Two FgLEU2 Genes with Different Roles in Leucine Biosynthesis and Infection-Related Morphogenesis in Fusarium graminearum

3-isopropylmalate dehydrogenase (IPMD) encoded by LEU2 is a key enzyme in leucine (Leu) biosynthetic pathway. Analysis of the genome sequence of Fusarium graminearum revealed two paralogous LEU2 genes (designated as FgLEU2A and FgLEU2B) in this fungus and the deduced amino acid sequences of FgLeu2A and FgLeu2B share 45% identity. Targeted disruption of individual FgLEU2A/B gene in F. graminearum assigned a more crucial role of FgLeu2A in Leu biosynthesis as disruption of FgLEU2A resulted in mutant (ΔFgLeu2A-10) that was Leu-auxotrophic and could not grow in minimal medium limited for amino acids, whereas FgLEU2B deletion mutant ΔFgLeu2B-2 was morphologically indistinguishable from the wild type strain PH-1. The growth defects of ΔFgLeu2A-10 could be overcome by exogenous addition of Leu at 0.25 mM. Double deletion of FgLEU2A and FgLEU2B (ΔFgLeu2AB-8) caused a more severe Leu-auxotrophic phenotype as the concentration of Leu exogenously added to medium to rescue the growth defect of ΔFgLeu2AB-8 should be raised to 1.25 mM, indicating a less important but nonnegligible role of FgLeu2B in Leu biosynthesis. Disturb of Leu biosynthesis caused by FgLEU2A deletion leads to slower growth rate, reduced aerial hyphal formation and red pigmentation on PDA plates and completely blocked conidial production and germination. All of the defects above could be overcome by Leu addition or complementation of the full-length FgLEU2A gene. ΔFgLeu2A-10 also showed significantly increased sensitivity to osmotic and oxidative stresses. Pathogenicity assay results showed that virulence of mutants lacking FgLEU2A were dramatically impaired on wheat heads and non-host cherry tomatoes. Additionally, a low level of deoxynivalenol (DON) production of ΔFgLeu2A-10 and ΔFgLeu2AB-8 in wheat kernels was also detected. Taken together, results of this study indicated a crucial role of FgLeu2A and a less important role of FgLeu2B in Leu biosynthesis and fungal infection-related morphogenesis in F. graminearum and FgLeu2A may serve as a potential target for novel antifungal development.

3-isopropylmalate dehydrogenase (IPMD) encoded by LEU2 is a key enzyme in leucine (Leu) biosynthetic pathway. Analysis of the genome sequence of Fusarium graminearum revealed two paralogous LEU2 genes (designated as FgLEU2A and FgLEU2B) in this fungus and the deduced amino acid sequences of FgLeu2A and FgLeu2B share 45% identity. Targeted disruption of individual FgLEU2A/B gene in F. graminearum assigned a more crucial role of FgLeu2A in Leu biosynthesis as disruption of FgLEU2A resulted in mutant (ΔFgLeu2A-10) that was Leu-auxotrophic and could not grow in minimal medium limited for amino acids, whereas FgLEU2B deletion mutant ΔFgLeu2B-2 was morphologically indistinguishable from the wild type strain PH-1. The growth defects of ΔFgLeu2A-10 could be overcome by exogenous addition of Leu at 0.25 mM. Double deletion of FgLEU2A and FgLEU2B (ΔFgLeu2AB-8) caused a more severe Leu-auxotrophic phenotype as the concentration of Leu exogenously added to medium to rescue the growth defect of ΔFgLeu2AB-8 should be raised to 1.25 mM, indicating a less important but nonnegligible role of FgLeu2B in Leu biosynthesis. Disturb of Leu biosynthesis caused by FgLEU2A deletion leads to slower growth rate, reduced aerial hyphal formation and red pigmentation on PDA plates and completely blocked conidial production and germination. All of the defects above could be overcome by Leu addition or complementation of the full-length FgLEU2A gene. ΔFgLeu2A-10 also showed significantly increased sensitivity to osmotic and oxidative stresses. Pathogenicity assay results showed that virulence of mutants lacking FgLEU2A were dramatically impaired on wheat heads and non-host cherry tomatoes. Additionally, a low level of deoxynivalenol (DON) production of ΔFgLeu2A-10 and ΔFgLeu2AB-8 in wheat kernels was also detected. Taken together, results of this study indicated a crucial role of FgLeu2A and a less important role of FgLeu2B in Leu biosynthesis and fungal infection-related morphogenesis in F. graminearum and FgLeu2A may serve as a potential target for novel antifungal development.

Introduction
Branched-chain amino acids (BCAAs) including leucine (Leu), valine (Val) and isoleucine (Ile) can be synthesized in bacteria, fungi and high plants, but not mammals. Mammals must acquire BCAAs from their diet and the absence of BCAAs biosynthetic pathway in mammals make the enzymes involved attractive targets for the development of herbicides and antimicroorganisms [1,2]. Recent studies have revealed that BCAAs, especially Leu, not only serve as fundamental substrates for protein synthesis but also have unique ability to initiate signal transduction pathways that modulate translation initiation [3,4].
Leu is synthesized as a branch that leads from 2-ketoisovalerate (2-KIV), the intermediate precursor of Val, to Leu and its biosynthesis contains four enzymatic reactions catalyzed by isopropylmalate synthase (IPMS, encoded by LEU4), isopropylmalate isomerase (IPMI, encoded by LEU1), isopropylmalate dehydrogenase (IPMD, encoded by LEU2) and branchedchain aminotransferase (BCAT, encoded by BCA1/2) [1]. The third Leu biosynthetic specific enzyme IPMD catalyzes the conversion of 3-isopropylmalate (3-IPPM) to 2-ketoisocaproate (2-KIC) and its encoding gene LEU2 is one of the best-studied BCAA biosynthetic genes because of its wide use as selected marker for gene-transformation experiments in Saccharomyces cerevisiae and several other yeast species [5,6].
In S.cerevisiae, functional analysis showed that disruption of LEU2 resulted in Leu-auxotroph and decreased resistance to acetic acid and to very high levels of ethanol stress of the mutants [7]. Very few documents characterized functions of LEU2 in filamentous fungi except for a relatively ancient and primary research in Aspergillus niger showed that this fungus elaborates two isoenzymes of IPMD and these two encoding genes were differentially expressed [8]. However, the detailed function of these two paralogous LEU2 genes in Leu biosynthesis or other cellular processes was not investigated in A. niger or other filamentous fungi.
Fusarium head blight (FHB) caused by the pathogenic ascomycete fungus Fusarium graminearum (teleomorph Gibberella zeae) is one of the most important diseases in cereal crops including wheat and barley worldwide [9][10][11]. FHB epidemics would not only cause huge yield losses but also bring health threaten and food safety concern to humans and animals as the fungus produces mycotoxins such as deoxynivalenol (DON) during infection [9,12]. Disease control of FHB and mycotoxin contaminations remains challenging due to the lack of resistant wheat cultivars to FHB and limited effective fungicides against F. graminaerum. Therefore, it is extremely urgent to develop novel, less-toxic and effective fungicides for the management of FHB. Enzymes involved in certain amino acids biosynthesis (e.g., BCAAs) have been regarded as attractive fungicide targets due to their absence in mammals and the important biological functions of the genes encoding for these enzymes in yeasts and filamentous fungi [13][14][15][16][17][18][19][20][21].
In this study, we found two paralogous LEU2 genes in the genome of F. graminearum, using targeted-gene disruption strategy we have constructed single and double deletion mutants of FgLEU2A/B gene in F. graminearum and investigated the different roles of each gene in Leu biosynthesis and in other important cellular processes. Our results are in consistence with previous findings that BCAAs biosynthesis is important event and FgLeu2A plays more important roles in Leu biosynthesis and infection-related morphogenesis and can be regarded as potential target for novel fungicide development.

Strains, growth conditions and phenotype assays
The Fusarium graminearum strain PH-1 (a Michigan field isolate, USA origin, FGSC 9075, NRRL31084 [22]) was used as the parental wild-type. To assess the mycelial growth and colony characteristics, the wild-type strain PH-1 and mutants were cultured on different medium including potato dextrose agar (PDA), yeast extract peptone glucose agar (YEPD), fructose gelatin agar (FGA) or minimal medium (MM) plates supplemented with or without different amino acids and incubated at 25˚C. Colony diameter was measured every 12 h. Fungal biomass analysis was conducted by collecting mycelia from 2-days-old liquid culture, and then washed with sterilized water, dried in 65˚C bake oven and determined by mycelial dry weight. Experiments were performed triplicate. Escherichia coli DH5α was cultured at 37˚C and used for bacterial transformations. For conidiation assay, five 5-mm mycelial plugs of the wild-type strain and mutants taken from the edge of a 3-day-old colony were inoculated in a 50-ml triangular flask containing 20 ml of MBL (40 g mung beans boiled in 1 L water for 20 min, and then filtered through cheesecloth) for conidiation. The flasks were incubated at 25˚C for 4 days in a shaker (180 rpm). The amount of conidia produced by the F. graminearum strains in MBL was counted using a hemacytometer. Additionally, conidial germination was carried out by re-suspending the conidia of the wild-type strain and mutants in 2% sterilized sucrose solutions supplemented with or without different amino acids at 25˚C for 4 h and 6 h, conidial germination was examined under a Nikon ECLIPSE E100 microscope (Nikon Co., Tokyo, Japan). Each experiment was carried out with three replicates.

Sequence analysis of FgLEU2A/B genes in F. graminearum
Using the amino acid sequence of S. cerevisiae IPMD Leu2 (Saccharomyces database accession numberYCL018W) as query, two paralogous FgLeu2A (FGSG_06675.3) and FgLeu2B (FGSG_10671.3) were identified through a homology search in the F. graminearum genome database (available at http://fungidb.org/fungidb/). Transcript of the F. graminearum FgLEU2A or FgLEU2B gene was detected by using reverse transcription PCR (RT-PCR) with primer pairs Leu2A-all-F + Leu2A-all-R or Leu2B-all-F + Leu2B-all-R (S1 Table), using cDNA and genomic DNA as templates, respectively. RNA was extracted from mycelia of the wild-type PH-1 using a TaKaRa RNAiso Reagent (TaKaRa Biotech. Co., Dalian, China), and used for reverse transcription with the primer oligo(dT) 18 using a PrimeScript™ RT reagent kit (TaKaRa Biotech. Co.,Dalian, China). PCR amplifications were conducted using the following parameters: initial denaturation at 95˚C for 3min, followed by 30 cycles of denaturation at 94˚C for 45 s, annealing at 53˚C for 45s, extension at 72˚C for 1.5 min, and final extension at 72˚C for 10 min. PCR products were detected on 1% agarose gels in Tris-acetate (TAE) buffer and photographed after staining with ethidium bromide. The resultant PCR product was purified, cloned and sequenced.

Construction of FgLEU2A/B deletion mutants
FgLEU2A deletion vector pBS-leu2A-Del was constructed by inserting two flanking sequences of FgLEU2A gene into the left and right sides of the hygromycin resistance gene HPH in the pBS-HPH1 vector [23]. First, a 651-bp upstream flanking sequence fragment of FgLEU2A was amplified by using primer pair A1 + A2 (S1 Table) from the genomic DNA of PH-1, and was inserted into KpnI-XhoI sites of the pBS-HPH1 vector to generate pBS-leu2A-up. Subsequently, a 941-bp downstream flanking sequence fragment of FgLEU2A amplified from PH-1 genomic DNA using the primers A3 + A4 (S1 Table) was inserted into BamHI-SacI sites of the pBS-leu2A-up vector to generate pBS-leu2A-Del. Similar to pBS-leu2A-Del, we constructed FgLEU2B deletion vector pBS-leu2B-Del (see S2A and S3A Figs). FgLEU2A/B double deletion vector pBS-leu2AB-Del was constructed by modifying pBS-leu2A-Del. The G418 sulfate resistance gene NEO amplified by using primer pair Neo-F + Neo-R from plasmid pBS-RP-Red-A8-NEO (S1 Table) was introduced to the XhoI-BamHI sites of pBS-leu2A-Del. Finally, the 3095-, 2708-or 3455-bp fragments containing leu2A/B-upstream-HPH(NEO)-leu2A/B-downstream cassette was obtained by PCR amplification with primer pair A1 + A4 or B1+ B4 using diluted plasmid pBS-leu2A/leu2B/leu2AB-Del as templates. The resultant PCR product was purified and stored at -20˚C for protoplast transformation.
The PEG-mediated protoplast fungal transformation was performed as described previously [24]. PDA medium supplemented with hygromycin (100 mg/L) was used to select the resistant transformants. After single spore isolation and PCR identification, transformants were kept at 4˚C for further experiments.

Complementation of FgLEU2A deletion mutant
To confirm that the phenotype of FgLEU2A deletion mutant is caused by disruption of the gene, the FgLEU2A deletion mutant ΔFgLeu2A-10 (hereafter named ΔFgLeu2A-10) was complemented with the full-length FgLEU2A containing its promoter and terminator regions. The complemented plasmid pCA-leu2A-C was constructed using the backbone of pCAMBIA1300. First, G418 sulfate resistance cassette was amplified from plasmid pBS-RP-Red-A8-NEO with primers Neo-F + Neo-R (S1 Table), and cloned into the XhoI-KpnI site of pCAMBIA1300 to create plasmid pCA-NEO. Then, a 3437-bp of full-length FgLEU2A gene including 997-bp upstream and 1313-bp terminator region was amplified using primers leu2A-com-F + leu2Acom-R (S1 Table) from the genomic DNA of wild-type. This fragment was cloned into the BamHI-XbaI sites of pCA-NEO to generate the complement plasmid pCA-leu2A-C. Finally, the 4440-bp fragment containing neo-full-length leu2A cassette was obtained by PCR amplification with primer pair Neo-F + leu2A-com-R using diluted plasmid of pCA-leu2A-C as templates (see s S2A Fig). PCR products were separated on 1% agarose gels in TAE buffer and photographed after staining with ethidium bromide. The resultant PCR product was purified and stored at -20˚C for protoplast transformation. Transformation of the protoplasts of ΔFgLeu2A-10 with the full-length FgLEU2A was conducted as described above except that G418 sulfate (100 mg/L) was used as a selection agent.

Amino acids or intermediate products assays
Exogenous amino acids or intermediate products were supplemented to FGA at different concentrations as indicated in figures. Five 5-mm mycelial plugs of the wild-type strain and mutants were taken from the edge of a 3-day-old colony and incubated in the plates described above. Experiments were performed in triplicate and plates were photographed two days after inoculation.

Plant infection assays
Conidia were harvested from 4-day-old cultures of the wild-type strain and each mutant, then washed with sterilized water and re-suspended in 0.01% (vol/vol) Tween 20 solution and adjusted to 10 5 conidia/ml. Plant infection assays were performed using single floret injection method as previously described [25]. Briefly, 10 μl of conidial spores (10 5 conidia/ml) was injected into a single floret in the central section spikelet of single flowering wheat heads of susceptible cultivar Jimai19at early anthesis, 10 spikes were used for each strain. Infected spikelets in each inoculated wheat head were recorded 10 days after inoculation. The experiment was repeated for three times.
To examine the ability to colonize cherry tomato, a 10 μl aliquot of conidial suspension (10 5 conidia/ml) was injected into the wounded tomato after surface sterilization. There are five replicates for each strain. Inoculated tomatoes were incubated at 25˚C and 100% humidity with 12 h of daylight, and were photographed 3 days after inoculation. The experiment was repeated for three times.

Determination of sensitivity of the FgLEU2A/B mutants to cell stresses
Mycelial growth tests were performed on FGA plates amended with 1.25mMLeu and supplemented with the following chemicals: osmotic stress NaCl, D-sorbitol, KCl and glycerol at 1.2 M; oxidative stress H 2 O 2 at 12 mM; cell wall inhibitors congo red at 0.4 mg/ml and caffeine at 5 mM; 0.005% cell membrane damager SDS. The percentage of inhibition of mycelia radial growth (PIMG) was calculated using the formula PIMG = [(C-N)/(C-5)] Ã 100, where C is the colony diameter of the non-treatment, and N is that of the chemical treatment. Each experiment was carried out with three replicates and the experiments were repeated twice.

DON production analysis
Diseased wheat kernels were harvested from inoculated spikelets 20 days after inoculation and used for DON production analysis using a previously described protocol [26,27].The amount of DON in each sample was determined using a high-performance liquid chromatographymass spectrometer/mass spectrometer (HPLC-MS/MS) system (Agilent1100-6410, Agilent Technologies, Palo Alto, Calif.). Mass spectrometric parameters were performed using a previously described detection method [28].
Additionally, the amount of F. graminearum ergosterol in each sample was also determined. Total ergosterol was extracted from infected wheat powder by using a previously published protocol [29]. Ergosterol concentrations were quantified with a high-performance liquid chromatography (HPLC) system, Waters 1525-2489. Ergosterol was separated at room temperature on a hypersil BDSC18 250 nm×4.6 nm, 5μm analytical column using 100% methanol (chromatography pure) as mobile phase. The detection wavelength was 282 nm. The identification of ergosterol was based on retention time and co-chromatography of commercial standard of ergosterol (Sigma-Aldrich, St. Louis, MO, USA).

Standard molecular methods
Fungal genomic DNA was extracted using a previously published protocol [30]. Plasmid DNA was isolated using a plasmid miniprep purification kit (BioDev Co., Beijing, China). Southern blot hybridization analysis of FgLEU2A/B genes in the wild-type and mutants was performed using probes as indicated in S2A and S3A Figs. Probe was labeled with digoxigenin (DIG) using a high prime DNA labeling and detection starter kit II according to the manufacturer's instructions (Roche Diagnostics, Mannheim, Germany).

Data analysis
Statistical comparisons were made between the wild type parent and the mutants by one-way ANOVAs conducted using the software package SPSS (version 13 for Windows, 2004). And for the relative gene expression data, one-sample t-test was conducted to test the difference between different time points or different strains using SPSS.
FgLEU2A is 1313-bp in length and encodes 367 amino acids. FgLEU2A has four predicted introns and are 53-bp in length located between the 52 th and 104 th nucleotide, 55-bp in length located between the 126 th and 180 th nucleotide, 50-bp in length located between the 253 h and 302 th nucleotide, and 54-bp in length located between the 1024 th and 1077 th nucleotide, respectively. FgLEU2B is 1086-bp in length and encodes 362 amino acids. FgLEU2B has no predicted intron. Sequencing of the 1101-bp and 1086-bp cDNA of FgLEU2A and FgLEU2B, respectively, verified the existence of the predicted introns.
Phylogenetic analysis was performed with the neighbor-joining method using MEGA 4.1 software [31] and revealed two separated branches of Leu2 (Group1 and Group 2) in fungi. Yeast species contain a single Leu2 clustered in Group 1 in their genome while in the filamentous fungal species; some contain single Leu2 while others possess two Leu2 orthologs in their genome. As shown in S1 Fig, both the Fusarium and Aspergillus species contains two Leu2 orthologs clustered in both Group 1 and Group 2, however, other fungal species including Magnaporthe oryzea, Ustilago maydis and Sporisorium reilianum also contain two Leu2 orthologs in their genome with both Leu2 clustered together in the Group 1 branch with yeast and Neurospora species. FgLeu2a is more closely to its S. cerevisiae and N. crassa homologues than to FgLeu2b.

Expression pattern of FgLEU2A and FgLEU2B in different cultures in F. graminearum
Real-time PCR assays showed that when cultured in low-nutrient medium MM containing no amino acids, both FgLEU2A and FgLEU2B genes were significantly (p<0.01) up-regulated in vegetative hyphae as compared to 12 h germlings, the relative expression level of FgLEU2A gene raised to 2.90-and 19.42-fold and FgLEU2B gene raised to 3.65-and 4.48-fold at 24 h and 36 h, respectively (Fig 1).

Generations of mutants of FgLEU2A and FgLEU2B in F. graminearum
Single deletion mutants of FgLEU2A were generated by transforming the 3095-bp fragment containing leu2A-upstream-HPH-leu2A-downstream to the protoplasts of the wild-type, resulting in more than 20primary hygromycin-resistant transformants. Complementation of FgLEU2A mutants was conducted by transforming the 4553-bp fragment containing NEO-leu2A-upstream-leu2A-downstream to the protoplasts of the FgLEU2A deletion mutants, resulting in more than 25primary G418 sulfate -resistant transformants. First, PCR analysis with primers A5 + A6 (S1 Table)  Single deletion mutants of FgLEU2B were generated by transforming the 3455-bp fragment containing leu2B-upstream-HPH-leu2B-downstream to the protoplasts of the wild-type, resulting in more than 30 primary hygromycin-resistant transformants. First, PCR analysis with primers B5 + B6 (S1 Table) was used for detection of FgLEU2B deletion mutants. This primer pair amplified a 1438-and 1778-bp fragment from wild type strain PH-1and FgLEU2B deletion mutant, respectively. The mutant ΔFgLeu2B-2 showing the deletion mutant size fragment was chosen for further analysis. Subsequently, reverse transcription PCR with the primer pair leu2B-RT-F + leu2B-RT-R amplified an expected 220-bp fragment from the wild-type progenitor PH-1, but not from the Double deletion mutants of FgLEU2A and FgLEU2B were generated by transforming the 2708-bp fragment containing leu2A-upstream-NEO-leu2A-downstream to the protoplasts of ΔFgLeu2B-2, resulting in more than 15 primary G418 sulfate-resistant transformants. First, PCR analysis with primers A5 + A6 (S1 Table) was used for detection of FgLEU2AB double deletion mutants. This primer pair amplified a 2167-and 1398-bp fragment from wild type strain PH-1and FgLEU2AB deletion mutant, respectively. The double deletion mutant ΔFgLeu2AB-8 showing the deletion mutant size fragment was chosen for further analysis. Subsequently, reverse transcription PCR with the primer pair leu2A-RT-F + leu2A -RT-R amplified an expected 246-bp fragment from the wild-type progenitor PH-1, but not from  Disruption of FgLEU2A leads to defects in mycelial and conidial morphogenesis in F. graminearum Mycelial morphogenesis was affected by disruption of FgLEU2A but not FgLEU2B. Mutants lacking FgLEU2A (ΔFgLeu2A-10 and ΔFgLeu2AB-8) could not survive in MM and FGA medium and showed a slower mycelial growth rate and reduced aerial hyphal formation on PDA plates as compared to wild type strain PH-1 andΔFgLeu2A-8C (Fig 2A-2C), although microscopy examination revealed no visible differences in hyphal structure between both deletion mutants and the wild type strain. Additionally, when cultured in liquid medium, the mycelial dry mass of ΔFgLeu2A-10 and ΔFgLeu2AB-8 was significantly (p<0.05) reduced in PDB and remained undetectable in FGB or MM (Fig 2D).
To further evaluate the effects of FgLEU2A and FgLEU2B on conidial morphogenesis, the wild type PH-1 and mutants were cultured in MBL and the amount of conidia was compared. Results showed that when cultured in MBL conidial formation was completely blocked in FgLEU2A deleted mutants and the defects in conidial formation could be partially rescued by exogenous supplementation of Leu at 0.25 mM and completely restored at 1.25 mM (Fig 3A). Conidial germination assay was conducted using a modified protocol, conidia were collected from MBL culture amended with 1.25 mM Leu, washed with sterilized water and incubated in 2% sucrose solutions with or without Leu. Very few conidia of ΔFgLeu2A-10 and ΔFgLeu2AB-8 could germinate in 2% sucrose solutions and germination rates of ΔFgLeu2A-10 and ΔFgLeu2AB-8 raised to the wild-type level by addition of 1.25 mM Leu (Fig 3B).
Deletion of FgLEU2B caused no visible differences in mycelial and conidial morphogenesis as compared to the wild-type strain PH-1. Complementation of ΔFgLeu2A-10 with full-length FgLEU2A gene rescued all the defects of the deletion mutant. Deletion of FgLEU2A results in reduced red pigmentation and downregulation of genes involved in red pigment formation in F. graminearum Apart from the defects in mycelial and conidial morphogenesis, disruption of FgLEU2A also leads to reduced accumulation of red pigment (Figs 2B and 4A). Further analysis of the expression level of genes involved in red pigmentation revealed that all five genes (PKS12, GIP1, GIP2, AURF, AURJ) were significantly down-regulated (p<0.05) in ΔFgLeu2A-10 and ΔFgLeu2AB-8 as compared to the wild-type strain PH-1 (Fig 4B). We also noticed that the double deletion mutant FgLeu2AB-8 showed a more severe defect in red pigmentation ( Fig 4A) and a lower relative expression level of the five red pigment biosynthetic genes than the single deleted mutant ΔFgLeu2A-10 ( Fig 4B).
FgLeu2A plays a more crucial role in Leu biosynthesis in F. graminearum  acids. The auxotroph of ΔFgLeu2A-10 and ΔFgLeu2AB-8 could be completely or partially recovered when cultured in full-nutrition medium YEPD or PDA plates. The auxotroph caused by the disruption of FgLEU2A could also be overcome by exogenous addition of Leu into the culture medium. When cultured in FGA amended with Leu at 0.25 mM, the growth defects of the single deletion mutant ΔFgLeu2A-10 disappeared, while the concentration to rescue the auxotroph of the double deletion mutant ΔFgLeu2AB-8 should be higher than 0.25 mM (Fig 5).
Auxotroph caused by FgLEU2A disruption is due to a block in the conversion of 3-IPPM to 2-KIC and Leu depletion in F. graminearum   FgLEU2A is involved in the adaption to cellular stresses in F. graminearum Results of the sensitivity assay showed that disruption ofFgLEU2A caused reduced tolerance to osmotic and oxidative stresses, while deletion of FgLEU2B brought no significant changes to osmotic and oxidative stresses tested. As shown in Fig 7, ΔFgLeu2A-10 and ΔFgLeu2AB-8 showed significantly (p<0.05) increased sensitivity to all the four osmotic stresses mediated by NaCl, KCl, glycerol or D-srobitol at 1.2M as well as oxidative stresses mediated by H 2 O 2 at 12 mM. The sensitivity of all the mutants to cell wall and cell membrane inhibitors did not change significantly as comparing to the wild-type strain PH-1.

FgLEU2A plays a more important role in plant infection and DON accumulation in F. graminearum
Virulence of F. graminearum strains was evaluated by point inoculating conidial suspension on flowering wheat heads and non-host cherry tomatoes. Results revealed the involvement of FgLEU2A but not FgLEU2B in full virulence. As shown in Fig 8A, ten days after inoculation, the wild type progenitor PH-1, ΔFgLeu2B-2, and ΔFgLeu2A-8C caused the typical scab symptoms in the inoculated as well as the nearby spikelets of flowering wheat heads. Under the same conditions, however, both FgLEU2A-deficent mutants lost aggressiveness in flowering wheat heads and the scab symptoms caused by ΔFgLeu2A-10 and ΔFgLeu2AB-8 were restricted in the inoculated spikelet. And for the non-host cherry tomatoes, the size of water- Two FgLEU2 Genes with Different Roles in Fusarium graminearum soaked rot lesion caused by ΔFgLeu2A-10 and ΔFgLeu2AB-8was significantly smaller as compared with the wild type progenitor PH-1, ΔFgLeu2B-2, and ΔFgLeu2A-8C (Fig 8B).
The amount of DON accumulation was also compared in the wild-type and the mutants, and results showed that when cultured on sterilized wheat kernels for 20 days, the amounts of DON produced by the wild-type PH-1 or the complemented strain were 4 folds higher than that produced byΔFgLeu2A-10 and 16 folds higher than that produced by ΔFgLeu2AB-8 (significantly reduced, p<0.05) (Fig 9).

Expression levels of FgLEU2A/B genes in the gene deletion mutants
Real-time PCR assays showed that the expression level of FgLEU2A in ΔFgLeu2B-2 was significantly (p< 0.05) higher than that in the wild-type parent PH-1 especially in the low-nutrition medium MM (fold change higher than 6.0) (Fig 10). Similar results were observed from other two FgLEU2B deletion mutants (data not shown). The expression levels of FgLEU2B in ΔFgLeu2A-10 did not change too much as compared with those in the wild-type parent PH-1 (Fig 10).

Discussion
Yeast species including Sacharomyces cerevisiae, Candia albicans possess single LEU2 gene encoding IPMD, the penultimate enzyme in the Leu biosynthesis. In this study, however, LEU2 gene duplication was found in the genome of several filamentous fungi including Two FgLEU2 Genes with Different Roles in Fusarium graminearum Aspergillus spp., Fusarium spp., Magnaporthe oryzae, Ustilago maydis and Sporisorium reilianum (S1 Fig). In F. graminaerum, paralogous genes encoding one enzyme have been previously reported. We have characterized functions of three cyp51 genes encoding 14αdemethylases and two ERG24 genes encoding sterol C-14 reductase of the ergosterol biosynthetic pathway, both enzymes are the targets of sterol biosynthesis inhibitors (SBIs). Targeted disruption of individual gene revealed different roles of these genes in mediating sensitivity to SBIs, however, deletion of single gene showed no effects on vegetative development or virulence [23,32]. In A. fumigates, Oliver et al., reported four putative ILV3 (ILV3A-D) genes encoding DHAD with only Ilv3A required in BCAAs biosynthesis and full virulence [33].  Recently, Liu et al. characterized two Cdc2 kinases orthologs with distinct functions in vegetative and infectious hyphae in F. graminaerum [34]. When comparing the genome sequences of S. cerevisiae and F. graminaerum, we found 154 F. graminearum multiple genes (gene duplication numbers from 2 to 6) corresponding 63 S. cerevisiae single genes. Further investigation into the functions of these duplicated genes would underlie the evolutionary mechanism of the fungus.
Previous studies showed that A.niger elaborates two differentially expressed LEU2 genes [8]. In F. graminearum, FgLEU2A and FgLEU2B were also differentially expressed in different cultures but the expression pattern of these two genes shares variations with A. niger LEU2 genes. Expression of both FgLEU2 genes could be detected during germination and hyphal formation phases and was greatly stimulated when cultured in low nutrition medium with FgLEU2A upregulated in higher levels comparing with FgLEU2B. Different reactions of FgLEU2A and FgLEU2B to nutrition starvation at transcriptional level indicated a more crucial role of FgLEU2A in Leu biosynthesis in F. graminearum.
Functions of two LEU2orthologs have not been investigated in filamentous fungi. In this study, targeted disruption of individual FgLEU2 gene in F. graminearum assigned different roles of FgLeu2A and FgLeu2B for Leu biosynthesis. The FgLEU2A deleted mutants, but not the FgLEU2B deletion mutant displayed Leu auxotroph. Exogenous supplementation of Leu or 2-KIC rescued the auxotroph of both ΔFgLeu2A-10 and ΔFgLeu2AB-8, indicating FgLeu2A, but not FgLeu2B, is dependent for Leu biosynthesis. Additionally, judged from the minimal concentration of exogenously-added Leu for satisfaction of different mutants, deletion mutants of FgLEU2B lacking Leu auxotroph does not make FgLeu2B absolutely useless in Leu biosynthesis. ΔFgLeu2AB-8 required a higher concentration Leu or 2-KIC from culture medium to overcome its auxotroph than ΔFgLeu2A-10, indicating a minor role of FgLeu2B in Leu biosynthesis in F. graminearum.
Apart from Leu biosynthesis, FgLEU2A and FgLEU2B also showed distinct roles in other cellular processes. Deletion mutants of FgLEU2 showed different growth rates and aerial hyphal formation on different medium, indicating diverse roles of FgLEU2A and FgLEU2B in the utilizations of different nutrients. Significant down-regulation of red pigmentation genes in the FgLEU2A deleted mutants provided the molecular explanations for the reduced red pigment accumulation in these deletion mutants and assigned a more important role for FgLEU2A than FgLEU2B in secondary metabolism. FgLEU2A disruption mutants showed increased sensitivity to osmotic and oxidative stresses comparing with the wild type strain PH-1. The increased sensitivity to osmotic and oxidative stresses have also been observed in our previously characterized ILV mutants including FgILV5, FgILV1, FgILV2 and FgILV6 [19][20][21], and as the cellular stresses assay were conducted using YEPD medium in which all the mutants could acquire enough Leu from culture medium and show no obvious grown defects, which indicating the involvement of FgLEU2 genes in the adaption to osmotic and oxidative stresses and this involvement is Leu-independent. Additionally, the inhibitory rate of ΔFgLeu2AB-8 was higher than that of ΔFgLeu2A-10, which is also in agreement with the different role weight of FgLEU2A and FgLEU2B in other cellular processes in F. graminearum.
Plant infection assays revealed reduced aggressiveness of FgLEU2A deleted mutants, which cannot be separated from the severe phenotypic defects exhibited by these mutants in vitro. The limited acquisition of Leu or 2-KIC from plant tissues the mutants encountered during colonization and mycelial spread within spikelets resulted in a series of infection-related defects including lower percentages of conidial germination, slower mycelial growth rate, reduced aerial hyphal formation, which may synthetically prevent the aggressiveness of the mutants within wheat heads. Second, it is well recognized that reactive oxygen species play crucial roles in host-pathogen interaction. When encountering pathogen attack, plants use the oxidative burst as an early defense reaction. FgLEU2Adeficient mutants showed reduced tolerance to the oxidative stress mediated by H 2 O 2 , which may be related to the failed colonization or reduced aggressiveness in plant tissues. Additionally, DON, the end product of the trichothecene biosynthetic pathway, has been identified as an important virulence factor and plays a crucial role in the spread of the fungus through the infected wheat head, the decreased DON production caused by the disruption of FgLEU2A may also account for the reduced aggressiveness in ΔFgLeu2A-10 and ΔFgLeu2AB-8 [35][36][37].
Previous studies in S. cerevisiae, C. albicans, Cryptococcus neoformans and Mycobacterium tuberculosis indicated that certain genes encoding BCAA biosynthetic enzymes(e.g., ILV2) were important for fungal survive and/or virulence [13,[15][16][17]. Recent reports in the rice blast fungus M. oryzae have revealed crucial roles of several BCAA biosynthetic genes in fungal development and full virulence in filamentous fungi [14,18].We recently identified several Val and Ile biosynthetic genes with distinct but crucial roles in mycelial and conidial morphogenesis, plant infection, and DON biosynthesis in the important wheat head scab fungus F.graminearum [19][20][21]. In this present study, a series of phenotypic defects displayed by the deletion mutants of FgLEU2A indicated the involvement of this Leu biosynthetic enzyme encoding gene in fungal development and infection in F. graminearum. The important functional roles of these BCAA biosynthetic genes, together with the fact that BCAA biosynthesis only exist in high plants, fungi and bacteria but not mammals, are making enzymes involved in BCAAs biosynthesis attractive and potential targets for the development of herbicides, fungicides and other antimicrobial compounds.  S1 Table for the primer sequences). (B) Reverse transcription PCR analysis of FgLEU2B expression in PH-1 and ΔFgLeu2B-2 using cDNA as template. NCK is a negative control without template cDNA in the PCR amplification. (C) Southern blot hybridization analysis of PH-1 and ΔFgLeu2B-2 using a 954-bp FgLEU2B upstream fragment as a probe. Genomic DNA preparation of each strain was digested with Nco I. (TIF) S1 Table. Oligonucleotide primers used in this study and their relevant characteristics. (DOCX)