The gene flow and mode of reproduction of Dothistroma septosporum in the Czech Republic

Abstract

Since 1911, dothistroma needle blight, caused by Dothistroma septosporum, has been recorded in most European countries. In the Czech Republic, the fungus has become an important disease of pines since 2000, especially Austrian pines in plantations of Christmas and ornamental trees. The aim of this study was to analyse the population structure, gene flow and mode of reproduction of this pathogen. Microsatellite and mating-type markers were analysed in a Dothistroma population in the southeastern part of the country using reference isolates from other European countries. The haplotypic diversity was high, with 87 unique and 13 shared haplotypes (probable clones) identified in 121 samples. Based on structure analysis, the isolates were divided into two populations, with an uneven distribution over the sampling sites. The grouping of the sites to the populations did not follow a geographical pattern because certain isolates that were sympatrically co-occurring at the same site were placed in different populations. Tests for random mating (the index of association and a parsimony tree-length permutation test) showed a significant clonal mode of reproduction in most cases, but the intrapopulation haplotypic diversity is unexpectedly high. Although a teleomorphic stage of D. septosporum has not been previously observed in the Czech Republic, the high intrapopulation haplotypic diversity can be explained by infrequent sexual reproduction consistent with the occurrence of both mating types.

Introduction

Dothistroma needle blight (DNB) is caused by two closely related species, Dothistroma septosporum and D. pini (the anamorphic stage of Mycosphaerella pini) and results in severe needlecast in pines and, to a lesser extent, other conifers. According to European and Mediterranean Plant Protection Organization (EPPO) and European law, DNB is a quarantine fungal disease in Europe. DNB is distributed worldwide and is reported in at least 63 countries (Watt et al., 2009), and more than 80 host species. DNB has emerged in Europe in the past 15–20 years and is currently widespread in most European countries (Watt et al., 2009). Although its European isolates belong primarily to Dothistroma septosporum, D. pini has been detected in samples from France (Ioos et al., 2010), Hungary, Russia and Ukraine (Barnes et al., 2011). Woods et al. (2005) consider DNB to be a disease supported by global climate change and emphasize the influence of locally enhanced precipitation on the spread of the disease throughout the temperate zone.

In the Czech Republic (CR), D. septosporum was first noted in 1999 on imported nursery stock of Pinus nigra from Hungary, and the first detection of DNB in a plantation was recorded the following year in 6-year-old Austrian pines in Christmas tree plantations (Jankovskýet al., 2004). The increasing number of new findings is explained by the substantial number of trees imported in the 1990s and by the use of infested stock from nurseries until 2000, when strict phytosanitary precautions were instituted throughout the CR. Even though DNB monitoring was conducted prior to the 1990s, it is possible that the disease occurred during that period. In the CR, DNB has been identified in 21 pine species, four spruce species and Douglas-fir (Bednářováet al., 2006); however, D. pini has not been recorded in the country to date (March 2012). Observations of the current DNB outbreak have been conducted throughout the CR, and the disease has been found mainly in Austrian pine plantations or on amenity trees in urban areas, whilst the infestation of native Scots pine stands is relatively rare (Jankovskýet al., 2004; Bednářováet al., 2006). Since 2008, new observations in Scots pine plantations and Scots and bog pine stands in protected areas have been reported; the occurrence of native pine populations is anticipated at such natural sites (Mikeska et al., 2008). Therefore DNB may be caused by an older population of the pathogen there, whereas Austrian pine plantations have been more recently infected by infested nursery stock.

Few studies have investigated the genetic structure of D. septosporum populations. In New Zealand, no genetic diversity was detected in the pathogen using RAPD and RAMS DNA markers (Hirst et al., 1999). Microsatellite markers for D. septosporum were developed by Barnes et al. (2008) but have not been used in any population analysis. Dale et al. (2011) recently used AFLP and mating type markers to reveal a high haplotypic diversity, a mixed mode of reproduction, and an equal mating-type segregation ratio in D. septosporum populations in British Columbia, Canada. A population with a high genetic variation has a strong evolutionary potential, and it is probable that such a population can rapidly adapt to new conditions (McDonald & Linde, 2002). Gene flow among populations and recombination are two crucial sources of new allelic and genotypic variation that can increase the genotypic or haplotypic diversity of a population (McDonald & Linde, 2002). In pleomorphic ascomycetes, the role of propagation by asexual (conidia) or sexual (ascospores) diaspores is crucial for genotypic and haplotypic diversity. However, the role of ascospores for infections in situ is unclear. In Europe, DNB is usually considered a human-mediated disease, and D. septosporum is hypothesized to have short-lived conidia with limited dispersal abilities (Gadgil, 1970; Peterson, 1973). However, the precise data for the CR region are incomplete.

The aim of this study was to analyse the population structure, gene flow and mode of reproduction of D. septosporum in the CR using microsatellite (SSR) and mating-type markers.

Materials and methods

Sampling

Localities of D. septosporum occurrence were monitored between 2005 and 2010. Fungal cultures were isolated from green needles with mature acervuli that were producing fertile conidia, and the sampling was usually performed from June to September. In localities with a relatively high number of isolates, the isolates were obtained from different trees or from the same host tree. The needles were incubated in moist chambers for 2 days to produce fresh conidia in the acervuli, and the presence of the conidia was evaluated under a microscope prior to isolation. Small slivers of needles with acervuli (5 mm long) were soaked in ready-to-use sodium hypochlorite solution (Sigma-Aldrich) for 3–5 min. The samples were then washed twice with 96% ethanol for 1 min and soaked for 5–10 min in sterile distilled water. The samples were placed on Petri dishes with malt extract agar (MEA; 30 g malt extract, 15 g agar and 5 g peptone per litre; Himedia) and incubated at 21°C in the dark; after 2–3 days of inoculation, the culture was transferred to a fresh Petri dish. A culture originating from a single acervulus from one needle was maintained as a single isolate; the additional single spore purification was not carried out. For long-term storage, the isolates were maintained as agar slant cultures with malt extract medium at 7°C in the dark.

In all, 103 isolates from 46 sampling sites were included in the analysis (Table 1).

Table 1. List of strains included in the study

Sample no.^a	Isolate no.	Country	Locality	Host	Host age^b	Habitat^c	Population
^aAB, sample kept as two samples due to an allele polymorphism in a single locus. ^bXT, Christmas tree size (≤2 m high); OXT, overgrown christmas tree; M, mature tree; ×, Pinus mugo and Pinus uncinata subsp. uliginosa not evaluated due to their scrub growth. ^cA, urban greenery; B, Christmas tree plantation in use or abandoned; C, artificial reforestation; D, natural reforestation.
24	1126	Finland	Ähtäri	Pinus sylvestris	OXT	D	2
76	1365	CR	Borkovická Blata	Pinus sylvestris	OXT	D	2
95	1385	CR	Bynina – Christmas tree plantation	Pinus nigra	XT/OXT	B	1
109 AB	1400	CR	Bynina – Christmas tree plantation	Pinus nigra	XT/OXT	B	1
82	1359	CR	Bynina – private garden	Pinus nigra	OXT	B	1
86	1357	CR	Bynina – private garden	Pinus nigra	OXT	B	1
8	1123	CR	Havlíčkův Brod	Pinus nigra	OXT	A	2
23	1124	CR	Havlíčkův Brod	Pinus nigra	OXT	A	2
9	1125	CR	Havlíčkův Brod – Mírovka	Pinus mugo	×	A	1
70	1338	Austria	Hollenstein an der Ybbs	Pinus sp.	OXT	D	1
94	1374	CR	Jakule	Pinus nigra	M	D	1
103	1377	CR	Jakule	Pinus nigra	OXT	D	1
89	1348	CR	Jandovka	Pinus sylvestris	M	D	2
97	1371	CR	Jandovka	Pinus sylvestris	OXT	D	2
34	1165	CR	Jarcová	Pinus nigra	OXT	A	2
35	1166	CR	Jarcová	Pinus nigra	OXT	A	2
36	1167	CR	Jarcová	Pinus nigra	OXT	A	2
87	1354	CR	Jasenice	Pinus nigra	OXT/M	C	1
96	1413	CR	Karolinka – glass factory	Pinus nigra	OXT/M	A	2
102	1405	CR	Karolinka – medical centre	Pinus mugo	×	A	1
108	1421	CR	Karolinka – medical centre	Pinus mugo	×	A	1
1	1110	CR	Koryčany	Pinus nigra	M	C	2
40 AB	1176	CR	Křtiny – arboretum	Pinus ponderosa	OXT	A	2
41	1177	CR	Křtiny – arboretum	Pinus ponderosa	OXT	A	1
42	1178	CR	Křtiny – arboretum	Pinus ponderosa	OXT	A	2
101	1407	CR	Křtiny – arboretum	Pinus ponderosa	OXT	A	2
49 AB	1218	CR	Karviná– Lázně Darkov	Pinus mugo	×	A	2
50 AB	1219	CR	Karviná– Lázně Darkov	Pinus mugo	×	A	2
51	1220	CR	Karviná– Lázně Darkov	Pinus mugo	×	A	1
28	1153	CR	Lhotka nad Bečvou	Pinus nigra	XT	B	1
13	1275	CR	Lidmilův mlýn, Sádek	Pinus mugo	×	A	1
22	1245	CR	Lidmilův mlýn, Sádek	Pinus mugo	×	A	1
26	1246	CR	Lidmilův mlýn, Sádek	Pinus mugo	×	A	1
47	1216	CR	Lidmilův mlýn, Sádek	Pinus nigra	OXT	C	1
64	1331	CR	Luhačovice	Pinus ponderosa	OXT/M	A	2
71	1332	CR	Luhačovice	Pinus ponderosa	OXT/M	A	1
72	1333	CR	Luhačovice	Pinus ponderosa	OXT/M	A	1
17	1250	CR	Mezina	Pinus nigra	XT	C	2
54 AB	1224	CR	Mezina	Pinus nigra	XT	C	1
55	1225	CR	Mezina	Pinus nigra	XT	C	2
56	1226	CR	Mezina at Bruntál	Pinus nigra	XT	C	1
99 AB	1384	CR	Mštenovice	Pinus nigra	M	C	1
16	1258	CR	Nová Pec	Pinus ponderosa	M	A	1
58	1231	CR	Nová Pec	Pinus ponderosa	M	A	1
59 AB	1232	CR	Nová Pec	Pinus ponderosa	M	A	1
84 AB	1367	CR	Nová Pec	Pinus mugo	×	A	2
91	1366	CR	Nová Pec	Pinus mugo	×	A	2
5 AB	1118	CR	Nové Hrady – Sušidla	Pinus mugo	×	A	1
6	1120	CR	Nové Hrady – Sušidla	Pinus nigra	OXT	A	1
11	1300	CR	Odry	Pinus nigra	XT	B	1
62	1295	CR	Odry	Pinus nigra	XT	B	1
63	1296	CR	Odry	Pinus nigra	XT	B	1
33	1163	CR	Oznice at Valašské Meziříčí	Pinus mugo	×	A	1
45 AB	1162	CR	Oznice at Valašské Meziříčí	Pinus mugo	×	A	2
18	1244	CR	Pernek	Pinus uncinata subsp. uliginosa	×	A	2
57	1230	CR	Pernek	Pinus uncinata subsp. uliginosa	×	A	2
73	1362	CR	Pernek	Pinus ponderosa	M	A	1
77	1361	CR	Pernek	Pinus ponderosa	M	A	1
83	1351	CR	Pernek	Pinus ponderosa	M	A	1
67	1327	Finland	Pyhtää	Pinus sylvestris	OXT/M	D	2
107	1420	CR	Rožnov pod Radhoštěm – library	Pinus mugo	×	A	1
37	1171	CR	Rožnov pod Radhoštěm – Písečná	Pinus nigra	OXT	A	2
38 AB	1173	CR	Rožnov pod Radhoštěm – Písečná	Pinus nigra	OXT	A	1
39	1174	CR	Rožnov pod Radhoštěm – Písečná	Pinus nigra	OXT	A	2
85	1353	CR	Rožnov pod Radhoštěm – Písečná	Pinus mugo	×	A	1
90	1364	CR	Rožnov pod Radhoštěm – Písečná	Pinus mugo	×	A	1
92	1346	CR	Rožnov pod Radhoštěm – Písečná	Pinus mugo	×	A	1
79	1368	CR	Rožnov pod Radhoštěm – Rybníčky	Pinus nigra	OXT	A	2
81	1370	CR	Rožnov pod Radhoštěm – Rybníčky	Pinus nigra	OXT	A	2
88 AB	1356	CR	Rožnov pod Radhoštěm – Rybníčky	Pinus nigra	OXT	A	2
69	1329	Norway	Rundhaug	Pinus sylvestris	OXT	D	2
3	1112	CR	Řícmanice – arboretum	Pinus cembra	OXT/M	A	1
4	1113	CR	Řícmanice – arboretum	Pseudotsuga menziesii	M	A	1
10	1127	CR	Řícmanice – arboretum	Pinus nigra	OXT/M	A	2
25	1128	CR	Řícmanice – arboretum	Pinus mugo	×	A	1
14	1270	CR	Sádek	Pinus nigra	OXT	A	1
15	1268	CR	Sádek	Pinus ponderosa	M	A	1
27	1247	CR	Sádek	Pinus ponderosa	M	A	1
30	1248	CR	Sádek	Pinus ponderosa	M	A	1
48	1217	CR	Sádek	Pinus ponderosa	M	A	1
61	1269	CR	Sádek	Pinus nigra	OXT	A	1
7	1121	CR	Soběslavská Blata	Pinus sylvestris	OXT	D	2
2	1111	CR	Strhaře	Pinus nigra	OXT	B	1
104 AB	1403	CR	Střitež nad Bečvou	Pinus mugo	×	A	1
68 AB	1328	Finland	Suonenjoki	Pinus sylvestris	OXT/M	D	2
80	1350	CR	Sušidla	Pinus sylvestris	OXT	D	2
43	1183	Slovenia	the river Soča valley	Pinus nigra	M	D	2
44 AB	1184	Slovenia	the river Soča valley	Pinus nigra	M	D	2
46	1186	Slovenia	the river Soča valley	Picea abies	M	D	2
100 AB	1411	CR	Valašské Klobouky	Pinus nigra	M	A	1
110	1380	CR	Valašské Klobouky	Pinus nigra	M	A	1
12	1294	CR	Vídeň at Velké Meziříčí	Pinus nigra	OXT	A	2
21	1242	CR	Vídeň at Velké Meziříčí	Pinus nigra	OXT	A	2
60 AB	1241	CR	Vídeň at Velké Meziříčí	Pinus nigra	OXT	A	1
20	1223	CR	Šance Dam	Pinus mugo	×	A	2
52	1221	CR	Šance Dam	Pinus mugo	×	A	2
53	1222	CR	Šance Dam	Pinus mugo	×	A	2
19	1234	CR	Zálší at Veselí nad Lužnicí	Pinus sylvestris	OXT	C	2
74	1358	CR	Zašová– Lužanka restaurant	Pinus ponderosa	OXT	A	1
29	1154	CR	Zubří– sawmill	Pinus ponderosa	OXT	A	1
31	1156	CR	Zubří– sawmill	Pinus ponderosa	OXT	A	2
32	1157	CR	Zubří– poultry farm	Pinus nigra	OXT	B	2
105 AB	1409	CR	Zubří– poultry farm	Pinus nigra	OXT	B	1

DNA extraction and PCR

DNA was extracted from freshly grown cultures using the DNeasy plant mini kit (QIAGEN). The identity of D. septosporum (as distinguished from D. pini) was determined by the species-specific PCR targeting of the beta-tubulin gene (Ioos et al., 2010). All of the sampled isolates were confirmed as D. septosporum. Labelled primers derived by Barnes et al. (2008) were used for the SSR analysis. The following primer pairs (loci) for the analysis were selected by M. Mueller (Finnish Forest Research Institute METLA, Vantaa, Finland, personal communication): Doth_F, Doth_G, Doth_J, Doth_K, Doth_L, Doth_S, Doth_DS1 and Doth_DS2. The forward primers were labelled with one of four fluorophores on the 5′ end (6-FAM, VIC, NED or PET). The combinations of primer pairs sharing the same fluorophores were chosen when the PCR products of the two primer pairs were clearly distinguishable according to their size. Each microsatellite locus was amplified by an independent PCR reaction (Barnes et al., 2008) using Taq DNA polymerase (Promega). The SSR locus sizes were analysed using the ABI PRISM 3100-Avant Genetic Analyzer (Applied Biosystems) and determined with genemarker v. 1.3 software.

In 18 isolates, two alleles were detected at a single locus – such isolates probably do not have a single spore origin. These isolates were analysed as two samples (A, B) differing in these allele sizes. In total, 121 samples were analysed. The core data set (111 samples) originated in the Czech Republic. The 10 remaining samples were from other European countries (Austria, Finland, Norway and Slovenia).

The mating-type genes were amplified with species-specific, mating-type-specific primers using procedures and primers developed by Groenewald et al. (2007). The PCR products were viewed under UV light on 1% agarose gels stained with SERVA DNA Stain G. The fragments were compared with a 100 bp DNA ladder (New England Biolabs) and scored as either mating type 1 (820 bp) or mating type 2 (480 bp; Groenewald et al., 2007).

Data analysis

The haplotypes were determined using arlequin v. 3.5.1.2 (Excoffier & Lischer, 2010). Two sets of haplotype data were created. The first data set included all of the haplotypes, whereas the second data set was purged of clones, with each individual haplotype present as only one copy (clone-corrected). Prior to the analysis of the mode of reproduction, population assignments were carried out to avoid the Wahlund effect (reduction of heterozygosity in a population caused by subpopulation structure, e.g. caused by different allele frequencies within subpopulations). The samples were grouped into populations according to structure– a Bayesian, model-based algorithm (Hubisz et al., 2009), and the number of populations resulting from the grouping was estimated according to Evanno et al. (2005). The haplotypic diversity was calculated on the full data set, whereas the genetic diversity and population differentiation statistics were computed on the clone-corrected data set. The methods used for assessing the mode of reproduction included both data sets to avoid a reduced probability of rejection of the null hypothesis of random mating in the clone-corrected data set (Milgroom, 1996). Each population was diagnosed for the number of polymorphic loci and number of unique haplotypes (occurring only once in the entire data set). multilocus v. 1.3b (Agapow & Burt, 2001) was used for the calculations of haplotypic diversity. Nei’s gene diversity (Nei, 1987) was computed as the average gene diversity over all of the loci in arlequin v. 3.5.1.2. The pairwise F_ST– measure of population differentiation based on allele frequency differences (Wright, 1949) and number of migrants Nm– derived from F_ST computation (Slatkin, 1991) between the populations, were also calculated. The presence of random mating was tested using the index of association and the parsimony tree-length permutation test. The index of association, I_A (Brown et al., 1980; Smith et al., 1993; Haubold et al., 1998; Agapow & Burt, 2001), was calculated using multilocus v. 1.3b. A randomization (10 000 repeats) of alleles among individuals, independently for each locus (Burt et al., 1996), was performed to assess any significant deviations from the gametic equilibrium expected under the null hypothesis of random mating (α = 0·05). To exclude the risk of a physical linkage between the loci, which would skew the results, an exact test between all of the pairs of loci in each population was performed using arlequin v. 3.5.1.2. Because no loci were found to be linked, no further precautions were taken before the analysis.

The parsimony tree-length permutation test (PTLPT) was used to compare the lengths of the most parsimonious trees obtained from the observed data to the lengths of comparable trees from randomized data sets in a manner similar to that used in the previous randomizations. Again, 10 000 randomizations were performed to assess the statistical significance associated with the test of the null hypothesis of random mating. Under random mating, the expected tree length is substantially longer, and more than one tree would be identified. In contrast, a clonal population should produce one well-resolved, significantly shorter tree (Burt et al., 1996). The data sets for this analysis were generated using multilocus v. 1.3b and subsequently analysed using paup v. 4.0b10 (Swofford, 2003).

To examine the dispersal patterns, the relationships between the sites were tested to evaluate the distribution of rare alleles and shared haplotypes. A local spatial analysis was performed, and the correlation between the geographic and genetic distance matrices was assessed. The allele frequencies were determined on the clone-corrected data set for each population (arlequin v. 3.5.1.2). The genetic-distance-matrix computation and a Mantel test were performed in GenAlEx v. 6.41 (Peakall & Smouse, 2006) to detect any possible geographic structure shown by the samples (i.e. isolation by distance). To study the within-site spatial patterns, the three sites with the highest number of samples from a single population were selected (Sádek, Rožnov p. Radhoštěm – Písečná and Nová Pec). A 2-D local spatial analysis (GenAlEx v. 6.41) was used on the full data set to determine whether the individuals were genetically correlated with their two nearest neighbours. This analysis was then repeated to determine whether the individuals were genetically correlated with their three nearest neighbours. If clonal reproduction is prevalent, samples located closer to each other are expected to be more genetically similar. The three sites were then combined in a Multiple Dclass analysis (GenAlEx v. 6.41) using the full data set. A total of 40 runs were performed, with the distance increasing by 5 km on each successive run.

Results

All eight of the microsatellite markers were polymorphic, and 81 different alleles were detected at the loci analysed. The eight preselected markers were determined to be sufficient for the population genetic analysis because the plot of genetic diversity versus the number of loci showed that six of the markers explained 99% of the observed variation. In all, 100 haplotypes were found. Of these, 87 were unique in the entire collection (i.e. present in only one copy); of the remaining 13 shared haplotypes, nine were found only within sites, and four were shared within and between sites. The population allele frequencies are shown in Table 2.

Table 2. Dothistroma septosporum population allele frequency based on eight microsatellite markers

Locus	Population allele frequency
	Allele (bp)	Pop1	Pop2
J	182	0·038462	0·104167
	186	0	0·083333
	188	0·250000	0·312500
	190	0·480769	0·479167
	192	0·230769	0
	204	0	0·020833
DS2	361	0·153846	0·041667
	371	0·115385	0·020833
	375	0·230769	0·020833
	377	0·153846	0·104167
	379	0	0·020833
	381	0·019231	0·020833
	385	0	0·020833
	387	0·019231	0·125000
	397	0	0·104167
	399	0	0·125000
	400	0	0·020833
	401	0·076923	0·020833
	403	0·038462	0
	407	0·192308	0·125000
	421	0	0·020833
	427	0	0·020833
	429	0	0·041667
	446	0	0·020833
	450	0	0·020833
	489	0	0·020833
	493	0	0·041667
	495	0	0·041667
K	336	0·557692	0·125000
	350	0	0·020833
	354	0·057692	0·041667
	356	0·192308	0·145833
	358	0·173077	0·104167
	360	0	0·062500
	362	0	0·041667
	364	0	0·041667
	366	0·019231	0·187500
	368	0	0·020833
	370	0	0·020833
	374	0	0·062500
	397	0	0·020833
	408	0	0·020833
	412	0	0·020833
	415	0	0·062500
G	172	0·019231	0
	179	0·019231	0·020833
	182	0·211538	0·333333
	184	0·750000	0·625000
	186	0	0·020833
F	172	0	0·020833
	174	0·288462	0·854167
	176	0·711538	0·125000
S	330	1·000000	0·770833
	346	0	0·020833
	364	0	0·208333
DS1	143	0	0·020833
	147	0	0·062500
	151	0·480769	0·166667
	153	0·019231	0·270833
	155	0·038462	0·104167
	157	0·115385	0·291667
	159	0·230769	0·020833
	161	0·115385	0·020833
	174	0	0·041667
L	304	0·038462	0
	308	0	0·020833
	321	0·846154	0·520833
	337	0	0·062500
	341	0·057692	0
	342	0	0·020833
	345	0	0·020833
	347	0	0·020833
	349	0	0·041667
	351	0·019231	0·125000
	357	0·019231	0·041667
	359	0	0·020833
	362	0	0·020833
	363	0·019231	0
	367	0	0·041667
	369	0	0·020833
	379	0	0·020833

structure analysis was performed in order to assign individual isolates to populations. Accordingly, samples were divided into two populations based on K parameter (Fig. 1), each with 29 sampling sites. Therefore, the two populations sympatrically co-occurred at 12 of the sampling sites. The placement of the samples into populations is shown in Table 1. Both populations have a similar proportion of polymorphic loci, and the percentage of variation is almost three times higher within populations than between them. The population statistics and characteristics are shown in Table 3. The observed haplotypic diversity was high, and the values of Nei’s gene diversity were 0·49 (Pop1) and 0·64 (Pop2). The F_ST and predicted Nm values between these populations are 0·2939 and 1·20124. These numbers, supported by results of a Mantel test, are indicative of prevalent range-limited spread by conidia and significant differentiation between two populations with very low numbers of migrants. Assessment of Nm is susceptible to violation of the underlying model for F_ST computation (Whitlock & McCauley, 1999) but nevertheless the computed number indicates very limited exchange of migrants between populations. The exact tests (arlequin v. 3.5.1.2) performed for each of the populations using the clone-corrected data set did not indicate that any of the loci are in linkage disequilibrium.

Figure 1.  Delta K plot of the structure analysis, showing K = 2 as the most probable number of populations in the data set.

Table 3. Population genetic statistics for Dothistroma septosporum populations

Population	Sample size				Diversity
	Full data	CC	Loci^a	No. of haplotypes^b	Haplotypic^c	Nei’s gene^d
CC: clone corrected. ^aProportion of polymorphic loci. ^bNumber of unique haplotypes. ^cHaplotypic diversity (multilocus v.1.3b) calculated using complete data set. ^dGene diversity calculated on CC data using arlequin v. 3.5.1.2; average gene diversity over loci.
Pop1	66	52	0·875	44	0·989744	0·494344
Pop2	55	48	1·00	43	0·993939	0·648715

In Pop1, including Czech and Austrian samples, significant correlation between the genetic and geographic distances was evidenced by a Mantel test conforming to assumed spatially-limited conidial spread. In comparison, Pop2 contained Czech and the remaining foreign samples from distant localities which were genetically similar. In the 2-D local spatial analysis, the sites with the highest number of samples in Pop1 (Sádek, Rožnov p. Radhoštěm - Písečná and Nová Pec sites) showed significant correlations with their two nearest neighbours as follows: four of five samples from Rožnov, all of the samples from Nová Pec and one of six from Sádek showed correlations. In the analysis of the three nearest neighbours, three of seven of the samples showed correlations at Rožnov, and three of four samples at Nová Pec. No correlations were found for Sádek. In the Multiple Dclass analysis, the correlation was the highest at the shortest distances and decreased as the distance increased, but it remained significant, even at the maximum distance tested.

Tests for random mating were performed using the index of association and a PTLPT analysis (Table 4) and generally showed a significant clonal mode of reproduction, even in the clone-corrected data set. However, the I_A index for Pop1 suggested random mating. With the exception of Pop1, the PTLPT tests showed results that were consistent with the I_A values, even in the clone-corrected data set.

Table 4. Random mating test of Dothistroma septosporum on full data set and clone-corrected (CC) data set

Population	Index of association (IA)^a		PTLPT^b
	IA	P value	L	L*	P value
^aValues obtained after 10 000 randomizations from multilocus. ^bParsimony tree-length permutation test (PTLPT); L = observed length of tree produced through parsimony analysis (number in parentheses corresponds to number of most parsimonious trees found); L* = length of the shortest tree in the randomized data (10 000 replicates).
Pop1	0·07102	0·145	73 (1)	86	<0·0001
Pop2	0·36831	<0·0001	104 (2)	129	<0·0001
Total	0·38554	<0·0001	181 (1)	229	<0·0001
Pop1 CC	0·0492321	0·5639	75 (2)	71	0·002985
Pop2 CC	0·286235	0·0005	104 (1)	116	<0·0001
Total CC	0·33665	<0·0001	176 (2)	199	<0·0001

Both mating types were found in the Czech sampling sites; in contrast, all of the Fennoscandian samples included only mating type 2. Mating type 2 occurred more frequently than mating type 1 in both Pop1 and Pop2. However, the ratio of mating type 1 to mating type 2 including samples from both populations did not differ significantly from a 1:1 segregation (χ² = 0·54, P = 0·462), and a chi-square analysis of the clone-corrected data yielded similar results (χ² = 0·87, P = 0·349). Nevertheless, both of the mating types were detected within a sampling site in 10 cases (21·7% of the sampling sites; Bynina – Christmas tree plantation, Jakule, Karolínka – medical centre, Luhačovice, Mštenovice, Pernek, Valašské Klobouky, Vídeň at Velké Meziříčí, Zubří– sawmill and Zubří– poultry farm). Among these sites, five had samples that all belonged to Pop1 (Bynina – Christmas tree plantation, Jakule, Karolínka – medical centre, Mštenovice and Valašské Klobouky) and one site had samples that belonged to Pop2 (Zubří– sawmill). The four remaining sites included samples that fell into both populations.

Discussion

The genetic separation of Dothistroma samples into two populations by structure analysis does not correspond to their geographical location. The highly geographically separated samples from Austria, Finland, Norway and Slovenia were placed into Pop2, as were 23 of the Czech sites. This result could be an artefact of the analysis resulting from insufficient sampling of these sparsely sampled distant locations. Alternatively, the Pop2 sites could represent descendants of the same former colonization event. However, samples placed in both populations occurred sympatrically at 12 of the Czech sites. For example, three sites in the town of Rožnov pod Radhoštěm (library, Písečná and Rybníčky), separated by a distance of 2 km, included samples from both populations. Moreover, the site sampled in detail (Písečná) included samples from both populations. The occurrence of both populations is primarily limited to urban greenery and arboretums, where the occurrence of the pathogen is strongly influenced by human activities and where the disease was introduced by the planting of infected nursery stock. The effects of human-mediated haplotype transfer are misleading in terms of the population assignment. It is probable that the conidia are dispersed only over short distances (maximum 150 m) by rain or splashing water (Peterson, 1973) and that the unreleased conidia in acervuli lose viability during the 4–6 months on the ground (Gadgil, 1970). Therefore, the role of the conidia in long-distance dispersal is limited. These assumptions were partially corroborated by the spatial analyses that showed a significant genetic correlation between the samples in Pop1 located in close proximity. The F_ST and Nm values indicate limited ability of the fungal diaspores to disperse over long distances as they show significant differentiation between populations, and marginal numbers of migrants between populations. The haplotypes shared among the samples (presumably clones) were distributed primarily within the sampling sites, and only four such haplotypes occurred at two different sites. Nová Pec and Pernek, sharing two common haplotypes, are approximately 4 km apart. The first possible clone was recorded only on Pinus ponderosa at both of the sites, whereas the second possible clone was detected on two pine species, Pinus mugo and P. uncinata subsp. uliginosa. The other example of identical haplotypes was detected at Lhotka nad Bečvou and Jasenice (5 km apart): both of the occurrences were found on P. nigra, and it is possible that these pines shared a common origin. These findings of shared haplotypes are explainable by human-mediated haplotype transfer. Nevertheless, one example of a shared haplotype occurred in Havlíčkův Brod and Křtiny, 90 km apart, a result that is surprising.

The observed overall haplotypic diversity was high, which agrees with the results of Dale et al. (2011). If the hypothesis of a relatively recent introduction into Europe is considered, this unexpected outcome could be explained either by occasional genetic recombination, by the colonization of an array of genetically distinct individuals, or by the simultaneous colonization of different sites by genetically diverse diaspores. The most probable explanation would include a combination of all these hypotheses. The division of the data set into two almost-equal numbers of samples assigned to each population may be explained by two separate introductions of D. septosporum into Europe. This hypothesis could be tested by an analysis of the genetic similarity between isolates from potential native populations located in either Central America or Nepal (Evans, 1984; Ivory, 1994). The first European record of D. septosporum was published in 1911 by Doroguine in the boreal zone around St. Petersburg, Russia (Doroguine, 1911). During the 1950s and 1960s, occasional occurrences of the species were recorded in Austria (Petrak, 1961), the UK (Murray & Batko, 1962) and France (Morelet, 1967). Although the presence of D. septosporum was confirmed in the Czech Republic during 1999 on imported stock and during 2000 in plantations (Jankovskýet al., 2004), the fungus may have persisted in the country for a longer period of time, and the accumulation of the observed haplotypic diversity could have occurred. The biology of the fungus allows it to respond sensitively to weather conditions (Woods et al., 2005; Welsh et al., 2009), and it is probable that the recent fluctuations in climatic conditions provided DNB better opportunities to spread. In general, it is probable that forest trees respond slowly to disease, because the generation time of trees is much longer than the timescale on which such pathogens can spread (Shaw & Osborne, 2011).

Although the random mating tests performed on the data indicate a dominant clonal mode of reproduction, the high level of intrapopulation haplotypic diversity can be explained by infrequent sexual reproduction. This explanation is in agreement with Dale et al. (2011). A teleomorphic stage of D. septosporum has not been observed in the CR, and reports of a teleomorphic stage in other European countries are unusual (Butin, 1985; Karadžić, 1989; Kowalski & Jankowiak, 1998). The asci and ascospores have only rarely been observed, although an extensive observation was performed (Kowalski & Jankowiak, 1998). Nevertheless, the occurrence of a teleomorphic stage was occasionally recorded in two neighbouring countries, Germany and Poland, with similar environmental conditions. Additionally, Groenewald et al. (2007) confirmed the presence of both mating types in Austria and Poland. The results presented here confirm the occurrence of both mating types at 10 sampling sites (21·7% of all of the sampling sites) with the possibility for sexual reproduction of the pathogen to occasionally occur. Four of these sites (Luhačovice, Pernek, Vídeň at Velké Meziříčí and Zubří– poultry farm) included samples placed in both Pop1 and Pop2. Such localities may serve as local hotspots of genotype diversity. The current haplotype diversity within these sites can be explained by possible mating and subsequent recombination, resulting in mixed genotypes that tend to assign to either Pop1 or Pop2. In addition, the I_A results for the clone-corrected data in Pop1 suggest random mating. Nevertheless, I_A is known to be less sensitive to detecting clonal mode than PTLPT analysis (Taylor et al., 1999). However, according to previous findings (Jankovskýet al., 2004; Bednářováet al., 2006), the spread of the fungus results primarily from the movement of infected planting stock. In urban areas and arboretums, it is probable that the diversity of D. septosporum is increased by the introduction of diverse fungal diaspores of multiple origins. Apart from these instances of long-distance human-mediated transfer, the local genotypic structure is consistent with a hypothesis of predominantly conidial, range-limited spread. This hypothesis is supported by the results of the Mantel test and the spatial analyses.

The future spread of DNB can be predicted as a slow dispersal across short distances via conidia if human-mediated transfer is reduced. From the practical point of view, the control of nurseries and plantations is necessary to prevent epidemic dispersal of the pathogen. The respective phytosanitary precautions are currently applied in many EU countries. Phytosanitary authorities should pay extensive attention to imported coniferous planting stock to prevent increase of haplotypic diversity of the pathogen due to import of diverse genotypes of D. septosporum from foreign countries.

Acknowledgements

This work was supported by the Ministry of Agriculture of the Czech Republic project no. QH81039 and by Internal grant agency of the Faculty of Forestry and Wood Technology, Mendel University in Brno, project no. IGA 30/2010. Lucie Zejdová is thanked for the help with construction of Figure 2.

Figure 2.  Map of the southeastern region of the Czech Republic, showing the sampling sites and the association of the isolates studied with the populations Pop1 and Pop2.