Short communication. Detection by multiplex PCR and characterization of nontoxigenic strains of Pseudomonas syringae pv. phaseolicola from different places in Spain

The efficient control of halo blight, caused by Pseudomonas syringae pv. phaseolicola, is primarily based on the use of pathogen-free seed. Detection of the pathogen in seeds is currently carried out with high-sensitive methods based on the detection by PCR of genes involved in the biosynthesis of phaseolotoxin, which was believed to be produced by all strains of the pathogen with epidemiological importance. However, field epidemics of halo blight in the county of Castilla y León, Spain, are often associated to nontoxigenic isolates of P. syringae pv. phaseolicola, which cannot be detected using current molecular and serological methods. The results presented in this work show the existence of nontoxigenic isolates of P. syringae pv. phaseolicola in areas other than Castilla y León, indicating the need to establish a reliable methodology for seed certification. A simple two-step methodology is presented with the aim to identify both types of isolates that is based on a multiplex enrichment PCR of seed soakates and on pathogenicity assays. Additional key words: BIO-PCR, halo blight, phaseolotoxin, Pseudomonas syringae pv. syringae, races.

inoculum, (ii) the lack of crop rotations and (iii) the ineff iciency of the chemical control (C.Asensio, personal communication).
The most relevant bacterioses of bean are those caused by the Gram negative bacteria Pseudomonas syringae pv.phaseolicola, P. syringae pv.syringae and Xanthomonas campestris pv.phaseoli, which result in important economical losses worldwide (Smith et al., 1988;Saettler, 1991).P. syringae pv.phaseolicola, which causes halo blight of bean, is a seed-borne pathogen present in all countries were beans are cultivated, and is responsible for serious crop losses, especially in temperate areas of the world, such as Spain.Contaminated seeds, either internally or externally, constitute the primary source of infection (Schwartz, 1989) and a very low level of inoculum can result in severe epidemics under favourable environmental conditions (Trigalet and Bidaud, 1978;Webster et al., 1983).Control of the disease is difficult and its efficient integrated control requires the use of seed certified free of the pathogen.Methods used for detection of this pathogen, however, have to be highly sensitive because a single contaminated seed among 2,000 or 5,000 can initiate an epidemic (Trigalet and Bidaud, 1978;Webster et al., 1983).
Several highly sensitive PCR-based methods have been designed for the detection of P. syringae pv.phaseolicola, and all of them rely on the detection of DNA sequences involved in the biosynthesis of phaseolotoxin (Tourte and Manceau, 1991;De la Fuente-Martínez et al., 1992;Prosen et al., 1993;Schaad et al., 1995;Audy et al., 1996).It is generally accepted that the genes for the biosynthesis of phaseolotoxin are present in all the strains of P. syringae pv.phaseolicola found causing field epidemics (Rudolph, 1995;Schaad et al., 1995).Additionally, this DNA appears to be highly specific because it has not been found in other bacterial bean pathogens, although it was also described in strains of P. syringae pv.actinidiae (Tamura et al., 1989(Tamura et al., , 2002)), which infects kiwi (Actinidia deliciosa Liang et Ferguson), and in a single strain of P. syringae pv.syringae isolated from vetch (Vicia sativa L.) (Tourte and Manceau, 1995).However, it was recently showed that more than 66% of the P. syringae pv.phaseolicola isolates causing field epidemics in Castilla y León, Spain, were nontoxigenic and did not contain DNA specific for the biosynthesis of phaseolotoxin (Rico et al., 2003;Oguiza et al., 2004).The strains of P. syringae pv.phaseolicola that lacked the phaseolotoxin gene ORF6 were designated as tox-, whereas those containing this gene were designated as tox+ (Rico et al., 2003).Consequently, this designation will be used from now on because it is helpful to distinguish those strains that can (tox+) or cannot (tox-) be detected by current PCR methods using primers directed to the phaseolotoxin cluster.Incidentally, all the P. syringae pv.phaseolicola tox-strains analyzed so far also lacked other genes of the phaseolotoxin biosynthesis cluster and failed to produce phaseolotoxin (Rico et al., 2003;Oguiza et al., 2004).Additionally, the tox-isolates also failed to react with one of the few commercial antibodies for the serological detection of P. syringae pv.phaseolicola (Rico et al., 2003), all of which are polyclonal.In consequence, the efficient certification of bean seed in Spain using current highly sensitive methods, such as PCR and serological methods, will depend on the possible distribution of tox-populations.The aim of this work was to examine the presence of tox-populations of P. syringae pv.phaseolicola in other bean producing areas of Spain.
In this work, the presence of Pseudomonas syringae in 22 bean seed samples obtained from experimental fields in Spain was examined.Seventeen samples were of cv.Alavesa from Álava, two samples were of cv.Blanca from Huesca, two samples of an unknown cultivar were from Galicia and one sample of cv.Jules from an unknown place was obtained from a Spanish Germplasm Bank (Plant Genetic Resource Centre, CRF-INIA, Alcalá de Henares).Seeds (15 to 100 g) were soaked overnight at 4ºC in sterile distilled water (1:1.5 w/v) and the water extract was recovered and centrifuged at 1,5550 × g for 5 min.The pellet was washed twice in 5 ml of distilled water and concentrated 20 times from the original extraction volume.Aliquots of 100 µl from this concentrated soakate were plated on Milk-Tween (MT) medium (Goszczynska and Serfontein, 1998), incubated overnight at 18ºC and used for enrichment-PCR analysis (Schaad et al., 1995).The bacterial growth from a single plate was washed with 3 ml of sterile distilled water, and 5 µl of the resulting suspension were used as template in a multiplex PCR reaction in a total volume of 25 µl which contained 20 pmol each primer, 0.15 mM each dNTP, 1X PCR reaction buffer, 1.5 mM MgCl 2 and 1 unit of Taq DNA polymerase (Biotaq; Bioline Ltd., London).Control strains and primers used are described in Tables 1 and 2. The primers used in the multiplex PCR were primer pair P5.1-P3.1 (Schaad et al., 1995;Zhang and Patil, 1997), specific of open reading frame (ORF) 6 from locus phtE of the phaseolotoxin biosynthesis cluster, and primer pair P3004L and P3004R, which amplify a 0.24 kb fragment (GenBank acc no.AJ568001) that has been observed only in nontoxigenic strains (Oguiza et al., 2004).Additionally, 100 µl aliquots of a 1/10 dilution of the concentrated seed soakate were cultured on MT plates for 24-48 h at 28ºC, and putative colonies belonging to P. syringae pvs.phaseolicola and syringae were identified by their fluorescence and the presence or absence of haloes due to the hydrolysis of casein (Goszczynska and Serfontein, 1998).The pathogenicity of the resulting isolates was tested on bean pods of cv.Canadian Wonder, which is susceptible to all known races of P. syringae pv.phaseolicola (Taylor et al., 1996), and cv.Perona as described previously (Harper et al., 1987).Only fluorescent colonies producing water-soaking or sunken-brown lesions on bean pods, characteristic of P. syringae pvs.phaseolicola and syringae, respectively, were retained.Reference strains were used for comparison (Table 1).
No colonies were observed in 17 out of the 22 seed lots examined, or the few encountered were not fluorescent.In all these cases no amplification bands were produced by multiplexed PCR, and these seed lots were not further assayed.The remaining f ive seed lots produced fluorescent colonies on MT or King's Medium B (King et al., 1954) plates, and three of them produced amplification bands upon multiplexed PCR.One of the lots, originating from Galicia, produced the specif ic 0.5 kb fragment after amplif ication with primers P5.1 and P3.1, whereas other two lots, both from cv.Alavesa, amplif ied the 0.24 kb fragment specific of P. syringae pv.phaseolicola tox-isolates (Table 3).The identification of the fluorescent colonies obtained from the diluted soakates from these five seed lots was performed essentially as described (Lelliott and Stead, 1987;Braun-Kiewnick and Sands, 2001) following the LOPAT scheme (Lelliott and Stead, 1987), with slight modifications (Cazorla, 1998;Rico et al., 2003).Degradation of casein was examined on MT plates (Goszczynska and Serfontein, 1998).The utilisation of D-mannitol, inositol, D-sorbitol, erythritol, L+tartrate, D+tartrate, L-lactate, arabinose, adonitol, glutarate and glucuronate as the sole carbon source was examined on Ayer's minimal medium (Lelliott and Stead, 1987) containing the carbohydrates to a final concentration of 0.1% (w/v).
All the fluorescent colonies obtained (20 colonies in total) were levan positive, oxidase and arginine dihydrolase negative, and elicited a hypersensitive reaction on tobacco (LOPAT group Ia); they were therefore identified as Pseudomonas syringae (Lelliott and Stead, 1987).Seven colonies from two seed lots from Álava showed a biochemical profile typical of P. syringae pv.syringae while the 13 colonies from one lot from Galicia and two lots from Álava, produced water soaked lesions on bean pods and showed the biochemical profile typical of P. syringae pv.phaseolicola (Lelliott and Stead, 1987;Braun-Kiewnick and Sands, 2001) (Table 3).These data agreed with the results of the enrichment-PCR amplifications of the three seed lots soakates with the primer pairs specific for tox+ or toxstrains of P. syringae pv.phaseolicola (see Table 3).
From the above 13 P. syringae pv.phaseolicola isolates, four isolates from the seed lot from Galicia did not utilise mannitol, produced the expected amplification fragments with primer pairs P5.1-P3.1 and PHA19-PHA95, and did not produce the 0.24 kb band with primers P3004L and P3004R (Fig. 1 and Table 3); therefore, they were classified as typical tox+ isolates.Assimilation of mannitol is usually negative for strains of P. syringae pv.phaseolicola that contain the DNA for the biosynthesis of phaseolotoxin (Marques et al., 2000), whereas the strains that do not contain this DNA can utilise mannitol as the sole carbon source (Rico et al., 2003).Additionally, primers PHA19 and PHA95 (Table 2) direct the amplification of a 0.48 kb internal fragment from the amidinotransferase gene amtA, which is also essential for phaseolotoxin biosynthesis (Marques et al., 2000;Hernández-Guzmán and Álvarez-Morales, 2001).The remaining nine isolates, 1 D-Mannitol was added as the sole carbon source to Ayer's minimal medium (Lelliott and Stead, 1987) to a final concentration of 0.1% (w/v).Those isolates that did not show significant growth after 1 week of incubation at 28ºC were considered negatives.
from two seed lots from Álava, could use mannitol as the sole carbon source, did not amplify the expected fragments with primers P5.1-P3.1 and PHA19-PHA95, and produced a band of 0.24 kb with primers P3004L and P3004R (Fig. 1 and Table 3); therefore, they were classified as typical tox-isolates.
The race assignation of the different isolates was also investigated by PCR analysis (Rico et al., 2003) using specific primers for avirulence genes avrPphB, avrPphE (Stevens et al., 1998) and avrPphF (Tsiamis et al., 2000), which were used as published.The four tox+ strains identified in this study were tentatively assigned to race 2 and 6, since they neither amplified genes avrPphF nor avrPphB.On the other hand, the nine toxstrains contained avrPphF, and could be assigned to any of the races 1, 5, 7 or 9 (Table 3).These data agree with the race structure reported for P. syringae pv.phaseolicola in Castilla y León, in that all tox-isolates contain avrPphF whereas most of the tox+ isolates lack this gene and belong to race 6 (Rico et al., 2003).
The identification of toxigenic and nontoxigenic populations of P. syringae pv.phaseolicola was greatly simplified by the analysis of seed soakates with the two-step strategy implemented in this work.The first step consisted on the analysis of soakates by a multiplex PCR with primers specif ic for tox+ and tox-P.syringae pv.phaseolicola strains.The second step entailed the plating of seed soakates onto semiselective medium; from the resulting growth, only those fluorescent colonies that produced characteristic lesions on bean pods were selected and further characterized.
Only a few of the seed lots examined yielded fluorescent bacteria, which were identified as P. syringae, which might be due to the low size of most samples (around 50 g of seeds) and/or their good sanitary condition.In all cases, the seed soakates that produced specific amplicons for tox+ or tox-P.syringae pv.phaseolicola populations also yielded characteristic tox+ or tox-isolates, respectively, suggesting the applicability of primers P3004L/R for the detection of nontoxigenic isolates.Toxigenic and nontoxigenic isolates were not found coexisting in any of the seed lots examined, although this can be due to the small sample size, because their coexistence in disease bean plants has been repeatedly observed (C.Asensio, personal communication).
Altogether, the results obtained in this work clearly indicate that tox-populations of P. syringae pv.phaseolicola can be found in different bean producing areas of Spain, besides Castilla y León.This implies that bean seed lots from Spain cannot be certified as free of this pathogen using routine highly sensitive methods, such as the PCR detection of DNA sequences for the biosynthesis of phaseolotoxin or serological methods with commercial antibodies.Infected bean seeds are the main primary source of P. syringae pv.phaseolicola in f ield epidemics; additionally, the use of seed certified free of the pathogen is the main method for the control of halo blight (Smith et al., 1988;Saettler, 1991).Therefore, the primers validated in this study can be used for the development of appropriate detection methods that can identify all the populations of the pathogen that are currently found in the country.

Figure 1 .
Figure1.Multiplexed PCR using the primer pair P5.1 and P3.1 in combination with P3004L and P3004R to identify P. syringae pv.phaseolicola tox+ and tox-isolates.PCR products were separated by electrophoresis in a 1% agarose gel.M: 1 kb DNA ladder (Promega).A8: two colonies from seed lot A8, cv.Alavesa from Álava; A18: 7 colonies from seed lot A18, cv.Alavesa from Álava; G: 5 colonies from a seed lot of an unknown cultivar from Galicia.Sizes are indicated in kb.The 0.5 kb and 0.24 kb fragments correspond to the amplicons obtained with primer pairs P5.1-P3.1 and P3004L-P3004R respectively.

Table 1 .
Bacterial strains used in this study

Table 2 .
Primers used in this study

Table 3 .
Identification of toxigenic and nontoxigenic P. s. pv.phaseolicola strains and P. s. pv.syringae strains in five seed lots