First characterization of the complete mitochondrial genome of fungal plant-pathogen Monilinia laxa which represents the mobile intron rich structure

Monilinia laxa is an important fungal plant pathogen causing brown rot on many stone and pome fruits worldwide. Mitochondrial genome (mitogenome) plays a critical role in evolutionary biology of the organisms. This study aimed to characterize the complete mitogenome of M. laxa by using next-generation sequencing and approaches of de novo assembly and annotation. The total length of the mitogenome of M. laxa was 178,357 bp, and its structure was circular. GC content of the mitogenome was 30.1%. Annotation of the mitogenome presented 2 ribosomal RNA (rRNA) genes, 32 transfer RNA genes (tRNA), 1 gene encoding mitochondrial ribosomal protein S3, 14 protein-coding genes and 15 open reading frame encoding hypothetical proteins. Moreover, the group I mobile introns encoding homing endonucleases including LAGLIDADG and GIY-YIG families were found both within coding regions (genic) and intergenic regions of the mitogenome, indicating an enlarged size and a dynamic structure of the mitogenome. Furthermore, a comparative mitogenomic analysis was performed between M. laxa and the three closely related fungal phytopathogen species (Botryotinia fuckeliana, Sclerotinia sclerotiorum and, S. borealis). Due to the number and distribution of introns, the large extent of structural rearrangements and diverse mitogenome sizes were detected among the species investigated. Monilinia laxa presented the highest number of homing endonucleases among the fungal species considered in the analyses. This study is the first to report a detailed annotation of the mitogenome of an isolate of M. laxa, providing a solid basis for further investigations of mitogenome variations for the other Monilinia pathogens causing brown rot disease.

. Circular map of the mitogenome of Monilinia laxa representing with introns (A) and, with group I mobile introns (HEGs) within genic regions (B) and intergenic regions (C). Genes are visualized by arrows, which are presented in a clockwise direction (forward). Blue arrows: protein-coding genes, orange arrows: hypothetical open reading frames, green arrows: genes of large and small ribosomal subunits, yellow arrow: gene encoding ribosomal protein 3, red arrows: genes of transfer RNAs, grey arrow: introns, pink arrows: HEGs within genic regions, light blue arrows: HEGs within intergenic regions. Circular mitogenomes were generated by using the Geneious 9.1.8 software 23 .

Results
General features and gene content of the mitogenome of the brown rot fungal pathogen Monilinia laxa. The mitogenome characterized in this study was submitted to NCBI GenBank with the accession number MN881998. The length of the mitogenome of M. laxa isolate Ni-B3-A2 was 178,357 bp, and included a large number of repeated sequences and many different introns (Fig. 1). The overall information about the mitogenome of M. laxa was as follows: T: 34.7%, C: 13.5%, A: 35.2%, and G: 16.6% and the content of GC is 30.1% with A + T-rich feature. The genome had 29 protein-coding genes (PCGs) including open reading frames for hypothetical proteins, and 14 of the coding genes were related to oxidative phosphorylation system and electron transport which were cob, cox1, cox2, cox3, nad1, nad2, nad3, nad4, nad4L, nad5, nad6, atp6, atp8, and atp9 ( Fig. 1; Table 1). Besides, 15 open reading frames encoded hypothetical proteins which were described as orf139, orf126, orf213, orf199, orf111, orf149, orf101, orf174, orf179, orf117, orf109, orf100, orf99, orf185. All the coding genes and open reading frames represented once except for orf99, which was named twice in the genome, but for non-homologous sequences.
All 29 annotated PCGs had the same direction, and their start codon was ATG ( Table 1). The preferred stop codons were TGA for the 11 PCGs, TAA for the 12 PCGs, TAG for the 4 PCGs, and TCG for 1 PCG (Table 1). Only nad6 started with the translation initiation codon ATT and stopped with the codon ACT (Table 1). This gene also contained the highest AT frequency and the lowest GC contents (22.6%) among all the PCGs ( Table 1).
Most of the genes were interrupted by introns (which are non-coding ones), as shown in Fig. 1 and Table 2. The 33 introns identified were in the core mitochondrial PCGs (Table 2). Seven introns were found in the cox2 gene, accounting for 65.4% of the total length of the gene, which was the most intron rich gene. Intron content changed among the genes (Table 2). However, nad3, nad4L, nad6, atp8-9 genes contained no intron ( Table 2).
Genes of the small and large ribosomal RNA (rRNA) subunits (rns and rnl, respectively) were identified. The sequence region of rnl was also invaded by introns encoding homing endonucleases (Fig. 1). Moreover, one ribosomal protein-coding gene (rps3) was determined.
Genome structure and order of the genes were presented in Fig. 1. All the genes identified and some information about their positions, products, lengths, GC contents, start and stop codons were described in Table 1.

Mobile introns in the mitogenome of Monilinia laxa.
A total of one hundred and nineteen different mobile introns were annotated in the mitogenome of M. laxa (Table 3). All encoding introns were characterized as the group I intron type, which encodes homing endonucleases (HE) ( Table 3). Among those, the eighty-nine belonged to LAGLIDADG family, and the thirty belonged to GIY-YIG family, and were distributed within genic regions as well as intergenic regions (Fig. 1). The start codons of these elements were highly variable, but stop codons were mostly TAA, TGA, TAG (Table 3). Moreover, only two representative different stop codons were identified as ACT and AGT (Table 3). All sequences were represented once, and homology was not found among the sequences within each family. Seventy of the HE genes occupied within the intragenic regions of the mitogenome (Fig. 1). Most of these introns occupied the cox1, cox2, cox3, cob, nad1, nad5 genes. The longest mobile intron sequence with 1,320 bp length was identified within cox2 gene (Table 3). On the other hand, some genes (atp9, atp8, nad3, nad4L, nad6) did not show a mobile intron invasion. Sequence lengths, start-stop codons and the main location in the genome were represented for the HEs in Table 3. Locations of the different families of group I introns were illustrated in Fig. 1. Transfer RNAs in the mitogenome of M. laxa. A total of 32 tRNAs associated with essential 19 amino acids were found in the mitogenome of M. laxa (Fig. 2). Coding for Cysteine amino acid was absent in the mitogenome of M. laxa. Several tRNAs were present with more than one copy: trn-Arg (4 copies), trn-Ser (3 copies), trn-Trp (3 copies), trn-Met (3 copies), trn-Lys (2 copies), trn-Gly (2 copies), trn-Asp (2 copies), trn-Leu (2) by representing different anticodon sequences (Fig. 3). Genes coding tRNAs were mostly clustered closely on the mitogenome (Fig. 1). One of the main tRNA cluster was observed in proximity of the rnl and rsp3 genes, both involved in the ribosome construction process. Due to presence/absence of the extra arms, as shown in Fig. 3, tRNA sequence lengths were variable and ranged between 71 bp (trnT, trnW, trnG, trnK, trnR, trnK) and 86 bp (trnS). comparative analyses between the mitogenomes of M. laxa and some closely related species. Mitogenomes Table 4). The GC content was the lowest in the mitogenome of B. fuckeliana (29.9%) and the highest in the mitogenome of S. borealis (32.9%) ( Table 4). All mitogenomes consisted of the core genes of mitogenomes ( Table 4). The number of tRNAs varied among the species (Table 4). Genome organization was represented in Fig. 4 The conserved gene orders were the same among the species as following; cox1, nad4, cob, atp9, nad1, atp8, atp6, nad2, nad3, cox2 (Fig. 4). Genome sizes differed among the four mitogenomes and ranged from 82 to 203 kb (Table 4). Sclerotinia borealis had the highest intron content, which makes the largest mitogenome size in comparison to the other species (Table 4). However, the mitogenome of M. laxa presented the highest content of mobile introns when compared to the other three closely related species (Table 4).
Repetitive sequences in the mitogenomes. Repetitive sequences detected in the mitogenomes were variable among the four species. Sclerotinia borealis and M. laxa presented the high number of repeats with total numbers 62 and 60, respectively ( Table 4). Numbers of repeats were 27 and 20 in the mitogenomes of S. sclerotiorum and B. fuckeliana, respectively. The longest repeats of more than 10 bp in length were (AT) 17 in the M. laxa mitogenome and this repetitive element was in an intron of cob gene. Botryotinia fuckeliana and S. borealis presented the same type of repetitive element, (T) 35 and (T) 34 , respectively (Table 4). However, the location of these repetitions changed in the two species. The repetitive element located in the intergenic region between the genes tRNA-Leu and tRNA-Ale in B. fuckeliana while in S. borealis the repetitive element was found in the intergenic region between the genes cox2 and nad4L. These longest repetitive sequences were found once in the

Discussion
Currently, more than 700 complete fungal mitochondrial genomes are available, but the mitogenomes of Monilinia species have not been reported in the organelle genome of the NCBI database. Mitogenome of M. fructicola was recently announced by Ma et al., 26 but since this genome was not found in NCBI-blast searches, we did not use it in this study. According to the NCBI organelle genome database search, the mitogenome of M. laxa (isolate Ni-B3-A2) with 178,357 bp is one of the largest fungal mitochondrial genomes. Expansion of the mitogenome size has been driven by the accumulation of introns, mobile introns such as HEs, hypothetical genes, and repeats regions. Thus, increasing information about mitogenomes of plant pathogenic fungi is quite valuable. Group I introns were firstly detected in this pathogen, and these mobile elements acting as ribozyme may contribute variations within this species. Ribosomal protein-coding genes are occasionally present in fungal mitogenomes [27][28][29] . Rps3 encodes protein S3, which contributes to small ribosome assembly and this gene was identified in the isolate of M. laxa in this study. Rps3 has been reported in several fungal mitogenomes, and its homolog genes have been found also in the nuclear genome for the others [28][29][30] . The sequence similarity and location of rps3 are quite variable among fungal species 28,30 . In some fungal species, such as Ophiostoma ulmi 31 , rnl was detected within rnl group I intron 28 . The complete structure of rps3 was not interrupted by any intron in the sequenced mitogenome of M. laxa. Rps3 is LAGLIDADG  252  AAT  TGA  Intergenic region   LAGLIDADG  375  ACT  TAA  Intergenic region   LAGLIDADG  405  ATA  TGA  Intergenic region   LAGLIDADG  378  TTT  TGA  Intergenic region   LAGLIDADG  420  AAT  TGA  Intergenic region   LAGLIDADG  366  AAG  TGA  Intergenic region   GIY-YIG  1,032  GAG  TGA  Intergenic region   Table 3. The distribution and main locations of the LAGLIDADG and GIY-YIG homing endonucleases in the mitogenome of Monilinia laxa.   www.nature.com/scientificreports/ highly interesting marker to evaluate evolutionary dynamics of fungal mitogenomes due to the high variability of its sequence (length, location, and rearrangement), presence, and invasion by homing endonucleases 28 . Alternative start codon (ATT) was identified for nad6 gene in the mitogenome of M. laxa in this study. These codons are suggested for mitochondrial DNA by the NCBI Genbank (https ://www.ncbi.nlm.nih.gov/Taxon omy/ Utils /wprin tgc.cgi#SG4). Similarly, possible initiation codons were reported as TTG for cox1 and nad4 in the mitogenome of fungal pathogen Scytalidium auriculariicola 32 and, TTG and GTG were presented as start codons for nad2 and cox3, respectively, in the mitogenome of the nematode endoparasitic fungus Hirsutella vermicola 33 . On the other hand, some possible start codons detected in the mitogenome of fungal phytopathogen Stemphylium lycopersici were considered as suspicious codons and suggested that those were not acting as a start codon and the ORFs with the alternative start codons may have been co-translated with the upstream exons 34 . However, there has not been any proof of such a co-translation in the mitogenomes. Besides, different stop codons were described for fungal mitogenomes, as shown in this study and some previous studies [32][33][34] . Alternative transcription/translation language of mitogenomes is an interesting point of the independent evolutionary history of mitochondria and still required to be explored.

Homing endonucleases family Lenght (bp) Start codon Stop codon Location
Many different HEs were discovered just in one isolate of M. laxa. All the sequences were non-homologous and presented different start-stop codons. Such diverse HE sequences might serve as good candidates for genome editing as reviewed by Stoddart 35 . Investigating the presence of any common HE gene in mitogenomes among the different isolates of the same species would be informative to uncover the stability of these elements at the species level. Further investigations have been performing to answer to this intriguing question.
Distribution of HEs changed among the mitochondrial genes. Moreover, the distribution of mobile introns within the same gene shows high diversity among species, as shown for cytb gene 36 . Mobile introns are one of the major sources for diversity and dynamics structures of mitogenomes 7,37 . Moreover, those elements may transfer horizontally among species 38 , as well as between mitogenome and nuclear genome 39 . Thus, it would be interesting to compare HEs among Monilinia species causing brown rot disease, and this will be pursued with our ongoing research. The mobile introns are highly interesting to understand mitogenome evolution within/among fungal species. Moreover, repetitive sequence structures varied among the species, and those elements could be used as molecular markers in population genetics and diversity analysis.
The number of tRNAs in fungal mitogenomes, even among species from the same family may vary 40 . The total tRNAs slightly differed among M. laxa and closely related species. Moreover, tRNAs of M. laxa presented different anticodons, varied lengths, and extra-arms. As an unnoticed perspective, detecting tRNAs, their structures, and related mutations within and among fungal species could be useful to investigate evolutionary changes and affected traits.
Monilinia laxa was compared with some closely related species. Previously, the mitogenome of S. borealis was compared to the known mitogenomes of helotialean fungi in a study conducted by Mardanov et al. 41 . Another research article was analyzed the mitogenomes of Phialocephala subalpina, S. sclerotiorum and B. cinerea 42 . In this study, M. laxa represented the highest level of intron content in comparison to the other three species. The sizes of four mitogenomes varied from 82 to 203 kb due to the different numbers of introns (Table 4). Monilinia laxa mitogenome (178 kb) has the second longest after S. borealis, while the smallest mitogenome belonged to B. fuckeliana. It has been observed that the genome size is directly correlated with the number and size of introns. The 14-essential protein-coding genes (atp6, 8-9, cob, cox1-3, nad1-6, and nad4L), two ribosomal RNA genes (rnl and rns) and 1 ribosomal protein coding gene (rps3) were observed in all these mitogenomes. However, the number of tRNAs varied between 31 and 33 across these species (Table 4). Gene orders except for anonymous ORFs and tRNAs were the same among the four mitogenomes studied.
The first mitogenome of M. laxa indicated a mobile intron rich structure in comparison to the closely related species, and it may differ within/between species of Monilina species. Our project is ongoing to obtain more mitogenome data for a large collection of M. laxa and M. fructicola.

Materials and methods
Fungal sample and DNA isolation. Isolates of M. laxa were obtained from brown-rot-diseased peach fruits in Turkey, and after pure culturing, were stored at -20 °C on filter papers. Species identification based on both morphological criteria and polymerase chain reaction (PCR) with species-specific primers 3 . The isolate used for mitogenome characterization was obtained from the city Nigde and named Ni-B3-A2.
One piece of filter paper (approximately 0.5 mm 2 ), from the long-term storage, was aseptically placed on to potato dextrose agar, incubated for one week at 23 °C in darkness. Mycelia from 7-day old culture were transferred to potato dextrose broth and incubated for 5-7 days at room temperature in a rotary shaker. Then, mycelium was harvested from the liquid using vacuum filtration. Total DNA was extracted by using Norgen Plant/ Fungi DNA Isolation Kit (Norgen, Canada) following the manufacturer's protocol. DNA quality and quantity were measured using a spectrophotometer (NanoQuant Infinite M200, Tecan) as well as fluorometer (Qubit 3.0, Thermo Fisher Scientific, USA) using the dsDNA high sensitive assay kit (Thermo Fisher Scientific, USA). Furthermore, genomic DNA was visualized on 1% agarose gel to check for any break/smear or multiple bands.
Whole genome sequence analyses. Sequencing libraries were constituted using Illumina platform with TruSeq Nano kit to acquire as paired-end 2 × 151-bps, with about a 350-bp insert size. The next-generation sequence was performed by an external service (Macrogen Inc., Next-Generation Sequencing Service, Geumcheon-gu, Seoul, South Korea) that provided the raw sequence data. By using Trimmomatic v.36 software 43 , adapters were removed from raw reads and low-quality reads were trimmed by the setting of the parameters as LEADING and TRAINING = 10 (If their quality score is below 10, cut the bases off the start of the reads), SLIDING WINDOW = 5:20 (look at starting at base 1 and a window of 5bps, if the average quality score drops