Optimized pupal age of Tenebrio molitor L. (Coleoptera: Tenebrionidae) enhanced mass rearing efficiency of Chouioia cunea Yang (Hymenoptera: Eulophidae)

Chouioia cunea Yang (Hymenoptera: Eulophidae) has been widely used for biological control of the fall webworm, Hyphantria cunea (Drury) (Lepidoptera: Arctiidae), in China. The yellow mealworm, Tenebrio molitor L. (Coleoptera: Tenebrionidae), an important resource insect species distributed worldwide, is considered to be a potential alternative host for mass rearing of C. cunea to the Chinese oak silkworm, Antheraea pernyi (Guerin-Meneville) (Lepidoptera: Saturniidae), which is currently used. In this study, we investigated the effects of host age on C. cunea mass rearing by measuring parasitism, development and adult fertility of C. cunea on T. molitor pupae of different ages. The results showed no significant differences in the percentage of parasitized hosts and developmental time of C. cunea in pupae of different ages. However, the number of C. cunea adults (137.2–154.7 adults per host) that emerged from 0, 1, and 2-day-old pupae was significantly higher than that from 4-day-old pupae. The lowest percentages of unemerged adults were found in 2-day-old (1.2%) and 3-day-old (1.4%) pupae, which were significantly lower than that of 4-day-old pupae (10.3%). The emergence of adult females from 0 to 2-day-old pupae (120.2–142.3 per pupa) was significantly higher than that from 4-day-old hosts (64.6). Adult females emerging from 2-day-old pupae carried significantly more eggs (258.2 eggs/female) than those from 0 and 1-day-old pupae (178.4–178.9 eggs/female). Our findings indicated that 2-day-old pupae of T. molitor were most suitable to rear C. cunea. Overall, this research provided valuable information to optimize pupae for the mass rearing of C. cunea on host T. molitor.

cunea, reaching a parasitism rate of 88%, than that in nonrelease control plots with a 4.7-12.9% parasitism rate 12 . In China, the mass production of C. cunea is primarily via the Chinese oak silkworm, also known as the tussah silkworm, Antheraea pernyi (Guerin-Meneville) (Lepidoptera: Saturniidae) 2,13 , and approximately 325.54 billion C. cunea adults have been released to two-thirds of the H. cunea infested area (235,000 ha) in China in biological control programs from 1986 to 2012 12 . Compared with other hosts, A. pernyi has various advantages 14 and has also been largely used for mass production of Trichogramma parasitoids for biological control of corn borers in northeastern China 15,16 . The mass rearing of C. cunea on A. pernyi, as well as its field release to suppress H. cunea, have been recognized as major biocontrol successes in China 2,17 .
Although A. pernyi is a good alternative host for rearing C. cunea, some challenges remain limiting its application. For example, in areas unsuitable for A. pernyi production, the transportation and storage of A. pernyi can be costly. In addition, bacterial diseases infecting A. pernyi pupae have posed severe threats to C. cunea mass rearing 18 . Therefore, it is essential to seek alternative host species for mass production and field application of C. cunea.
Previous research highlights that C. cunea can be reared on pupae of the silkworm, Bombyx mori L. (Lepidoptera: Bombycidae) 19 , Asian corn borer, Ostrinia furnacalis (Guenée) (Lepidoptera: Pyralidae) 20 , and the yellow mealworm, Tenebrio molitor L. (Coleoptera: Tenebrionidae) 21 . Tenebrio molitor can be easily reared with low cost, and the larvae are currently recognized as one of the most common foods for captive insectivore mammals, birds, reptiles, and amphibians 22,23 . Tenebrio molitor has also become popular as human food in some countries or regions due to its high nutritional value 24,25 , yielding many research efforts on food processing and safety concerns 26,27 .
A previous study reports that 75% of T. molitor pupae can be successfully parasitized by C. cunea, including deceased T. molitor pupae 21 , and 167.3 progeny per pupa are reported from treatment with a parasitoid:pupa ratio of 3:1, which indicate that T. molitor pupae are a promising alternative host for mass rearing C. cunea. However, little study has been conducted to optimize the rearing conditions such as determining host age and other factors of T. molitor pupae for mass rearing of C. cunea. Notably, host age is reported to have generaly a substantial influence on the development of parasitoids in their hosts [28][29][30] .
In this study, we investigated the parasitism and development of C. cunea parasitoids on T. molitor pupae at different ages (i.e., 0, 1, 2, 3 and 4-day-old pupae) and the effects of pupa age on fertility of C. cunea females. This study was expected to provide useful information to optimize the mass rearing of C. cunea on T. molitor.

Results
Effect of host age on parasitism of C. cunea. When 0, 1, 2, 3 and 4-day-old pupae of T. molitor were provided to C. cunea, no significant differences in percentage of parasitized pupae were found (F 4,15 = 1.95, P = 0.154) (Fig. 1). Significant differences in percentage of deformed T. molitor (presumably caused by parasitoid host feeding at T. molitor pupae stage) were detected among different host ages (F 4,15 = 3.44, P = 0.035). The highest percentage of deformed T. molitor adults emerged from 4-day-old hosts (30.0%), followed by 0 and 3-day-old pupae, and no deformed T. molitor adults emerged from 1 and 2-day-old pupae.
Effect of host age on C. cunea development. The results showed that host pupae of all ages were parasitized and that C. cunea emerged successfully. However, significant differences occurred in the total number of C. cunea obtained per host (N c ) among host ages (F 5,15 = 8.84, P = 0.001). One-day-old pupae produced the highest www.nature.com/scientificreports www.nature.com/scientificreports/ number of parasitoids (154.7), followed by 0, 2, and 3-day-old pupae, and 4-day-old pupae produced the lowest number of parasitoids (72.1) ( Table 1).
The highest percentage of unemerged C. cunea per pupae (P ue ) was observed in 4-day-old pupae (10.3%), followed by 0 and 1-day-old pupae. The lowest percentage of unemerged parasitoids was observed in 2 and 3-day-old pupae (1.2-1.4%) (F 5,15 = 3.40, P = 0.036). Although no significant difference occurred in the percentage of adult females reared per pupa (P f ) among different host ages (F 5,15 = 0.49, P = 0.746), the host age significantly affected the number of emerged female adults per host (N fe ) (F 5,15 = 5.90, P = 0.005). No significant differences in N fe were found between 1-day-old pupae and 0, 2 and 3-day-old pupae; however, the value was significantly higher than that from 4-day-old pupae. No significant differences were found in C. cunea developmental time among all host ages (F 4,15 = 2.91, P = 0.058) ( Table 1).
Effect of host age on fertility of female C. cunea. Significant differences were detected in number of all eggs in ovarioles per C. cunea female among all host ages (F 4, 15 = 10.30, P < 0.001). Generally, C. cunea females emerging from 2-day-old pupae carried the largest number of eggs (258.2), followed by those from 3 and 4-day-old pupae. The number of eggs carried per female was the lowest in parasitoids emerged from 0 and 1-day-old pupae (178.4-178.9) (Fig. 2).

Discussion
The results of the current study indicated the great potential of T. molitor pupae as alternative hosts for mass rearing of C. cunea. We highlight that C. cunea preferred parasitizing younger T. molitor pupae. When older pupae were provided, the percentages of unsuccessfully parasitized hosts and unemerged C. cunea remaining in hosts were higher, whereas more female adult parasitoids emerged from hosts when C. cunea females parasitized younger hosts. Our results showed that 95% of 1-day-old and 90% of 2-day-old pupae of T. molitor were successfully parasitized by C. cunea. This result consolidates the earlier research by Yang & Li 21 , where only 75% of T. molitor pupae were successfully parasitized by C. cunea without host optimization reported. Overall, 1 and 2-day-old pupae of T. molitor can be considered as highly suitable hosts for the mass rearing of C. cunea. In addition, C. cunea females emerging from 2-day-old host pupae showed the greatest fertility compared with that from the other host ages.  www.nature.com/scientificreports www.nature.com/scientificreports/ Compared with A. pernyi, T. molitor showed many advantages as host for C. cunea. For example, A. pernyi overwinters as diapausing pupae and generally takes six months to complete the cycle 31 , whereas T. molitor can breed continuously throughout the year and provide fresh pupae uninterrupted. Thus, T. molitor as host can help to reduce the storage cost for C. cunea mass rearing. Furthermore, other hosts such as A. pernyi and B. mori are economically important insects and can only be cultured in limited regions and seasons, which has limited the potential of improving mass production of C. cunea via A. pernyi and B. mori. Other hosts, such as O. furnacalis pupae, show potential for C. cunea mass production 20 , but for O. furnacalis, strict management is required in the field since it can be a major agricultural pest during mass production.
Our results also showed that approximately 30% of beetles emerged deformed from 4-day-old parasitized pupae, indicating the unsuccessful parasitism of C. cunea. Yang & Li also found that the percentage of deformed T. molitor beetles reached 50% when C. cunea parasitoid and host pupa were at the ratio of 1:3 21 . Deformed T. molitor beetles usually showed a pair of incomplete forewings but could still crawl and forage. The secretions from the venom gland of C. cunea may play a major role in causing deformities of parasitized beetles 32 . Similarly, previous research showed that 10% venom sac extract from Tetrastichus sp. (Hymenoptera: Eulophidae) artificially injected in O. furnicalis pupae caused 8.2% deformed moths with defects of spreading wings and that soon died 33 . In addition, the deformed beetles may be explained by contributions from other factors such as host resistance to parasitism and host feeding by C. cunea. Indeed, Yang & Xie also found that C. cunea adult females display host feeding after oviposition 34 . Zhu et al. indicated that the expression of 74 unigenes involved in T. molitor immune response was significantly altered after T. molitor pupae were parasitized by Scleroderma guani 35 . Although our study reported significant differences in the percentages of deformed T. molitor beetles among different host ages, further studies to elucidate the physiological mechanisms leading to these deformities in T. molitor are urgently needed.
The host age is reported to significantly affect the number of parasitoids per host, while a similar number of C. cunea can be reared on 0 and 2-day-old hosts. When C. cunea adults and their hosts were tested at a ratio of 2:1, the total number of C. cunea reared per host pupa reached 181.7 parasitoids 21 , a number higher than the 154.7 C. cunea reared on 1-day-old pupae in our study. Moreover, the developmental period of C. cunea in all host ages ranged from 21.8 to 23.6 days at 25 °C, which is longer than that reported by Yang & Li (17.2 to 22.9 days at 28 °C) 21 . Generally, our results indicated that the pupal age had no effect on developmental times of C. cunea. However, pupal age significantly affected the number of adult females emerging per host. The number of adult females emerged from 0 to 2-day-old pupae was approximately 2-fold of that emerged from 4-day-old pupae. The number of eggs carried by female parasitoids is a key index for quality evaluation and reflects the fertility of a parasitoid in biological control programs 36 . Our results indicated that the C. cunea females emerging from 2-day-old pupae carried more eggs than those that emerged from 0, 1, and 3-day-old pupae. Notably, Sun et al. reported a positive linear relationship between egg load and female body size of C. cunea 37 . However, our results indicated a balance occurred between the number of eggs carried and the number of C. cunea reared on pupae of different ages.
As T. molitor beetles can be easily reared worldwide with fewer limitations and ecological risks, the beetle can be considered as a promising host to enhance mass rearing of C. cunea. Overall, our study showed that 2-day-old pupae were the most suitable age for rearing C. cunea. This research adds valuable information to optimize the mass rearing of C. cunea on its alternative host T. molitor. Further studies are needed to develop efficient field release operations of C. cunea in biological control programs, as well as the technology for long-term storage of C. cunea parasitoids. Effect of host age on C. cunea parasitism. The experiment was conducted at 25 ± 1 °C, 65 ± 5% R.H. and in complete darkness in an incubator (versatile environmental test chamber, MLR-351H, SANYO Electric Co., Ltd., Japan). Previous research shows that the adult females of C. cunea mate before emerging from the host pupa of A. pernyi 34 and that the females oviposit 51.7% of their total eggs in the first day after emergence 37 . Therefore, during the experiment, two newly emerged (<5 h) and mated C. cunea adult females were introduced into a glass tube (diameter: 3.5 cm, length: 10 cm) with one T. molitor pupa (1, 2, 3, or 4-day-old) to allow parasitism 21 . After 24 hours, the parasitoids were removed, and the pupa from each age treatment was monitored daily to document the parasitoid emergence. The number of emerging T. molitor adults, deformed T. molitor adults (presumably caused by host feeding) and parasitized pupae were recorded. The numbers of parasitoids emerged and unemerged parasitoids from each host were recorded, and sexed. Parasitoid larvae remaining in each host pupa were also counted. The developmental time of C. cunea was documented from parasitism to adult emergence.

Hosts. Tenebrio molitor
www.nature.com/scientificreports www.nature.com/scientificreports/ Four replicates were conducted for each pupal age, and 10 pupae were examined per replicate; thus, a total of 40 pupae were tested at each pupal age.
Effect of host age on fertility of female C. cunea. A previous study indicated that the eggs of newly emerged females of C. cunea parasitizing Chinese oak silkworm pupae were nearly matured 37 . Therefore, the fertility of C. cunea females was evaluated by assessing the number of all eggs carried by each adult female immediately after emergence. For all tested ages of parasitized T. molitor pupae, 10 newly emerged C. cunea adult females were randomly selected for each replicate. The collected parasitoids were then dissected under a stereomicroscope (SMZ-168 series, Motic, China) to count the number of eggs per C. cunea female 40 . Each treatment was replicated 4 times. A total of 40 adult females emerged from each host age were examined.
Statistical analyses. For each biological parameter described above, data were analyzed using one-way ANOVA, and means were compared using Tukey's HSD test at P < 0.05. All data were subjected to a normality test (Shapiro-Wilk test) prior to ANOVA. Female progeny (%) and unemerged C. cunea (%) data were arcsine square root transformed, while count data were logarithm-transformed prior to the normality test. All the statistical analyses were performed using the SAS statistical software package (SAS Institute, Cary, NC, USA). The figures were plotted using OriginPro 2017 SR2.
Number of C. cunea adults reared per host was calculated based on the equation below: where N a is the total number of C. cunea adults per host, N fe is the number of emerged adult females per host, N me is the number of emerged adult males per host, N fue is the number of unemerged adult females per host, and N mue is the number of unemerged adult males per host. The total number of C. cunea reared per host was calculated based on the equation below: where N c is the total number of C. cunea reared per host, N a is the number of C. cunea adults reared per host, and N ll is the number of larvae per host.
Percentage of unemerged C. cunea reared per host was calculated based on the equation below: where P f is the percentage of adult females reared per host, N fe is the number of adult females emerged per host, N fue is the number of unemerged adult females per host, and N a is the number of C. cunea adults reared per host.