at night and 28 °C during the day, and relative humidity greater than 80%). Damage was measured using a visual damage index (DI) scale ranging from 1 to 9, where 1 = no visible symptoms and 9 = severely diseased with typical V-shaped chlorotic leaf edge lesions presenting blackened veins areas. Data were taken 8 and 11 days post infection (d.p.i.). A total of 10 plants were used for each fungal strain. Infestation with the insect pest Mamestra brassicae (Mb) was carried out under the controlled greenhouse conditions above described using 12-week-old-plants. Eggs were provided by the Laboratory of Entomology, Wageningen University, The Netherlands. Egg hatching larvae were fed with fresh leaves of B. oleracea (MBG-BRS0106) ad libitum. Mb inoculation was carried out depositing 5 7-10 days old larvae on the 5-6th true leaf of the plant. After 5 days, data on leaf area consumed per plant and the damage index were recorded: 1 = no damage, 2 = 1-10% leaf consumed, 3 = 11-20% leaf consumed, 4 = 21-30% leaf consumed, 5 = more than 30% of the leaf consumed. A total of 10 plants were used for each fungal strain.

Statistical analysis.
For each trait (i.e. growth promotion, tolerance to cold, inoculation with Xcc and infestation with Mb) a Student's t-test was used to compare the means of each fungal inoculation treatment with the control at P < 0.05 and P < 0.01; significant differences are denoted using asterisks. www.nature.com/scientificreports/