Introduction

Erwinia tracheiphila (Smith), the causal agent of cucurbit bacterial wilt, is transmitted by striped (Acalymma vittatum (F.)) and spotted (Diabrotica undecimpunctata howardi (Barber)) cucumber beetles. Bacterial wilt causes severe losses in many cultivated cucurbit crops, primarily in the genera Cucurbita and Cucumis. Economic losses of cucurbit crops from bacterial wilt can reach 75% (Zehnder et al. 1997), and muskmelon (Cucumis melo L.) and cucumber (Cucumis sativus L.) are among the most susceptible crops (Sherf and MacNab 1986). Epidemics have occurred primarily in the Midwest and Northeast U.S. as well as in southern Quebec, Canada (Sherf and MacNab 1986; Fleischer et al. 1999; Toussaint et al. 2013); the disease has been restricted primarily to these areas with the exception of a recent report of bacterial wilt on pumpkin and watermelon in New Mexico (Sanogo et al. 2011).

Overwintering of E. tracheiphila occurs in the foregut and hindgut of striped cucumber beetles (Garcia-Salazar et al. 2000). It is transmitted when E. tracheiphila-infested frass of adult beetles comes into contact with fresh wounds on leaves, stems, or floral nectaries (Sasu et al. 2010) and the bacteria invade the xylem. E. tracheiphila cells multiply in the xylem and can block water flow (Main and Walker 1971). Infected plants initially exhibit wilting of leaves near the infection site, followed by wilting of vines and eventual collapse and plant death.

Current management for cucurbit bacterial wilt relies primarily on insecticide applications (Cavanagh et al. 2009), but this approach is costly and endangers the health of humans, pollinators, insectivorous birds and other ecosystem service providers (Cavanagh et al. 2009; Potts et al. 2010). Although watermelon (Citrullus lanatus var. lanatus) is resistant to bacterial wilt and a few cultivars of cucumber (Cucumis sativus) are tolerant, resistant cultivars are not commercially available for most cucurbit crops. Therefore, a clearer understanding of the bacterial wilt infection process may yield insights to assist plant breeders in developing such cultivars. A starting point toward this goal is to understand the impact of host age on disease progress, as this may help breeders develop efficient protocols for screening candidate lines for resistance. Moreover, tracing patterns of pathogen movement in the xylem could result in a deeper understanding of disease development.

The impact of plant age on wilt symptom development has been investigated for several xylem-inhabiting bacterial pathogens. Plant age affects bacterial wilt and canker of tomato caused by Clavibacter michiganensis subsp. Michiganensis: tomato transplants up to the 17- to 18-leaf stage wilted and died after inoculation, whereas plants that were inoculated after that stage exhibited only mild symptoms (Sharabani et al. 2013). Similarly, Ralstonia solanacearum caused earlier symptoms, higher disease incidence, and greater disease severity of bacterial wilt of tomato in 2- to 3-week-old seedlings than in 5- to 6-week-old plants (Thomas and Upreti 2014). For cucurbit bacterial wilt, the relationship between plant age and symptom development has been examined only for the relatively resistant host genera Cucurbita and Citrullus. In pumpkin (Cucurbita pepo L.), which exhibits intermediate resistance to E. tracheiphila (Sherf and MacNab 1986), seedlings inoculated at the cotyledon stage were more susceptible than older plants and resistance increased sharply with plant age (Brust 1997a, 1997b). Seedlings of watermelon (Citrullus lanatus), which is highly resistant to E. tracheiphila (Watterson et al. 1971), exhibited more rapid symptom development than older plants and resistance increased with age (Watterson et al. 1971). However, no such relationships have been assessed for the highly susceptible cucurbits in the genus Cucumis, such as muskmelon or cucumber, which are at the greatest risk of economic losses from bacterial wilt.

The mechanisms by which E. tracheiphila causes bacterial wilt have not yet been investigated in detail. Absence of evidence for pectolytic enzyme production suggests that physical occlusion by the bacteria is the primary cause of xylem dysfunction (Main and Walker 1971; Watterson et al. 1971). Moreover, the presence of strands of ooze as infected stems are cut and drawn apart suggests the presence of extracellular polysaccharides, which could also contribute to xylem blockage, as with wilt by R. solanacearum, Pantoea stewartii and Xylella fastidiosa (Ayers et al. 1979; Beck von Bodman et al. 1998; Saile et al. 1997). Factors that contribute to the virulence of other bacterial xylem pathogens, such as biofilm formation, quorum sensing, outer membrane vesicle production and motility (Herrera et al. 2008; Ionescu et al. 2014; Koutsoudis et al. 2006; Meng et al. 2005), have not yet been examined in E. tracheiphila. Recently, observational studies using a constructed bioluminescent strain of E. tracheiphila provided the first evidence that the bacterium can move both upward in muskmelon seedlings and downward into the roots following inoculation (Vrisman et al. 2016). The objectives of the present study were to: i) quantify the impact of host age on development of wilt symptoms in muskmelon, and ii) trace the movement of E. tracheiphila in the xylem following inoculation.

Materials and methods

Plant growth conditions

Muskmelon (Cucumis melo cv. Athena) seeds were planted in a 1:1:1 matrix of peat moss, coarse perlite, and Metro-Mix 300 (Sun Gro Horticulture, Vancouver, BC, Canada). A single seed was planted in each 650 cm3 pot. The pots were incubated at 26 °C under a daily regimen of 14 h light and 10 h darkness under ambient relative humidity (RH) in growth chambers. Plants were watered daily and fertilized weekly (NPK: 15–5-15: Miracle Gro®, The Scotts Co., Marysville, OH).

Bacterial strain and inoculum preparation

E. tracheiphila strain SCR3 (Saalau Rojas and Gleason 2012) was used in this study. This strain is a spontaneous rifampicin-resistant derivative of an isolate from a symptomatic muskmelon plant grown in Iowa, U.S. Pathogenicity of this isolate was confirmed by puncture inoculation of the first true leaf of 2-week-old muskmelon plants in growth chamber trials (Saalau Rojas and Gleason 2012). To prepare cells for plant inoculation assays, SCR3 was recovered from −80 °C storage on solid nutrient agar peptone medium (de Mackiewicz et al. 1998) that was amended with rifampicin (75 μg/ml) (NAP-Rif). Cells were grown at 27 °C for 3 days, then transferred to fresh NAP-Rif medium and grown at 27 °C for another 3 days. Bacterial suspensions were prepared by recovering SCR3 colonies from the surface of solid NAP-Rif medium and suspending them in 10 mM phosphate-buffered saline (PBS) to a concentration of approximately 2.5 × 108 CFU/ml, based on a standard curve relating cell density to optical density at 540 nm.

Symptom expression experiments

Muskmelon seeds were planted 2, 4, and 6 weeks before inoculation in the first experiment and 2, 4, 6 and 8 weeks before inoculation in a second experiment, in order to create multiple cohorts of plants that varied in age at the time of inoculation. In each experiment, plant age at inoculation was considered a treatment, and each treatment included four single-seedling replicates. A 100-μl droplet of inoculum was applied at the base of the adaxial surface of the youngest fully expanded leaf, followed by puncturing the leaf at the site of the droplet with a 28.6-mm-diameter, 60-pin florist’s pin frog (Kenzan Pin Frog, sold by www.save-on-crafts.com). Next, the pipette tip was rubbed lightly against the punctured site in order to ensure maximal contact of the inoculum with the puncture wounds and an additional 100 μl of suspension was applied to the punctured site on the leaf, after which the pipette tip was again rubbed lightly on the site. Control plants were inoculated with PBS (10 mM) buffer in the same manner as described above. After inoculation, plants were incubated in a growth chamber at 26o C under a daily regimen of 14 h light and 10 h darkness and ambient RH. Numbers of wilted and asymptomatic leaves on each plant and their locations relative to the site of inoculation were determined daily until all E. tracheiphila-inoculated plants displayed wilt symptoms. Rate of wilting was determined by estimating the number of days for 50% of the plant to show wilt symptoms.

Pathogen movement experiments

Plants that were 2, 4, 6, and 8 weeks old were inoculated as described above, with plant age at inoculation considered as a treatment. On days 3, 7, 14, and 21 after inoculation, four plants of each treatment were chosen arbitrarily for destructive sampling. Stem segment samples (5 cm long and devoid of nodes) from each treatment were excised above and below the point of inoculation. In the first run of the experiment, all internodes were sampled for the plants that were 2 or 4 weeks old at inoculation (referred to here as 2-week-old and 4-week-old plants), whereas every third internode was sampled in older plants(referred to here as 6-week-old and 8-week-old plants). In the second run of the experiment, every internode was sampled from the 2- and 4-week-old plants, whereas every second internode was sampled from the 6- and 8-week-old plants; again, four plants were selected for each treatment. These internode stem segments were surface-sterilized by spraying them with 70% ethanol and then air dried on sterile paper towels. A stem segment was cut transversely at its midpoint, the cut surface was imprinted on the surface of NAP-Rif medium, and the resulting imprint was streaked for single colonies. The presence of colonies exhibiting E. tracheiphila morphology was recorded after 4 days of incubation at 27o C. The number of wilted and asymptomatic leaves per plant was also counted on each sampling date. Sampling was terminated in each treatment when all leaves had wilted. Bacterial movement in the upward direction vs. the downward direction was compared on the basis of the mean number of internodes from the site of inoculation to the furthest internode from which E. tracheiphila cells were recovered.

Data analysis

In the symptom expression experiments, disease progress was determined based on the area under the disease progress curve (AUDPC) using the trapezoidal method (Simko and Piepho 2012) and the means were compared with a Fisher’s least significant difference test using SAS 9.1 (SAS Institute Inc., Cary, NC, USA). For each treatment in the pathogen movement experiments, the mean number of internodes from which E. tracheiphila was recovered above vs. below the point of inoculation was compared using a Student’s t-test.

Results

Rate of wilting

Leaves on the inoculated muskmelon plants began to wilt as early as 4 days after inoculation. In both runs of the experiment, the inoculated leaves wilted rapidly regardless of plant age at inoculation (Table 1). In the first experiment, the plants that were 2 weeks old at inoculation wilted significantly (p < 0.05) faster than those that were 4 or 6 weeks old, as reflected in the mean number of days from inoculation until wilting of 50% of leaves and area under the disease progress curve (AUDPC) (Simko and Piepho 2012). Similarly, plants that were 4 weeks old at the time of inoculation wilted significantly faster than those that were 6 weeks old. In the second run of the experiment, although the plants in each age group wilted more slowly than similarly-aged plants in the first experiment, the plants that were younger at the time of inoculation again wilted faster than those that were older. Specifically, the 2- or 4-week-old plants wilted significantly (p < 0.05) faster than the 6- and 8-week-old plants based on the number of days required for plants within each treatment to display 50% of wilted leaves and on the AUDPC (Table 1). No wilting was observed on the control plants in either experiment.

Table 1 Impact of plant age at the time of inoculation with Erwinia tracheiphila on the rate of wilt and disease progression in muskmelon (Cucumis melo cv. Athena) plants in two growth chamber experiments
Full size table

Bacterial movement

In each of two replicate experiments to evaluate bacterial movement from the site of inoculation, E. tracheiphila cells were first recovered from stem samples 7 days postinoculation (dpi) and were isolated from sites both above and below the inoculation point (Fig. 1). There was evidence of upward movement of E. tracheiphila, determined by recovery of bacteria from stem segments above the inoculation site at 7, 14 and 21 dpi, and at sites as far as 5 and 8 internodes above the inoculation site in the first and second runs of the experiment, respectively. Stem growth above the inoculation point continued until 14 dpi on plants that had been 4, 6, or 8 weeks old at inoculation, but the pathogen was not recovered from the uppermost internodes of the stem even at 21 dpi, indicating that its movement in the upward direction was limited.

Fig. 1
figure 1

Directional movement of E. tracheiphila in stems of muskmelon (Cucumis melo cv. Athena) plants inoculated at different ages. The results of two replicate experiments are shown in (a) and (b). Data shown are the mean number of internodes from the site of inoculation to the point where E. tracheiphila cells were recovered (black bars) or were not recovered (hatched bars) from stem segments. Plants were 2, 4, 6, or 8 weeks old when the youngest true leaf was inoculated. The zero point on the y-axis indicates the point of inoculation. Positive numbers on the y-axis indicate the number of internodes above the inoculation point, whereas negative numbers indicate number of internodes below the inoculation point. Data are means of four replicates per treatment. * indicates sampling times for which the bacterial movement from the site of inoculation was significantly (Student’s t-test, p < 0.05) greater in the downward than upward direction. n = 4. # = indicates samples that were too desiccated to attempt to recover E. tracheiphila

Full size image

At each sampling time, E. tracheiphila moved much further downward than upward, with movement sometimes limited by reaching the lowest possible node (Fig. 1). In contrast to the upward movement to a maximum of 5 to 8 internodes above the inoculation site, bacteria moved downward as far as 9, 13, and 23 nodes by 7, 14 and 21 dpi, respectively, in the first run of the experiment regardless of plant age at inoculation (Fig. 1A), and to similar distances in the second run of the experiment (Fig. 1B). On plants inoculated at 6 or 8 weeks of age, E. tracheiphila cells were detected in nearly all of the stem segments sampled below the inoculation point at 21 dpi, and the distance of movement of the bacterium at 14 and 21 dpi was significantly (p < 0.05) greater below than above the point of inoculation.

In both experiments, the 2-week-old and 4-week-old plants died before reaching 14 and 21 dpi respectively; thus, data for these treatments are not presented for those time points. By 21 dpi in both runs of the experiment, the stems of all of the plants that were 6 weeks old at inoculation had withered and died above the inoculation point; isolations from that portion of the plant in the first run were not attempted because the pathogen does not survive in dead host tissue (Latin 2000). In general, E. tracheiphila reached the lowest node of each plant by the time of the plant’s death.

Discussion

Our results provide the first experimental evidence that the rate of wilting of Cucumis sp. crop by E. tracheiphila is impacted by host age, with young plants developing wilt symptoms significantly faster than older ones based on the time required for 50% of the leaves to wilt. Although grower guides frequently state that young cucurbit plants are more susceptible to infection than older plants (Brust 1997a; Watterson et al. 1971), experimental evidence supporting this assertion is absent for Cucumis spp. In the Brust (1997a) study, more wilt was observed in pumpkin seedlings that had been inoculated at the cotyledon stage than at the later stage of ≥1 true leaf, although most of the seedlings with true leaves never showed wilt and recovered from the inoculation. The general trend observed in the Brust study – increasing resistance with increasing plant age – agrees with results of the present study of muskmelon.

Knowledge that susceptibility to bacterial wilt decreases as muskmelon plants age has important management implications. For example, our evidence that plants are more susceptible to bacterial wilt when they are young can help to inform optimal timing for reduced-insecticide management strategies such as the deployment of row covers as protective barriers against cucumber beetles (Mueller et al. 2006; Saalau Rojas et al. 2011). In particular, although row covers should be deployed during the most susceptible period to suppress bacterial wilt, deciding when to remove them should factor in both the probability of pathogen transmission by cucumber beetles (Saalau Rojas et al. 2011) and plant phenology related to resistance. In addition, clarifying responses to inoculation as a function of seedling age should help plant breeders to optimize screening assays for bacterial wilt resistance. For example, using plants that are 4 weeks old at inoculation might be a cost-effective option because they are relatively small and thus require minimal growth space, and moderate in susceptibility when compared with 2-, 6-, and 8-week-old plants. Moreover, their requirement for about 8 dpi to show symptoms enables sufficient observation time to compare symptoms among breeding lines. However, additional, season-long experiments in the field would be needed in order to comprehend impact on yield and fruit quality when plants become infected at later stages of the growing season.

The mechanism responsible for a decreased wilting rate as muskmelon plants age is unclear. Ontogenic resistance could result from factors such as increased production of phytochemicals and/or the development of physical barriers that slow disease progress (Panter and Jones 2002). Alternatively, the pathogen may be diluted if plant growth exceeds pathogen growth, which could slow the rate of wilting. The correlation of plant age with plant size makes it difficult to separate the impacts of age versus size; this is particularly true for cucurbit crops, most of which increase rapidly in size during the early part of the growing season. It is possible that both ontogenic and plant size-related factors may operate to slow wilting in older, larger plants. Further experimentation will be required to unravel the causal mechanisms of this phenomenon.

Our experiments are among the first to trace the internal movement of E. tracheiphila following infection. In common with other xylem-limited vascular wilt diseases, symptom expression as indicated by visible wilt progressed distally beyond the nodes from which the pathogen was recovered, presumably as sieve plates became blocked and water flow ceased (Holland et al. 2014; McElrone et al. 2003). Such a blockage of the upward movement of water inside the vascular system may help to explain our observation that E. tracheiphila moved in a primarily downward rather than upward direction from the inoculation point. Vrisman et al. (2016) provided observational evidence that a bioluminescent-labeled strain of E. tracheiphila moved not only upward from the inoculated leaf of a muskmelon seedling but also downward into the roots. The mechanism driving bacterial movement against the xylem flow is unclear, but is consistent with the movement exhibited by two other xylem-limited vascular wilt pathogens: Xylella fastidiosa, the causal agent of Pierce’s disease of grape (Meng et al. 2005), and Acidovorax avenae subsp. citrulli, the causal agent of bacterial fruit blotch of cucurbits (Bahar et al. 2010). X. fastidiosa, a nonflagellated bacterial pathogen, was shown to spread in the xylem via motility mediated by type IV pili (Meng et al. 2005). Type IV pili, which are like grappling hooks, enable the bacterial cells to jerk forward along a surface in a form of motility known as twitching (Mattick 2002). X. fastidiosa mutants defective in these pili showed reduced downward colonization of the xylem (Meng et al. 2005). As a xylem-limited pathogen, E. tracheiphila may employ a similar mechanism of movement. This is consistent with the recent discovery that the putative type IV pili genes are conserved among Erwinia spp., including E. tracheiphila (Shapiro 2012). Our study, the first to quantify the relative movement of the pathogen upward and downward following infection, is a further step in understanding the movement of E. tracheiphila in the xylem and helps to set a foundation for evaluating the role of pilus genes in downward movement.