Erwinia psidii Rodrigues-Neto et al. 1987, is a Gram-negative plant-pathogenic bacterium, a member of the ‘amylovora’ group (Dye, 1981), which causes diseases in eucalyptus, guava and papaya crops in tropical and subtropical conditions (RodriguesNeto et al., 1987 Coutinho et al., 2011; Chai et al., 2017). Symptoms observed in these hosts are similar, which include dieback, leaf hydrosis (i.e., water-soaking) close to midrib, stem canker and wilt (Montoya-Estrada et al., 2019a; Borges-Junior et al., 2020). In eucalyptus forests planted with susceptible genotypes, the incidence of bacterial blight can reach 30% of the trees (Borges-Junior et al., 2020). Eucalyptus trees affected by the disease may have lower net photosynthetic rates, reduced growth and lower biomass production, thereby causing a drastic loss of productivity. There is no effective control against the pathogen, which has spread rapidly in eucalyptus forests, causing negative impacts on production (Lanna-Filho et al., 2021).
The biological aspects of E. psidii have been extensively studied (Teixeira et al., 2009; Hermenegildo et al., 2019), which has made it possible to establish standards for characterizing the genetic variability of the pathogen (Teixeira et al., 2009; Hermenegildo et al., 2021) and determine some components involved in virulence (Montoya-Estrada et al., 2019a; Pereira et al., 2021). However, studies on the bacterium's lifestyle are poor, which limits the understanding of the bacterial pathogen's life cycle in association with the host or under planted forest conditions. An important question related to E. psidii concerns the survival period of the bacterium on the myrtaceous phylloplane or weeds. A study on the persistence of the bacterium on phylloplane could explain the sudden and rapid occurrence of widespread outbreaks of the disease in eucalyptus forests. Understanding the survival of the pathogen on the phylloplane allows us to elucidate part of the bacterium's life cycle, which makes it possible to plan a more effective management program against bacterial blight.
Plant-pathogenic bacteria may survive in the phylloplane and serve as a reservoir for primary infections, which can cause epidemics in a plant population (Leben 1981; Martins et al., 2018; Daniel and Boher, 1985; Hildebrand et al., 2001). Erwinia psidii may have an epiphytic phase, which may result in the risk of spreading the pathogen in eucalyptus forests via wind-driven rain (Lanna-Filho et al., 2021). Furthermore, on contaminated cuttings the bacterium can be transported long distances and introduced into areas considered free of the pathogen. In this context, here we investigated the survival period of E. psidii on the phylloplane of elite eucalyptus genotypes, weeds commonly found in Brazilian eucalyptus forests and plants from the Myrtaceae family (Eugenia uniflora L., and Psidium guajava L.) that are found close to eucalyptus forests.
For this study, E. psidii strain CR01R was obtained from the bacterial culture collection of the Plant Bacteriology and Biocontrol Laboratory, Universidade Federal do Rio Grande do Sul, Porto Alegre, Brazil. The rifampicin-resistant (RifR) mutant (CR01R) was constructed based on previous studies carried out by Lanna-Filho et al. (2021). Mutant identity was confirmed using species-specific primers Ep2L and Ep2R (Silva et al., 2015). Erwinia psidii CR01R cultures grown for 12 h on 523 medium were resuspended in sterile distilled water. Cell suspensions were adjusted to OD540 = 0.2 (107 × CFU/mL) and sprayed (point of run-off) on Eucalyptus urophylla (SUZBA1175), E. urophylla (SUZBA1922), E. grandis × E. urophylla (FA6825), lantana (Lantana camara), brachiaria (Brachiaria decumbens), horseweed (Conyza bonariensis)and sourgrass (Digitaria insularis), guava (Psidium guajava L.) and pitanga (Eugenia uniflora L.) plants, with fully developed leaves. The plants were kept under greenhouse conditions at semi-controlled temperature and humidity.
At intervals of seven days after inoculation, leaves were collected randomly, weighed, placed in flasks containing 100 mL of PBS and sonicated (40 kHz, 120 W) for 1 min in an ultrasonic water bath to recover bacterial cells. Appropriate dilutions of leaf washings were plated on 523 medium containing rifampicin (100 μg/mL) and azoxystrobin (15 μg/mL). The broad-spectrum fungicide azoxystrobin (strobilurin) was added to 523 medium to inhibit fungal growth. Bacterial population size was estimated from plate counts made after 2 to 3 days’ incubation at 28°C and expressed as CFU/g leaf tissue. For each plant, three replicates were used, and each replicate included five plants (n = 15). The experiment was repeated twice for each eucalyptus clone, weeds and myrtaceae. For each sampling time, dilutions of leaf washings were plated in triplicate. Sampling was discontinued after the CR01R strain was not detected (species-specific primers Ep2L/Ep2R) on three consecutive sampling dates.
Colony-forming unit (CFU) data were transformed to log10 (×) prior to statistical analysis. For the elaboration of the bacterial growth curves, non-linear regression was applied based on the Sigmoidal, Gaussian and Exponential models. The area under the growth progress curve (AUGPC) was computed by each regression curve using the trapezoidal rule (Whittaker and Robinson, 1940) as follows: Ʃi[(Yi (Xi+1 - Xi)] + 0.5 (Yi+1 − Yi) (Xi-1 - Xi)], where Yi = bacterial population at the ith observation, and Xi =time in hours or days at the ith observation. Data were subjected to a t-test (p < 0.001) or ANOVA followed by a Tukey test (p < 0.05) using the software SigmaPlot 14.5 (Systat Software, San Jose, CA, USA).
Live cells E. psidii were recovered on the phylloplane of clone SUZBA1175 for up to 35 days. Moreover, on the phylloplane of clones SUZBA1922 and FA6825 the bacterium survived for up to 56 days (Figure 1a). When compared to the first sampling time (106 × CFU/g leaf tissue), at 7 days post-inoculation, the pathogen population decreased 1000-fold on the leaves of clone SUZBA1175. For both clones SUZBA1922 and FA6825, the bacterial population decreased 100-fold. On the leaf surface of clone SUZBA1175, the E. psidii population (0.9 × 101 CFU/g leaf tissue) was last detected 35 days post-inoculation. In the leaves of clones SUZBA1922 and FA6825, the presence of viable E. psidii cells was detected up to 56 days. AUGPC values showed a significant difference (p <0.001) in the E. psidii population on the clone phylloplane (Figure 1b). On the SUZBA1175 clone phylloplane, the AUPGC value was 27% lower for the E. psidii population compared to the SUZBA1922 clone (Figure 1b). When compared to clone FA6825, the AUPGC value was 21% lower for the pathogen population.
The survival period of E. psidii on the leaf surface of pitanga and guava plants was 77 (1.5 × 101 CFU/g leaf tissue) and 85 (6.0 × 101 CFU/g leaf tissue) days (Figure 2a), respectively. On both plants, the plant-pathogenic bacterium population showed a slight decline over all sampling dates. However, on guava the pathogen population was generally 10- to 100-fold higher on most sampling dates compared to pitanga. Based on AUGPC data, there was a significant difference (p < 0.001) for E. psidii populations on pitanga and guava leaves (Figure 2b). The data showed that the plant-pathogenic bacterium had greater persistence on guava leaves (AUGPC 197.62 ± 2.59) compared to pitanga leaves (AUGPC 266.36 ± 2.75). This means that the E. psidii population was 27% higher on the guava leaf surface than the pathogen population found on pitanga.
On weed leaf tissue, viable cells of plant-pathogenic bacterium were recovered for up to 7 days in B. decumbens or D. insularis (Figure 3a). On the other hand, E. psidii survived for 21 days on C. bonariensis or L. camara. Data from the 7th sampling recovered a bacterial population of 2.5 × 102 and 4.7 × 102 CFU/g leaf tissue on B. decumbens and D. insularis, respectively. This represents an amount 10000-fold lower than the total bacterial population (6.9 × 106 CFU/g leaf tissue) at the 1st sampling time. At 21 days, E. psidii populations of 2.4 × 102 and 5.1 × 101 CFU/g leaf tissue were recovered on C. bonariensis and L. camera, respectively. Over the sampling dates, a sharp decline in the bacterial population was observed in B. decumbens and D. insularis, while for C. bonariensis and L. camara the pathogen population had a slight decline 3 weeks post-inoculation. Based on AUGPC data, there was a significant difference (p < 0.001) for E. psidii populations on the phylloplane of all weeds (Figure 3b). AUGPC values showed that E. psidii populations had a negative impact in association with the leaf surface of B. decumbens (AUGPC 34.59 ± 0.58) and D. insularis (AUGPC 42.9 ± 1.13), because they had lower survival. Furthermore, AUGPC data were higher for the bacterial population on the phylloplane of C. bonariensis (AUGPC 75.28 ± 0.43) and L. camera (AUGPC 69.58 ± 0.64).
Our data clearly show that E. psidii can survive on the phylloplane of all tested plants, which can serve as an inoculum reservoir for the occurrence of new infections in forest or nursery conditions. This is a key point for understanding the epidemiology of bacterial late blight, because it demonstrates that the pathogen can be introduced into disease-free areas via asymptomatic cuttings. The reserve of bacterial inoculum on the phylloplane can also explain the sudden occurrence of widespread outbreaks of the disease in eucalyptus forests (Coutinho et al., 2011; Arriel et al., 2014) and guava orchards (Oliveira et al., 2000; Uesugi et al., 2001; Marques et al., 2007). Lanna-Filho et al. (2021) reported that E. psidii survived between 40 and 60 days on the eucalyptus phylloplane (E. saligna and E. urophylla × E. globulus), which supports the data obtained in this study. E. psidii can also survive as an endophyte, as reported by Montoya-Estrada et al.(b) (2019). The authors proved that the bacterium can colonize and move in the xylem vessels of eucalyptus clones, which can be an important inoculum reservoir for asymptomatic cuttings or trees in planted forest conditions.
Although the survival period of the plant-pathogenic bacterium was the same among the phylloplanes of the eucalyptus clones (Fig. 1a), the bacterial population rate varied significantly (Fig. 1b). There is clearly a relationship between the host and the epiphytic survival of the bacterium. The genotype and phenotype of the host plant play a major role in structuring the phyllosphere’s microbial communities (Bodenhausen et al. 2014; Wagner et al. 2016). In general, leaf chemistry, morphology, and developmental stage can significantly favor the establishment and survival of the bacterial pathogen on the phylloplane (Whipps et al., 2008; Ryffel et al. 2016; Li et al. 2021). In a previous study, Lanna-Filho et al. (2021) reported that the E. psidii population was higher on the phylloplane of the Eucalyptus urophylla × E. globulus hybrid when compared to E. saligna. Likely, restriction factors (microstructure and chemical characteristics) present on the E. saligna phylloplane caused negative impacts on the multiplication of the bacterium (Lanna-Filho et al. 2021).
Plant-pathogenic bacteria have a greater ability to survive on phylloplane plants homologous to their preferred host, compared to heterologous plants (Chang et al. 1992). Our data support this statement, as the plant-pathogenic bacterium exhibited greater survival on the pitanga and guava phylloplanes (Fig 2a) compared to its survival on weeds (Fig 3a). From a practical standpoint, pitanga and guava plants near eucalyptus forests may serve as potential sources of bacterial inoculum for epidemic outbreaks of bacterial blight. In addition to E. psidii existing in the resident (epiphytic) phase, it can also be in the pathogenic phase, causing symptoms in these fruit trees (Rodrigues-Neto et al. 1987; Caires et al. 2019). This is critical because, without the eradication of these fruit trees, E. psidii can persist indefinitely in both epiphytic and pathogenic phases, causing the inoculum to increase in eucalyptus forests.
Regarding weeds, E. psidii showed a lower ability to persist on the leaf surface. However, under forest conditions, weeds can serve as a brief reservoir of inoculum, acting as a "stepping-stone" from one plant to another. Besides reducing the distance between plants, this would increase the efficiency of the bacterium's dispersal process. Our results also show another interesting aspect, as there seems to be a positive relationship between E. psidii's ability to persist on the phylloplane and leaf morphology. Both L. camara and C. bonariensis are dicotyledonials and have similar leaf morphology, which probably enabled a longer (21 days) persistence of E. psidii on their phylloplanes. On the other hand, the B. decumbens and Digitaria sp. monocotyledons supported plant-pathogenic bacterium on their phylloplanes for a shorter period (7 days). Yan et al. (2022) showed that anatomical properties of leaf such as venous and stomatal density are determinant for the predilection and prevalence of certain specific microorganism groups. Furthermore, the chemical and nutritional properties of the leaf can also benefit the establishment and persistence of the bacterium by providing requirements for its survival (Aragón et al., 2017; Goswami et al. 2021).
Our results clearly show that E. psidii has an important epiphytic phase and can persist for days on host and non-host plants. This information brings relevant contributions to the implementation of strategies that aim to mitigate bacterial blight in nursery or forest conditions. The use of bacterial pathogen detection protocols on asymptomatic cuttings is a sine qua non condition to prevent transport and/or planting of contaminated cuttings in bacterium-free areas. Another important strategy is the eradication of plants such as pitanga and guava near eucalyptus forests. Only in this way is it possible to eliminate putative reservoirs of plant-pathogenic bacterium that can recurrently serve as sources of primary inoculum for the emergence of new infections. The weeds that were the focus of this study should also be eliminated to prevent them from being potential sources for shelter and survival of the pathogen.
The combined use of the previously mentioned strategies has an effective role in mitigating the occurrence of bacterial blight in eucalyptus forests, because they act directly on the bacterium's survival and pathogenicity phases. Although studies on the E. psidii × eucalyptus pathosystem have advanced in recent years, there is still a need for further research focused on developing pathogen detection protocols. Although cuttings and trees are asymptomatic, the plant-pathogenic bacterium may be present in very low titers. This can make pathogen detection difficult, potentially resulting in false negatives. Furthermore, robust research is needed to select potential genotypes resistant to bacterial blight. Although this can be a herculean task, resistance combined with other strategies can result in greater success in controlling the disease in the main eucalyptus-producing areas.