Involvement of a velvet protein ClVelB in the regulation of vegetative differentiation, oxidative stress response, secondary metabolism, and virulence in Curvularia lunata

The ortholog of Aspergillus nidulans VelB, which is known as ClVelB, was studied to gain a broader insight into the functions of a velvet protein in Curvularia lunata. With the expected common and specific functions of ClVelB, the deletion of clvelB results in similar though not identical phenotypes. The pathogenicity assays revealed that ΔClVelB was impaired in colonizing the host tissue, which corresponds to the finding that ClVelB controls the production of conidia and the methyl 5-(hydroxymethyl) furan-2-carboxylate toxin in C. lunata. However, the deletion of clvelB led to the increase in aerial hyphae and melanin formation. In addition, ΔClVelB showed a decreased sensitivity to iprodione and fludioxonil fungicides and a decreased resistance to cell wall-damaging agents and osmotic stress and tolerance to H2O2. The ultrastructural analysis indicated that the cell wall of ΔClVelB became thinner, which agrees with the finding that the accumulated level of glycerol in ΔClVelB is lower than the wild-type. Furthermore, the interaction of ClVelB with ClVeA and ClVosA was identified in the present research through the yeast two-hybrid and bimolecular fluorescence complementation assays. Results indicate that ClVelB plays a vital role in the regulation of various cellular processes in C. lunata.

Scientific RePORTS | 7:46054 | DOI: 10.1038/srep46054 brn1 may have some connection or a coordinated mediation mechanism in secondary metabolism. However, we do not know how both are connected 3 .
In addition to the virulence genes described above, various regulatory systems evidently control the regulation of the secondary metabolite biosynthesis in most fungi in response to the external environment 5 . The velvet family protein plays a key role in regulating secondary metabolism and the differentiation processes, such as fungal development and sporulation. It shares a common domain that is present in most parts of filamentous fungi. Different velvet protein members interact with each other in the nucleus 6 . As a key member of the velvet protein family, VelB has been researched in a few fungal species. In Aspergillus nidulans, the removal of velB leads to reduced secondary metabolites and sexual fruit body formation 6 . In Fusarium fujikuroi, FfVel2 (VelB ortholog) have similar functions in regulating fungal development and secondary metabolism 7 . Similar phenotypes of conidiation, melanin biosynthesis, hypersensitivity to oxidative stress, and virulence have been reported for the BcVelB (VelB ortholog) mutant 8 . This study aims to elucidate the functions of the VelB-ortholog ClVelB in C. lunata. In the current study, the deletion of clvelB showed a few distinct phenotypic characteristics compared with the VelB mutants in a few other fungi.

Results
Identification of the VelB ortholog in C. lunata. The ClVelB (accession number: KY435512) sequence was extracted from C. lunata genomic database (Dryad Digital Repository) using BlastP analyses with the sequence of A. nidulans VelB. The open reading frame of clvelB comprises 1,011 bp, does not contain introns, and encodes a 336-amino-acid protein. ClVelB falls in a group of dothideomycete VelB homologs, which is a sister to the eurotiomycete group including A. nidulans VelB, and the sordariomycete group including F. fujikuroias FfVel2 (Fig. 1A). The alignment of ClVelB with A. nidulans VelB (Fig. 1B) showed 90% positives and 51% identity (National Center for Biotechnology Information, BlastPAlign). Involvement of ClVelB in the regulation of hyphal growth, asexual development, and pigment formation in C. lunata. Target gene deletion strategy was employed by replacing clvelB with a hygromycin resistance (hph) cassette to investigate the biological functions of ClVelB in C. lunata (Fig. 2). The Southern hybridization pattern confirmed that homologous recombination occurs at the clvelB locus in Δ ClVelB. Complementation of the deletion mutant (ClVelB-C) was accomplished by the reintroduction of wild-type (WT) clvelB into the genome of Δ ClVelB. The radial growth rates of the mutants and WT on the complete medium VelB protein sequences were obtained from GenBank using A. nidulans. AnVelB as a query. AnVelB, C. lunata ClVelB, and Fusarium fujikuroi FfVelB are marked in yellow highlights. A blue oval shadow marks the single candidate ortholog. (B) ClVelB, AnVelB, and FfVelB were aligned using ClustalW. Conserved velvet superfamily domains are highlighted in red, asterisks mark identical residues, colons mark conserved residues, and periods indicate semi-conserved residues. ) were compared. Δ ClVelB had a significantly slower mycelial growth rate than WT and the complemented strain ClVelB-C on the CM medium ( Table 1). As the primary source of inoculum for host infections, conidia are formed during exposure to light. Time course experiments were performed to follow the onset of conidiation in the generated mutant under different illumination conditions. WT and ClVelB-C exhibited an obvious banding rhythm which reflected periods of conidiation under LD conditions, whereas that in Δ ClVelB was greatly reduced (Fig. 3A). The conidiation of WT was the most in the LL condition, the least in the DD condition, and a moderate number in the LD condition (Fig. 3C). However, the conidiation of Δ ClVelB sharply declined, and the differences in conidiation in the preceding three conditions were not as obvious as in WT and ClVelB-C (Fig. 3C).
WT and ClVelB-C sparingly developed aerial hyphae accompanied by high numbers of conidia in the LL condition. Although the hyphae of Δ ClVelB were not evidently different from those of the WT (Fig. 3D), Δ ClVelB exhibited "fluffy" colonies that are characterized by a cotton-like appearance (Fig. 3B) and produced fewer conidia (Fig. 3C). Overall, these results indicate that ClVelB controls the balance between aerial conidiation and hyphal growth, that is, it represses aerial hyphae growth and promotes conidiation in response to the light condition.
ClVelB regulates the melanization of mycelia. Similar to Botrytis cinerea BcVelB, the deletion of ClVelB leads to an increase in mycelial pigmentation 8 , and insufficient ClVelB leads to the increased melanization of mycelia, which is grown both on a solid (Fig. 4A) and in a liquid CM medium (Fig. 4B), indicating that ClVelB negatively regulates the mycelial pigmentation in C. lunata. Hyphal pigmentation develops faster in Δ ClVelB than in WT (Fig. 4B). By 68 h, all strains were darkly pigmented. We detected the expression of the PKS gene (pks18), the transcription factor gene (cmr1), and three synthase genes (brn1, brn2, and scd) related to the synthesis of DHN melanin in the WT and mutant to further confirm this observation ( Fig. 4C) 4,9 . qRT-PCR analyses showed that the expression levels of pks18 in Δ ClVelB were enhanced compared to those in WT. By 48 h, the expression of pks18 in Δ ClVelB has a 12.53-fold increase, which peaked at 60 h (57.82-fold). At 48 h, the expression of cmr1 has a 5.25-fold increase in Δ ClVelB compared to that in WT. brn1, brn2, and scd also showed high expression levels in Δ ClVelB compared to those in WT at both 48 and 60 h. For all the genes, the reintroduction of clvelB restored the WT expression levels. Overall, we conclude that ClVelB plays a negative regulation role in the synthesis of melanin. The pyroquilon and kojic acid inhibitors were used to study the influence on melanization and support our previous studies that the conidial and mycelial melanin of C. lunata is not the tyrosine-derived but DHN type 10 . While the colors of all the cultures (WT, Δ ClVelB, and ClVelB-C) grown on kojic acid remained the same, those grown on pyroquilon were changing from black to light brown (Fig. 5), bolstering previous research on melanization in C. lunata.
ClVelB is required to cope with oxidative stress. The growth rates of the mutants were quantified on media supplemented with stressors that induce osmotic stress (1.2 M NaCl, 1.2 M KCl), fungicides (10 μ g/ mL iprodione, 0.1 μ g/mL fludioxonil), and oxidative stress (2.0 or 4.0 mM H 2 O 2 ) to assess whether ClVelB is also essential to cope with various kinds of stresses. Under osmotic stress conditions, all mutants showed comparable growth rates. Δ ClVelB showed a slightly decreased resistance to osmotic stresses cultured in 1.2 M NaCl or 1.2 M KCl medium and a slight decreased sensitivity to the dicarboximide fungicide iprodione and phenylpyrrole fungicide fludioxonil (Fig. 6). The intercellular glycerol of fungus plays a significant role in responding to osmotic stress 11 . As shown in Fig. 7, Δ ClVelB exhibited a low basal level of glycerol accumulation and the expression of the gpd1 gene that is responsible for glycerol synthesis showed a similar trend, which partially explains why Δ ClVelB exhibited decreased resistance to osmotic stresses. Δ ClVelB showed high sensitivity to H 2 O 2 compared to the WT strain, reintroduction of WT clvelB gene into the mutant restored the tolerance of WT to oxidative stress (Fig. 8A).  The expression of the catalase gene cat3 that related to oxidative stress responses, exhibited an obvious difference between the WT and the clvelB mutant (Fig. 8B). In Δ ClVelB, cat3 decreased by approximately 3.7-fold before adding H 2 O 2 and approximately 3.3-fold after the addition of H 2 O 2 compared to the level in the WT strain. Collectively, the data indicate that ClVelB regulates oxidative stress responses by controlling the expression of cat3 gene. The mechanism should be researched further.
ClVelB regulates cell wall integrity. The deletion of clvelB led to a decrease in resistance to osmotic stresses, which indicates that ClVelB might regulate the integrity of the cell wall and/or the cell member. To prove this hypothesis, we tested the sensitivity of Δ ClVelB to cell wall damaging agents, namely, Congo red and Caffeine and to cell member damaging agent SDS. The results indicate that Δ ClVelB displayed a decreased resistance to these compounds to some extent (Fig. 9A). Studies have shown that Congo red could disturb the fungal cell wall by binding to cellulose and chitin 12 . Thus, we tested the expressions of the 1,3-beta-glucan synthase gene gls2 and MAPK gene slt2, which are homologous to the core element genes of Saccharomyces cerevisiae cell wall integrity (CWI) pathway, in the clvelb deletion mutant. The expression levels of gls2 and slt2 in Δ ClVelB were lower than those in WT (Fig. 9B), which agrees with the phenotype that Δ ClVelB showed decreased resistance to Congo red. More interestingly, we found that the deletion of clvelB led to the decrease of fungal cell width compared with WT (Fig. 9C). These results demonstrate that ClVelB might be related to the regulation of the CWI pathway in C. lunata. Effects of ClVelB on the hyphal hydrophobicity. In numerous fungal species, the cell surface of aerial hyphae shows a distinct hydrophobic feature 13 . The deletion of fgvelB leads to loss of function to maintain the hydrophobicity of the hyphal surface in Fusarium graminearum 14 . To confirm if clvelB has the same function in C. lunata, 20 μ l drops of 2.5% bromophenol blue solution or ddH 2 O were added to each strain surface. Both the 2.5% bromophenol blue solution and the ddH 2 O maintained spherical droplets on the surface of the Δ ClVelB colony without being absorbed or extended for more than 30 min, thereby demonstrating the strong hydrophobicity of the Δ ClVelB hyphae, which is similar to those of WT, and the complemented strain (Fig. 10). These results indicate that ClVelB did not contribute in regulating the hyphal hydrophobicity of C. lunata.
ClVelB regulates M5HF2C toxin biosynthesis. Reports indicate that VelB regulates the synthesis of secondary metabolites in numerous fungi 15 . Therefore, detecting the M5HF2C toxin production in Δ ClVelB is necessary. After culturing in Fries 3 medium for 30 days, the amount of M5HF2C produced by Δ ClVelB was 79.3% lower than that produced by WT (Fig. 11A). The expression of the M5HF2C biosynthesis related gene clt-1 was analyzed by qRT-PCR to further confirm that ClVelB acts as a positive regulator of M5HF2C toxin production. The expression level of clt-1 in Δ ClVelB decreased by 31.9% compared to that in WT, which was consistent with the profiles of M5HF2C production (Fig. 11B). The experiment results indicate that ClVelB played a major role in the regulation of M5HF2C biosynthesis in C. lunata. ClVelB is essential for virulence in C. lunata. Mycotoxin M5HF2C has been described as one of most important virulence factors in C. lunata 2 . We further assayed the infective ability of Δ ClVelB on maize leaves because the deletion of clvelB compromised the ability of C. lunata to produce M5HF2C. The penetration and establishment of primary lesions by Δ ClVelB were similar to those by WT. However, the infection proceeded differentially. The capability of Δ ClVelB to colonize the surrounding host tissue was impaired (Fig. 12). In any case, the lesion sizes on maize leaves inoculated with Δ ClVelB decreased significantly compared to those inoculated with WT, indicating that ClVelB was essential to the complete virulence in C. lunata.
Interaction of ClVelB with ClVeA and ClVosA in C. lunata. In A. nidulans, the positive control of secondary metabolism is accomplished through the physical interaction of VelB with another velvet-like protein VeA in the nucleus 6 , and VelB-VosA heterodimer has additional functions in trehalose biogenesis and spore viability 16 . A direct yeast two-hybrid (Y2H) method was used to ascertain the analogous protein-protein interactions of the C. lunata orthologs (ClVelB [336 aa], ClVeA [598 aa, accession number: KY435511], and ClVosA [302 aa, accession number: KY435513]). The full-length ClVelB protein was fused to the GAL4 activation domain, and the full-length proteins of ClVeA and ClVosA were respectively fused to the GAL4 binding domain. Then, yeast cells expressing different combinations were tested for ADE2 and HIS3 reporter gene activities. This experiment showed that ClVelB interacts with ClVeA and ClVosA (Fig. 13A). Bimolecular fluorescence complementation (BiFC) experiments with splitYFP-constructs were conducted to control the false positive fluorescence signal due to simple and close co-localization more stringently and further confirm the dimerization of ClVelB with ClVeA    and ClVosA. The BiFC analysis suggests that ClVelB can interact with ClVeA and ClVosA as homodimers in the cellular nuclei of tobacco (Fig. 13B).

Discussion
VelB has been reported to be a filamentous fungi-specific regulator that plays multifaceted roles in various biological processes, including fungal development, colonial morphology, and secondary metabolism. However, certain changes in the preceding roles have been found in different fungi. For example, the deletion of Ffvel2 led to decreased conidiation and hyphal growth in F. fujikuroi 7 . In this research, the clvelB deletion mutant also exhibited reduced growth rate (Table 1) and conidiation (Fig. 3C) but increased aerial hyphae formation (Fig. 3B). In F. graminearum, the disruption of velB caused the hydrophobicity change of the cell surface 14 . Instead, we found that the clvelB deletion mutant exhibited no effect on hydrophobicity (Fig. 10). A recent study of A. nidulans indicated that the conidia of the velB mutant showed decreased resistance to numerous H 2 O 2 and UV stresses,  which resulted in a low-level accumulation of trehalose in the mutant 16 . In the current study, we also found that the deletion of clvelB led to the slightly decreased resistance to a few stress agents, including NaCl and KCl (Fig. 6), which may be attributed to a lower basal accumulation of glycerol in Δ ClVelB compared with that in WT (Fig. 7). The reduced tolerance to stress agents in Δ ClVelB indicated a variation in the cell membrane or cell wall composition. Therefore, we tested the sensitivity of Δ ClVelB to the cell member damaging agent SDS and the cell wall damaging agents Caffeine and Congo red. In line with the expectations, Δ ClVelB showed a decreased resistance to these compounds (Fig. 9A), which is in agreement with the expressions of the 1,3-beta-glucan synthase gene gls2 and MAPK gene slt2 in the clvelB deletion mutant (Fig. 9B). Moreover, the cell wall of Δ ClVelB became thinner (Fig. 9C). These results demonstrate that ClVelB may regulate the cell wall composition and integrity in C. lunata, indicating that VelB can bind to the promoter region of the β -glucan synthase gene fksA to regulate the cell wall synthesis in A. nidulans 17 .
When tested for pathogenicity, Δ ClVelB produced smaller lesions than WT or the complemented strain (Fig. 12). With regard to virulence and basic metabolism of fungal cells, we also found that the disruption of clvelB affects the redox status, the clvelB deletion mutant is more sensitive to H 2 O 2, and the growth defects became more evident (Fig. 8A). Managing ROS is a determinant of fungal success in infecting host and in the basic cellular of fungal cells. In accordance with the more pronounced effect of H 2 O 2 on the radial growth rate of the clvelB mutant compared to that of the WT, a significant reduction in the expression level of cat3 was observed (Fig. 8B). Reactive oxygen species (ROS) plays a major role in pathogen-host interactions 18 . Under a pathogen attack, plants use the oxidative burst as an initial defense reaction. The fungus shows resistance against oxidative burst while infecting the host plant because C. lunata has effective ROS-detoxification systems, such as peroxidases and catalases 19 . Thus, the increased sensitivity of the clvelB mutant to oxidative stress might be partially related to the reduced virulence of the mutant on the host plant.
VelB proteins have been reported to regulate secondary metabolism in some fungi. In A. nidulans, the velB deletion mutant showed decreased sterigmatocystin (SM) production and synthesized a brownish pigment 6 . In F. graminearum, the FgVelB mutant produced a yellow pigment and a dramatically low level of DON 14 . In the current study, we observed that Δ ClVelB produced a significantly high level of melanin (Fig. 4A). Furthermore, the expressions of five DHN melanin biosynthesis genes were significantly up-regulated in Δ ClVelB (Fig. 4C). These  results indicate that VelB repressed melanin expression as previously described in Cochliobolus heterostrophus 20 . In contrast, the clvelB mutant produced a lower expression of M5HF2C toxin (Fig. 11A), which has been identified as one of the most important virulence factors of C. lunata 21 . ClVelB is essential for virulence to facilitate the colonization of the plant tissue. Notably, the penetration via germ tubes or infection cushions remains unaffected in the deletion mutant. Thus, no difference between the primary infections of Δ ClVelB and WT exists. However, the lesions of Δ ClVelB did not spread, suggesting that the mutant cannot kill the ambient host cells of the infection site. Predicting the reason for this result is difficult, and several factors are probably responsible. Hence, mycotoxin production and conidiation may contribute to virulence.
The velvet proteins of VeA, VelB, and VosA are fungi-specific transcription factors, which contain the velvet domain 22 . In numerous filamentous fungi, these proteins form different complexes that play distinct roles. Among them, VelB forms a heterodimer with VeA, which is required for secondary metabolites production and fungal deveploment 6 . In A. nidulans, the disruption of either velB or veA results in defects in the SM production and sexual fruiting body formation 22 . In the same way, FgVelB and FgVeA have similar roles in regulating fungal development, glycerol accumulation, DON synthesis, and pathogenicity 14 , which indicates that VelB cooperates with VeA to regulate fungal development and secondary metabolism. In A. nidulans, VelB contains neither a typical nuclear export signal (NES) nor a nuclear localization signal (NLS). Instead, the A. nidulans VeA protein includes a NES and a bipartite NLS in the N-terminal part. VeA is necessary for the efficient nuclear import of VelB. Earlier studies on A. nidulans have shown that the positive control of secondary metabolism can be achieved via the physical interaction of VelB with VeA in the nucleus 6 . VelB has additional functions in trehalose biogenesis and spore viability, which requires the VelB-VosA heterodimeric protein complex formation 23 . In A. nidulans, the VelB-VosA complex represses β -glucan synthesis by directly binding to the promoter regions of the cell wall biosynthetic genes in conidia and ascospores, thereby activating the formation of spore wall during sporogenesis 17 . The Y2H and BifC approaches confirmed that the C. lunata orthologs ClVelB interacted with ClVeA and ClVosA, and likely formed the complexes of ClVelB-ClVeA and ClVelB-ClVosA analogous to the situation found in other filamentous fungi 23 . Given that these complexes regulate numerous processes in fungal biology, we suspected that ClVelB may regulate the biosynthesis of M5HF2C toxin and DHN melanin in combination with ClVeA and interact with ClVosA to control the sporulation in C. lunata. Studying the roles of the ClVelB-ClVeA and ClVelB-ClVosA complexes in the different functions in C. lunata is interesting because ClVelB interacts with ClVeA and ClVosA in the Y2H and BiFC tests (Fig. 13). In conclusion, this study would help us understand the biological roles of C. lunata and may provide target sites for designing a new agent to control C. lunata and a few similar fungi.

Methods
Fungal strains, plant materials, and culture conditions. C. lunata WT strain CX-3, whose genome sequence is available (Dryad Digital Repository) 24 , was used as a progenitor for the transformation experiment in this study. The clvelB gene deletion strain was generated in the CX-3 genomic background. ClVelB-C was strain complemented with the WT clvelB gene. Unless mentioned otherwise, all strains were cultivated in Petri dishes containing solid synthetic CM medium. Cultures were incubated at 28  and F. fujikuroi FfVel2 (accession number: FN675836) were used to query the C. lunata genome database for orthologs. Fungal genomic DNA and total RNA were prepared as previously described to verify the existence and sizes of introns in clvelB 25 . DNA and cDNA amplification were performed using the primer pair VelB-FL-F and VelB-FL-R, respectively (Table S1). Phylogenetic tree was built using the MEGA 5.0 and alignment created using ClustalW.
clvelB gene deletion and complementation. We inserted two flanking sequences of clvelB into the two sides of the hph gene in pC1300 kh vector to construct the deletion vector 1300 kh-ClVelB-D (Fig. 2) 26 . clvelB was deleted using the ATMT method 3 . Hygromycin was added to the medium to a final concentration of 200 μ g/ml for selecting transformants, and putative clvelB deletion mutants were verified by the PCR and Southern hybridization tests.
The pC1300N vector contains the G418 resistance cassette comprising the G418 resistance gene under the control of its promoter and the TrpC terminator from A. nidulans was used for gene complementation. The full-length sequence of clvelB under the control of the promoter and the TrpC terminator were inserted into the HindIII-XbaI sites of pC1300N to create plasmid 1300N-ClVelB-C and to construct ClVelB complementary mutants 26 . The final plasmid carrying both the WT clvelB and the G418 resistance cassette, as well as the TrpC terminator, was used to transform the clvelB deletion mutant and subsequently create a clvelB-restoring strain using the ATMT method as described above except for the use of the G418 selection agent. The integration at the target sites and the complementation of the clvelB mutant were confirmed through the PCR and Southern hybridization analysis. The sequences of primers for gene disruption, complementation and PCR confirmation are shown in Table S1.
Analysis of mycelial development and conidiation. Mycelial development was observed under different conditions on a CM plate added with the corresponding agents that were suggested in the figure legends. Mycelial development was tested according to the description of the procedure 14 . The conidia that formed on the CM were harvested from the cultures of each strain with 5 ml of sterile ddH 2 O and were immediately counted with a hemocytometer. Each experiment was independently treated with three replications.
Microscopic observation of conidial and hyphal morphology. The conidial and hyphal morphologies of each strain were examined using the electron microscope Tecnai G2 Spirit Biotwin (FEI, USA) and Hitachi Sirion 200 scanning electron microscope (FEI, USA), respectively. The samples were prepared according to the description of the methods 14 .
Detection of intracellular glycerol content. Each strain was cultured in a liquid CM medium at 180 rpm for 72 h at 28 °C. After dealing with 1.2 M NaCl for 2 h, mycelia were collected and ground in liquid nitrogen. Mycelial powders (100 mg) were harvested to test the glycerol content using the glycerol assay kit (Chaoyan, Shanghai, China) according to the instructions of the manufacturer. Each experiment was independently replicated three times.
Oxidative stress sensitivity tests. Tests of sensitivity to H 2 O 2 and gene expression analyses were conducted as described 20,27 . Pigmentation. The pigmentation of hyphae on a solid medium and melanin types were tested according to the description of the procedure 20,28,29 . WT strain (CX-3), clvelB deletion mutant (Δ ClVelB), and its complemented strain (ClVelB-C) were cultured in a liquid CM and transferred to 1.5 ml Eppendorf tubes at 48, 60, and 68 h to test the pigmentation of hyphae. The samples at 48 and 60 h were used for qRT-PCR analyses of cmr1, pks18, brn1, brn2, and scd. The genes were expressed as fold change compared with that of WT at 48 h. Analysis of M5HF2C toxin production and expression level of clt-1. The mutants were cultured in Fries 3 medium for 30 days to determine whether they retained the ability to produce the virulence-related toxin M5HF2C. HPLC-MS analysis of the extract from C. lunata cultures was performed on an Agilent 1100 high-pressure liquid chromatography station to determine the amount of M5HF2C using a previously described protocol 2 . The mycelia of WT, Δ ClVelB, and ClVelB-C were inoculated into the liquid CM medium and cultured at 180 rpm for 72 h at 28 °C to determine the expression levels of clt-1. The total RNA was extracted and the expressive level of clt-1 was determined using qRT-PCR assays 30 . Each experiment was independently replicated three times.
Virulence assays. For infection assays, the fourth leaves of the susceptible maize HUANGZAO-4 seedlings at the seven-leaf stage were inoculated with 10 μ l droplets of conidial suspensions (1.0 × 10 6 conidia/ml). These inoculated leaves were incubated on two layers of Whatman 3MM filter papers moisturized with 10 mM of 6-benzyladenine  in Petri dishes at 28 °C for 96 h. This test was independently replicated three times. Y2H assay. The full-length cDNA sequences of clvelB, clveA, and clvosA were amplified to verify the probable interaction of ClVelB with ClVeA and ClVosA using Y2H assay. The clvelb cDNA was inserted into the EcoRI-BamHI sites of the pGADT7 vector containing the yeast GAL4 activation domain, and the cDNAs of clveA and clvosA were respectively inserted into the EcoRI-BamHI sites of the pGBKT7 vector containing the GAL4 binding domain (Clontech, Mountain View, CA, USA). The plasmid pairs of pGADT7-ClVelB/pGBKT7-ClVeA and pGADT7-ClVelB/pGBKT7-ClVosA were co-transformed into the S. cerevisiae reporter strain AH109 using the LiAc/SS-DNA/PEG transformation method 31 . The plasmid pairs of pGADT7-SV40/pGBKT7-53 and pGADT7-SV40/pGBKT7-Lam served as the positive and negative controls, respectively. The experiment was independently replicated three times.