IntroductionTop

Material and methodsTop

Plant material

In our study, we evaluated 48 durum wheat (T. turgidum spp. durum) accessions against an isolate of the hemibiotrophic fungus Z. tritici. Amongst them, 26 accessions were commercial cultivars, selected from the Red Andaluza de Experimentación Agraria (RAEA) (Castilla et al., 2019), and 22 were breeding lines, belonging to the wheat breeding program being developed at IFAPA (Spain). From the RAEA commercial cultivars studied, six were considered as control-checks due to their role as historical checks in Spanish durum wheat experiments. The other 22 were recently registered varieties (Table S1 [suppl]). After a first assessment of disease susceptibility, the historical check ‘Amilcar’ was selected as the reference susceptible control accession.

Nine wheat accessions with diverse disease severity (DS) rating scale were randomly selected for the image analysis assays. The DS rating scale goes from 0 to 5 (McCartney et al., 2002), where 0 = immune with no visible symptoms, 1 = highly resistant with hypersensitive flecking, 2 = resistant with small chlorotic or necrotic lesions and no pycnidial development, 3 = moderately resistant, characterized by coalescence of chlorotic and necrotic lesions with slight pycnidial development, 4 = susceptible with moderate pycnidial development and coalesced necrotic lesions, and 5 = very susceptible with large, abundant pycnidia and extensively coalesced necrotic lesions (Fig. 1). These accessions were: ‘LG Origen’, ‘BL 39’, ‘Sculptur’ (DS 3), ‘Athoris’, ‘BL 36’, ‘BL 33’ (DS 4), ‘Avispa’, ‘BL 34’ and ‘Amilcar’ (DS 5).

 

 

Figure 1.  Example of leaves infected with Septoria tritici blotch (STB), showing diverse disease severity (DS) scores. Leaves from breeding lines, commercial cultivars, and control-checks with different DS scores: DS 2, (a) ‘RGT Rumbadur’ and (b) ‘RGT Voilur’; DS 3, (c) ‘LG Origen’, (d) ‘BL 39’, and (e) ‘Sculptur’; DS 4, (f) ‘Athoris’, (g) ‘BL 36’, and (h) ‘BL 33’; DS 5, (i) ‘Avispa’, (j) ‘BL 34’, and (k) ‘Amilcar’. Disease severity is presented according to McCartney et al. (2002), see Material and Methods for rating scale.

 

 

Pathogen isolation and molecular characterization

Zymoseptoria tritici was isolated from a naturally infected field of Santaella, Córdoba (Spain). Infected leaves were cut into pieces and placed horizontally in a square culture dish (120 mm × 120 mm) containing wet filter paper to maintain high humidity for 24 or 48 h. Mature pycnidia were observed on the leaf surface, as well as the conidia, in a gelatinous matrix of cirri. We used a sterile needle and a dissecting microscope to collect cirri from the pycnidia and transfer them onto potato-dextrose-agar (PDA; Difco Laboratories Inc.) plates, supplemented with streptomycin (50 mg/L) and chloramphenicol (250 mg/L). PDA plates were incubated for 48 h at 20 °C in the dark. The colonies grew in the form of yeast-like spores. Once the colony began to darken, the spores were ready to harvest. These colonies were subsequently transferred with a sterile needle to PDA plates every 48 h to finally isolate the fungus Z. tritici avoiding cross-contamination from other microorganisms. Stocks of Z. tritici were obtained according to the procedure developed by Stewart & McDonald (2014) for fungal-isolate retrieval, with minor modifications. Colonies were collected using a sterile scalpel, cutting them into small pieces of PDA, and mixing them in a 250 mL conical flask containing 50 mL of yeast-malt-sucrose liquid medium (4 g of yeast extract, 4 g of malt extract, 4 g of sucrose, and 1 L of H2O). Flasks were then sealed with cotton stoppers and aluminum foil and placed on a shaker at 210 rpm and 18 °C in the dark. After 8 to 10 days of growth, 25 mL of spore concentrate was collected in a 50 mL Falcon tube and centrifuged at 3500 rpm (Eppendorf Centrifuge 5810R) at room temperature for 15 min to precipitate the spores. After removing the supernatant, the spore pellet was thoroughly mixed with distilled water. Fungal stocks were stored as microconidial suspensions at -80 °C with 30% glycerol until needed.

Molecular identification of Z. tritici isolate was based on the amplification and sequencing of the internal transcribed spacer (ITS) regions of ribosomal DNA (rDNA). Mycelium for DNA extraction was grown on PDB (Scharlab, S.L. Spain). Genomic DNA was extracted using the Plant DNeasy Mini kit (Qiagen, Germany) according to the manufacturer’s instructions. A 290 bp fragment of the ITS region was amplified using the pair primers ITS1 (5’ TCC GTA GGT GAA CCT GCG G 3’) and ITS2 (5’GCT GCG TTC TTC ATC GAT GC 3’) (White et al., 1990). PCR products were purified using the FavorPrepTM purification kit (FAVORGEN, China). DNA was Sanger sequenced at SCAI facility at University of Córdoba (Spain) with the forward and reserve primers as in PCR. Sequences were subjected to BLAST analysis at NCBI (https://blast.ncbi.nlm.nih.gov/Blast.cgi). The ITS sequences were deposited at the NCBI database under the SUB9540116 accession number.

 

Inoculation assays

Our inoculation procedure was developed as described by Stewart & McDonald (2014), with minor modifications. Seeds of the 48 durum wheat accessions mentioned above (commercial cultivars and breeding lines) were sown in 8 × 7 × 7 cm pots containing a mix (1:1, v/v) of commercial compost (Suliflor SF1 substrat; Suliflor Lithuania) and sand. Pots were placed in trays and incubated in a growth chamber at 21 °C and 70% relative humidity (RH), with 16 h of light. At the same time, fungal spores were retrieved from the spore suspension at -80 °C, before adding 150 µL of the spore suspension to a 250 mL conical flask containing 100 mL of yeast-malt-sucrose medium. The flask was then placed on a shaker as described above to obtain fresh spores as inoculum. After 16 days, seedlings were inoculated when the second and third leaves emerged (growth stage Z13; Zadoks et al., 1974) in a spore suspension prepared with distilled water, to which Tween-20 was added (0.1%). Spore concentration was measured using a hemocytometer, and conidia were adjusted to 107 spores mL-1. Four seedlings of each accession were inoculated with 30 mL of spore solution using a hand-sprayer until the solution ran off the leaves. A total of 192 (8 leaves used per accession) plants were inoculated in each biological replicate. Once leaves were totally dry, plants were sealed in clear plastic bags to provide 100% RH, maintained for 48 h in a growth chamber at 22/18 °C day/night with a 16 h photoperiod. After 48 h, the plastic bags were removed, and plants were kept in the same conditions with a humidity level of 75-80% using humidifiers to promote infection. The susceptible control check ‘Amilcar’ was inoculated following the same treatment, under the same conditions and with the same replications as the other accessions. This experiment was performed three times with similar results.

 

Disease assessment

At 20 days post inoculation, the second and third leaves of each plant were evaluated. The infection process was quantitatively scored as the percentage of necrotic leaf area (NLA), including both sporulating (Chartrain et al., 2004) and nonsporulating areas (Suffert et al., 2013; Gerard et al., 2017). The percentage of NLA also included chlorotic areas of accessions which did not develop necrosis. In addition, seedling reactions were qualitatively scored using the DS rating scale from 0 to 5 (McCartney et al., 2002). Reaction types 0-3 were considered resistant, while reaction types 4 and 5 were considered susceptible. Although reaction type 3 includes pycnidium development, it is considered resistant because the growth and sporulation of the fungus is quite restricted, and the chlorotic reaction is similar to the chlorotic blotches of reaction type 2. The same analysis was performed with adult plants (2 months) in a growth chamber to confirm the absence of variation in resistance with respect to the seedlings (Fig. S1 [suppl]).

Once all plants were scored, 2-cm sections of the middles of leaves with STB symptoms were cut and photographed using a digital camera (Canon Powershot SX710 HS) and a ruler as guidance for measurements. Leaves from three accessions with a DS of 2, 3, 4, and 5 were randomly selected in order to support the DS rating scale through image analysis using the imaging software NIS-Elements (vers. 4.50; Nikon Instruments Inc.). First, images were cropped into NLA sections of 25 mm2 bearing pycnidia. Then, the software permitted a color threshold analysis that distinguished pycnidia amongst the STB lesions. Subsequently, the number and area of pycnidia were measured using established size and circularity parameters. In order to double-check the images, manual selection of pycnidia was carried out in cases where the pycnidia were not selected according to the established parameters. Lastly, the total pycnidial area (mm2 ), number of pycnidia, and pycnidia size (mm2 ) were calculated. For pycnidium assessment, three cuts per accession were analyzed and in two independent experiments.

 

Data analyses

The experimental design was developed as random blocks. The percentage of NLA covered with pycnidia

was transformed according to the formula y = log(x), whereas pycnidium count and the number of pycnidia per mm2 of NLA were transformed according to the formula y = x1/2. ANOVA and Tukey’s honestly significant difference (HSD) test were performed on these data. Pycnidium size was analyzed using the Kruskal-Wallis test. Data analyses and figures were carried out using R software (R Core Team, 2020), NIS-Elements, and ImageJ (Schneider et al., 2012).

STB infection studies

The development and the aggressiveness of STB in 48 durum wheat accession is shown in Fig. 2. The accessions expressed, on average, 62.2% NLA, with most accessions falling in the range of 55-75%, indicating the great development capability of Z. tritici amongst them. Accession ‘BL 41’, which showed the highest NLA value (79%), belonged to the group of selected lines from our current breeding program, while line ‘BL 45’ showed the lowest NLA value (50%). Breeding lines presented, on average, 64% NLA. Commercial cultivars also expressed a high percentage of NLA, with an average value of 60.5% in this group. Most commercial cultivars had an NLA ranging from 55% to 75%, except for ‘LG Origen’ (33.3%), ‘Don Ortega’ (35.8%), ‘Sculptur’ (39.8%), ‘Nobilis’ (50.2%), and ‘RGT Rumbadur’ (50.2%). Cultivar ‘Teodorico’ presented the highest NLA value amongst commercial cultivars (76.4%), followed by ‘RGT Fernandur’ and ‘SY Leonardo’, with 75.4% and 74.8% NLA, respectively. Lastly, the control-check cultivars (selected from commercial cultivars) showed mean levels of NLA ranging from 42.9% (‘Kiko Nick’) to 77.7% (‘Amilcar’). This group of accessions presented similar NLA values to the breeding lines and the commercial cultivars, with an average NLA value of 61% (Fig. 2).

Disease severity scores showed differences among the accessions studied. Most of the breeding lines (18 accessions out of 22) presented a DS of 4 (susceptible), which denotes significant fungus-reproduction capability in the form of pycnidia in the necrotic lesions. In addition, two breeding lines, ‘BL 34’ and ‘BL 40’, expressed a DS of 5 (very susceptible) with high reproduction of the fungus and extensively coalesced lesions (Fig. 2). In contrast, breeding lines ‘BL 28’ and ’BL 39’ presented a DS score of 3 (moderately resistant), indicating that the development of Z. tritici produced lesions in the form of necrosis but to lesser extent and with limited presence of pycnidia. Commercial cultivars exhibited more variability in DS scores, showing accessions ranging from 2 to 5 (Fig. 2). This group presented more lines with a DS of 5 (very susceptible) than the breeding lines; however, in contrast, the included accessions that exhibited resistance (DS of 2), such as ‘RGT Rumbadur’ and ‘RGT Voilur’, only showed necrotic or chlorotic lesions without the presence of pycnidia. Commercial cultivars also included two accessions with moderate resistance (DS of 3): ‘LG Origen’ and ‘Sculptur’. Regarding control-check cultivars, each DS score (except for a DS of 2) featured two accessions; ‘Amilcar’ and ‘Avispa’ showed a DS of 5, ‘Athoris’ and ‘Euroduro’ showed a DS of 4, and ‘Kiko Nick’ and ‘Simeto’ showed a DS of 3. It should be noted that none of the accessions studied presented a DS of 0 (immune with no visible symptoms) or 1 (highly resistant with hypersensitive flecking). Lastly, all studied accessions were organized to analyze their distribution according to DS and group (Fig. 3). In total, 40 out of 48 accessions (83.3%) could be considered susceptible to STB as they showed a DS of 4 or 5, while only 8 out of 48 accessions (16.6%) exhibited resistance with a DS of 2 or 3 (Fig. 3).

 

 

Figure 2.  Septoria tritici blotch (STB) infection in durum wheat accessions. Mean percentage of necrotic leaf area (NLA), presented in columns, and disease severity (DS) rating scale, presented as numbers at the top of the figure for breeding lines, commercial cultivars, and control-checks. Accessions were arranged according to their mean percentage of NLA and classified by DS in panels. DS is presented according to McCartney et al. (2002), see Material and Methods for rating scale. Error bars represent the SE calculated from three independent experiments with eight replicates each.

 

 

Figure 3.  Classification of accessions according their disease severity (DS) (McCartney et al., 2002), characterized as breeding lines, commercial cultivars and control-checks.

 

 

Evaluation of STB symptoms through image analysis

Leaves from 9 accessions, which presented diverse DS symptoms ranging from 3 to 5, were selected for image analysis with the objective of better understanding pycnidial development through wheat genotypes. The peak of NLA was reached a couple of days before the peak of pycnidium development. An example of leaves showing DS scores of 2-5 amongst breeding lines, commercial cultivars, and control-check cultivars is shown in Fig. 1. Accessions with a DS score of 2 expressed some chlorotic and necrotic lesions without pycnidial development, being considered as resistant (Fig. 1a,b). However, in addition to accessions with a DS of 3 showing similar lesions to accessions with a DS of 2, they developed a few pycnidia (Fig. 1c-e). The accessions with a DS of 4 presented moderate necrotic and coalescent lesions, while the presence of pycnidia was increased compared to accessions with a DS of 3 (Fig. 1f-h). Lastly, accessions presenting a DS score of 5 developed a high density of necrotic lesions toward the leaf, in addition to a higher number of pycnidia compared to accessions with DS scores of 3 and 4 (Fig. 1i-k).

Images were then analyzed using image analysis software (NIS-Elements; Nikon Instruments Inc.) in order to obtain numeric parameters, to precisely describe the quantitative differences in STB symptoms amongst the accessions studied (Table 1). Images were cropped into NLA sections of 25 mm2 and then analyzed. ‘LG Origen’, ‘BL 39’, and ‘Sculptur’ were classified with a DS of 3 and, through image analysis, they showed mean NLA values of 0.92%, 0.83%, and 1.12% covered with pycnidia, respectively. In the same NLA sections analyzed, these accessions presented, on average, pycnidium counts of 65 (‘LG Origen’), 85 (‘BL 39’), and 90 (‘Sculptur’), implying mean values of 2.59, 3.40, and 3.59 pycnidia per mm2 of NLA, respectively (Table 1). Pycnidial parameters of accessions with a DS of 3 presented lower values in comparison with accessions with a DS of 4 or 5. ‘Avispa’, ‘BL 34’, and ‘Amilcar’ were classified with a DS of 5 and showed mean values of 11.93, 14.25, and 16.71 pycnidia per mm2 of NLA, with total counts, on average, of 298, 356, and 418, respectively. These accessions also expressed higher mean values of NLA covered with pycnidia compared to accessions with a DS of 3, whereby ‘Avispa’, ‘BL 34’, and ‘Amilcar’ presented values of 5.70%, 4.83%, and 5.34%, respectively (Table 1). ‘Athoris’,’BL 36’, and ‘BL 33’ were classified with a DS of 4 and presented moderate pycnidial development, with numeric values located between those of accessions with a DS of 3 and a DS of 5. These accessions presented, on average, 2.04% (‘Athoris’), 2.71% (‘BL 36'), and 2.50% (‘BL 33’) NLA covered with pycnidia. ‘Athoris’, ‘BL 36’, and ‘BL 33’ presented, on average, pycnidium counts of 173, 198, and 210 in the NLA sections analyzed, respectively, which translates to 6.92, 7.92, and 8.40 pycnidia per mm2 of NLA (Table 1). All accessions analyzed expressed statistically similar values of pycnidium size.

However, the remaining parameters showed statistical differences among the accessions. Mean pycnidium count and, therefore, mean number of pycnidia per mm2 of NLA were statistically different among accessions classified with a DS of 3, 4, and 5 (Table 1). Moreover, some accessions classified in the same DS group showed differences, such as ‘Avispa’ and ‘Amilcar’. Similarly, the percentage of NLA covered with pycnidia was statistically different in accessions with a DS of 3, 4, and 5. Only ‘BL 36’ (DS of 4) and ‘BL 34’ (DS of 5) were not statistically different. Table 1 also shows the parameters analyzed in terms of their mean values grouped by DS score, showing statistical differences for accessions with a DS of 3, 4, and 5 for all parameters evaluated except for pycnidium size. Pycnidium count increased, on average, from 80 in DS 3, to 194 in DS 4 and 357 in DS 5, thereby implying a similar increase in the number of pycnidia per mm2 of NLA, with values of 3.19 (DS of 3), 7.75 (DS of 4), and 14.30 (DS of 5). These increases in pycnidium count and the number of pycnidia per mm2 of NLA were correlated with a similar increase in the NLA covered with pycnidia, with average values of 0.96% in accessions with a DS of 3, 2.41% in accessions with a DS of 4, and 5.29% in accessions with a DS of 5.

 

 

Table 1.  Pycnidium parameters in NLA lesions of 25 mm2 in accessions with different DS scores [1].

DS = disease severity; NLA = necrotic leaf area. [1] Values are means ± standard deviation for three leaves evaluated through image analysis in three accessions classified per DS score. Data with the same letter within a column are not significantly different (HSD, p<0.05).

 

 

In our study, the incidence of STB in the durum wheat accessions was, in general, severe. Most of the durum wheat accessions studied presented 50% or higher mean values of NLA, which suggests an elevated development of Z. tritici in the leaf tissue amongst groups with diverse genetic origin (breeding lines and commercial cultivars). These results of high STB progression through diverse durum wheat accessions are in accordance with results in other studies (Ghaneie et al., 2012; Chedli et al., 2018). In addition to the high disease incidence in most of the commercial cultivars (and control-checks) studied (60% NLA or higher), it could be noted that they showed a DS score corresponding to them being susceptible (DS of 4) or very susceptible (DS of 5). These disease parameters imply that STB would produce not only severe infections through colonized leaf tissue, leading to yield losses, but also great persistence and reproduction capability in the form of pycnidia, increasing the sexual recombination of the fungus and its possibilities of overcoming wheat resistance (McDonald & Mundt, 2016), explaining the current incidence of STB in Spanish wheat-cropping areas, especially in Andalusia (Royo & Briceño-Félix, 2011). Breeding lines presented the same disease patterns as commercial cultivars and control-checks in terms of NLA and DS score; however, in comparison, they almost all (18 of 22 lines) presented a DS of 4 (susceptible). This DS score in breeding lines, together with the high levels of NLA mentioned, highlights the lack of resistant resources in current Spanish wheat-breeding programs to face Z. tritici; therefore, there is a necessity to research and introduce resistant wheat genetic material against this pathogen in tailored programs, as seen in other Mediterranean areas (Ghaneie et al., 2012; Ferjaoui et al., 2015; Kidane et al., 2017).

Although Z. tritici presented, in general, high disease incidence in our studied accessions, we found some promising sources of resistance amongst them, such as in some control-check cultivars. The occurrence of these resistance patterns in our selected control-check cultivars is explained by the fact that they were chosen because of their role as historical checks, underscoring the shortage of studies related to the resistance and susceptibility of Spanish durum wheat against STB from which we could have selected control-check accessions with a certain degree of resistance. There were some commercial and control-check cultivars which expressed reduced mean values of NLA in comparison with the remaining accessions studied, such as ‘LG Origen’ (33.3%), ‘Don Ortega’ (35.8%), ‘Sculptur’ (39.8%), and ‘Kiko Nick’ (42.9%). These cultivars (except for ‘Don Ortega’) also exhibited a DS of 3, suggesting that they not only restricted the colonization of leaf tissue by the fungus, but also presented a significant reduction in pycnidium development, similar to breeding lines ‘BL 28’ and ‘BL 39’ and the control-check cultivar ‘Simeto’, which, despite their higher mean values of NLA, also expressed a DS of 3, being promising resources of resistance (McCartney et al., 2002; Chartrain et al., 2004). Finally, two commercial cultivars, ‘RGT Rumbadur’ and ‘RGT Voilur’, exhibited a DS score of 2, the lowest DS score in our study, which denotes an absence of pycnidia in their developed lesions. Although these cultivars presented 50-60% mean values of NLA, their type of resistance prevents subsequent infections via an inhibition of pycnidium development and, thus, pycnidiospores, which spread the infection via rain splash to other plants. Moreover, this type of resistance could prevent the sexual reproduction of Z. tritici by inhibiting pseudothecium formation, which represents the main source of primary inoculum via ascospores dispersed by wind, being an essential resource to control STB in the field (Suffert et al., 2011). Thus, commercial cultivars evaluated which presented resistance patterns could develop an advantage in fields with high STB incidence. Both, commercial cultivars and breeding lines can provide a valuable source of resistance for breeding programs.

The combination of both evaluation methods used in our study, i.e., qualitative (DS; McCartney et al., 2002) and quantitative (NLA and presence or absence of pycnidia; Suffert et al., 2013), permitted the classification of our studied accessions on the basis of their resistance patterns (Fig. 2). However, quantitative traits among different accessions or even among diverse Z. tritici isolates may be difficult to visually evaluate in a growth chamber or in the field (Karisto et al., 2018). In order to precisely assess these quantitative differences, image analysis emerged as a powerful tool in some recent studies (Stewart & McDonald, 2014; Stewart et al., 2016; Karisto et al., 2018). In our work, we used a camera to obtain images of infected leaves, which were later analyzed with specific image software (Fig. 1). Due to this analysis being more thorough in comparison with visual scoring, it enabled the accurate evaluation of various quantitative parameters of STB disease, thereby allowing the classification of accessions on the basis of their DS score. Our results showed statistical differences in quantitative parameters for accessions classified with different DS scores and, as we expected, accessions with a DS of 3 represented the lowest values, considering their partial resistance with slight pycnidium development. Thus, accessions classified with a DS of 3 showed a lower percentage of NLA covered with pycnidia, as well as a lower pycnidium count and a lower number of pycnidia per mm2 of NLA, than accessions with a DS of 4, which then showed lower values than accessions with a DS of 5 (Table 1). This increase in the values of quantitative parameters from accessions with a DS of 3 to those with a DS of 4 and 5 mainly corresponded to an increase in pycnidium count, considering that the pycnidium size among accessions did not present statistical differences. However, our results also showed statistical differences for the quantitative parameters studied amongst accessions classified with the same DS score, which highlights the value of image analysis as tool not only in the detection of quantitative differences amongst accessions classified with the same DS score, but also in the possible classification of accessions on the basis of only qualitative traits (Stewart & McDonald, 2014; Stewart et al., 2016; Karisto et al., 2018). In future studies, our evaluation of STB symptoms is expected to follow the method developed by Stewart & McDonald (2014), whereby scanned images of infected leaves are analyzed through a macro process, leading to an automatic analysis of quantitative STB disease parameters in several leaves. In addition, considering the lack of knowledge about the Z. tritici-durum wheat interaction (Somai-Jemmali et al., 2017a), studies of the fungal infection process and its associated plant defense mechanisms in our resistant accessions represent the next step to a better understanding of durum wheat resistance to Z. tritici.

The present study evaluated the incidence of the fungus Z. tritici, which is the causal agent of one of the most important diseases in wheat, in both commercial cultivars and breeding lines currently available in Spain. The limited sources of resistance found in our 48 studied accessions, evaluated visually and through image analysis, highlights the importance of the incorporation of these sources of resistance against STB into future durum wheat breeding programs. In addition, due to the scarcity of studies related to the resistance and susceptibility of durum wheat against STB, our work represents a novel contribution to the selection of sources of resistance amongst cultivars which currently have valuable agronomic traits in Spain, leading to advantages when facing STB.