Introduction
Pyrrolizidine alkaloids, particularly 1,2-dehydropyrrolizidines (hereafter PAs), are powerful defensive compounds in several plant taxa, such as Eupatorieae and Senecioneae (Asteraceae), Boraginaceae, and Crotalarieae (Leguminosae).1,2 However, specialized herbivores, such as arctiine moths, danaine and ithomiine butterflies and chrysomeline beetles, among others, are able to cope with pro-toxic free-base PAs from these plants, converting them into non-toxic N-oxides.3,4 These insects sequester PA N-oxides, incorporating them into their tissues, and this renders chemical protection against predators and parasitoids.1,5 Arctiine moths and danaine and ithomiine butterflies also use PAs as precursors of male sexual pheromones, such as the dihydropyrrolizine hydroxydanaidal.6-8 PAs were also recorded in the grasshopper Zonocerus variegatus, the aphid Aphis jacobaeae, the bug Largus rufipennis, and the coccide Ceroplastes albolineatus.1 PA-containing insects are generally aposematic, i.e., their unpalatability is associated with warning signals alerting predators of this trait; warning signals can be visual, sonorous, or odorant.9
Among arctiine moths,10 PAs were found in the tribes Amerilini and Arctiini (subtribes Callimorphina, Arctiina, Pericopina, Ctenuchina, Euchromiina, and Phaegopterina), but not in Lithosiini and Syntomini.11 Three PA acquisition strategies were found in these moths: (i) PA-obtained as adults, (ii) PA-specialist feeding as larvae, and (iii) PA-generalist feeding as larvae.11 In the first syndrome, adults either feeding on nectar containing PAs or display a pharmacophagous behavior, or obtaining the alkaloid visiting withered PA-containing plants.2,12,13 This syndrome was widely found in Ctenuchina, Euchromiina, and Phaegopterina.11 In the second, larvae are generally monophagous, feeding on a single PA-host plant genus, from which they sequester the alkaloids; subtribes Callimorphina and Pericopina shared this syndrome.11,14 The third syndrome comprises polyphagous larvae, feeding both on non-PA and PA-host plants.5,11 Species from the subtribe Arctiina showed this syndrome.11 In the two last syndromes, adults can also sequester these alkaloids, even when feeding on PA-plants as larvae.11
Other interesting PA pathways in arctiine moths are the de novo biosynthesis of insect ester alkaloids (insect PAs) and the stereochemical inversion of 7R configuration to 7S pyrrolizidine rings.14-16 Both biochemical mechanisms are involved in the biosynthesis of dihydropyrrolizine male sexual pheromones.16-19 Insect PAs are esters biosynthesized from plant-acquired necine bases with necic acids of insect origin, and it has been suggested that they are biosynthesized via transesterification of plant PAs.19 In the insect PAs of the callimorphine type, necic acids are derived from isoleucine, while in the creatonotine type, they are derived from valine.16,20 Callimorphine type PAs were found in the Arctiina Arctia caja, Creatonotos transiens, Estigmene acrea, and Grammia geneura, in the Callimorphina Callimorpha dominula and Tyria jacobaea, and in the Pericopina Gnophaela latipennis and Hyalurga syma.14-16 PAs of the creatonotine type were found in the Arctiina Creatonotos transiens, Estigmene acrea, and Grammia geneura, and in the Callimorphina Utetheisa ornatrix.16 Insect PAs were not found in the Callimorphina Nyctemera annulata.21 No chemosystematic patterns of insect PAs seem to emerge from different subtribes of Arctiini. Stereochemical inversion must take place in Arctiini, since a general feature of male sexual pheromones seems to be the 7R configuration.6-8,16-19
Although Arctiini comprises around 11,000 species, only a small fraction was studied regarding PAs.16,22 For instance, in Pericopina, a single species, Hyalurga syma, was studied in relation to PA profile and chemical defense.14 This gap makes it difficult to search any patterns of PA acquisition syndromes or insect PA production in the arctiine phylogeny, i.e., it is difficult to infer about the evolution of PAs in the Arctiinae subfamily. We have observed the sequestration of PAs from host plants in the pericopine moth Scearctia figulina, adding more data to this poorly studied subject. Moreover, we carried out bioassays to test if S. figulina was protected against predation, as suggested for PA-containing insects.
Experimental
Studied organisms
The moth Scearctia figulina (Butler) (Erebidae: Arctiinae: Pericopina) is a Neotropical species that feeds on leaves of Heliotropium transalpinum Vell. (Boraginaceae) as larvae (Figure S1). Adults were found flying around their host plants. The genus Heliotropium has a Pantropical distribution with around 200 species, and H. transalpinum is a Neotropical species with wide distribution, from Mexico to Argentina, occurring in the open areas of Cerrado savanna and in tropical seasonal forests.23,24 These species, as well as other Boraginaceae species, show PAs in their tissues.14,25,26
Scearctia figulina sampling and rearing
Gregarious moth eggs (87 ± 11 eggs, n = 10, mean ± standard error) were sampled in individuals of H. transalpinum in an open area near the Animal Biology Department, Institute of Biology at the State University of Campinas, Campinas, São Paulo, Brazil (22°49'15.38"S, 047°04'8.87"W). After eclosion, gregarious larvae were kept in plastic containers (18 cm high, 11 cm diameter, 10-15 larvae per container) until pupation in an incubator at 27 ºC, L:D 12:12 photoperiod, with no relative humidity control. The containers were cleaned on a daily basis and old, eaten leaves of H. transalpinum were replaced by new, intact ones. Pupae were individualized in small containers (6 cm high, 5 cm diameter) until adult emergence. Eggs, fourth instar larvae, pupae, and adults of both sexes were sampled and immediately frozen at -20 ºC for PA analysis. The fourth instar larvae were sampled immediately after ecdyse to prevent feces in their midguts.
Pyrrolizidine alkaloid analysis
Ten samples of freeze-dried gregarious eggs, fourth instar larvae, pupae, and adults of both sexes reared in laboratory, and five samples of H. transalpinum leaves were quantified for PAs by colorimetric assay according to Trigo et al.14 The alkaloids were characterized by gas chromatography-mass spectrometry (GC-MS) using a mass fragmentation pattern and van den Dool & Kratz retention index (see Table S1).14,15,27-29 We assigned the absolute stereochemistry of PAs based only in retention index of chiral PAs reported in the literature.20,27
Chemical defense of Scearctia figulina: predation bioassays
The chemical defense in moths was investigated against three kinds of predators: the orb-weaving spider Nephila clavipes (Nephilidae), the wolf spider Lycosa erythrognatha (Lycosidae), and the chick Gallus gallus (Phasianidae). The two former are potential predators of S. fugilina in nature, and the latter is generally used as a predator model for visually hunting vertebrate predators.30 The license for research involving wild animals was provided by IBAMA-ICMBio (Ministério do Meio Ambiente, Brazil). The Ethics Committee for Animal Use of the University of Campinas approved all experimental procedures. Chicks were donated to free range farms by the end of the experiment. Bioassays followed experimental procedures of previous works from our research group.15,31
Statistical analysis
We checked if the PAs from the host plant H. transalpinum can explain the PAs in lab-reared S. fugilina, using a detrended correspondence analysis (DCA).32 Additionally, we checked if PA concentrations differed among developmental stages of lab-reared moths using an one-way analysis of variance (ANOVA).33
We compared predator responses (prey or release) among three predators and larvae, pupae, and between adults of both sexes using two approaches: (i) comparing the response of three predators in relation to adults, and (ii) comparing the response of L. erythrognatha and G. gallus in relation to larvae. No bioassays with larvae of N. clavipes were carried out because their larvae may not be in contact with this predator in the natural environment. We analyzed the frequency of individuals preyed or released, using a generalized linear model (GLM) with binomial distribution and logit link function, using the package "bbmle" in R 3.1.0 for Windows.34 In the first approach we performed a pair-pair comparison with Bonferroni correction,33 using α = 0.05 and k number of comparisons = 3; Bonferroni correction decreases the significant threshold to 0.0167.
Results
Pyrrolizidine alkaloids
We characterized 16 PAs in the system H. transalpinum-S. fugilina, and five were designated as unidentified (Table 1, Figure 1). The leaves of the host plant had predominantly riderine (IX, 7S,3'R) and 3'-acetylrinderine (XI), which together account for 90% of total PAs; supinine (IV, 3'R), 3'-acetylsupinine (VI), and 3'-acetylintermidine (X, 7R, 3'R) are present in low relative abundances (Table 1). However, moths did not show a similar profile. Only supinine was sequestered unchanged and maintained throughout the moths' life-cycle; traces of 3'-acetylintermedine were found only in larvae. The other host plant PAs were not present in moths. We observed that intermedine (VII, 7R, 3'R) and lycopsamine (VIII, 7R, 3'S) were the main alkaloids in the moths, reaching 70% of relative abundance; amabiline (V, 3'S), the necine base retronecine (I), and 7- and 9-senecioylretronecine-type PAs (II, III) were found in low amounts. Additionally, we found insect PAs of the callimorphine and creatonotine types in all developmental stages (Table 1). The predominant insect PA was callimorphine (XIV); other insect PAs were 7-deoxycallimorphine (XV), 7-deoxy-1,2-dihydrocallimophine (XVI), creatonotine (XII), and isocreatonotine (XIII). Based on the retention index, we assume that all insect PAs have a 7R configuration, and show the absolute configuration of necic acid moiety identical to those reported by literature (Figure 1).20 For the 7-deoxy-1,2-dihydrocallimorphine (XVI), the asymmetric center at C1 has R configuration.20 Insect PAs accounted for approximately 20% of PAs in S. figulina. Three unidentified PAs were found, and they accounted for 8% of relative abundance.

Figure 1 Pyrrolizidine alkaloids found in leaves of Heliotropium transalpinum and in eggs, larvae, pupae, and adults of Scearctia figulina.
Table 1 Relative abundance of pyrrolizidine alkaloids (mean ± standard error) in developmental stages of Scearctia figulina and in the leaves of its host plant, Heliotropium transalpinum
Pyrrolizidine alkaloida | RIb | [M]+ | Relative abundance / % | |||||
---|---|---|---|---|---|---|---|---|
Egg | Larva | Pupa | Male | Female | Host plant | |||
Retronecine (I) | 1484 | 155 | - | 1.25 ± 0.12 | 0.53 ± 0.11 | 0.46 ± 0.08 | 0.82 ± 0.09 | - |
Unidentified PA | 1858 | - | - | - | - | - | - | 0.83 ± 0.03 |
7-Senecioylretronecine type (III) | 1864 | 237 | - | 1.18 ± 0.21 | 0.72 ± 0.07 | 0.66 ± 0.06 | 0.57 ± 0.06 | - |
Isocreatonotine A (XIII) | 1877 | 255 | 0.69 ± 0.10 | 0.31 ± 0.08 | 0.27 ± 0.01 | 0.14 ± 0.01 | 0.30 ± 0.07 | - |
7-Deoxy-1,2-dihydrocallimorphine (XVI) | 1883 | 283 | 1.59 ± 0.20 | 1.51 ± 0.09 | 1.25 ± 0.21 | 1.23 ± 0.09 | 0.96 ± 0.06 | - |
7-Deoxycallimorphine (XV) | 1890 | 281 | 1.39 ± 0.31 | 1.99 ± 0.13 | 0.63 ± 0.07 | 1.94 ± 0.31 | 0.80 ± 0.05 | - |
9-Senecioylretronecine type (II) | 1895 | 237 | - | 1.43 ± 0.11 | 0.84 ± 0.10 | 1.09 ± 0.06 | 1.24 ± 0.10 | - |
Creatonotine A (XII) | 1938 | 255 | 1.50 ± 0.16 | 1.97 ± 0.16 | 1.73 ± 0.28 | 1.20 ± 0.16 | 1.4 ± 0.12 | - |
Supinine (IV) | 2020 | 283 | 3.76 ± 0.27 | 6.78 ± 0.33 | 7.13 ± 0.13 | 7.27 ± 0.44 | 8.06 ± 0.59 | 2.39 ± 0.40 |
Callimorphine (XIV) | 2024 | 297 | 12.47 ± 0.78 | 11.83 ± 1.02 | 13.75 ± 1.79 | 14.13 ± 1.70 | 12.58 ± 1.53 | - |
Amabiline (V) | 2027 | 297 | - | 2.5 ± 0.12 | 1.91 ± 0.36 | 2.71 ± 0.57 | 2.31 ± 0.34 | - |
3'-Acetylsupinine (VI) | 2108 | 325 | - | - | - | - | - | 5.32 ± 0.85 |
Unidentified PA | 2163 | - | - | - | - | - | - | 1.41 ± 0.29 |
Intermedine (VII) | 2167 | 299 | 52.99 ± 1.0 | 47.26 ± 0.37 | 48.01 ± 0.52 | 47.98 ± 2.15 | 48.41 ± 1.05 | - |
Lycopsamine (VIII) | 2175 | 299 | 18.10 ± 0.61 | 15.92 ± 0.42 | 17.57 ± 1.04 | 16.71 ± 0.51 | 17.04 ± 1.74 | - |
Rinderine (IX) | 2185 | 299 | - | - | - | - | - | 9.11 ± 0.55 |
3'-Acetylintermedine (X) | 2220 | 341 | - | 1.26 ± 0.08 | - | - | - | 1.29 ± 0.27 |
Unidentified PA | 2231 | - | 2.99 ± 0.29 | 2.05 ± 0.10 | 1.76± 0.53 | 1.95 ± 0.49 | 2.48 ± 0.50 | - |
3'-Acetylrinderine (XI) | 2245 | 341 | - | - | - | - | - | 79.65 ± 1.51 |
Unidentified PA | 2543 | - | 2.42 ± 0.33 | 2.08 ± 0.22 | 2.47 ± 0.16 | 1.40 ± 0.35 | 1.42 ± 0.41 | - |
Unidentified PA | 2581 | - | 2.10 ± 0.33 | 1.23 ± 0.19 | 1.70 ± 0.21 | 1.13 ± 0.36 | 1.61 ± 0.30 | - |
aRetronecine (I) was identified by coinjection of pure substances; all other PAs were characterized by comparison with literature retention indices and mass fragmentation patterns (see Table S1);
bretention indices.
DCA showed that PAs did not cluster host plants and moths; 3'-acetylrinderine (XI) may explain the host plant cluster, while intermedine (VII) explains the moth cluster (Figure 2). The lack of retronecine (I) in eggs (see Table 1) may explain why eggs were clustered apart from the other developmental stages of S. figulina.

Figure 2 Detrended correspondence analysis (DCA) for pyrrolizidine alkaloids (roman numbers; u: unidentified PA, see Table 1) in leaves of Heliotropium transalpinum and eggs, larvae, pupae, males, and females of Scearctia figulina.
We did not find any significant differences in PA concentrations in larvae (32.3 ± 2.7 µg of PAs per mg of dry weight), pupae (35.5 ± 3.0), males (41.4 ± 1.8), and females (34.0 ± 3.4) of lab-reared S. figulina (one-way ANOVA, F3,36 = 1.974, P = 0.135). Eggs were not quantified since PAs in gregarious eggs did not reach the threshold for the colorimetric analysis. The leaves of H. transalpinum had approximately fifty times less PAs than S. figulina (0.75 ± 0.05 µg mg-1, n = 5).
Chemical defense of Scearctia figulina against predators
All three predators consumed all the palatable preys offered. There is a significant difference in the response of three predators regardless of the sex of adult S. figulina (GLM, log-likelihood = −57.343, df = 2, χ2 = 22.786, P = 0.002). The orb-weaving spider N. clavipes released all 40 adults bioassayed (100%), while the wolf spider L. erythrognatha preyed upon 3 of 40 adults (92.5% release), and the chick G. gallus preyed upon 7 out of 40 (82.5% release). A post hoc pair-pair comparison showed that N. clavipes preyed on lesser S. figulina than L. erythrognatha (log-likelihood = −38.666, df = 1, χ2 = 15.518, P < 0.001) and G. gallus (log-likelihood = −38.000, df = 1, χ2 = 22.747, P < 0.001), but no differences occurred between the response of G. gallus and L. erythrognatha (log-likelihood = -38.229, df = 1, χ2 = 3.392, P = 0.065). No significant differences were found between sexes (log-likelihood = −30.617, df = 1, χ2 = 0.109, P = 0.741), and there was no interaction between predator response and sex (log-likelihood = −0.488, df = 2, χ2 = 0.258, P = 0.879). No significant difference was found between L. erythrognatha and G. gallus in relation to larvae of S. figulina (log-likelihood = −14.50, df = 1, χ2 = 1.412, P = 0.235). The wolf spider preyed upon one larva out of 13 bioassayed (92.3% release) and the chick preyed upon 4 out of 18 (77.8% release).
Discussion
The PA profile of H. transalpinum has already been described by Trigo et al.14 We did not find the PAs transalpinecine, subulacine, and their stereoisomers, as given by Medina et al.25 As other arctiines that feed on PA plants as larvae,12-14 our results showed that S. figulina also uptakes these defensive compounds from its host plant. Although no analyses were performed to check if PAs are present only in the N-oxide form, we have made this assumption, since it has occurred in other arctiines studied.3,4,15,16 The uptake of PA from plants may only occur in larvae of S. figulina, since we did not observe adults in PA sources.
Regarding PA patterns, the alkaloid profile in moths might be expected to be a fingerprint of the host plant profile. However, this does not occur in S. figulina, as well as in other PA specialist lepidopterans. When these insects feed on plants with 7S-configured PAs, they invert this chiral center to the 7R configuration.35 The raison-d'etre for this is male sexual pheromone. In danaine and ithomiine butterflies and arctiine moths, all dihydropyrrolizine male pheromones identified have 7R configuration.6-8 As the clades danaine/ithomiine and arctiine are not monophyletic,36 the 7R trait may have evolved twice, independently. Trigo et al.35 suggested that ancestral PA-plants with 7R alkaloids may have molded this trait. In PA specialist leaf beetles, such as Longitarsus and Platyphora species, dihydropyrrolizines are not present, and these beetles show 7S PAs sequestered from their host plants, even if some epimerization to 7R does occur.37,38 However, in S. figulina we did not find an androconial organ that produces pheromones. It is suggesting that epimerization in S. figulina might be due to a phylogenetic constraint, since the more basal species Amerila spp.11 have androconial organs and might produce PA-derived pheromones.39
Arctiine moths also biosynthesize their own PA metabolites; the insect PAs.16,19 These alkaloids are biosynthesized by the esterification of a necine base of plant origin (generally the 7R retronecine) with a necic acid of insect amino acid origin.16,19,40 Only two types of insect PAs were found: PAs of the callimorphine type, whose necic acid is derived from the isoleucine, and of the creatonotine type, whose necic acid is a valine derivative.16,20,40 Larvae, pupae, and adults of both sexes of S. figulina showed both insect PA types, suggesting that biosynthesis occurs in larvae, which is similar to other arctiines that obtain PAs from larval host plants.16 The function of insect PAs was first associated to the biosynthesis of dihydropyrrolizines, male sexual pheromones, via intramolecular transesterification of insect PA O9-ester to insect-PAO7-ester,19 although this kind of PA is present in both males and females. Our research group has already shown that insect PAs are also defensive compounds against predators.15,41 When arctiine moths feed on plants containing retronecine, which is innocuous and ineffective against predation,41 the biosynthesis of insect PAs would be a mechanism to maximize insect chemical defense. However, it is difficult to determine which would be the first function of insect PAs, precursor of sexual pheromones or defensive compounds. Interestingly enough, no free retronecine was found in host plant leaves. Therefore, the moth may first epimerize 7S to 7R PAs, and then, hydrolyze them to proceed to a further esterification into insect PAs. Additionally, the moth also deacetylates the PAs from the host plant. We can speculate that an O7-acetyl moiety would prevent the transesterification and the further biosynthesis of insect PAs.
Another unanswered question on insect PAs is: why do some arctiini species biosynthesize a single type of these alkaloids, while others biosynthesize both types? This is more noticeable when we compare the PA sequestration by the pericopines S. figulina and H. syma. They have a very similar life-style, feeding on the same host plant. The PA epimerization and deacetylation is similar in both species, but insect PA biosynthesis is quite different.14Searctia figulina biosynthesizes five insect PAs, of the callimorphine and creatonotine types, while H. syma biosynthesizes just one callimorphine type PA.14 A comparative study of insect PA biosynthesis may shed light on the evolutionary mechanisms underlying these findings.
The defensive role of PAs against predators is well known and has been extensively exploited in literature. A further discussion on this function in S. figulina would be rhetoric. However, assuming that PAs are responsible for S. figulina defenses, our results showed that different predators have different PA release thresholds. The orb-weaving spider N. clavipes was the more sensitive predator, releasing all bioassayed individuals. The wolf-spider was less responsive, followed by chicks. We can suggest that a possible high encounter rate of N. clavipes with PA-defended insects would account for the high responsiveness. As the orb-weaving spiders build their nets in forest corridors and patches,42 they may size and release many flying PA insects. The encounter rate of L. erythrognatha with PA insects would be lower, since they wander on floors of several environments,43 where PA-sequestering insects are not so common. Finally, chicks are a model of visually oriented predadors, such as birds. To what extent they share an evolutionary history with PA insects is unkown. In addition, like the wolf spider, chicks show a wandering habit, which decreases the probability of encountering PA insects.
Another point that deserves attention in relation to the defensive role of PAs in adults of S. figulina is a possible mimicry pattern with other moths once this species, the arctiine Episcea extravagans (Arctiini: Pericopina) and several josiini moths (Notodontidae: Dioptidae) share the same wing color pattern.44,45 We can hypothesize that E. extravagans, as S. figulina and other arctiines, is a PA-adapted insect using these compounds for defenses against predators. It is known that Josiini moths use Passiflora as larval host plants,45 which have cyanogenic glucosides in their leaves46 and they may use these compounds for defense against predators in the same way as other cyanide-feeder insects such as unpalatable Heliconius butterflies.46 If so, Müllerian mimicry can explain the similarities of the wing patterns of these insects.
Conclusions
Although our data on S. figulina have shed more light on PAs in arctiine moths, an extensive survey on PA acquisition syndromes and insect PA types from this subfamily should be carried out to suggest any evolutionary trends.
Supplementary Information
Supplementary information is available free of charge at http://jbcs.sbq.org.br as PDF file.
0103-5053-jbchs-27-08-1437-suppl01.pdf