European Journal of Plant Pathology

, Volume 137, Issue 1, pp 181–196

Histopathology of durable adult plant resistance to leaf rust in the Brazilian wheat variety Toropi

Authors

  • Caroline Wesp-Guterres
    • Programa de Pós-Graduação em FitotecniaUniversidade Federal do Rio Grande do Sul
    • Departamento de FitossanidadeUniversidade Federal do Rio Grande do Sul
  • Felipe André Sganzerla Graichen
    • Universidade Estadual de Mato Grosso do Sul
  • Márcia Soares Chaves
    • Embrapa—Centro Nacional de Pesquisa de Trigo
Article

DOI: 10.1007/s10658-013-0232-5

Cite this article as:
Wesp-Guterres, C., Martinelli, J.A., Graichen, F.A.S. et al. Eur J Plant Pathol (2013) 137: 181. doi:10.1007/s10658-013-0232-5

Abstract

Leaf rust, caused by the fungus Puccinia triticina is a major disease of wheat (Triticum aestivum) worldwide. This disease is prevalent in southern South America where the environmental conditions and high genetic variability of P. triticina favour epidemics. The primary means of controlling pathogenic P. triticina races has been through using wheat varieties containing race-specific resistance genes. The defence mechanisms involved in durable race non-specific resistance to P. triticina are probably distinct from those involved in non-durable race-specific resistance. We investigated the histological components of resistance to P. triticina present in three wheat genotypes: the race non-specific resistant Brazilian variety Toropi; the race-specific resistant line RL6010 Lr9; and the susceptible Brazilian variety BRS 194. Plants of these three genotypes were inoculated with P. triticina race MFP and tissue samples excised from flag leaves at various times after inoculation to assess the number of infective structures, frequency of cell death and the accumulation of autofluorescent cells and hydrogen peroxide. The genotypes showed different resistance mechanisms active at different times during the infection process. Our results for Toropi indicate that there was a reduction in the extent of formation of stomatal appressoria and all subsequent structures. During attempted penetration we also observed the production of autofluorescent compounds and late cell death, but not peroxide formation. This non-specific resistance to P. triticina involves both pre-haustorial and post-haustorial mechanisms which may be responsible for maintaining the low disease severity observed in this variety even under high inoculum pressure.

Keywords

Race non-specific resistance mechanismsPost-haustorial resistancePre-haustorial resistancePuccinia triticina

Introduction

The basidomycete fungus Puccinia triticina Erikss. & Henn. is the causative agent of wheat (Triticum aestivum L.) leaf rust, one of the principal wheat diseases occurring in practically all the areas where this cereal is cultivated (Huerta-Espino et al. 2011). Southern Brazil, Argentina, Paraguay and Uruguay form the Southern Cone Region of South America, a single P. triticina epidemic unit (Rajaram and Campos 1974) in which about 9 million hectares of wheat is planted each year (German et al. 2007). The occurrence of repeatedly severe leaf rust epidemics in Southern Cone Region are due to the presence of several factors conducive to the growth and dissemination of P. triticina, including favourable environmental conditions during extended periods of time, the high variability of this pathogen and high inoculum pressure. Such epidemics can result in yield-losses of up to 50 % if fungicides are not used to control leaf rust (German et al. 2007; Huerta-Espino et al. 2011).

Due to the reasons cited above, in such environments the most appropriate approach to control leaf rust and to stabilize the P. triticina population is the breeding and commercial use of cultivars with adequate levels of race non-specific adult plant resistance (Huerta-Espino et al. 2011). Durable resistance is even more important in developing countries, where it is difficult to substitute varieties as rapidly as the pathogen overcomes specific resistance (Trethowan et al. 2005). In the Southern Cone Region adult plant resistance to wheat leaf rust has been prioritized in countries where national wheat breeding programs are conducted, such as Brazil, Argentina, and Uruguay (German et al. 2007). This type of resistance was first proposed for potato blight (Niederhauser et al. 1954) and then applied to wheat leaf rust (Caldwell 1968), for which it was later described as a durable form of resistance (Johnson 1984). The genes conferring race non-specific resistance to P. triticina have partial and additive effects, which, while not preventing infection because the host essentially remains susceptible, slow the development of the disease (Marone et al. 2009), commonly known as “slow rusting”. However, when minor adult plant resistance (APR) slow rusting genes are present alone they do not confer adequate resistance, especially under high disease pressure, but combinations of four or five such genes usually result in near-immunity or at least a high level of resistance (Singh et al. 2011).

In plant interactions involving pathogenic rusts, resistance is generally due to the triggering of plant defence mechanisms before or after the haustoria formation (Heath 1981). In spite of recent progresses regarding host-pathogen interactions in diverse pathosystems it is believed that several unexplored resistance mechanisms still exist. All resistance is transitory when viewed in an evolutionary context (Ribeiro do Vale et al. 2001), which means that there is a continual demand for novel defence mechanisms to be incorporated into commercial plant genotypes. The observation and investigation of different types of plant reactions to pathogenic rusts can result in the discovery of resistance mechanisms acting at different stages of infection, which, when combined, can provide plants with multiple barriers to infection (Rubiales and Niks 2000). In addition, the detailed study of genetic and histological resistance mechanisms can generate valuable information regarding the most promising mechanisms for further biochemical and physiological investigation (Heath 1981).

The race non-specific resistant (partially resistant) wheat variety Toropi, bred in Brazil, was released in 1965 and extensively cultivated for 15 years, and has shown durable APR to leaf rust for more than 40 years (Rosa et al. 2011). Although Toropi is derived from the Frontana variety (Kohli and Skovmand 1997), which contains the slow rusting gene Lr34, there is no evidence that this gene is present in this variety (Barcellos et al. 2000). The Lr34 gene is known to confer durable APR, not only to leaf rust but also to wheat stripe rust caused by Puccinia striiformis f. sp. tritici (Singh 1992), powdery mildew caused by Blumeria graminis (Spielmeyer et al. 2005) and, depending on the genetic background of the wheat variety containing this gene (i.e. the Thatcher variety and its derivatives), also to wheat stem rust caused by Puccinia graminis f. sp. tritici (Dyck 1987). In Toropi, two recessive genes have been reported to be associated with APR and, while not yet having been fully described, have received the temporary designations Trp-1 and Trp-2 (Barcellos et al. 2000) and have been reported to be located on chromosomes 1A and 4D (Brammer et al. 1998) where no other T. aestivum APR genes are known to occur (McIntosh et al. 1995). Toropi seems to have one seedling resistance gene, dominant race-specific APR and two complementary race non-specific APR genes, with the latter two genes having conferred effective APR at locations in Canada and New Zealand (Rosa et al. 2011). The Toropi variety has also been shown to present other important agronomic characteristics besides durable APR, including increased phosphorous absorption, translocation and distribution (i.e. low-phosphorous tolerance) as well as tolerance to aluminum toxicity (Espíndula et al. 2009) and resistance to Fusarium head blight (Kohli 1989). However, the use of Toropi in genetic improvement programs has been limited.

Knowledge of the mechanisms responsible for the P. triticina resistance shown by Toropi could produce novel additional possibilities for the wider use of this genotype in wheat improvement programs. Furthermore, such knowledge may be useful for the identification of the genes involved in P. triticina resistance, thus aiding the development of efficient molecular markers for the assisted selection of commercial genotypes carrying more durable resistance traits.

The principal aim of our study was to assess the mechanisms involved in Toropi APR to P. triticina leaf rust by investigating the Toropi micro-phenotypes expressed during P. triticina infection. To this end, we evaluated the differentiation of fungal structures, the occurrence of plant cell death, and the presence of fluorescent compounds, autofluorescence and hydrogen peroxide in leaf tissues. Assessments were also made on the race-specific resistant line RL6010 Lr9 and the susceptible variety BRS 194.

Material and methods

Wheat genotypes, fungal race and plant inoculation

The wheat genotypes used were: the race non-specific resistant Brazilian variety Toropi; line RL6010 Lr9 (race-specific resistant control) a near isogenic line (Thatcher background) containing the Lr9 resistance gene; and the Brazilian variety BRS 194 (susceptible control). The experiments used Puccinia triticina Erikss. & Henn. race MFP as the pathogen (Long and Kolmer 1989). Totally expanded flag leaves were inoculated with a P. triticina race MFP spore suspension [distilled water, containing: spores, 105 − 106 ml−1; 0.01 % (v/v) Tween 20®] and incubated for 18 h at 20 °C, in the dark, in a moist chamber and then transferred to an environmental cabinet maintained at 20 °C and 100 % relative humidity with a 12-h photoperiod until symptoms appeared at about 15 days after inoculation.

All chemicals used in this study were of at least analytical quality and were purchased from the Sigma Chemical Corporation unless otherwise stated.

Sample preparation, evaluation of fungal infection, colonization and plant cell death

Experiments were set up in which at least triplicate 1.5 cm samples were excised from the middle of flag leaves from each genotype at the following hours after inoculation (hai): 48, 120 and 240 hai. All samples were fixed for 24 h in 0.15 % (w/v) trichloroacetic acid in ethanol:dichloromethane (3:1 v:v), stained by boiling for 5 min in 0.05 % (w/v) trypan blue in lactophenol:ethanol (1:2 v:v) and cleared in chloral hydrate:water (5:2 w:v), dehydrated in an ethanol sequence (80 % (v:v) for 30 min; 90 % (v:v) for 30 min; 100 % for 2 x 30 min), stained for 5 min with a saturated solution of picric acid in methyl salicylate (to stain papillae) and cleared for 15 min in methyl salicylate. This procedure was adopted since both general histological features and papillae could be visualized in the same slide.

The stained samples were placed on glass slides, mounted (adaxial side up) in methyl salicylate, covered with a cover slip, sealed with nail varnish and examined under bright-field and phase-contrast microscopy using an Olympus BX-41 optical microscope. We examined 100 to 300 readings for each replicate and calculated the following: % germinated spores = (germinated spores/observed spores) × 100; % stomatal appressoria = (stomatal appressoria/germinated spores) × 100; % sub-stomatal vesicles = (sub-stomatal vesicles/stomatal appressoria) x 100; % infective hyphae = (infective hyphae/sub-stomatal vesicles) × 100 %; % haustoria mother-cells = (haustoria mother-cells/ infective hyphae) × 100 and % haustoria = (haustoria/haustoria mother-cells × 100). Flag leaf samples collected 120 hai after inoculation were also assessed for the presence of P. triticina sporogenic tissue and the results expressed as % sporogenic tissue = (fungal colonies with sporogenic tissue/number of fungal colonies) × 100. Samples collected at 240 hai were assessed qualitatively, not quantitatively, because the intense colonization of the plant tissue prevented counting of the individual fungal structures. At the time points examined we also quantified plant cell death using trypan blue. Attempted colonization was considered to have occurred when stomatal appressoria were present, while successful penetration was shown by sub-stomatal vesicles (Jagger et al. 2011). The percentage of colonization attempts associated with plant cell death was calculated by dividing the number of colonization attempts resulting in the death of plant cells by the total number of colonization attempts.

Host formation of autofluorescent compounds, hydrogen peroxide and papillae

Plant phenolic autofluorescent compounds associated with P. triticina penetration attempts were evaluated in 50 infection units, consisting of at least a spore and a stomatal appressoria. We also evaluated the presence of autofluorescent compounds associated with sub-stomatal vesicles 48 hai and 120 hai after inoculation, this was achieved by fixing plant tissue samples for 24 hai in 3:1 (v:v) ethanol:dichloromethane containing 0.15 % (w/v) trichloroacetic acid and then storing the samples in aqueous glycerol (50 % (w/v) glycerol) until analysis. To detect autofluorescence, the fixed samples were cleared in 5:2 (w:v) chloral hydrate:water, mounted in 50 % glycerol and examined using blue-light epifluorescence microscopy using an Olympus BX 41 microscope fitted with a U-MWB2 excitation filter set consisting of a 460 nm to 490 nm dichroic beamsplitter and a 500 nm BA520 barrier filter. The percentage of infection units associated with the production of autofluorescent compounds was calculated in two different ways as: (i) the number of infection units associated with the production of autofluorescent compounds/number of units infected with at least a spore and a stomatal appressoria × 100; and (ii) the number of infective units associated with the production of autofluorescent compounds/number of penetrating infective units with the formation of sub-stomatal vesicle × 100. In both cases, we used three replicates and 100 to 300 readings for each replicate.

Hydrogen peroxide formation was evaluated at 24 hai, 48 hai, 72 hai, 96 hai and 120 hai. Three samples of each genotype were dipped into a solution of 0.1 % (w/v) 3,3′-diaminobenzidine-4 hydrochloride (DAB) in acidified water (pH 3.8) (Thordal-Christensen et al. 1997) and excess solution removed by vacuum filtration, the samples being placed in the light for 30 min and then in the dark for 5 h, after which they were fixed, cleared and mounted in 50 % (v/v) glycerol and evaluated under bright-field microscopy. The percentage of penetration attempts associated with the production of reactive oxygen species (ROS) was calculated by dividing the number of ROS-associated penetration attempts by the total number of penetration attempts, 50 evaluations being made for each sample.

Papillae were detected using the same methodology as that used to observe fungal structures (q.v. previous section). When present, phase contrast microscopy showed papillae as highly refractive appositions in plant cell walls in contact with fungal hyphae, mediated by the reaction with picric acid (Bender et al. 2000).

Data analysis

Data of germinated spores, stomatal appressoria, sub-stomatal vesicles, infective hyphae, haustoria mother-cell and haustoria during P. triticina infection on wheat genotypes Toropi, BRS 194 and line RL6010 Lr9 at 48 and 120 hai were submitted to ANOVA in a factorial arrange of 3 × 2. It was tested the single effects of genotype and time as well as the interaction among themselves. The discrimination of the variation amongst treatments was done by the Fisher’s Least Significant Differences Test (LSD) at 5 % of probability level. Chi-squared (χ2) was used to analyze the following percentage data: colony formation, plant cell death, autofluorescent cells and presence of hydrogen peroxide. The Pearson correlation coefficient (ρ) was used to test associations between the following variables: formation of colonies and sporogenic tissue; percentage of penetration attempts associated with cell death; production of autofluorescent compounds and ROS at 48 hai and 120 hai. Associations were considered significant when ρ-values were significant by the t-test (p = 0.05). The SAS statistical package v 8.0 was used to analyze the results (SAS Institute Inc. Cary, NC, USA).

Results

Evaluation of the P. triticina infection and colonization processes

The ANOVA revealed that, except for haustoria, there was a significant interaction between the three genotypes (Toropi, BRS 194 and line RL6010 Lr9) and the two time points (48 hai and 120 hai) for all the variables measured (percentage spore germination, appressorium formation, substomatal vesicle formation, differentiation of infective hyphae, haustorium mother cell formation). The simple effects of the two factors (genotype and time) were equally significant. The discrimination of variation between treatments was therefore performed through comparisons between the three genotypes for each time point examined (hai) and between the two time points (hai) for each genotype. Regarding haustoria formation there was significant difference only between genotypes (Table 1).
Table 1

Percentages of spore germination, stomatal apressoria, substomatal vesicles, infective hyphae, haustorium mother cells and haustoria during Puccinia triticina infection on wheat genotypes Toropi (partially-resistant), RL6010 Lr9 (race-specific resistant control) and BRS194 (susceptible control) at 48 hai and 120 hai

  

Genotype

 

Infective event

Time (hai)

Toropi

BRS 194

RL6010 Lr9

CV (%)

Spore Germination (%)

48

64.57

a

A

65.78

a

A

47.11

b

A

8.63

120

57.12

b

A

80.22

a

B

58.29

b

B

Stomatal Appressoria (%)

48

7.58

b

A

14.29

a

A

17.01

a

A

22.11

120

7.74

b

A

34.56

a

B

32.96

a

B

Substomatal vesicles (%)

48

20.89

b

A

55.63

a

A

23.57

b

A

19.38

120

33.16

b

A

35.64

b

B

80.81

a

B

Infective hyphae (%)

48

11.11

b

A

26.50

ab

A

49.44

a

A

33.29

120

0.00

b

A

83.01

a

B

88.95

a

B

Haustorium mother cells (%)

48

0.00

c

A

61.11

b

A

100.00

a

A

10.74

120

0.00

b

A

96.50

a

B

93.33

a

A

Haustoria (%)

 

0.00

b

 

99.33

a

 

93.05

a

 

11.03

Means followed by the same letter are not significantly different at the .05 level by the Fisher’s Least Significant Differences Test

Lowercase letters indicate comparisons between genotypes for each time (hai)

Uppercase letters indicate comparisons between times (hai) for each genotype

The percentage of spore germination was significantly different for genotypes (P > F = 0.0001) and time points (P > F = 0.0339) as well as their interaction (P > F = 0.0086). At 48 hai percentage germination was significantly lower for line RL6010 Lr9 than for the other two genotypes, which were not significantly different from each other. At 120 hai spore germination was higher for the susceptible cultivar BRS 194 than the other two genotypes, with no observed difference between Toropi and line RL6010 Lr9. Toropi showed no significant increase in percentage spore germination over time, however BRS 194 and line RL6010 Lr9 showed a significantly larger proportion of germinated spores at 120 hai than at 48 hai (Table 1).

The percentage of stomatal appressoria was significantly different between genotypes (P > F = 0.0001) and time points (P > F = 0.0001) and the interaction between these factors was also significant (P > F = 0.0034). The percentage of appressoria differentiation in each genotype followed a similar pattern for both time points, with Toropi showing the lowest values and BRS 194 and RL6010 Lr9 higher values. A comparison of the percentage appressoria formed over time for each of the three genotypes showed similar results to those observed for spore germination, Toropi, showing no significant increase in regard to the two time periods while the genotypes BRS 194 and RL6010 Lr9 showed a significant increase between 48 hai and 120 hai (Table 1).

The percentage of sub-stomatal vesicles formed was significantly different between genotypes (P > F = 0.0004), time points (P > F = 0.0010) and their interaction (P > F = 0.0001). At 48 hai BRS 194 showed a higher percentage of sub-stomatal vesicles than Toropi and RL6010 Lr9, which did not differ. At 120 hai the qualitatively resistant genotype RL6010 Lr9 produced significantly more sub-stomatal vesicles than Toropi and BRS 194. For RL6010 Lr9 there was a marked increase in the percentage of differentiated sub-stomatal vesicles over time, whereas the opposite was observed for BRS 194. Toropi showed no significant difference over time in regard to these vesicles (Table 1).

The percentage of infective hyphae formed was significantly different between genotypes (P > F = 0.0001), time points (P > F = 0.0013) and their interaction (P > F = 0.0041). At 48 hai RL6010 Lr9 showed the highest percentage of infective hyphae, while Toropi showed the lowest. At 120 hai BRS 194 and RL6010 Lr9 showed statistically similar percentages of infective hyphae, while Toropi showed none. A comparison of the percentage of infective hyphae over time for each of the three genotypes showed similar results to those observed for spore germination and appressoria formation. At 120 hai the susceptible BRS 194 and RL6010 Lr9 showed significantly more differentiation of infective hyphae than at 48 hai, while for Toropi there were no significant differences in infective hyphae formation over time (Table 1).

The percentage of differentiated haustorium mother cells (HMC) was significantly different between genotypes (P > F = 0.0001), time points (P > F = 0.0072) and their interaction (P > F = 0.0002). At 48 hai RL6010 Lr9 showed the highest percentage of HMC. At 120 hai the highest percentages of HMC occurred on BRS 194 and RL6010 Lr9, which were both statistically equal. No HMC were observed on Toropi. For each of the three genotypes, HMC formation of over time was similar to the results for spore germination, appressoria formation and infective hypha. At 120 hai there was more infective hypha differentiation on BRS 194 and RL6010 Lr9 than at 48 hai. On Toropi, however, the difference in infective hyphae formation over time was not significant, since no haustorium mother cells were observed at any of the examined times (Table 1).

The percentage of haustoria formed was significantly different only between genotypes (P > F = 0.0001). Time points (P > F = 0.1585) and the interaction with genotypes (P > F = 0.2049) were not significantly different. BRS 194 and RL6010 Lr9 showed statistically similar percentages of haustoria between themselves (99.33 % and 93.05 %, respectively) and higher than Toropi, which showed none.

The sampling performed at 240 hai in Toropi included segments of tissue with pustules and others without pustules. It could be seen that in the Toropi samples showing pustules, differentiation of fungal structures was similar to that observed in the susceptible BRS 194, with the Toropi samples showing intense growth of intercellular hyphae and many haustoria. However, in samples where no pustules were observed, some attempts at infection were associated with cell death. Although some Toropi samples showed the presence of hyphae without the formation of haustoria, most infection attempts seemed to be stopped in the early stages of the infectious process, indicating a pre-haustorial resistance mechanism in this genotype.

At 240 hai only 16 % of haustoria were associated with established P. triticina colonies in Toropi, while for BRS 194 this value was 88 % and there was intense growth of intercellular hyphae. Sporogenic tissue was identified by thickening of P. triticina mycelia and the presence of sporogenic cells containing immature or mature uredospores. These structures were formed within the host wheat tissue before rupture of the epidermis and release of the uredospores. The differentiation of sporogenic tissue was observed at 240 hai after inoculation. In Toropi just 9 % colonies showed differentiated sporogenic tissue, as compared with no colony or sporogenic tissue formation in the resistant RL6010 Lr9 and 32 % in the susceptible BRS 194.

Cell death, formation of autofluorescent compounds, hydrogen peroxide and papillae

Cells dying in response to inoculation with P. triticina stained intensely with trypan blue, unlike live cells, which were only slightly blue whether or not they were in contact with the fungus. No dead cells were seen in BRS 194 before 240 hai after inoculation (Fig. 1), the necrotic cells present at 240 hai were associated with intense tissue colonization and had probably died from nutrient exhaustion due to the progress of the infection and not because of any defence mechanism. In Toropi, 37 % of infection attempts were associated with cell death 120 hai (Fig. 1). Unlike the cell death observed in RL6010 Lr9, which was apparent 48 hai for all cells in which haustoria were produced (P = <0.0001) (Fig. 1), Toropi cell death seemed to take place at later stages of infection. In spite of the occurrence of late cell death in Toropi, a large proportion of the observed infection attempts appeared to be stopped by a pre-haustorial resistance mechanism during the initial phases of infection.
https://static-content.springer.com/image/art%3A10.1007%2Fs10658-013-0232-5/MediaObjects/10658_2013_232_Fig1_HTML.gif
Fig. 1

Percentage of infective units (germinated Puccinia triticina spores plus appressoria) associated with resistance response events (presence of cell death, papilla, reactive oxygen species and autofluorescent compounds) on three wheat genotypes at 48 h and 120 h after inoculation with P. triticina. The genotypes used were BRS 194 (susceptible), line RL6010 Lr9 (race-specific resistant) and the race non-specific resistant variety Toropi (partially-resistant). Bars indicate the standard error and hai = hours after inoculation

https://static-content.springer.com/image/art%3A10.1007%2Fs10658-013-0232-5/MediaObjects/10658_2013_232_Fig2_HTML.gif
Fig. 2

Responses to Puccinia triticina infection in wheat varieties Toropi (partially-resistant), BRS 194 (susceptible) and line RL6010 Lr9 (race-specific resistant). a, b and c Leaf flag symptoms 15 days after inoculation. d, e and f Presence of autofluorescent cells. Arrows indicate structures and reactions. g, h and i Peroxide formation. Arrows indicate spores. E = stomata, AP = stomatal appressoria, SSV = sub-stomatal vesicle. Magnification shown between parentheses. hai = hours after inoculation. Bars indicate size of the structures

https://static-content.springer.com/image/art%3A10.1007%2Fs10658-013-0232-5/MediaObjects/10658_2013_232_Fig3_HTML.gif
Fig. 3

a Theoretical model based on the assumption that 103 rust spores were deposited on the leaves of susceptible and partially resistant plants. The figure shows a comparison between the susceptible BRS 194 and the partially resistant Toropi. At each stage of infection the difference between the genotypes increases, with Toropi always showing lower established infection events than BRS 194. b Transverse section of a wheat leaf infected with Puccinia triticina. Arrows indicate the infection events and respective percentage reduction in Toropi in relation to BRS 194. Toropi showed an increased percentage reduction in infective events at each stage. GERM = Germination, AP = Appresorium, SSV = Sub-stomatal vesicle, IH = Infective hyphae, HMC = Haustorium mother cell, H = Haustorium. b The photomicrograph was edited to show all infection events at the same time

In BRS 194, at 36 hai autofluorescent cells were not observed while a relatively high proportion of infection attempts were associated with autofluorescent compounds in Toropi (27 %) and RL6010 Lr9 (41 %) (P = <0.0001). At 48 hai, where sub-stomatal vesicles occurred, indicating that penetration was successful, there was an increased percentage of autofluorescent cells associated with attempted colonization as compared to the percentage seen at 36 hai. The percentage of autofluorescent cells associated with attempted colonization 48 hai after inoculation was 89 % for Toropi and 100 % for RL6010 Lr9 (P < 0.0001). Furthermore, at 120 hai BRS 194 showed autofluorescence in about 10 % of the penetrated stomata, while for Toropi and RL6010 Lr9 100 % of cells associated with colonization showed very marked autofluorescence in which various cells presented evidence for the production of autofluorescent compounds (P < 0.0001) (Figs. 1 and 2d, e, f and 3).

The production of ROS varied temporal and quantitatively among genotypes (P < 0.0001) (Figs. 2g, h i, and 3). No peroxide production was detected in RL6010 Lr9 at 24 h after inoculation, with peak production occurring at 48 h and coinciding with haustoria formation, after which peroxide production reduced, occurring in about 7 % of penetration attempts. There was little peroxide production in BRS 194 at 120 h, whilst that which did occur was restricted to penetrated stomata (Fig. 2j). In Toropi there was no peroxide production at sites where infection attempts occurred (Fig. 2h).

There was a high correlation between cell autofluorescence in response to stomatal appressoria and sub-stomatal vesicles formation 48 hai after inoculation (ρ=0.99). In addition, there was a positive correlation between the presence of autofluorescent cells in response to sub-stomatal vesicle formation and cell death (ρ=0.69). At 120 hai there was a positive correlation between production of sporogenic tissue and the presence of colonies (ρ=0.85) and also between the presence of autofluorescent cells in response to stomatal appressoria formation and cell death (ρ=0.96). Although not strongly correlated, there was a certain degree of negative correlation between the presence of autofluorescent cells in response to stomatal appressoria formation (ρ= − 0.21), the presence of autofluorescent cells in response to sub-stomatal vesicle formation (ρ= − 0.32), ROS production (ρ= − 0.16) and the occurrence of cell death (ρ= − 0.25). Similar results were observed for the correlations involving the presence of sporogenic tissue and resistance response (Table 2). Papillae were not observed on cell walls of any of the three genotypes investigated.
Table 2

Pearson’s correlation coefficients between histological components of resistance to Puccinia triticina in wheat genotypesa at 120 h after inoculation

 

Autofluorescent compounds detected

Histological components of resistance

Presence of P. triticina colonies

Stomatal appressoria

Sub-stomatal vesicles

Production of reactive oxygen species

Plant cell death

Sporogenic tissue

0.85 (<0.001)b

0.12 (0.822)

0.04 (0.928)

−0.29 (0.354)

−0.16 (0.526)

P. triticina colonies

 

−0.21 (0.690)

−0.32 (0.538)

−0.16 (0.617)

−0.25 (0.319)

Autofluorescent compounds in stomatal appressoria

  

0.69 (0.127)

−0.93 (0.07)

0.96 (0.002)

Autofluorescent compounds in sub-stomatal vesicles

   

−0.33 (0.66)

0.61 (0.196)

Production of reactive oxygen species

    

−0.55 (0.06)

aThe genotypes used were the race non-specific resistant variety Toropi (partially-resistant), line RL6010 Lr9 (race-specific resistant control) and the variety BRS194 (susceptible control)

bValues between parentheses correspond to the probability that ρ = 0

Discussion

Race-specific resistance genes, which provide nearly complete resistance to a specific race of rust, are traditionally the most widely used genes in the development of wheat varieties resistant to leaf rust, although this type of resistance has low durability due to the rapid evolution of virulence in P. triticina races (Kliebenstein and Rowe 2009). However, in the search for more durable resistance, non race-specific resistance, which provides only partial resistance, is a promising alternative because it does not restrict the development of the pathogen and generally involves several genes producing small but incremental effects which result in more recalcitrant resistance because the pathogen must undergo many mutations for partial resistance to be eroded (Jagger et al. 2011; Huerta-Espino et al. 2011).

Our results regarding the infection events indicate that Toropi resistance to P. triticina occurs at the beginning of the plant-pathogen interaction process, starting with stomatal appressoria formation and becoming more effective as subsequent infection events occur.

Differences in temporal and spatial resistance responses were observed not only between the race non-specific resistant Toropi and the susceptible BRS 194, but also for the race-specific resistant RL6010 Lr9. Our data show that the main resistance mechanisms in Toropi are pre-haustorial and serve to reduce the formation of infective structures when compared to BRS 194 (Table 1), resulting in the lower infection efficiency observed in Toropi (Fig. 2b). Comparisons between 48 hai and 120 hai (Table 1) showed that, for Toropi, there was no progress of infection from the first to the second time point, while this was observed for the majority of the infection events in the other two genotypes. Except for sub-stomatal vesicles, RL6010 Lr9 and BRS 194 produced similar percentages of fungal structures up to haustoria differentiation (Table 1). This was as expected, since RL6010 Lr9 shows post-haustorial resistance and is thus as permissive as BRS 194 until haustoria formation.

As explained above, in our experiments we applied a suspension containing P. triticina spores (105−106 ml−1). We used the data produced to create a theoretical model based on the assumption that 103 spores were deposited on the leaves of susceptible and partially resistant plants. Figure 3a shows a comparison between the susceptible BRS 194 and the partially resistant Toropi using the actual data which we obtained in our study. The figure shows that, starting with a theoretical 1000 spores on both BRS 194 and Toropi, at each stage of infection (germination, appressoria formation, sub-stomatal vesicle formation etc.) the difference between the genotypes increased, with Toropi always showing lower established infection events than BRS 194. When this is transformed into percentage differences in respect to the susceptible BRS 194 it can be seen that the percentage reduction shown by Toropi increases at each stage (Fig. 3b). For example, at germination BRS 194 showed 800 germinated spores while Toropi showed 570, 29 % less than for BRS 194. Likewise, at haustoria formation BRS 194 showed 78 haustoria while Toropi showed just one, an approximately 99 % reduction. As infective hyphae grow they may produce various HMC, each of which can produce an haustorium. Since a single colony may be formed from different haustoria the number of colonies was considered to be 25 % the number of HMC. Differentiation between genotypes was accentuated after appressoria formation, when Toropi showed an 84 % reduction with regard to BRS 194. The same was true for sub-stomatal vesicles, infective hyphae and HMC with the reduction shown in Toropi being from 85 % to almost 99 % as compared to BRS 194. This model reproduces the effects observed in the field with Toropi over the past 50 years.

It has been shown that the a protein encoded by the barley Rpg1 gene, which has conferred durable resistance to many P. graminis f. sp. tritici races for over 60 years, is phosphorylated within 5 min after inoculation in response to avirulent, but not virulent, rust spores and that phosphorylation is required for disease resistance (Nirmala et al. 2010). The recognition and subsequent activation of genes involved in the durable resistance of Toropi may occur at very early stages, soon after the fungus has contact with the plant. This is supported by the hindrance in the evolution of the fungi structures since the beginning of the infection process, as it can be seen on Table 1, specially at appressoria formation, with a final effect being about 75 % reduction in relation to the other genotypes. The same has been observed in quantitative resistance of wheat to P. triticina in genotypes possessing the gens Lr12 e Lr13 (Bender et al. 2000) and P. striiformis, with the largest resistance effect being an approximately 50 % reduction in the formation of stomatal appressoria in the resistant genotypes as compared to the susceptible control. Reduction in rust appressoria formation was also reported in other plants such as Hordeum chilense (Rubiales and Niks 1992) or in sunflower (Prats et al. 2007). Subsequent structures have also been reported to have been reduced (Broers and López-Atilano 1996). At 48 hai, Toropi and RL6010 Lr9 presented about half the number of sub-stomatal vesicles as BRS 194. However, at 120 hai, Toropi and BRS 194 showed similar values, which were about half the value for RL6010 Lr9 (Table 1). Other workers have shown that, as compared with a susceptible genotype, the development of sub-stomatal vesicles can be retarded in resistant plants possessing the APR gene Yr18 (Elahinia 2008). In our study, this retardation seemed to occur in Toropi, reducing the number of infective hyphae subsequently formed, which were observed only up to 48 hai after inoculation in this APR genotype.

It is believed that the activation of plant defences, such as phytoalexins or pathogen-related proteins, can cause the disintegration of infective hypha during the advanced stages of infection. Similarly, the Lr34 and Lr46 genes have been associated with reduced intercellular hyphal growth but not with the hypersensitivity response or papilla formation (Rubiales and Niks 1995; Martínez et al. 2001). This phenomenon also occurs in wheat genotypes with durable resistance to P. striiformis, where a reduction in the number of sub-stomatal vesicles has been observed between 48 hai and 144 hai after infection (Broers and López-Atilano 1996). We observed only a few haustoria mother-cells, possibly because plant defence compounds may reduce differentiation of fungal structures and result in dysfunctional haustoria mother-cells leading to reduced haustorial differentiation and a lower incidence of infection (Sillero and Rubiales 2002). It is known that the quantitative mechanism of wheat resistance to P. striiformis acts during several phases of the development of this pathogen (Broers and López-Atilano 1996). These results indicate that principally pre-haustorial resistance mechanisms operate in Toropi, with a subsequent reduction in the formation of infective structures, and thus a lower efficiency of infection, as compared to the susceptible BRS 194.

The BRS 194 and RL6010 Lr9 genotypes produced a high percentage of infective hyphae showing differentiated haustorial mother-cells and haustoria, both of which not observed in Toropi (Table 1). However, no fungal colonies or sporogenic tissue occurred in RL6010 Lr9, which possesses race-specific, post-haustorial resistance associated with a hypersensitivity response (Montesanto et al. 2003). In our experiments, the race-specific resistant RL6010 Lr9, behaved mostly as the susceptible genotype BRS 194, except for germination and sub-stomatal vesicles, with all P. triticina infection attempts being associated with host cell death (Table 1). This type of reaction is characteristic of the hypersensitivity resistance conferred by major genes (Heath 1981). For Toropi, which exhibits partial resistance, we found cell death in 37 % of the infection attempts, indicating that different resistance mechanisms (both pre and post-haustorial) may be combined in this genotype. In Toropi, plant cell death occurred around 120 hai after infection, during the more advanced stages of the infectious process as compared to RL6010 Lr9. Other workers have also reported the occurrence of cell death in pathosystems involving rusts and plants with partial resistance, such as wheat and the stripe rust P. striiformis (Jagger et al. 2011), oat (Avena sativa) and the crown rust Puccinia coronata f. sp. avenae (Graichen et al. 2011), flax (Linum usitatissimum) and the flax rust Melampsora lini (Kowalska and Niks 1999).

Late cell death can be considered as an incomplete hypersensitivity response occurring in only some cells and which is expressed during the advanced phases of the infectious process (Kowalska and Niks 1999). In this type of non-hypersensitive cell death, ultra-structure alterations in the host plant cells begin about 3 days after inoculation and include the plasmolysis of cells in contact with infective hyphae or those forming haustoria (Jiang and Kang 2010). In our experiments, late cell death appears to have been induced by contact with infective hyphae or haustoria mother-cells without actual penetration. In Toropi, no haustoria were observed in dead cells, but it is not clear if this was due to the lack of formation of these structures or because haustoria were formed but their presence was masked by the general disorganization within the cell caused by cell death. In the susceptible BRS 194, at the advanced phase of the infection process at 120 hai after infection, a few necrotic cells were observed associated with intense colonization of the plant tissues, these cells probably having died from nutrient exhaustion due to the progress of the infection and were unrelated to resistance.

The accumulation of plant autofluorescent compounds in response to the penetration of plant cells by pathogenic fungi is known to occur in several interactions between plants and pathogens, with fluorescence being attributed to the presence of phenolic compounds that, when accumulated in mesophyll cells, can lead to lignification of plant cell walls, cell death and papillae formation (Rojas-Molina et al. 2007; Bozkurt et al. 2010). In the resistant genotypes assessed in our study we found that there was an increase in the production of plant autofluorescent compounds over time, with the autofluorescence we observed in Toropi cells resembling that described for autofluorescent cells associated with broad bean (Viciae faba) plants partially-resistant to infection with Uromyces viciae-fabae (Rojas-Molina et al. 2007). Similar autofluorescent compounds have been reported in oat genotypes partially-resistant to P. coronata and powdery mildew caused by Erysiphe graminis f. sp. avenae (Graichen et al. 2011; Parry and Carver 1986). In our study, it is possible that the cell death observed in Toropi was due to the accumulation of phenolics and not to a hypersensitivity reaction, this being supported by the fact that we found that the susceptible BRS 194 showed a much lower percentage (10 %) of autofluorescent cells than the resistant genotypes. Similar low percentages of autofluorescent cells have been reported in wheat genotypes susceptible to P. striiformis (Bozkurt et al. 2010). We found a high correlation between the presence of autofluorescent compounds and cell death, this also having been reported for specific resistance to P. striiformis in wheat (Bozkurt et al. 2010) and for resistance to P. coronata in oats (Graichen et al. 2011). Further investigations are needed to determine the type of phenolics formed in the genotypes studied by us.

Reactive oxygen species (ROS) constitute a further important plant defence mechanism in interactions between plants and pathogens, with ROS production being detectable only minutes after attack by virulent or avirulent strains of pathogenic fungi. The production of ROS occurs in two stages, the first occurring soon after contact between the fungus and the plant, irrespective of whether the plant is resistant or susceptible, while the second phase (explosive burst) only occurs in resistant plants and is initiated some hours after contact with the pathogen (Manickavelu et al. 2010). The ROS formed are highly reactive and toxic and, when present at high concentrations, not only induce a hypersensitivity reaction but also participate in signalling, the activation of defence genes and the initiation of changes in the plant cell wall leading to lignification, as well as having a direct antimicrobial action on the pathogen (Shetty et al. 2008).

In our experiments, the partially-resistant Toropi showed practically no peroxide formation at tentative infection sites. Absence of peroxide production has also been reported for the partially-resistant wheat variety Guardian during exposure to P. striiformis as well as for plants carrying the Lr34 gene when they were exposed to P. triticina (Melichar et al. 2008; Orczyka et al. 2010). The absence of peroxide accumulation may be the key determinant in cell death in Toropi, contrasting with the race-specific resistant RL6010 Lr9. For the race-specific resistant RL6010 Lr9 we found that peak ROS production, detected as peroxide (H2O2), occurred 48 hai after infection and coincided with haustoria formation, supporting work by other authors who reported that oxidative damage in this genotype was low and transient, reaching its peak 2 days after inoculation (Orczyka et al. 2010). Our results also support work showing that ROS production is decisive in the incompatibility response of the isogenic wheat line Lr10 to P. triticina (Manickavelu et al. 2010). In the susceptible BRS 194 we found that peroxide production was low and associated only with stomatal penetration occurring at 120 hai post-infection, with this late peroxide accumulation probably not occurring rapidly enough to impede infection. Other workers have also described the build-up of peroxide in the epidermal cells or stomata of susceptible plants 4 and 5 days after inoculation (Orczyka et al. 2010; Wang et al. 2010). Our results also show that peroxide production was negatively correlated with the formation of P. triticina colonies and sporogenic tissue in all the genotypes tested. Furthermore, the resistance of Toropi was not related to papillae formation, being this finding in accordance with what has been reported for wheat genotypes containing the APR gene Lr34 (Rubiales and Niks 1995).

These results demonstrate that different resistance mechanisms are involved in different types of resistance, as shown by the fact that in Toropi pre-haustorial mechanisms were clearly involved in the partial resistance process (although post-haustorial mechanisms may also have occurred) while resistance in RL6010 Lr9 was post-haustorial with the formation of autofluorescent compounds and peroxide production leading to the death of cells containing haustoria. A combination of these mechanisms resulted not only in the sequential reduction of the formation of fungal structures but also in the late cell death which was responsible for the low fungal infection efficiency observed in Toropi as compared to the susceptible BRS 194. However, some pustules did occur on some isolated tissue samples taken from Toropi flag leaves, presumably at sites where the defence mechanisms were not adequately expressed in terms of intensity or time of initiation.

Similarly, in einkorn wheat (Triticum monococcum L.), an ancient wheat which is now a rarely planted relic crop, resistance to P. triticina has been found to involve not only pre-haustorial papillae formation but also post-haustorial hypersensitivity responses and cell death if papillae formation is flawed and haustoria are formed, with some pustules being formed when both these mechanisms fail (Anker and Niks 2001). The wheat variety Kariega, which shows APR to P. striiformis stripe rust, also presents a combination of pre-haustorial resistance (lignification) and post haustorial resistance (non-hypersensitive cell death), as described by Moldenhauer et al. (2008). The involvement of different resistance mechanisms, active from spore deposition to haustoria formation, has also been described for the Lathyrus sativus/Uromyces pisi pathosystem (Vaz Patto and Rubiales 2009).

In Toropi, the main effect of the sequential restriction of the formation of infective structures is the reduced infection efficiency of Puccinia triticina, which is about 20 times lower in Toropi than in the susceptible BRS 194 (Fig. 3). This lower infection efficiency leads to a reduction in the number of pustules per plant (Fig. 2b) and hence fungal spore production, the epidemiological consequence of which is a reduced infection rate leading to a delay in the start of an epidemic or a weaker epidemic overall. The stability of resistance in Toropi to different races of P. triticina over time, coupled with the low severity of P. triticina infection observed in the field and the presence of different resistance mechanisms at a cellular level constitute strong evidence that this genotype really possesses a slow rusting mechanism.

Race-specific resistance is usually overcome within 5 years, especially as regards wheat rusts (Singh and Huerta-Spino 2001). Since its release in 1965, the Toropi variety has proved to have a high level of partial resistance to P. triticina, with slow rusting reaching maximum final severity of 5 % of the flag leaves.

One of the major challenges faced by plant breeders is obtaining wheat varieties with not only durable resistance to rusts but also desirable agronomic traits. However, in the last decade, soon after the diversity and genetic basis of slow rusting resistance became clearer, production started of high-yielding wheat lines showing near-immune levels of resistance, this being achieved by combining four or five additive minor resistance genes for both leaf rust and stripe rust (Singh et al. 2000). It is to be expected that genotypes combining different resistance mechanisms, preferentially of the pre-haustorial type, will potentially present more durable resistance to P. triticina than genotypes with resistance based only on hypersensitivity. Since such mechanisms act during different phases of the infectious process they offer multiple barriers which are not easily overcome by a simple mutation in the pathogen (Rubiales and Niks 2000; Kliebenstein and Rowe 2009), resulting in greater robustness of the plant defence response (Lipka et al. 2005). Our work corroborates these statements, and the techniques described by us may be useful tools for selecting genotypes likely to show durable resistance.

In Toropi, both partial resistance to P. triticina leaf rust and aluminum tolerance are located on chromosome 4D (Brammer et al. 1998; Raman et al. 2005) and are multigenic traits involving distinct mechanisms (Ryan et al. 2009). Resistance to non-adapted rust pathogens is polygenically inherited and is an active response which involves salicylic acid signalling and the production of reactive oxygen species that are often also common to host resistance responses (Ayliffe et al. 2011). Although our knowledge of the Toropi signaling mechanisms active in response to such stresses is still rather rudimentary, the overlap between abiotic and biotic stress responses can be considered as not only evidence of the plasticity of plant responses to environmental challenges but also the diversity of plant secondary metabolism (Glombitza et al. 2004). Recent work has revealed several types of molecules, including transcription factors and kinases, which are promising candidates for involvement in cross-talk between stress signalling pathways. Emerging evidence suggests that not only ROS signalling pathways but hormone signalling pathways regulated by abscisic acid, salicylic acid, jasmonic acid and ethylene play key rolls in cross-talk between abiotic and biotic stress signalling (Fujita et al. 2006).

The plasticity of plant responses has recently been demonstrated by Krattinger et al. (2009), who isolated the Lr34 gene and predicted that it encodes a pleiotropic drug resistance (PDR)-like ATP-binding cassette (ABC) transporter, with the same gene controlling multi-pathogen resistance to leaf rust (Lr34), stripe rust (Yr18) and powdery mildew (Pm38). We strongly believe that the detailed characterization of the mechanisms and genes involved in Toropi resistance to abiotic stresses such as aluminum toxicity and low phosphorous availability and biotic stresses such as leaf rust and Fusarium head blight will show that Toropi can be an important resource for worldwide wheat improvement programs. Combinations of Toropi APR resistance with the well characterized Lr34 gene (Krattinger et al. 2009; Dakouri et al. 2010) should also be possible, and could result in not only an increase in the durability of leaf rust resistance but also enhanced resistance to other wheat diseases such as stem rust, stripe rust and powdery mildew. New spatio-temporal studies on the mechanisms of durable resistance at both the biochemical and molecular levels in well characterized slow rusting genotypes like Toropi will contribute to the process of fixing high levels of partial resistance in agronomically superior genotypes.

Investigation of the major components governing resistance in plant pathogen interactions at the histological level using modern high-resolution light and electron microscopy may also permit the establishment of the connection between genetic control and the phenotypic expression of resistance. The parallel characterization of the biochemical and physiological metabolic processes involved in plant defence reactions and the determination of the timing of such reactions may also provide useful data for studies on the identification and characterization of differentially expressed genes in susceptible and partially resistant plant genotypes. Such approaches may allow optimization of the identification of genetic markers associated with important quantitative loci, which may produce data that could help in the identification of genetic sequences which are, de facto, involved in plant defence reactions.

Acknowledgments

The authors thank the Brazilian Federal Agency for the Coordination of the Improvement of Higher Education Personnel (Coordenação de Aperfeiçoamento de Pessoal de Nível Superior, CAPES) for a scholarship to the first author and the Brazilian National Council for Scientific Research and Development (Conselho Nacional de Desenvolvimento Científico e Tecnológico, CNPq) and the Foundation for Research Support of Rio Grande do Sul State (Fundação de Amparo à Pesquisa do Estado do Rio Grande do Sul, FAPERGS) for financial support.

Copyright information

© KNPV 2013