Present address: Syngenta Seeds Ltd, New Farm Crops, Whittlesford, Cambridge, CB2 4QT, UK.
Contributions of disease resistance and escape to the control of septoria tritici blotch of wheat
Version of Record online: 28 JUL 2009
© 2009 John Innes Centre, UK. Journal compilation © 2009 BSPP
Volume 58, Issue 5, pages 910–922, October 2009
How to Cite
Arraiano, L. S., Balaam, N., Fenwick, P. M., Chapman, C., Feuerhelm, D., Howell, P., Smith, S. J., Widdowson, J. P. and Brown, J. K. M. (2009), Contributions of disease resistance and escape to the control of septoria tritici blotch of wheat. Plant Pathology, 58: 910–922. doi: 10.1111/j.1365-3059.2009.02118.x
- Issue online: 8 SEP 2009
- Version of Record online: 28 JUL 2009
- disease escape;
- isolate-specific resistance;
- Mycosphaerella graminicola;
- septoria tritici blotch;
- Triticum aestivum
The contributions of disease escape and disease resistance to the responses of wheat to septoria tritici leaf blotch (STB) were analysed in a set of 226 lines, including modern cultivars, breeding lines and their progenitors dating back to the origin of scientific wheat breeding. Field trials were located in the important wheat-growing region of eastern England and were subject to natural infection by Mycosphaerella graminicola. STB scores were related to disease-escape traits, notably height, leaf spacing, leaf morphology and heading date, and to the presence of known Stb resistance genes and isolate-specific resistances. The Stb6 resistance gene was associated with a reduction of 19% in the level of STB in the complete set of 226 lines and with a 33% reduction in a subset of 139 lines of semidwarf stature. Greater plant height was strongly associated with reduced STB in the full set of lines, but only weakly in the semidwarf lines. Shorter leaf length was also associated with reduced STB, but, in contrast to earlier reports, lines with more prostrate leaves had more STB on average, probably because they tended to have longer leaves. Several lines, notably cvs Pastiche and Exsept, had low mean levels of STB which could not be explained by either escape traits or specific resistance genes, implying that they have unknown genes for partial resistance to STB.
For the last three decades Mycosphaerella graminicola (anamorph Septoria tritici), the causal agent of septoria tritici blotch (STB), has been the most important foliar disease of winter wheat in Western Europe and many other wheat-growing areas worldwide (Goodwin et al., 2003). The lack of highly resistant cultivars means that the only fully effective method of controlling this disease is by programmed application of fungicides. However, the emergence of widespread resistance to strobilurin (Fraaije et al., 2005) and, more recently, triazole fungicides (Brunner et al., 2008) has made selection for resistance to STB a high priority for wheat breeders.
Thirteen major genes for resistance to STB, Stb1 to Stb12 and Stb15, have been identified and mapped (Arraiano et al., 2001, 2007; Brading et al., 2002; Adhikari et al., 2003, 2004a,b,c; McCartney et al., 2003; Chartrain et al., 2005a,b, 2009). Each of these genes is effective against one or a few known isolates of M. graminicola. A gene-for-gene relationship was demonstrated for Stb6 (Brading et al., 2002), which is widespread in sources of resistance to STB worldwide (Chartrain et al., 2005c). Stb6 and Stb15 are the genes most commonly present in current European wheat germplasm (Arraiano & Brown, 2006). Several other isolate-specific resistances, which have been identified but not yet analysed genetically (Chartrain et al., 2004a), are present in European cultivars (Arraiano & Brown, 2006).
Stb1, which originated from wheat cv. Bulgaria 88 (Adhikari et al., 2004a) is, to date, the only known STB-specific resistance gene to be considered durable in the sense of Johnson (1984). This gene has been effective for over 25 years in widely grown cultivars across the STB-prone region of Indiana and adjacent states in the USA (Adhikari et al., 2004a). Stb4, originally from spring wheat cv. Tadinia, was effective in California for more than 15 years, but has recently become ineffective (Adhikari et al., 2004a). At present it is not known what the contribution of other STB-specific resistance genes to resistance in field conditions might be.
Quantifying resistance of cultivars to STB under standard field conditions is particularly difficult as it needs to be distinguished from disease escape. Resistance implies that a relatively low amount of disease is formed on a leaf by a given amount of inoculum. Escape implies that the amount of inoculum reaching the upper leaves is reduced or its arrival on those leaves is delayed. In general, taller plants and plants in which leaves are more widely spaced tend to have less disease because the vertical spread of spores up the plant is reduced (Bahat et al., 1980; Eyal, 1981; Danon et al., 1982; Baltazar et al., 1990; Jlibene et al., 1992). Furthermore, at any particular time of scoring, flag leaves which have been expanded for a greater length of time tend to have larger amounts of disease (Arama et al., 1999; Arraiano et al., 2006). This causes STB scores on flag leaves to be negatively correlated with plant height, distance between leaves and date of heading (van Beuningen & Kohli, 1990). It has also been proposed that leaf insertion angle or leaf prostrateness affects lateral movement of spores within the crop, such that more erect leaves transmit spores to other leaves at the same height from the ground for a greater period of time (Lovell et al., 1997, 2004).
This paper estimates the relative contributions of disease resistance and escape to control of STB in cultivars and breeding lines relevant to UK wheat breeding germplasm in field conditions. STB and several disease-escape traits were scored in a series of field trials so that, by adjusting STB scores on flag leaves for the effects of escape traits, an estimate of the foliar susceptibility of flag leaves was obtained. In addition, STB scores were related to the presence of Stb resistance genes and previously identified isolate-specific resistances. Finally, cultivars were identified as having partial resistance to STB that could not be accounted for by disease-escape traits or by the presence of particular Stb genes or isolate-specific resistances.
Materials and methods
Two-hundred and thirty-six wheat lines were tested in field trials (Arraiano & Brown, 2006). They included cultivars on the NIAB Recommended Lists of wheat cultivars in the late 1990s (NIAB, 1997–2000) and as many of their progenitors as could be obtained as far back as the start of scientific wheat breeding in the 1860s (Bonjean & Angus, 2001). Of these lines, 10 were very susceptible to yellow rust in the first year of trials, so were not included in later trials and their data not included in the analysis. Most lines were winter types released as commercial cultivars or breeding lines developed by European breeders. The presence of the resistance genes Stb6, Stb9 and Stb15 and of specific resistance to the M. graminicola isolates CA30JI, IPO001, IPO92006 and ISR398 in the lines was reported previously (Arraiano & Brown, 2006).
A total of 12 field trials was conducted over a period of 3 years (2001–03) in different sites in eastern England, the main area of wheat production in the UK. The sites were Downham Market (Cebeco Seed Innovations Ltd; trials Cebeco01, Cebeco02, Cebeco03) and Docking (Advanta Seeds Ltd; Advanta01, Advanta02) in Norfolk, Woolpit (Nickerson UK Ltd; Nickerson01, Nickerson03) in Suffolk, Great Abington (Syngenta Seeds Ltd; Syngenta01), Whittlesford (Syngenta Seeds Ltd; Syngenta02) and Abbots Ripton (SW Seed Ltd; SW02) in Cambridgeshire, Spalding (Elsoms Seeds Ltd; Elsoms01) in Lincolnshire and Doncaster (SW Seed Ltd; SW03) in South Yorkshire. In all experiments, fields were ploughed, disced and harrowed before planting. Each trial was sown in randomized complete blocks with two to three replications. In all trials, plots were squares of 1.5 × 1.0 m sown with 10–20 g of seeds. Seeds were dressed with Sibutol Secur (Bayer CropScience) to prevent all major seedborne diseases of winter wheat. All trials were sown in late October to promote natural infection with STB. Powdery mildew (Blumeria graminis f.sp. tritici) and yellow rust (YR; Puccinia striiformis f.sp. tritici) were controlled by a range of fungicides with activity against these diseases but little or no effect on STB, but control of YR was not effective in 2002. Weeds were controlled by pre- and post-emergence herbicides as appropriate. A growth regulator was applied in all years, according to the site’s standard crop-management procedure.
In 2001, disease on flag leaves was recorded as a single plot score. In 2002 and 2003, disease was scored on 10 separate main tillers, selected randomly in the middle rows of each plot. The percentage of flag-leaf area covered by lesions bearing pycnidia of M. graminicola was estimated when 10–30% of pycnidial coverage was detected on the flag leaf of two susceptible cultivars, Riband and Longbow. In four of the 12 trials, Elsoms01 and Nickerson01 in 2001, and Advanta02 and Syngenta02 in 2002, STB infection did not establish to adequate levels, so STB could not be scored in these trials. In the two other trials in 2002, Cebeco02 and SW02, both STB and yellow rust were scored as a percentage of the whole flag-leaf area covered by each disease (Table 1).
Data were collected for plant traits related to disease escape: heading date (HD); height of plant from the ground to the flag-leaf ligule in mid-May (Ht_midMay), at the time of heading (Ht_HD) and at maturity (GS85: Zadoks et al., 1974; Ht_FL); distance from the ground to the ligule of the second leaf (Ht_L2); length of the flag leaf (LL); and flag-leaf prostrateness (LP). Distance between the ligules of the flag and second leaves (FL_L2) was estimated from Ht_FL and Ht_L2. Owing to management constraints, not all of these characters were scored every year in every trial (Table 1).
HD of each plot was determined as the time when 50% of ears were half emerged (GS 55; Zadoks et al., 1974), and was recorded as days from the first of May. Height components were scored in three typical tillers for each plot. LL was recorded for 10 leaves per plot collected from one of the replicates at each site in 2003 and measured in the laboratory from the base to the tip of the leaf. LP data were recorded on tillers where the ears were half emerged on a six-point scale, adapted from the NIAB Variety Descriptions (DUS – distinctness, uniformity and stability) for barley: 1, erect; 2, semi-erect; 3, horizontal; 4, semi-recurved; 5, recurved; and 6, deflexed.
The presence of specific resistance genes and isolate-specific resistances in this set of lines was described previously (Arraiano & Brown, 2006).
As pycnidial coverage was scored on a percentage scale, the analysis was done on logit-transformed data. For STB, YR and LL, means for each plot were obtained using anova. The yellow rust epidemic in 2002 could not be fully controlled and may have affected STB scores. The logit-transformed score for yellow rust was therefore used as a covariate to obtain adjusted STB plot means in 2002.
For each trait, the mean for each line was estimated by generalized linear modelling (glm). The possible effects of trial, line and trial by line interaction were taken into account by fitting the model Site, Block × Site, Line and Line × Site in the glm procedure of genstat (2006).
Associations of disease severity with disease-escape traits and isolate-specific resistances were then estimated by glm. Owing to the large number of non-orthogonal terms which could be included in a model, the identification of a well-fitting yet parsimonious multiple regression model was done using the rsearch procedure in genstat. Choice of a model was based on the value of adjusted R2 and Mallows’Cp, where larger R2 and smaller Cp values indicated the best-fitting model (Goedhart, 2006). This analysis was done for the full set of 226 lines and also for a subset of 139 lines for which the height to the flag leaf was less than 68 cm, which included all Rht-D1b and Rht-B1b (formerly Rht2 and Rht1) semidwarf lines and very short lines with the tallness allele of those two genes (rht-D1b and rht-B1b).
Principal component (PC) analysis (PCA) was applied to a matrix of correlations between mean values of each trait for each line to identify new, orthogonal variables which accounted for variation in the dataset. Varimax factor rotation was then applied to the PC scores to identify a set of individual traits which were largely independent from one another and which accounted for as much as possible of the phenotypic variation. The correlation of these traits with susceptibility to STB was then tested. All analyses were done using the statistical package genstat.
Agronomic data analysis
In analyses of variance of agronomic traits in the 12 field trials, there were highly significant differences between the lines for all traits studied (Line term in Table 2). A significant line-by-site interaction (Line × Site term in Table 2) was also detected for all traits, but, in every case except YR02, it was considerably smaller than the main effect of line. The mean of YR02 within each plot was used as a covariate in the analysis of STB data in 2002, as described above. Otherwise, further analysis of traits was done on mean values for each line across all trials.
|Factor||STB (%)||YR02 (%)||HD (days in May)||Ht_midMay (cm)||Ht_HD (cm)||Ht_FL (cm)||Ht_L2 (cm)||LP||LL (cm)|
|Site × Block||88·9***||55·9***||42·9***||497.0***||837·4***||822·8***||668·2***||1·7**||–|
|Line × Site||0·9***||3·8***||8·2***||68.1***||41·4***||55·9***||43·7***||0·7***||–|
There were strong correlations between all of the height parameters and between flag-leaf length and prostrateness (Table 3). All four measurements of plant stature were strongly positively correlated. HD was strongly negatively correlated with Ht_midMay, presumably because earlier-flowering lines had made greater progress towards adult height by that date. LP was positively correlated with leaf length LL, implying that longer leaves tend to have a more prostrate attitude. In the full set of 226 lines, LL and LP were positively correlated with the height traits, but this association was weaker or even absent for 139 lines with mean Ht_FL <68 cm.
|All 226 lines|
|HD||−0·40***||0·08 ns||0·13 ns||0·17**|
|139 lines with Ht_FL ≤68 cm|
|HD||−0·61***||−0·12ns||−0·07 ns||0·01 ns|
|LP||0·34***||0·23**||0·14 ns||−0·01 ns||−0·15ns|
|LL||0·13ns||0·26**||0·20*||0·09 ns||0·14 ns||0·57***|
Analysis of all 226 lines
Further analysis of these correlations and their relationship to STB was done by PCA (Table 4). For the full set of lines, only the two first principal components (PC1 and PC2) had eigenvalues greater than 1 and accounted for 79.2% of the total variance, whereas the other five accounted for only 20.8% (Table 4). This implied that most of the variation in the traits scored reflected just two underlying variables. The first PC (PC1), accounting for 61.1% of the total variation, was mainly composed of three height parameters, Ht_FL, Ht_L2 and Ht_HD, plus LP and LL. This supported the conclusions from the pairwise correlations between traits (Table 3) that the different methods of scoring plant height reflected a single underlying variable and that height, leaf length and leaf prostrateness were positively associated. PC1 had a positive and highly significant correlation with STB, indicating that taller lines had less STB than shorter ones. For PC2, accounting for 18.1% of the variation, the main contribution was from heading date, contrasting with Ht_midMay, supporting the conclusion that earlier flowering lines tend to have earlier stem extension (Table 3). PC2 was weakly positively correlated with STB, indicating a tendency for later heading lines to have lower STB severity at the time of scoring. Although PC3 had an eigenvalue smaller than 1, it explained 13.1% of the variation and was much larger than the remaining four PCs. Its eigenvector indicated a positive correlation between leaf length and leaf prostrateness (Table 4), confirming the correlation between these parameters found in the correlations matrix (Table 3). This PC had a negative significant correlation with STB, indicating that after accounting for the positive correlation of height with LL and LP by PC1, lines of similar height with longer and more prostrate leaves tended to have more STB.
|All 226 lines|
|Loadings of parameters|
|Correlation with STBa||0·6***||0·2*||−0·3***||0·0 ns||−0·2*||0·0 ns|
|139 lines with Ht_FL ≤68 cm|
|Loadings of parameters|
|Correlation with STBa||0·2*||−0·3***||0·0 ns||0·1 ns||0·0 ns||−0·2*|
On a plot of the first three PCs, it was evident that the majority of variation was absorbed by three variables: a height variable (which could be represented by Ht_HD, Ht_FL or Ht_L2), heading date (HD) and a leaf variable which could be represented by either LL or LP (shown for all 226 lines in Fig. 1; the pattern for the semidwarf lines was similar). This conclusion was supported by varimax factor rotation (results not shown).
The correlation between STB and PC1 was supported by the consistently negative correlations between STB and the height parameters at each site and across sites (Table 5, Fig. 2a). Heading date was significantly negatively correlated with STB at four out of eight sites. Whilst mean STB was negatively correlated with heading date across sites (Fig. 2b), this correlation was not significant, in contrast to the correlation of STB with PC2, combining HD and Ht_midMay (Table 4). This implies that the correlation of STB with PC2 reflected plant height, indicated by the strong correlation of Ht_midMay with mature plant height (Table 3), rather than earlier HD. LP and LL were negatively correlated with STB, reflecting the combined tendency of taller lines to have both longer, more prostrate leaves and less STB (PC1), as indicated by PC3 (Table 4).
|All 226 lines|
|STB mean across sites||−0·4***||−0·7***||−0·7***||−0·7***||−0·1ns||−0·2**||−0·3***|
|139 lines with Ht_FL ≤68 cm|
|Cebeco03||−0·1 ns||−0·1 ns||−0·1 ns||0·0ns||0·1 ns||–||−0·2*|
|Nickerson03||–||−0·1 ns||−0·1 ns||−0·1ns||0·1 ns||–||−0·1 ns|
|SW03||–||0·0 ns||−0·2*||0·0ns||−0·1 ns||–||−0·1ns|
|STB mean across sites||−0·1ns||−0·2**||−0·3**||−0·3**||0·0ns||0·2*||0·1ns|
Analysis of lines with Ht_FL <68 cm
Similar relationships between the phenology traits and between them and STB were found for the subset of 139 lines of semidwarf stature (Ht_FL <68 cm). There were three PCs with eigenvalues larger than 1, accounting for 88.7% of the total variation (Table 4). PC1 accounted for 47.8% of the variation and, as in the whole set of 226 lines, the height parameters had the highest loadings. The correlation of height and consequently PC1 with STB was lower and less significant for the semidwarf lines than with the full set of lines, reflecting the lower variance in the height of this subset of lines. The main contributors to PC2, which explained 21.4% of the variation, were leaf prostrateness and length. PC2 was the dimension most highly correlated with STB, implying that semidwarf cultivars with longer, more prostrate leaves tended to have more STB. PC3 explained 19.5% of the variation and the highest contribution to it was heading date, contrasting again with Ht_midMay. This PC was not correlated with STB.
For this set of 139 lines of semidwarf stature, the correlations of STB with agronomic characters at individual trial sites were weaker than for the full set of lines, confirming the findings of the PCA (Table 5).
Stb resistance genes and isolate-specific resistances
Several generalized linear models of the lines’ mean STB scores were evaluated using the rsearch procedure in genstat. These models included known Stb genes (Stb6, Stb9 and Stb15), specific resistances controlled by unidentified genes (R_CA30JI, R_IPO001, R_IPO92006 and R_ISR398) and the seven agronomic traits scored. rsearch indicated that a model with four terms gave the best fit for all 226 lines (R2 = 49.0, Cp = 2.70) and a slightly different model was best for the 139 semidwarf cultivars (R2 = 16.3, Cp = 2.99; Table 6).
|Factor||d.f.a||Variance||Variance ratio||Pb||STB variationc (%)|
|All 226 lines|
|Height to flag leaf||1||60·1||200·6||***||45·0|
|Height in mid May||1||0·9||3·0||ns||0·4|
|139 lines with Ht_FL≤68 cm|
|Height to flag leaf||1||3·5||8·6||**||4·2|
The residuals from the model in which STB scores were fitted to all escape traits which had a significant effect could be treated as estimates of the susceptibility of each of the lines’ leaves to STB. For both the complete set of 226 and the 139 semidwarf lines, after removing the effect of variation in Ht_FL and LP, the only resistance gene significantly associated with reduced STB in field conditions was Stb6 (Table 6). For the modern lines of semidwarf stature, the presence of Stb6 was even more strongly associated with reduced STB than in the full set of 226 lines (P < 0.001). The presence of resistance gene Stb15, the only gene other than Stb6 present in numerous cultivars (Arraiano & Brown, 2006), was not associated with STB levels, and the effect of Stb9, although present in the semidwarf model, was not significant (Table 6). The estimates of foliar susceptibility were 19% (P < 0·001) and 33% (P = 0·002) lower on average in lines with Stb6 than in those without, for the complete set of 226 lines and the semidwarf lines, respectively (Fig. 3).
Two modern UK cultivars, Pastiche (Plant Breeding Institute, late 1980s) and Exsept (Nickerson UK Ltd, late 1990s), were identified as having strong partial resistance that could not be accounted for by disease-escape traits or the presence of isolate-specific resistances (Table 7, Fig. 3).
|NSL 94-6897||2||61||R||0||Caribo||3||75||3||Grosse Tete||1||92||R||6||Maris Bilbo||18||57||11|
|Riebesel 57/41||1||79||1||Et’le de Choissy||5||68||3||Hereward||7||59||R||6||Brimstone||9||67||11|
|Anduril||2||63||1||Maris Templar||7||64||3||Marne Desprez||4||79||6||Buster||10||62||11|
|Arminda||2||66||1||Maris Nimrod||3||69||4||Harrier||10||61||7||Maris Envoy||8||75||13|
|NSL 93-5372||3||60||1||Br. Teutonen||3||81||4||M. Widgeon||4||80||7||Durin||14||63||14|
|Sportsman||5||62||1||80 30 Versailles||2||87||4||Garnet||4||81||7||Sicco||8||72||14|
|Vilmorin 23||2||75||2||Chiddam Blanc||1||99||4||Melbor||2||89||R||8||Rialto||7||67||R||16|
|Heines Kolben||2||73||2||Moulin||7||60||4||Koga 1||3||82||R||8||Koga II||5||77||16|
|M. Ploughman||2||77||2||Mardler||9||56||R||4||WH929623-5||11||57||8||Camp Remy||11||67||17|
|Bastard II||1||81||2||Thatcher||2||80||4||White Fife||2||97||8||Jufy I||7||80||17|
|Cappelle Desp.||3||71||2||NSL WW13||5||65||4||Madrigal||11||57||8||Hedgehog||17||56||17|
|Blé Seigle||1||92||2||Elysee||4||72||R||5||Hatif Invers.||3||87||9||Norman||17||61||23|
|Buchan||6||58||2||Maris Ensign||4||77||5||Cleo||7||73||9||Heines Peko||5||89||23|
|Dynamo||8||55||3||Magellan||9||60||5||Joss Cambier||7||70||10||Prof. Delos||5||89||26|
|Aintree||5||67||3||Heines 110||3||84||5||Bon Fermier||3||83||R||10||Maris Beacon||12||74||31|
|NSL 92-5719||4||62||R||3||Equinox||10||55||5||Maestro||10||64||10||Maris Dove||9||75||R||39|
|Spark||4||70||3||Brock||9||59||5||Maris Freeman||9||70||10||Maris Pinion||16||75||41|
The research reported here identified cultivars with the potential to be sources of resistance to STB in bread wheat, by separating the effects of escape traits on disease levels from those of foliar resistance per se. Underpinning the analysis presented here is the assumption that a certain amount of inoculum is present in a plot when the epidemic is initiated and that, in a line with low severity of STB on its flag leaves, the amount of disease caused by that inoculum can be reduced in two ways. Disease resistance reduces the amount of disease per unit of inoculum, whilst disease escape reduces the amount of inoculum reaching the flag leaf or delays infection of that leaf. This assumption leads to the methodology used here, in which all reductions in STB that can be explained by linear regression on plant morphology and developmental traits are treated as disease escape, while the residuals from that regression are considered as resistance, although it is possible that those residuals could also include an element of escape caused by as-yet unknown traits.
The strong negative correlation between STB levels and plant height for the full set of 226 cultivars and breeding lines, measured in four ways (Table 5), accords with several previous studies (van Beuningen & Kohli, 1990; Baltazar et al., 1990; Jlibene et al., 1992; Camacho-Casas et al., 1995; Chartrain et al., 2004b; Simón et al., 2004). Among the 139 lines of semidwarf stature, however, the correlation was weaker and more variable between sites. Together, these results imply that the shift from tall to semidwarf stature has substantially increased levels of STB, but that, within modern, semidwarf cultivars, selection for increased height would have little benefit in providing escape from STB.
The strong correlation between the four separate measures of plant stature (Table 3) implies that the time required to score field trials in research on STB or other disease in which escape traits are implicated can be minimized by scoring a single variable related to final plant height rather than several related traits. Here, the trait that most strongly explained variation in STB levels was final height to flag leaf (Ht_FL), which is thus an appropriate height trait to be scored in north-west European conditions.
Other traits, including later heading, more prostrate leaves and reduced leaf length, made lesser contributions to disease reduction (Table 4). It has been proposed that later heading affects measurements of STB at a given time because it reduces the length of time for which the relevant leaf has existed and has thus been exposed to infection by M. graminicola at the time the trial is scored (Shaw & Royle, 1993). Here, however, the effect of heading date on STB levels was minor and inconsistent between trials, despite a wide range of variation between lines in mean heading date (28–47 days from 1 May, Tables 2 and 5). This supports the work of some authors (Arama et al., 1999; Simón et al., 2005) but not others (Chartrain et al., 2004;Arraiano et al., 2006; Zhang et al., 2007) and thus creates further uncertainty about the contribution of later heading to disease escape.
The correlations of leaf prostrateness and leaf length with STB levels (Tables 3–6, Fig. 2) were complicated by their interaction with plant height (Table 3). The positive correlation of leaf prostrateness and length with each other (Table 3) may largely have been a mechanical effect of the greater mass of longer leaves causing greater curvature, but this has not been tested. In the set of 139 semidwarf lines, both leaf traits were positively correlated with STB severity (Table 5). In the full set of 226 lines, however, they were positively correlated with height (Table 3), and since the effect of the very wide variation in plant height on septoria was so great (Table 6), means of prostrateness and length were negatively correlated with mean STB levels. The complex, three-way interaction between STB, height and leaf morphology was reflected in the PCA (Table 4). PC1 for all 226 lines, which had negative loadings of all four height variables and of leaf prostrateness and length, was positively associated with STB, implying that lines which were shorter (and thus had shorter, more erect leaves on average; Table 3) tended to have more STB (Table 4). PC3, however, largely consisted of negative loadings of the leaf traits and was negatively correlated with STB (Table 4), implying that after allowing for the effect of height on STB, lines which had longer, more prostrate leaves tended to have more STB. However, in the set of 139 semidwarf cultivars, variation in height had a much smaller effect on STB (Tables 5 and 6) and was much less significantly correlated with LP and LL (Table 3 and 4) so the positive correlation of prostrateness and length with STB (Table 5) was not complicated by interactions with height.
Lovell et al. (1997) proposed that plants with longer, more erect leaves would have more STB because upper leaf layers would overlap with lower leaves for a greater fraction of their lifetime so less kinetic energy would be required to move inoculum between leaf layers from basal leaves to the top of the canopy. The results presented here do not support this hypothesis because lines with more erect leaves tended to have less STB, not more (Tables 4–6), and because leaf length had no significant effect on STB levels after accounting for the effects of height and leaf. Genetic analysis is required to test whether reduction in STB is a direct effect of leaf erectness itself (Table 6), the result of genes for erectness being linked to those for STB resistance, or merely a chance association between the two traits in UK wheat germplasm.
Once the effects of disease-escape traits were accounted for, the variable which explained most variation in levels of STB was the presence or absence of the Stb6 resistance phenotype (Table 6). In an earlier study, all lines expressing this phenotype, in the form of specific resistance to the M. graminicola isolate IPO323, had the Stb6 resistance gene (Chartrain et al., 2005c). It is unlikely that the effect of Stb6 in contributing to reduced STB levels is the result of it being effective against a significant fraction of the M. graminicola population, because the frequency of avirulence was very low (less than 1% in a sample of over 500 M. graminicola isolates; Brown et al., 2003). Stb6 is present in many cultivars that have been used as sources of resistance to STB in world breeding programmes (Chartrain et al., 2005c) and may itself confer partial resistance to STB. It is conceivable that the effect of Stb6 is the residual effect of a ‘defeated’ gene-for-gene resistance on virulent pathogen genotypes, as has been reported in potato, rice and apple (Pedersen & Leath, 1988; Danial et al., 1994; Durel et al., 2000; Li et al., 2001). Alternatively, Stb6 may be linked to another gene which confers partial resistance. Genetic analysis is required to test these two hypotheses. It is possible that not all the sources of Stb6 in current wheat germplasm (Chartrain et al., 2005c) contribute equally to control of STB in the field, since some Stb6 cultivars are susceptible in field conditions (Fig. 3). No other resistance gene or phenotype contributed significantly to reduction of STB (Table 6).
The data presented here have implications for methods of selecting for reduced STB in wheat cultivars. It has been known for some time that tall, late-flowering wheat cultivars tend to have lower levels of STB (Eyal, 1981; van Beuningen & Kohli, 1990; Camacho-Casas et al., 1995; Chartrain et al., 2004;Arraiano et al., 2006). The information presented here shows that this holds in UK conditions, but, within the range of modern wheat cultivars, taller, later-flowering plants have only a slight advantage in terms of reduced STB, so selecting for these traits solely as a means of escaping the disease will probably have little benefit. It is therefore possible to select lines with good partial resistance to STB while selecting simultaneously for traits related to plant growth and development which are optimal in local conditions.
The resistance gene Stb6 is associated with a useful reduction in levels of STB in field conditions. Either the M. graminicola isolate, IPO323, which has the avirulence gene recognized by Stb6 (Kema et al., 2000), or markers linked to Stb6 (Brading et al., 2002; Eriksen et al., 2003) can be used to assist selection for resistance to STB. However, this gene alone is not sufficient to provide adequate resistance to STB in the field.
The data also imply that, when conducting genetic analysis on STB severity in adult plants, escape traits need to be separated, as far as possible, from disease resistance (Arraiano et al., 2006). This would enable one to determine whether quantitative trait loci (QTL) associated with reduced STB severity did so by increasing resistance or escape.
During the 1970s, there was a shift in the main facultative disease of wheat in the UK from stagonospora nodorum blotch (SNB), caused by Phaeosphaeria nodorum (anamorph Stagonospora nodorum), to STB. Concurrently, there was a substantial reduction in the level of atmospheric pollution by SO2. Moreover, the level of P. nodorum on cv. Squarehead’s Master in the first half of the 20th century rose in parallel with an increase in atmospheric partial pressure of SO2 (pSO2), so it was proposed that the fall in SNB and rise in STB could be caused in part by reduced pSO2 (Bearchell et al., 2005). Whilst the set of cultivars reported here does not comprehensively cover those grown in the UK up to about 1960, it is notable that Bersée, the most important cultivar in the UK in the 1940s, and Cappelle Desprez, which was grown on more than half the wheat-cropping area in the UK between 1955 and 1965 (Bonjean & Angus, 2001), had good partial resistance to STB, as did their progenitors Vilmorin 23, Vilmorin 27 and Hybride du Joncquois (Table 7). Many cultivars released by UK breeders from the 1960s onwards, notably Maris Kinsman and Maris Ranger, which were progenitors of all subsequent UK cultivars, and many of their descendants including such important semidwarf cultivars as Maris Bilbo, Longbow, Norman, Riband and Consort, were susceptible to STB (Table 7). Cappelle Desprez was susceptible to SNB, but substantial progress in breeding for resistance to SNB was made by the Plant Breeding Institute (Scott & Benedikz, 1986), whose cultivars dominated UK wheat production between 1975 and 2000. It is likely, therefore, that a switch in the major wheat cultivars in the UK from susceptibility to SNB and resistance to STB to the reverse situation accounted for a substantial proportion of the switch from P. nodorum to M. graminicola as the predominant facultative fungal pathogen of wheat, along with the increase in atmospheric pSO2.
The results presented here and in a previous paper (Arraiano & Brown, 2006) indicate that there is a wide range of partial resistance to STB among UK wheat cultivars. An important challenge to wheat breeders, therefore, is to combine resistance genes from different cultivars. Several lines have been identified with especially high levels of partial resistance, as distinct from disease-escape traits. Such cultivars, notably Pastiche, Exsept, Jena, Flame and Boxer, may be useful sources of resistance in breeding for STB resistance. In particular, Pastiche and Exsept have no known specific resistance genes or isolate-specific resistances (Arraiano & Brown, 2006), indicating that they must have genes for partial resistance which remain to be identified. It is notable that several of the lines with the greatest resistance, including Exsept, Boxer, Flame and the NSL lines listed in Table 7, were bred by Nickerson UK Ltd, which suggests that there may be useful partial resistance genes in that company’s breeding germplasm.
We thank the many members of our organisations who helped with scoring the field trials. This research was supported by Defra and the companies which were members of the project consortium through the Sustainable Arable LINK Programme.
- 2003. Identification and molecular mapping of a gene in wheat conferring resistance to Mycosphaerella graminicola. Phytopathology 93, 1158–64. , , ,
- 2004a. Molecular mapping of Stb1, a potentially durable gene for resistance to septoria tritici blotch in wheat. Theoretical and Applied Genetics 109, 944–53. , , et al. ,
- 2004b. Microsatellite markers linked to the Stb2 and Stb3 genes for resistance to septoria tritici blotch in wheat. Crop Science 44, 1403–11. , , ,
- 2004c. Molecular mapping of the Stb4 gene for resistance to Septoria tritici blotch in wheat. Phytopathology 94, 1198–206. , , , , , ,
- 1999. Heading date and resistance to septoria tritici blotch in wheat not genetically associated. Euphytica 106, 63–8. , , ,
- 2006. Identification of isolate-specific and partial resistance to septoria tritici blotch in 238 European wheat cultivars and breeding lines. Plant Pathology 55, 726–38. , ,
- 2001. Chromosomal location of a gene for resistance to septoria tritici blotch (Mycosphaerella graminicola) in the hexaploid wheat ‘Synthetic 6x’. Theoretical and Applied Genetics 103, 758–64. , , , ,
- 2006. Resistance of wheat to septoria tritici blotch (Mycosphaerella graminicola) and associations with plant ideotype and the 1BL-1RS translocation. Plant Pathology 55, 54–61. , , , ,
- 2007. A gene in European wheat cultivars for resistance to an African isolate of Mycosphaerella graminicola. Plant Pathology 56, 73–8. , , , , , ,
- 1980. Factors affecting the vertical progression of septoria leaf blotch in short-statured wheat. Phytopathology 70, 179–84. , , , ,
- 1990. Association between dwarfing genes Rht1 and Rht2 and resistance to Septoria tritici blotch in winter wheat (Triticum aestivum L. em Thell). Theoretical and Applied Genetics 79, 422–6. , , ,
- 2005. Wheat archive links long-term fungal pathogen population dynamics to air pollution. Proceedings of the National Academy of Sciences, USA 102, 5438–42. , , , ,
- 1990. Deviation from the regression of infection on heading and height as a measure of resistance to septoria tritici blotch in wheat. Plant Disease 74, 488–93. , ,
- 2001. The World Wheat Book: a History of Wheat Breeding. Andover, UK: Intercept. , ,
- 2002. A gene-for-gene relationship between wheat and Mycosphaerella graminicola, the Septoria tritici blotch pathogen. Phytopathology 92, 439–45. , , , ,
- 2003. Identifying and evaluating specific resistance genes for use in controlling septoria tritici blotch. In: KemaGHJ, Van GinkelM, HarrabiM, eds. Global Insights into the Septoria and Stagonospora Diseases of Cereals. Proceedings of the Sixth International Symposium on Septoria and Stagonospora Diseases of Cereals 2003, Tunis, Tunisia, 111–2. , , , , , ,
- 2008. Evolution of the CYP51 gene in Mycosphaerella graminicola: evidence for intragenic recombination and selective replacement. Molecular Plant Pathology 9, 305–16. , , ,
- 1995. Septoria tritici resistance and associations with agronomic traits in a wheat cross. Crop Science 35, 971–6. , , ,
- 2004a. Sources of resistance to septoria tritici blotch and implications for wheat breeding. Plant Pathology 53, 454–60. , , , ,
- 2004b. Partial resistance to Septoria tritici blotch (Mycosphaerella graminicola) in wheat cultivars Arina and Riband. Phytopathology 94, 497–504. , , , ,
- 2005a. Resistance of wheat line Kavkaz-K4500 L.6.A.4 to Septoria tritici blotch controlled by isolate-specific resistance genes. Phytopathology 95, 664–71. , , ,
- 2005b. Genetics of resistance to Septoria tritici blotch in the Portuguese wheat breeding line TE 9111. Theoretical and Applied Genetics 110, 1138–44. , , , , , ,
- 2005c. Presence of the Stb6 gene for resistance to septoria tritici blotch (Mycosphaerella graminicola) in cultivars used in wheat-breeding programmes worldwide. Plant Pathology 54, 134–43. , , ,
- 2009. Identification and location of Stb9, a gene for resistance to septoria tritici blotch in wheat cultivars Courtot and Tonic. Plant Pathology 58, 547–55. , , , ,
- 1994. Evolution of virulence patterns in yellow rust races and its implications for breeding for resistance in wheat in Kenya. Euphytica 80, 165–70. , , ,
- 1982. The relationship among stature, maturity class, and susceptibility to septoria leaf blotch of wheat. Phytopathology 72, 1037–42. , , ,
- 2000. Does the Vf gene maintain a residual resistance to apple scab despite its breakdown by Venturia inequalis race 6 strains. Acta Horticulturae 538, 575–80. , , , , ,
- 2003. Inheritance and localisation of resistance to Mycosphaerella graminicola causing septoria tritici blotch and plant height in the wheat (Triticum aestivum L.) genome with DNA markers. Theoretical and Applied Genetics 107, 515–27. , , ,
- 1981. Integrated control of septoria diseases of wheat. Plant Disease 65, 763–8. ,
- 2005. Role of ascospores in further spread of QoI-resistant cytochrome b alleles (G143A) in field populations of Mycosphaerella graminicola. Phytopathology 95, 933–41. , , et al. ,
- GenStat Release 9.1 2006. GenStat® for Windows Release 9.1. Hemel Hempstead, UK: VSN International.
- 2006. Procedure RSEARCH. In: PayneRW, ed. The Guide to GenStat® Release 9 Part 2: Statistics. Hemel Hempstead, UK: VSN International, 210–6. ,
- 2003. The Mycosphaerella sequencing initiative. In: KemaGHJ, Van GinkelM, HarrabiM, eds. Global Insights into the Septoria and Stagonospora Diseases of Cereals: Proceedings of the Sixth International Symposium on Septoria and Stagonospora Diseases of Cereals 2003, Tunis, Tunisia, 149–51. , , ,
- 1992. Field disease evaluation method for selecting wheats resistant to Mycosphaerella graminicola. Plant Breeding 108, 26–32. , , ,
- 1984. A critical analysis of durable resistance. Annual Review of Phytopathology 22, 309–30. ,
- 2000. Avirulence in the wheat septoria tritici leaf blotch fungus Mycosphaerella graminicola is controlled by a single locus. Molecular Plant-Microbe Interactions 13, 1375–9. , , ,
- 2001. Are the dominant and recessive plant disease resistance genes similar?: a case study of rice R genes and Xanthomonas oryzae pv. oryzae races. Genetics 159, 757–65. , , et al. ,
- 1997. Influence of crop growth and structure on the risk of epidemics by Mycosphaerella graminicola (Septoria tritici) in winter wheat. Plant Pathology 46, 126–38. , , , , ,
- 2004. Position of inoculum in the canopy affects the risk of septoria tritici blotch epidemics in winter wheat. Plant Pathology 53, 11–21. , , , , ,
- 2003. Chromosomal location of a race-specific resistance gene to Mycosphaerella graminicola in the spring wheat ST6. Theoretical and Applied Genetics 107, 1181–6. , , , ,
- NIAB (19972000). Recommended Lists for Winter Wheat. In: Cereal Variety Handbooks 1997–2000. Cambridge, UK: National Institute of Agricultural Botany.
- 1988. Pyramiding major genes for resistance to maintain residual effects. Annual Review of Phytopathology 26, 369–78. , ,
- 1986. Septoria. In: Annual Report of the Plant Breeding Institute 1986. Cambridge, UK: Plant Breeding Institute, 97–8. , ,
- 1993. Factors determining the severity of epidemics of Mycosphaerella graminicola (Septoria tritici) on winter wheat in the UK. Plant Pathology 42, 882–99. , ,
- 2004. Influence of plant height and heading date on the expression of the resistance to Septoria tritici blotch in near isogenic lines of wheat. Crop Science 44, 2078–85. , , ,
- 2005. Association between Septoria tritici blotch, plant height, and heading date in wheat. Agronomy Journal 97, 1072–81. , , , , , ,
- 1974. A decimal code for the growth stages of cereals. Weed Research 14, 415–21. , , ,
- 2007. Modeling the effect of cultivar resistance on yield losses of winter wheat in natural multiple disease conditions. European Journal of Agronomy 26, 384–93. , , ,