The Basic Leucine Zipper Transcription Factor PlBZP32 Associated with the Oxidative Stress Response Is Critical for Pathogenicity of the Lychee Downy Blight Oomycete Peronophythora litchii

In this study, we utilized the RNAi technique to investigate the functions of PlBZP32, which possesses a basic leucine zipper (bZIP)-PAS structure, and provided insights into the contributions of bZIP transcription factors to oxidative stress, the production of sporangia, the germination of cysts, and the pathogenicity of Peronophythora litchii. This study also revealed the role of PlBZP32 in regulating the enzymatic activities of extracellular peroxidases and laccases in the plant-pathogenic oomycete.

The transcription factors (TFs) of the basic leucine zipper (bZIP) family are multifunctional across pathogenic fungi. For example, MoAtf1 regulates the transcription of genes encoding laccases and peroxidases, the oxygen scavengers, and thus is required for full virulence in the rice blast fungus Magnaporthe oryzae (13). MoAP1 also mediates the oxidative stress response and is critical for growth, conidium formation, and pathogenicity in M. oryzae (14). In another plant-pathogenic fungus, Fusarium oxysporum, HapX mediates iron homeostasis; therefore, it is essential for rhizosphere competence and virulence of this pathogen (15). In the human pathogen Cryptococcus neoformans, Gsb1 is required for oxidative stress response, mating, and virulence (16). The Yap1-involved H 2 O 2 detoxification is also associated with virulence in Ustilago maydis (17), Candida albicans (18), Alternaria alternata (19), Verticillium dahliae (20), Aspergillus parasiticus (21), and Colletotrichum gloeosporioides (22). However, in Cochliobolus heterostrophus and Aspergillus fumigatus, CHAP1 and AfYap1 are associated with the sensitivity against H 2 O 2 and menadione but are not essential for virulence (23,24).
Oomycetes, which include many notorious plant pathogens, are fungus-like organisms; however, they are evolutionarily related to brown algae and belong to the kingdom Stramenopila (25). Systematic analysis identified conventional bZIPs and novel bZIP transcription factors in Phytophthora infestans (26), Phytophthora ramorum, and Phytophthora sojae (27,28). Several bZIPs from P. sojae are upregulated by H 2 O 2 treatment (28). Using stable gene silencing analyses, several P. infestans bZIPs were found to play roles in protecting the pathogen from hydrogen peroxide-induced injury (26). The novel P. infestans bZIP PITG_11668 was nucleus localized, suggesting that it is an authentic transcription factor (26). Pibzp1 from P. infestans was functionally characterized, which interacts with a protein kinase and is required for zoospore motility and plant infection (29).
A Per-ARNT-Sim (PAS)-containing bZIP transcription factor was annotated in P. sojae, named PsBZPc32 (28), but without functional characterization. In this study, we identified the PsBZPc32 ortholog in P. litchii named PlbZIP32. We analyzed the sequences of BZP32 orthologs, as well as the transcriptional profile of PlBZP32. We further characterized the function of the PlbZIP32 gene in oxidative stress response, asexual sporulation, and pathogenicity. Our results showed that PlBZP32 is required for the asexual development, oxidative stress response, and pathogenicity of P. litchii.

RESULTS
PlBZP32 gene belongs to a bZIP transcription factor family and is upregulated in zoospores, cysts, and late stages of infection. P. litchii BZP32 (PlBZP32) is the ortholog of P. sojae BZPc32, which encodes a bZIP transcription factor and possess a unique bZIP-PAS structure (30). There are 50 bZIP transcription factors in P. litchii based on a search with the Batch CD-search tool (CDD; in NCBI) (31), while PlBZP32 is the only one containing a bZIP-PAS structure (see Table S1 in the supplemental material). Here, we also searched the orthologous proteins from P. infestans, P. parasitica, P. ramorum, Pythium ultimum, Albugo candida, Saprolegnia parasitica, and Thalassiosira pseudonana for sequence alignment and phylogenetic analysis. PlBZP32 is a 392-amino acid (aa) protein with a bZIP domain located in aa positions 246 to 294 and a PAS domain located in aa positions 304 to 376 (see Fig. S1 in the supplemental material). We found that the bZIP domain-PAS structure is conserved in all the orthologs examined here (Fig. S1). Phylogenetic analysis showed that these kinds of bZIP transcription factors are widespread and conserved in oomycetes and algae, while Thalassiosira pseudonana is outside the oomycete group (Fig. 1).
To investigate the biological function of PlBZP32, we first examined the transcriptional profile of the PlBZP32 gene in various growth stages, including mycelia, sporangia, zoospores, cysts, germinating cysts, and oospores, and in infection stages, including 1.5, 3, 6, 12, 24, and 48 h postinoculation. Our results showed that PlBZP32 was upregulated in zoospores, cysts, and late stages of infection compared with that of the vegetative mycelial growth stage (Fig. 2), suggesting that PlBZP32 might function in these specific stages.
Generation of PlBZP32-silenced transformants. To generate the PlBZP32-silenced mutants, we used a polyethylene glycol (PEG)-mediated protoplast transformation with a pTORmRFP4 vector carrying an antisense full-length copy of PlBZP32. We obtained 182 transformants by antibiotic resistance screening. These transformants were confirmed by genomic PCR and subsequent quantitative reverse transcription PCR (qRT-PCR). Among them, three transformants (T15, T21, and T35) showed significant reduction in PlBZP32 transcriptional levels in mycelium and zoospore, compared with the wild-type strain SHS3 and the CK strain in which the backbone vector pTORmRFP4 was transformed ( Fig. 3A and B). Considering that homology-based transcriptional silencing might cause off-target effects, especially those affecting the expression of neighboring gene(s) (32,33), we searched the neighboring locus of PlBZP32 and found one gene (Pl105397; GenBank accession number MT396990) 875 nucleotides (nt) away from PlBZP32. Pl105397 encodes a structural maintenance of chromosomes (SMC) protein FIG 1 Phylogenetic analysis of PlBZP32 protein and its orthologs. The evolutionary history was inferred using the neighbor-joining method (47). Evolutionary analyses were conducted in MEGA7 (48). containing a Discs-large homologous regions (DHR) domain. The transcriptional level of Pl105397 in PlBZP32-silenced transformants was also examined, and no significant difference was achieved compared with wild-type and CK strains (Fig. 3C).
We checked the growth of these PlBZP32-silenced transformants in comparison with wild-type and CK strains. We measured the colony diameter of these strains at 3 days after inoculation and found that the silencing of the PlBZP32 gene did not affect the growth of P. litchii when cultured on carrot juice agar (CJA) medium ( Fig. 3D and E).
PlBZP32-silenced mutants were more sensitive to oxidative stress. To investigate whether PlBZP32 is involved in the response to oxidative stress, the PlBZP32silenced mutants and wild-type strain were exposed to H 2 O 2 concentrations of 2 and 5 mM, respectively. The PlBZP32-silenced mutants displayed a greater growth inhibition rate than the wild-type and CK strains (Fig. 4). These results suggested that PlBZP32 is required for resistance to oxidative stress.
We also analyzed the transcription levels of PlBZP32 after treatment with 5 mM H 2 O 2 . We found no significant changes in the transcription level of PlBZP32 from 5 minutes to 3 h posttreatment, compared with the untreated control (Fig. 5). These results suggests that the involvement of PlBZP32 in oxidative stress management might be not associated with a transcription-level regulation of this gene.
Silencing of PlBZP32 impaired the pathogenicity of P. litchii. The production and accumulation of reactive oxygen species, the oxidative burst, has been shown to occur in plant-pathogen interactions, such as H 2 O 2 and O 2Ϫ , directly acting as antimicrobial agents (34,35). Given that silencing of PlBZP32 decreased the resistance of P. litchii to H 2 O 2 , we next analyzed whether PlBZP32 is involved in the pathogenicity of this pathogen. PlBZP32-silenced mutants and the wild-type strain were inoculated on lychee leaves. At 12, 24, and 36 h after inoculation, we measured the diameter of lesions. The results showed that the lesions caused by PlBZP32-silenced mutants were significantly smaller than that caused by wild-type and CK strains (Fig. 6), indicating that PlBZP32 is required for the full pathogenicity of P. litchii.
Silencing of PlBZP32 affected the production of sporangia and germination of cysts. Transcriptional profiling analyses showed that the PlBZP32 gene was upregulated in asexual stages (zoospore and cyst) of P. litchii, so we also analyzed the production of sporangia and the germination rate of cysts. Our results showed that PlBZP32-silenced mutants produced approximately 40% to 50% more sporangia than wild-type and CK strains ( Fig. 7A and B) and each sporangiophore of PlBZP32-silenced mutants produced more branch tips and sporangia ( Fig. 7C and D). We tested the germination rate of cysts and found it was significantly decreased in the PlBZP32-silenced mutants. The germination rate of cysts is less than 50% in PlBZP32-silenced mutants, whereas it was approximately 80% in wild-type or nonsilenced strains (Fig. 8). However, the silencing of PlBZP32 did not significantly affect the length and germination of sporangia (see Fig. S2 and S3 in the supplemental material). Thus, PlBZP32 may negatively regulate the sporangium production and positively regulate the cyst germination. The colony diameters of the tested strains were measured. Wild-type and CK strains were used as controls. The growth inhibition rate was calculated using the following formula: (the diameter of control Ϫ the diameter of treated strain)/(the diameter of control) ϫ100%. The bar chart depicts means Ϯ SD derived from 3 independent repeats, and different letters represent a significant difference (P Ͻ 0.05) based on Duncan's multiple range test method. PlBZP32 disruption attenuates the activities of extracellular peroxidases and laccases. Peroxidase and laccase have been reported to be involved in the resistance of ROS and, thus, critical for fungal pathogenicity (14,36). Given that the PlBZP32silenced mutants displayed reduced pathogenicity compared with the wild-type (WT) and nonsilenced strains, we further analyzed the peroxidase and laccase activities in these strains.
Extracellular peroxidase activity was assessed based on Congo red (CR) degradation (37). As shown in Fig. 9A and C, diameters of the degradation halos caused by the three PlBZP32-silenced mutants were reduced compared with the controls. On the other hand, the activities of the extracellular laccase were measured by an oxidation assay of ABTS [2,2=-azinobis (3-ethylbenzothiazolinesulfonic acid)] (37). Three PlBZP32-silenced mutants showed a significantly decreased accumulation of ABTS, as visualized by dark purple staining around the mycelial mat, compared with the nonsilenced and wild-type strains ( Fig. 9B and D).

DISCUSSION
In oomycetes, bZIP transcription factors are a large protein family. Fifty bZIP transcription factors were identified in P. litchii; 71 bZIP transcription factors were predicted from P. sojae, of which 45 were confirmed by CDD (NCBI conserved  domain database) or the SMART database (28); 22 bZIPs were identified in P. ramorum (27); and 38 bZIPs were identified in P. infestans (26). Based on transcriptional analysis, several P. sojae infection-associated bZIPs were regulated by H 2 O 2 treatment (28). Several bZIP proteins from P. infestans were functionally characterized and only Pibzp1 was found essential for virulence (26,29). In our study, we identified and analyzed a bZIP transcription factor, PlBZP32, which is an ortholog of P. sojae BZPc32 and possesses a bZIP-PAS structure.
CRISPR/Cas9-mediated gene disruption in P. sojae was established in 2016 (38), and our group also succeeded in disrupting PlPAE4 and PlPAE5 genes using this technique (39). However, we failed to knock out PlBZP32 in P. litchii, suggesting that it might be an essential gene. Therefore, we investigated the function of PlBZP32 by a gene silencing strategy. We obtained only 3 PlBZP32-silenced transformants out of 182 transformants screened by antibiotic resistance. The silencing efficiency of this gene is much lower than that of PlMAPK10, as reported by our group previously (40). Out of all the G418-resistant transformants, only around 37% contain the empty vector, based on PCR verification. This gene was downregulated by 31% to 39% in the mycelia of the three PlBZP32-silenced transformants and 65% to 87% in the zoospores. The efficiency of downregulating this gene expression seems higher in zoospores than in mycelia, likely because the transcriptional level of PlBZP32 is highly expressed in zoospores compared to mycelia. The gene silencing strategy might affect other genes with high identity in the sequence. However, other P. litchii BZP-encoding genes showed less than 51% identity with PlBZP32; thus, we assumed that our targeted silencing of PlBZP32 may not affect the expression of other BZP-encoding genes. On the other hand, homologybased transcriptional silencing may also cause off-target effects on the neighboring gene(s) (32,33), but we managed to confirm that it did not happen to the annotated gene Pl105397 that was proximal to the PlBZP32 locus.
The production of sporangia and germination of cysts are very important for the asexual reproduction, dissemination, and infection of P. litchii. In this study, the PlBZP32-silencing mutants produced more sporangia but with reduced germination rate of cysts, and the sporangial germination rate was not affected by silencing of PlBZP32. These results suggested that PlBZP32 might be a negative regulator in the production of sporangia and a positive regulator in the germination of cysts.
We further found that PlBZP32 is required for the full virulence of P. litchii and its resistance to oxidative stress caused by H 2 O 2 . This may indicate that PlBZP32 regulates P. litchii virulence via resistance to oxidative stress. In supporting this hypothesis, we found that the enzymatic activities of ROS scavenger peroxidases and laccases were both decreased in the PlBZP32-silencing mutants. This finding is consistent with the report of M. oryzae MoAP1 (14). However, the direct target genes of the PlBZP32 transcription factor in regulating oxidative resistance and/or pathogenicity of P. litchii are unclear.
PAS domain-containing proteins detect a wide range of physical and chemical stimuli and associate with a series of signal transduction systems (41). PlBZP32 contains a PAS domain immediately following the bZIP domain, suggesting that it might interact with some small ligand through a different pathway than other bZIP transcription factors. Further study is needed for characterizing the function of the PAS domain in bZIP transcription factors.
Overall, our study identified a bZIP transcription factor responsible for P. litchii development, oxidative stress response, and pathogenicity. At present, we do not fully understand its regulatory mechanism, particularly the function of its PAS domain (and the corresponding ligands/signal molecules) and its downstream target genes. It is also worth further investigating the function diversity (if any) among oomycete species.
P. litchii strain and culture conditions. P. litchii strain SHS3 (wild type) was isolated from Guangdong Province, China, and cultured on carrot juice agar (CJA) medium (juice from 200 g carrot for 1 liter medium, 15 g agar/liter for solid media) at 25°C in the dark. The PlBZP32-silenced transformants were maintained on CJA medium containing 50 g/ml G418.
Sporangia were harvested by flooding the mycelia, which had been cultured on CJA media for 5 days, with sterile water, then filtering the subsequent suspension through a 100-m strainer (44). The suspension was incubated with sterile distilled water at 16°C for 2 h for releasing zoospores. For oospore collection, P. litchii strains were inoculated on CJA medium which was covered by an Amersham Hybond membrane. After 10 days, the aerial mycelia were separated from the medium by removing the Amersham Hybond membrane. Then, oospores were scraped from the medium surface and collected.
Transformation of P. litchii. The full-length open reading frame of PlBZP32 was amplified with primers PlBZP32-BsiwI-F (AAACGTACGATGGACTTCACGTCGCCTAATG) and PlBZP32-ClaI-R (AAAATCGATA ATGTCGCGGTCCACAC), using the cDNA from P. litchii strain SHS3 as the template. The amplified PlBZP32 fragment was ligated into the pTORmRFP4 vector digested with ClaI and BsiWI in antisense orientation (45,46).
Pathogenicity assays on lychee leaves. The 8-to 10-day-old soft expanded lychee leaves were collected from the same plant in an orchard in South China Agricultural University, Guangzhou, Guangdong Province, China. The 5-mm hyphal plugs of the PlBZP32-silenced mutants were inoculated on the abaxial side of lychee leaves in a dark box and then were placed in climate room under 80% humidity in the dark at 25°C for 36 h (44). The virulence of each transformant was tested with the wild-type strain (SHS3) and CK strain as positive controls. Lesion length (the longest diameter of the lesion) was measured at 36 h postinoculation. At least three independent repeats were performed for each instance with 12 leaves/repeats. Extracellular enzyme activity assays. The detection of peroxidase secretion and laccase activity was performed following the reported procedure (37). The experiments were repeated three times independently, with three technical replicates each time.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only.