Septoria tritici blotch (STB), caused by Zymoseptoria tritici (formerly Mycosphaerella graminicola and Septoria tritici), is a widespread and economically significant disease of wheat (Eyal, 1981; Eyal et al., 1985; 1987). STB is a serious threat to wheat production in the USA, particularly in the Pacific Northwest (Jackson et al., 2000) where the environment is favorable for disease development almost every year, and in other growing regions when spring rains are favorable for disease development. Wheat cultivars resistant to Z. tritici restrict or delay pathogen development and disease symptoms (Nelson and Marshall, 1990). The disease can be spread by air-borne ascospores or splash-dispersed conidia, infecting the upper leaves during the wheat-growing season (Eyal et al., 1987; Ponomarenko et al., 2011). In disease-conducive environments, epidemics can quickly cause significant reductions in yield on susceptible plants (Gilbert et al., 1998; Shaner and Finney, 1976). Histopathological analysis has revealed that Z. tritici infects its host plant through stomata. During infection, the hyphae remain strictly in the intercellular space in close contact with mesophyll cells (Cohen and Eyal, 1993; Duncan and Howard, 2000; Kema et al., 1996b). However, no specialized feeding structures like haustoria are formed. The first visible symptoms of the infection appear as chlorotic areas on the leaf surface about 10 days after inoculation (DAI) (Ponomarenko et al., 2011). These chlorotic areas later become necrotic lesions starting around 14 to 18 DAI. When the first signs of pycnidia appear, these structures emerge from the stomata and become filled with basket-shaped masses of mycelia (Cohen and Eyal, 1993; Duncan and Howard, 2000; Kema et al., 1996b).

Wheat cultivars are usually assessed for resistance to Z. tritici by visual estimation of symptoms. However, this method has several disadvantages. First, the fungus has a long latent period (up to 14 or 18 days), so the standard time for evaluating resistance to Z. tritici in wheat cultivars is between 24 and 28 DAI. Although scoring flag leaves of adult plants usually produces the best results (Adhikari et al., 2003), growing plants to maturity slows evaluation and limits the number of plants tested. Second, using an integer scale for rating resistance may vary from person to person, making it challenging to distinguish resistant plants phenotypically. Finally, the expression of resistance to Z. tritici under greenhouse and field conditions can be significantly influenced by the environment. Some of these difficulties can be minimized by detached-leaf assays (Arraiano et al., 2001a), by scoring plants in a greenhouse (Kema et al., 1996a) or by automating the phenotyping step (Stewart et al., 2016), but these require much more effort and reduced throughput. Therefore, a rapid, sensitive, and high-throughput method for quantifying fungal DNA would be helpful for accurately assessing pathogenicity and for large-scale testing of plant genetic resources for resistance to Z. tritici.

Several methods have been used for detecting and quantifying Z. tritici in susceptible and resistant wheat cultivars. Cohen and Eyal (1993) estimated mycelial growth histologically in the intercellular spaces of mesophyll cells of resistant and susceptible wheat cultivars. Although fungal growth increased dramatically in the susceptible response during the later stages of pathogenesis, no differences were observed at an early stage of the infection (8 DAI) (Kema et al., 1996b). The histopathological analyses also showed that a non-virulent mutant colonized sub-stomatal cavities less efficiently and displayed reduced intercellular growth and concomitantly less fungal growth in the apoplast of wheat leaves compared to untransformed controls (Stergiopoulos et al., 2003). In another approach, ß-glucuronidase-producing transformants of Z. tritici were used to examine fungal development in inoculated wheat leaves. In general, the rates of fungal growth and the estimated levels of fungal protein when the pycnidia matured appeared to be high in compatible interactions and moderate to low in incompatible interactions (Pnini-Cohen et al., 2000).

A multiplex polymerase chain reaction (PCR) method was used to detect and quantify four foliar fungal pathogens of wheat, including Z. tritici (Fraaije et al., 2001; Guo et al., 2006). However, those primer pairs for each foliar pathogen varied in sensitivity due to weak non-specific amplification and primer dimer formation during the PCR method. Although some of the methods reported previously are promising, they are not sensitive enough to monitor the slight changes in fungal DNA that may occur during pathogenesis. The real-time PCR (Heid et al., 1996) method has proven to be rapid, sensitive, and accurate enough to estimate the quantity of Z. tritici DNA in large numbers of inoculated wheat plants. Real-time PCR has been used successfully for the detection and quantification of a range of plant pathogens, such as bacteria (Brouwer et al., 2003; Humphris et al., 2015; Peňázová et al., 2020; Schaad et al., 1999; Weller et al., 2000), fungi (Bates et al., 2001; Brouwer et al., 2003; Chu et al., 2019; Gachon and Saindrenan, 2004; Milgate et al., 2023; Oliver et al., 2008; Ozdemir et al., 2020; Qi and Yang, 2002; Vandemark et al., 2002; Winton et al., 2003) and viruses (Balaji et al., 2003; Diaz-Lara et al., 2020). It has been used to estimate fungicide resistance in field populations of Z. tritici (Fraaije et al., 2005), to detect presymptomatic fungal biomass to guide fungicide applications (Guo et al., 2006; Tonti et al., 2019) and to identify early responses of candidate genes in resistant cultivars (Farsad et al., 2013).

More than 20 major genes for resistance to STB have been identified and mapped in wheat (Chartrain et al., 2005; Goodwin, 2007; Langlands-Perry et al., 2022; Somasco et al., 1996; Yang et al., 2022), and several have now been cloned (Hafeez et al., 2025; Saintenac et al., 2018, 2021). These Stb genes interact with the pathogen in a gene-for-gene manner (Brading et al., 2002) leading to large phenotypic differences between susceptible and resistant genotypes. In addition to the major genes, more than 170 quantitative trait loci (QTL) have been identified that are distributed on all wheat chromosomes (Brown et al., 2015). However, the effects of these genes on the pathogen while growing within its host are not well understood.

Breeding wheat for resistance to Z. tritici could benefit from a comprehensive analysis of fungal DNA accumulation in susceptible (compatible) and resistant (incompatible) interactions. The ultimate goal of this study was to use real-time PCR as a disease assessment method to differentiate between resistant and susceptible wheat cultivars and to enhance our understanding of STB development. We hypothesized that the quantity of fungal DNA increases rapidly after infection of susceptible wheat cultivars by Z. tritici, but that it should remain constant or increase only slightly in resistant cultivars. This change should be noticeable in the amount of fungal DNA present in total nucleic acids extracted from inoculated wheat leaves, potentially much earlier than symptom development. The specific objectives of this study were to: (i) develop an improved, quantitative real-time PCR (qPCR) method for quantifying Z. tritici in infected wheat leaves; (ii) use the qPCR method to monitor fungal DNA accumulation over a 27-day time-course in two susceptible and two resistant wheat cultivars with different specific resistance genes; (iii) apply the qPCR method to examine the effects of inoculum concentration and plant age on fungal DNA accumulation; and (iv) determine the relationship between fungal DNA and disease severity in a population of segregating resistant and susceptible recombinant-inbred lines (RILs) to compare the real-time PCR quantification with visual disease-assessment methods.

In these experiments, specific primers were designed for a particular region of the ß-tubulin gene that provided an efficient method to observe the changes in the amount of Z. tritici DNA during compatible and incompatible interactions in wheat leaves. Previous primers for multiplex and real-time quantitative PCR of Z. tritici based on sequences of the ß-tubulin gene (Fraaije et al., 2001; Guo et al., 2006) were specific for Z. tritici when tested on a range of unrelated fungi that also cause disease in wheat. Still, they gave a product of 555 bp, outside the 50-150-bp range recommended for real-time PCR. In addition, those primers were not specific for Z. tritici when tested against related Septoria, Mycosphaerella and Zymoseptoria species. A more elegant approach used the real-time PCR method and TaqMan probes to detect specific alleles of the cytochrome b gene for resistance or susceptibility to fungicides (Fraaije et al., 2005). Although that method can monitor fungal DNA (Keon et al., 2007), it is more complicated and expensive than the SYBR green technique. This newly developed real-time PCR method can accurately detect and quantify Z. tritici DNA to estimate disease resistance in wheat cultivars.

The primer set ß-tubulin-2 was highly specific for Z. tritici isolates except for S. triseti, the most closely related to Z. tritici of the species tested (S. B. Goodwin, unpublished data), and which therefore should be reclassified into the genus Zymoseptoria. The isolate of S. triseti tested was from Phalaris arundinacea (reed canarygrass), and this pathogen is specific for that host, so cross amplification should not be a problem when the ß-tubulin-2 primers are used on infected wheat. More importantly, no amplification was detected from the genomic DNA of key fungal and bacterial pathogens of wheat and the other Septoria and Mycosphaerella species tested from other hosts. Other major fungal pathogens of wheat are too distantly related for the primer sites to be conserved, as confirmed by blastn analyses of GenBank and genomic sequences, including all members of the Pucciniales and Erysiphales. These results suggest that the primer set ß-tubulin-2 was highly specific for use in the real-time PCR assay. High negative correlations between the DNA quantities estimated from standard curves and the C_T values further suggest that this method is reliable. In addition, no amplification was detected in DNA isolated from healthy or un-inoculated wheat leaves, confirming that the primer set ß-tubulin-2 was specific for Z. tritici.

There were notable changes in the amounts of fungal DNA found in the susceptible cultivars throughout the 27-day time course. Initially, during the first 3 DAI in both resistant and susceptible cultivars, there was an increase in fungal DNA, which was likely due to spore germination and epiphytic growth of the fungus in the enclosed inoculation chamber. This early stage of pathogenesis is essential for establishing disease, spreading the fungus to the apoplast, and colonizing neighboring sub-stomatal cavities (Duncan and Howard, 2000; Fones et al., 2017). Increase in fungal biomass even on resistant cultivars soon after inoculation is consistent with the epiphytic growth that is known to occur on resistant hosts under permissive growth chamber conditions (Fones et al., 2023). Subsequently, there was a gradual decrease in fungal DNA from about 2 or 3 DAI to 6 or 9 DAI. This decrease was probably due to a change in conditions from the sheltered environment of the inoculation chamber to the relatively high solar radiation and low humidity of an unprotected greenhouse bench at 3 DAI. Any fungal hyphae remaining outside the host at this stage were likely to die rapidly, which helps explain the observed decrease in fungal DNA.

During the latent phase of infection, small but statistically significant differences in fungal DNA were detected between the resistant and susceptible cultivars beginning at 12 DAI. The level of fungal DNA remained relatively consistent and low in the susceptible cultivars during the latent phase until symptoms became evident at around 18 DAI. After that, both symptoms and fungal DNA increased dramatically. A similar pattern occurred with the rust pathogen P. graminis f. sp. tritici, in which the increase in fungal DNA in susceptible interactions was concomitant with symptom expression (Mayama et al., 1975). However, this result contrasts with those from fungi such as the rice blast pathogen, M. oryzae, which showed a rapid and constant increase in fungal DNA beginning two days after inoculation (Qi and Yang, 2002). The data for Z. tritici suggest that a developmental switch may be initiated to cause the rapid increase in fungal DNA and onset of symptoms. This switch most likely occurs when the fungus transitions from its biotrophic to its necrotrophic stage of infection, and it could be triggered by toxins such as those produced by some species related to Z. tritici (Perrone et al., 2001). For example, cercosporin is created by members of the anamorphic genus Cercospora (Perrone et al., 2001). Another hypothesis is that the pathogen triggers programmed cell death in the host (Keon et al., 2007). Whatever the mechanism, these dynamic changes in Z. tritici DNA in the susceptible cultivars were significantly correlated with the onset of symptoms so can serve as an indicator of susceptibility. Similarly high correlations between disease susceptibility and Z. tritici DNA estimated by real-time PCR were reported by Guo et al. (2006), but that analysis did not include resistant cultivars.

Intriguingly, it was observed that there was a slight increase in disease severity and fungal DNA in the resistant cultivar W7984 at 24 DAI. These results support the previous findings that Z. tritici can cause infection in incompatible interactions, but at a much lower level than in compatible interactions (Chu et al., 2019; Kumar et al., 2015; Milgate et al., 2023; Oliver et al., 2008; Ozdemir et al., 2020; Ware et al., 2003). Even though such a low level of fungal infection may not be significant for disease epidemiology, it could still help maintain genetic diversity within populations. For instance, if incompatible strains can fertilize compatible members of the opposite mating type, it could initiate the sexual cycle.

A PCR-based method for quantifying Z. tritici DNA in plants is highly desirable due to the long latent period of STB disease. Previous studies have used disease severity and pycnidial density to identify cultivars with increased resistance to Z. tritici (Eyal et al., 1987; Goodwin, 2007; Brown et al., 2015; Ors et al., 2018). Evaluating genetic materials for resistance to Z. tritici under greenhouse and field conditions requires visual assessment, which usually takes 24 to 28 days after inoculation. In addition to the genetic component, the environment during the time of symptom expression, variation among evaluators, and use of an integer scale for disease rating (Gaunt et al., 1986; Rosielle and Brown, 1979) also can affect the accuracy of disease scoring and complicate the analysis and interpretation of data. On the other hand, using a specific ß-tubulin-2 primer set in real-time PCR at 12 DAI provides a less subjective estimate of disease severity, making it a more reliable method. However, differences in fungal DNA between resistant and susceptible cultivars at earlier time points were minor. Therefore, the real-time PCR approach, in its current form, is best suited for use in a secondary role in a breeding program to minimize the effect of environmental variations on disease expression and to provide an unbiased estimator of quantitative resistance. This study analyzed two resistant wheat cultivars, Tadinia and W7984, which possess specific resistance genes Stb4 and Stb8, respectively. These genes have been mapped to chromosomes 7D (Adhikari et al., 2004a) and 7BL (Adhikari et al., 2003) and probably were contributed by wheat’s B- and D-genome ancestors. Despite their separate evolutionary origins, both had a similar effect on fungal DNA accumulation during pathogenesis. This likely extends to most other STB resistance genes. Similar strong correlations between visual estimates of disease severity and DNA quantity also were shown for Z. tritici in bread (Gouache et al., 2009) and durum (Tonti et al., 2019) wheat.

Real-time PCR was used to distinguish between wheat cultivars and correctly segregated RILs into resistant and susceptible classes based on disease symptoms. The cultivars and RILs rated as susceptible to Z. tritici had more fungal DNA in their leaves than the resistant ones; the lowest fungal DNA was detected in the resistant cultivars and RILs. The Spearman rank correlations between pycnidial density and fungal DNA were positive and significant for both susceptible cultivars. However, four RILs and Yecora Rojo had higher fungal DNA than the resistant RILs, but the difference was insignificant. It is unclear why this discrepancy occurred, but it could be due to biological or operational factors. One possible explanation is that the plants with lower-than-expected fungal DNA received less inoculum, possibly because they grew more slowly than the other lines tested.

The pycnidial density scores were calculated based on several years of testing of adult plants, which were scored at 24 to 28 DAI (Adhikari et al., 2003, 2004). Plant samples for real-time PCR analysis were collected at 12 DAI, the earliest point at which significant differences were observed before the exponential increase in biomass seen in the time-course experiment. Since the plants were collected for DNA analysis before symptom expression, it was impossible to know whether they would have shown the same pycnidial densities as the same lines when tested as adult plants scored at 28 DAI. This uncertainty could result in some susceptible RILs being scored as resistant based on real-time PCR results alone. Delaying the sample collection by one or two days could eliminate this problem. However, the opposite error of resistant plants falsely scored as susceptible did not occur in our tests. Our results were consistent with those obtained with other previously described methods to quantify fungal growth in wheat cultivars (Chu et al., 2019; Cohen and Eyal, 1993; Keon et al., 2007; Oliver et al., 2008; Pnini-Cohen et al., 2000; Stergiopoulos et al., 2003). Highly positive and significant correlations between fungal DNA and disease severity also were observed in alfalfa to Aphanomyces euteiches (Vandemark et al., 2002), in wheat to Fusarium spp (Burlakoti et al., 2007; Kumar et al., 2015; Milgate et al., 2023; Ozdemir et al., 2020), and with several pathogens of Arabidopsis (Brouwer et al., 2003; Gachon and Saindrenan, 2004).

Our analyses revealed that there were more noticeable differences in fungal DNA when the tests were conducted on adult plants instead of at seedling stages. However, even the seedling tests provided good discrimination. One explanation for this result may be that adult plants provide a better substrate for fungal growth, and thus, more significant differences in fungal DNA yield. Another reason may be how the plants were harvested. In the seedling tests, the entire small plant was clipped off at the soil level, so the samples included the stems that typically do not show symptoms, plus leaf tissue that grew after the plant was inoculated. These may have decreased the amounts of fungal DNA relative to the total samples. In contrast, only the leaves were harvested from the much larger adult plants, and no additional leaves grew after inoculation, so these samples included only tissue that was inoculated. Excluding stems and the newest leaves from the seedling samples might better represent the actual level of fungal DNA and provide a more accurate separation between resistant and susceptible plants. Inoculum concentration also was critical, as no differences were apparent at the lower inoculum concentration by 12 DAI. Symptoms did appear on these plants, but they were delayed by several days relative to those inoculated with more spores. Therefore, delaying the assay by two or three days or increasing the concentration of the inoculum beyond the highest level tested in our experiments may lead to increased discrimination between resistant and susceptible wheat lines.

Although the quantitative real-time PCR technique should have broad applicability for analyzing resistance against Z. tritici in wheat, there are a few potential caveats. One situation where the qPCR approach might fail is with QTL of minor effect. The mechanisms of quantitative resistance might vary among loci, and those with small effects could cause too little of a reduction in fungal biomass to be detected. The two resistant and two susceptible lines tested each showed very similar responses, but more lines should be tested to be certain that this pattern is general. No durum wheat cultivars were tested, but they are likely to show the same responses since their main difference from bread wheat is the addition of the D genome (where the Stb4 gene is located) in the latter. The Stb8 gene is on wheat chromosome 7BL (Adhikari et al., 2003), one of the genomes that is shared with durum wheat, so it seems likely that this technique also will work well with durum wheat cultivars. The validation involved some experiments in which resistance was scored based on percent of infected, necrotic leaf area and others where it was based on pycnidia percentage, the two most common methods of scoring resistance against STB. Therefore, it seems likely that qPCR will work on other cultivars with different genes and with differing measures of resistance.

However, quantitative real-time PCR may have trouble if wheat lines are tolerant to STB, i.e., can be infected by the pathogen without showing disease symptoms (Jeger, 2023; Schafer, 1971) or by causing no loss of fitness, which is measured in agricultural systems by yield (Caldwell et al., 1958). Mikaberidze and McDonald (2020) recently identified what they considered to be tolerance in the Z. tritici – wheat pathosystem. However, unlike the classical definition of tolerance in plant pathology where there is pathogen proliferation and disease without yield loss (Caldwell et al., 1958), tolerance was defined instead as lines showing extensive necrosis without pycnidia production (Mikaberidze and McDonald, 2020) and effects on yield or other components of host fitness were not measured. Fungal biomass also was not estimated so it is not known whether the lines identified as tolerant according to their criteria allowed multiplication of the pathogen. The qPCR approach likely would identify lines with classical tolerance as susceptible if they allowed the pathogen to grow without producing symptoms (Schafer, 1971), or if the fungus grew and produced symptoms without affecting yield (Caldwell et al., 1958). Extensive necrosis without pycnidia production and with little fungal biomass could occur, for example, if a wheat cultivar was highly susceptible to a toxin produced by the fungus, which could diffuse through the leaves to kill tissue ahead of pathogen invasion. Under that situation the qPCR technique would likely identify tolerant cultivars as resistant. It would be very interesting to test the lines identified as tolerant by Mikaberidze and McDonald (2020) to quantify their amounts of Z. tritici biomass relative to those that were scored as resistant.

The real-time PCR assay also should be tested with different pathogen isolates. Variation in virulence (Eyal et al., 1985; Kema et al., 1996a) and specific adaptation by Z. tritici to vertically resistant wheat cultivars (Cowger et al., 2000; Jackson et al., 2000; Ababa and Mekonnen, 2024) may affect the fungal DNA produced by particular cultivar-isolate combinations. Accordingly, it is necessary to investigate the amount of fungal DNA in planta that is proportional to the development of disease symptoms induced by different virulent isolates or pathotypes of Z. tritici. A potential approach to enhance the technique could be to use a plant gene as a reference for normalization and endogenous control (Valsesia et al., 2005; Winton et al., 2002). In this case, fungal DNA would be estimated as a percentage of the total amplifiable DNA in each sample. This would control for pipetting errors, inaccurate estimates of DNA concentration, and sample-to-sample variance in the efficiency of PCR amplification (Winton et al., 2002).

However, none of these are likely to have affected the current results. For example, the low level of variability among technical replicates ( Figure 3 ) indicates that pipetting errors were infrequent. Variation among biological replicates was higher than for technical replicates but was still reasonably low ( Figures 4 , 6 ), so estimates of template DNA concentrations should have been accurate. Differences in amplification efficiency among samples were not estimated. Still, we believe they were small due to the relatively low variability among biological replicates and the smooth curves seen in the time courses. If amplification efficiencies varied among DNA samples, we would expect the points on the time-course experiments to be erratic. Instead, the points were consistent, except possibly for the 24-hour sample of Yecora Rojo. Although the results obtained from the current experiments were consistent and repeatable, including an endogenous control is an improvement that should be incorporated into future applications of this approach for estimating the quantity of Z. tritici DNA in infected wheat leaves.

In conclusion, this approach provided a rapid, robust, quantitative method to monitor the growth dynamics of Z. tritici during infection of wheat. The real-time PCR assay was more accurate and suitable for high-throughput analysis than other techniques, such as competitive PCR, which did not always correlate well with Fusarium head blight symptom expression (Nicholson et al., 1998). So far, at least 22 Stb genes for resistance to Z. tritici have been identified (Tidd et al., 2023). This real-time PCR assay can be used to evaluate segregating mapping populations for qualitative resistance governed by Stb genes and quantitative trait loci (QTL) involved in STB resistance (Adhikari et al., 2003; 2004a; 2004b; 2004c; Arraiano et al., 2001b; Brading et al., 2002; Goodwin, 2007; Goodwin et al., 2015; Liu et al., 2013; McCartney et al., 2002; Tidd et al., 2023; Yang et al., 2018). In addition to the known Stb genes many QTL have been identified, which also might reduce fungal DNA and could provide a more stable and long-lasting resistance in the field. Other analyses have shown that real-time PCR was more accurate and thorough for dissecting QTL for resistance to Puccinia coronata f. sp. avenae (crown rust) in the Ogle/TAM O-301 (OT) mapping population of oat compared to other assessment methods (Jackson et al., 2006, 2007). More recently, regression analysis and real-time PCR assays were utilized in multi-species winter wheat field trials in Australia to assess partial resistance and to estimate genotype yield in the presence of Fusarium crown rot caused by F. pseuodograminearum (Milgate et al., 2023). The real-time PCR method also was valuable for detecting significant differential expression patterns of defense-related genes between resistant and susceptible cultivars and between the resistant and susceptible RILs segregating for the Stb4 and Stb8 genes for resistance (Adhikari et al., 2007). Overall, the real-time PCR method developed in this study is precise, specific, cost effective, and repeatable and can be a valuable tool for making decisions about managing STB disease, to identify resistant progeny in breeding programs, for investigating disease epidemiology and plant-pathogen interactions, and in assessing disease risks. Additionally, it may provide new perspectives for developing durable resistance to Z. tritici in wheat and thus could help reduce the threat of this pathogen to food security.