Effect of Low Temperature and Wheat Winter-Hardiness on Survival of Puccinia striiformis f. sp. tritici under Controlled Conditions

Wheat stripe rust, caused by Puccinia striiformis f. sp. tritici (Pst), is one of the most important diseases of wheat worldwide. Understanding the survival of Pst during the overwintering period is critical for predicting Pst epidemics in the spring. Real-time quantitative PCR (qPCR) methods quantifying Pst DNA and RNA (cDNA) were developed and compared for the ability to quantify viable Pst in leaf tissues. Both qPCR of DNA and RNA can provide reliable measurement of viable Pst in plant tissues prior to the late sporulation stage for which qPCR of DNA gave a much higher estimate of fungal biomass than qPCR of RNA. The percentage of Pst biomass that was viable in detached and attached leaves under low temperatures decreased over time. Pst survived longer on attached leaves than on detached leaves. The survival of Pst in cultivars with strong winter-hardiness at 0°C and -5°C was greater than those with weak winter-hardiness. However, such differences in Pst survival among cultivars were negligible at -10, -15 and -20°C. Results indicated that Pst mycelia inside green leaves can also be killed by low temperatures rather than through death of green leaves under low temperatures. The relationship of Pst survival in attached leaves with temperature and winter-hardiness was well described by logistic models. Further field evaluation is necessary to assess whether inclusion of other factors such as moisture and snow cover could improve the model performance in predicting Pst overwintering potential, and hence the epidemic in spring.


Introduction
Wheat stripe rust, caused by Puccinia striiformis Westend. f. sp. tritici Erikss. (Pst), is one of the most important diseases of wheat worldwide [1,2]. Annual cycles of stripe rust on winter wheat can be divided into four stages: oversummering, infection of seedlings in autumn, overwintering, and spring epidemic. As an obligate parasite, Pst has to survival under unfavorable weather in living host tissues [3,4], and therefore oversummering and overwintering are two bottlenecks for Pst epidemic development.
Temperature is an important factor influencing Pst development, including its survival during the winter [1,[5][6][7][8][9][10][11][12][13][14][15][16]. The low limit of air temperature for Pst overwintering (November to February in China and North America) is believed to be around -6 to -7°C (monthly mean) [3,6,11,[17][18][19] and could be as low as -10°C when wheat plants are under snow [3,5,17]. Pst development might halt when temperature is below -10°C [8]. Wheat-growing regions in China are divided into overwintering and non-overwintering regions for Pst based on winter temperatures [17,20]. Winter-hardiness of wheat cultivars and atmospheric moisture levels may also affect Pst overwintering [3,21]. Previous findings regarding the effect of environmental factors on Pst overwintering were based field observations, which reflect the overall effects of many external factors on Pst overwintering.
Studying Pst overwintering is hampered by the inability of assessing the amount of fungal biomass and its viability in wheat leaves efficiently. Recent development in PCR technology has enabled rapid detection and quantification of pathogens in plant tissues. The amount of Pst in leaves in early stages of infection prior to visual symptoms can be quantified using real-time quantitative PCR (qPCR) of DNA [22]. RNA is expected to be a better choice over DNA as an indicator of metabolic changes in response to external factors, hence indicating the viability. Quantification of RNA has recently been developed for plant pathogens including Phytophthora parasitica [23], Pectobacterium atrosepticum [24], Puccinia triticina [25][26], P. striiformis [25,27], Verticillium dahliae [28], and V. longisporum [28].
The current study was to 1) study the stability of DNA and RNA in Pst urediniospores stored under different conditions, 2) develop qPCR methods to quantify Pst DNA and RNA (cDNA) in plant tissues, and evaluate these two methods for quantifying Pst inside leaves at different stages of its development, 3) study Pst survival under several constant low temperatures using detached and attached leaves of one cultivar and 4) investigate the effect of cultivar winter-hardiness on Pst survival at low temperatures.

Fungal strains
A single Pst strain (CYR32-Chinese Yellow Rust 32, originally from Longnan, Gansu province, China) was used to develop qPCR methods. In addition, several other common wheat fungal pathogens were used to validate primer specificity, including P. triticina (from Hebei Agricultural University, China), P. graminis f. sp. tritici (from Shenyang Agricultural University, China), Fusarium graminearum, Erysiphe graminis f. sp. tritici, and Bipolaris sorokiniana (from Northwest A&F University, China).

Pst multiplication and inoculation
CYR32 was multiplied in Taibai Observation Experimental Station, Shaanxi, China. Fresh urediniospores of CYR32 were inoculated onto seedlings of wheat cultivar Mingxian 169, a cultivar susceptible to all known Pst races. Seeds were sown in pots of 10-cm in diameter and 12-cm in depth, containing cow dung compost and soil (1:2, V/V). Ten days after sowing, seedlings with two leaves were inoculated with a fresh urediniospore suspension (50 mg urediniospoes/mL in deionized distilled water) of Pst CYR32 [29]. A fine paintbrush was used to daub spore suspension onto the adaxial surface of leaves until the whole leaf surface became wet without run-off [21,29]. The seedlings were kept in a dew chamber at 8-10°C in dark for 24 h immediately after inoculation and then grown on benches in a greenhouse at 14-17°C with a photoperiod of 10-14 h [29]. The plants in each pot were separated from plants in other pots with a transparent plastic cylinder to prevent contamination.

DNA extraction
Wheat leaves (100 mg) were quickly frozen and transferred into a mortar pre-cooled with liquid nitrogen. Sixty microliters of Tris-HCl (pH 8.0) was added to the mortar and ground, and the resulting powder was transferred into a 1.5 mL eppendorf tube. Urediniospores were used as standards in qPCR; 10 mg urediniospores were put into a 1.5 mL eppendorf tube. Thirty microliters of Tris-HCl buffer (pH 8.0) was added into the tube. The tube was incubated for 5 min on ice to let Tris-HCl buffer permeate into urediniospores and was quickly frozen in liquid nitrogen. The urediniospores-Tris-HCl ice lump (S1 Fig) was transferred to a mortar precooled with liquid nitrogen and ground, and the resulting powder was transferred into a new 1.5 mL centrifuge tube.
Once the powder was obtained, the following steps were used to extract DNA: 1) add 600 μL CTAB buffer (heated at 65°C before use) to the 1.5 mL eppendorf tube, vortex for 2 min and incubate for 40 min in a water bath at 65°C (invert the tube every 5 min); 2) incubate for 5 min on ice; 3) add 600 μL phenol: chloroform: isoamyl alcohol (25:24:1), and shake for 20 min; 4) centrifuge for 10 min at 15255× g at room temperature; 5) collect the aqueous phase and re-extract with 500 μL chloroform: isoamyl alcohol (24:1), and shake for 20 min; 6) repeat step 4; 7) collect the aqueous phase, add 50 μL of 3 M sodium acetate (pH 5.2) and 500 μL of isopropyl alcohol, and incubate for 30 min at -20°C; 8) centrifuge for 10 min at 15255× g at 4°C, wash the pellet with 75% ethanol and then with absolute ethanol, air-dry and dissolve it into 20 μL TE buffer (pH 8.0); 9) add 1.2 μL RNase (1 mg/mL), and incubate for 30 min at 37°C to remove RNA; and 10) measure DNA concentration using Nano Drop RND-2000 spectrophotometer (Thermo, USA), check DNA integrity through electrophoresis run in 1.0% agarose gel and visualize under AlphaImager System (CA, USA) after staining in an ethidium bromide solution (0.5 μg/mL) for 15 min.
cDNA synthesis cDNA was obtained through reverse transcription of RNA using PrimeScript RT Reagent Kit (Perfect Real Time) (TaKaRa, Japan). RNA was reversely transcribed at 37°C for 15 min, followed by heating at 85°C for 5 s. All steps followed the manufacturer's instructions. Real-time quantitative PCR qPCR reactions were performed using an UltraSYBR Mixture kit (CWBIO, China). Each reaction contained 12.5 μL Ultra SYBR Mixture, 1.0 μL primer mix (containing 10 μM of each primer), 2 μL of cDNA or DNA (100 ng RNA and 140 ng DNA), and RNase free water up to 25 μL. The amplification conditions were as following: a denaturation step at 95°C for 3 min followed by 40 amplification cycles consisting of denaturation at 95°C for 30 s, annealing at 60°C for 30 s, and extension at 72°C for 30 s, with an amplification melting curve at 95°C for 1 min, 55°C for 1 min and followed by 81 cycles of 20 s from 55°C to 95°C with a step of 0.5°C. It was conducted on a real-time Thermal Cycler iQ5 (Bio-Rad, USA). The primer pair of Pst EF1 was used. The iQ5 software 2.1 was used to estimate Cq value and PCR efficiency at the default settings. The mean Cq value and amplification efficiency (%) were calculated from three technical replicates for all individual samples.

Stability of DNA and RNA in Pst urediniospores
DNA and RNA were quantified from three types (treatments) of CYR32 urediniospores over time: 1) dead urediniospores at ambient temperature, 2) fresh urediniospores at ambient temperature and 3) fresh urediniospores stored at -20°C. DNA and RNA were quantified at day 0, 4, 8, and 12. For each type of urediniospores, 5 mg of urediniospores were weighed and kept in a single tube. To kill the urediniospores, tubes with urediniospores were placed in a water bath at 60°C for 1 h. The effect of this water bath treatment on the viability of urediniospores was confirmed on 2.5% water agar (S3 Fig) as described previously [31]. On each of the four sampling events, DNA and RNA were extracted from urediniospores subjected to three treatments and measured using both spectrophotometer Nano-Drop RND-2000 (Thermo Scientific, USA) and qPCR. For each treatment at each sampling time, there were three replicates.

Comparison of qPCR methods for quantifying Pst in inoculated leaves
Ten-day-old seedlings of cv. Mingxian 169 were inoculated with CYR32 urediniospores as described above. Leaf samples were collected 1 day before inoculation, and 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 20, and 30 days post inoculation. On each sampling time point, six first-leaves were randomly collected. As the objective was to quantify Pst biomass inside leaves, surface urediniospores were washed with sterile distilled water three times before DNA or RNA extraction. Three of the six leaves were used for RNA extraction and the others for DNA extraction. Both DNA and RNA were quantified using qPCR as described above. The experiment was done three times.

Pst survival in detached and attached leaves under low temperature
General experimental protocols are given in Table 1. Urediniospores were inoculated onto seedlings for use in all experiments, and the inoculation procedures were the same as described above. Initially, the survival of Pst in detached and attached leaves was studied only on Mingxian 169.
For detached leaf assay, Mingxian 169 was grown in ten pots (10 × 12 cm per pot) with 15-20 seedlings per pot. Ten days after sowing, seedlings with two leaves were inoculated with suspension of fresh urediniospores as described above. When symptoms of stripe rust were first observed (ca. 12 days after inoculation) (S4 Fig), 30 leaves with symptoms were randomly Pst survival in detached leaves collected from seedlings. Individual leaves were placed separately into 2.0 mL centrifuge tubes. These tubes were placed into a low-temperature-freezing cabinet THD-3010 (Tianheng, China), set to one of the five temperatures (0, -5, -10, -15, and -20°C). Three leaves (tubes) were randomly taken out from the cabinet for qPCR RNA assay at 0 (at the time of sampling), 6,12,18,24,30,36,42,48 and 54 h at each temperature ( Table 1). The experiment was done three times. For attached leaf assay, the experiment was conducted in the same way as the detached leaf experiment except that seedlings with initial Pst symptoms were incubated in a low-temperature-freezing cabinet THD-3010 (Tianheng, China). Wheat leaves were cut off at four time points depending on the temperature; the exact sampling time was determined based on the results of the detached leaf assay (Table 1). This experiment was done three times.
To assess whether seedlings were still alive after incubation for 30 h at -15°C and -20°C, 48 h at -10°C, 54 h at 0°C and -5°C, the seedlings were moved to 4°C for 1 day, and then kept in a greenhouse (14-17°C with a photoperiod of 10-14 h) for one week before assessment for plant growth.

Winter-hardiness evaluation of wheat cultivars
Seeds of wheat cultivars Mingxian 169, Xiaoyan 22, 01-314, Xinong 979, Lantian 15 and Jingnong 411 were sown in pots of 10-cm in diameter and 12-cm in depth, containing cow dung compost and soil (1:2, V/V). Twelve days after sowing, seedlings with three leaves were treated in low-temperature-freezing cabinet THD-3010 (Tianheng, China), set to one of the five temperatures (0, -5, -10, -15, and -20°C) for 36 h, respectively. The seedlings were kept at 4°C for 24 h after treatment and then moved to an artificial climate incubator with temperature of 18/ 23°C, photo period of 16/8 h, relative humidity of 60-80% and light intensity of 8000-10000 Lx for 15 d. The mortality of wheat leaves was calculated. The winter-hardiness was expressed using the temperature which causes 50% mortality of wheat leaves (LT50) [32].

Pst survival in six cultivars under low temperatures
Pst survival under low temperatures was further studied on six wheat cultivars covering three levels of winter-hardiness: Mingxian 169 and Xiaoyan 22 (low winter-hardiness), Xinong 979 and 01-314 (moderate winter-hardiness) and Lantian 15 and Jingnong 411 (high winter-hardiness) ( Table 2). Two pots of seedlings (15-20 seedlings per pot) with initial Pst symptoms of each cultivar were treated in the same way as described above for the attached leaf assay except that sampling times were 0, 12, 18 and 24 h at -15 and -20°C (Table 1). This experiment was done three times.

Data analysis
To compare quantification of Pst DNA and RNA in three types of urediniospores or in inoculated wheat leaves, a repeated measurement ANOVA was carried out in which the DNA and RNA were the main treatment factor. Repeated measurement ANOVA is necessary because of temporal correlation between assessments made over time on the same treatments. A post-hoc test for multiple treatment comparisons was carried with the Student-Newman-Keuls test using the pooled residual error from the ANOVA whenever significant treatment or interaction effects were revealed by ANOVA. For the comparisons on the inoculated wheat leaves, two separate ANOVAs were carried out: the first included only samples from 0 to 11 days after inoculation and the second included all time points. To reduce variance heterogeneity, data were logarithmically transformed. The replicate experiments were used as a block factor in ANOVA.
Viable Pst biomass in leaves was estimated as the amount of urediniospore based on qPCR of RNA (cDNA). The concentration of viable Pst in leaves was calculated as the ratio between quantified Pst RNA and leaf weight. The percentage of Pst (PPS) that survived a specific low temperature treatment at time x was calculated as: where A 0 and A x are Pst concentrations in leaves at zero and x hour, respectively. Repeated measurement ANOVA and the Student-Newman-Keuls test were used to assess the Pst survival on detached or attached leaves under different temperatures over time. To compare Pst survival on the detached and attached leaves, the leaf status (detached or attached) was treated as a factor in the repeated measurement ANOVA. In this analysis, only common sampling times between the attached and detached leaf experiments were included. Similarly, repeated measurement ANOVA and the Student-Newman-Keuls test were used to assess the effect of cultivar winter-hardiness and temperature on Pst survival on attached leaves over time. Three levels of winter hardiness were used in the ANOVA: low, moderate or high winterhardiness. In all ANOVAs, replicate experiments were used as a block factor.
A key objective of the present study was to model the temporal dynamics of Pst survival at low temperatures, taking into account cultivar winter-hardiness. To study Pst survival over time on attached leaves at a given temperature, a two-parameter logistic model was fitted to each of 30 combinations of temperature and cultivar: where r and m are the rate of Pst mortality and time to 50% mortality, respectively. Average PPS values over three replicate experiments were used to fit this logistic model to each combination of temperature and cultivar. This model assumes a 100% viable population at time zero. Finally, another logistic model was used to relate r or m to temperature in which winter-hardiness of cultivars was treated as a factor in fitting the logistic model.
All data analysis was carried out using software R (version 2.11.1).

Stability of DNA and RNA in Pst urediniospores
There were significant differences between DNA and RNA concentrations in three storage conditions (P < 0.001). The difference resulted primarily from large variation in RNA concentration over time (Fig 1). Compared with RNA, DNA concentration remained relatively stable over time (Fig 1). DNA integrity did not change much with time for all conditions (Fig 2A-2D). The 28S, 18S, 5.8S and 5S RNA were intact on day 0 (Fig 2E), but considerably degraded on day 4 for dead urediniospores (Fig 2F lane 1) with both 28S and 18S almost totally degraded on day 8 (Fig 2G lane 1). RNA extracted from viable urediniospores showed slight degradation on both day 4 and 8 ( Fig 2F-2G lane 2). On day 12, RNA from viable urediniospores kept at ambient temperatures degraded more severely than that from urediniospores stored at -20°C ( Fig 2H lane 2-3). The bands for 28S and 18S were clear for urediniospores stored at -20°C on all assessment days (Fig 2E-2H lane 3). The amplified fragment size was 248 bp and 159 bp from DNA and cDNA of Pst using EF1 primer pair, respectively (S2 Fig). Standard curves based on RNA (cDNA) and DNA were directly estimated using software iQ5 (Fig 3); the coefficient of determination (R 2 ) for standard curves was greater than 0.98 in all quantification runs. Amplification efficiency ranged from 95.1% to 101.2% for RNA (cDNA) and from 97.3% to 99.7% for DNA. Five mg of Pst urediniospores under three storage conditions for 0-12 days were assayed using the qPCR of DNA and RNA (cDNA) (Fig 4). The amount of quantified Pst significantly differed (P < 0.01) between the two methods [qPCR of DNA and RNA (cDNA)]. However, both methods could not differentiate dead, fresh and dormant urediniospores except for a much lower amount of quantified Pst by qPCR of RNA for dead urediniospores at the ambient temperature (Fig 4).   Quantification of Pst in leaves with qPCR of DNA and RNA Initially, quantified DNA and RNA followed a similar trend-increasing gradually over time and reaching a peak around day 10 ( Fig 5). The overall level of DNA was greater (P < 0.05) than RNA from day 7 onwards. The greater increase in DNA from day 7 to day 10, relative to RNA, led to significant (P < 0.01) interactions between DNA-RNA differences and time (0-11 days). On day 20 and 30, RNA concentration decreased to nearly zero, in contrast to the sharp increase in DNA concentration (Fig 5). This increase in DNA and decrease in RNA on day 20 and 30 led to an even greater difference between DNA and RNA (P < 0.01) and also the interaction between DNA-RNA difference and time (P < 0.001) when all data were analyzed together. On day 20, more than 50% leaf area senesced (

Pst survival in detached and attached leaves under constant temperatures
Overall, the percentage of Pst that survived in detached and attached leaves decreased (P < 0.001) over time at all test temperatures (Fig 6A-6B). Pst survival differed greatly (P < 0.001) between attached and detached leaves. A greater (P < 0.001) percentage of Pst survived in attached leaves than in detached leaves after prolonged incubation: at 0°C during 42-54 h, at -5°C during 42-48 h, at -10°C during 36-42 h and at -15°C and -20°C after 18 h. This difference was most pronounced at 0°C (Fig 6B). Viable Pst was still detected after incubation at 0°C and -5°C for 54 h in both detached and attached leaves. The percentage of Pst that survived was 10.1% and 4.0% at 0°C and -5°C, respectively after incubation for 54 h in detached leaves; the corresponding values in attached leaves were 25.8% and 5.5%, respectively. The amount of viable Pst was below the limit of detection after incubation at -10°C for 48 h, at  Quantified Puccinia striiformis f. sp. tritici in inoculated wheat leaves (cv. Mingxian 169) from day 0 to day 30 after inoculation using qPCR of DNA and RNA (cDNA). Urediniospores on the leaf surface were removed through washing leaves with distilled water. On day 20, more than 50% leaf area senesced, and on day 30, the leaves were completely senesced. The experiment was done three times. The vertical bar of a mean value represents the standard deviation among the three repeats of the experiment; significant treatment differences were based on the pooled residual error in the repeated measurement ANOVA. The level of quantified Pst DNA with "**" on 8, 9, 10, 20 and 30 days was significantly different from that based on the RNA quantification at P = 0.01. There were no significant (P > 0.05) differences between quantified levels of Pst DNA and RNA on other days.

Pst survival in six cultivars under low temperatures
At 0°C and -5°C, the percentage of viable Pst increased with increasing cultivar winter-hardiness (Fig 7A-7B). However, such differences between cultivars were very small at other test temperatures (Fig 7C-7E). The temporal dynamics of viable Pst at a given temperature was well described by a logistic model with two parameters: rate of Pst mortality (r) and time to 50% mortality (m). This model explained more than 99% of the total variability in average viable Pst over time for all 30 datasets (five temperatures × six cultivars). Both r and m were related to temperature though the relationship of m with temperature ( Fig 8A) was stronger than that of r with temperature ( Fig 8B). Overall, m decreased with decreasing temperature; differences in m among cultivars were primarily at 0°C and -5°C. In contrast, there were large differences in r across all test temperatures; however, there appeared to be no discernable pattern of r in relation to either of cultivar or winter hardiness.
The relationship between m and temperature was well described by a logistic model with four parameters (Fig 8A):

Discussion
This is the first published study that quantifies viable Pst in wheat leaves using qPCR of RNA. RNA is usually more easily degraded than DNA. Quantified DNA concentration based on qPCR may remain stable for longer than the corresponding estimated RNA. However, whether qPCR quantification of RNA is better than the corresponding qPCR of DNA as an indicator of viable fungal structure depends on the type of fungal structure. For urediniospores, both qPCR of RNA and DNA cannot distinguish dormant from dead spores. In contrast, qPCR of RNA, but not DNA, can determine whether hyphae inside leaves are viable. Two possible reasons may explain why qPCR of RNA cannot distinguish dead urediniospores from viable ones. First, RNA is closely linked to protein synthesis (hence concentration), which varies considerably according to the organism's metabolic needs [33]. In the present study, viable spores were not maintained under conditions that are conducive to spore germination since the objective was to assess the detection of viable (yet non-germinating spores) using qPCR methods. The overall level of metabolism in non-germinating spores may be below the detection level of the present qPCR method. Second, the present qPCR method only quantifies mRNA for a particular gene-elongation factor 1, which relates to the production of cytoskeleton [34][35][36]. In non-germinating spores, metabolic activity related to cytoskeleton production may be very low and cannot be detected by the present qPCR method. Further research is needed to determine whether viable non-germinating spores could be distinguished from dead ones based on qPCR quantification of other genes.
There were no significant differences between the two qPCR methods (RNA and DNA) in the amount of quantified Pst in inoculated wheat leaves from 0 to 11 days after inoculation. However, qPCR of DNA resulted in a much greater estimate of Pst than qPCR of RNA post sporulation (12 to 30 days). Usually, DNA can remain intact for a longer time after cell death than RNA; therefore, qPCR of DNA may overestimate the amount of Pst because of dead urediniospores or mycelia. Thus, both qPCR of RNA and DNA can be used to estimate Pst amount during the initial latent period but only qPCR of RNA should be used post sporulation to estimate viable Pst.
The percentage of Pst that is viable in detached and attached leaves decreased over time under low temperatures. Low temperature is believed to reduce pathogen survival in infected leaves [1]. As expected, Pst survival is greatly reduced when temperature is below -10°C. However, some reports claimed that Pst can survive as long as the infected foliage remains green during winter [3][4]. The present result indicated that Pst in attached green leaves did not survive at -20°C after 30 h, however the plants were still alive at this condition. Saulescu and Braun [37] also reported that plants may remain alive even up to 48 h at -18 to -20°C. Under the condition of this study, we conclude that Pst mycelia inside green leaves can be killed by low temperature and that the decrease in viable Pst is not all due to death of leaf tissues. It should be noted that seedlings were transferred to low temperatures without adequate acclimatization time in the present study. Wheat plants that have grown to the tillering stage and been slowly acclimated to low temperatures may survive better, which may also influence Pst survival. Thus, it would be interesting to study whether wheat plant age and acclimatization time could affect Pst survival.
Winter-hardiness of wheat cultivars may affect Pst overwintering [3,21]. The present results suggested that the percentage of viable Pst was greater on cultivars with strong winter-hardiness than those with weak winter-hardiness mainly at 0°C and -5°C. Under the conditions of this study, both temperature and cultivar winter-hardiness affect Pst survival in winter, but temperature has more impact on determining the Pst survival.
Many models have been developed to describe the relationship between winter temperature and stripe rust epidemics based on field observations [6,7,11,38] and were further revised [16] to predict the spring epidemic. For the first time, we have developed models relating Pst mortality to temperature and cultivar winter-hardiness. These models predict the shortest and longest Pst survival on cultivars of low and high winter-hardiness, respectively, at a particular temperature. It should be noted that we assumed that there is no mortality when temperature > 0°C which was shown previously [5].
In summary, we have developed a method based on qPCR of RNA to quantify viable Pst biomass in leaf tissues. This qPCR-RNA method can be used in detecting and quantifying the pathogen prior to symptoms, which may be used as an input for disease prediction. A mathematical model was derived to relate Pst survival to temperature and cultivar winter-hardiness, which can be used predict Pst survival under fluctuating conditions. However, further data on field Pst survival are needed to assess whether inclusion of other factors (e.g. moisture and snow cover) could improve model predictions of Pst survival over the winter. For this purpose, we are sampling wheat leaves in early spring from plants showing symptoms in the autumn at many locations to quantify viable Pst biomass. These data will be used to evaluate and, if necessary, improve this temperature-only model. Availability of such a predictive model will be valuable to assist growers in timing fungicide applications in given geographical regions. The vertical bar of each mean value represents the standard deviation of the three mean values of the three experiments; significant treatment differences were based on the pooled residual error in the repeated measurement ANOVA. The group of cultivar with similar winter-hardiness with different letters at the same temperature differed significantly at P = 0.05. (TIFF)