Antibiotics were purchased from Solarbio (Beijing China). calcofluor white (CFW) was obtained from Sigma-Aldrich (United States), 4',6-diamidino-2-phenylindole (DAPI) was obtained from Beyotime (Shanghai, China). Lysis enzymes are obtained from Sigma (United States), snailase is obtained from Solarbio (Beijing China), and lywallzyme is obtained from Guangdong Microbial Culture Collection Center (Guangdong, China).

All strains used in this study are listed in Table 1. The C. empetri MEFC009 and all mutant strains were grown on PDA. Cultivation in shake flasks was carried out in seed medium (MKS: soluble starch 15g/L, sucrose 10g/L, cotton seed meal 5g/L, peptone 10g/L, KH₂PO₄ 1g/L, CaCO₃ 2g/L, and pH 6.5) and FR901379 production medium (MKF: glucose 10g/L, corn starch 30g/L, peptone 10g/L, (NH₄)₂SO₄ 6g/L, KH₂PO₄ 1g/L, FeSO₄ 0.3g/L, ZnSO₄ 0.01g/L, CaCO₃ 2g/L, and pH 6.5).

The cell wall and nuclei of the hyphae of C. empetri MEFC009 were stained by CFW and DAPI, respectively (Chazotte, 2011). The hyphae of C. empetri MEFC009 was immersed in 200μl phosphate buffered saline (PBS pH 7.4, 0.1mol/L), then stained by 0.8μg/ml of DAPI or 0.05μg/ml of CFW for 5min. The hyphae were washed twice by PBS to remove the stain and observed by confocal microscopy (FluoView FV1000). In addition, 0.8μg/ml of DAPI and 0.05μg/ml of CFW were added together to stain the nuclei and cell wall of the hyphae simultaneously. The excitation and emission wavelength of CFW and DAPI is 345/405nm with blue fluorescence.

All primers used in this study are listed in Supplementary Table S1. The selectable marker gene nat was amplified from pPK2natGFPD by PCR using primers PtrpC/nat-R (Gu et al., 2018). The coding DNA sequence (CDS) of neo was synthesized fused with promoter Ppgk and terminator Tpgk by fusion PCR. The markers nat and neo were cloned into the binary vector pCambia1300 using a One-Step Cloning Kit (Vazyme, Nanjing, China) to construct the recombinant plasmids pPM-3, and pPM-4 (Hršelová et al., 2015). The recombinant A. tumefaciens LBA4404 harboring plasmids pCambia1300, pPM-3, and pPM-4 were constructed. The Agrobacterium-mediated transformation of C. empetri MEFC009 was performed as previously described (Yamada et al., 2009).

To evaluate the function of promoters PgpdA, Ppgk, and PgpdAt, the cassettes containing the above promoters were constructed, respectively. The DNA cassettes of PgpdA-sgfp-TtrpC-hph and PgpdAt-sgfp-TtrpC-hph were amplified from plasmid pAN52-4 (Punt et al., 1991) and pXH2-1 (Huang et al., 2014) by PCR using primer pairs PgpdA-F/TtrpC-R and PgpdAt-F/TtrpC-R, respectively. The fragment of sgfp of the cassette of Ppgk-sgfp-Tpgk-hph was amplified from plasmid pXH2-1, and then fused with Ppgk, Tpgk, and hph by fusion PCR, resulting in cassette Ppgk-sgfp-Tpgk-hph.

To knock out the ku80 gene, the flanking 5' and 3' DNA of the ku80 gene were amplified by PCR from the genome of C. empetri MEFC009 using primer pairs Uku80-F/Uku80-(hph)-R and Dku80-(hph)-F/Dku80-R, and fused with hph marker by fusion PCR. The gene-targeting cassette was amplified using primers Uku80-CS-F/Dku80-CS-R. The ku80 deletion cassette using neo marker was constructed through the same approach. The flanking 5' and 3' DNA of pks11.2 gene were amplified by PCR from the genome of C. empetri MEFC009 using primer pairs Upks11.2-F/Upks11.2-R and Dpks11.2-F/Dpks11.2-R and fused with hph marker by fusion PCR. The gene-targeting cassette was amplified by primers Upks11.2-CS-F/Dpks11.2-CS-R.

The transformation mediated by protoplast was developed in C. empetri MEFC009 according to the protocol previously described with modification (Marcone et al., 2010). The fresh hyphae from the PDA plate were broken by Superfine homogenizers (Fluko) and cultivated in 50ml of MKS medium for 2days, 220rpm, 25°C. Then hyphae were broken again and inoculated in a new shack flask with 10% proportion, cultivated for 1day at 25°C. The hyphae were harvested by filtration through Miracloth (Calbiochem), washed with 0.6M MgSO₄, and immersed in the enzymatic solution (0.6% of snailase, 0.6% lysis enzymes, and 0.7% of lywallzyme) for 4h at 30°C to prepare protoplasts. The protoplasts were collected by filtered through Miracloth and centrifuged at 1,500rpm at 4°C for 20min. After washed with 1M D-sorbitol buffer and STC (1M D-sorbitol, 10mM Tris-HCl pH 8.0, 50mM CaCl₂) successively, the protoplasts were resuspended in STC used for transformation.

To optimize the regeneration medium for protoplasts, the protoplasts suspension was diluted to 10⁴cells/ml with STC and H₂O, respectively. A 100μl portion of the protoplast dilution was mixed with the regeneration medium and pour into PDAS plates with different concentrations of D-sorbitol: 0, 0.4, 0.6, 0.8, and 1.2M.

As for transformation, approximately 1μg of DNA fragments were mixed with 100μl protoplast solution (1.5×10⁷ protoplasts/ml) and 50μl of ice-cold PSTC (40% PEG 4000, 1M D-sorbitol, 10mM Tris-HCl pH 8.0, and 50mM CaCl₂), incubating on ice for 25min. Then 1ml of PSTC was added and incubated for 20min at room temperature. Thereafter, the protoplasts mixture was mixed with 20ml of liquid top agar (PDB with 0.5% agarose and 0.8M D-sorbitol), spread on PDAS amended with the antibiotic, and cultured for 5–7days at 30°C.

C. empetri MEFC009 and mutant C. empetri MEFC009-Δku80 were inoculated in 50ml of MKF medium at 25°C for 10days. Then 1ml of fermentation culture was taken every 2days and extracted with equal volumes of methanol by ultrasonic crushing for 1h. Four independents were set for each experiment. The amount of FR901379 was determined by high performance liquid chromatography (HPLC) monitored at 210nm. The separation was carried out on a C₁₈ reversed phase column (Agilent ZORBAX SB-C18 column, 4.6×150mm, 5μm) at a flow rate of 1ml/min at 30°C, using acetonitrile (ACN)/H₂O as elution solvents. A linear gradient from 5 to 100% ACN in H₂O containing 0.1% (v/v) TFA in 20min was used.

Hyphae are the most common resource for preparing protoplast. The morphology of vegetative hyphae is the key factor related to the difficulty of protoplast transformation, such as septate to aseptate and uninucleate to multinucleate. In the genetic engineering of the fungi species with aseptate multinucleate hyphae, heterokaryons are spontaneously formed during transformation. And, it is difficult to screen out the homozygote with engineered nucleus (Kaminskyj and Hamer, 1998). The cell wall of the vegetative hyphae of C. empetri MEFC009 was stained by CFW. As shown in Figure 1, the hyphae of C.empetri are branched and septate. The length of each cell is about 7–8 μm. The nuclei were stained in bright blue by DAPI. We can observe that the nuclei were equidistantly distributed in hyphae. When stained with DAPI and CFW simultaneously, the nuclei and diaphragm of the hyphae could be observed. It indicated that only one nucleus was present in each septate hyphae cell. This type of vegetative hyphae is the most suitable for genetic transformation.

A selectable marker is the essential element for genetic transformation, which is usually also a limiting factor for genetic engineering in filamentous fungi. There are a few selectable markers that have been developed in the genetic transformation of filamentous fungi (Ramsden et al., 2011). However, these selectable markers are not always universal in different species because of the different sensitivity to antibiotics. The resistance gene of hygromycin B, hph, has been confirmed functional for the transformation of C. empetri F-11899 (Yamada et al., 2009). To verify the selectable markers applicable to C. empetri MEFC009, the sensitivity to four selectable antibiotics were tested. The growth of C. empetri MEFC009 were inhibited on potato dextrose agar (PDA) individually supplemented with 100μg/ml of hygromycin B, geneticin, and nourseothricin, but grew normally on plates supplemented with 500μg/L of pyrithiamine. It demonstrated that C. empetri MEFC009 is sensitive to hygromycin B, geneticin, and nourseothricin.

The plasmids containing the resistant genes of hygromycin B (hph), geneticin (neo), or nourseothricin (nat) were respectively transformed into C. empetri MEFC009 using ATMT method as previously described (Yamada et al., 2009). The transformants were successfully screened out on PDA plates containing 100μg/ml of corresponding antibiotics (Figure 2). The results of PCR showed that the resistant genes hph, neo, and nat were integrated into the chromosome of the transformants, respectively, (Supplementary Figure S1). Therefore, the resistant genes hph, neo, and nat are available selectable markers for the transformation of C. empetri.

Although the genetic transformation of C. empetri MEFC009 was achieved through ATMT method, it is still inefficient, no more than 10 transformants obtained in each experiment. In addition, the transformation process is laborious, time-consuming, and poorly repeatable. Protoplasts-mediated transformation is the most convenient and popular method for the transformation of filamentous fungi (Prabha and Punekar, 2004). The generation of high-quality protoplasts is the major challenge to achieve efficient transformation (Roth and Chilvers, 2019). However, enzymatic digestion conditions of the cell walls are different for each fungal strain. The details should be optimized for each case such as hydrolase composition, isotonic solution, regeneration medium, etc. (Ruiz-Díez, 2002).

We optimized the protoplast transformation according to the previously developed method (Marcone et al., 2010). The hyphae of C. empetri MEFC009 were cultivated in PDB medium using a two-step inoculation method. The optimized enzyme solution of 0.6% of snailase, 0.6% lysis enzymes, and 0.7% of lywallzyme was used for cell wall digestion, which produced 1.5×10⁶ protoplasts/ml at 30°C for 2–4h (Figure 3A). The protoplasts were purified and resuspended at 1.5×10⁷ protoplasts/ml in STC solution. In the following PEG-CaCl₂ mediated transformation, 1μg of DNA fragment of hph-sgfp was mixed and incubated with 100μl protoplasts solution. The regeneration medium was optimized, and the protoplasts could regenerate on the PDA plates with 0.8M and 1.0M D-sorbitol (Figure 3B). Beginning with 1.5×10⁶ protoplasts, approximately 100–120 positive transformants were obtained in a reaction using hph marker (Figure 3C). The efficiency is significantly higher than that of the ATMT method and is sufficient for functional genomic research and metabolic engineering.

For a specific fungal host, the efficient promoters are the key elements to achieve high-level expression of target genes (Blumhoff et al., 2013). The promoters of glyceraldehyde-3-phosphate dehydrogenase (gpd) and phosphoglycerate kinase (pgk) of A. nidulans have widely been used in numerous filamentous fungi (Streatfield et al., 1992; Cao et al., 2012). The function of promoters Ppgk, PgpdA (from A. nidulans), and PgpdAt (from A. terreus; Huang et al., 2014) were evaluated in C. empetri MEFC009 using sGFP (synthetic green fluorescent protein) as the reporter. As shown in Figure 4, the DNA cassettes containing the promoters (PgpdA, PgpdAt, or Ppgk), sgfp, TtrpC terminator, and hph were transformed into C. empetri MEFC009, respectively. The integration of DNA cassettes in the genome was confirmed by genomic PCR. The hyphae of transformants were randomly picked out and observed under the fluorescence microscope. The transformants harboring PgpdA, PgpdAt, and Ppgk evaluation cassettes displayed significant green fluorescent, while the parental strain did not show any visible fluorescence (Figure 4). These results demonstrated that promoters PgpdA, PgpdAt, and Ppgk successfully drove sgfp expression in C. empetri MEFC009. According to the intensity of fluorescence, PgpdA and PgpdAt showed similar promoter activity, which is significantly stronger than that of Ppgk.

Gene-targeting is the basis for various genetic engineering techniques, including gene knockout, promoter replacement, and site-specific expression of heterogeneous genes, etc. (Ninomiya et al., 2004). However, homologous recombination (HR) efficiency in filamentous fungi is extremely low because the NHEJ pathway dominates the repair of DNA double-strand breaks (Huang et al., 2016). Learning from previous research, we tried to improve the frequency of gene targeting by disrupting the NHEJ pathway. The ku80 gene was identified according to the annotation of genome sequencing of C. empetri MEFC009. As shown in Figure 5A, 1.5kb region in ku80 gene was replaced by the hph selectable marker. The transformants were screened using PDAS-H plates (PDA with 0.8M D-sorbitol and 100 μg/ml of hygromycin B). Forty-six transformants with hygromycin B resistance were randomly selected for further genotyping. Only in two transformants, the gene-targeting cassette was directly integrated at the ku80 loci (Figure 5B). Therefore, the frequency of HR was about 4% in C. empetri.

The phenotypic characteristics of C. empetri MEFC009-Δku80-hph were assessed to evaluate the influences of ku80 deletion. The ku80 deletants showed the same growth rates, colonial morphology, and pigmentation on PDA plates (Figure 5C). Furthermore, no deviation in FR901379 productivity was observed in the shake flask fermentation (Figure 5D). To evaluate the HR frequency in the ku80 deletant, the hph integrated in ku80 loci was replaced with neo through another round of gene-targeting (Figure 5E). The transformants were screened by PDAS containing 100μg/ml of geneticin and 11 transformants were randomly selected for genotyping. The results of genomic PCR showed that the hph sequence was successfully replaced with neo sequence in all 11 transformants (Figure 5F). These results demonstrated that the HR frequency has been improved from 4 to 100% by deleting ku80 in C. empetri.

The 1,8-dihydroxynaphthalene (DHN) melanin is the most common type of conidial pigments in ascomycetous fungi, which is polymerized from polyketide naphthopyrone (Eisenman and Casadevall, 2012). The colony of C. empetri MEFC009 displays black color on PDA plates. Identifying the synthesis gene of melanin would be a perfect practice for testing the efficiency of genetic modification. According to the results of antiSMASH analysis, there are 18 PKSs in the genome of C. empetri MEFC009 (Blin et al., 2019). Among them, PKS11.2 was predicted to be responsible for the synthesis of melanin, which has four potential catalytic domains: beta-ketoacyl synthase (KS), an acyltransferase (AT), two acyl carrier sites (P1 and P2), and a thioesterase (TE)/Claisen cyclase (CYC; Figure 6A). The phylogenetic tree indicates that PKS11.2 appears highly homologous to other identified DHN-melanin PKSs, BcPKS13 of Botrytis cinerea (Schumacher, 2016), WdPKS1 of Wangiella (Exophiala) dermatitidis (Feng et al., 2001), and BoPKS1 of Bipolaris oryzae (Moriwaki et al., 2004; Figure 6B). To identify the function of pks11.2, the gene-targeting cassette harboring hph selectable marker was transformed in the NHEJ-deficient mutant C. empetri MEFC009-Δku80-neo. All the transformants growing on selection plates exhibited white colonies (Supplementary Figure S2; Figure 6C). It indicated that the pks11.2 was indeed involved in the biosynthesis of melanin and was disrupted by gene-targeting in all transformants. The genotype was further confirmed by genomic PCR, that the selectable marker hph was indeed directly integrated at pks11.2 loci in all eight randomly selected transformants (Figure 6D). Therefore, C. empetri MEFC009-Δku80-neo was indeed an excellent chassis cell for gene-targeting.