Transfection of entomopathogenic Metarhizium species with a mycovirus confers hypervirulence against two lepidopteran pests

Significance

The entomopathogenic members in genus Metarhizium have been used to control agricultural, grassland, and medical pests around the world. Here, we identified a mycovirus, MfPV1, in Metarhizium flavoviride isolated from small brown planthopper (Laodelphax striatellus). Transfection of the commercial strains of Metarhizium anisopliae and Metarhizium pingshaense with MfPV1 enhanced virulence of the fungi to diamondback moth (Plutella xylostella) and fall armyworm (Spodoptera frugiperda), two important lepidopteran pests. MfPV1-infected fungi produced more conidia and up-regulated pathogenesis-related genes that encode adhesin-like protein, hydrolyzed protein, and destruxin synthetase. This study reports a mycovirus that can confer hypervirulence in Metarhizium species and may facilitate the development of entomopathogenic fungus with enhanced virulence as an environmentally safe alternative to chemical insecticides.

Abstract

Although most known viruses infecting fungi pathogenic to higher eukaryotes are asymptomatic or reduce the virulence of their host fungi, those that confer hypervirulence to entomopathogenic fungus still need to be explored. Here, we identified and studied a novel mycovirus in Metarhizium flavoviride, isolated from small brown planthopper (Laodelphax striatellus). Based on molecular analysis, we tentatively designated the mycovirus as Metarhizium flavoviride partitivirus 1 (MfPV1), a species in genus Gammapartitivirus, family Partitiviridae. MfPV1 has two double-stranded RNAs as its genome, 1,775 and 1,575 bp in size respectively, encapsidated in isometric particles. When we transfected commercial strains of Metarhizium anisopliae and Metarhizium pingshaense with MfPV1, conidiation was significantly enhanced (t test; P-value < 0. 01), and the significantly higher mortality rates of the larvae of diamondback moth (Plutella xylostella) and fall armyworm (Spodoptera frugiperda), two important lepidopteran pests were found in virus-transfected strains (ANOVA; P-value < 0.05). Transcriptomic analysis showed that transcript levels of pathogenesis-related genes in MfPV1-infected M. anisopliae were obviously altered, suggesting increased production of metarhizium adhesin-like protein, hydrolyzed protein, and destruxin synthetase. Further studies are required to elucidate the mechanism whereby MfPV1 enhances the expression of pathogenesis-related genes and virulence of Metarhizium to lepidopteran pests. This study presents experimental evidence that the transfection of other entomopathogenic fungal species with a mycovirus can confer significant hypervirulence and provides a good example that mycoviruses could be used as a synergistic agent to enhance the biocontrol activity of entomopathogenic fungi.

Chemical pesticides have been used to control insect pests since the 1940s; however, their adverse effects on nontarget organisms, residues on food crops and groundwater, and development of insect resistance to chemicals have forced scientists to develop alternative ecofriendly measures such as biocontrol, the exploitation of living agents to combat harmful organisms (1, 2). Biocontrol agents are an important component of a sustainable pest management program because effective agents can be chosen to persist in the environment with low ecological impact (3). Entomopathogenic fungi have been widely used to control agricultural, grassland, medical, and veterinary insect pests (4, 5). Species of Metarhizium (Clavicipitaceae: Hypocreales) can infect more than 200 species of arthropods in over 50 families (6). They produce spores (conidia) that can germinate on the surface of the insect, penetrate its cuticle, and grow inside the insect. Then, they produce toxins, which presumably aid in suppressing host immune defenses, eventually killing the host, and fending off potential microbial competitors (7, 8). After death of the host, the fungus will grow out of the integument and produce aerial conidia, which are disseminated into the environment and may infect more insects (8). Pesticides developed from species of Metarhizium remain active for a long period, are not harmful to the environment and nontarget organisms, but they can take a long time to kill the insect, are unstable, and their control efficacy is low (9, 10).

Mycoviruses have been discovered in a wide range of fungal species, and some are usually associated with symptomless infections or with reduced growth rate or virulence to higher organisms of the pathogenic fungus (11–13). This reduction in virulence led to the idea that mycoviruses can serve as a biocontrol measure against fungal pathogens as in the case of chestnut blight caused by the fungus Cryphonectria parasitica in North America (14, 15). Some pathogenic fungi can live as a beneficial or nonpathogenic endophyte when infected with a mycovirus as do Sclerotinia sclerotiorum or Pestalotiopsis theae (16, 17). Few known mycoviruses, however, promote fungal growth and are associated with a hypervirulent or other extreme phenotypes. The presence of a viral 6.0-kbp dsRNA in Nectria radicicola, the fungus that causes ginseng nectria root rot, was found to be associated with high levels of virulence, sporulation, laccase activity, and pigmentation in this fungus (18). A newly discovered example is a novel ambigrammatic mycovirus that enhances the virulence of Puccinia striiformis f. sp. tritici, which causes wheat stripe rust (19).

Many mycoviruses have also been discovered in entomopathogenic fungi. About 21.3% of a worldwide collection of isolates of the entomopathogenic fungus Beauveria bassiana were found to harbor various mycoviruses and other dsRNA elements (20). Some of the isolates infected with polymycoviruses were found to be mildly hypervirulent against the greater wax moth (G. mellonella). There is also evidence that species of Metarhizium harbor large numbers of mycoviruses, which generally reduce mycelial growth, conidial production, and virulence of the host fungus by affecting its ability to tolerate heat and UV-B radiation and to penetrate insect cuticles (21, 22). However, information on mycoviruses in entomopathogenic fungi and their effects on the host phenotype remains limited. In the present study, we isolated and characterized a novel mycovirus Metarhizium flavoviride partitivirus 1 (MfPV1) from M. flavoviride isolated from small brown planthopper (SBPH, Laodelphax striatellus), an important insect vector for several plant viruses (23, 24). When we used MfPV1 to transfect virus-free commercial strains of two other Metarhizium species, conidiation and virulence of the transfected isolates were enhanced. This study reports a mycovirus in an entomopathogenic fungus that can endow significant hypervirulence to other fungi and opens the possibility that mycoviruses can elevate the activity of biocontrol agents.

Results

Identification of a Partitivirus by RNA-seq and Characteristics of the Virus Genome.

Of 35,443,732 total raw reads obtained from RNA-seq of the RNA library constructed with SBPH samples, 32,766,442 clean reads of 150 nucleotides (nt) were obtained after removing adaptor sequences and low-quality reads (SI Appendix, Table S1). A total of 6,690 clean reads, which accounted for 0.02% of the total reads, were mapped to the RNA1 and RNA2 of partitiviruses (SI Appendix, Fig. S1A). After de novo assembly of the reads into large contigs by Velvet 1.2.10, a Blastx search of the NCBI nonredundant protein sequences databases using the contigs as queries, two unknown viral contigs (1,730 and 1,553 nt in length, respectively) having high identities with partitiviruses were found. Then, we confirmed the validity of the two unknown contigs by RT-PCR and determined the terminal sequences by 5′ and 3′ RACE for obtaining full-length viral genome sequences (SI Appendix, Fig. S1B).

The complete viral genome contained two segments of double-stranded RNA (dsRNA), 1,775 nt for dsRNA1 (GenBank accession MH143600), and 1,575 nt for dsRNA 2 (GenBank accession MH143601) in length. Each dsRNA segment was predicted to have a single ORF that encodes a 62.6-kD putative RNA-dependent RNA polymerase (RdRp) comprising 538 amino acids (aa) and a 47.4-kD putative capsid protein (CP) comprising 440 aa (Fig. 1A). The identity of deduced aa sequences of dsRNA1 and dsRNA2 with those of approved species in family Partitiviridae was 28.6 to 70.9% for RdRp and 20.2 to 59.5% for CP, with the highest identities to the species in genus Gammapartitivirus (SI Appendix, Table S2). All the identities were much lower than the species demarcation thresholds in the family Partitiviridae (25). Conserved nucleotide sequences at the 5′ and the 3′ terminal of the genomic RNAs were identical to that of gammapartitiviruses (Fig. 1B and SI Appendix, Fig. S1C). In addition, a 65-nt and an 88-nt stem-loop structure was present in the 5′-untranslated region (5′UTR) of dsRNA1 and dsRNA2, respectively (Fig. 1C). Phylogenetic trees based on the deduced aa sequences of both RdRp and CP were constructed using MEGA 11. In both trees, the sequences from the virus that we identified were clustered in a clade of gammapartitiviruses (Fig. 1D). At the aa level, the RdRp encoded by dsRNA1 has six conserved motifs (III to VII) within the genus Gammapartitivirus (Fig. 1E). The results presented above suggest that this virus belongs a separate species in the genus Gammapartitivirus.

Fig. 1.

Analysis of the genomic characteristics and the classification of MfPV1. (A) A diagram shown the two proteins, RNA-dependent RNA polymerase (RdRP) and coat protein (CP), encoded by the two dsRNAs, respectively, in genome of MfPV1. (B) Sequence alignment of 5′ untranslated regions (UTR) or 3′ UTR between dsRNA1 and dsRNA2 of MfPV1 genome. (C) Prediction of the structure of 5′ UTR in dsRNA1 or dsRNA2 using the UNAFold web server (www.unafold.org). (D) Phylogenetic analysis among the species in genus Alphapartitivirus, Betapartitivirus, Gammapartitivirus, Deltapartitivirus, and Cryspovirus using the deduced amino acids of RdRP and CP. (E) Analysis of six conserved domains in RdRP among the nine species in genus Gammapartitivirus using the deduced amino acids.

Identification of the Virus Host and Purification of Viral Particles.

Species in genus Gammapartitivirus so far comprise only viruses isolated from various fungal species (25). Fungal isolate KF18 which was isolated from dead SBPHs (Fig. 2A) exhibited globose colony with irregularly brown mycelia and green conidia at the edge of the colony on potato dextrose agar (PDA) medium (Fig. 2B). Phylogenetic analysis based on sequences of ITS, TEF 1, RPB1, and RPB2 amplified from strain KF18 with other related fungi showed that the isolate has been clustered together with M. flavoviride var. minus ARSEF 1764, minus ARSEF 2037, minus ARSEF1099 (SI Appendix, Fig. S2), suggesting isolate KF18 belongs to M. flavoviride. In addition, isometric, nonenveloped, spherical viral particles of about 28- 30 nm in diameter were observed from purified virus particles prepared from the mycelia of Mf_KF18 by TEM (Fig. 2C), similar to the particles as reported for other species in family Partitiviridae (26). Besides that, the result of RT-PCR using the primer set ORF1-F/R and ORF2- F/R showed that the existences of RNA1 and RNA2 of the virus both in strain Mf_KF18 and purified viral fractions (Fig. 2 D and E). Thus, we propose this virus as “Metarhizium flavoviride partitivirus 1” (MfPV1).

Fig. 2.

Isolation and transfection of MfPV1 in Metarhizium. (A) The pictures show the dead small brown planthoppers (SBPHs) (Left and Middle panels), or the cadavers (Right panel), with the green fungi observed around the dead SBPHs. (B) Morphology of M. flavoviride, KF18 (the 18th clone isolated from the SBPH cadavers obtained from Kaifeng, Henan province). KF18 was found to grow slowly and generate the green spore in the reproductive phase. (C) Image of the virions of MfPV1. The isometric, nonenveloped, spherical particles of 28 to 30 nm in diameter were observed using transmission electron microscopy (TEM) in the purified viral fractions. (D) RT-PCR detection of MfPV1 in KF18 (Mf_KF18) using the specifical primers derived from the dsRNA1 and dsRNA2, respectively. The sizes of PCR productions from dsRNA1 and dsRNA were 1,717 and 1,382, respectively. Water was used for the PCR templets as the negative control. (E) RT-PCR determination of MfPV1 in Mf_KF18 virions using the same primers as D. Water was used for the PCR templets as the negative control. B represents blank loading. (F) Separation of dsRNAs from the Mf_KF18 and Ma_114445/MfPV1 isolates (see below) in the agarose gel. Several bands were detected in the Mf_KF18 but only two bands found in the Ma_114445/MfPV1 isolates. dsRNAs extracted from the Ma_114445 were used as the negative control. (G) Screen of positive culture in virion-transfected protoplasts of Metarhizium anisopliae by RT-PCR using specifical primers as D. The RNAs extracted from different samples were used for cDNA synthesis as the PCR templets. M, marker; line 1, positive culture; line 2, virus-free strain Ma_114445; line 3, Mf_KF18.

dsRNAs in Mf_KF18 and Transfection of MfPV1 to Other Metarhizium spp.

A total of seven bands with corresponding sizes between 900 and 1,900 bp as estimated by dsDNA markers were obtained in strain KF18 dsRNA preparations (Fig. 2F: line 2), suggesting Mf_KF18 might host several virus species. To elucidate any biological impacts of MfPV1 and other potential viruses in Mf_KF18, we first used several treatments to eliminate viruses from Mf_KF18, including single-conidium subculture, hyphal-tip transfer, and treatment with ribavirin or cycloheximide, but they failed to eliminate the viruses (SI Appendix, Fig. S3). Then, we transfected the purified MfPV1 particles into protoplasts of virus-free M. anisopliae Ma_114445 and Metarhizium pingshaense Mp_336563 to obtain MfPV1-infected isogenic isolates. Colonies were regenerated from selected protoplasts on PDA and confirmed to be positive for MfPV1 RdRp and CP by RT-PCR (Figs. 2G and 4A). The transfection rates were 55.0% for M. anisopliae and 61.5% for M. pingshaense. After the colonies were subcultured several times, RT-PCR showed that the MfPV1 was stably maintained in the virus-infected isolates (Ma_114445/MfPV1 and Mp_336563/MfPV1). Alignments of CDS of MfPV1 genomes in Ma_114445/MfPV1 and those from RNA-seq revealed only one base-pair difference in RNA1 and three in RNA2, but no difference in aa sequences both in RNA1 and RNA2 (SI Appendix, Fig. S4 A-D). As expected, the dsRNA from Ma_114445/MfPV1 or Mp_336563/MfPV1 yielded two dsRNA bands (Fig. 2F: line 3, 4B: line 2), and no dsRNA bands were found for Ma_114445 or Mp_336563 (Fig. 2F: line 4, 4B: line 3). The targeted dsRNA bands were then recovered and amplified, confirming the presence of the virus (SI Appendix, Fig. S4E).

Fig. 4.

Effects of MfVP1 on the pathogenicity of M. pingshaense. (A) Screen of positive culture in virion-transfected protoplasts of M. pingshaense (Mp_336563) by RT-PCR using specifical primers of MfPV1. The RNAs extracted from different samples were used for cDNA synthesis as the PCR templets. M, marker; line 1, positive culture; line 2, virus-free strain Mp_336563; line 3, Mf_KF18; W, water. (B) Examination of dsRNAs from the Mp_336563/MfPV1 isolates in the agarose gel. Two bands, representing two dsRNAs, were observed in the Mp_336563/MfPV1 strain. dsRNAs extracted from the Mp_336563 strain were used as the negative control. (C) Morphological difference of Mp_336563 (a strain of M. pingshaense) with or without MfPV1. Compared with the Mp_336563, Mp_336563/MfPV1 showed more stronger of white colony and displayed more mycelium. The picture was taken at 3, 5, and 14 d of clone growth, respectively. (D and E) Effects of MfVP1 on the spore yield (D) and (E) biomass of M. anisopliae. (F) Measurement of the mortality rates of Plutella xylostella larvae treated by 0.05% Tween-80 combined with the spores of Mp_336563 or Mp_336563/MfPV1 at 1 to 5 d postinoculation. The mortality rates of P. xylostella larvae treated by 0.05% Tween-80 were used as the negative control. In (D and E), data are means ± SD of three biological replications. *P < 0.05; **P < 0.01 (Student’s t test). In (F), data are means ± SD of three biological replications.

Changes in Biological Characteristics of MfPV1-Transfected Isolates of M. anisopliae and M. pingshaense.

For M. anisopliae, the colony diameter of the virus-free strain Ma_114445 and the transfected strain Ma_114445/MfPV1 only had a slight difference during a culture period over 11 d (SI Appendix, Fig. S5A). However, by day 5, their colony morphologies differed; Ma_114445 had a yellow and white colony with a wavy margin, in contrast to the smooth-edged, green and gray colony of the transfected strain Ma_114445/MfPV1 (Fig. 3A). Both colonies retained these features until they conidiated, when the mycelia of strain Ma_114445/MfPV1 became denser and formed a thicker layer of conidia than that of Ma_114445 with sparse conidia (Fig. 3B). On day 21, significantly more conidia were produced by Ma_114445/MfPV1 than by Ma_114445 (approximately 5.4 times more, Fig. 3C; t test: P = 0.0024). However, the mycelial biomass of the virus-infected strain after 5 d was significantly lower than that of the virus-free strain (approximately 11.45% less, Fig. 3D, t test: P = 0.0019). On the hypertonic medium containing 1/4 SDAY with either 1 M NaCl or 1 M sorbitol, the mean diameter of the virus-infected colonies on day 5 was larger than that of the virus-free strain. Similarly, when the conidia were placed on 1/4 SDAY with 400 μg/mL SDS, the virus-infected colonies were larger than those of the virus-free strain, but the colony diameters of the virus-free strain were larger on 1/4 SDAY amended with 500 μg/mL Congo red (SI Appendix, Fig. S5 B and C), suggesting that chemical tolerance of the virus-infected strain was enhanced.

Fig. 3.

Effects of MfVP1 on the growth and pathogenicity of M. anisopliae. (A) Morphological difference of Ma_114445 (a strain of M. anisopliae) with or without MfPV1. Compared with the Ma_114445, Ma_114445/MfPV1 showed more obvious green and gray colony with smooth edge. The picture was taken on the fifth day of clone growth. (B and C). Effects of MfVP1 on the morphology (B) and spore yield (C) of M. anisopliae. Comparison of the morphologies between the Ma_114445 and Ma_114445/MfPV1 strains at the 21 d post clone inoculated into the medium (B). Statistical analysis of the spore yield between the Ma_114445 and Ma_114445/MfPV1 strains (C). (D) Statistical analysis of the mycelium biomass between the Ma_114445 and Ma_114445/MfPV1 strains on the fifth day of clone growth. (E–G). Virulence assay demonstrated the enhancement of insecticidal ability to P. xylostella larvae by Ma_114445/MfPV1 strain. Measurement of the mortality rates of P. xylostella larvae treated by 0.05% Tween-80 combined with the spores of Ma_114445 or Ma_114445/MfPV1 at 1 to 5 d postinoculation. The mortality rates of P. xylostella larvae treated by 0.05% Tween-80 were used as the negative control (E). The dead P. xylostella larvae were used to isolate the Ma_114445 strains in PDA medium (F), which were further exanimated the MfPV1 by RT-PCR with specifical primers (G). (H) Virulence assay demonstrated the enhancement of insecticidal ability to S. frugiperda larvae by Ma_114445/MfPV1 strain. In (C and D), data are means ± SD of three biological replications. **P < 0.01 (Student’s t test). In (E and H), data are means ± SD of three biological replications.

In virulence bioassay using diamondback moth (P. xylostella), the highest mortality rate for the larvae of the negative control was 13.3% by day 5, but that of the larvae treated with conidia of Ma_114445/MfPV1 on days 1 to 5 was 2.2%, 4.4%, 22.2%, 41.1%, and 54.4%, respectively, compared with 0%, 0%, 10.0%, 24.4%, and 41.1%, respectively, for Ma_114445 (Fig. 3E). The mortality rates of larvae treated with conidia of Ma_114445/MfPV1 were significantly higher than those of Ma_114445 at 2, 3, 4, 5 dpi (32.4% higher at 5 dpi; ANOVA, P < 0.05, Fig. 3E). In addition, white mycelia appeared on the dead larvae after inoculation with either Ma_114445 or Ma_114445/MfPV1 in high humidity (Fig. 3F), suggesting the larvae had been killed by the fungal strains. When each strain on the dead larvae was tested by RT-PCR, the results showed that the virus was still present in Ma_114445/MfPV1 strain but not in Ma_114445 strain (Fig. 3G), again suggesting that MfPV1 was responsible for the higher mortality rates of the larvae.

When we evaluated the mortality of larvae of fall armyworm (Spodoptera frugiperda) after feeding on corn leaves that had been soaked in the conidial suspension of either Ma_114445/MfPV1 or Ma_114445, the mortality rates of larvae on Ma_114445/MfPV1-treated corn leaves were significantly higher (ANOVA, P < 0.05) than those on Ma_114445-treated leaves at 3 dpi (68.9% and 48.3% respectively), 4 dpi (83.3% and 63.3% respectively), 6 dpi (93.3% and 78.3% respectively), and 7 dpi (95.0% and 81.67% respectively). The mortality rate of the negative control was only 12.2% by day 7 (Fig. 3H). The corresponding LT₅₀ was 3.85 d for Ma_114445 and 2.99 d for Ma_114445/MfPV1.

Similarly, the transfection of the virus-free strain Mp_336563 of M. pingshaense with MfPV1 did not affect the host growth rate (SI Appendix, Fig. S6). The white colonies of the transfected and virus-free strains were similar in morphology before day 3 but differed by day 5; the infected strain remained white, but a yellow center had developed in the white virus-free colony (Fig. 4C). Conidial and biomass production by Mp_336563/MfPV1 were also significantly greater than in the virus-free strain (Fig. 4 D and E). In the virulence bioassay using P. xylostella, the mortality rate for the larvae in the negative control was 14.2% by day 5; the mortality rates of larvae treated with conidia of Mp_336563/MfPV1 were significantly higher (ANOVA, P < 0.05) than those of Mp_336563 at 2 dpi (35.6% vs 19.2%, respectively), 3 dpi (55.6% vs 39.2%), 4 dpi (70.0% vs 50.0%), and 5 dpi (77.5% vs 63.3%; 32.4% higher, Fig. 4F). The corresponding LT₅₀ was 4.43 d for Mp_336563 and 3.22 d for Mp_336563/MfPV1.

Identification of Differentially Expressed Genes (DEGs) in M. anisopliae in Response to MfPV1 Infection and Validation of Selected Genes Using RT-qPCR.

To learn more about potential molecular mechanisms underlying the virus-induced hypervirulence in the host fungus, we analyzed the RNA-seq data from MfPV1-infected and virus-free strains of M. anisopliae cultured in potato dextrose broth (PDB) medium with cicada slough addition. The relative percentage of reads was similar in all samples (SI Appendix, Table S3). Results of the read alignment revealed that >80% of the reads were mapped to the reference genome (SI Appendix, Table S4). The square of Pearson correlation coefficient (r) was 0.928, 0.978, and 0.956 for three replicates of Ma_114445, and 0.972, 0.955, 0.93 for Ma_114445//MfPV1 (SI Appendix, Fig. S7A). Three samples of Ma_114445 were clustered together with those of Ma_114445/MfPV1, whereas the intergroup was dispersed in the PCA (SI Appendix, Fig. S7B). All the above analyses of RNA-seq data indicated that the data were reliable. Using P-value < 0.05 and fold-change >1.2 as criteria for differential expression, we identified 1,872 DEGs for strain Ma_114445/MfPV1, with 1,136 of these (60.68%) up-regulated and 736 (39.32%) down-regulated (Fig. 5A and Dataset S1). Hierarchical clustering revealed similar expression patterns among the three biological replicates of strains Ma_114445/MfPV1 and Ma_114445 (SI Appendix, Fig. S7C). In the GO analysis, the annotations for 1,388 of the DEGs were from three major functional ontologies (biological process, cellular component, and molecular function); 415 were annotated as biological process with 392 up-regulated and 23 down-regulated; 637 were annotated as molecular function with 372 up-regulated and 265 down-regulated; and 336 were annotated as cellular component with 235 up-regulated and 101 down-regulated. Based on low to high of P-value, the top four enriched terms were cofactor binding (GO:0048037, molecular function), heme binding (GO:0020037, molecular function), and tetrapyrrole binding (GO:0046906, molecular function) (SI Appendix, Fig. S7D and Dataset S2).

Fig. 5.

Analysis of the DEGs of M. anisopliae induced by the MfPV1 using RNA sequencing and detection of three conidiation-related genes of M. anisopliae cultured on PDA medium for 12 d by RT-qPCR. (A) Volcano plot shows the upregulation and downregulation of genes in M. anisopliae stimulated by the MfPV1. Upregulation of the genes functioning relative to the pathogenicity of M. anisopliae are marked by black font. (B and C) Validation of the six pathogenesis-related genes by RT-qPCR at 3 d postinoculation (B). The Mad1 gene, a virulence factor induced to express highly in early phase of Metarhizium infection, was confirmed by RT-qPCR at 6, 12, 24, 30, and 36 h, respectively (C). (D) Detection of the three conidiation-related genes by RT-qPCR in Ma_114445 and Ma_114445/MfPV1 strains cultured on PDA medium for 12 d. In (B–D), the data are means ± SD of at least three biological replications. * P < 0.05; ** P < 0.01; *** P < 0.001 (Student’s t test).

The expression of several genes related to Metarhizium pathogenicity, including genes involved in destruxin synthesis and the genes encoding adhesin-like protein (Mad1), a protease that hydrolyzes the insect cuticle, and a collagen-like protein (MCL1), was up-regulated (SI Appendix, Table S5). Six of the up-regulated DEGs (MAN_02447, MAN_05588, MAN_06794, MAN_07719, MAN_03711, MAN_10464) were selected to assess their expression by RT-qPCR using gene-specific primers. In agreement with the RNA-seq results, the RT-qPCR results showed that all six pathogenesis-related genes were up-regulated significantly in Ma_114445/MfPV1 cultured in cicada induced medium at 3 dpi. Especially high was the expression of Mad1, approximately 7–120 times higher than the other selected DEGs (Fig. 5B). Thus, we validated its expression in the MfPV1-infected and MfPV1-free strains cultured in PDB from 6 to 36 h after transfection as described above. The results confirmed that Mad1 expression was higher in the MfPV1-infected strain than in the virus-free strain from 12 to 36 h (Fig. 5C). We also validated the expressions of the three central regulators (BrlA, AbaA, and WetA) of the conidiation pathway in filamentous fungi by RT-qPCR using gene-specific primers in M. anisopliae, which were cultured on PDA medium for 12 d. The results showed significantly up-regulated expressions of them in MfPV1-infected strain (Fig. 5D).

Discussion

The mycovirus MfPV1 that we isolated from M. flavoviride in L. striatellus enhanced the conidiation and virulence of M. anisopliae and M. pingshaense after they were transfected with purified virions. The virus is phylogenetically most closely related to species in the genus Gammapartitivirus, family Partitiviridae (Fig. 1D), and shares the highest aa identity for RdRp with Penicillium stoloniferum virus S (70.9%) and for CP with Discula destructiva virus 2 (59.5%) (SI Appendix, Table S2), both of them are lower than the species demarcation criteria for family Partitiviridae (≤90% aa identity for RdRp and/or ≤80% for CP) (25, http://ictvonline.org/), suggesting that MfPV1 should be classified as a species in the genus Gammapartitivirus.

Partitiviruses are transmitted intracellularly by seeds in plants, by oocysts in protozoa, or by hyphal anastomosis, cell division, and sporogenesis in fungi (25). All the tested conidia of Metarhizium KF_18 remained positive for MfPV1 after we tried to eliminate the mycovirus via single-conidium isolation (SI Appendix, Fig. S3). Because of the ineffectiveness of hyphal tip isolation to eliminate partitivirus particles or dsRNA (27, 28), it was hard to remove MfPV1 from its native host M. flavoviride by single-conidium or hyphal-tip subculture, and ribavirin or cycloheximide treatments (SI Appendix, Fig. S3). Transfection of other hosts with purified partitivirus particles to obtain isogenic isolates has been reported (29–31), so here we transfected virus-free protoplasts of commercial strains Ma_114445 of M. anisopliae and Mp_336563 of M. pingshaense with MfPV1 using PEG (Figs. 2G and 4A). After subculturing the virus-transfected strains for at least 10 generations, the morphologies of colonies and conidial production were consistent with the previous generation, suggesting stable presence of MfPV1 in the transfected fungal strains. In future studies, we will transfect other entomopathogenic fungi with MfPV1 in efforts to improve their virulence.

In general, partitiviruses seem to be associated with symptomless infections of their fungal hosts (27, 32). However, some of the viruses have deleterious effects on the host fungus, resulting in slower growth, reduced conidiation, altered morphology, and disrupted sexual reproduction, which can lead to hypovirulence in some fungal phytopathogens (30, 33, 34). In contrast, one gammapartitivirus, Talaromyces marneffei partitivirus 1 can enhance the virulence of its native host T. marneffei, an opportunistic pathogen of mammals, presumably by upregulating virulence factors and suppressing host RNAi machinery (35). B. bassiana polymycovirus-1 (BbPmV-1, genus Polymycovirus, family Polymycoviridae) and B. bassiana nonsegmented virus 1 (BbNV-1, genus Unirnavirus, family Unirnaviridae) reduce the growth of their host B. bassiana, but have a mild hypervirulent effect against the greater wax moth (Galleria mellonella) (21). However, purified particles of BbPmV-1 and BbNV-1 could not be introduced by transfection into protoplasts of the commercial, virus-free strain B. bassiana ATCC 704040, a biocontrol agent for various arthropod pests. In this study, mycelial growth of the MfPV1-transfected M. anisopliae strain did not differ clearly from that of the strain without the virus, but its culture morphology differed, and it produced more conidia (Fig. 3 A–C) and had significantly higher virulence against larvae of two important agricultural pests, P. xylostella (Fig. 3E) and S. frugiperda (Fig. 3H). Similar results were also obtained for the MfPV1-transfected strain of M. pingshaense (Fig. 4F). Thus, the present study demonstrates that a mycovirus can confer substantial hypervirulence to other species of an entomopathogenic fungus and is an important finding in the field of biological control. Thus, mycoviruses might be used to enhance the virulence of commercially available Metarhizium strains and provide a viable alternative to using genetic engineering to improve the biocontrol efficacy of entomopathogenic fungi.

In our transcriptomic analysis, mRNA levels for diverse genes differed in MfPV1-infected Ma_114445/MfPV1 compared to those in the virus-free isolate (Dataset S1). Among 1,136 up-regulated genes, 11 are involved in infection processes (SI Appendix, Table S5). We selected six of these 11, encoding Mad1, Pr1H, Pr2, Mcl1, lipase, and destruxin synthetase and validated their upregulation using RT-qPCR (Fig. 5B). Adhesin protein MAD1 is important for the attachment of conidia to the host surface; disruption of Mad1 blocks adhesion of conidia of M. anisopliae to insect surfaces and greatly reduces its virulence in caterpillars, suggesting that MAD1 is a potential virulence factor (36). In our transcriptomics results, the expression of Mad1 was up-regulated in virus-infected strain by 1.66 times (SI Appendix, Table S5) and was always higher than in virus-free strain by 12 h after culturing (Fig. 5C). Later in the infection process, Metarhizium species secrete proteolytic, chitinolytic, and lipolytic enzymes that degrade major constituents of the cuticle (37–39). Overexpression of the gene encoding subtilisin-like proteases (Pr1) from M. anisopliae in the hemolymph of Manduca sexta activates the prophenoloxidase system, which is involved in proteolytic cleavage during a cascade of trypsin-like enzymatic reactions. The combined toxic effects of Pr1 and the reaction products of phenoloxidase cause larvae inoculated with a Pr1-overexpressed fungus to die 25% faster and reduce their food consumption by 40% compared to rates after inoculation with the wild-type fungus (40). The expression of Pr1H, Pr2, and a gene for lipase up-regulated in the virus-infected strain by 1.39, 11.80, and 1.87 times, respectively (SI Appendix, Table S5). Metarhizium can synthesize a collagen-like protein (Mcl1), an immune modulator that suppresses the insect host defense system (41). In this study, the mRNA level of Mcl1 was 4.07 times higher in the virus-infected strain than in the virus-free strain. The gene encoding destruxin synthetase in the virus-infected strain was also up-regulated (1.87 times higher than in the virus-free strain). Destructin produced by Metarhizium is a bioactive secondary metabolite that accelerates death of the infected insect by weakening its immune defense, damaging its muscular system and the malpighian tubules, affecting excretion, and eventually impairing feeding and mobility (42). All of these results suggest that the up-regulated expression of numerous virulence-related genes in the virus-infected strain is responsible for the hypervirulence to the target pests. Further studies are required to determine the mechanism for enhancing the expression of virulence-related genes in Metarhizium.

Metarhizium species may be introduced in the field as conidia or mycelia fragments in aqueous solutions, and they then enter the host either directly or indirectly (43). The complex process of conidiation in filamentous fungi involves many aspects such as the differentiation of specialized structures, regulation of gene expression, and responses to environmental conditions (44). A central regulatory pathway (BrlA-AbaA-WetA), which is involved in conidiogenesis and other processes, is functionally conserved in phyto- and entomopathogenic fungi (45–47). In B. bassiana, BrlA, AbaA, and WetA are important regulators of conidiation, conidial maturation, and virulence (48, 49). Mycoviruses can affect fitness of the fungi by enhancing or reducing the sporulation. The L1 dsRNA of N. radicicola could stimulate sporulation and increasing virulence of N. radicicola (19). The mycovirus Phytophthora infestans RNA virus 2 stimulates sporangia production in Phytophthora infestans, which causes potato late blight (50). In the present study, MfPV1-infected M. anisopliae produced five times more conidia than the virus-free strain (Fig. 3C). Our transcriptomic experiment used M. anisopliae strains cultured in cicada-amended PDB medium. The cicada slough can induce the expression of virulence genes in M. anisopliae because it serves as a substitute for insect cuticle to stimulate the expression of pathogenicity genes in entomopathogenic fungi (51). Transcriptomic data analysis showed that sporulation-related genes were not up-regulated because the sporulation of fungi cultured in the liquid medium needs air as a necessary factor (52). When we detected the expressions of the three central regulators (BrlA, AbaA, and WetA) of conidiation by RT-qPCR, using the fungal cultures on PDA medium for 12 d, all three genes are significantly up-regulated (Fig. 5E), which indicates that MfPV1 achieved sporulation stimulation likely through upregulation the central regulatory pathway of sporulation in M. anisoplia.

In conclusion, we have shown here that the virulence of two species of Metarhizium from commercial biological control products is enhanced after they were transfected with the mycovirus MfPV1 that we isolated from M. flavoviride. The transfected species were hypervirulent against two lepidopteran pests and produced higher levels of conidia and expression of pathogenesis-related genes. The mycovirus thus has the potential to enhance the biocontrol activity of entomopathogenic fungi as an environmentally safe alternative to chemical insecticides.

Method Details

Insects and Fungal Isolates.

Small brown planthopper (SBPH, L. striatellus Fallén) (Hemiptera: Delphacidae) were collected from Kaifeng, Henan Province, China, in June 2016 and maintained on rice seedlings (cv. Wuyujing 3) in insect-proof cages at 27 °C with 16 h light/8 h dark. P. xylostella L. (Lepidoptera: Plutellidae) and S. frugiperda J. E. Smith (Lepidoptera: Noctuidae), both economically important agricultural pests, were provided by the laboratory of entomology in our institute. Strain Mf_KF18 of M. flavoviride sensu lato was isolated from dead SBPHs. Commercial strain Ma_114445 of M. anisopliae (Metschnikoff) Sorokin and Mp_336563 of M. pingshaense Q.T. Chen and H.L. Guo were purchased from BNCC (BeNa Culture Collection, China). Strains Ma_114445/MfPV1 and Mp_336563/MfPV1 were virus-harboring strains transformed by polyethylene glycol 4000 (PEG-4000) in this study.

RNA Sequencing, Assembly, and Analysis.

In 2016, we identified numerous dead adult SBPHs that were covered with a green fungus during their feeding period in insect-proof cages in our laboratory. We randomly selected 10 living adults for RNA sequencing (RNA-seq). Total RNA was extracted from the sample using TRIzol reagent (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions. The Ribo-Zero Magnetic Kit (Epicentre, Madison, WI) was used to deplete ribosomal RNA from transcriptomes for RNA-seq; then, a library was constructed with a TruSeq total RNA Sample Prep Kit (Illumina, San Diego, CA). The Illumina HiSeq X Ten platform with PE150 bp was used for the RNA-seq. The CLC Genomics Workbench 9.5 was used for de novo assembly of RNA-seq data. The assembled contigs were subsequently screened against the NCBI databases using BLASTn and BLASTx searches with default options. Two contigs caught our attention because of their distinctness, so we selected them for further analysis.

Sequencing and Analyzing the Viral Genome.

To further characterize any viruses represented by the distinct contigs identified from the RNA sequencing of the RNA from the 10 SBPHs, we designed specific primers using Primer Premier 5 (Premier Biosoft Interpairs, Palo Alto, CA) for RT-PCR and RACE-PCR (SI Appendix, Table S6) to acquire the full genomic sequence of the virus(es). The sequences of the extreme ends of the genomic RNAs were determined employing the 5′- and 3′-RACE System for Rapid Amplification of cDNA Ends kits (Thermo Fisher Scientific, Waltham, MA). The PCR products were separated by electrophoresis in 1.2% agarose gels and visualized by ethidium bromide staining, and then purified with the Wizard SV Gel and PCR Clean-Up System (Promega, Madison, WI). The PCR products were cloned into the PEASY-T5 cloning vector (TransGen Biotech, Beijing, China) and then inserted into Trans-T1 Chemically Competent Cells (TransGen). The positive clones were sequenced by Sanger sequencing at Sangon Biotech Co., Ltd (Shanghai, China).

Contigs were assembled using the DNAMAN (v6) program (Lynnon Biosoft, San Ramon, CA). Open reading frames (ORFs) were deduced using ORFfinder (https://www.ncbi.nlm.nih.gov/orffinder/). Sequence identities were computed using LALIGN program of EMBL’s EBI European Bioinformatics Institute (EMBL-EBI) (53) with the default settings. Terminal sequences were aligned using MEGA 11 (v11.0.13) and the Muscle method (54). The UNAFold Web Server (http://www.unafold.org) was used to predict the secondary structure (55). For phylogenetic analysis, sequences were aligned using the Muscle program in MEGA 11, and then phylogenetic trees were generated using the neighbor-joining (NJ) method with a P-distance model and 1,000 bootstrap (54).

Isolation and Identification of Fungi from Dead SBPHs.

For isolating the sporulating fungus, the dead insects were surface-sterilized with 0.2% v/v sodium hypochlorite, then placed on PDA amended with 100 μg/ mL ampicillin and 50 μg/ mL kanamycin, and incubated at 28 °C with 12 h light/12 h dark for 7 to 14 d. Single colonies were transferred to a new dish of PDA and cultured for 7 d. Total DNA was extracted from each fungal culture using a genomic DNA purification kit (Promega). Partial regions of four nuclear loci, including internal transcribed spacer (ITS), translation elongation factor 1-α (TEF1), RNA polymerase II first, and second subunit (RPB1 and RPB2), were amplified using corresponding primer pairs (56, 57) (SI Appendix, Table S6). The amplicons were purified and sequenced at Sangon Biotech Co., Ltd. The obtained sequences were analyzed using DNAMAN (v6, Lynnon Biosoft).

Purification and Observation of Virus Particles by TEM.

Mycelial plugs of strain Mf_KF18 were incubated in PDB on a shaking incubator at 28 °C for 7 d. After harvest, 30 g of mycelia was ground to a fine powder in liquid nitrogen; virions were purified from the crude extract using sucrose gradient centrifugation as previously described (58). The virus-containing zone (20 to 30% sucrose) was then centrifuged at 35,000 rpm for 3 h. The pellets were then suspended in 0.01 M PB (pH 7.4), pipetted on carbon-coated 200-mesh copper grids, and negatively stained with 1% w/v uranyl acetate in 0.01 M PB, and then observed by an H-7700 transmission electron microscope (Hitachi, Tokyo, Japan) at 80 kV.

dsRNA Extraction and Enzymatic Treatments.

A mycelial plug of the respective fungal culture was grown in PDB with shaking at 150 rpm at 28 °C for 7 d. Mycelia were harvested and ground to a fine powder in liquid nitrogen. A previously described procedure for dsRNA extraction was referred (59). dsRNA pellet was washed three times, then suspended in 100 µL DEPC-H₂O, and treated with DNase I and S1 nuclease (TaKaRa, Dalian, China) according to the instructions. The digestion products were separated using 1.2% agarose gel and visualized by ethidium bromide staining, then purified with the Wizard SV Gel and PCR Clean-Up System (Promega).

Elimination of the Mycovirus from Strain Mf_KF18.

Virus-free isogenic fungal isolates are needed to test the effects of mycoviruses on a host fungus. Four treatments were used to eliminate RNA viruses from Mf_KF18: 1) Hyphal tips were cultured as described by Kamaruzzaman et al. (60). 2) Conidia were serially diluted and plated to produce single-spore colonies as described by Li et al. (61). 3) Protoplasts were prepared and regenerated as described by Kamaruzzaman et al. (60). 4) Conidia were plated on PDA amended with 500 μM ribavirin (1-β-d-ribofuranosyl-1,2,4-triazole-3-carboxamide), a guanine nucleoside analog that is a broad-spectrum inhibitor of RNA virus replication (62). RNA was extracted from each putative virus-free colony and tested for the virus by RT-PCR using specific primers (SI Appendix, Table S6).

Protoplast Preparation and Transfection with the Virus.

Conidia (1 × 10⁸/mL) of virus-free strains Ma_114445 of M. anisopliae and Mp_336563 of M. pingshaense were incubated overnight in PDB to generate hyphae. Protoplasts were prepared by incubating the hyphae in a solution of 0.02% w/v driselase and w/v 0.02% snailase (Solarbio, Beijing, China) in 0.7 M NaCl for 30 min at 28 °C on a shaker at 100 rpm to release protoplasts. The mixture was filtered through Miracloth (Merck, Billerica, MA) and then centrifuged at 5,000 rpm for 10 min; the resulting protoplast pellet was washed twice with 0.7 M NaCl and resuspended in 0.5 mL of STC buffer (1 M sorbitol, 50 mM Tris pH 8.0, and 50 mM CaCl₂·2H₂O). Purified VLPs (5 µL) were added to 100 µL of protoplasts (about 1 × 10⁸/mL) for a transfection assay; PEG-mediated transfection tests were conducted according to previous reports (14, 60). All transfected, regenerated isolates were subcultured at least three times to confirm the stability of the mycovirus in the host.

Biological Characterization of MfPV1-Infected Strains of M. anisopliae and M. pingshaense.

To compare virus-infected and virus-free strains of M. anisopliae and M. pingshaense for any phenotypic differences, we evaluated colony morphology, growth, conidiation, biomass, and virulence in P. xylostella and S. frugiperda. A mycelial plug approximately 0.5 cm in diameter was removed from the edge of actively growing colonies and placed in the center of a PDA plate. Colony diameter was measured every 24 h. Colonies were also photographed at 5, 7, 9, and 21 dpi to compare colony characteristics of the colonies. Conidia were harvested in sterile water containing 0.05% Tween-80 and counted using a hemocytometer (63). For weighing mycelia, a mycelial plug of the respective strains was incubated in PDB at 150 rpm at 28 °C for 5 d, then the mycelium was harvested, dried, and weighed using a precision balance (the accuracy to 1 mg). To evaluate chemical stress tolerance of the isolates, 2 µL of a conidial suspension (1 × 10⁷/mL) was spotted onto 1/4 Sabouraud dextrose agar with yeast extract medium (1/4 SDAY: glucose 10 g, tryptone 2.5 g, yeast extract 5 g) amended with MNB (10 μg/mL), Congo red (500 μg/mL), NaCl (1 M), sorbitol (1 M), and SDS (400 μg/mL), and then incubated for 5 d as above. Colony diameters were measured to calculate the mycelial growth (23). Five replicates were used for each strain. Means for a variable for the virus-infected and virus-free strains were compared for significant differences at P < 0.05 using a paired t test in SPSS v22.0 (IBM, Armonk, NY).

Virulence of virus-infected and virus-free strains was bioassayed using second-instar larvae of P. xylostella that had been immersed in a conidial suspension of the fungus (5 × 10⁷ conidia/mL 0.05% Tween-80) for 60 s then incubated at 20 °C with 16 h light/8 h dark or second-day larvae of S. frugiperda that were allowed to feed on corn leaves that had been immersed in conidial suspension (5 × 10⁸ conidial/mL0.05% Tween-80) for 60 s then incubated at 28 °C with 16 h light/8 h dark. Larvae were treated with 0.05% Tween-80 as the negative control. Any dead larvae were counted every 24 h, and the median lethal time (LT50) was estimated. Each fungal strain was tested using 30 larvae of P. xylostella and 24 larvae of S. frugiperda, and the experiment was done three times. Means for a variable for the negative control, virus-infected, and virus-free strains were compared for significant differences at P < 0.05 using ANOVA and post hoc least significant difference (LSD) test for multiple comparisons in SPSS v22.0 (IBM).

RNA-seq of MfPV1-Infected and Virus-Free Strains of M. anisopliae and RT-qPCR Validation of Selected DEGs.

Total RNA was extracted from M. anisopliae strains Ma_114445 and Ma_114445/MfPV1 that had been cultured for 3 d in cicada induced medium (PDB supplemented with 0.2% cicada slough). Total RNA from three biological replicates was pooled for each RNA-seq library. cDNA library was constructed and then sequenced on an Illumina Novaseq platform according to the manufacturer’s instructions, with 150 bp paired-end reads generated (Novogene, Beijing, China). Raw data in the FASTQ format were processed using the open-source software fastp (HaploX Biotechnology, Shenzhen, China) (64); clean reads were obtained by removing reads containing adapter, ploy-N, and low-quality reads from raw data. Pearson correlation coefficient (r) and principal component analysis (PCA) were performed using ggplot2 packages in R based on the fragments per kilobase per million mapped fragments (FPKM) of all genes in each RNA-seq sample. The heatmap was performed using ggplot2 and pheatmap packages in R package. An index of the reference genome was built and paired-end clean reads aligned to the reference genome using Hisat2 v2.0.5 (65). Genes with an adjusted P-value ≤0.05 and fold-change ≥1.2 found by DESeq2 were assigned as DEGs. A gene ontology (GO) enrichment analysis of the DEGs was implemented using the R package clusterProfiler (http://bioconductor.org/packages/release/bioc/html/clusterProfiler.html), in which gene length bias was corrected (66). The fold-change in the expression of the DEGs based on the RNA-seq data was then validated by RT-qPCR using gene-specific primers designed using the software Oligo 7.60 (OLIGO, Colorado Springs, CO) (SI Appendix, Table S6). Total RNA was used to construct template cDNA for the RT-qPCR using Hifair III first Strand cDNA Synthesis SuperMix (Yeasen, Shanghai, China) according to the manufacturer’s instructions. The Hieff UNICON Universal Blue qPCR SYBR Green Master Mix (Yeasen) was used for the RT-qPCR in a QuantStudio 6 Flex Real Time PCR system (Thermo Fisher Scientific). Transcript levels of the selected genes were normalized against the expression of the gene for glyceraldehyde-3-phosphate dehydrogenase (GAPDH). The relative quantitative method (2^−ΔΔCT) was used to infer the fold-change in expression of the selected genes (67).

RT-qPCR Validation of Sporulation-Related Genes.

Total RNA was extracted from M. anisopliae strains Ma_114445 and Ma_114445/MfPV1 that had been cultured for 12 d on PDA medium. Sporulation-related genes were validated by RT-qPCR performed as described above. Gene-specific primers were designed using the software Oligo 7.60 (OLIGO) (SI Appendix, Table S6).

Supplementary Material

Appendix 01 (PDF)

Dataset S01 (XLSX)

Dataset S02 (XLSX)

Data, Materials, and Software Availability

Statistical analysis was conducted using the SPSS Statistics (version 19) program with one-way ANOVA and a post hoc LSD test (for multiple comparisons; P < 0.05). For two-group comparisons, two-tailed Student’s t test was conducted with a P value smaller than 0.05 being considered statistically significant. The two cDNA sequences of MfPV1 dsRNAs have been deposited in GenBank under Accession Nos. MH143600 (68) and MH143601 (69). The RNA-seq raw data have been deposited in NCBI under GEO Accession No. GSE266301 (70). All other data are included in the manuscript and/or supporting information.