Evolutionary histories determine DNA barcoding success in vascular plants: seven case studies using intraspecific broad sampling of closely related species.

BACKGROUND: Four plastid regions, rpoB, rpoC1, matK, and trnH-psbA, have been recommended as DNA barcodes for plants. Their success in delimiting species boundaries depends on the existence of a clear-cut difference between inter- and intraspecific variability. We tested the ability of these regions to discriminate among closely related species in seven genera of flowering plants with different generation times (trees, perennials, and annuals). To ensure a maximum coverage of intraspecific diversity, and therefore to better evaluate the resolution power of each barcode, we applied a population genetics approach by sampling three to 45 individuals per species over a wide geographical range. RESULTS: All possible combinations between loci were analysed, which showed that using more than one locus does not always improve the resolution power. The trnH-psbA locus was most effective at discriminating among closely related species (Acer, Lonicera, Geranium, and Veronica), singly or in combination. For Salix, Adenostyles, and Gentiana, the best results were obtained with the combination of matK, rpoB, and trnH-psbA. No barcoding gap was found within six genera analysed, excepting Lonicera. This is due to shared polymorphisms among species, combined with very divergent sequences within species. These genetic patterns reflect incomplete lineage sorting and hybridization events followed by chloroplast capture. CONCLUSIONS: Our results strongly suggest that adding trnH-psbA to the two obligate DNA barcodes proposed by the CBOL plant-working group (matK and rbcL) should be mandatory for closely related species. In our sampling, generation time had no influence on DNA barcoding success, as the best and worst identification successes were found for the two tree genera (Acer, 64 % success and Salix, 86 % failure). Evolutionary histories are the main factor influencing DNA barcoding success in the studied genera.

Effects of pyrrolizidine alkaloids and sesquiterpenes on snail feeding.

We determined in the laboratory the feeding response of two populations of the generalist herbivorous snail Arianta arbustorum (Helicidae) towards the composite Adenostyles alliariae and towards various allelochemicals. These were: a pyrrolizidine alkaloid (PA) extract of Adenostyles leaves; senecionine (a PA present in Adenostyles); retrorsine (a PA not present in Adenostyles) and two sesquiterpene (ST) fractions from Adenostyles: a mixture of the STs adenostylone and neoadenostylone, and deacyladenostylone. Tertiary PAs and PA N-oxides were tested separately. For each allelochemical, we tested whether it was deterrent or whether it induced changes of feeding behaviour (i.e. whether it had pre- or postingestive effects), and whether the effects were more pronounced with younger (smaller) snails. The tertiary PA extract from Adenostyles was deterrent, especially for young snails, but did not induce changes of feeding behaviour. Tertiary PA senecionine was deterrent for young snails only and induced changes of feeding behaviour. Also, consumption of untreated Petasites was higher after this treatment. Tertiary PA retrorsine was not deterrent, but induced changes of feeding behaviour. The PA N-oxides showed no activity against the snails. The mixture of adenostylone and neoadenostylone was deterrent and induced feeding aversions. Deacyladenostylone was highly deterrent, but did not induce changes of feeding behaviour. At the Jura site, PA content of Adenostyles was lower than at the Black Forest site. The snails from Jura consumed much less Adenostyles than the snails from Black Forest, and also ate a little less of the treated leaf discs. The PAs which are encountered by the snails in their natural food plants (PA extract and senecionine) were more deterrent than retrorsine (a novel compound). This suggests that the snails have mechanisms for the rejection of allelochemicals which they encounter in their natural food plants, but not for novel allelochemicals. The results suggest two hypotheses regarding the function of the allelochemicals in Adenostyles: (1) The allelochemicals act mainly on very young snails. (2) PAs render Adenostyles toxic, while STs act as feeding deterrents.