Complex plant quality—microbiota–population interactions modulate the response of a specialist herbivore to the defence of its host plant

This is a self-archived version of an original article. This version may differ from the original in pagination and typographic details. Author(s): Title: Year: Version: Copyright: Rights: Rights url: Please cite the original version: CC BY 4.0 https://creativecommons.org/licenses/by/4.0/ Complex plant quality—microbiota–population interactions modulate the response of a specialist herbivore to the defence of its host plant © 2022 The Authors. Functional Ecology published by John Wiley & Sons Ltd on behalf of British Ecological Society Published version Minard, Guillaume; Kahilainen, Aapo; Biere, Arjen; Pakkanen, Hannu; Mappes, Johanna; Saastamoinen, Marjo Minard, G., Kahilainen, A., Biere, A., Pakkanen, H., Mappes, J., & Saastamoinen, M. (2022). Complex plant quality—microbiota–population interactions modulate the response of a specialist herbivore to the defence of its host plant. Functional Ecology, 36(11), 2873-2888. https://doi.org/10.1111/1365-2435.14177 2022 Functional Ecology. 2022;36:2873–2888. | 2873wileyonlinelibrary.com/journal/fec Received: 26 November 2021 | Accepted: 26 August 2022 DOI: 10.1111/1365-2435.14177 RESEARCH ARTICLE Complex plant quality— microbiota– population interactions modulate the response of a specialist herbivore to the defence of its host plant Guillaume Minard1,2,3 | Aapo Kahilainen1,4 | Arjen Biere5 | Hannu Pakkanen6 | Johanna Mappes1,7 | Marjo Saastamoinen1,8 1Organismal and Evolutionary Biology Research Programme, University of Helsinki, Helsinki, Finland; 2Université de Lyon, Lyon, France; 3Ecologie Microbienne, UMR CNRS 5557, UMR INRA 1418, VetAgro Sup, Université Lyon 1, Villeurbanne, France; 4Finnish Environment Institute, Biodiversity Centre, Helsinki, Finland; 5Department of Terrestrial Ecology, Netherlands Institute of Ecology (NIOOKNAW), Wageningen, The Netherlands; 6Department of Chemistry, University of Jyväskylä, Jyväskylä, Finland; 7Department of Biological and Environmental Science, University of Jyväskylä, Jyväskylä, Finland and 8Helsinki Institute of Life Sciences, University of Helsinki, Helsinki, Finland This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited. © 2022 The Authors. Functional Ecology published by John Wiley & Sons Ltd on behalf of British Ecological Society. Correspondence Guillaume Minard Email: [email protected] Funding information Academy of Finland, Grant/Award Number: 265641, 273098 and 320438; H2020 European Research Council, Grant/Award Number: 637412 Handling Editor: Sergio Rasmann Abstract 1. Many specialist herbivores have evolved strategies to cope with plant defences, with gut microbiota potentially participating to such adaptations. 2. In this study, we assessed whether the history of plant use (population origin) and microbiota may interact with plant defence adaptation. 3. We tested whether microbiota enhance the performance of Melitaea cinxia larvae on their host plant, Plantago lanceolata and increase their ability to cope the defensive compounds, iridoid glycosides (IGs). 4. The gut microbiota were significantly affected by both larval population origin and host plant IG level. Contrary to our prediction, impoverishing the microbiota with antibiotic treatment did not reduce larval performance. 5. As expected for this specialized insect herbivore, sequestration of one of IGs was higher in larvae fed with plants producing higher concentration of IGs. These larvae also showed metabolic signature of intoxication (i.e. decrease in Lysine levels). However, intoxication on highly defended plants was only observed when larvae with a history of poorly defended plants were simultaneously treated with antibiotics. 6. Our results suggest that both adaptation and microbiota contribute to the metabolic response of herbivores to plant defence though complex interactions. KEYWORDS herbivore, Lepidoptera, microbiota, plant defence, trophic interactions 13652435, 2022, 11, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/1365-2435.14177 by University Of Jyväskylä Library, Wiley Online Library on [03/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 2874 | Functional Ecology MINARD et al. 1 | INTRODUCTION In most ecosystems, communities are regulated through a network of trophic interactions (Fretwell, 1987). In these networks, herbivore– plant interactions play a central role and often drastically impact the dynamics of the rest of the community (e.g. bottomup interactions with predators, parasites or topdown interactions with detritivores), as well as matter fluxes and nutrient cycles (Fretwell, 1987; Metcalfe et al., 2014). Both partners of this interaction are involved in a coevolutionary arms race, in which the plant evolves new traits against herbivores, while in turn, herbivores adapt to these traits to be able to consume the plant (Ehrlich & Raven, 1964). Coevolutionary theory predicts that such forces lead to the diversification of both partners, and symmetric diversification patterns have been observed between plants and herbivores across many empirical data (Futuyma & Agrawal, 2009). Being sessile, plants are easily exposed to herbivory. Instead, plants have developed an array of chemical and physical traits as defence mechanisms against herbivores that act either constitutively or after induction [reviewed by Aljbory & Chen (2018)]. Many plants produce secondary metabolites that aim to reduce the performance of the herbivores through direct (repellent, toxins) or indirect defence (attract another trophic level such as a predator or a parasitoid of the herbivore). Among compounds involved in direct defence, terpenoids are by far the most diverse and ubiquitous class with more than 30,000 different molecules (Mithöfer & Boland, 2012). They are followed by alkaloids (~12,000 molecules) and phenolic compounds (~9000 molecules). The functions of defensive metabolites are diverse and often poorly characterized. As an example, some of them alter the digestion, induce oxidative stress and cytotoxicity (Bhonwong et al., 2009; Treutter, 2005). Due to their common dual antiherbivory and antimicrobial activity, recent studies have suggested that in certain cases, the main target of defensive compounds may not be the herbivore itself but instead its gut microbiota (Hammer & Bowers, 2015). Modification of the herbivore's microbiota could then negatively impact the physiology and consequently the performance of the herbivore feeding on the defended host plant (Hammer & Bowers, 2015; Mithöfer & Boland, 2012). Conversely, the gut microbiota may also be involved in the detoxification of defensive compounds, and thus contribute to the adaptation of specialist herbivores to highly defended host plants (Hammer & Bowers, 2015). For example, the microbiota of pine weevils and bark beetles help their host to degrade pine tissues rich in terpenoid resins or phenolic compounds (Berasategui et al., 2017; Cheng et al., 2018). These coleopterans can, thus, only interact efficiently with their host plant when they form a holobiont composed of the insect host and specialized microorganisms. Similar patterns may also be evident in Lepidoptera. Recent studies have, however, suggested that Lepidoptera harbour a transient and highly variable microbiota with limited impact on host performance and primary metabolism (Duplouy et al., 2020; Hammer et al., 2017; Whitaker et al., 2016). Conversely, few studies have indicated that this relatively transient microbiota can contribute to lepidopteran immunity (Duplouy et al., 2020; Galarza et al., 2021; Mason et al., 2019; Yoon et al., 2019), that some microbes might lead to transgenerational responses and facilitate host plant shifts (Voirol et al., 2020), or in contrast elicit host plant defences (Wang et al., 2017). Interactions regarding potential detoxification of plant defensive compounds, however, remain poorly studied. In order to further understand the potential relevance of host plant defence– herbivore– microbiota interactions, we used a wellcharacterized plant– herbivore study system represented by the host plant ribwort plantain (Plantago lanceolata) and its specialist herbivore, the Glanville fritillary (Melitaea cinxia) butterfly. The bioactive metabolites of Plantago lanceolata have previously been investigated (Tamura & Nishibe, 2002) and P. lanceolata is known to synthesize aucubin and catalpol, two monoterpenoids that belong to the class of iridoid glycosides (IGs) that carry antibacterial, antifungal and antipredatory properties (Baden & Dobler, 2009; Davini et al., 1986; Marak et al., 2002; Reudler et al., 2011). The molecules are not active while formed in the plant and are compartmentalized within the vacuole. Whenever a herbivore or a pathogen degrades the plant cell, the plant will release the IGs that could then react with the βglucosidases of either the plant, the herbivore or the pathogen to form reactive aglycones (Kim et al., 2000; Pankoke et al., 2013). In another system, it has been demonstrated that the aglycone links proteins through an activated dialdehyde that forms protein adducts and reacts with the side chain of lysine, an essential amino acid (Mander & Liu, 2010). This reaction decreases the amount of lysine available, denatures digestive enzymes and decreases the nutritional value of the food to herbivores and pathogens. The same dialdehyde structure appears after activation of aucubin by βglucosidases and consistently results in decreased levels of lysine (Kim et al., 2000; Konno et al., 1999). Some specialist herbivores have, however, adapted to cope with iridoid glycosides, and even store them for their own defence against parasites and predators (Bowers & Puttick, 1986). Previous studies in M. cinxia have shown that prediapause larvae contain between 0.2% and 2% of IGs, mostly catalpol (Nieminen et al., 2003; Suomi et al., 2001, 2003). All studies so far have been assessing an entire individual, and thus, it remains unclear whether larvae actively transport and store the IG compounds in specific tissues or whether the compounds are maintained within the gut. However, as 1day starved larvae as well as adult butterflies (which do not feed anymore on the host plant) contain equivalent levels of IGs, it has been suggested that the compounds are actively sequestered (Nieminen et al., 2003; Suomi et al., 2003). This may contribute to the higher tolerance of larvae to their natural enemies when they are feeding on highly defended plants (Laurentz et al., 2012; Nieminen et al., 2003; Reudler et al., 2011). Studies assessing the impacts of high IG levels in host plants on larval performance vary between no effects to a small increase in weight and development rate (Laurentz et al., 2012; Reudler et al., 2011; Saastamoinen et al., 2007). In addition to IGs, P. lanceolata also produce the high levels of a phenolic compound, the phenylethanoid glycoside verbascoside (alternatively named acteoside) (Tamura & Nishibe, 2002), 13652435, 2022, 11, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/1365-2435.14177 by University Of Jyväskylä Library, Wiley Online Library on [03/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License | 2875 Functional Ecology MINARD et al. which antimicrobial and antiherbivory properties remain controversial (Fazly Bazzaz et al., 2018; Holeski et al., 2013; Pardo et al., 1993; Reichardt et al., 1988). In this study, our first objective was to investigate whether the microbiota of M. cinxia affect the performance of larvae. We selected prediapause larvae since they live as gregarious families on a single host plant (contrarily to postdiapausing larvae) and their condition partially explains their survival probability during diapause (Duplouy et al., 2018; Kahilainen et al., 2018; Kuussaari & Singer, 2017; Tack et al., 2015). Our second objective was to evaluate whether any differences in performance are mediated by effects of the microbiota on the ability of the host larvae to cope with the defences of their host plants. Our third objective was to assess whether the composition of the larval gut microbiome and its effect on larval performance differ among larvae whose parents originated from populations with a different history of host plant use. In order to investigate the impact of larval microbiota on larval performance and the ability of larvae to cope with IGs, we addressed the following three questions: (1) Do microbiota confer an advantage to their host larvae? For this, we used leaf antibiotic treatments to manipulate the larval microbiome and tested whether larvae feeding on untreated leaves had higher performance (survival, development rate, biomass production) than larvae feeding on antibioticstreated leaves. (2) Can any advantages conferred by microbiota be related to mitigation of the impact of ingested host IGs? For this, we used selection lines of P. lanceolata, in which plants had been selected for high or low constitutive levels of leaf IGs (highIG and lowIG lines) (Marak et al., 2000). We assessed the metabolomes (including levels of IGs and lysine) of larvae fed on antibioticstreated and untreated leaves of highIG and lowIG plants to test (i) whether larvae fed highIG lines had higher levels of IGs and reduced levels of lysine, (ii) whether microbiota mitigated a reduction in lysine levels and (iii) whether larvae feeding on untreated highIG plants had higher performance than larvae of antibioticstreated highIG plants. (3) Does the parental origin of the larvae affect the microbiome composition of the larvae and its effect on larval performance and ability to cope with IGs? For this, we used M. cinxia larvae whose parents originated from populations in two distinct regions (Eckerö and Sund) within the Åland archipelago in Finland that differ in their evolutionary history of host use. Populations from Eckerö have a history of encountering lower defence levels due to the presence of two host plant species, Veronica spicata and Plantago lanceolata, that generally contain low and high levels of IGs respectively. By contrast, larvae originating from Sund have a history of encountering only the more highly defended host plant P. lanceolata. We studied the composition of their microbiota and performance after being exposed to contrasted IGs concentrations. We thus specifically tested (i) whether the assembled gut bacterial and fungal microbiome of larvae differs between parental population origin and (ii) whether larvae originating from regions with a stronger evolutionary history with highIG plants (Sund) benefit more from their associated microbiome when feeding on highIG plants than larvae from regions with a weaker evolutionary history with highIG plants (Eckerö). 2 | MATERIALS AND METHODS 2.1 | Plant selection lines and rearing The seeds of the P. lanceolata selection lines were obtained from the Netherlands Institute of Ecology (NIOOKNAW). The two lines of plants have been selected to produce either high or low concentrations of Iridoid Glycosides (IG), as previously described (Marak et al., 2000). A total of six genotypes (three from each line) were selected from a pool of 20 genotypes (see Supplementary methods). The leaves of these plants were used to feed the larvae during a period of 27 days according to the experimental design described below (Figure 1). To track any variation that may have occurred during the experiment, samples from the host plants were randomly harvested from six plants every 3 days, immediately freezedried and stored at −80°C for metabolomics analysis. 2.2 | Experimental design and measurements Parental lines of the butterflies were sampled in the spring of 2016 at the end of larval diapause (fifth and sixth instar larvae) in two distinct regions (Eckerö and Sund) located in two large islands of the Åland archipelago, Finland. No permit was required to collect this species. Within each region, the individuals were collected from three habitat patches belonging to the same semiindependent habitat network (Hanski et al., 2017). Larvae from the two regions have previously been shown to be genetically differentiated (Nair et al., 2016) and they also show ecological differences: habitats from Eckerö are colonized by two host plant species of the butterfly, P. lanceolata and Veronica spicata, of which the butterfly prefers V. spicata in these regions. The habitats from Sund contain only P. lanceolata and the butterflies also slightly prefer this species when both host plants are provided in a choice test (Kuussaari et al., 2000). It is important to note that the two plant species produce the same IG compounds, but IG levels in V. spicata are more than twofold lower than in P. lanceolata (Saastamoinen et al., 2007). For the present experiment, five females from Sund and four females from Eckerö were mated with a male collected from the same region but from a different local population (in order to avoid inbreeding). Females were allowed to lay eggs and the larvae hatching from these eggs (females lay clutches of eggs) were used for the feeding assay: A fullfactorial design in which a group of 20 individuals from each family was assigned to each treatment (see Figure 1 for a scheme of the experimental design). In total, 2160 larvae were used (6 plant families × 9 herbivore families × 2 antibiotics treatments × 20 larvae). In order to avoid clutch effects, larvae from different clutches were randomly distributed across the treatments. Each group was fed with leaves from one of the six selected plant genotypes that were either supplemented with 200 μl of water (Control) or 200 μl of an antibiotic solution (Antibiotic) described below. The plant leaves were harvested, rinsed with water and a piece of 1.7 cm2 of plant leaves was deposited on a sterile petri dish containing 20 larvae. Leaves 13652435, 2022, 11, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/1365-2435.14177 by University Of Jyväskylä Library, Wiley Online Library on [03/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 2876 | Functional Ecology MINARD et al. from the same plant genotype were collected from at least three individual plants to provide food to the larvae and the same plant was not used in two successive days in order to limit the enhancement of plant defensive response. Water or antibiotic solutions were deposited at the surface of the plant leaf piece. This procedure was executed every day for each petri dish. The antibiotic solution contained three antibacterial and two antifungal compounds and was prepared according to the previously published method with some modifications (Chung et al., 2013). The composition of the antibiotic solution was 2 × 10−4 g ml−1 of neomycin sulphate, 1 × 10−3 g ml−1 aureomycin and 6 × 10−5 g ml−1 streptomycin and 8 × 10−4 g ml−1 methyl paraben, 6 × 10−4 g ml−1 sorbic acid. As we were able to divide each initial larval family (eggs from a single female) into twelve 20 larvae groups, each family was fed with all of the six selected genotypes of P. lanceolata with or without antibiotics (i.e. splitfamily design). Different egg clutches were used for this analysis. The development and survival of the larvae to the second instar was recorded in order to characterize larval lifehistory traits. The larvae that survived and reached the third instar were sampled, snap frozen and stored at −80°C. The mass of a group of three third instar larvae per family per treatment was recorded, and the rest of the larvae were used for metabarcoding, qPCR and metabolomics analysis. The larvae were reared in groups and group size was recorded as it is known to impact larval performance in the gregarious phases of the larval development (Saastamoinen, 2007). All the samples were prepared under a flow hood with sterile material and proper protection in order to limit contaminations. The larvae were dissected in 1X PBS (GIBCO) on a platform refrigerated with ice <0°C to limit IG conversions. Dissected guts were stored in 1X PBS and carcasses (including the dissection buffer) were stored without any additional buffer. Dissected body parts were immediately frozen in liquid nitrogen and stored at −80°C. Guts of three M. cinxia individuals per family, plant genotype and treatment were pooled. Those pools were used to identify the microbial communities colonizing each conditions through bacterial and fungal metabarcoding approaches and qPCR quantifications (see details in Supplementary methods). In addition, three 0.5cm2 leaf pieces per plant genotype were analysed with the same approach. FIGURE 1 Experimental design. (a) Larval families were obtained from parents collected as larvae in Åland islands. Males and females that were mated came from different habitat patches belonging to different population networks. Five and four families were obtained from individuals collected in Eckerö and Sund, respectively. (b) Each larvae represent a group of 20 individuals that were daily fed with a leaf of Plantago lanceolata collected from a plant genotype selected to produce either high (H2, H3, H9) or low (L4, L6, L7) levels of plant defence (iridoid glycosides or IGs). (c) The larval performance and survivorship of larvae were recorded. Prokaryotic and eukaryotic microbiota of plant leaves and larval gut were investigated through metabarcoding and qPCR. The metabolism linked to plant defence and the insect response was investigated through 1HNMR and LCESI MS 13652435, 2022, 11, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/1365-2435.14177 by University Of Jyväskylä Library, Wiley Online Library on [03/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License | 2877 Functional Ecology MINARD et al. The metabolome of the larvae and the plant, carcasses from the pooled individuals were analysed with ESIMS and 1HNMR. Metabolome from 3.5 mg of freezedried plant leaves from each genotype was analysed with 1HNMR (see details in Supplementary methods). The experimental design and measurements are summarized in Figure 1. 2.3 | Statistical analysis All analyses were computed using R (R Core Team, 2016). Factors significantly impacting the plant metabolites and the larval microbiota or metabolites were tested with a permutational multivariate analysis of variance (adonisANOVA) with the package vegan (Oksanen et al., 2013). The relative concentrations of metabolites were scaled and centred prior to the analyses in order to limit the impact of highly abundant metabolites. Global variations in the plant metabolite composition across plant IG selection lines (‘Plant line’), plant genotypes within selection lines (‘Plant genotype’) or time was conducted through redundancy analysis (RDA). Differences in the specific plant metabolites between plant lines, plant genotypes and time were assessed using a linear model (LM). OTUs explaining the most part of the variation in the larval microbiota were represented through principal coordinate analysis (PCoA) based on Bray– Curtis dissimilarities in community composition between samples. Microbial and metabolite composition across larvae were represented with bar plots and heat maps. Comparisons of metabolites and microorganisms present in larval gut samples across the two sites of parental origin of the larvae (‘Larval origin’), the two plant lines and the two antibiotic treatments were conducted with the package lme4 (Bates et al., 2017, p. 4) by computing linear mixed models (LMM) with two random effects (i) plant genotype nested within plant line and (ii) larval family nested within larval origin. The impact of covariates (i.e. the two sites of parental origin of the larvae, the two plant lines and the two antibiotic treatments) on larval lifehistory traits was compared by using an LMM (mass) or a generalized linear mixed model (GLMM) with a Poisson distribution and a loglink function (development time) or a binomial distribution and a probit link function (survival). Plant genotype nested within plant line and larval family nested within larval origin were used as random effect. Tukey post hoc tests were conducted to test for differences between levels with the emmeans package (Lenth et al., 2022). 3 | RESULTS 3.1 | Do microbiota confer an advantage to their host larvae? 3.1.1 | Broadspectrum antibiotic treatment induces an erosion of the microbiota towards specific taxa The bacterial and fungal gut communities of the larvae were mainly affected by the antibiotic treatment (Table 1). The fungal community being, however, more affected than the bacterial community with a drastic reduction of the community similarity variation after treatment (R2 = 0.086 and 0.237, respectively, for bacteria and fungi; Figure 2a,e; Figure S3). When considering the OTU levels, bacterial and fungal communities were mostly influenced by the variation in CommunityaFactors df PseudoF R2p Bacteria Origin 12.586 0.023 0.003** Category 12.369 0.021 0.009** Treatment 19.767 0.086 0.001*** Origin × Category 11.192 0.011 0.202 Origin × Treatment 11.370 0.121 0.111 Category × Treatment 11.711 0.150 0.042* Origin × Category × Treatment 10.762 0.007 0.719 Residuals 94 0.826 Fungi Origin 10.349 0.003 0.901 Category 10.186 0.001 0.992 Treatment 131.670 0.237 0.001*** Origin × Category 10.504 0.004 0.757 Origin × Treatment 11.324 0.010 0.226 Category × Treatment 11.104 0.008 0.305 Origin × Category × Treatment 11.572 0.012 0.173 Residuals 94 0.725 aBray– Curtis dissimilarity distances. . p ≤ 0.1, *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001. TABLE 1 Analysis of the factors influencing the global composition of the larval microbiota 13652435, 2022, 11, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/1365-2435.14177 by University Of Jyväskylä Library, Wiley Online Library on [03/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 2878 | Functional Ecology MINARD et al. (d) (a) (b) ** * **** ** * * ****** ** ** ****** ****** * ** + Antibiotic -Antibiotic Eckerö Sund Abundance Otu0002|Cedecea (log10) Plant Lines + Antibiotic -Antibiotic Eckerö Sund Abundance Otu0005|En terobacteriaceae (log10) Plant Lines + Antibiotic -Antibiotic Eckerö Sund ecnadnubAgol(ainiwrE|3000utO 10) Plant Lines + Antibiotic -Antibiotic Eckerö Sund ecnadnubAmuiropsodalC|5000utOxelpmoc(log10) Plant Lines (c) (g) (f) + Antibiotic -Antibiotic Eckerö Sund AbundanceOtu0001|Us tilaginaceae (log10) Plant Lines MDS1 MDS2 Eckerö H + Antibiotic Eckerö H -Antibiotic Eckerö L-Antibiotic Eckerö L+ Antibiotic Sund H + Antibi otic Sund H -Antibiotic Sund LAntibiotic Sund L+ Antibi otic Eckerö H + Antibi otic Eckerö H -Antibiotic Eckerö L-Antibiotic Eckerö L+ Antibiotic Sund H + Antibiotic Sund H -Antibiotic Sund L-Antibiotic Sund L+ Antibiotic MDS1 MDS2 1.0 0.5 0.0 -0.5 -1.0 -0.5 0.00.51.0 10.0 7.5 5.0 2.5 0.0 10.0 7.5 5.0 2.5 0.0 HLHL 10.0 7.5 5.0 2.5 0.0 10.0 7.5 5.0 2.5 0.0 HLHL 10.0 7.5 5.0 2.5 0.0 10.0 7.5 5.0 2.5 0.0 HLHL 10.0 7.5 5.0 2.5 0.0 10.0 7.5 5.0 2.5 0.0 HLHL (e) 5 4 3 2 0 HLHL 1 5 4 3 2 0 1 2 1 0 -1 -4 -2 -2 02 13652435, 2022, 11, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/1365-2435.14177 by University Of Jyväskylä Library, Wiley Online Library on [03/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License | 2879 Functional Ecology MINARD et al. the abundance of three and two OTUs, respectively, that were also the most abundant taxa retrieved from the larval gut (Figure S2). These OTUs were identified as Cedecea sp. (Otu0002), Erwinia sp. (Otu0003) and an unclassified Enterobacteriaceae (Otu0005) for the bacteria, and as Ustilaginaceae unclassified (Otu0001) and Cladosporium complex (Otu0005) for the fungi. Reduction in the density of these dominant bacterial OTUs between antibiotic treated and nontreated larvae was only observed for larvae originating from the region Eckerö within the archipelago, while larvae from both regions (Eckerö and Sund) experienced reduction in the density of the most dominant fungal OTUs (Figure 2 b – g ; Table S2). 3.1.2 | Larvae show little differences in performance and survivorship Larvae fed with antibiotic did not consistently show divergence in their development time and mass (p = 0.340 and 0.427, respectively, see details Table 2) in comparison with nontreated larvae. Only survival was significantly impacted by the treatment with a slight increase in survival of individuals from Eckerö fed with lowIG plant and antibiotics (Figure 3a). The larval body mass (estimated after collection at L3) was not impacted by the antibiotic treatment or by the IG selection line category of the host plant they were fed with (Table 2, Figure 3b). Larvae originating from the different regions showed no difference in survival, development time (until L3) or body mass. The development time of larvae fed with highIG lines was significantly longer (Table 2), but this difference was smaller than one day on average (12.8 ± 0.28 days for the HighIG lines and 12.0 ± 0.26 for the lowIG lines) and contrast Tukey tests (that are more conservative) did not evidence no significant differences in development among individuals belonging to both regions (Figure 3c). 3.2 | Can any advantages conferred by microbiota be related to mitigation of the impact of ingested host IGs? 3.2.1 | IG selection leads to broad plant metabolome variations Overall, selection for highIG and lowIG explained 27% of the metabolite variation among the plants (Table 3; Figure S4). In addition, the composition of the plant metabolites varied among the plant genotypes and changed over time (Table 3; Figure S4). The two defensive compounds (aucubin, catalpol) targeted by the selection process were confirmed to be lower in lowIG lines compared to highIG lines (difference for aucubin relative concentration: CI 95% = −0.045 to −0.020; F1,50 = 31.97; p = 7.4 × 10−7; catalpol: CI 95% = −0.055 to −0.033; F1,50 = 84.94; p = 2.33 × 10−12). A different defensive compound (i.e. verbascoside) showed also variation in response to the selection treatment but was not regarded in this study (Figure S4). 3.2.2 | Larvae from different origins show a signature of protein alteration after being fed with high levels of iridoid glycosides and this effect is accentuated in the presence of microbiota Comparisons of the metabolite composition in larvae was conducted on pooled carcasses of three individuals after gut removal. Generally, levels of aucubin and catalpol stored within larval carcasses were too low to be efficiently assessed through 1HNMR spectra (Figure S5). Therefore, these analyses were performed using LCESIMS. The results showed that larvae stored low concentrations of IGs (36.56 ± 39.90 and 21.97 ± 20.89 ng.mg−1 or 10−4% for aucubin and catalpol, respectively). These values can, however, be slightly underestimated since the larvae were dissected on ice, and thus, some metabolite conversion may have occurred resulting in decreased IG concentration in the carcasses. None of the studied factors significantly impacted the aucubin concentration within larval carcasses (Table 4), whereas larvae that had fed on highIG lines stored more catalpol than those that had fed on lowIG lines (Figure 4a,b). This result suggests that the larval storage of IG in the early developmental instars is weak but depends on the concentration of these compounds in the host plant. The activation of IGs consists of a deglycosilation, which leads to the release of glucose and a reactive aglycone that binds lysine residues. We found that the detectable levels of lysine were lower in larvae from Eckerö fed with leaves from highIG lines compared to those fed with leaves from lowIG lines (Table 4; Figure 4c). These results suggest that the iridoid glycosides are at least partially activated in the larvae and that they efficiently reduce the amount of accessible lysine within the individuals. In addition, there was an interaction between plant IG line and the antibiotic treatment; individuals form Eckerö that had been fed on lowIG line host plants harboured a higher amount of lysine when their microbiota were altered by antibiotics compared with untreated individuals (Figure 4c), while no difference was FIGURE 2 Variation in microbiota. (a) The nonmetric multidimensional scaling represents similarities (Bray– Curtis distances) in the bacterial communities across treatment groups. Abundance data are shown for the three most important bacterial OTUs namely (b) Cedecea sp.— Otu0002, (c) Erwinia sp.— Otu0003 and (d) an unclassified Enterobacteriaceae— Otu0005. (e) The nonmetric multidimensional scaling represents similarities (Bray– Curtis distances) in the fungal communities across treatment groups. Abundance data are shown for the two most important bacterial OTUs namely (f) an unclassified Ustilaginaceae— Otu0001, (g) Cladosporium sp.— Otu0005. Plant selection lines that contain highIG are referred as ‘H’ while those containing lowIG are referred as ‘L’. Each dot represents a pool of three individuals from the same petri dish. Different point shapes represent pools of individuals fed with the same plant genotype. Different point colours represent pools of individuals that come from different families. Statistical significances from Tukey post hoc tests are reported for p < 0.05 ‘*’, p < 0.01 ‘**’ and p < 0.001 ‘***’ 13652435, 2022, 11, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/1365-2435.14177 by University Of Jyväskylä Library, Wiley Online Library on [03/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 2880 | Functional Ecology MINARD et al. observed between antibiotic treated and untreated individuals fed with plants producing high levels of IGs. This suggests either that the microbiota contribute to the activation of the IGs in those larvae or that part of the lysine acquired through the food is used by the microbiota of the larvae. It is noteworthy, however, that we cannot exclude the potential direct effect of the antibiotic treatment. Interestingly, this effect was only observed in the larvae originating from Eckerö region that has a history of poorly defended plant consumption (i.e. Veronica spicata) but not on the larvae from Sund that naturally feed on more highly defended plants (i.e. Plantago lanceolata). TABLE 2 Factors influencing the larvae larval performance and survivorship Variable Factoradf χ2p Development to L3 Plant lines (highIG vs. lowIG) 17.612 0.006** Antibiotic treatment 10.911 0.340 Larval origin (Sund vs. Eckerö) 12.329 0.126 P. lines × Ab. treatment 12.127 0.145 P. lines × Larv. origin 10.202 0.652 Ab. treatment × Larv. origin 10.008 0.927 P. lines × Ab. treatment × Larv. origin 10.333 0.564 Survival to L3 Plant lines (highIG vs. lowIG) 10.516 0.473 Antibiotic treatment 113.257 0.0003*** Larval origin (Sund vs. Eckerö) 10.067 0.796 P. lines × Ab. treatment 11.431 0.232 P. lines × Larv. origin 12.633 0.105 Ab. treatment × Larv. origin 10.070 0.792 P. lines × Ab. treatment × Larv. origin 18.383 0.004** df Fp Mass (L3 larvae) Plant lines (highIG vs. lowIG) 1,4.074 0.573 0.491 Antibiotic treatment 1,84.846 0.637 0.427 Larval origin (Sund vs. Eckerö) 1,7.357 1.549 0.251 P. lines × Ab. treatment 1,85.035 1.054 0.308 P. lines × Larv. origin 1,84.933 0.654 0.421 Ab. treatment × Larv. origin 1,84.933 2.379 0.127 P. lines × Ab. treatment x Larv. origin 1,85.126 0.505 0.479 aLarval families nested by origin and plant genotype nested by category were used as random factors. *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001. FIGURE 3 Survival and performance of the larvae across the treatment groups. (a) The survival rate, (b) the dry mass and (c) the development time to L3 have been measured either on average for the whole population (survival) for pools of three individuals (dry mass) or individually (development time). Plant selection lines that contain highIG are referred as ‘H’ while those containing lowIG are referred as ‘L’. Different point shapes represent individuals fed with the same plant genotype. Different point colours represent individuals that comes from different families. Statistical significances from Tukey post hoc tests are reported for p < 0.01 ‘**’ (a) (b) (c) ** + Antibiotic -Antibiotic Eckerö Sund lavivru S) noitroporp( Plant Lines + Antibiotic -Antibiotic Eckerö Sund Dry mass (mg) Plant Lines + Antibiotic -Antibiotic Eckerö Sund Developmenttime to L3 (Days) Plant Lines 1.0 0.5 0.0 1.0 0.5 0.0 LL HHLL HHLL HH 10.0 7.5 5.0 2.5 0.0 10.0 7.5 5.0 2.5 0.0 30 20 10 0 30 20 10 0 13652435, 2022, 11, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/1365-2435.14177 by University Of Jyväskylä Library, Wiley Online Library on [03/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License | 2887 Functional Ecology MINARD et al. Kim, D.- H., Kim, B.- R., Kim, J.- Y., & Jeong, Y.- C. (2000). Mechanism of covalent adduct formation of aucubin to proteins. Toxicology Letters, 114(1), 181– 188. h t t p s : / / d o i . o r g / 1 0 . 1 0 1 6 / S 0 3 7 8 - 4 2 7 4 ( 9 9 ) 0 0 2 9 5 - 7 Konno, K., Hirayama, C., Yasui, H., & Nakamura, M. (1999). Enzymatic activation of oleuropein: A protein crosslinker used as a chemical defense in the privet tree. Proceedings of the National Academy of Sciences of the United States of America, 96(16), 9159– 9164. Kuussaari, M., Singer, M., & Hanski, I. (2000). Local specialization and landscapelevel influence on host use in an herbivorous insect. Ecology, 81(8), 2177– 2187. https://doi.org/10.1890/00129658(2000)081[2177:LSALL I]2.0.CO;2 Kuussaari, M., & Singer, M. C. (2017). Group size, and egg and larval survival in the social butterfly Melitaea cinxia. Annales Zoologici Fennici, 54(1– 4), 213– 223. https://doi.org/10.5735/086.054.0119 Lampert, E. C., & Bowers, M. D. (2010). Host plant influences on iridoid glycoside sequestration of generalist and specialist caterpillars. Journal of Chemical Ecology, 36(10), 1101– 1104. https://doi. o r g / 1 0 . 1 0 0 7 / s 1 0 8 8 6 - 0 1 0 - 9 8 4 9 - 4 Laurentz, M., Reudler, J. H., Mappes, J., Friman, V., Ikonen, S., & Lindstedt, C. (2012). Diet quality can play a critical role in defense efficacy against parasitoids and pathogens in the glanville fritillary (Melitaea cinxia). Journal of Chemical Ecology, 38(1), 116– 125. https://doi. o r g / 1 0 . 1 0 0 7 / s 1 0 8 8 6 - 0 1 2 - 0 0 6 6 - 1 Lee, J., Hwang, S., & Cho, S. (2016). Immune tolerance to an intestineadapted bacteria, Chryseobacterium sp., injected into the hemocoel of Protaetia brevitarsis seulensis. Scientific Reports, 6(1), 1– 14. https://doi.org/10.1038/srep3 1722 Lenth, R. V., Buerkner, P., Herve, M., Love, J., Miguez, F., Riebl, H., & Singmann, H. (2022). Emmeans: Estimated marginal means, aka leastsquares means (1.7.41). https://CRAN.Rproje ct.org/packa ge=emmeans Lighthart, B. (1988). Some changes in gut bacterial flora of fieldgrown Peridroma saucia (Lepidoptera: Noctuidae) when brought into the laboratory. Applied and Environmental Microbiology, 54(7), 1896– 1898. Lindstedt, C., Talsma, J. H. R., Ihalainen, E., Lindström, L., & Mappes, J. (2010). Diet quality affects warning coloration indirectly: Excretion costs in a generalist herbivore. Evolution, 64(1), 68– 78. https://doi. org/10.1111/j.15585646.2009.00796.x Mander, L., & Liu, H.- W. (. B.). (2010). Comprehensive natural products II: Chemistry and biology. Elsevier. Marak, H. B., Biere, A., & Damme, J. M. M. V. (2000). Direct and correlated responses to selection on iridoid glycosides in Plantago lanceolata L. Journal of Evolutionary Biology, 13(6), 985– 996. https:// doi.org/10.1046/j.14209101.2000.00233.x Marak, H. B., Biere, A., & van Damme, J. M. M. (2002). Two herbivoredeterrent iridoid glycosides reduce the invitro growth of a specialist but not of a generalist pathogenic fungus of Plantago lanceolata L. Chemoecology, 12(4), 185– 192. https://doi.org/10.1007/PL000 12667 Mason, C. J., & Raffa, K. F. (2014). Acquisition and structuring of midgut bacterial communities in gypsy moth (Lepidoptera: Erebidae) larvae. Environmental Entomology, 43(3), 595– 604. https://doi. org/10.1603/EN14031 Mason, C. J., Ray, S., Shikano, I., Peiffer, M., Jones, A. G., Luthe, D. S., Hoover, K., & Felton, G. W. (2019). Plant defenses interact with insect enteric bacteria by initiating a leaky gut syndrome. Proceedings of the National Academy of Sciences of the United States of America, 116(32), 15991– 15996. https://doi.org/10.1073/pnas.19087 48116 Metcalfe, D. B., Asner, G. P., Martin, R. E., Espejo, J. E. S., Huasco, W. H., Amézquita, F. F. F., CarranzaJimenez, L., Cabrera, D. F. G., Baca, L. D., Sinca, F., Quispe, L. P. H., Taype, I. A., Mora, L. E., Dávila, A. R., Solórzano, M. M., Vilca, B. L. P., Román, J. M. L., Bustios, P. C. G., Revilla, N. S., … Malhi, Y. (2014). Herbivory makes major contributions to ecosystem carbon and nutrient cycling in tropical forests. Ecology Letters, 17(3), 324– 332. https://doi.org/10.1111/ele.12233 Minard, G., Kahilainen, A., Biere, A., Pakkanen, H., Mappes, J., & Saastamoinen, M. (2022). Complex plant quality— Microbiota— Population interactions modulate the response of a specialist herbivore to the defense of its host plant. Zenodo, https://doi. org/10.5281/zenodo.7045832 Minard, G., Tikhonov, G., Ovaskainen, O., & Saastamoinen, M. (2019). The microbiome of the Melitaea cinxia butterfly shows marked variation but is only little explained by the traits of the butterfly or its host plant. Environmental Microbiology, 21, 4253– 4269. https://doi. org/10.1111/14622920.14786 Mithöfer, A., & Boland, W. (2012). Plant defense against herbivores: Chemical aspects. Annual Review of Plant Biology, 63, 431– 450. h t t p s : / / d o i . o r g / 1 0 . 1 1 4 6 / a n n u r e v - a r p l a n t - 0 4 2 1 1 0 - 1 0 3 8 5 4 Nair, A., Fountain, T., Ikonen, S., Ojanen, P. S., & van Nouhuys, S. (2016). Spatial and temporal genetic structure at the fourth trophic level in a fragmented landscape. Proceedings of Royal Society B: Biological Sciences, 283(1831), 1– 8. https://doi.org/10.1098/Rspb.2016.0668 Näpflin, K., & SchmidHempel, P. (2016). Immune response and gut microbial community structure in bumblebees after microbiota transplants. Proceedings of the Royal Society B: Biological Sciences, 283(1831), 20160312. https://doi.org/10.1098/rspb.2016.0312 Newell, P. D., & Douglas, A. E. (2014). Interspecies interactions determine the impact of the gut microbiota on nutrient allocation in Drosophila melanogaster. Applied and Environmental Microbiology, 80(2), 788– 796. https://doi.org/10.1128/AEM.02742 - 13 Nieminen, M., Suomi, J., Van Nouhuys, S., Sauri, P., & Riekkola, M.- L. (2003). Effect of iridoid glycoside content on oviposition host plant choice and parasitism in a specialist herbivore. Journal of Chemical Ecology, 29(4), 823– 844. https://doi.org/10.1023/A:10229 23514534 Oksanen, J., Blanchet, F. G., Kindt, R., Legendre, P., Minchin, P. R., O'Hara, R. B., Simpson, G. L., Solymos, P., Stevens, M. H. H., & Wagner, H. (2013). vegan: Community ecology package (2.010). http://cran.rproje ct.org/web/packa ges/vegan/ index.html Pankoke, H., Buschmann, T., & Müller, C. (2013). Role of plant βglucosidases in the dual defense system of iridoid glycosides and their hydrolyzing enzymes in Plantago lanceolata and Plantago major. Phytochemistry, 94, 99– 107. https://doi.org/10.1016/j.phyto chem.2013.04.016 Pardo, F., Perich, F., Villarroel, L., & Torres, R. (1993). Isolation of verbascoside, an antimicrobial constituent of Buddleja globosa leaves. Journal of Ethnopharmacology, 39(3), 221– 222. Phalnikar, K., Kunte, K., & Agashe, D. (2019). Disrupting butterfly caterpillar microbiomes does not impact their survival and development. Proceedings of the Royal Society B: Biological Sciences, 286(1917), 20192438. https://doi.org/10.1098/rspb.2019.2438 R Core Team. (2016). R: A language and environment for statistical computing. https://www.Rproje ct.org/ Reichardt, P. B., Clausen, T. P., & Bryant, J. P. (1988). Phenol glycosides in plant defense against herbivores. In Biologically active natural products (Vol. 380, pp. 130– 142). American Chemical Society. https:// doi.org/10.1021/bk19880380.ch010 Reudler, J. H., Biere, A., Harvey, J. A., & van Nouhuys, S. (2011). Differential performance of a specialist and two generalist herbivores and their parasitoids on Plantago lanceolata. Journal of Chemical Ecology, 37(7), 765– 778. https://doi.org/10.1007/s1088 6 - 0 1 1 - 9 9 8 3 - 7 Reudler, J. H., Lindstedt, C., Pakkanen, H., Lehtinen, I., & Mappes, J. (2015). Costs and benefits of plant allelochemicals in herbivore diet in a multi enemy world. Oecologia, 179(4), 1147– 1158. https://doi. o r g / 1 0 . 1 0 0 7 / s 0 0 4 4 2 - 0 1 5 - 3 4 2 5 - 0 Reudler Talsma, J. H., Biere, A., Harvey, J. A., & van Nouhuys, S. (2008). Oviposition cues for a specialist butterfly— Plant chemistry and size. Journal of Chemical Ecology, 34(9), 1202– 1212. https://doi. o r g / 1 0 . 1 0 0 7 / s 1 0 8 8 6 - 0 0 8 - 9 5 1 9 - y 13652435, 2022, 11, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/1365-2435.14177 by University Of Jyväskylä Library, Wiley Online Library on [03/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License 2888 | Functional Ecology MINARD et al. Rosa, E., Minard, G., Lindholm, J., & Saastamoinen, M. (2019). Moderate plant water stress improves larval development, and impacts immunity and gut microbiota of a specialist herbivore. PLoS One, 14(2). https://doi.org/10.1371/journ al.pone.0204292 Rosa, E., Woestmann, L., Biere, A., & Saastamoinen, M. (2018). A plant pathogen modulates the effects of secondary metabolites on the performance and immune function of an insect herbivore. Oikos, 127(10), 1539– 1549. https://doi.org/10.1111/oik.05437 Saastamoinen, M. (2007). Lifehistory, genotypic, and environmental correlates of clutch size in the Glanville fritillary butterfly. Ecological Entomology, 32(2), 235– 242. https://doi. org/10.1111/j.13652311.2007.00865.x Saastamoinen, M., van Nouhuys, S., Nieminen, M., O'Hara, B., & Suomi, J. (2007). Development and survival of a specialist herbivore, Melitaea cinxia, on host plants producing high and low concentrations of iridoid glycosides. Annales Zoologici Fennici, 44(1), 70– 80. Suomi, J., Sirén, H., Jussila, M., Wiedmer, S. K., & Riekkola, M.- L. (2003). Determination of iridoid glycosides in larvae and adults of butterfly Melitaea cinxia by partial filling micellar electrokinetic capillary chromatography— Electrospray ionisation mass spectrometry. Analytical and Bioanalytical Chemistry, 376(6), 884– 889. https://doi. o r g / 1 0 . 1 0 0 7 / s 0 0 2 1 6 - 0 0 3 - 2 0 0 3 - 1 Suomi, J., Wiedmer, S. K., Jussila, M., & Riekkola, M.- L. (2001). Determination of iridoid glycosides by micellar electrokinetic capillary chromatographymass spectrometry with use of the partial filling technique. Electrophoresis, 22(12), 2580– 2587. https://doi.org/10.1002/15222683(20010 7)22:12<2 5 8 0 : : A I D - E L P S 2 5 8 0 >3.0.CO;27 Tack, A. J., Mononen, T., & Hanski, I. (2015). Increasing frequency of low summer precipitation synchronizes dynamics and compromises metapopulation stability in the Glanville fritillary butterfly. Proceedings of the Royal Society B, 282, 20150173. Tamura, Y., & Nishibe, S. (2002). Changes in the concentrations of bioactive compounds in plantain leaves. Journal of Agricultural and Food Chemistry, 50(9), 2514– 2518. Treutter, D. (2005). Significance of flavonoids in plant resistance and enhancement of their biosynthesis. Plant Biology (Stuttgart, Germany), 7(6), 581– 591. https://doi.org/10.1055/s2005873009 Voirol, L. R. P., Weinhold, A., Johnston, P. R., Fatouros, N. E., & Hilker, M. (2020). The legacy of a butterfly's parental microbiome on offspring performance. Applied and Environmental Microbiology, 86, 1– 16. https://doi.org/10.1128/AEM.00596 - 20 Wang, J., Peiffer, M., Hoover, K., Rosa, C., Zeng, R., & Felton, G. W. (2017). Helicoverpa zea gutassociated bacteria indirectly induce defenses in tomato by triggering a salivary elicitor(s). New Phytologist, 214(3), 1294– 1306. https://doi.org/10.1111/nph.14429 Wang, J., Yang, M., Song, Y., Acevedo, F. E., Hoover, K., Zeng, R., & Felton, G. W. (2018). Gutassociated bacteria of Helicoverpa zea indirectly trigger plant defenses in maize. Journal of Chemical Ecology, 44(7– 8), 690– 699. https://doi.org/10.1007/s1088 601809700 Whitaker, M. R. L., Salzman, S., Sanders, J., Kaltenpoth, M., & Pierce, N. E. (2016). Microbial communities of lycaenid butterflies do not correlate with larval diet. Frontiers in Microbiology, 7, 1– 16. https://doi. org/10.3389/fmicb.2016.01920 Xiang, H., Wei, G.- F., Jia, S., Huang, J., Miao, X.- X., Zhou, Z., Zhao, L.- P., & Huang, Y.- P. (2006). Microbial communities in the larval midgut of laboratory and field populations of cotton bollworm (Helicoverpa armigera). Canadian Journal of Microbiology, 52(11), 1085– 1092. https://doi.org/10.1139/w06064 Yoon, S. A., Harrison, J. G., Philbin, C. S., Dodson, C. D., Jones, D. M., Wallace, I. S., Forister, M. L., & Smilanich, A. M. (2019). Host plantdependent effects of microbes and phytochemistry on the insect immune response. Oecologia, 191(1), 141– 152. https://doi. o r g / 1 0 . 1 0 0 7 / s 0 0 4 4 2 - 0 1 9 - 0 4 4 8 0 - 3 Zhang, S., Shu, J., Xue, H., Zhang, W., Zhang, Y., Liu, Y., Fang, L., Wang, Y., & Wang, H. (2020). The gut microbiota in camellia weevils are influenced by plant secondary metabolites and contribute to saponin degradation. MSystems, 5(2), e0069219. https://doi.org/10.1128/ m S y s t e m s . 0 0 6 9 2 - 1 9 SUPPORTING INFORMATION Additional supporting information can be found online in the Supporting Information section at the end of this article. How to cite this article: Minard, G., Kahilainen, A., Biere, A., Pakkanen, H., Mappes, J., & Saastamoinen, M. (2022). Complex plant quality— microbiota– population interactions modulate the response of a specialist herbivore to the defence of its host plant. Functional Ecology, 36, 2873–2888. https://doi. org/10.1111/1365-2435.14177 13652435, 2022, 11, Downloaded from https://besjournals.onlinelibrary.wiley.com/doi/10.1111/1365-2435.14177 by University Of Jyväskylä Library, Wiley Online Library on [03/11/2022]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons License