Introduction

Non-native plants often become invasive when introduced outside their native ranges (Pimentel 2002). There are a number of hypotheses that have been proposed to explain invasion success including preadaptation (Baker 1965; Parker et al. 2003; Richards et al. 2006), escape from natural enemies (Williams 1954; Keane and Crawley 2002), propagule pressure (Williamson 1996), empty niches (Elton 1958; MacArthur 1970), invasional meltdown (Simberloff and Von Holle 1999), evolution of increased competitive ability (Blossey and Nötzold 1995), novel weapons (Callaway and Ridenour 2004), and diversity-invisibility (Elton 1958). Recently, there is an increasing emphasis on post-introduction evolution as an important determinant of invasion success (Sakai et al. 2001; Lee 2002; Cox 2004; Prentis et al. 2008; Suarez and Tsutsui 2008). There is accumulating evidence that invasive plants can undergo rapid adaptive evolution in their new range including the evolution of latitudinal or altitudinal clines (Maron et al. 2004, 2007; Keller et al. 2009), increased phenotypic plasticity (Lavergne and Molofsky 2007), or other attributes that improve colonization or competitiveness with native species (eg. Jain and Martins 1979; Prati and Bossdorf 2004; Kliber and Eckert 2005; Dlugosch and Parker 2008a, 2008b; Seifert et al. 2009, Xu et al. 2010).

Hybridization between species or genetically distinct populations of the same species can be an important factor leading to evolutionary change and successful invasions (Arnold 1997; Ellstrand and Schierenbeck 2000; Rieseberg et al. 2007; Schierenbeck and Ellstrand 2009). For instance, hybridization may produce novel genotypes that have a selective advantage in the introduced range. This could arise through producing traits that are intermediate between the two parents, recombining parental traits, or by producing transgressive (extreme) traits, some of which might be favorable in the new environment (Ellstrand and Schierenbeck 2000; Rieseberg et al. 2007). Hybridization also may be particularly advantageous in exotic plant populations because there will be an initial lack of local adaptation. The benefits of hybridization such as heterosis or the production of novel genotypes may then outweigh the cost of losing gene combinations for local adaptation in the former native range (Verhoeven et al. 2011). There are now a number of examples of inter- and intra-specific hybridization that have led to invasiveness (Schierenbeck and Ellstrand 2009; Gaskin et al. 2009; Travis et al. 2010; Mukherjee et al., pers. comm.).

Brazilian peppertree (Schinus terebinthifolius Raddi, Anacardiaceae) was introduced into Florida from South America in the 19th century as an ornamental plant (Morton 1978). Based on herbarium records and available literature, it began to escape cultivation in the 1950s, and is currently one of Florida’s most invasive weeds (Schmitz et al. 1991; Cuda et al. 2004, 2006). The initial discovery of a naturalized population of Brazilian peppertree in the Florida Keys 50–60 years after it was introduced confirms the long lag period often exhibited by woody weeds before they become invasive (Kowarik 1995). Lag periods for exponential growth and naturalization of horticultural plants like Brazilian peppertree can be correlated with marketing time (Pemberton and Liu 2009), and/or may be a function of propagule pressure and evolutionary change after introduction (Sakai et al. 2001).

Historical records and genetic evidence indicate that two genetic lineages of Brazilian peppertree were established in Florida, USA, one in Miami on the east coast and a second near Punta Gorda on the west coast (Nerhling 1944; Morton 1978; Workman 1979). Since arriving, the distributions of these two genotypes have greatly expanded, and they have extensively hybridized (Williams et al. 2005, 2007).

Brazilian peppertree was initially targeted for biological control in Hawaii (Krauss 1963; Yoshioka and Markin 1991), and later in Florida, USA (Bennett et al. 1990; Cuda et al. 2006). Recent surveys for natural enemies in Brazil resulted in the discovery of several insects, specifically thrips and psyllids, that are highly adapted to the two Brazilian peppertree genotypes and/or to their hybrids established in Florida (Manrique et al. 2008; Christ 2010). More importantly, natural enemies belonging to the thrips genus Pseudophilothrips and the psyllid genus Calophya turned out to be a complex of cryptic species (Cuda et al. 2009; Mound et al. 2010; Burckhardt et al. 2011).

In this review, the ecological significance of local adaptation of these natural enemies is discussed in the context of the different Brazilian peppertree populations and their hybrids.

Genetic structure of Brazilian peppertree

Both chloroplast sequence and nuclear microsatellite data indicate there were two separate introductions of Brazilian peppertree into Florida, one on the west coast (the A chloroplast haplotype) and one on the east coast (the B chloroplast haplotype) (Williams et al. 2005, 2007). GIS mapping indicated that individuals with the highest ancestry were found close to the introduction sites and became progressively more admixed with increasing distance away from the introduction sites (Williams et al. 2007). Although we do not know when the two lineages began hybridizing, most Brazilian peppertrees in Florida are now hybrids although there are still individuals in the Miami and Punta Gorda areas that retain a high proportion of ancestral DNA (ancestry coefficient, q > 0.90) from the original introductions (Williams et al. 2007).

Brazilian peppertree has strong phylogeographic structure in its native range with most localities having only one or several closely related chloroplast haplotypes differing by one to two mutational differences (Williams et al. 2005). A parsimony network revealed that the introduced haplotypes A and B were different from other haplotypes in Brazil (Fig. 1). Furthermore, haplotypes A and B are from allopatric populations, indicating that the Florida introductions are from two distinct regions in South America. The origin of the Punta Gorda haplotype A is southeastern Brazil (Williams et al. 2005), whereas the origin of the Miami haplotype B is northeastern Brazil (Williams, unpublished).

Fig. 1
figure 1

CpDNA haplotype network of Schinus terebinthifolius illustrating relationships between the different haplotypes. Each connecting line indicates one nucleotide difference and unlabeled nodes are inferred intermediates. Figure redrawn from Williams et al. (2005)

Full size image

Genetic structure of natural enemies

Pseudophilothrips spp.

An extensive investigation of the genetic structure of P. ichini and a recently described cryptic species Pseudophilothrips gandolfoi Mound, Wheeler and Williams (Mound et al. 2010) was conducted using the mitochondrial cytochrome oxidase I gene (COI). Seven haplotypes were found in P. ichini from Bahia to Santa Catarina, Brazil, but only a single haplotype was found for the recently described P. gandolfoi (Fig. 2), which appears to be confined to more inland populations of Brazilian peppertree in the state of Paraná (Mound et al. 2010). P. gandolfoi, previously referred to as Pseudophilothrips sp. near ichini (Manrique et al. 2008), is almost always associated with Brazilian peppertree populations characterized by haplotypes C and D and has very low survival on populations characterized by haplotype A from Brazil (Manrique et al. 2008).

Fig. 2
figure 2

Relationships of Pseudophilothrips species inferred using the neighbor-joining method and K2P pairwise distances of mitochondrial COI sequences. Bootstrap values are shown next to the branches. The scale bar indicates the number of base substitutions per site. Figure from Mound et al. (2010); reprinted with permission

Full size image

Calophya spp.

To date, the genetic structure of only a small sample of C. terebinthifolii from Santa Catarina (n = 20) was investigated with the COI using methods similar to those reported in Mound et al. (2010). During a recent survey trip in March 2010, psyllids collected on Brazilian peppertree in northeastern Brazil (Salvador, Bahia) were identified as a new species, Calophya latiforceps Burckhardt, using both morphological and molecular characters (Burckhardt et al. 2011).

After sequencing this new species (n = 4 individuals) at the COI gene and calculating the Kimura 2-parameter (K2P) genetic distance both between and within the two psyllid species Kimura (1980), the results indicated that all individuals tested from southern Brazil have identical mitochondrial COI haplotypes and that the four psyllids sequenced from Salvador only differ at 0.2–0.7% of their sequence. However, C. latiforceps from Salvador, Bahia, was found to be genetically different from C. terebinthifolii collected in southern Brazil with a 13.4% sequence divergence. The morphological and genetic evidence confirmed the two psyllids are distinct species (Burckhardt et al. 2011).

Performance of natural enemies on Brazilian peppertree genotypes

Pseudophilothrips spp.

Performance (survival, development time, and adult longevity) of P. ichini and P. gandolfoi has been investigated in the laboratory (Manrique et al. 2008). P. ichini was originally collected on Brazilian peppertrees in the city of Ouro Preto, Minas Gerais, Brazil in November 2007. Peppertrees in this region are characterized by haplotype A. In contrast, P. gandolfoi was collected on Brazilian peppertrees in the vicinity of Curitiba, Paraná, Brazil, in January 2007. Peppertrees here carry either haplotype C or D. Ouro Preto is located 830 km northeast of Curitiba.

The two Pseudophilothrips spp. differed in their ability to accept the Florida populations as their host plants. For instance, P. gandolfoi exhibited low survival (0–4%) and short adult longevity (<ten days) when reared on the original Florida populations characterized by haplotypes A and B, or their hybrids between the original invasive populations, whereas higher survival (~50%) and longevity (~30 days) were observed for P. ichini on these same haplotypes (Manrique et al. 2008).

Calophya spp.

Performance (% rearing success) of C. terebinthifolii was investigated on five native haplotypes of Brazilian peppertree in a laboratory study conducted in Brazil (Christ 2010). Calophya terebinthifolii performed significantly better on plants with haplotype A, which occur in Florida, than the other populations characterized by haplotypes O, D, K, and M (G-test, G = 7.63; P < 0.01) (Christ 2010). There was over a 75% success rate when this psyllid was raised on its natal plants with haplotype A, a 20% success rate on plants with haplotype O, and 0% success on plants characterized by haplotypes D, K, and M (Fig. 3). Because haplotypes M and O are only one base-pair different from haplotype K (D.A. Williams, unpublished data), these data suggest that populations of C. terebinthifolii also are highly adapted to specific Brazilian peppertree haplotypes. The performance of the newly described C. latiforceps from Salvador has not yet been tested on the different Brazilian peppertree haplotypes.

Fig. 3
figure 3

Performance of Calophya terebinthifolii on five different haplotypes of Brazilian peppertree. Data from Christ (2010). Sample sizes for haplotypes were A (n = 12), O (n = 5), K (n = 2), D (n = 2), and M (n = 1)

Full size image

Conclusions

Most of Florida’s Brazilian peppertrees are the result of intraspecific hybridization between two introductions (haplotypes A and B) from distinct source regions in Brazil (Williams et al. 2005, 2007; Mukherjee et al., pers. comm.). Common garden studies recently conducted in Florida suggest that these hybrid individuals have higher growth rates than the parental types, which may have facilitated the invasion (Geiger et al. 2011). Our research on the genetics and performance of two thrips and two psyllid natural enemies of Brazilian peppertree has shown that these insects exhibit what Harley and Forno (1992) referred to as ‘fine tuned’ adaptation to specific populations and genotypes of their host plants. On-going studies have not yet revealed a negative effect of hybridization per se on the performance of biological control agents. However, our studies indicate that it may be necessary to match biocontrol agents with the specific Brazilian peppertree geographic populations and/or their hybrids that occur in the introduced range.

Successful invasive species often are introduced multiple times from distinct source regions in the native range, so intraspecific hybridization may be quite common (Dlugosch and Parker 2008a, 2008b; Schierenbeck and Ellstrand 2009). Several studies also have shown that some pathogens and herbivores can be adapted to specific genotypes or populations of their host plants (Hasan 1972; Karban 1989; Karban and Strauss 1994; Gaskin and Schaal 2002; Goolsby et al. 2006; Ayres et al. 2009; Gaskin and Kazmer 2009). Our studies on the Brazilian peppertree invasion in Florida and fine scale adaptation of some of its natural enemies provide further evidence of this phenomenon. Future biological control programs could benefit from population genetic studies on both the invasive species and potential biological control agents, especially when the specific origin(s) of the weed are in doubt or there is evidence of differential attack by natural enemies.