Background

The genus Fusarium comprises a large group of saprophytic and phytopathogenic filamentous fungi, widely distributed in all terrestrial ecosystems. Some of them are well-known research models for basic biological processes, such as Fusarium oxysporum in studies of plant pathogenesis [1], or Fusarium fujikuroi because of its ability to produce gibberellins, growth-promoting plant hormones [2]. Like many other species of Fusarium, F. fujikuroi is capable of producing a large array of secondary metabolites in addition to gibberellins, including bikaverin [3], fusarins [4], fusaric acid [4], fusarubin [5], and carotenoids, among others [6]. Recently, the analysis of its genome and the deduced proteome uncovered an unexpected complexity for its secondary metabolism [7]. In recent years F. fujikuroi has become a preferred fungal model to investigate the biochemical basis for the biosynthesis and regulation of several of these compounds, such as gibberellins, bikaverin [8], and carotenoids [9].

The synthesis of secondary metabolites is usually modulated by external cues, such as light or nutrient availability. Light is an environmental signal used by fungi to control many aspects of their biology and life cycle [10]. Light also serves as an indicator of potentially hazardous stresses, such as exposure to air, desiccation and osmotic stress, heat shock, or UV exposure. In Fusarium sp., light stimulates the synthesis of carotenoids [11]. Thus, when dark-grown cultures are illuminated, carotenoids accumulate for several hours after light onset in Fusarium aquaeductuum [12] and F. fujikuroi [13], providing a characteristic orange pigmentation. All the structural genes of the carotenoid pathway have been identified and their functions were determined by targeted mutation or by biochemical assays [14]. The carRA, carB, carT, and carD genes encode a bifunctional cyclase/phytoene synthase, a desaturase, a torulene-cleaving dioxygenase, and an aldehyde dehydrogenase, respectively, involved in the sequential biosynthetic steps needed to produce the acid xanthophyll neurosporaxanthin (Additional file 1: Figure S1). The carRA and carB genes are linked in a cluster with the retinal-forming dioxygenase carX gene and with the carO rhodopsin gene [15]. Except carD, only moderately induced by light, carT and the carRA/carB/carX/carO cluster are strongly photoregulated [9].

Expression studies revealed that stimulation of carotenogenesis by light is achieved at the level of transcription, with a rapid increase in mRNAs of most of the structural genes during the first hour of illumination [11]. The dependence on wavelength of carotenoid photoinduction was investigated in F. aquaeductuum, showing an action spectrum consistent with the participation of a flavoprotein [16]. In Neurospora crassa, a fungus with a similar light-regulated carotenoid pathway, the photoreceptor that mediates this response is the White Collar complex (WCC), a heterodimer formed by the flavin photoreceptor WC-1 and its partner WC-2 [17]. Upon illumination, the WCC is activated by light and binds to the promoters of target photo-regulated genes to stimulate their transcription. The WCC is also responsible for other processes stimulated by light in N. crassa, and photocarotenogenesis is totally abolished in the absence of either of the two White Collar (WC) proteins [18].

The wc-1 ortholog gene has been investigated in F. fujikuroi [19] and in F. oxysporum [20], where it was named wcoA and wc1, respectively. In both species, the null mutants for this gene retain a certain level of photoinduction of carotenoids when grown under continuous light, indicating that the WC flavoprotein is not the only photoreceptor involved in photocarotenogenesis. A more detailed analysis has shown that the stimulation of carotenoid biosynthesis by light in F. fujikuroi is carried out in two stages, a fast one mediated by WcoA and a slower one mediated by another flavin photoreceptor, the DASH cryptochrome CryD [21]. The significant carotenoid accumulation in wcoA mutants under constant illumination can be attributed to CryD, whose photoactivity has been demonstrated experimentally [22]. Interestingly, the mutation of wcoA or cryD also affects the production of other secondary metabolites [19, 23], even in the dark in the case of wcoA, indicating other functions for these photoreceptors, in addition to the control of carotenogenesis.

Deep-orange mutants, which produce large amounts of neurosporaxanthin and other carotenoids, have been described in F. fujikuroi [24] and F. oxysporum [25]. Their increased carotenoid content is due to the enhanced expression of the structural genes carRA, carB [15, 26], carT [27], and carD [28] either in the light or in darkness. However, mRNA levels for the car genes still increase in response to light in the carS mutants of F. fujikuroi [27] or F. oxysporum [25]. As observed for wcoA, the carS mutation not only affects the synthesis of carotenoids, but also the production of other secondary metabolites, such as gibberellins or bikaverin [29, 30].

The gene carS encodes a protein of the RING finger (RF) family that includes a putative LON protease domain [25, 31]. This structural combination was previously described in the CrgA protein of the Zygomycota fungus M. circinelloides, whose mutants exhibit a similar pattern of overproduction of carotenoids [32]. RING-finger domains are characteristic of a family of ubiquitin ligases, which transfer ubiquitin to specific target proteins to modulate their activity or determine their fate. In the case of CrgA, its ability to inactivate MCWC-1b white collar protein through its specific mono- and di-ubiquitylation was demonstrated [32]. The molecular mechanism by which the CarS protein down-regulates the genes for carotenoid biosynthesis in Fusarium has not yet been investigated. However, although the Fusarium and Mucor species are taxonomically very distant, and the CarS and CrgA polypeptides exhibit considerable sequence divergence, the conservation of the structural domains suggests similar mechanisms of action.

To better understand the regulatory role of the CarS protein in F. fujikuroi and its possible relationship with the regulation by light, we have carried out ribonucleic acid sequencing (RNA-seq) analyses from mycelia of the wild type, a carS mutant and a carS-complemented strain, grown in the dark or exposed to light for 1 hour. The analyses were extended to the effects of light and mutation of carS in F. oxysporum. The results revealed that CarS and light control the expression of a large battery of genes in both species, among which those of the car cluster stand out for the magnitude of their response. Interestingly, the data reveal an outstanding overlap between the genes induced by light and derepressed in the carS mutant, suggesting regulatory relationships between CarS and light. Some of the commonly regulated genes are supposedly involved in oxidative stress.

Results

Experimental design and analysis of global transcriptomic data

Previous analyses showed low mRNA levels of the genes of the carotenoid pathway in the dark-grown mycelia and a rapid accumulation after illumination, reaching maximum levels after approximately 1 hour of exposure to light [11]. To better understand the experimental conditions, we verified whether shorter exposures to light allow a maximum response after 1 hour of incubation. Therefore, real-time q-PCR experiments, abbreviated RT-qPCR hereafter, were done comparing wild type mycelia treated with different light pulses with a total incubation of 60 min (5 min light + 55 min dark; 15 min light + 45 min dark, 30 min light + 30 min dark and 60 min light). The results showed that mRNA levels of the carRA and carB genes, used as controls for photoinduction, increased gradually with longer light exposures, reaching their higher levels after 1 hour of illumination (Fig. 1). Therefore, to check the effect of light on the transcriptome of F. fujikuroi, we compared dark cultures with parallel cultures exposed for 1 hour to light. To examine the effect of the carS mutation, we used the mutant SG39, which was previously found to have a mutation that changes a highly conserved cysteine residue of the CarS protein [31]. As an additional control, we used SG256, a stable carS+ complemented strain obtained in two steps from a SG39 transformant [31] (Additional file 1: Figure S2).

Fig. 1
figure 1

Effect of duration of light pulse on transcript levels of carRA and carB genes. RT-PCRs of genes carRA and carB were carried out in total RNA samples from the wild type grown in the dark and illuminated for 5 min, 15 min, 30 min, or 1 h. To make the results comparable, in the dark control (DC) and in the samples exposed to light pulses of less than 1 hour, the incubation was extended up to 1 hour in the dark, as indicated in the column labels. Data are referred to the data in the dark and show the mean and error of the mean of three biological replicates

Full size image

The basic parameters of the RNA-seq data generated in this study, such as G + C content, millions of raw reads and percentage of mapped reads, are presented in Additional file 2: Table S1. In the genome of F. fujikuroi [7], the annotated coding sequences account for 49.2% of the total sequence, with an average coding length of 1457 nucleotides, an average number of 2.8 exons per gene and an average for the length of the exons of 518 nucleotides. The overall G + C content is 47.4%, which rises to 51.5% in the protein coding sequences. The sequences obtained in our RNA-seq study had an average G + C content close to 52%, in good agreement with the overall value of the genome for the coding sequences. With respect to read quality, its analysis with the FastQC package yielded similar high-quality profiles for all the samples (representative example in Additional file 1: Figure S3).

Mapping of the raw sequence data to the annotated reference genome downloaded from Ensembl Fungi (http://fungi.ensembl.org/index.html) identified transcripts for 15,095 genes, with 15,106 isoforms, and assignable to 14,815 coding sequences. These numbers fit well with the number of identified transcription start sites, 15,101. Once a high read quality was verified, the gene expression levels in the different samples were calculated for each gene using the Tuxedo protocol (see Materials and Methods). The read counts for each gene were obtained from the mapping results and normalized to FPKM values (fragments per kilobase of exon per million of mapped reads), and the overall distribution of gene expression in each sample was represented in boxplots (Additional file 1: Figure S4), which showed comparable gene expression patterns between the different samples. In addition, the representation of the coefficient of variation between replicates as a function of the expression levels (FPKM) showed a similar distribution in all strains and conditions, with a higher variation for genes with low expression, and an increasing reproducibility for genes with higher expression levels (Additional file 1: Figure S5). Therefore, the mean of the two biological replicates was used hereafter as the expression value in each strain and experimental condition.

Effects of light and carS mutation on the F. fujikuroi transcriptome

The differentially expressed genes were determined using a two-fold change threshold and a p-value threshold of 0.05 when comparing each condition with its control. The numbers and proportions of the genes activated or repressed in response to light or in the carS mutant SG39, in the different combinations of interest, are summarized in Table 1. The light and the SG39 genotype had similar impacts on the transcriptome, with at least 4–5% of genes with 2× alterations up or down in their expression. The effect of light and SG39 genotype over the entire transcriptome was represented by volcano plots (Fig. 2) and scatter plots (Additional file 1: Figure S6). In the wild type, the number of genes activated by light was slightly higher than those repressed by light (Table 1), but the graphical representations revealed a predominance of activation among the genes with the largest changes in their expression (green dots in Fig. 2a and Additional file 1: Figure S6A). This is more clearly seen in a graphical representation of the 30 genes with the largest activating or repressing fold-changes (blue arrowheads in Additional file 1: Figure S7A). This indicates a more relevant up-regulating role for light in Fusarium. A similar result was noticeable in the case of the effect of the SG39 genotype in the dark: although the number of genes with at least a 2-fold change in their mRNA levels was even greater for those with a decreased expression in this mutant strain (821 vs 716, Table 1), genes with major changes were more abundant among those up-regulated (green dots in Fig. 2b and Additional file 1: Figure S6B; Additional file 1: Figure S7A).

Table 1 Number of genes whose expression changes more than two-fold above (activated) or below (repressed) the controls, under the indicated conditions
Full size table
Fig. 2
figure 2

Effect of light and carS mutation on transcript levels of the F. fujikuroi genes. a-d Volcano plot representations of global expression data in the indicated pairwise comparisons. The empty circles indicate the positions of the car genes shown in panel E. a Effect of light in the wild type (wild type in the dark vs. wild type after illumination). b Effect of the SG39 genotype in the dark (wild type in the dark vs. carS mutant SG39 in the dark). c Effect of light in the carS mutant (carS mutant SG39 in the dark vs. carS mutant SG39 after illumination). d Effect of the SG39 genotype after illumination (wild type after illumination vs. carS mutant SG39 after illumination). e Expression levels of the car genes in the wild type (WT), carS mutant SG39 (carS−) and complemented strain SG256 (C) in the dark (D) or after one-hour illumination (L)

Full size image

The number of genes influenced by light was much lower in the carS mutant SG39 (285 genes, Table 1) than in the wild strain (1259 genes) or the complemented strain SG256 (1120 genes). In addition, large changes in mRNA levels resulting from illumination were very scarce in the carS mutant (Fig. 2c and Additional file 1: Figure S6C). This suggests that the deregulation produced by the carS mutation in many genes in SG39 affects their ability to respond to light. This is also visualized in Volcano and scatter plots that show the effect of the carS mutation in the light (Fig. 2d and Additional file 1: Figure S6D). The number of genes activated or repressed in the carS mutant SG39 is sensibly reduced in light compared to darkness (509 vs 717 for up-regulation and 633 vs 821 for down-regulation, Table 1). In this case there is a marked reduction in genes with high induction in expression levels (compared green dots in Fig. 2b and d, and in Additional file 1: Figure S6B and D).

The expression data are very reproducible for genes more drastically influenced by light or by SG39 genotype. This is the case of the structural genes of carotenoid metabolism ggs1, carRA, carB, carT, carD, and carX, as well as the carO rhodopsin gene or the carS regulatory gene (empty circles in Fig. 2a and b). However, the relative impact of light or the carS mutation on these genes is mostly attenuated when the two regulatory conditions act at the same time (Fig. 2c and d). The car genes differ in their absolute expression levels but exhibit a similar regulatory pattern in response to light and carS− genotype (Fig. 2e).

Since it is likely that the carS mutant SG39 has other unrelated mutations, only those genes that recovered a wild-type expression level in the complemented strain SG256 were hereafter considered as affected by the carS mutation. To visualize the effect of carS complementation in the SG39 genetic background, we compared in Venn diagrams the sets of genes activated or repressed in SG39 with those with the opposite effect in SG256 (Additional file 1: Figure S8). This comparison showed that approximately 55% of the genes whose expression changed in SG39 recovered their basal expression in SG256, suggesting that many genes with changes in expression close to the 2× threshold were not related to the carS mutation but to random variations. A comparison on the effect of light on the wild type and SG256 showed a greater overlap for the genes induced by light than for those repressed by light (64% vs. 21%, referred to the number of genes regulated by light in the wild-type), suggesting a greater abundance of changes in gene expression due to random variations in the set of genes repressed by light.

Relationship between the effects of light and carS mutation

The sets of genes that increase their expression as a consequence of light or the carS mutation showed a significant overlap (Fig. 3a). According to an enrichment analysis based on Fisher’s exact test, given that genome of F. fujikuroi consists of 15,095 genes, it would be expected that approximately 20 genes coincided by chance if both regulatory phenomena, light and carS mutation, were independent. Instead we observed 241 overlapping genes that resulted in a significant enrichment of approximately 12× with a p-value < 2.2 × 10− 16. There were also more genes coincident between those repressed by light or the carS mutation (80) than expected by chance (15, Fig. 3b). Therefore, the data show a patent relationship between the genes activated or repressed by light and those activated or repressed in the carS mutant. However, overlaps were mostly non-existent in comparisons between sets of genes activated by light and repressed by the carS mutation, or vice versa (Additional file 1: Figure S9A and B), with matching gene numbers as expected by chance. When higher activation thresholds - up to 10× - were considered, the degree of coincidence between the sets of genes activated by light and the carS mutation increased to approximately 40% of the genes in both conditions (Fig. 3c).

Fig. 3
figure 3

Overlap between the effect of light and carS mutation in the transcriptome of F. fujikuroi. a-b Venn diagrams of F. fujikuroi genes activated or repressed by light or by the carS mutation. The effect of the carS mutation was corrected with the data of the SG256 complemented strain. c-d Venn diagrams of F. fujikuroi genes activated or repressed by light in the wild type or in the carS mutant SG39. In all the diagrams, the numbers indicate genes that correspond to the conditions mentioned above. The surfaces of the circles are proportional to the numbers of genes. Overlapping areas of the circles correspond to genes that coincide in the conditions compared. The sets of genes activated/repressed by the carS mutation in dark conditions overlap significantly with the sets of genes activated/repressed in the wild type (WT) by light according to p-values < 2.2 × 10− 16 and odds ratios of 12.18 and 5.03, respectively, calculated using Fisher’s exact test. e Effect of the threshold for expression change in the number of genes activated by light and by the carS mutation. The graph represents the genes activated by any of the two conditions and the coinciding ones. Percentages of coinciding genes are also represented (dotted line)

Full size image

As already mentioned, the number of genes activated by light was much lower in the carS mutant than in the wild strain. However, most of the genes that were still activated or repressed by light in the carS mutant were also activated or repressed by light in the wild type (Fig. 3d and e). These results indicate that there is a set of genes that are regulated by light independently of CarS function. On the other hand, no coincidence was detected when the sets of genes activated or repressed by light in one strain were compared with those that exhibited the opposite pattern in the other (Additional file 1: Figure S9C and D), which supports the alleged relationship between genes activated by light and by the carS mutation.

Identity of strongly differentially expressed genes

The genes influenced by light or by the carS mutation were sorted by their degree of activation or repression and identified in the proteome of F. fujikuroi (complete list of genes exceeding 2-fold change displayed in Additional file 3: Table S2A-D). There was a considerable difference between the highest levels of activation and repression by light: while the most up-regulated genes in the wild strain increased their expression about 600 fold, the expression of the most repressed genes only decreased about 20-fold (Additional file 1: Figure S7A, Additional file 3: Table S2A and B). The differences between the maximum activations and repressions were less remarkable in the case of the effect of the carS mutation (Additional file 1: Figure S7A, Additional file 3: Table S2C and D). In the lists of genes activated by light or by the carS mutation, those of the car cluster (carX, carRA, carB, and carO, Additional file 1: Figure S1) were among the most activated, with carO standing out as the fourth gene in induction by light and the third gene in induction in the carS mutant (Additional file 1: Figure S7A, and Additional file 3: Table S2A and C). This result confirms the car cluster as one of the major regulatory targets of light and CarS. The effects of light and carS mutation on car genes are summarized in Table 2. In accordance with previous data [27, 28], the transcriptional changes found for the carT and carD structural genes, unlinked to the car cluster, were less pronounced. However, one of the most affected genes was FFUJ_10321, which encodes a protein with 78% identity with the short-chain alcohol dehydrogenase-like protein Bli-4 of N. crassa, strongly induced by light in this fungus [33]. As indicated in the discussion, this gene has sequence similarity to retinol dehydrogenases and could be involved in the metabolism of carotenoids.

Table 2 Effect of light and carS mutation on the expression of relevant gene groups
Full size table

Genes with highest up-regulations by light include those for the photoreceptors Phr, CryD and VvdA (Table 2 and Additional file 3: Table S2A). The effect of light on these genes was formerly known [23, 34, 35], but the influence of CarS on their expression has not been investigated. The carS mutation did not affect the transcript levels of the vvdA gene and produced moderate stimulations on those of phr and cryD. In addition, the data confirmed the lack of effect of light on the expression of wcoA, which was also not influenced by the carS mutation.

Genes whose expression was strongly regulated by light or by CarS included several genes presumably related to different aspects of cellular responses to stress conditions (Table 2). The third and sixth genes among those with the strongest activations by light (fifth and 23th among those activated by the carS mutation) were FFUJ_09119 and FFUJ_O9320 (Additional file 3: Table S2A and C), which encode a FAD flavoprotein and a protein with a ferritin-like domain, respectively. The gene FFUJ_01993, encoding a putative peroxiredoxin, is also in a high position in both tables (positions 27 and 35, respectively). The possible relation of these genes with stress functions is considered in the discussion. The second most activated gene by light is FFUJ_11472 (25th among those activated by the carS mutation), which encodes a protein with a catalase domain. Two other genes with catalase domains, FFUJ_05128 and FFUJ_03407, were also remarkably affected by light and by the carS mutation (Table 2). For validation purposes, the RNA-seq data for these genes were substantiated essentially by the analyses of mRNA levels by RT-qPCR (representative examples shown in Fig. 4). Due to its possible involvement in the metabolism of carotenoids, the presumed retinol dehydrogenase FFUJ_10321 was also included in the RT-qPCR assays.

Fig. 4
figure 4

RT-qPCR analysis on the effect of light and carS mutation on several putative stress-related genes. RT-qPCRs were carried out with the RNA samples from the wild type (WT), carS mutant SG39 (carS−) and complemented strain SG256 (C) used in the RNA-seq study. Darker bars: mycelia grown in the dark (D); paler bars: dark-grown mycelia exposed to light for 1 h (L). Data are referred to the data of the wild strain in the dark and show the mean and error of the mean of the two biological replicates. The unpaired t-test was applied to determine statistical differences between data from dark and light-exposed cultures and between wild type/C and carS− cultures. Differences found to be statistically significant are indicated with asterisks according to P values of 0.0332 (*) and 0.021 (**)

Full size image

GO term enrichment analysis of activated genes by light and carS mutation

The high number of genes whose expression was altered by the carS mutation indicates that the function of the CarS protein is not limited to the control of enzymes related to carotenogenesis or stress, but also affects other biological processes in the fungus. To obtain more global information on the role of CarS in Fusarium and its relation to light, Gene Ontology (GO) terms were tentatively assigned to the genes that were up-regulated in either of these two conditions. The GO enrichment analysis provides significantly overrepresented GO terms in both sets of differentially expressed genes compared to the background of the genome. The assignment of GO terms was based on the identification of PFAM domains [36]. In the case of the F. fujikuroi proteome, at least one significant PFAM domain was found in about 85% of the genes activated by light and 87% of the genes activated in the carS mutant.

A GO term enrichment analysis performed on this set of genes was summarized with REVIGO [37] and is shown in Fig. 5. The most significant non-redundant GO term for the genes activated by light was “Phosphorelay signal transduction system”, and at a lower extent for other GO terms, some of them also related to signaling or regulation (genes for relevant GO terms listed in Additional file 4: Table S3A). Although activating levels are generally not very high, the appearance of functionally related protein groups is striking. Thus, the list includes six putative histidine kinases, a family of proteins that form two-component systems with partner regulatory proteins. It also stands out the presence in the Metabolic process category of three genes for putative molybdoenzymes of the sulfite oxidase family, which includes nitrate reductase-like enzymes, together with an enzyme involved in the synthesis of molybdopterin, a precursor of the molybdenum cofactor required for the function of these proteins. In this category also appeared the ortholog of the Frequency protein, a central regulator of circadian rhythmicity in N. crassa [38].

Fig. 5
figure 5

Global effects of light and carS mutation in the transcriptome of F. fujikuroi. Treemaps representations of Revigo non-redundant GO term enrichment analysis of the F. fujikuroi genes induced by light (left, wild type dark vs. wild type light) or by the carS mutation (right, wild type dark vs. carS mutant dark). Each rectangle area in the treemap represents the −log10 (p-value) for the corresponding GO-term

Full size image

In the case of the effect of the carS mutation (Fig. 5, right panel), the most significant non-redundant GO term was “Organic acid metabolism”, followed by “response to stress”. The first category also includes enzymes supposedly involved in stress, like the aforementioned catalase FFUJ_11472, or the cyanate liase FFUJ_02467, predicted to catalyze the conversion of the toxic cyanate to carbamate, which is spontaneously decomposed to ammonia and carbon dioxide. It is also worth mentioning the presence of three proteins related to DNA damage, the light-activated photolyase Phr (already mentioned in Table 2), a DNA repair enzyme (FFUJ_06571) and a DNA-damage checkpoint protein (FFUJ_13454), as well as some enzymes involved in the metabolism of lipids. In overall, the comparison with the categories of the genes influenced by light indicates that, despite the high coincidence between the genes controlled by light and by CarS, the CarS protein has a narrower range of influence. However, in agreement with the high correlation with the effect of light, 9 of the 17 genes activated by the carS mutation (shown in Additional file 4: Table S3) also appear in GO terms for the genes activated by light.

Effects of light and carS mutation on global gene expression in F. oxysporum

As an indication of the degree of conservation of the regulations by light and by CarS in Fusarium, we extended our study to F. oxysporum. In this species, the effect of the carS mutation was inferred from the analysis of two independent carS mutants, SX2 and SX3 [25]. In this case, to rule out possible effects due to secondary mutations, we only considered as genes influenced by the carS mutation those that exhibit at least a 2-fold activation or repression in the two mutant strains. The RNA-seq data and their basic characteristics are described in Additional file 2: Table S1B. The sequence data in relation to the annotated genome identified transcripts for 17,999 genes, with 18,097 isoforms, assignable to 17,687 coding sequences. As was already found for the F. fujikuroi data, the representation of the expression values in FKPM with box plot graphs (Additional file 1: Figure S10) showed similar global distributions between all samples, making them comparable. Therefore, as in F. fujikuroi, analyses of the data were carried out using the mean of the two replicates as expression value in each strain and experimental condition.

As already done for F. fujikuroi (Fig. 2), the effect of light and carS mutation over the F. oxysporum transcriptome (gene lists with at least 2-fold changes displayed in Additional file 5: Table S4) was represented in volcano plots (Fig. 6) and scatter plots (Additional file 1: Figure S11). The plots exhibited patterns similar to those of F. fujikuroi (compare Fig. 2a with Fig. 6a and Fig. 2b with Fig. 6b), with higher numbers of strongly up-regulated genes (green dots) than down-regulated genes (red dots) as a result of light or carS mutation. As in F. fujikuroi, the genes of the car cluster (dots 1–4 in Fig. 6) exhibited large increases in their transcript levels, accompanied to a lesser extent by the gene carT (dot 6), whereas only minor up-regulating effects were observed for ggs1, carD or carS (dots 5, 7 and 8). As predicted, the comparison of the transcriptional patterns between the two carS mutants used in the assay revealed minor differences (Fig. 6c), explained by a combination of random variations and the occurrence of different secondary mutations.

Fig. 6
figure 6

Volcano plot representations of expression data of the F. oxysporum genes under different regulatory conditions. a Effect of light in the wild type. b Effect of the carS mutation (carS mutant SX2, similar results were obtained with SX1). c Comparison of the two carS mutants used in the assay. The empty circles indicate the positions of the car genes

Full size image

Comparison between the effects of light and carS mutation on the F. oxysporum and F. fujikuroi transcriptomes

The impact of light on the F. oxysporum transcriptome was slightly lower compared to F. fujikuroi. Using the same 2-fold threshold for modified expression, 768 genes were induced (4.27% of all the genes, compared to 4.79% in F. fujikuroi) while only 473 were repressed (2,63%, compared to 3.54%). The effect of the carS mutation was also lower in F. oxysporum, with 470 induced genes (2.61%, compared to 4.74% in F. fujikuroi) and 217 repressed genes (1.21%, compared to 5.44%). The level of overlap between the genes induced by light and by the carS mutation (Fig. 7a) was also higher than expected by chance (17.4% for a minimal two-fold increase) but lower than that observed in F. fujikuroi (26.9%). The percentage of coincident genes increased also in F. oxysporum for higher changes in transcript levels, rising to 24.51% for a ten-fold threshold (Fig. 7a). A significant overlap was also found between genes repressed by light and by the carS mutation (11.9%), but no overlap was found between the genes with a decrease of at least ten-fold in their mRNA levels.

Fig. 7
figure 7

Global effects of light and carS mutation in the transcriptome of F. oxysporum. a Venn diagrams of F. oxysporum genes activated or repressed by light or by the carS mutation. The data correspond to two-fold and 10-fold thresholds for expression change. The numbers indicate genes corresponding to the conditions indicated above. The surfaces of the circles are proportional to the numbers of genes. The overlapping areas of the circles correspond to genes that coincide in the compared conditions. b Treemaps representations of Revigo non-redundant GO term enrichment analysis of the F. oxysporum genes induced by light (left) or by the carS mutation (right). Each rectangle area in the treemap represents the −log10 (p-value) for the corresponding GO-term

Full size image

The REVIGO summary of the ontology enrichment analysis carried out with the F. oxysporum genes activated by light or by the carS mutation (Fig. 7b) revealed different GO terms than in F. fujikuroi. In this case, the most significant non-redundant GO term for the genes activated by light was “Response to stimulus”, with lower significance levels for other terms, while “Small molecule metabolism” was the most significant among those induced in the carS mutant. These differences can be attributed, at least partially, to the lower number of genes in the F. oxysporum classifications and to the overlap between the different GO categories. As an indicator of the high similarity between both species related to light and CarS regulation, it was observed that among the genes classified in the GO terms with higher significance, 18 out of the 30 activated genes by light and 5 out of the 7 activated genes in the carS mutants in F. oxysporum (Additional file 4: Table S3B) were also found in the lists for the GO terms of F. fujikuroi (Additional file 4: Table S3A).

A comparison using the Bidirectional Best Hit algorithm [39] between the genomes of the two Fusarium species revealed 11,918 genes with possible orthologs in both strains (Additional file 6: Table S5), which will be referred to hereafter as common genes, while 3177 (21.0%) and 6081 (33.8%) genes were specific to the F. fujikuroi and F. oxysporum genomes, respectively. Light had less influence on the specific genes in both species; thus, of the sets of 3177 and 6081 genes, only 9.2 and 12.4% were activated by light and 11.5 and 16.7% were activated by the carS mutation, respectively (Fig. 8a). In relation to the common genes, when comparing the sets of genes activated by light in both species, we found 272 coincidences, whereas the number of genes coinciding among those repressed by light was only 64. The coincidence in the genes activated by the carS mutation was lower (74, Fig. 8a), suggesting differences in the regulatory roles played by the CarS protein in both Fusarium species. However, again the number of matches was greater for induction than for repression in the carS mutants.

Fig. 8
figure 8

Overlaps between F. fujikuroi and F. oxysporum shared genes activated or repressed by light or by the carS mutation. a Overlaps between activated or repressed genes by light or by carS mutation between F. oxysporum and F. oxysporum. Genes without orthologs in the other species are indicated with an asterisk. The overlap between the genes activated by light in F. fujikuroi and F. oxysporum is significant, with an odds ratio enrichment of 9.15 and a p-value < 2.2 × 10− 16. The overlap between the genes repressed by light in F. fujikuroi and F. oxysporum is also significant with, an odds ratio enrichment of 6.27 and a p-value < 2.2 × 10− 16 (p values computed with the Fisher’s exact test). b Representation of overlaps between the Venn diagrams shown above. Only genes with orthologs in the F. fujikuroi (F. fuj.) and F. oxysporum (F. oxy.) genomes are considered. c Box-plot representations of the fold changes in the expression of genes simultaneously activated by light and carS mutation (common light carS−) or only by light (common light) in F. fujikuroi and F. oxysporum compared to those activated specifically in only one of the species (specific light and specific carS−). There were only three genes activated only by the carS mutation in both Fusarium species (common carS−), and therefore the box-plot is not represented

Full size image

An overall representation of the matches between the genes regulated by light and CarS in F. fujikuroi and F. oxysporum in a Venn diagram revealed a greater degree of overlap for the activated genes than for the repressed genes (Fig. 8b). If the focus is on the 272 genes activated by light and the 74 genes activated in the carS mutant in both species, we found 67 genes that are activated in both regulatory conditions; however, no coincidences were found between the 64 genes repressed by light and the 13 genes repressed by the carS mutation. The relative changes in mRNA levels for the 67 genes were greater in the different species and regulatory conditions than in the genes that coincide only in the activation by light or by the carS mutation (Fig. 8c). This set of genes (Additional file 7: Table S6) includes the four genes of the car cluster, and the genes carT and ggs1, as well as the genes for the cryptochromes CryD and FFUJ_03105/FOXG_02060, and different enzymes related to stress or other aspects of cellular metabolism.

Discussion

RNA-seq technology is a powerful tool that allows the quantification of mRNA levels of all genes in the genome of an organism. We have used this methodology to investigate the impact on the global transcriptome of two regulatory conditions that control carotenogenesis in Fusarium: up-regulation by light and down-regulation by the RING-finger protein CarS. In addition, we have used the relationships between both regulatory effects to explore their possible regulatory connections. The effect of light has been investigated in other species, but this is the first report on a transcriptomic analysis on the function of a RING-finger protein in fungi. Our RNA-seq data showed that illumination for 1 hour changed at least two-fold the expression of about 8% of the Fusarium genes in our experimental conditions. This is a smaller impact than that observed in N. crassa [40], a major reference in studies of fungal photobiology. In this fungus, the number of genes whose expression changed at least two-fold by light reached 31% of those expressed. The analyses in N. crassa were done with four times of illumination, 15 min, 1 h, 2 h, and 4 h, what could explain at least in part the 4-fold difference in the number of affected genes between both species. A more detailed time course of light exposure could increase the number of affected genes in Fusarium.

The impact of the mutation of the carS gene on the Fusarium transcriptome is quantitatively not very different from that of light, with about 10% of the genes affected directly or indirectly using the same fold change threshold. In this case, comparisons with N. crassa are not possible, because carS-like mutants have not been described in this fungus and there is no clear carS ortholog in its genome. Either after illumination or as a cause of the carS mutation, activation predominates among the main effects in Fusarium expression. It seems very likely that the effects of light and carS mutation share control elements, as indicated by the significant overlap between the affected genes in both regulatory conditions. This conclusion is reinforced by the drastic decrease in the genes regulated by light in the carS mutant.

In a previous screening of genes regulated by the CarS protein, based on a subtractive hybridization protocol, only about 60 genes were identified [31]. This group, which included genes activated or repressed in a carS mutant, represented less than 0.5% of the annotated F. fujikuroi genes, which corresponded to an experimental efficiency at least an order of magnitude lower than that obtained with the RNA-seq method. This difference suggests that the data of subtractive hybridization are biased towards those genes with higher expression. In addition, in contrast to the efficiency of the RNA-seq technique, the subtractive hybridization study did not reach saturation, as revealed by the absence of some of the genes of the car cluster in the scrutiny. However, the present RNA-seq analysis confirmed some of the genes identified with this previous approach. These include the homologs of the rds1 and bli-4 genes. The case of the bli-4 homolog, FFUJ_10321, is particularly interesting. This 341-aa protein exhibits a significant similarity with retinol dehydrogenases (RDHs) of mammals, as RDH11, RDH12, RDH13, and RDH14 of Mus musculus (e.g., 31% identity along 301 residues of the 316 residues of RDH12 [Q8BYK4.1], E-value 6e-18). RDHs catalyze the conversions between retinol and retinal, and in F. fujikuroi such enzymatic activity could be related with the retinal produced by the CarX enzyme [41]. RDH13 (Q8NBN7.2, 31% identity with FFUJ_10321 along 210 residues) has received special attention among mammalian RDHs. RDH13 is located on the outside of the inner membrane of the mitochondria in the pigmented epithelium of the retina, where it presumably protects against oxidative damage caused by light [42]. In humans, RDH13 is also found in mitochondria and could be involved in protection against oxidative stress caused by retinaldehyde. There is evidence that RDH13 specifically protects against light-induced apoptosis in photoreceptors and, therefore, inhibits cell death associated with mitochondria.

Our RNA-seq analyses point to the car cluster as one of the main regulatory targets of the CarS protein. In fact, the genes of this cluster are among the most influenced by light. However, other genes strongly regulated by light, as those of the photoreceptors CryD and VvdA, are hardly affected by the carS mutation compared to light, providing examples of the separation of both regulatory mechanisms. CryD and VvdA play roles in the regulation of carotenogenesis by light [21] and the photoinduction of the cryD and vvdA genes was described previously [23, 35]. On the contrary, neither the light nor the carS mutation affect the expression of the gene for the main photoreceptor involved in the photoinduction of carotenoids, the WC-1 protein WcoA. The lack of effect of the carS mutation on the levels of wcoA mRNA does not rule out a possible role of CarS in the control of WcoA function, e.g., at the level of ubiquitination as described for its homolog CrgA on the Wc-1 protein Mcw-1 in M. circinelloides [32]. WcoA is the main transcriptional inducer of the genes of the car cluster, either in light or in darkness [21], so we may anticipate some type of regulatory connection with CarS, which has not been yet clarified.

The analysis of alterations in the transcriptome as a result of illumination or the absence of CarS in the dark reveals large changes in mRNA levels of several genes related to stress functions (Table 2). FFUJ_09119 encodes a FAD flavoprotein belonging to the family of pyridine nucleotide-disulphide oxidoreductases, a group that includes glutathione reductase, trypanothione reductase, lipoamide dehydrogenase, mercuric reductase, thioredoxin reductase, and alkyl hydroperoxide reductase. FFUJ_09320 contains a ferritin-like domain (E-value 1.70e-21) and it is homologous to the transcription factor Rds1, with a function in stress-related responses in Schizosaccharomyces pombe [43] and related to drug sensitivity in Saccharomyces cerevisiae [44]. FFUJ_01993 encodes a protein with an OsmC super family domain (5.83e-41), related to osmotic stress. This family also includes an organic hydroperoxide detoxification protein. Three other proteins strongly regulated by light and CarS contain catalase domains, and therefore could play roles in the defense against oxidative stress. FFUJ_11472 encodes a large protein consisting of a carboxy sulfite oxidase domain (E-value 2.88e-113) followed by a catalase domain (E-value 0). FFUJ_05128 has a catalase domain that covers most of the protein (E-value 0) and is very similar to a putative peroxisomal catalase from the yeast Candida utilis [45], with 54% identity along a stretch of 483 of the 577 residues of the protein. The last one is FFUJ_03407, a small protein (229 residues) with a putative catalase domain segment (residues 61 to 138, 2.40e-29). Light or CarS exerted a lesser influence on other predicted catalase genes of F. fujikuroi. Thus, FFUJ_03451 and FFUJ_11706, orthologs of the cat-1 and cat-2 catalase genes of N. crassa [46], were stimulated 3.2 and 2.1 fold by light but were not significantly influenced by the carS mutation.

Mitochondria are a primary target of H2O2 damage in fungi [47]. This could explain why some of the genes that are induced to a greater degree by light or carS mutation may be involved in the prevention of stress in the mitochondria. Interestingly, carotenoids, especially xanthophylls, have been reported as antioxidant agents in fungi [48]. In fact, different indirect data suggest that neurosporaxanthin has antioxidant activity. In N. crassa, the synthesis of this xanthophyll in response to light is greater in the absence of superoxide dismutase activity [49]. In F. aquaeductuum, the presence of H2O2 in the culture medium stimulates the synthesis of neurosporaxanthin in the dark [12]. Considering the presumed antioxidant properties of carotenoids, the coincidence in the regulation of the expression of carotenogenesis genes with some genes related to oxidative stress is very suggestive.

The GO analyses have revealed the participation of light and CarS in the control of some genes for proteins of potential interest in signaling regulations, such as six genes for predicted histidine kinases, also found under the GO term “Biological regulation” (Additional file 4: Table S3). These proteins participate in two-component systems that are very abundant in prokaryotes, but are also found in lower eukaryotes, where they participate in different signaling pathways [50]. Five of these genes exhibit modest activations by light, usually ca. 2–3-fold, but the putative histidine kinase gene FFUJ_10367 stands out with 33-fold activation under light. Two of them, FFUJ_02424 and FFUJ_13427, are related with the Nik-1 (Os-1) protein of N. crassa, involved in osmotic sensing. However, they exhibit a lower similarity with Nik-1 (17 and 19% identity) than the real Nik-1 ortholog, FFUJ_02727 (80% identity), and therefore they are allegedly involved in other unknown signal transduction mechanism. This is also true for the other putative histidine kinase genes, whose lack of sufficient similarity to functionally characterized orthologs in other organisms hinders the deduction on their putative regulatory functions. In the category of metabolic process, the presence of three putative genes of the nitrate reductase group stands out. In contrast to these genes, the assimilatory nitrate reductase gene FFUJ_12277, niaD (52,9% identity with niaD of A. nidulans) is repressed by light (3,5-fold reduction, Additional file 3: Table S2B). The occurrence of more than one nitrate reductase can be explained by their specialized participation in other metabolic processes, such as denitrification or ammonia fermentation, related with hypoxia and already described in F. oxysporum [51].

The Fusarium species, highly ubiquitous in nature, usually infect a limited range of plant species and possess specific virulence genes. A comparative study between the genomes of Fusarium graminearum, Fusarium verticillioides and F. oxysporum f. sp. Lycopersici revealed the existence of large lineage-specific genomic regions, in some cases covering whole chromosomes, that in F. oxysporum reach up to one quarter of the genome [52]. The existence of these species-distinct genomic regions, presumably involved in their host-specific pathogenicity, explains the high number of non-coincident genes between F. oxysporum and F. fujikuroi. Interestingly, species-specific genes were less affected by light and by CarS than shared genes, suggesting regulatory functions that precede pathogenic specializations in the Fusarium group.

In this work, we have used F. fujikuroi and F. oxysporum to evaluate the degree of conservation in different Fusarium species of transcriptional regulation by light and by CarS. The reason for choosing these species is the availability of carS mutants [25], not yet described in other Fusarium species. The comparison of the effect of light and the carS mutation in F. fujikuroi and F. oxysporum revealed a set of 67 genes activated by both regulatory conditions in both species, of which ca. 20% are uncharacterized proteins. The group of 67 genes exhibits higher transcriptional changes that in average and should be therefore particularly informative on the regulatory relations between the controls by light and the CarS protein in these fungi. With the exception of carD, all the structural genes associated to carotenoid metabolism are found in this group, confirming the major role of light and CarS in the regulation of carotenoid biosynthesis, possibly associated with the formation of photoactive CarO rhodopsin. The rest of the proteins are involved in very diverse functions, related to stress, signaling, and carbohydrate or lipid metabolism. The function of the uncharacterized proteins, or the occurrence of possible functional links between the identified proteins, should be subjects of future research.

Conclusions

Light and the loss of function of the carS gene exert mainly activating effects on the transcriptome of Fusarium fujikuroi, with a greater functional diversity in the case of genes influenced by light. The number of the latter decreases drastically in the carS mutant, indicating that CarS modulates the expression of many genes regulated by light. Approximately 27% of genes activated at least twice by light or by the carS mutation are coincident, raising this percentage to 40% for higher activation thresholds. The large overlap between the two sets of genes confirms regulatory connections between the control of gene expression by light and the CarS protein, which seems to play a role by keeping the mRNA levels of many light-inducible genes low in the dark. According to the magnitude of transcriptional changes, the genes for carotenogenesis are major regulatory targets of light and CarS. Other common regulatory targets are proteins with putative connections with stress responses. They include several genes with catalase domains, which is consistent with roles of light and CarS protein in the control of oxidative stress. Some genes that are strongly induced by light are not affected by the carS mutation, indicating separate mechanisms of action. In this group the genes for the photoreceptors CryD and VvdA, that play a role in the regulation of carotenogenesis, stand out. The effects of light and CarS mutation were also investigated in F. oxysporum, that exhibited a similar global response than F. fujikuroi, but with quantitative and qualitative differences in their respective sets of affected genes. However, 67 genes exhibited a remarkable up-regulation by light and by the carS mutation in both species, indicating the conservation of regulatory roles.

Methods

Strains and culture conditions

The strains of Fusarium fujikuroi were wild type IMI58289, obtained from the Imperial Mycological Institute (Kew, Surrey, England), the carS mutant SG39 [24] and the complemented strain SG256, isolated from a sectoring transformant obtained by introduction of the wild type carS gene in the mutant SG39 [31] (Additional file 1: Figure S2). The strains of Fusarium oxysporum were wild type f. sp. lycopersici 4287 (race 2), kindly provided A. Di Pietro (Universidad de Córdoba, Spain) and the 4287-derived carS mutants SX1 and SX2 [25].

Strains were cultured in DG minimal medium [composition per liter: 30 g glucose, 3 g NaNO3, 1 g KH2PO4, 0.5 g KCl, 0.5 g MgSO4·7H2O and microelements, [53]] or in DGasn medium (the same as DG medium but with 3 g asparagine instead of NO3Na). To obtain conidia of the strains of F. fujikuroi, they were grown on EG medium (composition per liter: 1 g glucose, 1 g yeast extract, 1 g NO3NH4, 1 g KH2PO4, 0.5 g MgSO4·7H2O, 16 g agar).

For RNA-seq or RT-qPCR analyses, 106 conidia were inoculated in 100 ml of DG medium in 500-ml Erlenmeyer flasks and incubated in the dark for 3 days in an orbital shaker. Subsequently, 25 ml samples of the cultures were transferred to Petri dishes under red safelight and incubated for 1 hour in the dark or under white light. Illumination was provided with a set of four fluorescent tubes (Philips TL-D 18 W/840) at a distance of ca. 60 cm, yielding a light intensity of 7 W m− 2. When indicated, shorter light pulses were given followed by a subsequent incubation in the dark for up to 1 hour. The process with F. oxysporum was the same, except that the 3-day cultures were incubated in 145-mm Petri dishes with 80 ml of DGasn medium. Mycelia samples were obtained by filtration, frozen immediately in liquid nitrogen, and stored at − 80 °C.

Expression analysis

The transcript levels of the investigated genes were analyzed by reverse transcription qPCR (RT-qPCR), achieved according to MIQE guidelines [54]. Relevant RT-qPCR methodological details not mentioned in this section are described in Additional file 8. Total RNA samples were extracted from 150 to 200 mg of ground mycelia and treated with DNAse with the NucleoSpin RNA kit (Macherey-Nagel, Düren, Germany). In the case of the effect of light pulse duration (data in Fig. 1), RNA was extracted with the RNeasy RNA isolation kit (Qiagen, Chatsworth, CA, USA). In the case of the effect of light and carS mutation (data in Fig. 4), the extraction procedure is described in the RNA-seq section. RNA concentrations were estimated with a Nanodrop ND-1000 spectrophotometer (Nanodrop Technologies, Wilmington, DE, USA). Concentration and A260/A280 absorbance ratios of the RNA samples are indicated in Additional file 8. Samples of 2.5 μg of RNA were retrotranscribed to cDNA with Transcriptor first-strand cDNA synthesis kit (Roche, Mannheim, Germany), and final cDNA concentrations were adjusted to 25 ng μl− 1. RT-qPCR analyses were performed in a LightCycler 480 real-time instrument (Roche). For amplification and detection, LightCycler 480 SYBR Green I Master (Roche) was used following manufacturer reaction protocol. Genes and primer sets (forward vs reverse in 5′- > 3′ orientation and amplicon length) were carRA (CAGAAGCTGTTCCCGAAGACA vs TGCGATGCCCATTTCTTGA, 65 bp), carB (TCGGTGTCGAGTACCGTCTCT vs TGCCTTGCCGGTTGCTT, 68 bp), FFUJ_10321 (GCTCGCAACCCGTCAATT vs ACGGCGACTTTGTTGATTAGGT, 60 bp), FFUJ_09320 (CAACGATGAGGACCGCTTTC vs AGTCGCGTGGCCGATTT, 61 bp), FFUJ_05128 (GCGCCTTCAGCAACAGACA vs TGACGGTCTGATTCGTAGGATTG, 74 bp), FFUJ_09119 (CCCGGAACCGTTCAATACTG vs TTATCTCCGTTCTGAGCCATCA, 66 bp), FFUJ_01993 (AATCCATCAACCCTCCTCAGATC vs TCTTTGTTCTGCTGGGTGCTT, 68 bp) and FFUJ_03407 (CTCCCAACAGCTTCGCTTACA vs TGTTGTCGCTGACCTGATACG, 69 bp). Coding sequences and amplicon locations are indicated in Additional file 9. Transcript levels for each gene were normalized against the tubulin beta chain gene FFUJ_04397 (CCGGTGCTGGAAACAACTG vs CGAGGACCTGGTCGACAAGT, 69 bp) and the glyceraldehyde 3-phosphate dehydrogenase (GAPDH) gene FFUJ_13490 (GTGACCTCAAGGGCGTTCTG vs CGAAGATGGAGTTTGTGTT, 84 bp). FFUJ_04397 was used as a reference gene for constitutive expression in former RT-qPCR studies in Fusarium fujikuroi (see, e.g., [19, 21, 23, 28, 35]), and the GAPDH gene was recently used as an internal control in RT-qPCR studies in F. graminearum [55, 56].

Statistical differences between selected RT-qPCR data in Fig. 4 were analyzed with the Student’s t-test (unpaired t-test) with the GraphPad Prism 7.0 software (http://www.graphpad.com), using P values of 0.0332 (*), 0.021 (**) and 0.002 (***).

RNA-seq procedures

Total RNA samples were extracted with Trizol (Invitrogen, Paisley, UK) using the protocol described by the manufacturer. The samples were processed with the Illumina protocol by the company Life Sequencing (Valencia, Spain), which consists of the following steps: (1) enrichment in poly-A RNA to discard ribosomal RNA or other non-coding RNAs; (2) fragmentation of RNA into pieces of 100 to 300 bp; (3) synthesis of double strand cDNA by reverse transcription with oligo dT; (4) addition of adapters to terminal ends; and (5) Selection of fragments of appropriate size and subsequent amplification by PCR. Sequencing was achieved with the Illumina HiSeq Platform in 50SE composition using the Single-End methodology.

For statistical and graphic analyses of the sequence data, we used the R programming language, based on our own scripts, and the Bioconductor data analysis libraries (http://www.bioconductor.org). The Integrative Genomics Viewer IGV application (IGV) version 2.3.57 was used for mapping visualization [57].

Bioinformatics analyses

The processing of RNA-seq sequencing data was carried out through the high-performance computing facilities provided by the CICA platform (Centro Informático Científico de Andalucía, Sevilla, Spain) using the Tuxedo protocol. For quality control, the software package FastQC was used. For short read mapping to reference genomes, transcript assembly and differential expression, the software tools Bowtie, TopHat, Cufflinks, Cuffmerge and Cuffdiff [58] were used with default parameters. Tophat was used to map short reads to the reference genome and Cufflinks to assemble sample specific transcripts. The results obtained with Cufflinks were integrated with the Cuffmerge tool to generate the complete transcriptome associated with our study. Gene expression levels were measured as FPKM (fragments per kilobase of exon per million of mapped reads) with the Cuffdiff tool, which normalizes the expression levels obtained in the sequencing eliminating biases due to differences in the length of the transcript and the size of the library [59]. The rest of the analysis and the visualization of the results were done with the CummeRbund package, obtained from the Bioconductor platform (https://www.bioconductor.org), which manages the files through R scripts.

To verify the quality of the sequencing data, the dispersion and squared coefficient variation (SCV) of the samples were analyzed. No substantial overdispersion was observed and SCV did not indicate a high degree of variability between the replicates. The differentially expressed genes were selected based on criteria combining a log2 fold change of 1 and a p-value of 0.05.

GO categories were generated for the genomes of F. fujikuroi and F. oxysporum by detecting protein domains using the PFAM database and applying the available pfam2GO correspondence. This allowed the annotation of approximately 50% of the genes. GO term enrichment analysis was performed using the Bioconductor R package topGO with a p-value of 0.05 computed by Fisher’s exact test. REVIGO (http://revigo.irb.hr) was used to eliminate the redundancy from enriched GO terms based on a measure of semantic similarity and visualize the results as treemaps [37].