Insights into the mechanism of cyanobacteria removal by the algicidal fungi Bjerkandera adusta and Trametes versicolor

Abstract Fungal mycelia can eliminate almost all cocultured cyanobacterial cells within a short time. However, molecular mechanisms of algicidal fungi are poorly understood. In this study, a time‐course transcriptomic analysis of algicidal fungus Bjerkandera adusta T1 was applied to investigate gene expression and regulation. A total of 132, 300, 422, and 823 differentially expressed genes (DEGs) were identified at 6, 12, 24, and 48 hr, respectively. Most DEGs exhibited high endopeptidase activity, cellulose catabolic process, and transmembrane transporter activity by using Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) analyses. Many decomposition genes encoding endopeptidases were induced a little later in B. adusta T1 when compared with previously investigated algicidal fungus Trametes versicolor F21a. Besides, the accumulated expression of Polysaccharide lyases8 (PL8) gene with peptidoglycan and alginate decomposition abilities was greatly delayed in B. adusta T1 relative to T. versicolor F21a. It was implied that endopeptidases and enzymes of PL8 might be responsible for the strong algicidal ability of B. adusta T1 as well as T. versicolor F21a.

been reported to exhibit algicidal ability (Han et al., 2011;Shu et al., 2016;Wang et al., 2010;Zeng, Wang, & Wang, 2015;Zeng et al., 2019). Among these, T. versicolor F21a and B. adusta T1 were considered as the two best algicidal fungi (Dai et al., 2018;Han et al., 2011;Zeng et al., 2015Zeng et al., , 2019. Previous studies have reported that both living and dead cyanobacterial cells first adhere to fungal mycelia before being eliminated by surrounding mycelia (Dai et al., 2018;Jia et al., 2010). It has been further demonstrated that the membranes of cyanobacterial cells and the pyrrole ring of chlorophyll a were extensively disrupted by mycelia of P. chrysosporium (Zeng et al., 2015). Transcriptomic and proteomic analyses of the algicidal mechanism of T. versicolor F21a showed that several biological processes, such as glucan 1,4-α-glucosidase activity, hydrolase activity, lipase activity, and endopeptidase activity, and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways, including glycolysis/gluconeogenesis, pyruvate metabolism, starch and sucrose metabolism, and amino acids biosynthesis, are involved in the elimination cyanobacterial cells (Dai et al., 2018;Gao et al., 2017). The expression of all Carbohydrate-Active enZYmes (CAZyme) genes significantly increased during the algicidal process in T. versicolor F21a (Dai et al., 2018;Gao et al., 2017). Several members of CAZyme, such as AA5, GH18, GH5, GH79, GH128, and PL8, might play key roles in the decomposition of cyanobacterial cells at different eliminating stages (Dai et al., 2018).
Although the underlying molecular mechanism of algicidal fungus T.
versicolor F21a was elucidated, there are no reports on the mechanism of other efficient algicidal fungi.
B. adusta is a widely distributed "white rot" fungus, which has been often associated with the decomposition of hardwoods (Moody, Dudley, Hiscox, Boddy, & Eastwood, 2018). The components of wood cell walls, such as cellulose, hemicellulose, and recalcitrant lignin, can be degraded by this fungus (Moody et al., 2018).
Besides, this fungus has been reported to decompose a wide range of environmental pollutants (Bouacem et al., 2018;Han et al., 2011;Sugawara, Igeta, Amano, Hyuga, & Sugano, 2019). In our previous study, B. adusta T1 was found to be one of the best algicidal fungi (Han et al., 2011). In this study, gene expression in the mycelia of B. adusta T1, cocultivated with and without cyanobacterial cells during the algicidal process, was compared by a time-serial transcriptomic analysis. Differentially expressed genes (DEGs) were used to identify key decomposition gene(s) and pathway(s) in B. adusta T1, and the results were compared with that of T. versicolor F21a reported in a previous study (Dai et al., 2018).

| Fungal and algal strains
The previously isolated fungus B. adusta T1 from Zijinshan Mountain was used in this study (Han et al., 2011). Cyanobacterial strain (Microcystis aeruginosa PCC7806) was provided by the Institute of Hydrobiology of the Chinese Academy of Sciences (Wuhan, China).

| Cocultivation of fungal mycelia and cyanobacterial cells
The cyanobacterial strain was cultivated at 25°C under 12-hr light and 12-hr dark cycles with ~90 μmol/m 2 s -1 of photons in BG-11 medium . Round fungal mycelium (seven mm in diameter) was inoculated onto a nine-cm plate, containing 15 ml of potato liquid medium, and incubated under static conditions for five days. Then, fungal mycelia were taken and transferred into 250-mL Erlenmeyer flasks containing 100 ml of algal solution or medium.
The cocultures were incubated at 25°C, 90 μmol photons/m 2 s -1 , and F I G U R E 1 Changes in the algicidal process of B. adusta T1. Note: (a) Images of cocultivation after 48 hr; CK, the cyanobacterial cells as control;T1, the cocultivation of cyanobacterial cells and B. adusta T1 mycelia; S-T1, the cocultivation of cyanobacterial cells and died fungal mycelia. (b) Changes in chlorophyll a content during the algicidal process 120 rpm to investigate differentially expressed fungal genes.

| RNA isolation and sequencing
Mycelia of B. adusta T1 were collected from cocultures after 6, 12, 24, and 48 hr of incubation. Two biological replicates of each treatment were used for RNA sequencing. Total RNA was extracted from each sample with TRIzol reagent following the manufacturer's instructions (Takara, Dalian, China). Then, crude RNA was digested via 10 U DNase I (TaKaRa, Japan) at 37°C for 30 min, and then, mRNA was isolated using Dynabeads® Oligo (dT) 25 (Life, America) following the manufacturer's instructions.
One hundred ng mRNA of each sample was used to construct a sequencing library using NEBNext® Ultra TM RNA Library Prep Kit (NEB, America). Paired-end sequencing of cDNA fragments (~300 bp) was performed using Illumina HiSeq 4,000 platform at BGI-Shenzhen, China.

| Quantitative PCR (qPCR) validation
qPCR was used to validate the gene expression calculated from RNA-Seq data. A few randomly selected lignocellulose-active enzyme genes were used in this study, and the β-actin gene of B.

F I G U R E 2
Multi-dimensional scaling of gene expression data. Note: 6h_ck, control sample at 6h; 6h_T, treatment sample at 6 hr; 12h_ck, control sample at 12 hr; 12h_T, treatment sample at 12 hr; 24h_ck, control sample at 24 hr; 24h_T, treatment sample at 24 hr; 48h_ck, control sample at 48 hr; 48h_T, treatment sample at 48 hr F I G U R E 3 Number of fungal DEGs during the algicidal process of B. adusta T1

| Elimination rate during the algicidal process
The algicidal process of B. adusta T1 was monitored via spectrophotometer. As shown in Figure

| RNA-Seq data generation and mapping
Mycelia of B. adusta T1 that was cocultivated with cyanobacterial cells at 6, 12, 24, and 48 hr were used for RNA sequencing.
Fungal mycelia without cyanobacterial cells at the same time point were used as a control. Good quality RNA was isolated and used for RNA sequencing ( Figure A1). A total of 63,437,015 pairs of raw reads (SRA accession: PRJNA543936) were generated (Table A2).
Approximately 96% of reads were retained after the removal of adaptor and low-quality bases (Table A2). More than 64% of reads were uniquely mapped to the reference genome by pipeline STAR (Table A2), suggesting that the results of mapping can be used for the identification of fungal DEGs.  (Table A1).

| Identification of fungal DEGs involved in the algicidal process
Similar expression patterns were observed between qRT-PCR and transcriptomic analysis ( Figure A3), indicating that DEGs identified by the transcriptomic analysis were suitable for further analyses.

| Annotation and enrichment analyses of fungal DEGs
After the comparison of candidate genes with Nr from NCBI, GO, and KEGG databases, DEGs were used to obtain enriched terms by Fisher's F I G U R E 4 GO term enrichment of fungal DEGs in the cellular component category exact test (p < .05). The GO terms of DEGs were enriched in the extracellular region, cell wall, signal recognition particle, proteasome core complex, prefold in complex, ribosome, and other cellular components categories ( Figure 4). Similarly, DEGs were found to be enriched on transport and catabolic processes in the biological process category, particularly cellulose catabolism and carbohydrate transport ( Figure 5).
Further, DEGs were enriched on decomposition and transporter activities in the molecular function category that included the activities of triglyceride lipase, serine-type peptidase, manganese peroxidase, carboxypeptidase, cellulose 1,4-β-cellobiosidase, β-glucosidase, aspartictype endopeptidase, α-amylase, glycolipid transporter, amino acid transmembrane transporter, and other ( Figure 6). The KEGG analysis showed that DEGs were enriched on glycerolipid metabolism, starch and sucrose metabolism, metabolism of xenobiotics by cytochrome P450, galactose metabolism, and ascorbate and aldarate metabolism in different stages of the algicidal process ( Figure 7).

| Composition and expression of CAZyme genes of B. adusta T1 and its comparison with that of T. versicolor F21a
A total of401 CAZyme genes were identified in the genome of B. adusta by hmmscan against the dbCAN database ( Table 1). The F I G U R E 5 GO term enrichments of fungal DEGs in the biological process category lignocellulose-active genes can be divided into 77 CAZyme modules ( Table 1). Most of the genes belonged to Glycoside Hydrolases (GH) family and Auxiliary Activities (AA) family. About 312 CAZyme genes were identified in the genome of T. versicolor F21a (Dai et al., 2018). The number of CAZyme genes in B. adusta T1 genome (401 CAZyme genes) was higher than that of T. versicolor F21a (312 CAZyme genes). Seventy CAZyme modules were detected in B. adusta T1, compared to 43 CAZyme modules in T. versicolor F21a in the previous study (Dai et al., 2018). However, the algicidal effects of T. versicolor F21a were slightly more efficient than that of B. adusta T1 (Han et al., 2011).
The enzymes of GH128, AA7, AA6, and GH109 were less efficient in cyanobacterial cell disruption. It is noteworthy that the accumulated expression of Polysaccharide lyases genes, particularly the PL8 module was highly up-regulated during the later stage of the algicidal process of B. adusta T1, which was much delayed when compared to T. versicolor F21a (Dai et al., 2018).

| Expression of other decomposition genes in B. adusta T1 and their comparison with that of T. versicolor F21a
Only a few serine-type peptidase, carboxypeptidase, and aspartictype endopeptidase, with strong ability in cyanobacterial cells disruption, were enriched in the DEGs list during the early stage of the algicidal process (6 hr) ( Figure 6). However, no strong decomposition enzyme was enriched during the later stage of the algicidal process until 24 hr ( Figure 6). During the later stage (24 hr), proteins F I G U R E 6 GO term enrichments of fungal DEGs in the molecular function category with aspartic-type endopeptidase activity and manganese peroxidase activity were the main decomposition enzymes ( Figure 6).
Various types of decomposition enzymes, such as threonine-type endopeptidase and serine-type endopeptidase, were induced after 48 hr of cocultivation. In this study, proteases with Protein ID jgi|Bjead1_1|36244|fgenesh1_kg.4_#_443_#_Locus8459v1_ medCvg1568.9s and jgi|Bjead1_1|342083|CE153752_10262, and jgi|Bjead1_1|110676|e_gw1.8.836.1 were observed to be the main degradation genes that might be involved in cyanobacterial cells disruption ( Figure 9). Thus, these proteases can play significant roles in the algicidal process. The decomposition genes showed delayed expression compared with that of T. versicolor F21a.

| D ISCUSS I ON
Although several fungi showed a strong algicidal activity (Han et al., 2011), the underlying molecular mechanisms for algicidal capacities are largely less investigated. Interestingly, a few fungi from the Polyporales order of Basidiomycota exhibited a strong algicidal activity (Han et al., 2011). Comparative genome analyses found that the genomes of white rot fungi contain more genes encoding plant cell wall degrading enzymes than that of brown rot and mycorrhizal fungi (Kohler et al., 2015;Tisserant et al., 2013). White rot fungi including the order Polyporales can degrade lignin as well as cellulose (Kohler et al., 2015). In the present study, we observed that the number of CAZyme genes and expressed CAZyme genes of B.
adusta T1 was great than that of T. versicolor F21a. However, the algicidal effects of B. adusta T1 were slightly less efficient than that High lignocellulose degradation ability of white rot fungi, in comparison with that of brown rot fungi and mycorrhizal fungi, can be attributed to the number of genes encoding plant cell wall degrading enzymes in fungal genomes as a result of long term natural selection (Kohler et al., 2015). The numbers of CAZyme genes were not directly correlated with algicidal abilities, which might be due to the fact that most algicidal fungi were isolated from terrestrial environments and lacked evolution selection pressure in the water system (Han et al., 2011).
Direct contact between fungal mycelia and cyanobacterial cells was required for eliminating cyanobacterial cells by fungi (Han et al., 2011;Jia et al., 2010). Previous studies showed that a few decomposition enzymes might play important roles in eliminating cyanobacterial cells by T. versicolor F21a. In particular, cellulase, β-glucanase, and protease were supposed to efficiently disrupt cyanobacterial cells by T.
versicolor F21a (Dai et al., 2018;Gao et al., 2017  MlrA encoding a protease, fungal aflatoxin-detoxifizyme could be a possible candidate enzyme involving in MC degradation. In order to investigate the mechanism for MC degradation in fungi, there is more work need to be done.

| CON CLUS IONS
In this study, the algicidal process of B. adusta T1 was investigated by a time-serial transcriptomic analysis, and the results were compared with

ACK N OWLED G M ENTS
This study was financially supported by the National Natural Science Foundation of China (Project Nos. 31601289, 31470465, and 51309003).

CO N FLI C T S O F I NTE R E S T
None declared.

E TH I C S S TATEM ENT
None required.

DATA AVA I L A B I L I T Y S TAT E M E N T
The raw paired-end sequences from the Bjerkandera adusta isolate The number of reads were expressed in pairs.

TA B L E A 3 (Continued)
F I G U R E A 1 Total RNAs extracted from mycelia co-cultivated with cyanobacterial cells (Treatment) and without cyanobacterial cells (Control) of 6, 12, 24, and 48 hr samples

F I G U R E A 2
Boxplots showing the distribution of the FPKM values of each sample. Note: 6h_ck, control sample at 6 hr; 6h_T, treatment sample at 6 hr; 12h_ck, control sample at 12 hr; 12h_T, treatment sample at 12 hr; 24h_ck, control sample at 24 hr; 24h_T, treatment sample at 24 hr; 48h_ck, control sample at 48 hr; 48h_T, treatment sample at 48 hr. "_0" and "_1" represent repeat samples