Introduction
The Saturniidae, also called the giant silkworm, is one of the most eye-catching lepidopteran families. They are usually large and have a great commercial value as ornamental or edible insects (Li et al., Reference Li, Zhang, Lin and Terenius2017). Their cocoons, which are made from silk fibres and sericin binder, are widely used as an important natural polymer composite (Chen et al., Reference Chen, Porter and Vollrath2012). The family comprises many genera and species, and both the larval and adult stages of Saturniidae show much morphological diversity, so that the phylogenetic interpretation of the family is difficult (Friedlander et al., Reference Friedlander, Peigler, Regier and Mitter2002). Nowadays, mitochondrial DNA sequence data are increasingly used to infer saturniid phylogenetic relationships (Chen et al., Reference Chen, Li, Chen, Wang, Li, Xia, Zeng, Li, Liu and Qin2014; Kim et al., Reference Kim, Kim, Jeong and Kim2018; Rubinoff and Doorenweerd, Reference Rubinoff and Doorenweerd2020). However, sequences of Asian species are still scarce, which greatly hindered further research (Chen et al., Reference Chen, Zhang, Jin, Yang, Li and Liu2022; Lu et al., Reference Lu, Huang, Naumann, Kitching, Xu, Sun and Wang2022; Nethavhani et al., Reference Nethavhani, Straeuli, Hiscock, Veldtman, Morton, Oberprieler and van Asch2022).
Within the Bombycoidea spinning behaviour is regarded as a plesiomorphic behaviour of Saturniidae (Chen et al., Reference Chen, Porter and Vollrath2012; Tsukada et al., Reference Tsukada, Kajiura, Shoumura and Satoh2012). In other Bombycoidea, such as the Sphingidae, the spinning behaviour is degenerated or completely lost, and is replaced by an underground-pupation stage (Lampe, Reference Lampe2010). This loss of spinning behaviour in different taxa has not been studied systematically. Most of the non-cocooning saturniid groups were reported from Africa, and only species of Salassa and Aglia were reported to be non-cocooning in Asia (Lampe, Reference Lampe2010). Hence to explore the evolution of spinning behaviour (cocooning), more mitogenomic evidence of non-cocooning saturniids is needed, particularly from Asian taxa.
Agliinae and Salassinae are two monogeneric subfamilies of Saturniidae whose life histories and mitogenomes are still poorly known (Regier et al., Reference Regier, Grant, Mitter, Cook, Peigler and Rougerie2008). The genus Aglia Ochsenheimer, 1810 has a Palaearctic distribution, whereas the genus Salassa Moore, 1859 is distributed in both the Palaearctic and Oriental realms (Lampe, Reference Lampe2010). The life histories of only one Aglia and two Salassa species have been reported (Lampe, Reference Lampe2010). Yet, although the pupation process in these two genera has not been described or studied in detail, Aglia and Salassa species appear to be non-cocooning under indoor rearing conditions (Lampe, Reference Lampe2010). Hence Agliinae and Salassinae are possibly the only non-cocooning Saturniidae in Asia. However, whether non-cocooning is a natural and universal behaviour in these two subfamilies requires further confirmation.
Against this background, the present study reports for the first time on the successful indoor rearing of several Aglia and Salassa species. Pupation behaviour of two Aglia and five Salassa species was observed, and mitogenomes of both genera were sequenced for the first time. Combining these mitogenomic and behavioural data may help to trace the evolutionary gain and loss of cocooning behaviour in the phylogenetic framework of the Saturniidae.
Materials and methods
Collection, rearing and observation of saturniid species
Specimens used in this study were collected in China between 2021 and 2022. Eggs of Aglia tau (Linnaeus, 1758), Aglia homora Jordan, 1911, Salassa shuyiae Zhang and Kohll, Reference Zhang and Kohll2008, Salassa mesosa Jordan, 1910, Salassa paratonkiniana Brechlin, 2009, Salassa daxuensis Brechlin, 2015 and Salassa arianae Brechlin & Kitching, 2010 were obtained from wild-caught adult females through light traps.
Larvae were reared and observed in mesh cages inside a room, where temperature was controlled by an air conditioner. Food of larvae was replaced every two days before the leaves dried. To observe their natural pupation behaviour, mature larvae were moved to large tree branches inserted in pots with soft soil and moss.
Sample identification and DNA extraction
The morphological identification was mainly based on Naumann and Lalhmingliani (Reference Naumann and Lalhmingliani2019) and Naumann, Brosch and Nässig (Reference Naumann, Brosch and Nässig2003). All samples were preserved in 85% ethanol, and stored at −20℃ in the Lepidoptera Collection of Nanjing Normal University. Genomic DNA was extracted from muscle tissue of each species using the TIANamp Genomic DNA Kit (TIANGEN, Beijing, China). DNA concentrations were determined by a Nanodrop 2000 spectrophotometer.
Mitogenome sequencing and assembly
DNA was sequenced using next-generation sequencing (NGS) on the Illumina NovaSeq platform. The library of each sample was constructed by TruSeqTM DNA Sample Prep Kit (insert size 400 bp). All libraries were sequenced based on the PE150 mode (paired-end, 2 × 150 bp). Around 4 GB of raw data from each sample were trimmed so that adaptor contamination, and short and low-quality reads were removed using Fastp (Chen et al., Reference Chen, Zhou, Chen and Gu2018). MitoZ was used for the assembly of the mitogenomes (Meng et al., Reference Meng, Li, Yang and Liu2019), with the kmers-megahit = 79, 99, 119, 141. To verify the accuracy of NGS sequencing results, COI gene fragments of all samples were sequenced using Sanger sequencing with the primer pair (LCO1490 as a forward primer, 5’-GGTCAACAAATCATAAAGATATTGG-3’; HCO2198 as a reverse primer, 5’-TAAACTTCAGGGTGACCAAAAAATCA-3’) (Folmer et al., Reference Folmer, Black, Hoeh, Lutz and Vrijenhoek1994). r-Taq polymerase (Takara, Beijing, China) was used for the polymerase chain reaction based on the method of Yanai et al. (Reference Yanai, Sartori, Dor and Dorchin2017).
Mitogenome annotation and analysis
The mitogenomes were annotated using the MITOS WebServer (Bernt et al., Reference Bernt, Donath, Juhling, Externbrink, Florentz, Fritzsch, Putz, Middendorf and Stadler2013) and MitoZ (Meng et al., Reference Meng, Li, Yang and Liu2019) and aligned with available sequences of Ephemera from GenBank using ClustalW in MEGA 11 (Tamura, Stecher and Kumar, Reference Tamura, Stecher and Kumar2021). The secondary structures of tRNAs were determined using tRNAscanSE 2.0 (Lowe and Chan, Reference Lowe and Chan2016) and the MITOS WebServer (Bernt et al., Reference Bernt, Donath, Juhling, Externbrink, Florentz, Fritzsch, Putz, Middendorf and Stadler2013). The circular mitogenome maps were portrayed using the visualisation module of MitoZ (Meng et al., Reference Meng, Li, Yang and Liu2019).
Four newly sequenced mitogenomes, viz. Aglia tau (Linnaeus, 1758), Aglia homora Jordan, 1911, Salassa shuyiae Zhang and Kohll, Reference Zhang and Kohll2008 and Salassa mesosa Jordan, 1910 (table 1) were used in the comparative analysis. Nucleotide composition and relative synonymous codon usage were assessed with MEGA11 (Tamura, Stecher and Kumar, Reference Tamura, Stecher and Kumar2021). The AT-skew and GC-skew were calculated by the formulas: AT-skew = (A − T)/(A + T) and GC-skew = (G − C) / (G + C) (Perna and Kocher, Reference Perna and Kocher1995). A sliding window analysis of 200 bp (step size 20 bp) was used to estimate nucleotide diversity (Pi) of 13 PCGs by DnaSP V6 (Rozas et al., Reference Rozas, Ferrer-Mata, Sánchez-DelBarrio, Guirao-Rico, Librado, Ramos-Onsins and Alejandro2017). K2P distances (Kimura, Reference Kimura1980) of the mitogenomes were estimated using MEGA11. The ratio of nonsynonymous/synonymous (Ka/Ks) substitution rates of 13 PCGs was calculated by DnaSP V6 (Rozas et al., Reference Rozas, Ferrer-Mata, Sánchez-DelBarrio, Guirao-Rico, Librado, Ramos-Onsins and Alejandro2017).
Table 1. GenBank accession numbers of mitogenome sequences of species used in this study with references on their life history and cocooning behaviour
Phylogenetic analysis
A total of 37 mitogenomes of four families were used for the phylogenetic reconstruction, including 34 species of 20 genera within Saturniidae as ingroup, and three species from three families as outgroup (Table 1).
The nucleotide sequences of 13 PCGs were aligned with MAFFT (Katoh and Standley, Reference Katoh and Standley2013) L-INS-i (accurate) strategy and codon alignment mode in PhyloSuite v1.2.2 (Zhang et al., Reference Zhang, Gao, Jakovli´c, Zou, Zhang, Li and Wang2020). The ambiguous sites were removed using Gblocks 0.91b (Talavera and Castresana, Reference Talavera and Castresana2007). Individual genes were then concatenated using PhyloSuite v1.2.2 (Zhang et al., Reference Zhang, Gao, Jakovli´c, Zou, Zhang, Li and Wang2020). The PCGs matrix, with all codon positions of the 13 protein-coding genes was used for the reconstruction of the phylogenetic relationships. For the phylogenetic reconstruction, Bayesian inference (BI) and maximum likelihood (ML) analyses were performed. The partition model of each dataset was selected by the ModelFinder based on Bayesian information criterion (BIC) and AICc (standard correction to Akaike information criterion) (Kalyaanamoorthy et al., Reference Kalyaanamoorthy, Minh, Wong, von Haeseler and Jermiin2017). BI phylogenies were reconstructed using MrBayes 3.2.6 through the online CIPRES Science Gateway (Miller, Wayne and Terri, Reference Miller, Wayne and Terri2011; Ronquist et al., Reference Ronquist, Teslenko, Mark, Ayres, Darling, Höhna, Larget, Liu, Suchard and Huelsenbeck2012), with the following settings: two parallel runs with four Markov chains were run for 10 million generations (with a sample frequency of 1000), with a burn-in of 25% trees. RAxML 8.2.0 was used for ML analysis with the GTR + Γ + I model and 1000 bootstrap replicates (Stamatakis, Reference Stamatakis2014). FigTree 1.4.2 (http://tree.bio.ed.ac.uk/software/figtree/; accessed on 5 July 2021) was used for the editing of phylogenetic trees.
Results
Life history of Aglia homora Jordan, 1911
Three host plants were used, viz. Juglans regia, Quercus glandulifera and Salix babylonica.
Eggs were obtained from wild adult females caught in China: Taibai Mountain, Baoji City, Shaanxi Province, 1000 m a.s.l., 19-V-2021.
Eggs were amber in colour (fig. 1A). L1 larva light green, body covered with small light yellowish thorns and with six long spiny thorns: one pair right behind the head, pointing forward, bifurcated apically; one pair slightly behind the anterior pair, longest amongst the six ones, pointing backward, bifurcated apically; one near the posterior end, bifurcated apically, pointing upward, and one located at the posterior end, shortest amongst the six, pointing backward (fig. 1B). Except for the shortest posterior thorn which is totally reddish, the other five thorns are reddish brown apically and basally, light-yellowish medially (fig. 1B). L2 larva generally similar to L1, but with several oblique white lines dorsally on abdomen, body surrounded by a circular white line on lateral margin; shape and colour pattern of the six long thorns same as L1 (fig. 1C). L3 larva similar to L2, but with a pair of red dots on lateral margin below the second pair of thorns (fig. 1D). L4 larva without small light yellowish thorns on the abdomen, but with white tubercles all over the body. There are only five, instead of six, long thorns, which are green with apical and median red dots. The lateral red dot in L3 becomes an eyespot in L4 larva (fig. 1E). L5 larva without thorns, lateral eyespots larger, body covered with tiny white tubercles, body colour turned reddish from green when fully matured before pupation (fig. 1F). Pupae brownish black, darker at high humidity, yellowish brown at low humidity; abdominal segments V–VII with a patch of dorsal spines (fig. 1G). Adult morphology: see original description (fig. 1H).
Figure 1. Life history of Aglia homora. (A) eggs. (B) L1 larva. (C) L2 larva. (D) L3 larva. (E) L4 larva. (F) L5 larva. (G) pupa. (H) female adult.
Of the three host plant species studied, J. regia is the most successful with a larval survival rate above 50%, larvae pupate in 30–35 days after hatching and are on average larger than on other host plants. Larvae that feed on S. babylonica have a survival rate of ca. 50%, pupate in 40–50 days after hatching and remain on average smaller than larvae feeding on J. regia. Although several larvae that fed on Q. glandulifera completed their life cycle successfully, their survival rate was less than 10%, their larval stage lasted more than 100 days, and their pupae remained smaller. Overall, larvae did not aggregate together, but clearly preferred tender leaves and avoided old leaves. To simulate the natural temperature conditions, pupae were kept at 5℃ from September 2021 to March 2022, after which they were kept at 15–20°C. With this latter temperature regime adults emerged within 25–35 days.
Life history of Aglia tau (Linnaeus, 1758)
Three host plants were used, viz. J. regia, Q. glandulifera and S. babylonica.
Eggs were obtained from wild adult females caught in China: Suiyang Town, Dongning City, Mudanjiang City, Heilongjiang Province, 600 m a.s.l., 24-V-2021.
Eggs were amber in colour (fig. 2A). L1 larva light green, body covered with multiple paired light yellowish thorns and with six long spiny thorns: one pair right behind the head, pointing forward; one pair, involving the longest thorns, slightly behind the anterior pair, pointing backward; one thorn near the posterior end and pointing upward; and finally, one thorn, i.e. the shortest one of the six, located at the posterior end and pointing backward (fig. 2B). Except for the shortest posterior thorn which is totally reddish and non-bifurcated, the other five thorns are apically bifurcated and share a similar colour pattern, viz. reddish brown apically and basally, light-yellowish medially (fig. 2B). L2 larva generally similar to L1, but with several oblique white dorsal lines on the abdomen and a circular white line surrounding the body. The shape and colour pattern of the six long thorns is the same as in L1 (fig. 2C). L3 larva similar to L4, body covered with tiny white dots, a pair of white dorsal thorns present on each abdominal segment. The six long thorns are relatively shorter in instars; lateral margin below the second pair of thorns with two red dots (fig. 2D, E). L5 larva without thorns, body covered with tiny white tubercles, the entire body turned reddish from green when fully matured before pupation (fig. 2F). Larval morphology similar to A. homora, but with thinner thorns, L4 and L5 larvae with smaller dots below the second pair of thorns. Pupae brownish black, darker at high humidity, yellowish brown at low humidity; abdominal segments V–VII with a patch of dorsal spines (fig. 2G). Adult morphology see original description (fig. 2H).
Figure 2. Life history of Aglia tau. (A) eggs. (B) L1 larva. (C) L2 larva. (D) L3 larva. (E) L4 larva. (F) L5 larva. (G) pupa. (H) male adult.
Of the four host plants studied, J. regia is the most successful, with a larval survival rate of ca. 80%, larvae pupate in 25–35 days after hatching and larvae are on average larger than on other host plants. Larvae fed on S. babylonica have a survival rate of ca. 50%, pupate 40–50 days after hatching, and are on average smaller than larvae fed on J. regia. Larvae fed on Q. glandulifera have a survival rate of ca. 50%, need at least 80 days to reach pupation time, and yield pupae that are smaller than on the other host plants. Finally, larvae fed on Acer truncatum all died in the L1 or L2 stage. Larval behaviour similar to A. homora. To simulate the natural temperature conditions, pupae were kept at 5℃ from September 2021 to March 2022, after which they were kept at 15–20°C. With this latter temperature regime adults emerged within 14–20 days.
Life history of Salassa shuyiae Zhang and Kohll, Reference Zhang and Kohll2008
Two host plants were used, viz. Camellia sinensis and Litsea pungens.
Eggs were obtained from wild adult females caught in China: Bawang Mountain, Changjiang Li Autonomous County, Hainan Province, 600 m a.s.l., 15-XI-2021.
Eggs white, oblate (fig. 3A). L1 larva green, flat, leaf-shaped in dorsal view; body margin marked by a light yellowish line and black setae. Such setae also occur along the midline; midline consists of a green band between a pair of white lines; head and posterior end sharp, with a light brownish apex (fig. 3B). L2 larva without setae and obvious light yellowish line on lateral margin; head, caudal region and midline densely covered with tiny tubercles; midline consists of a white band; near the midline at the posterior margin of each abdominal segment there are 1–3 black dots; head slender triangular shaped and curved upward, both head and posterior end with a sharp apex, light brownish apically (fig. 3C). L3, L4, and L5 larva generally similar to L2 larva (fig. 3D–F); except larger body, L4 and L5 larva have a large patch of black dots near the lateral margin of several medial abdominal segments (fig. 3E, F). The L5 larva has a relatively short head (fig. 3F). In the L6 larva the midline consists of a green band between a pair of white lines; head shaped as an equilateral triangle in dorsal view; both head and posterior end blunt (fig. 3G). When L6 larvae reach maturity, they change colour from green to red (fig. 3G). Pupae brownish black, darker at high humidity; yellowish brown at low humidity, terminal segment sharp and slender (fig. 3H). Adult morphology see original description (fig. 3I).
Figure 3. Life history of Salassa shuyiae. (A) eggs. (B) L1 larva. (C) L2 larva. (D) L3 larva. (E) L4 larva. (F) L5 larva. (G) L6 larva. (H) pupa. (I) female adult.
Eggs kept at 10–13℃, hatched within 45 days, in the period 20–28 December 2021. Hatching could be stimulated by water, mature eggs hatched within two hours after watering, keeping dry can delay the hatching process and increase failure rates. Seven host plants were used in this study. Amongst them, C. sinensis and L. pungens yielded similar results, with a survival rate above 50% until pupation. Yet, larvae feeding on L. pungens become slightly larger than larvae feeding on C. sinensis. Larvae reared at 17–20℃, instars I–VI last for ca. 10, 11, 11, 12, 20, and 30 days respectively. We also experimented with Ligustrum compactum, Ilex chinensis, C. sinensis, and Deyeuxia langsdorffii, all the larvae died before L2. During this study, larvae showed no aggregation behaviour, smaller larvae usually rest under leaves at the main vein, larger larvae rest on the branches. Pupae were kept in 25–30℃, watered once a week. Adults emerged after ca. 130 days, but only 10% emerged successfully, 70% got out of the pupae but did not unfold their wings, 20% died in pupae. This may have been caused by inappropriate temperature and humidity conditions.
Life history of Salassa mesosa Jordan, 1910
Three host plants were used, viz. C. sinensis, J. regia and L. pungens.
Eggs were obtained from wild adult female caught in China: Gulinqing, Maguan County, Wenshan Prefecture, Yunnan Province, 1200 m a.s.l., 31-VIII-2022.
Eggs white, oblate (fig. 4A). L1 larva similar to S. shuyiae, but body margin marked by a wider light yellowish line, head light brownish and bifurcated apically (fig. 4B). Compared with S. shuyiae, L2 larvae of S. mesosa with shorter head, which is slender triangular shaped with a sharp apex, body midline consists of a green band between a pair of white lines, lateral margin covered with small white denticles (fig. 4C); L3, L4, and L5 larvae generally similar to L2 larva (fig. 4D–F); except that they are larger, the light yellow margin gradually narrower in later instars, head of L3 and L4 larvae shortened into an equilateral triangle in dorsal view, but rounded in L5 (fig. 4D–F). Body colour darker in L5 larva, changing from dark green to brownish when fully matured before pupation (fig. 4F). Pupae brownish black, darker at high humidity, yellowish brown at low humidity (fig. 4G). Adult morphology see original description (fig. 4H).
Figure 4. Life history of Salassa mesosa. (A) eggs. (B) L1 larva. (C) L2 larva. (D) L3 larva. (E) L4 larva. (F) L5 larva. (G) pupa. (H) mating adults.
Eggs kept at 20℃ hatched within 15 days. Hatching process and larval behaviour similar to S. shuyiae. Of the three host plants used in this study, C. sinensis and L. pungens yielded similar results, viz. larvae pupate within 55 days with a survival above 70% until pupation. Larvae fed on J. regia pupate within 60 days with a survival rate of ca. 35%. We also tried Lindera glauca as host plant, but all the larvae died in L2. All the larvae pupated in December 2022. Next, the pupae were kept at 8°C for 90 days and watered once a week. Adults emerged in February 2023, with a success rate of 90%.
Life history of Salassa paratonkiniana Brechlin, 2009
Host plant: Juglans regia.
Eggs were obtained from wild adult females caught in China: Shimen Gorge, Ailao Mountain, Yuxi City, Yunnan Province, 2200 m a.s.l., 6-VII-2022.
Eggs white, oblate (fig. 5A). L1 similar to S. mesosa (fig. 5B). L2 larva distinguishable from S. mesosa through a black band just before the centre of body midline, and lateral margin marked by a purplish brown line (fig. 5C). L3 larva generally similar to L2 larva except for its larger body size (fig. 5D). The head of L4 and L5 larvae is shortened and rounded apically (fig. 5E, F). Their body colour changes from green to brownish when fully matured before pupation. Pupae brownish black, darker at high humidity, yellowish brown at low humidity (fig. 5G). Adult morphology see original description (fig. 5H).
Figure 5. Life history of Salassa paratonkiniana. (A) eggs. (B) L1 larva. (C) L2 larva. (D) L3 larva. (E) L4 larva. (F) L5 larva. (G) pupa. (H) female adult.
Eggs kept at 20℃, hatched within 15 days. Hatching process and larval behaviour similar to S. shuyiae. Larval survival rate was ca. 30%, and larvae pupated within 75 days. The larval colour patterns is far more variable than in the other species in this study. Pupae were kept at 8℃ for 135 days and watered once a week, after which they were kept at 20°C. With this latter temperature regime adults emerged within 75 days.
Life history of Salassa daxuensis Brechlin, 2015
Two host plants were used, viz. C. sinensis, J. regia.
Eggs were obtained from wild adult females caught in China: Bifeng Gorge, Ya'an City, Sichuan Province, 1000 m a.s.l., 26-VI-2022.
Eggs white, oblate (fig. 6A). L1 larva similar to S. mesosa (fig. 6B). L2 larva also similar to S. mesosa but without a black dot on the midline (fig. 6C). L3 and L4 larvae similar to L2 larva, with the same colour pattern and similar head (fig. 6D, E). Head of L5 and L6 larvae rounded, body colour as in L4 larva (fig. 6F, G). Colour of L6 larva changes from green to reddish brown when fully matured before pupation (fig. 6G). Pupae brownish black, darker at high humidity, yellowish brown at low humidity (fig. 6H). Adult morphology see original description (fig. 6I).
Figure 6. Life history of Salassa daxuensis. (A) eggs. (B) L1 larva. (C) L2 larva. (D) L3 larva. (E) L4 larva. (F) L5 larva. (G) L6 larva. (H) pupa. (I) female adult.
Eggs kept at 20℃, hatched within 20 days. Hatching process and larval behaviour similar to S. shuyiae. Survival rate was ca. 35% with J. regia as host plant and larvae pupated within 45 days, while survival rate was 10% with C. sinensis and larvae pupated within 100 days. We also tried Phellodendron amurense but larvae remained smaller on this host plant and died in the L5 stage. Pupae were kept at 8℃ for 120 days and watered once a week, after which they were kept at 18°C. With this latter temperature regime adults emerged within 75 days.
Life history of Salassa arianae Brechlin & Kitching, 2010
Host plant: Juglans regia.
Eggs were obtained from wild adult females caught in China: Ningshan County, Ankang City, Shaanxi Province, 1500 m a.s.l., 16-VI-2021.
Eggs white, oblate (fig. 7A). L1 larva similar to S. mesosa (fig. 7B). L2 larva similar to S. mesosa, but can be recognised by the head with a relatively blunt apex, body midline with two anterior black dots, and abdominal segments III–VI with paired dark dots posteriorly near midline (fig. 7C). L3 and L4 larvae similar to L2 larva, with the same colour pattern; head shaped as an equilateral triangle (fig. 7D, E). Colour pattern of L5 larva similar to L4 larva; head of L5 larva rounded. Larval body colour changes from green to reddish brown when fully matured before pupation (fig. 7F). Pupae brownish black, darker at high humidity, yellowish brown at low humidity (fig. 7G). Adult morphology see original description (fig. 7H).
Figure 7. Life history of Salassa arianae. (A) eggs. (B) L1 larva. (C) L2 larva. (D) L3 larva. (E) L4 larva. (F) L5 larva. (G) pupa. (H) female adult.
Eggs kept at 20℃, hatched within 16 days. Hatching process and larval behaviour similar to S. shuyiae. Survival rate was ca. 40% until pupation, within 80 days. Pupae were kept at 2–4℃ for 120 days and watered once a week, after which they were kept at 18°C. With this latter temperature regime adults emerged within 75 days.
Mitogenome structure of Aglia and Salassa
The circular complete mitogenome sequences of A. homora, A. tau, S. shuyiae and S. mesosa were 15,354 bp, 15,334 bp, 15,290 bp and 15,289 bp long respectively (fig. 8A–D). These mitogenome sizes are comparable to those of other Saturniidae (Table 1). All of them have 37 genes, including 13 protein-coding genes (PCGs), 22 tRNA genes and two rRNA genes (12S and 16S), with a large non-coding region known as the control region (CR) between s-rRNA and trnM. Nine PCGs and 14 tRNAs are located on the majority strand (H-strand), while the remaining 14 genes are encoded on the minority strand (L-strand) (Table 2). The four new mitogenomes have the typical gene content and arrangement of most Lepidoptera including Saturniidae (Table 2), with the gene order trnM-trnI-trnQ, which is the only difference from the ancestral gene order of insects viz. trnI-trnQ-trnM (Cao et al., Reference Cao, Ma, Chen and Yang2012).
Figure 8. Mitochondrial maps of four species. (A) Aglia homora. (B) Aglia tau. (C) Salassa shuyiae. (D) Salassa mesosa.
Table 2. Features of the mitogenomes of four saturniid species
Four PCGs (ND1, ND4, ND4L, and ND5) in the four new mitogenomes are not located on the H-strand, but on the L-strand, which is consistent with other saturniid mitogenomes (Table 2) (Chen et al., Reference Chen, Li, Chen, Wang, Li, Xia, Zeng, Li, Liu and Qin2014, Reference Chen, Zhang, Jin, Yang, Li and Liu2022). All PCGs start with conventional initiation codons ATN (ATG, ATT, ATA, and ATC), as in other saturniid species, except for the COI genes which start with CGA (Table 3). All the PCGs are terminated by typical TAN codons, i.e. TAG for ND1 and ND2 of S. mesosa, and TAA for all the other PCGs (Table 3).
Table 3. Initiation codon - termination codon of PCGs and anticodons of tRNAs of the mitogenomes of four saturniid species
The four mitogenomes contained 22 typical tRNA genes, whose sizes ranged from 56 bp (trnE of S. mesosa) to 71 bp (trnD of S. mesosa and trnK of A. tau, A. homora, S. shuyiae and S. mesosa) (Table 2). The secondary structures of the tRNA genes could be folded into the typical clover-leaf structure except for trnS1 in the four species and trnE in S. mesosa (fig. 9A–D) whose T-arm were replaced by a loop. This peculiar structure of trnS1 is commonly found in other saturniid mitogenomes (Lu et al., Reference Lu, Huang, Naumann, Kitching, Xu, Sun and Wang2022), but for trnE it was never reported before. The size of the CR between s-rRNA and trnM ranged from 332 bp (S. mesosa) to 340 bp (A. homora and A. tau) (Table 2).
Figure 9. Putative secondary structure of trnE. (A) Aglia homora. (B) Aglia tau. (C) Salassa shuyiae. (D) Salassa mesosa.
Phylogenetic analysis
Most previous phylogenetic studies of Saturniidae were based on morphological characters or protein coding sequences (Friedlander et al., Reference Friedlander, Horst, Regier, Mitter, Peigler and Fang1998; Regier et al., Reference Regier, Grant, Mitter, Cook, Peigler and Rougerie2008; Rubinoff and Doorenweerd, Reference Rubinoff and Doorenweerd2020). Regier et al. (Reference Regier, Grant, Mitter, Cook, Peigler and Rougerie2008) reconstructed the phylogenetic tree of Saturniidae using four protein coding genes viz. CAD, DDC, period and wingless. Their system is up to now the most commonly used with Agliinae and Salasinae as two monogeneric subfamilies, and Salassinae as sister group of Saturniinae. In this system, the 18 genera used in this study except Salassa and Aglia belong to the Saturniinae.
The present phylogenetic analysis is based on the 13 PCGs of the four new mitogenomes and the mitogenomes of all the saturniid genera available in GenBank (last visit on April 4, 2023) (Table 1). Three mitogenomes were selected for those genera for which more than three mitogenomes were available (Table 1). In total, the ingroup comprised 34 species from 20 genera, while the outgroup consisted of three species from three different families (fig. 10, Table 1). The resulting BI and ML trees had the same topology as shown in fig. 10 in which: (1) Neoris haraldi and Cricula andrei appear as single branches with sister group positions to all other ingroup taxa; (2) all these other ingroup taxa form a clade in which the Saturniini of the Saturniinae (including Actias, Antheraea, and Saturnia) consititute a well-supported subclade that is sister to a subclade comprising the other tribes of the Saturniinae along with the Agliinae and Salassinae (fig. 10). Yet these two subclades are only well-supported by BI, not by ML; (3) Within the non-Saturniini subclade, three further clades could be distinguished: one is the Attacini of the Saturniinae (including Attacus, Epiphora, Samia and Rhodinia), the second is the Agliinae + Salassinae, and finally the Bunaeini + Eochroini + Micragonini (fig. 10). Yet, only the Attacini clade was well-supported by both BI and ML, the two other clades were only well-supported by BI, not by ML. Hence, in contrast to previous studies (Friedlander et al., Reference Friedlander, Horst, Regier, Mitter, Peigler and Fang1998; Regier et al., Reference Regier, Grant, Mitter, Cook, Peigler and Rougerie2008; Rubinoff and Doorenweerd, Reference Rubinoff and Doorenweerd2020; Nethavhani et al., Reference Nethavhani, Straeuli, Hiscock, Veldtman, Morton, Oberprieler and van Asch2022), the present results do not support the monophyly of Saturniinae since the Agliinae and Salassinae were placed among the Saturniinae tribes (fig. 10). The most striking result of this work is the placement of Neoris haraldi and Cricula andrei, since both these species were formerly included in the Saturniini of the Saturniinae (Regier et al., Reference Regier, Grant, Mitter, Cook, Peigler and Rougerie2008; Rubinoff and Doorenweerd, Reference Rubinoff and Doorenweerd2020), whereas in the current analyses they are with strong support widely separated from the Saturniini (fig. 10).
Figure 10. Phylogenetic relationships of Saturniidae constructed by BI and ML methods based on DNA sequences of 13 mitogenomic PCGs. Posterior probability and bootstrap values on node are separated by a slash, red dots represent non-cocooning species, black dots represent cocooning species.
Discussion
Pupation behaviour of Aglia and Salassa
The five Salassa species reared in this study shared a similar pupation process (fig. 11A–F). The day before the larvae fully mature, their food intake decreases. Then the dorsal vessel mid-line becomes brown. Subsequently, a large amount of viscous excreta is discharged from their intestines, while the body colour changes from green to reddish brown (fig. 3G). After this colour transition, larvae spit out liquid from their mouth and spread it all over their body before they descend from trees to pupate in ground surface deposits (fig. 11A–F). The function of the liquid is unknown, but we speculate that it may protect the body during long-distance crawling. Larvae showed, indeed, strong crawling ability and can crawl far away from their host plant. Larvae do not dig into the soil, but look for cracks or press a shallow pit in surface deposits to create a space, after which they spin and pull the surrounding deposits to construct a pupal chamber (fig. 11A–F). They produce only a small amount of silk that does not form a closed cocoon (fig. 11A–F).
Figure 11. Pupation behaviours of Salassa and Aglia species. (A) Salassa shuyiae larva making pupal chamber. (B) Salassa shuyiae larva making pupal chamber. (C) pupa and pupal chamber of Salassa shuyiae. (D) Salassa mesosa larva making pupal chamber. (E) Salassa mesosa larva making pupal chamber. (F) Salassa mesosa pupa inside pupal chamber. (G) Aglia homora larva making pupal chamber. (H) Aglia homora larva making pupal chamber. (I) Aglia homora pupa inside pupal chamber.
The pupation behaviour of A. tau and A. homora was similar to that of the Salassa species, except for some morphological differences (fig. 11G–I). Mature larvae eat less, change from green to red, empty their body, cover their body with a liquid spit from their mouth, descend from their host plant and pupate in the ground (fig. 11G). In contrast to the overground pupation of Salassa, the two Aglia species pupated underground. Larvae burrowed 2–3 cm into the soil, reinforced their burrow slightly with a small amount of silk and then pupated inside it (fig. 11G–I). This behavioural difference is facilitated by particular morphological features of the Aglia larvae and pupae: Aglia pupae have multiple tiny spines on their abdominal segments, which can help them to move more efficiently within the soil (figs 2G, 4G). Mature larvae of Aglia are densely covered with white tubercles, making their body rough and apt for digging, in contrast to the smooth and soft larvae of Salassa (figs 3G, 11G).
Loss of cocooning behaviour in Saturniidae
As a common behaviour of Saturniidae and superfamily Bombycoidea, cocooning ought to be an ancestral behaviour of the Saturniidae (Craig, Reference Craig2003). As such, cocoon morphology and silk composition are highly variable between genera and species (Lampe, Reference Lampe2010) (fig. 12A–I).
Figure 12. Cocoons of several saturniid genera. (A) cocoon of Actias chrisbrechlinae. (B) cocoon of Saturnia pyretorum. (C) cocoon of Saturnia lesoudieri. (D) cocoon of Samia canningi. (E) cocoon of Attacus atlas. (F) cocoon of Cricula trifenestrata. (G) cocoon of Rhodinia fugax. (H) cocoon of Antheraea yamamai. (I) cocoon of Loepa diffundata.
Most of the non-cocooning saturniid genera occur in Africa (Lampe, Reference Lampe2010) and of the 10 African genera with mitogenome sequences in GenBank, six are non-cocooning (fig. 10). In Asia, however, only Aglia and Salassa show non-cocooning behaviour, as we confirm in the present study (fig. 11).
Given that not all Saturniidae show cocooning behaviour and that cocooning is supposed to be an ancestral feature, one may wonder whether the loss of cocooning behaviour evolved once or several times in the Saturniidae. According to Regier et al. (Reference Regier, Grant, Mitter, Cook, Peigler and Rougerie2008) Salassinae and Agliinae are unrelated, implying that cocooning behaviour was lost at least twice within the Saturniidae. Moreover, a third loss of cocooning behaviour may be associated with the Bunaeini. However, the present analysis suggests that the three non-cocooning taxa form a clade with the Eochroini and Micragonini (fig. 10). This implies that cocooning behaviour may have been lost only once in the common ancestor of the clade comprising Salassa + Aglia + Bunaeini + Eochroini + Micragonini, and was secondarily regained in the Micragonini. Yet, the taxon sampling in this study is still far too limited to present this alternative suggestion as well-corroborated and definitive. Hence a more comprehensive phylogenetic analysis is needed to resolve this issue and to further explore the puzzling phylogenetic placements of Neoris, Cricula, Salassa and Aglia proposed by the present work.
Acknowledgements
We would like to express our sincere gratitude to Mr. Zhenbang Xu, Shiyu Tong, Haolin Gan and Yangyang Gou for providing eggs and specimens for this study, Mr. Da Pan and Zhiteng Chen for their assistance in phylogenetic analysis.
Financial support
This work is funded by the National Natural Science Foundation of China (31750002, 32070475) and the Priority Academic Program Development of Jiangsu Higher Education Institutions (PAPD).
Competing interests
None.
