Seasonal and herbivore-induced dynamics of foliar glucosinolates in wild cabbage (Brassica oleracea)

Seasonal and herbivore-induced dynamics of foliar glucosinolates in wild cabbage (Brassica oleracea) Levels of plant secondary metabolites are not static and often change in relation to plant ontogeny. They also respond to abiotic and biotic changes in the environment, e.g., they often increase in response to biotic stress, such as herbivory. In contrast with short-lived annual plant species, especially those with growing periods of less than 2–3 months, investment in defensive compounds of vegetative tissues in biennial and perennial species may also vary over the course of an entire growing season. In garden experiments, we investigated the dynamics of secondary metabolites, i.e. glucosinolates (GSLs) in the perennial wild cabbage (Brassica oleracea), which was grown from seeds originating from three populations that differ in GSL chemistry. We compared temporal long-term dynamics of GSLs over the course of two growing seasons and short-term dynamics in response to herbivory by Pieris rapae caterpillars in a more controlled greenhouse experiment. Long-term dynamics differed for aliphatic GSLs (gradual increase from May to December) and indole GSLs (rapid increase until mid-summer after which concentrations decreased or stabilized). In spring, GSL levels in new shoots were similar to those found in the previous year. Short-term dynamics in response to herbivory primarily affected indole GSLs, which increased during the 2-week feeding period by P. rapae. Herbivore-induced changes in the concentrations of aliphatic GSLs were population-specific and their concentrations were found to increase in primarily one population only. We discuss our results considering the biology and ecology of wild cabbage. Keywords Brassica oleracea · Cabbage · Glucosinolates · Plant defence · Plant insect interactions · Secondary plant metabolites Communicated by Michael Heethoff. Introduction Electronic supplementary material The online version of this article (https ://doi.org/10.1007/s0004 9-018-0258-4) contains In herbaceous plants, foliar chemical defences tend to supplementary material, which is available to authorized users. increase within developmental stages, but also across the entire ontogenetic trajectory (Barton and Koricheva 2010). * Rieta Gols Often, concentrations of secondary metabolites vary over rieta.gols@wur.nl the growing season (Nelson et al. 1981; Velasco et al. 2007), Laboratory of Entomology, Wageningen University which in short-lived (e.g. annual) plants may encompass & Research, PO Box 16, 6700 AA Wageningen, their entire life cycle. In biennial and perennial species The Netherlands investment in foliar defensive compounds may not only German Centre for Integrative Biodiversity Research, vary within but also across seasons. Seasons are character- Leipzig, Germany ized by changes in abiotic factors such as light conditions Max Planck Institute for Chemical Ecology, Jena, Germany (day length, shading) and temperature, which also affect Centre for Ecology and Hydrology, Wallingford, UK secondary chemistry (Agerbirk et  al. 2001; Gouinguene Department of Terrestrial Ecology, Netherlands Institute and Turlings 2002; Akula and Ravishankar 2011). Plants of Ecology, Wageningen, The Netherlands further respond to biotic factors such as pathogen infection Department of Ecological Sciences, Section Animal and insect herbivory by increasing their levels of secondary Ecology, VU University Amsterdam, De Boelelaan 1085, metabolites, thereby minimising the investment in defence 1081 HV Amsterdam, The Netherlands Vol.:(0123456789) 1 3 78 R. Gols et al. until it is necessary (Karban and Baldwin 1997). The pro- caused by differences in allele frequencies at four loci duction of secondary metabolites can also be constrained (Mithen et al. 1995) (Fig. 1). In this study, temporal varia- by biosynthetic and ecological costs (Hamilton et al. 2001; tion in GSLs in relation to plant ontogeny were assessed in Strauss et al. 2002). Thus, levels of secondary metabolites common garden experiments over a 1- and 2-year period. in plants at a given time are the result of both genetic and Using the same seed batches, we also investigated short-term environmental factors. foliar GSL dynamics in response to herbivory, i.e. induc- One well-studied group of secondary metabolites are glu- ibility of GSLs, in response to feeding damage inflicted by cosinolates (GSLs), which are characteristically produced by a specialist chewing herbivore Pieris rapae L. (Lepidoptera: all plants in the Capparales including species in the cabbage Pieridae) that commonly feeds on these plants in the UK. and mustard family (Brassicaceae) (Fahey et al. 2001). This The latter experiments were conducted in a greenhouse to family includes important crop plant species such as various control levels of herbivory. cultivars of cabbage (Brassica oleracea) and oil seed spe- cies (e.g. B. napus, B. juncea), as well as the model plant for genetic and molecular research, Arabidopsis thaliana. A Materials and methods good model system in the Brassicaceae for studying tempo- ral variation in plant chemical defences is the wild cabbage, Plants and insects B. oleracea L. It is a perennial crucifer that grows naturally along the Atlantic coastlines of western Europe and is the Seeds of B. oleracea were collected from different individ- ancestor of all cultivated cabbage varieties (Wichmann et al. ual plants (> 15) in three wild cabbage populations growing 2008). The plant can live up to 10 years in the field and within 10–20 km of each other on chalk cliffs along the south produces shoots that do not fall from the plant in autumn coast of Great Britain, near Swanage in Dorset. The three but which may remain attached to the stem of the plant until wild populations were located at sites known as ‘Old Harry’ the next spring when new shoots are produced. Moreover, (‘OH’, 50°38′N, 1°55′W), ‘Kimmeridge’ (‘KIM’, 50°36′N, populations of wild cabbage growing in the south-west of 2°07′W), and ‘Winspit’ (‘WIN’, 50°35′N, 2°02′W). These England differ significantly in GSL composition and con- populations significantly differ in qualitative and quantitative centration (Mithen et al. 1995; Moyes and Raybould 2001; GSL characteristics (Fig. 1, Harvey et al. 2011). Gols et al. 2008). These differences in GSL composition To measure foliar GSL dynamics in response to her- among the populations correlate significantly with herbivore bivory, we used caterpillars of the small cabbage white but- pressure (Newton et al. 2009, 2010). The responses to vari- terfly, P. rapae L. (Lepidoptera: Pieridae), which is a spe- ation in the GSL profiles have been shown to be herbivore- cialist of brassicaceous plant species and is frequently found specific, which may explain how diversity in GSL profiles feeding on the wild cabbage populations in England (Moyes is maintained in these populations (Newton et al. 2010). et al. 2000; Newton et al. 2010). Larval feeding damage of Moreover, when wild cabbage plants originating from dif- this species produces irregular sized holes throughout the ferent populations were grown together in plots in a garden leaf. The caterpillars used in the experiments were obtained experiment, colonization by herbivores and the amount of from the general insect rearing facility at Wageningen Uni- damage that they caused were not only determined by GSL versity (WU) where they have been reared on Brussels chemistry of the focal plant but also by that of its neighbour- sprout plants (B. oleracea var. gemmifera cv. Cyrus) for ing plants (Bustos-Segura et al. 2017). many generations. However, in the studies above, GSLs were measured at single time points or were restricted to qualitative aspects Glucosinolate dynamics in leaves and seeds of the GSL profile (i.e. the presence or absence of specific of field‑grown plants (experiment 1) GSLs). GSLs have been reported to also exhibit seasonal variability (Agerbirk et al. 2001; Haribal and Renwick 2001; In a previous study, it was demonstrated that aliphatic GSL Velasco et al. 2007). Here, we investigated different tempo- concentrations in leaf tissues collected from plants growing ral aspects of GSL dynamics, i.e. (1) short-term dynamics in naturally in Dorset and in a greenhouse correlated positively, response to herbivory and (2) long-term dynamics in relation although their absolute levels were lower in greenhouse- to plant ontogeny and season in the perennial B. oleracea. grown plants (Mithen et al. 1995). In a pilot study, we meas- Concentrations of GSLs were assessed in plants grown from ured similar GSL concentrations in tissues collected from seeds of wild cabbage that originated from three selected plants grown in sand and peat soil (supplementary data, Dorset (UK) populations that have been demonstrated to Fig. S1). Based on these results, we predict that GSL con- differ qualitatively and quantitatively in their GSL profiles centrations measured in plants growing in a common garden (Gols et al. 2008; Harvey et al. 2011). Variation in qualita- in the Netherlands at the same latitude under similar clima- tive aliphatic GSL chemistry in B. oleracea is predominantly tological conditions as the wild cabbage plants in the UK 1 3 Seasonal and herbivore-induced dynamics of foliar glucosinolates in wild cabbage (Brassica… 79 Fig. 1 Biosynthesis of aliphatic Chain-length modification → glucosinolates (GSLs) a (modi- fied according to Mithen et al. Methionine 1995) and b presence/absence of functional alleles at the four GS-ELONG GS-PRO loci required for the biosynthe- sis of the specific GSLs found in leaves of the three wild GS-ELONG cabbage populations (Winspit, C3 C4 C5 Kimmeridge and Old Harry) homo-methionine dihomo-methionine trihomo- used in this study. Abbreviated methionine compounds in bold font are detected in the leaves of wild B. oleracea. In b GSL distributions 3-methylthiopropyl GSL are given as fractions of plants 4-methylthiobutyl GSL producing a specific GSL com- (3MTP) (4MTB) pound in the leaves with n the GS-OX GS-OX number of plants per population (data from Bustos-Segura et al. 2017) Mehtylsulfinyl- 3-methylsulfinylpropyl GSL 4-methylsulfinylbutyl GSL (3MSOP) (4MSOB) alkyl GSL GS-ALK GS-ALK 2-propenyl GSL 3-butenyl GSL Alkenyl GSL (2Prop) (3But) GS-OHP GS-OH R-2-hydroxy-3-butenyl GSL OH GSL 3-hyroxypropyl GSL (OH3But) Locus↓ GSL → 3MSOP2Prop 4MSOB3But2OHBut elong -- ++ + pro ++ -- - alk -+-+ + oh --- -+ Population all GS tot n Kimmeridge 0.96 1.00 0.86 0.81 1.00 0.72 190 Winspit 0.87 0.85 0.43 1.00 0.95 0.43 183 Old Harry1.000.920.661.000.800.54 169 would deviate little from those measured in plants growing a between-plant distance of one meter. Plants were watered at their original location in the UK. when necessary. No additional fertilisation was applied, nei- In experiment 1a, GSL dynamics in leaf tissues were ther were the herbivores that colonized the plants removed. determined over a 2-year period. Seeds were germinated Leaf tissues were sampled for GSL analysis from each plant in moist peat soil and seedlings were transferred to a com- three times: in September 2006 in and in May and August mon garden in May 2006 in a plot near the Netherlands of the following year. Individual plants were sampled by Institute of Ecology in Heteren. Seedlings were grown in cutting one-third of the distal part of 6–10 fully expanded nutrient enriched potting soil (Lentse potgrond no. 4, Lent, leaves of various ages. Leaves that had turned yellow were The Netherlands) in a layer ~ 30 cm deep on top of sand. excluded from sampling. Collected leaf tissues were flash- Twelve seedlings from the three populations, 36 plants in frozen in liquid nitrogen immediately after sampling, pooled total, were planted in alternating positions in two rows with and stored at − 20  °C until GSL analysis. In 2008 most 1 3 ← Side-chain modification 80 R. Gols et al. plants had few leaves and only produced flowers and seeds; per plant and used for GSL analysis. For comparison with therefore, leaf tissues were not collected in this year. the previous experiment with herbivores, extra sets of 8–10 In an additional experiment (1b) conducted in 2012 we control and herbivore-exposed plants were sampled at day measured GSL dynamics within a growing season. Here 16 only. As P. rapae was no longer available, plants were we used plants of the KIM and OH population only. Seeds infested with six neonate P. brassicae caterpillars that were were germinated and seedlings were grown in the same soil introduced onto the plant at day 0. as mentioned above. Seedlings were transferred to a gar- den plot adjacent to Wageningen University on May 23 and Glucosinolate analysis plants were sampled at June 15, July 12, August 23, Octo- ber 4 and December 5 in 2012. We also sampled plants in Tissues collected in experiment 1a and 1b were extracted May 2013 to confirm the results of the previous experiment. and analysed as follows: GSLs were cold-extracted with Plants were sampled only once: at the various time points 1.00 ml of 80% methanol solution containing 50.0 µM intact leave tissues were collected from die ff rent plant individuals. 4-hydroxybenzyl GSL as internal standard, desulfated with Ten to 12 plants were sampled per population at each time arylsulfatase (Sigma-Aldrich) on a DEAE Sephadex A 25 point. Ten leaf discs (Ø = 1.0 cm) were punched from vari- column. The eluted desulfoglucosinolates were separated ous leaves using a cork borer, flashfrozen in liquid nitrogen using high performance liquid chromatography (Agilent immediately after sampling, pooled and stored at − 20 °C 1100 HPLC system, Agilent Technologies, Waldbronn, Ger- until analysis. many) on a reversed phase C-18 column (Chromolith Perfor- mance RP18e, 100 × 4.6 mm, Merck, Darmstadt Germany) Glucosinolate dynamics in response to herbivory with an water-acetonitrile gradient (0–3% acetonitrile from in greenhouse‑grown plants (experiment 2) 0 to 3 min, 3–23% acetonitrile from 3 to 11 min, 23–33% acetonitrile from 11 to 13 min, followed by a washing cycle; −1 In a second experiment we measured GSL dynamics in flow 1 ml min ). Detection was performed with a photo- response to herbivory in a greenhouse were herbivory could diode array detector and peaks were integrated at 229 nm. be controlled. Seeds were germinated and seedlings were For peak identification, some samples were run on an LC- transferred to 2.1-l pots filled with peat soil (Lentse potgrond IonTrap-MS-system to determine [M–H]− in negative mode. #4). Plants were watered daily. Greenhouse conditions were R-2-hydroxy-3-butenenyl GSL was identified based on its set at 18–25 °C, 40–80% r.h. and a photoperiod of at least retention time determined in rapeseed samples. −2 −1 16 h. If the light dropped below 500 µmol photons m  s GSLs in tissues obtained in experiment 2 were extracted during the 16-h photoperiod, supplementary illumination and analysed in a different laboratory using a slightly dif- was applied (SON-T). When the plants were 4 weeks old, ferent method (van Dam et al. 2004). GSLs were extracted they were fertilized once a week with ‘Kristallon blauw’ twice with 1  ml of boiling 70% methanol solution, des- −1 (N:P:K:micro nutrients as 19:6:20:4) at 2.5 mg l , which ulfated with arylsulfatase (Sigma-Aldrich) on a DEAE was applied to the soil. Plants were 6 weeks old when they Sephadex A 25 column. The compounds were separated were used in experiments. on a DIONEX summit HPLC, (DIONEX, Sunnyvale, CA, To investigate the temporal changes in GSL levels in USA) on a reversed phase C-18 column (Alltima C-18, response to herbivory, GSLs were measured at four time 150 × 4.6 mm, 3 µm, Alltech, Deerfield, IL, USA) with an points; 0, 4, 8 and, 16 days following infestation with P. acetonitrile–water gradient (2–35% acetonitrile from 0 to −1 rapae caterpillars. Per population, 9–11 plants were each 30 min; flow 0.75 ml min ). Detection was performed with infested with 10 first instar P. rapae caterpillars divided a photodiode array detector and peaks were integrated at over three leaves. Caterpillars were free to move and feed 229 nm. 2-Propenyl GSL (ACROS, New Jersey, USA) was within a plant. Sampling consisted of punching leaf discs used as an external standard. (Ø = 1.7 cm), one per leaf, from 5 to 6 fully unfolded leaves. To calculate the concentrations of the different types of The leaf discs were pooled per plant, flashfrozen in liquid GSLs in both analyses, we used the generally accepted relative nitrogen immediately after sampling and stored at − 20 °C. response factors (Wathelet et al. 2004) for detection at 229 nm The same plants were sampled repeatedly during the induc- in relation to the respective standards. GSL compounds were tion period. classified based on their amino acid origin as indole GSLs In an additional experiment using only two populations, (derived from tryptophan) and aliphatic GSLs (derived from WIN and OH with ten plants per population, we determined methionine) (Halkier and Gershenzon 2006). Small amounts of the effect of repeated mechanical damage on GSL dynamics. 2-phenylethyl GSL (abbreviated to 2PE), the sole GSL derived Mechanical damage was inflicted on day 0, 4 and 8, and 16 from phenylanaline, were detected as well, but they were only by punching holes in 5–6 leaves, two holes per leaf, using a consistently found in the WIN population. We further divided cork borer (Ø = 1.0 cm). The removed tissues were pooled the aliphatic GSLs in subgroups according to their side-chain 1 3 Seasonal and herbivore-induced dynamics of foliar glucosinolates in wild cabbage (Brassica… 81 modification (Fig.  1a). Aliphatic GSL compounds were abbre- three plant populations (Figs. 2, 3). However, foliar con- viated as depicted in Fig. 1. Indole GSLs were abbreviated as: centrations of the different GSLs and classes of GSLs var - I3M = indolyl-3-methyl GSL; 1MOI3M = 1-methoxyindolyl- ied considerably among the three populations and these 3-methyl GSL; 4MOI3M = 4-methoxyindolyl-3-methyl GSL; also depended on the time of sampling (Table 1, Figs. 2, 4OHI3M = 4-hydroxyindolyl-3-methyl GSL. 3a). Population-specific differences in foliar GSLs were similar as described before for these populations (e.g. Har- Statistics vey et al. 2011). Briefly, leaf tissues of KIM plants contain relatively high concentrations of indole GSLs (85% of the GSL dynamics in plants grown in a common garden (experi- total GSL content in Fig. 2a), whereas OH and WIN leaf ment 1a) were measured repeatedly on the same plant and tissues contain, relatively and absolutely, high concentra- were analysed using a mixed model analysis of variance with tions of aliphatic GSLs [contributing on average 71 (OH) population, sampling time, and their interaction as fixed fac- and 85% (WIN) to the total foliar GSL content in Fig. 2b, tors and plant individual as a random variable. The effects c]. Overall GSL concentrations are the highest in the in the model were based on restricted likelihood estimation leaves of WIN plants. PLS multivariate analysis separated (REML) using SAS 9.3 (SAS Institute Inc., Cary, NC, USA), the plants according to their sampling date and population GSL concentrations were log-transformed to meet assump- origin (Fig. 3a). GSL concentrations, especially those of tions of normality and homoscedasticity. I3M, 1MOI3M, 2Prop and 3But, were much lower in leaf Multivariate statistics (i.e. principal components analysis, tissues sampled in May 2007 than in tissues sampled in PCA) were used to separate the plants in relation to popu- September 2006 and August 2007 (Figs. 2, 3a). The plants lation origin and sampling date based on their foliar GSL grow new foliage in spring and drop the old leaves. The concentrations. We used the projection to latent structures GSL dynamics in the newly grown leaves appear to follow (PLS by means of partial least squares projections) extension the same seasonal patterns as in the previous year. of the program (SIMCA 15.0, Umetrics, Umeå, Sweden), For the six most dominant compounds, both the effect which relates variables in the X matrix (GSLs) to variables of sampling time and population were significant with the in the Y matrix (population classes and sampling times). exception of OH3But for which the population effect was The program’s cross validation procedure evaluates the sig- not significant (Table  1). Interestingly, the population- nificance of each additional component (starting with none) sampling interaction term was significant for 2Prop and by comparing the goodness of (R ) and the predictive value 3But, but not for any of the indole GSLs. This means that (Q ) of the extended model with that of the reduced model the seasonal effect on indole GSLs is similar in the three (Eriksson et al. 2006). populations (low in spring, high in autumn), but that it We used orthogonal projections to latent structures differentially affected the alkenyl GSLs 2Prop and 3But (OPLS) to reveal linear time dependent changes in GSL depending on population origin. concentration for each population separately: (1) over the In experiment 1b, within-season GSL dynamics were growing season, (2) in response to P. rapae feeding dura- followed in more detail for WIN and KIM plants. The tion and (3) In response to repeated mechanical damage. increase from May to December was statistically signif- OPLS allows for separation of the systematic variation into icant for the aliphatic GSLs, 3But, OH3But and 2Prop predictive (i.e. explained by herbivory) and orthogonal vari- and this effect was stronger in WIN than in KIM plants ation. In these analyses time was included as a quantitative (Figs.  4, 5). The seasonal dynamics of the indole GSLs variable. were more idiosyncratic. For example, concentrations of In all of these analyses, per sample, concentrations of the I3M and to a lesser extent for 1MOI3M peaked in July individual GSL compounds, total aliphatic and total indole (Fig, 4), whereas concentrations of 4MOI3M increased GSLs, grand totals and % aliphatic GSLs served as variables continuously (Fig.  5). In both plant populations, GSL in the model. Data were log-transformed, mean-centred and concentrations did not differ between plants sampled in scaled to unit variance before they were subjected to the June 2012 and May 2013, with the exception of OH3But analysis. in WIN plants, of which concentrations were margin- ally higher in May 2013 than in June 2012 (F = 4.11, 1,28 P = 0.052). In addition to abiotic factors, natural infesta- Results and discussion tions with herbivores are likely to have confounded the effects that can strictly be attributed to plant ontogeny. Seasonal dynamics in glucosinolates concentrations Nevertheless, our data show that changes in glucosinolate chemistry over the season can be quite dramatic depending The five different aliphatic and three different indole GSLs on plant population and are GSL-class specific. that were detected in the leaf tissues were present in all 1 3 82 R. Gols et al. Fig. 2 Foliar glucosinolate Indole GSLs Aliphatic GSLs (GSL) concentrations measured repeatedly in the same plant OH3But 2Prop 4MOI3M individual during 2 years of 3MSOP 3But 1MOI3M growth in three wild cabbage populations, a KIM, b WIN and I3M 4MSOB c OH, originating in Dorset, England and grown in a com- KIM mon garden. Concentrations above the x-axis depict aliphatic 0 GSL concentrations, those below the x-axis depict indole 20 GSLs. Pants were sampled in September 2006 (first year of growth), and in May and August of the following year. For the full names of the GSL compounds see Fig. 1 and the “Glucosinolate analysis” section WIN OH Sept ‘06 May ‘07 Aug ‘07 Plant individual population; I3M and 1MOI3M responded the strongest to Glucosinolates dynamics in response to herbivory P. rapae feeding and changes in concentrations of 1MO3IM were much pronounced in KIM and WIN than in OH plants To tease apart the effect of herbivory from other effects, (Figs.  6, 7). In WIN and KIM plants, concentrations of short-term GSL dynamics were followed in plants grown 1MOI3M were more than 40 times higher after 16 days of under greenhouse conditions were herbivory could be con- P. rapae feeding compared to the levels measured at t = 0 trolled. All indole GSLs increased in response to herbivory when the caterpillars were introduced. Herbivore-induced but not to the same extent and this also depended on the changes in aliphatic GSL were only found in OH plants. 1 3 Glucosinolate concentration (µmol / g DW) 1 1 1 2 2 2 3 3 4 4 4 5 5 5 6 6 6 7 7 7 8 8 8 9 9 9 10 10 10 1 1 1 2 2 2 3 3 4 4 5 4 6 6 5 7 7 6 8 7 9 8 10 9 11 11 11 12 12 1 1 2 2 3 3 4 4 5 4 6 5 7 6 8 7 9 8 10 9 11 11 12 12 Seasonal and herbivore-induced dynamics of foliar glucosinolates in wild cabbage (Brassica… 83 Fig. 3 Multivariate analysis (PLS by means of partial least squares projections) of glucosinolates (GSLs) in the leaves. GSLs were measured in plants from three wild cab- bage populations KIM (blue symbols), OH (green symbols) and WIN (yellow–red symbols) originating in Dorset, England that were grown in a common garden and of which leaves were sampled in September 2006 (light coloured symbols) and in May and August of the following year (darker coloured symbols). The score plot a visualises the structure of the samples according to the first two latent variables. R X[1] and R X[2] depict the variance frac- tion explained by the latent vari- ables. The ellipse in the score plots defines the Hotelling’s T2 confidence region and provides a 95% confidence interval for the observations. The loading plot b defines the orientation of the PC planes with the original variables. The full names of the GSL compounds are given in Fig. 1 and the “Glucosinolate analysis” section. tot aliphat totals of aliphatic GSLs, tot indole total of indole GSLs, total grand totals of GSLs, % aliphat percentage aliphatic of total. Model statistics for foliar GSLs with three significant components: R X = 0.74, Q = 0.49. (Color figure online) In this population, aliphatic GSLs tended to increase when induce primarily the indole GSLs in Brassica species exposed to P. rapae feeding (Figs. 6, 7). The only compound (Bodnaryk 1992; Pontoppidan et al. 2005). However, the that decreased in response to herbivory was 2PE in WIN additional experiment in which we measured the changes plants that produced this compounds consistently albeit in in GSL concentrations in response to mechanical damage, very low concentrations. This compound is more common in revealed that only the aliphatic GSLs 3But and OH3But root tissues of Brassica species in which it is often the domi- significantly increased in response to mechanical dam- nant GSL and tends to increase in response to root herbivory age, whereas concentrations of the indole GSLs were not (van Dam et al. 2009). In root tissues of the populations affected (supplementary data, Figs. S2–3). The effects of used in this study, it can contribute 25–40% to the total GSL mechanical damage on GSL concentrations were relatively content (van Geem et al. 2016). small as the statistical model was only significant for WIN The results described above may have been somewhat and not for OH plants (Fig. S3). These results suggest that inflated by the mechanical damage caused by the repeated the changes in indole GSL concentrations are caused by sampling of the same plant, which has been shown to caterpillar feeding and not by mechanical damage. 1 3 84 R. Gols et al. Table 1 Significance levels of statistical tests (“Materials and meth- inducible indole GSLs in wild B. oleracea, which are much ods”) on seasonal long-term (across seasons) effects on the dynamics higher than has been thus far reported in any other brassica- of glucosinolate (GSL) concentrations in Brassica oleracea leaves ceous species, may play an important role against specialist Compound Pop Time Time × pop herbivores such as P. rapae, Plutella xylostella and Athalia rosae (Gols et al. 2008; Harvey et al. 2011; Abdalsamee and Aliphatic Müller 2012). These herbivore species are specialist feed- C3 ers on plants containing GSLs and are able to effectively  3MTP – – – circumvent exposure to toxic aliphatic and aromatic GSL  3MSOP – – – breakdown products (Müller et al. 2001; Ratzka et al. 2002;  2-Prop < 0.001 < 0.001 0.06 Wittstock et al. 2004). The effects of high concentrations of C4 indole GSLs that are present in some wild B. oleracea popu-  4MTP – – – lations on the performance of specialist insect herbivores  4MSOB – – – merits further study.  3But < 0.001 < 0.001 < 0.001  OH3But 0.21 < 0.001 0.33 Seasonal and herbivore‑induced differences  Total aliphat < 0.001 < 0.001 0.002 in glucosinolate dynamics among the cabbage Indole populations  I3M < 0.02 < 0.001 0.27  1MOI3M 0.002 < 0.003 0.006 GSL concentrations within and between classes varied  4MOI3M < 0.001 < 0.001 0.88 considerably among the populations and are consist-  Total indole 0.34 < 0.001 0.06 ent with previously reported results (Mithen et al. 1995; Italics depict significant effects. Seeds originated from three popu- Moyes et  al. 2000; Gols et  al. 2008). In addition, the lations (pop effect), WIN, KIM and OH, in Dorset, England. GSLs dynamics of GSLs varied as well among the populations, were classified as aliphatic with variable change length (C3 or C4) or both over the season and in response to herbivory. It is indolic. For the full names of the GSL compounds, see Fig. 1 and the not clear to what extent GSL-related differences among “Glucosinolate analysis” section the relatively small populations of wild B. oleracea plants growing along the English coastlines are the result of vari- Among the different classes, indole GSLs appear to be able selection from abiotic and biotic factors such as insect the most responsive to herbivory (Agerbirk et  al. 2009; herbivores and pathogens. Wichmann et al. (2008) demon- Hopkins et al. 2009; Textor and Gershenzon 2009), whereas strated that the spatial distribution of cabbage populations herbivore-induced changes in aliphatic GSLs appear to be in the county of Dorset has changed little over at least the more idiosyncratic (van Dam et al. 2009). The increase of past 70 years. Each of the populations used in this study indole GSLs from May to August in the field-grown plants is exposed to significant variation in abiotic and perhaps may reflect herbivore pressure which is also highest during biotic conditions. For example, the KIM population grows this period (Bustos-Segura et al. 2017). In nature, levels of Fig. 4 Within season dynamics WIN KIM of aliphatic (above the x-axis) 1MOI3M and indole GSLs (below the x-axis) in leaf tissues in plants 4MOI3M Indole originating from the KIM and GSLs I3M WIN population. Plants were grown in a common garden in 3But 2012 and sampled at June 15, 4MSOB July 12, August 23, October 4 and December 5. The full 2Prop Aliphatic names of the GSL compounds OH3But are given in Fig. 1 and the GSLs -20 “Glucosinolate analysis” 3MSOP section. Error bars depict the -40 mean standard error of the total aliphatic and indole GSL concentrations, respectively (n = 10–12) Sampling date 1 3 GSL concentration -1 (μmolg DW) Seasonal and herbivore-induced dynamics of foliar glucosinolates in wild cabbage (Brassica… 85 Fig. 5 Multivariate analysis of within year glucosinolate dynamics in leaf tissues in plants originating from two wild cabbage populations, a KIM, b WIN in Dorset, England. GSL data presented in Fig. 4 were subjected to OPLS by means of partial least squares projections. The horizontal axis coincides with day of the year (Julian data) from left to right. The full names of aliphatic GSLs com- pounds are given in Fig. 1 and the “Glucosinolate analysis” section. tot aliphat totals of ali- phatic GSLs, tot indole total of indole GSLs, total grand totals of GSLs, % aliphat percent- age aliphatic of total. Arrows point at variables of which the correlation coefficient with time is significantly different from 0 and at the same time these vari- ables contributed significantly to the separation of samples in relation to time based on model variable importance values (VIP). Variables with VIP > 1 are highly influential (Eriksson et al. 2006). Model statistics for KIM: overall significance statistical model F = 15.9, 2,56 P < 0.001; OPLS predictive sta- 2 2 tistics, R X = 0.417, R Y = 0.398, Q = 0.362; for WIN: overall model significance, F = 12.2, 2,56 P < 0.001; OPLS predictive sta- 2 2 tistics R X = 0.353, R Y = 0.364, Q = 0.303 on an exposed cliff top and is subject to strong prevail- This may explain why KIM plants largely rely on indole ing south-westerly winds, whereas the WIN population GSL defences that are only produced upon herbivory and, grows in a sheltered cove that only is exposed to gentle thus, infest in defence when under attack. In the WIN and southerly winds. OH plants, on the other hand, also grow OH population, herbivory is most likely more predictable in an exposed site, but the cliffs are facing east and it and, therefore, plants also have high levels of constitutive is thus to some extent sheltered from prevailing south- aliphatic GSL defences that increase with aging of the westerly winds. Wind exposure may affect certain plant foliage. Newton et al. (2010) observed temporal consist- traits directly, but it may also affect the ability of herbi- encies in herbivore distributions among GSL genotypes vores to find and exploit plants, thus influencing selec- across and within 12 wild cabbage populations growing tion exerted by these herbivores on plant defence traits. along a linear coastline gradient in Dorset, UK. GSL 1 3 86 R. Gols et al. 4OHI3M 4MOI3M Indole GSLs 1MOI3M I3M 2PE 4Pent 3But 2Pro Aliphatic GSLs -10 4MSOB OH3But -20 3MSOP -30 04 816 04 816 04 816 KIM WIN OH Duration of herbivory (days) Fig. 6 Glucosinolate dynamics in response to Pieris rapae feeding and indole GSL concentrations, respectively (n = 9–11). Pants were in leaf tissues of plants originating from three wild cabbage popula- grown in a greenhouse and sampled at t = 0, 4, 8 and 16  days fol- tions (KIM, WIN and OH). Concentrations above the x-axis depict lowing introduction of ten first instar P. rapae caterpillars. The full aliphatic GSL concentrations, those below the x-axis depict indole names of the GSL compounds are given in Fig.  1 and the “Glucosi- GSLs. Error bars depict the mean standard error of the total aliphatic nolate analysis” section (4Pent 4 pentenyl GSL) genotype differentially affected several herbivore species and individual plants can live for up to 10 years in the wild. which could explain how diversity in GSL chemistry is In this situation, selection pressure from antagonists may be maintained in brassicaceous plant species (Lankau 2007; much more predictable as the plants remain in situ within for Bidart-Bouzat and Kliebenstein 2008; Burow et al. 2010; many years. By contrast, many annuals, such as the ‘model’ Newton et al. 2010). Moreover, a recent study showed that species A. thaliana, are strongly r-selected (e.g. trade off increased variation in GSL chemistry among neighbouring resources for rapid growth against reduced defence) have plants correlated positively with herbivore diversity and very short-life cycles and tend to ‘move around’ to different negatively with plant damage (Bustos-Segura et al. 2017). early successional sites from year to year. In these situa- Longer-lived perennial plants, such as wild cabbage, tions selection pressures from insects and other antagonists which retain their foliage over winter and grow in stable are often highly stochastic and thus less predictable. There- habitats over extended time frames, are important to con- fore, the mechanisms maintaining genetic variation and the sider when studying the evolution, maintenance and genetic phenotypic expression in defence traits, such as GSLs in variation in traits related to resistance against insects. The brassicaceous plant species, may differ profoundly in annual interaction between GSLs and insects, with a few excep- and perennial plant species. A broad interspecific and inter- tions, is often studied in (short-lived) annual plants such as population analysis of quantitative variation in GSL in the A. thaliana and some Brassica and Sinapis species, as well Brassicaceae might confirm the broader applicability of gen- as cultivated Brassica species, which have been exposed to eralisations of defence expression in annual and perennial artificial selection (Gols and Harvey 2009; Hopkins et al. plants. 2009; Textor and Gershenzon 2009). Wild cabbage grows in populations that are very stable (Wichmann et al. 2008) 1 3 GSL concentration -1 (μmolg DW) Seasonal and herbivore-induced dynamics of foliar glucosinolates in wild cabbage (Brassica… 87 Fig. 7 Multivariate analysis of glucosinolate dynamics in leaf tissues at variables of which the correlation coefficient with time is signifi- in response to feeding by Pieris rapae caterpillars in plants originat- cantly different from 0 and at the same time these variables contrib- ing from three wild cabbage populations, a KIM, b OH, and c WIN uted significantly to the separation of samples in relation to time in Dorset, England. GSL data were subjected to OPLS by means of based on model variable importance values (VIP). Variables with VIP partial least squares projections. The horizontal axis coincides with > 1 are highly influential (Eriksson et  al. 2006). Model statistics for increased duration of feeding from left to right. Plants were grown KIM: overall significance statistical model F = 33.0, P < 0.001, 2,36 2 2 2 in a greenhouse and sampled repeatedly at t = 0, 4, 8 and 16  days OPLS predictive statistics R X = 0.347, R Y = 0.712, Q = 0.647; for since the introduction of ten first instar P. rapae caterpillars. The WIN: overall model significance, F = 49.3, P < 0.001; OPLS pre- 2,40 2 2 2 full names of aliphatic GSLs compounds are given in Fig.  1 and the dictive statistics R X = 0.317, R Y = 0.743, Q = 0.711; for OH: overall “Glucosinolate analysis” section. tot aliphat totals of aliphatic GSLs, model significance, F = 35.1, P < 0.001; OPLS predictive statistics 2,32 2 2 2 tot indole total of indole GSLs, total grand totals of GSLs, % aliphat R X = 0.325, R Y = 0.719, Q = 0.687 percentage aliphatic of total, 4Pent 4-pentenyl GSL. Arrows point 1 3 88 R. Gols et al. Open Access This article is distributed under the terms of the Crea- in defence chemistry of wild cabbage (Brassica oleracea). Oeco- tive Commons Attribution 4.0 International License (http://creat iveco logia 166:421–431 mmons.or g/licenses/b y/4.0/), which permits unrestricted use, distribu- Hopkins RJ, van Dam NM, van Loon JJA (2009) Role of glucosinolates tion, and reproduction in any medium, provided you give appropriate in insect–plant relationships and multitrophic interactions. Annu credit to the original author(s) and the source, provide a link to the Rev Entomol 54:57–83 Creative Commons license, and indicate if changes were made. Karban R, Baldwin IT (1997) Induced responses to herbivory. Univer- sity of Chicago Press, Chicago Lankau RA (2007) Specialist and generalist herbivores exert opposing selection on a chemical defense. New Phytol 175:176–184 References Mithen R, Raybould AF, Giamoustaris A (1995) Divergent selection for secondary metabolites between wild populations of Brassica Abdalsamee MK, Müller C (2012) Effects of indole glucosinolates on oleracea and its implications for plant–herbivore interactions. performance and sequestration by the sawfly Athalia rosae and Heredity 75:472–484 consequences of feeding on the plant defense system. J Chem Moyes CL, Raybould AF (2001) The role of spatial scale and intraspe- Ecol 38:1366–1375. https ://doi.org/10.1007/s1088 6-012-0197-4 cific variation in secondary chemistry in host–plant location by Agerbirk N, Olsen CE, Nielsen JK (2001) Seasonal variation in leaf Ceutorhynchus assimilis (Coleoptera: Curculionidae). Proc R Soc glucosinolates and insect resistance in two types of Barbarea vul- B Biol Sci 268:1567–1573 garis ssp. arcuata. Phytochemistry 58:91–100 Moyes CL, Collin HA, Britton G, Raybould AE (2000) Glucosinolates Agerbirk N, De Vos M, Kim J, Jander G (2009) Indole glucosinolate and differential herbivory in wild populations of Brassica olera- breakdown products and its biological effects. Phytochem Rev cea. J Chem Ecol 26:2625–2641 8:101–120. https ://doi.org/10.1007/s1110 1-008-9098-0 Müller C, Agerbirk N, Olsen CE, Boevé JL, Schaffner U, Brakefield Akula R, Ravishankar GA (2011) Influence of abiotic stress signals on PM (2001) Sequestration of host plant glucosinolates in the secondary metabolites in plants. Plant Signal Behav 6:1720–1731. defensive hemolymph of the sawfly Athalia rosae. J Chem Ecol https ://doi.org/10.4161/psb.6.11.17613 27:2505–2516 Barton KE, Koricheva J (2010) The ontogeny of plant defense and Nelson CJ, Seiber JN, Brower LP (1981) Seasonal and intraplant vari- herbivory: characterizing general patterns using meta-analysis. ation of cardenolide content in the California milkweed, Ascle- Am Nat 175:481–493. https ://doi.org/10.1086/65072 2 pias eriocarpa, and implications for plant defense. J Chem Ecol Bidart-Bouzat MG, Kliebenstein DJ (2008) Differential levels of insect 7:981–1010. https ://doi.org/10.1007/bf009 87622 herbivory in the field associated with genotypic variation in glu- Newton E, Bullock JM, Hodgson D (2009) Bottom-up effects of glu- cosinolates in Arabidopsis thaliana. J Chem Ecol 34:1026–1037 cosinolate variation on aphid colony dynamics in wild cabbage Bodnaryk RP (1992) Effects of wounding on glucosinolates in the coty - populations. Ecol Entomol 34:614–623 ledons of oilseed rape and mustard. Phytochemistry 31:2671–2677 Newton E, Bullock JM, Hodgson D (2010) Temporal consistency in Burow M, Halkier BA, Kliebenstein DJ (2010) Regulatory networks herbivore responses to glucosinolate polymorphism in populations of glucosinolates shape Arabidopsis thaliana fitness. Curr Opin of wild cabbage (Brassica oleracea). Oecologia 164:689–699 Plant Biol 13:347–352 Pontoppidan B, Hopkins R, Rask L, Meijer J (2005) Differential wound Bustos-Segura C, Poelman EH, Reichelt M, Gershenzon J, Gols R induction of the myrosinase system in oilseed rape (Brassica (2017) Intraspecific chemical diversity among neighbouring napus): contrasting insect damage with mechanical damage. Plant plants correlates positively with plant size and herbivore load but Sci 168:715–722. https ://doi.org/10.1016/j.plant sci.2004.10.003 negatively with herbivore damage. Ecol Lett 20:87–97. https :// Ratzka A, Vogel H, Kliebenstein DJ, Mitchell-Olds T, Kroymann J doi.org/10.1111/ele.12713 (2002) Disarming the mustard oil bomb. Proc Natl Acad Sci USA Eriksson L, Johansson E, Kettaneh-Wold N, Trygg J, Wold CW S 99:11223–11228 (2006) Multi- and megavariate data analysis. Part 1: basic princi- Strauss SY, Rudgers JA, Lau JA, Irwin RE (2002) Direct and ecological ples and applications. Umetrics AB, Umeå costs of resistance to herbivory. Trends Ecol Evol 17:278–285 Fahey JW, Zalcmann AT, Talalay P (2001) The chemical diversity and Textor S, Gershenzon J (2009) Herbivore induction of the glucosi- distribution of glucosinolates and isothiocyanates among plants. nolate–myrosinase defense system: major trends, biochemical Phytochemistry 56:5–51 bases and ecological significance. Phytochem Rev 8:149–170 Gols R, Harvey JA (2009) Plant-mediated effects in the Brassicaceae van Dam NM, Witjes L, Svatos A (2004) Interactions between above- on the performance and behaviour of parasitoids. Phytochem Rev ground and belowground induction of glucosinolates in two wild 8:187–206 Brassica species. New Phytol 161:801–810 Gols R, Wagenaar R, Bukovinszky T, van Dam NM, Dicke M, Bullock van Dam NM, Tytgat TOG, Kirkegaard JA (2009) Root and shoot JM, Harvey JA (2008) Genetic variation in defense chemistry in glucosinolates: a comparison of their diversity, function and wild cabbages affects herbivores and their endoparasitoids. Ecol- interactions in natural and managed ecosystems. Phytochem Rev ogy 89:1616–1626 8:171–186. https ://doi.org/10.1007/s1110 1-008-9101-9 Gouinguene SP, Turlings TCJ (2002) The effects of abiotic factors van Geem M, Gols R, Raaijmakers CE, Harvey JA (2016) Effects on induced volatile emissions in corn plants. Plant Physiol of population-related variation in plant primary and second- 129:1296–1307 ary metabolites on aboveground and belowground multitrophic Halkier BA, Gershenzon J (2006) Biology and biochemistry of glu- interactions. Chemoecology 26:219–233. https://doi.or g/10.1007/ cosinolates. Ann Rev Plant Biol 57:303–333s0004 9-016-0222-0 Hamilton JG, Zangerl AR, DeLucia EH, Berenbaum MR (2001) The Velasco P, Cartea ME, Gonzalez C, Vilar M, Ordas A (2007) Factors carbon-nutrient balance hypothesis: its rise and fall. Ecol Lett affecting the glucosinolate content of kale (Brassica oleracea 4:86–95 acephala group). J Agric Food Chem 55:955–962. https ://doi. Haribal M, Renwick JAA (2001) Seasonal and population variation in org/10.1021/Jf062 4897 flavonoid and alliarinoside content of Alliaria petiolata. J Chem Wathelet J-P, Iori R, Leoni O, Rollin P, Quinsac A, Palmieri S (2004) Ecol 27:1585–1594. https ://doi.org/10.1023/A:10104 06224 265 Guidelines for glucosinolate analysis in green tissues used for Harvey JA, van Dam NM, Raaijmakers CE, Bullock JM, Gols R biofumigation. Agroindustria 3:257–266 (2011) Tri-trophic effects of inter- and intra-population variation 1 3 Seasonal and herbivore-induced dynamics of foliar glucosinolates in wild cabbage (Brassica… 89 Wichmann MC, Alexander MJ, Hails RS, Bullock JM (2008) Histori- Wittstock U et al (2004) Successful herbivore attack due to metabolic cal distribution and regional dynamics of two Brassica species. diversion of a plant chemical defense. Proc Natl Acad Sci USA Ecography 31:673–684 101:4859–4864 1 3 http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Chemoecology Springer Journals

Seasonal and herbivore-induced dynamics of foliar glucosinolates in wild cabbage (Brassica oleracea)

Free
13 pages

Loading next page...
 
/lp/springer_journal/seasonal-and-herbivore-induced-dynamics-of-foliar-glucosinolates-in-ngvKiGhhBq
Publisher
Springer Journals
Copyright
Copyright © 2018 by The Author(s)
Subject
Life Sciences; Ecology; Entomology; Nature Conservation
ISSN
0937-7409
eISSN
1423-0445
D.O.I.
10.1007/s00049-018-0258-4
Publisher site
See Article on Publisher Site

Abstract

Levels of plant secondary metabolites are not static and often change in relation to plant ontogeny. They also respond to abiotic and biotic changes in the environment, e.g., they often increase in response to biotic stress, such as herbivory. In contrast with short-lived annual plant species, especially those with growing periods of less than 2–3 months, investment in defensive compounds of vegetative tissues in biennial and perennial species may also vary over the course of an entire growing season. In garden experiments, we investigated the dynamics of secondary metabolites, i.e. glucosinolates (GSLs) in the perennial wild cabbage (Brassica oleracea), which was grown from seeds originating from three populations that differ in GSL chemistry. We compared temporal long-term dynamics of GSLs over the course of two growing seasons and short-term dynamics in response to herbivory by Pieris rapae caterpillars in a more controlled greenhouse experiment. Long-term dynamics differed for aliphatic GSLs (gradual increase from May to December) and indole GSLs (rapid increase until mid-summer after which concentrations decreased or stabilized). In spring, GSL levels in new shoots were similar to those found in the previous year. Short-term dynamics in response to herbivory primarily affected indole GSLs, which increased during the 2-week feeding period by P. rapae. Herbivore-induced changes in the concentrations of aliphatic GSLs were population-specific and their concentrations were found to increase in primarily one population only. We discuss our results considering the biology and ecology of wild cabbage. Keywords Brassica oleracea · Cabbage · Glucosinolates · Plant defence · Plant insect interactions · Secondary plant metabolites Communicated by Michael Heethoff. Introduction Electronic supplementary material The online version of this article (https ://doi.org/10.1007/s0004 9-018-0258-4) contains In herbaceous plants, foliar chemical defences tend to supplementary material, which is available to authorized users. increase within developmental stages, but also across the entire ontogenetic trajectory (Barton and Koricheva 2010). * Rieta Gols Often, concentrations of secondary metabolites vary over rieta.gols@wur.nl the growing season (Nelson et al. 1981; Velasco et al. 2007), Laboratory of Entomology, Wageningen University which in short-lived (e.g. annual) plants may encompass & Research, PO Box 16, 6700 AA Wageningen, their entire life cycle. In biennial and perennial species The Netherlands investment in foliar defensive compounds may not only German Centre for Integrative Biodiversity Research, vary within but also across seasons. Seasons are character- Leipzig, Germany ized by changes in abiotic factors such as light conditions Max Planck Institute for Chemical Ecology, Jena, Germany (day length, shading) and temperature, which also affect Centre for Ecology and Hydrology, Wallingford, UK secondary chemistry (Agerbirk et  al. 2001; Gouinguene Department of Terrestrial Ecology, Netherlands Institute and Turlings 2002; Akula and Ravishankar 2011). Plants of Ecology, Wageningen, The Netherlands further respond to biotic factors such as pathogen infection Department of Ecological Sciences, Section Animal and insect herbivory by increasing their levels of secondary Ecology, VU University Amsterdam, De Boelelaan 1085, metabolites, thereby minimising the investment in defence 1081 HV Amsterdam, The Netherlands Vol.:(0123456789) 1 3 78 R. Gols et al. until it is necessary (Karban and Baldwin 1997). The pro- caused by differences in allele frequencies at four loci duction of secondary metabolites can also be constrained (Mithen et al. 1995) (Fig. 1). In this study, temporal varia- by biosynthetic and ecological costs (Hamilton et al. 2001; tion in GSLs in relation to plant ontogeny were assessed in Strauss et al. 2002). Thus, levels of secondary metabolites common garden experiments over a 1- and 2-year period. in plants at a given time are the result of both genetic and Using the same seed batches, we also investigated short-term environmental factors. foliar GSL dynamics in response to herbivory, i.e. induc- One well-studied group of secondary metabolites are glu- ibility of GSLs, in response to feeding damage inflicted by cosinolates (GSLs), which are characteristically produced by a specialist chewing herbivore Pieris rapae L. (Lepidoptera: all plants in the Capparales including species in the cabbage Pieridae) that commonly feeds on these plants in the UK. and mustard family (Brassicaceae) (Fahey et al. 2001). This The latter experiments were conducted in a greenhouse to family includes important crop plant species such as various control levels of herbivory. cultivars of cabbage (Brassica oleracea) and oil seed spe- cies (e.g. B. napus, B. juncea), as well as the model plant for genetic and molecular research, Arabidopsis thaliana. A Materials and methods good model system in the Brassicaceae for studying tempo- ral variation in plant chemical defences is the wild cabbage, Plants and insects B. oleracea L. It is a perennial crucifer that grows naturally along the Atlantic coastlines of western Europe and is the Seeds of B. oleracea were collected from different individ- ancestor of all cultivated cabbage varieties (Wichmann et al. ual plants (> 15) in three wild cabbage populations growing 2008). The plant can live up to 10 years in the field and within 10–20 km of each other on chalk cliffs along the south produces shoots that do not fall from the plant in autumn coast of Great Britain, near Swanage in Dorset. The three but which may remain attached to the stem of the plant until wild populations were located at sites known as ‘Old Harry’ the next spring when new shoots are produced. Moreover, (‘OH’, 50°38′N, 1°55′W), ‘Kimmeridge’ (‘KIM’, 50°36′N, populations of wild cabbage growing in the south-west of 2°07′W), and ‘Winspit’ (‘WIN’, 50°35′N, 2°02′W). These England differ significantly in GSL composition and con- populations significantly differ in qualitative and quantitative centration (Mithen et al. 1995; Moyes and Raybould 2001; GSL characteristics (Fig. 1, Harvey et al. 2011). Gols et al. 2008). These differences in GSL composition To measure foliar GSL dynamics in response to her- among the populations correlate significantly with herbivore bivory, we used caterpillars of the small cabbage white but- pressure (Newton et al. 2009, 2010). The responses to vari- terfly, P. rapae L. (Lepidoptera: Pieridae), which is a spe- ation in the GSL profiles have been shown to be herbivore- cialist of brassicaceous plant species and is frequently found specific, which may explain how diversity in GSL profiles feeding on the wild cabbage populations in England (Moyes is maintained in these populations (Newton et al. 2010). et al. 2000; Newton et al. 2010). Larval feeding damage of Moreover, when wild cabbage plants originating from dif- this species produces irregular sized holes throughout the ferent populations were grown together in plots in a garden leaf. The caterpillars used in the experiments were obtained experiment, colonization by herbivores and the amount of from the general insect rearing facility at Wageningen Uni- damage that they caused were not only determined by GSL versity (WU) where they have been reared on Brussels chemistry of the focal plant but also by that of its neighbour- sprout plants (B. oleracea var. gemmifera cv. Cyrus) for ing plants (Bustos-Segura et al. 2017). many generations. However, in the studies above, GSLs were measured at single time points or were restricted to qualitative aspects Glucosinolate dynamics in leaves and seeds of the GSL profile (i.e. the presence or absence of specific of field‑grown plants (experiment 1) GSLs). GSLs have been reported to also exhibit seasonal variability (Agerbirk et al. 2001; Haribal and Renwick 2001; In a previous study, it was demonstrated that aliphatic GSL Velasco et al. 2007). Here, we investigated different tempo- concentrations in leaf tissues collected from plants growing ral aspects of GSL dynamics, i.e. (1) short-term dynamics in naturally in Dorset and in a greenhouse correlated positively, response to herbivory and (2) long-term dynamics in relation although their absolute levels were lower in greenhouse- to plant ontogeny and season in the perennial B. oleracea. grown plants (Mithen et al. 1995). In a pilot study, we meas- Concentrations of GSLs were assessed in plants grown from ured similar GSL concentrations in tissues collected from seeds of wild cabbage that originated from three selected plants grown in sand and peat soil (supplementary data, Dorset (UK) populations that have been demonstrated to Fig. S1). Based on these results, we predict that GSL con- differ qualitatively and quantitatively in their GSL profiles centrations measured in plants growing in a common garden (Gols et al. 2008; Harvey et al. 2011). Variation in qualita- in the Netherlands at the same latitude under similar clima- tive aliphatic GSL chemistry in B. oleracea is predominantly tological conditions as the wild cabbage plants in the UK 1 3 Seasonal and herbivore-induced dynamics of foliar glucosinolates in wild cabbage (Brassica… 79 Fig. 1 Biosynthesis of aliphatic Chain-length modification → glucosinolates (GSLs) a (modi- fied according to Mithen et al. Methionine 1995) and b presence/absence of functional alleles at the four GS-ELONG GS-PRO loci required for the biosynthe- sis of the specific GSLs found in leaves of the three wild GS-ELONG cabbage populations (Winspit, C3 C4 C5 Kimmeridge and Old Harry) homo-methionine dihomo-methionine trihomo- used in this study. Abbreviated methionine compounds in bold font are detected in the leaves of wild B. oleracea. In b GSL distributions 3-methylthiopropyl GSL are given as fractions of plants 4-methylthiobutyl GSL producing a specific GSL com- (3MTP) (4MTB) pound in the leaves with n the GS-OX GS-OX number of plants per population (data from Bustos-Segura et al. 2017) Mehtylsulfinyl- 3-methylsulfinylpropyl GSL 4-methylsulfinylbutyl GSL (3MSOP) (4MSOB) alkyl GSL GS-ALK GS-ALK 2-propenyl GSL 3-butenyl GSL Alkenyl GSL (2Prop) (3But) GS-OHP GS-OH R-2-hydroxy-3-butenyl GSL OH GSL 3-hyroxypropyl GSL (OH3But) Locus↓ GSL → 3MSOP2Prop 4MSOB3But2OHBut elong -- ++ + pro ++ -- - alk -+-+ + oh --- -+ Population all GS tot n Kimmeridge 0.96 1.00 0.86 0.81 1.00 0.72 190 Winspit 0.87 0.85 0.43 1.00 0.95 0.43 183 Old Harry1.000.920.661.000.800.54 169 would deviate little from those measured in plants growing a between-plant distance of one meter. Plants were watered at their original location in the UK. when necessary. No additional fertilisation was applied, nei- In experiment 1a, GSL dynamics in leaf tissues were ther were the herbivores that colonized the plants removed. determined over a 2-year period. Seeds were germinated Leaf tissues were sampled for GSL analysis from each plant in moist peat soil and seedlings were transferred to a com- three times: in September 2006 in and in May and August mon garden in May 2006 in a plot near the Netherlands of the following year. Individual plants were sampled by Institute of Ecology in Heteren. Seedlings were grown in cutting one-third of the distal part of 6–10 fully expanded nutrient enriched potting soil (Lentse potgrond no. 4, Lent, leaves of various ages. Leaves that had turned yellow were The Netherlands) in a layer ~ 30 cm deep on top of sand. excluded from sampling. Collected leaf tissues were flash- Twelve seedlings from the three populations, 36 plants in frozen in liquid nitrogen immediately after sampling, pooled total, were planted in alternating positions in two rows with and stored at − 20  °C until GSL analysis. In 2008 most 1 3 ← Side-chain modification 80 R. Gols et al. plants had few leaves and only produced flowers and seeds; per plant and used for GSL analysis. For comparison with therefore, leaf tissues were not collected in this year. the previous experiment with herbivores, extra sets of 8–10 In an additional experiment (1b) conducted in 2012 we control and herbivore-exposed plants were sampled at day measured GSL dynamics within a growing season. Here 16 only. As P. rapae was no longer available, plants were we used plants of the KIM and OH population only. Seeds infested with six neonate P. brassicae caterpillars that were were germinated and seedlings were grown in the same soil introduced onto the plant at day 0. as mentioned above. Seedlings were transferred to a gar- den plot adjacent to Wageningen University on May 23 and Glucosinolate analysis plants were sampled at June 15, July 12, August 23, Octo- ber 4 and December 5 in 2012. We also sampled plants in Tissues collected in experiment 1a and 1b were extracted May 2013 to confirm the results of the previous experiment. and analysed as follows: GSLs were cold-extracted with Plants were sampled only once: at the various time points 1.00 ml of 80% methanol solution containing 50.0 µM intact leave tissues were collected from die ff rent plant individuals. 4-hydroxybenzyl GSL as internal standard, desulfated with Ten to 12 plants were sampled per population at each time arylsulfatase (Sigma-Aldrich) on a DEAE Sephadex A 25 point. Ten leaf discs (Ø = 1.0 cm) were punched from vari- column. The eluted desulfoglucosinolates were separated ous leaves using a cork borer, flashfrozen in liquid nitrogen using high performance liquid chromatography (Agilent immediately after sampling, pooled and stored at − 20 °C 1100 HPLC system, Agilent Technologies, Waldbronn, Ger- until analysis. many) on a reversed phase C-18 column (Chromolith Perfor- mance RP18e, 100 × 4.6 mm, Merck, Darmstadt Germany) Glucosinolate dynamics in response to herbivory with an water-acetonitrile gradient (0–3% acetonitrile from in greenhouse‑grown plants (experiment 2) 0 to 3 min, 3–23% acetonitrile from 3 to 11 min, 23–33% acetonitrile from 11 to 13 min, followed by a washing cycle; −1 In a second experiment we measured GSL dynamics in flow 1 ml min ). Detection was performed with a photo- response to herbivory in a greenhouse were herbivory could diode array detector and peaks were integrated at 229 nm. be controlled. Seeds were germinated and seedlings were For peak identification, some samples were run on an LC- transferred to 2.1-l pots filled with peat soil (Lentse potgrond IonTrap-MS-system to determine [M–H]− in negative mode. #4). Plants were watered daily. Greenhouse conditions were R-2-hydroxy-3-butenenyl GSL was identified based on its set at 18–25 °C, 40–80% r.h. and a photoperiod of at least retention time determined in rapeseed samples. −2 −1 16 h. If the light dropped below 500 µmol photons m  s GSLs in tissues obtained in experiment 2 were extracted during the 16-h photoperiod, supplementary illumination and analysed in a different laboratory using a slightly dif- was applied (SON-T). When the plants were 4 weeks old, ferent method (van Dam et al. 2004). GSLs were extracted they were fertilized once a week with ‘Kristallon blauw’ twice with 1  ml of boiling 70% methanol solution, des- −1 (N:P:K:micro nutrients as 19:6:20:4) at 2.5 mg l , which ulfated with arylsulfatase (Sigma-Aldrich) on a DEAE was applied to the soil. Plants were 6 weeks old when they Sephadex A 25 column. The compounds were separated were used in experiments. on a DIONEX summit HPLC, (DIONEX, Sunnyvale, CA, To investigate the temporal changes in GSL levels in USA) on a reversed phase C-18 column (Alltima C-18, response to herbivory, GSLs were measured at four time 150 × 4.6 mm, 3 µm, Alltech, Deerfield, IL, USA) with an points; 0, 4, 8 and, 16 days following infestation with P. acetonitrile–water gradient (2–35% acetonitrile from 0 to −1 rapae caterpillars. Per population, 9–11 plants were each 30 min; flow 0.75 ml min ). Detection was performed with infested with 10 first instar P. rapae caterpillars divided a photodiode array detector and peaks were integrated at over three leaves. Caterpillars were free to move and feed 229 nm. 2-Propenyl GSL (ACROS, New Jersey, USA) was within a plant. Sampling consisted of punching leaf discs used as an external standard. (Ø = 1.7 cm), one per leaf, from 5 to 6 fully unfolded leaves. To calculate the concentrations of the different types of The leaf discs were pooled per plant, flashfrozen in liquid GSLs in both analyses, we used the generally accepted relative nitrogen immediately after sampling and stored at − 20 °C. response factors (Wathelet et al. 2004) for detection at 229 nm The same plants were sampled repeatedly during the induc- in relation to the respective standards. GSL compounds were tion period. classified based on their amino acid origin as indole GSLs In an additional experiment using only two populations, (derived from tryptophan) and aliphatic GSLs (derived from WIN and OH with ten plants per population, we determined methionine) (Halkier and Gershenzon 2006). Small amounts of the effect of repeated mechanical damage on GSL dynamics. 2-phenylethyl GSL (abbreviated to 2PE), the sole GSL derived Mechanical damage was inflicted on day 0, 4 and 8, and 16 from phenylanaline, were detected as well, but they were only by punching holes in 5–6 leaves, two holes per leaf, using a consistently found in the WIN population. We further divided cork borer (Ø = 1.0 cm). The removed tissues were pooled the aliphatic GSLs in subgroups according to their side-chain 1 3 Seasonal and herbivore-induced dynamics of foliar glucosinolates in wild cabbage (Brassica… 81 modification (Fig.  1a). Aliphatic GSL compounds were abbre- three plant populations (Figs. 2, 3). However, foliar con- viated as depicted in Fig. 1. Indole GSLs were abbreviated as: centrations of the different GSLs and classes of GSLs var - I3M = indolyl-3-methyl GSL; 1MOI3M = 1-methoxyindolyl- ied considerably among the three populations and these 3-methyl GSL; 4MOI3M = 4-methoxyindolyl-3-methyl GSL; also depended on the time of sampling (Table 1, Figs. 2, 4OHI3M = 4-hydroxyindolyl-3-methyl GSL. 3a). Population-specific differences in foliar GSLs were similar as described before for these populations (e.g. Har- Statistics vey et al. 2011). Briefly, leaf tissues of KIM plants contain relatively high concentrations of indole GSLs (85% of the GSL dynamics in plants grown in a common garden (experi- total GSL content in Fig. 2a), whereas OH and WIN leaf ment 1a) were measured repeatedly on the same plant and tissues contain, relatively and absolutely, high concentra- were analysed using a mixed model analysis of variance with tions of aliphatic GSLs [contributing on average 71 (OH) population, sampling time, and their interaction as fixed fac- and 85% (WIN) to the total foliar GSL content in Fig. 2b, tors and plant individual as a random variable. The effects c]. Overall GSL concentrations are the highest in the in the model were based on restricted likelihood estimation leaves of WIN plants. PLS multivariate analysis separated (REML) using SAS 9.3 (SAS Institute Inc., Cary, NC, USA), the plants according to their sampling date and population GSL concentrations were log-transformed to meet assump- origin (Fig. 3a). GSL concentrations, especially those of tions of normality and homoscedasticity. I3M, 1MOI3M, 2Prop and 3But, were much lower in leaf Multivariate statistics (i.e. principal components analysis, tissues sampled in May 2007 than in tissues sampled in PCA) were used to separate the plants in relation to popu- September 2006 and August 2007 (Figs. 2, 3a). The plants lation origin and sampling date based on their foliar GSL grow new foliage in spring and drop the old leaves. The concentrations. We used the projection to latent structures GSL dynamics in the newly grown leaves appear to follow (PLS by means of partial least squares projections) extension the same seasonal patterns as in the previous year. of the program (SIMCA 15.0, Umetrics, Umeå, Sweden), For the six most dominant compounds, both the effect which relates variables in the X matrix (GSLs) to variables of sampling time and population were significant with the in the Y matrix (population classes and sampling times). exception of OH3But for which the population effect was The program’s cross validation procedure evaluates the sig- not significant (Table  1). Interestingly, the population- nificance of each additional component (starting with none) sampling interaction term was significant for 2Prop and by comparing the goodness of (R ) and the predictive value 3But, but not for any of the indole GSLs. This means that (Q ) of the extended model with that of the reduced model the seasonal effect on indole GSLs is similar in the three (Eriksson et al. 2006). populations (low in spring, high in autumn), but that it We used orthogonal projections to latent structures differentially affected the alkenyl GSLs 2Prop and 3But (OPLS) to reveal linear time dependent changes in GSL depending on population origin. concentration for each population separately: (1) over the In experiment 1b, within-season GSL dynamics were growing season, (2) in response to P. rapae feeding dura- followed in more detail for WIN and KIM plants. The tion and (3) In response to repeated mechanical damage. increase from May to December was statistically signif- OPLS allows for separation of the systematic variation into icant for the aliphatic GSLs, 3But, OH3But and 2Prop predictive (i.e. explained by herbivory) and orthogonal vari- and this effect was stronger in WIN than in KIM plants ation. In these analyses time was included as a quantitative (Figs.  4, 5). The seasonal dynamics of the indole GSLs variable. were more idiosyncratic. For example, concentrations of In all of these analyses, per sample, concentrations of the I3M and to a lesser extent for 1MOI3M peaked in July individual GSL compounds, total aliphatic and total indole (Fig, 4), whereas concentrations of 4MOI3M increased GSLs, grand totals and % aliphatic GSLs served as variables continuously (Fig.  5). In both plant populations, GSL in the model. Data were log-transformed, mean-centred and concentrations did not differ between plants sampled in scaled to unit variance before they were subjected to the June 2012 and May 2013, with the exception of OH3But analysis. in WIN plants, of which concentrations were margin- ally higher in May 2013 than in June 2012 (F = 4.11, 1,28 P = 0.052). In addition to abiotic factors, natural infesta- Results and discussion tions with herbivores are likely to have confounded the effects that can strictly be attributed to plant ontogeny. Seasonal dynamics in glucosinolates concentrations Nevertheless, our data show that changes in glucosinolate chemistry over the season can be quite dramatic depending The five different aliphatic and three different indole GSLs on plant population and are GSL-class specific. that were detected in the leaf tissues were present in all 1 3 82 R. Gols et al. Fig. 2 Foliar glucosinolate Indole GSLs Aliphatic GSLs (GSL) concentrations measured repeatedly in the same plant OH3But 2Prop 4MOI3M individual during 2 years of 3MSOP 3But 1MOI3M growth in three wild cabbage populations, a KIM, b WIN and I3M 4MSOB c OH, originating in Dorset, England and grown in a com- KIM mon garden. Concentrations above the x-axis depict aliphatic 0 GSL concentrations, those below the x-axis depict indole 20 GSLs. Pants were sampled in September 2006 (first year of growth), and in May and August of the following year. For the full names of the GSL compounds see Fig. 1 and the “Glucosinolate analysis” section WIN OH Sept ‘06 May ‘07 Aug ‘07 Plant individual population; I3M and 1MOI3M responded the strongest to Glucosinolates dynamics in response to herbivory P. rapae feeding and changes in concentrations of 1MO3IM were much pronounced in KIM and WIN than in OH plants To tease apart the effect of herbivory from other effects, (Figs.  6, 7). In WIN and KIM plants, concentrations of short-term GSL dynamics were followed in plants grown 1MOI3M were more than 40 times higher after 16 days of under greenhouse conditions were herbivory could be con- P. rapae feeding compared to the levels measured at t = 0 trolled. All indole GSLs increased in response to herbivory when the caterpillars were introduced. Herbivore-induced but not to the same extent and this also depended on the changes in aliphatic GSL were only found in OH plants. 1 3 Glucosinolate concentration (µmol / g DW) 1 1 1 2 2 2 3 3 4 4 4 5 5 5 6 6 6 7 7 7 8 8 8 9 9 9 10 10 10 1 1 1 2 2 2 3 3 4 4 5 4 6 6 5 7 7 6 8 7 9 8 10 9 11 11 11 12 12 1 1 2 2 3 3 4 4 5 4 6 5 7 6 8 7 9 8 10 9 11 11 12 12 Seasonal and herbivore-induced dynamics of foliar glucosinolates in wild cabbage (Brassica… 83 Fig. 3 Multivariate analysis (PLS by means of partial least squares projections) of glucosinolates (GSLs) in the leaves. GSLs were measured in plants from three wild cab- bage populations KIM (blue symbols), OH (green symbols) and WIN (yellow–red symbols) originating in Dorset, England that were grown in a common garden and of which leaves were sampled in September 2006 (light coloured symbols) and in May and August of the following year (darker coloured symbols). The score plot a visualises the structure of the samples according to the first two latent variables. R X[1] and R X[2] depict the variance frac- tion explained by the latent vari- ables. The ellipse in the score plots defines the Hotelling’s T2 confidence region and provides a 95% confidence interval for the observations. The loading plot b defines the orientation of the PC planes with the original variables. The full names of the GSL compounds are given in Fig. 1 and the “Glucosinolate analysis” section. tot aliphat totals of aliphatic GSLs, tot indole total of indole GSLs, total grand totals of GSLs, % aliphat percentage aliphatic of total. Model statistics for foliar GSLs with three significant components: R X = 0.74, Q = 0.49. (Color figure online) In this population, aliphatic GSLs tended to increase when induce primarily the indole GSLs in Brassica species exposed to P. rapae feeding (Figs. 6, 7). The only compound (Bodnaryk 1992; Pontoppidan et al. 2005). However, the that decreased in response to herbivory was 2PE in WIN additional experiment in which we measured the changes plants that produced this compounds consistently albeit in in GSL concentrations in response to mechanical damage, very low concentrations. This compound is more common in revealed that only the aliphatic GSLs 3But and OH3But root tissues of Brassica species in which it is often the domi- significantly increased in response to mechanical dam- nant GSL and tends to increase in response to root herbivory age, whereas concentrations of the indole GSLs were not (van Dam et al. 2009). In root tissues of the populations affected (supplementary data, Figs. S2–3). The effects of used in this study, it can contribute 25–40% to the total GSL mechanical damage on GSL concentrations were relatively content (van Geem et al. 2016). small as the statistical model was only significant for WIN The results described above may have been somewhat and not for OH plants (Fig. S3). These results suggest that inflated by the mechanical damage caused by the repeated the changes in indole GSL concentrations are caused by sampling of the same plant, which has been shown to caterpillar feeding and not by mechanical damage. 1 3 84 R. Gols et al. Table 1 Significance levels of statistical tests (“Materials and meth- inducible indole GSLs in wild B. oleracea, which are much ods”) on seasonal long-term (across seasons) effects on the dynamics higher than has been thus far reported in any other brassica- of glucosinolate (GSL) concentrations in Brassica oleracea leaves ceous species, may play an important role against specialist Compound Pop Time Time × pop herbivores such as P. rapae, Plutella xylostella and Athalia rosae (Gols et al. 2008; Harvey et al. 2011; Abdalsamee and Aliphatic Müller 2012). These herbivore species are specialist feed- C3 ers on plants containing GSLs and are able to effectively  3MTP – – – circumvent exposure to toxic aliphatic and aromatic GSL  3MSOP – – – breakdown products (Müller et al. 2001; Ratzka et al. 2002;  2-Prop < 0.001 < 0.001 0.06 Wittstock et al. 2004). The effects of high concentrations of C4 indole GSLs that are present in some wild B. oleracea popu-  4MTP – – – lations on the performance of specialist insect herbivores  4MSOB – – – merits further study.  3But < 0.001 < 0.001 < 0.001  OH3But 0.21 < 0.001 0.33 Seasonal and herbivore‑induced differences  Total aliphat < 0.001 < 0.001 0.002 in glucosinolate dynamics among the cabbage Indole populations  I3M < 0.02 < 0.001 0.27  1MOI3M 0.002 < 0.003 0.006 GSL concentrations within and between classes varied  4MOI3M < 0.001 < 0.001 0.88 considerably among the populations and are consist-  Total indole 0.34 < 0.001 0.06 ent with previously reported results (Mithen et al. 1995; Italics depict significant effects. Seeds originated from three popu- Moyes et  al. 2000; Gols et  al. 2008). In addition, the lations (pop effect), WIN, KIM and OH, in Dorset, England. GSLs dynamics of GSLs varied as well among the populations, were classified as aliphatic with variable change length (C3 or C4) or both over the season and in response to herbivory. It is indolic. For the full names of the GSL compounds, see Fig. 1 and the not clear to what extent GSL-related differences among “Glucosinolate analysis” section the relatively small populations of wild B. oleracea plants growing along the English coastlines are the result of vari- Among the different classes, indole GSLs appear to be able selection from abiotic and biotic factors such as insect the most responsive to herbivory (Agerbirk et  al. 2009; herbivores and pathogens. Wichmann et al. (2008) demon- Hopkins et al. 2009; Textor and Gershenzon 2009), whereas strated that the spatial distribution of cabbage populations herbivore-induced changes in aliphatic GSLs appear to be in the county of Dorset has changed little over at least the more idiosyncratic (van Dam et al. 2009). The increase of past 70 years. Each of the populations used in this study indole GSLs from May to August in the field-grown plants is exposed to significant variation in abiotic and perhaps may reflect herbivore pressure which is also highest during biotic conditions. For example, the KIM population grows this period (Bustos-Segura et al. 2017). In nature, levels of Fig. 4 Within season dynamics WIN KIM of aliphatic (above the x-axis) 1MOI3M and indole GSLs (below the x-axis) in leaf tissues in plants 4MOI3M Indole originating from the KIM and GSLs I3M WIN population. Plants were grown in a common garden in 3But 2012 and sampled at June 15, 4MSOB July 12, August 23, October 4 and December 5. The full 2Prop Aliphatic names of the GSL compounds OH3But are given in Fig. 1 and the GSLs -20 “Glucosinolate analysis” 3MSOP section. Error bars depict the -40 mean standard error of the total aliphatic and indole GSL concentrations, respectively (n = 10–12) Sampling date 1 3 GSL concentration -1 (μmolg DW) Seasonal and herbivore-induced dynamics of foliar glucosinolates in wild cabbage (Brassica… 85 Fig. 5 Multivariate analysis of within year glucosinolate dynamics in leaf tissues in plants originating from two wild cabbage populations, a KIM, b WIN in Dorset, England. GSL data presented in Fig. 4 were subjected to OPLS by means of partial least squares projections. The horizontal axis coincides with day of the year (Julian data) from left to right. The full names of aliphatic GSLs com- pounds are given in Fig. 1 and the “Glucosinolate analysis” section. tot aliphat totals of ali- phatic GSLs, tot indole total of indole GSLs, total grand totals of GSLs, % aliphat percent- age aliphatic of total. Arrows point at variables of which the correlation coefficient with time is significantly different from 0 and at the same time these vari- ables contributed significantly to the separation of samples in relation to time based on model variable importance values (VIP). Variables with VIP > 1 are highly influential (Eriksson et al. 2006). Model statistics for KIM: overall significance statistical model F = 15.9, 2,56 P < 0.001; OPLS predictive sta- 2 2 tistics, R X = 0.417, R Y = 0.398, Q = 0.362; for WIN: overall model significance, F = 12.2, 2,56 P < 0.001; OPLS predictive sta- 2 2 tistics R X = 0.353, R Y = 0.364, Q = 0.303 on an exposed cliff top and is subject to strong prevail- This may explain why KIM plants largely rely on indole ing south-westerly winds, whereas the WIN population GSL defences that are only produced upon herbivory and, grows in a sheltered cove that only is exposed to gentle thus, infest in defence when under attack. In the WIN and southerly winds. OH plants, on the other hand, also grow OH population, herbivory is most likely more predictable in an exposed site, but the cliffs are facing east and it and, therefore, plants also have high levels of constitutive is thus to some extent sheltered from prevailing south- aliphatic GSL defences that increase with aging of the westerly winds. Wind exposure may affect certain plant foliage. Newton et al. (2010) observed temporal consist- traits directly, but it may also affect the ability of herbi- encies in herbivore distributions among GSL genotypes vores to find and exploit plants, thus influencing selec- across and within 12 wild cabbage populations growing tion exerted by these herbivores on plant defence traits. along a linear coastline gradient in Dorset, UK. GSL 1 3 86 R. Gols et al. 4OHI3M 4MOI3M Indole GSLs 1MOI3M I3M 2PE 4Pent 3But 2Pro Aliphatic GSLs -10 4MSOB OH3But -20 3MSOP -30 04 816 04 816 04 816 KIM WIN OH Duration of herbivory (days) Fig. 6 Glucosinolate dynamics in response to Pieris rapae feeding and indole GSL concentrations, respectively (n = 9–11). Pants were in leaf tissues of plants originating from three wild cabbage popula- grown in a greenhouse and sampled at t = 0, 4, 8 and 16  days fol- tions (KIM, WIN and OH). Concentrations above the x-axis depict lowing introduction of ten first instar P. rapae caterpillars. The full aliphatic GSL concentrations, those below the x-axis depict indole names of the GSL compounds are given in Fig.  1 and the “Glucosi- GSLs. Error bars depict the mean standard error of the total aliphatic nolate analysis” section (4Pent 4 pentenyl GSL) genotype differentially affected several herbivore species and individual plants can live for up to 10 years in the wild. which could explain how diversity in GSL chemistry is In this situation, selection pressure from antagonists may be maintained in brassicaceous plant species (Lankau 2007; much more predictable as the plants remain in situ within for Bidart-Bouzat and Kliebenstein 2008; Burow et al. 2010; many years. By contrast, many annuals, such as the ‘model’ Newton et al. 2010). Moreover, a recent study showed that species A. thaliana, are strongly r-selected (e.g. trade off increased variation in GSL chemistry among neighbouring resources for rapid growth against reduced defence) have plants correlated positively with herbivore diversity and very short-life cycles and tend to ‘move around’ to different negatively with plant damage (Bustos-Segura et al. 2017). early successional sites from year to year. In these situa- Longer-lived perennial plants, such as wild cabbage, tions selection pressures from insects and other antagonists which retain their foliage over winter and grow in stable are often highly stochastic and thus less predictable. There- habitats over extended time frames, are important to con- fore, the mechanisms maintaining genetic variation and the sider when studying the evolution, maintenance and genetic phenotypic expression in defence traits, such as GSLs in variation in traits related to resistance against insects. The brassicaceous plant species, may differ profoundly in annual interaction between GSLs and insects, with a few excep- and perennial plant species. A broad interspecific and inter- tions, is often studied in (short-lived) annual plants such as population analysis of quantitative variation in GSL in the A. thaliana and some Brassica and Sinapis species, as well Brassicaceae might confirm the broader applicability of gen- as cultivated Brassica species, which have been exposed to eralisations of defence expression in annual and perennial artificial selection (Gols and Harvey 2009; Hopkins et al. plants. 2009; Textor and Gershenzon 2009). Wild cabbage grows in populations that are very stable (Wichmann et al. 2008) 1 3 GSL concentration -1 (μmolg DW) Seasonal and herbivore-induced dynamics of foliar glucosinolates in wild cabbage (Brassica… 87 Fig. 7 Multivariate analysis of glucosinolate dynamics in leaf tissues at variables of which the correlation coefficient with time is signifi- in response to feeding by Pieris rapae caterpillars in plants originat- cantly different from 0 and at the same time these variables contrib- ing from three wild cabbage populations, a KIM, b OH, and c WIN uted significantly to the separation of samples in relation to time in Dorset, England. GSL data were subjected to OPLS by means of based on model variable importance values (VIP). Variables with VIP partial least squares projections. The horizontal axis coincides with > 1 are highly influential (Eriksson et  al. 2006). Model statistics for increased duration of feeding from left to right. Plants were grown KIM: overall significance statistical model F = 33.0, P < 0.001, 2,36 2 2 2 in a greenhouse and sampled repeatedly at t = 0, 4, 8 and 16  days OPLS predictive statistics R X = 0.347, R Y = 0.712, Q = 0.647; for since the introduction of ten first instar P. rapae caterpillars. The WIN: overall model significance, F = 49.3, P < 0.001; OPLS pre- 2,40 2 2 2 full names of aliphatic GSLs compounds are given in Fig.  1 and the dictive statistics R X = 0.317, R Y = 0.743, Q = 0.711; for OH: overall “Glucosinolate analysis” section. tot aliphat totals of aliphatic GSLs, model significance, F = 35.1, P < 0.001; OPLS predictive statistics 2,32 2 2 2 tot indole total of indole GSLs, total grand totals of GSLs, % aliphat R X = 0.325, R Y = 0.719, Q = 0.687 percentage aliphatic of total, 4Pent 4-pentenyl GSL. Arrows point 1 3 88 R. Gols et al. Open Access This article is distributed under the terms of the Crea- in defence chemistry of wild cabbage (Brassica oleracea). Oeco- tive Commons Attribution 4.0 International License (http://creat iveco logia 166:421–431 mmons.or g/licenses/b y/4.0/), which permits unrestricted use, distribu- Hopkins RJ, van Dam NM, van Loon JJA (2009) Role of glucosinolates tion, and reproduction in any medium, provided you give appropriate in insect–plant relationships and multitrophic interactions. Annu credit to the original author(s) and the source, provide a link to the Rev Entomol 54:57–83 Creative Commons license, and indicate if changes were made. Karban R, Baldwin IT (1997) Induced responses to herbivory. Univer- sity of Chicago Press, Chicago Lankau RA (2007) Specialist and generalist herbivores exert opposing selection on a chemical defense. New Phytol 175:176–184 References Mithen R, Raybould AF, Giamoustaris A (1995) Divergent selection for secondary metabolites between wild populations of Brassica Abdalsamee MK, Müller C (2012) Effects of indole glucosinolates on oleracea and its implications for plant–herbivore interactions. performance and sequestration by the sawfly Athalia rosae and Heredity 75:472–484 consequences of feeding on the plant defense system. J Chem Moyes CL, Raybould AF (2001) The role of spatial scale and intraspe- Ecol 38:1366–1375. https ://doi.org/10.1007/s1088 6-012-0197-4 cific variation in secondary chemistry in host–plant location by Agerbirk N, Olsen CE, Nielsen JK (2001) Seasonal variation in leaf Ceutorhynchus assimilis (Coleoptera: Curculionidae). Proc R Soc glucosinolates and insect resistance in two types of Barbarea vul- B Biol Sci 268:1567–1573 garis ssp. arcuata. Phytochemistry 58:91–100 Moyes CL, Collin HA, Britton G, Raybould AE (2000) Glucosinolates Agerbirk N, De Vos M, Kim J, Jander G (2009) Indole glucosinolate and differential herbivory in wild populations of Brassica olera- breakdown products and its biological effects. Phytochem Rev cea. J Chem Ecol 26:2625–2641 8:101–120. https ://doi.org/10.1007/s1110 1-008-9098-0 Müller C, Agerbirk N, Olsen CE, Boevé JL, Schaffner U, Brakefield Akula R, Ravishankar GA (2011) Influence of abiotic stress signals on PM (2001) Sequestration of host plant glucosinolates in the secondary metabolites in plants. Plant Signal Behav 6:1720–1731. defensive hemolymph of the sawfly Athalia rosae. J Chem Ecol https ://doi.org/10.4161/psb.6.11.17613 27:2505–2516 Barton KE, Koricheva J (2010) The ontogeny of plant defense and Nelson CJ, Seiber JN, Brower LP (1981) Seasonal and intraplant vari- herbivory: characterizing general patterns using meta-analysis. ation of cardenolide content in the California milkweed, Ascle- Am Nat 175:481–493. https ://doi.org/10.1086/65072 2 pias eriocarpa, and implications for plant defense. J Chem Ecol Bidart-Bouzat MG, Kliebenstein DJ (2008) Differential levels of insect 7:981–1010. https ://doi.org/10.1007/bf009 87622 herbivory in the field associated with genotypic variation in glu- Newton E, Bullock JM, Hodgson D (2009) Bottom-up effects of glu- cosinolates in Arabidopsis thaliana. J Chem Ecol 34:1026–1037 cosinolate variation on aphid colony dynamics in wild cabbage Bodnaryk RP (1992) Effects of wounding on glucosinolates in the coty - populations. Ecol Entomol 34:614–623 ledons of oilseed rape and mustard. Phytochemistry 31:2671–2677 Newton E, Bullock JM, Hodgson D (2010) Temporal consistency in Burow M, Halkier BA, Kliebenstein DJ (2010) Regulatory networks herbivore responses to glucosinolate polymorphism in populations of glucosinolates shape Arabidopsis thaliana fitness. Curr Opin of wild cabbage (Brassica oleracea). Oecologia 164:689–699 Plant Biol 13:347–352 Pontoppidan B, Hopkins R, Rask L, Meijer J (2005) Differential wound Bustos-Segura C, Poelman EH, Reichelt M, Gershenzon J, Gols R induction of the myrosinase system in oilseed rape (Brassica (2017) Intraspecific chemical diversity among neighbouring napus): contrasting insect damage with mechanical damage. Plant plants correlates positively with plant size and herbivore load but Sci 168:715–722. https ://doi.org/10.1016/j.plant sci.2004.10.003 negatively with herbivore damage. Ecol Lett 20:87–97. https :// Ratzka A, Vogel H, Kliebenstein DJ, Mitchell-Olds T, Kroymann J doi.org/10.1111/ele.12713 (2002) Disarming the mustard oil bomb. Proc Natl Acad Sci USA Eriksson L, Johansson E, Kettaneh-Wold N, Trygg J, Wold CW S 99:11223–11228 (2006) Multi- and megavariate data analysis. Part 1: basic princi- Strauss SY, Rudgers JA, Lau JA, Irwin RE (2002) Direct and ecological ples and applications. Umetrics AB, Umeå costs of resistance to herbivory. Trends Ecol Evol 17:278–285 Fahey JW, Zalcmann AT, Talalay P (2001) The chemical diversity and Textor S, Gershenzon J (2009) Herbivore induction of the glucosi- distribution of glucosinolates and isothiocyanates among plants. nolate–myrosinase defense system: major trends, biochemical Phytochemistry 56:5–51 bases and ecological significance. Phytochem Rev 8:149–170 Gols R, Harvey JA (2009) Plant-mediated effects in the Brassicaceae van Dam NM, Witjes L, Svatos A (2004) Interactions between above- on the performance and behaviour of parasitoids. Phytochem Rev ground and belowground induction of glucosinolates in two wild 8:187–206 Brassica species. New Phytol 161:801–810 Gols R, Wagenaar R, Bukovinszky T, van Dam NM, Dicke M, Bullock van Dam NM, Tytgat TOG, Kirkegaard JA (2009) Root and shoot JM, Harvey JA (2008) Genetic variation in defense chemistry in glucosinolates: a comparison of their diversity, function and wild cabbages affects herbivores and their endoparasitoids. Ecol- interactions in natural and managed ecosystems. Phytochem Rev ogy 89:1616–1626 8:171–186. https ://doi.org/10.1007/s1110 1-008-9101-9 Gouinguene SP, Turlings TCJ (2002) The effects of abiotic factors van Geem M, Gols R, Raaijmakers CE, Harvey JA (2016) Effects on induced volatile emissions in corn plants. Plant Physiol of population-related variation in plant primary and second- 129:1296–1307 ary metabolites on aboveground and belowground multitrophic Halkier BA, Gershenzon J (2006) Biology and biochemistry of glu- interactions. Chemoecology 26:219–233. https://doi.or g/10.1007/ cosinolates. Ann Rev Plant Biol 57:303–333s0004 9-016-0222-0 Hamilton JG, Zangerl AR, DeLucia EH, Berenbaum MR (2001) The Velasco P, Cartea ME, Gonzalez C, Vilar M, Ordas A (2007) Factors carbon-nutrient balance hypothesis: its rise and fall. Ecol Lett affecting the glucosinolate content of kale (Brassica oleracea 4:86–95 acephala group). J Agric Food Chem 55:955–962. https ://doi. Haribal M, Renwick JAA (2001) Seasonal and population variation in org/10.1021/Jf062 4897 flavonoid and alliarinoside content of Alliaria petiolata. J Chem Wathelet J-P, Iori R, Leoni O, Rollin P, Quinsac A, Palmieri S (2004) Ecol 27:1585–1594. https ://doi.org/10.1023/A:10104 06224 265 Guidelines for glucosinolate analysis in green tissues used for Harvey JA, van Dam NM, Raaijmakers CE, Bullock JM, Gols R biofumigation. Agroindustria 3:257–266 (2011) Tri-trophic effects of inter- and intra-population variation 1 3 Seasonal and herbivore-induced dynamics of foliar glucosinolates in wild cabbage (Brassica… 89 Wichmann MC, Alexander MJ, Hails RS, Bullock JM (2008) Histori- Wittstock U et al (2004) Successful herbivore attack due to metabolic cal distribution and regional dynamics of two Brassica species. diversion of a plant chemical defense. Proc Natl Acad Sci USA Ecography 31:673–684 101:4859–4864 1 3

Journal

ChemoecologySpringer Journals

Published: May 10, 2018

References

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off