Identiﬁcation and Characterisation of Pseudomonas savastanoi pv. savastanoi as the Causal Agent of Olive Knot Disease in Croatian, Slovenian and Portuguese Olive ( Olea europaea L.) Orchards

: Strains of Pseudomonas savastanoi pv. savastanoi ( Pss ), isolated from infected olive trees ( Olea europaea L.) in three European countries (Croatia, Slovenia and Portugal) were identiﬁed and characterised according to their colony morphology, physiological and biochemical features. According to the LOPAT scheme, 38.6% of Pss isolates were grouped in the Ib cluster. The Portuguese Pss strains were fully consistent with the typical LOPAT proﬁle for this bacterium. Conversely, most Slovenian Pss strains showed delayed oxidase activity, whilst Croatian Pss strains did not produce any ﬂuorescent pigment when grown in vitro. For Pss molecular identiﬁcation, both end-point and real-time PCR were used, as well as MALDI–TOF, which was additionally used for proteomic analysis and the subsequent species identiﬁcation of a number of strains that showed deviations from expected LOPAT results. Pss was conﬁrmed as a causal agent of olive knot disease in 46.6% of olive orchards screened. Overall, these data suggests a possible correlation of certain Pss features with the geographical origin and the ecological niche of Pss isolates.


Introduction
Pseudomonas syringae (P. syringae) is a species complex of bacteria currently known to include at least 15 species, 13 phylogroups, and more than 60 pathovars [1][2][3]. Most of the strains belonging to the P. syringae species complex are plant pathogens or were originally detected in the agro-environment; however, several strains have also been isolated from habitats unrelated to agriculture [4]. Pseudomonas savastanoi (P. savastanoi) is one of the plant pathogenic species belonging to the P. syringae complex and has established infections in a wide array of wild and cultivated plants, including both herbaceous and woody species. However, each pathovar has a specific host range, as expected by definition [5]. The main host of Gammaproteobacterium P. savastanoi pv. savastanoi (Pss) is the olive tree (Olea europaea L.). Olives trees are grown in a large area worldwide, and are a typical feature of Mediterranean countries such as Portugal, Croatia and Slovenia, whose productions account for around 722, 150 and 33 thousand tons of olives per year, respectively [6]. Olive oil is important for its benefits to human health, due to its high content of secondary plant metabolites and antioxidants [7]. Olive trees are susceptible to many abiotic and biotic stresses that negatively impact yield and the organoleptic and/or biochemical traits of

Distribution and Incidence of Olive Knot Disease Symptoms in Surveyed Countries
More than 600 hectares covering 58 different olive orchards were surveyed, with a total of 206 samples collected. The incidence of olive knot disease on various olive varieties grown in the surveyed orchards is presented in Table 1. In Croatia, the incidence of olive knot disease was observed on the most prevalent olive varieties, predominantly represented by the most commonly grown olive varieties: Leccino, Pendolino and Istarska bjelica. Moreover, an incidence of 100% of olive knot disease was observed on Nocciara, Italian rosinjola, Buža momjanska, Karbonera, Moražola, Cipressino, Pendolino piangente, Bianchera di Udine and Arbequina varieties, which are grown to a lesser extent, in approximately 3% of olive orchards. Some autochthonous Croatian varieties, such as Buža and Istarska bjelica, were present in more than half of all surveyed areas, and the incidence of olive knot disease was 32% and 30% in the surveyed orchards. Although grown in 13% of sampled orchards, the olive variety Karbonaca exhibited around 83% of disease incidence.
In Slovenian surveyed orchards, the most abundant variety was Istarska bjelica, grown in 75% of orchards. Introduced Italian varieties Leccino and Frantoio were grown in fewer than half of the surveyed orchards, and the incidence of symptoms of olive knot disease was approximately 66% and 100%, respectively. Olive varieties Ascolana tenera, Arbequina, Itrana, Mata, Leccione and Pendolino were grown in 12.5% of orchards. For these varieties, olive knot disease symptoms were found on Mata, Itrana, Leccione, Pendolino, Arbequina and Ascolana tenera, whilst no symptoms were ever observed on the Istarska bjelica variety.
In Portuguese olive orchards, the most prevalent symptomatic variety was Verdeal (60% of sampled orchards), followed by Cobrançosa and Cordovil in 40% of surveyed orchards. The next more prevalent varieties (present in 20% of olive orchards) with symptoms of olive knot disease were: Negruxa de Coimbra, Negruxa, Madura, Negruxa de Tras os Montes and Xuinha/Azeitona de Coimbra. All these varieties in all surveyed Portuguese olive orchards were sampled due to the presence of the typical olive knot disease symptoms.
Overall, the most severe infections were observed on the olive varieties Frantoio, Arbequina, Maurino and Pendolino. Symptoms varied according to the average tumour diameter. In Croatia, Slovenia and Portugal, the largest average tumour diameters were observed on infected plants of the Picholine (1.40 ± 0.20 cm), Mata (3.00 ± 0.00 cm) and Madural (1.03 ± 0.13 cm) varieties, respectively. The lowest values of average knot diameter in Croatia, Slovenia and Portugal were observed on the Buža momjanska (0.60 ± 0.10 cm), Ascolana tenera (0.75 ± 0.25 cm) and Cordovil (0.80 ± 0.00 cm) varieties. The absence of olive knot disease was determined in the Ascolana tenera variety in Croatia and the Istarska bjelica variety in Slovenia.

Biochemical (LOPAT), Physiological and Molecular Tests for Characterisation of Pss Isolates from Infected Olive Materials
Pss isolates and their corresponding geographical origins are listed in Table 2. Out of 206 sampled knots collected in 2021 and 2022, in 88 (i.e., 42.7%), the associated bacteria were identified as Pss using a real-time PCR method or a standard PCR assay (Figures S1 and S2). For Pss isolates from Slovenia, the positive PCR results were confirmed by a mass spectrometry method (MALDI-TOF) (Supplementary S1). The pathogen was thus confirmed to be present and infective in 27 (46.6%) out of a total of 58 olive orchards surveyed in Slovenia, Croatia and Portugal (Table 2).   When tested using a real-time PCR assay, 32 out of 102 (31.4%) and 2 out of 40 (5%) of the bacterial isolates from infected materials from Croatia and Portugal, respectively, were confirmed to be Pss ( Figure S1). The presence of Pss was also confirmed in varieties Buža, Pendolino, Karbonaca, Leccino, Italian rosinjola, Nocciara, Frantoio, Maurino, Porečka rosulja and two unknown varieties in Croatian orchards. On infected plant material from Portugal, Pss was isolated only from knots that had developed on Madural and Cordovil varieties. Other cultivars from Portugal exhibited typical Pss symptoms but bacterial isolation was unsuccessful due to the advanced ageing of the knots.
A total of 54 out of 60 (90%) bacterial isolates from Slovenia were identified as P. savastanoi by PCR ( Figure S2). The six isolates that tested negative for Pss came from the Itrana variety and were subsequently identified as belonging to non-pathogenic species of Curtobacterium flaccumfaciens (Hedges) Collins and Jones and Bacillus megaterium de Bary (Supplementary S1 and Table S1).
The MALDI-TOF spectra derived from three Croatian Pss strains isolated from olive varieties Leccino (17L), Frantoio (B19F) and Nocciara (P15N) and Pss Slovenian strains isolated from knots on varieties Arbequina (A1-1), Frantoio (F3-4) and Leccino (L1-1) were selected for analysis due to variabilities in the LOPAT scheme test results for these strains ( Figure S3). These strains were homogeneous and clearly identified as Pss by the Bruker software, even though the protein profile of P. syringae pv. syringae-the closest genetic relative-was very similar. Typical spectra of some local Pss isolates, along with spectra originating from the Pss reference strain and from P. syringae pv. syringae, are shown in Figure 1, revealing the uniformity of Pss strains and their close resemblance to the P. syringae pv. syringae spectrum. Two small protein peaks-corresponding to approximately 7500-8000 m/z, and between 10,000 to 11,000 m/z-in Pss isolate samples were lacking in the P. syringae pv. syringae strain spectra.
In Figure 2, a comparison is shown between the colony morphology o the CFPB5075 strain-used here as a reference for Pss grown on King' B (KB) agarised medium-and of representative Pss isolates from Croatia, Slovenia and Portugal. All isolates exhibited similar colony morphology, with a mostly rounded or slightly oval shape. Margins were smooth or corrugated and brighter than the colony centre. Isolates of Pss formed colonies with pale white to yellow pigment on KB solid medium. In particular, colonies from Croatian and Slovenian strains were mostly white-greyish, whilst those from Portuguese strains had a yellowish pigmentation. Colony size ranged from 0.5 to 3.0 mm in diameter. Most Pss strains from Slovenia formed colonies of 0.5 to 1.0 mm in diameter after three days of incubation on KB. The diameter of Pss isolates from Croatia ranged from 1.0 to 3.0 mm, whilst those of Portuguese Pss isolates ranged from 1.0 to 2.0 mm after two days of growth ( Figure S4).
In the biochemical characterisation, 34 out of 88 (38.6%) isolates tested negative for oxidase, pectolytic activity and the presence of arginine dihydrolase (Tables 3 and S2). Based on these results, these isolates were classified as group Ib, as expected [29]. The exceptions were Slovenian isolates, which were generally delayed positive for oxidase activity. Only four Slovenian isolates (i.e., 13.3% of the total Slovenian strains coded as L6-2, L5-1, F4-2 and At2-4) were negative for oxidase activity. Most Croatian isolates did not produce fluorescent pigment under UV light when grown on KB solid media. Conversely, all isolates from Slovenia and Portugal tested positive for the in vitro production of fluorescent pigments when grown on solid KB medium. In Figure 2, a comparison is shown between the colony morphology o the CFPB5075 strain-used here as a reference for Pss grown on King' B (KB) agarised medium-and of representative Pss isolates from Croatia, Slovenia and Portugal. All isolates exhibited similar colony morphology, with a mostly rounded or slightly oval shape. Margins were smooth or corrugated and brighter than the colony centre. Isolates of Pss formed colonies with pale white to yellow pigment on KB solid medium. In particular, colonies from Croatian and Slovenian strains were mostly white-greyish, whilst those from Portuguese strains had a yellowish pigmentation. Colony size ranged from 0.5 to 3.0 mm in diameter. Most Pss strains from Slovenia formed colonies of 0.5 to 1.0 mm in diameter after three days of incubation on KB. The diameter of Pss isolates from Croatia ranged from 1.0 to 3.0 mm, whilst those of Portuguese Pss isolates ranged from 1.0 to 2.0 mm after two days of growth ( Figure S4).  In the biochemical characterisation, 34 out of 88 (38.6%) isolates tested negative for oxidase, pectolytic activity and the presence of arginine dihydrolase (Tables 3 and S2). Based on these results, these isolates were classified as group Ib, as expected [29]. The exceptions were Slovenian isolates, which were generally delayed positive for oxidase activity. Only four Slovenian isolates (i.e., 13.3% of the total Slovenian strains coded as L6-  Tobacco HR + + + + more than 70% of isolates were positive for tested characteristic; -more than 70% of isolates were negative for tested characteristic; DP-delayed positive.

Discussion
The field research described in this study included data collected from over 600 hectares of olive orchards surveyed for olive knot disease in Croatia, Slovenia and Portugal. Interestingly, variety Istarska bjelica from Slovenian orchards did not show symptoms of olive knot disease, whereas in Croatia, symptoms were observed on this variety. This observation suggests that the Pss infection of Istarska bjelica variety in Slovenia is somehow suppressed.
Notably, higher total phenolic content-considered a major defense mechanism against Pss-has been observed in different plant tissues of Istarska bjelica compared to other olive varieties [30,31]. In a recent study of 21 different varieties, all showed a degree of susceptibility to olive knot disease [32], although Istarska bjelica was not included. Interestingly, the variety Itrana in Slovenian orchards showed symptoms of olive knot disease, but the presence of Pss was not detected by PCR. The only non-pathogenic bacteria identified by MALDI-TOF were C. flaccumfaciens and B. megaterium. The effects of the coinfection of these bacterial species with Pss needs to be investigated to determine whether they exert a suppressive effect on Pss growth in infected olive trees [33,34]. Whilst typical symptoms of olive knot disease were observed in all surveyed Portuguese orchards, Pss as the causal pathogen was confirmed only in two samples. However, the isolation of bacteria from these samples was difficult due to the severe aging and cracking of the knots. Despite the challenges, this observation raises the question of whether antagonistic organisms in the surveyed Portuguese region may have infected those trees subsequent to Pss infection and consequently suppressed further proliferation of Pss. The presence of Pss has been detected in Portuguese orchards that also harbour various bacterial species identified as antagonists of Pss [34], which may contribute to the suppression of Pss growth in the cases we have described.
The epidemiology of olive knot disease in other significant olive-growing countries such as Croatia and Slovenia has been poorly studied up to now. In Slovenia, the first molecular identification of Pss was reported in 2016 in symptomatic varieties Leccino, Maurino, Pendolino, Frantoio and Ascolana tenera [35], whilst the first detailed molecular characterisation of a Croatian strain was reported in 2019 [36]. The latter study determined the whole genome sequence of Pseudomonas sp. strain ST1 (belonging to P. amygdali species), isolated from olive knots from Croatia [33,36]. However, there remains a lack of LOPAT characterisation data on Pss that inhabit the olive orchards of Croatia, Slovenia and Portugal, creating a need to evaluate whether such Pss isolates have biochemical profiles consistent with pathogenic, tumour-inducing olive-plant bacterial species.
The isolation of bacteria from knots, together with their molecular identification, confirmed that these symptoms are caused by Pss in 46.6% of the sampled orchards. Negative results for the presence of Pss in sampled plant materials might be caused by aged, dry and cracked knots or may be due to the absence of Pss in orchards that have plants displaying symptoms of olive knot disease from prior infections. All isolated Pss strains included in this study originate from olive-growing micro-regions, so the biochemical similarities between them are expected, as was observed in the case of levan-positive strains from a limited population distribution in Italy, for example [21,22]. However, some differential characteristics can be determined within micro-regions (even between strains isolated from the same plant [29,37]), similar to what was observed in the present study. For molecularly confirmed Pss isolates, LOPAT results were recorded that deviated from those expected for segregated strains. Additionally, differences between Pss strains isolated from the same orchard were noted. All isolates from Portugal (coded as PT15 and PT17) showed typical LOPAT results for olive strains [37]. Pss formed mostly round or slightly oval shaped colonies, with white-greyish to yellowish pigmentation, with a diameter range from 0.5 to 3 mm. However, some variability in Pss colony morphology was observed ( Figure S4), which has previously been ascribed to adaptation to different environmental conditions or horizontal gene transfer [37,38].
Although the differences in the phenotype of Pss colonies were expected, the results of the LOPAT biochemical identification of Pss isolates showed a clear separation, or even grouping, of most Slovenian strains. The majority of Slovenian strains exhibited delayed positive oxidase activity. Another grouping, which contained most of the Croatian isolates, was characterised as having a lower occurrence of fluorescent pigment on solid KB medium in vitro. However, the mass spectroscopy analysis showed high similarities in the protein profiles of selected Slovenian and Croatian strains (Figure 1). These results are interesting, given that Pss was confirmed on a single plant species (olive), wherein similarities in biochemical features are to be expected [39]. However, from a genetic point of view, some segregation of strains occurs as a result of different geographical origins or factors acting within a specific ecological niche, which might also occur on a micro-location level. Although the fluorescent pigment was absent in Croatian strains, this feature is considered variable for Pss bacteria [21,38,40]. It has been observed that fluorescence does not always occur in so-named 'fluorescent' Pseudomonads [41], and our results support this. Similar results were determined for some Australian and Japanese strains [20,21]. However, it is important to mention that the absence of fluorescent pigment on KB medium usually occurs when strains are levan-positive, which did not occur in our case.
In light of the interesting results obtained in this study, more genetic and phylogenetic research should be done to better understand the pathogen's virulence in relation to geographical origin and olive cultivars. This study also contributes to the mapping of the geographical distribution of pathovar Pss, with the aim to update available databases that still present the distribution of three different pathovars of P. savastanoi tumour-inducing bacteria (Pss, Psn and Psf ) as one tumour-inducing species. Clear differentiation is obligatory, due to the established differences in the virulence of these woody-host pathogens [15,42]. Furthermore, the study of the susceptibility of Croatian and Slovenian autochthonous varieties to olive knot disease should be examined. These autochthonous olive cultivars could elicit some resistance traits, which are as yet unknown. According to more recent studies, including ours, an update of the LOPAT scheme should be considered, due to deviations from the expected grouping of Pss isolates [18][19][20][21][22]. In the literature [42], it is stated that strains of Pss originating from the former Yugoslavia region are amongst the least-virulent strains, whilst strains from Portugal are of intermediate pathogenicity, which demonstrates the existence of virulence differences between Pss strains according to geographical origin. These observations warrant further examination of the role of geographical and environmental factors of Pss strain grouping in the future studies.  Table 1). Samples of symptomatic plant material were sealed in plastic bags, correctly labeled and kept in a portable fridge. After field research, samples were stored at +4 • C until processed.

Bacterial Isolation and Characterisation
The bacterial isolation was performed from knots of randomly collected infected olive branches. Olive knots were washed under tap water to remove dust, followed by surface disinfection with 70% ethanol and left to dry. After disinfection, knots were aseptically cut into small pieces and placed into sterile distilled water for 30 min. The resulting suspensions were serially diluted 10-fold, two times. A volume of 100 µL of each sample dilution was placed on King's B (KB) solid growth media supplemented with cycloheximide (100 mg/mL). The morphology of the obtained colonies was visualised by a Trinocular Microscope BOE 2200.520 (BOECO, Hamburg, Germany) at 4.00× magnification. The biochemical and physiological characterisation of bacterial colonies was carried out according to the LOPAT scheme [18,43]. For this purpose, pure bacterial cultures were spread on growth media. Bacterial cultures were characterised based on fluorescence (366 nm wavelength) on KB, measured after 24-48 h of incubation, Gram staining using 3% potassium hydroxide (KOH) and the LOPAT scheme. A bacterial inoculum of 10 8 CFU/mL was used for physiological and biochemical tests. For determination of levan production, pure cultures were streaked onto 5% sucrose nutrient agar (SNA) and incubated for 3-5 days for the development of white mucoid colonies in positive isolates. The oxidase reaction test was performed using commercial MAST ID TM Oxidase strips (Mast House, UK) following the manufacturer's instructions. Pectolytic activity was determined on disinfected potato slices in Petri dishes with moistened filter paper, whilst the presence of arginine dihydrolase was determined using Thornley's medium 2A with pH adjusted to 7.2. The total volume used was 5 mL. A colour change from pink to red within four days was recorded as a positive reaction. The hypersensitivity reaction was carried out on tobacco variety Nicotiana tabbacum L. cv. Samsun using syringae without a needle for inoculation with bacterial suspensions. The complete collapse of impregnated leaf tissue within 24 h was recorded as a positive result. The morphological characterisation of bacterial colonies included measuring diameter (mm) after 24 h, shape, pigmentation, margin shape and smoothness. All isolates from Croatia and Portugal (of which there were 35 and 2, respectively) were characterised. For the characterisation of Slovenian isolates, 30 Pss strains obtained from different olive varieties were selected at random.

Molecular Identification of Pss
The specific detection of Pss was achieved by real-time PCR using TaqMan ® probes according to the method of Tegli et al. [24] after growing bacteria on KB and PVF-1 media incubated at 26 • C for 48 h. Bacterial DNA was isolated using a Maxwell ® RSC Instrument (Promega, Madison, WI, USA) and Maxwell ® RSC Cultured Cells DNA Kit (Promega, Madison, WI, USA), following the manufacturer's instructions. Primers and probes sequences were as follows: PsvRT-F 5 CGGATTTGGTTTGCGGGGTA 3 , PsvRT-R 5 AATGGGGT-GACACTAAAAATTGTGAA 3 , PsvRT-P 5 VIC-CTCGTGCGATCTAAACAGCCGTAGC-QSY 3 .
All obtained isolates and the reference strain were included in this survey. Reference strain CFBP5075 (Italy) was purchased from the National Institute of Agricultural Research, INRA (Paris, France). The PCR reaction was performed in an Applied Biosystems Quant Studio 1 PCR machine (Thermo Fisher Scientific, Waltham, MA, USA). MicroAmp ® Optical 96-well reaction plates from Applied Biosystems ® (Waltham, MA, USA) were used. The reaction mixture volume was 50 µL using the following reagent concentrations: 1x TaqMan ® Universal PCR Master Mix, 300 nM of each primer (PsvRT-F and PsvRT-R), 100 nM of probes (PsvRT-P) and 3 µL of DNA.
For Slovenian isolates, the identification of Pss from olive knots was based on a standard PCR assay [22]. Bacterial DNA was extracted from pure cultures by the boiling method. A loop-full of bacteria grown on KB medium for 24 h at 26 • C was suspended in 500 µL of sterile water, heated to 95 • C for 10 min and immediately cooled on ice. Dilutions (1:1000) were prepared in sterile TE buffer. For the amplification of the iaaL gene, we used IAALF and IAALR primers designed by Penyalver et al. [23], which produced a 454 bp-long amplified fragment. PCR assays were performed in a final volume of 25 µL in a reaction mixture containing 1x GoTaq PCR buffer (Promega), 1.

Protein Mass Spectra Analysis
Protein profiles of six Pss strains from Croatia and Slovenia, and the reference strain CFBP5075 (Italy), were examined by MALDI-TOF mass spectrometry. The analyses were performed on a Bruker Microflex LT mass spectrometer (Bruker Daltonics, Billerica, MA, USA), according to the manufacturer's instructions and using direct transfer sample preparation with standard α-cyano-4-hydroxycinnamic acid (HCCA) bacterial cell extraction [44]. The obtained spectra were compared to those of the Pss reference strain and their closest phylogenetic neighbour, P. syringae pv. syringae. Briefly, the bacteria were grown on KB medium for 24 h at 28 • C. A small fraction of a single colony from each strain was then transferred to a corresponding spot on the Bruker 96-well target plate and treated with 1 µL HCCA reagent. After plate insertion and following the MALDI-TOF sample processing procedure, 40 sub-spectra for each of the 40 randomised positions within the spot were collected and presented as one main spectrum for each spot/bacterial strain. A mass range of 2000 to 20,000 Da was used for the analysis. The mass spectra profiles were identified from the mass spectra profiles using the Bruker MBT Compass HT software.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/plants12020307/s1, Supplementary S1: Slovenian PCR positive Pss isolates confirmed using mass spectrometry; Table S1: Bacteria strains isolated from olive knots in Slovenia identified using mass spectrometry; Table S2: Characterisation of molecularly identified Pss isolates; Figure S1: Amplification plots of positive Pss isolates from Croatia and Portugal. Figure S2: Gel electrophoresis of end-point PCR identified Pss isolates from Slovenia. Figure S3: Biochemical and physiological characteristics of Pss isolates. Figure S4: Colony morphology of tested Pss isolates.