Introduction
The use of natural enemies as biocontrol agents is an important alternative to the environment, health and resistance issues associated with the use of chemical insecticides (Greany & Carpenter, Reference Greany and Carpenter1998). However, this requires the production of large numbers of beneficial insects at low costs for augmentative and inoculative releases. The development of artificial diets could considerably reduce the costs of mass propagation compared with conventional rearing methods (Glenister, Reference Glenister, Coll and Ruberson1998; Glenister & Hoffmann, Reference Glenister, Hoffmann, Ridgway, Hoffmann, Inscoe and Glenister1998; Ruberson & Coll, Reference Ruberson, Coll, Coll and Ruberson1998; Thompson, Reference Thompson1999; Wittmeyer & Coudron, Reference Wittmeyer and Coudron2001).
Improvement of artificial diets can be convoluted, tedious and often underappreciated in mass rearing of beneficial insects. The main method for optimizing artificial diet is to measure a few preselected biochemical and (or) physiological parameters and to test the effect of changes in diet formulation on insect performance (Adams, Reference Adams2000; Wittmeyer & Coudron, Reference Wittmeyer and Coudron2001; Coudron et al., Reference Coudron, Wittmeyer and Kim2002; Coudron & Kim, Reference Coudron and Kim2004). Typically, diet components are changed one at a time and insect performance is tested after each change. This endeavor is time-consuming, taking years to decades to optimize a diet, with many attempts ending in failure.
Another approach, using n-dimensional mixture designs (Lapointe et al., Reference Lapointe, Evens and Niedz2008), identified a set of response-optimized meridic diets that contain fewer ingredients than the previous commercial diet for Diaprepes abbreviates (L.) (Coleoptera: Curculionidae), and followed that with a geometric design combined with response surface models to identify major nutritive components of the diet (Lapointe et al., Reference Lapointe, Evens, Niedz and Hall2010). Tan et al. (Reference Tan, Wang and Zhang2013) investigated an artificial diet for raising Orius sauteri (Poppius) (Heteroptera: Anthocoridae) using a microencapsulation technique. They tested 25 combinations of ingredients using an orthogonal experimental design and identified five optimal combinations based on different biological and physiological characters. The results of a follow-up test of locomotory and respiratory capacity indicated that respiratory quotient, metabolic rate, and average creeping speed were all influenced by varying dietary ingredients.
The use of gene expression could be a more direct method to accelerate diet development and identify deficiencies in diet formulations. Nutrigenomics examines how nutrition affects gene expression patterns and offers not only a molecular means to measure an insect's response to changes in the food stream but also provides information on diet limitations (Yocum et al., Reference Yocum, Coudron and Brandt2006).
Arma chinensis (Fallou) (Hemiptera: Pentatomidae) is a predaceous insect species that can effectively suppress a wide range of agricultural and forest insect pests in the orders Lepidoptera, Coleoptera, Hymenoptera and Hemiptera (Gao et al., Reference Gao, Wang and Yu1993; Chai et al., Reference Chai, He, Jiang, Wu, Pan, Hu and Ruan2000; Liang et al., Reference Liang, Zhang, Song and Peng2006; Yan et al., Reference Yan, Li, Peng, Zhou and Gao2006a, Reference Yan, Tang, Zhang and Wangb; Chen et al., Reference Chen, Zhang, Zhang, Tian, Xu and Li2007; Gao, Reference Gao2010; Zou et al., Reference Zou, Wang, Zhang, Zhang, Zhang and Chen2012). An insect-free artificial diet comprised of pig liver and tuna was developed for A. chinensis (Zou et al., Reference Zou, Wu, Coudron, Zhang, Wang, Liu and Chen2013a). Several life history parameters were diminished for A. chinensis reared on the artificial diet compared to a natural food source like the pupae of Chinese oak silk moth (COSM) Antheraea pernyi (Guérin-Méneville) (Lepidoptera: Saturniidae). Fecundity and egg viability was lower, and developmental time from 2nd instar to adult and the preovipositional period were significantly longer for diet-fed A. chinensis. Nymphal weight, body length, adult longevity, survival from 2nd instar to adult, and fertility increased, while sex ratio (♂ : ♀) decreased, with the rearing of consecutive generations on the diet. Additionally, the longevity of adults reared on the artificial diet was significantly longer than of those reared on pupae. As a result, the cost to rear A. chinensis on the artificial diet approached 2.0 times the cost of rearing A. chinensis on pupae of A. pernyi (Zou et al., Reference Zou, Coudron, Wu, Gu, Xu, Zhang and Chen2015).
The molecular mechanisms underlying the nutritive impact of the artificial diet of A. chinensis health have been investigated. The biological pathways associated with differentially expressed genes (DEGs) between the pupae-fed and diet-fed treatments were identified by mapping 13,872 DEGs and annotated sequences to the reference canonical pathways in KEGG (Kanehisa et al., Reference Kanehisa, Goto, Kawashima, Okuno and Hattori2004; Zou et al., Reference Zou, Coudron, Liu, Zhang, Wang and Chen2013b). In total, 5879 sequences were assigned to 239 KEGG pathways. The pathways most represented by the DEGs were metabolic pathways (891, 15.16%) and pathways in cancer (215, 3.66%).
One group of DEGs that were upregulated in diet-fed vs. prey-fed A. chinensis were enriched for seven pathways related to fat metabolism, including adipocytokine signaling pathway, pyruvate metabolism, fatty acid biosynthesis, glycerolipid metabolism, fat digestion and absorption, fatty acid metabolism and fatty acid elongation. These potentially signal excess dietary lipid that may be remediated by reducing the pig liver, tuna and chicken egg in the diet formulation. Another group of DEGs that were upregulated was enriched for four pathways related to starch and sugar metabolism, including carbohydrate digestion and absorption, and fructose and mannose metabolism, sugar-lipase-3, glucose transporter, and insulin and mTOR signaling pathway. These potentially signal excess dietary sugar and carbohydrates that may be remediated.
Most of the up-regulated DEGs associated with fat and sugar metabolism are also related to vitamins, including ascorbate and aldarate metabolism, vitamin digestion and absorption, folate biosynthesis, pantothenate and coenzyme A biosynthesis, nicotinate and nicotinamide metabolism, biotin metabolism, retinol metabolism, thiamine metabolism, vitamin B6 metabolism and riboflavin metabolism, as a singular substance and the results reported here.
The high instance of a DEG relating to cancer is a challenging result due in part to the dearth of information on cancer in insects. However, we speculate that canned tuna fish could have caused the enrichment of DEGs related to cancer pathways because some ingredients are carcinogens, such as acrylamide (Christova-Bagdassarian et al., Reference Christova-Bagdassarian, Tishkova and Vrabcheva2012).
Knowing the physiological roles of the DEGs enable us to predict effects of some dietary ingredients and subsequently propose formulation improvements to artificial diets.
Our objective was to compare the performance of A. chinensis reared on the previous diet with its performance when reared on a modified version of that diet wherein genome-directed formulation changes had been made to the diet that included reduced sugar, carbohydrate, biotin, nicotinamide, pyridoxine, riboflavin, thiamine, vitamin C, L-glutamine, chicken egg, pig liver, and canned tuna, increased calcium pantothenate, folic acid, and wheat germ oil, and added Ecuadorian shrimp.
Materials and methods
A. chinensis
The A. chinensis used to establish a laboratory colony for this study were obtained in July 2015 from Qian'an county (44° 57′ N, 124° 14′ E, 139 m), Songyuan city, Jilin province. Adults were held in 450 ml paper cups, one mated pair per cup, and supplied with distilled water absorbed in a cotton swab. Two soybean seedlings at the two cotyledon stage were provided in the adult rearing chambers in water-filled 6 cm × 1.5 cm-glass tubes. Water was added to the tubes every 2 or 3 days. Female A. chinensis laid egg masses on the surface of cage net or soybean leaves. Egg masses of A. chinensis were collected from the laboratory colony and transferred onto # 5 qualitative filter paper moistened with distilled water in a 9 cm diam. Petri dish under the conditions described above. Distilled water was added to the filter paper once every day. Egg hatch occurred in ca. 6 days. First instars were placed in 310 ml transparent plastic cups containing a piece of moist absorbent cotton. Molt to 2nd instar occurred in ca. 3 days after egg hatch. Immatures were maintained on COSM pupae and held at 27 ± 1 °C, 16 : 8 (L : D) and 75 ± 5% RH for ca. four generations prior to this study.
Food preparation
COSM pupae were purchased from a supermarket in Tianjin and stored at 4 °C until fed to A. chinensis. No additional preparation was required. Pupae were not used beyond 15 days of storage.
A comparison of the first artificial diet formulation (FAD) and a reformulation of that diet based on transcriptome information (RFD) are shown in tables 1 and 2. Both diets are comprised of a chemically defined portion (table 1) plus supplemental materials (table 2). The list of ingredients presented in tables 1 and 2 represents the final composition per 220 and 210 ml of FAD and RFD, respectively. The ingredients were blended together, the mixture adjusted to pH 6.8 and stored at 4 °C prior to use. The diet was encapsulated in a composite sheet (15 cm × 10 cm), constructed of Parafilm® and plastic film (Heyuan Evergreen Plastics Mfg. Co., Ltd. Taiwan), sterilized with a 5% solution of sodium hypochlorite, and formed into 40 µl hemispherical domes (Greany & Carpenter, Reference Greany and Carpenter1998; Coudron et al., Reference Coudron, Wright, Puttler, Brandt and Rice2000). The diet domes were refrigerated at 4 °C for up to 7 days prior to feeding, and new diet was prepared weekly.
Chemically defined ingredients in 220 ml of the first artificial diet (FAD) and 210 ml of the reformulated diet (RFD).
Natural product ingredients in 220 ml of the first artificial diet (FAD) and 210 ml of the reformulated diet (RFD).
a K-Lex brand (Alaska, U.S.A.).
b Canned Century brand in soybean oil (Thai Union Manufacturing Co., Ltd., Thailand).
In response to the DEGs in diet-fed vs. prey-fed A. chinensis (Zou et al., Reference Zou, Coudron, Liu, Zhang, Wang and Chen2013b) the following changes were made to the FAD formulation: the quantity of biotin, nicotinamide, vitamin B6, thiamine, riboflavin, and vitamin C were reduced and calcium panthothenate and folic acid were increased in RFD compared with FAD in order to address the upregulated genes related to vitamins; L-glutamine was reduced in order to address the upregulated genes in the pathway of alanine, aspartate and glutamate metabolism; sucrose was reduced in order to address the upregulated genes in the pathways related to starch and sugar metabolism; chicken egg and pig liver were reduced to address the upregulated genes related to fat metabolism; wheat germ oil was increased to address the low performance of males (Yousef et al., Reference Yousef, Abdallah and Kamel2003); and Ecuadorian shrimp replaced half of the tuna fish in order to decrease carcinogens that may have been in the canned fish.
Rearing with the reformulated diet
Individual 2nd instars were placed in 300 ml paper cups and maintained through the adult stage as described above. Each day the nymphs and adults were provided with fresh distilled water and encapsulated diet. The 2nd, 3rd, 4th, 5th instars and adult pairs were given 1, 1, 2, 3, and 8 diet domes (40 µl/dome), respectively. Diet domes fed to nymphs and adults were changed every day. A. chinensis was reared for six consecutive generations exclusively on the reformulated diet.
Life history evaluation
Daily observations were made to record changes in development for individuals reared on the reformulated artificial diet. Two days after emergence adults were weighed. Weight measurements were made using a Sartorius BP 211D (Sartorius AG, Göttingen, Germany) balance. Eggs were collected and counted from each mated pair daily. Egg hatch (i.e., viability) was determined by counting the number of 1st instars hatched 8 days after oviposition. The number of cannibalized eggs was verified by the empty eggshells. Fecundity was determined by counting the total number of eggs oviposited per female during its entire life and fertility was determined by counting the number of females that laid fertile eggs (Rojas et al., Reference Rojas, Morales-Ramos and King2000; Wittmeyer et al., Reference Wittmeyer, Coudron and Adams2001). For fertility calculations, females that did not lay eggs because of death due to cannibalism during the 15 days after pairing were removed.
Data analysis
Two sample t-tests for means were used to compare developmental time, body weight, body length, preovipositional period, adult longevity, and total fecundity between different treatments in the same generation. The proportion of viable eggs, survival from egg to adult, cannibalism, sex ratio and fertility were compared between different treatments in the same generation by the Chi-squared tests of 2 × 2 tables. SAS (version 8.0) was used to analyze the data.
Results
Development time
A comparison of life history parameters for A. chinensis reared on the RFD showed no significant differences in developmental times between males and females of A. chinensis in any of the treatments. Therefore, developmental time comparisons were done without sex distinction.
The preovipositional period of females was extended significantly in RFD-fed insects compared with those reared on FAD (F1, dF = 98; t = −2.10 and P = 0.038; F3, dF = 98; t = −6.29 and P < 0.001; F4, dF = 98; t = −13.71 and P < 0.001; F5, dF = 98; t = −5.13 and P < 0.001; F6, dF = 98; t = −5.16 and P < 0.001). The female reared on RFD required approximately 1.12–5.52 days longer preoviposition time than those fed on FAD. However, the preovipositional period of females was 0.86 day shorter in RFD-fed insects compared with those reared on FAD in F2, but there was no significant difference between them (dF = 98; t = 1.26 and P = 0.210) (table 6).
There was no significant difference for time to egg hatch in all generations (F2, dF = 98; t = −0.60 and P = 0.550; F3, dF = 98; t = −1.94 and P = 0.056; F4, dF = 98; t = −1.84 and P = 0.069; F5, dF = 98; t = 0.31 and P = 0.760; F6, dF = 98; t = 0.15 and P = 0.879). There was no significant difference in developmental time for 1st-instar nymphs in all treatments (F2, dF = 98; t = 1.73 and P = 0.087; F3, dF = 98; t = 1.34 and P = 0.184; F4, dF = 98; t = 0.44 and P = 0.659; F5, dF = 98; t = 0.77 and P = 0.445; F6, dF = 98; t = 1.00 and P = 0.320). However, developmental time was significantly longer for 2nd-instar nymphs reared on RFD in F1, F3, F4, F5, and F6 than those reared on FAD (F1, dF = 98; t = −7.81 and P < 0.001; F3, dF = 98; t = −3.56 and P = 0.001; F4, dF = 98; t = −7.86 and P < 0.001; F5, dF = 98; t = −5.79 and P < 0.001; F6, dF = 98; t = −4.92 and P < 0.001). But the second generation RFD-fed 2nd instars completed development approximately 0.5 day earlier than those reared on FAD (dF = 98; t = 2.44 and P = 0.017). Developmental time was significantly longer for 3rd-instar nymphs reared on RFD in F1, F2, F3, F4, F5, and F6 than those reared on FAD (F1, dF = 98; t = −7.47 and P < 0.001; F2, dF = 98; t = −10.73 and P < 0.001; F3, dF = 98; t = −5.19 and P < 0.001; F4, dF = 98; t = −10.11 and P < 0.001; F5, dF = 98; t = −11.84 and P < 0.001; F6, dF = 98; t = −8.19 and P < 0.001). Developmental time was significantly longer for 4th-instar nymphs reared on RFD in each generation than those reared on FAD (F1, dF = 98; t = −5.33 and P < 0.001; F2, dF = 98; t = −7.37 and P < 0.001; F3, dF = 98; t = −2.42 and P = 0.017; F4, dF = 98; t = −17.74 and P < 0.001; F5, dF = 98; t = −19.55 and P < 0.001; F6, dF = 98; t = −11.30 and P < 0.001). For 5th-instar nymphs, the developmental time was significantly longer for those reared on RFD in F2, F4, F5, and F6 than those reared on FAD (F2, dF = 98; t = −7.39 and P < 0.001; F4, dF = 98; t = −11.76 and P < 0.001; F5, dF = 98; t = −11.87 and P < 0.001; F6, dF = 98; t = −9.30 and P < 0.001). But there were no significant differences in developmental times for the first and the third generations RFD-fed 5th instars than those reared on FAD (F1, dF = 98; t = −1.97 and P = 0.052; F3, dF = 98; t = −0.67 and P = 0.506) (table 3).
Egg and nymphal developmental time (days) of Arma chinensis reared on a former artificial diet (FAD) (Zou et al., Reference Zou, Wu, Coudron, Zhang, Wang, Liu and Chen2013a) and reformulated diet (RFD) for six generations.
Values are mean ± SE. Asterisks identified the significant difference in development time between FAD and RFD in the same generation (P < 0.05).
a Egg mass.
b 1st instar clutch.
Weight
There was no significant difference in egg weight between RFD-fed and FAD-fed A. chinensis in F2, F3, and F4 (F2, dF = 98; t = 0.46 and P = 0.646; F3, dF = 98; t = 1.12 and P = 0.266; F4, dF = 98; t = 1.48 and P = 0.141). Egg weights were significantly higher for FAD-fed A. chinensis than of RFD-fed A. chinensis in F5 and F6. These differences, although statistically significant, were small (F5, dF = 98; t = 2.05 and P = 0.043; F6, dF = 98; t = 3.18 and P = 0.002) (table 4).
Egg and nymphal weights (mg) of Arma chinensis reared on a former artificial diet (FAD) (Zou, Reference Zou2013; Zou et al., Reference Zou, Coudron, Wu, Gu, Xu, Zhang and Chen2015) and reformulated diet (RFD) for six generations.
Eggs or nymphs per treatment was weighted individually at 2 d after egg deposition for eggs and 2 d after eclosion for nymphs. Values are mean ± SE. Asterisks identified the significant difference in weight between FAD and RFD in the same generation (P < 0.05).
The 1st-instar nymphs weighed significantly higher for RFD-fed A. chinensis than of FAD-fed A. chinensis in F2 (dF = 98; t = −8.17 and P < 0.001). However, there was no significant difference in weight for 1st-instar nymphs between RFD-fed nymphs and FAD-fed nymphs in other generations (F3, dF = 98; t = 0.99 and P = 0.326; F4, dF = 98; t = −0.47 and P = 0.643; F5, dF = 98; t = 0.84 and P = 0.406; F6, dF = 98; t = 0.27 and P = 0.785) (table 4).
The 2nd-instar nymphs weighed significantly higher for RFD-fed A. chinensis than of FAD-fed A. chinensis in F1 and F2 (F1, dF = 98; t = −5.47 and P < 0.001; F2, dF = 98; t = −12.21 and P < 0.001). However, for F3, F4, F5 and F6, the 2nd-instar nymphs weighed significantly lower for RFD-fed A. chinensis than of FAD-fed A. chinensis (F3, dF = 98; t = 7.44 and P < 0.001; F4, dF = 98; t = 11.12 and P < 0.001; F5, dF = 98; t = 9.89 and P < 0.001; F6, dF = 98; t = 9.09 and P < 0.001) (table 4).
The 3rd-instar nymphs weighed significantly higher for RFD-fed A. chinensis than of FAD-fed A. chinensis in F1 (dF = 98; t = −5.39 and P < 0.001). However, the 3rd-instar nymphs weighed significantly lower for RFD-fed A. chinensis than of FAD-fed A. chinensis in F2, F3, F4, F5, and F6 (F2, dF = 98; t = 7.73 and P < 0.001; F3, dF = 98; t = 10.66 and P < 0.001; F4, dF = 98; t = 8.33 and P < 0.001; F5, dF = 98; t = 13.80 and P < 0.001; F6, dF = 98; t = 17.82 and P < 0.001) (table 4).
The 4th-instar nymphs weighed significantly higher for RFD-fed A. chinensis than of FAD-fed A. chinensis in F1 (dF = 98; t = −3.53 and P < 0.001). For F3, the 4th-instar nymphs were heavier for RFD-fed A. chinensis than of FAD-fed A. chinensis, but there was no significant difference between them (dF = 98; t = −0.86 and P = 0.395). The 4th-instar nymphs weighed significantly lower for RFD-fed A. chinensis than of FAD-fed A. chinensis in F2, F4, F5, and F6 (F2, dF = 98; t = 4.48 and P < 0.001; F4, dF = 98; t = 16.48 and P < 0.001; F5, dF = 98; t = 20.14 and P < 0.001; F6, dF = 98; t = 22.10 and P < 0.001) (table 4).
The 5th-instar nymphs weighed significantly lower for RFD-fed A. chinensis than of FAD-fed A. chinensis in F1, F3, F4, F5 and F6 (F1, dF = 98; t = 8.58 and P < 0.001; F3, dF = 98; t = 5.04 and P < 0.001; F4, dF = 98; t = 15.29 and P < 0.001; F5, dF = 98; t = 12.58 and P < 0.001; F6, dF = 98; t = 14.93 and P < 0.001). However, there was no significant difference in weight between RFD-fed 5th-instar nymphs and FAD-fed 5th-instar nymphs in F2 (dF = 98; t = 1.51 and P = 0.135) (table 4).
The female body weights were significantly higher for FAD-fed A. chinensis than of RFD-fed A. chinensis in each generation (F1, dF = 98; t = 11.50 and P < 0.001; F2, dF = 98; t = 3.90 and P < 0.001; F3, dF = 98; t = 4.64 and P < 0.001; F4, dF = 98; t = 9.49 and P < 0.001; F5, dF = 98; t = 8.39 and P < 0.001; F6, dF = 98; t = 4.69 and P < 0.001). The male body weights were significantly higher for FAD-fed A. chinensis than of RFD-fed A. chinensis in each generation except for F2 (F1, dF = 98; t = 8.83 and P < 0.001; F3, dF = 98; t = 2.85 and P = 0.005; F4, dF = 98; t = 6.41 and P < 0.001; F5, dF = 98; t = 9.38 and P < 0.001; F6, dF = 98; t = 5.68 and P < 0.001). For F2, there was no significant difference in male body weight between RFD-fed A. chinensis and FAD-fed A. chinensis (dF = 98; t = −0.71 and P = 0.480). Adult females and males reared on RFD weighed an average of 8.09 to 19.62 mg and 4.75 to 11.29 mg less than those reared on FAD (table 6).
Length
There was no significant difference in egg length between RFD-fed and FAD-fed A. chinensis in F2 and F3 (F2, dF = 98; t = 1.31 and P = 0.192; F3, dF = 98; t = 1.79 and P = 0.076). Egg lengths were significantly shorter for RFD-fed A. chinensis than of FAD-fed A. chinensis in F4, F5 and F6. These differences, although statistically significant, were small (F4, dF = 98; t = 5.31 and P < 0.001; F5, dF = 98; t = 5.03 and P < 0.001; F6, dF = 98; t = 6.30 and P < 0.001) (table 5).
Egg and nymphal body length (mm) of Arma chinensis reared on a former artificial diet (FAD) (Zou, Reference Zou2013; Zou et al., Reference Zou, Coudron, Wu, Gu, Xu, Zhang and Chen2015) and reformulated diet (RFD) for six generations.
Egg and nymphal body length values were measured individually at 2 d after egg deposition for eggs and 2 d after eclosion for nymphs. Values are mean ± SE. Asterisks identified the significant difference in body length between FAD and RFD in the same generation (P < 0.05).
When fed RFD, the body length of F2 A. chinensis 1st-instar nymphs was significantly shorter than those fed FAD (dF = 98; t = 3.42 and P < 0.001). However, the body length of 1st-instar nymphs was significantly longer for RFD-fed A. chinensis than those fed FAD in F3 (dF = 98; t = −2.30 and P = 0.023). There was no significant difference in body length for 1st-instar nymphs between RFD-fed nymphs and FAD-fed nymphs in other generations (F4, dF = 98; t = 0.71 and P = 0.482; F5, dF = 98; t = 0.62 and P = 0.537; F6, dF = 98; t = 0.77 and P = 0.445) (table 5).
For 2nd instar nymphs, there was no significant difference in body length between RFD-fed nymphs and FAD-fed nymphs in the first four generations (F1, dF = 98; t = −0.96 and P = 0.341; F2, dF = 98; t = −0.34 and P = 0.736; F3, dF = 98; t = 0.55 and P = 0.582; F4, dF = 98; t = 1.39 and P = 0.168). Body lengths of 2nd instar nymphs were significantly shorter for RFD-fed A. chinensis than of FAD-fed A. chinensis in F5 and F6. These differences, although statistically significant, were small (F5, dF = 98; t = 2.02 and P = 0.046; F6, dF = 98; t = 2.63 and P = 0.010) (table 5).
The RFD-fed 3rd-instar nymphs were a little shorter than those fed FAD in F1, but there was no significant difference between them (dF = 98; t = 1.41 and P = 0.162). However, the 3rd-instar nymphs were significantly shorter for RFD-fed A. chinensis than of FAD-fed A. chinensis in other generations (F2, dF = 98; t = 8.79 and P < 0.001; F3, dF = 98; t = 8.24 and P < 0.001; F4, dF = 98; t = 4.17 and P < 0.001; F5, dF = 98; t = 11.14 and P < 0.001; F6, dF = 98; t = 7.68 and P < 0.001) (table 5).
For F2, the 4th-instar nymphs were longer for RFD-fed A. chinensis than of FAD-fed A. chinensis, but there was no significant difference between them (dF = 98; t = −1.54 and P = 0.127). The RFD-fed 4th-instar nymphs were significantly longer than those fed FAD in F3 (dF = 98; t = −2.93 and P = 0.004). However, the 4th-instar nymphs were significantly shorter for RFD-fed A. chinensis than of FAD-fed A. chinensis in F1, F4, F5 and F6, (F1, dF = 98; t = 4.55 and P < 0.001; F4, dF = 98; t = 18.35 and P < 0.001; F5, dF = 98; t = 32.35 and P < 0.001; F6, dF = 98; t = 19.78 and P < 0.001) (table 5).
The RFD-fed 5th-instar nymphs were a little shorter than those fed FAD, but there was no significant difference between them in F2 (dF = 98; t = 1.85 and P = 0.067). For other generations, the RFD-fed 5th-instar nymphs were significantly shorter than those fed FAD (F1, dF = 98; t = 11.00 and P < 0.001; F3, dF = 98; t = 5.51 and P < 0.001; F4, dF = 98; t = 31.46 and P < 0.001; F5, dF = 98; t = 10.52 and P < 0.001; F6, dF = 98; t = 13.24 and P < 0.001) (table 5).
The body length of RFD-fed female was significantly shorter than those fed FAD for each generation in a manner similar to the female body weight (F1, dF = 98; t = 6.50 and P < 0.001; F2, dF = 98; t = 8.84 and P < 0.001; F3, dF = 98; t = 12.43 and P < 0.001; F4, dF = 98; t = 16.17 and P < 0.001; F5, dF = 98; t = 18.62 and P < 0.001; F6, dF = 98; t = 17.79 and P < 0.001). The body length of RFD-fed male was significantly shorter than those fed FAD for each generation (F1, dF = 98; t = 18.46 and P < 0.001; F2, dF = 98; t = 11.80 and P < 0.001; F3, dF = 98; t = 11.31 and P < 0.001; F4, dF = 98; t = 16.06 and P < 0.001; F5, dF = 98; t = 23.19 and P < 0.001; F6, dF = 98; t = 10.88 and P < 0.001). Adult females and males reared on RFD were 0.52 to 1.22 mm and 0.82 to 1.41 mm shorter on average than those reared on FAD (table 6).
Adult weight (mg), body length (mm), preoviposition (days), total fecundity and longevity (days) of Arma chinensis reared on a former artificial diet (FAD) (Zou, Reference Zou2013; Zou et al., Reference Zou, Wu, Coudron, Zhang, Wang, Liu and Chen2013a, Reference Zou, Coudron, Wu, Gu, Xu, Zhang and Chen2015) and reformulated diet (RFD) for six generations.
Values are mean ± SE. Asterisks identified the significant difference between FAD and RFD in the same generation (P < 0.05).
a Measured 2 days after emergence of adults.
b Adults died naturally.
c Average eggs per fertile female.
Survival
The survival from 1st to 2nd-instar nymphs of F2, F3, and F6 for RFD-fed A. chinensis was significantly lower than those of nymphs reared on FAD (F2, dF = 1; χ 2 = 13.67 and P < 0.001; F3, dF = 1; χ 2 = 5.02 and P = 0.025; F6, dF = 1; χ 2 = 4.84 and P = 0.028). These differences, although statistically significant, were small. There was no significant difference in survival from 1st to 2nd-instar nymphs of F4 and F5 between RFD and FAD treatments (F4, dF = 1; χ 2 = 1.39 and P = 0.238; F5, dF = 1; χ 2 = 2.43 and P = 0.119) (table 7).
Egg viability, survival and sex ratio of Arma chinensis reared on a former artificial diet (FAD) (Zou et al., Reference Zou, Wu, Coudron, Zhang, Wang, Liu and Chen2013a, Reference Zou, Coudron, Wu, Gu, Xu, Zhang and Chen2015) and reformulated diet (RFD) for six generations.
Asterisks identified the significant difference between FAD and RFD in the same generation (P < 0.05).
a Proportion of eggs not cannibalized that successfully hatched.
b Proportion of 1st instars developing to 2nd instars.
c Proportion of 2nd instars developing to adults.
There was no significant difference in the survival from 2nd instar to adult in F1 and F3 for RFD-fed and FAD-fed treatments (F1, dF = 1; χ 2 = 0.48 and P = 0.488; F3, dF = 1; χ 2 = 0.01 and P = 0.906). The survival from 2nd instar to adult of F2, F4, F5 and F6 for RFD-fed A. chinensis was significantly lower than for those reared on FAD (F2, dF = 1; χ 2 = 7.89 and P = 0.005; F4, dF = 1; χ 2 = 18.71 and P < 0.001; F5, dF = 1; χ 2 = 8.47 and P = 0.004; F6, dF = 1; χ 2 = 13.45 and P < 0.001) (table 7).
Cannibalism did not occur in nymphs because of isolation. However, adults reared on both FAD and RFD in all generations were observed cannibalizing eggs. There were significantly more eggs from RFD-fed A. chinensis that were cannibalized by adults than those from FAD-fed in F2 (dF = 1; χ 2 = 596.37 and P < 0.001). The egg cannibalism decreased from F3 to F6 in RFD-fed treatments and was significant lower for RFD-fed A. chinensis than those fed FAD in F4, F5 and F6 (F4, dF = 1; χ 2 = 7.77 and P = 0.005; F5, dF = 1; χ 2 = 19.92 and P < 0.001; F6, dF = 1; χ 2 = 20.20 and P < 0.001). However, there was no significant difference in egg cannibalism between RFD-fed A. chinensis and FAD-fed A. chinensis in F3 (dF = 1; χ 2 = 1.21 and P = 0.272) (table 8).
Cannibalism of eggs and adults of Arma chinensis reared on a former artificial diet (FAD) (Zou et al., Reference Zou, Wu, Coudron, Zhang, Wang, Liu and Chen2013a, Reference Zou, Coudron, Wu, Gu, Xu, Zhang and Chen2015) and reformulated diet (RFD) for six generations.
Asterisks identified the significant difference between FAD and RFD in the same generation (P < 0.05).
a Proportion of females that oviposited. Dead males were replaced with virgin males of the same age.
b Proportion of eggs cannibalized.
c Total number of adult pairs.
d Proportion of females cannibalized in the total number of adult pairs (n c), verified by the empty corpse of the females.
e Proportion of males cannibalized in the total number of adult pairs (n c), verified by the empty corpse of the males.
There was no significant difference in the proportion of cannibalized females between RFD-fed adults and FAD-fed adults in F1, F3, F4, F5, and F6 (F1, dF = 1; χ 2 = 1.50 and P = 0.221; F3, dF = 1; χ 2 = 0.03 and P = 0.869; F4, dF = 1; χ 2 = 1.39 and P = 0.239; F5, dF = 1; χ 2 = 1.95 and P = 0.163; F6, dF = 1; χ 2 = 0.13 and P = 0.720). However, there were less cannibalized females in RFD-fed than in FAD-fed insects in F2 (dF = 1; χ 2 = 21.99 and P < 0.001). A similar situation occurred in males of all groups. There was no significant difference in the proportion of cannibalized males between RFD-fed adults and FAD-fed adults in F1, F3, F4, and F6 (F1, dF = 1; χ 2 = 0.68 and P = 0.411; F3, dF = 1; χ 2 = 0.19 and P = 0.663; F4, dF = 1; χ 2 = 1.00 and P = 0.318; F6, dF = 1; χ 2 = 1.01 and P = 0.316). However, there were less cannibalized males in RFD-fed than in FAD-fed insects in F2 and F5 (F2, dF = 1; χ 2 = 24.84 and P < 0.001; F5, dF = 1; χ 2 = 4.83 and P = 0.028) (table 8).
The male adults lived longer than female adults in all of the treatments. The longevity of female adults reared on RFD was significantly shorter than for those reared on FAD in F2, F3, F4, F5 and F6 (F2, dF = 98; t = 6.85 and P < 0.001; F3, dF = 98; t = 5.14 and P < 0.001; F4, dF = 98; t = 7.59 and P < 0.001; F5, dF = 98; t = 7.66 and P < 0.001; F6, dF = 98; t = 5.40 and P < 0.001). RFD-fed females lived approximate 7.42–14.36 days shorter, respectively, than those reared on FAD. However, there was no significant difference in female longevity between RFD-fed adults and FAD-fed adults in F1 (dF = 98; t = 1.32 and P = 0.190). The longevity of male adults reared on RFD was significantly shorter than for those reared on FAD in F2, F3, F4, and F6 (F2, dF = 98; t = 2.16 and P = 0.033; F3, dF = 98; t = 4.02 and P < 0.001; F4, dF = 98; t = 7.36 and P < 0.001; F6, dF = 98; t = 3.39 and P = 0.001). RFD-fed males lived approximate 7.20–15.94 days shorter, than those reared on FAD. However, there were no significant differences in male longevity between RFD-fed adults and FAD-fed adults in F1 and F5 (F1, dF = 98; t = 0.57 and P = 0.570; F5, dF = 98; t = 0.65 and P = 0.518) (table 6).
Reproductive capacity
The difference in total fecundity between RFD-fed and FAD-fed treatment in F1, F3, F4, F5, and F6 was highly significant (F1, dF = 98; t = 5.93 and P < 0.001; F3, dF = 98; t = 3.90 and P < 0.001; F4, dF = 98; t = 8.05 and P < 0.001; F5, dF = 98; t = 9.40 and P < 0.001; F6, dF = 98; t = 9.39 and P = 0.001). RFD-fed females laid significantly less eggs than those reared on FAD. For F2, there was no significant difference in total fecundity between RFD-fed and FAD-fed treatment (dF = 98; t = 0.49 and P = 0.625) (table 6).
The viability of F2, F5, and F6 eggs from RFD-fed females was significantly higher than those of females reared on FAD (F2, dF = 1; χ 2 = 18.86 and P < 0.001; F5, dF = 1; χ 2 = 18.31 and P < 0.001; F6, dF = 1; χ 2 = 4.96 and P = 0.026). However, the egg viability of F3 from RFD-fed females was significantly lower than those of females reared on FAD (dF = 1; χ 2 = 59.93 and P < 0.001) and there was no significant difference in egg viability of F4 between RFD-fed A. chinensis and FAD-fed A. chinensis (dF = 1; χ 2 = 0.05 and P = 0.826). Further, the viability decreased for eggs from RFD-fed females over F3 to F6 generations (table 7).
There was no significant difference in the proportion of fertile females (i.e., females that oviposited eggs) between RFD-fed females and FAD-fed females in F1, F2, and F3 (F1, dF = 1; χ 2 = 0.16 and P = 0.691; F2, dF = 1; χ 2 = 1.49 and P = 0.222; F3, dF = 1; χ 2 = 1.74 and P = 0.187). The proportions of fertile females decreased from F4 to F6, and there was significant difference in the proportion of fertile females between RFD-fed females and FAD-fed females (F4, dF = 1; χ 2 = 9.66 and P = 0.002; F5, dF = 1; χ 2 = 8.01 and P = 0.005; F6, dF = 1; χ 2 = 8.75 and P = 0.003) (table 8).
The food source impacted the sex ratio (♂ : ♀). There were significantly more females than males in the RFD-fed F1 than those reared on FAD (dF = 1; χ 2 = 8.03 and P = 0.005). However, there was no significant difference in sex ratio between RFD-fed and FAD-fed treatments from F2 to F6 (F2, dF = 1; χ 2 = 0.09 and P = 0.768; F3, dF = 1; χ 2 = 0.75 and P = 0.386; F4, dF = 1; χ 2 = 0.03 and P = 0.870; F5, dF = 1; χ 2 = 0.03 and P = 0.870; F6, dF = 1; χ 2 = 1.04 and P = 0.309) (table 7).
Discussion
From a perspective of nutrigenomics, nutrients act as dietary signals that are detected by cellular sensors and influence gene and protein expression and, subsequently, metabolite production (Müller & Kersten, Reference Müller and Kersten2003). The mechanism of action of nutrients is strongly related to their capacity to modulate gene expression. However, in terms of gene expression-based biomarker development, progress to date has been limited. Although a number of potential gene expression-based nutrient sensitive biomarkers have been identified, these suffer from potential confounding effects that undermine their value. This reflects a fundamental problem with the specificity of single genes as biomarkers since expression of most individual genes are regulated by more than one environmental factor. Expression profile ‘signatures,’ defined as the characteristic patterns of differential gene expression, can help to overcome this problem and be used to measure responses to different levels of nutrients (Elliott, Reference Elliott2008).
Transcriptomes in responses to different foods had been analyzed in Chrysomya megacephala (Fabricius) (Diptera: Calliphoridae) (Zhang et al., Reference Zhang, Yu, Yang, Song, Hu and Zhang2013), Locusta migratoria (Linnaeus) (Orthoptera: Acrididae) (Spit et al., Reference Spit, Zels, Dillen, Holtof, Wynant and Broeck2014), Coleomegilla maculata (DeGeer) (Coccinellidae: Coleoptera) (Allen, Reference Allen2015) and Spodoptera spp (Lepidoptera: Noctuidae) (Roy et al., Reference Roy, Walker, Vogel, Chattington, Larsson, Anderson, Heckel and Schlyter2016). However, an effort to reformulate diets according to transcriptome data have not been reported.
We report prominent performance improvements over six generations that resulted from the formulation changes based on transcriptome data. Those included increased adult body weight, shortened preoviposition period, and higher egg viability. Other improvements, such as shortened developmental time and increased weight during nymphal stages, were intermittent and interspersed with no change or a comparable decrease in performance within a generation and across generations. The most prominent decline in performance over six generations was lowered fecundity and survival from 2nd instar to adults.
The improvements we observed are encouraging. However, overall the formulation changes did not result in the level of improvement we want to achieve. If we accept the premise that transcriptome information accurately identifies nutritional shortfalls then it remains possible that changes we made may not have been extensive enough (e.g., requires more reduction of sugar and lipid, etc.) or that a more complex design may be needed to maximize the ingredients (Lapointe et al., Reference Lapointe, Evens and Niedz2008) identified by the transcriptome information. It is also possible that Ecuadorian shrimp was not a good substitute for tuna fish.
Another possibility is that transcriptome results provide partial information about nutritional deficiencies but lack important post-transcriptional information that would be revealed via proteome analysis (Greenbaum et al., Reference Greenbaum, Colangelo, Williams and Gerstein2003; de Godoy et al., Reference de Godoy, Olsen, Cox, Nielsen, Hubner, Fröhlich, Walther and Mann2008; Maier et al., Reference Maier, Güell and Serrano2009). Thus, expression profile ‘signatures’, differences in protein expression and metabolic profiling in diet-fed and prey-fed A. chinensis may provide valuable dietary insight that could complement, as well as extend, the transcriptome information and thereby aid in directing formulation improvements. These will be explored in future studies.
Acknowledgements
The authors thank all reviewers for their useful comments. Support for this project was provided by National Natural Science Foundation of China (31401806), Natural Science Foundation of Tianjin (16JCZDJC33600), National Key R&D Program of China (2017YFD0201000) and the USDA Agricultural Research Service project Insect Biotechnology Products for Pest Control and Emerging Needs in Agriculture (5070-22000-037-00-D). USDA is an equal opportunity provider and employer. Mention of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the USDA.
