Evidence for both phylogenetic conservatism and lability in the evolution of secondary chemistry in a tropical angiosperm radiation

Abstract

Over evolutionary timescales, shifts in plant secondary chemistry may be associated with patterns of diversification in associated arthropods. Although foundational hypotheses of plant-insect codiversification and plant defense theory posit closely related plants should have similar chemical profiles, numerous studies have documented variation in the degree of phylogenetic signal, suggesting phytochemical evolution is more nuanced than initially assumed. We utilize proton nuclear magnetic resonance (1H NMR) data, chemical classification, and genotyping-by-sequencing to resolve evolutionary relationships and characterize the evolution of secondary chemistry in the Neotropical plant clade Radula (Piper; Piperaceae). Sequencing data substantially improved phylogenetic resolution relative to past studies, and spectroscopic characterization revealed the presence of 35 metabolite classes. Broad metabolite classes displayed strong phylogenetic signal, whereas the crude 1H NMR spectra featured evolutionary lability in chemical resonances. Evolutionary correlations were detected in two pairs of compound classes (flavonoids with chalcones; p-alkenyl phenols with kavalactones), where the gain or loss of a class was dependent on the other's state. Overall, the evolution of secondary chemistry in Radula is characterized by strong phylogenetic signal of broad compound classes and concomitant evolutionary lability of specialized chemical motifs, consistent with both classic evolutionary hypotheses and recent examinations of phytochemical evolution in young lineages.

Streamlined Discovery of Antifungal Compounds in Costa Rican Piper Species Using Weighted Gene Correlation Network Analysis of crude 1H NMR Data

“…peracuminatum compound 2 was associated with another module (royal blue) that represented part of the dihydrochalcone motif, which is also present in antifungal compounds isolated from P. mollicomum 15 and P. aduncum 16 , and that can be postulated as an important structural determinant for yeast inhibition.…”

“…14 This approach has been successfully applied for the identification of metabolic changes resulting from ecological factors or treatments, and to evaluate structure level differences in secondary metabolism across taxa. 15,16 Previously, we assessed a diverse collection of crude Piper extracts to survey their growth inhibitory capabilities against four organisms, each serving as a proxy for different facets of common plant-pathogen interactions. Given that our 1 H NMR network-based approach is useful for the direct identification of spectral features, we envisioned that it could be applied within the biological screening framework to rapidly guide the discovery of bioactive natural products that are relevant to plant-mediated trophic interactions.…”

“…This approach offers the opportunity to further accelerate the discovery of bioactive natural products by targeting the 1 H NMR resonances affiliated with the biological activity of the crude extracts, thereby overcoming the bottleneck of repeatedly assaying fractions throughout the isolation process to track bioactive compounds. [17][18][19][20] However, the investigation of plant-pathogen interactions in this genus has remained mostly limited to the characterization of specific antifungal compounds across individual species. A recent review highlighted the prevalence of amides and phenylpropanoids as the most common type of antifungal compounds in the genus, which could indicate the conservation of specific modes of chemical defenses.…”

Host repertoires and changing insect–plant interactions

“…One reason is simply that plant characteristics do not always follow phylogeny. Plant secondary compounds have long been recognised as the important determinants of plant–herbivore interactions (Dethier, 1941; Fraenkel, 1959; Ehrlich & Raven, 1964), but even if there is a broad phylogenetic conservatism in plant secondary chemistry, their distribution can still vary within groups in complex ways (Uckele et al ., 2020). In terms of plant defences, there are examples not only of defence escalation, but also of de‐escalation and convergence (Agrawal et al ., 2009; Agrawal, 2010; Edger et al ., 2015; Defossez et al ., 2018; Livshultz et al ., 2018; Volf, 2018).…”