Functional Analysis of Keto-Acid Reductoisomerase ILVC in the Entomopathogenic Fungus Metarhizium robertsii

Ketol-acid reductoisomerase (ILVC) is the second enzyme in the branched-chain amino acid (BCAA) biosynthesis, which regulates many physiological activities in a variety of organisms from bacteria to fungi and plants. In this work, function mechanisms of ILVC in Metarhizium robertsii Metchnikoff (Hypocreales: Clavicipitaceae) were explored with site-directed mutagenesis, reductase activity assays and transcriptomics analysis. The reductase activity assays showed that ILVC from phytopathogenic fungi exhibited significantly higher activities than those from entomopathogenic fungi but lower than those from yeast. Site-directed mutagenesis and enzymatic activities of MrILVC with different active-site mutants (Arg-113, Ser-118, Asp-152, Asp-260, and Glu-264) confirmed that active sites of MrILVC are conserved with plant and bacterial ILVCs. Deleting MrilvC causes the complete failures of vegetative growth and conidial germination, feeding with branched-chain amino acids (BCAAs) recovers the fungal growth but not conidial germination, while both characteristics are restored when supplemented with yeast extract. Compared to ΔMrilvC cultured in czapek agar (CZA), plenty of genes involved in the biosynthesis of antibiotics and amino acids were up- or down-regulated in the wild type or ΔMrilvC feeding with either BCAAs or yeast extract. Further analysis showed some genes, such as catalase A, participate in mycelial growth and conidial germination was down-regulated in ΔMrilvC from CZA, revealing that MrILVC might control the fungal development by gene regulation and BCAAs or yeast extract could play partial roles of MrILVC. This study will advance our understanding of ILVC function mechanisms in fungi.


Introduction
Valine, leucine, and isoleucine form the small group of branched chain amino acids (BCAAs), which are synthesized in bacteria, plants, and fungi, but not in animals [1,2]. The biosynthesis of these amino acids shows high similarity between organisms of the different kingdoms of life, of which valine and isoleucine synthesis are carried out by the same enzymes, and leucine is created from α-ketoisovalerate (a transamination precursor of valine) [3]. However, striking differences on the capabilities to synthesize or degrade branched-chain amino acids exist, and BCAAs in different groups of organisms also have some unique functions and features, revealing the various function mechanisms of BCAAs among different organisms [4].
ILVC is a bifunctional enzyme that catalyzes two quite different reactions, but occur at a common active site, acting both as an isomerase and as a reductase [11]. ILVCs have a GxGxx(G/A)xxx(G/A) motif as a binding site for NADP(H), and Mg 2+ is required for binding NADP(H) [9,12,13]. Site-directed mutagenesis and activity assays showed that plant and bacterial ILVCs have evolved different mechanisms of induced fit to prepare the active site for catalysis [9,12]. Previous sequence comparison revealed that those active residues are highly conserved across plant, fungal, and bacterial ILVCs, but the gene or genes responsible for ILVC activity in fungi have not been investigated in any detail. In addition, functional analysis of the ILVC has been carried out in bacteria, plants, and fungi; however, transcriptome analyses for the blocking of BCAA biosynthesis remain unclear [6,14].
In our previous study, we found that ILVC is associated with conidial germination and fungal pathogenicity in the insect-pathogenic fungus Metarhizium robertsii Metchnikoff (Hypocreales: Clavicipitaceae), formerly classified as Metarhizium anisopliae, and the ilvC-deleted mutant failed to germinate on CZA complemented with three BCAAs. Complementation of BCAAs cannot recover the conidial germination for the ilvC-deficient strain [14]. This study seeks to further characterize ilvC in M. robertsii by phenotypic and comparative analyses of its deletion mutants. Relative catalytic activities of ilvC mutants in five active-site residues were assayed and catalytic activities of ILVCs from different fungi were compared.

Site-Directed Mutagenesis, Protein Expression and Purification
Site-directed mutagenesis was performed with TaKaRa MutanBest kit (TaKaRa, Beijing, China). Briefly, the ilvC sequence was amplified by PCR from the cDNA template and PCR products were inserted into pOT2 plasmid. pOT2-Ilvc plasmid was amplified with designed mutant primers by PCR, and blunted using Blunting Kination Enzyme Mix, then ligated using ligation solution I in the kit. Plasmids were transformed into Escherichia coli Migula (Enterobacterales: Enterobacteriaceas) DH5α (TransGen, Beijing, China) and verified by DNA sequencing. Five amino acid residues contacting both NADP(H) and Mg 2+ are conserved among bacteria, fungi, and plants; thus, their 5 active-site residue mutageneses were performed as mentioned above, respectively [15].
The recombinant ILVC protein was expressed in E. coli [16]. Briefly, ilvC with a 6× Histag sequence at the C-terminus was amplified by PCR from the cDNA template or pOT2-Ilvc, and PCR products were inserted into pET-28b (+) vector (Novagen, Beijing, China), then transformed into E. coli strain BL21 (DE3)-competent cells (TransGen, Beijing, China). ILVC protein expression was induced by the addition of isopropyl β-D-thiogalactoside (IPTG) to a final concentration of 0.5 mM and purified with Ni-NTA agarose (Qiagen, Chatsworth, CA, USA) using a nickel-ion affinity column (Qiagen). Protein purity was monitored by SDS-PAGE.

Reductase Activity Assays
The enzymatic activity assays for ILVC protein were performed [15,17,18]. Firstly, ILVC protein (1 µg) was pre-incubated in the activity assay buffer (50 mM Tris-HCl pH 7.0, 1 mmol DTT, 200 mmol NADPH, 10 mmol MgCl 2 ) at room temperature for 30 min. Secondly, 2-Acetolactate was added to the reaction mixture. The enzymatic activity was determined by monitoring the rate of NADPH oxidation at 340 nm. Six independent experiments were repeated. In this study, reductase activity of ILVC from S. cerevisiae, A. fumigatus, F. graminearum, Beauveria bassiana (Hypocreales: Cordycipitaceae) and M. robertsii were determined by protein expression and purification from E. coli.

The Sensitivity to Herbicides
Mycelial growth tests were performed on SDAY plates supplemented with the following AHAS inhibitors at the concentrations of 10 µg mL −1 : tribenuron-methyl, rimsulfuron, metsulfuron-methyl, halosulfuron-methyl, chlorimuronethyl, imazethapyr and imazapyr (Henan Tianfu Chemical Co., Ltd., Henan, China). The percentage of inhibition of mycelial radial growth was calculated using the formula: mycelial growth inhibition = [(C − N)/ (C − 5)] × 100, where C is colony diameter of the control, and N is that of the treatment with AHAS inhibitors [6]. Each experiment was carried out with six replicates.

Quantitative RT-PCR Analysis of MrilvC
100-µL aliquots of a conidial suspension collected from strains cultured on SDAY were spread on SDAY at 28 • C in the dark; at 18 h, 48 h, and 96 h, cultured strains were collected. Conidial suspensions were put on cicada (Cryptotympana atrata Westwood (Hemiptera: Cicadidae)) hind wings [19]. For the samples of M. robertsii during the infection progress, conidia were used for cuticle infection by dipping the fifth-instar Galleria mellonella L. (Lepidoptera: Pyralidae) larvae into a conidial suspension (1 × 10 7 conidia mL −1 ) for 10 s. Then, the infected larvae were maintained in a stable photoperiod with 16 h of light and 8 h of dark at 28 • C without feeding. Infected G. mellonella were collected at 24 h, 48 h, and 96 h. Quantitative RT-PCR was performed [20]. In brief, TRIzol reagent (Invitrogen, Carlsbad, CA, USA) was used for RNA extraction, and complementary DNA were synthesized with a gDNA remover kit (Toyobo, Osaka, Japan) as the manufacturer's instructions. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as the control. Gene amplification was achieved using the Bio-Rad CFX96 real-time PCR detection system (Bio-Rad, Hercules, CA, USA). Transcripts of MrilvC were normalized to the control gene, and the 2 −∆∆Ct method was used to calculate the relative expression level (mean ± standard deviation) of each gene. Three independent experiments were performed for each gene.

Phenotypic Assays
A 4-mm-diameter agar plug of each strain from the border of a 5-day-old colonies on SDAY plates was centrally attached onto CZA (0.3% NaNO 3 , 0.1% K 2 HPO 4 , 0.05% MgSO 4 ·7H 2 O, 0.05% KCl, 0.001% FeSO 4 ·7H 2 O, 3% sucrose, and 1.5% agar, w/v), CZA + AA (CZA with 1 mM Ile, Leu, and Val) or CZA + yeast (CZA with 1% yeast extract powder, w/v) at 28 • C under a 12 h:12 h light/dark cycle. The radial growth (colony diameter) of the vegetative mycelia was measured daily from the back sides of the plates with vernier caliper. For the assessment of different strain conidial yield capacity, 100-µL aliquots of a conidial suspension were spread on CZA, CZA + AA, and CZA + yeast. From day 3 onwards, 5-mm diameter colony disks were excised at random from the plates daily and washed off into 2 mL of 0.05%. The concentration of the conidial suspension was determined by a hemocytometer under a microscope.

Transcriptomics Analysis
A 4-mm-diameter agar plug of each strain from the border of a 5-day-old colony containing both mycelium and conidium on SDAY plates was centrally attached onto CZA, CZA + AA, and CZA + yeast plates for 24 h when both mycelial growth and conidial germination happened. RNA samples, library construction, and sequencing were performed [21]. Total RNA was extracted and treated with recombinant DNase I (rDNase I; Sigma, St Louis, MO, USA) to remove any potential genomic DNA contamination. To maximize target coverage, equal amounts of total RNA from the three replicates of strains were pooled as a sample library for RNA-Seq library construction, and then a total of 5 µg of total RNA were prepared for RNA-Seq.
Total RNA was purified with oligo(dT) beads and fragmented into small pieces, treated with RNase H for RNA removement after the complementary DNA synthesized. The cDNA library was constructed and sequenced using the Illumina Hi-Seq 4000 platform with a single-end (single reads of 50 bases) sequencing strategy at Beijing Genomics Institute (BGI, Shenzhen, China). Clean reads were obtained by removing the raw reads, then used to map the M. robertsii genome (https://www.ncbi.nlm.nih.gov/genome/?term=Metarhizium+ robertsii (accessed on 31 January 2020)). The gene expression level was normalized using the fragments per kilobase per million mapped reads (FPKM) method [22]. Rigorous algorithms were applied to identify differential expressed genes (DEGs) with a false discovery rate (FDR) ≤ 0.001 and an absolute value of the log 2 ratio ≥ 1 as thresholds [23]. Gene ontology (GO) functional analysis (Blast2GO, https://www.blast2go.org/ (accessed on 13 March 2020)) and Kyoto Encyclopedia of Genes and Genome (KEGG) pathway enrichment analysis was performed to further understand the biological functions of the identified DEGs. In addition, a total of 20 genes involved in conidial germination or mycelial growth in this study were selected to perform quantitative RT-PCR validation. All primers used in this study were listed in Table S1.

Statistical Analysis
Data were analyzed by one-way analysis of variance, followed by Dunnett's multiple comparison and Tukey's test using the SPSS v23.0 software. All results were expressed as the mean ± standard error of the mean (SD). p < 0.05 was considered statistically significant.

Phylogenetic Analysis of ILVC from Eukaryotes
Phylogenetic analysis showed that ILVC protein from fungi, plants, and bacteria were clustered in respective clades. Analysis of the domain architecture showed that ILVC protein from plants had the shortest conserved domain (ILVC superfamily) and bacterial ILVC sequences were the longest in length among the eukaryotes ( Figure 1A). Five amino acid residues contacting both NADP(H) and Mg 2+ are also conserved in fungi, bacteria, and plants, indicating that all these ILVCs have either reductase or isomerase activity [11]. In M. robertsii, residues contacting both NADP(H) and Mg 2+ are Arg-113, Ser-118, Asp-152, Asp-260, and Glu-264 ( Figure 1B).

Enzymatic Activity of ILVC from Different Fungi and MrILVC with Active-Site Mutant
As shown in Figure 2A, ILVC from phytopathogenic fungi (A. fumigatus and F. graminearum) exhibited significantly higher activities than those from entomopathogenic fungi (B. bassiana and M. robertsii) but lower than that from S. cerevisiae. All the mutants (R113A, S118A, D152G, D260K, and E264A) exhibited significantly reduced activities compared to those of WT ( Figure 2B), confirming that the roles of those residues on ILVC activities were conserved among different fungi [15].

ILVC Contributes to Mycelial Growth and Conidial Germination
MrilvC expression sharply increased expression at the stage of conidial germination and decreased rapidly from 60 h ( Figure 4A). Additionally, MrilvC expression, increased rapidly after infection with conidia, was highly expressed from 96 h-I when M. robertsii has entered into the hemocoel.
Conidia of the ∆MrilvC mutant were unable to germinate on CZA medium supplied with all three BCAAs (Ile, Leu and Val), but were able to germinate on CZA medium supplemented with yeast extract ( Figure 4B). The MrilvC-deficient strain was unable to grow on CZA plates, but could grow on CZA medium supplemented with BCAAs or yeast extract despite the colony being smaller than that of the WT (Figure 4C,D). In addition, in contrast to the WT and complemented transformant, the MrilvC-deficient strain showed a dramatic decrease in the conidial yields in CZA medium supplemented with BCAAs or yeast extract ( Figure 4E).

RNA-Seq Analysis of ∆MrilvC on Different Medium
Totals of 10,934 (WT) and 10,869 (CZA) genes were detected in the WT and ∆MrilvC mutant grown on CZA plates; 11,011 (CZA + AA) and 11,012 (CZA + yeast) genes expressed in the ∆MrilvC mutant grow on CZA + AA and CZA + yeast plates ( Figure 5A). Of these genes, 312 and 247 genes were found to be specifically expressed in WT and CZA, respectively, indicating the potential roles of MrILVC in their expression. In addition, 82, 87, and 111 genes were expressed specifically in CZA, CZA + AA and CZA + yeast samples, respectively. Compared with the WT, 1524 genes were significantly up-regulated while 2556 genes were significantly down-regulated after the deletion of MrilvC in M. robertsii grown on CZA plates ( Figure 5B). It is worth mentioning that, compared with the WT, all genes encoding key enzymes of BCAA biosynthesis were significant up-regulated in the CZA, indicating that the disruption of BCAA metabolism contributes to the expression levels of those key enzymes involved in BCAA biosynthesis ( Figure S1). Moreover, a total of 3165 and 4020 genes were differentially expressed between CZA + AA/CZA + yeast and CZA samples, respectively, and 2536 DEGs were found between CZA + AA and CZA + yeast samples. Based on sequence homology, the sequences significantly enriched 22 functional groups ( Figure 5C). Catalytic activity, binding, cellular anatomical entity, intracellular and cellular process, and metabolic process were the most commonly annotated terms in each of the three GO term categories. The DEGs of WT/CZA were enriched in 17 KEGG pathways (p < 0.05) ( Table 1). Of these pathways, biosynthesis of antibiotics and biosynthesis of amino acids were the most highly enriched KEGG pathways, indicating that the severity affection of normal biosynthesis in M. robertsii is caused by the MrILVC deletion. The biosynthesis of antibiotics and biosynthesis of amino acids were also the most enriched pathways in CZA + AA/CZA and CZA + yeast/CZA, but not in CZA + yeast/CZA + AA. As mentioned above, plenty of DEGs, the expression levels of these genes were validated using quantitative RT-PCR, from WT/CZA (116), CZA + AA/CZA (95), and CZA + yeast/CZA (106) were enriched in the biosynthesis of amino acids. In addition to DEGs involved in BCAA biosynthesis, some DEGs participated in the biosynthesis of other amino acids, such as lysine, proline, and arginine (Table S2). Glycerolipid metabolism  -31  37  -ko00052  Galactose metabolism  -22  --ko00330  Arginine and proline metabolism  -41  --ko00460  Cyanoamino acid metabolism  -27  36  -ko00300 Lysine biosynthesis -14 --ko00520 Amino sugar and nucleotide sugar metabolism ---109 ko00072 Synthesis and degradation of ketone bodies --11 -ko00450 Selenocompound metabolism --12 -ko01200 Carbon metabolism --96 -ko00240 Pyrimidine metabolism --31 -WT: The WT strain cultured on CZA plates; CZA: ∆MrilvC mutant grown on CZA plates; CZA + AA: ∆MrilvC mutant grown on CZA + AA plates; CZA + yeast: ∆MrilvC mutant grown on CZA + yeast plates.

MrILVC Affects Expression of Genes Involved in Mycelial Growth and Conidial Germination
Venn analysis showed that a total of 459 DEGs were found between conidial germinated and inhibited samples ( Figure 6A and Table S3). Among of them, more than 100 DEGs were significantly enriched in seven function classes with function classification, for which biosynthesis of antibiotics (40 DEGs) was the most highly enriched pathway. Based on the expression patterns, DEGs with higher expression levels in conidial germinated than inhibited samples were listed in Table 2. Among of them, catalase A and SAM-dependent methyltransferase genes played important roles in the fungal conidial germination [24,25].  Venn analysis of DEGs showed 2037 DEGs co-existed in WT/CZA, CZA + AA/CZA, and CZA + yeast/CZA, and were significantly enriched in 22 function classes ( Figure 6B and Table S4). DEGs with higher expression levels in mycelial growth than inhibited samples were listed in Table 3. Some genes involved in mycelial growth, such as Alcohol dehydrogenase, catalase A and chitinase were lower expressed in CZA than other samples [26][27][28]. In addition, 20 genes involved in conidial germination or mycelial growth were shown to be up-regulated in WT/CZA, CZA + AA/CZA, and CZA + yeast/CZA from RNA-Seq data. Quantitative RT-PCR analysis revealed 19, 18, and 18 genes were significantly upregulated in WT/CZA, CZA + AA/CZA, and CZA + yeast/CZA, respectively, supporting the validity of our transcriptomics results ( Figure S2).

Discussion
ILVCs can be separated into two classes based on the length of their polypeptides: class I proteins with~340 amino acid residues, and class II proteins with more than 520 residues [12,13]. In our study, ILVCs in fungi had various lengths from 400 to 500 residues, which were located between the plant and the bacteria in length. Domain architecture analysis showed that all ILVCs have the IlvC domain, and plant ILVCs have a C-terminal extension, play dual roles in forming the active site, but are missing in ILVCs of both bacteria and fungi, confirming that ILVCs from bacteria and fungi belong to the class I group and suggesting they may have similar function mechanisms which were different from plant ILVCs [13]. Moreover, multiple-sequence alignment of ILVCs showed that residues constituting NADP(H) and Mg 2+ binding sites are well conserved, while the overall length of each KARI protein is different.
ILVCs from different organisms have evolved different mechanisms of induced fit to prepare the active site for catalysis [12]. Enzymatic activity assays showed that ILVC from phytopathogenic fungi exhibited significantly higher activities than those from entomopathogenic fungi but lower than that from S. cerevisiae, suggesting that there is a direct link between the ILVC catalysis activity and the characteristics of fungal life cycles. ILVC and its orthologous proteins have either reductase or isomerase activity, but only the reductase activity was explored in this study because of the difficulty and reaction complex of isomerase action [11,15]. Structure-based multiple sequence alignment of ILVCs from bacteria and plants revealed that five amino acid residues contacting both NADP(H) and Mg 2+ are conserved [15]. In our study, sequence alignment and enzymatic activities of MrILVC with different active-site mutants confirmed that those residues are also conserved in fungal ILVCs.
In this study, all tested AHAS inhibitors showed significant inhibitory effects with 10~40% mycelial growth inhibition on M. robertsii, revealing that the use of entomopathogenic fungi to control pests should be avoided during the AHAS inhibitors' used area. Consistent with the observation in F. graminearum, deleting ilvC showed significant resistance to tribenuronmethyl, rimsulfuron, halosulfuron-methyl, chlorimuronethyl, imazethapyr and imazapyr, suggesting that those six AHAS inhibitors might inhibit the fungal ILVC activity [6].
In this study, qRT-PCR showed MrilvC expression sharply increased expression at the stage of conidial germination and infection, which is consistent with the impairment of the full virulence on insects treated with direct injection or cuticle infection in our previous study [14]. BCAAs account for more than 20% of total protein amino acids and are an important constituent of the amino acid pool in organisms [29]. Transcriptomics analysis showed lots of genes involved in both metabolism and biosynthesis of amino acids were differentially expressed after the MrilvC deletion, which implied that the blocking of BCAAs impaired protein functioning in those pathways.
Catalase A, a monofunctional catalase in fungi, played crucial roles in various aspects of cell physiology and cellular differentiation, including conidial germination and mycelial growth [25,30]. In M. robertsii, the expression of catalase A was down-regulated after the deletion of MrilvC, while it returned to an expression level similar to the WT in ∆MrilvC feeding with yeast extract. In addition, some genes involved in conidial germination or mycelial growth, such as Alcohol dehydrogenase, SAM-dependent methyltransferase, and chitinase, showed similar expression patterns with catalase A. From these results, MrilvC played critical roles in fungal development by regulating those gene expressions, and characteristics of ∆MrilvC could be partially restored to the WT with gene regulation by feeding with BCAAs or yeast extract. It is worth mentioning that deleting MrilvC led to the noticeable increase in the expression of genes involved in BCAA biosynthesis, and we speculated that BCAA levels could affect its biosynthesis and the disruption of BCAA biosynthesis by deleting MrilvC, resulting in those genes being significant up-regulated; more experiments are needed to explore those patterns.
In conclusion, our data indicate that MrILVC plays critical roles in controlling conidial germination and mycelial growth by the regulation of metabolism and biosynthesis of amino acids. ILVCs from different fungi revealed distinct reductase activity. Active sites of MrILVC are conserved with plant and bacterial ILVCs; one mutation among those exhibited significantly reduced activities. Data from this study advance our understanding of the function of ILVCs in entomopathogenic fungi, which would contribute to future BCAA metabolism investigations.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/jof7090737/s1, Figure S1: Expression levels of genes involved in BCAA pathways from different samples, Figure S2: Quantitative RT-PCR validation of differential expressed genes identified in RNA-Seq analysis. WT: The WT strain cultured on CZA plates; CZA: ∆MrilvC mutant grow on CZA plates; CZA + AA: ∆MrilvC mutant grow on CZA + AA plates; CZA + yeast: ∆MrilvC mutant grow on CZA + yeast plates, Table S1: Primers used in this study, Table S2: Expression of genes involved in biosynthesis of amino acids, Table S3: Differential expressed genes between conidial germination and inhibited samples, Table S4: Differential expressed genes between mycelial growth and inhibited samples.