SSR marker analysis of Monilinia fructicola from Swiss apricots suggests introduction of the pathogen from neighbouring countries and the United States

Authors


E-mail: andrea.patocchi@acw.admin.ch

Abstract

Monilinia fructicola is a quarantine fungal pathogen in Europe, but many major stone fruit growing countries in Europe have reported its presence recently. In Switzerland, the fungus was first found in a single apricot orchard in 2008. This study confirms the presence of M. fructicola in nine out of 22 commercial orchards in Canton Valais, Switzerland. Five simple sequence repeat markers (SSRs) were developed for M. fructicola and samples from Switzerland, Spain, Italy, France and the United States were analysed and compared in order to assess the genetic diversity of the pathogen, identify the origin of the disease, and verify if the fungus reproduces sexually in Europe. In the 119 European samples analysed, 12 different haplotypes were found, indicating a relatively high genetic diversity of the pathogen considering that the first report in Europe was 10 years ago. Three haplotypes found in Europe were identical to those found in the American samples (two from the east coast and one from the west coast). Population structure analysis suggests that the European population is derived from at least two ‘invasion’ events probably originating from the US (one from the east coast, the other from the west coast). Preliminary evidence of sexual reproduction of M. fructicola in Europe is reported.

Introduction

Brown rot is one of the most important fungal diseases of commercially grown stone fruits worldwide (Ogawa et al., 1995). The disease is primarily caused by three Monilinia spp.: M. laxa, M. fructigena and M. fructicola. A fourth species, Monilia polystroma, was discovered recently and was confirmed to cause brown rot in pome and stone fruits in Japan and Europe (van Leeuwen et al., 2002; Petróczy & Palkovics, 2009). Monilinia spp. infect blossoms, twigs and fruits, causing blossom blight, twig canker and fruit rot, and resulting in economic losses due to crop loss.

Until recently M. laxa and M. fructigena were the major pathogens causing brown rot in Europe. However, in 2001 M. fructicola was discovered in peach orchards in the southeast (Rhône valley) of France (EPPO, 2002a). Since then, M. fructicola has been found in several other European countries, including Austria (EPPO, 2002b), Spain (de Cal et al., 2009), the Czech Republic (Duchoslavováet al., 2007), Italy (Pellegrino et al., 2009), Switzerland (Hilber-Bodmer et al., 2010) and Germany (EPPO, 2010). The pathogen was reportedly eradicated from Austria in 2006 (EPPO, 2006). Surveys performed on imported stone fruits indicate that trade can be responsible for spreading M. fructicola. Stone fruits imported into Switzerland from the US and France tested positive for M. fructicola (Bosshard et al., 2006), and peaches imported into Hungary from Spain and Italy also tested positive for M. fructicola (Petrózcy & Palkovics, 2006). Consequently, highly dispersible and abundant spores of infected, imported fruits may establish the disease in new locations (Bosshard et al., 2006).

In contrast to M. laxa and M. fructigena, M. fructicola regularly undergoes sexual recombination in certain regions like California (Michailides et al., 2007). This may allow the fungus to combine resistances to fungicides which developed independently from each other. In the US and other countries, M. fructicola isolates with resistance to methyl benzimidazole carbamates such as benomyl, thiophanate-methyl (Ma et al., 2003) and carbendazim (Lim et al., 2006), sterol demethylation inhibitors (DMIs) like propiconazole (Schnabel et al., 2004; Luo & Schnabel, 2008), and respiration inhibitor fungicides such as azoxystrobin and boscalid (Amiri et al., 2010) have already emerged.

Pathogens can spread across borders in Europe. This can occur by means of trade (as indicated above) or through natural spore dispersion in rainy and windy environments (Nagarajan & Singh, 1990). Molecular markers for fingerprinting pathogen populations can be very useful to track disease epidemics and to identify means by which the pathogen spreads or is being spread. Microsatellite, or simple sequence repeat (SSR) markers, are highly polymorphic, multi-allelic, co-dominant, and PCR-based. They are commonly used to assess the genetic diversity in a given population and to study the genetics of populations (Tenzer et al., 1999; Gobbin et al., 2003). Until now, this type of molecular marker has not been developed for any of the Monilinia spp. The few genetic diversity studies so far performed on Monilinia spp. were based on random amplified polymorphic DNA (RAPD) markers (M. laxa, Gell et al., 2007; M. fructicola, Lim et al., 2006), ITS sequences (M. laxa and M. fructicola, Snyder & Jones, 1999) or inter-simple sequence repeat (ISSR) markers (M. fructicola, Ma et al., 2003; Luo & Schnabel, 2008; Fan et al., 2010).

The objectives of this study were: (i) to survey M. fructicola in the main apricot growing area of Switzerland (Canton Valais); (ii) to develop SSR markers to characterize populations of M. fructicola from Switzerland, US, France, Italy and Spain in order to identify the possible origin of the disease in Europe and particularly in Switzerland; and (iii) to investigate if the pathogen reproduces sexually in Europe.

Materials and Methods

Monilinia fructicola reference isolates

A total of 57 isolates of M. fructicola from the US states of Georgia (24 isolates), South Carolina (32 isolates) and California (one isolate) were available for this study (Table 1). From Europe, 8, 20, 21 and 18 isolates were available from France, Spain (Ebro Valley), Italy (Piemonte region) and Switzerland (collected from Canton Valais in 2008), respectively. An additional 52 isolates of M. fructicola were obtained during this study from Canton Valais, Switzerland in 2009 (Table 1).

Table 1. Monilinia fructicola isolates used for this study
No. isolatesCountry (state)Region or countyNo. orchardsNo. samples per orchardHostYear of isolation
  1. aIsolates collected for this study.

24US (Georgia)Crawford, Hall, Macon, Peach45–7Peach2002, 2003
32US (South Carolina)Anderson, Edgefield, Saluda, York64–7 (one orchard only one sample)Peach2001, 2003
1US (California)Unknown11Apricot2001
8FranceSoutheast81Peach (6), cherry and apricot (one each)2000–2003 and 2008
20SpainEbro Valley121–3Peach2006–2008
21ItalyPiemonte29–12Peach2008
18SwitzerlandValais118Apricot2008
52aSwitzerlandValais91–6 (one orchard 31)Apricot2009

2009 sampling and single-spore isolation

Fruit and mummies showing symptoms were collected between the end of July and the end of August from 22 commercial apricot orchards in Switzerland and placed individually in plastic bags. The fruit were kept at 4°C until further processing, but in general no longer than 5 days. With the exception of three orchards, where brown rot severity was very low, no less than 20 fruit (one per tree) were collected from each orchard. Most fruit came from a single orchard (orchard J), where M. fructicola was detected in 2008 (Hilber-Bodmer et al., 2010). A total of 721 samples were collected from the fruit and mummies. A sample is defined in this study as mycelium and spores that developed on the fruit from a single infection event. Most samples came from single fruits but 44 visually distinct samples came from 22 fruits and three from one fruit. Mycelium and spores were scraped from fruit or mummies with a pre-sterilized wooden toothpick and transferred to a potato dextrose agar (PDA) Petri dish (90 mm). After 7–10 days of incubation at 20°C in darkness, colonies were transferred to fresh PDA dishes.

For single-spore isolation of M. fructicola isolates, conidia developing on PDA were harvested using a pre-sterilized wooden toothpick, resuspended in 1 mL of double distilled water (ddH2O), and vortexed. The conidia suspension was adjusted to 20–40 spores mL−1 and 250 μL were evenly distributed onto the surface of a PDA dish. After 4 days of incubation at 20°C a single colony was transferred to a fresh dish. All PDA was amended with 500 mg L−1 streptomycin.

DNA extraction for species identification

The Monilinia spp. were identified to the species level using molecular techniques. DNA was extracted from mycelium obtained from the same location on the fruit where mycelium was taken for the fungal isolation, using a wooden toothpick. DNA was extracted according to Kawasaki (1990) with minor modifications. Mycelium was transferred into a tube containing 50 μL Kawasaki-Buffer, incubated for 20 min at 80°C, and shaken at 800 rpm in an Eppendorf Thermomixer Comfort (Eppendorf). After a short spin (15 s at 16 100 g) the supernatant was used in the multiplex PCR assay according to Côtéet al. (2004). Double distilled water was used as the negative control and previously extracted DNA from M. laxa, M. fructigena and M. fructicola from pure cultures as positive controls. The total PCR reaction volume was 10 μL, containing 2 μL crude extract of Monilinia spp., 5 μL of the Multiplex PCR Master Mix (QIAGEN), 1 μL · ddH2O and 0·1 μm of each of the four primers (MO368-5, MO368-8R, MO368-10R, Laxa-R2). Amplifications were carried out in a SensoQuest labcycler (SensoQuest). The initial step of 95°C for 15 min was followed by 28 cycles of 94°C for 1 min, 60°C for 1 min and 72°C for 1 min and 1 cycle of 10 min at 72°C. PCR products were electrophoresed on 1·4% agarose gels amended with ethidium bromide and visualized under UV light.

SSR development for M. fructicola

Simple sequence repeat markers were developed following the protocol of Brunner & Frey (2004) using the repeats AG, AT, CTT and TCG starting from DNA of M. laxa and M. fructicola. DNA was extracted from 7-day-old potato dextrose broth (PDB) cultures (inoculated with conidia from pure cultures) using the DNeasy Plant Mini kit (QIAGEN) following manufacturer’s instructions. Primer pairs flanking the SSR sequence were designed with the program primer3 (Rozen & Skaletsky, 2000). The names of the SSR markers were prefixed with ‘CH’ indicating the country where the markers were developed (i.e. Switzerland), followed either by ML or MFc indicating the origin of the DNA used to develop the markers (i.e. M. laxa or M. fructicola, respectively), while the number indicates the sequenced clone.

Simple sequence repeat markers (Table 1) were multiplexed using the Multiplex PCR kit (QIAGEN). Forward primers were labelled as following: CHML5 and CHMFc5 were labelled with 6FAM; CHMFc1, CHMFc4 and CHMFc12 were labelled with NED, VIC and PET, respectively. The PCR mix (final volume 10 μL) consisted of 2 μL DNA, 5 μL Multiplex PCR Master Mix, 1 μL of Q-Solution, 1 μL sterile water and 1 μL of the primer mix (final concentrations: CHMLSSR5f/r 1·875 μm, CHMFc1f/r 5 μm, CHMFc4f/r 1·25 μm, CHMFc5f/r 1·875 μm, CHMFc12f/r 3·75 μm). PCR was performed in a thermocycler (SensoQuest) using the following cycling conditions: 15 min at 95°C for initial denaturation; 35 cycles of 40 s at 94°C, 90 s at 50°C, 90 s at 72°C; ending with 30 min at 60°C and a final hold at 10°C. Fragment analysis was performed on the ABI PRISM® 3100 DNA capillary sequencer (Applied Biosystems). A total of 0·8 μL of a 1:20 diluted PCR product was transferred to 15 μL of formamide containing 0·25 μL of the fluorescent GeneScan™-500-LIZ™ size standard (Applied Biosystems). The reactions were then denatured for 5 min at 95°C, rapidly cooled in the freezer, and loaded on the sequencer. GeneMapper™ v. 4·0 (Applied Biosystems) was used for data analysis.

Allele patterns were compared to distinguish haplotypes. First, fingerprints of samples from the US were compared to each other. The names of the different haplotypes were formed by the prefix ‘US’ followed by a number. Then the French, Italian, Spanish and Swiss samples were compared (in this order) to identify haplotypes. The names of haplotypes not identified in the samples from the US were named using a prefix indicating the country where the new haplotype was found for the first time followed by progressive numbers.

Population analysis

Bayesian analysis as implemented in the software structure v. 2·3·3 (Pritchard et al., 2000) was used to investigate the population genetic structure and individual admixture proportions. Two different datasets were used in the analyses: the first was composed of only a single representative for each haplotype, the second was composed of all different haplotypes found in each population, i.e. those haplotypes that were found in more than one population were represented more than once in this dataset. Samples from the same countries (European samples) or states (US samples) were defined as populations for the sampling location prior. Runs were performed with all combinations of the following parameters: admixture versus no admixture, prior on sampling location versus no such prior, independent versus correlated allele frequencies among populations. Five runs each for K = 1–6 with 100 000 MCMC repetitions after a burn-in period of 50 000 repetitions were performed. The true number of clusters was inferred from deltaK (Evanno et al., 2005). In addition, the deltaK results were confirmed by Calinski & Harabasz’ pseudo-F (1974) as implemented in genodive (Meirmans & van Tienderen, 2004).

The possible extent of clonality was assessed with the Index of Association (IA) using multilocus v. 1·3b (Agapow & Burt, 2001) applied to both datasets. The IA value was computed using 50 000 simulated datasets. For outcrossing populations, indices close to zero should be observed whereas, because gametic disequilibrium is expected in asexual or inbreeding populations, values with significant deviation from zero would be expected in populations without sexual recombination.

Results

Identification of Monilinia spp. from Switzerland to the species level

Species-specific PCR amplification was successfully conducted for 648 of 721 samples. Results indicated the presence of M. laxa (373 samples), M. fructigena (222 samples) and M. fructicola (43 samples) in Swiss orchards. The 10 fruit with mixed infections revealed the combinations M. fructicola and M. laxa (six times), M. fructicola and M. fructigena (twice), and all three species (once; Table 2). In total, 52 samples (8%) obtained from nine orchards tested positive for M. fructicola. Only three of 248 samples from orchard J tested positive for M. fructicola and seven samples did not yield PCR fragments.

Table 2.   Isolates of Monilinia spp. collected in Canton Valais (Switzerland) in 2009
 No. samples
Monilinia laxa373
M. fructigena222
M. fructicola43
M. fructicola and laxa6
M. fructicola and fructigena2
M. laxa and fructigena1
M. fructicola, laxa and fructigena1
Total648

SSR marker development and test of variability

Eighty-one and 96 clones of the M. laxa and M. fructicola SSRs enriched libraries were sequenced, respectively. Comparison of the sequences allowed identification of 46 and 66 unique sequences. For 15 M. laxa and 12 M. fructicola sequences it was possible to design primers flanking the sequence repeats. The primers were tested using DNA of the Monilinia spp. from which the marker was developed. Amplification was successfully obtained from 12 M. laxa and four M. fructicola SSR markers. Reciprocal tests indicated that only the M. laxa SSR CHML5 was able to produce amplicons with M. fructicola DNA and that none of the M. fructicola SSRs generated amplicons with M. laxa DNA (data not shown).

Variability of the five SSRs producing amplicons from M. fructicola DNA (Table 3) was assessed by testing the markers on 57 M. fructicola isolates originating from the US. CHMFc1 was the most variable SSR marker (30 different alleles identified), while CHMf12 was the least variable (only one allele, a second allele was identified in the European samples). Markers CHML5, CHMFc4 and CHMFc5 had intermediate variability with five, four and five alleles, respectively (Table 3). Among the 57 isolates from the US a total of 51 unique haplotypes were detected (Table 4).

Table 3.   Primer sequences, repeated sequence, allele ranges, number of alleles and accession number of the sequences of the microsatellite (SSR) markers
MarkerPrimer seq. forward (5′–3′) and label usedPrimer seq. reverse (5′–3′)RepeatsAllele rangeNo. allelesAccession no.
CHML5TGTTCCTCGGGAGTATGAGGTCACCAGCGGTACCATATCCCTC TTC TTC (6)217–2465JF813829
CHMFc1TGATGAAGGGAGGTGTAAAGGACAGGATCCCTCTTCCCAACGATT (18) GAGTT (5)126–20230JF813825
CHMFc4GGCCTTAGCTGTCAGCATTCATTAGGCAGTCGGCTACACGGAGAGAT (4)216–2374JF813826
CHMFc5GAAGCAATTCCGGAGGAAACGAAGTTAACCTTGGAAAGACAAGCGA (15)74–1045JF813827
CHMFc12AAGGGAAAGGCAAGGCTTAGTCAGCAACGATAACCTCTCGTCAA (3)156–1602JF813828
Table 4.   Number of Monilinia fructicola haplotypes and their geographic distribution
CountryNo. of isolatesNo. of haplotypesHaplotype
US06US24US34US39US40US41US45US50FRA01FRA02FRA03FRA04FRA05SP01ITA01ITA02CH01
  1. CA: California; SC: South Carolina; GE: Georgia.

  2. aUS haplotypes not indicated in the table were found only once.

US5751a1CA2SC1SC3GE2SC/GE2SC2GE1GE         
France861       12211    
Spain2057         10111   
Italy214  4    2      114 
CH 08185        6622   2 
CH 095272       3 4732  31
CH 08 + 097082       966932  51

Genetic diversity of European M. fructicola isolates

SSR marker CHML5 revealed one allele (220 bp in length), CHMFc1 three alleles (169, 173 and 176 bp), CHFc4 two alleles (216 and 223 bp), CHMFc5 three alleles (85, 86 and 102 bp) and CHMFc12 only one allele (160 bp) from European isolates (Table 5). The 102 bp allele of CHMFc5 is unique to some European isolates.

Table 5.   Allele (in bp) composition of Monilinia fructicola haplotypes found in European samples
SSR markersHaplotypes
US06US34US50FRA01FRA02FRA03FRA04FRA05SP01ITA01ITA02CH01
CHML5220220220220220220220220220220220220
CHMFc1176173173176176176169169176169169169
CHMFc4216223216223223216223216216223216223
CHMFc58586861028510210210286868585
CHMFc12160160160160160160160160160160160160

Among the 119 European samples 12 different haplotypes were discovered (Tables 4 & 5). Among the eight isolates from France, six haplotypes were found, namely FRA01 to -05 and US06. Haplotype US06 was also identified in the isolate from California. Five different haplotypes (US06, FRA03 to -05 and SP01) were found among the 20 samples from Spain. SP01 was unique to Spain. The most frequent haplotypes identified were FRA03 (10 samples) and US06 (seven samples). Four haplotypes were identified from 21 Italian isolates. Four isolates had haplotypes identical to US34, two were identical to US50, while 11 and four isolates had the haplotypes ITA01 and ITA02, respectively. ITA01 was found only in Italian samples. Haplotypes US34 and US50 were previously identified in the isolates from South Carolina and Georgia, respectively. Five haplotypes were discovered among the 18 Swiss isolates collected in 2008 (all found in orchard J). All of them were also found in French (haplotypes FRA01 to -04) and Italian (ITA02) isolates. The two most frequent haplotypes were FRA01 and FRA02 (six samples each). Seven haplotypes (US06, FRA01, FRA03 to -05, ITA02 and CH01) were identified among the 52 Swiss samples collected in 2009. The most frequent haplotype was FRA05 (32 samples from five orchards). This haplotype was not detected in the 2008 survey. Up to four different haplotypes were found in a single orchard.

Compared to the 2008 samples, FRA02 was not found in 2009 and three new haplotypes were identified (US06, FRA05 and CH01). Haplotype CH01 was only found in Switzerland. The three M. fructicola samples collected in 2009 from orchard J revealed the same haplotypes (FRA01, FRA04 and ITA02) as the samples collected in 2008 from the same orchard. The haplotypes most widely distributed in Europe and identified in France, Spain and Switzerland were US06 and FRA03 to -05.

Population analysis

The structure analyses revealed K = 2 as the most probable number of clusters in 13 out of 16 tested parameter combinations and in all those using the location prior. The same K was also supported by Calinski & Harabasz’ pseudo-F (1974). Group 1 of these two genetically distinct groups is formed by the western US population California and the European populations Spain, France and Switzerland, whereas group 2 is composed of the eastern US populations Georgia and South Carolina. Group membership of Italy is intermediate due to its mixed composition, with most individuals being very similar to a few members of the eastern US population and the remaining being similar to members of group 1 (Fig. 1).

Figure 1.

 Bayesian clustering inferred by structure for K = 2 (using the no admixture model with correlated allele frequencies and sampling locations as prior information) with population Q-matrix (top), representing the mean cluster contribution per population, and individual Q-matrix (bottom) composed of bars with K different fragments of length proportional to each of the inferred clusters. Native = sample locations where Monilinia fructicola is endemic; introduced = sample locations from putative introduction sites in Europe. US_CA = California, US; US_GE = Georgia, US; US_SC = South Carolina, US; ES_EV = Spain, Ebro Valley; FR_SE = France, South East; CH_VA = Switzerland, Valais; IT_PI = Italy, Piemont.

The Index of Association analyses for both datasets (a single representative per haplotype versus one representative per observed haplotype in each population) showed no evidence of clonality in all but the population from South Carolina, where small but significant deviations from zero were observed (dataset 1: IA = 0·19, = 0·05; dataset 2: IA = 0·17; < 0·05). This suggests that recombination may be common in this species and that it likely occurs in all European countries.

Discussion

Monilinia fructicola appears to have increased in occurrence in Switzerland. A survey conducted in 2008 verified the presence of the pathogen in 4·5% of all samples, and M. fructicola was found in only one orchard (orchard J) of 60 orchards with a total occurrence of 1·6% (Hilber-Bodmer et al., 2010). In 2009, 8% of all samples tested positive for M. fructicola and the pathogen was found in nine of 22 (40·9%) orchards. However, it is not clear if the increase in occurrence coincides with an increase of spread. The fact that all M. fructicola samples from 2008 were from orchard J and none from packinghouses (packinghouse-derived fruit from up to 57 orchards represented 70% of the 2008 samples) does not necessarily mean that M. fructicola was not present in the packinghouses. Apricots were cooled to 8°C in the packinghouses upon delivery and it is possible that this temperature selected for M. laxa and M. fructigena prior to sampling. In vitro studies have shown that in contrast to M. fructicola, mycelium of M. laxa and M. fructigena can still grow at 8°C (A. Patocchi, unpublished data). In order to avoid the above-mentioned potential bias, the survey was repeated but this time all samples came from freshly harvested fruits and mummies of various orchards in Canton Valais.

Five SSRs with different degrees of polymorphism were developed, one from M. laxa and four from M. fructicola. CHMFc1 with the composed sequence repeat (GATT)x (GAGTT)y was the most polymorphic and amplified 30 different alleles from 57 US samples. This type of repeat is often highly polymorphic, because it can develop 1 bp differences in nucleotide sequences. In contrast, the least variable SSR, CHMFc12 with the repeat sequence of TCAA, yielded only two alleles. The sequenced clone contained only three repeats of this SSR which, together with the low variability observed in the sample here, suggests that CHMFc12 may be of recent origin. Analysis of additional samples of M. fructicola from additional countries where the pathogen is endemic may help to clarify this hypothesis.

The major hurdles for developing primers for putative SSR markers were the absence or short length of the sequences flanking the cloned SSR repeat regions. The design of flanking primers for SSR repeats was limited to 15 of 46 available unique M. laxa sequences and 12 of 66 M. fructicola sequences. PCR tests showed that only 12 and four primer pairs for M. laxa and M. fructicola, respectively, produced reliable amplicons. A similar success rate for SSR development, using the same method, was reported by Brunner & Frey (2004).

All five developed SSRs were multiplexed in a single PCR reaction. The different range of the alleles of the SSR markers CHML5 (217–246 bp) and CHMFc5 (74–104 bp) enabled the use of the same labelling dye (6-FAM). The possibility of defining the haplotype of a sample with a single PCR reaction considerably reduced the costs and the time needed for the analysis (Patocchi et al., 2009).

Transferability of SSR markers between M. laxa and M. fructicola was very low. Only SSR CHML5 developed from M. laxa produced amplicons from M. fructicola DNA and none of the M. fructicola SSR markers produced amplicons from M. laxa DNA (data not shown). The low transferability of SSRs from one species to another may be a result of high genetic diversity between Monilinia species. Relatively high genetic diversity between M. laxa and M. fructicola was observed by Snyder & Jones (1999) and Gril et al. (2010) in studies using ITS sequences and AFLP markers, respectively. Low level of transferability of SSR markers between related species has already been reported by Dutech et al. (2007).

The five SSR markers, which to the authors’ knowledge are the first ones developed for M. fructicola, allowed a first glimpse at the structure of the European M. fructicola populations. Compared to the US population the genetic diversity in the European populations is low. However, considering the relatively low number of isolates per European country included in this study, a surprisingly high number of haplotypes were discovered per country (4–8) for a disease that has apparently only been recently introduced (Table 4). This may indicate that the pathogen was present in Europe well before its discovery in France in 2001 (EPPO, 2002a), giving it time to diversify. Such diversification may have occurred due to sexual reproduction, accumulation of random mutations, or through gene flow (i.e. multiple invasion events).

The population structure analysis suggests that at least two independent invasion events occurred from the US to Europe. Most European M. fructicola isolates (i.e. Spain, France and Switzerland) show admixture proportions that seem to have originated from California populations. In contrast, the Italian population was composed of individuals with admixture proportions typical for California and some populations of the Eastern US. The results of the population structure analysis are also supported by the discovery of the Californian haplotype US06 in samples originating from Spain, France and Switzerland and the haplotype US34 as well as US50 from South Carolina and Georgia found only in the Italian samples. The recent study of Fan et al. (2010) further supports the hypothesis that one of the sources of M. fructicola found in Europe may be California. Fan et al. (2010) produced a dendrogram based on inter-simple sequence repeats (ISSRs) data including the French reference isolate Ft which was used also in the present study. In the study of Fan et al. (2010) the French isolate grouped together with six other Californian isolates, and in this study all isolates from France clearly present the characteristic of the Californian sample (Fig. 1).

The Index of Association analysis suggests that recombination may be common in this pathogen both in its native range as well as in Europe. Interestingly, eight of the 12 haplotypes found in the European samples can be inferred from the (theoretical) mating of haplotypes ITA02 and FRA01 (Table 6; note that ITA02 is one of the haplotypes grouped together with the Californian haplotype by structure analysis). All possible haplotypes derived from this hypothetical cross were found in Switzerland (Table 4). However, it is unlikely that (with or without sexual reproduction) the disease spread from Switzerland to other countries because Switzerland exports very few fruits or plant material into the European Union. In contrast, Switzerland imports large quantities of fruits potentially hosting M. fructicola (i.e. peaches, nectarines, apricots and cherries) from France, Spain and Italy, suggesting that the pathogen may have been introduced by means of trade. The import of M. fructicola-infected fruits from France and California to Switzerland was previously documented (Bosshard et al., 2006). The recombination event(s) may have taken place in France, considering that six of the eight haplotypes from the hypothetical cross were found in only eight isolates from France. Alternatively, the ‘theoretical’ mating proposed above may have happened in California, where sexual recombination was documented to occur. However, in this case all the derived haplotypes then had to be introduced into Europe. This hypothesis looks less likely than a recombination event which took place in a European country and from there spread into other European countries. Thus far, M. fructicola apothecia have not been documented in a European country.

Table 6.   Hypothetical cross between haplotypes ITA02 and FRA01 and derived progenies
SSRsCHML5CHMFc1CHMFc4CHMFc5CHMFc12Haplotypes
  1. aAllele sizes are indicated in bp.

Parent 1220a16921685160ITA02
Parent 2220176223102160FRA01
Parental 122016921685160ITA02
Parental 2220176223102160FRA01
Progeny 122017622385160FRA02
Progeny 222017621685160US06
Progeny 322016922385160CH01
Progeny 4220169216102160FRA05
Progeny 5220176216102160FRA03
Progeny 6220169223102160FRA04

Acknowledgements

The authors are grateful to Dr R. Ioos (Laboratoire National de la Protection des Végétaux, Malzeville, France), Dr A. de Cal (Instituto Nacional de Investigación y Tecnologia Agraria y Alimentaria INIA, Madrid, Spain) and Dr C. Pellegrino (Centro AGROINNOVA, Grugliasco, Italy) for providing M. fructicola isolates from France, Spain and Italy, respectively. We thank Dr Gabriella Parravicini and Sarah Bryner (Swiss Federal Institute of Technology, Zürich, Switzerland) for technical assistance in the development of the SSR markers; and Dr Patrick Brunner (Swiss Federal Institute of Technology, Zürich, Switzerland) for the fruitful discussions on the population genetics of M. fructicola.

Ancillary