Genome-Wide Identification of bZIP Transcription Factor Genes and Functional Analyses of Two Members in Cytospora chrysosperma

The basic leucine zipper (bZIP) transcription factor (TF) family, one of the largest and the most diverse TF families, is widely distributed across the eukaryotes. It has been described that the bZIP TFs play diverse roles in development, nutrient utilization, and various stress responses in fungi. However, little is known of the bZIP members in Cytospora chrysosperma, a notorious plant pathogenic fungus, which causes canker disease on over 80 woody plant species. In this study, 26 bZIP genes were systematically identified in the genome of C. chrysosperma, and two of them (named CcbZIP05 and CcbZIP23) significantly down-regulated in CcPmk1 deletion mutant (a pathogenicity-related mitogen-activated protein kinase) were selected for further analysis. Deletion of CcbZIP05 or CcbZIP23 displayed a dramatic reduction in fungal growth but showed increased hypha branching and resistance to cell wall inhibitors and abiotic stresses. The CcbZIP05 deletion mutants but not CcbZIP23 deletion mutants were more sensitive to the hydrogen peroxide compared to the wild-type and complemented strains. Additionally, the CcbZIP23 deletion mutants produced few pycnidia but more pigment. Remarkably, both CcbZIP05 and CcbZIP23 deletion mutants were significantly reduced in fungal virulence. Further analysis showed that CcbZIP05 and CcbZIP23 could regulate the expression of putative effector genes and chitin synthesis-related genes. Taken together, our results suggest that CcbZIP05 and CcbZIP23 play important roles in fungal growth, abiotic stresses response, and pathogenicity, which will provide comprehensive information on the CcbZIP genes and lay the foundation for further research on the bZIP members in C. chrysosperma.


Introduction
Transcription factors (TFs) are involved in all kinds of biological processes, such as cellular growth, differentiation, and the response to environmental factors, through modulating the expression of downstream genes. Their roles, including specific promoter sequences binding to transcriptional inhibition or activation, interaction with other TFs or molecular chaperones, as well as post-translational modifications, are crucial to regulating the transcription of target genes [1][2][3][4]. According to the similarities of primary and/or three-dimensional structure of the DNA-binding and multimerization domains, TFs can be categorized into different families, for example, the basic leucine zipper (bZIP) proteins, fungal-specific Zn2Cys6 proteins, Cys2His2 (C 2 H 2 ) zinc finger proteins, MADS-box proteins, helix-loop-helix (HLH) proteins, homeobox proteins, and so on [5][6][7]. Among them, the bZIP proteins, widely distributed in eukaryotes, are regarded as one of the central regulators functioning in various biological processes in pathogenic fungi, such as fungal plant immunity and promote the infection on host poplar [50]. Therefore, the TFs are the essential components to connect the upstream signals and downstream effectors, which can be regarded as the potential targets to reveal the molecular pathogenic mechanism of plant pathogens. However, little is known about the functions of TFs in C. chrysosperma.
In this study, we identified and characterized 26 bZIP genes in C. chrysosperma, namely, CcbZIP01 to CcbZIP26. Among them, CcbZIP05 and CcbZIP23, significantly down-regulated in CcPmk1 deletion mutant, were selected for further analyses. The results showed that both CcbZIP05 and CcbZIP23 contribute to the fungal development, stress response, and pathogenicity of C. chrysosperma. These results provide a foundation for further study of other bZIP members in C. chrysosperma.

Identification of bZIP Transcription Factors in C. chrysosperma
The draft genome sequence of C. chrysosperma was sequenced by our lab. (Unpublished data, NCBI GenBank accession number JAEQMF000000000.) The bZIP protein sequences of various fungal species were downloaded from NCBI (https://www.ncbi.nlm.nih.gov/, accessed on 11 October 2019) or FTFD (http://ftfd.snu.ac.kr/index.php?a=view, accessed on 24 December 2019) and were used as queries for BLASTP searches (E-value cutoff < 1 × 10 −5 ) against the C. chrysosperma genome. The resulting bZIP domains were used to generate motif by using the hidden Markov model (HMM) with E-value threshold ≤ 500, and then searched in the C. chrysosperma genome [51]. All output genes (removing the repetitive sequences) manually confirmed the existence of the bZIP domain (InterPro ID #IPR004827) by using the InterProScan (http://www.ebi.ac.uk/interpro/, accessed on 26 December 2019). The sequences containing the bZIP domain were regarded as candidate CcbZIPs for the further analyses.

Gene Structural Characterization and Phylogenetic Analysis
The theoretical isoelectric points and molecular weights of the identified gene products were predicted by using the ExPASy proteomics server (http://web.expasy.org/ protparam/, accessed on 30 December 2019). The genome sequences of the bZIP genes and the corresponding CDS sequences were committed to the Gene Structure Display Server (GSDS, http://gsds.cbi.pku.edu.cn/, accessed on 27 December 2020) to analyze the number and alignment of introns and exons.
Multiple sequence alignments of the identified bZIP TF sequences were performed using MEGA6 [52], followed by minor modifications using GeneDoc [53]. A phylogenetic analysis was performed with raxmlGUI1.3 software [54], using the maximum likelihood method, and a bootstrap test with 1000 iterations. TBtools (V0.66836) was used to compile phylogenetic trees and relative expression level in heatmap [55]. The Pathogen Host Interaction Base (PHIB) (http://phi-blast.phi-base.org/, accessed on 24 December 2020) was used to identify bZIP homologs that had been functionally analyzed.

Conditions for Fungal Strains' Growth and Treatment
The wild-type (WT) strain (CFCC 89981) of C. chrysosperma was derived from the Forest Pathology Laboratory of Beijing Forestry University (Strain No: G-YS-11-C1) [56]. All strains were cultured on potato dextrose agar (PDA, 200 g potato, 20 g dextrose, 15 g agar, and 1 L water) medium at 25 • C. Potato dextrose broth (PDB, 200 g potato, 20 g dextrose, and 1 L water) was used to culture mycelium ready for DNA and RNA extraction. To analyze the differences in vegetative growth among the strains, growth patterns were monitored and colony diameters were measured daily post-inoculation (dpi) on PDA plates and continued for 3 days. To calculate the conidial production, the number of pycnidia was counted at 30 dpi. To assess the stress response, small agar blocks were cut from the edges of 2-day-old cultures and placed on PDA plates supplemented with different stress agents including 0.05 M NaCl, 0.04 M KCl, 0.8 M Sorbitol, 20 µg/mL calcofluor white (CFW), 200 µg/mL Congo red (CR), and 0.01% Sodium dodecyl sulfate (SDS), then incubated at 25 • C in the dark for 60 h. In the oxidative stress test, a 1 × 10 5 spores/mL suspension was added into the melted PDA medium (about 45~50 • C) and then poured into 9-cm plates with a drop of 5 M or 10 M H 2 O 2 on a filter in the center of the medium plate. The diameter of the inhibition zone was measured at 60 hpi. In order to visualize the chitin deposition and distribution, conidia of each strain were inoculated on 1-mm PDA slides for 1 day and then stained with 10 mg/mL CFW. All tests were repeated three times, and all data were analyzed in SPSS 16.0 by One-Way ANOVA and Duncan's range test to measure specific differences between strains. Significant changes between different treatments were calculated with p < 0.05.

Generation of CcbZIP Gene Deletion Mutants
The CcbZIP genes were disrupted using the split-marker system. In the case of CcbZIP05, firstly, the upstream (~1.3 kb) and downstream (~1.2 kb) flanking sequences of CcbZIP05 were PCR amplified with specific primer pairs CcbZIP05-5Ffor/rev and CcbZIP05-3Ffor/rev, as shown in the Table S1. Then, the hygromycin B resistance cassette (HPH), which includes sequences of approximately~20 bp that overlap with the 5 and 3 flanking sequences, was amplified by the specific primer pairs CcbZIP05-5Frev+/HY-R and YG-F/CcbZIP05-3Ffor+. Then, the upstream and downstream fragments and two-thirds of the hygromycin cassette were fused by overlapping PCR using the primers CcbZIP05-5Ffor/HY-R and YG-F/CcbZIP05-3Frev, respectively. The resulting two overlapped fragments were transformed into the protoplasts of C. chrysosperma by using the PEG-mediated transformation, as described previously [57]. Finally, the gene deletion mutants were screened by PCR with the primer pairs External-CcbZIP05for/rev and Internal-CcbZIP05for/rev, then confirmed by Southern blotting analysis according to the manufacturer's instructions (DIG-High Prime DNA Labeling and Detection Starter Kit I).
To complement the CcbZIP05 deletion mutant, fragments containing upstream~1.5 kb of local promoter sequence, open reading frame, and downstream~0.2 kb of the sequence were cloned from gDNA using primer CcbZIP05-Compfor/Comprev. The obtained PCR products and the geneticin-resistant cassette were co-transformed into the protoplasts of the deletion mutant strain. Successful complementation was confirmed by PCR with the primer pair External-CcbZIP05for/rev and Internal-CcbZIP05for/rev. The CcbZIP23 deletion mutants were obtained in the same way and all the primers mentioned are listed in Table S1.

Pathogenicity Tests
For the pathogenicity test, healthy annual twigs of Populus euramericana (a susceptible Populus species) were used. C. chrysosperma initiated infection through wounds; therefore, the selected branches were scalded with a 5-mm diameter hot iron rod and then inoculated with 5-mm agar plugs taken from the leading edge of colonies of wild-type, gene deletion mutant, and complementary strains, respectively. The inoculated twigs were placed on top of trays, kept humid with distilled water, and incubated at 25 • C in a day/night cycle. The lesions of the inoculated twigs were photographed and measured at 3 dpi and 6 dpi, respectively. The experiments were repeated three times, with at least 25 twigs for each strain.

RNA Extraction and RT-qPCR Analysis
To analyze the expression levels of CcbZIP05 and CcbZIP23 in CcPmk1-deletion mutants, 5-mm agar plugs of wild-type, ∆CcPmk1, and complemented strains were added into PDB supplemented with sterilized poplar twigs, respectively, and then incubated at 25 • C with shaking at 150 rpm for 2 days. To assess the expression levels of candidate effector genes in ∆CcbZIP05 and ∆CcbZIP23 mutants, the same method was used. To examine the expression of putative chitin synthase-related genes (CHS), the strains were growth in PDB at 25 • C with shaking at 150 rpm for 3 days.
Total RNA was extracted with TRIzol reagent (Invitrogen, USA) according to the manufacturer's instructions. The total RNA was treated with DNase before reverse transcription. Agarose gel electrophoresis was used to examine the quality of total RNA and to estimate its concentration. First-strand cDNA was synthesized from 1 µg RNA with ABScript II cDNA Fist-Strand Synthesis Kit (ABclonal, China), according to the manufacturer's instructions. The qRT-PCR assay was conducted with 2X Universal SYBR Green Fast qPCR Mix (ABclonal, China) using an ABI 7500 real-time PCR system (Applied Biosystems, USA). The CcActin gene served as the endogenous control for all qRT-PCR analyses. All samples were independently subjected to three replicate experiments. The relative expression of genes was calculated by using the 2 −∆∆Ct method.

Identification of bZIP Genes in C. chrysosperma
To characterize the bZIP members of C. chrysosperma, we downloaded the bZIP protein sequences of eight fungal species, including 22 bZIPs in M. oryzae, 15 bZIPs in Botrytis cinerea, 20 bZIPs in Verticillium dahliae, 15 bZIPs in Saccharomyces cerevisiae, 25 bZIPs in Colletotrichum gloeosporioides, 17 bZIPs in Sclerotinia sclerotiorum, 15 bZIPs in Cryphonectria parasitica, and 26 bZIPs in F. graminearum. Then we extracted the bZIP domain sequences of these 155 bZIPs and used them as templates to search through the C. chrysosperma genome database with HMMER, and 56 putative hits were obtained. Subsequently, the acquired 56 sequences were manually confirmed for the presence of bZIP domain by using InterProScan on the EBI web server. Finally, 26 putative sequences were identified, which contained the bZIP domain including bZIP_1 domain (PF00170), bZIP_Maf (PF03131), or bZIP_2 (PF07716). These genes were named CcbZIP01 to CcbZIP26 according to their locus order. The length of these 26 CcbZIP proteins ranged from 163 to 633 aa, which contained 0 to 3 introns, and theoretical molecular weight ranged from 18.24 to 69.87 KDa, with predicted pI values ranging from 4.635 to 11.082. Among them, the largest difference in length between the cDNA and gDNA sequences of Ccbzip04 was 1709 base pairs (bp) ( Table 1).

Analysis of Conserved Domains and Motif in CcbZIP Proteins
The classical bZIP domain usually consists of a basic region and a leucine-zipper region. The basic region is highly conserved, with an N-X 7 -R/K motif and the variable leucine-zipper region with a leucine or other hydrophobic amino acids every seven amino acids. To investigate the characteristics of the CcbZIP domain, multiple amino acid sequences of the bZIP domain were aligned, as shown in Figure 1. The core asparagine (N) residues of the bZIP domains of CcbZIP04 and CcbZIP19 were substituted with aspartic acid (D), while they were substituted with valine (V) in CcbZIP12 and isoleucine (I) in CcbZIP24. Except for the four CcbZIPs mentioned above, all the remaining CcbZIP domains contained invariant N-X 7 -R motif. The first heptad repeat Leu was highly conserved in the leucine zipper region, except for CcbZIP14, which was replaced by Met. However, the following heptad repeat Leu was often replaced by other bulky hydrophobic amino acids, e.g., the second heptad repeat Leu in CcbZIP13, -20, and -26 was replaced by Met.

Analysis of Conserved Domains and Motif in CcbZIP Proteins
The classical bZIP domain usually consists of a basic region and a leucine-zipper region. The basic region is highly conserved, with an N-X7-R/K motif and the variable leucine-zipper region with a leucine or other hydrophobic amino acids every seven amino acids. To investigate the characteristics of the CcbZIP domain, multiple amino acid sequences of the bZIP domain were aligned, as shown in Figure 1. The core asparagine (N) residues of the bZIP domains of CcbZIP04 and CcbZIP19 were substituted with aspartic acid (D), while they were substituted with valine (V) in CcbZIP12 and isoleucine (I) in CcbZIP24. Except for the four CcbZIPs mentioned above, all the remaining CcbZIP domains contained invariant N-X7-R motif. The first heptad repeat Leu was highly conserved in the leucine zipper region, except for CcbZIP14, which was replaced by Met. However, the following heptad repeat Leu was often replaced by other bulky hydrophobic amino acids, e.g., the second heptad repeat Leu in CcbZIP13, -20, and -26 was replaced by Met.  Black, medium gray, and light gray indicate the conserved percent of 100%, >80%, and >60%, respectively. The short, black lines in the middle represent a typical bZIP domain; asterisks show the conserved amino acids of bZIP domains. The bottom part is the sequence logo formed by bZIP domain from CcbZIP genes, and the larger letters of the amino acid residues, the more frequently they appear at the same site.
In addition, three CcbZIP members contained other predicted domains except for the bZIP domain, which may have different specific functions. CcbZIP03 contained an Aft1_HRA domain (IPR021755), Aft1_OSM domain (IPR020956), and Aft1_HRR domain (IPR021756) at the N-terminal of the sequence. CcbZIP07 contained a Hap4_TF_heteromerisation domain (IPR018287) at the N-terminal of the sequence. CcbZIP25 contained a PAP1 domain (IPR013910) at the C-terminal of the sequence ( Figure S1).

Intron Numbers and Their Distribution Pattern in CcbZIPs
The evolutionary imprint of some gene families can be revealed by the intron/exon arrangement [58]. To further understand the structural characteristics of the CcbZIP gene, the intron arrangement pattern and insertion sites of CcbZIPs were identified using the GSDS. The different intron distribution patterns are displayed in Figure 2. Among them, 7 CcbZIP genes contained one intron, 11 CcbZIP genes contained two introns, and 3 CcbZIP genes contained three introns. The intron lengths ranged from 52 to 1477 bp; 58% of the introns were less than 100 bp and only two introns were larger than 1000 bp.
(IPR021756) at the N-terminal of the sequence. CcbZIP07 contained a Hap4_TF_hete erisation domain (IPR018287) at the N-terminal of the sequence. CcbZIP25 contain PAP1 domain (IPR013910) at the C-terminal of the sequence ( Figure S1).

Intron Numbers and Their Distribution Pattern in CcbZIPs
The evolutionary imprint of some gene families can be revealed by the intron/ arrangement [58]. To further understand the structural characteristics of the CcbZIP g the intron arrangement pattern and insertion sites of CcbZIPs were identified using GSDS. The different intron distribution patterns are displayed in Figure 2. Among th 7 CcbZIP genes contained one intron, 11 CcbZIP genes contained two introns, and 3 Cc genes contained three introns. The intron lengths ranged from 52 to 1477 bp; 58% o introns were less than 100 bp and only two introns were larger than 1000 bp. Previous reports have shown that the insertion position of introns affects gen pression [59]. Since the basic region holds a special status in the bZIP domain, which bind to specific DNA sequences to regulate gene expression, the distribution patter introns within the basic region is important for its functions. We analyzed the patte intron positions within the basic, hinge, and leucine regions, and 13 CcbZIP genes tained one intron in the bZIP domain region. More than half of the introns' (CcbZIP 03, -10, -11, -16, -23, -26) insertion sites were located in the N-X7-R motif ( Figure 3). I guingly, the intron insertion sites of CcbZIP23 and CcbZIP26 were located in the Previous reports have shown that the insertion position of introns affects gene expression [59]. Since the basic region holds a special status in the bZIP domain, which may bind to specific DNA sequences to regulate gene expression, the distribution pattern of introns within the basic region is important for its functions. We analyzed the pattern of intron positions within the basic, hinge, and leucine regions, and 13 CcbZIP genes contained one intron in the bZIP domain region. More than half of the introns' (CcbZIP01, -03, -10, -11, -16, -23, -26) insertion sites were located in the N-X 7 -R motif ( Figure 3). Intriguingly, the intron insertion sites of CcbZIP23 and CcbZIP26 were located in the core arginine (R) in the N-X 7 -R motif, which are both inserted between the second and third codon. arginine (R) in the N-X7-R motif, which are both inserted between the second and third codon. The red letter means splicing sites are located between two codons. Black, medium gray, and light gray indicate the conserved percent of 100%, >80%, and >60%, respectively.

Phylogenetic Analysis of bZIP Gene Family
To investigate the evolutionary relationship of the CcbZIP genes with other fungal bZIP genes, a phylogenetic analysis of bZIP genes in nine fungal species was performed. A phylogenetic tree was produced from 181 fungal bZIP sequences (including the 26 CcbZIP proteins in C. chrysosperma) (Table S2), and the defined clades consisting of homologous members are listed in Figure S2, respectively. These 181 bZIPs were divided into eight clades and were assigned to A through H. Overall, the 26 CcbZIP genes were evenly distributed in all these clades, and CcbZIP23 and CcbZIP26 belonged to the same subcluster in the clade D. The bZIP superfamily has been reported to have accumulated from a single eukaryotic gene ancestor and has undergone multiple independent amplifications [60]. None of the bZIPs from C. gloeosporioides and S. sclerotiorum was distributed in clade E, indicating the late divergence of those bZIP genes in these fungi.
Meanwhile, we identified the pathogenicity functions of the respective homologs of the 26 CcbZIP genes in C. chrysosperma using the online website PHIB (Table S3). The results showed that 11 of CcbZIP homologous genes were required for the virulence in different fungi, and one of them (MoAP1 from M. oryzae, CcbZIP25 homologous) was indispensable for the fungal pathogenicity.

Phylogenetic Analysis of bZIP Gene Family
To investigate the evolutionary relationship of the CcbZIP genes with other fungal bZIP genes, a phylogenetic analysis of bZIP genes in nine fungal species was performed. A phylogenetic tree was produced from 181 fungal bZIP sequences (including the 26 CcbZIP proteins in C. chrysosperma) (Table S2), and the defined clades consisting of homologous members are listed in Figure S2, respectively. These 181 bZIPs were divided into eight clades and were assigned to A through H. Overall, the 26 CcbZIP genes were evenly distributed in all these clades, and CcbZIP23 and CcbZIP26 belonged to the same subcluster in the clade D. The bZIP superfamily has been reported to have accumulated from a single eukaryotic gene ancestor and has undergone multiple independent amplifications [60]. None of the bZIPs from C. gloeosporioides and S. sclerotiorum was distributed in clade E, indicating the late divergence of those bZIP genes in these fungi.
Meanwhile, we identified the pathogenicity functions of the respective homologs of the 26 CcbZIP genes in C. chrysosperma using the online website PHIB (Table S3). The results showed that 11 of CcbZIP homologous genes were required for the virulence in different fungi, and one of them (MoAP1 from M. oryzae, CcbZIP25 homologous) was indispensable for the fungal pathogenicity.

Construction of CcbZIP Deletion Mutants
Mitogen-activated protein kinase (MAPK) signaling cascades are highly conserved in eukaryotes and play essential roles in developmental processes and various cellular responses [62]. Our previous works revealed that the MAKP gene CcPmk1 plays crucial roles in fungal virulence in C. chrysosperma through regulating the expression of downstream genes [48]. Thus, we queried the expression data of the 26 CcbZIP genes through the transcriptomic data of CcPmk1 deletion mutant and wild-type (NCBI SRA database with accession numbers SRR12262932, SRR12262933, SRR12262934 (wild-type 1-3), and SRR12262935, SRR12262936, SRR12262937 (ΔCcPmk1-1-3)) (Table S5) [47] and found that only CcbZIP05 and CcbZIP23 were significantly down-regulated in the CcPmk1-deficient mutant compared to the wild-type ( Figure 5A), which was also validated by using qRT-PCR ( Figure 5B), indicating that they were modulated by CcPmk1 and might be involved in fungal pathogenicity. Therefore, CcbZIP05 and CcbZIP23 were selected for further functional analyses.
In order to characterize the functions of CcbZIP05 and CcbZIP23 on fungal development and pathogenicity, we carried out gene knockout by homologous recombination.

Construction of CcbZIP Deletion Mutants
Mitogen-activated protein kinase (MAPK) signaling cascades are highly conserved in eukaryotes and play essential roles in developmental processes and various cellular responses [62]. Our previous works revealed that the MAKP gene CcPmk1 plays crucial roles in fungal virulence in C. chrysosperma through regulating the expression of downstream genes [48]. Thus, we queried the expression data of the 26 CcbZIP genes through the transcriptomic data of CcPmk1 deletion mutant and wild-type (NCBI SRA database with accession numbers SRR12262932, SRR12262933, SRR12262934 (wild-type 1-3), and SRR12262935, SRR12262936, SRR12262937 (∆CcPmk1-1-3)) (Table S5) [47] and found that only CcbZIP05 and CcbZIP23 were significantly down-regulated in the CcPmk1-deficient mutant compared to the wild-type ( Figure 5A), which was also validated by using qRT-PCR ( Figure 5B), indicating that they were modulated by CcPmk1 and might be involved in fungal pathogenicity. Therefore, CcbZIP05 and CcbZIP23 were selected for further functional analyses.
Briefly, the target bZIP gene in the wild-type was replaced by the hygromycin marker gene. Successful single integration deletion mutants were generated for both two CcbZIP genes (named ΔCcbZIP05 and ∆CcbZIP23), and complementation of the two deletion mutants was performed respectively (named ΔCcbZIP05-4/C and ∆CcbZIP23-1/C) ( Figure S3). The qRT-PCR was used to measure the relative expression levels of CcbZIP05 and CcbZIP23 in the wild-type, ΔCcPmk1 mutants, and complemented strains. The CcActin gene was used as the reference gene. The experiments were repeated three times. The data were analyzed using Duncan's range test. The asterisks indicate significant differences (p < 0.05).

CcbZIP05 and CcbZIP23 Are Important for the Development of C. chrysosperma
To evaluate the functions of CcbZIP05 and CcbZIP23 on mycelia growth, the ΔCcbZIP05 and ΔCcbZIP23 deletion mutants were grown on PDA plates. Compared with the wild-type, the ΔCcbZIP05 and ΔCcbZIP23 mutants exhibited a significantly smaller colony diameter, and the growth of the complemented strains ΔCcbZIP05-4/C and ΔCcbZIP23-1/C on PDA media was restored to wild-type levels ( Figure 6A,B). In addition, we observed the growth morphology of the mycelium under the light microscope. The results showed that the mycelia of the ΔCcbZIP05 and ΔCcbZIP23 mutants had more frequent branches than wild-type and complemented strains. In addition, the mycelial branch growth direction of ΔCcbZIP23 mutants became twisted and coiled ( Figure 6C). Interestingly, we found that ΔCcbZIP23 but not ΔCcbZIP05 mutants would accumulate obvious pigment in the PDA plates compared to the wild-type and complemented strains ( Figures 6D, S4 and S5). (B) The qRT-PCR was used to measure the relative expression levels of CcbZIP05 and CcbZIP23 in the wild-type, ∆CcPmk1 mutants, and complemented strains. The CcActin gene was used as the reference gene. The experiments were repeated three times. The data were analyzed using Duncan's range test. The asterisks indicate significant differences (p < 0.05).
In order to characterize the functions of CcbZIP05 and CcbZIP23 on fungal development and pathogenicity, we carried out gene knockout by homologous recombination. Briefly, the target bZIP gene in the wild-type was replaced by the hygromycin marker gene. Successful single integration deletion mutants were generated for both two CcbZIP genes (named ∆CcbZIP05 and ∆CcbZIP23), and complementation of the two deletion mutants was performed respectively (named ∆CcbZIP05-4/C and ∆CcbZIP23-1/C) ( Figure S3).

CcbZIP05 and CcbZIP23 Are Important for the Development of C. chrysosperma
To evaluate the functions of CcbZIP05 and CcbZIP23 on mycelia growth, the ∆CcbZIP05 and ∆CcbZIP23 deletion mutants were grown on PDA plates. Compared with the wild-type, the ∆CcbZIP05 and ∆CcbZIP23 mutants exhibited a significantly smaller colony diameter, and the growth of the complemented strains ∆CcbZIP05-4/C and ∆CcbZIP23-1/C on PDA media was restored to wild-type levels ( Figure 6A,B). In addition, we observed the growth morphology of the mycelium under the light microscope. The results showed that the mycelia of the ∆CcbZIP05 and ∆CcbZIP23 mutants had more frequent branches than wildtype and complemented strains. In addition, the mycelial branch growth direction of ∆CcbZIP23 mutants became twisted and coiled ( Figure 6C). Interestingly, we found that ∆CcbZIP23 but not ∆CcbZIP05 mutants would accumulate obvious pigment in the PDA plates compared to the wild-type and complemented strains ( Figure 6D, Figures S4 and S5).
Subsequently, we analyzed the conidial production of each strain. The result showed that ΔCcbZIP23 mutants formed fewer pycnidia than wild-type and ΔCcbZIP23-1/C strains ( Figure 6D,E), while ΔCcbZIP05 mutants had no difference in the number of pycnidia compared with the wild-type ( Figure S3). These results showed that CcbZIP05 and CcbZIP23 were involved in the fungal growth, and CcbZIP23 regulated the conidial production and pigment formation in C. chrysosperma. The asterisks indicate significant differences (* p < 0.05, ** p < 0.01).

CcbZIP05 and CcbZIP23 Are Required for the Stress Responses
To evaluate the functions of CcbZIP05 and CcbZIP23 in cell wall integrity and stress responses, the fungal growths of the wild-type, ΔCcbZIP05/ΔCcbZIP23, and Subsequently, we analyzed the conidial production of each strain. The result showed that ∆CcbZIP23 mutants formed fewer pycnidia than wild-type and ∆CcbZIP23-1/C strains ( Figure 6D,E), while ∆CcbZIP05 mutants had no difference in the number of pycnidia compared with the wild-type ( Figure S3). These results showed that CcbZIP05 and CcbZIP23 were involved in the fungal growth, and CcbZIP23 regulated the conidial production and pigment formation in C. chrysosperma.

CcbZIP05 and CcbZIP23 Are Required for the Stress Responses
To evaluate the functions of CcbZIP05 and CcbZIP23 in cell wall integrity and stress responses, the fungal growths of the wild-type, ∆CcbZIP05/∆CcbZIP23, and complemented strains on the PDA supplemented with cell wall interfering agents (CFW, CR, and SDS), osmotic interfering agents (NaCl, KCl, and sorbitol), and oxidative stress agent (H 2 O 2 ) were analyzed. As shown in the Figure 7, the growth inhibition rate of the ∆CcbZIP05 deletion mutants was significantly reduced on the PDA plate supplied with CFW, CR, and sorbitol compared with the wild-type, while no distinguished differences were found between the ∆CcbZIP05 mutants and wild-type grown on the PDA plate supplemented with SDS, NaCl, and KCl. The data are summarized in Figure 7B. Additionally, the ∆CcbZIP05 mutants were more sensitive to 5 M H 2 O 2 or 10 M H 2 O 2 stress than the wild-type and complemented strains ( Figure 7B). Similarly, the growth of the ∆CcbZIP23 mutants showed significantly lower growth inhibition than WT on PDA plate supplied with CFW, CR, SDS, and sorbitol, while the ∆CcbZIP23 mutants enhanced their tolerance to salt stresses (NaCl and KCl) compared to the wild-type and complemented strains ( Figure 7B). Intriguingly, ∆CcbZIP23 mutants displayed comparable phenotypes compared to the wild-type and complemented strains on the PDA plates supplied with 5 M H 2 O 2 or 10 M H 2 O 2 ( Figure 7B) . These results suggest that CcbZIPs are involved in stress responses, while they may show convergent and distinguished roles, etc.  Chitin is an essential component of the fungal cell wall, which plays important roles in hyphal growth and fungal morphogenesis. Therefore, we compared the chitin deposition among each strain by using the CFW staining. As shown in Figure 8A, B, obvious chitin deposition was observed at the septa and hyphal tips in the wild-type, ∆CcbZIP23, and complemented stains, while the chitin was not significantly accumulated at the hyphal tips of the ∆CcbZIP05 mutants. Additionally, a significantly increased number of septa was found in ∆CcbZIP23 mutants compared to the wild-type, ∆CcbZIP05, and complemented strains ( Figure 8A). Subsequently, we calculated the expression of the putative chitin synthase-encoding genes (CcChs1, CcChs2, CcChs3, CcChs6, CcChs7, CcChs8), which had been analyzed in our previous works [48]. The results showed that the expression of CcChs2 was significantly increased while the expression of CcChs1, CcChs3, CcChs6, and CcChs7 was significantly reduced in the ∆CcbZIP05 mutants ( Figure 8C). As for the ∆CcbZIP23 mutants, different expression patterns were found. The expression of CcChs1, CcChs3, and CcChs6 was significantly up-regulated, while the expression of CcChs2, CcChs7, and CcChs8 was significantly down-regulated ( Figure 8C). The qRT-PCR was used to measure the expression of chitin synthase-encoding genes in the wildtype, ΔCcbZIP05 and ΔCcbZIP23 mutants, and complemented strains. The CcActin gene was used as the reference gene. The experiments were repeated three times. The data were analyzed using one way ANOVA and Duncan's range test. The asterisks indicate significant differences (p < 0.05).

CcbZIP05 and CcbZIP23 Are Involved in Pathogenicity
To explore whether the functions of CcbZIP05 and CcbZIP23 are related to the path- qRT-PCR was used to measure the expression of chitin synthase-encoding genes in the wild-type, ∆CcbZIP05 and ∆CcbZIP23 mutants, and complemented strains. The CcActin gene was used as the reference gene. The experiments were repeated three times. The data were analyzed using one way ANOVA and Duncan's range test. The asterisks indicate significant differences (p < 0.05).

CcbZIP05 and CcbZIP23 Are Involved in Pathogenicity
To explore whether the functions of CcbZIP05 and CcbZIP23 are related to the pathogenicity of C. chrysosperma, we inoculated the wild-type, ∆CcbZIP05 and ∆CcbZIP23 mutants, and complemented strains on the poplar twigs. The lesion areas on poplar branches inoculated with the ∆CcbZIP05 and ∆CcbZIP23 mutants were significantly smaller than those inoculated with the wild-type and complemented strains at 6 dpi ( Figure 9A,B). In consideration of the reduced growth rate of ∆CcbZIP05 (29.9%) and ∆CcbZIP23 (51.7%), the lesion areas on poplar branches inoculated with the ∆CcbZIP05 and ∆CcbZIP23 mutants were reduced by about 42.3% and 68.9% compared to the wild-type and complemented strains, respectively. Therefore, the results indicated that the reduced virulence of ∆CcbZIP05 and ∆CcbZIP23 mutants might partly be resulted from the defects in fungal growth. The asterisks on the bars indicate a significant difference between CcbZIP05 deletion mutants, CcbZIP23 deletion mutants, and the wild-type (p < 0.05). In each experiment, 20 healthy poplar twigs were inoculated with wild-type, gene deletion mutants, and complementary strains, respectively. The experiment was repeated three times.

CcbZIP05 and CcbZIP23 Differentially Regulate the Expression of Putative Effector Genes
Transcription factors can regulate the expression of downstream targets to activate or suppress their functions. Our previous studies found that the CcPmk1 and the Gti1/Pac2 transcription factor CcSge1 could regulate the expression of putative effector genes [48,49]. Therefore, we calculated the expression of putative effector genes in ΔCcbZIP05 and Figure 9. Pathogenicity test of ∆CcbZIP05 and ∆CcbZIP23 mutants on poplar twigs. (A) Infection symptoms on detached poplar twigs inoculated with the wild-type, gene deletion mutants, and complemented strains. The inoculated twigs were photographed at 6 dpi. CK represents twigs inoculated with PDA plugs after scalding. (B) Lesion areas were determined on the inoculated twigs. The asterisks on the bars indicate a significant difference between CcbZIP05 deletion mutants, CcbZIP23 deletion mutants, and the wild-type (p < 0.05). In each experiment, 20 healthy poplar twigs were inoculated with wild-type, gene deletion mutants, and complementary strains, respectively. The experiment was repeated three times.

CcbZIP05 and CcbZIP23 Differentially Regulate the Expression of Putative Effector Genes
Transcription factors can regulate the expression of downstream targets to activate or suppress their functions. Our previous studies found that the CcPmk1 and the Gti1/Pac2 transcription factor CcSge1 could regulate the expression of putative effector genes [48,49]. Therefore, we calculated the expression of putative effector genes in ∆CcbZIP05 and ∆CcbZIP23 mutants including the CcCAP1 (Cysteine-rich secretory proteins, Antigen 5, and Pathogenesis-related 1 protein), glycoside hydrolase genes (GME1202_g, GME5006_g, GME6980_g, GME10300_g), and Nepl-like gene (NLP GME2220_g). As shown in Figure 10, the expression of GME2220_g, GME5006_g, and GME10300_g was significantly decreased in ∆CcbZIP23 mutants, but they showed comparable expression level among wild-type, ∆CcbZIP05 mutants, and complemented strains. However, the expression of CcCAP1 and GME6980_g was significantly up-regulated in the ∆CcbZIP23 mutants compared to the wildtype. Intriguingly, a different expression pattern of GME1202_g was found in ∆CcbZIP05 and ∆CcbZIP23 mutants, which was significantly down-regulated in the ∆CcbZIP05 mutants but significantly up-regulated in the ∆CcbZIP23 mutants ( Figure 10). ∆CcbZIP05 and ∆CcbZIP23 mutants, which was significantly down-regulated in the ∆CcbZIP05 mutants but significantly up-regulated in the ∆CcbZIP23 mutants ( Figure 10).

Discussion
Transcription factors (TFs) regulate the expression of downstream genes and are involved in a variety of key cellular functions. They are considered to be important components of the signal transduction pathway and are the crucial link between signal flow and the expression of the target gene. The bZIP TFs are one of the largest and the most diverse TF families, which are widely and exclusively distributed in eukaryotes. Many key biological processes require bZIP TFs, such as various stress responses, fungal growth, primary and secondary metabolism, and pathogenicity of phytopathogenic fungi [9,46,63]. However, no available data of bZIP TFs in C. chrysosperma were reported. Hence, in this study, we systematically identified a comprehensive set of 26 bZIP TF members in the C. chrysosperma genome, and functionally characterized the roles of CcbZIP05 and CcbZIP23 in the fungal growth, stress responses, and pathogenicity, which will provide insights for functional research of other bZIP genes in C. chrysosperma. Figure 10. CcbZIP05 and CcbZIP23 differentially regulate expression of putative effector genes. The qRT-PCR was used to measure the expression levels of putative effector genes in the wild-type, ∆CcbZIP05 and ∆CcbZIP23 mutants, and complemented strains. The CcActin gene was used as the reference gene. The experiments were repeated three times. The data were analyzed using one-way ANOVA and Duncan's range test. The asterisks indicate significant differences (p < 0.05).

Discussion
Transcription factors (TFs) regulate the expression of downstream genes and are involved in a variety of key cellular functions. They are considered to be important components of the signal transduction pathway and are the crucial link between signal flow and the expression of the target gene. The bZIP TFs are one of the largest and the most diverse TF families, which are widely and exclusively distributed in eukaryotes. Many key biological processes require bZIP TFs, such as various stress responses, fungal growth, primary and secondary metabolism, and pathogenicity of phytopathogenic fungi [9,46,63].
However, no available data of bZIP TFs in C. chrysosperma were reported. Hence, in this study, we systematically identified a comprehensive set of 26 bZIP TF members in the C. chrysosperma genome, and functionally characterized the roles of CcbZIP05 and CcbZIP23 in the fungal growth, stress responses, and pathogenicity, which will provide insights for functional research of other bZIP genes in C. chrysosperma.
Since the lifestyle of fungi heavily depends on their adaptations to their environment, replicated genes may lead to new functions and improve adaptations in a changing environment. In pathogenic fungi, improved nutrient uptake and more efficient catabolism, resistance, and adaptation to host infection can also be obtained through the expansion of gene families [64]. The bZIP superfamily is reported to have evolutionarily evolved from a single eukaryotic gene ancestor and has experienced multiple independent expansions [60]. Phylogenetic analysis of CcbZIPs and other selected fungal bZIP TFs showed that they were evenly distributed in eight clades (A-H), suggesting the bZIP gene family was present when these fungi were undifferentiated. Only the leucine zipper region of CcbZIP23 and CcbZIP26 lacked the following heptad repeats of the leucine (L), and the core arginine (R) had a phase 1 intron (Figures 1 and 3). In eukaryotes, the evolution of introns may have a role in the functional evolution of paralogs, and more introns usually imply more complex regulation [58,59]. Genetic structure analysis revealed differences in the number of introns in CcbZIPs (Figure 2). The CcbZIP genes contained 0 to 3 introns, but the maximum number of introns was lower than that of C. minitans (max. six introns) and U. virens (max. four introns). It has been shown that introns are lost faster than they are gained after segmental replication [65], indicating a putative gene duplication event happened in a recent period in C. chrysosperma compared to the C. minitans and U. virens. In addition, eight bZIP domains were inserted by introns in the N-X 7 -R region. Similar results were found in the bZIP domains of C. minitans and U. virens [23,25].
Previous studies showed that some bZIP genes were required for host infection in M. oryzae, and they also exhibited increased expression levels in the infection stages, such as MoATF1, MoHAC1, MoAP1, MoBZIP10, and MoMETR [22,38]. According to our expression profile data, 15 CcbZIP genes showed significant changes at the early infection stages, indicating the involvement of these genes in pathogenicity ( Figure 4). Additionally, some of their homologous genes are involved in fungal virulence. For example, overexpression of meaB (CcbZIP02 homologue) can reduce the production of aflatoxin B1 and thus attenuate the virulence of Aspergillus flavus [66]. VdAtf1 (CcbZIP03 homologue) is involved in the virulence of V. dahliae by mediating nitrogen metabolism [67]. VdHapX (CcbZIP07 homologue) controls iron homeostasis and is also crucial for the virulence of V. dahliae [68]. AaMetR (CcbZIP20 homologue, a methionine biosynthesis regulator) contributes to virulence and oxidative stress tolerance in Alternaria alternata [69]. However, the CcbZIP genes that did not differentially express during the initial infection stages might also contribute to the fungal virulence. For example, the CcbZIP25 homolog gene Yap1 is required for fungal virulence and stress response in C. gloeosporioides and M. oryzae [70,71], but the Bap1 (Yap1 homolog) has no impact on pathogenicity and differentiation in B. cinerea [30]. In this study, CcbZIP05 and CcbZIP23 were involved in the fungal virulence of C. chrysosperma. The ∆CcbZIP05 and ∆CcbZIP23 mutants showed significantly reduced lesion sizes on poplar twigs compared to the wild-type. However, previous studies showed that deletion of VDAG_08640 (CcbZIP05 homologue) in V. dahliae showed no significant defects in fungal growth, stress responses, pathogenicity, and microsclerotia formation [72]. Moreover, similar results were found in MGG_07305 (CcbZIP05 homologue) of M. oryzae [22]. Additionally, deletion of MGG_00587 (CcbZIP23 homologue) of M. oryzae also did not affect fungal growth, appressorium formation, and fungal pathogenicity, while significantly increased conidiation was observed in the MGG_07305 deletion mutants [38]. Here, we found that deletion of CcbZIP23 significantly compromised the conidiation. The results suggest that the bZIP orthologs in different fungal species may display distinct functions.
Pmk1 is a pathogenicity-related component of the MAPK signaling pathway in almost all the pathogenic fungi and plays crucial roles in fungal development, pathogenicity, and stress response through the regulation of other genes [48]. In Schizosaccharomyces pombe, the Pmk1 MAPK pathway is involved in cell wall integrity through regulating the Atf1 expression [73]. In M. oryzae, a bZIP gene, MGG_00587, was significantly down-regulated during early appressorium development (4 h) in the ∆pmk1 mutant compared to wildtype [74]. In C. chrysosperma, CcbZIP05 and CcbZIP23 were significantly down-regulated in ∆CcPmk1 ( Figure 5). Therefore, it is considered that Pmk1 can regulate the expression of downstream TF genes, including the bZIP TFs. Here, we found that the ∆CcbZIP05 and ∆CcbZIP23 mutants showed some similar phenotypes as the ∆CcPmk1 deletion mutant, such as the reduced fungal growth, conidiation, stress response, and increased branching. On the other hand, some different defects were observed in ∆CcbZIP23 mutants compared to the ∆CcPmk1 mutants. For example, CcbZIP23 deletion mutants accumulated obvious pigment in the plates, increased resistance to the CFW, CR, and sorbitol stresses, and reduced fungal virulence while the ∆CcPmk1 mutants were nonpathogenic [48]. These results suggest that CcbZIP05 and CcbZIP23 may also be regulated by other components in addition to CcPmk1.
Fungal secondary metabolites are essential in the competition, defense, and development of fungi [75,76]. Transcriptional regulation plays an important role in the biosynthesis of fungal secondary metabolites [77]. In Aspergillus flavus, the bZIP TF AflRsmA regulates Aflatoxin B 1 (AFB 1 , a potent carcinogen) biosynthesis through the oxidative stress pathway. Overexpression of AflrsmA increases AFB 1 production [46]. Similarly, the deletion of Afap1 (a bZIP TF) significantly reduces the production of AFB 1 [78]. In Aspergillus nidulans, activation of RsmA (a YAP-like bZIP gene) greatly increases the production of secondary metabolites through binding to the two sites of the AflR promoter region (a C6 TF involved in the production of the carcinogenic and anti-predation secondary metabolites, namely, sterigmatocystin) [79]. In this study, deletion of CcbZIP23 significantly increased the pigmentation accumulation in the plates ( Figure 6 and Figure S5). These results indicate that CcbZIP23 is involved in the secondary metabolism in C. chrysosperma, but the regulatory mechanism remains to be investigated.
Previous studies have shown that the fungal cell wall is composed mainly of chitin and glucan. Chitin synthesis is mainly mediated by chitin synthase. In plant pathogenic fungi, there are generally seven or eight chitin synthase-encoding genes [80,81]. In M. oryzae, three CHS genes (CHS1, CHS2, and CHS6) are involved in conidiation, pathogenicity, and stress responses [80]. In F. graminearum, ∆FgChs2 and ∆FgChs5 mutants (CHS2, CHS6 homologous gene, respectively) show significantly reduced mycelial growth and conidiation, but they show increased sensitivity to CR [81]. In Neurospora crassa, the chs-1 RIP mutant displays abnormal branching, swollen hyphal tips, and reduced growth rate; in transformants expressing CHS3-GFP or CHS6-GFP constructs, GFP signals are detected mainly at the hyphal tip, suggesting that they may be involved in polarized growth [82]. In our study, we found that the expression of chitin synthase-encoding genes was differentially regulated by CcbZIP05 and CcbZIP23, which may partly contribute to the defects in chitin distribution, fungal growth, conidiation, stress responses, and pathogenicity in ∆CcbZIP05 and ∆CcbZIP23 mutants.
It is well known that pathogens will deliver numerous effectors into plant cells or the apoplast to suppress host immunity and then promote the infection or colonization [83]. Many reports showed that the TFs could regulate the expression of downstream effector genes [84,85]. In this study, we calculated the expression of several putative effector genes in ∆CcbZIP05 and ∆CcbZIP23 mutants, and their homologues were found to be involved in the fungal pathogenicity such as the glycoside hydrolase family 12 [86], Nep1-like proteins [87,88]. We found that CcbZIP05 and CcbZIP23 could regulate the expression of glycoside hydrolase 12 family genes, NLP gene, and CcCAP1, which may partly contribute to fungal virulence. In conclusion, we identified a total of 26 bZIP TF family members in the C. chrysosperma genome and analyzed the functions of CcbZIP05 and CcbZIP23, which played important roles in fungal development, stress responses, and virulence of C. chrysosperma. The results provide a comprehensive view of the CcbZIP TF family and provide a basis for further studies on the function of other bZIP genes.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/10 .3390/jof8010034/s1. Figure S1: Distributions of conserved domains in CcbZIP genes. All CcbZIP genes manually confirmed the existence of domains by using the InterProScan on EBI web server (http://www.ebi.ac.uk/interpro/, accessed on 26 December 2019). Different domains are highlighted in different color. Figure S2: Phylogenetic analysis of bZIP members from nine fungi. The maximum likelihood trees were constructed based on 181 full-length protein sequences. The phylogenetic tree was constructed using raxmlGUI1.3 software and the bootstrap test carried out with 1000 iterations. The bZIP members are clustered into eight clades (A-H) indicated by colored branches. The CcbZIPs are denoted by red circles. The sequences were collected from organisms as follows: Cc, Cytospora chrysosperma; Bc, Botrytis cinerea; Cg, Colletotrichum gloeosporioides; Cp, Cryphonectria parasitic; Fg, Fusarium graminearum; Mo, Magnaporthe oryzae; Sc, Saccharomyces cerevisiae; Ss, Sclerotinia sclerotiorum; Vd, Verticillium dahliae. Figure S3: Construction and confirmation of CcbZIP05 and CcbZIP23 disrupted and complemented mutants. (A) The abridged general view of the generation of deletion mutants for the CcbZIP05 gene. (B) The abridged general view of the generation of deletion mutants for the CcbZIP23 gene. Southern blot analysis was conducted to determine the single integration events, which are indicated on each schematic map. Figure S4: Colony morphologies and pycnidia formation. (A) The colony morphologies and pycnidia formation of the wild-type, ∆CcbZIP05 mutants, and complemented strains after 30 days of growth on PDA plates. (B) Quantification of pycnidia production in each strain. Figure S5: Colony morphologies of wildtype, ∆CcbZIP23 mutants, and complemented strain grown on the PDA supplemented with cell wall interfering agents (CFW and SDS) and osmotic interfering agents (NaCl, KCl, and sorbitol) at 30 days. Table S1: Primers used in this study. Table S2: Sequence information of bZIP genes in this study. Table S3: Homologous genes of CcbZIPs in the PHI database.