Bacterial canker, caused by Pseudomonas syringae pv. actinidiae on kiwifruit (Actinidia spp.), is one of the major diseases that occasionally causes severe damage in orchards and extensive economic losses in Japan (Serizawa et al., 1989; Takikawa et al., 1989), Korea (Ko et al., 2002; Koh, 1995; Koh et al., 1994, 2012), Italy (Balestra et al., 2009; Scortichini, 1994), and many other countries (Balestra et al., 2010, 2011; Bastas and Karakaya, 2012; Everett et al., 2011; Vanneste et al., 2011b). In Korea, bacterial canker on kiwifruit was first observed at Jeju in 1988, and has rapidly spread since (Koh et al., 1994; Lee et al., 2005). This pathogenic bacterium has been isolated from both commercial (A. deliciosa and A. chinensis) and wild (A. arguta and A. kolomikta) (McCann et al., 2013; Scortichini et al., 2012) kiwifruits. This bacterium comprises four distinct clades (McCann et al., 2013). Three clades correspond to their geographical origins of isolation, which are Japan, Korea, and New Zealand. The fourth clade is responsible for the recent bacterial canker outbreak in Italy and New Zealand, and is now distributed globally. Pathogenic bacteria in the fourth clade have also been observed in Korea (Koh et al., 2012).

Bacterial canker disease symptoms on kiwifruit include brown-black leaf spots, often surrounded by yellowish halos, blossom necrosis, die-back or blight of twigs, reddening of lenticels, and bleeding cankers on stems, with whitish to red bacterial ooze (Scortichini et al., 2012). The bleeding cankers on kiwifruit stems begin to appear in February or March, and reach their maximum incidence in late April or May (Ko et al., 2002). Symptoms on kiwifruit leaves begin to appear in early April or May (Ko et al., 2002; Koh et al., 1994). P. syringae pv. actinidiae is found on both symptomatic and asymptomatic tissues, indicating that it can survive as an epiphyte (Vanneste et al., 2011c). Temperatures within the range of 12°C to 18°C and humid conditions could favor multiplication of the bacterium, allowing it to disperse systemically from infected sites to uninfected ones (Scortichini et al., 2012). Temperatures above about 35°C could suppress multiplication and dispersal of the pathogen. Freezing plant tissue during the cold winter could be a major factor contributing to the development of severe bacterial canker symptoms in early spring, although the pathogenic bacterium does not display ice nucleation activity (Ferrante and Scortichini, 2014; Ko et al., 2000a, 2000b; Serizawa et al., 1989).

To manage bacterial canker on kiwifruit, applying streptomycin and streptomycin sulfate·oxytetracycline from mid-April to early May, copper hydroxide from mid-January to early February, and kasugamycin from mid-March to early April controlled the disease in Korea (Koh et al., 1999). Spraying bactericidal compounds, such as streptomycin and copper formulations, is also a highly reliable means of chemically controlling bacterial canker in other countries (Cameron and Sarojini, 2014; Nakajima et al., 2002). Trunk injection of streptomycin has also been used in Korea and has appeared effective in treating infected vines (Koh et al., 1996). However, streptomycinresistant isolates of P. syringae pv. actinidiae were observed in Japan (Han et al., 2003b; Nakajima et al., 1995). Although streptomycin-resistant isolates have not been observed in Korea (Han et al., 2003b), reducing antibiotic (including streptomycin) application is important to preventing antibiotic resistance. Foliar application of copper hydroxide or copper-antibiotic formulated pesticides after mid-April causes severe phytotoxicity (Koh et al., 1996). Consequently, an integrated approach that includes precisely scheduled spray treatments with effective and environmentally friendly bactericides, preventative measures aimed at reducing the bacterial inocula and possible infection pathways, and enhancement of host resistance against the disease seems to be the best solution for coexistence with the disease (Ko et al., 2002; Scortichini et al., 2012). Development of resistant cultivars and biological pesticides including bacteriophages, antimicrobial peptides, and compounds that induce or activate plant defense mechanisms, is in progress (Cameron and Sarojini, 2014; Scortichini et al., 2012).

As there is no effective chemical measure for managing bacterial canker on kiwifruit from May to the harvest season in Korea (Ko et al., 2002; Koh et al., 1996, 1999), growers generally remove canes and trunks with bacterial ooze to prevent dispersion of the pathogen early in the growing season. Cutting off the canes and trunks contributes to overall yield loss. Although the initial infection on the leaves of healthy plants in the spring can lead to death by the end of the year, the symptoms on canes seldom appear from late spring to the harvest season (Koh, 1995; Koh et al., 1994; Scortichini et al., 2012). These observations indicate that infected plants could produce during the year of infection. Infections in a given year rarely cause yield loss during that year. The death or removal of canes and trunks where bacterial ooze occurred early in the growing season could contribute to yield loss caused by the bacterial canker. The disease incidence (DI) of kiwifruit early in the growing season is strongly correlated with the yield loss during that year. Maximum incidence of bacterial canker could be an indirect way to estimate yield loss caused by bacterial canker.

The maximum incidence of bacterial canker on kiwifruit canes generally occurred in late April or early May in Korea (Ko et al., 2002). The pathogen causing bacterial canker can easily transfer to other plants within and between orchards in all seasons, except for the hot summer, and the detectable symptoms on canes and trunks occur after sufficient colonization and multiplication of infected or migrated pathogens (Ko et al., 2000b; Koh, 1995; Koh et al., 1994; Scortichini et al., 2012). Successful colonization and multiplication causes necrotic lesions under the canes’ or trunks’ bark, which cannot be detected from the outside view. Symptoms on kiwifruit stems observed early in the growing season do not always appear at sites where successful pathogen infections occurred during the last growing and overwintering seasons. Some symptoms appear at sites where successful infections occurred more than one year previously. The observed maximum incidence in late April or early May is the result of accumulated damage on kiwifruit stems, caused by colonization and multiplication of the pathogen. Maximum incidence is not always the result of infection by the pathogen during the last growing and overwintering seasons, and would vary in direct proportion to the accumulated damage. We conducted the present study to develop a model to calculate the accumulated damage on kiwifruit canes caused by bacterial canker during the last growing and overwintering seasons; and finally estimate the maximum incidence of bacterial canker on kiwifruit canes at early growing season.

We developed a model to estimate the accumulated potential damage on kiwifruit canes caused by bacterial canker during the last growing and overwintering seasons (D-PSA-K) and determine the maximum incidence of bacterial canker. The D-PSA-K model can be applied based on hourly and daily mean air temperature data.

To test the effect of temperature on bacterial canker development in kiwifruit canes, we used the P. syringae pv. actinidiae KACC16843 isolate. The isolate does not belong to the fourth clade responsible for the recent outbreak in Italy and New Zealand. Isolates belonging to the fourth clade have already been observed in Korea. The mode of pathogenic action of the isolates often collected in Korea is different from the mode of those belonging to the fourth clade (Han et al., 2003a; Koh et al., 2012; Vanneste et al., 2011a). The fourth clade showed significantly lower pathogenicity on Hayward than on the other cultivars, including Hort16A (Balestra et al., 2009). For Hayward, bacterial canker caused by the fourth clade may be rare compared with the disease caused by the Korean clade, including the isolate used in this study. In other cultivars, including Hort16A, the amount of pathogenicity of the fourth clade will need to be tested to develop a more precise model for the cultivars.

We tested the effect of temperature on bacterial canker development in kiwifruit canes using overwintered Hayward canes, but without leaves or roots. If whole plants with leaves and roots had been tested, inoculation using the culture suspension might not have resulted in such long necrotic lesions under the bark as plant growth during the growing season could compensate for the necrosis caused by inoculation. Additional factors, including the diameter and age of canes, could also affect necrotic lesion development on canes. To that end, we did not test the effect of temperature on bacterial canker development on various kiwifruit canes in nature, nor did we compute the equation representing the relationship between temperature and lesion development on kiwifruit canes. The equation developed in this study is only relevant under the assumption that there was no growth and that the diameters of all kiwifruit canes were about 10 mm. Consequently, the resulting equation would describe the relationship between temperature and lesion development on hypothetical kiwifruits, and might not be useful for estimating the lesion progress or the time of symptom appearance on kiwifruit.

The amount of lesion development on hypothetical kiwifruits that do not grow can be considered equal to the potential maximum damage caused by the colonization and multiplication of infected or migrated pathogens. The resulting equation could be used to estimate the amount of potential damage within the temperature ranges that we tested. In addition, a test using whole plants would not be suitable for developing a model, as it is almost impossible to have many canes in whole plants with no significant differences among them. Whereas the longest distances between inoculated parts and exuded spots of bacterial ooze in new canes were 20-66 cm 28 days after inoculation in April (Serizawa and Ichikawa, 1993a), necrotic lesion development under the bark of overwintered canes was not observed at 15°C 3 and 6 days after inoculation (Fig. 1). Ko et al. (2000a, 2000b) also showed that the length of lesions caused by the pathogen on overwintered canes at 15°C during 18 days is 5-8 mm. Ko et al. (2000a) explained that it would come from the difference in speed of sap flow.

Whereas temperatures within the range of 12°C to 18°C favor disease development caused by the pathogen on new canes (Serizawa and Ichikawa, 1993b), the temperatures did not favor necrotic lesion development on overwintered canes (Fig. 1). It could come from the difference in speed of sap flow and resistance of tissue between new and overwintered canes. Once inside the plant, infections caused by the pathogen can occur on the majority of tissues and rapidly become systemic through the bacterial migration inside the xylem (Spinelli et al., 2011). High speed of sap flow at more than 20°C would enable rapid migration of the pathogen and the pathogen could cause long necrotic lesions at more than 20°C. In addition, the pathogen clades that are mainly isolated and cause bacterial canker in Korea and Japan can grow at 30°C (Choi et al., 2014) whereas the symptoms caused by the pathogen were not discovered on new canes at 25°C (Serizawa and Ichikawa, 1993b). It indicated that the pathogen could invade susceptible tissues and cause the symptoms at 30°C.

The research of Ko et al. (2000a, 2000b), designed to investigate the effect of freezing on bacterial canker development in kiwifruit cane, was also conducted using overwintered Hayward canes without leaves or roots. As there was no growth under the tested freezing conditions, results from a test using whole plants might have been similar to those from our test using kiwifruit canes without leaves and roots. The equation obtained by re-analyzing the data reported by Ko et al. (2000b) could show the relationship between freezing and bacterial canker lesion development on the canes of normal kiwifruits, within the range of −10°C to −3°C.

As D-PSK-K estimates the amount of necrotic lesions on kiwifruit canes caused by bacterial canker during the last growing and overwintering seasons, the best way for evaluating the accuracy of model would be the comparison of the estimated and the observed amount of necrotic lesions on kiwifruit canes. However, the necrotic lesion cannot be detected from the outside view and we cannot measure the amount of necrotic lesions without unrecoverable damage on kiwifruit canes. It is actually impossible to observe the amount in kiwifruit orchards. In order to evaluate the accuracy of model, we had to compare the estimated with something that can be measured and could be directly proportional to the amount of necrotic lesions in kiwifruit orchards. We evaluated the accuracy of D-PSA-K by comparing the observed maximum DI on kiwifruit canes with the damage estimated by the model in this study.

In both Hayward and Hort16A, the D-PSA-K model estimated high accumulated damage on kiwifruit canes caused by bacterial canker during the overwintering and growing seasons, when the observed maximum DI was large (Table 1). D-PSA-K could roughly predict the amount of disease occurrence in a kiwifruit orchard, relative to the past disease occurrences in that orchard. For the Hayward cultivar, D-PSA-K estimated the accumulated damage as 7.3-11.1 times the observed maximum DIs, and the relationship equation between the accumulated damage (D) and the observed maximum DI. These results were obtained from linear regression analysis using the data in Table 1, where DI = 0.1158D and R² = 0.89. The results indicated that the maximum DI for the Hayward cultivar could be predicted with high accuracy using D-PSA-K, and that the accumulated potential damage calculated using D-PSA-K could estimate the actual damage caused by the bacterial canker well. We did not observe such a relationship for Hort16A. These differences might be caused by differences in resistance to low temperatures and pathogen differences between Hayward and Hort16A (Balestra et al., 2009; Ferrante and Scortichini, 2014; Koh et al., 2010). Differences in the resistance of kiwifruit cultivars to the bacterial canker in Korea are not clearly understood. Outbreaks of bacterial canker causing the destruction of an entire orchard have been observed on Hort16A in Korea, but no such outbreaks on Hayward have been reported over the past two decades (Koh et al., 2010). These results indicate that Hort16A might be more susceptible to bacterial canker caused by the pathogen belonging to the Korean clade than is Hayward. In addition, A. deliciosa, including Hayward, is more frost tolerant than A. chinensis, as represented by Hort16A (Ferrante and Scortichini, 2014). For Hort16A, D-PSA-K also estimated the accumulated damage as 0.96 or 6.52 times the observed maximum DI, indicating that Hort16A might be more susceptible to bacterial canker in Korea than is Hayward. D-PSA-K might be used to estimate trends of real disease occurrences, although the model could only accurately estimate the maximum DI for Hayward. Some adjustments of D-PSA-K, including consideration of differences between Hayward and other cultivars in terms of disease resistance and frost tolerance, might enable more precise estimation for the other cultivars.

As reported by the Rural Development Administration of Korea in 2009, Hayward was cultivated on 918 ha, whereas the other cultivars were cultivated on 132 ha in 2008 in Korea. Hayward is the most widely cultivated in Korea and, as a result, D-PSA-K would be useful for most kiwifruit orchards there. D-PSA-K would be also useful for summarizing disease occurrence on kiwifruit stems at regional or national scales, assuming that all kiwifruits cultivated in Korea are Hayward.

In conclusion, we developed a model to estimate the accumulated potential damage on kiwifruit canes caused by bacterial canker during the previous overwintering and growing seasons. The D-PSA-K model could estimate maximum DI on the stems of Hayward, as an indirect measure of yield loss caused by the bacterial canker in a given year. Although the model could estimate the maximum DI only for Hayward, D-PSA-K might also be used to estimate disease occurrence in a kiwifruit orchard compared to past occurrences, and to estimate the trends of real disease occurrences in other cultivars. D-PSA-K could help kiwifruit growers select new sites for kiwifruit cultivation and establish an improved production plan by predicting the loss in kiwifruit production due to bacterial canker, using past or predicted weather data.