Biological characters of Trichogramma dendrolimi (Hymenoptera: Trichogrammatidae) reared in vitro versus in vivo for thirty generations

Trichogramma dendrolimi which is an economically important biological control agent were reared for 30 generations on a modified artificial medium and natural host. Biological characters were assessed and compared with parasitoids reared in vivo. Pupation rate and normal adults rate of in vitro–reared parasitoids were significantly higher compared with in vivo–reared parasitoids. The adult emergence rate, number of adults produced, and fecundity of T. dendrolimi reared in vitro were lower than those reared in vivo. The percentage of females and longevity did not vary between the two rearing methods. The overall fitness of the parasitoids reared artificially from the first to the 20th generation was higher than of those reared from the 21st generation onwards. No differences were observed in the fitness parameters of parasitoids reared in vivo across any of the 30 generations. The results suggest that the modified artificial medium used in this study is suitable for the continuous rearing of T. dendrolimi for at least 20 generations, and has the potential for the mass production of these parasitoids for use in biological control. Such a substrate could be examined for use in rearing other parasitoid species that are important in biological control.

of nutrients between the artificial medium and a natural host over 10 generations 16 . Therefore, the present study determined the quality of the modified rearing medium by comparing biological parameters, including rates of parasitism, pupation, adult emergence and normal adults, number of adults, percentage of females, fecundity, and longevity, of T. dendrolimi reared on artificial medium and those reared on A. pernyi eggs for 30 generations.

Comparison of in vitro-and in vivo-reared T. dendrolimi over 30 generations. Rates of pupation
and normal adults from generation G1 to G30 were significantly higher in the in vitro culture compared with in vivo culture (Fig. 1), although there was no difference in the rate of normal adults at G25 (Table 1).
By contrast, the mean rate of adult emergence from pupae, number of adults produced, and fecundity of T. dendrolimi reared on artificial medium were lower than for those reared on A. pernyi eggs over the 30 generations ( Fig. 2) except for G5, where no differences occurred in the number of adults produced and in their fecundity (Table 1).
In the in vivo culture, emergence rates were over 90% in all the generations, whereas they were all under 80% in the in vitro culture. The number of adults and fecundity of the in vivo-reared G30 generation was almost twice that of the in vitro-reared insects.
The percentage of females and longevity rates of T. dendrolimi adults grown in vitro versus in vivo were similar across the 30 generations (Fig. 3). The percentage of females in G1, G5, G25, and G30 was significantly greater when reared in vivo, and there were significant differences in lifespan of G5, G10 and G20 insects between two rearing methods. However, no differences were found in either lifespan or the percentage of females in in vivoversus in vitro-reared insects in the other generations ( Table 1).  Table 2. For in vitro-reared insects, there were significant differences in all parameters except for the percentage of females (Table 3). Emergence rate, normal adults rate, number of adults produced, and fecundity of G30 insects decreased by 37.18%, 11.94%, 46.77%, and 37.01%, respectively, compared with G1 parasitoids.

Efficacy of in vitro versus
For in vivo-reared insects, no differences occurred across generations in any parameter except the rate of emergence and number of normal adults at G25, and females longevity at G20-G30 (Table 3).

Discussion
To aid the successful and economic rearing of parasitoids for use in biological control programs, there is a need to compare the biological characteristics of in vitro-reared insects with those reared in factitious and/or natural hosts, not only over a single generation, but also over continuous generations, especially when focusing on the quality of parasitoids reared on artificial media 8,17 . Therefore, the current study analyzed the biological traits of T. dendrolimi successively reared in vitro and in vivo for 30 generations. Before T. dendrolimi adult emergence began, each egg card was cut open to aid emergence and allow the insects to spread their wings, resulting in fewer deformed adults being obtained on artificial hosts. In fact, the number of normal adults (effective adults) reared in vitro were less than in vivo. In addition, similar percentage of females and lifespan between the two culture systems demonstrated that it is possible to use the artificial medium tested here in to reproduce T. dendrolimi and the efficacy is similar with the natural host eggs.
The lower emergence rates, fewer adults produced and fecundity observed with in vitro culture versus in vivo culture and previous biochemical results of the adults produced in vitro, which showed decreased protein concentrations compared with those reared in vivo 16 , indicated that the medium might not be wholly suitable for mass rearing in terms of its nutrient load compared with natural host eggs. Thus, parasitoids reared on artificial medium appeared to be inferior compared with those reared on A. pernyi eggs. The emergence rates of both T. dendrolimi and Trichogramma chilonis reared in vitro were 90% of that reared in vivo 18 , and in vitro-reared females of Trichogramma australicum produced significantly more progeny than did females reared on natural or factitious hosts for only one generation 9 . Trichogramma minutum were reared for 10 generations on an artificial diet that resulted in more deformed females, but also in adults that lived longer, parasitized more Helicoverpa Figure 2. Emergence rate, number of adults produced and fecundity of Trichogramma dendrolimi reared in vitro and in vivo for 30 generations. Means (±SE) were calculated from five replicates. Arcsin transformation (= asin (sqrt (x/100))) where x is a percentage, and Log10 transformation (= log10(x)) where x is data of fecundity and number of adults produced. Data with an asterisk differed significantly according to pairedsample t-tests at P = 0.05.
SCientifiC RepoRTS | (2017) 7:17928 | DOI:10.1038/s41598-017-17915-9 zea eggs and had similar emergence rates compared with insects reared in vivo on H. zea eggs, suggesting, while this medium is nutritionally adequate, additional work is required for it to be suitable for use in mass rearing programs 8 .
Compared with the performances of T. dendrolimi reared on the artificial host egg EC-II 19 , the modified medium used herein supported the production of more than 20 generations, with more pupae, normal adults, and females. Dai et al. reported the continuous rearing of T. dendrolimi on an artificial diet comprising pupal holotissue of A. pernyi (30%), egg yolk (14%), skimmed milk (26%), and distilled water (30%) for 41-50 generations, with 60-80% pupation and emergence rates, while obtaining 8% malformed adults 20 . However, Grenier and De Clercq suggested that it is not advisable to maintain entomophagous insects on synthetic diets for many generations, because they may suffer from non-intentional selection, inducing a reduction in genetic variability and deterioration in performance 6 .
Parasitoids of G1-G20 showed fitness to the artificial host compared with G21-G30, with stronger and more numerous normal adults, higher reproductive capacity, -live longer and stable percentage of females. However, almost all the biological parameters decreased after the 20th generation. Nutritional deficiency of the artificial medium should be the major reason of the intergenerational defects. Other than nutrition, inbreeding, which is common in parasitoids, might be one explanation for this decline in fitness [21][22][23] . Although inbreeding depression in Trichogramma appears unlikely 24,25 , the decline in genetic quality could arise because the in vitro-reared adults were bred in a small egg card that increased opportunities for inbreeding. Meanwhile, deformed adults that were not removed during the continuous rearing process could have also resulted in the population decline 21 . Therefore, for successful mass production, it would be necessary to rejuvenate the parasitoids every few generations. However, the frequent introduction of new strains for in vitro mass production would: (i) require allowing each new strain to adapt to laboratory conditions within a few generations; (ii) carry the risk of misidentification of the introduced strain or species; and (iii) risk introducing pathogens or hyperparasitoids to the breeding system 6 .
In conclusion, based on the biological parameters examined here and combined with previous biochemical analyses 16 , our studies indicate that the modified artificial medium was suitable for the development of T. dendrolimi from eggs to adults, and supported adult survival and continuous reproduction for at least 20 generations. Liu et al. reported there is no significant difference in parasitism and development of T. dendrolimi reared in the lyophilized diet that either stored under −16 °C for 522 days or stored at room temperature for 140 days in comparison with that developed in fresh diet 26 . It has been estimated that the costs for producing Trichogramma on artificial media are 50% lower than on their factitious or natural hosts 20 . The artificial medium could be used to overcome shortages in the supply of lepidopteran host eggs and reduce parasitoid production costs. It also has the potential for use as an artificial host for the large-scale production of this parasitoid. However, there is a need to test the efficacy of the parasitoids reared using this medium against target pests in the field. Future studies should also examine the continuous rearing of other species of natural enemies, and the impact on their quality, using the methods described herein.

Materials and Methods
Stock culture. Trichogramma dendrolimi was originally collected from the Institute of Plant and Environment Protection, Beijing Academy of Agricultural and Forestry Sciences, Beijing, China, the same as Lü et al. 3,4,16 reported previously. In the laboratory, T. dendrolimi stock cultures were reared on eggs of A. pernyi as a factitious host. Climatic conditions were 27 ± 1 °C, 75% ± 5% relative humidity (RH) and a 16:8 L:D photoperiod.
Artificial medium preparation. The artificial medium used in this study was the modified artificial medium developed by Lü et al. 4 (Table 4). The artificial medium was prepared as described previously 3 . Pupal hemolymph was obtained from live A. pernyi pupae that were immersed in a water bath at 60 °C for 10 min to avoid melanization of the hemolymph. After sterilization of the pupae surface with 75% ethanol,  Table 2. Biological parameters of Trichogramma dendrolimi reared in vitro and in vivo for 30 generations. Data are expressed as the means ± SE based on five replicates (data of fecundity and longevity expressed as the means ± SE based on 30 replicates). Arcsin transformation (= asin (sqrt (x/100))) where x is a percentage; Log10 transformation (= log10(x)) where x is data of fecundity, longevity, and number of adults produced. Means followed by the same letter in the same column were not significantly different at the 5% level (Tukey's test  the hemolymph was collected by pressing the pupae under sterile conditions. Neisenheimer's salt solution was prepared with NaCl 7.5 g, KCl 0.1 g, CaCl 2 0.2 g, NaHCO 3 0.2 g, and 1000 mL of distilled water and used after sterilization.

In vitro In vivo In vitro In vivo In vitro In vivo In vitro In vivo In vitro In vivo In vitro In vivo
Preparation of artificial egg cards and A. pernyi egg cards. For artificial egg cards, the preparation referred to the work by Lü et al. 16 . Twenty semispherical domes (2 mm-3 mm in diameter × 3 mm high) were produced by pressing a heated glass rod onto one half of a sheet (8 cm × 7 cm) of polyethylene and polypropylene copolymer film (30 µm thick) through a plastic semispherical mold. After the sheets were sterilized by UV irradiation, 4 µL medium was transferred to each dome using a pipette. The half of the sheet containing the domes (convex side) to be exposed to oviposition by the parasitoid was folded over the other half and sealed using a plastic sealer, so that the concave side and the bottom piece of film provided sufficient space to allow aeration for parasitoid development. Afterward, the external surface of the egg cards was treated with 10% (wt/vol) polyvinyl alcohol to stimulate oviposition. For A. pernyi egg cards, twenty eggs were attached to cardboard strips (2 cm × 2 cm) with 10% (wt/vol) polyvinyl alcohol.
In vitro and in vivo rearing. For the first generation, twenty artificial egg cards and twenty A. pernyi egg cards were placed in a plastic tray (30 cm × 20 cm × 5 cm) for exposure to T. dendrolimi adults of the same generation for 24 h. Parasitoids of both sexes were released in the trays using a ratio of parasitoids to artificial eggs of 6:1. Sex ratios were approximately 8:1 in all five replicates (one tray corresponded to one replicate). Trays were placed in climatic incubators set at 27 ± 1 °C, 75% ± 5% RH and a 16:8 L:D photoperiod. After 24 h of exposure, the egg cards were taken out and the wasps were removed. From the third day after parasitization, the egg cards were monitored for parasitoid development under a binocular microscope. Before T. dendrolimi adult emergence began, each egg card was inserted into a glass tube (3 cm in diameter × 9 cm high), which was covered by a cotton cloth with a rubber band. Each artificial egg card was cut open to aid emergence before adults emergence. When adults emerged, a new egg card was inserted into the glass tube for rearing of the next generation. Adults of two egg cards from each tray were used for biological assessment.
Biological parameters assessed. During the in vitro and in vivo rearing process, the number of larvae, pupae, total adults, and male versus female adults were counted; the fecundity of female adults produced from both cultures was recorded every five generations. Rates of pupation and adult emergence, the proportion of normal adults (i.e., adults with normal wings and abdomen) referred to Lü et al. 3,4 and percentage of females were calculated as follows:  Fecundity was measured by holding newly emerged females over a glass tube (3 cm in diameter × 9 cm high) containing an artificial egg card or A. pernyi egg card with 35 eggs. The number of larvae and pupae were counted until the death of the female and the mean fecundity of each female was then calculated. The date of emergence and death of the females were also recorded to calculate longevity. Thirty in vitro-and thirty in vivo-reared females were tested every 5 generations. Statistical analysis. Mean pupation, emergence, normal adults and female adults percentage, mean number of adults produced, fecundity and longevity for this parasitoid reared in in vitro and in vivo were compared using one-way analysis of variance (ANOVA) and Tukey's test across continuous rearing T. dendrolimi for 30 generations. Paired-sample t-tests were used to analyze differences between the artificial medium (in vitro) and A. pernyi eggs (in vivo) groups. Before analysis, percentage data were arcsine square root-transformed, data of fecundity, longevity, and number of adults produced were log10-transformed to fit a normally distributed. In all experiments, differences among means were considered significant at P < 0.05. Statistical analyses were conducted by using SPSS 17.0 software (SPSS Inc. Chicago, IL, USA).
Data availability. All data generated or analysed during this study are included in this published article.