Furostanol Saponins and Ecdysteroids from Plants of the Genus Helleborus as Phagostimulants and Predator Deterrents for Larvae of Two Monophadnus Sawfly Species

Sawfly species of the genus Monophadnus are specialised on Ranunculaceae plants from which the larvae can sequester furostanol saponins into the haemolymph, mainly (25R)-26-[(α-L-rhamnopyranosyl)oxy]-22α-methoxyfurost-5-en-3β-yl-O-β-D-glucopyranosyl-(1→3)-O-[6-acetyl-β-D-glucopyranosyl-(1→3)]-O-β-D-glucopyranoside (compound 1). In this work, TLC, GC-MS, and HPLC-DAD-ESI/MS analyses together with feeding, repeated simulated attacks, and ant deterrence bioassays were conducted to extend the chemoecological knowledge about two sawfly species specialised on H. foetidus L. (Monophadnus species A) and H. viridis L. (Monophadnus species B). Larvae of Monophadnus species B were mostly feeding on the squares treated with the n-butanol fraction from H. foetidus, compound 1 being its primary non-nutritional stimulant. In contrast, all H. viridis fractions stimulated feeding, with n-hexane marginally more active. β-sitosterol within n-hexane was determined as the nutritional stimulant. Quantitative analyses demonstrated that leaves of H. viridis but not H. foetidus contain the ecdysteroids 20-hydroxyecdysone and polypodine B. Moreover, the haemolymph of Monophadnus species B larvae reared on H. viridis contained the glycosides of polypodine B and 20-hydroxyecdysone at a concentration of 2.5 to 6.8 µmol/g fresh weight of haemolymph. This concentration is several thousand times higher than the concentration range of the aglycones in their host plant (3.63 × 10−4 to 2.23 × 10−4 µmol total ecdysteroids/g fresh weight of leaves), suggesting bioaccumulation. The larvae of both species fed on H. foetidus do not show any traces of ecdysteroids in their haemolymph, indicating a facultative role of these compounds in their defence as well as their inability to endogenously synthesise these compounds. The haemolymph containing ecdysteroids was a significant feeding deterrent against Myrmica rubra L. ant workers (one of their natural predators) at 0.8 mg/mL. The larvae kept effective deterrent levels of glycosylated ecdysteroids (≅175 mM) between simulated attacks on days 1 and 2, but the levels clearly decreased on day 3 (≅75 mM). Most larvae (89%) survived a first attack but only 23% a consecutive second one. As a conclusion, we report for the first time that two Monophadnus species feeding on H. viridis sequester phytoecdysteroids into the larval haemolymph in the form of glycosides. In addition, compound 1 possesses defensive and phagostimulant activities, and we present evidence for a combined effect of furostanol saponins and ecdysteroids as repellents against ants.


Introduction
Plants face a constant threat of being eaten, driving the evolution of diverse defence strategies.One such strategy is "becoming poisonous", i.e., the biosynthesis of secondary metabolites toxic to herbivores [1].There is indirect evidence that poisonous plants communicate their toxicity to animals, largely achieving successful co-existence, as long as herbivores can choose to eat other non-toxic plants [2].
Driven by the selective pressures of competition and predation, numerous herbivorous insects have undergone dietary niche differentiation by specialising on toxic plants.This shift offers a dual benefit: Firstly, it reduces competition from non-specialised herbivores.Secondly, the toxicity of the host plant provides a form of associational defence against both dedicated predators and larger herbivores that might inadvertently consume the insects alongside their fodder [3], although some insect species are themselves toxic to cattle [4].Furthermore, some insects also reuse harmful plant compounds to their own benefit in what is known as "sequestration", the selective uptake and storage of plant allelochemicals (toxins) for defence [5].However, detoxification or excretion are also possible strategies [6].Notably, herbivores and non-herbivores can synthesise de novo defence chemicals which are similar to those produced by plants ("convergent evolution") [7].
Sawfly larvae of the tribe Phymatocerini (Hymenoptera: Tenthredinidae: Blennocampinae) are generally specialised on toxic plants of the orders Liliales and Ranunculales [8].Several plant species within these orders account for a significant number of poisonings in cattle [2].Previously, we described how phymatocerine Monophadnus larvae (Figure 1) sequester a steroidal furostanol saponin (compound 1; Figure 2) present in Ranunculaceae leaves into their haemolymph, the blood-like fluid of invertebrates [9], and how the haemolymph is exuded by local disruption of the integument and release of haemolymph droplets upon a predator's bite [10][11][12].This defence strategy has been called "easy bleeding".

Introduction
Plants face a constant threat of being eaten, driving the evolution of diverse defence strategies.One such strategy is "becoming poisonous", i.e., the biosynthesis of secondary metabolites toxic to herbivores [1].There is indirect evidence that poisonous plants communicate their toxicity to animals, largely achieving successful co-existence, as long as herbivores can choose to eat other non-toxic plants [2].
Driven by the selective pressures of competition and predation, numerous herbivorous insects have undergone dietary niche differentiation by specialising on toxic plants.This shift offers a dual benefit: Firstly, it reduces competition from non-specialised herbivores.Secondly, the toxicity of the host plant provides a form of associational defence against both dedicated predators and larger herbivores that might inadvertently consume the insects alongside their fodder [3], although some insect species are themselves toxic to cattle [4].Furthermore, some insects also reuse harmful plant compounds to their own benefit in what is known as "sequestration", the selective uptake and storage of plant allelochemicals (toxins) for defence [5].However, detoxification or excretion are also possible strategies [6].Notably, herbivores and non-herbivores can synthesise de novo defence chemicals which are similar to those produced by plants ("convergent evolution") [7].
Sawfly larvae of the tribe Phymatocerini (Hymenoptera: Tenthredinidae: Blennocampinae) are generally specialised on toxic plants of the orders Liliales and Ranunculales [8].Several plant species within these orders account for a significant number of poisonings in cattle [2].Previously, we described how phymatocerine Monophadnus larvae (Figure 1) sequester a steroidal furostanol saponin (compound 1; Figure 2) present in Ranunculaceae leaves into their haemolymph, the blood-like fluid of invertebrates [9], and how the haemolymph is exuded by local disruption of the integument and release of haemolymph droplets upon a predator's bite [10][11][12].This defence strategy has been called "easy bleeding".

Introduction
Plants face a constant threat of being eaten, driving the evolution of diverse defence strategies.One such strategy is "becoming poisonous", i.e., the biosynthesis of secondary metabolites toxic to herbivores [1].There is indirect evidence that poisonous plants communicate their toxicity to animals, largely achieving successful co-existence, as long as herbivores can choose to eat other non-toxic plants [2].
Driven by the selective pressures of competition and predation, numerous herbivorous insects have undergone dietary niche differentiation by specialising on toxic plants.This shift offers a dual benefit: Firstly, it reduces competition from non-specialised herbivores.Secondly, the toxicity of the host plant provides a form of associational defence against both dedicated predators and larger herbivores that might inadvertently consume the insects alongside their fodder [3], although some insect species are themselves toxic to cattle [4].Furthermore, some insects also reuse harmful plant compounds to their own benefit in what is known as "sequestration", the selective uptake and storage of plant allelochemicals (toxins) for defence [5].However, detoxification or excretion are also possible strategies [6].Notably, herbivores and non-herbivores can synthesise de novo defence chemicals which are similar to those produced by plants ("convergent evolution") [7].
Sawfly larvae of the tribe Phymatocerini (Hymenoptera: Tenthredinidae: Blennocampinae) are generally specialised on toxic plants of the orders Liliales and Ranunculales [8].Several plant species within these orders account for a significant number of poisonings in cattle [2].Previously, we described how phymatocerine Monophadnus larvae (Figure 1) sequester a steroidal furostanol saponin (compound 1; Figure 2) present in Ranunculaceae leaves into their haemolymph, the blood-like fluid of invertebrates [9], and how the haemolymph is exuded by local disruption of the integument and release of haemolymph droplets upon a predator's bite [10][11][12].This defence strategy has been called "easy bleeding".Compound 1 is a feeding deterrent against ants (one of their natural predators) at the concentration found in the larval haemolymph of two Monophadnus species (at least 1.2 µmol/g FW) [9].Although this compound seems to play a major role in the chemical defence of Monophadnus larvae, other plant secondary metabolites found in the R-type Ranunculaceae such as γ-lactones derivatives [13] (Figure 3) and ecdysteroids [14] (Figure 4) may also be involved in the chemoecological relationship between Monophadnus larvae and their host plants, mainly of the genera Helleborus, Ranunculus, Clematis, and Pulsatilla.
Plants 2024, 13, x FOR PEER REVIEW 3 of 19 Compound 1 is a feeding deterrent against ants (one of their natural predators) at the concentration found in the larval haemolymph of two Monophadnus species (at least 1.2 µmol/g FW) [9].Although this compound seems to play a major role in the chemical defence of Monophadnus larvae, other plant secondary metabolites found in the R-type Ranunculaceae such as γ-lactones derivatives [13] (Figure 3) and ecdysteroids [14] (Figure 4) may also be involved in the chemoecological relationship between Monophadnus larvae and their host plants, mainly of the genera Helleborus, Ranunculus, Clematis, and Pulsatilla.Insects that acquire defensive chemicals (allelochemicals) from their host plants might benefit evolutionarily from recognising these plants, at least partially, by their chemical cues [15].This idea is supported by research showing that the steroid alkaloid which is sequestered by the larvae of Rhadinoceraea nodicornis (Konow, 1886) also acts as a phagostimulant for the larvae [16].
Therefore, we here aim to investigate whether sawfly species of the genus Monophadnus specialised on plants of the genus Helleborus sequester additional plant secondary metabolites besides compound 1 into their larval haemolymph.We further want to explore whether these sequestered metabolites act as feeding stimulants (phagostimulants) for the larvae and as deterrents against predators.The results of this study will serve to better understand the chemoecological relationship of Monophadnus larvae with their host plants (Helleborus spp.) and possibly extend the discussion to other sawfly species with similar defensive "easy-bleeding" behaviour.

Comparative Chemical Analyses of Host-Plant and Larval Haemolymph Extracts
The haemolymph of the different larval samples revealed considerable variation in the types of metabolites present, as shown in Table 1.All samples contained saponins; in addition, Monophadnus species A had sterols whilst Monophadnus species B had ecdysteroids in their haemolymph.The types of metabolites detected in the haemolymph of the two Monophadnus species were also found in their respective host plants suggesting potential bioaccumulation processes.Plants 2024, 13, x FOR PEER REVIEW 3 of 19 Compound 1 is a feeding deterrent against ants (one of their natural predators) at the concentration found in the larval haemolymph of two Monophadnus species (at least 1.2 µmol/g FW) [9].Although this compound seems to play a major role in the chemical defence of Monophadnus larvae, other plant secondary metabolites found in the R-type Ranunculaceae such as γ-lactones derivatives [13] (Figure 3) and ecdysteroids [14] (Figure 4) may also be involved in the chemoecological relationship between Monophadnus larvae and their host plants, mainly of the genera Helleborus, Ranunculus, Clematis, and Pulsatilla.Insects that acquire defensive chemicals (allelochemicals) from their host plants might benefit evolutionarily from recognising these plants, at least partially, by their chemical cues [15].This idea is supported by research showing that the steroid alkaloid which is sequestered by the larvae of Rhadinoceraea nodicornis (Konow, 1886) also acts as a phagostimulant for the larvae [16].
Therefore, we here aim to investigate whether sawfly species of the genus Monophadnus specialised on plants of the genus Helleborus sequester additional plant secondary metabolites besides compound 1 into their larval haemolymph.We further want to explore whether these sequestered metabolites act as feeding stimulants (phagostimulants) for the larvae and as deterrents against predators.The results of this study will serve to better understand the chemoecological relationship of Monophadnus larvae with their host plants (Helleborus spp.) and possibly extend the discussion to other sawfly species with similar defensive "easy-bleeding" behaviour.

Comparative Chemical Analyses of Host-Plant and Larval Haemolymph Extracts
The haemolymph of the different larval samples revealed considerable variation in the types of metabolites present, as shown in Table 1.All samples contained saponins; in addition, Monophadnus species A had sterols whilst Monophadnus species B had ecdysteroids in their haemolymph.The types of metabolites detected in the haemolymph of the two Monophadnus species were also found in their respective host plants suggesting potential bioaccumulation processes.Insects that acquire defensive chemicals (allelochemicals) from their host plants might benefit evolutionarily from recognising these plants, at least partially, by their chemical cues [15].This idea is supported by research showing that the steroid alkaloid which is sequestered by the larvae of Rhadinoceraea nodicornis (Konow, 1886) also acts as a phagostimulant for the larvae [16].
Therefore, we here aim to investigate whether sawfly species of the genus Monophadnus specialised on plants of the genus Helleborus sequester additional plant secondary metabolites besides compound 1 into their larval haemolymph.We further want to explore whether these sequestered metabolites act as feeding stimulants (phagostimulants) for the larvae and as deterrents against predators.The results of this study will serve to better understand the chemoecological relationship of Monophadnus larvae with their host plants (Helleborus spp.) and possibly extend the discussion to other sawfly species with similar defensive "easy-bleeding" behaviour.

Comparative Chemical Analyses of Host-Plant and Larval Haemolymph Extracts
The haemolymph of the different larval samples revealed considerable variation in the types of metabolites present, as shown in Table 1.All samples contained saponins; in addition, Monophadnus species A had sterols whilst Monophadnus species B had ecdysteroids in their haemolymph.The types of metabolites detected in the haemolymph of the two Monophadnus species were also found in their respective host plants suggesting potential bioaccumulation processes.

Composition of Essential Oil from Helleborus spp. Leaves
Overall, the essential oils of H. viridis and H. foetidus consisted of a high proportion of unidentifiable sesquiterpenoids and diterpenoids (Table 2).The profile of volatiles in the leaves of H. foetidus L. was more complex than the one in H. viridis, where only a few of them were identified.We observed in both samples the presence of a peak with a 96 m/z base peak at 4.2-4.3min which was assigned to protoanemonin (compound 3, C 5 H 4 O 2 ; 96.08 Da) in accordance with the literature data [17].The identity of this peak was confirmed by injection of the synthetic standard (Section 4.2).This compound was found in traces in other Helleborus [17,18].
Helleborus viridis and H. foetidus also shared the presence of benzeneacetaldehyde and (E)-2-Hexenal.Hexenals and hexenols are released as a result of lipoxygenase activity when leaves are damaged mechanically (i.e., cut prior to hydrodistillation) [17].The result is an enrichment of the essential oil in (Z)-3-hexenol and (E)-2-hexenal, popularly known as "leaf aldehydes".They are known to have important chemoecological roles in insect-plant and insect-insect communication [19,20].

Phagostimulant Activity of Plant Extracts, Fractions, and Isolates in Monophadnus spp. Larvae
The bioassays established that the crude methanol extract of both Helleborus spp.leaves stimulated larval feeding activity (p < 0.01), thus confirming that we achieved the extraction of the phagostimulant principles from the plant material (Figure 5a).In the case of H. foetidus, these principles were mostly soluble in the n-butanol fraction (p < 0.01) and to a lesser extend in water (p < 0.05) (Figure 5a).The larvae did not show a preference for treatment over control papers when using n-hexane or chloroform fractions (p > 0.05).Therefore, we selected the n-butanol fraction for further subfractionation by column chromatography, as this showed the strongest and most consistent preference with more larvae on treatments in 11 out of 12 dishes.
In the case of H. viridis (Figure 5b), all fractions were equally preferred to their controls, with 11 out of 12 replicates showing more larvae feeding on treatments than controls.Due to time limitations, we had to select one fraction to investigate further, so we picked the one where the difference between the mean number of larvae feeding on the treatment and control was greatest, which was the n-hexane fraction.
GC-MS analyses (Table 3) showed no differences between the n-hexane fractions of H. viridis and H. foetidus in terms of classes of major identifiable compounds (steroids/triterpenes and pigments), while more clear differences were observed by LC-MS analysis of the nbutanol fractions.Among the n-butanol fractions of H. foetidus, subfraction F was the one and only preferred (Figure 5c); this was then found to contain the saponin compound 1 and sugars.Subfraction D contained compound 1 only, which was also apparently active although short of being significant.Among the n-hexane column chromatography subfractions of H. viridis, the most preferred was C (Figure 5d); this was then found to contain β-sitosterol (Table 3).Table 3. Major phytochemical classes identified in the fractions and subfractions tested for phagostimulant activity.
Compound 1-from the active phagostimulant subfraction of the H. foetidus n-butanol fraction-was highly active (Figure 6).The saponin is also sequestered and responsible-at least in part-for the deterrence of the larval haemolymph [9]; thus, it is a multifaceted component in the Monophadnus-Helleborus system.We did not test any of the nutritional compounds identified in the active subfractions of Helleborus spp.(β-sitosterol and sugars), as they are known to act as unspecific phagostimulants for many insects [23,24].Total weight of paper targets available was on average 80 mg per replicate (40 mg controls, 40 mg treatments), so larvae still had both choices available at the end of the experiment.

HPLC-UV-ESI/MS-MS of Monophadnus spp. Larval Haemolymph and Helleborus spp. Leaf Extracts
Ecdysteroids were detected and dereplicated with the help of HPLC-UV-ESI/MS-MS (Figures 7 and 8).The methanolic extract from H. viridis leaves contained ecdysteroids mainly in form of polypodine B (Rt = 32.20 min) and 20-hydroxyecdysone (32.41 min).These compounds were not detected in the methanolic extract of H. foetidus leaves (Figure 7).Ecdysteroids were detected and dereplicated with the help of HPLC-UV-ESI/MS-MS (Figures 7 and 8).The methanolic extract from H. viridis leaves contained ecdysteroids mainly in form of polypodine B (Rt = 32.20 min) and 20-hydroxyecdysone (32.41 min).These compounds were not detected in the methanolic extract of H. foetidus leaves (Figure 7).Comparative analyses of sawfly haemolymph and plant leaves revealed that more polar derivatives of the plant ecdysteroids (Rt = 30.29 and 31.45min, respectively) were present in Monophadnus species B. The MS spectra of these ecdysteroid derivatives (Figure 8) show at least one loss of a 162 m/z fragment which may correspond to a hexose moiety, thus accounting for the increase in polarity.
The concentration of ecdysteroid glycosides in Monophadnus species B was ca.10,000fold higher than their aglycones in the plant in terms of the fresh weight of the samples (Table 4).The amount of ecdysteroid glycosides per mg dry haemolymph resulted to be 0.054 ± 0.  The concentration of ecdysteroid glycosides in Monophadnus species B was ca.10,000-fold higher than their aglycones in the plant in terms of the fresh weight of the samples (Table 4).The amount of ecdysteroid glycosides per mg dry haemolymph resulted to be 0.054 ± 0.030 µmol for polypodine B glycoside and 0.049 ± 0.029 µmol of 20-hydroxyecdysone glycoside (N = 3), achieving 62.70 ± 43.56 µM and 56.75 ± 40.38 µM in the fresh haemolymph, respectively.

Cross-Rearing Experiment
After the cross-rearing experiment, the glycosylated ecdysteroids could be detected in the haemolymph of Monophadnus species A when the larvae were fed on leaves of H. viridis only.Conversely, rearing Monophadnus species B on its non-native host plant H. foetidus resulted in no traces of glycosylated ecdysteroids in their haemolymph (Figure 9).These observations rule out an endogenous synthesis of ecdysteroids from other steroidal precursors found in the plant.

Cross-Rearing Experiment
After the cross-rearing experiment, the glycosylated ecdysteroids could be detected in the haemolymph of Monophadnus species A when the larvae were fed on leaves of H. viridis only.Conversely, rearing Monophadnus species B on its non-native host plant H. foetidus resulted in no traces of glycosylated ecdysteroids in their haemolymph (Figure 9).These observations rule out an endogenous synthesis of ecdysteroids from other steroidal precursors found in the plant.

Ecdysteroid Levels in Repeatedly Collected Haemolymph
The results showed that ecdysteroid concentrations were similar on day 1 and day 2 but significatively decreased on day 3 (Table 5).In parallel, the survival of the larvae also dramatically decreased from day 2 to day 3.

Ecdysteroid Levels in Repeatedly Collected Haemolymph
The results showed that ecdysteroid concentrations were similar on day 1 and day 2 but significatively decreased on day 3 (Table 5).In parallel, the survival of the larvae also dramatically decreased from day 2 to day 3.

Feeding Deterrent Effect on Ants of Haemolymph and Sequestered Plant Metabolites
All samples were endowed with significant deterrent effects at a concentration of 0.8 mg/mL.Moreover, the isolated compound retained its activity even at concentrations ten times lower (Table 6). ) not significant, Wilcoxon matched-pairs signed-ranks test.(-) Not tested.( a ) Data from our previous work [9].Total numbers of ants feeding on test and control droplets are given between parentheses.

Entomological Considerations
Species names such as M. monticola (Hartig, 1837), M. longicornis (Hartig, 1837), and M. latus (Costa, 1894) have been used to describe species feeding on Helleborus spp.[25][26][27].According to the current taxonomy of the genus, these sawflies belong to M. taegeri (Lacourt and Noblecourt, 2020) [28,29], but the actual taxonomy of this genus appears to be more complex [28].The Monophadnus larvae we collected from the two different Helleborus species showed very clear and consistent morphological (Figure 1) and behavioural differences (species A being nocturnal, B diurnal).We therefore have grounds for considering our two Monophadnus, labelled A and B, to be separate species, but the nomenclature remains to be determined by a wider revision of the genus Monophadnus which contains ca. 25 species worldwide [30].

Comparative Analysis of All Major Secondary Plant Metabolite Classes in Helleborus and Monophadnus Species
A preliminary comparative chemical screening of insect and plant materials using different instrumental chromatography techniques showed furostanol saponins present in both host-plant species and in the haemolymph of the two sawfly species (Table 1), which agrees with our previous work [9].Previous works reported on the occurrence of phytoecdysteroids in certain Ranunculaceae species only [14].Indeed, ecdysteroids were detected in H. viridis and the haemolymph of Monophadnus larvae reared on its leaves but not in H. foetidus.Fatty acids that were present in the plant samples are virtually absent in the haemolymph of the larvae.The presence of other ubiquitous dietary compounds such as sugars, phenolic compounds, and β-sitosterol in haemolymph did not show a consistent pattern.
Therefore, saponins and ecdysteroids were shortlisted as potential feeding stimulants (phagostimulants) for the sawflies and/or as a chemical defence mechanism.A further strong candidate for these activities was ranunculin (compound 2, Figure 2).This hypothesis is biologically significant as this compound has been reported to occur in all five host-plant genera for Monophadnus species-M.aequalis (MacGillivray, 1908) feeds on Anemone spp., M. alpicola (Benson, 1954) on Pulsatilla spp., M. monticola (Hartig, 1837) on Helleborus spp.and Ranunculus spp., M. nigriceps (F.Smith, 1874) and M. spinolae (Klug, 1816) on Clematis spp., and M. pallescens (Gmelin, 1790) on Ranunculus spp.[31][32][33]-plus in three other genera of the Ranunculaceae, whereas this compound is not detected in seventeen other Ranunculaceae genera [34].Thus, host-plant use by Monophadnus sawflies is significantly associated with the occurrence of ranunculin in the plants (p = 0.001, Fisher exact probability test; N = 25 Ranunculaceae genera).Unfortunately, the reactivity and instability of this compound's derivatives precluded feeding assays at room temperature lasting for hours.Our analysis of essential oils extracted by hydrodistillation from the host plants revealed a significant difference in the compound's concentration between H. viridis and H. foetidus (trace amounts vs. 0.6%, respectively) (Table 2).Consequently, it is challenging to definitively attribute a chemoecological role of ranunculin for Monophadnus larvae.

Phagostimulants and Chemical Defence in the Two Helleborus Species with Monophadnus species B Larvae
Feeding experiments were carried out to find out the plant metabolites responsible for the selective preference of Monophadnus species B larvae for Helleborus foetidus and H. viridis leaves (Figure 5).
When offered the crude extract and different polarity fractions from H. foetidus, the feeding of the larvae was significantly stimulated by the n-butanol extract and water to a lesser extent.A subsequent assay was conducted with its column chromatography subfractions.Only one showed significant stimulation, which contained compound 1 and sugars.The latter are nutritional compounds known to act as unspecific phagostimulants for many insects [23,24], so only compound 1 was further assayed resulting to be endowed with significant phagostimulant activity (Figure 6).This adds another chemoecological role to this saponin which was already reported as a sequestered deterrent by Monophadnus spp.larvae [9].
When offered the crude extract and different polarity fractions from H. viridis leaves, the larvae were stimulated by all of them with the same statistical significance.The nhexane fraction was selected for further fractionation, as it was marginally more active.Only one subfraction showed significant stimulation and contained β-sitosterol, a nutritional compound known to act as an unspecific phagostimulant for many insects [23,24].However, we selected it for further deterrence assays, and it showed an activity at 0.8 mg/mL (Table 6).
Besides compound 1 [9], the characteristic γ-lactones of Ranunculaceae R-type species (Figure 3) and ecdysteroids (Figure 4) are potentially also involved in the chemical defence of the larvae due to their blistering and moulting effects, respectively [35,36].However, the short life and extreme reactivity of the γ-lactones pose technical challenges for their experimental handling.Ecdysteroids are analogues of invertebrate steroid hormones which interfere in the process of insect moulting and are also known to be a deterrent to invertebrate herbivores [37].Ecdysteroids occur in plants as a complex mixture, but the major compounds are usually ecdysone, its 20-hydroxylated derivative-also known as ecdysterone, polypodine A, or β-ecdysone-(compound 5; Figure 4) and polypodine B (compound 6; Figure 4) [38].Several Helleborus species can contain the two latter compounds, but H. foetidus belongs to those few evergreen Helleborus species not synthesising ecdysteroids in detectable quantities [37].
The results from the cross-rearing experiments (Figure 9) followed by targeted analyses revealed that larvae of both Monophadnus species can sequester the ecdysteroids by a glycosylation process if the aglycones are present in the offered Helleborus leaves (Figures 7 and 8).
Therefore, the larvae of the two Monophadnus species primarily feeding on H. foetidus-which does not synthesise ecdysteroids-are tolerant to ecdysteroids when these are experimen-tally added to the diet, and the larvae are able to both glycosylate and accumulate these compounds in their haemolymph when feeding on H. viridis together with compound 1.

Role of Ecdysteroids in Chemical Defence of Monophadnus spp. Larvae
The HPLC-UV/DAD-MS analyses of both plant extracts and insect haemolymph in specific conditions for the detection of both ecdysteroids and saponins revealed that two aglycones (20-OH-ecydsone and polypodine B) present in the plant extracts are present in the larvae haemolymph in the form of glycosides (Figures 7 and 8).They also showed the presence of compound 1 in all samples.The concentration of ecdysteroid glycosides in Monophadnus species B was ca.10,000-fold higher than their aglycones in the plant (Table 4).Thus, it is likely that this sawfly species glycosylates the ecdysteroids and maintains high levels of these compounds in the haemolymph, which suggests an active sequestration process.These levels were maintained after at least one simulated attack (Table 5).The biological activities of many glycosylated ecdysteroids are significantly lower than those of their corresponding aglycones, likely because the presence of this substituent impairs the interaction with the ecdysteroid receptor(s) [39].Therefore, glycosylation may help to avoid interference of the ecdysteroids with the insect's hormone balance, but this would need further experiments to be confirmed.Anyhow, these ecdysteroids seem like a facultative addition to the defence of Monophadnus larvae, nevertheless representing a useful contribution to larvae protection which is primarily based on the presence of compound 1.
In our bioassays, we found a significant deterrent activity to ants using 0.08 mg/mL (166 µM) of 20-hydroxyecdysone and 0.8 mg/mL of the haemolymph (equivalent to ca. 4-10 µM of glycosylated ecdysteroids) both with similar deterrent activity (Table 6).If we suppose that the glycosides are fully hydrolysed within the ant and that polypodine B is as active as 20-hydroxyecdysone, then the haemolymph seems to be 16-41 times more active than what the ecdysteroids account for, which may be due to the contribution of the furostanol saponins acting as a primary defence deterrent achieving 70% of deterrence at 0.08 mg/mL (72 µM) [9].Whether ecdysteroids and saponins act additively or synergistically requires further experiments.
The concentration of ecdysteroids in fresh haemolymph was maintained after the first simulated attack but decreased more than 60% after the second one during the survival test (Table 5).This could be due to either a decrease in the mass of ecdysteroids in haemolymph or a dilution phenomenon.The concentrations of ecdysteroids decrease both in terms of their µM and µmol/mg haemolymph dry weight, thus suggesting the first phenomenon.The experiment also suggests that the larvae survived the loss of significant amounts of haemolymph if provoked once but not twice on two consecutive days.The reduced fitness of the larvae may then also affect sequestration of plant secondary metabolites.
Predatory insects with biting-chewing mandibles constitute the type of predators to which the easy bleeding of deterrent haemolymph is most effective, but vertebrate predators such as birds also seem to be deterred [40].In the field, workers of the ant Lasius niger L. were observed to retreat immediately after biting into a larva of Monophadnus species B, thereby getting into contact with its haemolymph; the ants subsequently cleaned their mandibles for several minutes [9].Thus, testing the sawfly haemolymph on ants in laboratory bioassays reflects predator-prey interactions that occur under natural conditions.

Future Research and Concluding Remarks
We here demonstrated for first time that glycosylated ecdysteroids play an adjuvant role in the chemical defence of Monophadnus spp.larvae and that compound 1 is the main phagostimulant and more potent deterrent compound present in both the larvae's host plant and their haemolymph.In Monophadnus spp.larvae, these ecdysteroids are not stored in specialised glands like in other insects [41] but freely circulate in the larval haemolymph [42].Therefore, glycosylation seems to protect the insect from their deleterious effects by increasing the water/aqueous medium solubility of the parent compounds, thus facilitating their accumulation in the haemolymph whilst diminishing their ability to cross membranes as well as diminishing their affinity to ecdysteroid receptors [39].
The glycosylation of dietary ecdysteroids by the larvae should rely on the existence of enzymes responsible for the process.Where these are located in Monophadnus spp.larvae and what their specificity is are matters that warrant further research.In other insects, "nontargeted" UDP-glucosyltransferases are responsible for the detoxification and elimination of a wide range of endogenous and exogenous compounds including toxins [43].It is also well known that "ecdysteroid-targeted" enzymes are present in baculoviruses.The baculovirus enzyme ecdysteroid UDP-glucosyltransferase (EGT) disrupts the hormonal balance of the insect host by catalysing the conjugation of ecdysteroids, the moulting hormones, with the sugar moiety from UDP-glucose or UDP-galactose [43].The natural balance between insects and viruses allows many infected insects to bypass the feeding arrest preceding moulting, leading to larger size before eventual ecdysis [44].Only future research will unveil if Monophadnus species have their own enzyme-targeted or "nontargeted"-or if they have harnessed a symbiosis with EGT + baculoviruses to handle the high concentrations of ecdysteroids present in their diet.
Another point that warrants further scrutiny is the clarification of the insect taxonomy.As discussed in the introduction, we have morphological evidence that suggests that Monophadnus species A and B are two separate species.Our work demonstrated that there is a clear difference in their behavioural responses in feeding experiments, with species A being nocturnal and species B being active in their behavioural responses during the day.Both can feed on both H. viridis and H. foetidus species, able to sequester ecdysteroids wherever they are present in the diet as well as compound 1.The physiological maintenance of distinct sequestration processes between closely related species is known from sawflies of the genus Athalia [42] and, more generally, from other insects such as leaf beetles (Chrysomelidae) [41].Deterrent assays combining ecdysteroid glycosides and saponins in specific quantities should be carried out to prove if both types of compounds act in synergy towards defending the larvae against predators.
Overall, our results corroborate the chemoecological role of compound 1 in Monophadnus species reared on Helleborus species as both a feeding stimulant and the main chemical deterrent against predators described in our previous publication [9].Our present study adds new layers of complexity to the easy-bleeding phenomenon by demonstrating the facultative contribution of dietary plant ecdysteroids to the chemical defence system of Monophadnus larvae already based on the occurrence of compound 1.

Insect and Plant Samples
Larvae of two Monophadnus species were collected from several sites in Switzerland on Helleborus foetidus and Helleborus viridis (Table 7).In accordance with our previous publication [8], we designated Monophadnus species A as the one naturally feeding on H. foetidus and Monophadnus species B as the one naturally feeding on H. viridis (Figure 1).Voucher specimens of insect samples used in this study are kept in the collections of the Senckenberg Deutsches Entomologisches Institut (Müncheberg, Germany) and the Royal Belgian Institute of Natural Sciences (Bruxelles, Belgium).
In the laboratory, the larvae were fed fresh leaves of their respective host plants.Haemolymph droplets were collected in glass capillaries after gently piercing the larval integument with forceps.The haemolymph was suspended in ethanol and stored at −20 • C in the dark.Each ethanolic extract was sonicated in an ice bath for 15 min and then vortexed for 1 min.After centrifugation (5 min, 4000 rpm), the supernatant was transferred to a new vial and the precipitate re-extracted twice with methanol under the same conditions.The combined supernatants were dried under N 2 and subsequently re-dissolved in methanol up to a concentration of 1 to 3 mg/mL.2) was obtained as described by Prieto et al. [22].
Protoanemonin (compound 3; Figure 3)-a product of the hydrolysis of the naturally occurring ranunculin (compound 2; Figure 3)-was synthesised according to Grundmann and Kober [45], as follows.Angelica lactone (5.7 g) was dissolved in CS 2 (6.25 mL), and bromine (2 mL) was added dropwise at −20 • C until the solution became colourless.The solution was dried under vacuum, and the residue was dissolved in 40 mL ether.A trace of hydroquinone was added to prevent dimerisation.Then, two equivalents of quinoline (14 g), a tertiary base, were added dropwise at −20 • C, and the reaction mixture was stored overnight at the same temperature.The next day, the excess of quinoline was extracted from the reaction mixture with acidic water; the organic layer filtrate yielded a solid consisting of quinoline hydrobromide.The residue was washed thoroughly with the solvent, which was subsequently eliminated in a rotavapor under low-pressure conditions.The liquid residue was further distilled in a rotary bulb-to-bulb device, and the fraction boiling between 65 • C and 80 • C at 12 torr was collected in a cool trap.Subsequent vacuum distillation yielded the pure compound (bp 68 • C, 8 torr). 1 H-NMR (CDCl 3 ) of the oily, yellow compound corresponded to the data reported in the literature for protoanemonin [35], with traces of anemonin, a product of protoanemonin dimerisation (compound 4; Figure 3).

Plant Extraction
Freshly collected plant material was chopped and macerated in methanol at room temperature for one week.After concentration under reduced pressure, the crude extract was suspended in methanol 10% v/v aq. and successively partitioned with n-hexane, chloroform, and n-butanol to obtain four corresponding fractions.After removing the solvent in a rotavapor, each fraction was freeze-dried to remove any traces of water, the residue was dissolved in methanol (HPLC grade) and centrifuged, and the clear supernatant was used for analysis.

Hydrodistillation and Analysis of Essential Oils from Plant Samples
Fresh leaves (40-60 g) of each Helleborus species were collected and subjected to hydrodistillation in a Clevenger apparatus for 2.5 h to obtain the essential oil.The samples were centrifugated (14,000 rpm, Eppendorf centrifuge), and the clear supernatants were stored at −80 • C.
The essential oils were dissolved in hexane and analysed by GC/EIMS with a Varian CP-3800 gas chromatograph equipped with a DB-5 capillary column (30 m × 0.25 mm; coating thickness 0.25 µm) and a Varian Saturn 2000 ion mass detector (Varian, Inc., Walnut Creek, CA, USA).Analytical conditions: injector and transfer line temperatures of 220 • C and 240 • C, respectively; oven temperature programmed from 60 • C to 240 • C at 3 • C/min; carrier gas helium at 1 mL/min; injection of 0.2 µL (10% hexane solution); and split ratio 1:30.The identification of the constituents was based on comparison of the retention times with those of authentic samples as well as on computer matching against NIST98 [46] and Adams [47] libraries' spectra and a home-made library mass spectra from pure substances.

Comparative Analytical Screening of Plant and Insect Samples
A preliminary screening of the samples of larval haemolymph and crude leaf extracts was performed by thin-layer chromatography using n-butanol-acetic acid-water (4:1:5) as the mobile phase and silica gel F 254 as the stationary phase (Merck, Darmastadt, Germany) and GC-EIMS in a Thermo Quest Trace GC 2000 chromatograph (Thermo Finnigan, San Jose, CA, USA) coupled to a Thermo Finnigan Trace MS mass spectrometer, equipped with a split-splitless injector and a Thermo Finnigan AS 2000 autosampler.A fused silica capillary column (DB-XLB, 15 m × 0.25 mm) was used.Samples were injected (1 µL) in split mode (1:95).The injector was heated to 320 • C, while the transfer line was set at 350 • C and the source temperature at 250 • C. The temperature profile started at 50 • C, followed by a 10 • C/min ramp to 320 • C which was held constant for 15 min, and finally a 10 • C/min ramp up to 360 • C, the thermal maximum of the column.EIMS spectra were acquired in scan mode in a 50-650 m/z range.The identification of individual constituents was achieved by matching mass spectral data with those held in the NIST98 [46] and a home-made library mass spectra from pure substances.
The ESI/MS conditions were positive ion mode, source voltage 4.5 kV, sheath gas (N 2 ) at flow rate 60 au, curtain gas (N 2 ) at 9.00 au, source current 80 A, capillary voltage 3 V, and capillary temperature 280 • C. Full scan spectra were obtained in the positive ion mode, with a scan time of 1 s from m/z 250 to 1400.The ion trap was running in automatic gain control with a maximum injection time of 200 ms.For the MS 2 analyses, the most intense molecular ions were isolated with a width of 1 m/z unit and fragmented by using an activation amplitude of 25%.
Quantitative analyses were performed using both MS and UV data in parallel.Aliquots of four different concentrations of standards (0.01 to 0.25 mg/mL) were injected into the analytical system, eluted, and monitored with the UV-ESI/MS detectors under the conditions detailed above.Injections of plant extracts and standards were performed in triplicate, on three different days.The response of the ESI/MS detector to different concentrations of ecdysteroids was linear in the range from 0.01 to 0.25 mg/mL (r = 0.9998).Haemolymph was injected in duplicate, and three of such analyses were performed on different days.Peak areas were integrated and related to the amount of injected external standard.

Feeding Bioassays
In a "cross-feeding experiment", larvae of the two Monophadnus species were reared on leaves of their non-native host plants as per Table 1: Monophadnus species A (sample code 7, Table 7) on H. viridis and larvae Monophadnus species B (sample codes 4 and 11, Table 1) on H. foetidus.This showed that the two sawfly species accepted both plants as feeding hosts with over 70% survival on both.Preliminary experiments to find a methodology for feeding trials discovered that while Monophadnus species B was active in the daytime and we were able to follow its behaviour in trials over time, species A was nocturnal, and its lack of activity made it unsuitable for a behavioural trial.As Monophadnus species B accepts and grows well on both Helleborus species, we chose to use this species for our bioassays.
The phagostimulant activity of plant extracts and isolates in Monophadnus species B larvae was investigated in a series of replicated bioassays in which groups of larvae were placed in a dish with a choice of paper squares with a single test solution or a control, and feeding behaviour was monitored.For the initial experiments with H. viridis and H. foetidus extracts, we tested 12 replicates of the whole extract plus 12 replicates of each fractional extract (made in sequence from non-polar to polar solvents: hexane, chloroform, butanol, water, dried, and redissolved in methanol).We used 10 larvae per dish with the H. viridis trials and 9 per dish with H. foetidus (fewer were available); each larva was only used once.The larvae were offered a grid of four 1.5 × 1.5 cm squares of filter paper in sealed dishes.All squares were treated with 20 µL of a 1 M sugar solution (without which larvae would not feed on paper); two squares were subsequently treated with 20 µL of extracts or isolates in different concentrations and the other two squares with solvent only (control).A scoping bioassay with different concentrations of whole-leaf extract showed that we obtained a significant feeding response from larvae using 0.25 g fresh weight of leaf material per mL of methanol.We therefore used this as our base concentration, having established a fresh weight-to-dry weight equivalency, and knowing the dry weight proportions of our fractional extracts made up all our test solutions in relation to this.
The number of larvae feeding on control and treatment squares in each dish was visually checked and recorded every 1.5 h for 6 h.Our scoping bioassay had suggested that feeding responses were maximised at 6 h but that this was before the surface of the paper targets had been significantly consumed.Data from the 6 h readings were processed by the Wilcoxon signed-rank test to assess whether feeding activity was significantly higher on extract-treated papers than controls for each treatment at each time-this ranks the difference in numbers between extract and control and tests for a consistent pattern of preference between the two.
Subsequently, we repeated this experimental protocol after the subfractionation of the most active fraction of the H. viridis (n-hexane) and the H. foetidus (n-butanol), testing 12 replicates of each subfraction in the same way.We had fewer larvae available for this so were only able to use 6 or 7 larvae in each set of replicates.Again, subfraction concentrations were made by dissolving the dried subfraction in methanol to a concentration equivalent to 0.25 g whole-leaf extract per mL methanol.
To test the isolated saponin (compound 1, 0.5 mg/mL in methanol), we used 8 larvae per dish but were only able to run 7 replicates.We were able to measure the feeding activity more precisely in this experiment by weighing the dry treated and control filter paper squares before and after the experiment and analysing the weight loss for treatments and controls per dish.As weight is a continuous measure, we analysed the average weight loss using a Hotelling's T 2 test.

Larval Survival and Concentration of Ecdysteroids in a Repeatedly Collected Haemolymph Assay
In the laboratory, 18 sham procedure control group larvae and 18 test larvae of Monophadnus species B were daily fed fresh leaves of H. viridis, weighed, and checked

Figure 1 .
Figure 1.Larvae of Monophadnus species A (left, picture by Alison M. Barker) and Monophadnus species B (right, picture by Jean-Luc Boevé) attached to the underside of a leaf of Helleborus foetidus L. and Helleborus viridis L., respectively.

Figure 1 .
Figure 1.Larvae of Monophadnus species A (left, picture by Alison M. Barker) and Monophadnus species B (right, picture by Jean-Luc Boevé) attached to the underside of a leaf of Helleborus foetidus L. and Helleborus viridis L., respectively.

Figure 1 .
Figure 1.Larvae of Monophadnus species A (left, picture by Alison M. Barker) and Monophadnus species B (right, picture by Jean-Luc Boevé) attached to the underside of a leaf of Helleborus foetidus L. and Helleborus viridis L., respectively.

Figure 5 .Figure 6 .Figure 5 .
Figure 5. Feeding bioassays of Monophadnus species B on paper squares treated with the initial crude methanol extract and its increasing polarity solvent fractions (a,b) and selected subfractions of the active n-butanol fraction of Helleborus foetidus (c) and the n-hexane fraction of Helleborus viridis (d) leaves.Orange and yellow colours indicate larvae feeding on fraction-treated and solvent control squares, respectively, at 6 h.(Pig) Pigments.(*) p < 0.05; (**) p < 0.01, Wilcoxon signed-rank tests.

Figure 5 .Figure 6 .Figure 6 .
Figure 5. Feeding bioassays of Monophadnus species B on paper squares treated with the initia methanol extract and its increasing polarity solvent fractions (a,b) and selected subfraction active n-butanol fraction of Helleborus foetidus (c) and the n-hexane fraction of Helleborus vi leaves.Orange and yellow colours indicate larvae feeding on fraction-treated and solvent squares, respectively, at 6 h.(Pig) Pigments.(*) p < 0.05; (**) p < 0.01, Wilcoxon signed-rank t

Figure 7 .
Figure 7. HPLC/MS chromatograms of Helleborus spp.n-butanol extracts, haemolymph of Monophadnus species B larvae feeding on Helleborus viridis, and haemolymph of Monophadnus species A larvae feeding on Helleborus foetidus.

Figure 8 .
Figure 8. Annotated MS spectra of the ecdysteroids detected in the haemolymph of Monophadnus species B feeding on Helleborus viridis.

Figure 9 .
Figure 9. TLC analysis (mobile phase: BAW; reagent: cerium sulphate; picture taken under 254 nm UV light in inversed B/W for visual enhancement) of the haemolymph of Monophadnus species A (lanes denoted after sample codes 7 and 8; see Section 4.1) and species B (lanes denoted after sample codes 3, 4, 11, and 12; see Section 4.1) reared either in Helleborus viridis or Helleborus foetidus.The yellow circles highlight areas where ecdysteroid metabolites can be detected.The small numbers between dashed lines correspond to different metabolites found in that zone.

Figure 9 .
Figure 9. TLC analysis (mobile phase: BAW; reagent: cerium sulphate; picture taken under 254 nm UV light in inversed B/W for visual enhancement) of the haemolymph of Monophadnus species A (lanes denoted after sample codes 7 and 8; see Section 4.1) and species B (lanes denoted after sample codes 3, 4, 11, and 12; see Section 4.1) reared either in Helleborus viridis or Helleborus foetidus.The yellow circles highlight areas where ecdysteroid metabolites can be detected.The small numbers between dashed lines correspond to different metabolites found in that zone.

Table 1 .
Presence of different classes of phytochemicals in Monophadnus spp.haemolymph as compared with their Helleborus spp.host-plant species.

Table 2 .
Composition of the essential oil of Helleborus spp.leaves analysed by GC/EIMS.

Table 4 .
Concentration of ecdysteroids in the methanolic extract of Helleborus spp.leaves (N = 5) and in larval haemolymph of Monophadnus species B (N = 2) feeding on Helleborus viridis.Values (mean ± SD) are expressed in µmol compound/g fresh weight of sample.

Table 4 .
Concentration of ecdysteroids in the methanolic extract of Helleborus spp.leaves (N = 5) and in larval haemolymph of Monophadnus species B (N = 2) feeding on Helleborus viridis.Values (mean ± SD) are expressed in µmol compound/g fresh weight of sample.

Table 5 .
Survival of larvae and concentration of ecdysteroids in haemolymph of a sample population (18 larvae on day 1) of Monophadnus species B following repeated haemolymph withdrawal.

Table 7 .
Larvae of Monophadnus spp.collected in Switzerland on Helleborus spp.-grown leaves of H. foetidus and H. viridis were collected in the same areas where Monophadnus species A and B larvae were found, respectively, and identified by Alison M. Barker (vouchers deposited at CABI Center, Delémont, Switzerland). Full