Differential gene expression in response to Fusarium oxysporum infection in resistant and susceptible genotypes of flax (Linum usitatissimum L.)

Background Flax (Linum usitatissimum L.) is a crop plant used for fiber and oil production. Although potentially high-yielding flax varieties have been developed, environmental stresses markedly decrease flax production. Among biotic stresses, Fusarium oxysporum f. sp. lini is recognized as one of the most devastating flax pathogens. It causes wilt disease that is one of the major limiting factors for flax production worldwide. Breeding and cultivation of flax varieties resistant to F. oxysporum is the most effective method for controlling wilt disease. Although the mechanisms of flax response to Fusarium have been actively studied, data on the plant response to infection and resistance gene candidates are currently very limited. Results The transcriptomes of two resistant and two susceptible flax cultivars with respect to Fusarium wilt, as well as two resistant BC2F5 populations, which were grown under control conditions or inoculated with F. oxysporum, were sequenced using the Illumina platform. Genes showing changes in expression under F. oxysporum infection were identified in both resistant and susceptible flax genotypes. We observed the predominant overexpression of numerous genes that are involved in defense response. This was more pronounced in resistant cultivars. In susceptible cultivars, significant downregulation of genes involved in cell wall organization or biogenesis was observed in response to F. oxysporum. In the resistant genotypes, upregulation of genes related to NAD(P)H oxidase activity was detected. Upregulation of a number of genes, including that encoding beta-1,3-glucanase, was significantly greater in the cultivars and BC2F5 populations resistant to Fusarium wilt than in susceptible cultivars in response to F. oxysporum infection. Conclusions Using high-throughput sequencing, we identified genes involved in the early defense response of L. usitatissimum against the fungus F. oxysporum. In response to F. oxysporum infection, we detected changes in the expression of pathogenesis-related protein-encoding genes and genes involved in ROS production or related to cell wall biogenesis. Furthermore, we identified genes that were upregulated specifically in flax genotypes resistant to Fusarium wilt. We suggest that the identified genes in resistant cultivars and BC2F5 populations showing induced expression in response to F. oxysporum infection are the most promising resistance gene candidates. Electronic supplementary material The online version of this article (10.1186/s12870-017-1192-2) contains supplementary material, which is available to authorized users.


(Continued from previous page)
Conclusions: Using high-throughput sequencing, we identified genes involved in the early defense response of L. usitatissimum against the fungus F. oxysporum. In response to F. oxysporum infection, we detected changes in the expression of pathogenesis-related protein-encoding genes and genes involved in ROS production or related to cell wall biogenesis. Furthermore, we identified genes that were upregulated specifically in flax genotypes resistant to Fusarium wilt. We suggest that the identified genes in resistant cultivars and BC 2 F 5 populations showing induced expression in response to F. oxysporum infection are the most promising resistance gene candidates.
Keywords: Linum usitatissimum, Flax, Biotic stress, Fusarium oxysporum, High-throughput sequencing, ROS, 1,3-betaglucanase, Cell wall, Background Flax (Linum usitatissimum L.) is a widely distributed crop, which is used for fiber and oil production [1]. Genetic polymorphism of L. usitatissimum and related species is well characterized [2][3][4][5][6][7] and could be used for the breeding of improved cultivars. Although potentially highyielding flax varieties have previously been developed, biotic and abiotic stresses can markedly decrease flax production. Therefore, the molecular mechanisms underlying the responses of flax to unfavorable environments are intensively studied. In this regard, changes in the expression of stress-responsive genes and microRNAs have been detected in flax plants under abiotic stresses, such as drought [8], salinity and alkalinity [9,10], nutrient imbalance [11], and high concentrations of aluminum ions [12,13].
Among biotic stresses, Fusarium oxysporum f. sp. lini is recognized as one of the most devastating flax pathogen. It causes wilt disease, which is one of the major limiting factors for flax production in most of the flaxgrowing areas worldwide. Epidemics of the disease can result in an 80% to 100% loss in yield [14]. Breeding and cultivation of flax varieties resistant to F. oxysporum is the most effective method for controlling wilt disease, and in this regard, evaluation of flax germplasm for resistance to Fusarium wilt has revealed accessions with potential utility in breeding programs [15]. Furthermore, the search for genes conferring resistance to Fusarium infection is currently underway, and amplified fragment length polymorphism (AFLP) analysis of a flax mapping population derived from doubled haploid lines has already led to the identification of two quantitative trait loci associated with resistance to Fusarium wilt [16]. However, the genes that define resistance to Fusarium in some flax genotypes remain unknown.
Alterations that occur in flax plants under Fusarium infection have been actively studied and, in some cases, the molecular mechanisms underlying responses have been elucidated. The role of pathogenesis-related (PR) proteins, including chitinase and β-1,3-glucanase, in response to Fusarium has been revealed. Upregulation of chitinase genes has been identified in flax plants under F. oxysporum infection [17]. Flax lines with ectopic expression of the β-1,3-glucanase gene or overexpression of endogenous β-1,3-glucanase gene show enhanced resistance to F. oxysporum and F. culmorum [18,19]. Moreover, those flax plants with overexpressed β-1,3glucanase have increased contents of antioxidants, phenolics, and polyamines, as well as alterations in cell wall biopolymer composition [18][19][20]. Enhanced resistance via an increase in antioxidant activity has also been observed in transgenic flax plants with increased contents of flavonoids, carotenoids, or other terpenoids [21][22][23]. Furthermore, the involvement of antioxidants and cell wall components in the flax response to Fusarium has been demonstrated in different plant material, including cell cultures, seeds, and seedlings. Oxidative burst, activation of lipid peroxidation, and phenylpropanoid metabolism have been observed in flax cells under interaction with F. oxysporum [24]. The contribution of the antioxidant potential of phenylpropanoids, which accumulate in seeds, and pectin content in flax resistance to Fusarium have also been identified [25], as have the changes in pectin metabolism in flax seedlings under Fusarium infection [26]. Changes in the expression of genes participating in stress response, defense response, metabolism regulation, and, in particular, the phenylpropanoid pathway have been detected in flax plants during the early stages of Fusarium infection [27], and it has been suggested that an increase in methyl salicylate level in flax plants in response to F. oxysporum is associated with activation of the phenylpropanoid pathway [28]. The role of polyamines in response to Fusarium has also been revealed [29]. RNA-seq of flax plants after infection with F. oxysporum allowed identification of changes in the expression of genes involved in signal transduction, regulation of transcription, hormone signaling, reactive oxygen species (ROS) regulation, secondary metabolism, and other processes [17]. Thus, it has been variously established that PR-proteins, antioxidants, and cell wall components are involved in the flax response to Fusarium infection.
In the present study, we used high-throughput sequencing of transcriptomes to evaluate the changes in flax gene expression under F. oxysporum infection in resistant and susceptible flax cultivars and BC 2 F 5 populations, the latter of which were obtained from crosses between resistant and susceptible flax cultivars and then selected for both resistance to F. oxysporum and phenotypical similarity with the susceptible parent for several generations. This approach allowed us to identify candidate genes conferring resistance to F. oxysporum infection in L. usitatissimum.

Plant material
Experiments for identification of flax cultivars with resistance and susceptibility to F. oxysporum have previously been performed at the All-Russian Research Institute for Flax (Torzhok, Russia). Based on the obtained results, two resistant (Dakota and #3896) and two susceptible (AP5 and TOST) cultivars were selected for examination in the present study. In addition, hybrids of cultivars resistant and susceptible to F. oxysporum were obtained at the same institute. The susceptible cultivar AP5 was crossed with both resistant cultivars (Dakota and #3896), and the resulting F 1 plants were backcrossed to AP5. Subsequently, selection against a provocative background (soil inoculated with an isolate of Fusarium oxysporum f. sp. lini) was performed and plants that were resistant but phenotypically similar to the AP5 cultivar were selected. A further backcross was then conducted and resistant plants similar to AP5 were again selected. Thereafter, self-pollination of BC 2 F 1 individuals and selection of resistant families that were phenotypically similar to AP5 were performed for five generations. As a result, BC 2 F 5 populations resistant to F. oxysporum were obtained: #3896 × АР5 (recurrent parent АР5) and Dakota × АР5 (recurrent parent АР5).
Thus, two cultivars with resistance (Dakota and #3896) and two cultivars with susceptibility (AP5 and TOST) to F. oxysporum, as well as resistant BC 2 F 5 populations (#3896 × АР5 and Dakota × АР5), were used in our study. Seeds were initially sterilized in 70% ethanol for 1 min and in 1% sodium hypochlorite for 20 min, and then rinsed 15 times in sterile deionized water. The plants were grown in sterile 16 mm × 150 mm glass tubes on Murashige-Skoog medium in a growth chamber at 22°C with a 16 h day and 8 h night.
Fusarium oxysporum f. sp. lini pathogenic isolate #39 from the phytopathogen collection of the All-Russian Research Institute for Flax was grown on potato dextrose agar for 5 days prior to inoculation. This was the same isolate that was applied for selection of resistant families after crosses between Dakota and AP5 and #3896 and AP5. Seven-day-old flax plants were inoculated with 1 ml of a 10 5 per ml preparation of F. oxysporum spores (fungal infection) or with 1 ml of sterile water (control). After 48 h, when necrosis of roots had appeared, root tips (approx. 5 mm in length), in which the infection initially occurred, were collected and frozen in liquid nitrogen. It was previously shown that F. oxysporum displays a clear preference for the root tips of flax. Two days after inoculation, the fungus was mainly distributed around the root tips, with significantly less presence in the elongation zone and in lateral branches [30]. Accordingly, for studying the early infection stages, we selected root tips as the most preferable experimental material. In total, approximately 120 infected plants and 120 plants grown under control conditions were obtained for four cultivars and two BC 2 F 5 populations.

Library preparation and transcriptome sequencing
Total RNA was extracted from pooled plant samples using an RNeasy Plant Mini Kit (Qiagen, USA). Each pool included 10-12 plants of each cultivar/population under control conditions or fungal infection. Thus, 24 RNA samples were extracted in two replicates under either fungal infection or control conditions: Dakota, #3896, AP5, TOST, BC 2 F 5 #3896 × АР5, and BC 2 F 5 Dakota × АР5. RNA concentration and quality were evaluated using a Qubit 2.0 fluorometer (Life Technologies, USA) and Agilent 2100 Bioanalyzer (Agilent Technologies, USA). Highquality RNA samples (RNA integrity number not less than 8.0) were used for cDNA library preparation with polyAbased mRNA capture using a TruSeq Stranded Total RNA Library Prep Kit (Illumina, USA). The quality of the 24 obtained libraries was evaluated using the Agilent 2100 Bioanalyzer. The libraries were sequenced using a Next-Seq 500 high-throughput sequencer (Illumina) and paired-end reads (80 + 80 nucleotides) were obtained.

High-throughput sequencing data analysis
Illumina reads were trimmed and filtered using Trimmomatic [31] and then F. oxysporum reads were filtered out by mapping to the F. oxysporum reference genome and transcriptome (NCBI assembly/WGS identifier ASM14995v2/ AAXH01) using bowtie2. The remaining reads were used for transcriptome assembly using Trinity 2.4.0 with the default parameters [32]. The assembly was performed (1) for each cultivar/population, (2) for each BC 2 F 5 population jointly with the corresponding parents, and (3) for all sequenced flax samples together.
The quality of assemblies was assessed with N50, ExN50, and L50 statistics using QUAST 4.5 and Trinity utilities. Trasncripts that were less than 200 nucleotides in length were excluded from subsequent analysis. The transcripts were also analyzed with BUSCO to evaluate the completeness of the assembly [33]. Trinotate pipeline was then used for the annotation of the assembled transcripts (http://trinotate.github.io/). The derived transcripts were analyzed for the presence of open reading frames (ORFs) using TransDecoder [34]. The transcripts and the predicted proteins were aligned to the UniProt database using blastx and blastp, respectively. The protein sequences were scanned for the presence of PFAM domains using HMMER [35,36]. On the basis of these data, a local SQLite database was constructed and transferred to Trinotate. Finally, the transcripts were annotated using the Gene Ontology (GO), KEGG, and COG databases.
Then reads were mapped to the assembled transcripts (all nine assemblies) and quantified using bow-tie2 [37] and rsem [38]. Read counts per transcript and per gene were calculated. The derived read count data were analyzed using edgeR [39]. After normalization using the TMM method, we attempted to identify the following gene responses: Only genes with a CPM greater than 2.0 for at least three samples were used for further analysis. The t-test was used to determine p-values. False discovery rate (FDR) values were derived using the Benjamini-Hochberg p-value adjustment procedure. The gene set enrichment analysis (GSEA) with GO data was performed using Goseq [40]. For this analysis, we used lists of the top 50, 100, 200, 500, 1000, and 2000 upregulated or downregulated genes, separately. Different GO terms were enriched     Genes showing changes in expression under F. oxysporum infection were identified in both resistant and susceptible flax genotypes. GO analysis was performed for the top 100 up-and down-regulated genes. In the flax genotypes with resistance to F. oxysporum (resistant cultivars and BC 2 F 5 populations), the top 100 upregulated genes were related to NAD(P)H dehydrogenase activity, oxidoreductase activity, respiratory chain complex I, and mitochondrial parts ( Table 2). In the susceptible cultivars, the top 100 upregulated genes were related to other categories, including translation, ribosome, biosynthetic process, and cytosolic part (Table 3). GO analysis of the top 100 downregulated genes also revealed differences between resistant and susceptible genotypes: in resistant genotypes, the genes were related to microtubule, kinesin complex, cytoskeletal part, cell junction, clathrin coat, and adenyl nucleotide binding (Table 4); in susceptible genotypes, the genes were related to cell wall, external encapsulating structure, transferase activity, transferring glycosyl groups, polysaccharide metabolic process, water channel activity, intrinsic component of membrane, and xyloglucan metabolic process (Table 5).
Genes specifically upregulated in resistant genotypes of flax in response to F. Oxysporum For identification of candidate genes responsible for resistance to F. oxysporum in flax, we searched for genes that were up-or down-regulated in resistant cultivars and BC 2 F 5 populations under F. oxysporum infection, but did not show a change in expression (or showed less change) in susceptible cultivars. The full results are presented in Additional file 1 and the results for the top 30 differentially expressed genes (excluding two unknown genes) are presented in Table 6. Note: One gene can belong to more than one GO category. FDRfalse discovery rate Upregulation was revealed for genes encoding SRG1 (senescence-related gene 1) protein, UDP-glycosyltransferase 73C3 (UGT73C3), AAA-ATPase ASD, mitochondrial (AATPA), glucan endo-1,3-beta-glucosidase, MYB transcription factors, ERD dehydrins, and Auxin-responsive protein SAUR, among others. We suggest that the identified genes with specifically induced expression in response to F. oxysporum infection in resistant cultivars and resistant BC 2 F 5 populations are the most promising resistance gene candidates. GO terms with the most significant differences between flax genotypes resistant and susceptible to the fungus, and the expression profiles of related genes are presented in Additional file 2. In resistant cultivars and populations, genes involved in the response to biotic stimulus and stress, defense response, antioxidant activity, and cell wall organization or biogenesis were more strongly upregulated than in the susceptible cultivars.

Discussion
Plant mechanisms of response to Fusarium infection include synthesis of PR proteins and antimicrobial compounds, production of ROS, and changes in cell wall structure [41][42][43][44][45]. In the present study, we evaluated the changes in gene expression in response to F. oxysporum infection in resistant and susceptible flax cultivars and resistant BC 2 F 5 populations. The advantage of our study is the use of the two BC 2 F 5 populations, which were obtained from crosses between the examined resistant and susceptible cultivars, and which are resistant to F. oxysporum but phenotypically similar to the susceptible parent. This approach allowed us to compare the changes in gene expression under F. oxysporum infection in resistant and susceptible genotypes and to identify genes that were specifically induced in resistant flax plants in response to the infection.
Significant downregulation of genes involved in cell wall organization or biogenesis was detected in response to F. oxysporum in susceptible cultivars (AP5 and TOST). However, we observed no similar trend in resistant cultivars and populations. It could be suggested that, in susceptible cultivars, changes in apoplast structure in response to F. oxysporum are more pronounced. The role of cell wall compounds in the response of flax [26,46] and other plant species [47][48][49] to F. oxysporum has been revealed previously. However, the present study is the first to identify the differential expression of genes related to cell wall organization or biogenesis in flax cultivars and BC 2 F 5 populations with different resistance to Fusarium wilt.
A number of the top 100 upregulated genes in resistant cultivars are related to NAD(P)H oxidase activity. In susceptible cultivars, we also revealed upregulation of NAD(P)H oxidase-related genes; however, most of these were not included in the top 100 upregulated genes. NAD(P)H oxidases are involved in ROS signaling and stress response in plants, and are one of the sources of ROS that are induced in response to pathogen attack and involved in early defense responses via an oxidative burst [45,[50][51][52][53][54][55][56][57]. NADPH oxidase upregulation and early oxidative burst have been revealed in a resistant banana cultivar in response to F. oxysporum infection [56,58,59]. In flax plants, we observed a similar trend. ROS signaling associated with NAD(P)H oxidases could be one of the mechanisms constituting the L. usitatissimum defense response to F. oxysporum. We accordingly suggest that NAD(P)H oxidases could be promising candidates for proteins that participate in the defense response against F. oxysporum in flax.
The genes that were induced more strongly in the resistant genotypes of flax compared with the susceptible genotypes in response to F. oxysporum infection are involved in crucial biological processes, including transcription regulation, auxin signaling, stress response, and photosynthesis. The most significant upregulation was observed for genes encoding SRG1 protein, UGT73C3, AATP5, glucan endo-1,3-beta-glucosidase (beta-1-3-glucanase), and epidermis-specific secreted glycoprotein EP1. Among these proteins, beta-1-3-glucanase is the most well-known fungal-responsive protein in flax. This enzyme hydrolyzes beta-1,3-glucans of the cell wall in fungi, and the role of this protein in plant defense against pathogens is well known [60][61][62][63][64]. In flax, increased resistance to F. oxysporum and F. culmorum has been observed in transgenic flax lines containing the potato beta-1,3-glucanase gene, and in plants overexpressing the beta-1,3-glucanase gene [18,19]. We also revealed the upregulation of this gene in flax plants in response to F. oxysporum infection, and the changes were observed to be stronger in the cultivars and BC 2 F 5 populations showing resistance to Fusarium wilt (Fig. 1). In resistant genotypes under F. oxysporum infection, the induction of beta-1,3-glucanase expression was more pronounced compared with that in susceptible genotypes (p < 0.01, Mann-Whitney test). Moreover, under the stress conditions, the expression level of beta-1,3-glucanase was significantly higher in resistant cultivars and populations (p < 0.05), whereas under control conditions, there was no significant difference between the resistant and susceptible genotypes.
Thus, in response to F. oxysporum infection, we revealed changes in the expression of genes that encode PR proteins, and also in genes that are involved in ROS production and cell wall structure change. Furthermore, we identified genes that were specifically upregulated in flax genotypes resistant to Fusarium wilt. Special attention should be given to these genes in further searches for resistance gene candidates. Our work complements previously obtained results on flax response to F. oxysporum infection and provides a basis for detailed investigations of flax defense mechanisms against Fusarium wilt.

Conclusions
In the present study, we used high-throughput sequencing to search for genes involved in the early defense response of L. usitatissimum against infection by the fungus F. oxysporum. To this end, we first used resistant and susceptible flax cultivars and F. oxysporum-resistant BC 2 F 5 populations, which were obtained from crosses between the resistant and susceptible flax cultivars. An analysis of gene expression revealed diverse patterns of differentially expressed genes for resistant and susceptible flax genotypes. Genes involved in response to biotic stimulus and stress, defense response, antioxidant activity, and cell wall organization or biogenesis were more strongly upregulated in the resistant genotypes than in the susceptible genotypes. Moreover, we identified genes that were specifically induced in genotypes resistant to Fusarium wilt in response to F. oxysporum infection. These genes are the most promising candidates for genes conferring resistance to F. oxysporum infection in L. usitatissimum.