1.Introduction
The interplay between plants and insects represents one of nature’s most common species interactions (Bernays and Chapman 1994; Stork 2018; Joneset al. 2022). Insects rely upon their host plants, as they use them as a food source, for reproduction or as a shelter (Futuyma and Moreno 1988; Kergoat et al. 2017). When feeding on a particular plant, herbivorous insects are exposed to a variety of different structural and molecular components that can affect their cellular homeostasis and in further consequence their fitness (for a summary of different plant defense mechanisms see e.g. Fürstenberg-Hägg et al. 2013). The efficient utilization of a host, thus, requires genetic and physiological “adjustments” to the different properties of the respective plant(s) (Ehrlich and Raven 1964; Després, David and Gallet 2007; Heidel-Fischer and Vogel 2015; Simon et al. 2015; Xi, Guo and Hu 2024). Within the order Lepidoptera, the majority of butterflies (superfamily Papilionoidea) are purely herbivorous. The associations with a host can be very exclusive, with a specialist insect species using only one or a few host plant species (Bernays and Graham 1988; Via 1990; Bernays 1991; Forister et al. 2015; Hardy et al.2020). In contrast, phytophagous generalists are characterized by a broader host repertoire that can cover multiple plant families and sometimes even different orders (polyphagy; Bernays and Graham 1988; Via 1990; Forister et al. 2015; Hardy et al. 2020). While specialists should be able to develop strong adaptations to their hosts, generalists encounter a much broader diversity of physically and chemically different hosts and should require a more flexible set of adaptations.
As suggested by Ehrlich and Raven (1964), the broad host repertoire of polyphagous butterflies can be achieved by including plants that are chemically similar. Chemical similarity in plants can either result from shared ancestry or convergent evolution (Janz and Nylin 1998; Janz, Nyblom and Nylin 2001; Kergoat et al. 2005; Heidel-Fischeret al. 2009). Adaptations to one specific host could, thus, also serve as preadaptations for another plant with a similar chemical profile and allow an expansion of the host repertoire (Janz and Nylin 2008; Agosta and Klemens 2008; Agosta et al. 2010). This idea has received some support from simulation studies (Araujo et al.2015) and at the level of gene expression (Heidel-Fischer et al.2009; Celorio-Mancera et al. 2023). Using the polyphagous comma butterfly (Polygonia c-album ), Heidel-Fisher et al. (2009) showed that similarities in the gene expression profiles on different hosts could be explained by a close phylogenetic relatedness or the same growth form of the plants. Despite these similarities in the gene expression profiles, the overall patterns indicated a very host-specific transcriptional response in the larval gut (Heidel-Fisher et al.2009). Such host-specific gene expression profiles in P. c-albumwere further supported by Celorio-Mancera et al. (2013). Based on this, it has been suggested that adaptations to host plants can profitably be viewed as “modules of co-expressed genes and their resulting phenotypic appearance ” (cf. Celorio-Mancera et al.2023). The latter study, which examined the overall gene expression patterns on different plants in P. c-album and related species, furthermore, showed that these specific modules can be conserved over evolutionary time and can show quite distinct gene expression profiles that differ in the degree of overlap among plants (Celorio-Manceraet al. 2023). These similarities or differences could potentially be used as an indicator for overlapping mechanisms to cope with a plant’s properties.
Another important factor when describing environment-specific expression patterns is how quickly insects can adapt their transcriptional response to a new host. A high degree of plasticity and rapid responses to a host switch can ensure that juvenile stages are able to react quickly to a changing environment (cf. Xiao et al. 2019). In many phytophagous insects in which the host is consumed during their larval stages, the actual host plant is determined by the mother (König et al. 2016; see also Nylin and Janz 1993; Gripenberg et al. 2010; Refsnider and Janzen 2010). It is still largely unknown if and to which extent larvae of such species can leave their host plant and look for alternatives, in case of deteriorating plant quality or oviposition mistakes (but see Nylin et al. 2000; Schäpers et al.2016). Although such scenarios are not well documented, larval movement within plants should often be necessary for similar reasons. Plastic adaptations to different plants and plant parts without a long temporal lag could ensure that larvae survive and complete their development without severe fitness costs. Thus, an experimental shift of host plant during larval development should give some indication of how difficult an evolutionary host plant shift would be for caterpillars. Furthermore, previous studies have mainly focused on gene expression in larvae that were reared on the same host plant during their entire development (for a summary see Birnbaum and Abbot 2020). The transcriptional profiles could, thus, not exclusively be associated with the direct responses to a certain host but also were the product of different developmental trajectories linked to a particular plant (cf. Celorio-Mancera et al. 2023). Experimental shifts could now help to further identify genes and transcriptional adjustments that are crucial for the use of a specific host plant.
Polygonia c-album (Linnaeus, 1758), the comma butterfly, is a polyphagous butterfly species in the family Nymphalidae that feeds on trees, bushes and herbs from four different orders: Rosales, Fagales, Malphigiales and Saxifragales (Nylin 1988). Within these orders the repertoire of the comma comprises Urtica , Humulus andUlmus (Urticaceae and other “urticalean rosids”), Betulaand Corylus (Betulaceae) as well as Salix (Salicaceae) andRibes (Grossulariaceae) (Nylin 1988). For some of these plants, quite characteristic chemical profiles have been described. Urticaceae, for instance, are rich in alkaloids and phenols, leaves of the Salicaea contain flavonoids and phenolic glycosides, while the presence of cyanogenic compounds has been reported for the genus Ribes (Hegnauer, 1973). Ribes is particularly unusual as a host, as its use is rather rare among butterflies. Within the family of Nymphalidae (~6,100 species; Van Nieukerken et al. 2011), for instance, Ribes is only consumed by a few species of the genusPolygonia (Gamberale-Stille et al. 2019), which together with the distinct gene expression patterns (Celorio-Mancera et al. 2013, 2023) indicates very specific and exclusive adaptations to a likely chemical divergent host. Such plant-specific profiles now represent a good prerequisite for investigating the plasticity of adaptations.
In this study, the larvae’s ability to adjust to a new host within one generation was investigated. Special attention was paid to how, and how fast, polyphagous butterfly larvae can react to a new plant in terms of gene expression. Earlier studies already showed that larvae can be moved to another host plant during their development without major effects on their performance (Söderlind 2012). It is now assumed that also the underlying transcriptional patterns can be adjusted very rapidly in response to the new environment. Based on the patterns found in previous studies, it was expected that diet shifts between hosts that strongly overlap in their gene expression profiles should be easier (i.e. faster and with fewer responding genes). In contrast, a slower and more drastic adjustment is assumed for switches between plants with very opposing patterns (e.g. Urtica vs. Ribes ; see Celorio-Manceraet al. 2023). To test this, two host switch experiments were combined with a performance screening. It was predicted that:
(1) There is a temporal course of transcriptional adjustments. Shortly after the host switch, the gene expression profiles will still correspond to the first host, but can change quickly to regain cellular homeostasis on the new host plant. (2) There is a lower number of transcriptional differences between hosts with similar chemical properties. It is assumed that overlapping processes are involved when feeding on host plants that show similarities in their chemical composition, either due to a close phylogenetic relatedness or a shared growth form (e.g. Switch A). These overlaps will result in a lower number of differentially expressed genes. (3) Switching to chemically more different hosts is more challenging. (in contrast to (2)) The use of hosts with very distinct structural and chemical properties (e.g. Switch B), often requires more transcriptional and physiological changes to regain cellular homeostasis that will manifest in a higher number of differentially expressed genes. (4) Differences in gene expression profiles will have a physiological correspondence. Depending on the extent of required adjustments, the switch to certain hosts can be associated with higher costs, which can result in a reduced larval performance.