Flight Muscle and Wing Mechanical Properties are Involved in Flightlessness of the Domestic Silkmoth, Bombyx mori

Flight loss has occurred in many winged insect taxa. The flightless silkmoth Bombyx mori, is domesticated from the wild silkmoth, Bombyx mandarina, which can fly. In this paper, we studied morphological characteristics attributed to flightlessness in silkmoths. Three domestic flightless B. mori strains and one B. mandarina population were used to compare morphological components of the flight apparatus, including wing characteristics (shape, forewing area, loading, and stiffness), flight muscle (weight, ratio, and microscopic detail) and body mass. Compared with B. mandarina, B. mori strains have a larger body, greater wing loading, more flexible wings and a lower flight muscle ratio. The arrangement in microscopy of dorsal longitudinal flight muscles (DLFMs) of B. mori was irregular. Comparative analysis of the sexes suggests that degeneration of flight muscles and reduction of wing mechanical properties (stiffness) are associated with silkmoth flightlessness. The findings provide important clues for further research of the molecular mechanisms of B. mori flight loss.


Introduction
Insects occur worldwide and their distributions have been shaped by their ability to fly. Insect flight may have evolved more than 400 million years ago [1]. Flight plays a crucial role in mating, reproduction, finding food, and escaping from predation. However, many winged insects have secondarily become flightless during their evolution [2,3]. In 1854, Wollaston documented that 200 of 550 beetles in the Madeiran archipelago had lost their ability to fly [4,5]. Some orders, such as Grylloblattodea and Siphonaptera, are entirely flightless [6]. In walking sticks, grasshoppers, earwigs, caddisflies, and scorpionflies, flightlessness is common [3,6]. Roff estimated that 5% of pterygotes are flightless [6]. In Lepidoptera species, the flight ability of migratory and non-migratory populations of the monarch butterfly is different [7].
Wings and flight muscles are crucial in insect flight. Flight muscle is the power engine of flying insects, and wings generate the aerodynamic forces required for flight. In many flightless insect species, the flight apparatus has been altered. The flight muscle of flightless grasshopper, Barytettix psolus is reduced compared to locust, Schistocerca gregaria, which capable of flight. The hemithorax of Schistocerca gregaria is filled with large, heavily tracheolated muscles, while it is almost empty in flightless Barytettix psolus [8]. In winter moth, Nyssiodes lefuarius, the flight muscles of flightless   Wings were removed using scissors and photographed with a digital camera. Wing area and wing shape were measured using ImageJ 1.47v software. Since it is likely that forewings play a major role in flight of moths and butterflies in generating aerodynamic forces [7,28], and the hind wings are mainly used to maintain balance, we quantified forewing characteristics only. Wing loading (mg/mm 2 ) was defined as the ratio: body mass (mg)/forewing area (mm 2 ). Wing shape was estimated with the parameter aspect ratio in ImageJ software according to the major and minor axes (major axes/minor axes).
Wing mechanical properties affect the wing deformation and aerodynamic force production of insects. Wing stiffness is a major characteristic of mechanical properties, and it is usually expressed in terms of storage (elastic) modulus (E') [29]. Storage modulus is a dynamic mechanical analysis parameter that evaluates the recoverable deformation energy of materials. It is also known as elastic modulus (in dynamic mechanical analysis). The E' of forewings was measured using Dynamic Mechanical Analyzer (DMA-Q800, TA Instruments, USA). Forewings were removed just before the test to ensure the samples were fresh [30]. We trimmed the same part of wings into 1 cm × 0.5 cm rectangles and fixed the regular film slice between two grips of the instrument. A frequency range of 1-100 Hz was used to determine the storage (elastic) modulus E', in a multi-frequency-strain module at 0.1% strain [30,31]. A lower or higher E' indicates relatively more flexible or stiff wings.

Microscopy of Flight Muscle
The thorax of silkmoths was dissected and fully covered in optimum cutting temperature (OCT) compound (SAKURA Tissue-Tek O.C.T. Compound, Torrance, CA, USA), then snap frozen in liquid nitrogen (−196 • C). The embedded thorax was sectioned to 15 um using freezing microtome (HM525 NX, Thermo Scientific, Waltham, MA, USA), and stained with hematoxylin and eosin. Images were taken at 10× magnification with a microscope (DP80, Olympus, Tokyo, Japan). Muscle fiber areas were measured using ImageJ 1.47v software. We counted the total number of myofibers of dorsal longitudinal flight muscles (DLFMs) artificially in drawing software of the windows10 system.

Statistical Analysis
As the four silkworm populations have different genetic backgrounds and distribute in distinct branches of the phylogenetic tree [23], we treat them as four independent populations. We used Analysis of Variance (ANOVA) method (IBM SPSS v. 22) to analyze the effects of domestication on the flight apparatus (body mass, flight muscle mass, flight muscle ratio, wing shape, wing area, wing loading). We firstly used a multi-way ANOVA model (morphology = population + sex + population × sex) to evaluate the effects of population and sex on each of the morphological measures. Further, one-way ANOVA tests were employed to assess population effects on the morphology in males and females, respectively. Tukey HSD post hoc tests were performed for the comparisons of mean values of populations (the level of significance was 0.05). To validate whether a morphology affects the flight of silkmoth, we mainly considered the three pairs comparisons between flying B. mandarina and each of the three flightless B. mori. Morphological measures that showed significant differences in all of these three pairs of flight-flightlessness comparisons were considered related to the flight of silkmoth.
Both male and female B. mori lost their flight ability, which shows that gender did not play a crucial role in the loss of silkmoth flight ability (it does not rule out an influence). Therefore, without considering the influence of gender, we compared male B. mori and female B. mandarina (the latter is usually heavier than the former) to validate whether the body mass and body-mass related morphologies (wing loading (body mass/forewing area) and flight muscle ratio (thorax mass/body mass)) are essential for the flightlessness of silkmoths.
For the area of DLFMs, the mean area, and the total number of myofibers of DLFMs, we used a Student's t-test to determine whether differences were significant between domestic and wild silkmoth (the level of significance was 0.05).

Population Divergence of Morphological Traits
Domestication shaped body type and wing morphology of B. mori (Figure 1). Sex and population have significant effects and interactions on measurements of body mass, flight muscle mass, flight muscle ratio, wing shape, wing area and wing loading ( Table 2). Comparative results of mean values (Tukey HSD post hoc tests) suggested that body mass, wing loading and flight muscle ratio of B. mandarina were significantly different with each of the three B. mori populations (except for body mass of J106 males that were similar to B. mandarina males) (Table 3; Figure 2). At least one of the three B. mori populations showed similarities with B. mandarina in the measurements of aspect ratio (wing shape), flight muscle mass and forewing area (Table 3; Figure 3).  Table 1 and plotted in Figures 2 and 3. Bars = 1 cm.  Table 1 and plotted in Figures 2 and 3. Bars = 1 cm.   Table 1 and plotted in Figures Table 3. The significance level is 0.05 (*p < 0.05, **p < 0.01, ***p < 0.001). The body weight (mg) of B. mori was larger than that of B. mandarina in males (A) and females (B). The wing loading of B. mori was significantly larger than that of B. mandarina in males (C) and females (D). The flight muscle ratio of B. mori was significantly lower than that of B. mandarina in (E) males and (F) females.
Tukey HSD post hoc tests are shown in Table 3. The significance level is 0.05 (*p < 0.05, **p < 0.01, ***p < 0.001). The body weight (mg) of B. mori was larger than that of B. mandarina in males (A) and females (B). The wing loading of B. mori was significantly larger than that of B. mandarina in males (C) and females (D). The flight muscle ratio of B. mori was significantly lower than that of B. mandarina in (E) males and (F) females.    (Table 4; Figure 4A). The wing loading of B. mandarina females was larger than J106 and similar to 872 (Table 4; Figure 4B). These results suggested that body weight and wing loading are not key factors of silkmoth flightlessness.
The measurements storage modulus (E') were more variable in the high-frequency range, but the E' of all three B. mori groups floated in the same zone in both sexes ( Figure 5A,B). The E' of B. mandarina was always higher than that of B. mori ( Figure 5A,B), meaning that the wings of B. mandarina are stiffer and better able to resist deformation. The stiffer wings of B. mandarina maybe have the potential to generate greater lift forces than the softer wings of B. mori.   The measurements storage modulus (E') were more variable in the high-frequency range, but the E' of all three B. mori groups floated in the same zone in both sexes ( Figure 5A, B). The E' of B. mandarina was always higher than that of B. mori ( Figure 5A, B), meaning that the wings of B. mandarina are stiffer and better able to resist deformation. The stiffer wings of B. mandarina maybe have the potential to generate greater lift forces than the softer wings of B. mori.   Table 4. The significance level is 0.05 (*p < 0.05, **p < 0.01, ***p < 0.001). (A) Body weight (mg) of B. mori males was significantly lower than B. mandarina females. (B) The wing loading of B. mandarina females was significantly larger than that of J106 males and similar to that of 872 males. The wing loading of Dazao was significantly larger than that of B. mandarina females. (C) The flight muscle ratio of B. mori males was significantly larger than that of the B. mandarina females. The measurements storage modulus (E') were more variable in the high-frequency range, but the E' of all three B. mori groups floated in the same zone in both sexes ( Figure 5A, B). The E' of B. mandarina was always higher than that of B. mori ( Figure 5A, B), meaning that the wings of B. mandarina are stiffer and better able to resist deformation. The stiffer wings of B. mandarina maybe have the potential to generate greater lift forces than the softer wings of B. mori.
To illustrate whether flight muscle ratio is a key factor in determining silkmoth flightlessness, we compared the flight muscle ratio of B. mandarina females to that of B. mori males. The results showed that the flight muscle ratio of B. mandarina females was significantly lower than that of B. mori males ( Table 4; Figure 4C), which implied that the ratio is not essential in B. mori flight loss.
We used microscopy to examine the most prominent muscle class, the dorsal longitudinal flight muscles (DLFMs). In the adult thorax of B. mandarina, the DLFMs were composed of two sets of muscle fibers, each set of fibers was separated into six groups of fibers (fascicles) by perimysium ( Figure 6A). In B. mori, no clearly separated group of fibers was observed in either of the two sets of muscle fibers ( Figure 6B). Some perimysia of B. mori DLFMs seem to be absent. The area of DLFMs, the mean area and the total number of myofibers of DLFMs were lightly reduced in B. mori, but the wild and domestic silkmoths were otherwise similar ( Figure 6C. Student's t-test, p area = 0.127, p mean area = 0.370, p number = 0.092). This observation suggested that the arrangement of DLFMs of B. mori is irregular. This case is similar to previous observations in flightless hawkmoths [9] and indicates that the irregular DLFMs may have weakened the function of the flight muscle and contributed to B. mori flightlessness.

Discussion
The evolution of flight has contributed to insect diversification [32]. Flight ability enables insects to disperse, forage and avoid predation. Nevertheless, flightless insects occur in nearly all of the winged orders [5]. Research on flight loss promotes understanding of species adaptation and evolution.

Discussion
The evolution of flight has contributed to insect diversification [32]. Flight ability enables insects to disperse, forage and avoid predation. Nevertheless, flightless insects occur in nearly all of the winged orders [5]. Research on flight loss promotes understanding of species adaptation and evolution.
The morphological characteristics differed between B. mori and B. mandarina, as well as between males and females. Sexual dimorphism is common in insects. For instance, body shape differs between males and females in Drosophila [33]. In butterflies and moths, sexual dimorphism occurs often, leading to different body color, body size, body composition (e.g., relative thorax size), wing size and wing shape in males and females [34,35]. Gender also has a significant effect on the morphological characteristics of silkmoths, but flightlessness is not a dimorphic character in B. mori since both sexes are flightless.
Our results show that the interaction between population and sex affected flight-related morphologies. The origins (genetic backgrounds) of the B. mori used in the study were different, suggesting that they might have experienced different selective pressure during their domestication. For example, the 872 strain is a commercial race and fecundity might be a preferred direction of domestication. The J106 strain is a landrace and easy breeding is more important for them. In these cases, both artificial selection and sexual selection played crucial roles. Throughout the life cycle, females typically allocate more energy for reproduction and males usually allocate more energy for fighting for mating opportunities [34]. In this way, their morphologies would be affected differently. We believe that the effects of sexual and artificial selection on flight-related morphologies is the reason for the interaction between sex and the population. Flight loss of B. mori probably occurred under artificial selective pressure rather than sexual selective pressure. Thus, we focused on the morphological differences between B. mandarina and B. mori.
The morphological features of body type, wings and flight muscles differ between B. mori and B. mandarina. These include body mass, wing loading, wing mechanical properties, and flight muscle ratio. During domestication, silkworms were selected for greater mass to increase silk production [26]. With increased body mass, the wing loading (body mass/forewing area) of B. mori increased. Research on birds and Lepidoptera demonstrated a negative correlation between flight ability and wing loading [20,22]. In butterfly Pararge aegeria, acceleration capacity was positively correlated with wing loading and body mass [36]. A high or low wing loading and body mass does not always result in poor flight performance. Flightless domestic silkmoths have larger body mass and wing loading than flying wild silkmoths. However, the flying female B. mandarina had a larger body mass and greater wing loading than flightless B. mori males. This suggests that wing loading and body mass are not the key factors of silkmoth flight loss.
Flapping wings generate aerodynamic forces in insect flight. The wing shape of flying insects changes considerably in spanwise and chordwise directions [37]. The motion of wings and their three-dimensional shape have a significant effect on lift forces [38][39][40]. In a robotic insect experiment, the aerodynamic forces decreased monotonically as the flexibility of wings increased [19]. In contrast, flexible wings produced larger aerodynamic force than rigid wings in hawkmoths and bumblebees [15,18]. These studies showed that flexibility or stiffness of insect wings does not always indicate enhancement or reduction of aerodynamic forces. Rajabi and Gorb believe that a balance between flexibility and stiffness is needed [41]. The flexibility of wings should be kept in a suitable range. Wings that are too soft cannot resistant aerodynamic forces and excessively rigid wings cannot form dimensional shape. The storage modulus (E') of domestic B. mori was lower than B. mandarina. E' usually reflects the stiffness of materials [29].
The lower E' of B. mori wings indicated lower stiffness (more flexibility). We suppose that the reduced flexibility of B. mori wings compromises the balance between stiffness and flexibility and reduced the capacity to generate lift. This change might have affected the flight ability of B. mori and was involved in silkmoth flightlessness.
Wingbeats require considerable energy [42][43][44], which is provided by the flight muscles. Dysfunction of flight muscles can lead to weakened flight ability. For example, the degeneration of flight muscles in Drosophila leads to flightlessness or reduced flight ability [45][46][47]. We found that B. mori had a reduced proportion of flight muscles. However, in a comparison between female B. mandarina and male B. mori, female B. mandarina had a significantly lower flight muscle ratio. This suggests that the lower flight muscle ratio of B. mori was not responsible for flightlessness. The perimysia of B. mori flight muscle seems to be absent, which implies a degeneration of this powerful engine. The structure of the perimysium provides an important mechanical function in skeletal muscles [48], such as the transmission of forces, passive elasticity, and stiffness of muscles [48][49][50]. In Nyssiodes lefuarius (Lepidoptera: Geometridae), the dorsal longitudinal muscles of flightless females have no clearly separated bundles in contrast to flying males [9]. This situation is similar to the DLFMs of the domestic silkmoth. The degeneration of silkmoth DLFMs might have affected the construction of flight muscles and led to an insufficient energy supply.
To increase silk production, larger silkworms have been selected for breeding. Based on our understanding of silkworm domestication, we believed that excess body weight is the major reason for silkworm flightlessness. However, the view is purely anecdotal and might be misleading further research on silkmoth flightlessness. By measuring and comparing the morphology of the flight apparatus of silkmoths, we demonstrated that body weight, flight muscle ratio, wing loading, and wing mechanical properties were different between wild and domestic silkmoth. They might affect silkmoth flight ability. However, comparisons between flying females and flightless males demonstrated that the body weight, wing loading and flight muscle ratio were not attributed to silkmoth flightlessness. Then, we speculated that flight muscle structure and wing mechanical properties (stiffness) were key aspects in flight loss.
To date, most studies of the relationship between morphology and the loss of flight have been conducted on insects that have undergone natural selection (in wild field). This study expanded knowledge of natural selection examples to domestic insects (undergoing artificial selection). The findings provide important clues for further research on the molecular mechanisms of B. mori flight loss. Morphological data have limitations for explaining complex issues related to loss of flight in species. Additional studies on physiology and molecular biology would increase our understanding of energy metabolism and the molecular mechanism of silkmoth flightlessness.

Conclusions
We measured and compared flight apparatus that could influence silkmoth flight ability and verified that flight muscle and wing mechanical properties (stiffness) are essential for silkmoth flightlessness. The measurements are useful for understanding silkmoth flight loss. The result offered a dependable direction for future research in the flight loss of the silkmoth. Despite the findings, further research should be conducted to determine whether the energy supply is sufficient. Genes involved in flight muscle development should be examined in the wild and domestic silkmoth (e.g., expression and nucleotide sequence of genes).