Introduction

Invasions of exotic pests and pathogens constitute a growing threat to agriculture, human health and natural environments. Biological invasion is a process with a series of steps, which may include transport, colonization and establishment, a lag time, range expansion and spread. Invasive species may evolve during both their initial establishment and their subsequent range expansion, especially in response to the selection generated by the novel environment.

By comparing the genetic diversity of invasive species, between their native and introduced areas, we can gain insight into the invasion process. The genetic diversity of many invasive species has been studied in the past decade (Tsutsui et al., 2000; Sakai et al., 2001; Kolbe et al., 2004), and revealed different patterns. While most species show evidence of founder effects and genetic bottlenecks, a few actually exhibit an increase in genetic diversity during the invasion process.

Pinewood nematode (Bursaphelenchus xylophilus (Steiner et Burhrer) Nickle) is an extremely dangerous invasive species, which causes a destructive forest disease—pine wilt disease. Since the 1970s, pine wilt disease has been prevalent in Asia, including areas in Japan, Korea, Hong Kong, Taiwan and Mainland China. It has caused tremendous economic losses in these regions. In Mainland China, pinewood nematode was first found in Zijingshan (the Purple Mountain), Nanjing, Jiangsu province, in 1982 (Cheng et al., 1983). Since then, it has been found in other provinces, including Ma'anshan and Chuzhou in Anhui province (1988), Shenzhen in Guangdong province (1988), Yantai in Shandong province (1991), Zhoushan in Zhejiang province (1991), Enshi in Hubei province (1999), Zunyi in Guizhou province (2001) and Fuling in Chongqing (2003) (Yang et al., 2003). Although the Chinese government has adopted stringent control measures, the disease is still spreading rapidly northward and westward in the southern and southeastern parts of China. It has killed a large number of pine trees and destroyed many pine woods (mainly of Pinus massoniana, which is a dominant pine species in southern China), and has caused great economic losses and had serious impacts on local forests, landscapes and ecological environments.

B. xylophilus is widely distributed in North America (Canada, United States and Mexico) (CABI/EPPO, 1999). In the United States, it has been found in 36 states, including all the Great Plains states except North Dakota. It is a native species in North America (Kanzaki and Futai, 2002). As the type specimens of B. xylophilus were found in the United States, we take this as its native region. In its native region, pinewood nematode does not kill pine trees and rarely causes economic losses (Dwinell and Nickle, 1989), although it was reported to damage alien pines in Columbia, MO, USA (Dropkin and Foudin, 1979). The question as to why pinewood nematode spread uncontrollably and became highly pathogenic to pine trees when it was introduced into the Asian regions remains unanswered.

With such a successful invasive species, exploring the invasion process and its genetics is important. However, until now, most studies have focused on comparing the molecular differences between B. xylophilus and its sister species (de Guiran and Bruguier, 1989; Webster et al., 1990; Harmey and Harmey, 1993; Iwahori et al., 1998; Beckenbach et al., 1999; Kanzaki and Futai, 2002; Zheng et al., 2003). Few studies have examined the genetic differences within the species, except Zhou's latest study (Zhou et al., 2007). In this paper, we focus on two questions: how the genetic diversity of pinewood nematode has changed during the invasion process, and how pinewood nematode spread among geographical regions in China.

Materials and methods

Pinewood nematode samples

We collected 134 samples of B. xylophilus from different regions in 2005. We isolated 119 Chinese samples from infested wood of dying pine trees sampled randomly from 28 sites in seven provinces. Each site represents a county in which pine wilt disease has been found, based on the published data of the Chinese Forestry Ministry. We isolated 13 American samples and two Japanese samples from materials intercepted by the quarantine departments. Each sample was isolated from wood chip. In addition, we collected seven samples of a sister species (B. mucronatus Mamiya (1986), an ubiquitous native species) from different localities to provide an outgroup. Table 1 lists all the samples and their sources. Figure 1 shows the sampling sites in China.

Table 1 List of samples and localities of B. xylophilus (1–34) and B. mucronatus (35–41)
Full size table
Figure 1
figure 1

Map of geographic distribution of B. xylophilus (shadow parts) with sampling sites (black dots) and range of geographic populations (circles) in China.

Full size image

Nematode isolation and culture

We extracted live nematodes from infested wood by using the Baermann's funnel technique, which involves immersion of small pieces of wood in water. Nematodes migrating from the chopped wood into the water are collected from the closed bottom of the funnel. Using a stereomicroscope, we transferred the nematodes to a glass slide for microscopic examination. After identifying nematodes by their morphological features under a high-powered microscope, we selected about 100 nematodes from each sample to be cultured on a fungal mat of Botrytis cinerea grown on potato dextrose agar culture medium in a Petri dish at 24–25 °C for about 20 days.

DNA extraction and molecular identification

We picked up the cultured nematodes, sterilized them with 3% H2O2 for 10 min and then washed them four or five times with distilled water. We transferred a 100 μl suspension (more than 20 000 individuals) into extraction buffer and ground it with a pestle and mortar. We used a phenol/chloroform procedure to extract DNA. Following ethanol precipitation, we resuspended DNA in TE buffer (pH 8.0) and stored it at −20 °C. Detailed methods were derived from Molecular Cloning (Sambrook et al., 1989).

To ensure that the DNA samples of B. xylophilus and B. mucronatus were pure and not mixed with DNA of other nematodes, we used the molecular identification method internal transcribed spacer (ITS)–polymerase chain reaction (PCR)–restriction-fragment length polymorphism (RFLP). The primer sequences were those used by Iwahori (Iwahori et al., 1998). We used the restriction endonuclease HinfI (New England BioLabs Inc., Ipswich, MA, USA) to digest the product of PCR. The RFLP patterns in each species were distinctly different (Iwahori et al., 2000). The pattern of B. xylophilus was characterized by band sizes of 150, 155, 250 and 255 bp, whereas the pattern of B. mucronatus had band sizes of 110, 160, 242 and 404 bp.

Molecular method selection

Various methods have been used in genetic studies of pinewood nematode, such as PCR–RFLP (Iwahori et al., 1998; Zheng et al., 2003; Takemoto et al., 2005), DNA sequencing (Beckenbach et al., 1999; Kanzaki and Futai, 2002) and random amplification primer DNA (RAPD) techniques (Zheng, 1998). However, none of these three methods would have been appropriate for our study of pinewood nematode. PCR–RFLP is limited by its low polymorphism. DNA sequencing reveals that the rate of gene evolution is too low to adequately address intraspecific genetic variation. We previously tried to compare DNA sequences of 14 geographical subpopulations of B. xylophilus from different regions of China, including mitochondrial DNA COI gene and ribosomal DNA gene ITS region, and the result showed that no distinct differentiation among these populations (consensus of nucleotide sequences >99.5%) (X-Y Cheng, unpublished data). RAPD analysis is limited by its poor reproducibility.

Among the alternative methods available, amplified fragment length polymorphism (AFLP) analysis is attractive, because of its high polymorphism, its high reproducibility and because it is easy to develop without any previous knowledge of the genome. It is therefore extensively used in research in plants, bacteria, fungi and animals (Keiper and McConchie, 2000; Ribeiro et al., 2002), and is also widely used in studies of invasive species (Amsellem et al., 2000; Darling et al., 2004) and nematodes (Folkertsma et al., 1996; Semblat et al., 1998). To detect the genetic diversity of pinewood nematode, we have previously established a suitable AFLP protocol by screening 52 pairs of primers selected from 1480 primer combinations for pinewood nematode analysis (Cheng et al., 2005). We used a selection of these primers in this study.

AFLP fingerprinting

We carried out AFLP reactions according to the standard protocol with double restriction (EcoRI, MseI), ligation, preamplification and selective amplification (Vos et al., 1995). Each PCR reaction was accompanied by a negative control to monitor contamination. The restriction enzymes were from New England BioLabs Inc., and T4 DNA ligase and PCR reagents were from Promega Inc. (Madison, WI, USA). All oligonucleotides were synthesized by Sangon Co. Ltd (Shanghai, China). The sequences of the EcoRI-adapter and MseI-adapter and primers for preamplification followed those of Vos et al. (1995). Other reagents were from Amresco Inc. (Solon, OH, USA).

We selected 10 primer pairs from those identified by Cheng et al. (2005). The primer combinations were as follows EcoRI-AAG, EcoRI-AAC, EcoRI-TGA, EcoRI-TTG, EcoRI-TAC, EcoRI-AGA, EcoRI-AGT, EcoRI-TTT, EcoRI-AAG and EcoRI-TAG in combination with MseI-CGT and MseI-CAA.

We labeled MseI primers fluorescently with 6-FAM at the 5′ end. Although labeling MseI primers may not be as common as labeling EcoRI primers, similar fingerprints could be obtained with either the EcoRI or MseI primer labeled in Vos et al. (1995). Most of the detected fragments were EcoRI–MseI, because amplification of the MseI–MseI fragments was inefficient in the presence of the EcoRI primers—that is, there was preferential amplification of EcoRI–MseI fragments compared with MseI–MseI fragments in the AFLP reaction (Vos et al., 1995).

We separated PCR products on 5% denaturing polyacrylamide gels on an ABI Prism 3100 genetic analyzer (Applied Biosystems, Foster, CA, USA), with internal molecular weight size standards (GENESCAN-500 ROX). We scored bands semiautomatically by using GENOTYPER software and checked them by hand. To assess any bias (Bonin et al., 2004), we confirmed the genotypes of seven samples by at least two independent repeats (starting from DNA extracts), which revealed the amplification patterns to be highly reproducible.

Data analyses

For each sample and each primer combination, we determined the number of AFLP fragments and the size of each fragment within the range 50–500 bp. Only bands with peak value more than 200 were chosen for statistical analysis. Bands of similar sizes were assumed to be homologous. We calculated allelic frequencies for the marker alleles associated with each fragment (presence (1) and absence (0)). As AFLP loci can be scored as dominant diploid markers, we estimated null allele and then overall allele frequencies. Here, we assumed that the population of pinewood nematode was in Hardy–Weinberg equilibrium, since pinewood nematode has full bisexual reproduction and is transmitted widely through its habitat by vector insects (Monochamus alternatus). For each primer pair, we determined the number of polymorphic and monomorphic bands. The percentage of polymorphic loci estimated was based on the percentage of loci not fixed for one allele (P95). Following Lynch and Milligan's (1994) recommendation that low-frequency markers should be removed from the data set for dominant data, we used only fragments with an observed frequency of less than 1–3/N (the total sample size) for analysis (Lowe et al., 2004).

We evaluated genetic diversity within each population by using three indices: the percentage of polymorphic loci, Nei's genetic diversity and the Shannon's information index.

We calculated Nei's diversity (h) using the following formula (Lowe et al., 2004):

where xi was the population frequency of each allele (1,0) at locus i. The population mean diversity was estimated by the average over all markers.

We calculated the Shannon's information index (I) (Lowe et al., 2004) as a measure of gene diversity across all loci according to the formula:

where pi was also the proportion of the ith allele in the population.

As allelic richness can be heavily influenced by variation in sample size, sample size must be standardized with the lowest sample number and equal-sized samples randomly selected from the other populations. We repeated this 10 times. We calculated the mean and standard errors.

To estimate the changes of genetic diversity of pinewood nematode populations in the invasive process, we needed samples with different invasive histories. However, such samples are difficult to obtain, so we used the occurrence history of pine wilt disease instead of the history of pinewood nematode invasion (not considering the lag time). We assigned five-time stages according to the occurrence time of pine wilt disease in China: <5 years (occurred after 2000), 5–10 years (in 1996–2000), 11–15 years (in 1991–1995), 16–20 years (in 1985–1990) and >20 years (before 1985) (according to the time of samples collected). By comparing the difference in genetic diversity of the different groups, we hoped to clarify the changes in genetic variation in pinewood nematode populations during an invasion.

We also calculated Nei's unbiased genetic distance. We used the genetic distance matrix to construct a dendrogram by using the unweighted pair group method with arithmetic mean (UPGMA) and neighbor-joining (NJ) methods performed by the software PHYLIP 3.62 (Felsenstein, 2000). The statistical robustness of each node was estimated by bootstrapping with 1000 replicates.

We used the software SPSS for windows version 10.0 (SPSS Inc., 1999) for all statistical analyses.

Results

AFLP polymorphism

The 10 primer pairs yielded 2410 AFLP markers (across the 141 samples of nematodes), of which 1138 were polymorphic loci. The average polymorphism of loci generated by a primer pair was 46.84%, with a range from 34.90 to 53.82% (Table 2). The error rate estimated from the repeated genotyping was in the range 2.8–3.2%.

Table 2 Polymorphism revealed for each primer pair
Full size table

Comparison of genetic diversity (P, h, I) between Chinese populations and American populations

To compare the genetic diversity of native populations (American) and invasive populations (Chinese), we used all AFLP loci. The values were adjusted for variation in sample size through multiplication by 2n/(2n−1). To compare the genetic diversity of Chinese populations with that of American populations, 13 samples were randomly selected from the Chinese populations and calculated the mean and standard errors from 10 replicates. The means of three indices of genetic diversity (that is, percentage of polymorphic loci (P), mean gene diversity (h) and Shannon's index (I)) of Chinese population were a little higher than those of the American population (Table 3). The results suggest that there was no distinct difference in genetic diversity between the native and invasive populations.

Table 3 Comparison of genetic diversity of B. xylophilus identified with 2410 AFLP loci between American and Chinese populations
Full size table

Comparison of genetic diversity (P, h, I) among Chinese geographic populations

We also compared genetic diversities of geographic populations in China. We divided the Chinese pinewood nematode populations into four large geographic groups (Figure 1). These were southern populations (P1, including all samples from Guangdong), eastern seaboard populations (P2, including samples from the seaboard and island of Zhejing), eastern inland populations (P3, including all samples from Anhui and Jiangsu, and two sample from Zhejing) and western populations (P4, including samples from Congqing, Guizhou and Hubei). To reduce the errors occurring because of sample size, on the basis of the lowest number of samples among populations, we sampled the same number randomly from the other three populations and repeated the sampling 10 times. The results are shown in Table 4.

Table 4 Comparison of genetic diversities (P, h, I) of B. xylophilus among different geographic populations in China
Full size table

The results show that obvious differences in genetic diversity existed among geographic populations in China. The genetic diversity of the southern populations was distinctly higher than that of the other populations. The genetic diversity of the eastern seaboard populations was obviously lower than that of the eastern inland populations and western populations. The latter two had similar genetic diversity.

Comparison of genetic diversity (P, h, I) among groups with different invasive histories

We compared the genetic diversities of the five-time groups (identified above). Again, on the basis of the lowest number of samples among groups, we sampled the same number randomly from the other four groups and repeated sampling 10 times. The results are shown in Table 5.

Table 5 Comparison of genetic diversities (P, h, I) among groups of different invasive years of B. xylophilus in China
Full size table

The results show that, except for the earliest group (which occurred only in Nanjing in Jiangsu province), no significant difference existed among groups according to one-way analysis of variation (d.f.=3, 36; P: F=1.898, P=0.132; h: F=1.977, P=0.119; I: F=1.620, P=0.190). The results indicated that the genetic diversity of pinewood nematode populations did not obviously change during the invasive process. This means that the phenomenon of genetic diversity being lessened by genetic drift and founder effects was not apparent (Figure 2).

Figure 2
figure 2

Changes in genetic diversity (P, h, I) of B. xylophilus in groups of different invasive stages. P, percentage of loci polymorphism; h, Nei's gene diversity; I, Shannon's information index.

Full size image

Genetic relationship among samples of pinewood nematode

We carried out cluster analysis with PHYLIP 3.62 on the basis of Nei's unbiased genetic distance. We found similar topological structure using UPGMA and NJ methods (Figure 3).

Figure 3
figure 3

Dendrogram of 141 samples of B. xylophilus and B. mucronatus based on presence/absence of amplified fragment length polymorphism bands. Bootstrap values above 50% are shown next to corresponding nodes. Numbers above lines are support for tree by unweighted pair group method with arithmetic mean (UPGMA) method and numbers below lines are for tree by neighbor-joining (NJ) method.

Full size image

The dendrogram clusters the 141 into four groups: Bm group, US group, GD–AH–ZJ group and GD–JS group. Using one sample of B. mucronatus as an outgroup, seven samples of B. mucronatus clustered together and formed a group at the base of the tree. All samples of B. xylophilus clustered together, consistent with the status of B. xylophilus and B. mucronatus as distinct species. In the B. xylophilus group, the 13 American samples formed a distinct group. The Chinese samples were divided into two large groups. One was formed mainly by samples from Jiangsu. The other was formed mainly by samples from Zhejiang and Anhui (except samples from Mingguang, which clustered with Jiangsu). Guangdong samples were scattered across the two groups. Samples from Hubei, Guizhou and Congqing joined in the Jiangsu groups. The two Japanese samples clustered with the Zhejing–Anhui group.

Discussion

Change in genetic diversity of pinewood nematode populations during the invasive process

Many studies have shown that introduced species usually experience a loss of genetic variation during the invasive process, presumably because of genetic drift and founder effects (Tsutsui et al., 2000; Sakai et al., 2001). However, some species exhibit an increase in genetic diversity during a biological invasion (Kolbe et al., 2004). In this study, we compared the genetic diversity between native and invasive populations of pinewood nematode and among groups with different invasive histories. The results showed no obvious change in genetic diversity between native and invasive populations, nor among groups. This result may mean that the invasive process in pinewood nematode involves no founder effect or genetic bottleneck.

An explanation for this pattern may be that pinewood nematode introductions involved multiple sources and large quantities of the pest, which would allow considerable variation to be transferred from the source populations. This explanation is plausible, since a mass of materials carrying pinewood nematode and its vector insects are intercepted by Chinese customs inspection each year. The sustained rich genetic diversity may facilitate successful invasion.

Relationship between B. xylophilus and B. mucronatus

From the dendrogram (Figure 3), we can see that all samples of B. xylophilus and B. mucronatus clustered separately. This pattern supports the suggestion that the two species are distinct, even though they are very similar to each other in morphology and biology, and have been regarded as ‘a super species’ (de Guiran and Bruguier, 1989) or a pinewood nematode species complex (Webster et al., 1990; Abad et al., 1991).

Kanzaki and Futai (2002) studied the phylogenetic relationships of the B. xylophilus group by DNA sequence analysis. It was assumed that B. mucronatus originated in the eastern part of Asia and B. xylophilus originated from a population of B. mucronatus remaining in North America. In our study, samples from the United States formed a single cluster, at the base of the tree near the B. mucronatus group. This arrangement supports Kanzaki's assumption that B. xylophilus originated in North America (Kanzaki and Futai, 2002).

Source of Chinese invasive populations and possible spreading routes of pinewood nematode in China

Previously, it has been reported that the Japanese invasive population clearly originates from America (de Guiran and Bruguier, 1989; Tares et al., 1992; Iwahori et al., 2002). Nevertheless, where the Chinese invasive population came from, whether directly from the native US population or from other invasive populations adjoining China (such as Japan and Korea), is still in doubt. There are three speculations. One is that the Chinese pinewood nematode population was introduced from America. That theory would fit with the place where pinewood nematode and the disease were first found in China: on Zijingshan (the Purple Mountain), near where the Purple Mountain Observatory of the Chinese Academy of Sciences is located. In the early 1980s, large-scale instruments were imported from America for reconstruction of the observatory, and it has been proposed that pinewood nematode was introduced in the wooden packing materials from an American source. The second hypothesis is that the Chinese pinewood nematode population originated from Japan. That would fit with their similarities in pathogenicity (Li et al., 1983). The third hypothesis, that a part of Chinese population was from America and others from Japan, is suggested by the results of RAPD analysis (Zheng, 1998). In our study, owing to the limited sampling from the native populations and the other invasive populations outside China, we cannot be sure where the Chinese invasive population came from. However, the Chinese pinewood nematode populations are genetically closer to the Japanese populations than to the American populations, so it appears that at least part of the Chinese invasive populations was from Japan. Our results cannot show where the Jiangsu populations came from. The dendrogram does suggest that the Chinese populations comprise multiple invasions from different source populations.

The dendrogram divides the Chinese populations into at least two large groups, which can be interpreted as two invasion routes (Figure 4):

1.:

Guangdong → Anhui and Zhejiang

2.:

Guangdong → Jiangsu → Hubei, Chongqing, Guizhou and Mingguang (Anhui)

Figure 4
figure 4

Two possible major spread routes of B. xylophilus in China.

Full size image

Of course, we cannot exclude the other possible minor pathways—for example, that pinewood nematode was transmitted directly from Japan or the United States to somewhere in China.

Under the above hypothesis, Guangdong was the source region for the pinewood nematode invasion into China. This means that pinewood nematode was first introduced into Guangdong and then spread to Jiangsu, Anhui and Zhejiang. This speculation is in conflict with the time that pine wilt disease emerged in China. We propose two scenarios to explain this. One is that Guangdong was the first location of pinewood nematode invasion, even though pinewood nematode was first found in Nanjing in Jiangsu province. Since Zijingshan in Nanjing is a famous place of historic interest and scenic beauty, the tree damage and death would have rapidly attracted attention. Before that time, the pinewood nematode may have entered China and killed pine trees in other areas without being noticed. The supposition that the pinewood nematode infection first entered China in Guangdong is based, not only on molecular data, but also on the history of trade. Guangdong was the earliest open area in China, at the end of the 1970s, and it possesses the largest ports in China. Most imported cargos had to pass through Guangdong during the early stages of China's opening to foreign trade. Since the end of the 1970s, large numbers of electric instruments and cars from Japan have flooded into China, and it is possible that pinewood nematode was introduced from Japan accompanying the wooden wrappers and templates. Pinewood nematode was first transmitted to the economically developed areas, such as Jiangsu, Zhejing and Ma'anshan in Anhui, spread in these areas and was then transmitted to other areas. Moreover, pine wilt disease was epidemic in Hong Kong in the late 1970s (Corlett, 1999); the native pine trees (P. massoniana) in the Hong Kong region suffered extensively. Shenzhen in Guangdong is adjacent to Hong Kong, and pinewood nematode could have been transmitted from Hong Kong to Shenzhen by its vector insects (M. alternatus). It is, therefore, most likely that pinewood nematode was first introduced into Guangdong and then spread to other areas of China. Another scenario is that Zhijingshan in Nanjing was really the earliest location of the pinewood nematode invasion, but that the initial population was eliminated by the measure of clearing out dead trees, which has been implemented in every year since 1982 in that locality. The current population of Nanjing could be the result of repeated invasions because of the transmission of pinewood nematode between regions, and would therefore be different from the initial invasive population. Consequently, the initial invasive history would not be reflected by the dendrogram.

From the dendrogram, we can infer that a new spreading center has formed in Jiangsu. As Jiangsu is an economically developed region, frequent trade and heavy traffic may have accelerated the rates of transmission and spread of pinewood nematode. Some government projects, such as, electric power construction, may have especially accelerated the transmission of pinewood nematode from Jiangsu to other undeveloped regions, because the main electrical wires, cables and electrical materials are from Jiangsu province. Accompanying these materials was a large amount of wood, in the form of stocks and stow woods that were cast off on the way. Some of these wooden materials may contain large amounts of live pinewood nematodes. This traffic may be one of the reasons why the range of pinewood nematode is still expanding in China.

Our results imply that reinforcement of quarantine is essential to control the spread of pinewood nematode. This should include both international quarantine and especially domestic quarantine, which is largely ignored.