• covariation;
  • defense chemistry;
  • hybridization;
  • secondary metabolites;
  • transgressive segregation


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
  • Hybridization can lead to novel qualitative or quantitative variation of secondary metabolite (SM) expression that can have ecological and evolutionary consequences.
  • We measured pyrrolizidine alkaloid (PA) expression in the shoots and roots of a family including one Jacobaea vulgaris genotype and one Jacobaea aquatica genotype (parental genotypes), two F1 hybrid genotypes, and 102 F2 hybrid genotypes using liquid chromatography–tandem mass spectrometry (LC-MS/MS).
  • We detected 37 PAs in the roots and shoots of J. vulgaris, J. aquatica and the hybrids. PA concentrations and compositions differed between genotypes, and between roots and shoots. Three otosenine-like PAs that only occurred in the shoots of parental genotypes were present in the roots of F2 hybrids; PA compositions were sometimes novel in F2 hybrids compared with parental genotypes, and in some cases transgressive PA expression occurred. We also found that PAs from within structural groups covaried both in the roots and in the shoots, and that PA expression was correlated between shoots and roots.
  • Considerable and novel variation present among F2 hybrids indicates that hybridization has a potential role in the evolution of PA diversity in the genus Jacobaea, and this hybrid system is useful for studying the genetic control of PA expression.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

The role of hybridization in evolutionary processes, including the generation of novel traits, introgression of traits between species, and even speciation, has received widespread attention (Stebbins, 1959; Arnold, 1992; Rieseberg & Carney, 1998; Abbott et al., 2009). In recent years it has become apparent that hybridization can lead to the generation of novel molecular and morphological phenotypes (Rieseberg et al., 2003; Kim et al., 2008). Such phenotypes can persist over evolutionary time and can even lead to speciation among hybrid lineages (Seehausen, 2004; Soltis & Soltis, 2009). At the metabolic level, hybridization can impact the diversity of secondary metabolites (SMs) in plants (Orians, 2000). SMs are important for mediating interactions between plants and their environment (Iriti & Faoro, 2009), and the composition of plant SMs can play a role in determining the evolutionary success of populations and species (e.g. Burow et al., 2010).

In the first (or F1) hybrid generation, most phytochemicals are expressed at concentrations that are either similar to those of one of the parents or intermediate between those of the two parents (Orians, 2000). However, recombination in F2 and later generation hybrids is expected to increase variation in phytochemical expression among different F2 genotypes. Transgressive segregation can occur, such that some F2 genotypes may vary outside the range observed in parental genotypes, and provide key variation upon which selection can act during the process of adaptation (Rieseberg et al., 1999, 2007). One of the drawbacks of many studies that quantify SM expression by hybrids is that only mean values are reported for each hybrid class (i.e. F1, F2, or backcross; e.g. Hallgren et al., 2003; O’Reilly-Wapstra et al., 2005). When genotypes are pooled within classes, transgressive phenotypes may not be identified. Also, many studies fail to carry out replicate measurements on genotypes within parental or hybrid classes, and such studies therefore fail to measure and test for genetically controlled variation in SM expression within these classes. In this study, we investigated variation among > 100 replicated F2 hybrids, which allowed us to conduct appropriate statistical testing to identify differences between and among hybrid and parental genotypes.

SM accumulation can be influenced by a number of factors, including genetics, abiotic factors (such as nutrient and light availability), biotic factors (including competition, herbivory and disease), and interactions between these factors (Lankau & Kliebenstein, 2009; Kirk et al., 2010). However, little is known about the mechanisms behind these complex regulatory systems. Recent work on the genomics and ecology of model and nonmodel species has started to shed light on the control of SM expression. For example, studies of glucosinolate expression in Arabidopsis thaliana have identified four major genetic loci responsible for the expression of 14 different glucosinolates (Kliebenstein, 2009). In addition to the regulatory complexity within individuals, there is considerable variation in SM profiles both within and among plant populations (e.g. Burow et al., 2010). Furthermore, more attention has been paid to SMs in above-ground plant parts than in below-ground plant parts, even though the latter are probably equally important to a species’ ecology, and there is often interaction or coordination between the expression of SMs in above-ground plant tissues and that in below-ground plant tissues (van Dam et al., 2009).

Species in the genus Jacobaea (syn. Senecio, Asteraceae) have been used to investigate the evolutionary basis of SM diversity in plants, because they contain a diverse but structurally related group of alkaloids that play a role in biotic interactions (e.g. Hartmann, 1999; Hol & van Veen, 2002; Macel & Vrieling, 2003; Macel et al., 2005; Kowalchuk et al., 2006). Twenty-six pyrrolizidine alkaloids (PAs) have been reported from 24 species of Senecio sect. Jacobaea (Pelser et al., 2005), although the recent development of more sensitive analytical methods has allowed the detection of a greater number of structural PA variants in the same species (Joosten et al., 2009, 2010, 2011). In Jacobaea species, all PAs except for senecivernine are derived from senecionine N-oxide. Senecionine N-oxide is synthesized in the roots, transported to the shoots via the phloem, and diversified into other PA structures (Hartmann & Toppel, 1987; Sander & Hartmann, 1989; Hartmann et al., 1989). Structurally derived PAs are thought to be produced from the precursor senecionine N-oxide via a limited number of steps (Hartmann & Dierich, 1998; see a schematic diagram representing putative PA biosynthetic pathways in Fig. S2). Apart from structural diversification, PAs do not undergo any turnover or degradation (Sander & Hartmann, 1989; Hartmann & Dierich, 1998). PAs can occur in plants in two forms: tertiary amine (free base) and N-oxide (Rizk, 1991; Wiedenfeld et al., 2008; Joosten et al., 2011). The proportion of tertiary amine is different among PAs and between genotypes. In Jacobaea plants, the tertiary amine form is usually present among higher proportions in jacobine-like PAs than among senecionine-like and erucifoline-like PAs. However, the mechanisms by which one form is converted to the other are not well understood (Joosten et al., 2011).

PA composition and concentration vary greatly among and within Jacobaea species (Witte et al., 1992; Macel et al., 2002, 2004; Pelser et al., 2005). Four different PA chemotypes of Jacobaea vulgaris are reported to occur; these include jacobine, erucifoline, mixed and senecionine chemotypes (Witte et al., 1992; Macel et al., 2004). Field studies and controlled bioassays that incorporate herbivores indicate that plant resistance to herbivorous invertebrates is correlated with plant PA concentration and composition (Leiss et al., 2009; Macel & Klinkhamer, 2010). Individual PAs have different deterrent effects on generalist herbivores (Macel et al., 2005), and also have different stimulatory effects on the oviposition of the specialist herbivore Tyria jacobaeae (the cinnabar moth; Macel & Vrieling, 2003). Furthermore, free base PAs appear to have different effects on generalist herbivores compared with their corresponding N-oxides (van Dam et al., 1995; Macel et al., 2005). These cumulative findings indicate that PA diversity is ecologically important with respect to interactions between plants and herbivores.

Interspecific hybridization is widespread in the Senecio genus, including section Jacobaea (e.g. Vincent, 1996). For example, hybridization between Senecio squalidus and Senecio vulgaris led to the origin of three new fertile hybrid taxa, and S. squalidus itself is a hybrid species resulting from a cross between Senecio aethnensis and Senecio chrysanthemifolius (Abbott & Lowe, 2004; James & Abbott, 2005; Abbott et al., 2009). There are many other well-documented cases of hybridization between Senecio species (e.g. Beck et al., 1992; Hodalova, 2002; Lopez et al., 2008), including natural hybridization between J. vulgaris (formerly Senecio jacobaea L.) and J. aquatica (formerly Senecio aquaticus L.) which occurs in The Zwanenwater Nature Reserve in the Netherlands (Kirk et al., 2004).

Jacobaea vulgaris (tansy ragwort or common ragwort) is native to Europe and west Asia but is invasive in North America, Australia and New Zealand. Jacobaea aquatica (marsh ragwort) is closely related to, but not a sister species of, J. vulgaris (Pelser et al., 2003). The two species are ecologically distinct. Jacobaea vulgaris often occurs in dry, sandy soil with little organic matter and J. aquatica is found in wet habitats in soils that are high in organic matter. The two species are attacked by different guilds of herbivorous insects in the field. Different susceptibility to a generalist herbivore has been observed (Kirk et al., 2004, 2010). Putative hybrids from the Zwanenwater (the Netherlands), initially identified in 1979 based on highly variable and usually intermediate flower and leaf lobe morphology compared with J. vulgaris and J. aquatica, were confirmed to be hybrids between these two species using molecular genetic markers and PA composition (Kirk et al., 2004). The natural hybrid population is highly backcrossed with J. vulgaris, and F1 hybrids are uncommon in the natural population (Kirk et al., 2004, 2005). In contrast to J. vulgaris, J. aquatica lacks jacobine-like PAs but is rich in senecionine-like PAs (Kirk et al., 2010). A previous study that characterized PA composition of natural hybrids and artificial F1 hybrids of the two species showed that PA expression was affected by species and environment interactions (Kirk et al., 2010).

To obtain a hybrid family, we selected a J. vulgaris genotype of the jacobine chemotype, which is rich in jacobine-like PAs, and a J. aquatica genotype. We established an artificial J. vulgaris × J. aquatic family, which includes two parental genotypes, two F1 hybrids, and c. 100 different F2 hybrid genotypes. These are all kept in tissue culture and can be reproduced at length. The hybrid system to a great extent overcomes the problem of unavailability of the relevant pure PAs for the study of the effects of individual alkaloids or PA combinations. Joosten et al. (2011) demonstrated the genetically controlled presence of tertiary PAs in Jacobaea species using this system. Kirk et al. (2011) reported transgressive segregation of primary and secondary metabolites in the F2 hybrids of this cross using NMR-based metabolomics.

In this study, we aimed to investigate whether hybridization can generate new PA variation in this system and to gain an initial understanding of how PA accumulation is genetically regulated based on the pattern of PA variation. We focused on differences in PA expression among segregating hybrids originating from a single cross between two parental genotypes, and we grew plants under standard conditions to eliminate the effect of environment on PA expression. The methods used in this study differed from those used in previous work in two respects: first, the large numbers of genotypes and replications resulted in a very large sample size; second, we measured PAs by LC-MS/MS, which is highly sensitive and can detect the two forms of PAs simultaneously (Joosten et al., 2011). We addressed the following questions: Do F2 hybrids produce novel PAs? Does any F2 hybrid genotype show evidence of transgressive variation (over-expression or under-expression) with regard to the concentrations of total PA, a structural group of PAs, or any individual PAs? Does hybridization produce novel PA compositions among F2 genotypes? Is there covariation in the expression of individual PAs? Are there correlations between the accumulation of PAs in the roots and that in the shoots?

Materials and Methods

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Study system

Jacobaea vulgaris Gaertn. subs. dunensis, Jacobaea aquatica (Hill) Gaertn. et al. subs. aquatica (parental species, parents), and F1 and F2 hybrids of these species were used in this study. Jacobaea vulgaris seeds were collected at Meijendel Nature Reserve (52°7′54′′N, 4°19′46′′E, the Netherlands), and J. aquatica seeds were collected at the Zwanenwater Reserve (52°48′38′′N, 4°41′7′′E, the Netherlands). Seeds of the two species were sterilized, were germinated in glass vials, and were maintained in tissue culture. Replicate genotypes (clones) from each parental species were subsequently grown in pots in climate rooms (humidity 70%; light 16 h at 20°C; dark 8 h at 20°C). Before blooming, the potted plants were kept in a cold room (humidity 70%; light 8 h at 4°C; dark 16 h at 4°C) for c. 10 wk to produce vernalization. Crosses were performed by rubbing flower heads together (both species are self-incompatible; Kirk et al., 2005, 2010). Two rayed F1 offspring were selected from this initial cross, and were reciprocally crossed with each other to produce two sets of offspring. A number of F1 crosses were made, and we selected the family that produced the greatest number of viable F2 genotypes. From the selected F1 cross, we obtained one set of 56 F2 individuals, and a second set (from the reciprocal cross) of 46 F2 individuals. The parental, F1 and F2 individuals were maintained in tissue culture and were cloned in order to obtain replicate genotypes for the experiments described here. These cloned individuals are referred to as ‘genotypes’ hereafter. The hybrid status of F1 and F2 individuals used in this study was confirmed using amplified fragment length polymorphism (AFLP) and single nucleotide polymorphism (SNP) markers (K. Vrieling et al., unpublished data).

Plant growth

We aimed to use six cloned replicates per F2 genotype and c. 12 cloned replicates per parental and F1 genotype; however, a few plants died or grew poorly in tissue culture, and were therefore not included in the experiment. Plants were propagated by tissue culture and were potted in 1.3-l pots filled with 95% sandy soil (collected from Meijendel), 5% potting soil (Slingerland Potgrond, Zoeterwoude, the Netherlands) and 1.5 g l−1 Osmocote slow release fertilizer (Scott®, Scotts Miracle-Gro, Marysville, Ohio, USA; N : P : K = 15 : 9 : 11). Plants were kept in a climate room for 6 wk (humidity 70%; light 16 h at 20°C; dark 8 h at 20°C). In total, we grew > 600 individual plants including replicates of the two parental, two F1 and 102 F2 genotypes.

Plant harvesting

Plants were harvested after 6 wk. Whole plants were gently removed from the potting medium. Shoots were separated from roots with scissors just above the root crown, and roots were rinsed with water. Roots and shoots from each plant were immediately wrapped in a piece of aluminum foil and kept in a cooler with liquid nitrogen until harvesting was completed, and then were stored at −80°C until freeze-drying. In total, we harvested the shoots and roots from 609 plants. Each parental and F1 hybrid genotype was replicated 11 or 12 times. F2 hybrid genotypes were replicated three to six times. In most cases there were six replicates per F2 genotype; however, in a few cases some replicates were lost as a consequence of plant death or poor growth. Samples were freeze-dried for 1 wk under vacuum with a collector temperature of −55°C (12-liter Freeze Dry System, Labconco Free Zone®, Labconco Corporation, Kansas City, Missouri, USA). The dry weights of shoots and roots were measured, and plants were ground into fine powder and stored at −20°C until PA extraction.

Pyrrolizidine alkaloid extraction and analysis

Approximately 10 mg of powdered plant material was extracted with 1 ml of 2% formic acid. Heliotrine, monocrotaline and monocrotaline N-oxide were added as internal standards to the extraction solvent at a concentration of 1 μg ml−1. The plant extract solution was shaken for 30 min. Solid plant material was removed by centrifugation at 720 × g for 10 min and filtered through a 0.2-μm nylon membrane (Acrodisc 13-mm syringe filter, Pall Life Sciences, Ann Arbor, MI, USA). An aliquot of the filtered solution (25 μl) was diluted with water (975 μl) and injected into the LC-MS/MS system.

A Waters Acquity ultra performance liquid chromatographic (UPLC) system coupled to a Waters Quattro Premier XE tandem mass spectrometer (Waters, Milford, MA, USA) was used for PA analysis. Chromatographic separation was achieved on a Waters Acquity BEH C18 150 × 2.1 mm, 1.7 μm UPLC column, run with a water/acetonitrile linear gradient containing 6.5 mM ammonia at a flow of 0.4 ml min−1. The gradient started at 100% water and during analysis the acetonitrile percentage was raised in 12 min to 50%. The column was kept at 50°C and the injection volume was 10 μl. The MS system was operated in positive electrospray mode. Data were recorded in multiple monitoring mode (MRM) using two selected precursor ions to product ion transitions per compound. Cone energy was 40 V and collision energy settings were optimized for the individual compounds. In Table 1 an overview is given of the mass spectrometric settings used for the detection of the relevant PAs. The samples were run in a randomized order divided over five series. For each compound the sum of the two peak areas was normalized against the peak area of the internal standard heliotrine. Quantification was performed against a standard solution (100 μg l−1) of the PAs in a diluted extract of Tanacetum vulgare (tansy). The extract of T. vulgare material was prepared in the same way as the other extracts and was used to mimic a PA-free plant extract. The standard solution was injected every 30 samples, and the averaged response of each compound was used for quantification. Seventeen individual PA standards (see Joosten et al., 2011, for the source of the standards) were available for this study, representing over 80% of the total amount of PAs present in the majority of plants extracts. For those compounds for which no reference standard was available, a semiquantitative (indicative) value could be obtained by comparison with the most closely related analogue (e.g. an isomer). Identification of these PAs was based on their retention time, molecular mass and fragmentation pattern and on comparison with PA standards and/or literature data. Data processing was conducted with Masslynx 4.1 (Waters Corporation, Milford, MA, USA).

Table 1.   Pyrrolizidine alkaloids (PAs) detected in Jacobaea aquatica, Jacobaea vulgaris and hybrids
GroupPACodeRetention time (min)Precursor mass (m/z)Fragment mass 1; 2 (m/z)Collision energy 1; 2 (eV)Standard used for quantification
  1. Retention times and selected mass spectrometric conditions are given.

Senecionine-like PAs (simple senecionine-related derivatives)Senecioninesn9.93336.294.0; 120.040; 30sn
Senecionine N-oxidesnox6.97352.294.0; 120.040; 30snox
Integerrimineir9.72336.294.0; 120.040; 30ir
Integerrimine N-oxideirox6.83352.294.0; 120.040; 30irox
Retrorsinert8.49352.294.0; 120.040; 30rt
Retrorsine N-oxidertox6.01368.294.0; 120.040; 30rtox
Usaramineus8.29352.294.0; 120.040; 30rt
Usaramine N-oxideusox5.89368.294.0; 120.040; 30rtox
Riddelliinerd7.91350.294.0; 138.040; 30rd
Riddelliine N-oxiderdox5.48366.294.0; 118.040; 30rdox
Seneciphyllinesp9.16334.294.0; 120.040; 30sp
Seneciphylline N-oxidespox6.36350.294.0; 138.040; 30spox
Spartioidinest8.96334.2120.0; 138.030; 30sp
Spartioidine N-oxidestox6.36350.294.0; 138.040; 30spox
Acetylseneciphyllineacsp11.80376.2120.0; 138.030; 30acsp
Acetylseneciphylline N-oxideacspox8.86392.294.0; 118.040; 30acspox
Seneciverninesv10.09336.294.0; 120.040; 30ir
Jacobine-like PAs (jacobine-related derivatives)Jacobinejb7.89352.2120.0; 155.030; 30jb
Jacobine N-oxidejbox5.49368.2120.0; 296.030; 25jbox
Jacolinejl6.13370.294.0; 138.040; 30jb
Jacoline N-oxidejlox4.39386.294.0; 120.040; 30jbox
Jaconinejn8.75388.294.0; 120.040; 30jb
Jaconine N-oxidejnox5.77404.294.0; 138.040; 30jbox
Jacozinejz7.23350.294.0; 138.040; 30jb
Jacozine N-oxidejzox5.11366.294.0; 118.040; 30jbox
Dehydrojaconinedhjn7.86386.294.0; 120.040; 30jb
Erucifoline-like PAs (erucifoline-related derivatives)Erucifolineer7.56350.294.0; 120.040; 30er
Erucifoline N-oxideerox4.80366.294.0; 118.040; 30erox
Acetylerucifolineacer10.18392.294.0; 118.040; 30er
Acetylerucifoline N-oxideacerox7.17408.294.0; 120.040; 30erox
Otosenine-like PAs (otosenine-related derivatives)Senkirkinesk7.31366.2122.0; 168.030; 25sk
Otosenineot5.60382.2122.0; 168.030; 25sk
Onetineone4.35400.2122.0; 168.030; 30sk
Desacetyldoroninedesdor6.26418.2122.0; 168.030; 30sk
Floroseninefs8.35424.2122.0; 168.035; 30sk
Floridaninefd6.79442.2122.0; 168.030; 30sk
Doroninedor9.01460.2122.0; 168.030; 30sk

Data analysis

We checked for maternal effects on both quantitative and qualitative variation with regard to F2 genotypes from different maternal F1 parents within the reciprocal cross (data not shown). As no significant maternal effects were found, F2 genotypes from both maternal parents were pooled for the analysis.

Analysis of PA qualitative variation  The genotype-dependent presence of florosenine, floridanine and doronine in the roots and shoots was tested using binomial general linear models in which PA concentration values were coded as either 0 (absent) or 1 (present) and genotype was designated as the fixed factor. We carried out qualitative analyses incorporating these three PAs because they were the only PAs that were absent in some samples. All other PAs were always present.

Analysis of PA quantitative variation  We classified the PAs identified in this study into four types according to their structural characteristics and biosynthetic pathways (see Figs S1, S2; Pelser et al., 2005): senecionine-like PAs, jacobine-like PAs, erucifoline-like PAs and otosenine-like PAs (Table 1). Senecivernine and senkirkine were not grouped with any other PAs by Pelser et al. (2005). However, based on the experimental data obtained in our PA measurements, senecivernine expression was closely correlated with the expression of senecionine-like PAs, and senkirkine expression was similarly correlated with that of otosenine-like PAs. Senecivernine and senkirkine were therefore grouped, respectively, with senecionine-like PAs and otosenine-like PAs for the purposes of analysis.

We used ANOVAs to determine whether PA quantities in roots and/or shoots were dependent on genotype. We defined each PA as a separate dependent variable. We also used ANOVAs to determine whether the four structural groups of PAs, free bases, N-oxides, and total PA were dependent on genotype. The data were log-transformed. We tested for normal distribution and homogeneity of the variance using the residuals from the models. Differences between the hybrids and parental genotypes were evaluated from the data in regression coefficient matrices of the models. In each matrix, the estimated coefficient of a hybrid indicated whether it had a lower or higher amount of PA than one of the parents, and the P-value showed whether the difference was significant (Crawley, 2005). The hybrids were compared with each of the two parents separately.

There were a number of variables (see details in Table S3) that did not meet the assumptions for a linear model. We examined among-genotype differences in these variables using Kruskal–Wallis tests for which PA concentrations were defined as independent variables and genotype was defined as the factor. The data were log-transformed to achieve homogeneity of the variance among genotypes. Differences between hybrid and parental genotypes were evaluated using multiple comparisons after Kruskal–Wallis tests, for which either of the parents was defined as the control (Giraudoux, 2010).

The type of quantitative PA variation (in hybrids compared with parents) was classified as follows: under-expression (U; concentration in the hybrid significantly less than that in both parents); dominant to the parent with lower expression (Dl; concentration in the hybrid not different from that in the parent with lower expression and significantly different from that in the other parent); intermediate to the parents (Im; concentration in the hybrid intermediate to but significantly different from those in the two parents); dominant to the parent with higher expression (Dh; concentration in the hybrid not different from that in the parent with higher expression and significantly different from that in the other parent); over-expression (O; concentration in the hybrid significantly greater than those in both parents); not different from the parents (ND; concentration in the hybrid not significantly different from those in either parent).

Analysis of PA composition  Differences in PA composition were evaluated using relative concentrations of individual PAs. The relative concentration was calculated as follows: (absolute concentration of an individual PA or a group of PAs)/(total PA concentration) × 100. The relative concentration data were not normally distributed and the variances among the genotypes were not homogeneous. We therefore tested for differences in relative PA concentration among genotypes using Kruskal–Wallis tests and nonparametric multiple comparisons (Giraudoux, 2010).

Differences in PA composition among genotypes and between the shoots and roots were evaluated using an Adonis test, which is a nonparametric MANOVA (Oksanen et al., 2010). Genotype and plant part (shoots or roots) were defined as factor variables. We visualized variation in PA composition using a nonmetric multidimensional scaling (NMDS) method, which is analogous to principal components analysis (PCA) or multidimensional scaling (MDS) but without distribution assumptions (Goslee & Urban, 2007). As in a PCA or MDS plot, each point in the NMDS plot represents an individual sample, and points that are close together indicate that those samples have similar PA compositions. NMDS can avoid the arch and compressed pattern that occurs in PCA when data include samples that have few components in common (Quinn & Keough, 2002).

Cluster and correlation analysis  A hierarchical cluster analysis of individual PAs in shoots and roots was carried out to identify similarities in the expression of different PAs. The data used in this analysis were log-transformed absolute PA concentrations. The hierarchical cluster analysis was carried out using the likelihood linkage analysis method (Kojadinovic, 2010). We tested for correlations between PA concentrations in the shoots and roots using Spearman correlation tests (on absolute concentrations). P-values were adjusted for multiple comparisons using sequential Bonferroni methods.

All analyses were conducted in R version 2.10.0 (R Development Core Team, 2009).


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

PA qualitative variation

In total, we detected 37 PAs in the shoots and roots of the parents and the F1 and F2 hybrids. We classified each PA into one of four structural groups: senecionine-like PAs, jacobine-like PAs, erucifoline-like PAs or otosenine-like PAs (Table 1). Otosenine-like PAs do not occur as N-oxides. PAs of other types were present and detected in both forms, except for dehydrojaconine and senecivernine, which were only detected in the free base form.

Most parental PAs were always present in the offspring, although some only in trace amounts (< 0.1 μg g−1 DW). Three PAs, florosenine, floridanine and doronine, were present in J. aquatica shoots, but were absent in J. vulgaris shoots and were absent (or present in trace amounts) in the roots of both parents. These three PAs were present in the shoots and roots of the two F1 hybrids. They were absent in the shoots and/or roots of some F2 genotypes, but were present in much higher concentrations in some F2 plants compared with the parents (Tables 2, S1, S2). The presence of all three of these PAs was genotype-dependent both in the shoots and in the roots (shoots and roots tested separately, in all cases: df = 105; χ2 > 600, < 0.01).

Table 2.   Qualitative variation of three otosenine-like pyrrolizidine alkaloids (PAs) in the roots and shoots of two F1 and 102 F2 hybrids between Jacobaea aquatica and Jacobaea vulgaris
PAsJ. aquaticaJ. vulgarisF1-AF1-BF2
  1. All other PAs reported in this study were always present in parents, F1 hybrids, and F2 hybrids. Numbers indicate the number of F2 genotypes in which a particular PA was absent or present. If a certain PA was present in the roots or shoots of a single replicate, we scored that PA as present in that genotype. If the PA was not found in any of the replicates, it was regarded as absent in the genotype. Trace indicates concentrations < 0.1 μg g−1 DW.


PA quantitative variation

We analysed quantitative variation in the concentration of 34 individual PAs (excluding florosenine, floridanine, and doronine), the sum concentrations of the four PA groups, the sum concentration of free bases and N-oxides, and total PA concentration. All variables were genotype-dependent (ANOVA or Kruskal–Wallis test; separately for shoots and roots; in all cases: df = 105; < 0.01).

Jacobaea aquatica had a lower total PA concentration than J. vulgaris in shoots. Both of the F1 genotypes were intermediate to the parents. F2 genotypes were also, on average, intermediate to the parents. However, a 20-fold difference in genotypic mean total PA concentration (334.0–6835.0 μg g−1 DW) was observed among F2 hybrid genotypes (Fig. 1 and Table S1).


Figure 1. Frequency distribution of genotypic mean concentrations (μg g−1 DW) of total pyrrolizidine alkaloid (PA), senecionine N-oxide, jacobine N-oxide, erucifoline N-oxide and otosenine in the shoots and roots of the 102 F2 hybrid genotypes between Jacobaea aquatica and Jacobaea vulgaris. The positions of the symbols above the bars indicate genotypic mean values for the two parental genotypes and the two F1 genotypes. Upward triangle, J. aquatic; downward triangle, J. vulgaris; rectangle, F1-A; diamond, F1-B. The genotypic mean concentration is the average value of the three to six replicates from the same genotype.

Download figure to PowerPoint

There was also great variation in the quantities of particular groups of PAs and individual PAs (Figs 1, S3 and Table S1). In F2 hybrid shoots, transgressive segregation (statistically significant under-expression or over-expression) of PA expression occurred in 7.5% of cases for concentrations of individual PAs and also in 7.5% of cases for concentrations of PA groups or total PA concentration (Fig. 2 and Table S3). Among the F2 hybrids, 14 genotypes had significantly lower total PA concentration compared with the parents, and no F2 genotypes had significantly higher total PA concentration. Otosenine-like PAs (group sum) were over-expressed in the shoot of one F2 hybrid genotype, as a result of the over-expression of desacetyldoronine and otosenine. Over-expression of erucifoline-like PAs (group sum), erucifoline, and its N-oxide was observed in some F2 hybrids. Over-expression of several minor PAs, including riddelliine, riddelliine N-oxide and jacozine N-oxide, occurred in a few F2 genotypes (Fig. 2 and Table S3).


Figure 2. Classification of pyrrolizidine alkaloid (PA) quantitative variation in the shoots and roots of Jacobaea aquatica, Jacobaea vulgaris, two F1 hybrids (closed bars) and 102 F2 hybrids (open bars) relative to the parental genotypes. Hybrid genotypes were classified into six types according to expression of an individual PA, group of PAs or total PA: U, under-expression; significantly less than those of both parents; Dl, dominant to the parent with lower expression; not different from that of the parent with lower expression and significantly different from that of the other parent; Im, intermediate to the parents; intermediate to the parents but significantly different from both parents; Dh, dominant to the parent with higher expression; not different from that of the parent with higher expression and significantly different from that of the other parent; O, over-expression; significantly greater than those of both parents; ND, not significantly different from those of the parents. The graphs show percentage of hybrids divided over the different types. See details in Supporting Information, Table S3.

Download figure to PowerPoint

Similar patterns of PA expression variation were found in hybrid roots. Extremely high or low concentrations of individual PAs only occurred in 6.2% of all tests. Some minor PAs such as retrorsine, retrorsine N-oxide, riddelliine, seneciphylline, acetylerucifoline and acetylerucifoline N-oxide were over-expressed in a few F2 genotypes. Transgressive concentrations of PA groups and transgressive total PA concentration were rarer (only 0.7% across tests including PA groups and total PA concentration) in the roots compared with the shoots (Fig. 2 and Table S3).

Variation in PA composition

PA composition differed in the shoots of the two parental genotypes. Senecionine-like PAs were dominant in J. aquatica, and jacobine-like PAs were dominant in J. vulgaris. In the roots of J. aquatica, > 96% of the total PA belonged to the senecionine group. In contrast to the shoots, senecionine-like PAs were also dominant in the roots of J. vulgaris, and comprised c. 60% of the total PA, while jacobine-like PAs comprised c. 30% and otosenine-like PAs comprised 5%. Erucifoline-like PAs were found only in low concentrations (Fig. 3a–d).


Figure 3. Relative concentrations of major pyrrolizidine alkaloids (PAs) in the shoots and roots of Jacobaea aquatica, Jacobaea vulgaris, and F1 and F2 hybrids. Relative concentrations represent the percentages of total PA concentration in a sample. The PAs shown in the graphs are the 10 PAs with the highest relative concentrations across all samples. Error bars are standard errors. The graph for F2 is based on the mean relative concentrations of individual PAs for all samples of the F2 genotypes and the other graphs represent individual samples from the same genotype. Jacobaea aquatica, one genotype, 12 replicates; Jacobaea vulgaris, one genotype, 12 replicates; F1-A, one genotype, 11 replicates; F1-B, one genotype, 12 replicates; F2, 102 genotypes, three to six replicates per genotype. Abbreviations for PAs are defined in Table 1.

Download figure to PowerPoint

The shoots of the two F1 hybrids showed a mixed pattern compared with the parents; concentrations of senecionine-like and jacobine-like PAs were approximately equal. The roots of F1 hybrids contained a greater variety of PAs than those of J. aquatica. They contained > 10% jacobine-like PAs, and also contained some other PAs including erucifoline and otosenine. However, the relative concentration of senecionine-like PAs remained high at c. 80% or more (Fig. 3e–h). The shoots and roots of F2 hybrids on average showed patterns similar to those of the F1 hybrids (Fig. 3i,j), but individual F2 hybrids showed variable patterns (Fig. S4).

Differences in PA composition between genotypes were significant in both shoots and roots, and differences between the shoots and roots were also significant (two-factor Adonis test; genotype: df = 05, r2 = 0.31, P = 0.01; plant part: df = 1, r2 = 0.36, = 0.01). The relative concentrations of major PAs and of PA groups were genotype-dependent (Kruskal-Wallis test; in all cases: df = 105; < 0.01). Shoots tended to contain greater relative concentrations of jacobine-like PAs than roots, while roots had higher relative concentrations of senecionine-like PAs than shoots. The shoot and root samples could therefore be divided into two groups with regard to PA composition (Fig. S4).

Covariation between individual PAs and shoot/root correlations

We investigated correlations between individual PAs both in the shoots and in the roots. Hierarchical cluster analysis (HCA) was used to visualize the covariation between PAs. Based on the clustering results, the PAs in the shoots could be divided into four groups. Interestingly, these groups correspond to the structural groups shown in Table 1, such that PAs from the same structural group clustered together (see structural groups in Table 1). However, there were some exceptions. Usaramine, spartiodine and their corresponding N-oxides are senecionine-like PAs but were not clustered with other senecionine-like PAs. Also, jacozine N-oxide clustered with erucifoline-like PAs instead of jacobine-like PAs (Fig. 4a). Furthermore, we found that the free base form of each PA often clustered with its corresponding N-oxide (Fig. 4a, Table S4). A similar pattern was found with regard to the cluster analysis of the PA concentrations in the roots (Fig. 4b).


Figure 4. Hierarchical clusters of individual pyrrolizidine alkaloids (PAs) in shoots (a) and roots (b) of Jacobaea aquatica, Jacobaea vulgaris, and F1 and F2 hybrids. The data used in this analysis were the log-transformed absolute concentrations of individual PAs. Jacobaea aquatica, one genotype, 12 replicates; Jacobaea vulgaris, one genotype, 12 replicates; F1-A, one genotype, 11 replicates; F1-B, one genotype, 12 replicates; F2, 102 genotypes, three to six replicates per genotype. Abbreviations for PAs are defined in Table 1.

Download figure to PowerPoint

We compared the concentrations of individual PAs, PA groups and total PA between shoots and roots. Concentrations of all individual PAs were significantly positively correlated between roots and shoots. Consequently, the concentrations of total PA and of all four groups were also correlated between these two tissues (Table 3).

Table 3.   Spearman rank correlations between pyrrolizidine alkaloid (PA) concentration in shoots and roots of Jacobaea aquatica (one genotype), Jacobaea vulgaris (one genotype), F1 hybrids (two genotypes) and F2 hybrids (102 genotypes)
  1. In all cases: df = 607, < 0.01.

Senecionine-like PAsSenecionine0.53Senecionine N-oxide0.42
Intergerrimine0.58Intergerrimine N-oxide0.51
Retrorsine0.41Retrorsine N-oxide0.44
Usaramine0.54Usaramine N-oxide0.80
Riddelliine0.22Riddelliine N-oxide0.29
Seneciphylline0.49Seneciphylline N-oxide0.45
Spartiodine0.54Spartiodine N-oxide0.60
Acetylseneciphylline0.54Acetylseneciphylline N-oxide0.40
Jacobine-like PAsJacobine0.77Jacobine N-oxide0.83
Jacoline0.82Jacoline N-oxide0.85
Jaconine0.83Jaconine N-oxide0.83
Jacozine0.49Jacozine N-oxide0.66
Erucifoline-like PAsErucifoline0.50Erucifoline N-oxide0.46
Acetylerucifoline0.24Acetylerucifoline N-oxide0.38
Otosenine-like PAsSenkirkine0.35Florosenine0.77
SumPA free bases0.57PA N-oxides0.46
Senecionine-like PAs0.44Jacobine-like PAs0.86
Erucifoline-like PAs0.50Otosenine-like PAs0.49
Total PA0.55  


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Novelty resulting from hybridization

In agreement with our expectations, we found that some F2 hybrid genotypes exhibited extreme expression of some PAs, and novel patterns of overall PA composition. We found evidence for qualitative novelty: three acetylated otosenine-like PAs (florosenine, floridanine and doronine) were present in the roots of F1 and some F2 genotypes, but never or only in trace amounts in the roots of the parents, although all three PAs were present in the shoots of J. aquatica (Tables S1, S2). Florosenine was also reported to be novel to F1 hybrids in a recent study by Kirk et al. (2010), although the detection method used by these authors was less sensitive than that used in this study. The expression of a parental SM in novel tissues can lead to new ecological and evolutionary consequences. For example, PAs have been shown to have different effects on the growth of root-associated micro-organisms (Kowalchuk et al., 2006), and the addition of a novel compound in the roots of hybrids might impact interactions with symbiotic or pathogenic microbes.

Some otosenine-like PAs such as desacetyldoronine were over-expressed in the shoots of some F2 hybrids, and in 10 F2 hybrids this structural group comprised > 20% of the total PA present. To our knowledge, otosenine-like PAs have not been previously reported as a major component of the bouquet of PAs in J. vulgaris or J. aquatica. In addition, overall PA compositions were different in some F2 hybrid genotypes compared with the parents. The two parental genotypes were well separated according to the NMDS analysis, and differed especially with regard to the relative amount of senecionine-like and jacobine-like PAs in shoots. Many F2 hybrid genotypes showed PA compositions that were intermediate to those of the parental genotypes (Fig. S4). However, some F2 hybrid shoots contained a higher relative proportion of erucifoline-like PAs. These F2 hybrids showed different patterns from those found in the shoots of either parental genotype, in which jacobine-like PAs or senecionine-like PAs were dominant. PAs can have individual effects on aboveground herbivores, or synergistic effects that depend on interactions among multiple PAs within a bouquet (Macel et al., 2005). The ecological role of erucifoline-like PAs is not well understood, but alteration of aboveground PA composition might have implications in terms of susceptibility to generalist and specialist herbivores. Novelty in PA composition among F2 genotypes illustrates that hybridization might increase the diversity of PA expression within the Jacobaea genus. It is also possible that altered PA expression can affect the fitness of natural hybrids, and can in turn mediate population dynamics within natural hybrid populations. These are interesting avenues for further research.

Differences between shoots and roots

Some interesting differences between PA compositions in the shoots and roots were observed. Generally, shoots contained higher proportions of jacobine- and erucifoline-like PAs and lower proportions of senecionine- and otosenine-like PAs compared with roots (Figs 3, S4 and Tables S1, S2). Moreover, shoots contained greater proportions of biosynthetically derived PAs than the roots (Fig. S4), while the roots contained higher total PA concentrations (Figs 1, S3 and Tables S1, S2). The mechanisms by which these patterns are established are not yet clear. In another study, a few J. vulgaris genotypes derived from natural populations also showed similar patterns (Joosten et al., 2009, 2011). However, the ecological implications of different PA compositions and concentrations in roots and shoots remain uncertain. Recent work has shown that jacobine-like PAs are more important than other PA groups for mediating interactions between Jacobaea plants and an aboveground generalist herbivore (western flower thrips; Leiss et al., 2009; D. Cheng et al., unpublished data; but also see Kowalchuk et al., 2006). If jacobine-like PAs are more important in mediating aboveground interactions than belowground interactions, it is logical that they should be sequestered to a great extent in aboveground plant parts. Otosenine-like PAs generally accumulate more in the roots (Table S1, S2, Figs 3, S3). However, the role of otosenine-like PAs in mediating belowground interactions has never been investigated.

Variation patterns and their implications for genetic regulation and biosynthesis

Previous studies have shown that genes that code for the presence of SMs usually have a dominant mode of inheritance: if one or both of the parents produced a particular metabolite, hybrids almost always produced it (Rieseberg & Ellstrand, 1993; Orians, 2000). This was also the case in our study with regard to the expression of PAs in Jacobaea hybrids; F1 and F2 hybrids always produced all PAs found in the parental individuals. Quantitative variation of SM expression followed a pattern of continuous variation, which suggests that concentrations of individual PAs and of structural groups are controlled by multiple genes. These genes may include loci coding for the enzymes involved in biosynthetic pathway and/or regulatory genes. The interaction between such genes may show dominant, overdominant, recessive, additive, or epistatic effects on PA expression; however, the number of loci involved in PA diversification and accumulation and their modes of action and interaction cannot be elucidated based on the results of this study. Quantitative trait locus (QTL) analysis of PA expression will allow us to investigate such genetic effects, and to identify interactions between loci.

We observed that expression of PAs within structural groups was correlated (Fig. 4 and Table S4), while PAs from different structural groups (except senecionine-like and erucifoline-like PAs) showed greater independence. This pattern appeared both in the shoots and roots (Fig. 4, and Table S4). This suggests that the up- or down-regulation of enzymatic pathways involved in the biosynthesis of derived structural groups (i.e. erucifoline-, jacobine- and otosenine-like PAs) may be active processes, but diversification within structural groups is more passive. In other words, once the pathway leading to the biosynthesis of PAs from a particular structural group (e.g. jacobine-like PAs) is turned on, several different PAs from within that group (jacobine, jacozine, jacoline, etc.) are synthesized in a codependent manner. Furthermore, the high correlation between the PA free bases and their corresponding N-oxides indicates that the conversion of PAs between the two forms may be a passive, concentration-dependent, and PA-structurally specific process (also see Joosten et al., 2011).

In spite of the differences in PA compositions between shoots and roots, these two tissues showed positive correlations with regard to the absolute concentrations of PAs. This pattern can be explained by processes of PA synthesis and accumulation in Jacobaea (Senecio) plants. The concentration of a particular PA in the shoots and/or roots is determined by a number of steps: synthesis of the backbone structure senecionine N-oxide, which occurs mostly in the roots of Jacobaea (Senecio) plants; structural transformation, which occurs primarily in the shoots; and translocation and storage of PAs. Root-to-shoot translocation of PAs occurs exclusively via the phloem. Once they are synthesized, PAs do not undergo any degradation or turnover. They are slowly but steadily distributed within the plant (reviewed by Hartmann & Ober, 2000). Therefore, it is not surprising that there were positive and highly significant correlations between PA concentrations in the shoots and roots.

In conclusion, understanding the mechanisms and consequences of such patterns of PA variation may provide fascinating clues with regard to biosynthetic pathways, evolutionary constraints, and the ecological role of these SMs. Furthermore, the hybrid system described in this study is a useful tool for understanding the ecological role of PA variation, because a great diversity of PA patterns is found among segregating hybrids. We detected 37 individual PAs in above- and belowground plant parts, including both free base and N-oxide forms of many PAs, using LC-MS/MS. We found qualitative and quantitative differences in the patterns of PA variation in segregating hybrids compared with parental genotypes. Moreover, we revealed that PAs from within structural groups covary, and there are significant correlations between the accumulation of PAs in the shoots and that in the roots.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Dandan Cheng thanks the China Scholarship Council (CSC) for financial support. We thank Cilke Hermans, Karin van der Veen-van Wijk, Richard Fens, Meike Klinkhamer and Henk Nell for their technical assistance, Eddy van der Meijden for discussions about the experimental design and the writing of the manuscript, Tom de Jong for the introduction to the statistical software R and Lotte Joosten for help and suggestions for PA extraction and measurements. We also thank three anonymous referees for their suggestions.


  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information
  • Abbott RJ, Brennan AC, James JK, Forbes DG, Hegarty MJ, Hiscock SJ. 2009. Recent hybrid origin and invasion of the British Isles by a self-incompatible species, oxford ragwort (Senecio squalidus L., Asteraceae). Biological Invasions 11: 11451158.
  • Abbott RJ, Lowe AJ. 2004. Origins, establishment and evolution of new polyploid species: Senecio cambrensis and S. eboracensis in the British Isles. Biological Journal of the Linnean Society 82: 467474.
  • Arnold ML. 1992. Natural hybridization as an evolutionary process. Annual Review of Ecology and Systematics 23: 237261.
  • Beck E, Scheibe R, Schlutter I, Sauer W. 1992. Senecio × Saundersii Sauer and Beck (Asteraceae), an intermediate hybrid between S. keniodendron and S. keniensis of MT Kenya. Phyton-Annales Rei Botanicae 32: 938.
  • Burow M, Halkier BA, Kliebenstein DJ. 2010. Regulatory networks of glucosinolates shape Arabidopsis thaliana fitness. Current Opinion in Plant Biology 13: 348353.
  • Crawley MJ. 2005. Statistics: an introduction using r. London, UK: John Wiley & Sons, Ltd.
  • van Dam NM, Tytgat TOG, Kirkegaard JA. 2009. Root and shoot glucosinolates: a comparison of their diversity, function and interactions in natural and managed ecosystems. Phytochemistry Reviews 8: 171186.
  • van Dam NM, Vuister LWM, Bergshoeff C, de Vos H, van Der Meijden ED. 1995. The ‘raison d’etre’ of pyrrolizidine alkaloids in cynoglossum officinale deterrent effects against generalist herbivores. Journal of Chemical Ecology 21: 507523.
  • Giraudoux P. 2010. Pgirmess: data analysis in ecology R package version 1.4.8. URL
  • Goslee SC, Urban DL. 2007. The ecodist package for dissimilarity-based analysis of ecological data. Journal of Statistical Software 22: 119.
  • Hallgren P, Ikonen A, Hjalten J, Roininen H. 2003. Inheritance patterns of phenolics in F1, F2, and back-cross hybrids of willows: implications for herbivore responses to hybrid plants. Journal of Chemical Ecology 29: 11431158.
  • Hartmann T. 1999. Chemical ecology of pyrrolizidine alkaloids. Planta 207: 483495.
  • Hartmann T, Dierich B. 1998. Chemical diversity and variation of pyrrolizidine alkaloids of the senecionine type: biological need or coincidence? Planta 206: 443451.
  • Hartmann T, Ehmke A, Eilert U, von Borstel K, Theuring C. 1989. Sites of synthesis, translocation and accumulation of pyrrolizidine alkaloid N-oxides in Senecio vulgaris L. Planta 177: 98107.
  • Hartmann T, Ober D. 2000. Biosynthesis and metabolism of pyrrolizidine alkaloids in plants and specialized insect herbivores. In: Leeper FJ, Vederas JC, eds. Biosynthesis: Aromatic Polyketides, Isoprenoids, Alkaloids. Heidelberg, Germany: Springer, 207243.
  • Hartmann T, Toppel G. 1987. Senecionine N-oxide, the primary product of pyrrolizidine alkaloid biosynthesis in root cultures of Senecio vulgaris. Phytochemistry 26: 16391643.
  • Hodalova I. 2002. A new hybrid Senecio × Slovacus from the S. nemorensis group (Compositae) in the West Carpathians. Biologia 57: 7582.
  • Hol WHG, van Veen JA. 2002. Pyrrolizidine alkaloids from Senecio jacobaea affect fungal growth. Journal of Chemical Ecology 28: 17631772.
  • Iriti M, Faoro F. 2009. Chemical diversity and defence metabolism: how plants cope with pathogens and ozone pollution. International Journal of Molecular Sciences 10: 33713399.
  • James JK, Abbott RJ. 2005. Recent, allopatric, homoploid hybrid speciation: the origin of Senecio squalidus (Asteraceae) in the British Isles from a hybrid zone on Mount Etna, Sicily. Evolution 59: 25332547.
  • Joosten L, Cheng D, Mulder PPJ, Vrieling K, van Veen JA, Klinkhamer PGL. 2011. The genotype dependent presence of pyrrolizidine alkaloids as tertiary amine in Jacobaea vulgaris. Phytochemistry 72: 214222.
  • Joosten L, Mulder PPJ, Klinkhamer PGL, van Veen JA. 2009. Soil-borne microorganisms and soil-type affect pyrrolizidine alkaloids in Jacobaea vulgaris. Plant and Soil 325: 133143.
  • Joosten L, Mulder PPJ, Vrieling K, van Veen JA, Klinkhamer PGL. 2010. The analysis of pyrrolizidine alkaloids in Jacobaea vulgaris; a comparison of extraction and detection methods. Phytochemical Analysis 21: 197204.
  • Kim M, Cui ML, Cubas P, Gillies A, Lee K, Chapman MA, Abbott RJ, Coen E. 2008. Regulatory genes control a key morphological and ecological trait transferred between species. Science 322: 11161119.
  • Kirk H, Cheng D, Choi Y, Vrieling K, Klinkhamer P. 2011. Transgressive segregation of primary and secondary metabolites in in F2 hybrids between Jacobaea aquatica and J. vulgaris. Metabolomics. doi: 10.1007/s11306-011-0301-8.
  • Kirk H, Choi YH, Kim HK, Verpoorte R, van der Meijden E. 2005. Comparing metabolomes: the chemical consequences of hybridization in plants. New Phytologist 167: 613622.
  • Kirk H, Macel M, Klinkhamer PGL, Vrieling K. 2004. Natural hybridization between Senecio jacobaea and Senecio aquaticus: molecular and chemical evidence. Molecular Ecology 13: 22672274.
  • Kirk H, Vrieling K, Van Der Meijden E, Klinkhamer PGL. 2010. Species by environment interactions affect pyrrolizidine alkaloid expression in Senecio jacobaea, Senecio aquaticus, and their hybrids. Journal of Chemical Ecology 36: 378387.
  • Kliebenstein DJ. 2009. A quantitative genetics and ecological model system: understanding the aliphatic glucosinolate biosynthetic network via QTLs. Phytochemistry Reviews 8: 243254.
  • Kojadinovic I. 2010. Hierarchical clustering of continuous variables based on the empirical copula process and permutation linkages. Computational Statistics & Data Analysis 54: 90108.
  • Kowalchuk GA, Hol WHG, van Veen JA. 2006. Rhizosphere fungal communities are influenced by Senecio jacobaea pyrrolizidine alkaloid content and composition. Soil Biology & Biochemistry 38: 28522859.
  • Lankau RA, Kliebenstein DJ. 2009. Competition, herbivory and genetics interact to determine the accumulation and fitness consequences of a defence metabolite. Journal of Ecology 97: 7888.
  • Leiss KA, Choi YH, Abdel-Farid IB, Verpoorte R, Klinkhamer PGL. 2009. NMR metabolomics of thrips (Frankliniella occidentalis) resistance in Senecio hybrids. Journal of Chemical Ecology 35: 219229.
  • Lopez MG, Wulff AF, Xifreda CC. 2008. Natural hybrids in Senecio (Asteraceae): new records from Argentina. Plant Biosystems 142: 185190.
  • Macel M, Bruinsma M, Dijkstra SM, Ooijendijk T, Niemeyer HM, Klinkhamer PGL. 2005. Differences in effects of pyrrolizidine alkaloids on five generalist insect herbivore species. Journal of Chemical Ecology 31: 14931508.
  • Macel M, Klinkhamer PGL. 2010. Chemotype of Senecio jacobaea affects damage by pathogens and insect herbivores in the field. Evolutionary Ecology 24: 237250.
  • Macel M, Klinkhamer PGL, Vrieling K, van der Meijden E. 2002. Diversity of pyrrolizidine alkaloids in Senecio species does not affect the specialist herbivore Tyria jacobaeae. Oecologia 133: 541550.
  • Macel M, Vrieling K. 2003. Pyrrolizidine alkaloids as oviposition stimulants for the cinnabar moth, Tyria jacobaeae. Journal of Chemical Ecology 29: 14351446.
  • Macel M, Vrieling K, Klinkhamer PGL. 2004. Variation in pyrrolizidine alkaloid patterns of Senecio jacobaea. Phytochemistry 65: 865873.
  • Oksanen J, Blanchet FG, Kindt R, Legendre P, O’Hara RB, Simpson GL, Solymos P, Stevens MHH, Wagner H. 2010. R package Vegan 1.17-4. URL
  • O’Reilly-Wapstra JM, Potts BM, McArthur C, Davies NW, Tilyard P. 2005. Inheritance of resistance to mammalian herbivores and of plant defensive chemistry in a Eucalyptus species. Journal of Chemical Ecology 31: 519537.
  • Orians CM. 2000. The effects of hybridization in plants on secondary chemistry: implications for the ecology and evolution of plant-herbivore interactions. American Journal of Botany 87: 17491756.
  • Pelser PB, de Vos H, Theuring C, Beuerle T, Vrieling K, Hartmann T. 2005. Frequent gain and loss of pyrrolizidine alkaloids in the evolution of Senecio section jacobaea (Asteraceae). Phytochemistry 66: 12851295.
  • Pelser PB, Gravendeel B, van der Meijden R. 2003. Phylogeny reconstruction in the gap between too little and too much divergence: the closest relatives of Senecio jacobaea (Asteraceae) according to DNA sequences and AFLPs. Molecular Phylogenetics and Evolution 29: 613628.
  • Quinn GP, Keough MJ. 2002. Experimental design and data analysis for biologists. Cambridge, UK: Cambridge University Press.
  • R Development Core Team. 2009. R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL
  • Rieseberg LH, Archer MA, Wayne RK. 1999. Transgressive segregation, adaptation and speciation. Heredity 83: 363372.
  • Rieseberg LH, Carney SE. 1998. Plant hybridization. New Phytologist 140: 599624.
  • Rieseberg LH, Ellstrand NC. 1993. What can molecular and morphological markers tell us about plant hybridization. Critical Reviews in Plant Sciences 12: 213241.
  • Rieseberg LH, Kim SC, Randell RA, Whitney KD, Gross BL, Lexer C, Clay K. 2007. Hybridization and the colonization of novel habitats by annual sunflowers. Genetica 129: 149165.
  • Rieseberg LH, Raymond O, Rosenthal DM, Lai Z, Livingstone K, Nakazato T, Durphy JL, Schwarzbach AE, Donovan LA, Lexer C. 2003. Major ecological transitions in wild sunflowers facilitated by hybridization. Science 301: 12111216.
  • Rizk AM. 1991. Naturally occurring pyrrolizidine alkaloids. Boca Raton, FL, USA: CRC Press.
  • Sander H, Hartmann T. 1989. Site of synthesis, metabolism and translocation of senecionine N-oxide in cultured roots of Senecio erucifolius. Plant Cell Tissue and Organ Culture 18: 1931.
  • Seehausen O. 2004. Hybridization and adaptive radiation. Trends in Ecology & Evolution 19: 198207.
  • Soltis PS, Soltis DE. 2009. The role of hybridization in plant speciation. Annual Review of Plant Biology 60: 561588.
  • Stebbins GL. 1959. The role of hybridization in evolution. Proceedings of the American Philosophical Society 103: 231251.
  • Vincent PLD. 1996. Progress on clarifying the generic concept of Senecio based on an extensive world-wide sample of taxa. In: Hind DJN, Beentje HJ, eds. Compositae: systematics. Proceedings of the international compositae conference. Kew, UK: Royal Botanic Gardens, 597611.
  • Wiedenfeld H, Roeder E, Bourauel T, Edgar J. 2008. Pyrrolizidine alkaloids: structure and toxicity. Bonn, Germany: V&R unipress GmbH.
  • Witte L, Ernst L, Adam H, Hartmann T. 1992. Chemotypes of 2 pyrrolizidine alkaloid-containing Senecio species. Phytochemistry 31: 559565.

Supporting Information

  1. Top of page
  2. Summary
  3. Introduction
  4. Materials and Methods
  5. Results
  6. Discussion
  7. Acknowledgements
  8. References
  9. Supporting Information

Fig. S1 Structures of the pyrrolizidine alkaloids (PAs) found in shoots and roots of Jacobaea aquatica, Jacobaea vulgaris, and F1 and F2 hybrids.

Fig. S2 Putative biosynthetic pathways for diversification of pyrrolizidine alkaloids (PAs) in the Jacobaea section.

Fig. S3 Frequency distribution of genotypic mean concentrations of pyrrolizidine alkaloids (PAs) from four structural groups in the shoots and roots of 102 F2 hybrid genotypes between Jacobaea aquatica and Jacobaea vulgaris.

Fig. S4 Pyrrolizidine alkaloid (PA) composition in the shoots and roots plotted by two-dimension nonparametric multidimensional scaling (NMDS) and the loadings.

Table S1 Pyrrolizidine alkaloid (PA) concentrations in the shoots of Jacobaea aquatic (one genotype), Jacobaea vulgaris (one genotype), F1 hybrids (two genotypes) and F2 hybrids (102 genotypes)

Table S2 Pyrrolizidine alkaloid (PA) concentrations in the roots of Jacobaea aquatic (one genotype), Jacobaea vulgaris (one genotype), F1 hybrids (two genotypes) and F2 hybrids (102 genotypes)

Table S3 Quantitative variation of pyrrolizidine alkaloids (PAs) in the shoots and roots of two F1 and 102 F2 hybrids relative to parental genotypes (one genotype each of Jacobaea aquatica and Jacobaea vulgaris)

Table S4 Coefficients (rs) of Spearman rank correlation between the individual pyrrolizidine alkaloid (PA) concentrations in the shoots and roots of Jacobaea aquatic (one genotype), Jacobaea vulgaris (one genotype), F1 hybrids (two genotypes) and F2 hybrids (102 genotypes)

Please note: Wiley-Blackwell are not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing material) should be directed to the New Phytologist Central Office.

NPH_3841_sm_FigS1-S4.doc301KSupporting info item
NPH_3841_sm_TableS1-S4.doc617KSupporting info item