Reagents

Laboratory common reagents such as ethanol, sucrose, paraformaldehyde, hematoxylin, eosin, agarose and electrophoresis reagents, etc. were purchased from Beijing Bioland Technology Company. Mercuric chloride (0.1%) for surface sterilization of freshly collected C. sinensis specimens was a gift from the Institute of Microbiology, Chinese Academy of Sciences. EcoRI endonuclease is a product of New England BioLabs, United States. The Universal DNA Purification kit was a product of TIANGEN BIOTECH Company, China. The DNeasy Plant Mini Kit was a product of Qiagen Company, Germany. The Gel Extraction Kit was a product of Omega Bio-Tek, United States. The Taq PCR reagent kit and Vector NTI Advance 9 software were purchased from Invitrogen, United States. Calmodulin was a product of Abicom (Shanghai, China).

Collection of C. sinensis specimens and ascospores

Immature C. sinensis specimens were purchased from local markets from the Hualong (located at 36°13’N, 102°19’E) and Yushu areas (located at 33°01’N, 96°48’E) of Qinghai Province of China (3800–4600 m’ altitude) in mid-May and characterized by a plump caterpillar body and very short stroma (1.0–2.0 cm) [19, 27]. Mature C. sinensis specimens were collected in mid-June and characterized by a plump caterpillar body and long stroma (>5.0 cm) and by the formation of an expanded fertile portion close to the stromal tip, which was densely covered with ascocarps (Fig 1). Governmental permission was not required for C. sinensis purchases in local markets, and the collections of C. sinensis specimens from sales by local farmers fall under the governmental regulations for traditional Chinese herbal products.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Cultivation of mature C. sinensis specimens in paper cups and collection of ascospores.

Mature C. sinensis specimens were cultivated in our Xining laboratory (altitude of 2,200 m) (Fig 1A). The fully ejected ascospores were collected using the double layers of autoclaved weighing papers (Fig 1B). Numerous semiejected ascospores adhere to the outer surface of an ascus (Fig 1C, after removing the upper layer of autoclaved weighing papers for collection of the fully ejected ascospores) during the massive ejection of ascospores. The stromal fertile portion (SFP) densely covered with numerous ascocarps is labeled with “]”.

https://doi.org/10.1371/journal.pone.0270776.g001

The specimens were washed thoroughly on site in running water with gentle brushing, soaked in 0.1% mercuric chloride for 10 min for surface sterilization and washed 3 times with sterile water. The thoroughly cleaned specimens were immediately frozen in liquid nitrogen on site and kept frozen during transportation to the laboratory and during storage prior to further processing [19, 27].

Some of the mature C. sinensis specimens were harvested along with the outer mycelial cortices and soil surrounding the caterpillar body and replanted in paper cups in soil obtained from C. sinensis production areas (Fig 1A) and were cultivated in our laboratory (altitude 2,200 m) in Xining City, Qinghai Province of China [67, 68]. Because of the phototropism of natural C. sinensis, we kept the windows fully open, allowing sufficient sunshine and a natural plateau breeze blowing over the cultivated specimens in the paper cups. The room temperature was maintained naturally, fluctuating with the lowest temperature at 18–19°C during the night and the highest temperature at 22–23°C in the early afternoon. The humidity of our laboratory was maintained by spraying of water using an atomizer twice a day in the morning and evening.

Fully ejected ascospores of C. sinensis were collected using double layers of autoclaved weighing paper (Fig 1B). During massive ascospore ejection, numerous ascospores adhered to the outer surface of asci, as shown in Fig 1C after removing the upper layer of autoclaved weighing papers for collection of the fully ejected ascospores, and failed to be brushed away using an autoclaved brush; hence, these ascospores were instead gently scratched off using a disinfected inoculation shovel or ring and referred to as semiejected ascospores.

The 2 types of ascospores were cleaned by 2 washes with 10% and 20% sucrose solutions and 10-min centrifugation at 1,000 rpm (desktop centrifuge, Eppendorf, Germany); the supernatant was discarded after each centrifugation. The pellets (ascospores) were subsequently washed with 50% sucrose solution and centrifuged for 30 min, and the ascospores that floated on the liquid were collected [67]. The fully and semiejected ascospores were stored in a -80°C freezer prior to further processing.

Histological examination of the SFP, ascocarps and ascospores of C. sinensis

The fully ejected ascospores of C. sinensis were diluted in normal saline, placed on a glass slide, and air-dried for histological examination under an optical microscope (Model BX51, OLYMPUS, Japan) without staining. The ascospores were fixed on a glass slide with 4% paraformaldehyde for 1 h, incubated in 0.01% calmodulin for 5 min for the visualization of septa, washed 3 times with PBS, and observed under a fluorescence microscope with UV epi-illumination (Model XZ51, OLYMPUS, Japan) [68].

The mature C. sinensis stromata collected during the massive ejection of ascospores were immersed in 10% formalin for fixation and subjected to dehydration in 50%, 70% and 95% ethanol for 1 h each [69]. The SFP tissues (densely covered with ascocarps) were embedded in paraffin and sliced to 5-μm thickness (Model TN7000 Microtome, Tanner Scientific, Germany). The ascus slices were stained with hematoxylin-eosin and observed under optical and confocal microscopes (Model Primo Star and Model LSM780, ZEISS, Germany).

Extraction and preparation of genomic DNA

The frozen test samples, including the stroma, SFP (densely covered with ascocarps), and the 2 types of ascospores of C. sinensis, were individually ground to a powder in liquid nitrogen. The mycelia of pure fungi, including H. sinensis and Samsoniella hepiali (synonym Paecilomyces hepiali [70]) (gifts from Prof. Guo Y-L [26]), Geomyces pannorum, Penicillium chrysogenum and Pseudogymnoascus roseus (gifts from Prof. Zhang Y-J), and Tolypocladium sinensis (provided by Prof. Wu J-Y), were also individually ground to powder in liquid nitrogen; these fungi have been frequently detected in natural C. sinensis, and some of them reportedly show differential dominance in the stroma and caterpillar body of C. sinensis [23, 26–29, 34, 36–39, 71–73]. Genomic DNA was individually extracted from these powder samples using the DNeasy Plant Mini Kit (Qiagen) according to the manufacturer’s manual [19, 26].

Universal primers and genotype- and species-specific primers

Table 1 lists the sequences of the IST5/ITS4 universal primers, fungal species-specific primers and the genotype-specific primers for GC- and AT-biased genotypes of O. sinensis. The positions of the genotype-specific primers are shown in Fig 2. All primers were synthesized by Invitrogen Beijing Lab. or Beijing Bomaide Technology Co.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Alignment of the ITS sequences of GC- and AT-biased genotypes of O. sinensis with multiple transition point mutations.

GT represents genotype. Genotypes #1–3 are GC-biased O. sinensis genotypes, and Genotypes #4–6 and #15–17 are AT-biased O. sinensis genotypes. The sequence segments shown in blue correspond to the primers designed based on the sequences of GC-biased genotypes, and those in red correspond to primers designed based on the sequences of AT-biased genotypes. The underlined “GAATTC” site shown in green is the enzymatic site of the EcoRI endonuclease, and it is present in the GC-biased sequences at nucleotides 294–299 in Genotype #1 but absent in the AT-biased sequences due to a single-base mutation (GAATTT). “(RC)” denotes the reverse complement sequence of the primers; “-” represents identical bases; and spaces indicate unmatched sequence gaps.

https://doi.org/10.1371/journal.pone.0270776.g002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. ITS5/ITS4 universal primers and genotype- and species-specific primers used for the PCR amplification and sequencing of ITS segments.

https://doi.org/10.1371/journal.pone.0270776.t001

Among multiple pairs of species-specific primers that were designed for G. pannorum, H. sinensis, P. chrysogenum, P. roseus, S. hepiali and T. sinensis, the following primer pairs were selected according to their specificity and amplification efficiency as examined through PCR amplification and sequencing using the genomic DNA templates isolated from the fungal mycelia (listed in Table 1): Prp2/Prp5 for G. pannorum and P. roseus, Pcp3/Pcp7 for P. chrysogenum, Php4/Php6 for S. hepiali, and Tsp1/Tsp3 for T. sinensis. The ITS sequences of P. roseus (AY608922) and G. pannorum (JF320819 and DQ189229) are 98–99% homologous and were amplified using the same pair of primers, Prp2/Prp5 (Table 1).

PCR protocol for the amplification of ITS segments

The genomic DNA templates and aforementioned universal, genotype- and species-specific primers were used in PCR assays using a PCR instrument (Bio-Rad, United States) and Taq PCR reagent kit (Invitrogen) to amplify the ITS1-5.8S-ITS2 segments using the following touch-down protocol: (1) 95°C for 5 min; (2) 36 cycles of 95°C for 30 sec, annealing temperature for 30 sec (the annealing temperature was initially set to 70°C and decreased stepwise by 0.3°C in each cycle), and 72°C for 1 min; (3) 72°C for 10 min and final incubation at 4°C [26, 27]. The PCR amplicons were examined by agarose gel electrophoresis and sequencing.

Amplicon sequencing, cloning-based sequencing and sequence analysis

Each of the targeted PCR amplicons obtained from the aforementioned genomic DNA templates was recovered from agarose gels using a Gel Extraction Kit (Omega Bio-Tek) and purified using a Universal DNA Purification kit (TIANGEN) according to the manufacturer’s manuals [26, 27]. The purified amplicons were subjected to EcoRI endonuclease (BioLabs) digestion, which specifically digested the GC-biased genotypes of O. sinensis but not the AT-biased genotypes (cf. the underlined “GAATTC” site in green shown in Fig 2) and analyzed by agarose gel electrophoresis. The purified amplicons were sequenced either directly or after cloning by Invitrogen Beijing Lab.

For cloning, the amplicon was inserted into a PCR2.1 vector (SHQIMBIO, Shanghai, China), which was then transfected into DH5α cells, and the cells were coated on agar containing 100 μg/mL ampicillin or kanamycin in Petri dishes. The Petri dishes were cultured at 37°C, allowing the growth of cells; 30 white colonies were selected per dish, transferred to liquid culture medium and grown at 37°C in a shaking incubator. The expanded clones were sequenced using the M13F/M13R primers (cf. Table 1) (Invitrogen or Beijing Bomaide Technology Co.). The sequences were analyzed using Vector NTI Advance 9 software (Invitrogen) [26, 27].

Phylogenetic analysis of ITS sequences

All 17 genotypes of O. sinensis available in the GenBank database and obtained from this study were analyzed by Nanjing Genepioneer Biotechnologies Co. to reveal their phylogenetic relationships. A Bayesian majority-rule consensus tree was inferred using MrBayes v3.2.7a software (the Markov chain Monte Carlo [MCMC] algorithm; http://nbisweden.github.io/MrBayes/) with a sampling frequency of 10³ iterations after discarding the first 25% of samples from a total of 1.1x10⁸ iterations [74].

Two types of C. sinensis ascospores

To mimic the wild environment of the Qinghai-Tibetan Plateau, mature C. sinensis specimens were cultivated in paper cups in our Xining laboratory (altitude 2,200 m) (Fig 1A). White ascospores started to become visible on the surface of asci after approximately one week of cultivation. The fully ejected ascospores were collected on the double layers of autoclaved weighing paper (Fig 1B), and the numerous semiejected ascospores that tightly adhered to the outer surface of asci (Fig 1C after removing the upper layer of autoclaved weighing papers for collection of the fully ejected ascospores) were collected by gentle scraping using a disinfected inoculation shovel or ring [67].

Microscopy observations of the SFPs, ascocarps and ascospores of C. sinensis

Fig 3 shows the histology of the fully ejected C. sinensis ascospores without staining (upper panel; 40x) and after staining with calmodulin to visualize the septa of the multicellular structure of the ascospores (lower panel; 400x).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Microscopy images of fully ejected ascospores of C. sinensis without staining (upper panel; 40x) or after staining with 0.01% calmodulin for visualization of the septa of multicellular ascospores (lower panel; 400x).

https://doi.org/10.1371/journal.pone.0270776.g003

Fig 4A shows a confocal image (bar, 500 μm) of a transverse section of the fertile portion of the C. sinensis stroma, which is densely covered with multiple ascocarps. Fig 4B shows an optical microscopy image (10x) of several ascocarps, and Fig 4C shows a close-up optical microscopy image (40x) of an ascocarp. Ascocarps were stained with hematoxylin-eosin containing multiple ascospores, which revealed uneven staining of nuclei (dark blue–purple), consistent with the multicellular heterokaryosis of the C. sinensis ascospores reported by Bushley et al. [75]. Fig 4D and 4E provide confocal images close to the opening of the perithecia, showing ascospores gathering toward the opening of the perithecium (Fig 4D; bar, 50 μm) and a semiejected ascospore hanging out and adhered to the opening of the perithecium (Fig 4E; bar, 20 μm). Arrows in Fig 4B and 4E indicate semiejected ascospores.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Microscopy images of the SFP, ascocarps and ascospores of C. sinensis.

Fig 4A is a confocal image of a transverse section of the SFP (bar, 500 μm). Fig 4B is an optical microscopic image (10x) of several C. sinensis ascocarps stained with hematoxylin-eosin. Fig 4C is a close-up optical image (40x) of an ascocarp stained with hematoxylin-eosin. Fig 4D and 4E are close-up confocal images showing ascospores gathering toward the opening of the perithecium (Fig 4D; bar, 50 μm) and a semiejected ascospore hanging out of the opening of the perithecium (Fig 4E; bar, 20 μm). Arrows in Fig 4B and 4E indicate semiejected ascospores.

https://doi.org/10.1371/journal.pone.0270776.g004

ITS sequences amplified using universal primers

ITS sequences were amplified from the immature and mature stroma, SFP (with ascocarps), and fully and semiejected ascospores of C. sinensis using the universal primers ITS5/ITS4. The targeted amplicons of 630⁺ bp are 99% homologous to sequence AB067721 of H. sinensis (Genotype #1 of O. sinensis) [18, 23–24, 26–27, 31, 33, 35, 41, 45].

ITS sequences amplified using species-specific primers

ITS sequences were amplified from the genomic DNA templates obtained from immature and mature stroma, SFP (with ascocarps), and fully and semiejected ascospores of C. sinensis using the S. hepiali-specific primers Php4/Php6. The targeted amplicons of 460⁺ bp in the sample lanes shown in Fig 5 were recovered and sequenced, and the sequences were 100% homologous to the EF555097 sequence of S. hepiali [26]. As determined through PCR amplification of the same quantity of genomic DNA, significantly lower abundance of the moieties of the ITS amplicons was obtained after amplification from genomic DNA of the fully ejected ascospores (lanes 1 and 3 of Fig 5) than after amplification of genomic DNA of the semiejected ascospores (lanes 2 and 4).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Agarose gel electrophoresis of the PCR amplicons obtained from genomic DNA of the fully and semiejected ascospores of C. sinensis using the Samsoniella hepiali-specific Php4/Php6 primers.

Lane M shows the molecular weight standard. Lanes 1 and 3 display the amplicon moieties amplified from the genomic DNA of the fully ejected ascospores. Lanes 2 and 4 show the amplicons amplified from the genomic DNA of the semiejected ascospores.

https://doi.org/10.1371/journal.pone.0270776.g005

In other PCR experiments, the ITS sequences of various fungi were amplified from the SFP (with ascocarps) and fully and semiejected ascospores of C. sinensis using the following species-specific primer pairs (cf. Table 1): Prp2/Prp5 for G. pannorum (JF320819) and P. roseus (AY608922), Pcp3/Pcp7 for P. chrysogenum (DQ336710), and Tsp1/Tsp3 for T. sinensis (DQ097715) (cf. Table 1). Amplicon sequencing did not detect the sequences of G. pannorum, P. roseus, or T. sinensis. Agarose gel electrophoresis showed that the amplicons obtained using the primers Pcp3/Pcp7 migrated at a speed similar to the positive control moiety of P. chrysogenum, but sequencing analysis did not reveal the ITS sequence of P. chrysogenum. Thus, species-specific PCR primers failed to detect the ITS sequences of G. pannorum, P. chrysogenum, P. roseus, or T. sinensis in the SFP (with ascocarps) and fully and semiejected ascospores of C. sinensis.

Profiling of multiple O. sinensis genotypes in the compartments of C. sinensis using a cloning-based sequencing approach

Multiple genotypes of O. sinensis were further examined using genomic DNA obtained from the immature and mature stromata, SFP (with ascocarps), and 2 types of ascospores of C. sinensis. Multiple genotype-specific primers were designed and tested for specificity and amplification efficiency, and 5 pairs were selected for amplification of the ITS1-5.8S-ITS2 sequences of the GC- and AT-biased O. sinensis genotypes (cf. Table 1). The universal primers ITS5/ITS4 and the genotype-specific primer pairs Hsprp1/Hsprp2, Hsprp1/Hsprp3, or Hsprp2/Hsprp3 were highly homologous to GC-biased Genotypes #1–2 and/or #7. The primer design did not consider GC-biased Genotypes #3 and #8–12 because the sequences were obtained from C. sinensis specimens collected from Nyingchi of Tibet and the Indian-Nepal side of the C. sinensis production areas, which are geographically distant from the sample collection locations in Qinghai Province used for our study. The genotype-specific primer pairs HsATp1/ITS4, HsATp1/HsATp2, and HsATp1/HsATp3 were selected to amplify the AT-biased sequences of O. sinensis (cf. Table 1 and Fig 2). The amplification specificity and efficiency of the primer pairings have been previously tested repeatedly for the genotypes of O. sinensis. The PCR amplicons were subjected to cloning-based sequencing to explore the multiple mutant genotypes.

The primer pair HsATp1/ITS4 is highly homologous to almost all AT-biased genotype sequences (100%/100%) but slightly less homologous to Genotype #5 (100%/95%) and AB067719-type sequences (95%/100%). The primer pair HsATp1/HsATp2 is highly homologous to sequences of Genotypes #5, #15 and #17 (100%/100%) but less specific for Genotypes #4 and #16 (100%/80%). The primer pair HsATp1/HsATp3 is highly homologous to Genotype #4 and #16 sequences (100%/100%) but less specific for Genotypes #5, #15 and #17 sequences (100%/85%). The homology of the primer HsATp1 to the Genotype #6 sequence is unknown because all Genotype #6 sequences available in GenBank are short, probably due to the secondary structure/conformation within the ITS1 sequence close to its 5’ end and the ITS2 sequence close to its 3’ end. The primer HsATp2 is 100% homologous to sequence EU555436 of Genotype #6, but HsATp3 is less specific (85%) for EU555436. Thus, based on pretest and formal test results of the primer pairs in terms of the amplification specificity and efficiency, using all of these primer pairs in multiple PCR runs is essential for ensuring the successful amplification of all known genotypes of O. sinensis existing in the stroma, SFP (with ascocarps), and ascospores of C. sinensis.

Table 2 lists the differential occurrence of the O. sinensis genotypes with numerous transition point mutations detected in the immature and mature stromata, SFP (with ascocarps), and the 2 types of ascospores of C. sinensis. GC-biased Genotype #1 H. sinensis was identified in all compartments of C. sinensis. GC-biased Genotype #2 of O. sinensis was detected in both immature and mature stromata [27, 28].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Differential occurrence of multiple genotypes of O. sinensis, S. hepiali and AB067719-type fungus in the compartments of natural C. sinensis.

https://doi.org/10.1371/journal.pone.0270776.t002

Zhang et al. [55] and Cheng et al. [56] studied C. sinensis specimens collected from the Nyingchi area in Tibet and detected variable H. sinensis sequences corresponding to Genotype #3 of O. sinensis [7–12]. Chen et al. [58] detected a variable H. sinensis sequence AJ488254 (Genotype #7 of O. sinensis) in the stroma of a C. sinensis specimen (#H1023) collected from Qinghai Province of China, but Genotype #1 H. sinensis AJ488255 in the caterpillar body of the same specimen. In the current study, however, GC-biased Genotypes #3 and #7–12 were not detected in any compartments of natural C. sinensis collected from the Hualong and Yushu areas of Qinghai Province.

The cloning-based ITS amplicon sequencing approach showed the differential occurrence of the AT-biased sequences of Genotypes #4–6 and #15–17 of O. sinensis in the immature and mature stromata, SFP (with ascocarps), and 2 types of ascospores of C. sinensis (cf. Table 2). Table 2 also shows that S. hepiali and the AB067719-type fungus coexisted in the stromata, SFP (with ascocarps), and ascospores of C. sinensis.

Cloning-based amplicon sequencing also detected Genotypes #13–14 of O. sinensis. The Genotype #13 KT339190 sequence was detected in semiejected ascospores, whereas the Genotype #14 KT339178 sequence was detected in fully ejected ascospores (cf. Table 2). Table 3 compares the ITS1, 5.8S gene, and ITS2 segment sequences of the mutant genotypes with those of Genotype #1 H. sinensis (Group-A [66]) and the AB067719-type Group-E fungus. ITS1 of KT339190 and ITS2 of KT339178 are 100% homologous to those of Genotype #1 H. sinensis (AB067721). The 5.8S-ITS2 segments of KT339190 and ITS1-5.8S segments of KT339178 are 99–100% homologous to those of the AB067719-type fungus but 64.2–94.9% similar to those of the AB067721 of Genotype #1. As shown in Table 3 and Fig 6, the O. sinensis offspring Genotypes #13–14 resulted from large DNA segment reciprocal substitutions and genetic material recombination between the genomes of the 2 parental fungi, Group-A Genotype #1 H. sinensis and the AB067719-type Group-E fungus.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Schematic representation of the ITS segment sequences of the parental fungi (H. sinensis and the AB067719-type fungus) and O. sinensis offspring Genotypes #13–14.

The green bars indicate the ITS1 segment; the pink bars refer to the 5.8S gene; the blue bars represent the ITS2 segment. AB067719 [32] and KT339197 discovered in the current study represent the AB067719-type Group-E fungus and are shown with lighter bars. AB067721 [32] and KT339196 discovered in the current study represent Genotype #1 H. sinensis (Group-A by [66]) and are shown with darker bars. Alignment of the AB067719 and AB067721 sequences is shown between the lighter bars for AB067719 and the darker bars for AB067721. KT339190 and KT339178 represent O. sinensis offspring Genotypes #13 and #14, respectively, showing large DNA segment reciprocal substitutions and genetic material recombination between the genomes of the 2 parental fungi, H. sinensis and the AB067719-type fungus.

https://doi.org/10.1371/journal.pone.0270776.g006

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Sequence similarities of the ITS1, 5.8S and ITS2 segments of Genotype #1 and AB067719-type sequences compared with the multiple genotypes of O. sinensis.

https://doi.org/10.1371/journal.pone.0270776.t003

AB067719-type fungus in natural C. sinensis was grouped as a Group-E fungus by Stensrud et al. [66], unlike O. sinensis of Groups A–C. This fungus and 13 other sequences were annotated in the GenBank database as “C. sinensis” or “O. sinensis” under GenBank taxon 72228 for C. sinensis and O. sinensis. GenBank also collected more than 900 sequences highly homologous to AB067719, including Alternaria sp., Ascomycota sp., Aspergillus sp., Avena sp., Berberis sp., Colletotrichum sp., Cordyceps sp., Cyanonectria sp., Dikarya sp., Fusarium sp., Gibberella sp., Hypocreales sp., Juglans sp., Lachnum sp., Nectria sp., Nectriaceae sp., Neonectria sp., Penicillium sp., and many uncultured endophytic fungi. Thus, the identity of AB067719-type fungus in natural C. sinensis needs to be further determined through culture-dependent approaches and multigene and whole-genome sequencing of a purified AB067719-type fungus.

Phylogenetic relationship of the genotypes of O. sinensis

Kinjo & Zang [32] and Stensrud et al. [66] discussed the phylogenetic relationships of Genotype #1 (Group-A) and Genotypes #4–5 (Groups B–C) of O. sinensis. Further phylogenetic analysis of mutant genotypes #1–12 of O. sinensis, which share the same H. sinensis-like morphological and growth characteristics, reflected the research progress that had been achieved at the time [7–12, 16–19, 25, 27, 28, 55–59]. In this study, we found additional AT-biased Genotypes #15–17 and GC-biased Genotypes #13–14 that show large DNA segment reciprocal substitutions and genetic material recombination (cf. Tables 2 and 3 and Fig 6). The sequences of all 17 O. sinensis genotypes were subjected to phylogenetic analysis using a Bayesian algorithm (Fig 7).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Bayesian phylogenetic analysis of multiple genotypes of O. sinensis.

Five ITS sequences of the whole genomes (ANOV01021709, LKHE01000582, LWBQ01000008, JAAVMX010000002 and JAAVMX010000019) of H. sinensis strains (Co18, 1229, ZJB12195 and IOZ07) and 71 ITS sequences of 17 genotypes of O. sinensis were analyzed. The Bayesian majority-rule consensus tree was inferred using MrBayes v3.2.7a software (Markov chain Monte Carlo [MCMC] algorithm) [75]. GC-biased Genotypes #1–3 and #7–14 of O. sinensis are indicated in blue alongside the tree, and the AT-biased Genotypes #4–6 and #15–17 of O. sinensis are indicated in red alongside the tree. The AB067719-type Group-E sequences (5 sequences) are indicated in purple alongside the tree as an outgroup control.

https://doi.org/10.1371/journal.pone.0270776.g007

GC-biased genotypes of O. sinensis are indicated in blue alongside the tree in Fig 7, including Genotypes #1–3 and #8–14 and the ITS sequences from 5 whole-genome sequences of H. sinensis strains 1229, CC1406-203, Co18, IOZ07 and ZJB12195.

The AT-biased genotypes of O. sinensis are indicated in red alongside the tree and cluster into a clade that contains 2 branches. Cluster-A includes Genotypes #5–6 and #16–17, and Cluster-B includes Genotypes #4 and #15.

The AB067719-type Group-E fungus is side-noted in purple in Fig 7 as an outgroup control and is located outside of the blue GC-biased branch and the red AT-biased genotype clade.

Genotype #13 formed a phylogenetic leaf clustered close to the GC-biased branch. This leaf was distant from one of its parental fungi, Group-A Genotype #1 H. sinensis, but farther away from another parental fungus, the AB067719-type Group-E fungus.

Genotype #14 formed another phylogenetic leaf at a position closer to the parental AB067719-type Group-E fungus than to another parental fungus, Group-A Genotype #1 H. sinensis.

GC-biased Genotype #1 and AT-biased Genotypes #5–6 and #16, as well as S. hepiali and the AB067719-type fungus, were detected in the ascospores of C. sinensis. Genotypes #5–6 and #16 were located within AT-biased Cluster-A in the Bayesian tree (Fig 7). Genotypes #4 and #15 form AT-biased Cluster-B in the Bayesian tree and were detected in the stroma and SFP (with ascocarps) but not in the ascospores of C. sinensis (cf. Table 2, Fig 7).

Differential occurrence of multiple genotypes of O. sinensis and dynamic alterations of the genotypes during C. sinensis maturation

In addition to the 12 O. sinensis genotypes previously discovered [15, 16, 18, 24, 27, 28, 30–32, 47, 55–58, 64, 66], the current study identified 5 new genotypes of O. sinensis: AT-biased Genotypes #15–17 with multiple transition point mutations and GC-biased Genotypes #13–14 that exhibit the characteristics of large DNA segment reciprocal substitutions and genetic material recombination in natural C. sinensis. These genotypes occurred differentially in the stroma, SFP (with ascocarps), and 2 types of ascospores of C. sinensis (cf. Table 2). The biomasses of the GC- and AT-biased genotypes undergo dynamic alterations in an asynchronous, disproportional manner during C. sinensis maturation, as demonstrated using a Southern blotting approach without DNA amplification [18, 27].

The ITS sequences of Genotype #1 were the most easily amplifiable and detectable using universal primers, but this finding may not necessarily imply dominance of the Genotype #1 DNA template in the test samples [7]. Southern blotting analysis without DNA amplification demonstrated that AT-biased genotypes were dominant in the stroma of natural C. sinensis during maturation, and GC-biased Genotype #1 was never the dominant fungus in the stroma [27, 28]. The easy pairing and elongation of the Genotype #1 ITS sequences using universal primers were observed due to (1) the sequence identity levels and (2) the absence of secondary structure/conformation in their sequences that could interfere with PCR amplification and sequencing [7]. In early molecular studies on C. sinensis (1999–2006), mycologists reported the detection of a single sequence of H. sinensis and therefore incorrectly judged H. sinensis to be the sole anamorph of O. sinensis.

Chen et al. [34] first reported the molecular heterogeneity of C. sinensis-associated fungi of the genera Hirsutella, Paecilomyces and Tolypocladium using a PCR amplicon cloning-sequencing methodology. However, insufficient attention was given to the “all-or-none” qualitative research technique and findings, and instead, the disproportionate amplicon clones selected for examination of the ITS sequences of different fungi were overemphasized, which unfortunately led to the improper conclusion that H. sinensis was the sole anamorph of O. sinensis. As demonstrated herein, we applied multiple pairs of genotype- and species-specific primers and cloning-based amplicon sequencing approaches with selection of at least 30 white colonies per Petri dish in cloning experiments, which enabled us to profile the components of the heterogeneous metagenome in the immature and mature stromata, SFP (with ascocarps), and ascospores of natural C. sinensis. In particular, this approach allowed us to identify those genotypes and fungi with low abundance and those with secondary structure/conformation in their DNA sequences, which may affect primer binding and DNA chain elongation during PCR amplification and sequencing [7, 10].

GC-biased Genotypes #1 and #2 coexisted in the stroma of natural C. sinensis, and their amplicon abundance was low in the immature stroma (1.0–2.0 cm in stromal height) [27, 28]. Genotype #1 showed a greatly elevated abundance in the maturing stroma (4.0–4.5 cm in height without an expanded fertile portion close to the stromal tip), and its abundance then plateaued in the mature stroma (6.5–7.0 cm in height, showing the formation of an expanded fertile portion close to the stromal tip, which is densely covered with ascocarps) (Fig 8; modified from Fig 6 of [28] in Chinese). In contrast to the maturation pattern of Genotype #1, the abundance of Genotype #2 remained at a low level in the maturing stroma and increased markedly in the mature stroma to a level that exceeded the abundance of Genotype #1. The cooccurrence of Genotypes #1 and #2 in the stroma with distinct maturation patterns indicates the genomic independence of the 2 GC-biased genotypes, which is proven by the absence of Genotype #2 sequences in the whole-genome sequences of 5 Genotype #1 H. sinensis strains [7–12, 49–53, 63, 64].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 8. Dynamic alterations of the abundance of the amplicons of Genotypes #1 and #2 of O. sinensis in the stromata of C. sinensis during maturation (modified from Fig 6 of [28]).

Immature C. sinensis had very short stroma of 1.5 cm. Maturing C. sinensis had stroma of 4.0 cm without an expanded fertile portion close to the stroma tip. Mature C. sinensis had long stroma of 7.0 cm and showed the formation of an expanded fertile portion close to the stroma tip, which was densely covered with ascocarps.

https://doi.org/10.1371/journal.pone.0270776.g008

In this study, we did not detect GC-biased Genotypes #3 and #7–12 of O. sinensis in the compartments of natural C. sinensis specimens collected from the Hualong and Yushu areas of Qinghai Province. This may indicate that these GC-biased genotypes of O. sinensis may exist in the natural C. sinensis specimens in different production areas.

A Bayesian phylogenetic analysis (cf. Fig 7) demonstrated that the AT-biased genotypes were clustered into 2 branches, Clusters A and B. AT-biased genotypes from both branches were detected in the stroma and SFP of C. sinensis, but Genotypes #4 and #15 of Cluster-B were absent in the ascospores, consistent with those discovered by Li et al. [59]. Genotype #4 of AT-biased Cluster-B was the dominant AT-biased genotype in the immature stroma during the asexual growth stage and coexisted with less dominant Genotypes #5–6 of AT-biased Cluster-A [16, 17]. The abundance of Genotype #4 of Cluster-B markedly decreased in the mature C. sinensis stroma and remained at a low level in the SFP (with ascocarps) during the sexual growth stage. In contrast, the abundance of Genotype #5 of AT-biased Cluster-A increased reciprocally and significantly in the mature C. sinensis stroma and SFP and predominated in the O. sinensis ascospores [16, 17, 68]. In addition, Genotype #15 of Cluster-B predominated in the SFP (with ascocarps) of C. sinensis prior to ascospore ejection and drastically declined after ejection [68]. The differential cooccurrence and alternative predominance of GC- and AT-biased genotypes of O. sinensis indicate their distinct physiological functions at different maturational stages of the C. sinensis lifecycle and their genomic independence, belonging to independent O. sinensis fungi.

Engh [30], Kinjo & Zang [32], Wei et al. [47] and Mao et al. [57] reported the detection of either Genotype #4 or #5 of O. sinensis, without codetection of GC-biased Genotype #1 O. sinensis, in some natural C. sinensis specimens collected from geographically distant areas and in cultivated C. sinensis. These findings likely indicate that the studied specimens might be in different maturational stages and that the amplicons of other cocolonized O. sinensis genotypes and fungal species might be overlooked during PCR amplification and sequencing due to lower-abundance amplicons caused by the experimental designs without using genotype- and species-specific primers or cloning-based amplicon sequencing strategies or due to secondary structures present in the sequences of O. sinensis genotypes, which may interfere with primer binding, DNA chain elongation, and sequencing [7].

In this study, we detected several AT-biased genotypes of Cluster-A and GC-biased Genotypes #1 H. sinensis and #13–14 in ascospores that cooccurred with S. hepiali and the AB067719-type fungus in the teleomorphic ascospores of natural C. sinensis. However, AT-biased Cluster-B genotypes were absent in ascospores but present in the SFP (with ascocarps). Pseudogymnoascus roseus, Geomyces pannorum, Tolypocladium sinensis, and Penicillium chrysogenum were not detected in the SFP (with ascocarps) and ascospores of C. sinensis despite their differential predominance in the stroma and caterpillar body of natural C. sinensis found in 2 previous mycobiota studies [38, 39].

The sequences of Genotypes #2–17 of O. sinensis, however, were not present in the genome assemblies (ANOV00000000, JAAVMX000000000, LKHE00000000, LWBQ00000000 and NGJJ00000000) of 5 Genotype #1 H. sinensis strains, Co18, IOZ07, 1229, ZJB12195 and CC1406-203, and instead belonged to the genomes of independent O. sinensis fungi [7–12, 27, 28, 49–54, 63, 64, 68]. These genomic independence findings support the hypotheses of independent O. sinensis fungi and an integrated microecosystem for natural C. sinensis [14, 24]. During the complex lifecycle of natural C. sinensis, its metagenomic members appear to function mutually and symbiotically at different development-maturation stages in a dynamically alternating manner from the immature to the maturing and then the mature stages and from initial asexual growth to sexual growth and reproduction.

Complex anamorph-teleomorph connection of O. sinensis

The anamorph-teleomorph connection of O. sinensis has been the subject of decades-long academic debate [7–12, 29, 60]. Wei et al. [41] hypothesized that H. sinensis is the sole anamorph of O. sinensis based on the 3 aforementioned sets of evidence; this hypothesis satisfied the requirements of the first and second criteria of Koch’s postulates, but unfortunately not the third and fourth criteria, making the hypothesis uncertain [7–12, 29, 54, 63, 64]. Although this hypothesis has been widely appreciated, Wei et al. [41] demonstrated significant genetic diversity of O. sinensis teleomorphs (cf. the far phylogenetic distance with a low bootstrap value between those of natural C. sinensis specimens G3 and S4, shown in Fig 6 of [41]).

Bushley et al. [75] discovered the multicellular heterokaryotic structure of the hyphae and ascospores of C. sinensis, which contain multiple mononucleated, binucleated and trinucleated cells. Li et al. [59], the research partners in the study [75], reported the identification of both GC-biased Genotype #1 and AT-biased Genotype #5 in 8 of 15 clones derived from the 25-day culture of a C. sinensis mono-ascospore. Based on insufficient and contradictory evidence, these researchers overinterpreted that all AT-biased genotypes were “ITS pseudogene” components of the genome of Genotype #1 H. sinensis, in agreement with the sole anamorph hypothesis for H. sinensis [41]. However, AT-biased genotype sequences are not present in the 5 whole-genome assemblies (ANOV0000000, JAAVMX0000000000, LKHE00000000, LWBQ00000000 and NGJJ00000000) of Genotype #1 H. sinensis strains Co18, IOZ07, 1229, ZJB12195 and CC1406-203 [7–10, 49–53, 64]. The genomic independence evidence disproved the “ITS pseudogene” hypothesis for the AT-biased genotypes of O. sinensis proposed by Li et al. [59].

In contrast, Barseghyan et al. [73] proposed a dual-anamorph hypothesis for O. sinensis involving Tolypocladium sinensis and H. sinensis using macro- and micromycological approaches, but it leaves an unresolved question regarding the genotype of O. sinensis with H. sinensis-like morphological and growth characteristics. Engh [30] hypothesized that C. sinensis and T. sinensis form a fungal complex in the natural insect-fungal complex, and the C. sinensis sequence AJ786590 was disclosed and uploaded to the GenBank database by Stensrud et al. [65]. Stensrud et al. [66] clustered AJ786590 and other sequences into Group-B of O. sinensis using a Bayesian phylogenetic approach and concluded that Group-B (AT-biased Genotype #4) sequences are phylogenetically distinct from GC-biased Group-A of O. sinensis (Genotype #1 H. sinensis).

Kinjo & Zang [32], Engh [30], Wei et al. [47] and Mao et al. [57] detected AT-biased Genotype #4 or #5 of O. sinensis from natural C. sinensis specimens collected from different geographic regions or from cultivated C. sinensis. Zhang et al. [55] and Cheng et al. [56] reported the detection of GC-biased Genotype #3 of O. sinensis in natural C. sinensis collected from Nyingchi in Tibet. Chen et al. [58] reported the detection of GC-biased Genotype #7 of O. sinensis from the stroma of C. sinensis specimens collected from Qinghai Province of China. These mutant genotypes of O. sinensis share the same H. sinensis-like morphological and growth characteristics without the cooccurrence of Genotype #1 H. sinensis, although they may share a common evolutionary ancestor [30, 32, 55–58, 66]. In the current study, we detected AT-biased Cluster-A genotypes (cf. Fig 7) and GC-biased Genotypes #1 and #13–14 in the teleomorphic ascospores of C. sinensis, along with S. hepiali and the AB067719-type fungus, and additional AT-biased Cluster-B genotypes in the stroma and SFP (with ascocarps) of C. sinensis.

Li et al. [48] recently reported the differential occurrence and transcription of the mating-type genes of the MAT1-1 and MAT1-2 idiomorphs in 237 H. sinensis strains, indicating genetic and transcriptional inability to perform self-fertilization under homothallic or pseudohomothallic reproduction, as proposed by Hu et al. [49] and Bushley et al. [75]. Although Bushley et al. [75] reported the transcription of both the MAT1-1-1 and MAT1-2-1 genes of the MAT1-1 and MAT1-2 idiomorphs, the MAT1-2-1 transcript of H. sinensis strain 1229 harbored unspliced intron I, which contains 3 stop codons. This type of transcript presumably produces a truncated and dysfunctional MAT1-2-1 protein missing the majority portion of the protein encoded by exons II and III [48, 76]. These findings constitute a coupled transcriptional-translational mechanism of H. sinensis reproductive control, in addition to the controls at the genetic and transcriptional levels. The differential occurrence and transcription of the mating genes of the MAT1-1 and MAT1-2 idiomorphs indicate that H. sinensis requires a sexual partner to perform physiologically heterothallic reproduction in the lifecycle of natural C. sinensis regardless of whether H. sinensis is monoecious or dioecious [48, 76]. For example, H. sinensis strains 1229 and L0106 appear to reciprocally produce functional MAT1-1-1 and MAT1-2-1 transcripts and presumably complementary mating proteins [75, 77], possibly constituting a pair of sexual partners for physiological heterothallism [48, 76]. In addition, the transcripts of mating-type genes in S. hepiali strain Feng [78] were coincidentally found to be complementary to those of H. sinensis strain L0106 [77]. The coincident transcriptomic findings reveal a possible transcriptomic mechanism for fungal hybridization, which triggers further investigation because H. sinensis is closely associated with a small quantity of S. hepiali in the compartments of natural C. sinensis, often resulting in difficulties in fungal isolation and purification even by top-notch mycology taxonomists [79].

Wei et al. [47] reported an industrial artificial cultivation project in which 3 anamorphic H. sinensis strains, 130508-KD-2B, 20110514 and H01-20140924-03, were reportedly used as inoculants. The successful cultivation of C. sinensis is important for supplementing the increasingly scarce natural resources of C. sinensis and for meeting the third criterion of Koch’s postulates, adding academic value to supplement the 3 aforementioned sets of evidence [41]. However, these authors reported the identification of the sole teleomorph of O. sinensis in the fruiting body of cultivated C. sinensis, and it belonged to AT-biased Genotype #4. Thus, the apparent fungal mismatch between the inoculants used in artificial cultivation and the final cultivated product resulted in failure to meet the requirement of the fourth criterion of Koch’s postulates. In addition to the detection of teleomorphic AT-biased Genotype #4 in cultivated C. sinensis, Wei et al. [47] also reported the detection of teleomorphic GC-biased Genotype #1 in natural C. sinensis. Because the sequence of AT-biased Genotype #4 is absent in the 5 whole-genome assemblies (ANOV00000000, JAAVMX000000000, LKHE00000000, LWBQ00000000 and NGJJ00000000) of the GC-biased Genotype #1 H. sinensis strains Co18, IOZ07, 1229, ZJB12195 and CC1406-203 [49–53], the scientific evidence reported by Wei et al. [47] disproves the sole anamorph and sole teleomorph hypotheses for O. sinensis, which were proposed 10 years ago by the same group of key authors [41]. The species mismatch reported by Wei et al. [47] may imply (1) that the researchers could have overlooked the Genotype #4 sequence in one or all of the anamorphic inoculant strains, which would confirm previous findings that GC-biased Genotype #1 H. sinensis (Group-A) and AT-biased genotypes (Groups B and C) formed a species complex [66] and that the actual causative agent is a fungal (species) complex containing several O. sinensis genotypes and S. hepiali [79]; or (2) that secondary or sequential infections by the true causal fungus/fungi (Genotype #4 and possibly other genotypes of O. sinensis) occurred in the course of artificial cultivation. A third possibility is that a series of preprogrammed, nonrandom mutagenic conversions of GC-biased Genotype #1 to AT-biased Genotype #4 occurred during artificial cultivation without exception in all cultivated C. sinensis products (and vice versa after the ejection of the ascospores); however, this possibility seems the least likely.

Zhang et al. [80] summarized nearly 40 years of artificial cultivation experience with failed induction of fruiting body and ascospore production in research-oriented academic settings, either in fungal cultures or after infecting insects with fungal inoculants. Distinct from the success of artificial C. sinensis cultivation in product-oriented industrial settings, as reported by Wei et al. [47], Hu et al. [49] inoculated 40 larvae of Hepialus sp. with 2 pure H. sinensis strains (Co18 and QH195-2) via the injection of a mycelial mixture through the second larval proleg, which caused death and mummification of the larvae but failed to induce the production of a fruiting body. Li et al. [79] inoculated 400 larvae of Hepialus armoricanus with 4 groups of inoculants (n = 100 larvae per inoculant): (1) conidia of H. sinensis, (2) mycelia of H. sinensis, (3) purified ascospores of C. sinensis, and (4) a mixture of 2 wild-type fungal strains, CH1 and CH2, isolated from the intestines of healthy living larvae of Hepialus lagii Yan [81]. Both strains CH1 and CH2 showed H. sinensis-like morphological and growth characteristics and contained GC-biased Genotype #1 H. sinensis and S. hepiali as well as highly abundant AT-biased Genotype #5 of O. sinensis and the markedly less abundant Genotypes #4 and #6 [79]. The application of a mycelial mixture of strains CH1 and CH2 as the inoculant resulted in a favorable infection-mortality-mummification rate of 55.2±4.4%, indicating 15–39-fold greater potency than the infection-mortality-mummification rates (1.4–3.5%; P<0.001) achieved after inoculation with the conidia or mycelia of H. sinensis or ascospores of C. sinensis. These findings reported by Li et al. [48, 79], Wei et al. [47], Hu et al. [49], and Zhang et al. [80] suggest that GC-biased Genotype #1 H. sinensis may not be the sole true causative fungus in natural C. sinensis and that inoculation synergy of the symbiotic fungal species may be needed to initiate the development of the stromal primordia and fruiting bodies, and sexual partners, regardless of whether they are the same or interspecific species, may be needed to induce the transition from asexual to sexual growth and reproduction and the maturation of the teleomorphic ascocarps and ascospores during the lifecycle of natural and cultivated C. sinensis.

The microcycle conidiation of C. sinensis ascospores using culture protocols favoring the growth of H. sinensis has also been reported [42–44]. Bushley et al. [75] demonstrated that C. sinensis ascospores are multicellular and heterokaryotic, consisting of multiple monokaryotic, bikaryotic and trikaryotic cells. Our study herein demonstrated that the fully ejected C. sinensis ascospores contain GC-biased Genotypes #1 and #14, AT-biased Genotypes #5–6 and #16 within AT-biased Cluster-A, S. hepiali and the AB067719-type fungus (cf. Table 2 and Fig 7). These findings suggest a multicellular heterokaryotic structure and genetic heterogeneity of ascospores [7–10]. However, the culture-dependent approach detected only GC-biased Genotype #1 and AT-biased Genotype #5 after 25-day liquid inoculation of mono-ascospores [59], suggesting the possibility of overlooking several AT-biased genotypes and GC-biased Genotypes #14, probably due to nonculturability of those genotypes or inappropriately designed and executed molecular techniques. The nonculturable nature of most genotypes of O. sinensis fungi examined to date [60–62] calls into question whether all genotypes of O. sinensis in all cells of multicellular heterokaryotic ascospores are capable of undergoing conidiation through in vitro microcycle conidiation under the commonly used experimental conditions that favor the growth of Genotype #1 H. sinensis. The lack of molecular information on conidia in these studies [42–44] makes the fungal/genotypic identity of the conidia uncertain due to the H. sinensis-like morphological and growth characteristics shared by multiple genotypes of O. sinensis [30, 32, 47, 55–59, 75, 79] and the Hirsutella-like morphology shared by numerous fungal species in the families Ophiocordycipitaceae and Clavicipitaceae and the genera Polycephalomyces and Harposporium [82].

The study described herein further revealed fungal and genotypic heterogeneity in the multicellular heterokaryotic ascospores of C. sinensis. AT-biased Cluster-A genotypes, GC-biased Genotype #1 H. sinensis and Genotypes #13 or #14 (showing large DNA segment reciprocal substitutions and genetic material recombination) in either type of ascospores, S. hepiali and the AB067719-type Group-E fungus naturally cooccur in teleomorphic C. sinensis ascospores.

Differential occurrence of Genotypes #4 and #15 of AT-biased Cluster-B

Studies of natural and cultivated C. sinensis have detected Genotype #4 of O. sinensis, but not coexisting GC- and AT-biased genotypes of O. sinensis, including AJ786590 discovered by Engh [30] and uploaded to GenBank by Stensrud et al. [65], AB067741–AB067749 identified by Kinjo & Zang [32], KC305891–KC305892 discovered by Mao et al. [57], and 2 sequences (“Fruiting body 3” and “Mycelia 3”, sequences unavailable in GenBank) with high homology to AB067749 and KC305892 found by Wei et al. [47]. These studies did not report the maturation stages of natural and cultivated C. sinensis specimens that were used as the study materials. However, other studies using genotype-specific primers, amplicon cloning and biochip-based SNP mass spectrometry genotyping techniques reported the detection of Genotype #4 of O. sinensis in natural C. sinensis coexisting with Genotype #1 H. sinensis with or without Genotypes #5–6 [16–18, 24, 27, 28, 59, 68].

Genotype #4 of AT-biased Cluster-B was found to predominate in the stroma of immature C. sinensis and to decline during C. sinensis maturation (cf. Fig 7) [16, 17, 27, 28, 63, 64, 68]. Genotype #4 and the newly discovered Genotype #15 of AT-biased Cluster-B are present in the SFP (with ascocarps) but not in ascospores of natural C. sinensis (cf. Table 2) [68].

Fungal factors involved in the control of the development, maturation and ejection of C. sinensis ascospores

In this study, we report the discovery of 2 types of ascospores of C. sinensis, namely, fully and semiejected ascospores. In contrast to the well-described fully ejected ascospores, semiejected ascospores show tight adhesion to the outer surface of asci, hanging out of the perithecial opening (cf. Figs 1C, 4B and 4E). The distinct ejection behaviors of ascospores are associated with divergent fungal compositions. In addition to the cooccurrence of GC-biased Genotype #1, AT-biased Cluster-A genotypes and the AB067719-type fungus in the 2 types of ascospores, the semiejected ascospores contain Genotype #13 of O. sinensis and a greater abundance of S. hepiali, whereas the fully ejected ascospores contain Genotype #14 and a markedly lower abundance of S. hepiali (cf. Fig 5 and Table 2). Genotypes #13–14 show alternating reciprocal substitutions of large DNA segments between the genomes of 2 parental fungi, namely, Genotype #1 H. sinensis (Group-A by [66]) and the AB067719-type Group-E fungus (cf. Table 3 and Fig 6), indicating that the biological processes of plasmogamy, karyogamy, chromosomal intertwining interaction, and genetic material recombination occur differentially between the 2 parental fungi regardless of the processes of hereditary variation caused by fungal hybridization or parasexual reproduction [9, 48]. The divergent fungal components with altered abundances may participate in the control of the development, maturation and ejection of C. sinensis ascospores.

Genotypes #6 and #16 of AT-biased Cluster-A were detected in fully ejected ascospores but not in semiejected ascospores using genotype-specific primers and cloning-based amplicon sequencing (cf. Table 2; Fig 7). The differential occurrence of these 2 AT-biased genotypes may need to be further verified due to the possible existence of secondary structures/conformation in the DNA sequences that may affect primer binding and DNA chain elongation during PCR amplification and sequencing [7–10]. In fact, Genotypes #6 and #16 were detected in both types of ascospores using the biochip-based MassARRAY SNP mass spectrometry technique [68].