Flowers can provide a protected and nutrient-rich environment to the epiphytic microflora, thus representing a sensible entry point for pathogens such as Pseudomonas syringae pv. actinidiae (Psa). This bacterium can colonize both male and female Actinidia flowers, causing flower browning and fall, and systemic invasion of the host plant, eventually leading to its death. However, the process of flower colonization and penetration into the host tissues has not yet been fully elucidated. In addition, the presence of Psa in the pollen from infected flowers, and the role of pollination in the spread of Psa requires confirmation.
The present study employed a Psa strain constitutively expressing the fluorescent GFPuv protein, to visualize in vivo flower colonization. Microscopy observations were performed by means of confocal laser scanning and wide-field fluorescent microscopy, and were coupled with the study of Psa population dynamics by quantitative PCR (q-PCR). The pathogen was shown to colonize stigmata, move along the stylar furrow, and penetrate the receptacles via the style or nectarhodes. Once the receptacle was invaded, the pathogen migrated along the flower pedicel and became systemic. Psa was also able to colonize the anthers epiphytically and endophytically. Infected male flowers produced contaminated pollen, which could transmit Psa to healthy plants. Finally, pollinators (Apis mellifera and Bombus terrestris) were studied in natural conditions, showing that, although they can be contaminated with Psa, the pathogen’s transmission via pollinators is contrasted by its short survival in the hive.
Among the production constraints of kiwifruit, the bacterial canker, caused by Pseudomonas syringae pv. actinidiae (Psa), has been the major limiting factor in the main growing areas worldwide since its pandemic outbreak in 20081,2,3. The disease affects various species of the genus Actinidia, including the two most important varieties for commercial purposes, A. chinensis var. deliciosa and A. chinensis var. chinensis4,5,6. In Italy, the first outbreak of bacterial canker was reported in 19947, but serious economic damage started being observed with the spread of a highly virulent pathogen, named biovar 3, genetically separated from previously identified biovars8,9,10,11,12.
Many aspects of Psa biology and epidemiology still require in-depth investigation. Among them, the elucidation of the primary entry points and mechanisms of spread from plant to plant are of crucial importance. A few short communications have listed plant organs in which the pathogen was detected13,14. However, the detection of an endophytic pathogen within a specific plant organ, not accounting for its ability to move through plant tissues, does not imply a primary apoplastic invasion through the same organ. In addition, these studies were mainly performed in A. chinensis var. chinensis, allowing the isolation of Psa from a number of natural openings and wounds such as symptomatic lenticels, lesion margins, buds, fruit stalks, leaf petioles, and flowers13,14.
Typical flower symptoms consist of browning of petals and sepals in early stages of infection. Subsequently, Psa-infected blossoms may wither before opening, or fall off soon after fruit set. However, such symptoms are not specific to Psa, being similar to those of blossom blight caused by Pseudomonas viridiflava and P. syringae pv. syringae15,16,17,18.
Although female flowers can provide bacteria with a protected and nutrient-rich environment14, the process of flower colonization by Psa and its penetration into the host tissues has not been yet fully elucidated. The precise knowledge of the mechanisms and time of the invasion of the host tissue is essential to understand the disease cycle and to develop effective control strategies. Despite the fact that more than 100 bacterial species are known to be plant pathogens, pollen-mediated transmission of bacterial pathogens has only been postulated for Xanthomonas arboricola pv. juglandis19,20, Erwinia amylovora21 and Pseudomonas syringae22, even though they are not the only pathogens which may infect flowers, and compelling evidence for pollen-mediated bacterial spread in real conditions is still missing. In Erwinia amylovora, which is the most widely studied bacterial pathogen leading to systemic infection by flower colonization23,24, the production of contaminated pollen by infected flowers was achieved only in experimental conditions, by applying a highly concentrated (108 cells mL−1) bacterial suspension directly on anthers21, while the production of contaminated pollen in natural conditions has never been described.
In addition, E. amylovora and other bacterial pathogens may rely on pollinating insects for secondary spread25,26. Several insect species, such as honeybees (Apis mellifera) and bumblebees (Bombus spp.), contribute to kiwifruit pollination27. However, although Actinidia spp. flowers are fragrant, their attractivity to pollinators is poor, and some of their characteristics (pendulous position, high pollen:ovule ratio, large multibranched stigmatic surfaces, scarce production of nectar) suggest that wind may be an important agent of pollen transfer28,29.
To obtain export-quality fruit, more than 2000 pollen grains should reach the stigma leading to fruits with 1000–1400 viable seeds27. Since pollination dictates the size and the quality of the kiwifruit berries, pollen collection and artificial pollination have become a widespread practice to enhance fruit production and size30,31.
Therefore, understanding the importance of pollen and pollination for the dissemination of Psa from plant to plant is a crucial step for the control of bacterial canker.
Several studies showed that Actinidia spp. pollen or commercial pollen preparations can be contaminated by Psa13,32,33,34,35,36. Preliminary results on the pollen-mediated vectoring of Psa were reported by Tontou et al.37. However, limited information is currently available on a number of crucial aspects regarding the actual presence of Psa in the pollen in field conditions, and the consequent epidemiological risk. In fact, although the viable Psa population on pollen grains was quantified in some experiments, crucial pieces of information (such as the minimal Psa population on pollen needed to effectively transfer Psa on flowers, the infection threshold of flowers, or the overall pollen-mediated infection efficacy) are still missing.
The present study elucidates the different phases of flower colonization by Psa till the systemic invasion of the host plant. Moreover, the effect of infection on fruit production and colonization by Psa was evaluated. Finally, this study provides the first evidence on the production of infected pollen by diseased male plants, and the role of pollen and pollination in the dissemination of Psa.
The colonization of the different flowers parts was studied in vivo and real-time by combining the use of a green fluorescent protein (GFP)-labeled Psa strain (CFBP7286-GFPuv) with fluorescent stereomicroscopy and confocal laser scanning microscopy. These methodologies allowed the univocal identification of the bacteria and the in vivo, non-invasive observation of plant tissues13,24. With the possible exception of fire blight21,38 (E. amylovora), this is the first, non-speculative report of pollen-mediated transmission of a phytopathogenic bacterium.
Materials and methods
Experiments in controlled conditions were performed on kiwifruit cultivars Hayward (female, A. chinensis var. deliciosa), Tomuri (male, A. chinensis var. deliciosa), Hort16A (female, A. chinensis var. chinensis) and CK2 (male, A. chinensis var. chinensis), using 4-year-old potted plants maintained in plastic pots (40 cm diameter) with a mixture of peat and sand (1:1, v/v). Standard irrigation and fertilization was applied. As mineral fertilizer, Poly-Feed (16N-8P-32K) by Fertica S.A. was used (18 g per plant applied two times). During the experiments, the plants were kept in glasshouse conditions under natural light. Temperature and relative humidity were, respectively, 22–24 °C and 70–80%. After inoculation, the relative humidity was raised to 100% for the following 48 h.
The Psa strain CFBP7286-GFPuv was used for artificial inoculation. This genetically engineered strain, retaining the virulence of the corresponding wild type (CFBP7286, biovar 3), expresses a green fluorescent protein and kanamycin resistance13. The inocula were prepared by cultures grown in liquid Luria-Bertani medium under moderate shaking at 27 °C, then resuspending the bacteria in 10 mM MgSO4. The titer of bacterial suspensions was determined by plating 10-µL drops of serial 1:10 dilutions on Luria-Bertani medium containing 15 g L−1 of agar amended with cyclohexamide (100 mg L−1). In experiments using infected samples from commercial fields, Psa identity in serial dilutions was confirmed by q-PCR39.
Experimental inoculation of Actinidia spp. flowers
Several inoculation methods of a Psa suspension at 106 colony-forming units (cfu) mL−1 were used for each plant cultivar: (i) spraying the whole flower, (ii) dipping the stigmata or anthers in the bacterial suspension, or (iii) pipetting 20 µL of the same bacterial suspension on the calyx. Sixty flowers were inoculated with each of the three different inoculation methods. Water-treated flowers were used as the control. All the female flowers were artificially pollinated 6 h before inoculation.
The tissues (stigmata/anthers, ovarium, and other flower parts) of six spray-inoculated flowers were separated at 0, 3, 24, 48, 72, 96, and 120 h after inoculation, and the epiphytic Psa population was measured on them. The samples were vigorously washed for 5 min in 5 mL of sterile 10 mM MgSO4 solution.
To assess endophytic Psa populations, the same tissue samples were surface-sterilized by washing in ethanol (70%, 3 min) and NaOCl (0.1%, 3 min), followed by two rinses in sterilized water. Then they were homogenized in 10 mM MgSO4 and the bacterial population was assessed as previously described. The complete efficacy of the sterilization method was confirmed by comparing the epiphytic Psa population of unsterilized and sterilized organs. To assess Psa endophytic population in the flower-bearing cane (i.e., stalk), a wood section of 1 cm in length on either side of the flower pedicel was excised and analyzed. Each experiment was repeated three times.
Psa presence on flowers and pollen in infected orchards
The sampling of flowers from male and female plants was performed in three commercial kiwifruit orchards in Faenza region, Italy (44°17'00.0“N 11°53'00.0“E), with a disease incidence of 50% (assessed as the proportion of vines showing symptoms).
The incidence of Psa and other kiwifruit pathogens (P. syringae pv. syringae, P. viridiflava) was determined on 600 samples of A. chinensis var. chinensis (cv. Hort16A) and A. chinensis var. deliciosa (cv. Hayward) female flowers showing browning symptoms. On flowers infected by Psa, the bacterial population of this pathogen was quantified.
The natural occurrence of Psa on asymptomatic male flowers and pollen was determined on closed (BBCH 53–55) and open flower samples (BBCH 60) (100 for each type and for the two A. chinensis varieties).
Anthers (from samples before dehiscence stage in closed flowers) or pollen (from fully open flowers) were aseptically separated from the remaining tissues of ten-flower samples. Before dissection of closed flowers, the external surface was sterilized as previously described. Pollen was washed with 20 mL of sterile water for 15 min on a rotary shaker at 120 rpm until full rehydration. Subsequently, the samples were precipitated by centrifugation (90 s at 10,000 × g), and the supernatant was used for bacterial enumeration.
Pollen germination was assessed as the percentage of granules emitting a tube after 5 h at 25 °C under shaded light, in a medium containing 10% sucrose and 15 g L−1 agar.
In 2016, to verify the possibility of contamination of pollen via an endophytic invasion, three 1-year -old A. chinensis var. chinensis potted male vines were spray-inoculated at bud break with CFBP7286-GFPuv (108 CFU mL−1 in 10 mM MgSO4). Plants were kept in controlled conditions till flower development. At the beginning of flowering, ten opening flowers were collected from each plant and anthers were aseptically excised. Presence of Psa was assessed by serial dilution plating and q-PCR, as previously described.
Flower colonization by pollen-vectored Psa
Pollen samples (0.5 g) of A. chinensis var. chinensis (cvs. CK2 and Belen) and A. chinensis var. deliciosa (cv. Tomuri) were artificially inoculated with 1.5 mL of a suspension of CFBP7286-GFPuv (107 CFU mL−1 in 10 mM MgSO4) obtained from fresh cultures in liquid LB medium. The wet contaminated pollen was dried at room temperature for 72 h. For each pollen sample, the bacterial population was monitored over 5 days. Before using contaminated pollen for inoculation of flowers, Psa population was determined and adjusted to 4.1 × 106 cfu g−1 by adding uncontaminated pollen. Pollen viability and germination were assessed as previously described.
Hand-pollination was performed on 4-year-old A. chinensis var. chinensis (cv. Hort16A) and A. chinensis var. deliciosa (cv. Hayward) potted plants, by dusting Psa-free pollen with a wet brush on each flower. In a second group of plants, infected pollen was used instead. Nine plants divided in three replicates were used for each treatment. Endophytic and epiphytic Psa populations were determined 5 and 10 days after inoculation in the different flower organs and in the bearing cane. The same experiments were repeated using naturally infected A. chinensis var. chinensis (cvs. CK2 and Belen) and A. chinensis var. deliciosa (cv. Tomuri) pollen. In this case, the Psa population was determined and adjusted to 1.2 × 104 cfu g−1.
Pollen-mediated vectoring of Psa in natural conditions
To mimic the pollen-mediated vectoring of Psa in natural conditions, flowering branches of male A. chinensis var. chinensis (cv. Belen) were detached from infected plants in a commercial orchard. Twelve branches were transferred in controlled semi-field conditions, maintained in water and placed for 5 days in proximity (1 m) to nine healthy A. chinensis var. chinensis (cv. Hort16A) potted plants divided in three replicates. To prevent uncontrolled diffusion of Psa in the environment by insects or water leach, plants were placed under a tunnel built with anti-aphid net (Antiafide 20/10 Mesh 50, Artes Politecnica s.r.l, Vicenza, Italy). The tunnel dimensions were 12 × 3.50 × 5.50 m. The floor was covered with an impermeable plastic film.
Psa load on naturally infected pollen was quantified as previously described, and it resulted to be 3 × 106 cfu g−1.
A similar experiment was repeated on A. chinensis var. deliciosa (cv. Hayward) using infected Tomuri vines.
Incidence was calculated as the percentage of flowers presenting a systemic endophytic Psa population inside the stalk (i.e., flower-bearing canes). Epiphytic and endophytic Psa populations were determined 5 and 10 days after inoculation in the different flower organs and in the bearing cane.
Systemic colonization of the host plant upon flower inoculation
To monitor the dynamics of systemic invasion upon flower infection, on six A. chinensis var. chinensis (cv. Hort16A) and six A. chinensis var. deliciosa (cv. Hayward) potted plants, a single flower cluster per plant was inoculated by using pollen from male plants of the same variety (cvs. Belen and Tomuri, respectively), contaminated by CFBP7286-GFPuv (4.1 × 104 CFU g−1).
The evolution of flower symptoms, fruit set, development, and fall were monitored weekly. The endophytic migration of Psa was assessed 4 months after inoculation. Tissue samples were taken from cane tops, grafting points, rootstock wood and roots. Tissue samples were surface-sterilized by washing in ethanol (70%, 3 min) and NaOCl (0.1%, 3 min), followed by two rinses in sterilized water. Successively, they were homogenized in 10 mM MgSO4 and the bacterial population was assessed as previously described. The experiment was repeated twice.
Psa vectoring by pollinators
The experiments were performed in 2 consecutive years in commercial orchards located in the area of Latina (41°28′03.35″N 12°54′13.32″E), with 15 to 20% disease incidence (assessed as the percentage of vines showing symptoms). Five commercial bumblebee hives (Natupol® Koppert B.V., The Netherlands) were placed in each orchard at 75% flowering. Five days after beehive positioning, bumblebee capture was performed for 3 consecutive days. Each sampling consisted of ten bumblebees captured after return to the beehive. Pollen harvested by bumblebees was also collected and analyzed. Psa was isolated and identified as previously described for other substrates.
To verify Psa survival on honeybees and in honeybee hives, a set of tubes was closed with microbiological filters. In each tube, a dead insect, 0.2 g wax, or 20 mg A. chinensis var. deliciosa (cv. Tomuri) pollen were placed, after inoculation with 10 µL of a Psa suspension with a bacterial titer of 106, 108 or 1010 cfu mL−1. The biological samples were maintained in a honeybee hive, together with a data logger recording temperature and relative humidity, and Psa population was monitored.
Following inoculation with bacterial suspensions or with artificially contaminated pollen, Psa distribution over flower tissues, and migration through them was monitored by microscopical observation.
Sample preparation and initial observations were performed using a motorized Nikon SMZ25 (Nikon, Tokyo, Japan) with a zoom ratio of 25:1 (zoom range: 0.63 × - 31.45 × ) fitted with a BHS (GHS) filter set. The GHS filter set (excitation light 450–490 nm, emission 510 nm) was used to visualize the GFPuv-labeled bacteria.
Optical sections were obtained with a Nikon C1-S (Nikon, Tokyo, Japan) confocal laser scanning microscope (CLSM) and equipped with an Argon laser. A 40, 60 and 100 × Nikon PlanApo objectives and the BHS (GHS) filter set were used for image acquisition. Images were acquired and analyzed by the NIS-Elements C Microscope Imaging Software.
Psa populations, at different time points and in different A. chinensis flowers parts, were analyzed by one-way analysis of variance (ANOVA) followed by means separation with Fisher’s least significant difference (LSD, P < 0.05). Incidence of male and female flowers contaminated by Psa in different A. chinensis variety was analyzed by Z-test (P < 0.05). For statistical analysis the software STATISTICA 7.0 (StatSoft Inc, Tulsa, USA) was used.
Population of Pseudomonas syringae pv. actinidiae on Actinidia spp. flowers
Psa was able to establish a detectable population on all the flowers parts of both A. chinensis var. chinensis and A. chinensis var. deliciosa female flowers (Fig. 1a, b). The highest population was observed on stigmata and styles with an average value of 3.6 ± 2.1 × 105 cfu for the whole duration of the experiment. Other flower parts, in A. chinensis var. deliciosa, hosted such a high population only transiently. In contrast, in A. chinensis var. chinensis, Psa population constantly ranged between 104 and 106 on all flower tissues (Fig. 1a, b). On male flowers, Psa attained a lower population, showing a substantial decrease after 48 h corresponding with tissue senescence (especially anthers; Fig. 1c, d). No differences were observed between A. chinensis var. chinensis and A. chinensis var. deliciosa male flowers.
The inoculation methods influenced the endophytic colonization of the different flower organs. Within floral tissues, the pathogen population was generally higher in A. chinensis var. chinensis tissues than in A. chinensis var. deliciosa (Supplementary Fig. S1). After inoculation by spray, the highest bacterial populations were found in the flower stalk. In female flowers, the dipping-inoculation of the stigma led to colonization of the flower pedicel and stalk only in A. chinensis var. chinensis, whereas on A. chinensis var. deliciosa, Psa was isolated from those flower parts only occasionally. After the infection of the ovarium, the pathogen was mainly found in the gynoecium and calyx (Supplementart Fig. S1).
On male flowers, the isolation of the pathogen from the flower tissues was achieved only after spray inoculation. Also on male flowers, the highest Psa populations were isolated from A. chinensis var. chinensis.
After spray inoculation, 75% of A. chinensis var. chinensis and 53% of A. chinensis var. deliciosa female flowers were infected stricto sensu (i.e., they harbored an endophytic Psa population). The infection of male flowers occurred, in both species, in approximately 60% of the inoculated samples.
The microscopic observations show that the bacterium, after the colonization of the stigma, moves along the style and reaches the ovarium surface both in A. chinensis var. deliciosa (Fig. 2) and A. chinensis var. chinensis (Fig. 3). Psa can colonize the outer surfaces of the style by multiplying among the papillae and invading the stylar groove (Figs. 2b, c and 3d). However, it can also move along the style without a direct colonization of the stylar groove (Fig. 3e, f). Moreover, when bacteria establish an endophytic population, they can penetrate the conductive stylar tissue and move along the tracheids (Supplementary Fig. S2). On the ovarium surface, high pathogen populations were observed especially in A. chinensis var. chinensis (Fig. 1). In addition, penetration into the ovarium from the style has also been found (Fig. 2d, e).
Growth of Psa on pollen and pollinated flowers
The infection of anthers, even though not supporting higher pathogen populations than other tissues, resulted in contamination of pollen grains (Figs. 1c, d and 4). Anther colonization was often scattered, with small and dispersed areas of each anther colonized by the bacterium, possibly indicating an external, stochastic source of inoculum (Fig. 4b, c). In other cases, Psa colonization becomes visible only after anther dehiscence suggesting a possible internal infection of the anther via pathogen systemic migration along the filament (Fig. 4f). Whenever pollen contamination was observed, the infection of tapetum was also found (Fig. 4g–i). At dehiscence, infected anthers released in the stroma both contaminated pollen grains and free Psa cells (Fig. 4g, h, k). Only a small number of pollen grains were contaminated by Psa cells. In all these cases, Psa was found associated with the pollen grain surface (Fig. 4h–j). Infection of pollen grains (i.e., penetration of Psa into the microgametophytes) was never observed.
When spray-inoculated at bud break, before flower development, A. chinensis var. chinensis male vines produced infected anthers in (3 ± 1.2)% of flowers.
Psa was able to survive on pollen grains (Fig. 5a) for at least 5 days, as confirmed by microscopical observation. After the pollination of A. chinensis var. chinensis flowers with infected pollen, Psa initially colonized the stigma and, thereafter, all the floral tissues, reaching the stalk in less than 10 days (Fig. 5b).
Following the exposure of healthy A. chinensis var. chinensis plants to infected male vines, the natural inoculation via contaminated pollen led, in 10 days, to 100% of contaminated flowers with (48 ± 12)% of them resulting in systemic invasion of the plant with an average endophytic Psa population of 104 cfu g−1 FW.
In A. chinensis var. deliciosa, natural inoculation by contaminated pollen resulted in (44 ± 18)% of flowers harboring an epiphytic Psa population, whereas (25 ± 13)% of flower infections led to systemic invasion.
Systemic colonization of plant tissues after pollen-mediated inoculation
After the growth on stigma and stylar furrow, which offers a nourishing and protected environment, Psa established an endophytic population inside the style (Supplementary Fig. S2 and Fig. 5) and the ovarium (Figs. 2 and 5). In particular, Psa was localized at the conjunction between the conductive stylar tissue and the central placenta, thus suggesting the penetration of the ovarium via migration inside the style. Penetration via nectarthodes was not observed in these experiments. Once inside the ovarium, Psa can colonize the vascular bundles, which may lead to the migration to flower pedicels. The migration of Psa via flower pedicel was observed by microscopical analysis (Fig. 2f) and resulted in systemic infection of the plant (Fig. 5b). Leaf spots were observed 2 months after flower inoculation.
In rare cases, highly infected female flowers produced droplets of bacterial exudate emerging with the nectar. Under high humidity conditions, the bacterial exudates dropped along the pedicel till reaching the abscission zone which, at flower drop, may represent a possible entry point.
Four months after inoculation of a single flower, the bacterium could be found in all the tissues (canes, grafting point, rootstock, roots), independent of the source of pollen infection (natural or experimental) and its bacterial load (Table 1), showing both acropetal and basipetal migration.
In the same plants, all fruit dropped within 75 days after flowering, when the fruit diameter was below 40 mm. In contrast, control (healthy) plants only dropped 21% of their fruit load in 84 days (end of monitoring period).
Flower blight incidence and contamination of pollinators in infected orchards
On A. chinensis var. chinensis female flower samples showing browning symptoms, Psa incidence was 82%, with an average population of 3.3 ± 1.6 × 108 cfu per flower. In contrast, 17% of these flowers were infected with either P. syringae pv. syringae or P. viridiflava. Psa was also found on 40% of asymptomatic flowers, but with an average population 1000 times lower than in symptomatic flowers (Table 2). The Psa incidence and epiphytic population were substantially similar between A. chinensis var. chinensis and var. deliciosa female flowers (Table 2).
On male flowers, anthers (from closed flowers) and pollen (from open ones) were separately analyzed and Psa was detected in 56% and 100% of the open flowers in A. chinensis var. chinensis var. deliciosa, respectively (Table 2). In closed flowers, the presence of Psa is limited to the outer side of the flower, while the anthers are generally not reached by the bacterium (Table 2). However, Psa was found in the anthers or pollen with a significant bacterial population (average 2.2 ± 1.3 × 102 cfu g−1), independently of flower phenological stage and presence of symptoms. This pollen was characterized by a very low viability, in no case higher than 5%.
During the 3 years of monitoring (2014–2016) of infected A. chinensis var. deliciosa orchards, about 20% of the worker bumblebees were contaminated with Psa (Table 3). On these individuals, the average Psa population in the different years was 2.8 ± 2.6 × 103 cells per insect. Significant Psa populations were also detected on the pollen harvested by bumblebees and stored in the hive (Table 3). About 40% of the pollen samples were contaminated by Psa (Table 3). On experimentally inoculated honeybee bodies, wax, and pollen maintained in a honeybee hive, viable Psa units were not detected 48 h after inoculation.
Colonization of flowers by Psa
Psa was able to densely colonize flowers. The styles hosted the highest populations ( ≥ 105 cfu), possibly because of nutrients and moisture secreted by the stigma. Male flowers are less conducive for Psa growth.
A. chinensis var. chinensis female flowers supported more uniform bacterial populations in the different flower parts, compared to A. chinensis var. deliciosa (Fig. 1). Since Psa attains similar populations on the styles of both A. chinensis varieties, the chemical composition of flower secretions is probably equally favorable to Psa growth. Morphological differences may explain the more widespread colonization of Psa on A. chinensis var. chinensis. In this variety, the denser trichome coverage on leaves, shoots, and flowers may promote moisture preservation, which is a key factor for flower-colonizing bacterial communities40. Furthermore, trichomes may also represent an entry point for the pathogen13.
Contamination of male flowers also resulted in the production of contaminated pollen. The contamination is substantially lower in closed than in open flowers, suggesting that pollen is generally contaminated by external inoculum sources (e.g., rain, wind, insects). However, contaminated pollen was also found in 5–10% of the closed flowers, where an external source of contamination can be excluded. Moreover, the experiment performed by inoculating, before flowering, male vines with CFBP7286-GFPuv confirmed this finding. Therefore, a systemic acropetal migration of Psa to anthers from other flower tissues could be responsible for this result. Nonetheless, an epiphytic contamination at bud break of developing sepals and petals and the migration of the pathogen to the inner tissues of flowers till reaching the anthers cannot be completely excluded. Finally, Psa populations on pollen were independent to the Psa population on flowers.
The apparently different incidence of Psa in open A. chinensis var. chinensis and A. chinensis var. deliciosa male flowers and pollen is probably due to a different disease pressure inside the orchards and to the different climatic conditions at flowering. In fact, A. chinensis var. deliciosa flowering takes place approximately 20 days after A. chinensis var. chinensis, when the climatic conditions (higher temperatures, lower relative humidity, lower occurrence of rainfall, and leaf wetting) may be less conducive for epiphytic Psa growth.
Invasion of host tissues
Unlike fungal pathogens, bacteria are generally unable to actively penetrate plant tissues, and rely on wounds and natural openings, such as stomata, lenticels, and nectarthodes41 to access the host’s apoplast. Flowers have been previously suggested to be among the major entry points for Psa, although no experimental evidence was available42,43,44. In this view, flower infection would lead to systemic invasion of the whole plant.
In this work, high bacterial populations, showing that Psa was multiplying, were found on stigmata, suggesting that the stigma is a strong candidate for initial flower infection. Supporting this view, the direct dip-inoculation of stigmata caused the highest pathogen population within the flower pedicel. Subsequently, the bacterium moves along the stylar furrow and stylar connections toward the ovarium, where further multiplication and penetration to the apoplast may occur. The ovarium surface is, in fact, rich in trichomes and represents a suitable environment for the pathogen. In contrast, direct penetration from the calyx into the ovarium is unlikely, since there are no natural connections between these organs.
Bacterial plant pathogens may have an obligate infection path through flower tissues. E. amylovora, for instance, migrates along the style and the stylar furrow24 before accessing the ovary tissue. In contrast, Psa can also penetrate into the apoplast from the stylar surface (Supplementary Fig. S2). In this case, the bacterium was found in the conductive tissues, but not in the tracheids. This observation may suggest that the pathogen is unable to move through the xylem sieves of small vessels.
Further migration to the flower pedicel and flower stalk has also been observed. The invasion of flower stalks eventually leads to systemic infection.
Male flowers collected in the field may also be infected by the pathogen, with similar susceptibility and symptomatology to female flowers. However, the infection path through their tissues is not as clear as in female flowers. Since the presence of Psa in the flower stalk (Fig. 1) may be due to accidental contamination of the stalk itself during the experimental inoculation, the systemic penetration of the bacterium via male flowers still needs further verification.
Ecological significance of pollen-mediated Psa spread and role of pollinators
Closed, asymptomatic flowers collected from symptomatic plants showed to be infected with Psa and to produce contaminated pollen, thus demonstrating that Psa may endophytically migrate to developing flowers in infected plants, leading to the contamination of anthers and the production of contaminated pollen. Paradoxically, Psa was more frequently found (although with lower populations) in asymptomatic flowers, and on pollen derived from asymptomatic flowers, than on symptomatic ones. It may be suggested that symptomatic flowers drop from the vine as its metabolism is no longer effective. This hypothesis is further supported by the observation that, in controlled conditions, all flowers infected via contaminated pollen dropped within 75 days after pollination.
Evidence of pollen contamination with Psa was previously provided32,36. However, the attempts to establish the ecological significance of pollen-vectored Psa in the development of bacterial canker33,37 only led to circumstantial evidence, either because of the artificial conditions of laboratory experiments, or because of the inability to track the source of infection and the bacterium during the various stages of pathogenesis in field conditions. Crucial information was missing about ecological aspects (such survival on flower parts and pathogen infection pathways), presumably affecting the chance of pollen-mediated Psa spread in field conditions.
Our research aimed to fill the information gap between flower inoculation and whole plant colonization, by visualizing the infection path exploited by Psa to penetrate the host plant’s apoplast. Our observations strongly suggest pollen-mediated flower contamination as one of the main infection pathways in field conditions.
Psa was found to survive on honeybees and in hives45. However, the importance of honeybee-mediated infection in field conditions is not clear. In this work, limited survival of Psa was observed in hive conditions; notably, the initial population in each sample had little influence on the survival time, suggesting that hive temperature (35 °C) and/or antimicrobial compounds, naturally present on the biological samples taken from the hive, may prevent Psa growth34. In addition to honeybees, significant Psa populations were also found on bumblebees foraging in infected fields. Overall, although direct evidence for insect-mediated Psa vectoring is still missing, the bacterial populations in hives and on insects are compatible with those employed for successful plant inoculation by hand-pollination (approx. 104 cfu g−1 pollen), but only for a short time after contamination. Thus, pollinators may be involved in short-range, secondary Psa spread.
This work demonstrates that flower tissues, and particularly stigmata, are crucial sites for Psa growth and penetration into the host tissues. Moreover, our work reports the first evidence of production of contaminated pollen by asymptomatic closed male flowers from infected plants and the role of pollen-mediated dispersal of a bacterial plant pathogen in natural conditions. Thus, plant protection strategies should consider the cultural practices involving flowers (for instance, inspection of pollen-donor plants, sanitation of pollinators, or the use of biological control agents competing with Psa in the floral niche) to minimize the risk of disease.
Vanneste, J. L. Recent progress on detecting, understanding and controlling Pseudomonas syringae pv. actinidiae: a short review. N. Z. Plant Prot. 66, 170–177 (2013).
Donati, I. et al. New insights on the bacterial canker of kiwifruit (Pseudomonas syringae pv. actinidiae). J. Berry Res. 4, 53–67 (2014).
Greer, G. & Saunders, C. The costs of Psa-V to the New Zealand kiwifruit industry and the wider community. In: Report to kiwifruit VineHealth. Lincoln University press, New Zealand: Agribusiness and Economics Research Unit, 2–9 (2012).
Balestra, G. M., Renzi, M. & Mazzaglia, A. First report of bacterial canker of Actinidia deliciosa caused by Pseudomonas syringae pv. actinidiae in Portugal. New Dis. Rep. 22, 10 (2010).
Koh, Y. J., Kim, G. H., Jung, J. S., Lee, Y. S. & Hur, J. S. Outbreak of bacterial canker on Hort16A (Actinidia chinensis Planchon) caused by Pseudomonas syringae pv. actinidiae in Korea. N. Z. J. Crop Hort. Sci. 38, 275–282 (2010).
Vanneste, J. L. et al. First report of Pseudomonas syringae pv. actinidiae, the causal agent of bacterial canker of kiwifruit in France. Plant Dis. 95, 1311 (2011).
Scortichini, M. Occurrence of Pseudomonas syringae pv. actinidiae on kiwifruit in Italy. Plant Pathol. 43, 1035–1038 (1994).
Balestra, G. M., Mazzaglia, A., Quattrucci, A., Renzi, M. & Rossetti, A. Current status of bacterial canker spread on kiwifruit in Italy. Aust. Plant Dis. Notes 4, 34–36 (2009).
Ferrante, P. & Scortichini, M. Identification of Pseudomonas syringae pv. actinidiae as causal agent of bacterial canker of yellow kiwifruit (Actinidia chinensis Planchon) in Central Italy. J. Phytopathol. 157, 768–770 (2009).
Spadaro, D. et al. [The arrival of kiwifruit canker in Piedmont]. Inf. Agrar. 66, 58–59 (2010).
Everett, K. R. et al. First report of Pseudomonas syringae pv. actinidiae causing kiwifruit bacterial canker in New Zealand. Australas. Plant Dis. Notes 6, 67–71 (2011).
Vanneste, J. L. The scientific, economic, and social impacts of the New Zealand outbreak of bacterial canker of kiwifruit (Pseudomonas syringae pv. actinidiae). Ann. Rev. Phytopatology 55, 377–399 (2017).
Spinelli, F., Donati, I., Vanneste, J. L., Costa, M. & Costa, G. Real time monitoring of the interactions between Pseudomonas syringae pv. actinidiae and Actinidia species. Acta Hortic. 913, 461–465 (2011).
Ferrante, P. et al. The importance of the main colonization and penetration sites of Pseudomonas syringae pv. actinidiae and prevailing weather conditions in the development of epidemics in yellow kiwifruit, recently observed in central Italy. J. Plant Pathol. 94, 455–461 (2012).
Young, J. M., Cheesmur, G. J., Welham, F. V. & Henshall, W. R. Bacterial blight of kiwifruit. Ann. Appl. Biol. 112, 91–105 (1988).
Everett, K. R. & Henshall, W. R. Epidemiology and population ecology of kiwifruit blossom blight. Plant Pathol. 43, 824–830 (1994).
Balestra, G. B. & Varvaro, L. Pseudomonas syringae pv. syringae causal agent of disease on floral buds of Actinidia deliciosa (A. Chev) Liang et Ferguson in Italy. J. Phytopathol. 145, 375–378 (1997).
Rossetti, A. & Balestra, G. M. Pseudomonas syringae pv. syringae on kiwifruit plants: epidemiological traits and its control. In: Fatmi M. et al. (eds) Pseudomonas syringae pathovars and related pathogens – Identification, Epidemiology and Genomics. Springer, Dordrecht, 65–68 (2008).
Ark, P. A. Further evidence of pollen dissemination of walnut blight. Phytopathology 34, 329–334 (1944).
Ercolani, G. L. [Individuazione di Xanthomonas juglandis (Pierce) Dowson in Emilia]. English, Abstract, Conclusion and Figure Captions. Phytopathol. Mediterr. 2, 1–10 (1962).
Wilson, M., Sigee, D. C. & Epton, H. A. S. Erwinia amylovora infection of hawthorn blossom: I. The anther. J. Phytopathol. 127, 1–14 (1989).
Mansvelt, E. L. & Hattingh, M. J. Scanning electron microscopy of pear blossom invasion by Pseudomonas syringae pv. syringae. Can. J. Bot. 65, 2523–2529 (1987).
Wilson, M., Sigee, D. C. & Epton, H. A. S. Erwinia amylovora infection of hawthorn blossom: II. The stigma. J. Phytopathol. 127, 15–28 (1989).
Spinelli, F., Ciampolini, F., Cresti, M., Geider, K. & Costa, G. Influence of stigmatic morphology on flower colonization by Erwinia amylovora and Pantoea agglomerans. Eur. J. Plant Pathol. 113, 395–405 (2005).
Johnson, K. B., Stockwell, V. O., Burgett, D. M., Sugar, D. & Loper, J. E. Dispersal of Erwinia amylovora and Pseudomonas fluorescens by honey bees from hives to apple and pear blossoms. Phytopathology 83, 478–484 (1993).
McArt, S. H., Koch, H., Irwin, R. E. & Adler, L. S. Arranging the bouquet of disease: floral traits and the transmission of plant and animal pathogens. Ecol. Lett. 17, 624–636 (2014).
Howpage, D., Spooner-Hart, R. N. & Vithanage, V. Influence of honey bee (Apis mellifera) on kiwifruit pollination and fruit quality under Australian conditions. N.Z. J. Crop Hortic. Sci. 29, 51–59 (2001).
Testolin, R., Vizzotto, G. & Costa, G. Kiwifruit pollination by wind and insects in Italy. N.Z. J. Crop Hortic. Sci. 19, 381–384 (1991).
Howpage, D., Vithanage, V. & Spooner-Hart, R. Pollen tube distribution in the kiwifruit (Actinidia deliciosa A. Chev. C. F. Liang) pistil in relation to its reproductive process. Ann. Bot. (Lond.) 81, 697–703 (1998).
Gonzalez, M. V., Coque, M. & Herrero, M. Papillar integrity as an indicator of stigmatic receptivity in kiwifruit (Actinidia deliciosa). J. Exp. Bot. 46, 263–269 (1995).
Hopping, M. E. & Jerram, E. M. Supplementary pollination of tree fruits. II. Field trials on kiwifruit and Japanese plums. N.Z. J. Agric. Res. 23, 517–521 (1980).
Vanneste, J. L. et al. Detection of Pseudomonas syringae pv. actinidiae in kiwifruit pollen samples. N.Z. Plant Prot. 64, 246–251 (2011).
Stefani, E. & Giovanardi, D. Dissemination of Pseudomonas syringae pv. actinidiae through pollen and its epiphytic life on leaves and fruits. Phytopathol. Mediterr. 50, 489–496 (2011).
Everett, K. R. et al. Heat treatments to kill Pseudomonas syringae pv. actinidiae on contaminated pollen. N.Z. Plant Prot. 65, 8–18 (2012).
Biondi, E. et al. Pseudomonas syringae pv. actinidiae detection in kiwifruit plant tissue and bleeding sap. Ann. Appl. Biol. 162, 60–70 (2013).
Gallelli, A., Talocci, S., L’Aurora, A. & Loreti, S. Detection of Pseudomonas syringae pv. actinidiae, causal agent of bacterial canker of kiwifruit, from symptomless fruits and twigs, and from pollen. Phytopathol. Mediterr. 50, 462–472 (2011).
Tontou, R., Giovanardi, D. & Stefani, E. Pollen as a possible pathway for the dissemination of Pseudomonas syringae pv. actinidiae and bacterial canker of kiwifruit. Phytopathol. Mediterr. 53, 333–339 (2014).
Sigee, D. C. Bacteria as plant pathogens. In: Bacterial Plant Pathology: Cell and Molecular Aspects. D. C. Sigee, ed. Cambridge University Press, Cambridge. pp. 1–12.
Gallelli, A., Talocci, S., Pilotti, M. & Loreti, S. Real-time and qualitative PCR for detecting Pseudomonas syringae pv. actinidiae isolates causing recent outbreaks of kiwifruit bacterial canker. Plant Pathol. 63, 264–276 (2014).
Pusey, P. L. The role of water in epiphytic colonization and infection of pomaceous flowers by Erwinia amylovora. Phytopathology 90, 1352–1357 (2000).
Gimenez-Ibanez, S. & Rathjen, J. P. The case for defence: plant versus Pseudomonas syringae. Microbes Infect. 12, 428–437 (2010).
Balestra, G. M. [Use of copper formulations to limit bacterial diseases in kiwifruit. (Speciale Actinidia)]. Riv. di Fruttic. e di Ortofloric. 66, 35–41 (2004).
Serizawa, S. & Ichikawa, T. Epidemiology of bacterial canker of kiwifruit. 2. The most suitable times and environments for infection on new canes. Ann. Phytopathol. Soc. Jpn. 59, 460–468 (1993).
Serizawa, S., Ichikawa, T., Takikawa, Y., Tsuyumu, S. & Goto, M. Occurrence of bacterial canker of kiwifruit in Japan: description of symptoms, isolation of the pathogen and screening of bactericides. Ann. Phytopathol. Soc. Jpn. 55, 427–436 (1989).
Pattemore, D. E., Goodwin, R. M., McBrydie, H. M., Hoyte, S. M. & Vanneste, J. L. Evidence of the role of honey bees (Apis mellifera) as vectors of the bacterial plant pathogen Pseudomonas syringae. Australas. Plant Pathol. 43, 571–575 (2014).
The work was funded by the European Union’s Seventh Framework Program for research, technological development, and demonstration under grant agreement no 613678 (DROPSA—Strategies to develop effective, innovative, and practical approaches to protect major European fruit crops from pests and pathogens). Dr. Giovanni Fiori is acknowledged for technical support.
Conflict of interest
The authors declare that they have no conflict of interest.
Electronic supplementary material
About this article
Pathogen-induced changes in floral scent may increase honeybee-mediated dispersal of Erwinia amylovora
The ISME Journal (2019)
Acta Horticulturae (2019)
Inheritance of seed and rhizosphere microbial communities through plant–soil feedback and soil memory
Environmental Microbiology Reports (2019)
Plant Microbiome and Its Link to Plant Health: Host Species, Organs and Pseudomonas syringae pv. actinidiae Infection Shaping Bacterial Phyllosphere Communities of Kiwifruit Plants
Frontiers in Plant Science (2018)