The wild-type strain of Coniothyrium minitans, ZS-1 (CCAM 041057), which produces pycnidia and conidia on potato dextrose agar (PDA) dishes and abundant conidia in liquid shake culture (Cheng et al., 2003), was used to construct the T-DNA insertional library. ZS-1TN1812, an abnormal mutant, was originally screened from the library. Sclerotinia sclerotiorum strain Ep-1PNA367, which was derived from the hypovirulent strain Ep-1PN by single-ascospore-isolation (Xie et al., 2006), was used to examine the antifungal ability of C. minitans mutants. Colletotrichum higginsianum strain Ch-1 (IMI349061) was donated by Dr Yangdou Wei (University of Saskatchewan). All strains were cultured on PDA at 20–22°C. Agrobacterium tumefaciens strain EHA105 was used to transform the wild-type strain ZS-1 as described by Li et al. (2005) and Gong et al. (2007). Escherichia coli strain JM109 was used to construct transformation vectors (plasmids) and propagate the plasmids (Gong et al., 2007). The wild-type strain, ZS-1, its mutants, Ch-1, and S. sclerotiorum strain Ep-1PNA367 were cultured at 20°C on PDA or PDB, and maintained as PDA slants. A. tumefaciens and E. coli were cultured in LB medium at 28 and 37°C, respectively, and maintained in LB containing 15–20% glycerol at −20°C.

The growth rate, colony morphology, conidial production, and parasitic ability of ZS-1TN1812 were determined using the methods described by Qin et al. (2011). The mutant was allowed to grow on PDA plates to determine the growth rate and conidial production. To determine the colony morphology, a light microscope was used to observe the hyphal tips. Sclerotia of S. sclerotiorum were used to examine the parasitic ability of the mutant.

To determine if ZS-1TN1812 could produce antifungal substances, the mutant was dual-cultured with S. sclerotiorum strain Ep-1PNA367 on PDA. ZS-1TN1812 was activated on PDA plates, hyphal agar discs were punched out with a hole punch, and transferred on to fresh PDA plates. The inoculated plates were incubated at 20°C for 4 days and then hyphal agar discs were punched from the activated colonies of S. sclerotiorum strain Ep-1PNA367 and were placed at the opposite side of ZS-1TN1812-inoculated plates. The co-inoculated plates were further incubated at 20°C for 7 days. As a control, the wild-type strain ZS-1 was dual-cultured with Ep-1PNA367. This experiment was repeated more than five times.

The antifungal ability of mutant ZS-1TN1812 was further examined on rapeseed leaves. Mutant ZS-1TN1812 was cultured in 100 mL PDB in a 250 mL flask, shaking at 150 rpm at 20°C for 12 days. The hyphae were removed by passing through three layers of filter paper and possible hyphal debris was removed by passing through a bacterial filter (0.22 μm). The filtrate was collected and stored at 4°C before use. Rapeseed leaves were sprayed with the filtrate until fully wet. After drying out at room temperature, the leaves were inoculated with activating hyphal agar discs of S. sclerotiorum strain Ep-1PNA367. The inoculated leaves were then placed in a tray and maintained at 100% humidity at room temperature for 1 to 3 days. The lesion diameters induced by strain Ep-1PNA367 were measured with a ruler. The filtrate from the wild-type strain ZS-1 was used instead of the mutant strain ZS-1TN1812, as controls. This experiment was repeated three times.

To characterize the antifungal substances produced by ZS-1TN1812, filtrate was either incubated in a water bath at 60, 80, 100, and 120°C for 30 min or treated with proteinase K (0.05 mg/mL) at 37°C for 2 h following the method described by Jin et al. (1996). The proteinase K digestion was stopped by adding 0.5 μM phenylmethanesulfonyl fluoride (PMSF). To detect the antifungal activity of the treated filtrate, 10% (v/v) of filtrate was amended in 18 mL PDA in a Petri dish (Φ = 90 mm) and activating hyphal agar of S. sclerotiorum was inoculated at the center of the plate. The plates were placed at 20°C for 2 days. The non-treated filtrate and PDA, without any emendation, were used as controls. Furthermore, C. higginsianum was used instead of S. sclerotiorum to repeat this experiment, but the plates were incubated at 20°C for 108 h. This experiment was repeated four times.

The mycelia mass of the mutants and wild-type strain ZS-1 were harvested 4 days after growing on cellophane membrane laid on PDA (CM-PDA, pH 6.2–6.5) plates, and were used for genomic DNA extraction. Genomic DNA extraction was performed with the CTAB method (Sambrook and Russell, 2001). For Southern blot analysis of the T-DNA insertion in ZS-1TN1812 and copy number determination of CmSIT1 in C. minitans, the protocols were performed according to Gong et al. (2007) with minor modifications. HPH and CmSIT1 fragment were used as probes amplified with primer pairs HPH-SP/HPH-AP and N1812LBFP1(EX)/ N1812LBFP2(EX) respectively.

The sequences flanking the insertional T-DNA were acquired using the inverse polymerase chain reaction (iPCR) technique following the methods described by Meng et al. (2007). Genomic DNA of ZS-1TN1812 was extracted. The primer pairs LB1/Pttrpc01 and LB3/Pttrpc01 are listed in Table 1. The PCR product was ligated to the TA cloning vector, pMD18-T, and then transformed into E. coli JM109 cells. The positive clones were sequenced at Beijing Sunbiotech Co., Ltd, and sequences were analyzed with DNAMAN version 5.2.9 software (Lynnon Biosoft). The DNA sequence was used to design additional specific primers N1812SalIL1/N1812LB1, N1812SalIL2/N1812LB2, N1812HindIIIR1/N1812RB1, and N1812HindIIIR2 /N1812RB2 (see Table 1) for inverse PCR amplification to obtain additional flanking genomic DNA. The DNA sequences obtained were used to search a local C. minitans genome database with the BLASTn program and then the T-DNA insertion in ZS-1NT1812 was localized and confirmed. Based on the BLASTn result, the T-DNA was inserted at the promoter region of a gene named CmSIT1. The full-length cDNA of CmSIT1 was obtained by RT-PCR amplification.

To investigate the possible functions of CmSIT1, the full length of the putative protein was used to search the GenBank database using the BLAST program of the NCBI website. Multiple alignments of CmSIT1 amino acids were conducted using the Clustal X version 2.0 program (Larkin et al., 2007). The 3D structure of CmSIT1 was predicted using the Phyre2 server (Kelley and Sternberg, 2009; Kelley et al., 2015) and RaptorX web server (Källberg et al., 2012).

To understand the expression pattern of CmSIT1, the total RNA was extracted from the mycelial mass, which was harvested after incubating on PDA for 48, 72, 96, or 120 h. The extraction of total RNA was conducted using TRIzol®130 Plus RNA Purification Kit (Invitrogen, USA) and potential DNA contamination was removed by RNase-free DNase I treatment (TaKaRa, Dalian, China), according to the manufacturer's instructions. First-strand cDNA was synthesized using RevertAid™ First Strand cDNA Synthesis Kit (MBI, Fermentas, USA), following the manufacturer's instructions. Total cDNA abundance in the samples was normalized using the actin gene as a control, which was amplified by primer pairs Actin-FP/Actin-RP. Primers pairs SIT1SP (EX)/ SIT1AP (EX) and PEX14-3′RACE2/P-AP (EX) for RT-PCR amplification (Table 1).

Since the expression of CmSIT1 in the wild-type strain ZS-1 was undetectable with RT-PCR when growing on PDA and the T-DNA insertion led to high expression of CmSIT1, we suspected that the abnormal phenotype of ZS-1TN1812 was due to the overexpression of CmSIT1. Therefore, an overexpression vector pOESIT1 was constructed and transformed into the wild-type strain ZS-1 using the Agrobacterium-mediated transformation system as described by Li et al. (2005). Primer SIT1-SP (OVER-T)/SIT1-AP(OVER-T) were used to obtain the full length cDNA of CmSIT1.

To determine whether CmSIT1 could be expressed when C. minitans is contacting its host or non-host fungi, strain ZS-1 of C. minitans, strain Ep-1PNA367 of S. sclerotiorum, and strain Ch-1 of C. higginsianum were grown on CM-PDA plates for 3 days. The mycelial mass (about 0.25 g) of C. minitans was obtained and mixed with either an equal mycelial mass of strain Ep-1PNA367 or strain Ch-1 in a sterilized mortar. The mixed mycelial mass was amended with 1 μL ddH₂O, and then ground finely with a pestle. About 200 μL homogenate was sampled and spread on CM-PDA plates and allowed to grow at 20°C for 2, 4, or 5 days. The mycelial mass of the culture mixture was sampled to extract RNA for determining the transcripts of CmSIT1 with quantitative real time PCR (qRT-PCR). The first strand cDNA was reverse-transcribed with the kit described above and used for quantitative PCR (qPCR) analysis with the CFX96 Real-Time PCR Detection System (Bio-Rad, USA). The reaction volume was 20 μL and each sample had three replicates. The program was as follows: denaturation at 95°C for 1 min, followed by 49 amplification cycles of 95°C for 10 s and 58°C for 30 s. The melt curve was generated to verify the specificity of the amplification (from 65 to 95°C with an increment of 0.5°C per cycle, with each cycle held for 5 s). The mycelial mass of strain ZS-1 was ground and spread on cellophane membrane laid on PDA and grown 2, 4, or 5 days and then harvested for RNA extraction and qRT-PCR analysis as controls, and the expression of C. minitans actin was used as a reference gene. The primer sequences QPCR-FP1/ QPCR-RP1 and CmACT289/CmACT419 were listed in Table 1. This experiment was performed four times.

To test the fungal strain for iron stress tolerance, different concentrations of ferric chloride (1, 1.5, 2, or 2.5 mM) were added in 20 mL PDA in 90 mm Petri dishes. An activating hyphal agar plug of C. minitans was inoculated at the center of each plate. The plates were cultured at 20°C in an incubator for 12 days. The PDA plates, without any emendation, were used as controls. After 12 days, photographs of colonies were taken and colony diameters were measured. This experiment was repeated four times.

To determine the possible accumulation of iron in mutant and wild-type colonies of C. minitans, iron concentrations in hyphal masses were measured. Both the mutant ZS-1TN1812 and the wild-type strain ZS-1 were grown on CM-PDA plates for 2 days, and then the mycelial mass was harvested for iron determination. For each sample, the 100 mg mycelial mass was finely ground under liquid nitrogen with a mortar and pestle. The hyphal powder was transferred to a clean 5 mL-tube with 2 mL ice-cold PBS buffer. The mixture was further subjected to sonication with an ultrasonic cleaner (SB5200DT, Wuhan, China) for 20 min, on ice-water. Then, the liquid in the 5 mL-tube was centrifuged at 12,000 rpm for 10 min at 4°C. The clear supernatant was transferred into a clean 2 mL-tube. The supernatant samples were either stored at −80°C or subjected to iron concentration determination. The Quantichrom™ Iron Assay Kit (Bioassay Systems, USA) was used to determine the iron concentration, following the manufacturer's instructions. The iron concentration was measured at a wavelength of 590 nm with an automatic micro plate reader (BOX 998, BioTek® Instruments, Inc. USA). PBS was used instead of a sample as the blank control. This experiment was performed 3 times.

SAS version 8.1 (SAS Institute, Inc., Cary, NC, USA) was used to analyze the variation between treatments for each experiment using an ANOVA. When significant treatment effects were found, means of different treatments were compared using the protected least significant difference test at P = 0.05.

ZS-1TN1812 grew slowly in sectors on PDA plates, with the growth rate being about 1.4 mm/d, while the wild-type strain, ZS-1, was about 3.0 mm/d. The hyphal tips of mutant ZS-1TN1812 branched excessively, and the hyphae in the colony centers were frequently vacuolated. The mutant, ZS-1TN1812, may have produced a few pycnidia, but no conidia were formed in the pycnidia, while 1.3 × 10⁹ conidia were produced in each 9 cm Petri dish by strain ZS-1 14 days post-incubation. The colony of the mutant strain developed on PDA, with a yellow to brown appearance, while the typical colony of the wild-type strain, ZS-1 was filled with dark pycnidia (Figure 1). Furthermore, mutant ZS-1TN1812 could not parasitize and cause the decay of S. sclerotiorum sclerotia after 30 days, while strain ZS-1 decayed the sclerotia completely.

C. minitans is able to produce antifungal substances (AFS) under low pH conditions, but this ability is weakened when growing on PDA medium (McQuilken et al., 2002; Yang et al., 2008). However, ZS-1TN1812 could produce a large amount of antifungal substances. When dual-cultured with strain Ep-1PNA367 on PDA, ZS-1TN1812 inhibited the growth of S. sclerotiorum, the two colonies did not intermingle with each other, and the inhibition zone was maintained indefinitely. While the wild-type strain ZS-1 colonies intermingled with the colonies of S. sclerotiorum, no inhibition zone could be observed. Similar inhibition was observed when dual-culturing on low-pH MCD medium (data not shown). Furthermore, the filtrate of ZS-1TN1812 could suppress the infection of S. sclerotiorum on rapeseed leaves, while the filtrate of the wild-type strain ZS-1 could only slightly delay the infection of S. sclerotiorum (Figure 2).

C. minitans' AFS may contain different components. We detected the antagonistic ability of AFS against S. sclerotiorum (host) and C. higginsianum (non-host). When 2 mL filtrate of ZS-1TN1812 was amended into 18 mL PDA medium, the filtrate almost fully inhibited the hyphal growth of S. sclerotiorum and C. higginsianum. However, when the filtrate was treated with high temperature or proteinase K, the antagonistic ability of the filtrate to S. sclerotiorum and C. higginsianum was significantly different. The antagonistic ability of 100°C-treated filtrate to S. sclerotiorum was not significantly different from non-treated filtrate, but the antagonistic ability was fully lost when the filtrate was treated with 120°C heat. The antagonistic ability was also maintained when the filtrate was treated with proteinase K. These data suggested that the antifungal component against S. sclerotiorum is not likely to be a protein.

Interestingly, although the non-treated filtrate produced by ZS-1TN1812 could inhibit the growth of C. higginsianum, the filtrate completely lost the activity when bathed at 60°C or treated with proteinase K (Figure 3). The phenomenon suggested that the antifungal component against C. higginsianum is very likely to be a proteinous substance. Thus, the antifungal substances produced by ZS-1TN1812 may contain both a proteinous component and a non-proteinous component, with S. sclerotiorum being sensitive to the non-proteinous substances, while C. higginsianum is sensitive to the proteinous substance.

Southern blot analysis showed only one T-DNA insertion in ZS-1TN1812 (Figure 4A). A DNA segment of ~1,800 bp was successfully amplified with inverse PCR from the region flanking the left side of T-DNA, after the genomic DNA of ZS-1TN1812 was digested with Kpn I (Figure 4B). The DNA fragment obtained was sequenced and used to search a local C. minitans ZS-1 genome database. The right flank of the inserted T-DNA was confirmed with PCR amplification (Figure 4C). We found that there were two putative genes flanking the T-DNA insertion. These two genes were arranged in the opposite direction and the un-translation space between the two genes was 997 bp (Figure 4E). Flanking the right side of the T-DNA insertion was a putative gene encoding a protein that is a homolog of peroxisomal membrane anchor protein 14 (PEX14), and flanking the left side was a putative gene encoding a protein that is a homolog of the siderochrome iron transporter (SIT). The two putative genes were named CmPEX14 and CmSIT1, respectively.

To probe if these two genes were expressed in the wild-type and mutant strains, the total RNA was extracted from the mycelial mass collected from colonies growing for 2 to 5 days on PDA plates and RNA samples were examined with RT-PCR amplification. The results showed that CmPEX14 was constitutively expressed in both the wild-type strain and in ZS-1TN1812, suggesting that the T-DNA insertion in the mutant did not significantly destroy CmPEX14 (Figure 4F). Surprisingly, the expression of CmSIT1 in the wild-type strain was not detected from these RNA samples, but was constitutively detected in ZS-1TN1812 (Figure 4F). Thus, the CmSIT1 was activated by insertion of T-DNA, and we supposed that the abnormal phenotype of mutant was possibly due to the overexpression of CmSIT1.

The full length CmSIT1 is 2,510 bp, with seven exons and the full length of cDNA is 2,052 bp, coding for a putative protein with 683 amino acid residues (GenBank Acc. No. MF447899). Southern blot analysis showed only one copy of CmSIT1 in the genome of C. minitans (Figure 4D). This putative protein may have 13 transmembrane domains, as predicted with TMHMM Server v. 2.0 (http://www.cbs.dtu.dk/services/TMHMM/). Blastp analysis using the NCBI database showed that CmSIT1 had a conserved domain (amino acid 124–amino acid 490) of the MFS transporter (cd06174) (Figure 5A). This conserved domain was also identified with the Phyre2 and RaptorX programs, showing that the heterogens in the predicted binding site contained metal ions and most are iron ions. A structural model of CmSIT1 was constructed with the E. coli multidrug transporter MdfA1pw4:A (Heng et al., 2015) as a template and the iron-binding site was displayed (Figure 5B). Thus, CmSIT1 is most likely a MFS transporter for iron uptake.

Siderophore-mediated Fe³⁺ uptake in fungi is accomplished through SIT transporters, a subfamily of the major facilitator superfamily. Blastp analysis showed that CmSIT1 has homologs with other fungi and it was highly homologous to the MFS transporter of Penicillium expansum (KGO39626.1), with 31% (183/585) identity, 54% similarity (319/585), and a 2% (17/585) gap. CmSIT1 also has considerable identity (25%, 141/569) and similarity (48%, 274/569) to the SIT1 protein of the S. cerevisiae strain S288c (NP_010849.3). The identity and similarity of CmSIT1 and other selected proteins were listed in Table 2. A phylogenetic tree analysis was constructed based on reference proteins from A. fumigatus, P. expansum, S. cerevisiae, and S. sclerotiorum (Figure 5C), and CmSIT1 was clustered with Sitp1 of S. cerevisiae, Sit1 of A. fumigates and MFS, a general substrate transporter of P. expansum. This further suggested that CmSIT1 is a SIT transporter.

The full length cDNA of CmSIT1 was cloned into the pCXH vector, transformed into A. tumefaciens EHA105, and then transformed into the wild-type strain ZS-1. The phenotypes, including growth rate, colony morphology, and production of antifungal substances, of the over-expression transformants were similar to that of ZS-1TN1812. The colony morphologies of over-expression transformants were shown in Figure 6A and the growth suppression of S. sclerotiorum was shown in Figure 6B. The CmSIT1 expression of overexpression transformants was also confirmed with RT-PCR (Figure 6C). Thus, the consistent expression of CmSIT1 could cause abnormal growth and excessive synthesis of antifungal substances.

Iron homeostasis is very important for most organisms. We suspected that the abnormal phenotype of ZS-1TN1812 is likely to be caused by excess uptake of the iron element since CmSIT1 encodes a putative siderophore iron transporter. To test this hypothesis, both the wild-type and mutant strains were grown on PDA medium amended with different concentrations of ferric chloride and the results showed that the growth of the wild-type strain was inhibited when growing on 1 mM FeCl₃-amended PDA medium. When the concentration of FeCl₃ was increased, the inhibition was more significant, with wild-type strain growth almost completely inhibited when the concentration of FeCl₃ in medium was 2.5 mM. The growth of ZS-1TN1812 in 2 mM FeCl₃-amended PDA medium was not significantly different from that of non-amended PDA medium. However, complete inhibition of growth was observed when growing on 2.5 mM FeCl₃-amended PDA medium. Interestingly, the colony morphology of the wild-type strain developed on 2 mM FeCl₃-amended PDA medium was similar to that of ZS-1TN1812 on PDA medium (Figures 7A,B).

As an essential element, it is not surprising that iron could be detected in the hyphal mass of strains ZS-1 and ZS-1TN1812. However, the iron level in ZS-1TN1812 was significantly higher than in strain ZS-1 when growing on the PDA plate, with the iron concentration at 37.2 μM in the mutant and only 16.3 μM in the wild-type strain ZS-1. When growing on 2.0 mM FeCl₃-amended PDA, the iron levels in strain ZS-1 and strain ZS-1TN1812 were significantly higher, at 78.5 and 106.2 μM (Figure 7C), respectively. This result combined with the results above suggested that iron-uptake ability in ZS-1TN1812 was significantly increased.

Although, the expression of CmSIT1 is not detectable via RT-PCR when growing on PDA medium alone, it may be expressed when dual-cultured with other microorganisms, since siderophores are important molecules in competition with other organisms for iron. To test this hypothesis, C. minitans was allowed to contact its host, S. sclerotiorum, or non-host, C. higginsianum, and the expression of CmSIT1 was determined using qRT-PCR. The results showed that when contacting S. sclerotiorum, CmSIT1 was significantly expressed at 2 days post-inoculation (dpi), and the expression declined at 4 and 5 dpi, but the expression level was still significantly higher than when growing on PDA alone. When C. minitans and non-host C. higginsianum grew on PDA, CmSIT1 was highly expressed at 2 dpi and the high level of expression was maintained even at 4 dpi, but dropped significantly at 5 dpi (Figure 8). This result suggested that CmSIT1 may play a role during parasitizing, such as competing for iron with its host, S. sclerotiorum.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. The reviewer JM-A and handling Editor declared their shared affiliation.