Efficacy of Soft-Rot Disease Biocontrol Agents in the Inhibition of Production Field Pathogen Isolates

The Dickeya and Pectobacterium bacterial species cause blackleg and soft-rot diseases on potato plants and tubers. Prophylactic actions are essential to conserve a high quality of seed potato tubers. Biocontrol approaches are emerging, but we need to know how efficient biocontrol agents are when facing the natural diversity of pathogens. In this work, we sampled 16 production fields, which were excluded from the seed tuber certification scheme, as well as seven experimental parcels, which were planted with seed tubers from those production fields. We collected and characterized 669 Dickeya and Pectobacterium isolates, all characterized using nucleotide sequence of the gapA gene. This deep sampling effort highlighted eleven Dickeya and Pectobacterium species, including four dominant species namely D. solani, D. dianthicola, P. atrosepticum and P. parmentieri. Variations in the relative abundance of pathogens revealed different diversity patterns at a field or parcel level. The Dickeya-enriched patterns were maintained in parcels planted with rejected seed tubers, suggesting a vertical transmission of the pathogen consortium. Then, we retained 41 isolates representing the observed species diversity of pathogens and we tested each of them against six biocontrol agents. From this work, we confirmed the importance of prophylactic actions to discard contaminated seed tubers. We also identified a couple of biocontrol agents of the Pseudomonas genus that were efficient against a wide range of pathogen species.


Introduction
Several species of the genera Dickeya and Pectobacterium are causative agents of the blackleg and soft-rot diseases in Solanum tuberosum stems and tubers, respectively [1]. In Europe, the species currently isolated from blackleg and soft-rot lesions on potato plants encompass Pectobacterium atrosepticum, Pectobacterium parmentieri, Pectobacterium brasiliense, Pectobacterium polaris, Dickeya dianthicola and Dickeya solani [2]. D. solani emerged twenty years ago [3]. Some other species such as P. punjabense, P. versatile and P. parvum were less frequently detected in lesions of potato plants [4][5][6]. Over the past years, important efforts in genome sequencing of bacterial isolates deposited in collections and those collected by new samplings contributed to a more accurate delineation of Dickeya and Pectobacterium species (examples in [7][8][9][10]).
From this knowledge in taxonomy and genomics, different molecular tools have been developed for pathogen diagnosis, hence for identifying and discarding the contaminated plant materials from the production process of certified seed potato tubers [11]. For instance, in France, a quality scheme was set up along production of tuber seeds from the in vitro propagation of potato varieties until tuber harvest in the field. Before and after tuber harvest, intensive inspections are carried out in the frame of the official control and certification scheme implemented by SOC (French control and certification body). Before harvest, 1% of symptomatic plants is the highest threshold value of blackleg disease incidence for maintaining a given field in the production scheme. In post-harvest controls, the threshold value is set at 0.2% of dry and wet rots, expressed as a weight percentage of tubers (for details see http://frenchseedpotato.com, accessed on 20 November 2022). In France, each year between 2013 and 2017, which is the sampling period in this study, around 9000 lots were harvested from over 20,000 ha. Over this period, the mean rate of rejected plots was 0.85% per year. A maximum rate reached 2.1% in 2016 with particularly rainy conditions, which are more favorable to the development of the disease. The lowest rate was 0.3% in 2017, most likely because of drier environmental conditions (data from FN3PT http://frenchseedpotato.com, accessed on 20 November 2022).
In parallel to the pathogen diagnosis and prophylaxis approach, research has been carried out to identify potato varieties and genetic determinants involved in response to Dickeya and Pectobacterium pathogens [12,13]. Some others allowed the discovery of a wide variety of biological agents including phages and bacteria [14][15][16][17]. Because Dickeya and Pectobacterium pathogen populations are diverse and dynamic over time and space, we need to understand their structure in potato fields and evaluate the efficiency of the potential biocontrol agents using a refreshed collection of pathogen isolates.
In this work, we sampled 16 production fields that were rejected for producing seed potato tubers. We isolated 463 pathogens and compared the structure of the pathogen populations of these fields. In addition, we harvested asymptomatic tubers from thirteen of these rejected fields and we planted them the next year in experimental parcels. We isolated 206 additional pathogens and then we compared the structure of the pathogen populations in production fields (year N) and experimental parcels (year N + 1). Finally, 41 isolates that were representative of the collected species in the rejected potato fields were tested for their sensitivity against six biocontrol agents Bacillus simplex BA2H3, Pseudomonas brassicacearum PA1G7 and PP1-210F, Pseudomonas fluorescens PA3G8, Pseudomonas lactis PA4C2 and Pseudomonas sp. PA14H7 [14]. These biocontrol agents were previously identified after a screening of 10,000 bacterial isolates for their capacity to inhibit the growth of P. atrosepticum CFBP6276 and D. dianthicola RNS04.9 [14]. Hence, we could evaluate the biocontrol efficiency of these biocontrol agents against a wider range of pathogens that are representative of the species sampled in potato fields.

Isolation and Characterization of the Dickeya and Pectobacterium Pathogens
Between 2013 and 2017, 16 potato production fields (France) were sampled. These production fields were excluded from certified lots of seed tubers because they exhibited more than 1% of symptomatic plants (blackleg disease). In order to study Pectobacterium and Dickeya populations at the field scale, around 30 plants with blackleg symptoms were collected in each field, and a single bacterial isolate was retained from each of the 463 plants, resulting in 463 pathogen isolates.
Out of these 16 production fields, thirteen (P1 to P13) were used for plantation assays. From 800 to 4000 asymptomatic tubers were harvested (year N) in each field and then planted the next year (year N + 1) at the experimental station of Comité Nord (Achicourt, France). Out of the 13 experimental parcels, seven exhibited enough plants with blackleg symptoms for sampling: 206 diseased plants were collected (around 30 per parcel) and then 206 pathogens were isolated.
Our sampling approach revealed pathogen diversity at a field or parcel level but not at a plant individual level. In the case of production fields and experimental parcels, pectinolytic bacteria were isolated from lesions on crystal violet pectate medium [18]. Bacterial isolates were purified on agar plates and characterized at the genus (Dickeya and Pectobacterium) and species levels using the PCR primers listed in Table S1 [19][20][21][22][23][24]. Characterization of the D. solani and D. dianthicola isolates from the production fields P1 to P17 was already done in a previous paper [25].

Growth Inhibition Assays with Biocontrol Agents
Among the pathogen isolates sampled in potato fields between 2013 and 2017, 41 isolates representative of the species diversity were used in growth inhibition assays in triplicate with each of the six biocontrol agents: Bacillus simplex BA2H3, Pseudomonas brassicacearum PA1G7 and PP1-210F, Pseudomonas fluorescens PA3G8, Pseudomonas lactis PA4C2 and Pseudomonas sp. PA14H7 [14]. One hundred µL of each pathogen culture at 10 7 CFU/mL were spread on a Petri dish in triplicate. After drying, 10 µL of each biocontrol agent culture at 10 9 CFU/mL was spotted on the center of each Petri dish. After drying, the bacteria were incubated at 25 • C for 24 h. Then, the inhibition halo was measured for each pathogen strain-biocontrol strain combination.

Statistical Analyses
In the in vitro antibiosis tests, statistical analyses were performed using Rstudio (RStudio team, 2021.09.1 version, Boston, MA, USA, https://www.rstudio.com/, accessed on 20 November 2022) software for Windows using the packages FSA and agricolae. A Kruskal-Wallis test (alpha 0.05) followed by a Dunn test with Benjamini-Hochberg correction (alpha 0.05) was performed to compare each pathogen.

Growth Inhibition Assays with Biocontrol Agents
Among the pathogen isolates sampled in potato fields between 2013 an isolates representative of the species diversity were used in growth inhibition triplicate with each of the six biocontrol agents: Bacillus simplex BA2H3, Pseudom sicacearum PA1G7 and PP1-210F, Pseudomonas fluorescens PA3G8, Pseudom PA4C2 and Pseudomonas sp. PA14H7 [14]. One hundred μL of each pathogen 10 7 CFU/mL were spread on a Petri dish in triplicate. After drying, 10 μL of ea trol agent culture at 10 9 CFU/mL was spotted on the center of each Petri dish. Af the bacteria were incubated at 25 °C for 24 h. Then, the inhibition halo was me each pathogen strain-biocontrol strain combination.

Statistical Analyses
In the in vitro antibiosis tests, statistical analyses were performed usin (RStudio team, 2021.09.1 version, Boston, MA, USA, https://www.rstudio.com on 20 November 2022) software for Windows using the packages FSA and a Kruskal-Wallis test (alpha 0.05) followed by a Dunn test with Benjamini-Hoc rection (alpha 0.05) was performed to compare each pathogen.
Our sampling approach (one isolate from one symptomatic plant, repeat 30 times in each experimental field) allowed us to access patterns of pathogen p in each field. Two to six species were identified in each field. Clearly, tw emerged among the analyzed fields, those enriched in Pectobacterium (P1, P3, P and the others in Dickeya species, either D. solani or D. dianthicola. Parcels P1,  Figure 1. Species assignation. The gapA sequence was used for species assignation: (a) 463 isolates collected from 16 potato fields, which were rejected for seed tuber production because of blackleg symptoms, and (b) 206 isolates collected from seven parcels exhibiting blackleg symptoms after plantation of tuber seeds collected from symptomatic fields.
Our sampling approach (one isolate from one symptomatic plant, repeated around 30 times in each experimental field) allowed us to access patterns of pathogen population in each field. Two to six species were identified in each field. Clearly, two patterns emerged among the analyzed fields, those enriched in Pectobacterium (P1, P3, P4 and P14) and the others in Dickeya species, either D. solani or D. dianthicola. Parcels P1, P3 and P4 contain only Pectobacterium isolates, while P14 contains also Dickeya isolates. In these four parcels enriched in Pectobacterium species, the relative abundance of the dominant species, namely P. atrosepticum in P1, P3 and P4 and P. parmentieri in P14, varies from 40% to 63%. In Dickeya-enriched patterns, D. dianthicola was dominant in P2, P7, P9, P12, P15 and P17 and D. solani in P5, P6, P8, P10, P11 and P13. Among the Dickeya-enriched patterns, the dominant species may represent up to 96% of the isolated strains.
In Pectobacterium species, the gapA nucleotide sequence revealed the presence of different alleles that give an insight into the infra-specific diversity of these pathogens. In the collected isolates, we identified two or three alleles in P. atrosepticum, P. carotovorum, P. odoriferum and P. parmentieri, and six in P. brasiliense, eight in P. versatile and 10 in P. polaris. Different alleles of the same species could coexist in the same production field (Table S2): for instance, three gapA alleles were identified in P. parmentieri isolates recovered from field P3, four gapA alleles in P. brasiliense from field P15 and five gapA alleles in P. polaris from field P1, suggesting the coexistence of different strains of the same species in the same field.

Species Patterns in Experimental Parcels
Out of the 16 production fields, thirteen (P1 to P13) were used for plantation assays using the harvested asymptomatic tubers. These production fields exhibited either a All seven sampled experimental parcels showed a Dickeya pattern ( Figure 2). Seed tubers used for planting these experimental parcels were collected from production fields that also exhibited a Dickeya pattern, suggesting a vertical transmission of pathogens via tuber seeds. By comparing the species pattern (the relative abundance of species) of each pair of production fields and planted experimental parcel (P2 and P2rep, P6 and contain only Pectobacterium isolates, while P14 contains also Dickeya isolates. In these four parcels enriched in Pectobacterium species, the relative abundance of the dominant species, namely P. atrosepticum in P1, P3 and P4 and P. parmentieri in P14, varies from 40% to 63%. In Dickeya-enriched patterns, D. dianthicola was dominant in P2, P7, P9, P12, P15 and P17 and D. solani in P5, P6, P8, P10, P11 and P13. Among the Dickeya-enriched patterns, the dominant species may represent up to 96% of the isolated strains. In Pectobacterium species, the gapA nucleotide sequence revealed the presence of different alleles that give an insight into the infra-specific diversity of these pathogens. In the collected isolates, we identified two or three alleles in P. atrosepticum, P. carotovorum, P. odoriferum and P. parmentieri, and six in P. brasiliense, eight in P. versatile and 10 in P. polaris. Different alleles of the same species could coexist in the same production field (Table S2): for instance, three gapA alleles were identified in P. parmentieri isolates recovered from field P3, four gapA alleles in P. brasiliense from field P15 and five gapA alleles in P. polaris from field P1, suggesting the coexistence of different strains of the same species in the same field.

Species Patterns in Experimental Parcels
Out of the 16 production fields, thirteen (P1 to P13) were used for plantation assays using the harvested asymptomatic tubers. These production fields exhibited either a All seven sampled experimental parcels showed a Dickeya pattern (Figure 2). Seed tubers used for planting these experimental parcels were collected from production fields that also exhibited a Dickeya pattern, suggesting a vertical transmission of pathogens via tuber seeds. By comparing the species pattern (the relative abundance of species) of each pair of production fields and planted experimental parcel (P2 and P2rep, P6 and P6rep,

Efficiency of Biocontrol Agents Facing Natural Diversity of Pathogens
We evaluated whether the biocontrol agents B. simplex BA2H3, P. brassicacearum PA1G7 and PP1-210F, P. fluorescens PA3G8, P. lactis PA4C2 and Pseudomonas sp. PA14H7 were active against the pathogens isolated from the sampled potato fields. We collected 738 measures of growth inhibition assays by testing in triplicate the six biocontrol agents against 41 pathogen isolates, all collected in production fields: five isolates of D. dianthicola, five of D. solani, seven of P. atrosepticum, five of P. brasiliense, five of P. parmentieri, four of P. versatile, three of P. carotovorum, three of P. polaris, two of P. odoriferum, one of P. punjabense and one of P. parvum (a list in Table S2).

Efficiency of Biocontrol Agents Facing Natural Diversity of Pathogens
We evaluated whether the biocontrol agents B. simplex BA2H3, P. brassicacearum PA1G7 and PP1-210F, P. fluorescens PA3G8, P. lactis PA4C2 and Pseudomonas sp. PA14H were active against the pathogens isolated from the sampled potato fields. We collecte 738 measures of growth inhibition assays by testing in triplicate the six biocontrol agen against 41 pathogen isolates, all collected in production fields: five isolates of D. dian thicola, five of D. solani, seven of P. atrosepticum, five of P. brasiliense, five of P. parmentier four of P. versatile, three of P. carotovorum, three of P. polaris, two of P. odoriferum, one of P punjabense and one of P. parvum (a list in Table S2).
Considering all pathogen isolates, Pseudomonas sp. PA14H7 and P. brassicacearu PA1G7 and PP1-210F exhibited the highest growth inhibition capacity, while P. fluorescen PA3G8, P. lactis PA4C2 and Bacillus simplex BA2H3 were less efficient (Figure 3). In the next step, we refined our analysis of the collected 738 halo measurements b considering each pathogen species for each biocontrol agent (Figure 4 and 5). Pseudomona sp. PA14H7 and P. brassicacearum PA1G7 and PP1-210F were able to antagonize th growth of a broad spectrum of species (Figure 4). In the presence of Pseudomonas s PA14H7, the more sensitive pathogen species (classes a or b in statistical analysis) wer D. dianthicola, D. solani, P. atrosepticum and P. parvum, while the other Pectobacterium spe cies appeared as less sensitive (classes c, d or e in statistical analysis). In the presence of P brassicacearum strains PA1G7 and PP1-210F, the more sensitive species were D. dianthicol D. solani, P. atrosepticum, P. parvum and P. punjabense, whereas P. versatile also appeared a sensitive in the presence of P. brassicacearum strains PA1G7. Other results obtained wit biocontrol strains B. simplex BA2H3, P. lactis PA4C2 and P. fluorescens PA3G8 are pre sented in Figure 5. These strains are less efficient for inhibiting the growth of the 41 teste pathogens. In the next step, we refined our analysis of the collected 738 halo measurements by considering each pathogen species for each biocontrol agent (Figures 4 and 5). Pseudomonas sp. PA14H7 and P. brassicacearum PA1G7 and PP1-210F were able to antagonize the growth of a broad spectrum of species (Figure 4). In the presence of Pseudomonas sp. PA14H7, the more sensitive pathogen species (classes a or b in statistical analysis) were D. dianthicola, D. solani, P. atrosepticum and P. parvum, while the other Pectobacterium species appeared as less sensitive (classes c, d or e in statistical analysis). In the presence of P. brassicacearum strains PA1G7 and PP1-210F, the more sensitive species were D. dianthicola, D. solani, P. atrosepticum, P. parvum and P. punjabense, whereas P. versatile also appeared as sensitive in the presence of P. brassicacearum strains PA1G7. Other results obtained with biocontrol strains B. simplex BA2H3, P. lactis PA4C2 and P. fluorescens PA3G8 are presented in Figure 5.   Figure 3 were analyzed different way. For each biocontrol agent, we calculated mean value and variation for all halos lected with pathogen isolates belonging to the same species. Statistical differences (Kruskal-W test multiple comparison (Dunn), p-value adjusted with the Benjamini-Hochberg method, alp 0.05) are indicated by different letters (a to e); horizontal bar indicates mean value.

Discussion
Several species-specific tools were developed for identifying and quantifying Dickeya and Pectobacterium pathogen species in plant samples, as well as for characterizing the collected bacterial isolates [11]. In this work, we mainly used a unique marker gapA for characterizing all the collected isolates from blackleg lesions. These symptomatic tissues

Discussion
Several species-specific tools were developed for identifying and quantifying Dickeya and Pectobacterium pathogen species in plant samples, as well as for characterizing the collected bacterial isolates [11]. In this work, we mainly used a unique marker gapA for characterizing all the collected isolates from blackleg lesions. These symptomatic tissues were sampled in potato tuber production fields, which were excluded from tuber-seed production because of a high prevalence of blackleg symptoms. A strength of this PCRsequencing tool is the absence of a priori on the taxon to be identified [24]. Another interest is the generation of nucleotide sequences that may be stored and used in alignments and relation trees with reference sequences according to an up-to-date taxonomy. For instance, this approach allowed us to identify a rare taxon P. punjabense in our collection of field isolates before the development of a dedicated species-specific tool [5]. The gapA marker is appropriate to uncover emerging or uncharacterized taxon.
In this work, we took advantage of the gapA marker for comparing Dickeya and Pectobacterium species patterns of potato tuber production fields. We focused our study on fields that were excluded from tuber-seed production because of a high prevalence of symptomatic plants (>1%). This work provided an understanding of what diversity of pathogens is discarded by the prophylactic strategy that is set up along the tuber-seed certification process. According to the relative abundance of the taxons in the 16 fields we sampled, we identified two major patterns: the Pectobacterium and Dickeya patterns. These patterns are characterized by a predominance of one species that represents at least 40% of the isolated pathogens. In this study, the predominant species were D. solani, D. dianthicola, P. atrosepticum and P. parmentieri. Together with predominant species, companion species were identified in each sampled field. Their numbers varied from one to five, among the following species: D. solani, D. dianthicola, P. atrosepticum and P. parmentieri, and among some less frequently isolated species such as P. brasiliense, P. carotovorum, P. odoriferum, P. parvum, P. polaris, P. punjabense and P. versatile. The predominance of some species could mirror a greater capability to exploit the potato host (hence a greater aggressiveness), and/or a greater competitiveness against other Dickeya and Pectobacterium pathogens and microbiota. Because only 16 fields were sampled, we could not exploit our data for testing statistical hypotheses of association or exclusion between species. We are currently expanding our samplings to investigate species relationships.
Several cases of predominant species and synergistic or antagonist relationships between Dickeya and Pectobacterium pathogens were reported. These relationships were considered at a species level, when behavior is shared by all (or almost all) strains of the same species, or at an infra-species level, when only one strain (or a few) exhibits a particular trait. In Norway, a study on 34 seed tuber lots revealed P. atrosepticum as a predominant species [26]. In Finland, a long-term survey (over 14 years) of seed tubers showed a predominance of P. carotovorum (52% of all samples) and, a more frequent co-occurrence of Pectobacterium-Pectobacterium species rather than Pectobacterium-Dickeya species [27]. In France, a long-term fields survey (over 10 years), revealed Pectobacterium species as predominant over Dickeya species in collected lesions and a non-random distribution of Dickeya and Pectobacterium predominant species in field survey [25]. In the USA, a field survey revealed D. dianthicola and P. parmentieri as predominant species; virulence assays in parcels showed D. dianthicola as more virulent than P. parmentieri [28]. A synergistic effect of the co-inoculation of the two species resulted in increased disease severity compared to single-species inoculation [28]. In Morocco, sampling revealed P. brasiliense and D. dianthicola as predominant species in distinctive production fields [29].
The predominant species reflects local contingencies that could be influenced by the impact of climatic factors: the optimal growth temperature is higher for Dickeya species as compared to Pectobacterium species, as discussed by Degefu et al. (2021) [27]. Relationships between D. solani and D. dianthicola appeared complex: competitive and synergistic behaviors were reported as depending on the colonized plant tissues. In greenhouse assays, D. dianthicola outcompeted D. solani in aerial parts of potato plants, while the two species co-existed in tubers [25]. In the case of D. solani, different properties were highlighted as potentially contributing to settlement in potato agrosystems: a capacity to initiate symptoms with a low inoculum in relation to a particular regulation of the pelED promoter [30]; a capacity to exploit a variety of nutrient resources [31], a capacity to produce a wide spectrum of anti-microbial compounds, improving its competitive fitness against other Pectobacterium and Dickeya species and microbiota [32,33]. Some other reports suggest that the production of antibiotics and their cognate resistance genes could also influence the dynamics of Pectobacterium and Dickeya pathogens by modifying the relative abundance of taxons and clones [34,35].
Facing the diversity of Pectobacterium and Dickeya pathogens, the main strategy deployed in all countries is a prophylactic approach consisting of discarding symptomatic parcels and lots of seed tubers. The identification and introgression of plant alleles for decreasing the sensitivity of S. tuberosum is a promising strategy, still in infancy [12,13]. Over the past years, intensive efforts allowed the identification of biocontrol agents: phages and bacteria [14][15][16][17]. A challenging issue is identifying biocontrol agents controlling a wide diversity of Pectobacterium and Dickeya species. Cocktails of phages or bacteria have been proposed to solve this issue [14][15][16], but an increased cost of production and patents could be expected. Another non-exclusive approach is to search for a biocontrol agent targeting a wide range of Dickeya and Pectobacterium species. In this work, we challenged six bacterial biocontrol agents against a wide diversity of pathogen isolates that we collected from potato fields. We showed that three of them exhibited a wide range spectrum in growth inhibition of pathogens. The biocontrol strain Pseudomonas sp. PA14H7 was very active, including against the predominant species that we identified in our sampling. The mechanism of action of biocontrol activity is under investigation.

Conclusions
This work highlighted distinctive patterns of Dickeya and Pectobacterium species in sampled production fields, which were excluded from the certification process of potato tuber seeds. Moreover, the gapA analysis revealed the infraspecific diversity at the plot scale and underlines the complex relationship existing between strains involved in the symptom expression. These patterns are maintained in parcels planted with rejected seed tubers, highlighting the importance of a certification scheme for rejecting contaminated tubers. Representatives of the Dickeya and Pectobacterium diversity in the potato fields were challenged by several biocontrol agents allowing the identification of a couple of biocontrol agents that were efficient against a wide range of pathogen species.