Exploring interactions between Beauveria and Metarhizium strains through co-inoculation and responses of perennial ryegrass in a one-year trial

Collection of plants and insects

Plants and seeds of perennial ryegrass (Lolium perenne L.), and living and dead (Coleoptera: Scarabaeidae) larvae were collected during November–December 2018 (roots and larvae) and March–April 2019 (roots and seeds) from nine pasture sites near roads or railways in the Los Rios region, Chile (Figs. 1A–1E). The sampling design included three latitudinal transects from the Andes to the Coastal Cordillera: one transect in the south (a–c), one transect in the middle (d–f) and one transect in the northern part (g–i) of the region. Through this strategy, we attempted to sample sites from different soil series and grazing pasture managements. Thus, 180 plant individuals were collected in total (3 transects × 3 plots × 10 plant individuals × 2 collection periods), as well as 14 dead and 164 living white grubs corresponding to the larvae of scarabaeids that feed on organic matter and the roots of ryegrass. For the sites h, f and d we conducted this research under sampling permits from Juan Carlos Marin, Luis Toloza and Carlos Villagra, respectively.

Fungal isolation from plants and larvae at field sites

Ryegrass seeds and roots were first surface washed with running water to remove material adhering to the surface. Subsequently, they were immersed in 1% sodium hypochlorite (NaClO) for 5 min and then thoroughly washed five times with sterile distilled water. Root fragments and seeds as well as fungal structures growing on field-dead larvae were aseptically placed onto 90 mm × 14.5 mm Petri dishes containing 2% water agar supplemented with antibiotics (0.4 g/200 mL penicillin G sodium salt and 0.9 mg/200 mL streptomycin sulfate) and incubated at room temperature. Living larvae were cleaned superficially in a 1% sodium hypochlorite solution and rinsed five times with sterile distilled water and subsequently killed by thermic shock and maintained in a humid chamber for 14 days to allow the growth of entomopathogenic fungi.

Fungal growth was checked periodically by observation with a stereo microscope. Once mycelia appeared either from seeds, roots and/or larvae, they were transferred to Petri dishes containing potato dextrose agar (PDA) (Merck KGaA, Darmstadt, Germany) and the cultures were incubated at room temperature until the colonies reached a diameter of 20 to 30 mm. Fungal strains were identified on the basis of their colony, morphological and reproductive characteristics after subculturing on potato dextrose agar (PDA) plates (von Arx, 1982). Isolates morphologically identified as Beauveria and Metarhizium were subsequently sequenced to obtain a more accurate identification.

DNA extraction, PCR and sequencing

The mycelium was removed from the surface of fresh cultures grown in PDA and deposited in a 2-mL microcentrifuge tube and pulverized in liquid nitrogen with a pestle. To remove external fungal infections on ryegrass seeds, 1 g of seeds was sterilized as described above and pulverized in liquid nitrogen by grinding with a mortar and pestle. The total genomic DNA was obtained with an E.Z.N.A. Fungal DNA Mini Kit (Omega Bio-tek, Norcross, GA, USA), according to the manufacturer’s instructions. To amplify the nuclear segment including the ITS1–5.8S–ITS2 region (ITS), we used the primer combination ITS1F (Gardes & Bruns, 1993) and ITS4 (White et al., 1990). We used a touchdown PCR with an initial step of 94 °C for 3 min, followed by 10 cycles of 94 °C for 30 s, with the annealing temperatures starting at 60 °C for 45 s (decreasing by 1 °C per cycle), then 72 °C for 1 min 15 s for elongation, followed by 26 cycles of 94 °C for 30 s, 50 °C for 45 s, elongation at 72 °C for 1 min 15 s and a final extension at 72 °C for 7 min. The same primers used for the PCR were used to sequence both strands of DNA in an automated genetic analyser (ABI3500; Applied Biosystems, Foster City, CA, USA) of the AUSTRAL-omics core facility at the Universidad Austral de Chile (www.australomics.cl, accessed 1 July 2021). Sequencher 4.1 (Gene Codes Corporation, Ann Arbor, MI, USA) was used to assemble and edit the sequence data.

Antagonism assays, BLAST, sequence alignments and phylogenetic analyses

To detect whether there was an inhibitory effect between the isolated EIPF strains (Table S1), a dual-culture antagonism tests on 2% PDA. For initial identification, ITS sequences from all these strains were compared against sequences from GenBank by BLAST (Altschul et al., 1990). To carry out the experimental assays, we selected the strains Beauveria vermiconia NRRL B-67993 (P55_1) and Metarhizium aff. lepidiotae NRRL B-67994 (M25_2), which have been deposited at the NRRL Culture Collection (https://nrrl.ncaur.usda.gov/, accessed 1 July 2021). To infer the phylogenetics of the fungal strains, we assembled an alignment (Dataset 1 for the genus Beauveria and Dataset 2 for the genus Metarhizium) including all available sequences from both genera and the corresponding outgroups in GenBank (https://www.ncbi.nlm.nih.gov/, accessed 1 July 2021).

Both datasets were aligned separately with MAFFT v7.407 under the E-INS-i option (Katoh & Standley, 2013). Subsequently, the optimal models for each dataset were identified with PhyML 3.0 (Guindon et al., 2010) available at the ATGC: Montpellier Bioinformatics platform (http://www.atgc-montpellier.fr, accessed 1 July 2021). In addition, maximum likelihood phylogenetic trees were inferred with RAxML 7.0.3 (Stamatakis, 2006) and the GTR+I+G model of nucleotide substitution for both datasets. Tree searches were carried out with 1,000 rounds generated from distinct maximum parsimony starting trees. Bootstrap support for individual branches was assessed with 1,000 replicates (Felsenstein, 1985). Phylogenetic trees were viewed and rooted with FigTree v1.4.4 (Rambaut, 2009).

Experimental design and harvest

Seeds from nine cultivars of perennial ryegrass commercialized in Chile were tested for germination and microbial load (fungal and bacterial contamination) (Table S2). The seeds were sterilized as described above and 100 seeds per perennial ryegrass cultivar were placed in Petri dishes containing water agar (SIGMA, St. Louis, MO, USA) at 2% and incubated for 14 days at ~23 °C. Observations of seed germination and microbial contamination were performed at 7 and 14 days after inoculation. According to these assays, the cultivar RODEO Edge was selected for this study as it had the highest seed germination percentage and the lowest presence of microorganisms.

Twenty-four perennial ryegrass mini-pastures were established in 120-L containers 40 cm in diameter at the beginning of May 2020 (Fig. 1F). The experiments were mounted in an open-walled greenhouse with a transparent plastic roof; this way, the temperature and environmental humidity corresponded to those of the local climatic conditions (average annual temperature, 12 °C; rainfall up to 2,500 mm, mostly in the winter (see Huber (1970)) present in the EEAA belonging to the Universidad Austral de Chile located in sector Cabo Blanco, 4 km north of Valdivia (Chile). The soil comes from a naturalized grassland vegetation and corresponds to the Andisol type (Typic Hapludand) belonging to the Valdivia series with a high organic matter content, a low pH and high phosphorus retention (CIREN, 2003). Before the experiments started, a chemical analysis of the soil was carried out at the Soil Laboratory of the Universidad Austral de Chile (Table S3). Fertilization was carried out 70 days after planting, in which N, P and K were supplied by a commercial granulated mixture of urea (1 g), triple superphosphate (15 g) and Basacote Green Tec Truf K (13.9 g) in each mini-pasture. After 205 days, N continued to be supplied to each mini-pasture as urea (0.8 g), once after each cutting, until experiment finished.

In each pot, 3.5 g of seeds that had been sterilized as described above were distributed evenly across the surface and covered with a thin layer of soil. Twelve randomly selected pots were co-inoculated with the EIPF strains; the remaining 12 pots corresponded to the control treatment (lacking EIFP strains). For fungal inoculation, strains were routinely grown in 500-mL Erlenmeyer flasks containing 200 g of sterilized rice at 23 °C for 14 days. Conidial suspensions were prepared by adding 250 mL of sterile distilled water plus 0.1% Tween 80 to the Erlenmeyer flasks. The conidial suspensions of each fungal strain were then homogenized and filtered into a plastic bottle through a parafilm with very small perforations. The conidia concentration in each fungal suspension was determined under a light microscope in a Neubauer chamber and adjusted to a concentration of 1 × 10⁸ conidia/mL (modified from Jaber & Enkerli, 2016). Using the soil drench technique, the surface of the pots was drenched with 250 mL of a previously prepared 1:1 conidia suspension of each strain (modified from Greenfield et al., 2016). During the growing season, the mini-pastures were regularly watered with 2,000–3,000 mL of tap water per pot to reach 80–85% field capacity (FC), representing water availability conditions. At the beginning of summer, a water restriction scenario representing the summer season was imposed on six mini-pastures with and six without inoculation with EIPF. The limit of irrigation was 40–45% FC. At the start of the water restriction treatment, the soil moisture level was monitored by TEROS 11 moisture sensors (METER Group; Pullman Inc., Chicago, IL, USA) installed at 20 cm depth.

Consecutive harvests were carried out by hand when the growing degree-days reached 270 degrees (to obtain plants with three leaves). Plants were cut to 4 cm high to ensure good regrowth in the mini-pastures. In the aboveground biomass obtained through harvesting, fresh weight and leaf area index measurements were made. The harvested material was deposited in ovens and left to dry for 3 days at 60 °C to determine the dry weight. In order to carry out microscopic and molecular analyses, six plants from each mini-pasture were randomly sampled every 2 months after the start of the experiment in May (autumn), corresponding to June (autumn), August (winter) and October (spring) (normal water availability), and 1 month under drought stress in January and March (summer). In each sample, six plants from each treatment were carefully collected at random, deposited in labelled hermetic plastic bags and stored in the laboratory at –20 °C for further analysis.

Quantification of root colonization by EIPF

EIPF-inoculated and non-inoculated perennial ryegrass plants were collected from the mini-pastures and placed separately into plastic bags to be transported to the laboratory. The roots were separated and thoroughly washed with distilled water until all the soil particles adhering to the roots were removed. The fresh weight and dry weight in fine roots ranged from 0.02 to 0.3 g on average within the same sample. Once dried, they were weighed and placed in 2-mL Eppendorf tubes. Subsequently, the roots samples were pulverized in liquid nitrogen with the help of a pestle and the genomic DNA was extracted with the DNAeasy PowerSoil Pro Kit (QIAGEN, Hilden, Germany). The extracted genomic DNA was stored at –20 °C for later use.

The relative amount of fungal DNA in these samples was determined by performing quantitative PCR reactions with a B. vermiconia NRRL B-67993-specific primer pair, a M. aff. lepidiotae NRRL B-67994-specific primer pair and a perennial ryegrass-specific primer pair. Subsequently, the efficiency-corrected Ct values for each fungal strain were subtracted from those of the perennial ryegrass. The efficiency correction is required for accurate quantitative PCR analysis, as recently suggested by Ruijter et al. (2021). PCR primers were designed with Primer-BLAST (Ye et al., 2012) (Table 1), in which the B. vermiconia NRRL B-67993-specific primer pair (BEA) binds to ITS2, the M. aff. lepidiotae NRRL B-67994-specific primer pair (MET) binds to ITS1 and the LOL primer pair binds to the ITS2 of perennial ryegrass.

For PCR amplification, ThermoFisher SYBR Green Power Up (ABGene; www.thermofisher.com) was used in a final volume of 20 μL, following the manufacturer’s instructions. Real-time PCR was performed in a QuantStudio 3 Thermocycler (Applied Biosystems, Foster City, CA, USA). The qPCR assays were performed using reactions in triplicate in a total volume of 20 µL containing 10 µL of ThermoFisher SYBR Green Power Up, 0.4 µL of the forward and reverse primers at 0.2 µM, 7.2 µL of nuclease-free water and 2 µL of genomic DNA. The parameters for qPCR were a first step of activation at 50 °C for 2 min, 2 min at 95 °C and 40 cycles of denaturation at 95 °C for 15 s, and alignment and extension at 60 °C for 30 s. To estimate the amount of DNA in the MET and BEA target fragments within the root samples and its variation over time (Samples 1 to 5), relative DNA quantification was carried out. In the first instance, an attempt was made to use the relative quantification of the fungi that assumed an amplification efficiency between 90% and 110%, then compared the exponential curves of ΔCt (2^ΔCt) directly (Livak & Schmittgen, 2001). However, because it is an environmental material, we had difficulty achieving the optimal values required for all primer pairs and samples. For this reason, we used the Miner software (Zhao & Fernald, 2005) to obtain the efficiency and Ct values for qPCR from individual PCR reactions. These values were used to calculate the relative abundance (R_0) of the target genes (MET and BEA) normalized to the plant gene (LOL) considering the average efficiency for each primer pair.

For microscopic examination of fungal root colonization, fine roots of perennial ryegrass were rinsed in water as described above, and fixed in a 1:3 mixture (vol/vol) of Chloroform and Ethanol containing Trichloroacetic acid (1.5 g/L). Subsequently, roots were washed with phosphate buffer and stained with Wheat Germ Agglutinin-Alexafluor 488 (Molecular Probes; Invitrogen, Waltham, MA, USA) as described in Deshmukh et al. (2006). Images were obtained with a Leica TCS SP5 confocal microscope using the brightfield channel and a Green Fluorescent Protein (GFP) filter set (excitation wavelength 488 nm, bandpass 505–530 nm) for the detection of Wheat Germ Agglutinin-Alexafluor 488. Final images were prepared as maximum projections of z-stacks, if not stated otherwise.

Data analysis

First, we checked whether the fungal inoculation procedure resulted in higher colonization rates and if there were differences in the abundance between the two inoculated fungal strains. We used the relative amount of fungi in the roots, as determined by qPCR performed with DNA from the root samples.

The variability of the relative abundance of DNA was analyzed by means of a linear mixed model (following a normal distribution of the data), using the variables time (month of sampling), fungal strain (M. aff. lepidiotae and B. vermiconia) and water treatment (80–85% FC and 40–45% FC) as fixed variables, whereas the factor “mini-pasture” was incorporated as a random variable. In the same way, a linear mixed model was used to explain the variability in the biomass production of the mini-pastures, according to the fresh weight, dry weight and leaf area. The fixed variables for the biomass model were time, fungal inoculation (inoculated mini-pastures vs. control mini-pastures) and water treatment. The “mini-pasture” factor was incorporated as a random variable. ANOVA was used to determine the statistical significance of the data, with P-values less than 0.05 considered significant. Tukey’s test was used to identify the levels between which significant differences were obtained for the variable “time”.

Linear mixed models for each response variable were constructed by using the lmer function of the lme4 data package (Bates et al., 2015) and validated according to the Akaike criterion (AIC). Both linear mixed models and ANOVA were performed in R Studio (R Core Team, 2018). Graphical representation of the data was carried out in GraphPad Prism 8.

EIPF isolates, antagonism assay and phylogenetic identification

Four Beauveria and 33 Metarhizium strains were isolated from insect larvae collected from Sites b, d, e, h and i (see Fig. 1B and Table S4). No Beauveria and Metarhizium isolates were obtained from roots and seeds of perennial ryegrass. From each sample, one strain was selected for antagonism testing, resulting in two Beauveria and 19 Metarhizium strains. All dual cultures that formed more than one colony per plate were discarded from the measurements. A similar growth pattern was observed for all Metarhizium vs. Beauveria strains, showing a percentage inhibition between 50% and 60%. In contrast, the Beauveria strains had different growth patterns: strain S01_6 showed growth differences in dual cultures compared with the control, whereas strain P55_1 had homogeneous growth in dual cultures, with a percentage inhibition of 47%. (Table S1). Finally, the combination of B. vermiconia NRRL B-67993 (P55_1) and M. aff. lepidiotae NRRL B-67994 (M25_2) was selected because of the lower percentage of inhibition compared with the controls.

Our ITS phylogenetic analyses showed that the strain NRRL B-67993 was 100% identical to the collection ARSEF 2922 (type material of B. vermiconia), whereas the strain NRRL B-67994 has an isolated position, forming a cluster with the collection ARSEF 7488 (type material of M. lepidiotae) supported by an 83% bootstrap value (Fig. 2).

Primer specificity and quantitative PCR (qPCR) performance

The primers used in the qPCR assays for the relative quantification of B. vermiconia, M. aff. lepidiotae and perennial ryegrass are shown in Table 1. All three primer pairs were free from non-specific amplification, and the amplicons were confirmed to be of the correct size. In addition, the ITS primers were specific to perennial ryegrass and did not amplify any of the Beauveria and Metarhizium strains tested.

EIPF colonization of perennial ryegrass roots as detected by fluorescence microscopy and real-time PCR

Real-time PCR assays showed amplification of fungal DNA in inoculated plants (Table S5), and no fungal DNA was detected in non-inoculated control plants.

The linear mixed model indicated the significant effect of the variable time (months) on the colonization of ryegrass roots by EIPFs (Table 2). Figure 3A shows the relative abundance of B. vermiconia NRRL b-67993 (P55_1) and M. aff. lepidiotae NRRL b-67994 (M25_2) prior to the incorporation of the water restriction treatment (soil moisture at 40–45% FC), corresponding to the months of June (autumn), August (winter) and October (spring). During this period, all the mini-pastures were maintained at the same soil moisture (80–85% FC). From June to August, a significant decrease in the abundance of both fungi was observed (Tukey’s test, P = 0.001), but from August to October, the abundance increased significantly (Tukey’s test, P = 0.007). No significant differences were observed when we compared the abundance of both fungi in any of these 3 months. In December (early summer), the mini-pastures were divided into two blocks. One block was maintained with a water supply corresponding to a soil moisture content of 80–85% FC, and the other block was supplied with less water to reach a soil moisture content of 40–45% FC. According to the linear mixed model, time and fungal strain were the determining variables in ryegrass root colonization in January and March, corresponding to the summer season (Table 2). Under the 80–85% FC humidity treatment, both strains significantly decreased their relative abundance from January to March (Tukey’s test, P = 0.005 and P = 0.011 for M. aff. lepidiotae NRRL b-67994 (M25_2) and B. vermiconia NRRL b-67993 (P55_1), respectively). However, in both January and March, M. aff. lepidiotae NRRL b-67994 (M25_2) was significantly more abundant than B. vermiconia NRRL b-67993 (P55_1) (Tukey’s test, P = 0.044 and P = 0.014, respectively). Under the 40–45% FC treatment, the strains showed significant differences in relation to their relative abundance only in March (Tukey’s test, P = 0.026), with M. aff. lepidiotae NRRL b-67994 (M25_2) again being more abundant than B. vermiconia NRRL b-67993 (P55_1); additionally, their abundance at this moisture level was significantly higher than under the 80–85% FC moisture treatment (P = 0.008) (Fig. 3B). When we compared the total fungal rDNA (rDNA M. aff. lepidiotae NRRL b-67994 (M25_2) + rDNA B. vermiconia NRRL b-67993 (P55_1)), a significant decrease in total abundance was observed from January to March under the 80–85% FC humidity treatment (P = 0.001), which, in turn, was significantly lower than that under the 40–45% FC humidity treatment (P = 0.005), but only in March (Fig. 3B). The colonisation of perennial ryegrass roots was confirmed by staining chitin-containing fungal cell walls with WGA-Alexafluor. In root samples from inoculated plants, hyphal structures could be observed in all analyzed roots, sometimes forming extensive networks on the root surface (Figs. 3C and 3D). Hyphae were also frequently observed to grow inter- and intracellularly in the rhizodermis (Fig. 3E and Fig. S1).

Impact of EIPFs on aboveground perennial ryegrass biomass in mini-pastures

In contrast to the rDNA analyses, biomass measurements were made in July (winter), September (early spring), October (spring), November (spring–summer transition), December (early summer), January (summer), February (summer) and March (summer–autumn transition) (Table S6). Data from July to December (Figs. 4A, 4C and 4E) correspond to the period of soil moisture at 80–85% FC, and data from January to March included 50% of the mini-pastures under water stress. For biomass production, the linear mixed model indicated that the variable “time” was highly significant for the annual variation in fresh weight, dry weight and leaf area, and the interaction of time and inoculation treatment was significant for dry weight (Table 2). The growth pattern of ryegrass was characterized by significant monthly increases in fresh and dry weights towards spring (July–November), which subsequently decreased towards summer (December), contrary to how the leaf area decreased from July to November and increased towards December. Specifically, Tukey’s test revealed that in July, the fresh weight was significantly higher in the mini-pastures inoculated with the EIPFs (P = 0.015), but showed a significantly lower leaf area than the control mini-pastures (P = 0.032), although the dry weight in October was significantly lower in the control mini-pastures than in the inoculated mini-pastures (P = 0.013). From January to March (Figs. 4B, 4D and 4F), the linear mixed model indicates that the hydric treatment was the most explanatory for biomass variation (Table 2), since in all 3 months, significant differences were observed between moisture levels for both fresh and dry weight, and for leaf area only in February. Specifically, in January, fresh and dry weights were significantly lower in the mini-pastures with a lower water supply (Tukey’s test, P = 0.007 for both variables). However, in February and March, mini-pastures with lower soil moisture produced more biomass than mini-pastures with higher moisture (Tukey’s test, P-values for February and March: P = 0.001 and P < 0.001 for fresh weight, and P < 0.001 and P = 0.001 for dry weight, respectively). Leaf area was significantly lower in the water-restricted mini-pastures during February (Tukey’s test, P < 0.001). No significant differences were observed when we compared control vs. inoculated mini-pastures under the same water treatment and in the month, but significant differences were observed when we compared control mini-pastures between months for each water treatment, and when we compared inoculated mini-pastures under the same experimental conditions. In Figs. 4B, 4D and 4F, these statistically significant differences can be observed according to the letters representing the significance codes.