The WT strains of A. cristatus (CGMCC 7.193) used in this study were isolated from Fuzhuan brick tea produced by the Yiyang Tea Factory in Yiyang, China. The strains were grown on low-concentration sodium chloride solid MYA media (20 g of malt extract, 5 g of yeast extract, 30 g of sucrose, 50 g of sodium chloride, and 1,000 mL of water) at 28°C to induce sexual development and high-concentration sodium chloride solid media (20 g of malt extract, 5 g of yeast extract, 30 g of sucrose, 170 g of sodium chloride, and 1,000 mL of water) at 37°C to induce asexual development. SD/−Trp, SD/−Leu, SD/−Trp/−Leu, SD/−Trp/− Leu/-His and SD/−Trp/−Leu/-His/−Ade (purchased from Beijing Cool Lab Technology, China). The colonies were imaged using a Canon EOS 7D Mark II camera (Canon, Tokyo, Japan).

The Achog1 gene fragment was ligated into the expression vector pGEX-6p-1, and GST fusion protein expression and transformation were performed. Then, the total proteins of A. cristatus cultivated under normal osmotic stress and hypertonic stress were extracted, and pull-down assays were carried out. The proteins were subsequently detected via LC–MS/MS. Proteins were verified by yeast two-hybrid assays. To facilitate subsequent analysis, the samples of A. cristatus cultivated under normal stress and hypertonic stress were named WT-D and WT-H, respectively.

Phylogenetic analysis was performed using MEGA 6.06 software with the conserved domains of homologous AcHog1 proteins (Tamura et al., 2013). ClustalW was employed for multiple sequence alignment with the default values. A maximum likelihood with a bootstrap value of 1,000 was used to generate a phylogenetic tree.

Homologous recombination was employed to delete the whole open reading frame (ORF) of Achog1. An Achog1 deletion cassette containing hph as a selective marker was constructed via the fusion of the 5′-untranslated region (5’-UTR) and the 3′-untranslated region (3’-UTR). Briefly, the 5’-BamHI-XhoI UTR and 3’-SpeI-XbaI UTR were amplified using specific primer pairs (Supplementary Table S1) from genomic DNA of the WT strain and cloned into the pDHt/sk-hyg cloning sites to generate the final Achog1-L-pDHt/sk-hyg-Achog1-R knockout vector. To complement the Achog1-null mutant, the Achog1 gene with its own promoter was amplified from the genomic DNA with the primer pair qc-hog1-F/R (Supplementary Table S1) and then inserted into pDHt/sknt at the enzyme sites HindIII/KpnI.

The plasmids were confirmed via PCR, restriction enzyme digestion and sequencing. The knockout plasmids were transformed into the WT strains, and complementary plasmids were transformed into the ΔAchog1 knockout strains by Agrobacterium tumefaciens-mediated transformation (ATMT), as previously described (Tan et al., 2018). The transformants were selected on MYA media that included 50 ug/mL hygromycin B or 80 ug/mL geneticin (G418) and were confirmed via PCR (the primers used are listed in Supplementary Table S1).

The transgene copy number in ΔAchog1 was estimated by using RT-PCR based on the methods of Song et al. (2002). Using the genomic DNA of ΔveA, a standard curve was established, in which the copy number was determined by Southern blot analysis (Tan et al., 2018). The genes encoding glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and HYG were selected from the A. cristatus genome as candidate reference genes (the primers used are listed in Supplementary Table S1).

RNA-seq was performed on the ΔAchog1 and WT strains. A conidial suspension (concentration 1 × 10⁶ conidia/mL) was used for inoculation. Three biological replicates were grown on MYA media comprising 17% NaCl at 37°C. The WT strain produced a large number of conidial heads, but the ΔAchog1 knockout strain did not after 39 h of culture. Based on these observations, all the samples were collected from the cellulose membrane, flash frozen in liquid nitrogen and stored at-80°C.

Total RNA was extracted as described previously using TRIzol (Gilbert et al., 2016). An Agilent Bioanalyzer 2,100 was subsequently used to assess the concentration and quality of the total RNA. mRNA libraries were prepared as described previously (Parkhomchuk et al., 2009). The quality of the final cDNA libraries was verified using an Agilent Bioanalyzer 2,100. The libraries were sequenced using an Illumina NovaSeq 6,000 (Illumina, United States). After filtering the raw data and checking the sequencing error rate and GC content distribution, the clean reads used for the subsequent analysis were obtained. HISAT2 software was used to quickly and accurately align the clean reads with the reference genome to obtain the positioning information. Differentially expressed genes were analysed using DESeq2 (Love et al., 2014); gene ontology (GO) functional enrichment and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway enrichment of the differentially expressed genes were performed by using ClusterProfiler software (Kanehisa and Goto, 2000; Young et al., 2010).

Total RNA was extracted at the tested time point. Then, 2 mg of RNA was utilized to synthesize cDNA using a RevertAid First Strand cDNA Synthesis Kit (Thermo Fisher Scientific, catalogue#K1622). RT-qPCR was performed using a CFX96 Real-Time PCR Detection System (Bio-Rad Laboratories, Hercules CA, United States) in a total volume of 10 μL, which consisted of 5 μL of SsoFast EvaGreen SuperMix (Bio-Rad, catalogue# 172–5,201), 1 μL of each primer (10 pmol/ml) and 1 μL of template. GAPDH was selected as a candidate reference gene. The primers employed in the RT-qPCR experiments were designed using the Primer 3 online program,¹ and the resulting RT-qPCR products were tested via agarose gel electrophoresis. Each primer pair was tested with serial dilutions of cDNA to determine the linear range of the RT-qPCR assays. Three biological replicates were analysed. All the RT-qPCR primers used are listed in Supplementary Table S1.

The homolog of hog1, SI65-07698, was identified based on the genome of A. cristatus (accession number [JXNT01000000]) (Ge et al., 2016). SI65_07698 is 1,689 bp in length, includes six introns and encodes a predicted protein of 366 amino acids showing 98% identity to S. cerevisiae Hog1 and the stress-activating kinase SakA of A. nidulans. The results of multisequence alignment revealed that Hog1 was very conserved across Aspergillus species (Supplementary Figure S1A). The predicted SI65_07698 protein was found to contain a Pkc-like superfamily domain (Supplementary Figure S1C), which was consistent with the conserved domain of hogA in A. nidulans. SI65_07698 was named Achog1 based on these analysises. It is worth noting that halophilic Aspergillus contains 3 Hog1 homologs, which are differentially regulated during different salinity conditions (Rodríguez-Pupo et al., 2021). A. cristatus does not contain a second Hog1 homolog, as it occurs in most aspergilli (Kawasaki et al., 2002; Garrido-Bazán et al., 2018). Achog1 deletion mutants were obtained by gene replacement with the hygromycin B phosphotransferase gene (hph) as a selective marker to verify the function of Achog1 in A. cristatus. Six ΔAchog1 transformants were ultimately obtained (Supplementary Figure S2B shows five of the transformants). The expression of Achog1 in ΔAchog1 strain was detected and the results showed that Achog1 was not expressed in ΔAchog1 strain, which further indicated that Achog1 had been successfully knocked out (Supplementary Figure S2D, the last column). Real-time fluorescent PCR showed that the numbers of hph in ΔAchog1 were single copies (Supplementary Figure S3 and Supplementary Table S2). Eighteen ΔAchog1-C strains with the WT-like phenotype were verified by PCR. The results are shown in Supplementary Figure S2C.

Aspergillus cristatus could produce pure sexual/asexual spores under hypotonic and hypertonic stress. It was speculated that there was an association between osmotic stress and sporulation. Therefore, the sexual and asexual development of WT and ΔAchog1 was observed under different osmotic stresses by inoculating an equal amount of conidia (concentration: 1 × 10⁶ conidia/mL). When hyphae of the WT strain formed conidiophores, most hyphae of ΔAchog1 only formed vesicles at 39 h on MYA media containing 3 M NaCl (Figures 1A,B, indicating that deleting Achog1 delays the germination of conidia. A large number of grey-green conidia were produced by WT, while only a small number of conidia was produced by ΔAchog1 in the center of the colony as the incubation time increased (Figure 1C). The conidial number produced by WT was nearly 2 times that produced by the ΔAchog1 strain (Figure 1D). No significant difference was found in sexual development between WT and ΔAchog1 on hypotonic media (Figure 1C: the first column). However, hyphal curling appeared in ΔAchog1 after culture with 3 M NaCl for 7 days, and a small amount of golden yellow cleistothecia appeared in the ΔAchog1 strain after culturing for 14 days (Figure 2), which was an interesting phenomenon. These results indicated that Achog1 positively regulated asexual sprouting and might inhibit sexual development under hypertonic stress in A. cristatus.

It is well known that one of the main functions of hog1 is to respond to hyperosmolarity in other fungi. The conidia of WT and ΔAchog1 (concentration: 1 × 10⁶ conidia/mL) were cultivated on MYA media with 3 M sorbitol and sucrose to confirm whether Achog1 is involved in hypertonic stress. The growth of ΔAchog1 was significantly slower than that of WT on MAY media with 3 M sucrose, and ΔAchog1 also grew slowly on media with 3 M NaCl and sorbitol (Figures 3A,B). Furthermore, the expression of genes related to hypertonic stress was detected. The results showed that the expression of genes related to hypertonic stress in ΔAchog1 was lower than that of in WT under hypertonic stress (Figure 3D). Fewer conidia were produced by ΔAchog1 (Figure 3C). The results of sensitivity tests showed that Achog1 could positively regulate the response to hypertonic stress, and the sensitivity of ΔAchog1 to different osmotic stress regulators was in descending order: 3 M sucrose >3 M NaCl >3 M sorbitol.

A study showed that hog1 mediated the oxidative stress pathway in C. albicans (Thomas et al., 2013). The conidia of WT and ΔAchog1 (concentration: 1 × 10⁶/mL) were cultivated in MYA with different concentrations of H₂O₂ (0 mM, 10 mM, 30 mM, and 50 mM) to verify whether Achog1 was involved in oxidative stress. The colony diameter of ΔAchog1 decreased, and its edges were irregular (Figures 4A,B), suggesting that the tolerance of ΔAchog1 to oxidative stress was significantly reduced. We detected the expression of genes related to oxidative stress. The results showed that the expression of genes related to oxidative stress in ΔAchog1 was lower than that of in WT under oxidative stress (Figure 4C). When the Achog1 gene was complemented, the sensitivity to H₂O₂ was essentially restored. These results indicated that Achog1 positively regulated the response to oxidative stress.

It was shown that the ΔAfhog1 deletion mutant strain grew slower than the wild type strain in medium containing high PH in A. fumigatus (Ma et al., 2012). To test whether Achog1 was involved in PH stress, equal amounts of conidia (1.0 × 10⁶ conidia/mL) of WT, ΔAchog1, and ΔAchog1-C were inoculated on MYA media containing 5% NaCl with different pH values (pH 3, pH 4, pH 5, pH 8, and pH 11). ΔAchog1 grew normally in media at pH 3–5, its growth was just slower than that of the WT and ΔAchog1-C strains after culturing for 6 days at 28°C. However, the growth of ΔAchog1 significantly slowed down in the media at pH 8, and ΔAchog1 could not grow at pH 11 (Figures 5A,B). Hyphae samples cultured in the medium with pH 8 were collected to detect the expression of genes related to pH stress. The results showed that the expression of genes related to pH stress in ΔAchog1 was lower than that of in WT (Figure 5C); indicating that Achog1 is mainly involved in alkali stress.

The colonies of the WT strain were brown-black in the center and yellow at the edge on MYA media. However, the colonies of the ΔAchog1 strain were brown, and their edges were pale yellow. The pigments produced by the ΔAchog1-C strains were similar to those produced by the WT (Figure 6A). Genes involved in pigment synthesis, such as SI65_05588 (ayg1), SI65_05589 (arp2), SI65_05591 (abr1) and SI65_05592 (abr2), were downregulated according to RNA-seq data (Figure 6B). Furthermore, we detected the expression of four genes related to pigment synthesis. The results showed that the expression of these genes in ΔAchog1 strain was lower than that of in WT (Figure 6C). These results suggested that the Achog1 gene played a role in regulating pigment synthesis in A. cristatus.

A pull-down mass spectrometry experiment was performed to determine the proteins in WT-D and WT-H (Figure 7A). The total ion chromatogram and evaluation chart of the identified proteins showed that the mass spectrometry data were normal, indicating that the identified proteins were accurate (Supplementary Figure S4). A total of 940 proteins and 2,980 peptides were identified (Supplementary Tables S8, S9). A total of 530 proteins were upregulated, and 291 proteins were downregulated (proteins with a fold change (FC) ≥ 2 were upregulated, and those with an FC ≤ 1/2 were downregulated; Figure 7B). Comparing WT-H with WT-D, 483 were present only in WT-H, and 251 were present only in WT-D, demonstrating that there were significant differences in the proteins interacting with AcHog1 between the normal stress and hypertonic stress conditions.

The top 20 proteins with the highest scores are listed in Supplementary Table S3. Among them, 5 proteins are uncharacterized, and the remaining 15 proteins are known functions. For example, the protein A0A1E3B957 is 6-phosphogluconate dehydrogenase. It has been reported that 6-phosphogluconate dehydrogenase is mainly involved in the salt stress response in plants (Nemoto and Sasakuma, 2000; Huang et al., 2003). AcHog1 interacted with 6-phosphogluconate dehydrogenase in samples incubated in MYA containing 3 M NaCl, suggesting that AcHog1 could regulate 6-phosphogluconate dehydrogenase to participate in the salt stress response in A. cristatus. RodA/RodB-related conidia (Pedersen et al., 2011), Ste20 and Cla4 located upstream of the HOG1 pathway were also found in WT-H (Dan et al., 2001; Joshua and Höfken, 2019). The data proved AcHog1 could interact with 6-phosphogluconate dehydrogenase and might be involved in the salt stress response. In addition, Achog1 could interact with RodA/RodB, Ste20 and Cla4 and Achog1 might be also involved in the production of conidia and the response of HOG pathway.

GO enrichment analysis showed that the differentially expressed proteins were mainly involved in various metabolic processes, asexual development processes and stress responses (Figure 7C). Nine proteins (Supplementary Table S4) related to asexual sporulation and stress responses were randomly selected for yeast two-hybrid validation. PGBKT7-AcHog1 and PGADT7 plasmids containing 9 proteins were co-transformed into yeast AH109 receptive cells. After 5 days of culturing in SD/−Leu/−Trp, a single colony was cultured on SD/−Trp/−Leu/-His and SD/−Trp/−Leu/-His/−Ade medium containing X-α-Gal and AbA for 5 days. PGBKT7-53 + PGADT7-T was chosen as a positive control (Figure 8A), and PGBKT7-lam + PGADT7-T was a negative control (Figure 8G). Both the colonies in the positive control and in the experimental groups (Figures 8B–F,H–K) were blue, indicating that all the selected proteins could interact with AcHog1 (Figure 8). These results showed that AcHog1 might regulate asexual sporulation and the response to osmotic stress and oxidative stress in A. cristatus. KEGG enrichment analysis showed that 14 differentially expressed proteins were enriched in the MAPK pathway and were significantly upregulated (Figure 7D). The AA1E3BER4 (AcSko1) protein is located downstream of AcHog1 in the HOG pathway in WT-H, and its interaction with AcHog1 was verified by using yeast two-hybrid assays (Figures 8B,K). Therefore, we speculated that Sko1 was the target protein of AcHog1, which will be a focus of our future studies.

To investigate the roles of Achog1 in A. cristatus, RNA-seq of ΔAchog1 and WT was performed at selected time points. Conidiophores and chain-like conidia appeared on the WT strain; however, only some vesicles were formed in ΔAchog1, when the ΔAchog1 and WT strains were cultivated on MYA media containing 3 M NaCl for 39 h (Figure 1A). Therefore, mycelial samples of the cultures at 39 h were collected for RNA-seq. All sequences obtained by sequencing were uploaded to the NCBI database: Sequence Read Archive,² and get the accession number: PRJNA877470. A summary of the RNA-seq data is reported in Supplementary Table S5. In this study, a total of 1,074 differentially expressed genes (Supplementary Table S10) were obtained from the Achog1 transcriptome data (|log2(FC)| > = 1 and padj < = 0.05). Among these differentially expressed genes, 480 were upregulated, and 594 were downregulated. The top 20 genes upregulated and downregulated in ΔAchog1 are listed in Supplementary Tables S6, S7, respectively. Among them, SI65_00200 (phiA) is a cell wall protein (Schachtschabel et al., 2012); SI65_01919 encodes a low-affinity glucose transporter, and a study showed that low-affinity glucose required the coordinated activities of the HOG and glucose signaling pathways in S. cerevisiae (Tomás-Cobos et al., 2004). The protein kinases Dsk1 (SI65_03917) contributed to cell cycle and pre-mRNA splicing (Tang et al., 2012). In addition, genes responding to the HOG pathway and related to sporulation were screened from the RNA-seq data by referring to the relevant literature reports and the genome database of Aspergillus. The results were shown in Table 1. It was noted that genes related to asexual development were downregulated, such as brlA (SI65_02778) and wetA (SI65_00383) (Tsai et al., 1999), were involved in asexual sporulation and downregulated in the ΔAchog1 mutant. We speculated that Achog1 could affect the conidia production of A. cristatus by down-regulating the expression of these genes.

GO enrichment analysis showed that the differentially expressed genes were mainly enriched in carbohydrate metabolism, transmembrane transport, hydrolase activity, binding and other processes. Carbohydrate metabolic process (GO: 0005975), hydrolase activity (GO: 0016798, GO: 0016798) and oxidoreductase activity (GO: 0016614) were significantly upregulated. Transmembrane transport (GO: 0055085) was significantly downregulated (Figure 9A). The differentially expressed genes were mainly enriched in various metabolic pathways and MAPK pathways by KEGG enrichment analysis. Galactose metabolism (ani00052) and amino sugar and nucleotide sugar metabolism (ani00520) were significantly upregulated, glycerophospholipid metabolism (ani00564) and glyoxylate and dicarboxylate metabolism (ani00564) were significantly downregulated (Figure 9B). In addition, all the genes enriched in the MAPK pathway were downregulated. We speculated that Achog1 also played an important role in hydrolase activity, various metabolic processes and MAPK pathways in A. cristatus.

To verify the reliability of the expression changes in the transcriptome, 6 genes were selected for RT-qPCR detection (Figure 9C). The gene expression patterns of the gene analysis were similar in the RT-qPCR and RNA-seq, indicating that the transcriptome sequencing results obtained in this study were reliable.