4.Discussion
Earlier results from Polygonia c-album and related species have suggested that adaptations to host plants can be understood as modules of host-specific gene expression and the resulting phenotypic traits (Celorio-Mancera et al. 2023). In the present study, it could now be shown that the larvae of such polyphagous butterflies can also “switch” between these modules during their individual development to adjust their molecular and cellular processes to a new host plant. As predicted, 2 hours after the switch, the larvae still showed gene expression patterns that corresponded to Host 1 (i.e. category 2), while Host 2-specific profiles, that were expected from the comparison of the non-switched controls (i.e. category 1), were predominantly found after 17 hours (in support of Hypothesis 1 in the Introduction). The extent of transcriptional adjustment depends strongly on the similarity of the host plants (in support of Hypotheses 2 and 3). For some comparisons even after 17 hours only very few differentially expressed genes were found, that corresponded to expression profiles expected for the second host. The overall low number of differentially expressed genes between some plants may result from shared modules and, consequently, the use of the same or similar downstream processes. This agrees with a high similarity of the host plants used in Switch A. Urtica dioica andUlmus glabra both belong to the “urticalean rosids”, whileSalix caprea and Ulmus glabra have the same growth form. It was shown that both a close phylogenetic relationship, as well as similar physical appearance can explain similarities in the chemical properties and defense mechanisms (Feeny 1976) and, thus, suggest the involvement of overlapping modules. Such similarities in the gene expression response to these hosts were already indicated by Heidel-Fischer et al. (2009) and Celorio-Mancera et al.(2012, 2023). The comparatively small number of transcriptional differences that were found in switches between Urtica andSalix is moreover consistent with the overlapping gene expression profiles that were previously shown by Celorio-Mancera et al . (2023).
In other switches, a high number of expression differences would, in turn, indicate the involvement of more specific and exclusive gene expression profiles and developmental pathways. The high number of transcriptional differences that were found in switches that involvedRibes and especially between the phylogenetically and chemically diverse Urtica and Ribes (Switch B), is consistent with the overall opposing gene expression profiles between these two hosts (Celorio-Mancera et al. 2023).
Based on these transcription profiles, it could be assumed that switches containing Ribes indeed represent a more complex situation. The differences between the different directions of the host switches, further indicated that switches to Ribes appear to be more challenging than the reverse. This could be related to the special chemical composition of Ribes . The leaves of Ribes species are described to contain cyanogenic compounds (Hegnauer 1973), which can act as a potent chemical defense against several herbivores (Bellotti and Arias 1993; Fürstenberg-Hägg et al . 2013). The gene expression patterns found here support this assumption. Most of the genes with a significantly different expression on Ribes were mainly involved in metabolic processes (Table S1). This was further confirmed by a closer examination of the upregulated genes onRibes in comparison to Urtica . In addition, an increased expression was recorded in genes that were directly or indirectly associated with metabolism and detoxification (e.g. Aldo-keto reductase, Glutathione s-transferase 1-like).
In agreement with Celorio-Mancera et al . (2023) an upregulation of the proline dehydrogenase was found, which besides the metabolic degradation of proline is also involved in energy recovery and response to (micro)environmental stress (Phang, Liu and Zabirnyk 2010; Servet 2012). Furthermore, an upregulation of spermine oxidase like onRibes could be shown again. Spermine oxidase is involved in the metabolic processing of polyamines (Chang et al. 2021). Moreover, it was also described to play a role in the response to oxidative stress, which, in addition to the upregulation of a Methionine sulfoxide reductase (HMEL008564-PA) may indicate a role in the reaction to cyanogenic glycosides found in Ribes leaves.
An involvement of serine proteases could also be observed. Proteases have been shown to play an important role in the breakdown of various plant proteins and, thus, contribute to the digestion and nutrient absorption of host plant material (Celorio-Mancera et al. 2013; Chikate et al. 2013). Moreover, proteases are involved in overcoming chemical plant defenses like protease inhibitors (Ryan 1990). For this, larvae can not only express a wide range of different proteases (e.g. serine proteases and serine-type endopeptidase) but also regulate their expression according to a specific host plant (Chouguleet al. 2005; Celorio-Mancera et al. 2013; Chikate et al. 2013; Vogel et al. 2014). Such host-specific involvement of proteases has previously been shown for P. c-album(Celorio-Mancera et al . 2013). Our results further support these previous findings, and strengthen the assumption that proteases are crucial candidates for the usability and processing of a broad host repertoire.
Besides an upregulation of these genes, increased expression has also been found in genes that are responsible for the development of the gut peritrophic matrix, which are rather a physiological adaptation to the usability of a host (Celorio-Mancera et al 2023; this study). The peritrophic matrix is a membranous layer that encases the food bolus and thus separates the ingested material from the gut epithelium (Lehane 1997). By this it represents an important protective barrier, as it prevents plant toxins and other harmful substances from entering the other body tissue (Puninean et al. 2010, Celorio-Mancera et al. 2013). An upregulation of the peritrophic matrix development can now either indicate a strengthening of the gut wall as a defensive mechanism, or the initiation of processes that repair damages from interactions with toxic components (see Celorio-Mancera et al.2013 for details; see also Pérez-Hedo et al. 2012). Previous studies already showed differential expression of genes involved in structural constituents of the cuticle as a response to Ribes(Celorio-Mancera et al. 2013). An increased production of chitin to strengthen or repair the intestinal tissue, thus, seems to be required for the usability of Ribes after a switch.
The downregulation of Sialin-like after a switch to Ribes can be interpreted as a transient response to a previous upregulation onUrtica . This is in agreement with its upregulation onUrtica shown in the earlier study by Celorio-Mancera et al. (2023). Sialin-like is likely involved in nitrate transport andUrtica is characterized by a high nitrate content (Hegnauer 1973; Taylor 2009). When switching larvae to Ribes , they quickly have to react to compensate for the lower nitrate concentration. This also agrees with the upregulation of transmembrane proteins on Urtica.Transmembrane proteins are responsible for the exchange of substances into and out of the cell and can, thus, play an important role in the maintenance of cellular concentrations as well as signaling.
The “unexpected” differences after the switches (i.e. category 3), were mostly involved in metabolism. However, a considerable proportion also included genes that played a role in signaling and responses to different (stress) stimuli, which further indicates a stress reaction to the switch itself. The repeated observation of the upregulation of some specific genes further implies their role in adaptation to a particular host plant. In contrast to the present project, previous studies measured the gene expression of larvae that were reared on the same host plant during their entire development (Heidel-Fischer et al. 2009; Celorio-Mancera et al. 2013, 2023). Here, changes in the transcriptional profiles were measured in short time intervals after the switch to a new host. The up- or downregulation of a particular gene can, thus, be interpreted as a direct response to a specific host, further confirming its contribution to the digestibility and usability of the plant. These genes now represent important candidates that could be targeted during knockout experiments to further determine their role in host plant adaptations.
The discrepancy in the gene expression between Urtica andSalix in the two experiments may be due to annual variation in host quality. The two switch experiments were carried out in different years. Differences in temperature as well as in the amount of precipitation are likely to influence the physical and chemical properties of host plants (De Bruyn, Scheirs and Verhagen 2002; Scheirs and de Bruyn 2005; Liu et al. 2023). These variations could then cause the differences in the transcriptional response between years.
Despite the transcriptional differences, host switches did not affect the performance of the larvae (i.e. Hypothesis 4 was not supported). Previous studies on sequential diet shifts in P. c-album already showed that caterpillars can switch easily between Urtica andSalix without major costs for their performance (Söderlind 2012). The present study now also showed that in switches that were assumed to be more challenging, larvae could still adjust to the new host without significant physiological and developmental deviations. Even when moved to Ribes larval performance did not differ from those in other treatments. This suggests that also at a phenotypic level, larvae can flexibly adapt to the new host environment.
The ability to adjust the transcriptional and phenotypic response demonstrates that the usability of a host is not determined by the experience of larvae during their early development. This quick adjustment of gene expression allows the caterpillar to flexibly switch to an alternative host and, therefore, to compensate for poor host quality due to, for instance, rapidly changing environmental conditions or oviposition mistakes. The larvae of P. c-album have been described to be relatively mobile and can cover substantial distances even as neonates (Nylin et al. 2000, Schäpers et al.2016). It is, thus, possible that the caterpillars under some circumstances can move to an alternative host plant if necessary (cf. Gamberale-Stille et al 2014). Moreover, this plasticity can make caterpillars more resistant to changes within a host plant. Reduced host quality, seasonal changes of the chemical composition, as well as an increase in potential defensive chemicals (e.g. induced defenses) may require altered or different cellular and physiological processes for the further use of a plant.
The described plasticity in gene expression is, however, perhaps more important in an evolutionary context. The findings support the assumption that the colonization of new plants does not require entirely new adaptations, but that existing adaptations to established hosts can act as preadaptations and facilitate the colonization of novel plants (ecological fitting; Janzen 1985; Agosta 2006; Agosta et al.2010). Nevertheless, the switch to some hosts still seems to be more complicated than to others and it takes longer until the state of homeostasis is reached when larvae are switched to such plants. Although we did not find any significant performance costs of the hosts switches, it should not be ruled out that physiological trade-offs were involved in the evolution of the host repertoire of Polygonia butterflies at some point. In line with this interpretation, the challenging switch to Ribes in the present study seems to be reflected in evolutionary patterns, as butterflies outside of Polygonia rarely have been able to colonize this host.
In conclusion, investigating the immediate response to a host switch helps characterizing genetic modules (sensu Celorio-Mancera et al. 2023) associated with host plant metabolism and detoxification, in that it disentangles direct responses to the plant from overall downstream effects. In addition, we propose that host switch experiments in the laboratory followed by transcriptomic investigations can be a valuable tool to examine not only plasticity in host use but also subtle and/or transient trade-offs in the evolution of host plant repertoires.