Giant transposable elements drive genetic diversification in Aspergillus oryzae, a domesticated fungus traditionally used in food fermentation Authors: Yuya Hamanaka, Takuya Katayama, Yuta Chiken, Kana Nishiguchi, Katsushi Yamaguchi, Shuji Shigenobu, Jun-ichi Maruyama Published: 2026 ABSTRACT SummaryStarships are giant transposable elements recently reported in fungi that carry cargo genes and drive rapid adaptation. While inter-species variation by Starships has been investigated, their role in intra-species variation remains unclear. Here, we investigated Starships in Aspergillus oryzae, a filamentous fungus differentiated from mycotoxigenic Aspergillus flavus and subsequently domesticated for various applications in food fermentation, such as sake, miso, and soy sauce production in Japan. Comparative analysis between A. flavus and A. oryzae revealed that cargo gene gains/losses during domestication likely contribute to safe and stable food production. Furthermore, cargo gene composition varied even among A. oryzae strains harboring the same Starship classes, suggesting its correlation with specific strain usage. These findings indicate that Starships facilitate rapid evolution and diversification under artificial selection by acquiring and rearranging cargo genes. Our study highlights Starships as key vehicles driving genetic diversification in the domestication of A. oryzae for multipurpose industrial applications.Keywords: Aspergillus oryzae, food fermentation, brewing, starship, giant transposable elements, intra-species variation, domestication, artificial selection Introduction Mobile genetic elements are DNA sequences that move within the genome and are associated with adaptive evolution.1,2 Transposons, one of the mobile genetic elements, are DNA sequences typically ranging from 100 to 10,000 bp that can move to different parts of the genome.3,4 They were first discovered in maize,5 and have since been found in almost all organisms, including animals, plants, fungi, bacteria, and archaea.6,7,8,9,10,11 Transposons can alter fitness by causing loss/gain of gene function or genome rearrangement, such as generating novel regulatory networks or promoting genetic diversity.12,13 Furthermore, there are known cases in which transposons are involved in the domestication process, wherein species possessing traits beneficial to humans are selected.14,15 In bacteria, a distinct type of large mobile genetic element known as integrative and conjugative element (ICE) has been identified.16,17 Unlike typical transposons, ICEs range from 20 to 500 kbp and carry multiple cargo genes that affect the phenotypes of bacterial cells.16 Like conventional transposons, ICEs are involved in adaptive traits such as antibiotic resistance through horizontal transfer among species,18,19 playing an important role in generating genomic diversity and promoting adaptive evolution. In fungi, adaptive evolution is driven by genomic variation via transposons.20,21 Recently, giant transposable elements called Starships (ranging from 15 to 700 kbp) were found in the subphylum of filamentous fungi Pezizomycotina.22,23 These elements function as fungal analogs of bacterial ICEs. Starships contain tyrosine recombinase (YR) encoding genes at the 5′ terminus of the elements, called “captain genes,” and also contain some genes called “cargo genes.”22,23 It is considered that the YR excises genomic elements from the chromosome and reintegrate them into other chromosome regions.24 Starships are estimated to be horizontally transferred among inter-genus or inter-species, and contribute to increased host fitness in many filamentous fungal species by conferring traits such as heavy metal and formaldehyde resistance, as well as virulence.24,25,26,27,28,29,30 Although Starships promote the evolution of inter-species adaptive traits,25,27,29 their role in the adaptive evolution driven by intra-species strain variation remains largely unexplored. Aspergillus oryzae is a domesticated filamentous fungus that plays an important role in diverse food fermentation processes, including traditional Japanese fermentations such as sake, soy sauce, and miso production.31 A. oryzae was differentiated as a lineage from Aspergillus flavus, a producer of the carcinogenic mycotoxin aflatoxin, and was subsequently domesticated for food fermentation.32,33 This process was associated with the loss of aflatoxin production and duplication of α-amylase genes.34,35,36 A. oryzae includes numerous industrial strains (Tane Koji strains, TK strains) developed for specific applications, and these strains are classified into multiple phylogenetic groups, exhibiting high genetic diversity.37 However, the genomic basis of strain diversity in A. oryzae is still poorly understood. Recently, Starship elements have been predicted in A. oryzae38,39; however, their roles have not been well annotated. To investigate the relationship between traits and genomic variation by Starships, we generated chromosome-level assemblies of A. oryzae strains from phylogenetically different groups by long-read sequencing and conducted comparative analysis. Here, we found that A. oryzae has ten classes of Starship elements with diverse distributions among A. oryzae and A. flavus. In the same class, some of the Starship elements were heterogeneous between A. oryzae and A. flavus and among A. oryzae strains, suggesting that these diversities involve inter- and intra-species variation in A. oryzae. Results Genome assembly and comparative analysis of genomic structures among A. oryzae strains Chromosome-level assemblies were generated to comparatively analyze the relationship between genomic variation and strain heterogeneity in A. oryzae. The genomes of 82 A. oryzae industrial strains, named TK strains, have been previously comprehensively sequenced using short-read sequencing and classified into eight clades.37 Strains belonging to clades D and H were not available; therefore, five strains, TK-31, TK-34, TK-35, TK-41, and TK-49, belonging to clades B, A, C, G, and E, respectively, were selected for genome sequencing (Figure S1A). In addition to these strains, the RIB40 strain (belonging to clade F), for which the whole-genome sequence has been determined,40 was re-sequenced as a control. The de novo assembly of the six A. oryzae strains showed 37.3–39.2 Mbp in 8–13 contigs (Figure S1B; Table S1). In addition, 98.0%–98.5% of the BUSCO values indicated high completeness. In particular, the genome sequences of TK strains were more contiguous than that obtained with sequencing using short-read technology.37 The chromosome of assemblies in TK strains was defined based on whole chromosomal alignment between the RIB40 strain and five TK strains (Figure S1C). Most contigs from the TK strain aligned uniquely to contigs of the RIB40 strain (Figure 1, Figure 2, Figure 3, Figure 4, Figure 5, Figure 6, Figure 7S1C), suggesting that chromosomal structures were nearly conserved among the A. oryzae strains. However, some contigs separately aligned to multiple contigs of the RIB40 strain, and thus, chromosomal terminal regions (0.3–1.0 Mbp) were swapped between the TK strains and RIB40 (Figure S1C: black arrowheads). Comparison of the same chromosomal regions between the genomic sequences of A. oryzae strains using synteny analysis (for example, Chr. 4 in RIB40 and TK-35; Figure S2A) revealed that eight chromosomal regions of approximately 10–290 kbp were positioned on different chromosomes between the strains (Figure 1A; regions 1–8). Despite being located at different positions on the chromosome, these regions of the same class exhibited high sequence similarity among A. oryzae strains, although some strains had distinct sequences in part (Figure 1B). In some of the chromosomal regions, two or three were present in a single strain, while some strains did not have such chromosomal regions (Figure 1A). Notably, the DUF3435 domain-encoding gene was positioned at the end of these chromosomal regions (Figure S2A). Recently, the DUF3435-encoding gene was reported to commonly exist in Starships.22,23 Therefore, we designated the chromosomal regions containing DUF3435-encoding genes as Starship Element in A. oryzae (SEA1–8). Syntenic genomic regions positioned in different chromosomes of A. oryzae (A) Chromosome maps of the eight regions in different chromosomes of A. oryzae strains (regions 1–8) and non-syntenic regions in TK-31 (regions 9 and 10). The colored regions indicate syntenic regions. (B) Circular synteny maps of regions 1–8 and their flanking regions. Red arrows indicate the chromosomal region in Figure 1A, and black arrows indicate its flanking regions. Asterisks represent contig end. Note that there is no synteny in the flanking regions. Functional predictions of SEAs (A) Heatmap of gene frequency of SEAs on the basis of domain structures. Gene domain structures were predicted with InterProScan, and the frequencies of genes within each SEA class were quantified. (B) (left) α-amylase genes in A. oryzae strains. Gene IDs and chromosomal positions of amyA and amyB/C are listed. (right) Synteny maps of strains containing amyB/C in SEAs. Red and light blue arrows indicate DUF3435-encoding genes and α-amylase genes, respectively. Note that in every SEA region harboring α-amylase, transposons were consistently found in the flanking regions (jockey-like retrotransposon, yellow blocks; tnpA, purple blocks; tao1, green blocks). (C) A synteny map of SEA3. Red and yellow arrows indicate DUF3435-encoding genes and glycosyltransferase 8 (GT8) encoding genes, respectively. GT8 genes were found positioned next to DUF3435-encoding genes. Four of eight SEA3 have GT8 genes. (D) A synteny map of SEA6. Yellow arrow indicates GH12 cellulase or GH10 xylanase encoding genes. GH12 and GH10 genes were present only at the terminal region of the TK-49 strain. Short-read mapping data for presence or absence of SEAs in A. oryzae (A) Short-read sequences (BioProject no. PRJDB7763)37 mapped to sequences of TK-41 SEA1. Industrial uses of each strain based on Watarai et al.37 are indicated. Red highlights/arrows represent DUF3435-encoding gene. The light blue arrow indicates α-amylase gene, and light blue highlight represents α-amylase gene and its flanking regions. TK-41 SEA1 contains jockey-like retrotransposon (yellow blocks), tnpA transposon (purple blocks), and tao1 transposon (green blocks). Note that mapped reads were detected in the regions corresponding to the α-amylase gene and transposons; however, this mapping is likely attributable to the presence of multiple similar sequences within the genome. In contrast, reads mapped across the entire α-amylase gene together with its flanking regions were present only in clade G, which includes TK-41. (B) Short-read sequences mapped to sequences of TK-34 SEA3. Red highlights/arrows represent DUF3435-encoding gene. Yellow highlights/arrows represent GT8 genes. Note that in clades A, C, G, and H, reads were mapped to the regions containing the GT8 gene. (C) Short-read sequences mapped to sequences of TK-49 SEA6. Red highlights/arrows represent DUF3435-encoding gene. Yellow highlights/arrows represent GH10 and GH12 genes. Note that in clades D, E, and H, reads were mapped to the regions containing the GH10 and GH12 genes. Distribution of 10 classes of SEAs in filamentous fungi (A) Heatmap of number of SEAs in filamentous fungal strains. Based on BLASTn analysis, genomic regions with hits above 30% in “query cover” and with DUF3435-encoding genes were considered to indicate the presence of SEAs. (B) Synteny analysis of SEA2 containing A. flavus-specific genes. SEA2 in TK-41 and Galactica in A. flavus AF13 (Chr1, 607,240–916,837 bp) were compared. Note that the NRPS gene and flanking regions of A. flavus were absent in A. oryzae. (C) Synteny analysis of SEA4 and SEA9 containing A. oryzae-specific genes. SEA4 in TK-35 and A. flavus La3279 (Chr4, 3,364,977–3,447,713 bp), and SEA9 in TK-31 and A. flavus AF13 (Chr6, 3,335,948–3,376,442 bp) were compared, respectively. Note that the NLR gene and methyltransferase gene were present in A. oryzae, but absent in A. flavus. Short-read mapping data for presence or absence of SEAs in A. oryzae and A. flavus (A) Short-read sequences (BioProject no. PRJDB7763)37 mapped to sequences of A. flavus AF13 chromosome 1 (607,240–916,837). Red highlights/arrows represent DUF3435-encoding gene. The yellow arrow/highlight indicates an NRPS gene. Note that reads were not mapped to NRPS gene and its flanking regions in A. oryzae. (B) Short-read sequences mapped to sequences of TK-35_SEA4. Red highlights/arrows represent DUF3435-encoding gene. Yellow highlights/arrows represent NLR genes. Note that reads were not mapped to regions containing NLR genes in A. flavus. (C) Short-read sequences mapped to sequences of TK-31_SEA9. Red highlights/arrows represent DUF3435-encoding gene. Yellow highlights/arrows represent methyltransferase genes. Note that reads were not mapped to regions containing methyltransferase genes in A. flavus. Putative horizontal transfer between A. oryzae and M. purpureus (A) Synteny map of SEA6 in A. oryzae and M. purpureus. Genome sequences of SEA6 in A. oryzae (TK-34 and TK-49) and chromosomes 3 and 4 in M. purpureus KUPM5 strain (acc. no. ASM2599979v1) were compared. Red arrows indicate DUF3435-encoding genes, and yellow arrows represent GH10 or GH12 genes. Note that GH10 and GH12 genes were only present within SEA6 of A. oryzae TK-49 and Chr3 of M. purpureus. (B) Gene similarity between M. purpureus KUPM5 and A. oryzae TK-49. The open reading frame sequences of the genes in TK-49 were subjected to a BLASTn search against M. purpureus, and the identity was calculated. Each plot represents each hit against KUPM5. Red plot indicates genes in SEA6 of TK-49. (C) Evolutionary model of transposition of SEA6 in A. oryzae. The most plausible evolutionarily model, minimizing the number of gene gain and loss events, was estimated. Horizontal transfer was considered only between clades, excluding transfer among strains. (Left) Phylogenic tree of DUF3435-encoding gene in SEA6. Following the inferred evolutionary order, the origin of SEA6 (blue), the horizontal transfer of SEA6 (brown), and the loss of GH10 and GH12 genes (red) are indicated using numbers. (Right) Phylogenetic tree of A. oryzae strains modified from Figure S1. Phylogenetic tree of M. purpureus strains and A. oryzae RIB40 were constructed. The blue/green lines indicate that SEA6 was retained in A. oryzae strains belonging to clades C, D, and E, and that the element was horizontally transferred to M. purpureus. The blue line represents SEA6 containing GH10 and GH12 genes, whereas the green line represents SEA6 without GH10 and GH12 genes. Conceptual model of SEAs spreading among A. oryzae strains The ancestral A. flavus lineage carried several SEAs, including SEA2, SEA4, SEA7, SEA8, and SEA9. Their chromosomal positions in the ancestral A. flavus genome were inferred from those in representative A. flavus strains. During differentiation from A. flavus to A. oryzae and its subsequent domestication, some SEAs were vertically inherited. Additional SEAs, including SEA1, SEA3, SEA5, SEA6, and SEA10, were acquired in the A. oryzae, whereas some SEAs appear to have been lost. Likewise, some cargo genes such as the NRP cluster genes, were lost; whereas, others, including NLR, methyltransferase, α-amylase, GT8, and GH10/12 genes, were acquired. Furthermore, genomic rearrangements also occurred through horizontal transfer among A. oryzae strains and translocation within individual genomes. These processes generated variation in SEA and cargo-gene composition among strains, some of which may subsequently have been selected during their use for specific industrial applications. Representative strains used for sake, soy sauce, and miso production, therefore, exhibit characteristic SEA compositions. SEA6 was also horizontally transferred from A. oryzae to M. purpureus. Classification of giant transposable Starship elements in A. oryzae In the above investigation, we focused on regions existing at different chromosomal positions among the strains, but this may potentially ignore chromosomal regions containing DUF3435-encoding genes. Hence, we comprehensively explored genes containing the DUF3435 domain (PF11917 or IPR021842) and identified 66 DUF3435-encoding genes in the A. oryzae genomes of the six strains. Phylogenetic analysis classified the 66 DUF3435-encoding genes into 12 clades (Figure S2B). Genes belonging to 10 clades of 12 clades were included in SEA1–8 classes, in which SEA4 class contained additional two clades of DUF3435-encoding genes (SEA4 #2 and #3 in addition to SEA4 #1). Two chromosomal regions containing remaining clades of DUF3435-encoding genes were found specifically in the TK-31 strain; these two regions were designated as SEA9 and SEA10 classes (Figure 1A). Eventually, 48 SEAs (across SEA1 to SEA10 classes) were identified among the six A. oryzae strains (Figure S3). The SEAs ranged from 6.3 kbp to 288.8 kbp and had an average GC content of 45.0% (Table S2), while the regions classified into the same SEAs were largely homologous despite some deletion or insertion of genes (Figure S3). Terminal inverted repeat (TIR) and direct repeat (DR) sequences are often found at both ends of Starship elements and are involved in facilitating the transposition process.23 Hence, TIR and DR sequences were predicted in the identified SEAs, and their sequences were found to differ depending on the SEA class (Figure S4), supporting the classification of the SEAs. Previous studies have revealed that Starship elements contain the DUF3435-encoding genes at their terminal parts, which are called “captain,” as they mediate the translocation of the Starship element containing cargo genes.22,23 Consistent with these findings, most of the identified SEAs contained the DUF3435-encoding genes at their terminal parts (Figure S3). Although some strains possessed two or three DUF3435-encoding genes in SEA4, the DUF3435-encoding genes belonging to the SEA4 #1 clade (Figure S2B) were defined as captain genes based on their positions in the regions (Figure S3). Starship elements were classified based on the DUF3435-encoding gene phylogeny.38 BLASTn search of DUF3435-encoding genes against the starbase, which is a database of Starship elements,41 revealed that SEA1, SEA3, SEA6, SEA7, and SEA8 belonged to the Prometheus-family, SEA4 and SEA9 belonged to the Enterprise-family, SEA2 belonged to the Galactica-family, and SEA5 belonged to the Phoenix-family. Notably, SEA10 showed no sequence similarity with elements in the starbase. Prediction of functions of the SEAs To examine the expression of genes located within SEAs, we performed RNA-seq analyses using three A. oryzae strains (RIB40, TK-35, and TK-41), which collectively harbor a broad range of SEA classes (SEA1–SEA8). Although not all cargo genes located within SEAs were transcriptionally expressed, their overall expression levels were comparable to those of genes located in other chromosomal regions (Figures S5A–S5C), suggesting that SEA-associated genes are not transcriptionally silenced and may have biological functions. To predict the functions of the identified SEAs, the genes included in the elements were annotated based on the domain structures of the proteins they encode (Data S1). The composition of cargo genes differed among the SEA classes (Figure 2A), suggesting that SEAs play different functions based on their classes. DUF3723, patatin-like phospholipase (PLP), ferric reductase (FRE), and NUDIX domain-containing genes, which are often found in Starship elements,23 were frequently included in the SEAs (Figure 2A). In a previous study, the Starship element HEPHAESTUS of Paecilomyces variotii was shown to comprise several genes that mediate metal resistance, while MITHRIDATE of P. variotii and A. nidulans were found to comprise genes that contribute to formaldehyde resistance.24,27 However, these genes were not included in the SEAs. Similar to Starship elements, we found the SEAs to frequently contain PLP genes and HET domain-containing genes, which are involved in heterokaryon incompatibility (Figure 2A).42 Thus, the SEAs may be involved in regulating heterokaryon incompatibility, which has been detected in A. oryzae.43 Furthermore, the SEAs were often found to contain DNA-binding proteins and protein kinases (Figure 2A), suggesting their potential involvement in signal transduction. Although regions belonging to the same SEAs were highly homologous, unconserved regions containing various genes were found between several SEA pairs (Figure S3). To explore the potential relationship between Starship cargo genes and the industrial applications of A. oryzae, we focused on genes that may contribute to fermentation processes. Interestingly, some of the SEA regions contained α-amylase-encoding genes that are closely related to food fermentation (Figure 2B). The α-amylase genes of A. oryzae are classified into two groups, namely amyA and amyB/C, based on the presence of Tc1/Mariner class putative transposable element tnpA (acc. no. AB072434.1) downstream of amyB/C,44,45 and those included in the SEA region belonged to the latter group. The positions of amyB/C genes were diverse among the strains. Particularly, two amyB/C genes were divergently oriented in SEA1 of TK-41 and SEA5 of TK-35 (Figure 2B). BLASTn analysis revealed that transposons tnpA and tao1 (acc. no. AB021710.1)46 and jockey-like retrotransposon (found in Drosophila similar to mammalian LINEs)47 were positioned adjacent to the α-amylase genes (Figure 2B). The jockey-like retrotransposons were positioned at the end of the no synteny regions containing α-amylase genes, indicating that amyB/C were translocated into specific SEAs via jockey-like retrotransposons. To examine whether regions containing α-amylase genes in the SEA were shared among A. oryzae strains, short-read sequence data of TK strains37 were mapped to the sequences of corresponding SEAs. The amyB/C genes and flanking regions in SEA3 of TK-34, SEA4_1 of TK-31, and SEA5 of TK-35 have almost no clade-specific sequences, thus, we used SEA1 of TK-41 as an example. By mapping analysis against to SEA1 of TK-41, the entire SEA1 region of TK-41 was conserved among the strains in clade G (Figure 3A), indicating that the SEA region contributes to an increase in the copy number of the α-amylase gene in strains of the same clade. Furthermore, to investigate whether the α-amylase genes located within SEA1 are functional, the two divergently oriented α-amylase genes in SEA1 of TK-41 were deleted (ΔTK41_S1-amy; Figure S6A), and α-amylase gene expression and enzymatic activity were examined. Because the nucleotide sequences of the α-amylase genes are nearly identical, the RT-qPCR analysis reflects the total transcripts of all α-amylase genes present in TK-41. The expression level of ΔTK41_S1-amy strain was decreased compared to that of the parental strain (Figure S6B). In addition, the α-amylase activity, measured by halo formation on starch agar media, was lower in the ΔTK41_S1-amy strain than in the parental strain (Figure S6C). These results support that the α-amylase genes located within SEA1 of TK-41 are transcriptionally expressed and functional. In addition, genes encoding glycosyl transferase 8 (GT8) were detected in SEA3 of some strains (Figure 2C). To investigate the conservation of SEA3 containing GT8-encoding genes among A. oryzae strains, mapping analysis was performed. GT8 genes were distributed among the strains in clades A, C, G, and H, the mapping pattern throughout the elements was similar within the strains of the same clade (Figure 3B). Interestingly, among the 52 strains harboring GT8 genes, 27 strains are used for miso production and 27 strains are used for sake production, with some strains used for both applications,37 suggesting a possible association between the presence of GT8 genes and strain usage in these fermentation processes. Phylogenetic analysis of the DUF3435-encoding genes in SEA3 revealed separate clades between SEA3 with and without GT8 genes, and the topology of this phylogenetic tree differed from strain phylogeny based on genome scale SNPs (Figure S7A). The GT8 gene-containing SEA3 is thought to have been transposed in A. oryzae through the following sequential events: (i) origin of GT8 gene-containing SEA3 was positioned as the common ancestor of clades C and G. Clades D, E, F, and H appeared to have lost GT8 gene-containing SEA3, while the element was retained in clades C and G (Figure S7A; blue line). Based on the mapping results in Figure 3B, all clade C strains including TK-35, which was sequenced in this study, are likely to possess SEA3; therefore, GT8 gene-containing SEA3 was vertically inherited into clade C. (ii) The GT8 gene-containing SEA3 from clade G was transferred horizontally to clade C (Figure S7A; green line), leading to the presence of two copies of GT8 gene-containing SEA3 in the corresponding clade. (iii) The additional GT8 gene-containing SEA3 in clade C, acquired from clade G, was transferred horizontally to clade A (Figure S7A; green line). These results indicated that the GT8 gene-containing SEA3 was propagated mainly via horizontal transfer among A. oryzae strains. Furthermore, RNA-seq analysis showed that GT8 genes of TK-35 and TK-41 were expressed (Figures S5B and S5C), suggesting that GT8-encoding genes in SEA3 are transcriptionally active and may retain biological functions in A. oryzae. In addition, to investigate whether expression of cargo genes in SEA3 differed between donor and recipient strains of horizontal transfer, we compared the expression profiles between the putative donor TK-41 (SEA3, clade G) and recipient TK-35 (SEA3_2, clade C). The shared cargo genes were transcriptionally expressed in both strains (Figure S5D). This observation is consistent with the hypothesis that SEAs retained transcriptionally active cargo genes upon horizontal transfer, which may have contributed to strain diversification associated with industrial use of A. oryzae. Furthermore, genes encoding a putative cellulase belonging to the glycoside hydrolase family 12 (GH12) and a putative xylanase belonging to the glycoside hydrolase family 10 (GH10), which contribute to polysaccharide degradation, specifically existed in SEA6 of TK-49 (Figure 2D). RT-qPCR analysis detected expression of the GH10 and GH12 genes, although GH12 transcripts were detected at relatively low levels (Figure S5E), suggesting that these genes are not transcriptionally silenced and may retain biological functions. GH10- and GH12-encoding genes were distributed among strains in clades D, E, and H, and strains outside of the clades (TK-24 and TK-27) (Figure 3C). Interestingly, eight of these ten strains with GH10 and GH12 genes are used for soy sauce brewing,37 suggesting that GH10 and GH12 genes were retained in the strains selected for soy sauce brewing. Among the SEAs, some cargo genes were different within the same SEA class (Figure S3), indicating that during the process of strain selection according to specific applications, the retention or loss of genes related to particular functions within the SEAs may generate intra-species variation. Distribution of SEAs in Aspergillus species To investigate the distribution of SEAs among Aspergillus species, their presence was examined by homology analysis of the genes included in the SEAs. SEAs were found in A. flavus, which is phylogenetically close to A. oryzae,48 and Aspergillus sojae, which is used for soy sauce fermentation,46 but not in other Aspergillus species such as Aspergillus nidulans, Aspergillus niger, and Aspergillus fumigatus (Figure 4A). In A. flavus, SEA2, SEA4, and SEA9 were found in multiple strains, whereas SEA7 and SEA8 were found only in La3279 and CA14 strains, respectively (Figure 4A). In contrast, SEA1, SEA3, SEA5, SEA6, and SEA10 were not detected in A. flavus (Figure 4A). SEA1, SEA3, SEA5, SEA7, and SEA9 were found in A. sojae. However, no SEAs were found in Aspergillus parasiticus, which is phylogenetically close to A. sojae49 (Figure 4A). Notably, although A. oryzae is phylogenetically farther from A. sojae than A. flavus, SEA1, SEA3, and SEA5 were shared between A. oryzae and A. sojae but not A. flavus and A. parasiticus (Figure 4A). Considering that A. oryzae and A. sojae are used in food fermentation, these SEAs may contribute to this industrial process. However, because no SEAs are conserved in Aspergillus luchuensis (Figure 4A), which is also used for Japanese traditional distilled liquor brewing,50,51 the distribution of SEAs appears to differ among industrially utilized Aspergillus species. To investigate the conservation of SEA regions between A. oryzae and A. flavus, a synteny analysis was performed (Figure 4B). Phylogenetic analysis revealed that the DUF3435-encoding genes in SEA2 and Galactica of the A. flavus AF13 strain were phylogenetically close (Figure S2B); however, unconserved regions with different gene contents (e.g., the non-ribosomal peptide synthetase (NRPS) encoding gene G4B11_000237,22 which is absent in A. oryzae) were found between SEA2 of TK-41 and Galactica of AF13 (Figure 4B). Upon mapping against A. flavus AF13, whole sequences of Galactica were found to be conserved among A. flavus strains; however, the NRPS-encoding gene was not detected in the A. oryzae strains (Figure 5A; yellow arrow and highlight). Short-read mapping analysis revealed that the NRP cluster containing NRPS was either partially (clades A, B, F, and G) or entirely (clade E) absent (Figure 5A). According to the phylogenetic analysis of the DUF3435-encoding genes of SEA2, the SEA2 was thought to have been transposed in A. oryzae through the following sequential events (Figure S7B): (i) origin of SEA2 was positioned as the common ancestor of A. oryzae and A. flavus (Figure S7B; blue line). (ii) NRPS was lost in A. oryzae during differentiation from A. flavus to A. oryzae (Figure S7B; green line), and (iii) SEA2 was horizontally transferred in an independent manner at least four times among A. oryzae strains. In contrast, SEA4 of A. oryzae contains nucleotide-binding oligomerization domain (NOD)-like receptor (NLR)-encoding genes, which are not present in SEA4 of A. flavus (Figure 4C). These NLR-encoding genes possess a structure comprising a nucleoside phosphorylase domain (IPR000845) at the N terminus, an NB-ARC domain (IPR002182) at the central region, and tetratricopeptide repeats (TPRs) (IPR019734) at the C terminus. Phylogenetic analysis of the DUF3435 #1 gene of SEA4 revealed that the SEA4 was thought to have been transposed in A. oryzae through the following sequential events (Figure S7C): (i) origin of SEA4 was positioned as the common ancestor of A. oryzae and A. flavus, and SEA4 was vertically inherited into A. oryzae strains from A. flavus (Figure S7C; blue and green lines); (ii) SEA4 without NLR genes was horizontally transferred from A. flavus into clade D; (iii) during differentiation from A. flavus, A. oryzae acquired NLR genes; and (iv) the NLR gene-containing SEA4 appears to have been horizontally transferred independently at least three times among clades in A. oryzae; however, the directionality of transfer remains unresolved (Figure S7C). The NLR genes were accompanied by a third DUF3435-encoding gene in SEA4 (DUF3435 #3), and no NLR genes were present in the regions lacking DUF3435 #3 (Figure S3; SEA4), raising the possibility that they were co-transferred with DUF3435 #3 into SEA4. Upon mapping against A. oryzae TK-35, these genes were found to be conserved in A. oryzae strains; however, A. flavus strains did not possess NLR genes (Figure 5B). In addition to the NLR genes in SEA4, a methyltransferase (IPR029063) encoding gene was also found to be A. oryzae-specific in SEA9 (Figure 4C). Mapping against A. oryzae TK-31 revealed that the methyltransferase gene was conserved in A. oryzae strains belonging to clade B; however, A. flavus strains did not possess the methyltransferase gene (Figure 5C). Phylogenetic analysis of the DUF3435-encoding genes in SEA9 revealed that the SEA9 is inferred to have been transposed through the following sequential events (Figure S7D): (i) origin of SEA9 was positioned as the common ancestor of A. oryzae and A. flavus (Figure S7D; blue line), and (ii) SEA9 was retained only in clade B of A. oryzae and lost in the other clades of A. oryzae (Figure S7D; green line), and the methyltransferase gene was acquired at once during the differentiation from A. flavus (Figure S7D). These findings suggest that SEAs originally present in A. flavus are mainly vertically inherited to A. oryzae and then transferred horizontally among A. oryzae strains, and that the SEAs have been evolved via gene gains/losses to have species-specific roles. Potential horizontal transfer of SEAs beyond genera Investigation of SEA distributions in other filamentous fungal species also revealed the presence of SEA6 in Monascus purpureus KUPM5 (Figure 4A). Synteny analysis further revealed that SEA6 of A. oryzae strains exhibited high sequence similarity with the corresponding regions of chromosomes 3 and 4 in M. purpureus (Figure 6A). When whole transcript sequences were compared between A. oryzae and M. purpureus, the transcripts from SEA6 showed much higher similarity than other genes (Figure 6B), suggesting that SEA6 was transferred between these species via horizontal transfer. In addition, GH10 and GH12 genes were found in the putative SEA6 region of M. purpureus (Figure 6A). These genes were also found in SEA6 of TK-49 strain (Figure 2D). Phylogenetic analysis based on the nucleotide sequences of the DUF3435-encoding genes in A. oryzae and M. purpureus supported the following sequential evolutionary model (Figure 6C). (i) First, ancestral A. oryzae strains of clades C, D, and E acquired SEA6 containing GH10 and GH12 genes. (ii) Based on the phylogeny of DUF3435-encoding genes, SEA6 was transferred horizontally to clade H strains from the ancestor of clade D strains. (iii) SEA6 containing GH10 and GH12 genes was also horizontally transferred from the ancestors of clade D strains to M. purpureus. (iv) After the divergence of clades C and D, GH10 and GH12 genes were lost in clade C. (v and vi) From clade C, SEA6 without GH10 and GH12 genes were transferred horizontally to clade A/B and M. purpureus, respectively. Therefore, the origin of SEA6 containing GH10 and GH12 genes was inherited by multiple transfers among A. oryzae strains and from A. oryzae to M. purpureus. These results indicate that, following the loss of GH10 and GH12 genes in a part of A. oryzae strains, SEA6 without GH10 and GH12 genes were transferred horizontally among A. oryzae strains and across genera from A. oryzae to M. purpureus. Discussion Transposons can alter the fitness of organisms by modifying the genome structure in response to their specific environments.52,53 In bacteria, the relationship between transposons and adaptive fitness has been extensively studied. ICEs are transposons that carry multiple genes within their regions and contribute to evolutionary adaptations, such as the acquisition of antibiotic resistance.16,19 In fungi, a functionally analogous class of elements called Starships has recently been identified. In this study, we used the domesticated filamentous fungus A. oryzae as a model and revealed that structural genome changes mediated by Starships are associated with domestication and intra-species strain variations related to their industrial applications. First, we generated chromosome-level genome sequences (8–13 contigs) of six representative A. oryzae strains from different clades. Comparative analysis of genomic structures among A. oryzae strains in phylogenetically different clades revealed that the genome structure was largely conserved, consistent with previous studies.54,55 Remarkably, synteny analysis revealed the presence of Starship elements in A. oryzae (SEAs) ranging from 10 to 290 kbp. In short-read-based assemblies, the detection of Starships is challenging because of their large size.22 However, by generating long-read-based assemblies, we were able to identify Starship elements in multiple industrial strains. SEAs were classified into 10 classes based on the phylogeny of DUF3435-encoding genes and repertoires of cargo genes. Despite belonging to the same class, the chromosomal positions varied among the strains, suggesting that SEAs spread among A. oryzae strains primarily through horizontal transfer and translocation within the genome. Few studies have extensively examined the chromosomal positions of Starships within the same species. Gluck-Thaler et al. recently reported that Starships promote strain heterogeneity in the fungal pathogen A. fumigatus and observed both conserved and variable chromosomal positions of Starship elements.30 Furthermore, Starship element distributions are not always correlated with strain phylogeny in A. fumigatus. In A. fumigatus, a sexual life cycle has been observed,56 and crosses between different strains in the environment may have led to phylogeny-independent distributions of Starship elements. In contrast, although mosaic genome structures resulting from ancestral sexual reproduction have been reported in A. oryzae,37 no sexual cycle has been identified in this species to date.57 There is a brewing technique known as tomokoji (or tomodane) (co-cultivation of multiple strains),58 and it is possible that such environments have facilitated the horizontal transfer of Starship elements between strains. In addition, Urquhart et al. demonstrated that a Starship element could transpose between different fungal species by co-culture.59 Thus, in A. oryzae, artificial selection during intra-species strain variation for fermentation purposes may have led to differences in Starship presence/absence and chromosomal positions among strains of the same species. To investigate whether SEAs were specific to A. oryzae or conserved in other Aspergillus species, we performed a BLASTn search. SEAs were found to be distributed only in A. oryzae, A. flavus, and A. sojae among the analyzed Aspergillus species. Interestingly, A. oryzae and A. sojae, which are domesticated species used for industrial applications,46,49 possess more DUF3435-encoding genes than A. flavus, which is consistent with the tendency for domesticated species to harbor more Starships.39 SEA1, SEA3, and SEA5 were found only in A. oryzae and A. sojae, but not in their ancestral species A. flavus and A. parasiticus (Figure 4A), suggesting that these regions may have been horizontally transferred between A. oryzae and A. sojae during domestication. SEA2, SEA4, SEA7, SEA8, and SEA9 were found in A. flavus (Figure 4A); thus, these SEAs were likely vertically inherited from the ancestral species A. flavus to A. oryzae. For instance, in SEA2/Galactica, the sequence identity between A. flavus strains exceeded 99%, suggesting high conservation. In contrast, in A. oryzae, recombination events and partial deletions of genomic regions or genes were observed within the same SEA class, implying that the diversification of SEAs occurred during the differentiation and domestication processes into A. oryzae. In SEA1, which is absent in A. flavus but uniquely present in A. oryzae, the region contains AogprK2/AO090166000020 (Figure S3; SEA1), encoding a class VI G protein-coupled receptor, and it is known that AogprK2 negatively regulates sclerotia formation.60 In A. oryzae, the majority of strains have no ability to form sclerotia,61 which is thought to have arisen as a result of domestication,57 and AogprK2 located within SEA1 may be one of the contributing factors. Thus, cargo genes commonly carried by Starships that are absent in A. flavus but present in A. oryzae may underlie traits specific to A. oryzae. SEAs contained diverse cargo genes that varied among SEA classes and strains (Figure S3; Data S1). RNA-seq analysis of three A. oryzae strains (RIB40, TK-35, and TK-41), which collectively harbor a broad range of SEA classes (SEA1–SEA8), showed that cargo genes within SEAs were not transcriptionally silenced compared with genes in other chromosomal regions (Figures S5A–S5C). This suggests that some of the cargo genes in SEAs are maintained in a transcriptionally active state and may retain biological functions. A comparative analysis showed that both A. flavus-specific and A. oryzae-specific cargo genes were present in the SEAs. For example, while the final product remains unidentified, an NRP (non-ribosomal peptide) biosynthetic gene cluster present in SEA2 of A. flavus is partially/entirely absent in A. oryzae strains (Figures 4B and 5A). Phylogenetic analysis of the DUF3435-encoding gene suggested that the NRPS gene was lost in A. oryzae during differentiation from A. flavus or domestication (Figure S7B). Because A. oryzae strains used in fermentation were selected for the loss of aflatoxin production,37 it is plausible that a similar selective process contributed to the elimination of toxin-related genes during the differentiation or domestication. In contrast, the A. oryzae-specific NLR gene was found in SEA4, and methyltransferase gene was found in SEA9 (Figure 4C). NLR genes are associated with programmed cell death, such as heterokaryon incompatibility in fungi.62,63 Heterokaryon incompatibility is thought to play a role in limiting the reception of harmful genetic elements such as transposons and mycoviruses between different individuals via cell fusion.64,65 In A. oryzae, heterokaryon incompatibility has been detected.43 In SEA4, the NLR gene is absent in A. flavus; however, it is present in multiple strains of A. oryzae except for clade D (Figures 5B, S3, and S7C). These NLR genes may enhance allorecognition, which contributes to genomic stability and leads to constant brewing. However, the absence of SEA4-associated NLR gene in clade D strains suggests other factors may also contribute to heterokaryon incompatibility. Additionally, DUF3435 #3 was always located adjacent to the NLR gene (Figure S3), suggesting the coexistence of additional Starship elements within SEA4. This suggests that Starship elements serve as vehicles for additional transposable elements and promote adaptation. Methyltransferases typically affect the transcription of numerous genes (such as those related to secondary metabolism, sexual/asexual development, and virulence) in Aspergillus sp.66,67,68,69 The methyltransferase gene found exclusively in A. oryzae within SEA9 may represent a unique genetic feature of clade B strains (Figure 5C). Although clade B strains are used for sake brewing, sake-brewing strains are also found in other clades. Functional characterization of methyltransferase-encoding gene in SEA9 may help clarify clade B-specific traits. These findings suggest that the presence/absence of SEAs and A. oryzae-specific cargo genes reflects selective pressures that favored strains with enhanced fermentation and metabolite production capacities during the differentiation and domestication processes. Next, the composition of the cargo genes in the SEAs was analyzed. Synteny analysis of SEAs within the same class revealed that most cargo genes were conserved; however, extra genes were irregularly distributed in the SEAs of specific strains. To explore the relationship between SEA cargo genes and the industrial applications of A. oryzae, we focused on α-amylase, GT8, GH10, and GH12 genes because of their predicted fermentation-related functions and strain-specific distributions. For example, the α-amylase genes were located in distinct Starship elements such as SEA1, SEA3, SEA4, and SEA5 depending on the strains (Figure 2B). In these α-amylase containing regions, jockey-like retrotransposons were consistently found at the termini, possibly mediating the transfer of α-amylase gene-containing regions into SEAs (Figure 2B). In bacteria, transposons carrying genes that enhance pathogenicity or confer antibiotic resistance are inserted into ICEs, which are thought to facilitate adaptation to new niches.70,71 Similar to bacterial ICEs, in SEAs, the acquisition of α-amylase genes through transposons may have contributed to the adaptation of A. oryzae to the brewing industry. A. oryzae possesses multiple copies of the α-amylase encoding gene, and the copy number of them varied by clade (Figure 2B).37 Specifically, TK-41 in clade G have four α-amylase genes, with two copies of the α-amylase gene positioned at SEA1. Copy number of α-amylase genes in A. oryzae is almost clade-dependent as other clade G strains are similar to TK-41.37 Deletion of the two divergently oriented α-amylase genes within SEA1 of TK-41 reduced both α-amylase gene expression and enzymatic activity (Figure S6), demonstrating that cargo genes carried by SEAs are functional. Together with their strain-specific distribution, these findings support the possibility that functional cargo genes within SEAs may have contributed to the traits relevant to the domestication and strain variation of A. oryzae. In addition to TK-41, in other strains where the α-amylase gene is located within an SEA, the copy number of the α-amylase gene was also conserved within each clade,37 supporting that SEAs contribute to increase in the copy number of the α-amylase genes. TK-49, which is used for soy sauce brewing, has two copies of α-amylase genes, and they are absent from the Starship elements. In soy sauce brewing, where soybean rather than starch is used as the substrate,72 high α-amylase activity are not necessary. Collectively, Starship elements may have contributed to the modulation of the α-amylase gene copy number according to industrial needs. In addition to the α-amylase gene, other strain-specific genes were found within the same class of SEAs. In SEA3, GT8 genes were specifically present in clades A, C, G, and H (Figures 2C and 3B). It is noteworthy that clades A, C, and G include A. oryzae strains used for miso production, where high salt concentrations are encountered during fermentation.73 BLASTn analysis of the GT8-encoding sequences revealed similarity to genes encoding galactinol synthase (EC 2.4.1.123). The galactinol synthase enzyme synthesizes galactinol from myo-inositol and UDP-galactose.74 In plants, galactinol is known to be involved in tolerance to oxidative, heat, and salt stresses.75,76,77 Although the function of GT8 genes in fungi remains largely unclear, these genes may be associated with broader adaptation to fermentation-associated environmental stresses, including salt stress. In addition, GT8 genes in SEA3 of TK-35 and TK-41 were transcriptionally expressed, as shown by RNA-seq analysis (Figures S5B and S5C), suggesting that they are not transcriptionally silenced and may retain biological functions. In SEA6, the TK-49 strain contained genes encoding GH10 xylanase and GH12 cellulase (Figure 2D). Mapping to the TK-49 genome revealed that the GH10 and GH12 genes were specifically present in clades D, E, and H (Figure 3C). Many strains belonging to these clades are used for soy sauce brewing,37 suggesting a possible association between the presence of GH10/GH12 genes and strain application. Xylanases and cellulases are generally involved in the degradation of plant cell wall polysaccharides, including components of soybeans used in soy sauce production.78,79,80,81 Therefore, GH10 and GH12 genes represent candidate factors that may contribute to fermentation-related traits. RT-qPCR analysis detected expression of both GH10 and GH12 genes, although GH12 transcripts were detected at relatively low levels (Figure S5E), suggesting that these genes are not transcriptionally silenced and may retain biological functions. It is possible that these genes were preferentially maintained in strains used for soy sauce brewing; however, this hypothesis requires experimental validation. These findings suggest that SEAs would have rearranged their genetic components by acquiring genes favored through artificial selection, resulting in intra-species variation for multiple brewing applications. Consistent with this possibility, RNA-seq analyses revealed that cargo genes within SEA regions were transcriptionally expressed and were not generally silenced relative to genes in other chromosomal regions (Figures S5A–S5C). Furthermore, genes shared between the putative donor strain TK-41 and recipient strain TK-35 within the GT8-containing SEA3 were expressed in both strains (Figure S5D), suggesting that SEA cargo genes can continue to be transcriptionally expressed following horizontal transfer. Interestingly, sequences that were highly homologous to SEA6 were identified on chromosomes 3 and 4 of M. purpureus (Figure 6A). M. purpureus is traditionally used in Asia for the fermentation of foods such as tofuyo (Japanese fermented tofu).82,83 GH10 and GH12 genes, which were identified in the SEA6 region of A. oryzae TK-49, were also found on chromosome 4 of M. purpureus. Phylogenetic comparison between the DUF3435-encoding genes within SEA6 and the single-nucleotide polymorphism-based phylogeny of A. oryzae strains suggested that SEA6 originated near the roots of clades C, D, and E, underwent horizontal transfer to M. purpureus, subsequently lost the GH10 and GH12 genes, and was then transferred horizontally to M. purpureus (Figure 6C). Under laboratory conditions, hyphal fusion between auxotrophic strains of A. oryzae and Monascus species has been demonstrated.84 Although A. oryzae and M. purpureus are rarely used together in the same fermentation process, both are commonly employed in the production of fermented foods in east Asia. This geographical and ecological overlap raises the possibility that SEA6 may have been transferred horizontally between the two species via contamination or other environmental interactions. These findings, which are in line with reports on other Starship elements, provide evidence that SEAs are capable of horizontal transmission across genus boundaries. We propose a conceptual model summarizing the roles of Starships in the differentiation and domestication to A. oryzae (Figure 7). Five SEA classes were detected in A. flavus lineages, one ancestral lineage of which differentiated into A. oryzae. During differentiation from an ancestral A. flavus lineage to A. oryzae and subsequent domestication, five additional SEA classes were acquired, whereas some SEAs appear to have been lost in individual strains or clades. In addition, some SEA cargo genes, including methyltransferase and NLR genes, were acquired, whereas SEA-associated NRP genes potentially involved in mycotoxin production were lost, generating A. oryzae-specific genetic diversification. Notably, fermentation-associated SEA cargo genes, including α-amylase, GT8, and GH10/12 genes, were also acquired, have facilitated strain variation during selection for specific industrial applications. Furthermore, SEAs have undergone intragenomic transposition and horizontal transfer among A. oryzae strains, and SEA6 appears to have been horizontally transferred to M. purpureus, a species belonging to a different genus. Together, these findings suggest that Starship elements serve as vehicles for the acquisition, loss, and rearrangement of genetic components, thereby generating genomic diversity during the differentiation and domestication of A. oryzae and facilitating strain variation. In yeast, an organism domesticated for multiple purposes, such as bread, sake, and wine, similar to A. oryzae, comprehensive genomic and phenotypic analyses have reported that copy number variants and gene content alterations are associated with the driving forces of domestication.85 A. oryzae may have adapted to an environmental niche, domestication, by employing Starship elements as vehicles for genetic components through the acquisition and rearrangement of cargo genes, thereby promoting strain variation for multipurpose applications. Elucidating the functions of unannotated-cargo genes within SEAs will provide further evidence for this scenario. In this study, we demonstrated that the diversity of Starship elements and their cargo genes between A. oryzae and A. flavus, as well as among A. oryzae strains may have contributed to the differentiation and domestication of A. oryzae. Our findings propose a eukaryotic model that diversity of genetic components within large mobile genetic elements drives the adaptation to specific environmental niches. Limitations of the study Our genomic comparative analysis revealed the presence/absence of Starship elements among A. oryzae strains; however, because only the genome sequences of representative strains from each clade were compared, intra-clade variation could not be adequately assessed. In addition, although we identified ten classes of Starship elements in this study, genome sequences for some clades (e.g., clades D and H) have not yet been sequenced at chromosomal level, raising the possibility that additional uncharacterized classes of Starship elements remain undiscovered. Furthermore, while we found the evidence of horizontal transfer of Starship elements between phylogenetic clades, there are limitations in determining the directionality of transfer between donor and recipient strains. Clarifying the directionality of Starship transfer will require integrated phylogenetic and population-level analyses, together with expanded long-read genome sequencing of A. oryzae to facilitate more comprehensive comparative investigations. Resource availability Lead contact Requests for further information and resources should be directed to and will be fulfilled by the lead contact, Jun-ichi Maruyama (amarujun@g.ecc.u-tokyo.ac.jp). Materials availability Materials including fungal strains and plasmids generated in this study will be provided from the lead contact. Data and code availability •Raw reads, genome assembly and annotation data of A. oryzae strains are deposited in the DDBJ as accession number BioProject PRJDB37648 (key resources table). The RNA-seq data of A. oryzae strains is available as the DDBJ BioProject PRJDB37768 and PRJDB42568 (key resources table). Raw reads, genome assembly and annotation data of A. oryzae strains are deposited in the DDBJ as accession number BioProject PRJDB37648 (key resources table). The RNA-seq data of A. oryzae strains is available as the DDBJ BioProject PRJDB37768 and PRJDB42568 (key resources table). •This study does not report original code. This study does not report original code. •Any additional information required to reanalyze the data reported in this study is available from the lead contact upon request. Any additional information required to reanalyze the data reported in this study is available from the lead contact upon request. Acknowledgments This study was supported by the Japan Society for the Promotion of Science (JSPS) KAKENHI (grant numbers 18H02123 to J.M., 24H00496 to J.M., 23H04547 to J.M. and 25KJ1154 to Y.H.), JST SPRING (grant number JPMJSP2108 to Y.H.), Sapporo Bioscience Foundation (to J.M.) and the Sasakawa Scientific Research Grant (to Y.H.). This work was supported by the NIBB Collaborative Research Program (18-468, 19-439, 20-405 and 21-315) to J.M. We would like to thank H. Asao, A. Akita, and M. Matsumoto for next-generation sequencing at NIBB. We would like to thank Editage (www.editage.com) for English language editing. Author contributions Y.H., T.K., S.S., and J.M. conceived of and designed the study. Y.H., T.K., Y.C., K.N., and K.Y. performed experiments. Y.H., K.T., Y.C. K.Y., S.S., and J.M. performed bioinformatic analysis. Y.H., T.K., K.Y., S.S., and J.M. wrote the manuscript. All authors approved the manuscript. Declaration of interests The authors declare no competing interests. STAR★Methods Key resources table REAGENT or RESOURCESOURCEIDENTIFIERAntibodiesEscherichia coli DH5αThermo Fisher ScientificCat# EC0112Chemicals, peptides, and recombinant proteinsProteinase K SolutionKANTO CHEMICAL CO., INCCat# 34060-96RNase A from Bovine PancreasNACALAI TESQUE, INC.Cat# 30141-14Pyrithiamine hydrobromideSigma-AldrichCas# 534-64-5Critical commercial assaysQIAGEN Genomic-tipsQIAGENCat# 10223Short Read Eliminator Kit XLPacific BiosciencesCat# SS-100-111-01Oxford Nanopore Ligation Sequencing KitOxford Nanopore TechnologiesCat# SQK-LSK109IsogenNIPPON GENECat# 311-02501RNeasy mini kitQIAGENCat# 74104Qubit™ 1X dsDNA High Sensitivity (HS) Assay KitThermo Fisher ScientificCat# Q33231PrimeScript™ RT reagent Kit with gDNA Eraser (Perfect Real Time)TaKaRaCat# RR047ATB Green™ Premix Ex Taq™ II (Tli RNaseH Plus)TaKaRaCat# RR820SPrimeSTAR® HS DNA PolymeraseTaKaRaCat# R010AIn-Fusion® Snap Assembly Master MixTaKaRaCat# 638947KOD FX NeoTOYOBOCat# KFX-201SmaITaKaRaCat# 1085ABst1107ITaKaRaCat# 1028ADeposited dataLong read sequence data for RIB40 and TK strainsThis studyDDBJ: PRJDB37648Short read sequence data of TK strainsWatarai et al.37DDBJ: PRJDB7763Short read sequence data for RNA-seq analysis of A. oryzae RIB40 strain in liquid mediumThis studyDDBJ: PRJDB37768Short read sequence data for RNA-seq analysis of A. oryzae strains in PD agar mediumThis studyDDBJ: PRJDB42568Experimental models: Organisms/strainsA. oryzae: Strain RIB40Machida et al.40NBRC 100959A. oryzae: Strain TK-31Watarai et al.37N/AA. oryzae: Strain TK-34Watarai et al.37N/AA. oryzae: Strain TK-35Watarai et al.37N/AA. oryzae: Strain TK-41Watarai et al.37N/AA. oryzae: Strain TK-49Watarai et al.37N/AA. oryzae: ΔTK41_S1-amy strainThis studyN/AOligonucleotidesThe list of oligonucleotides used in this study is shown in Table S3N/AN/ARecombinant DNApRGEgTK41_TE1amy2NDThis studyN/ASoftware and algorithmsMinKNOW v23.04.3Oxford Nanopore TechnologiesN/ATapestry v1.0.1Davey et al.86https://github.com/johnomics/tapestry?tab=readme-ovpyGenomeViz v0.4.4Shimoyama87https://github.com/moshi4/pyGenomeVizFastp v0.20.1Chen et al.88RRID: SCR_016962HISAT2 v2.2.1Kim et al.89RRID: SCR_015530Samtools v1.14Li et al.90RRID: SCR_002105FeatureCounts v2.0.1Liao et al.91RRID: SCR_012919Hifiasm v0.19.5-r587Cheng et al.92RRID: SCR_021069Funannotate v1.8.6Palmer and Stajich93RRID: SCR_023039Porechop v0.2.4Bonenfant et al.94RRID: SCR_016967NanoFilt v2.8.0Coster et al.95RRID: SCR_016966Flye v2.8.0-b1695Lin et al.96RRID: SCR_017016Pilon v1.24Walker et al.97RRID: SCR_014731AntiSMASH fungal version 7.0Blin et al.98RRID: SCR_022060BUSCO v5.4.4Manni et al.99RRID: SCR_015008InterProScan v5.22-61.0Jones et al.100RRID: SCR_005829BWA v0.7.18-r1243-dirtyLi et al.101RRID: SCR_010910GATK v4.2.0.0McKenna et al.102RRID: SCR_001876RAxML-NG v1.2.2Kozlov et al.103RRID: SCR_022066MUMmer v3.1Kurtz et al.104RRID: SCR_018171Circos v0.69-6Krzywinski et al.105RRID: SCR_011798Mafft v7.520Katoh et al.106RRID: SCR_011811FigTree v1.4.4University of Edinburgh; Rambaut LabRRID: SCR_008515R v4.2.2R Core Teamhttps://www.R-project.org/ggplot2WickhamRRID: SCR_014601chromoMapAnand and Rodriguez et al.107https://lakshay-anand.github.io/chromoMap/index.htmlOtherPacBioPacific BiosciencesN/AMinION Mk1BOxford Nanopore Technologieshttps://nanoporetech.com/ja/products/sequence/minion#gns.searchValue=mk1b&searchf1.value=Store&searchq.query=mk1bIllumina Novaseq 6000 Sequencing SystemIlluminahttps://www.illumina.com/systems/sequencing-platforms/novaseq.html; RRID: SCR_016387Qubit 4 FluorometerThermo Fisher ScientificCat# Q33226Multi-beads ShockerYasui KikaiCat# MB455UThermal Cycler Dice® Real Time System IVTaKaRaCat#TP1000 Experimental model and study participant details A. oryzae strains A. oryzae RIB40,40 TK-31, TK-34, TK-35, TK-41, and TK-49 strains37 were used in this study. Method details Genomic DNA extraction in A. oryzae To extract genomic DNA, conidia were inoculated into DPY medium [20 g/L dextrin, 5 g/L yeast extract, 10 g/L polypeptone, 5 g/L KH2PO4, and 0.5 g/L MgSO4·7H2O (pH 5.5)] and incubated at 30°C for 1 day. After the 1 day incubation, they were flash frozen in liquid nitrogen. Thereafter, frozen mycelia were homogenized using a Multi-beads Shocker (Yasui Kikai, Osaka, Japan). Genomic DNA of RIB40 was extracted from the homogenized mycelia using QIAGEN Genomic-tips (QIAGEN, Venlo, Netherlands), according to the manufacturer's instructions, and then used for sequencing using the PacBio SMRT sequencing (HiFi) method (Pacific Biosciences, Menlo Park, CA, USA). For sequencing using MinION (Oxford Nanopore Technologies, Oxford, UK), genomic DNAs of TK strains were extracted as described previously with some modifications.108 The homogenized mycelia were suspended in the extraction buffer [0.5% SDS, 50 mM EDTA, and 0.1 mg/ml proteinase K (pH 8.0)] and incubated at 60°C for 30 min. After debris removal by centrifugation, the supernatant was treated twice with phenol/chloroform/isoamyl alcohol (25:24:1) and once with chloroform/isoamyl alcohol (24:1). The extract was mixed with 4 μL of 5 mg/ml RNase A solution and incubated at 37°C for 1 h. The genomic DNA was precipitated using ethanol, and the precipitant was picked using a glass bar, washed with 70% ethanol, and dissolved in 100 μL of TE buffer. DNA concentration was measured using a Qubit 4 fluorometer (Thermo Fisher Scientific, Waltham, MA, USA). Genome sequencing For genome sequencing of the RIB40 strain, HiFi libraries were prepared using the SMRTbell Template Prep Kit 3.0 (Pacific Biosciences), according to the manufacturer’s instructions. The libraries were sequenced using the PacBio Sequel IIe platform. For genome sequencing of the five TK strains, short DNA fragments were removed using the Short Read Eliminator Kit XL (Pacific Biosciences). Library preparation and sequencing were performed using the Oxford Nanopore Ligation Sequencing Kit (SQK-LSK109; Oxford Nanopore Technologies), MinION flow cells (R9.4.1; Oxford Nanopore Technologies), and MinION Mk1B (Oxford Nanopore Technologies), according to the manufacturer's instructions. Base calling and quality control were performed using MinKNOW v23.04.3, with default parameters. RNA sequencing For genome annotation, the RIB40 strain was incubated in DPY liquid medium at 30°C for 1 day. Frozen mycelia were homogenized using a multi-bead shocker (Yasui Kikai, Osaka, Japan). To confirm the expression of genes located within the SEAs, the RIB40, TK-35, and TK-41 strains were cultured on PD medium overlaid with cellophane at 30°C for 2 days. The harvested mycelia were frozen and subsequently homogenized using a mortar and pestle. Total RNA was extracted using Isogen (NIPPON GENE, Tokyo, Japan) and cleaned using an RNeasy Mini Kit (QIAGEN) following the manufacturer's instructions. Library preparation and sequencing using a NovaSeq 6000 were performed commercially by Novogene (Beijing, China). For RNA-seq data analysis in the RIB40, TK-35, and TK-41 strains, the sequence reads were trimmed and filtered using fastp v0.20.1,88 and the filtered reads were mapped to the genome sequence of the corresponding A. oryzae strains using HISAT2 v2.2.1.89 The output SAM files were converted into BAM files and indexed using Samtools 1.14.90 The number of reads mapped to each gene was counted using featureCounts v2.0.1,91 and the transcripts per million (TPM) was calculated. Log10(TPM + 1) values were visualized as a heatmap. Genome assembly and annotation HiFi long reads obtained by PacBio sequencing were used to generate a de novo assembly of A. oryzae RIB40 using hifiasm v.0.19.5-r58792 with the default settings. Misassembled short contigs (smaller than 10-kb and repeat sequences) were excluded. Genome assemblies were annotated with the RNA sequencing data using the Funannotate pipeline v1.8.1693 and InterProScan v5.22-61.0100 with default parameters. Long reads of five TK strains (TK-31, TK-34, TK-35, TK-41, and TK-49) obtained by Nanopore sequencing were cleaned using porechop v0.2.4 for adaptor trimming94 and nanofilt v2.8.0 for quality filtering.95 The de novo assemblies were generated using flye v2.8.3-b169596 with a nano-raw option. The assemblies were then polished with the corresponding short-read sequences (BioProject No. PRJDB7763) using Pilon v1.24.97 Genome assemblies of the TK strains were also annotated using Funannotate pipeline93 without RNA sequencing data. Biosynthesis gene clusters (BGCs) were predicted using antiSMASH fungal version 7.0.98 The genome completeness of the assemblies and gene annotations were evaluated with BUSCO v.5.4.499 using the fungi_odb10 dataset. Telomeric repeats were identified and visualized using the weave software of the Tapestry package v.1.0.1, with -d 30 and -t TTAGGGTCAACA parameters.86,109 To identify DUF3435-encoding genes, protein sequences of A. flavus AF13 (Acc. No. GCA_014117485.1), A. flavus CA14 (Acc. No. GCA_014784225.2), A. flavus NRRL3357 (Acc. No. GCA_009017415.1), A. flavus NRRL21182 (Acc. No. GCA_002443195.2), A. niger CBS117618 (Acc. No. GCA_009176385.1), A. fumigatus Af293 (Acc. No. GCA_000002655.1), A. parasiticus CBS117618 (Acc. No. GCA_009176385.1), A. nidulans FGSC A4 (Acc. No. GCA_000011425.1), and A. luchuensis IFO4308 (Acc. No. GCA_016861625.1) were annotated InterProScan (v. 5.22-61.0).100 Protein sequences of A. flavus La3279 (Acc. No. GCA_030515275.1), A. sojae NBRC4239 (Acc. No. GCA_000226655.2), and A. sojae SMF134 (Acc. No. ASM827498v1) are publicly unavailable; these strains were annotated with the Funannotate pipeline using genome sequences of these strains. DUF3435-encoding genes were detected in the presence of PF11917 or IPR021842. Phylogenetic analysis To analyze the phylogenetic relationship between A. oryzae RIB40 and TK strains, a phylogenetic tree of A. oryzae and A. flavus was constructed. The publicly available short-read genomes of 35 strains in A. oryzae and A. flavus were used in this study. Short-read genomes were adapter-trimmed and quality-filtered using fastp.88 Next, filtered-reads were mapped to the A. flavus Afla-Guard (Acc. No. ASM1289687v1)110 by BWA-MEM101 and converted from sam files into bam files using SAMtools. Using GATK v4.2.0.0102 “Haplotype Caller” option, SNPs to Afla-Guard were called and filtered using the parameters previously described.55 Phylogenetic relationship was analyzed using RAxML-NG103 with the GTR+G+ASC_LEWIS model and 100 bootstrap replicates. To analyze the phylogenetic relationships of Monascus purpureus, nucleotide sequences of the beta-tubulin, beta-actin, and calmodulin genes were connected and aligned using MAFFT, and phylogenetic analysis was performed using RAxML-NG with default parameters. To construct a phylogenetic tree of DUF3435-encoding genes within the same class of SEAs, nucleotide sequences of DUF3435-encoding genes in representative strains were searched against the NCBI database using BLASTn analysis. Alignment by MAFFT and phylogenetic analysis were performed using RAxML-NG with default parameters. Whole genome alignment To identify the synteny between A. oryzae TK-31, TK-34, TK-35, TK-41, and TK-49 genomes and the RIB40 genome, MUMer aligner v3.1104 was used as previously described.55 Briefly, Nucmer was used to align with “-c 1000” parameter. The output files were filtered using delta-filter with the “-l 10000” parameter and alignments were extracted using show-coords. The alignments were written as bundle files. Using bundle files, circular synteny maps were visualized with Circos v0.69-6.105 Synteny analysis of genomic comparison was aligned and visualized using the pgv-mummer v.0.4.4, software of pyGenomeViz package v0.4.4, with default parameters.87 Large translocated regions were visualized using R package chromoMap v4.1.1.107 Large translocated regions and their flanking regions (15–20 kbp) were compared and visualized using the same method as in Nucmer and Circos. Identification of Starship elements in A. oryzae (SEA) Using gene prediction data, DUF3435-encoding genes were identified in six A. oryzae strains (RIB40, TK-31, TK-34, TK-35, TK-41, and TK-49). For phylogenetic analysis of DUF3435-encoding genes, open reading frame (ORF) sequences of DUF3435-encoding genes in A. oryzae, A. flavus (AF13, CA14, NRRL3357, La3279 and NRRL21182), A. sojae (NBRC 4239 and SMF134), A. parasiticus (CBS 117618), A. nidulans (FGSC A4), A. niger (CBS 101883), A. fumigatus (Af293), A. luchuensis (IFO4308), and Monascus purpureus (KUPM5; Acc. No. GCA_025999795.1) were aligned using MAFFT v7.520106 and analyzed using RAxML-NG.104 The phylogenetic tree was visualized using FigTree software. To determine the ends of these regions, the same alleles were compared among genomic sequences of A. oryzae strains. Structures of SEAs or genomic syntenies within SEAs were visualized using the pgv-simpleplot or pgv-mummer software of the pyGenomeViz package v0.4.4. The annotation of genes within the SEAs was surveyed using InterProScan. Identification of TIR and DR sequences For each SEA, genomic sequences 10 kb upstream and downstream of the ends of the SEAs were retrieved and aligned using MAFFT. Direct repeat (DR) sequences were defined as > 2 bp sequences with the same strand at the end of the SEA sequence. Terminal inverted repeat (TIR) sequences were defined as sequences > 2 bp with the opposite strand at the end of the SEA sequence. Conservation of SEAs among A. oryzae and A. flavus strains To determine the conservation of SEAs among A. oryzae and A. flavus strains, short-read sequences (BioProject No. PRJDB7763, PRJNA607981, PRJNA639008, PRJNA483051, and PRJNA992514) were mapped to representative SEA sequences. The mapping depth values for the SEA regions of representative strains were normalized by dividing them by the median depth of the chromosome on which each the SEA was located and are presented as a bar graph ranging from 0 to 2. A value of approximately one indicates the presence of the corresponding region. Comparison of SEAs between A. oryzae and other filamentous fungal species For searching the presence of SEAs in other filamentous fungal species, whole sequences of ten SEAs of representative strains (SEA1 of RIB40, SEA2 of RIB40, SEA3 of RIB40, SEA4_1 of RIB40, SEA5 of TK-41, SEA6 of TK-34, SEA7 of TK-34, SEA8 of TK-35, SEA9 of TK-31, and SEA10 of TK-31) were used to conduct an NCBI BLASTn search using “organism” options with “Aspergillus oryzae (taxid:5062)”, “Aspergillus flavus (taxid:5059)” or excluding “A. oryzae and A. flavus”. Genomic regions with BLASTn hits above 30% in “Query Cover” and with DUF3435-encoding genes searched above were considered to indicate the presence of SEAs. Analysis of gene horizontal transfer between A. oryzae and M. purpureus To predict the horizontal transfer of SEA6, the nucleotide sequences of whole genes in A. oryzae TK-49 containing SEA6 were compared with those of M. purpureus KUPM5 using NCBI BLASTn analysis with an E-value threshold of 1e-50. For BLASTn hits, dot plots (x-axis, gene length (bp); y-axis, identity (%)) were plotted. Genomic synteny blocks were visualized using pgv-mummer. Putative genes identified to be transferred horizontally between TK-49 and M. purpureus are shown as red dots. Investigation of the function of genes in SEAs To classify the genes included in the SEAs, their domain structures were predicted using InterProScan. The number of genes with the same domains was counted. Then, the number of genes with domains frequently detected (DUF3435 [IPR021842 or PF11917], DUF3723 [IPR022198 or PF12520], Patatin-like phospholipase (PLP) [IPR002641 or PF01734], Ferric_reduct [IPR013121, PF08030, or IPR013130] and NUDIX [IPR015797, PF00293, or IPR000086]), heterokaryon incompatibility-related domains (WD40 G-beta repeat [IPR015943, IPR001680, PF00400, IPR020472 or IPR019775], NACHT [PF05729, IPR007111, or IPR031359] and HET [PF06985 or IPR010730]), signaling-related domains (DNA-binding; Zinc finger [IPR007087, IPR013087, or PF13920], Zn(2)Cys(6) fungal-type [PF00172 or IPR001138], RING [IPR017907, IPR013083, IPR001841, PF13923, PF00097, or PF13445], MADS [IPR002100], Homedomain-like [IPR009057, PF00046, or IPR001356], F-box [IPR001810], Pkinase [IPR011009, PF17667, IPR000719, PF00069, or IPR008271]), and other domains (Methyltransferase [PF13489, PF13649, PF08241, PF13847, PF08242] and Arrestin [IPR024391 or PF02752], Ankyrin repeats [PF13606, PF13857, PF00023, PF12796, PF13637, or IPR002110], DnaJ [PF00226 or IPR001623], Helicase [IPR014001, PF00271 or IPR001650] and α-amylase [GH13; IPR015340, PF09260, IPR006046, or PF00128]) were visualized as heatmaps drawn using R package ggplot2. DNA manipulation Escherichia coli DH5α strain was used for DNA manipulation. Polymerase chain reaction (PCR) for plasmid construction was performed using PrimeSTAR® HS DNA Polymerase (TaKaRa Bio, Ohtsu, Japan). The In-Fusion HD Cloning kit (TaKaRa Bio, Ohtsu, Japan) was used for plasmid construction. Genomic PCR was performed using KOD FX Neo (TOYOBO). Primers used in this study are listed in Table S3. Construction of knockout strain The U6 promoter and a target sequence for 5'-flanking region of α-amylase gene were amplified from pRGE-gwAup111 using the SmaIIF1-PU6-F2nd/gTK41_TE1_amy-PU6-R1st primers and ligated with SmaI-digested pRGE-gRT6,111 yielding a genome-editing plasmid pRGEgTK41_TE1amy1st. The U6 promoter and a target sequence for 3'-flanking region of α-amylase gene were amplified from pRGE-gwAup using the Bst-PU6F/gTK41_TE1_amy-PU6-R2nd primers, and the guide RNA and U6 terminator were amplified from pRGEgwAup using the guide-TU6-F/Bst-TU6R primers. The two fragments were ligated with Bst1107I-digested pRGEgTK41_TE1amy1st, yielding a genome-editing plasmid pRGEgTK41_TE1amy2nd. Transformation of A. oryzae TK-41 strain was performed as previously described.111 The genome-editing plasmid pRGEgTK41_TE1amy2nd and the oligonucleotide TK41_TE1_amyDN-ssDD-F were used for transformation. Dextrin-peptone (DP) medium (20 g/L dextrin, 10 g/L polypeptone, 5 g/L KH2PO4, and 0.5 g/L MgSO4·7H2O) was used for pre-culture. Czapek-dox (CD) medium (3 g/L NaNO3, 2 g/L KCl, 1 g/L KH2PO4, 0.5 g/L MgSO4·7H2O, and 0.002 g/L FeSO4·7H2O, and 20 g/L sorbitol [pH 5.5]) was used to select the transformants. For selection of strain carrying the pyrithiamine-resistant ptrA marker plasmid, 0.1 μg/ml pyrithiamine was added. To remove the genome-editing plasmid from the transformants, dextrin was added to CD medium instead of sorbitol. RT-qPCR analysis Modified DPY medium (10 g/L starch from potato, 10 g/L polypeptone, 5 g/L yeast extract, 5 g/L KH2PO4, and 0.5 g/L MgSO4·7H2O, pH 5.5) was used for TK-41 culture. To detect expression of the GH10 gene in TK-49, a modified DPY medium containing xylan from corn core instead of starch was used. To detect expression of the GH12 gene in TK-49, modified DPY medium containing carboxymethyl cellulose instead of starch was used. Mycelia were collected from liquid culture at 30°C for 1 day and immediately frozen in liquid nitrogen. Thereafter, frozen mycelia were homogenized using a Multi-beads Shocker (Yasui Kikai, Osaka, Japan). Total RNA was extracted, and cDNA synthesis was performed using PrimeScript™ RT reagent Kit with gDNA Eraser (Perfect Real Time) (TaKaRa). RT-qPCR was performed using TB Green™ Premix Ex Taq™ II (Tli RNaseH Plus) (TaKaRa) and a Thermal Cycler Dice® Real Time System IV (TaKaRa). Primers used for RT-qPCR are listed in Table S3. Measurement of α-amylase activity To measure an α-amylase activity, conidial suspensions (5.0 ×104/5 μL) were spotted onto modified starch-containing DPY agar medium and incubated in the dark at 30°C. After 3 days culture, the plates were stained with iodine, and then halo sizes were measured. Quantification and statistical analysis p values were calculated using Student’s t-test. p values less than 0.05 were considered to indicate a significant difference. ∗ p < 0.05, and ∗∗∗ p < 0.001. Footnotes Supplemental information can be found online at https://doi.org/10.1016/j.isci.2026.117405. Supplemental information Document S1. Figures S1–S7, and Tables S1–S3 Data S1. List of genes in SEAs, related to Figure S3 and S5 Data S2. Uncropped DNA gel electrophoresis images, related to Figure S6A References Associated Data Supplementary Materials Document S1. Figures S1–S7, and Tables S1–S3 Data S1. List of genes in SEAs, related to Figure S3 and S5 Data S2. Uncropped DNA gel electrophoresis images, related to Figure S6A Data Availability Statement •Raw reads, genome assembly and annotation data of A. oryzae strains are deposited in the DDBJ as accession number BioProject PRJDB37648 (key resources table). The RNA-seq data of A. oryzae strains is available as the DDBJ BioProject PRJDB37768 and PRJDB42568 (key resources table). Raw reads, genome assembly and annotation data of A. oryzae strains are deposited in the DDBJ as accession number BioProject PRJDB37648 (key resources table). The RNA-seq data of A. oryzae strains is available as the DDBJ BioProject PRJDB37768 and PRJDB42568 (key resources table). •This study does not report original code. This study does not report original code. •Any additional information required to reanalyze the data reported in this study is available from the lead contact upon request. Any additional information required to reanalyze the data reported in this study is available from the lead contact upon request.