Species Diversity and Virulence Potential of the Beauveria bassiana Complex and Beauveria scarabaeidicola Complex

Beauveria is a very important fungal resource. Some Beauveria species have great economic and ecological value. Through surveying Beauveria in China and Thailand over the past 4 years, 15 Beauveria spp. were collected and identified. Three new species—namely, B. polyrhachicola, B. songmingensis, and B. subscarabaeidicola—were described and illustrated based on morphological characteristics and molecular data. The phylogenetic positions of the 15 species were evaluated according to phylogenetic inferences based on six loci (nrSSU, nrLSU, TEF, RPB1, RPB2, and Bloc). Nine species of Beauveria in our study were isolated from adult scarab beetles (Coleoptera: Scarabaeidae). The pathogenicity of the isolates from the B. bassiana complex and B. scarabaeidicola complex was determined with three bioassays using B. mori and T. molitor larvae, in addition to Protaetia brevitarsis adults. The results indicated that the B. bassiana complex isolates had great potential in the biocontrol of the three insects; by contrast, the B. scarabaeidicola complex isolates showed obvious host specificity and low virulence.


INTRODUCTION
Beauveria is a very important fungal resource, with some species having great economic and ecological value (Zimmermann, 2007;Rehner et al., 2011;. Beauveria bassiana (Bals.-Criv.) Vuill. and B. brongniartii (Sacc.) Petch are well-known environmentally safe alternatives to using chemical pesticides to control agricultural pests (Zimmermann, 2007;Rehner et al., 2011). Beauveria pseudobassiana S.A. Rehner and Humber has also been shown to have great potential in the biocontrol of numerous insect pests . The entomopathogenic fungi Beauveria spp. are a class of environmentally friendly fungal pathogens that play an important role in controlling insect populations in nature (Luo et al., 2018;McKinnon et al., 2018). Some Beauveria species, as endophytes or soil and rhizosphere inhabitants, have been considered for potential use as biocontrol agents against plant pathogens by concerned practitioners, such as agriculturalists and plant pathologists. These species can produce an array of bioactive metabolites that limit the growth of some fungal plant pathogens and induce plant systemic resistance against the pathogenic bacterium (Ownley et al., 2010).
Beauveria bassiana is the most widely used fungus available commercially for controlling agricultural and forestry pests (Li et al., 2011). Products based on this species have been developed in many countries around the world (Goettel et al., 2005;Faria and Wraight, 2007;Li et al., 2011). However, a growing body of molecular evidence has demonstrated that B. bassiana, originally known as a generalist with a global distribution, encompasses cryptic lineages adapted to specific hosts or ecologies (Li et al., 2011;Rehner et al., 2011). Many initially identified B. bassiana isolates may belong to any of the species in the B. bassiana complex, such as B. rudraprayagi Y. Agrawal, Mual and Shenoy, B. staphylinidicola (Kobayasi and Shimizu) B. Shrestha, Kepler and Spatafora, and B. peruviensis D.E. Bustamante, M.S. Calderon, M. Oliva, and S. Leiva (Rehner et al., 2011;Agrawal et al., 2014;Kepler et al., 2017;Bustamante et al., 2019). Therefore, the abovementioned mycoinsecticide formulations of B. bassiana are not likely all based on B. bassiana.
Beauveria scarabaeidicola (Kobayasi) S.A. Rehner and Kepler is widely distributed in Oceania and Asia and named after its host adult beetle (Coleoptera: Scarabaeidae). It was originally described as Cordyceps scarabaeicola occurring in its sexual morph on an adult scarab beetle in New Guinea (Kobayasi and Shimizu, 1976). Cordyceps scarabaeicola has also been reported occasionally from many Asian countries, including China, Japan, and Korea (Shrestha et al., 2014). In an important phylogenetic study of Beauveria, a new entomopathogenic species, B. sungii S.A. Rehner and R.A. Humber, was described as a scarabkilling pathogen (hosts of all B. sungii isolates were identified as scarabs) (Rehner et al., 2011). Later, however, Shrestha et al. (2014) demonstrated that the telemorphic stage of B. sungii was C. scarabaeicola based on morphological and phylogenetic evidence. Because C. scarabaeicola was described earlier than B. sungii, Kepler et al. (2017) recommended B. scarabaeicola as the name of this species. Recently, Chen et al. (2019) proposed a new species, B. yunnanensis, a Chinese species parasitic on Lepidoptera pupa buried in soil that was a sister lineage to B. scarabaeicola.
During surveys of entomopathogenic fungi from different regions in Yunnan Province, China, and Chiang Rai Province, Thailand, over the past 4 years, approximately 15 Beauveria spp. were found and identified ( Table 1). In this study, we aimed to: (1) reveal the hidden species diversity of the B. bassiana complex and B. scarabaeidicola complex based on phylogenetic analyses and morphological observation and (2) assess the biocontrol potential of species in the B. bassiana complex and B. scarabaeidicola complex through pathological tests on the lepidopteran Bombyx mori and the coleopteran Tenebrio molitor larvae as well as Protaetia brevitarsis adults.

Soil and Specimen Collection
All the soil samples and the majority of Beauveria specimens were collected from Yunnan Province in China. Some specimens were collected from Chiang Rai Province in Thailand. Soil samples and specimens were noted and photographed in the fields, and then carefully placed in plastic containers at low temperature. Afterward, they were carried to the laboratory and stored at 4 • C before examination and isolation.

Fungal Isolation and Culture
Beauveria strains were isolated from the soil samples using the Tenebrio molitor baiting method (Keyser et al., 2015). Conidia developing on insect cadavers were transplanted onto plates of potato dextrose agar (PDA; potato 200 g/L, dextrose 20 g/L, agar 20 g/L) and cultured at 25 • C. Teleomorph specimens were rinsed with tap water, washed with sterile distilled water, and then dried on sterile filter paper. To obtain axenic cultures, white tissue inside the sclerotia of the teleomorph specimens was removed and inoculated onto PDA plates using a sterilized dissecting knife. Colonies of the isolated filamentous fungi appearing in the culture were transferred onto fresh PDA media. The purified fungal strains were maintained in a culture room at 25 • C or transferred to PDA slants and stored at 4 • C. Specimens were deposited in the Yunnan Herbal Herbarium (YHH) at the Institute of Herb Biotic Resources of Yunnan University. Cultures were stored in the Yunnan Fungal Culture Collection (YFCC) at the Institute of Herb Biotic Resources of Yunnan University.

Morphological Observations
Specimens were examined using an Olympus SZ61 stereomicroscope (Olympus Corporation, Tokyo, Japan). Cultures on PDA slants were transferred to PDA plates and then incubated at 25 • C for 14 days. For morphological evaluation, microscope slides were prepared by placing mycelia from the cultures on PDA medium blocks (5 mm diameter) and then overlaid with a coverslip. Medan dye solution was used to observe asci and ascospores. Other structures were mounted in water. Micro-morphological observations and measurements were performed using a light microscope (CX40, Olympus Corporation, Tokyo, Japan) and a scanning electron microscope (Quanta 200 FEG, FEI Company, Hillsboro, United States). Length to width ratios are given as Q. Mean values for length, width, and Q are indicated by L m , W m , and Q m , respectively.

DNA Extraction, PCR and Sequencing
Specimens and axenic living cultures were prepared for DNA extraction. Genomic DNA was extracted using the Genomic DNA Purification Kit (Qiagen GmbH, Hilden, Germany) according to the manufacturer's protocol. The primer pair nrSSU-CoF and nrSSU-CoR was used to amplify a fraction of the nuclear ribosomal small subunit (nrSSU) (Wang et al., 2015). Primer pair LR5 and LR0R (Vilgalys and Hester, 1990;Rehner and Samuels, 1994) was used to amplify a fraction of the nuclear ribosomal large subunit (nrLSU) and EF1α-EF and EF1α-ER (Bischoff et al., 2006;Sung et al., 2007) for the translation elongation factor 1α (TEF). For amplification of the largest and second largest subunits of the RNA polymerase II (RPB1 and RPB2), PCR primer pairs RPB1-5 F/RPB1-5 R and RPB2-5 F/RPB2-5 R (Bischoff et al., 2006;Sung et al., 2007)   Boldface: data generated in this study. T ex-type material.
Frontiers in Microbiology | www.frontiersin.org were employed. The Bloc fragment was amplified using primer pair B5.1F/B3.1R . All the PCR reactions were performed in a final volume of 50 µL and contained 25 µL of 2 × Taq PCR Master Mix (Tiangen Biotech Co. Ltd, Beijing, China), 0.5 µL of each primer (10 µM), 1 µL of genomic DNA, and 23 µL of RNase-free water. Target gene amplification and sequencing were performed according to the methods described in our previous study .

Phylogenetic Analyses
Phylogenetic analyses were based on six gene (nrSSU, nrLSU, TEF, RPB1, RPB2, and Bloc) sequences. The sequences were retrieved from GenBank and combined with those generated in our study. Taxon information and GenBank accession numbers were provided in Supplementary Table 1 and Table 1.
Sequences were aligned using MAFFT v.7. 1 After alignment, the sequences of the genes were concatenated. Conflicts among the six genes were tested using PAUP * 4.0b10 (Swofford, 2002). The results revealed that the phylogenetic signals in the six genes were not in conflict. The data partitions were defined for the combined dataset using PartitionFinder V1.1.1 (Lanfear et al., 2012). Phylogenetic analyses were conducted using BI and ML methods employing MrBayes v3.1.2 and RaxML 7.0.3, respectively (Ronquist and Huelsenbeck, 2003;Stamatakis et al., 2008). The BI analysis was run on MrBayes v3.1.2 for five million generations using a GTR+G+I model determined by jModelTest version 2.1.4 (Darriba et al., 2012). GTR+I was selected as the optimal model for ML analysis, and 1,000 rapid bootstrap replicates were performed on the dataset. The first analysis based on the combined five-gene (nrSSU+nrLSU+TEF+RPB1+RPB2) dataset was performed using the following taxa: Akanthomyces, Amphichorda, Ascopolyporus, Beauveria, Blackwellomyces, Cordyceps, Gibellula, Hevansia, Samsoniella, and Simplicillium. Two taxa of Trichoderma were designated as outgroups. The second analysis based on the combined four-gene (TEF+RPB1+RPB2+Bloc) sequences was performed using Beauveria taxa.
We applied a (phylo-) genetic distance matrix calculation for the combined four-gene (TEF+RPB1+RPB2+Bloc) sequences to assess species boundaries in the B. bassiana complex and B. scarabaeidicola complex ( Table 2). The pairwise genetic distances of most Beauveria lineages (Supplementary Table 2) were measured based on the Kimura 2-parameter model using MEGA6 software (Tamura et al., 2013).

Virulence Assay of Beauveria spp. Isolates
A total of 10 Beauveria spp. isolates from the B. bassiana complex and B. scarabaeidicola complex were tested for their pathogenicity to B. mori and T. molitor larvae in addition to P. brevitarsis adults. Conidia for each isolate were obtained from 4-week-old cultures grown on malt extract agar plates, suspended in a sterile aqueous solution of 0.01% Tween 80, and mixed vigorously until homogeneous conidial suspensions were produced. Quantification of the conidia was performed using a hemocytometer under a light microscope at 400 × magnification. All of the suspensions were adjusted to 1 × 10 8 conidia/ml. The tested insects were individually placed in sterilized rearing boxes and 10 µl of conidial suspension was applied to the surface of each insect. A diet was provided for each insect and renewed as needed. Control groups were treated with the same volume of a sterile aqueous solution of 0.01% Tween 80. The test was replicated three times with 50 insects per replicate. All of the test groups were kept at 25 • C under a 12:12 h photoperiod cycle. The numbers of dead insects were recorded every 12 h for a 30 day period, which was used to determine the percentage of mortality. Cadavers were removed, immediately surface-disinfected, and individually placed and maintained in rearing box chambers. Mycelium samples from cadavers were aseptically removed and cultured on PDA for microscope examination, DNA extraction, and TEF sequencing to confirm that mortality was caused by the inoculated fungal strain.

Sequencing and Phylogenetic Analyses
The combined five-gene dataset included sequences from 123 fungal taxa. The final dataset consisted of 5,001 bp of sequence data (nrSSU 1,138 bp, nrLSU 910 bp, TEF 1,047 bp, RPB1 781 bp, and RPB2 1,125 bp). Eleven well-supported clades were recognized based on both Bayesian inference (BI) and maximum likelihood (ML) analyses of the combined five-gene dataset of 123 taxa from Cordycipitaceae and Trichoderma, which accommodate species of the genera Akanthomyces, Amphichorda, Ascopolyporus, Beauveria, Blackwellomyces, Cordyceps, Gibellula, Hevansia, Samsoniella, Simplicillium, and Trichoderma (Supplementary Figure 1). The phylogenetic analyses also revealed the species diversity of the B. bassiana complex and B. scarabaeidicola complex in Beauveria clades. This suggested that the groups composed of the B. bassiana complex and B. scarabaeidicola complex should be genetically composed of at least four species (Supplementary Figure 1). Phylogenetic analyses based on combined partial TEF+RPB1+RPB2+Bloc sequences consisting of 59 fungal taxa resolved most Beauveria lineages in separate terminal branches (Figure 1). This revealed a similar tree and clustering topology, as shown in Supplementary Figure 1. It was proposed that the two strains YFCC 859 and YFCC 867, which formed a strongly supported clade, might be a new species in the B. bassiana complex, named B. polyrhachicola (Figure 1). Our analyses also revealed that two newly discovered species, B. songmingensis and B. subscarabaeidicola, were phylogenetically clustered with B. yunnanensis and B. scarabaeidicola, but they were clearly distinguished from the latter two by forming two separate clades in the B. scarabaeidicola complex (Figure 1 and Supplementary Figure 1). The genetic divergence comparisons showed that the minimum threshold (p-distance) to distinguish genetic species in Beauveria was 0.010 for the combined four-gene (TEF+RPB1+RPB2+Bloc)

Morphological Features
The morphological characteristics of the three new species as well as photomicrographs of morphological structures are shown in Figures 2-4. The detailed fungal morphological descriptions are provided in the Taxonomy section.

Conidial Viability of the Beauveria bassiana Complex and Beauveria scarabaeidicola Complex Isolates
Percentage of conidial germination was used to determine conidial viability. The conidial viability of the B. bassiana complex isolates was high, but the highest value (x = 95%) was found on isolates of YFCC 844 from soil (see Table 3). The values of the conidial viability of the isolates in the B. scarabaeidicola complex were significantly lower than those in the B. bassiana complex. Only four B. scarabaeidicola complex isolates with conidial viability were greater than 65%, and their conidial viability values were not significantly different (P < 0.05).

Virulence of the Beauveria bassiana Complex and Beauveria scarabaeidicola Complex Isolates
Ten isolates had conidial viability greater than 65%. These isolates were then selected for pathogenicity tests against B. mori and T. molitor larvae as well as P. brevitarsis adults ( Table 4). The B. bassiana complex isolates were shown to have great potential for use in the management of various insect pests; by contrast, the B. scarabaeidicola complex isolates showed obvious host specificity and low virulence. All tested isolates in the B. bassiana complex inflicted mycoses on B. mori and T. molitor larvae and caused over 80% mortality, whereas those in the B. scarabaeidicola complex did not. It was determined that the 10 isolates were pathogenic to P. brevitarsis adults but demonstrated different levels of virulence. Like the conidial viability, the mortalities of P. brevitarsis adults caused by the B. scarabaeidicola complex isolates were significantly lower than those of isolates in the B. bassiana complex (P < 0.05), strengthening the hypothesis that the virulence of certain entomopathogenic fungi is related to their conidial viability (Butt et al., 1994;Fernandes et al., 2007). Additionally, B. bassiana YFCC 844, which was isolated from soil and exhibited the highest conidial viability, showed high virulence against B. mori and T. molitor larvae, as well as P. brevitarsis adults, causing (94.00 ± 1.15)% mortality against B. mori larva, (95.33 ± 1.45)% mortality against T. molitor larva, and (79.00 ± 1.53)% mortality against P. brevitarsis adults (  Asexual morph: Colonies on PDA reached 20-38 mm in diameter after 14 days at 25 • C, white, circular, velutinous, and closely appressed to the agar surface; reverse yellowish white. Odor indistinct. Vegetative hyphae septate, branched, hyaline, smooth-walled, 1.2-2.3 µm wide. Conidiogenous cells, long cylindrical to long flask shaped, solitary or occurring in dense lateral clusters, base cylindrical to ampulliform and 1.4-3.0 µm wide, sympodially branched neck tapering into a long, slender, denticulate rachis, produced laterally on aerial hyphae or from subtending cells, 11.8-40.9 × 1.4-3.0 µm. Conidia 2.0-3.8 × 1.7-2.6 µm, Q = 1.0-1.8 µm (L m = 2.7 µm, W m = 2.1 µm, Q m = 1.3), globose, subglobose, slightly ellipsoid, oblong, or cylindrical, hyaline, aseptate, walls smooth and thin. Notes: Regarding phylogenetic relationships, B. polyrhachicola forms a distinct lineage in the B. bassiana complex, and it is closely related to B. peruviensis, B. staphylinidicola, B. bassiana, and B. rudraprayagi (Figure 1). Morphologically, B. polyrhachicola is similar to B. bassiana, B. kipukae, B. pseudobassiana, B. varroae, and B. peruviensis in terms of the shape and size of the conidia (Rehner et al., 2011;Bustamante et al., 2019). However, B. polyrhachicola can be distinguished from them by its long conidiogenous cells (11.8-40.9 × 1.4-3.0 µm).

Beauveria songmingensis H. Yu & Y. Wang, sp. nov. Figure 3
MycoBank number 841451. Etymology: named after the location Songming County where this species was collected.

DISCUSSION
It is generally agreed that distinguishing individual Beauveria species can be difficult using only morphological characters, as several species in the genus are morphologically cryptic species. In this study, we conducted a comprehensive investigation of the cryptic species diversity of the B. bassiana complex and B. scarabaeidicola complex. The molecular phylogeny clearly suggested the existence of distinct species in the B. bassiana complex and B. scarabaeidicola complex that we accordingly propose as new species: B. polyrhachicola (Figure 2), B. songmingensis (Figure 3), and B. subscarabaeidicola (Figure 4). Beauveria polyrhachicola is practically indistinguishable in morphology from other members of the B. bassiana complex. The shape and size of the conidia and the colony color of B. polyrhachicola, among other morphological features, have been observed in B. bassiana, B. rudraprayagi, B. staphylinidicola, and B. peruviensis (Rehner et al., 2011;Agrawal et al., 2014;Kepler et al., 2017;Bustamante et al., 2019). In the B. scarabaeidicola complex, the macromorphology of B. scarabaeidicola, B. songmingensis, and B. subscarabaeidicola is very similar, and thus species cannot be distinguished visually. The macroscopic and microscopic observations performed during our investigation revealed the extensive overlap in morphological characters and the lack of distinctive phenotypic variation, supporting the notion of cryptic species in a species complex.
At present, multi-locus phylogenetic analyses have gained importance in delimiting the species within the entomopathogenic fungi Beauveria. Rehner et al. (2011) divided B. bassiana s. lat. and B. brongniartii s. lat. into several cryptic species and described six new species based on the Bloc nuclear intergenic region and three nuclear genes encoding elongation factor 1-a (TEF), RNA polymerase II largest subunit (RPB1), and RNA polymerase II second largest subunit (RPB2). Subsequently, more than seven new species and new combinations were confirmed using combined analysis of the four-locus sequence data (Zhang et al., 2012;Chen et al., 2013Chen et al., , 2017Chen et al., , 2018Agrawal et al., 2014;Robène-Soustrade et al., 2015;Imoulan et al., 2016). In more recent studies, six species were added to the genus based on multilocus (nrSSU, nrLSU, TEF, RPB1, and RPB2) sequence data: B. acridophila, B. blattidicola, B. diapheromeriphila, B. locustiphila, B. scarabaeidicola, and B. staphylinidicola (Kepler et al., 2017). In this study, we analyzed most species of the newly circumscribed genus Beauveria based on phylogenetic inferences of six nuclear molecular markers (nrSSU, nrLSU, TEF, RPB1, RPB2, and Bloc). Phylogenetic analyses based on the five-gene (nrSSU, nrLSU, TEF, RPB1, and RPB2) dataset and the combined four-gene (TEF+RPB1+RPB2+Bloc) sequences produced trees with similar topologies that resolved most Beauveria lineages in separate terminal branches (Figure 1 and Supplementary  Figure 1). The results of the present work indicate that the first dataset was conducive to determining the phylogenomic relationships between Beauveria and its related genera, and the use of the latter was essential to establish robust Beauveria species boundaries, particularly the B. bassiana complex and B. scarabaeidicola complex.
Scarab beetles are leaf and root feeding pests of grasses, grains, sugarcane, strawberry, potato tubers, and young nursery plants (Crocker et al., 1996;Yokoyama et al., 1998). Based on the published literature, there are about six Beauveria spp. that parasitize adult scarab beetles: B. asiatica, B. bassiana, B. brongniartii, B. majiangensis, B. pseudobassiana, and B. scarabaeidicola (Rehner et al., 2011;Kepler et al., 2017;Chen et al., 2018;Khonsanit et al., 2020;. Here, we identified an extension of the members to also include B. malawiensis, B. songmingensis, B. subscarabaeidicola, and B. yunnanensis, as shown in Figure 1. Chen et al. (2019) emphasized that hosts of B. yunnanensis isolates were Lepidoptera pupae. However, our morphological observations of specimens from a type locality of B. yunnanensis indicated that their hosts were adult scarab beetles. Moreover, the host of B. yunnanensis was not shown in their publication (Chen et al., 2019). It seems that the host of Lepidoptera pupa is doubtful. There is reason to believe that members of the B. scarabaeidicola species complex are host-specific.
Not all scarab-killing pathogens are suitable for mycoinsecticide formulations that control scarab beetles. Our data suggested that the B. scarabaeidicola complex isolates showed low virulence. In addition, mortalities in P. brevitarsis adults caused by the B. scarabaeidicola complex isolates were significantly lower than those of isolates in the B. bassiana complex. Additional research is needed to determine the effectiveness of other species before future consideration of isolates for biological pest control.

CONCLUSION
The B. bassiana complex and B. scarabaeidicola complex, as special groups in the genus Beauveria, are rich in species diversity and have a wide distribution in nature. The B. bassiana complex, which is made up of five species, is a cosmopolitan group of soilborne necrotrophic arthropod-pathogenic fungi that have been shown to have great potential for the management of various insect pests. The B. scarabaeidicola complex is composed of pathogens specific to scarab beetles, and it is found on leaf litter or buried in soil. Species in this complex are morphologically highly similar and can hardly be distinguished macroscopically. In this study, we reported the discovery and description of three new species: B. polyrhachicola, which was found in the B. bassiana complex, and B. songmingensis and B. subscarabaeidicola, which were found in the B. scarabaeidicola complex. In addition, 10 species of Beauveria were found to be parasitic on scarab beetles. However, not all members are suitable for mycoinsecticide formulations for controlling scarab beetles. Our data suggested that the B. scarabaeidicola complex isolates showed obvious low virulence. Additionally, the mortality of Protaetia brevitarsis adults caused by the B. scarabaeidicola complex isolates was significantly lower than that of isolates in the B. bassiana complex.