Introduction

Pea (Pisum sativum L.) seeds are used in a wide variety of forms. Approximately one half of the world’s production of dry peas is used to feed livestock while the remaining half is used for human consumption, mainly in developing countries. Dry mature seeds can also be processed to produce starch and protein concentrates. Pea has been the most widely produced grain legume in Spain since 2005 and its production for non-human consumption is increasing. Castilla y León, located on the North Central Spanish Meseta, with an average elevation of 750 m above sea level, is the Spanish Region with the highest pea production, with 92,380 ha devoted to this crop in 2009. The relatively cool climate in this Region requires the use of frost-tolerant cultivars and autumn sowing to obtain a satisfactory yield (Caminero 2002). In Castilla y León the pea crop is only profitable when planted in autumn and frost-tolerant cultivars are used. Early sowing allows for optimal use of water resources and avoids the drought period in the spring-summer season (Caminero 2002).

Early sowing and frost injury can increase the incidence of diseases in pea, in particular those caused by bacteria (Taylor et al. 1989; Hollaway and Bretag 1995; Reeves et al. 1996; Hollaway et al. 2007). The prevalence of these diseases represents a major constraint to the use of this crop (Carrouée 1997; Hollaway et al. 2007). Bradbury (1986) described several bacterial genera as potential agents of pea diseases, namely: Agrobacterium, Bacillus, Erwinia, Pseudomonas, and Xanthomonas. Lawyer and Chun (2001) pointed out that the world’s most important bacterial pea disease is bacterial blight caused by Pseudomonas syringae pv. pisi Sackett (Ppi) (syn. P. pisi Sackett), although sporadically it can be caused by Pseudomonas syringae pv. syringae van Hall (Psy), at times with similar symptomatology making it difficult to distinguish between the two agents by visual inspection. Moreover, because the biochemical profiles of these pathovars are very similar, identification based exclusively on biochemical characteristics is uncertain. Likewise, other Pseudomonas species, such as Pseudomonas viridiflava, Pseudomonas fluorescens and some soft-rotting Pseudomonas, have been isolated from pea plants with bacterial blight symptoms. However, among these latter species, pathogenicity on pea has been proven only for P. viridiflava in New Zealand, France and Spain (Taylor and Dye 1972; Grondeau et al. 1992; Martín-Sanz et al. 2010).

Bacterial blight of pea caused by Pseudomonas syringae pv. pisi (Ppi) was first identified in 1915 in Colorado, USA (Sackett 1916), but it was subsequently found in all countries in which pea production is important (Lawyer and Chun 2001). Seven different pathogenic races have been described on the basis of the differential response after inoculation in eight pea genotypes (Bevan et al. 1995). The frequency of each race varies between regions, the predominant pea cultivars cultivated in each region is probably one of the main factors determining these differences; either way race 2 seems to be the most frequent race worldwide (Lawyer and Chun 2001). Bacterial brown spot of pea caused by Pseudomonas syringae pv. syringae (Psy) was first described in 1966 in Wisconsin, USA (Hoitink et al. 1967). The information on this species as a pea pathogen is still limited but it appears to cause significantly lower yield losses than Ppi and the incidence of this disease is associated with early sowing (Lawyer and Chun 2001). The total yield loss caused by these two pathovars is considered to be one of the limiting factors in pea cropping (Hollaway et al. 2007), with disease incidence increasing as pea production becomes greater and the practice of early sowning becomes more frequent.

Both pathovars, Psy and Ppi, are primarily transmitted by infected seeds, although Psy has a wide host range with a possible resident phase in crop or weed species (Lawyer and Chun 2001). Cold and wet weather conditions favour the spread of the disease since these bacteria mainly penetrate the host plant through lesions which can originate from frost injury. Furthermore, these bacteria can act as ice nucleation particles under cold conditions increasing frost injury effects and lesions (Hirano and Upper 2000). For instance, in France outbreaks of bacterial blight of pea have been generally described after frost or mechanical injury, while without these factors no disease was observed even though contaminated seeds were sown in the field (Samson et al. 1997). The symptomatology caused by both species, Ppi and Psy, in fields is very similar. Frequently, after frost periods, elliptical water-soaked areas, which become olive-green and finally a brown area, are observed in aerial plant parts. These lesions often encircle the stem and may extend several centimetres and infect both stipules and leaflets. Severe infection of the stem can cause plant death (Hollaway et al. 2007).

According to López and Montesinos (1996) Ppi and Psy were recently described as pea pathogens in Spain. Pseudomonas syringae pv. pisi was first found as a pea pathogen in Spain in 1991 in experimental fields (Valladolid, Castilla y León) (Elvira-Recuenco and Taylor 2001), but only since 2002 has this bacteria been repeatedly but locally described by the Crop Protection Service in Castilla y León (J.L. Palomo, unpublished). Preliminary data indicate that the prevalence of bacterial diseases in pea has increased as the size of the pea crop increased. Considering the steady increase in the area devoted to early pea crops in Castilla y León, it is expected that the prevalence of these bacterial diseases will increase, eventually becoming a serious threat and a limiting factor for crop production.

The aims of this research were to identify the main agents of bacterial disease in pea in North-Central Spain (the highest pea-producing area in this country) and surrounding areas, to identify the variability in pathogenicity of Psy, to determine the most frequent pathogenic races for Ppi, and to find possible sources of resistance to these pathogens in high yielding and adapted genotypes.

Materials and methods

Sampling

Samples were collected from 2004 to 2008 in pea fields in Castilla y León and some adjacent Regions, namely, Navarre, the Basque Country, and the Province of Madrid. Together they represent 70% of the area planted with peas in Spain, and autumn sowing is predominant. Samples were collected from pea plants with symptoms of bacterial disease (water-soaked or brown areas on stem and leaves sometimes associated with wilt). Nearly all pea cultivars collected corresponded to dry pea used as livestock feed. Infected tissue from stems and leaves were removed, stored individually in sterile plastic bags maintained at 4°C, and processed within 24 h. Samples were surfaced sterilized with 0.5% NaOCl for 1 min, and then rinsed several times in sterilized distilled water. Plant tissues were macerated, spread on King’s B agar medium (KB) (King et al. 1954) and incubated for 48 h at 24°C. Pure cultures of bacteria were obtained from all the samples by selecting single colonies. Isolates were stored in 30% glycerol at −70°C.

Biochemical identification

Isolates were identified following the methodology described by Braun-Kiewnick and Sands (2001): Gram test (KOH), Hugh–Leifson oxidation(O)/fermentation(F) test, fluorescence on KB medium, LOPAT test (Levan production, oxidase reaction, potato soft rot, arginine dihydrosilase activity, and tobacco hypersensitivity), and the use of homoserine as carbon source. The following strains were used as controls: Pseudomonas syringae pv. pisi race 4 HRI-W 895A (Horticultural Research International, Wellesbourne, U. K.) as positive control for the fluorescence, homoserine, levan and tobacco tests; Pseudomonas cichorii CRD*99/378 (Centro Regional de Diagnóstico, Salamanca, Spain) as positive control for the oxidase activity; Pseudomonas fluorescens CECT 378 (Colección Española de Cultivos Tipo, Valencia, Spain) as positive control for arginine dihydrosilase and potato soft rot; and Erwinia rhapontici NCPPB 3766 (National Collection of Plant Pathogenic Bacteria, York, U.K.) as positive control for the Hugh-Leifson test. Sterilized water was used as negative control for all tests. Duplicate tubes or plates were run for each test, and tests were repeated at least twice.

Molecular identification

A single colony from each isolate was used for identification. Bacteria were collected on the end of a micropipette tip and resuspended in TE buffer (10 mM Tris-HCl, pH 8.0, 1 mM EDTA) in polymerase reaction tubes. Polymerase chain reactions (PCR) were carried out using a primer set specific for the AN3 and AN7 markers of Pseudomonas syringae pv. pisi (Arnold et al. 1996) (Table 1) in a thermal cycler (Perkin Elmer 9600) using the following procedure: initial strand separation at 95°C (15 min), 40 cycles of 72°C (1 min), 94°C (30 s), 60°C (1 min), final extension at 72°C (5 min). The amplifications were carried out in a volume of 25 μl containing 10 ρM of each primer, 1X Buffer II AmpliTaq Gold (Applied Biosystems), 0.2 mM of each dNTP, 1.5 mM of MgCl2 and 1U AmpliTaq Gold polymerase.

Table 1 Primers used to identify Pseudomonas syringae pv. pisi and P. syringae pv. syringae
Full size table

Two primers which amplify a sequence of 752 bp of the syrB gene (Sorensen et al. 1998) were used to identify Pseudomonas syringae pv. syringae (Table 1). This primer pair amplifies the syrB gene fragment in the pathovars atrofaciens, aptata and syringae of P. syringae, but do not in Pseudomonas syringae pv. pisi (Sorensen et al. 1998). Pathovars atrofaciens and aptata have not been described as pea pathogens (Bradbury 1986). The amplification procedure was similar to that described above except that PCR conditions were: initial strand separation at 95°C (15 min), 35 cycles of 72°C (3 min), 94°C (30 s), 60°C (1.5 min), final extension at 72°C (10 min).

Each amplification reaction was carried out at least twice. PCR products were electrophoresed in 1.5% TBE 1X agarose gels and detected by staining with ethidium bromide (0.2 ŋg/ml). The size of the PCR products was estimated using a 100 bp fragment ladder (Biotools). In all gels, several controls were included. For the AN3 and AN7 markers the reference strains of the seven Ppi races (Bevan et al. 1995), and for the gene syrB the CFBP 1769 strain of Psy (isolated on pea) (Collection Française de Bactéries Phytopathogènes, Angers, France), were used as positive control, respectively. The product of a reaction tube without bacterial template was used as negative control.

Pathogenicity test

All isolates and controls were inoculated in stems and leaves of the pea cultivar Gracia following the procedure described by Elvira-Recuenco et al. (2003). Gracia is one of the most extensively grown feed pea cultivar in Spain and it is susceptible to all races of Ppi and to Psy (Martín-Sanz 2008). The isolates which produced negative results on Gracia were inoculated on the pea cv. Kelvendon Wonder, which is also susceptible to all Ppi races (Bevan et al. 1995). Furthermore, in order to confirm pathogenesis, the isolates were inoculated in the genotype from which they were collected, when known. Two replicates of five plants each were used for each inoculation. From one of the plants which showed disease symptoms after these inoculations a new isolation was made to confirm Koch’s postulates. The identity of the new isolates was tested by the Gram reaction, glucose metabolism in the Hugh-Leifson medium, fluorescence on KB, LOPAT test, homoserine utilization as carbon source, and the molecular tests described in the previous section.

Furthermore, the isolates were inoculated into immature lemons since Psy is pathogenic in this fruit but Ppi is not (Mazarei and Kerr 1990). Lemons were surface sterilized with ethanol 70%, rinsed in distilled sterilized water, and then sterilized entomological needles were used to inoculate each isolate by depositing 20 μl of bacterial culture (108 ufc/ml) on puctures made at ten sites per isolate. Inoculated fruits were kept in a humid chamber at 25°C for 7 days, and the infection response was recorded daily.

Pathogenic variability

All the isolates that were pathogenic in the test described above were tested further on the eight pea cultivars listed in Table 2 to identify the Ppi pathogenic races (Bevan et al. 1995) and also on the P. abyssinicum accession JI2202, described as resistant to Ppi by Elvira-Recuenco et al. (2003). The methodology described by Elvira-Recuenco et al. (2003) was used with minor modifications. Ten seedlings per pot (containing a sterilized 2:1 mixture of peat and sand) were grown to the three-leaf stage at 20°C and 14 h light. Then, they were transferred to a climatic chamber and grown at 22/18°C (day/night) with 16 h light. Humidity was maintained at 100% until 24 h after inoculation, then humidity was progressively decreased to 70%. The inoculations were carried out on stems by stabbing the main stem–stipule junction of the youngest node with an entomological mounting pin containing bacteria scraped from the surface of an agar plate. Disease symptoms were scored 10 days after inoculation. In all assays, negative controls of sterilized water, and positive controls inoculated with Ppi type races and the Psy CFBP 1769 strain were included.

Table 2 Pea cultivars used to identify P. syringae pv. pisi races, and pea resistance genes according to Bevan et al. (1995)
Full size table

Resistance to Ppi and Psy among pea cultivars

Forty-nine pea cultivars which showed bacterial blight symptoms in the field during the collection periods, other feed pea cultivars widely used in Spain, and high-yielding breeding lines adapted to Spanish agro-ecological conditions were evaluated under controlled conditions for their resistance to Ppi and Psy in stems. The methodology used was the one described in the previous section, and previously used in other studies to evaluate resistance to pea bacterial blight (Taylor et al. 1989; Bevan et al. 1995; Hollaway and Bretag 1995; Elvira-Recuenco and Taylor 2001; Elvira-Recuenco et al. 2003). A total of ten seedlings per cultivar or breeding line were evaluated per strain in two replicates. The Ppi races used are listed in Table 3. In addition the highly virulent isolate CFBP 1769 of Psy (Martín-Sanz 2008) was also used. The evaluation was carried out as follows. For P. syringae pv. pisi (Elvira-Recuenco et al. 2003) a susceptible reaction was a clear water-soaked area surrounding the inoculation point; a resistant reaction was a localized necrotic area surrounding the inoculation point (hypersensitive response), and partial resistance was a localized necrotic area within water-soaked areas. For P. syringae pv. syringae (Martín-Sanz 2008) a highly resistant reaction (0) consisted of small localized necrosis in the infection point, a resistant reaction (1) consisted of necrosis spots less than 1 cm in diameter, a moderately resistant reaction (2) showed necrotic lesions at the inoculation point between 1 and 2 cm in diameter, a susceptible reaction (3) consisted of dark brown necrotic depressed lesions more than 2 cm long and reaching the next node in the apical direction, and a highly susceptible reaction (4) was similar to reaction type 3 but lesions were more than 3 cm long with a clear extension into the apical part of the plant.

Table 3 P. syringae pv. pisi and P. syringae pv. syrinage strains used in the inoculation experiments
Full size table

Results

Identification of bacterial isolates

A total of 298 isolates were obtained from field-grown pea plants with disease symptoms. Most isolates (276) were collected in the Castilla y León Region, while the remaining samples were collected in the surrounding regions of Navarre, Madrid and the Basque Country. The plant material collected corresponded to cultivars, landraces and breeding lines that represented most of the pea feed cultivars grown in the main pea producing areas in Spain.

The isolates were preliminarily characterized by their biochemical-physiological responses. Most of them (225, 77.5%) showed characteristics which fitted the expected pattern for P. syringae (Table 4). Among them, 201 isolates (67.45%) were capable of metabolizing homoserine, and thus, according to Braun-Kiewnick and Sands (2001) were preliminarily identified as P. syringae pv. pisi. The homoserine negative isolates were preliminarily identified as P. syringae pv. syringae. The remaining isolates did not fit the expected response patterns of P. syringae pathovars, nor were pathogenic on susceptible pea tester cultivars, except for three isolates which were further identified as P. viridiflava (Martín-Sanz et al. 2010).

Table 4 Reaction of Pseudomonas syringae isolates in biochemical tests and preliminary identification
Full size table

All isolates were tested in PCR to amplify DNA sequences specific to Ppi or Psy. Among the 225 isolates classified as P. syringae (either pv. pisi or pv. syringae), 222 isolates had a single amplifiable marker band; amplified DNA of 110 isolates was characteristic of the AN3 (132 bp) or AN7 (272 bp) Ppi marker. All these isolates were homoserine positive. The 99 isolates with the amplifiable AN7 marker also showed fluorescence under UV light in KB medium, while the 11 with the amplifiable AN3 marker did not show any fluorescence. From 112 out of the 225 isolates, a 752 bp band corresponding to the syrB gene absent in Pseudomonas syringae pv. pisi could be amplified by PCR (Sorensen et al. 1998). Among these latter isolates, 91 were homoserine positive and 21 homoserine negative (Table 5). From the three remaining isolates initially classified as P. syringae and all the isolates with other biochemical profiles, and identified as other species than P. syringae, none of the three marker bands could be amplified by PCR. Some examples of AN3, AN7 and syrB amplifications are shown in Fig. 1.

Table 5 Summary of the identification of the 225 Pseudomonas syringae isolates
Full size table
Fig. 1
figure 1

DNA fragments from P. syringae isolates and collection strains. 1, size marker 100 bp ladder de (Biotools); 2–6, fragments amplified with the AN7 primers; 7–11, fragments amplified with AN3 primers; 12–20, fragments amplified with syrB primers. In the following list the isolates are preceded by the letter P and the race in which they were included is indicated between parentheses. P. syringae pv. pisi, 2, P134 (2); 3, P123 (6); 4, P119 (4); 5 CFBP4766 (6); 6, CFBP4485 (2); 7, P68 (4); 8 P89 (5); 9, P214 (4); 10, CFBP1688 (1); 11, CFBP4768 (7); P. syringae pv. syringae, 12, P59; 13, P94; 14, P125; 15, P141; 16, P262; 17, SITA960; 18, CFBP1392; 19, NCPPB3505; 20, NCPPB3509

Full size image

Pathogenicity tests

All the 222 isolates from which any of the marker bands could be amplified by PCR also showed disease symptoms when inoculated in the cultivar Gracia. The isolates from which either the AN3 or AN7 marker (110) could be amplified, also generated the typical water-soaked lesions similar to the ones produced by the Ppi control race strains (Fig. 2). The isolates from which the syrB gene marker (112) could be amplified, generated brown necrotic lesions in the stems and leaves associated with apical death, similar to the symptoms generated by the Psy control strain (Fig. 3). This agreement in pea pathogenicity and molecular characteristics attributed to Psy led us to consider all the isolates which amplified the syrB gene marker as Pseudomonas syringae pv. syringae, irrespective of their response to homoserine. Furthermore, all the isolates from which the syrB gene marker was amplified, generated the same necrotic spots in immature lemons as the Psy control strain, while the isolates from which this marker could not be amplified, were not pathogenic in lemon as were also the Ppi control race strains. The remaining 3 isolates were non-pathogenic on either Gracia or Kelvendon Wonder and were not included in further tests. The isolates which were pathogenic in Gracia were tested in the pea material (cultivar, landrace, breeding line) from which they were isolated, and they were able to develop disease symptoms in all cases. New bacterial samples were isolated from these infected genotypes, and all these re-isolates showed the same biochemical and molecular characteristics as the corresponding original isolate, thus fulfilling Koch’s postulates.

Fig. 2
figure 2

Susceptible response of Kelvendon Wonder to P. syringae pv. pisi showing a typical water-soaked area

Full size image
Fig. 3
figure 3

Apical death caused by P. syringae pv. syringae in the pea cultivar Fortune

Full size image

The isolates which induced blight symptoms (water-soaked areas) in the cultivar Gracia were inoculated into the set of pea cultivars used to identify the Ppi pathogenic races (Bevan et al. 1995). This test allowed for the attribution of 105 isolates out of 110 to a specific Ppi race: 59.1% (65) isolates were identified as race 4, 20.0% (22) as race 2, 11.8% (13) as race 6, 3.6% (4) as race 5, and 0.9% (1) as race 3. The other five isolates (4.6%) showed a response pattern not previously described; they were incapable of developing disease in Fortune, Sleaford Triumph and the accession JI2202 of Pisum abyssinicum and capable in all other cultivars. According to this new response pattern, these five isolates should be included in a new pathogenic race of P. syringae pv. pisi, which following the serial numeration would be race 8. Table 5 summarizes the identification results.

The isolates from which the syrB marker of Psy could be amplified and which generated brown necrotic spots in the cultivar Gracia developed disease symptoms in all tester pea cultivars and in the P. abyssinicum accession, except in Kelvendon Wonder. Approximately in 90% of the inoculated susceptible plants the disease developed during the experiment time to the level of stem apical death, irrespective of the pea genotype tested. Thus, there were no appreciable differences in the pathogenic response to all these isolates and pea tester materials. In Kelvendon Wonder the lesions were restricted to the inoculation point; thus, this cultivar can be considered as resistant to Psy.

Response of pea materials to infection by Ppi and Psy

Table 6 shows the results obtained after the inoculation of 49 cultivars and breeding lines of pea with the seven races of Ppi and the strain CFBP 1769 of Psy. With regard to Ppi, none of the pea materials showed resistance to all races, although three cultivars (Cherokee, Corallo, Lincoln) were resistant to six of the races (no resistance to race 6 was found), and five cultivars were susceptible to all races (10.2%). The resistance-susceptibility response to the races was used to deduce the presence of specific resistance genes according to the gene-for-gene relationships described by Bevan et al. (1995) for pea-Ppi. Almost half of the pea genotypes tested (47%) were resistant to race 1, 3 and 7 and they had the R3 gene. Overall 30.6% of the genotypes had the combination R2+R3 (resistance to races 1, 2, 3, 5 and 7), 2% the R2+R4 (resistance to races 1, 2, 4, 5 and 7), 4.1% the R3+R4 (resistance to races 1, 3, 4, 5 and 7), and 6.1% the R2+R3+R4 (resistance to races 1, 2, 3, 4, 5 and 7). Thus, all resistant genotypes had the R3 gene. In relation to Psy, the most frequent response was that of moderate resistance (level 2) shown by 22 genotypes (44.9%); another 22 genotypes were susceptible or highly susceptible (levels 3 and 4), while only five genotypes (cultivars Messire, Attika, Cherokee, Iceberg and Isard) were highly resistant or resistant (levels 0 and 1, respectively).

Table 6 Response of pea cultivars to inoculation with reference strains of P. syringae pv. pisi (Ppi) and P. syringae pv. syringae (Psy), and bacteria isolated from field-grown pea samples
Full size table

Discussion

One of the aims of the research was to identify the causal agents of the bacterial diseases observed in pea fields in Castilla y León and surrounding areas in Spain. Most of the 298 bacterial isolates collected from pea plants were pathogenic when inoculated in pea tester varieties susceptible to P. syringae. Among the 222 pathogenic isolates 50.5% were identified as Psy and 49.5% as Ppi. The occurrence of Psy as causal agent of pea diseases has been considered less important than the occurrence of Ppi, and Psy infections seemed to be restricted to autumn-winter sowing of peas (Lawyer and Chun 2001). In most of the studies to identify the causal agents of bacterial diseases, Psy was not mentioned (Taylor et al. 1989; Hollaway and Bretag 1995; Reeves et al. 1996; Cirvilleri et al. 1998). However, in the early1970’s Psy had already been described as the main disease agent for some years in a few early-sown pea fields in New Zealand and in the USA (Butler and Fenwick 1970; Taylor and Dye 1972). Recently, both Ppi and Psy have been found in early-sown pea fields in New South Wales and Victoria, Australia (Hollaway et al. 2007). Ppi and Psy have seldom been described as pea pathogens in Spain and Ppi was first found as a pea pathogen in 1991 (Elvira-Recuenco and Taylor 2001; López and Montesinos 1996). According to data of the Crop Protection Service the incidence of bacterial diseases has dramatically increased in North-Central Spain, especially in the Castilla y León Region and surrounding areas, associated with the increase of both crop acreage and the practice of early sowing (J.L. Palomo, unpublished).

Thus, P. syringae is clearly the predominant agent of bacterial disease on pea plants associated with an early sowing date in late autumn-early winter in Castilla y León, with both pv. syringae and pv. pisi being almost equally present. There are no previous reports in which Psy had been found in such high proportion among isolates from pea fields. These results point to the fact that, at least under some circumstances or some regions, Psy can be as important as Ppi in inciting pea diseases. This information is important because the breeding efforts for resistance to bacterial diseases in most countries have focused solely on Ppi. Thus, the knowledge on resistance to Psy and its heredity in pea is scarce.

Molecular markers have been used to identify P. syringae pathovars and races pathogenic on legume and other plant species (Arnold et al. 1996; Sorensen et al. 1998; González et al. 2003; Scortichini et al. 2003) and the advantages and limitations of molecular characterization in identification of plant pathogenic bacteria have been reviewed (Alvarez 2004). The identification of Psy and Ppi was first carried out on the basis of biochemical characteristics. A classical way to differentiate these two pathovars was the use of homoserine as C source, described as positive for Ppi and negative for Psy (Hildebrand 1972). However, further data indicated that between 15% and 25% of the isolates identified as Psy were homoserine positive (Mazarei and Kerr 1990; Grondeau et al. 1992; Braun-Kiewnick and Sands 2001). The percentage of homoserine positive Psy isolates found in this study was much greater, 81.25%. Thus, this biochemical characteristic is not a useful criterion to distinguish between Psy and Ppi. On the other hand, all isolates finally classified as Psy had the PCR amplifiable syrB marker and generated disease symptoms (brown necrotic areas) different from Ppi symptoms (water-soaked areas) under growth chamber conditions. The genes syrB and syrD, which are related to the production and transport of bacterial toxins and found in a single copy in the Psy genome, are found in most of the Psy isolates from different plant species (Quigley and Gross 1994; Sorensen et al. 1998; Scortichini et al. 2003). Our results support that at least syrB is a good marker for identifying Psy, although there are some Psy pathogenic isolates that do not amplify syrB (Scortichini et al. 2003). In order to ascertain whether the high frequency of homoserine positive Psy isolates and the high frequency of Psy as a causal agent of pea disease in North-Central Spain are associated, further tests will be necessary.

Arnold et al. (1996) described the AN3 and AN7 markers obtained from RAPD markers as specific for Ppi. Among our isolates, the AN3 marker (132 pb, see Table 1) was amplified from races 3, 4, and 5, while AN7 (272 pb) was amplified from races 2, 4, 6, and the putative new race 8, but from none of the Psy isolates or from any reference strain of other Pseudomonas species. As Arnold et al. (1996) have shown, there is no correlation between the DNA marker amplified and race or fluorescence phenotype. Thus, for instance, in the Arnold et al. (1996) study and in our study Ppi race 4 isolates were fluorescent or non-fluorescent and amplified either AN3 or AN7.

Pea tester varieties were used to identify the different Ppi races among the isolates. The predominant race was race 4 which represented 59.1% of the Ppi isolates, followed by race 2 (20.0%), race 6 (11.8%), race 5 (3.6%) and race 3 (0.9%). Five isolates showed a response pattern which was not previously described suggesting that an additional race 8 could be present in Ppi. These five isolates amplified AN7 and showed fluorescence in KB. Previous data in Spain are limited to the description of eight isolates classified as race 4 and another eight to race 6 (Elvira-Recuenco and Taylor 2001). Taylor et al. (1989) found that race 2 was predominant (74.6%) in a world collection of 146 isolates, followed by race 4 (10.9%), race 6 (7.5%), and race 1 (5.48 %), while races 3 and 5 were each represented by a single isolate. Similar results were found among isolates from pea seeds in Great Britain (Reeves et al. 1996). In Sicily (Italy) the only race found was race 6 (Cirvilleri et al. 1998). In Australia race 3 was found to be predominant (64%) among 65 isolates (race 6, 31%; race 2, 5%) and these results were most likely due to the fact that the predominant pea cultivars were susceptible to races 3 and 6 (Hollaway and Bretag 1995). Our results could be due to a similar response since a minority (12.2%) of pea cultivars and lines found in Castilla y León are resistant to race 4 and none to race 6 (Table 6). The frequency distribution of Ppi pathogen races has been correlated to the presence of resistance genes in the most frequently cultivated pea cultivars in a particular region or county (Hollaway et al. 2007). Our results in North-Central Spain partially agreed with this hypothesis, for instance, the low frequency of Ppi race 3 agrees with the presence of R3 gene in the majority of the pea cultivars grown. However, the relatively low frequency of race 6 contrasts with the absence of resistance genes to this race in the cultivated materials. Furthermore, race 6 was present in pea materials in Spain before 2001 (Elvira-Recuenco and Taylor 2001). On the other hand, race 4 has been described as highly virulent (Elvira-Recuenco et al. 2003) which agrees with the high frequency of this race among our isolates and with the fact that this pathogen was isolated even from two cultivars (Cherokee and Corallo) with the corresponding resistance gene (Table 6). This result suggests that this resistance gene does not confer complete resistance under field conditions. In fact, in experimental field plots inoculated with race 4, the most common response observed in pea genotypes resistant to this race was partial resistance (Martin-Sanz, unpublished).

The isolates identified as Psy were highly pathogenic in all pea tester varieties with the exception of Kelvendon Wonder. Approximately 90% of the pea plants showed apical death at the end of the observation time, while in Kelvendon Wonder the necrosis was localized near the infection point. Thus this cultivar can be considered as resistant to Psy. No previous description of resistance in pea to this pathovar has been found, except in two cultivars for human consumption, namely “Elf” and “Champion from England” (Lawyer and Chun 2001)

The search for resistance among the pea materials cultivated in Spain and potentially useful in breeding programs was carried out by inoculating these pea cultivars with type race strains of Ppi and with Psy. Only five out of 49 pea genotypes showed a satisfactory level of resistance (levels 0 and 1) against Psy in growth chamber experiments. Isard was found to be the most promising source of resistance since this cultivar showed resistance under both growth chamber and field conditions, and for two consecutive years in natural infection field trials. The other resistant cultivars showed some disease symptoms under more variable environmental field conditions, thus their resistance can only considered to be incomplete. The search for resistance sources for Ppi races revealed that the cultivars Cherokee, Corallo or Lincoln are suitable materials since they carry resistance genes (R2+R3+R4) which give resistance to all Ppi races including the new race 8 (data not shown), except to race 6. The inheritance of these genes has been described as dominant monogenic (Bevan et al. 1995; Hunter et al. 2001) which eases the breeding process. Pisum abyssinicum is also a very promising source of race nonspecific resistance (Elvira-Recuenco et al. 2003) including race 6 and also additional resistance to the aggressive race 4, although the type of inheritance of this resistance from P. abyssinicum is not fully known, and its potential resistance has not yet been tested under field conditions in Spain.

Due to the steady increase in pea production in Spain and the increase in early-sowing in the main production areas like the Castilla y León Region, it is very likely that the incidence of bacterial disease will increase. In fact, in the last two seasons of this study the effect of bacterial pathogens in pea was clearly the most important constraint of the crop. Resistant cultivars could play an important role in reducing the losses caused by these pathogens and for this reason crossings to include them in the breeding program of the ITACyL have already started.