Phylogenetic Position of Shiraia-Like Endophytes on Bamboos and the Diverse Biosynthesis of Hypocrellin and Hypocrellin Derivatives

The main active ingredients of the fruiting bodies of Shiraia bambusicola and Rubroshiraia bambusae are Hypocrellins, belonging perylenequinones with potential photodynamic activity against cancer and microbial diseases. However, the strains of S. bambusicola and R. bambusae do not produce hypocrellins in culture, so resource exploitation of natural products was seriously restricted. In this study, a series of novel Shiraia-like fungal endophyte strains, with varying sporulation ability and synthesizing diverse secondary metabolites, was isolated from different bamboos. Based on phylogenetic analyses and morphological characteristics of the endophytes, Pseudoshiraia conidialis gen. et sp. nov. is proposed. The secondary metabolites of different fruiting bodies and strains have been comprehensively analyzed for the first time, finding that the endophytic strains are shown not only to produce hypocrellins, but also other perylenequinonoid compounds. It was noteworthy that the highest yield of total perylenequinone production and hypocrellin A appeared in P. conidialis CNUCC 1353PR (1410.13 mg/L), which was significantly higher than any other wild type P. conidialis strains in published reports. In view of these results, the identification of Shiraia-like endophytes not only confirm the phylogenetic status of similar strains, but will further assist in developing the production of valuable natural products.


Introduction
Hypocrellins belong to the perylenequinonoid family of compounds. They are very important photosensitizers and have attracted broad attention because of their light-induced antitumour, antifungal and antiviral activities [1][2][3][4][5][6][7]. In China, hypocrellins have been used medicinally to treat skin diseases for many years [8]. In recent years, hypocrellins and their derivatives have been incorporated into polymer micelles or nanoparticles for the treatment of methicillin-resistant Staphylococcus aureus infections [9] and cancer therapy [10][11][12]. In addition to benefiting the pharmaceutical industry, hypocrellins also have extensive potential applications in the agricultural, cosmetic, food and feed industries [13][14][15].
Shiraia-like endophytes associated with plants have been identified and shown to produce hypocrellins [16,[26][27][28][29]. Based on phylogenetic analysis, these strains generally clustered in the family Shiraiaceae [25,26], but their explicit taxonomic status could not be established due to the lack of morphological characteristics.
Since 2008, we have been characterizing Shiraia-like endophytes from bamboos, and have explored several interesting strains with conidial production, and diverse natural products. The aim of this study is to accurately establish the taxonomic status of Shiraialike endophytes based on morphological characteristics and phylogenetic analysis, and comprehensively analyze the secondary metabolites of these strains.

Isolates
Fruit bodies of Shiraia bambusicola were collected from Hangzhou, Zhejiang, China and those of Rubroshiraia bambusae from Yulong County, Yunnan, China. Fungal endophytes were isolated from asymptomatic tissues of bamboos (Poaceae: Bambusoideae) in various localities in China (Table 1). The methods of isolation are described in Shen et al. [30] and Zhou et al. [31]. All isolates were cultured on Potato Dextrose Agar (PDA, containing 200 g/L potato, 20 g/L dextrose and 20 g/L agar) at 25 • C for 14 days to observe the morphology. The specimens were deposited in the Fungarium of the College of Life Science, Capital Normal University, Beijing, China (BJTC) and the China Forest Biodiversity Museum of the Chinese Academy of Forestry (CAF), and ex-type living cultures were deposited in the China Forestry Culture Collection Center (CFCC) and Capital Normal University Culture Collection Center (CNUCC).
"-" indicating data unavailable. The strains of new species in this study are emphasized in bold. * Ex-holotype or ex-epitype cultures.

Morphological Analysis
Measurements and photographs of characteristic structures were made according to methods described by Liu et al. [32], and for each structure 30 measurements were made. Microscopic preparations were made in clear H 2 O, observed and photographed using a Nikon SMZ-1000 dissecting microscope (DM), an OLYMPUS light microscope (LM) or a Hitachi S-4800 scanning electron microscope (SEM). Colony characters and pigment production on PDA incubated at room temperature were noted after 14 d. Colony colors were taken from ColorHexa (https://www.colorhexa.com/, accessed on 18 December 2020). Growth rates were measured after 7 and 14 d.

Phylogenetic Analysis
The new sequences were submitted to the GenBank database and other sequences included in this study were downloaded from GenBank (Table 1) based on recent publications [25,38]. The DNA sequences generated with forward and reverse primers were aligned to obtain consensus sequences using EditSeq version 5.00. A partition homogeneity test was done to determine the congruence of gene fragments [39,40]. Subsequent alignments were generated using online MAFFT tools (https://www.ebi.ac.uk/ Tools/msa/mafft/, accessed on 24 November 2020), and edited using Gblocks 0.91b (http://www.phylogeny.fr/one_task.cgi?task_type=gblocks, accessed on 24 November 2020), selecting all options for a less stringent selection.
Maximum parsimony (MP) analysis was performed on the multi-locus alignment including two loci (ITS and LSU) with PAUP v.4.0b10 [41], using the heuristic search option with tree bisection and reconstruction (TBR) branch swapping and 1000 random sequence additions. Maxtrees were 1000, branches of zero length were collapsed and all multiple parsimonious trees were saved. Clade stability was assessed in a bootstrap analysis with 1000 replicates, each with 10 replicates of random stepwise addition of taxa. Tree statistics (TL), consistency index (CI), retention index (RI), rescaled consistency index (RC) and homoplasy index (HI) as the descriptive tree statistics were calculated for the generated trees.
For the Bayesian analysis, a Markov Chain Monte Carlo (MCMC) algorithm was conducted to reconstruct the single locus and multi-locus phylogenetic trees with Bayesian posterior probabilities in MrBayes v. 3.1.1 [42]. For the Bayesian analysis, models of nucleotide substitution were determined by MrModeltest v.2.3 [43] for each gene and included in the analyses (ITS, LSU, TEF and TUB2: GTR + I + G, SSU: HKY + I). The analyses of four chains were conducted for 10,000,000 generations with the default settings and sampled every 100 generations, halting the analyses at the average standard deviation of split frequencies of 0.01. The first 25% trees were discarded as the burn-in phase of the analyses and the posterior probabilities (PP) were obtained from the remaining trees.
Maximum likelihood (ML) analysis of the dataset was carried out using RAxML 8.0.14 [44][45][46] and the GTRGAMMI substitution model with parameters unlinked. The ML bootstrap replicates (1000) were computed in RAxML using a rapid bootstrap analysis and search for the best-scoring ML tree.
Trees were viewed in Treeview [47] and edited in Coreldraw v.X4 (Corel Corporation, Canada). ML bootstrap values (MLBS) and MP bootstrap values (MPBP) equal to or greater than 50% and Bayesian posterior probability (PP) equal to or greater than 0.95 are given at each node (Figure 2). The combined alignment and phylogenetic tree were submitted at TreeBASE (www.treebase.org, accessed on 24 November 2020; study S28492). Trees were viewed in Treeview [47] and edited in Coreldraw v.X4 (Corel Corporation, Canada). ML bootstrap values (MLBS) and MP bootstrap values (MPBP) equal to or greater than 50% and Bayesian posterior probability (PP) equal to or greater than 0.95 are given at each node (Figure 2). The combined alignment and phylogenetic tree were submitted at TreeBASE (www.treebase.org (accessed on 24 November 2020); study S28492).

Submerged Cultivation and Secondary Metabolite Extraction
The endophytic fungi isolates were cultured on PDA at 26 • C for 7 days. Based on colony colour, three strains representing different morphs were selected for further experiments. Five plugs (5 mm in diameter) of growing culture plus the adhering mycelium were added to 250 mL Erlenmeyer flasks containing 150 mL of Potato Dextrose Broth media (PDB, containing 200 g/L potato and 20 g/L dextrose). All liquid cultures were kept at 26 • C for 10 d with shaking (180 rpm).
Fresh mycelia of the fungal strains Shiraia bambusicola CNUCC 0172, CNUCC 0122 and CNUCC MJ1 were cultured on PDA at 26 • C for 10 days. Five plugs (5 mm in diameter) of growing culture plus the adhering mycelium were subsequently added to 150 mL PDB. The liquid cultures were kept at 26 • C for 10 d with shaking (180 rpm).
Fresh mycelia of the fungal strain Rubroshiraia bambusae CNUCC 1000 were cultured on PDA at 16 • C for 30 days. Five plugs (5 mm in diameter) of growing culture plus the adhering mycelium were subsequently added to 150 mL PDB. The liquid cultures were kept at 16 • C for 40 d with shaking (180 rpm).
The fermented mycelia of each fungus were filtered and dried at 45 • C. The dry powder (0.1 g) of ascostromata and mycelia of S. bambusicola and R. bambusae was accurately weighed and ultrasonic extracted for 30 min with 5 mL methanol. The dry powder of endophytic mycelia was treated in the same manner.

HPLC-DAD-MS Analysis
HPLC-DAD-MS analysis was performed using a Shimadzu LC-20AD liquid chromatography (LC) system coupled with a diode array detector (DAD) and an electrospray ionization-ion-trap-time-of-flight (ESI-IT-TOF) mass spectrometer (MS) (Shimadzu, Kyoto, Japan). For analytical purposes, a Kromasil 100-5 C18 (250 × 4.6 mm, 5 µm) column was used. The mobile phase was composed of water containing 0.1% formic acid (A) and methanol (B), and the gradient of eluent B started at 5% and gradually increased to 90% over 90 min at a flow rate of 1 mL/min. The MS conditions refer to Niu et al. [48].

Phylogeny
The multi-locus phylogenetic analysis included 68 ingroup samples, and used Pleospora herbarum (CBS 191.86) as outgroup. The dataset of five loci comprised 3 831 characters including the alignment gaps, of which 622 characters were parsimony-informative, 172 parsimony-uninformative and 3037 constant. A best scoring RAxML tree is shown in Figure 2, the maximum parsimony and Bayesian tree confirmed the tree topology obtained with maximum likelihood.
The results showed that among 14 strains isolated in this study, 10 strains were clustered together with endophyte group A as listed in Dai et al. [25], forming a highly supported clade (BS = 100, BP = 100, PP = 1.00). The strains CNUCC 0122, CNUCC 0172 and CNUCC MJ1 together with BJTC HOU999 were clustered in the clade of S. bambusicola. The strains CNUCC 1000 together with BJTC HOU1000 were clustered in the clade of R. bambusae.

Taxonomy
Based on phylogenetic analyses and morphological characteristics, a novel species belonging to a new genus was recognized in this study.  (Figures 3 and 4). Diagnosis. Pseudoshiraia conidialis differs from Shiraia bambusicola by small, cylindrical or ellipsoidal conidia, without septa.

Discussion
In our previous study, a large number of strains were isolated from S. bambusicola and R. bambusae (Data not shown). However, hypocrellins were found only in fruiting bodies of S. bambusicola and R. bambusae, and cultured strains did not produce these chemicals (Figure 3). More recent studies have demonstrated that a few Shiraia-like endophytes isolated from bamboo tissues could produce hypocrellins, and even the strains isolated from the stromata of S. bambusicola could also produce hypocrellins [27,30,50,[53][54][55]. But the taxonomic statuses of these strains were very confusing. Such as the strains Shiraia sp. SUPER H168, Shiraia sp. slf14, Shiraia sp. S9, S. bambusicola UV-62 and S. bambusicola ZH-5-1 were isolated from bamboo tissues as endophytes or isolated from the stromata of S. bambusicola [27,50,53,56,57]. Although they were regarded as Shiraia spp., the strains Shiraia sp. SUPER H168, Shiraia sp. slf14 and Shiraia sp. S9 were clustered with Shiraia-like endophytes but not with Shiraia in the phylogenetic trees ( Figure 2) [25,50]. Therefore, the taxonomic statuses of other strains like these need to be re-identified.
Morakotkarn et al. [26] and Dai et al. [25] isolated 22 Shiraia-like endophytic strains from bamboo tissues. Based on phylogenetic analysis, these strains were divided into two groups, one which could produce hypocrellins, and the other not. Among them, the group with hypocrellins was clustered with these known production strains (Figure 2), forming a highly supported clade. These belong to the newly described species Pseudoshiraia conidialis.
Unfortunately, cultures from most of the corresponding endophytes only had low production yield, and several attempts were made to increase hypocrellins production, such as addition of Triton X-100 surfactant to the submerged cultures [49], exposing cultures of Shiraia sp. to light at various wavelengths [58] or light/dark shift [59], or co-cultivation of Shiraia sp. with Pseudomonas fulva [50,60]. Our previous work [8] attempted to use gamma rays to mutate P. conidialis zzz816, boosting the HA production to increase to 414.9%.
The conidium-producing and hypocrellin-generating strain CNUCC 1353PR was isolated and characterized in this study. To the best of our knowledge, hypocrellin production in P. conidialis CNUCC 1353PR is significantly higher than other wild type P. conidialis strains in public [49][50][51][52]. Through comparative analysis of the metabolites of ascostromata and mycelia of S. bambusicola and R. bambusae, we discovered that HA, HB and SA only appeared in the ascostromata, but not in mycelia of cultured strains. In addition, the P. conidialis strains contain more diverse perylenequinonoid compounds, and in addition to HA, HB and SA, also produced EA, EB and EC; demonstrated by the four peaks (R t 12 min, 16 min, 17.5 min and 23.6 min) in the HPLC chromatogram of the mycelia of P. conidialis CNUCC 1353PR (Figure 5I), and the UV-Vis spectra indicated that they all belonged to perylenequinones (Supplementary Figure S1). These results displayed that P. conidialis CNUCC 1353PR could be a potential industrial strain for perylenequinone production.
Other sources of perylenequinone production could be found. For example, strain MSX60519, isolated from dry leaf litter, was also found to produce hypocrellins [16], and Li et al. [61] explored an endolichenic strain, Phaeosphaeria sp. 20081120, which could produce HA, HB and other perylenequinones. Meng et al. [62] investigated metabolites produced by an endophytic fungus identified as Penicillium chrysogenum isolated from Fagonia cretica, and also found hypocrellins. However, these strains did not cluster with Shiraiaceae ( Figure 2). Although the species of other family may also produce hypocrellins, further studies and verification are needed.

Conclusions
In this study, a new genus Pseudoshiraia in Shiraceae was established, and a series of species with high output of hypocrellins were exploited for the first time. Furthermore P. conidialis CNUCC 1353PR contains multiple types of perylenequinones, not only hypocrellins but also elsinochromes.