Kogure's pioneering work

Makita Kogure (Fig. 1) was a Japanese entomologist who graduated from Kyushu Imperial University and later served as the director of the Nagano-Prefecture Sericultural Experiment Station (NSES), where he conducted extensive studies on the silkworm PR. After leaving the station, he became a professor at the Tokyo University of Agriculture and Technology.

Fig. 1.

Portrait of Dr. Professor Makita Kogure (1899–1964).

Table 1.

Abbreviations used in this review.

In 1927 Nagase and Kobayashi, while working as senior technical researchers at the NSES, reported that the light conditions during egg incubation had a significant effect on the diapause incidence of the subsequent progeny. They accidentally stumbled upon this phenomenon. At that time, the production of non-diapause eggs in bivoltine races was generally carried out under continuous dark (DD) conditions at a low temperature (15°C). However, they observed unexpected diapause production when the eggs were inadvertently kept under continuous light (LL) conditions (Nagase and Kobayashi, 1927). Kogure took over their study upon his arrival at the NSES and conducted an extensive investigation of the silkworm PR from the perspective of environmental biology. He examined the effects of day length, photo-susceptibility, effective light intensity and wavelength, and temperature influence. Kogure primarily used two bivoltine races, Shohaku (a Japanese race) and Showa (a Cighinese race). He studied the diapause incidence using approximately 65,000 silkworms that had been subjected to various light and temperature conditions during their developmental stages. In 1933, Kogure presented a historic paper on the silkworm PR, which was also his doctoral thesis. This paper is widely cited in textbooks on insect photoperiodism as one of the pioneering works. In the paper, he cited the work of Marchovitch (1924) and used the term “photoperiodism”, which may be the first appearance of this term in zoological literature in Japan. He obtained SD photoperiodic response curves (PRCs) when the eggs were incubated at 20 and 15°C (Fig. 2). This was the first demonstration of PRC in the papers on insect photoperiodism.

Kogure investigated the PR of not only the egg stage but also the larval stage. He demonstrated that PR was observed during the larval period when the incubation temperature was below 20°C. LL conditions during the early larval stage, from the 1st to 3rd instar, induced significantly more diapause producers than DD conditions, as observed during the egg stage. In addition, during the late larval stage (4th–5th instar), LL conditions induced significantly more non-diapause producers than DD conditions. This indicates a change of PR from SD to LD response, both of which produce non-diapause producers under their respective conditions (Saunders, 2002), in the middle of the larval period. Kobayashi et al. (1986a, b; 1990) confirmed this PR change from SD to LD response during the larval stage using the tropical race Ringetsu. The authors discussed the directional change of PR based on transfer experiments of light-dark cycles. In Heliothis zea, pupal diapause is induced only by a reverse transfer from LD during the egg stage to SD during the larval stage (Adkisson and Roach, 1971). In the race Ringetsu, SD condition throughout the larval stage induced diapause, and transfer from LD to SD was not required to produce diapause induction. This PR change from SD to LD response differed depending on the silkworm race used. A bivoltine race, Daizo (p50), showed SD response during the egg stage (Egi et al., 2014), but showed LD response throughout the whole larval period, including the 1st instar (Shimizu and Hasegawa, 1988). In this race, exposure to three or four SD cycles during the 5th instar was enough to elicit diapause induction with 100% incidence (Hasegawa and Shimizu, 1987).

Fig. 2.

Photoperiodic response curves (PRCs) for embryonic diapause induction in the silkworm bivoltine race, Shohaku. Three batches of eggs were incubated at 24, 20, and 15°C under different photoperiods during the egg stage. All batches of the larvae were reared under the same condition. Drawn from the data of Kogure (1933).

Temperature influence

Kogure and his coworker (Kogure and Kobayashi, 1928; Kogure, 1930, 1933) found that ambient temperature remarkably influences the PR in the race Shohaku. Increasing incubation temperatures shifted PRCs of the SD response upward (Fig. 2). The critical day length did not change depending on temperatures. At 24°C incubation, no clear PR was observed with high diapause incidence. The loss of PR by high-temperature incubation was observed in other bivoltine races (Cui et al., 2021). Under LL almost 100% diapause was observed at temperatures examined except 15°C, and under DD the diapause incidence depended on the temperature with 100% diapause at high temperatures (Fig. 3). Kogure (1933) reported that the threshold value of light intensity evoking diapause induction under LL condition depended on the incubation temperature (see the section PHOTORECEPTION MECHANISM). Furthermore, he found also that high temperature during young larval stages, 1st to 3rd instar, induced significantly more diapause than low temperature, as observed during the egg stage. However, during the 4th and 5th instar of the late larval stage, high temperature induced more non-diapause than low temperature, indicating that the silkworms changed their temperature response for the diapause induction during the middle larval stage. Kobayashi et al. (1986b) confirmed this temperature response change using the Ringetsu race with clearer data.

Fig. 3.

Effect of incubation temperature on the diapause incidence under LL and DD condition during egg stage. The bivoltine race, Shohaku, was used in the experiments. All batches of the larvae were reared under the same condition. Drawn from the data of Kogure and Kobayashi (1928).

Nutritional influence

When bivoltine races are incubated at temperatures above 25°C during the egg stage and reared on mulberry leaves, they usually become diapause producers regardless of the photoperiod given during the larval period (Kogure, 1933; Takamiya, 1974; Kinoshita et al., 1982). However, when they were provided with artificial diets in place of mulberry leaves, they exhibited robust LD response throughout the entire larval period, irrespective of the high-temperature incubation (Sumimoto, 1974; Takamiya, 1974; Shimizu, 1982a, Tsuchida and Yoshitake, 1983a). The PRCs obtained by many authors have shown LD response but differed in the shape and critical day length depending on the race and artificial diets they used. The insertion of SD regime into even early larval instars (1st–3rd), produced a significant incidence of diapause (Takamiya, 1974). The replacement of an artificial diet with mulberry leaves during a single instar, along with the insertion of SD regime, induced diapause (Takamiya, 1974; Tsuchida and Yoshitake, 1983a). The replacement of mulberry feed during the 4th instar showed the most pronounced effect on diapause production. Simultaneously replacing mulberry feed and inserting SD regime into a single instar had a synergistic effect on diapause production (Takamiya, 1974). These observations suggest that both dietary chemical influence and photoperiodic signal might bring about the same biochemical impact, thereby integrating diapause induction.

There are many metabolic and physiological differences between silkworms reared on artificial diets and those reared on mulberry leaves (Dong et al., 2017). Many authors (Kato, 1974; Shimizu, 1982a; Tsuchida and Yoshitake, 1983) discussed the possible involvement of chemical compounds of mulberry leaves in this phenomenon. As this was observed, regardless of whether the artificial diets contained mulberry leaf powders or not, the involved compounds are presumed to be easily decomposed by the preparation procedure of the diets, such as autoclaving or heat treatment. In Pieris brassicae, Dumortier and Brunnarius (1989) discovered that when artificial diets instead of their host plant were provided to the insects, the photoperiodic clock shifted from a circadian to an hourglass mode. They proposed the potential involvement of dietary substances in their observation.

Bull and Adkisson (1960, 1962) reported that the dietary fat content remarkably affected the diapause incidence of the pink bollworm, Pectiophora gossypiella. The author of this review found also that adding 5% of soybean oil (Nacalai Tesque Inc.), which contained abundant polyunsaturated fatty acids (PUFA) of linolenic and linoleic (Ito and Nakasone, 1969), to an artificial diet significantly increased the diapause incidence in a bivoltine race reared under a LD (L20: D4) condition during the larval stages (unpublished observation), implying that PUFA content in the diets might affect the diapause induction. Silkworm diapause eggs just after oviposition had a fatty acid pattern characteristic of high proportions of PUFAs in the phospholipids compared with non-diapause eggs (Miura and Shimizu, 1989; Shimizu, 1992). Increasing PUFA in fatty acid compositions is considered to be an adaptive component of winter diapause to maintain cell membrane fluidity (Cossins and Bowler, 1987; Khani et al., 2007; Reynolds et al., 2012; Vukašinović et al., 2013). Generally, rearing insects at low temperatures and under SD photoperiods of diapause-inducing conditions increases the ratio of PUFA in lipids (Harwood and Takata, 1965; Valder et al., 1969; Nakasone and Horie, 1978; Behren, 1984; Baldus and Mutchmor, 1988; Hodkova et al., 2002; Hemmati et al., 2017). Some insect clock genes, such as per and tim, have been suggested to be involved in lipid accumulation (Omura et al., 2016; Zhu et al., 2019). In Drosophila melanogaster, composition change of PUFAs in the neurons expressing TRPA1 of a molecular thermosensor, altered the temperature preference behavior by modulating neuronal activity (Zheng, 2013; Suito et al., 2020).

Kogure's pioneering work showed that the embryonic diapause in the silkworms is determined by light and temperature experienced at each developmental stage. Textbooks describing insect photoperiodism have identified the silkworm as a SD response insect with photosensitivity during the egg and young larval stages (Danilevsky, 1965; Beck, 1980; Danks, 1987; Saunders et al., 2002; Denlinger, 2022). However, Kogure (1933) demonstrated not only the SD response during the early larval stages but also the LD response during the later larval stages. Few insects change the PR during the same developmental stage, though there are reports of PR changes during different periods (Saunders, et al., 2002). Kogure (1933) also found the change of low and high temperature effect on diapause incidence during the larval period, which was confirmed by Kobayashi et al. (1986b).

These changes appear to be reasonably adaptable to adjust the life cycle of bivoltine races to the seasonal transition in the temperate zone. In bivoltine races, SD and low temperature (early spring information) during the egg and early larval stage, followed by LD and high temperature (early summer information) during the late larval stage, induce non-diapause progeny of the next generation. Furthermore, LD and high temperature (summer information) during the egg and early larval stage, followed by SD and low temperature (autumn information) during the late larval stage, induce diapause progeny of the second generation. The mechanism behind the PR change during the larval period remains unknown. Regrettably, research on silkworm PR at the population level has shown no progress since Kogure's initial work. Further investigation into the dominance and accumulation of PR at various developmental stages is required.

Light sensitivity and effective wavelength

Kogure (1933) studied the light sensitivity for diapause induction and found that the photosensitivity was observed just after blastokinesis during embryonic development. He observed that the threshold light intensity that produces LL response was between 0.04 and 0.06 lux at 20°C incubation and between 0.1 and 0.14 lux at 15°C at egg stage. Shimizu and Kato (1984) found that it was between 0.1 and 1.0 lux during the larval period when a bivoltine race was reared on an artificial diet. The threshold value for the induction of LD response was between 0.0084 and 0.039 µw/cm² with irradiation of monochromatic light of 525 nm under a photoperiod of L17:D7 (Nakajima et al., 1972).

Although strict action spectra recorded according to the proper procedure have not been obtained for the silkworm, several researchers have investigated the spectral sensitivity of PR. Kogure (1933) used colored glass plates and found that red and orange-yellow light had little or no effect on diapause incidence, whereas violet light showed a distinct effect compared with other light. Nakajima and his coworkers (1972, 1981) used modern optical equipment to investigate the effects of different wavelengths on diapause incidence. They found that wavelengths in the blue-green (411–547 nm) region were effective, with a maximum sensitivity peak at 527 nm. The sensitivity falls rapidly above 588 nm, with red light being virtually ineffective. Tsuchida and Yoshitake (1979) also showed that light wavelengths from 400 to 540 nm were effective for LD response, and those above 580 nm were ineffective. Although these reports did not include any strict action spectrum, such as that shown in Megoura viciae (Lees, 1981), the involvement of a receptor pigment with blue-green absorption properties has been considered.

Localization of photoreceptor

The photoreception pathway of light input for PR has been found to differ among insect species, without any phylogenetic relationship (Numata et al., 1997; Saunders, 2012). In the silkworm, it has been shown that the photoperiodic receptor resides in the larval head, rather than the stemmata, using black cover painting and local supplementary illuminations with chemiluminescent paste (Shimizu and Hasegawa, 1988). The possibility of a translucent clypeus acting as a window to admit light to the pupative cerebral photoreceptor in the 1st and 2nd instar larva has been discussed. Claret (1966) described a similar idea of a window function of the clypeus in a cabbage butterfly, Pieris brassicae.

Fig. 4.

In vitro photoperiodic programming of diapause induction in the silkworm. Pairs of Br-SG complexes with CC-CA were isolated from 5th instar-day 1 females destined to become non-diapause producers and cultured for 4 days under LD of L20:D4 or SD of L8:L16. Then, the complexes were transferred into 5th instar-day 5 females destined to become non-diapause producers. The moths receiving SD-exposed complexes produced significantly more diapause eggs than those receiving LD-exposed complexes (Hasegawa and Shimizu, 1987). This illustration was redrawn from Shimizu (2018) with permission of Japanese Society of Sericulture Science.

Bowen et al. (1984) demonstrated that excised Brretrocerebral complexes of the hornworm moth, Manduca sexta, can be photoperiodically programmed under in vitro culture conditions, suggesting that both the photoreceptor and clock system of diapause induction are Br-centered. Soon after Bowen's report, in vivo and in vitro culture experiments were conducted by Hasegawa and Shimizu (1987) using Br-subesophageal ganglion (SG) complexes with endocrine glands, corpora cardiaca (CC)-corpora allata (CA), in the silkworm (Fig. 4). The results demonstrated that photoperiodic induction of the silkworm can be programmed under in vivo and in vitro culture conditions, implying that the components of the photoreceptor, clock, and counter system are located in the Br-SG complex.

As mentioned below, the silkworm Br has been found to control the secretion of DH from SG (Fukuda, 1952; Matsutani and Sonobe, 1987; Shimizu et al., 1997). It has been suggested that extra-retinal photoperiodic receptors of insects are closely associated with the regulatory site of hormone secretion in the Br (Steel and Lees, 1977; Shiga and Numata, 2007), and therefore, it is likely that the photoreceptor of the PR resides in the silkworm Br, not in SG. In lepidopterans, many reports have shown Br-centered photoreception in PR (Williams et al., 1965; Claret, 1966; Segue and Veith, 1976; Bowen et al., 1984).

Vitamin A requirement

When silkworms were reared on a vitamin A-deprived artificial diet, they exhibited a loss of larval phototaxis and ERG (electroretinograph) response in both their stemmata and compound eyes (Shimizu et al., 1981). Furthermore, the rearing of larvae on the vitamin A-deprived diet resulted in the disappearance of PR in the larval stage. However, the introduction of β-carotene (provitamin A) or vitamin A to the diet restored the PR (Shimizu and Kato, 1984; Hasegawa and Shimizu, 1988). The impacts of carotenoid and vitamin A deficiency were only noticeable under a low light intensity of 1 lux, as opposed to higher intensities exceeding 5 lux (Shimizu, 1982a; Shimizu and Kato, 1984). This correlation between light intensity and the effects of vitamin A deficiency implies that this phenomenon is associated with the photoreception process itself, rather than downstream of diapause induction.

Various arthropods, including three mites, Tetranchus urticae (Veerman and Hell, 1978; Veerman, 1980; Bosse and Veerman, 1996), Ambyseius potentiiae (Veerman et al., 1983) and A. cucumeris (Overmeer et al., 1989), and three insects, B. mori (Shimizu and, Kato, 1984), Apanteles glomeratus (Veerman et al., 1985), and P. brasscae (Claret and Volkoff, 1992), have demonstrated dependencies on carotenoids or vitamin A in their PRs. Additionally, liams et al. (2019) demonstrated that retinal and its derivatives are essential for photoreception in the monarch butterfly, Danaus plexippus, by knocking out NINAB1, which produces the enzyme that converts carotenoids into retinal. Hasegawa and Shimizu (1988) found that the silkworm Br of 5th instar larvae contained cis- and all-trans 3-OH-retinal (0.01 pmol per brain in combination) and all-trans-retinal (0.03 pmol per brain), which are chromophores of the insect visual pigments (Seki and Vogt, 1998). Furthermore, cis- and all-trans 3-OH-retinal were found in the Br and the compound eyes, whereas only all-trans-retinal was detected in the Br. These pieces of evidence suggest that 3-OH-retinal in the larval brain might be formed from all-trans retinal and function as the chromophore of opsin-like pigment (Gartner and Towner, 1995; Seki et al., 1998).

Boceropsin: A candidate of photoperiodic receptor pigment

Shimizu et al. (2001) hypothesized that an opsin-like pigment expressed in the silkworm Br could be a photoperiodic receptor and attempted to clone the cDNA encoding the opsin protein from the organ. They successfully cloned a cDNA (Genbank Accession No. AB064496) encoding an opsin protein consisting of 381 amino acids, which was conserved in insect visual pigments. This protein was predicted to be a G-protein-coupled opsin with seven transmembrane segments, and it had the 327th lysine (K)-residue that presumably bound the chromophore by a Schiff base (Gartner and Towner, 1995). Southern blotting experiments indicated that this gene is present in a single copy, and it was designated Boceropsin (Bombyx cerebral opsin). Miao et al. (2005) showed that this Boceropsin gene is located on the 15th chromosome. This was the first finding of an opsin expressed in the central nervous system of an insect. Molecular phylogenetic analysis revealed that this novel opsin belonged to the group of long-wavelength (green) photoreceptors. A database search showed that it had the highest degree of homology (88%) with a long-wavelength receptor (Bomopsin 1: BAB12276) expressed in the silkworm compound eyes. It also had a high degree of homology (84%) with Manopin 1 (L78080) from the sphingid moth M. sexta, which has an absorption peak at 520 nm (Chase et al., 1997; Briscoe, 2008).

RT-PCR analysis revealed that Boceropsin was expressed in the larval Br, but not in SG, first thoracic ganglia, or compound eye. Whole-mount immunohistochemistry was conducted using mouse antiserum to locate the cells that produce Boceropsin in the Br (Shimizu et al., 2001). The results showed that some cells (15–20 mm in diameter) in the Br exhibited clear immunoreactivity. Specifically, two types of bilateral Boceropsin immunoreactive cells were found in the dorsal-anterior protocerebrum (DAP) cells and ventral-anterior protocerebrum cells (VAP). The DAP cells had large somata and exhibited more intense immunostaining compared with other immunoreactive cells. Boceropsin immunoreactivity was also observed bilaterally in dorsal-posterior protocerebrum cells and lateral-posterior tritocerebrum cells, but not in SG or thoracic ganglia. A detailed histochemical observation by confocal laser microscopy confirmed these results (Shimizu, 2018).

Yu et al. (2008) conducted a systematic analysis of microRNAs (miRNAs) within the silkworm genome sequence. They successfully identified and cloned 35 previously undiscovered miRNAs. Among these, Bmo-miR-92 was identified as an ortholog of Drosophila dme-miR-92a expressed in the fly's Br primordium. Furthermore, they found that Boceropsin is a potential target of Bmo-miR-92, and noted a relatively low expression of bmo-miR-92 during the trachea-appearing embryonic stage. The authors speculated that this particular stage might hold essential implications for the functioning of Boceropsin. Notably, Kogure (1933) previously observed that the onset of light-sensitivity related to diapause induction occurs during the trachea differentiation stage following blastokinesis. This parallel is worth considering regarding Boceropsin's function.

In addition to Boceropsin, the author of this review cloned the cDNA encoding the long-wavelength receptor opsin in the larval Br of a wild moth, Antheraea pernyi, and cabbage butterfly, P. rapae. The deduced amino acid sequence of the new opsin of A. pernyi was 381 (BAC66474), and that of P. rapae was 382 (BAD06459), amono acids long, exhibiting homology with Boceropsin of 87% and 76%, respectively. Since the discovery of Boceropsin, the expression of some opsins has been reported in other insect Brs. Velarde et al. (2005) reported pteropsin, which is expressed in the lateral and dorsal central portion of the anterior cerebral region of Apis mellifera. This non-visual opsin of honeybees belongs to the Opn3 family, including mosquito Opn3, but its function remains unknown (Terakita and Nagata, 2014). Craig Montell and his coworkers found that three rhodopsins, Rh5, Rh6, and Rh7, were expressed in the Br neurons of D. melanogaster (Sokabe et al., 2016; Ni et al., 2017). Drosophila Rh7 with a broad spectrum with a longer wavelength tail reaching 500 nm was found to function as a circadian photoreceptor, as well as CRY1 (Ni et al., 2017; Sakai et al., 2017; Sokabe et al., 2017). In an aphid, Acyrthosiphon pisum, Collantes-Alegre et al. (2018) found that three species of opsin (Ap-Anthropsn, Ap-SWO4, Ap-C-Ops) were expressed in several cells of the protocerebrum of the insect. Among these cerebral opsins, Drosophila RH6 showed a relatively high homology (58%) with Boceropsin in terms of amino acid sequence, but the other opsins exhibited a lower homology of about 30%.

Studying the photoreceptor in PR of organisms appears to be an effective strategy to uncover the enigmatic mechanism of the photoperiodism. Several authors (Goto and Numata, 2009; Saunders, 2012, 2020; Barbera et al., 2021; Goto, 2022) have suggested that two types of photopigments may be involved in insect photoperiodic reception based on the external coincidence model (ECM) proposed by Pittendrigh and Minis (1964). One is Cryptochrome 1 (CRY1), a blue-light photopigment of the circadian system, and the other is the opsin pigment involved in the photoinductive process of diapause or non-diapause determination. Regarding Bombyx cry1, it has only been reported that KO of Bombyx cry1 does not affect diapause induction at 25°C incubation (Homma et al., 2022), while some pieces of evidence have suggested the involvement of opsin pigment in the photoreception. Boceropsin, expressed in the defined cells of Br, has been proposed as a likely candidate of photoperiodic receptor pigment (Shimizu et al., 2001).

In the silkworm, genome analysis identified six homo-logs of opsin (Velarde et al., 2005): BmUVop (UV-sensitive opsin of Drosophila Rh3 ortholog), BmBLop (blue-sensitive opsin, Rh5), BmLop1 (long wavelength1 opsin, Rh6), BmLop2 (long wavelength2 opsin, Rh6), BmUnop (broadband opsin, Rh7), and Bombyx pteropsin (a homolog of Opn3 group opsin). Among these, BmLop1 corresponds to Bomopsin1, a long-wavelength receptor of the compound eyes (Shimizu et al., 1998), and BmLop2 corresponds to Boceropsin. Based on the insight into the location of other insects' cerebral opsins, BmBLop, BmUnop, and Bombyx pteropsin are considered to be potent candidates for cerebral photoperiodic pigments in addition to Boceropsin, although the expression of these opsins has not yet been demonstrated in the silkworm. In the cricket Modicogryllus siamensis, three opsins that are expressed in the compound eyes were suggested to play roles in photoperiodic regulation (Tamaki et al., 2013).

It is generally accepted that the dorsal protocerebrum of the insect Br is an essential region that regulates endocrine control for diapause induction (Raabe, 1989; Ichikawa, 1991; Shiga and Numata, 2007; Shiga, 2023). The protocerebrum of the silkworm Br was found to control the release of DH from the CC-CA complex (Matutani and Sonobe, 1987; Shimizu et al., 1997). In the anterior region of protocerebrum, there are neurons expressing corazonin (Crz) (Tsuchiya et al., 2021) and clock-related proteins (Sehadová et al., 2004), which are suggested to be important molecules involved in the diapause induction, as described below. Reliable classical experiments suggest that the photoperiodic receptors are restricted to the dorsal protocerebrum region of A. pernyi (Williams, 1969) and the dorsal anterior region of M. viciae (Lees, 1964; Steel and Lees, 1977; Gao et al., 1997). Consequently, it is reasonable to speculate that the DAP cells expressing Boceropsin are the most likely photoreceptor cells involved in the silkworm photoperiodism. Further investigations, including neurophysiological and genetic experiments, are required to verify the role of Boceropsin.

Identification, production, and secretion of DH

Alongside upstream studies of the silkworm PR, Japanese researchers conducted systematic investigations of the hormonal effector system in the PR. Kinsaku Hasegawa (1951) and Soichi Fukuda (1951a, b) presented papers in the same volume of the Proceedings of the Japan Academy, which demonstrated the release of DH (factor) from SG into the hemolymph. Although they both contended for the priority of this discovery, Hasegawa attempted to identify the chemical structure of DH, and Fukuda aimed to reveal the control mechanism of DH secretion (Yaginuma, 2015). Fukuda was an insect endocrinologist who discovered the function of the prothoracic gland by studying the silkworm (Fukuda, 1940a, b; Ohnishi and Ishizaki, 1990).

With great effort, using a large number of silkworm pupae, the structure of DH has been identified as an amidated peptide of 24 amino acid residues, belonging to the insect FXPRL amide peptide family (Imai et al., 1991; Sato et al., 1992; Shen et al., 2018). It shares the same C-terminus five amino acid sequence with several other neuropeptides, including pheromone biosynthesis activating encoded in a single reading frame (Kawano et al., 1992; Sato et al., 1993). In situ hybridization with the cDNA encoding the common polyprotein precursor was performed, and 12 cells from three clusters were stained in SG: the three clusters were localized in the mandibular (anterior), maximal (medial), and labial (posterior) neuromere (Sato et al., 1994). Shiomi et al. (2015) established a TALEN-based KO mutant of DH-PBAN and showed that DH signaling is essential for embryonic diapause induction. Morphological polymorphism of the moth wing is determined also by DH action during the early pupal developmental stage (Tsurumaki et al., 1999; Yamanaka et al., 2000). A gene encoding DH was identified within other Lepidopterans lacking the embryonic diapause, and various physiological actions of DH were reported (Denlinger, 2022).

Fukuda and Takeuchi (1967) performed histological and surgical experiments and demonstrated that a pair of putative DH-producing cells were located on both sides of the midventral line of SG at a level approximately two-thirds from its anterior end. Immunocytochemical studies using DH and PBAN antiserum revealed three clusters of neurosecretory cells located on the ventral midline of SG: four cells in the anterior, six cells in the medial, and two cells in the posterior cluster (Ichikawa et al., 1995). The number, size, and position of the cells were identical to those observed by cDNA hybridization experiments reported by Sato et al. (1994). Surgical removal of the cluster cells revealed that the posterior cells had diapause induction activity, whereas median cells had pheromonotropic activity (Ichikawa et al., 1996). These posterior cells (hereafter called DH cells) corresponded to the neurosecretory cells that Fukuda and Takeuchi (1967) observed as being the pupative DH cells.

The level of DH in SG gradually increased during larval development and peaked at the early pupal stage in both diapause and non-diapause-producing silkworms (Kitagawa et al., 2005). It decreased markedly in diapause producers compared with non-diapause producers at the middle pupal stage, implying the release of DH into the hemolymph in diapause producers. Additionally, DH was found to act on developing ovaries through a G protein-coupled receptor during pupal-adult development to produce the diapause eggs (Homma et al., 2006). It was reported that DH activates trehalose in the ovary cells and the oocytes (Kamei et al., 2011; Cui et al., 2021).

Lucifer yellow-filled DH cells projected bilaterally symmetric dendritic branches to the anterior half of SG, and the axon of the cell passed through the Br and then entered the nervus corporis cardiaci3 (NCC3) to spread varicose terminal branches in CC-CA (Ichikawa et al., 1995). Immunocytochemical studies performed by Sato et al. (1998) showed the same results. Shimizu et al. (1997) confirmed the DH secretion from CC-CA by conducting anatomical experiments, including the expiration of SG and Br-SG complexes with CC-CA. In summary, these observations demonstrated that DH produced in the somata of the DH cells located in SG is transported to CC-CA via NCC3, and finally released into the hemolymph.

Long-term chronic recordings were also made from the axonal tract NCC-3 of DH cells during pupal-adult development (Ichikawa, 2003; Ichikawa and Kamimoto, 2003; Ichikawa and Suenobu, 2003). DH cells in the diapause producers showed firing activities throughout pupal-adult development, whereas the same cells in the non-diapause producers maintained an inactive state until the last quarter of the pupal period. The ionic mechanism of excitability in DH cells was analyzed with intrasomatic recordings (Miyazaki, 1980). The action potential in the soma is primarily dependent on the presence of calcium (Ca) ions, and the excitability of the soma is characterized by the less pronounced potassium (K) conductance, which causes the large and prolonged Ca-dependent action potential. Injection of a Ca-chelating agent, ethylene glycol bis tetraacetic acid (EGTA), into the pupae of diapause producers induced non-diapause eggs (Xian et al., 1997), whereas the injection of KCl in non-diapause producers induced diapause eggs (Yoshitake, 1954). The hemolymph K concentration of the silkworm was reported to rapidly elevate during the larval-pupal development (Shimizu, 1982b). These observations suggest that developmental changes in hemolymph cations concentration might affect DH secretion (Kitagawa et al., 2005).

Control of DH secretion by brain

Fukuda (1952) conducted some surgical manipulation experiments using pupae and concluded that, in nondiapause producers, the release of DH was inhibited by the Br through the esophageal connectives, but in diapause producers, it was stimulated by the Br: he proposed both inhibiting and stimulating action of the Br in DH secretion (Raabe, 1989). Moreover, Tsuchida and Yoshitake (1983b) proposed inhibiting action of the Br, and Morohoshi and Oshiki (1969) stimulating action by performing translation experiments of the Br-SG complex. Matsutani and Sonobe (1987) carried out microsurgical manipulation of the Br to reveal the controlling region of DH secretion. They found that cortices lateral to the median line on the dorsal side of the protocerebrum were related to the inhibitory control of DH secretion and that cortices of anterolateral areas on the ventral side of the protocerebrum participated in the stimulatory control. Shimizu et al. (1997) demonstrated that microsurgical transection of the protocerebrum from other cerebral parts induced diapause egg production in non-diapause producers, whereas the operation in diapause producers did not affect diapause production. From these observations, the authors concluded that the protocerebrum of the Br plays a crucial role in the inhibitory action on the DH secretion in non-diapause producers.

GABAergic control of DH secretion

The GABAergic control of DH secretion was demonstrated in pharmacological studies by Shimizu and his colleagues (Shimizu et al., 1989; Hasegawa and Shimizu, 1990; Shimizu et al., 1997). They used in vitro culture experiments to demonstrate the pharmacological effects of γ-aminobutyric acid (GABA) and picrotoxin (PTX), a GABA antagonist, on DH secretion from isolated Br-SG complexes or SG. Ichikawa (2003) observed the inactivity of firing from the axonal tract (NCC-3) of DH cells in non-diapause producer pupae. Further, he found that the injection of PTX into nondiapause producer young pupae induced firing excitation in the inactive DH cells, supporting the hypothesis that DH cells are regulated by GABAergic signals. One Br of the Daizo race was reported to contain approximately 0.32 nmol of GABA just after pupation (Yang et al., 1992). A considerable number of neurons of Br and SG were immunoreactive to a GABA-antiserum (Shimizu, 1998). There have been some reports describing the involvement of GABAergic regulation in the diapause induction of other insects (Webb and Denlinger, 1998; Zhang et al., 2013; Nikonorov et al., 2018).

Recently, Tsuchiya et al. (2021) proposed a hierarchical pathway consisting of GABAergic and subsequent Crz signals that modulate diapause induction. Crz is an undecapeptide involved in various physiological functions, including cardiostimulation, regulation of silk spinning rate and diapause induction in insects (Tanaka et al., 2002; Shiga et al., 2003; Predel et al., 2007). They reported that the injection of Crz into pupae of non-diapause producers induced diapause eggs. In addition to this noteworthy finding they observed that the expression of the plasma membrane GABA transporter (GAT) gene was 10–100 fold higher in the Br-SG complex in diapause producers than in non-diapause producers. GAT transports GABA from the synaptic cleft back into the presynaptic neuron to terminate GABAergic inhibition (Umesh and Gill, 2002). Dorsolateral Crz neurons were found to express PTX-sensitive GABA receptors in the Br. Furthermore, they discovered that KO mutants of GAT became mostly non-diapause producers, regardless of the diapause-producing condition. As described below, the expression of GAT is regulated by the circadian clock in the silkworm (Cui et al., 2021). These observations strongly suggest that GAT is a crucial factor in the diapause production of the silkworm through the circadian control.

Circadian clock genes

Numerous studies have investigated the role of circadian clock genes in the photoperiodism of insects (Ikeno et al., 2011; Goto, 2013; Meuti et al., 2015; Mukai and Goto, 2016). Clock genes have been identified in the silkworm (Iwai et al., 2006; Xiang et al., 2018), and KO of per and tim genes was reported to cause the loss of circadian behavioral activities (Ikeda et al., 2019; Nartey et al., 2021).

Recently, three groups reported the effects of KO of the per gene on silkworm PR using different races (Cui et al., 2021; Ikeda et al., 2021; Tobita and Kiuchi, 2022). Ikeda et al. (2021) generated a per KO strain using TALEN-mediated genome editing technology with the Kosetsu race. They observed the loss of the circadian rhythm for hatching and eclosion in the KO strain and a concurrent disruption of the temporal expression pattern of per and tim in the larval head. They also observed that the wild strain showed LD PR during the larval stage, whereas the KO strain showed 100% nondiapause production under both LD and SD conditions (Fig. 5A). Cui et al. (2021) also generated a KO strain of the per gene using the TALEN method with the Dazao race. They found that the KO disrupted the temporal expression patterns of the circadian clock genes compared with those of the wild-type strain, and reduced diapause induction under LL conditions during the embryonic stage at 20°C (Fig. 5B). Another research group, Tobita and Kiuchi (2022), generated KO mutants of per gene using the CRISPR/Cas9 method with the Daizo race. In the KO mutant, the diapause-producing action of SD during the larval stage and that under LL conditions during the embryonic stage were completely lost (Fig. 5C, D). In summary, the KO of per caused nondiapause production regardless of the diapause-inducing photoperiodic conditions and prevented the silkworm PR. Per KO mutants are unable to distinguish not only light condition but also temperature condition (Homma et al., 2022).

Fig. 5.

Effect of per gene KO on the PR. (A, B) were redrawn from the data of Ikeda et al. (2019) and Cui et al. (2021), respectively, and (C, D) were both redrawn from the data of Tobita and Kiuchi (2022). In each experiment, the silkworm race (R), embryonic stage conditions (ESC) and larval stage condition (LSC) were as follows. In experiment (A), R: the Kosetsu race, ESC: under LL at 20°C, LSC: under SD (L12:D12) or LD (L16:8) at 25°C. In experiment (B), R: the Dazao race, ESC: under DD or LL at 20°C, LSC: under 12L:12D at 25°C. In experiment (C), R: the Daizo race, ESC: under DD at 25°C, LSC: under SD (L8:D16) or LD (L20:D4). In experiment (D), R: the Daizo race, ESC: under DD or LL at 18°C, LSC: under 12L:12D at 25°C. Open, hatched, and closed bars show proportion of moths that oviposited non-diapause, mixed, and diapause eggs in the same batch, respectively. WT: wild-type strain, KO: knockout strain.

Similar results were obtained by KO of other core circadian genes such as tim, Clock, and cycle (Tobita and Kiuchi, 2022). This evidence suggests that the circadian clock is not just related to the PR induction through the pleiotropic function of one or a few genes, but is fully integrated into the mechanism (Denlinger, 2017). In the per KO mutant, Cui et al. (2021) reported that the transcription levels of GAT and GABA transaminase (GABAT) in the pupal Br-SG were downregulated compared with those in the wild-type strain. Notably, the level of GAT gene transcription was remarkably downregulated to approximately one-tenth in the per KO silkworm. The authors also observed upregulation of the transcription of the glutamic acid decarboxylase (GAD) gene and ionotropic GABA receptor (GRD) gene in the mutant. These alterations in the transcription of the genes resulted in an elevated amount of GABA released into the synaptic cleft.

Among the core genes of the circadian clock in the silkworm, per, tim, and cry2 are known as negative regulators, while Clock and cycle are known as positive regulators in the feedback loop, based on the transcription and translation of the clock genes (Brady et al., 2021; Goto, 2022). Cui et al. (2021) reported that the circadian clock system directly upregulated the expression of iontropic GABA receptor (GRD) in the pupal stage through Cycle.

In bivoltine races, incubation at a high temperature (25°C) resulted in 100% diapause occurrence irrespective of the light conditions. Nonetheless, both the group of Cui et al. (2021) and that of Homma et al. (2022) observed that KO mutants of the per gene exhibited nearly 100% non-diapause even at the high temperature incubation under DD. Homma et al. (2022) obtained similar results in the KO strains of other clock genes tim, Clock, cycle, and cry2. This diapause disappearance was significant in KO mutants of per and time. Further, they found that both PTX and DH injection into the pupae of these KO strains restored the diapause production. These observations suggest that the circadian clock, which is involved in diapause induction, is hierarchically located upstream of the GABAergic pathway. Liu et al. (2021) found that bmo-miR-6497-3p, one of the miRNAs, may be involved in the diapause induction of the silkworm by affecting the expression of the circadian clock genes during the blastokinesis phase. They discussed the epigenetic regulation of the circadian clock genes mediated by the miRNA in the diapause induction under different temperatures.

Relevant genes

The molecular receptor responsible for temperature cues involved in insect dormancy and PR remains unclear. Shiomi and his coworkers proposed that Bombyx (Bm) transient receptor potential ankyrin 1 (TRPA1) may function as a thermosensor in the diapause induction (Sato et al., 2014). A hypothetical model representing the hierarchical pathway from the perception of temperature cue by BmTRPA1 to DH release through GABAergic regulation has been proposed by Tsuchiya et al. (2021). In insects, TRPA1 is a Ca²⁺ -permeable, non-selective cation channel localized to the plasma membranes and found to be expressed in nociception receptor neurons and Br (Kang et al., 2019). They reported that the knockdown of BmTRPA1 through embryonic RNAi influenced diapause incidence and DH release during pupal-adult development in the Kosetsu race. Additionally, they found that the KO of BmTRPA1 induced LD response during the larval stage even at a high temperature incubation at 25°C under DD, which caused 100% diapause in the wild-type strain (Yokoyama et al., 2021). This means BmTRPA1 is not necessary for the function of the circadian clock underlying the PR during the larval period (Ikeda et al., 2021; Tobita and Kiuchi, 2022). On the other hand, in the KO strain, there were significant temporal changes in the expression profiles of the clock genes compared to the wild strain and no daily rhythm was observed, suggesting an interaction between BmTRPA1 and the circadian clock (Homma et al., 2022). Thermoperiod can synchronize the circadian clocks as a Zeitgeber, and substitute for photoperiod in the diapause induction of insects (Saunders, 2014). The clarification of BmTRPA1's role in these phenomena can be expected.

Akitomo et al. (2017) conducted a microarray analysis of expressed genes in the larval Br and found that the Crp18a1 (cytochrome P450 18a1) and Kr-h1 (Krüppel homolog 1) genes were upregulated in diapause-producing silkworms compared with non-diapause producers. The Crp18a1 gene encodes a key enzyme for steroid hormone inactivation (Guittard et al., 2011), and Kr-h1 encodes a zinc-finger-type transcription factor that mediates the repression of metamorphosis (Kayukawa et al., 2014). In a more recent study, Egi and Sakamoto (2022) proposed that the programming process of diapause induction by maternal photoperiodic and thermal conditions may involve juvenile hormone acid methyltransferase (Jhamt) and proton-coupled folate transporter (Pcft).

The epigenetic control of PR in insects has also been observed recently (Trionnaire et al., 2013; Reynolds, 2017). For example, in Nasonia vitripennis, knocking down DNA methyltransferase (Dnmt1a) or pharmacologically blocking DNA methylation disrupted PR of diapause induction, suggesting that DNA methylation is associated with the photoperiodism of the wasp (Pegoraro et al., 2016). In the silkworm, there has been one pharmacological study examining epigenetic control, in which the dietary administration of 5-azacytidine, a DNA methyltransferase inhibitor, did not affect PR during the larval stage (Egi et al., 2017).