Introduction

Pseudomonas syringae pv. lachrymans, a pathogen of cucumber (Cucumis sativus L.), is one of the pathovars belonging to the heterogeneous bacteria species Pseudomonas syringae (Young et al. 1996). It causes angular leaf spot (ALS), a very common cucumber disease. The symptoms of ALS consist of vein-limited water-soaked lesions, which later become necrotic, resulting in significant yield reduction (Bradbury 1986). Although this disease is ascribed to a single pathovar, there are numerous strains of this pathovar that vary from each other at biological, biochemical and genetic levels. Moreover, similar symptoms on cucumber may be caused by strains belonging to pathovar syringae (Olczak-Woltman et al. 2007).

Recent studies of P. syringae reveal that various strains differ dramatically in genome size, GC content, plasmid presence, and pathogenesis-related gene repertoire including effector-encoding genes (Baltrus et al. 2011; O‘Brien et al. 2011). These variations illustrate P. syringae genome plasticity and indicate the dynamic nature of this species. Analysis of the 16S variable region and of the Intergenic Spacer 1 (ITS1) region used to be commonly employed in bacterial strain differentiation and identification. In addition, intraspecific typing of bacteria can be done using the Multi-locus Sequence Typing (MLST) method. MLST is preferred for determination of genomic relatedness among bacterial strains because it allows comparison of sequences of housekeeping genes that encode proteins necessary for the microorganism’s survival (Maiden et al. 1998; Urwin and Maiden 2003). MLST can be applied to Pseudomonas syringae (Sarkar and Guttman 2004). Another approach allowing bacterial isolate identification is based on PCR amplification of bacterial repetitive sequences (rep-PCR technique). In this approach Enterobacterial Repetitive Intergenic Consensus (ERIC) sequences and Repetitive Extragenic Palindromic (REP) sequences are analyzed. These are short, repetitive DNA sequences with highly conserved central inverted repeats that are dispersed throughout the genomes of diverse bacterial species (Versalovic et al. 1991). Designing universal rep-PCR primers (ERIC and REP) allows generation of highly reproducible fingerprints that can differentiate strains below the level of species (Louws et al. 1994).

The aim of this study was to determine the genetic diversity among strains of Pseudomonas sp. collected from different cucurbit plants using the MLST method, ITS1 and rep-PCR primers. The molecular identification was verified by pathogenicity tests performed on cucumber and tobacco plants. Additionally, the aim of this study was to determine an efficient molecular technique for identification of various strains of P. syringae pv. lachrymans. We also looked for presence of plasmids in the collected bacterial strains and investigated whether there was a correlation between the plasmid’s presence and the strain’s increased pathogenicity and/or virulence.

Materials and methods

Bacterial strains

The bacterial collection consisted of 22 Pseudomonas strains. Nine strains were isolated from cucumber leaves that were showing angular leaf spot symptoms, collected in Poland (Olczak-Woltman et al. 2007). The others were obtained from the European bacterial culture collections (Table 1). All strains were stored on nutrient agar slopes at 4 °C and only pure cultures were used in molecular analysis and assays.

Table 1 Bacterial culture collection – name, place and plant of isolation, the HR results on tobacco and cucumber pathogenicity and virulence tests with DSI and symptoms description
Full size table

Cucumber pathogenicity and virulence tests

To verify the strain pathogenicity on cucumber and to define the virulence level, the seeds of susceptible to ALS Cucumis sativus inbred line B, selected from variety ‘Borszczagowski’, were individually sown into plastic pots filled with peat moss. The plants were maintained in a growth chamber at 25 °C during the day and 22 °C at night, with 16 h illumination under high pressure sodium lamps providing light intensity of 50 W m−2.

Inoculum was prepared by growing bacterial cells for 24 h on King B agar plates at 28 °C in the dark. The resulting colonies were suspended in sterile distilled water and adjusted to OD600 = 0.050 previously established to be equal to concentration of 1 × 107 CFU ml−1. Cucumber plants at the 2nd to 3rd leaf stage were inoculated by spraying the abaxial side of each leaf (Klement et al. 1990). Sterile water was used as a control. After inoculation, plants were kept in the dark at 100 % RH for 24 h, and then for 6 days in the growth chamber under light and with 90 % relative humidity. After 7 days, inoculated leaves were scored for disease severity and appearance of symptoms. The evaluation using a 1–9 disease severity index (DSI) where the most resistant plants were given the score of nine and the most susceptibility plants were given the score of one, was performed according to the protocol previously described (Olczak-Woltman et al. 2008). The pathogenicity tests of each strain were performed using a total of 24 plants (with four replications of six plants in each replication).

Pathogenicity verification

The strain pathogenicity was verified using Nicotiana benthamiana as a host. The seeds of N. benthamiana were sown into plastic pots filled with peat moss. The growth chamber conditions and bacterial inoculum preparation were the same as for the cucumber pathogenicity and virulence tests. Nine-weeks-old tobacco plants were inoculated by infiltration of the abaxial side of each leaf (five separate points per leaf) with bacterial suspension. Infiltration using sterile water was used as a control. Each strain was tested on a separate tobacco plant. Additionally, some tobacco plants were inoculated with four different strains on the same leaf. Plants were kept in the dark at 100 % relative humidity for 24 h, and then for 48 h in the chamber with light (50 W m−2) and 90 % relative humidity. The lesions were scored for the presence of hypersensitive reaction (HR) 3 d after inoculation.

Bacterial DNA isolation

After growing the cultures for 24 h in LB liquid medium on a rotary shaker at 28 °C, total genomic DNA was extracted from 1 ml of culture using the DNA Genomic-tips 100/G kit (Qiagen, Germany), following the manufacturer’s instructions. The DNA concentration was estimated by electrophoresis of 2 μl DNA, together with a DNA marker, on a 1 % agarose gel with 1× TEA buffer and stained with ethidium bromide.

Plasmid detection

Plasmid bands were detected and visualised using Eckhardt gel electrophoresis technique (Eckhardt 1978; modified by Maj et al. 2013). The bacterial cultures were grown overnight in LB liquid medium on a rotary shaker at 28 °C. The free-living cells contained in 1 ml of LB culture at OD620 = 0.20 were harvested by centrifugation. The pellet was suspended in 0.5 ml of cold water and layered onto 1 ml of ice-cold 0.3 % sodium sarcosinate, centrifuged again, and the pellet was resuspended in the mixture (40 μl) of 10 mM Tris–HCl, 10 mM EDTA, and 20 % Ficoll (MT 400 000). The samples were held on ice for 10 min. A 0.75 % agarose gel with 0.5× TBE buffer was prepared by loading 25 μl of 10 % SDS containing xylene cyanole FF (1 mg ml−1) into each origin slot and backtracking at 100 V for 30 min. Next, 10 μl of lysis solution (10 mM Tris–HCl, 10 mM EDTA, 0.4 mg ml−1 boiled RNase type A, 1 mg ml−1 lysosyme, and 1 mg ml−1 bromophenol blue) was mixed with each sample. Then 25 μl of such sample was loaded into the gel. The electrophoresis was run for 15 h and the gel was strained with ethidium bromide (0.5 μg ml−1) for 20 min. Strain 1448A of P.syringae pv. phaseolicola, that contains two plasmids of known size (Joardar et al. 2005), was used as a positive control in plasmids visualization on the gel.

Amplification of the ITS1 region

DNA amplification was performed using D21 and D22 primers (Manceu and Horvais 1997) according to the protocol of Sorensen et al. (1998) with modifications of Olczak-Woltman et al. (2007). The PCR amplifications were performed using a PTC-200 thermocycler (MJ Research, USA) with the following parameters: template denaturation at 94 °C for 1.5 min, primer annealing at 62 °C for 1 min, and DNA extension for 2 min at 72 °C. The PCR was repeated for 30 cycles, with a final extension step of 10 min at 72 °C. Amplicons were detected by electrophoresis on a 1 % agarose gels with 0.5× TBE buffer, stained with ethidium bromide and photographed using an UV trans-illuminator.

Amplification of the selected MLST loci

Identification of P. syringae pv. lachrymans strains was done by PCR amplification using primers for the following housekeeping genes: aconitate hydratase B (acnB), citrate synthase (cts), glyceraldehyde-3-phosphate dehydrogenase (gapA), gyrase B (gyrB), phosphofructokinase (pfk), phosphoglucoisomerase (pgi) (Sarkar and Guttman 2004; Hwang et al. 2005). In addition, we performed an amplification of harpin elicitor (hrpZ) (Inoue and Takikawa 2006) using the same conditions as for the MLST loci. The characteristics of the used MLST loci and the harpin elicitor, i.e. the name, encoding protein, melting temperature, length, primer sequence, amplicon length and the author are listed in Table 2. Reactions (repeated twice) were carried out for all strains in the collection, in the same PCR and detection conditions as described for amplification of the ITS1 region.

Table 2 The ITS1, hrpZ gene fragment and MLST primers used in this study
Full size table

Amplicons sequencing

The successful PCR reactions, in which the resulting target PCR product was obtained, were collected and commercially sequenced (Genomed, Poland). Contigs were assembled using Sequencher 4.5 Software (Gene Codes, USA). Consensus sequences were submitted to the GenBank and accession numbers are provided in Table 3.

Table 3 Consensus sequences submitted to the NCBI GenBank with accession numbers
Full size table

Strain differentiation using ERIC and REP fingerprinting

Primer sequences corresponding to ERIC elements (ERIC1R: 5′-ATGTAAGCTCCTGGGGATTCAC-3′ and ERIC2: 5′-AAGTAAGTGACTGGGGTGAGCG-3′) and corresponding to REP elements (REP1R-I: 5-′IIIICGICGICATCIGGC-3′ and REP2-I: 5′-ICGICTTATCIGGCCTAC-3′) were used (De Bruijn 1992). The reaction mixture (20 μl) consisted of 10× Dream Taq buffer (Thermo Scientific, Lithuania), 0.4 μM of dNTPs Mix, 0.75 μM of each primer, 1.0 U of DreamTaq DNA polymerase and 4 μl of bacterial suspension. The PCR amplifications were performed in a PTC-200 thermocycler (MJ Research, USA) with the following parameters: initial denaturation at 95 °C for 7 min, template denaturation at 94 °C for 1 min, primer annealing at 52 °C (ERIC) or at 41 °C (REP) for 1 min, and DNA extension for 8 min at 65 °C. The PCR was repeated for 30 cycles, with a final extension step of 15 min at 65 °C. Amplicons were detected by electrophoresis on a 1.5 % agarose gels with 0.5× TBE buffer, stained with ethidium bromide and photographed using a digital camera and a UV trans-illuminator.

Data analysis

Polymorphisms generated by MLST and ITS1 in PCR reactions and visualised after sequencing were analysed using Geneious R7 software (Biomatters, USA). Genetic similarity (%) among the bacterial strains was calculated and the alignments were made using ClustalW method. Dendrograms were constructed using UPGMA clustering method. Polymorphisms generated by ERIC- and REP- PCR reactions were analysed using GelCompar II v.6.6 software (Applied Maths, USA). Genetic similarity (%) of bacterial strains patterns was calculated using the Dice coefficient and a dendrogram was constructed using the UPGMA clustering method.

Sequences were aligned to genome sequences of P. syringae strains collected in the JGI database (DOE Joint Genome Institute, USA) using BLAST algorithm.

Results

Hypersensitivity reaction on tobacco leaves

All but four strains produced a HR on N. benthamiana leaves 3 d after infiltration, confirming their pathogenicity (Table 1, Fig. 1). Strains BG283, PD2760, WH1/01, and WH2/01 did not produce a HR on tobacco.

Fig. 1
figure 1

The results of the HR test performed on Nicotiana benthamiana with visualisation of differences in pathogenicity. A tobacco leaf three days after infiltration. This leaf was infiltrated with four different strains, all classified as P.s. pv. lachrymans: 814/98 of high virulence level, LMG5070 and BG966 of medium virulence level, and BG283 of low virulence level

Full size image

Pathogenicity and virulence tests on cucumber leaves

The bacterial strains exhibited diverse pathogenicity and virulence levels when tested on susceptible cucumber accession line B (Fig. 2 a–f). P. syringae pv. lachrymans strains 814/98 and CCM2857 were the most virulent and caused typical angular leaf spot symptoms on cucumber leaves, i.e. water-soaked, later necrotic lesions with a large chlorotic halo, as well as the bacterial ooze on the abaxial side of the inoculated leaves (Fig. 2a). P. s. pv. lachrymans strains LMG5070 and BG966 were far less virulent on cucumber leaves, causing fast drying water-soaked lesions, and dry, light-coloured, papery lesions, with numerous pinpoint chlorotic lesions (Fig. 2b). Strain CCM2858 was a very virulent one, causing characteristic drying of the part of cucumber leaf, but although the lesions were very large in size, they differed dramatically from the lesions caused by other strains. Strain 2905 of pathovar syringae and strains BG913, WW17/01, WW27/01, and WW254/02 caused numerous pinpoint chlorotic or light-coloured necrotic lesions on cucumber leaves. In the case of strain BG913, these symptoms were the most severe, and in case of WW17/01 and WW27/01 less severe (Fig. 2d). Similar symptoms were caused by LMG5070 and BG966 strains. Strains CCB37/09, PD2021, and WK1/02 caused on cucumber large, dry, light-coloured necrotic lesions without chlorosis (Fig. 2e). Similar lesion, although smaller in size, were caused by strains PD3662, WH6/01, WM2/02, and YPG1293. Strains BG283 (Fig. 2c), PD2760, WH1/01, WH2/01 (Fig. 2f), and WW4/02 produced only very weekly chlorotic lesions, HR reactions, or did not produce any symptoms on cucumber leaves.

Fig. 2
figure 2

The results of the pathogenicity and virulence tests performed on cucumber leaves of susceptible line B. Pictures were taken under day light (upper line) and under UV light (lower line—red leaves). The following strains are presented: a – Psl 814/98, b – Psl BG966, c – Psl BG283, d – Pss WW27/01, e – Pss WK1/02, f – Pf WH2/01. All strains were tested on line B - susceptible to angular leaf spot. Leaves were sprayed with inoculum containing 1 × 107 CFU ml−1 and results were recorded 7 days after inoculation

Full size image

Amplification of the ITS1 region

Amplification of the 16S-23S ITS1 region allowed us to obtain amplicons of 560–574 bp long in all strains. However, despite of several attempt sequencing of this fragment was possible only in case of 10 out of 22 strains.

Amplification of the hrpZ gene fragment

Amplification of the hrpZ gene fragment produced 810 bp long amplicon. However, despite several attempts, this fragment was obtained and sequenced for only four strains: 814/98, BG283, CCM2857, and PD3662.

Amplification of the MLST loci

Amplification of the 7 MLST loci resulted in informative amplicons that further facilitated grouping of strains. The obtained amplicons ranged from 530 to 970 bp, depending on the MLST locus (Table 2). Although the obtained amplicon quality was satisfactory, in the case of acn, pgi, and rpoD amplification failed in 1 strain, in case of gyrB and pfk - in 2 strains; in case of gapA – amplification of 3 strains failed. Amplicons for cts were obtained in all strains. The obtained amplicons were sequenced and the consensus sequences of all alleles were submitted to the NCBI GenBank (the consensus sequences symbols and numbers are provided in Table 3).

Strains differentiation by ERIC and REP fingerprinting

ERIC and REP showed significant diversity in the banding patterns of all strains. Each pattern consisted of approximately 11–22 bands, ranging in size from 130 to over 3,000 bp. The strains’ banding patterns differed from each other, although in case of ERIC (Figs. 4 and 5) and REP (data not shown) strains LMG 5070 and BG966 showed identical band patterns.

Sequences alignment

The obtained ITS1, hrpZ, and MLST sequences were aligned to sequences deposited in the JGI base. The alignment of 10 obtained ITS1 amplicons allowed us to divide strains into four groups. Based on the ITS1 loci analysis, strain 814/98 exhibited 99 % of genetic similarity to lachrymans strain Pla107 deposited in the JGI base. Strains LMG5070 and BG966 showed the highest genetic similarity to lachrymans strain Pla106. Strains CCM2858, PD2021, and WW4/02 exhibited the highest genetic similarity to strains of pathovar syringae, and the referenced strain 2905 confirmed its classification in this pathovar. Strains PD2760, WH1/01, and WH2/01 turned to be similar to Pseudomonas fluorescens.

The hrpZ amplicons were obtained only for strains 814/98, BG283, CCM2857, and PD3662. The alignment to the sequences deposited in the JGI database suggested that all these strains belong to pathovar lachrymans. In the case of 814/98 and CCM2857 the alignment indicated the highest similarity to lachrymans strain Pla107.

The MLST alignment allowed a very accurate strain grouping. All obtained MLST amplicons classified strains into the same pathovar, either lachrymans or syringae, or into P.fluorescens. Strain 814/98 according to loci gapA, gyrB, rpoD, cts, acn, pgi, and pfk was classified as pathovar lachrymans, with the highest similarity to strain Pla107. The same results were obtained for CCM2857, except that for locus pgi an amplicon was not generated. Strain BG283 exhibited the highest genetic similarity to strain Pla107 when using loci gyrB, rpoD, cts, and pgi. There were no amplicons for loci acn and pfk; and for locus gapA exhibited the highest similarity to pathovar syringae. This suggests that this strain may be classified as such but does not have to belong to pathovar lachrymans. Strains LMG5070 and BG966 in all tested MLST loci showed the highest similarity to pathovar lachrymans strain Pla106 and to pathovar tomato. Strain 2905 confirmed the highest similarity to pathovar syringae in all tested MLST loci. The same result was obtained for strain WW4/02 (although this does not correlate with the pathogenicity tests) and for strain WK1/02 (except for gyrB locus, where it was similar also to pathovar pisi). Most of the strains that exhibited high similarity to pathovar syringae in 5 or 6 loci (i.e. BG913, CCB37/09, CCM2858, WW17/01, WW27/01, WW245/02) showed similarity to pathovar aptata at the pfk locus, and to pathovar pisi at the acn locus (strains WW17/01, WW27/01, and WW254/02). Strain PD2021 showed high similarity to pathovar syringae at 5 loci, failed to produce amplicons at the acn locus and showed high similarity to pathovar pisi at the pfk locus. Four strains: PD3662, WH6/01, WM2/02, and YPG 1293, exhibited similarity to various pathovars at various MLST loci, that is to lachrymans (Pla106 and/or Pla107), syringae, tomato, maculicola, aesculi, actinidiae, or P.fluorescens. Therefore classification of these strains into one certain pathovar was not possible. Three strains in our collection (PD2760, WH1/01, and WH2/02) turned out to be P.fluorescens strains or at least exhibited the highest genetic similarity to this species at 6 MLST loci for PD2760, 5 loci for WH1/01 and 3 MLST loci for WH2/01.

Analysis of genetic similarity

Sequences of all MLST loci were used to construct phylogenetic trees (dendrograms) of genetic similarity. Based on the constructed dendrograms and alignment results collected strains were divided into four main groups (Figs. 5, 6 and 7). The first main group consisted of strains that belonged to pathovar lachrymans (strains 814/98, BG283, BG966, CCM2857, and LMG5070). These strains may be further divided into two separate clusters. The first cluster consisted of strains 814/98, BG283 and CCM2857, genetically very similar to each other, as well as to reference lachrymans strain Pla107 (JGI database). However, they differed in pathogenicity on cucumber, as strain BG283 exhibited only very week virulence. The second cluster consisted of two strains of P.s. pv. lachrymans: BG966 and LMG5070. These strains were very similar to each other and to reference lachrymans strain Pla106 (Figs. 6 and 7). The second main group is composed of pathovar syringae strains. This group consists of BG913, CCB37/09, CCM2858, PD2021, WK1/02, WW17/01, WW27/01, WW254/02, and WW4/02 and the reference syringae strain 2905 (Figs. 6 and 7). In this group, strain WW4//02 is an outlier, and strains WW17/01 and WW27/01, as well as PD2021 and CCB37/09 form two sub-groups (Fig. 7); this classification correlates with pathogenicity test results. The third main group consisted of three strains of P.fluorescens. The strains PD2760, WH1/01, and WH2/01 were classified as P.fluorescens based on the high genetic similarity to sequences of this species deposited at the JGI database. These strains exhibit high genetic similarity to each other, and form a separate branch on dendrograms (Fig. 5, 6 and 7). The last group (PD3662, WH6/01, WM2/02, and YPG 1293) contains strains that do not exhibit similarity to any given pathovar. However, in pathogenicity tests these strains show similarities, and MLST assays form a separate dendrogram branch for them (Fig. 7).

Presence of plasmids

Either plasmids or mega-plasmids were detected in 10 strains (Fig. 3). However, each strain contained a different set of plasmids. The presence or lack of plasmids, and their number and/or length, did not correlate with pathogenicity tests or with molecular characterizations. Strain 814/98 possessed four plasmids estimated to be about 40 kb, 50 kb, 100 kb, and 200 kb in size. The other strains possessed one plasmid each: strains BG283, BG966, and WW17/01 had a plasmid about 40–50 kb in size, strain BG913 about 80–90 kb, and strain WW4/02 about 100 kb in length. Strains WK1/02, WW254/02, CCB37/09, and PD2021 likely possessed a mega-plasmid which showed as a smeared high molecular weight band on the gel. Plasmid detection and visualisation was performed in a comparison to referenced strain P.s. pv. phaseolicola 1448A (Fig. 3). There was no obvious association between plasmid presence and strain pathogenicity or virulence.

Fig. 3
figure 3

The plasmid bands isolated from the collected Pseudomonas strains together with a positive control. Eckhardt gels for the 23 Pseudomonas sp. strains. Lane 1 – a positive control Pseudomonas syringae pv. phaseolicola 1448A strain (contains two plasmids of know size – 51,711 kb and 131,950 kb), than the collected strains: lane 2 – Psl 814/98, lane 3 – Psl CCM2857, lane 4 – Psl BG283, lane 5 – Psl LMG5070, lane 6 – Psl BG966, lane 7 – Pss WK1/02, lane 8 – Pss WW254/02, lane 9 – Pss BG913, lane 10 – Pss CCB37/09, lane 11 – Pss CCM2858, lane 12 – Pss PD2021, lane 13 – Pss 2905, lane 14 – Pss WW17/01, lane 15 – Pss WW27/01, lane 16 – Pss WW4/02, lane 17 – Pf PD2760, lane 18 – Pf WH1/01, lane 19 – Pf WH2/01, lane 20 – P.sp. PD3662, lane 21 – P.sp. WM2/02, lane 22 – P.sp. WH6/01

Full size image

Discussion

Genetic characterization and clustering, as well as fingerprinting of plant pathogenic bacteria by various molecular methods has been the subject of numerous papers (Versalovic et al. 1994; Weingart and Völksch 1997; Sarkar and Guttman 2004; Inoue and Takikawa 2006; Olczak-Woltman et al. 2007; Lee et al. 2012; Martín-Sanz et al. 2012). These studies showed that the MLST technique and rep-PCR are useful as methods for identification and classification of strains belonging to Pseudomonas syringae (Hwang et al. 2005; Louws et al. 1994). Therefore we decided to use these methods to group Pseudomonas strains isolated primarily from cucumber and kept in our Department’s collection.

The genetic diversity of Pseudomonas strains isolated from cucumber and other cucurbit plants was assessed using the MLST technique, rep-PCR method, and amplification of the ITS1 region. The last, however, is not very useful in P. syringae pv. lachrymans diagnostics due to the multiple copies of the ITS1 region and the intergenomic DNA variation in this region.

The MLSTs of seven housekeeping genes chosen for use in this study have been known to reveal higher genetic diversity within collection of Pseudomonas strains than rep-PCR fingerprinting (Ferrante and Scortichini 2010). In this study we confirmed, that the MLST method is indeed a powerful tool for accurate strain clustering that correlates well with both pathogenicity and virulence assessments.

One of the objectives of this study was identification of an efficient technique for differentiation of pathovar lachrymans strains from among other strains isolated from cucumber plants. Moreover, we investigated the presence of plasmids in pathovar lachrymans and looked for a correlation between plasmid presence and virulence. We grouped and characterized the pathovar lachrymans strains with regard to pathogenicity and molecular tests results. We want to stress that it was not easy to collect strains of pathovar lachrymans because many strains isolated from Cucumis species turned to belong to pathovar syringae.

According to the results, the sequencing, clustering and alignment analysis of our collected cucurbit strain collection allows their division into four main groups. The first main group consists of strains that belong to pathovar lachrymans. These strains differ in the level of virulence and ERIC- and REP-PCR fingerprinting patterns (Figs. 2 and 4). Strains 814/98 and CCM2857 are the most virulent. The analysis of the MLST amplicons showed their high genetic similarity. The product of amplification of the hrpZ gene fragment was present in both strains. Although they differ in the presence or absence of plasmids (four in 814/98 and none in CCM2857), they can be treated as a basic cluster for pathovar lachrymans based on the pathogenicity tests. The alignment with the JGI database sequence indicated high similarity to the lachrymans strain Pla107 that was characterized as highly virulent (Baltrus et al. 2011). The high virulence of both strains and difference in plasmids content suggest no correlation between plasmids presence and virulence for pathovar lachrymans.

Fig. 4
figure 4

The ERIC PCR fingerprinting pattern of the collected Pseudomonas strains. Lane M – 100 bp DNA molecular weight standard. Lines 1–21 – the following bacterial strains: lane 1 – Psl 814/98, lane 2 – Psl CCM2857, lane 3 – Psl BG283, lane 4 – Psl LMG5070, lane 5 – Psl BG966, lane 6 – P.sp. PD3662, lane 7 – P.sp. WM2/02, lane 8 – P.sp. WH6/01, lane 9 – Pss 2905, lane 10 – Pss CCM2858, lane 11 – Pss BG913, lane 12 – Pss PD2021, lane 13 – Pss CCB37/09, lane 14 – Pss WW4/02, lane 15 – Pss WW17/01, lane 16 – Pss WW27/01, lane 17 – Pss WK1/02, lane 18 – Pss WW254/02, lane 19 – Pf PD2760, lane 20 – Pf WH1/01, lane 21 – Pf WH2/01

Full size image

Two strains: LMG5070 and BG966 that showed high molecular similarity to each other and caused similar, but week symptoms on cucumber, did not produce the hrpZ gene fragment. Sequence alignment to strains deposited in the JGI database indicated significant similarity to pathovar lachrymans strain Pla106 (both for the MLST and the ITS1) described as less virulent and genetically separate from strain Pla107 (Baltrus et al. 2011). As before, despite of several attempts, it was not possible to isolate the same plasmids in both of these strains. In contrast to Pla107 and strains 814/98 and CCM2857, both strain Pla106 and strains LMG5070 and BG966 showed very high genetic similarity to pathovar tomato. We suggest therefore, that there are two different types of P.s. pv. lachrymans strains, that correspond to the Pla107 or Pla106 type respectively (Baltrus et al. 2011). Strains Pla107-like cause more developed symptoms on cucumber leaves, and exhibit some genetic similarity to pathovar syringae. Strains Pla106-like, are molecularly similar to pathovar tomato, are much less virulent on cucumber, and cause only weak symptoms.

Strain BG283 posed some problems with classification. This strain produced the hrpZ gene fragment, indicating high similarity to Pla107 strain type and in the MLST-based dendrogram was grouped into pathovar lachrymans type Pla107, but possessed one plasmid of the same length as BG966 (type Pla106), and in the rep-PCR based dendrogram was grouped into type Pla106 (Figs. 5, 6 and 7). However, since it exhibited extremely low level of virulence on cucumber and no HR on tobacco leaves (Fig. 1) this classification is not clear. Perhaps strain BG283 is lachrymans, but lost its virulence, or maybe it even does not belong to pathovar lachrymans.

Fig. 5
figure 5

The dendrogram of the genetic similarity (%) of the collected strains based on the ERIC PCR fingerprinting. The dendrogram was constructed using GelCompare II software, using the Dice coefficient and the UPGMA clustering method

Full size image
Fig. 6
figure 6

The dendrograms of the genetic similarity (%) of the collected bacterial strains based on the selected MLST loci. a – acn, b – gapA, c – gyrB, d – rpoD. Each dendrogram was constructed using Geneious R7 software, with the Tamura-Nei evolutionary model and the UPGMA clustering method

Full size image
Fig. 7
figure 7

The dendrogram of the genetic similarity (%) of the collected bacterial strains based on the results of all seven analyzed in this study MLST loci together. The dendrogram was constructed using Geneious R7 software, with the Tamura-Nei evolutionary model and the UPGMA clustering method

Full size image

The second main group consists of strains that are unequivocally pathovar syringae. The occurrence of this pathovar on cucumber, and its characteristics consisting of atypical numerous pinpoint chlorotic lesions, or dry necrotic lesions without water-soaking has been described previously (Olczak-Woltman et al. 2007), and the classification into pathovar syringae has been verified by amplification of the syrB gene fragment (Olczak-Woltman et al. 2007). In this study ten strains were classified as pathovar syringae (Fig. 7). These strains may further be divided into sub-groups. The reference strain P.s. pv. syringae 2905, together with cucumber strains BG913, WW17/01, WW27/01, and WW254/02 cause numerous pinpoint, chlorotic or lightly-coloured necrotic lesions on cucumber leaves. These symptoms were most intensive for BG913 and less intensive in the case of WW17/01 and WW27/01 (Fig. 2d). Interestingly, partly similar symptoms were caused by pathovar lachrymans strains LMG5070 and BG966.

Strains CCB37/09, PD2021, and WK1/02 caused large, dry, lightly-coloured necrotic lesions without chlorosis on cucumber leaves (Fig. 2e). Strain WK1/02 has been previously classified as more distant from the 2905 strain than the other pathovar syringae strains (Olczak-Woltman et al. 2007). The other two strains (CCM2858 and WW4/02) genetically similar 100 % to strains of pathovar syringae deposited at the JGI database, differ dramatically from the others in terms of the symptom type on cucumber. At the one end of the spectrum, strain CCM2858 was one of the most virulent in the collection and, at the other end, strain WW4/02 displayed nearly no virulence. Finally, there are three strains in our collection that caused either very week or no specific symptoms on cucumber, caused no HR on tobacco, and were classified into P.fluorescens, as they showed high genetic similarity to sequences of this species deposited in the JGI database.

The MLST loci used in this study were used previously as phylogenetic markers. They appear to provide a sound method for classification of Pseudomonas strains isolated from cucumber. The gyrB-based phylogenetic tree for the local isolates was topologically almost identical to the tree based on the rep-PCR fingerprinting, while the phylogenetic tree based on the rpoB and rpoD gene sequences revealed clearly different patterns of variation (Sarris et al. 2012). Our results indicate that although the gyrB is indeed one of the best markers for P.syringae pv. lachrymans, each method produces a somewhat different phylogenetic tree (Figs. 5 and 6) and perhaps the most reliable tree is obtained when several phylogenetic trees are combined into one (Fig. 7). Based on the ERIC and BOX PCR amplifications, strains belonging to P.s. pv. syringae appeared as the most heterogeneous and intermingled in relation to host plant, year and place of isolation (Weingart and Völksch 1997; Scortichini et al. 2003; Kalużna et al. 2010). We conclude that the best reflection of the strains biological properties, i.e. pathogenicity and virulence, are displayed when applying the MLST method.

There was some confusion and unexpected results when we analyzed the plasmid content. There are publications that suggest pathogenicity islands are plasmid-borne (Jackson et al. 1999; Gibbon et al. 1999; Ma et al. 2007). However, in this study some highly virulent strains had no plasmid, and some non-virulent strains had a plasmid (Fig. 3). This leads to a conclusion that in the case of Pseudomonas syringae pv. lachrymans, additional analysis should be performed to better describe the plasmid/virulence relationship.

Overall, the results of this study indicate that pathogenicity tests combined with molecular analysis are accurate in differentiation and classification of plant pathogenic bacteria. We demonstrated that P. syringae pv. lachrymans is genetically diverse, which was exhibited by both pathogenicity and molecular tests. In case of some strains isolated from cucumber it is difficult to say which pathovar they belong to, because the sequences of various loci exhibit similarity to genomes of different strains. In general, the obtained results confirm the genome architecture diversity and dynamics of P.syringae shown by Baltrus et al. (2011). There is a great need to sequence more genomes of different pathovars of the plant pathogen Pseudomonas syringae to further study its complex composition.