Rust, caused by Puccinia dracunculina, is the main foliar disease of open-field tarragon (Artemisia dracunculus) crops in Israel. As not much is known about the biology or epidemiology of this pathogen, the long-term objective of the current study was to accumulate the knowledge needed to develop an effective, environmentally friendly means of adequately managing the disease. Puccinia dracunculina is an autoecious brachy-form pathogen, but it is not known whether the life cycle is completed under field conditions in Israel. Field observations and greenhouse studies revealed that although the telial stage is produced, the pathogen overwinters in the uredinial stage. In vitro experiments were used to quantify the temperature and wetness requirements for urediniospore germination and to calculate the daily duration of conducive weather (DDCW); DDCW was defined as the number of hours during which temperature ranged between 15 and 25°C and RH was >90%. Cumulative DDCW values (CDDCW) were a good predictor of disease under natural conditions in two growing seasons. Disease onset occurred when CDDCW values reached a level of 10 and the relationship between log CDDCW values and season-long severity values (in logit) was highly significant, explaining 90·6% of the variation.
Tarragon (Artemisia dracunculus), also known as dragon’s-wort, is an economically important perennial herb of the Asteraceae family. It is used fresh or dried and processed for flavouring and fragrances, as well as in traditional medicine. The leaves are thin, elongated, glossy and dark green. Flowering is in the summer and the inflorescence, in yellow florets, develops at the top of the branches. Two varieties, French and Russian tarragon, are known. French tarragon has a better taste, but seed production is rare. Propagation is solely by cuttings, so disease-free material must be carefully selected to avoid spread of diseases through nurseries. Russian tarragon can be cultivated from seeds, but its taste is considered inferior (Palevitch & Yaniv, 1991; Werker et al., 1994).
In Israel, French tarragon cultivars are cultivated in two distinct production systems. The first is for the production of fresh herbs. Seedlings are transplanted into the soil in walk-in tunnels or in polyethylene-covered greenhouses and production continues all year round. When plants are 25–50 cm in height (usually within 1–2 months after planting), the upper part of the canopy is harvested manually using sharp knives. Subsequent harvesting is performed every 3–5 weeks, the frequency depending mainly on temperature and solar irradiation. The cuttings, 10–20 cm in length, are packed and sent to market. The second production system is for the dried herb market. Seedlings are transplanted into the soil in open fields in beds, three to four rows per bed, and grown for three to four consecutive seasons. New crops are transplanted in early spring (February to March), when sprouting is initiated in second-year or older crops. New stems and leaves form promptly and when the foliage fills the gaps between rows and plant height is about 30–50 cm, usually in May, the upper part of the canopy is mechanically harvested. The harvest is taken to a factory where the foliage is cleaned, chopped into small pieces, dried, processed and packed in small tubs. Subsequent harvesting is performed within 2–3 months and there are usually three to four production cycles in a growing season. In the autumn (October to November), plant growth rate decreases markedly, the leaves fall and the plants enter dormancy (December to February). The process starts again in the following season.
The primary foliar pathogen threatening open-field tarragon production in Israel is tarragon rust (Gamliel & Yarden, 1998), caused by the fungus Puccinia dracunculina. The pathogen was first described by Fahrendorff (1941). Puccinia dracunculina is an autoecious brachy-form pathogen, and thus all of its stages form on one host. In production fields, uredinial sori are observed, usually at the end of the first or during the second production cycle. As time passes, the disease intensifies and abundant uredinial sori develop on the leaves and stems, causing drying and defoliation of leaves, and reductions in growth rate and harvest quality. It is not uncommon for the third harvest to be so damaged that growers have to forego the fourth production cycle because of inadequate quantity and quality of foliage. In the autumn, with the drop in temperature, telial sori are produced en masse. As the biology of P. dracunculina in Israel has not been studied, it is not known how the pathogen overwinters in production fields or whether it completes its life cycle. As a consequence, attempts to manage the disease rely solely on chemical control. Spraying is initiated before disease onset and is applied thereafter at predetermined intervals. To avoid contamination of the marketable produce with fungicide residues, growers refrain from spraying for 2–3 weeks before each harvest, allowing the fungicide residues to degrade. On average, two to four sprays are applied in each growing cycle, i.e. 6–12 in a growing season. A profound understanding of the biology and epidemiology of the disease is currently unavailable, hindering the development of a rational disease management strategy. The long-term objective of the current study was to develop an effective, environmentally friendly means of adequately managing tarragon rust in open fields. The specific objectives were to: (i) explore the mechanisms by which the pathogen overwinters in production fields; (ii) quantify the environmental requirements of the uredinial stage of the pathogen; and (iii) document the dynamics of disease development in the production field.
Materials and methods
Observation, experimentation and plant sampling were carried out in a commercial tarragon field in Mevo-Hama (32°42′53″N, 35°39′15″E, 350 m a.s.l.) in the years 2008 and 2009. French tarragon transplants (Hishtil Nurseries) were planted in April 2007 in beds. The distance between beds was 0·7 m; each bed was 1·2 m wide and consisted of four rows. Planting stand density was 9 plants m−2. The plants were irrigated using a buried (10-cm-deep) drip irrigation system and cultivated and maintained according to the practices commonly employed by commercial growers in the area. Fungicides for the suppression of rust were not applied in the experimental part of the field.
Overwintering of the pathogen
Four possible overwintering mechanisms of P. dracunculina were examined: (i) completion of its brachy-form sexual stage: the teliospores form in the autumn and germinate during the winter, producing basidiospores that will later penetrate the next new leaves, forming pycnia and then new uredinial sori; (ii) survival of hyphae inside the lower living parts of the dormant plants; (iii) survival of urediniospores inside sori on dry leaves; or (iv) overwintering on other Artemisia hosts that are not dormant in the winter.
Laboratory and greenhouse studies of rust life cycle
Leaves and stems bearing telial sori were sampled from the experimental field in the autumn and winter seasons of 2007–2008 and 2008–2009. Attempts were made to induce teliospore germination by floating 25-mm2 dry leaf or stem pieces carrying telia in tap water in a small Petri dish at 5°C for 2–10 weeks (Anikster, 1986), and then transferring them to 18°C for germination testing. Leaf pieces bearing telial sori were used to inoculate young tarragon plants as follows: pots with tarragon plants were covered with PVC cylinders, covered at the top with filter paper. Leaf pieces were attached to the inside of the PVC cylinder allowing basidiospores (if formed) to drop down onto the plant. These plants were subsequently covered with a plastic bag to maintain humidity for 36–72 h at 18°C and were later maintained in a greenhouse at 20 ± 2°C.
Survival of hyphae
The possibility of hyphae overwintering inside the lower living parts of the dormant plants was examined in plants sampled from the Mevo-Hama commercial field from January to March in 2008 and 2009. Each year, 28 infested tarragon plants were uprooted from the commercial field and brought to Tel Aviv University. The soil surrounding the lower parts and roots of 14 plants was washed off in running tap water and the plants were replanted in a local soil. The other 14 plants were planted in buckets with the original soil surrounding their lower parts and roots. Seven plants from each group were maintained inside the net-house and the others were maintained outside the net-house. In April, plants started to sprout and appearance of rust on the new leaves was recorded. Tarragon plants originating from a nursery and grown under the same conditions served as controls.
Survival of urediniospores
The possibility of urediniospores overwintering inside sori on dry leaves was investigated by checking the germinating ability of urediniospores during the winter. Dry and green leaves and stems bearing abundant uredinial sori were cut in October 2009, put in netted plastic bags and kept at Tel Aviv University. The bags were maintained under one of six different sets of environmental conditions: (i) in a roof-covered net-house; (ii) in a shaded net-house, the bags covered with soil; (iii) in an incubator at 15°C; (iv) in a greenhouse at 21°C; (v) outside the net-house in the shade, exposed to precipitation; or (vi) in a net-house, exposed to sunlight. During the winter, samples were taken twice a month from each net bag. Urediniospore suspension was prepared from naturally infected leaves. Urediniospores were spread on 2·5% water agar and incubated for 24 h at 18°C. Germination was examined using a light microscope.
The possibility that P. dracunculina overwinters on different Artemisia species was tested by inoculating wild Artemisia monosperma and cultivated Artemisia arborescens. Six seedlings of each species (growing in a small bucket) were inoculated in the greenhouse (21 ± 2°C), including two buckets with tarragon plants as a control. After inoculation, all of the buckets were incubated in a moist chamber for 24 h. Appearance (or absence) of rust on the different plants was recorded.
Environmental conditions for germination of P. dracunculina urediniospores
The temperature and wetness duration required by P. dracunculina urediniospores for germination were studied in vitro in a repeated study in a controlled environment. Tarragon plants exhibiting severe P. dracunculina rust symptoms, collected from the experimental field in Mevo-Hama, served as the source of urediniospores. Infected leaves were gently brushed with a soft-bristled paintbrush and the urediniospores were collected in 55-mm Petri dishes. Drops of 0·1 mL containing fresh urediniospores (105 urediniospores mL−1) were placed onto Petri dishes containing water agar (2%) and spread with a glass rod. The dishes were then placed in incubators at six different temperatures (5, 10, 15, 20, 25 and 30°C) in the dark. After 24 h of incubation, urediniospore germination was determined by means of a light microscope (×100 magnification). Urediniospores were considered to have germinated if the length of the germ tube was at least twice that of the urediniospore itself. Because urediniospore aggregation may reduce germination, only separate urediniospores were considered. There were four replicates (Petri dishes) per treatment (temperature) and 25 urediniospores were inspected per replicate.
For the wetness-duration experiments, drops of urediniospores suspended in 0·1% water agar were placed on glass microscope slides that were then placed in closed Petri dishes over moistened paper. The Petri dishes were incubated at 18°C in the dark. At different times after the start of the experiment (0, 2, 4, 6, 10 and 24 h), the Petri dishes were opened and the drops dried within 5–10 min. Urediniospores were stained with cotton blue and germination was examined after 24 h as described above. The controlled-environment experiments were repeated once. The relationship between percentage urediniospore germination, temperature and wetness duration was quantified using regression analyses.
Dynamics of disease development in the production field
The dynamics of rust development were recorded in the Mevo-Hama field in 2008 and 2009. In both years, the crop was harvested three times: on 6 May, 2 August and 15 October in 2008, and on 18 May, 10 July and 28 September in 2009. The period between the initiation of sprouting in the spring (early to mid-March in both years) and first harvest is referred to hereafter as the first production cycle; the period between the first and second harvests is referred to as the second production cycle; and the period between the second and third harvests is referred to as the third production cycle. Harvesting was carried out by a mechanical harvester (De Pietri), cutting the foliage to a height of c. 10 cm above the soil; the cut foliage was transferred to the cargo tank of the harvester. In 2009, the upper plant parts remaining after the second harvest were blazed by means of a burner attached to a tractor and fuelled by commercial cooking gas. Blazing was performed to eradicate the pathogen and infected leaves before the following production cycle. Most plants survived the heat and sprouted normally with no damage.
Four field plots (6 m long, 0·7 wide) were marked in an unsprayed section of the commercial field. The plots were inspected periodically by two individuals, starting soon after the initiation of sprouting each year (early to mid-March) and continuing at 7- to 21-day intervals until the last harvest. Rust severity (the percentage of foliage exhibiting typical rust symptoms) was assessed visually on a whole-plot basis, and the scores were averaged. For some analyses, the logit-transformed rust severity values (y) were calculated as logit (y) = ln (y/(100–y)). Temperature and relative humidity (RH) were recorded hourly in the field by two data loggers (Hobo-Pro, Onset Computer Corp.) installed at a height of c. 1 m above ground. The foliage was considered wet at RH >90%. The temperature and wetness requirements for urediniospore germination defined in the above in vitro experiments were used to calculate the daily duration of conducive weather (DDCW) in the field for the 2008 and 2009 growing seasons. DDCW was defined as the daily number of hours during which temperature and wetness duration enabled germination of ≥80% of the urediniospores. Cumulative DDCW values (CDDCW) were calculated by accumulating the daily values starting at a biofix (i.e. the date of computation initiation; Llorente & Montesions, 2004) of 1 March.
Overwintering of the pathogen
Laboratory and greenhouse studies of rust life cycle
Only a few of the attempts to use induced teliospore germination to infect tarragon plants gave positive results. Generally, one cell of the teliospores germinated; rarely did both produce a basidium (Fig. 1, stage III). The diploid nucleus migrated to the basidium (Fig. 1, stage III, l), underwent meiosis and produced four haploid nuclei which later migrated separately to the four basidiospores formed on the basidium sterigmata (Fig. 1, stage III, m). Basidiospores were then ejected, and those falling on a tarragon leaf penetrated the epidermis. The fungus formed a pycnial cluster (Fig. 1, stage 0, c,d), and uredinial sori were produced (Fig. 1, stage II, h,i), surrounding the cluster. However, intensive searching of the tarragon field throughout the study did not reveal any traces of the pycnial stage (Fig. 1, stage I).
Survival of hyphae
No rust appeared on tarragon plants uprooted from the field and grown in local light soil in buckets. Three uredinial sori were observed on two out of 14 plants that were uprooted from the commercial field with the surrounding soil, indicating that overwintering may start from spores remaining on or in the soil from dry plant debris. Control plants remained disease free for the entire duration of the experiment.
Survival of urediniospores
Dry and green tarragon leaves covered with uredinial sori were kept in netted plastic bags placed in a net-house and germination ability assessed. In April 2010, urediniospores originating from leaves sampled in the previous autumn and kept at 15°C germinated at rates ranging from 1 to 5%, suggesting that overwintering urediniospores may serve as a source of primary inoculum for the following spring’s epidemics.
The possibility of overwintering on different Artemisia species was tested. Only the commercial variety of cultivated tarragon plants serving as controls became infected (>100 sori/plant). Neither of the two Artemisia species showed any signs of infection.
Environmental conditions for germination of P. dracunculina urediniospores
In the laboratory experiments, germination of P. dracunculina urediniospores was relatively low, never exceeding 25%. Urediniospores did not germinate at 5°C; the optimal temperature for spore germination was 21°C and the upper temperature limit for germination was 31°C (as estimated by extrapolation from the regression equation describing the relationship between temperature and urediniospore germination). About 80% of the germinating urediniospores did so at temperatures ranging from 15 to 25°C (Fig. 2a). Urediniospores did not germinate in dryness: a minimum 2 h of wetness was necessary for initiation of germination, and the proportion of germinating spores increased gradually with increasing wetness duration up to an asymptote at c. 11 h. About 80% of the germinating urediniospores did so at a wetness duration of 10 or more hours (Fig. 2b). The temperature and wetness requirements for urediniospore germination were used to calculate the DDCW, defined as the number of hours during which temperature ranged between 15 and 25°C and RH was >90%.
Dynamics of disease development in the production field
The first rust symptoms were observed in the field on 22 May 2008, at the beginning of the second production cycle. The disease developed slowly during this cycle, reaching a severity of c. 20% by mid-July, just before harvest. In the third cycle, from August to October, disease severity was high, ranging from 40 to 60% (Fig. 3a). The pattern of disease progression in 2009 was different: the first disease symptoms were observed earlier in the first production cycle, on 12 Apr 2009. During the second production cycle, the disease developed rapidly, reaching a severity of c. 60% by mid-July. Blazing the foliage soon after the second harvest almost completely eliminated the disease throughout most of the third production cycle. A few isolated sori were observed by late September, just before the third harvest (Fig. 3b).
DDCW values varied markedly between production cycles and the two growing seasons. In the first production cycle of 2008, DDCW was >0 h on only 3 days, and it never exceeded 4 h. In the second production cycle, DDCW ranged from 4 to 7 h on 35·5% of the days and from 8 to 11 h on 20·6% of the days. The weather in the third production cycle in 2008 was more conducive to the disease, with 62·1% of the days having DDCW values between 8 and 11 h (Fig. 4a). In 2009, in the first production cycle DDCW was >0 h on 25 days, with 3 days above 4 h. In the second production cycle, DDCW ranged from 4 to 7 h on 18% of the days, from 8 to 11 h on 30·3% of the days, and ≥12 h on 22·7% of the days. The weather in the third production cycle was highly conducive as well, with 43·5% of the days having DDCW values ranging from 8 to 11 h and 40·1% of the days having DDCW≥12 h (Fig. 4b).
The main differences in weather between the 2 years were apparent during the first disease cycle. This can be illustrated by comparing the CDDCW values calculated for this period. First disease symptoms were observed when CDDCW reached a log value of 1; this occurred in mid-May in 2008, and 40 days earlier in 2009. In the second and third production cycles, the pattern of CDDCW dynamics was comparable in both years (Fig. 5). CDDCW was a good predictor of disease development: the coincidence between log CDDCW values and the season-long severity values (in logit) was highly significant and 90·6% of the variation in disease severity (in both years) was explained by this parameter (Fig. 6).
Because the edible part of the tarragon plant is its foliage, even minor rust infection can lead to a significant decrease in quality and marketability of the product. Accordingly, maintaining disease-free foliage is crucial for high-quality production. Furthermore, as inoculum remaining on the plant parts after harvest serves as the source of initial inoculum for subsequent harvests, preventing its build-up would allow for a fourth production cycle. Indeed, in the experimental plots where fungicides for disease suppression were not applied, disease severity increased to levels that resulted in low-quality yields in the second and third (in 2008) harvests.
Puccinia dracunculina is autoecious: its entire life cycle is restricted to the tarragon plant. The life cycle is brachy-form, meaning that the aecial stage is missing. The whole life cycle was completed under controlled greenhouse conditions only, where teliospore germination was induced. In the Israeli field (in the present experiments), tarragon rust formed only the uredinial and telial stages: the sexual stages were not found in the field during the growing season or in the winter. The epidemiological significance of these findings is that teliospores have no role in transferring the disease to the next season. Germination of urediniospores was observed on leaves sampled in October and kept at 15°C until April, indicating that overwintering urediniospores may serve as a source of primary inoculum for the following spring’s epidemics.
A similar partial life cycle is found in Israel for sunflower rust, caused by Puccinia helianthi, an autoecious rust that, in more northern countries, is known to form all spore types on the plant. In Israel, however, pycnia and aecia are not seen on sunflower plants in the field (D. Shtienberg, unpublished data). The cultivated sunflower is sown in the spring (March–April). In May, the first uredinial sori are seen. As the rust epidemic advances, without fungicide treatment, all of the leaves are eventually covered with uredinial sori. Similar to tarragon rust, in sunflower, the entire life cycle occurs only in the greenhouse after special treatment to induce teliospore germination.
Environmental conditions for completing an entire rust life cycle do exist in Israel, as evidenced in garlic attacked by the autoecious rust Puccinia allii, which forms all spore types on the same plant (Anikster et al., 2004). Similarly, in a very common wild leek (Allium ampeloprasum) in Israel, almost every plant is attacked by this fungus. At the beginning of winter, pycnia and aecia appear, and later, uredinia and telia are seen. Cultivated garlic, unless sprayed with fungicide, will be covered by uredinial and telial sori. It thus becomes clear that dry leaves covered with telia that remain on and under the soil surrounding the wild leek plants in the summer are responsible for disease initiation in the following year.
The fungus P. dracunculina, which attacks cultivated tarragon fields in Israel, has the potential to finish its brachy-form cycle, as was proven here in the laboratory and greenhouse. However, an intensive search of the field during the two successive years of the study did not detect any pycnia on young leaves. It was concluded that the sexual stage does not exist in the field, and that the massive formation of teliospores leads to a dead end. On the other hand, the urediniospores (in sori on dried leaves) remained viable in the field from October to April at a temperature of 15°C. In addition, dry plants that were uprooted with the original soil in January and brought to Tel Aviv University began to sprout in April, and a few uredinial sori appeared on the new leaves. Thus, the urediniospores retained in and on the soil during the winter may serve as the source for the new epidemics in the following spring.
Spore germination, as affected by weather conditions, is an important factor affecting the development of rust diseases (Stuckey & Zadoks, 1989; de Vallavieille-Pope et al., 1995; Furuya et al., 2009). In our in vitro experiments, P. dracunculina urediniospores germinated at between 5 and 31°C, with an optimum between 15 and 22°C. Wetness was required for germination and optimal wetness duration for germination was ≥10 h (Fig. 2). These findings are in close agreement with those reported for other Puccinia species in other hosts, such as P. allii in onion (Gilles & Kennedy, 2003; Furuya et al., 2009), P. substriata var. indica in pearl millet (Tapsoba & Wilson, 1997), P. allii in leek (Gilles & Kennedy, 2003), P. recondita and P. graminis in wheat (Kramer & Eversmeyer, 1992) and P. helianthi on sunflower (Shtienberg & Vintal, 1995). A similar approach was presented by Pfender (2003), who combined temperature and moisture duration in one factor, termed wet degree-hours (DHw). DHw units were accumulated as summed degree-hours >2·0°C when moisture was present. Data were used to develop a weather-based infection model for stem rust of perennial ryegrass caused by P. graminis subsp. graminicola.
Adequate suppression of P. dracunculina is a feasible goal, which can be achieved by reducing the amount of initial inoculum and by restricting the rate of pathogen proliferation. The results of the current study may provide the knowledge needed to achieve these goals. It was found that P. dracunculina does not complete its sexual cycle in Israel, and uredinial sori overwintering in dormant plants serve as the source of initial inoculum for the following season. Accordingly, attempts should be made to suppress the uredinial stage until the plants enter dormancy, as a high level of disease at the end of one season is expected to contribute a large amount of initial inoculum to the following season. Suppression can be achieved, for example, by application of fungicides to the foliage developing in the autumn after the last harvest or by non-commercial cutting in the autumn aimed at removing infected tissues from the area. As tarragon is harvested several times within a growing season, uredinial sori remaining on the lower plant parts after harvesting serve as the source of initial inoculum for the following production cycle. Thus, attempts should also be made to improve sanitation within seasons. This can be achieved in several ways. In 2009, for example, the upper plant parts remaining after the second harvest were blazed and this action eradicated the pathogen from the lower plant parts. In the production cycle that followed, the disease was practically eliminated. Although the profitability and environmental consequences of this method should be carefully examined, this result illustrates the significance of initial inoculum in subsequent disease progression.
The rate of pathogen proliferation may be restricted, for example, by optimal timing of fungicide sprays. Success of chemical control relies on precise timing of the first spray of the season (soon before or after disease onset) and on accurate timing of subsequent sprays (when the environment is conducive to the pathogen). Results of the current study may serve to develop a forecasting system for this purpose. As indicated, disease onset occurred when CDDCW values reached a level of 10 (log in Fig. 5). This value can be used an indicator to initiate spraying. Had this threshold been used in 2008, sprays would not have been applied at all in the first production cycle, presumably saving on three unnecessary applications. This approach is the basis for numerous successful and widely used prediction systems. For example, accumulation of ‘severity values’ was the basis for initiating sprays against Phytophthora infestans by the blitecast system (MacKenzie, 1981), against Alternaria solani by the fast system (Madden et al., 1978), and against Didymella rabiei (Shtienberg et al., 2005). The suitability of environmental conditions during the season may also be quantified by CDDCW values and used to determine the need for implementing subsequent sprays. However, the specific threshold to be used was not determined in the current study and should be defined experimentally.
The authors wish to thank Uri Lotzinski, Shaul Ginszberg and Shaul Graph for their valuable contributions in managing the field experiments. The research was supported by SDA Spice, Sde Eliyahu, the Fresh Herbs Farmers Association and the Plant Marketing and Production Board.