**Gene Characterization and Enzymatic Activities Related to Trehalose Metabolism of In Vitro Reared** *Trichogramma dendrolimi* **Matsumura (Hymenoptera: Trichogrammatidae) under Sustained Cold Stress**

#### **Xin Lü \*, Shi-chou Han, Zhi-gang Li, Li-ying Li and Jun Li \***

Guangdong Key Laboratory of Animal Conservation and Resource Utilization, Guangdong Public Laboratory of Wild Animal Conservation and Utilization, Institute of Zoology, Guangdong Academy of Sciences, 105 Xingang Road West, Guangzhou 510260, China; hansc@giabr.gd.cn (S.-c.H.); leegdei@163.com (Z.-g.L.); liyingl32@163.com (L.-y.L.) **\*** Correspondence: greenhopelv@163.com (X.L.); junl@giabr.gd.cn (J.L.)

Received: 23 September 2020; Accepted: 4 November 2020; Published: 7 November 2020 -

**Simple Summary:** Trehalose is a non-reducing disaccharide that presents in a wide variety of organisms, where it serves as an energy source or stress protectant. Trehalose is the most characteristic sugar of insect hemolymph and plays a crucial role in the regulation of insect growth and development. *Trichogramma* species are economically important egg parasitoids, which are being mass-produced for biological control programs worldwide. Many *Trichogramma* species could be mass reared on artificial mediums (not insect eggs), in which components contain insect hemolymph and trehalose. These in vitro-reared parasitoid wasps were strongly affected by cold storage, but prepupae could be successfully stored at 13 ◦C for up to 4 weeks. The aims of the present study were to determine the role of trehalose and the relationship between trehalose and egg parasitoid stress resistance. Our study revealed that (1) trehalose regulated the growth under sustained cold stress; (2) prepupal stage is a critical developmental period and 13 ◦C is the cold tolerance threshold temperature; (3) in vitro reared *Trichogramma dendrolimi* could be reared at temperatures of 16 ◦C, 20 ◦C, and 23 ◦C to reduce rearing costs. This finding identifies a low cost, prolonged development rearing method for *T. dendrolimi*, which will facilitate improved mass rearing methods for biocontrol.

**Abstract:** *Trichogramma* spp. is an important egg parasitoid wasp for biocontrol of agriculture and forestry insect pests. Trehalose serves as an energy source or stress protectant for insects. To study the potential role of trehalose in cold resistance on an egg parasitoid, cDNA for trehalose-6-phosphate synthase (TPS) and soluble trehalase (TRE) from *Trichogramma dendrolimi* were cloned and characterized. Gene expressions and enzyme activities of *TdTPS* and *TdTRE* were determined in larvae, prepupae, pupae, and adults at sustained low temperatures, 13 ◦C and 16 ◦C. *TdTPS* and *TdTRE* expressions had similar patterns with higher levels in prepupae at 13 ◦C and 16 ◦C. *TdTPS* enzyme activities increased with a decrease of temperature, and *TdTRE* activity in prepupae decreased sharply at these two low temperatures. In vitro reared *T. dendrolimi* could complete entire development above 13 ◦C, and the development period was prolonged without cold injury. Results indicated trehalose might regulate growth and the metabolic process of cold tolerance. Moreover, 13 ◦C is the cold tolerance threshold temperature and the prepupal stage is a critical developmental period for in vitro reared *T. dendrolimi*. These findings identify a low cost, prolonged development rearing method, and the cold tolerance for *T. dendrolimi*, which will facilitate improved mass rearing methods for biocontrol.

**Keywords:** trehalase; trehalose metabolism; in vitro rearing; cold stress; *Trichogramma*

#### **1. Introduction**

Sugars are used for energy production and are stored as glycogen in the body fat, or as trehalose in the hemolymph [1–3]. *Trichogramma* are egg parasitoid wasps that obtain diverse nutrients, including sugars, from their host eggs during development. *Trichogramma dendrolimi* Matsumura is an important biological control agent that has been mass produced on eggs of *Corcyra cephalonica* (Stainton) and *Antheraea pernyi* (Guérin-Méneville) for biological control programs in China [4,5]. Artificial host eggs are now used to mass produce *T. dendrolimi* [6,7]. Lü et al. [8–10] developed an artificial medium containing trehalose for the continuous rearing of *T. dendrolimi* and revealed that trehalose was an essential ingredient of the artificial media. Biochemical characteristics, including trehalose content and trehalase activity in *T. dendrolimi,* continuously reared on artificial medium (in vitro) versus those reared on *A. pernyi* eggs (in vivo), were also studied [11]. The quality of in vitro reared *T. dendrolimi* was strongly affected by cold storage, but prepupae could be successfully stored at 13 ◦C for up to 4 weeks [12].

The developmental temperature threshold can vary, not only among insects, but also among populations [13,14]. *Trichogramma* spp. show different reactions to low temperatures when reared on different hosts or media [12,13]. For *T. dendrolimi*, the developmental threshold temperature was different among geographical populations and hosts: 10.34 ◦C for south population reared on *Philosamia cynthia ricini* [13]; 10.1 ◦C for the south population reared on *C. cephalonica* [14]; 5.34 ◦C/5.1 ◦C, 5.82 ◦C/5.42 ◦C, 11.03 ◦C/14.83 ◦C, and 12.37 ◦C/11.58 ◦C for the north population (egg, larva, prepupa and pupa) reared on *A. pernyi* [15,16]. The north population reared on *A. pernyi* was unable to complete the entire development (stopped developing at prepupal stage) at 10 ◦C, but was able to complete the entire development above 15 ◦C [15]. Compared with the in vivo (on *A. pernyi*) reared *Trichogramma*, in vitro (on artificial medium) reared *Trichogramma* of the south population was more affected by cold storage [12]. All of these experimental populations have been lab adapted. Limited information is available about inducing cold stress on the lab adapted strain of this particular insect species, and molecular mechanisms of trehalose metabolism, or the relationship between trehalose and egg parasitoid stress resistance [17–21]. Based on the facile sampling and reproducibility of the condition of in vitro reared *T. dendrolimi*, trehalose metabolism related enzymes were explored by investigating their changes in gene expression and corresponding enzyme activities to study the effect of cold stress conditions on egg parasitoids.

Trehalose is involved in the regulation of parasitoid growth and development. The disaccharide sugar trehalose serves as an energy source or stress protectant for parasitoids [22,23]. It promotes longevity, fecundity [24,25], and cold tolerance [18,26,27], and provides, energy needed to search for, and parasitize, hosts [28]. Trehalose is synthesized by trehalose-6-phosphate synthase (TPS, EC 2.4.1.15) and trehalose-6-phosphate phosphatase (TPP, EC 3.1.3.12) in the body fat, and is hydrolyzed by trehalase (TRE, EC 3.2.1.28) to yield two glucose molecules in the hemolymph [19,20,29–31]. The activity of these three enzymes also affects insect physiology and development. Trehalose and trehalase are closely associated with growth and development throughout insect life cycles [2,23]. However, we should notice the adult yields and quality of in vitro rearing some parasitoids (e.g., the tachinid *Exorista larvarum*) did not drop when the artificial medium without insect material, which trehalose has been replaced with sucrose and sucrose, was even deleted, without drops in adult yields [32,33].

TPS and TPP genes have two functional conserved domains similar to yeast genes and are homologs of yeast Tps1 (Ots A) and Tps2 (Ots B), respectively [34]. In insects, TPS is a fused gene [35] and two exons are involved in encoding trehalose synthetase [36]. Trehalase catalyses, the irreversible hydrolysis of trehalose to glucose, which is the only known pathway of trehalose utilization [2]. TRE is essential for energy metabolism and is important in insect growth and molting [37]. Trehalose may function as a cryoprotectant to stabilize proteins at low temperatures [2,18], and may also protect insects from external interference, assist in successful completion of metamorphosis, and aid survival in adverse environments [38]. In addition, trehalose is important in the regulation of insect growth and development and serves as an energy source and stress protectant [19,38–41].

TPS has been cloned, characterized, and purified from many insect species. It was first cloned from *Drosophila melanogaster* [42]. TRE has been identified in insect species, such as *Rhodnius prolixus*, *Nilaparvata lugens*, *Spodoptera exigua*, *Omphisa fuscidentalis,* and *Harmonia axyridis*[19,30,43–46]. Although TPS and TRE are important key enzymes for many insects, in parasitoid wasps, only the trehalase cDNA from *Pimpla hypochondriaca* has been cloned [47]. No trehalose metabolism genes in *Trichogramma*, or even in any egg parasitoids, have been characterized. To understand the role of trehalose in the cold resistance of in vitro reared *T. dendrolimi*, two genes (*TdTPS* and *TdTRE*) were identified, and cloned the full-length cDNA of *T. dendrolimi* reared on artificial medium using transcriptome data from *T. dendrolimi* reared on eggs of *A. pernyi*. The changes in gene expression and corresponding enzyme activities in four developmental stages (larva, prepupa, pupa, and adult) as each stage developed at low temperatures were also recorded. The trehalose metabolism in the life cycle of *T. dendrolimi* was systematically investigated in relation to cold hardiness.

#### **2. Materials and Methods**

#### *2.1. Insects*

*Trichogramma dendrolimi* were provided by Engineering Research Center of Natural Enemies, Institute of Biological Control, Jilin Agricultural University, Changchun, China. In the laboratory, *T. dendrolimi* stock cultures were reared on eggs of *A. pernyi* as a factitious host. Rearing conditions were 27 ◦C ± 1, 75% ± 5 relative humidity (RH) and a 16:8 h (L:D) photoperiod.

#### *2.2. Preparation of Artificial Medium and Insect Rearing*

The artificial medium used in this study was the modified artificial medium developed by Lü et al. [9]. It comprised 3 mL of the pupal hemolymph of *A. pernyi*, 2.5 mL egg yolk, 1 mL 10% malted milk solution, 1 mL Neisenheimer's salt solution, 0.1 g trehalose (Sigma, St. Louis, MO, USA), and 1.5 mL sterile water. The preparation of artificial egg cards was done as described by Lü et al. [11].

Artificial egg cards were placed in a plastic tray (20 cm × 10 cm × 3 cm) for 24 h exposure to *T. dendrolimi* adults of the same batch. Parasitoids of both sexes were released in the trays using a 6:1 ratio of parasitoids to artificial eggs. Sex ratios were approximately 8:1 (female:male) in all three replicates (one tray corresponded to one replicate). Trays were placed in climatic incubators (Yamato, Tokyo, Japan) set at 27 ◦C ± 1, 75% ± 5 RH, and a 16:8 h (L:D) photoperiod. After 24 h of exposure, the wasps were removed, and the egg cards were transferred to the temperature treatments [9].

#### *2.3. Experimental Set-Up, Sample Collection, and Biological Parameters Assessment*

In a pre-experiment, in vitro reared *T. dendrolimi* were unable to complete the entire development (from egg to adult) at 10 ◦C, and stopped developing at prepupal or pupal stages. The optimum storage condition for these parasitoid wasps are prepupae that can be stored at 13 ◦C for up to 4 weeks without affecting reproductive quality [12]. The present experiment had two factors: temperature (13 ◦C ± 1 (optimum storage temperature), 16 ◦C ± 1 (above the optimum storage temperature) and 27 ◦C ± 1 (optimum development temperature)), and developmental stage (larva, prepupa, pupa, and adult). In a preliminary experiment, based on the transparency of the egg cards, they can be easily monitored daily for parasitoid development using a binocular microscope. *T. dendrolimi* developing to new larvae, prepupae, pupae, or adults were collected on ice and maintained at −80 ◦C for gene expression and biochemical assessments.

In the follow-up rearing experiment, to investigate the developmental quality of *T. dendrolimi* reared on artificial medium at different temperatures (27 ◦C ± 1, 23 ◦C ± 1, 20 ◦C ± 1, 16 ◦C ± 1, and 13 ◦C ± 1), the developmental durations (from oviposition to adult emergence) of the eggs, larvae, prepupae, and pupae, number of male adults and total adults observed per egg card were examined. Pupation rate, adult emergence rate (based on pupal numbers), numbers of normal adults produced

(i.e., adults not having an enlarged abdomen and/or unexpanded wings) and the adult sex ratio (female proportion) was calculated as follows:


#### *2.4. Total RNA Extraction and Cloning of the Full-Length cDNA*

Total RNA was extracted from *T. dendrolimi* adults using a TransZol Up Plus RNA Kit (TransGen, Beijing, China). The RNA integrity and concentration were checked by agarose gel electrophoresis and spectrophotometry (NanoDrop2000, Wilmington, DE, USA), respectively. The fragments of *TdTPS* and *TdTRE* were obtained by transcriptome sequencing of *T. dendrolimi* (Hiseq 2000, Illumina, Beijing, China). Full-length sequences of TPS and TRE were obtained by 5 and 3 rapid amplification of cDNA ends (RACE) with the SMARTTM RACE Kit (TaKaRa, Tokyo, Japan), according to manufacturer instructions. 5 and 3 RACE were performed by nested PCR including Universal Primer Mix (UPM) and Nested Universal Primer (NUP) along with gene-specific primers (GSP) (Table 1). The PCR conditions were as follows: initial at 94 ◦C for 5 min, 32 cycles of 30 s at 94 ◦C, 30 s at 60 ◦C, 2 min at 72 ◦C, and a final extension at 72 ◦C for 10 min. The products were examined by agarose gel electrophoresis, purified using a SanPrep Column DNA Gel Extraction Kit (Sangon, Shanghai, China), ligated into a Pucm-T vector (Sangon, Shanghai, China), and sequenced by Sanger's method.



UPM, Universal Primer Mix; NUP, Nested Universal Primer; F, Forward; R, Reverse.

#### *2.5. Sequence Analysis*

The amino acid sequences of *T. dendrolimi* in the fasta format was used to query the sequence database of the National Center for Biotechnology Information (NCBI) to identify proteins with primary sequence similarity to *TdTPS* and *TdTRE*. Multiple sequence alignment was constructed using MEGA 7 [48] with the CLUSTAL V method [49,50]. Phylogenetic trees were constructed using the neighbor-joining (NJ) method [51]. *D. melanogaster* was used as the out-group, and the stability of the tree was assessed via bootstrapping with >1000 replicates.

#### *2.6. Expression of TdTPS and TdTRE*

Total RNA was extracted from 0.1 g of *T. dendrolimi* at larval, prepupal, pupal, and adult stages using a TransZol Up Plus RNA Kit (TransGen, Beijing, China). First-strand cDNA was synthesized from 1 μg total RNA using a FastQuant RT Kit With gDNase (Tiangen, Beijing, China).

Steps to construct linearized plasmid standards were as described previously with some modifications [52]. First, products *TdTPS* and *TdTRE* were extracted and purified with an agarose gel Extraction Kit (Sangon, Shanghai, China). Second, each gene was cloned separately using the Pucm-T Cloning Vector Kit (Sangon), according to the manufacturer instructions. Third, positive clones screened by PCR were processed for plasmid isolation using a Plasmid Extraction & Purification Kit (Sangon) and confirmed by Sanger sequencing (Sangon). Fourth, plasmids were completely linearized by *EcoR* I digestion for 4.5 h at 37 ◦C and confirmed by checking band patterns in the agarose gel. Fifth, linearized plasmids were quantified using a NanoDrop2000 spectrophotometer (NanoDrop), and copy numbers were calculated for all standards by the following formula [53]:

$$\text{Copies/} \mu\text{L} = \frac{(6.02 \times 10^{23} \text{ copies}) \times (\text{plasma concentration g/} \mu\text{L})}{(\text{number of bases}) \times (660 \text{ dollars}/\text{base})}$$

Finally, the standard DNA (template) was prepared in a dilution series from 10−<sup>3</sup> to 10−<sup>10</sup> (copies/5 μL) for qPCR. qPCR was performed using a SYBR Green Mix Kit (Tiangen, Beijing, China) to measure the cycle number (Ct) of each dilution in duplicate. Each PCR reaction was mixed with 10 μL SYBR Green Mix, 6 μL dd H2O, 2 μL cDNA, and 1 μL of each primer (10 μM). Cycling conditions for all standards were described as above followed with dissociation curve analysis. Standard curves were generated as linear regression between Ct and log10 starting copy number of standard DNA. The Ct values were reported by the MX3000P MXPro program (Agilent Technologies, Palo Alto, USA). Amplification efficiency, slopes, and correlation coefficient (R2) were automatically calculated by the program.

To study the gene expression profiles during the four life stages, absolute quantitative PCR (AQ-PCR) was conducted to estimate their starting copy numbers. Gene specific primers, qTdTPSF/R, and qTdTREF/R, were designed according to the full-length cDNAs and these are listed in Table 1. RT-PCR was performed to obtain gene targets in the following cycling condition: initial denaturation 3 min at 95 ◦C followed by 40 cycles including 5 s at 95 ◦C, 10 s at 55 ◦C, 15 s at 72 ◦C. In the case of the *T. dendrolimi* samples, qPCR Ct values of *TdTPS* and *TdTRE* expression profiles were used to estimate starting copy numbers based on their standard curves.

#### *2.7. Enzyme Activity Measurements*

To obtain crude extracts for enzyme activity study, each sample (0.0100 ± 0.0002 g of larvae, pupae, prepupae, and adults) was homogenized at 0 ◦C (TGrinder OSE-Y20 Homogenizer, Tiangen, Beijing, China) after adding 2000 μL of 20 mM phosphate buffered saline (PBS, pH 5.8). The homogenates were centrifuged at 10,000× *g* at 4 ◦C for 20 min (CP100MX, Hitachi, Tokyo, Japan), and cuticle debris was removed. Supernatants in PBS were maintained at −80 ◦C to analyze TPS and TRE activity.

To determine TPS activity, the method of Dual et al. [54] was used. The qualitative analysis of trehalose was performed by thin layer chromatography (TLC); the quantitative analysis of trehalose synthase activity was investigated by examining the difference in glucose that was formed from the hydrolysis of maltose by TPS in the presence and absence of α-glucosidase and measured by DNS. The reaction mixture, containing 150 μL of crude extract and 100 μL substrate (10% maltose) was inactivated in a water bath at 60 ◦C for 1 h, and then in boiling water bath for 10 min. After adding 85% phosphoric acid to adjust pH to 4.2, 1 mL of diluted maltase (alpha-glucosidase) was added to the mixtures. The mixtures were inactivated in a water bath at 60 ◦C for 1 h and then in a boiling water bath for 10 min. Diluting and volumetrizing the reaction solution, adding 1:2 (V reaction solution: V DNS method), boiling water bath for 10 min, cooled in an ice bath, then the absorbance was measured at 550 nm.

TRE activity was measured as described previously by the 3,5-dinitrosalicylic acid colorimetric method (DNS method) with absorbance measured at 540 nm [55,56]. One unit (U) of enzyme activity was defined as the amount of enzyme capable of releasing 1 mg of reducing sugar per minute. The reaction mixture consisted of 500 μL crude extract and 500 μL DNS solution. The reaction was stopped by heating in boiling water for 5 min, then 4 mL of a pH 5.8 KH2PO4-NAOH buffer solution was added.

#### *2.8. Statistical Analyses*

Each treatment was performed using three biological and three technical replicates. Multifactor analysis (PROC GLM) of variance was conducted to evaluate temperatures and effect of developmental stage on the gene expression and enzyme activities of parasitoids using Tukey's test. The qualities of biological parameters were compared using one-way analysis of variance (ANOVA) and multiple comparisons of means were conducted using Tukey's test. Before analysis, percentage data were arcsine square root−transformed, and the data on the number of adults produced were log10-transformed to fit a normal distribution. For absolute quantification analysis, the number of molecules was expressed as the mean of the log10-converted value ± standard error. In all experiments, differences among means were considered significant at *p* < 0.05. Statistical analyses were conducted using SPSS 22.0 software (SPSS Inc., Chicago, IL, USA).

#### **3. Results**

#### *3.1. Cloning and Characterization of Full-Length TdTPS and TdTRE cDNAs*

To identify the *TdTPS* and *TdTRE*, cDNA fragments involved in trehalose metabolism were identified through *T. dendrolimi* transcriptome data (Hiseq 2000, Illumina, Santiago, USA). Based on the cDNA fragments, specific primers were performed and full-length cDNA of *TdTPS* and *TdTRE* was obtained by RACE-PCR. The full length *TdTPS* gene has 3189 bp, and the cDNA has a 2358 bp open reading frame (ORF), encoding a polypeptide of 785 amino acids with an estimated molecular weight of approximately 88.60 kDa and a pI of 6.56. The ORF was identical to the homolog from *Trichogramma pretiosum* (100%) (Table 2). Sequence analysis showed that the deduced amino acid sequence includes a conserved TPS domain (aa 1–478, E = 1 <sup>×</sup> 10−133) and a TPP domain (aa 515–739, <sup>E</sup> <sup>=</sup> <sup>7</sup> <sup>×</sup> <sup>10</sup><sup>−</sup>34). Multiple protein alignment indicated that the *TdTPS* protein contains six conserved motifs (Figure 1). In addition to these specific motifs, *TdTPS* proteins contain the highly conserved domains, HDYHLML and DGMNLV, the same as TPS genes previously reported, such as in *D. melanogaster*, *Catantops pinguis,* and *Delia antiqua* [20,34,41]. Phylogenetic analysis showed that the *TdTPS* was more closely related to those of other Hymenoptera species (*T. pretiosum*, *Nasonia vitripennis,* and *Copidosoma floridanum*), and could be assigned to the same subgroup (Figure 2). The cloned TPS gene was designated as *TdTPS* and deposited into GenBank (MT108781).

**Figure 1.** Alignment of *TdTPS*. Multiple alignment of trehalose-6-phosphate synthase (TPS) protein sequences from different insect species. Identical amino acid residues are shown in purple. The conserved motifs and the signatures are underlined and boxed, respectively.


**Table 2.** The identities of *TdTPS* and *TdTRE* genes from different insects with *T. dendrolimi.*

**Figure 2.** Phylogenetics of *TdTPS*. Phylogenetic tree constructed using the neighbor-joining (NJ) method. Percentage bootstrap values larger than 40 are shown on each branch. *PcTPS*: *Polistes canadensis*, XP\_014609582; *PdTPS*: *Polistes dominula*, XP\_015172546; *ObTPS*: *Osmia bicornis*, XP\_029055554; *MrTPS*: *Megachile rotundata*, XP\_003702415; *PgTPS*: *Pseudomyrmex gracilis*, XP\_020289281; *CcTPS*: *Cephus cinctus*, XP\_015588847; *OaTPS*: *Orussus abietinus*, XP\_012281922; *NvTPS*: *Nasonia vitripennis*, XP\_016837588; *CfTPS*: *Copidosoma floridanum*, XP\_014213166; *TpTPS*: *Trichogramma pretiosum*, XP\_014221069; *DmTPS*: *Drosophila melanogaster*, ABH06641.1.

The full length *TdTRE* cDNA consisted of 2228 bp, including an 1878 bp open reading frame, encoding 625 amino acids with a predicted molecular weight of 73.2 kDa and a pI of 6.40. Basic

Local Alignment Search Tool (BLAST) analysis revealed that *TdTRE* is 55.95–97.61% identical in structure to other known insect TRE forms. *TdTRE* is also most similar to the TRE from *T. pretiosum* (97.61%) (Table 2). The deduced amino acid sequence of *TdTRE* contains one conserved motif (YYLMRSQPPLLIPM) and a signal peptide sequence (Figure 3). Moreover, *TdTRE* had the same signature motifs, PGGRFREFYYWDSY and QWDYPNAWPP. For phylogenetic analysis, *TdTRE* was clustered with *T. pretiosum* TRE (Figure 4). Based on the sequence identity with known soluble form trehalase genes, *TdTRE* was identified as a soluble trehalase and deposited into GenBank (MT108782).

**Figure 3.** Alignment of *TdTRE*. Multiple alignment of soluble trehalase (TRE) protein sequences from different insect species. Identical amino acid residues are shown in purple. The conserved motifs, the signatures, and signal peptide sequence are underlined, red boxed, and black boxed, respectively.

**Figure 4.** Phylogenetics of *TdTRE.* Phylogenetic tree constructed using the neighbor-joining (NJ) method. Percentage bootstrap values larger than 40 are shown on each branch. *NvTRE*: *Nasonia vitripennis*, XP\_008215783; *TsTRE*: *Trichomalopsis sarcophagae*, OXU30694; *CsTRE*: *Ceratosolen solmsi marchali*, XP\_011497766; *TpTRE*: *Trichogramma pretiosum*, XP\_014236786; *CfTRE*: *Copidosoma floridanum*, XP\_014216724; *OaTRE*: *Orussus abietinus*, XP\_012271873; *HsTRE*: *Harpegnathos saltator*, XP\_011144292; *PgTRE*: *Pseudomyrmex gracilis*, XP\_020280302; *MpTRE*: *Monomorium pharaonis*, XP\_028048276; *SiTRE*: *Solenopsis invicta*, XP\_011170317; *DmTRE*: *Drosophila melanogaster*, NP\_726025.1.

#### *3.2. Standard Curve*

Absolute quantification determines the actual copy numbers of target genes (*TdTPS* and *TdTRE*) by relating the Ct value to a standard curve and amplifying serial dilutions of plasmid standards by qPCR. The Ct values were measured and plotted against known copy numbers of the standard sample. The reaction efficiency and linearity for the serially diluted standards were of good quality for both genes (Figure 5). The standard curve covered a linear range of seven orders of magnitude. The slope (−3.3550 and <sup>−</sup>3.3396) and the correlation coefficient (R<sup>2</sup> = 0.9992 and 0.9990) of the standard curve indicated that this assay could be used to quantify target RNA in *T. dendrolimi*.

**Figure 5.** Standard curve for RT-qPCR amplification of standard sample. (**a**) Amplification plots for *TdTPS* and *TdTRE*; (**b**) standard curves of real-time PCR of *TdTPS* and *TdTRE*, using the method of absolute quantitative analysis, showing the testing in triplicate of a 10-fold dilution series containing a standard sample ranging from 9.97 <sup>×</sup> 106 to 9.97 and 1.16 <sup>×</sup> <sup>10</sup><sup>7</sup> to 1.16 <sup>×</sup> <sup>10</sup><sup>1</sup> copies per.

#### *3.3. E*ff*ect of Temperature on the Expression of TdTPS and TdTRE during Development*

Absolute quantification PCR (AQ-PCR) experiments to measure *TdTPS* and *TdTRE* absolute expression in four developmental stages at different temperatures revealed that there was interaction between two factors on the mRNA levels of the genes (Table 3). Table 4 shows that the levels of the two genes were more highly expressed at the optimum storage temperature (13 ◦C) in all developmental stages compared to the expression at the other treatment temperatures. The *TdTRE* transcripts were highly expressed at 16 ◦C. In contrast, at the optimum development temperature (27 ◦C), the expression levels of *TdTPS* and *TdTRE* were low in all developmental stages. *TdTPS* was highly expressed in prepupae when *T. dendrolimi* developed at the three treatment temperatures, and *TdTPS* expression level was also higher in adults when they developed at 16 ◦C. For trehalase, the absolute expression levels of *TdTRE* were very low in the larval stage at all temperatures, and much higher in the prepupal stage at 27 ◦C and 16 ◦C.


**Table 3.** Multifactor variance analysis of effects of two factors on gene absolute expression and enzymes activity of *Trichogramma dendrolimi* reared on artificial medium.

A, temperature; B, developmental stage.

**Table 4.** The gene expression of *TdTPS* and *TdTRE* of in vitro-reared *Trichogramma dendrolimi* reared at different temperatures and developmental stages.


Mean ± SE were calculated from three replicates. Mean ± SE followed by the same capital letter within a row were not significantly different (Tukey's test: *p* > 0.05); Mean ± SE followed by the same lowercase letter within a column were not significantly different (Tukey's test: *p* > 0.05).

#### *3.4. Changes in Enzyme Activities*

*TdTPS* and *TdTRE* enzyme activities were compared at different temperatures and different developmental stages, respectively. Comparison among temperatures indicated that *TdTPS* activities were similar at normal and cold temperatures. In larval and adult stages, *TdTRE* showed much higher activities at 16 ◦C and 13 ◦C. However, the *TdTRE* activity in prepupae showed the opposite result. In the pupal stage, *TdTRE* activities were similar at the three temperatures (Tables 3 and 5).

At 27 ◦C, *TdTPS* activity in the developmental stages was similar. *TdTRE* had significantly higher activity in the prepupal and pupal stages and then declined in the adult stage. The enzyme activity of *TdTPS* and *TdTRE* had a similar trend at 16 ◦C and 13 ◦C in all *T. dendrolimi* developmental stages. Enzyme activities of *TdTRE* in prepupae declined sharply when they developed at 13 ◦C and 16◦ C.


**Table 5.** Activity of the enzymes involved in trehalose metabolism of in vitro-reared *Trichogramma dendrolimi* reared at different temperatures and developmental stages.

Mean ± SE were calculated from three replicates. Mean ± SE followed by the same capital letter within a row were not significantly different (Tukey's test: *p* > 0.05); Mean ± SE followed by the same lowercase letter within a column were not significantly different (Tukey's test: *p* > 0.05).

#### *3.5. In Vitro Rearing at Di*ff*erent Temperatures*

The developmental durations and biological parameters at the five temperatures are shown in Table 6. The development of *T. dendrolimi* at 13 ◦C, 16 ◦C, 20 ◦C, and 23 ◦C prolonged for 21 days, 17 days, 8 days, and 5 days, respectively compared with those reared at the optimum temperature (27 ◦C). There were no significant differences in pupation rate, emergence rate, female proportion and number of normal adults among 16 ◦C, 20 ◦C, 23 ◦C, and 27 ◦C (*F*4,14 = 4.070, 135.396, 2.430, and 64.434, *p* = 0.033, 0.000, 0.116 and 0.000, respectively). The biological parameters (except female proportion) of *T. dendrolimi* reared on artificial medium were significantly affected by temperature. These parameters at 13 ◦C was lowest compared to that of other test temperatures.

**Table 6.** Developmental quality of in vitro reared *T. dendrolimi* at different temperatures.


Mean ± SE values were calculated from three replicates. Mean ± SE followed by the same lower case letter within a column are not significantly different (Tukey's test: *p* < 0.05).

#### **4. Discussion**

In this study, only the TPS gene of *T. dendrolimi* was obtained. The deduced amino acid sequence reveals that *TdTPS*, similar to the *DaTPS* gene in *Delia antiqua* [20], has two conserved functional domains that include an N-terminal TPS domain and a C-terminal TPP domain [43]. This result supports the conclusion that the insect TPS is a fused gene [57]. There are two signature motifs (HDYHL and DGMNLV) of TPS protein sequences for insects, plants, bacteria, fungi, and nematodes. Multiple protein alignment results show that, besides the signature motifs, *TdTPS* has other conserved motifs. Two types of trehalase exist in insects, a soluble trehalase and a membrane-bound trehalase with a transmembrane domain near the C-terminus [2,29,44,58,59]. Based on the conserved motif and specific signatures in deduced amino acid sequences, we identified the existence of the soluble form of trehalase (*TdTRE*) in in vitro reared *T. dendrolimi*.

Trehalose occurs in all insect species and is the most characteristic sugar in the insect hemolymph [57]. However, it has not been detected in the developmental stages of some species [2,23]. In the present study, *TdTPS* and *TdTRE* were expressed in larval, prepupal, pupal, and adult stages, whether they developed at the optimum development temperature or at sustained low temperatures. The results suggest that sustained low temperature has a strong effect on the expression level of trehalase genes in *T. dendrolimi*, which may facilitate the utilization of trehalose. A comparison of the expression levels of *TdTPS* and *TdTRE* at the tested temperature showed that sustained low temperature can upregulate their expression, but the expression of *TdTRE* was much lower than *TdTPS*. These indicated that the anabolism of trehalose was greater than its catabolism during low temperature development. Meanwhile, the reduction of trehalase activity inhibits the trehalose catabolism, which indirectly helps the trehalose accumulation. *TdTPS* might play a more important role than *TdTRE* in cold induction.

Prepupal stage is a critical developmental period for *Trichogramma*. The nutrient (host egg liquid) has been consumed by the end of larval stage and digestion of this food allows for the accumulation of energy used by the following pupal and adult stages. There were studies that have reported that the prepupa is the best stage for short-term storage of *Trichogramma* spp. reared in vivo [12]. In a previous study, sustained temperatures at 13 ◦C for 4 weeks was the optimum short-term storage condition for prepupae of in vitro reared *T. dendrolimi* [12]. Here, we explain the results of the previous study in terms of molecular physiology. *TdTPS* had a higher expression level in prepupae at the three test temperatures. Meanwhile, *TdTPS* had a high expression level in all stages when *T. dendrolimi* developed at low temperatures, especially at 13 ◦C. The enzymatic activity of *TdTPS* showed a similar change during the entire development. With a decrease of temperature, the activity increased. *TdTRE* showed a high expression level, but the enzymatic activity decreased at low temperatures in the prepupal stage. This trend was opposite to that observed at 27 ◦C. These findings indicate that the gene regulation for *TdTRE*, the soluble trehalase, might not determine the enzyme activity directly. However, there is no evidence to prove whether other soluble trehalase or membrane-bound trehalase genes exist in *T. dendrolimi* yet. Large-scale expression of TPS by insects before pupation promotes the synthesis of trehalose, leading to a high level of trehalose in the pupa stage. Under sustained low temperature, trehalase activity was inhibited in the prepupae since trehalose accumulation was probably required during this period to meet the energy required for chitin synthesis in the pupa and for adult emergence, suggesting its potential role in molting from pre-pupae to pupae. However, *TdTRE* is more active than *TdTPS* during cold stress. In vitro reared *T. dendrolimi* does not diapause under sustained low temperature stress, but cold only prolongs the development period after which development proceeds normally. In addition, the trehalose supplemented in artificial medium [9] needs to be consumed. On the other hand, in vitro reared *T. dendrolimi* could complete development above 13 ◦C because the trehalose added to the artificial medium may improve its cold tolerance. This protected the parasitoids, especially the young larvae, from low temperature injury. Trehalose was synthesized and accumulated at the same time for metabolism and utilization during the period. This indicates that trehalose might regulate growth of in vitro reared *T. dendrolimi* and the metabolic process of cold tolerance.

Trehalose concentration in insect blood hemolymph is not under homeostatic regulation. It is based on environmental conditions, physiological state, and nutrition. *Trichogramma* complete their development in the host egg before adult emergence. During development, the accumulation of trehalose helps the insects to resist environmental temperature stress and the stress of host malnutrition. A comparison of the trehalose contents and trehalase activity of *T. dendrolimi* produced in vitro and in vivo showed that the adults produced in vitro had higher trehalose content and trehalase activity over 10 generations [11]. Therefore, the responses and resistance of egg parasitoids to environmental stress may be different from those of other insects.

Future work will need to focus on the effects of different hosts (nutrition) on trehalose metabolism in *T. dendrolimi*, and determine how the trehalase genes are regulated under stress conditions.

#### **5. Conclusions**

Trehalose synthetase and soluble trehalase genes were identified from *T. dendrolimi* reared on an artificial medium. Sustained low temperature stress had different effects on trehalose metabolism related enzyme genes and enzyme activities of *Trichogramma*. The anabolism and catabolism of trehalose maintained a dynamic balance in the process of metabolism. Trehalose indeed accumulated as an energy source to be used in adverse conditions. *TdTPS* and *TdTRE* may be considered as an energy source and responsive enzymes for cold resistance in *Trichogramma*. The prepupa stage is a key period of *Trichogramma* development, in which the expression of genes involved in trehalose metabolism and corresponding enzyme activities undergo substantial changes. It' is found that 13 ◦C appears to be the cold tolerance threshold temperature for in vitro reared *T. dendrolimi*. Since it remains unknown whether or not the in vitro reared *T. dendrolimi* could diapause without negatively affecting their reproductive parameters, we suggest that in vitro reared *T. dendrolimi* could be reared at temperatures of 16 ◦C, 20 ◦C, and 23 ◦C to reduce rearing costs.

**Author Contributions:** Conceptualization, X.L., J.L., and L.-y.L.; methodology, X.L.; formal analysis, X.L.; resources, S.-c.H. and Z.-g.L.; data curation, X.L.; writing—original draft preparation, X.L.; writing—review and editing, J.L., S.-c.H., and L.-y.L.; visualization, X.L. and Z.-g.L.; supervision, L.-y.L. and J.L.; project administration, X.L. and J.L.; funding acquisition, X.L., J.L., S.-c.H., and Z.-g.L. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the Key-Area Research and Development Program of Guangdong Province (grant no. 2020B020223004), GDAS Special Project of Science and Technology Development (grant nos. 2020GDASYL-20200301003, 2020GDASYL-20200104025, and 2018GDASCX−0107), National Natural Science Foundation of China (grant no. 31501702), Science & Technology Planning Project of Guangdong (grant no. 2019B030316018).

**Acknowledgments:** We would like to thank Patrick De Clercq provided guidance; Chuan-bu Gao prepared the figures; Jun-gang Qian assisted with the experiment.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

© 2020 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Article* **Effect of Prey Species and Prey Densities on the Performance of Adult** *Coenosia attenuata*

**Deyu Zou 1, Thomas A. Coudron 2, Lisheng Zhang 3, Weihong Xu 1, Jingyang Xu 1, Mengqing Wang 3, Xuezhuang Xiao <sup>4</sup> and Huihui Wu 4,\***


**Simple Summary:** The predaceous fly *Coenosia attenuata* Stein has received attention because of its ability to effectively suppress a wide range of agricultural pests, such as fungus gnats, whiteflies and leaf miners. An effective level of control requires large numbers of *C. attenuata* to be available at low cost for release. Adult fungus gnats and drosophilids are now the main prey used to rear *C. attenuata* adults. However, previous studies showed *C. attenuata* fertility is lower when fed drosophilids compared to fungus gnats. The current study investigated the performance of *C. attenuata* adults when reared on different densities of adult *Drosophila melanogaster* Meigen or *Bradysia impatiens* (Johannsem). Results showed that the optimal prey density in the mass rearing of adult *C. attenuata* was 12–24 adult *B. impatiens* daily per predator. Additionally, *C. attenuata* adults suffered more wing damage, at some of the prey densities, when reared on *D. melanogaster* compared to *B. impatiens*. This information will be used to optimize rearing methods and decrease the cost of mass rearing in *C. attenuata*.

**Abstract:** Mass production of *Coenosia attenuata* Stein at low cost is very important for their use as a biological control agent. The present study reports the performance of *C. attenuata* adults when reared on *Drosophila melanogaster* Meigen or *Bradysia impatiens* (Johannsem). Different densities (6, 9, 15, 24 and 36 adults per predator) of *D. melanogaster* or (6, 12, 24, 36 and 48 adults per predator) of *B. impatiens* were used at 26 ± 1 ◦C, 14:10 (L:D) and 70 ± 5% RH. The results concluded that *C. attenuata* adults had higher fecundity, longer longevity and less wing damage when reared on *B. impatiens* adults compared to *D. melanogaster* adults. Additionally, *C. attenuata* adults demonstrated greater difficulty catching and carrying heavier *D. melanogaster* adults than lighter *B. impatiens* adults. In this case, 12 to 24 adults of *B. impatiens* daily per predator were considered optimal prey density in the mass rearing of adult *C. attenuata*.

**Keywords:** *Coenosia attenuata*; mass rearing; wing damage; *Bradysia impatiens*; *Drosophila melanogaster*; fecundity

#### **1. Introduction**

The predaceous fly *Coenosia attenuata* Stein (Diptera: Muscidae), also known as "tiger fly", "killer fly" or "hunter fly" [1–4], is native to Southern Europe [4,5] and has been reported to have spontaneously colonized a number of crops outdoors and in greenhouses in many countries worldwide [5–21]. It has received attention because of its ability to

**Citation:** Zou, D.; Coudron, T.A.; Zhang, L.; Xu, W.; Xu, J.; Wang, M.; Xiao, X.; Wu, H. Effect of Prey Species and Prey Densities on the Performance of Adult *Coenosia attenuata*. *Insects* **2021**, *12*, 669. https://doi.org/10.3390/ insects12080669

Academic Editor: Allen Carson Cohen

Received: 31 May 2021 Accepted: 21 July 2021 Published: 23 July 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

effectively suppress a wide range of agricultural pests, such as fungus gnats (Diptera: Sciaridae), whiteflies (Hemiptera: Aleyrodidae), leaf miners (Diptera: Agromyzidae), winged aphids (Hemiptera: Aphididae), leafhoppers of the genera *Eupteryx* (Hemiptera: Cicadellidae) and *Empoasca* (Hemiptera: Cicadellidae), midges (Diptera: Chironomidae), moth flies (Diptera: Psychodidae), shore flies (Diptera: Ephydridae) and fruit flies (Diptera: Drosophilidae) [7,8,13,22–35]. The wide range of prey used as food make the tiger-fly an attractive alternative to conventional control methods.

Intact wings play an important role in the life of *C. attenuata* adults. Adults of *C. attenuata* catch their prey while in flight and pursue targets at the range of 23–212 mm. Hence, they employ an interception strategy that is more energy efficient to intercept targets, which allows *C. attenuata* to cope with the extremely high line-of-sight rotation rates and thus prevents overcompensation of steering [36]. Adults of *C. attenuata* use mean flight speeds of 0.69 ms−1, mean wingbeat frequency of 306, 19 Hz and acceleration of mean peak 9.3 ms−<sup>2</sup> to intercept prey [36]. The flight of *C. attenuata* individuals is affected by environmental factors, adjusting in response to changes in temperature, the number of prey flights and conspecific density [37]. Therefore, wing damage will cause negative effects on the life of *C. attenuata* adults.

Mass rearing of *C. attenuata* is important given the environmental, health and resistance issues associated with the use of chemical insecticides. To achieve an effective level of control, however, requires the production of a large number of *C. attenuata* at low cost. Adults of fungus gnat and drosophilid are now the main prey used to rear *C. attenuata* adults [17,23,38,39]. Rearing drosophilids is quick, easy and not particularly expensive. However, they were used primarily as a complement to the fungus gnat diet because *C. attenuata* fertility is lower when fed drosophilids compared to fungus gnats [23]. The reason why the performance of *C. attenuata* reared on drosophilids is lower than those reared on fungus gnats have not been assessed. The present study reports our finding that *C. attenuata* adults had less wing damage, higher fecundity and longer longevity when reared on *Bradysia impatiens* (Johannsem) (Diptera: Sciaridae) compared to *Drosophila melanogaster* Meigen (Diptera: Drosophilidae).

#### **2. Materials and Methods**

#### *2.1. Vinegar Flies*

The vinegar fly, *D. melanogaster* were reared on bananas in open plastic canisters (about 1200 cm3) in tissue bags (40 × 30 cm2, 0.4-mm mesh openings) closed with binder clips. Adults were introduced into the tissue bags and the adults of the following generation started to emerge after ca. 11 days. The colony was maintained in a laboratory incubator and held at 26 ± 1 ◦C, 14:10 (L:D) and 70 ± 5% RH.

#### *2.2. Fungus Gnats*

A colony of fungus gnats was initiated with about 400 *B. impatiens* adults captured from a greenhouse at Wuqing Experiment Station (Tianjin, China). Fungus gnats were reared using the method very similar to that reported by Zou et al. (2021) [39]. Modifications were made to simplify and improve the processes of rearing and collecting fungus gnats for use in bioassays. Briefly, 300 mL of black peat (Lvdimeijing Science and Technology Co., Ltd., Beijing, China) and 55 to 60 g of dry kidney bean powder were placed in an open plastic box (25.5 × <sup>19</sup> × 7.8 cm3). The mix was then moistened with 250 mL of tap water and 0.2-cm thick layer of moist coir (Shanghai Galuku Agricultural Science and Technology Co., Ltd. Shanghai, China; desalted, EC = 0.5, family pack, common grade) was placed on the top of the mix. Then the open plastic box was placed in a tissue bag (50 × 35 cm2, 0.4-mm mesh openings). In this case, 400 to 500 newly emerged adult fungus gnats were placed in the tissue bag and closed with a binder clip. Fresh rearing medium was prepared daily and new cultures were set up daily.

The new fungus gnat adults deposited eggs on the media consisted of black peat, tap water and kidney bean powder. Newly hatched larvae fed on the media and adults

emerged after 18–22 d. The colony was maintained in a laboratory incubator and held at 26 ± 1 ◦C, 14:10 (L:D) and 70 ± 5% RH.

#### *2.3. Tiger-Fly*

The *C. attenuata* used to establish a laboratory colony in this study were collected at Leizhuangzi flower farm of Tianjin, China. Adults were provided an oviposition tissue cage (60 × <sup>55</sup> × 50 cm3), in which an open plastic box (29 × <sup>20</sup> × 7 cm3) containing black peat, tap water, kidney bean powder and eggs of *B. impatiens*. A 0.5-cm thick layer of moist coir was placed on the top of the rearing media and used for oviposition. *B. impatiens* and *D. melanogaster* adults were supplied as prey daily. Five to 6 days later, the plastic boxes containing eggs of *C. attenuata* and rearing media were removed to another cage and second and third instar larvae of *B. impatiens* were added to the box to feed larvae of *C. attenuata*. Distilled water was added to the box when the media became dry. About 20 to 21 days later, adults of *C. attenuata* emerged. The colony was maintained in an artificial climate chamber and held at 26 ± 1 ◦C, 14:10 (L:D) and 70 ± 5% RH.

#### *2.4. Performance of Tiger-Fly Adults Reared on Different Prey Species at Different Prey Densities*

Five female/male pairs of newly emerged adults of *C. attenuata* (<24-h-old), in the first generation that originated from field-caught adults, were transferred into tissue cages (60 × <sup>55</sup> × 50 cm3) containing an open 90-mm-diameter Petri dish containing of 0.7-cm thick layer of moist coir for oviposition. The moist coir was replenished after collecting egg daily. Two 110 cm strings were hung inside each cage and served as a perch for adult predators. In this case, 6, 9, 15, 24 and 36 adults of *D. melanogaster* and 6, 12, 24, 36 and 48 adults of *B. impatiens* (<24-h-old) were provided daily per predator adult. Prey were used only once and fresh prey were added daily. Five female/male pairs of *C. attenuata* adults (in one cage) were tested until death in each treatment and replicated 9 times. In total, 45 pairs of *C. attenuata* adults were tested in each treatment within a phytotron held at 26 ± 1 ◦C, 14:10 (L:D) and 70 ± 5% RH. Wing damage in *C. attenuata* adults was measured daily using a Mitutoyo 500-196-30 digital caliper (Mitutoyo, Kawasaki, Japan). Wing damage occurred along the long axis of wing. The extent of damage was calculated as a percentage using the length of damaged wing/the total length of wing × 100%. Both wings were assessed and the average was taken. The numbers of surviving and killed prey were recorded daily. Eggs deposited in the moist coir were collected and counted from each cage daily using a 00-sized paintbrush. They were then placed in 60-mm-diameter Petri dishes containing a single 55-mm-diameter filter paper, moistened with distilled water, sealed with Parafilm and inverted to keep the eggs moist. Hatch occurred ca. 6 days after oviposition and egg viability was calculated. Distilled water was sprayed to each tissue cage two times in the morning and afternoon per day. Tiger-flies were maintained in this manner until death.

#### *2.5. Comparation of Body Weight and Body Length in C. attenuata, D. Melanogaster and B. Impatiens*

The adults of *C. attenuata* caught prey in flight. So carrying prey with different body weights may cause different levels of wing damage for adults of *C. attenuata*. Adults of *C. attenuata* (<24-h-old) and *D. melanogaster* (<24-h-old) were placed in tissue bags (40 × 30 cm2) and held in a Siemens BCD-501W fridge (Siemens, Nanjing, China) at −<sup>20</sup> ◦<sup>C</sup> for 2 min before taking body length and weight measurements. Adults of *B. impatiens* (<24-h-old) were handled in the same way with *C. attenuata* and *D. melanogaster* and held at −20 ◦C for 3 min before using. Adult body length was measured (in resting position) from the apex of the head to the wing tip for *C. attenuata* and *D. melanogaster*. For *B. impatiens*, body length was measured (in resting position) from the apex of the head to the abdomen tip. Body length measurements were made using a Mitutoyo 500-196-30 digital caliper (Mitutoyo, Kawasaki, Japan). Weight measurements were made using a Sartorius BP 211D (Sartorius AG, Göttingen, Germany) balance. In this case, 30 females and 30 males were measured for each species.

#### *2.6. Statistical Analyses*

One-way ANOVA with subsequent Tukey's HSD test at α = 0.05 was used to compare the proportion of damaged wing, number of prey killed, preovipositional period, total fecundity between different prey densities, body weight and body length between insect species. To avoid possible mistakes due to multiple testing of the same data base, the *p*-values were Bonferroni corrected. Two sample t-tests for means were used to compare proportion of damaged wing, preovipositional period and total fecundity between prey species. These comparisons were carried out on day 4 and the last day. The proportion of eggs successfully hatched was compared between treatments by the Chi-square test at α = 0.01. All the statistical tests were carried out using SAS version 9.4.

#### **3. Results**

#### *3.1. Comparison of Wing Damage of C. attenuata When Reared on Different Prey at Different Prey Densities*

The mean proportion of damaged wings of *C. attenuata* females fed on *D. melanogaster* and males fed on *B. impatiens* or *D. melanogaster* did not differ significantly between prey densities at early ages (day 4) (*F*4, 220 = 1.27, *p* = 0.2826; *F*4, 220 = 1.55, *p* = 0.1886; *F*4, 220 = 2.4, *p* = 0.0513, respectively) (Figure 1B–D). However, *C. attenuata* females fed 48 adults of *B. impatiens* daily lost more wings compared with those fed 12 adults of *B. impatiens* daily on day 4 (*F*1, 88 = 9.39, *p* = 0.0029, Bonferroni-corrected *p* = 0.005) (Figure 1A). The wing damage increased with age in every case, but at different rates depending on prey species and density. When fed with *B. impatiens*, *C. attenuata* females showed significant differences at late age (31 days) between prey densities of 12 and 24 (*F*1, 88 = 13.15, *p* = 0.0004, Bonferroni-corrected *p* = 0.005), or between prey densities of 36 and 48 (*F*1, 88 = 10.16, *p* = 0.002, Bonferroni-corrected *p* = 0.005) (Figure 1A). For *C. attenuata* males, there were no significant differences between prey densities at late age (19 days) (*F*4, 220 = 0.63, *p* = 0.6425) (Figure 1B). When fed with *D. melanogaster*, both females and males of *C. coenosia* did not show significantly different wing damage between prey densities at later age (8 days) (*F*4, 220 = 2.08, *p* = 0.0848; *F*4, 220 = 2.07, *p* = 0.0856, respectively) (Figure 1C,D).

**Figure 1.** Mean proportion of damaged wings of all *Coenosia attenuata* when fed different prey at different densities daily per predator adult (mean values with 95% confidence intervals; error bars: 95% CI): (**A**) females fed 6, 12, 24, 36 and 48 adults of *Bradysia impatiens*; (**B**) males fed 6, 12, 24, 36 and 48 adults of *Bradysia impatiens*; (**C**) females fed 6, 9, 15, 24 and 36 adults of *Drosophila melanogaster*; (**D**) males fed 6, 9, 15, 24 and 36 adults of *Drosophila melanogaster*.

Females and males of *C. attenuata* had much shorter longevity when fed *D. melanogaster* adults compared to those fed *B. impatiens* adults (maximum age of 12 for *D. melanogaster* and of 46 for *B. impatiens* in females; maximum age of 10 for *D. melanogaster* and of 35 for *B. impatiens* in males). Mean proportion of wing damage in flies fed with *B. impatiens* at age 8 ranged from 3.25 to 22.42% in females and from 4.68 to 9.58% in males, while in flies fed with *D. melanogaster* it ranged from 12.28 to 21.56% in females and from 9.16 to 19.77% in males (Figure 1).

*C. attenuata* females fed 6 adults of *D. melanogaster* daily lost significantly more wings than those fed 6 adults of *B. impatiens* daily on day 4 (*t* = −2.935, *df* = 88, *p* = 0.0043) (Figure 1A,C). However, there was no significant difference in wing damage for *C. attenuata* females fed 24 adults of *D. melanogaster* compared with those fed 24 adults of *B. impatiens* daily, or for *C. attenuata* females fed 36 adults of *D. melanogaster* compared with those fed 36 adults of *B. impatiens* daily on day 4 (*t* = −1.934, *df* = 88, *p* = 0.0564; *t* = −0.447, *df* = 88, *p* = 0.6561, respectively) (Figure 1A,C). *C. attenuata* females lost significantly more wings when fed adults of *D. melanogaster* daily compared to *B. impatiens* at the prey density of 6 and 36 on day 8 (*t* = −4.556, *df* = 88, *p* < 0.0001; *t* = 2.803, *df* = 88, *p* = 0.0062, respectively). However, there was no significant difference in wing damage for *C. attenuata* females when fed 24 adults of *D. melanogaster* compared with those fed 24 adults of *B. impatiens* on day 8 (*t* = −0.874, *df* = 88, *p* = 0.3847) (Figure 1A,C).

There was no significant difference in wing damage for *C. attenuata* males fed 6 adults of *D. melanogaster* compared with those fed 6 adults of *B. impatiens*, or for *C. attenuata* males fed 36 adults of *D. melanogaster* compared with those fed 36 adults of *B. impatiens* on day 4 (*t* = −1.35, *df* = 88, *p* = 0.1805; *t* = −1.1, *df* = 88, *p* = 0.2743, respectively) (Figure 1B,D). *C. attenuata* males lost significantly more wings when fed adults of *D. melanogaster* daily compared to *B. impatiens* at the prey density of 24 on day 4 (*t* = −2.302, *df* = 88, *p* = 0.0237) and the prey density of 6 on day 8 (*t* = −2.631, *df* = 88, *p* = 0.01). However, there was no significant difference in wing damage for *C. attenuata* males when fed adults of *D. melanogaster* daily compared to *B. impatiens* at the prey density of 24 or 36 on day 8 (*t* = −0.362, *df* = 88, *p* = 0.7181; *t* = −0.482, *df* = 88, *p* = 0.6311, respectively) (Figure 1B,D).

#### *3.2. Number of Prey killed by C. attenuata Reared on Different Prey at Different Prey Densities*

*C. attenuata* adults killed all *B. impatiens* adults when fed 6 prey daily per predator adult except for the last 3 days, which suggests 6 adults of *B. impatiens* are not enough for *C. attenuata* adults (Figure 2). Most of *B. impatiens* adults were killed by *C. attenuata* adults when fed 12 prey daily per predator adult. The number of prey killed by *C. attenuata* adults fluctuated when fed 24, 36 and 48 prey daily per predator adult. The numbers of prey killed per predator daily were 15.33 to 23.77, 25.99 to 35.70 and 33.00 to 47.45 for *C. attenuata* adults fed 24, 36 and 48 prey daily per predator adult, respectively. The number of prey killed daily per predator decreased in the last few days with the increase of age and the mean proportion of damaged wings for *C. attenuata* adults fed 12, 24, 36 and 48 prey daily per predator adult (Figure 2). The number of *B. impatiens* adults killed daily per predator adult differed significantly between prey densities at early ages (day 4) (*p* < 0.0001 for 9 comparisons), except for densities of 24 vs. 36 (*F*1, 16 = 5.61, *p* = 0.0308, Bonferronicorrected *p* = 0.005) (Figure 2). At later age (day 31), the number of *B. impatiens* adults killed daily per predator adult differed significantly between prey densities (*p* < 0.0001 for 9 comparisons), except for densities of 36 vs. 48 (*F*1, 4 = 0.26, *p* = 0.6392, Bonferroni-corrected *p* = 0.005) (Figure 2).

*C. attenuata* adults killed 3.17 to 4.52 *D. melanogaster* adults when fed 6 prey adults daily per predator adult (Figure 3). The number of prey killed by *C. attenuata* adults fluctuated when fed 9, 15, 24 and 36 prey daily per predator adult. The numbers of prey killed daily per predator were 4.27 to 6.54, 4.37 to 7.29, 7.57 to 11.98 and 9.58 to 15.55 for *C. attenuata* adults fed 9, 15, 24 and 36 prey daily per predator adult, respectively. The number of prey killed per predator daily decreased in the last few days with the increase of age and the proportion of broken wings for *C. attenuata* adults in all treatments. The number of *D. melanogaster* adults killed daily per predator adult differed significantly with prey densities of 6 vs. 9 (*F*1, 16 = 10.64, *p* = 0.0049, Bonferroni-corrected *p* = 0.005), 9 vs. 36 (*F*1, 16 = 17.21, *p* = 0.0008,

Bonferroni-corrected *p* = 0.005) and with the prey densities of 6 vs. 15, 6 vs. 24, 6 vs. 36 and 15 vs. 36 (all *p* < 0.0001) at early ages (day 4). No significant differences were found in the other 4 comparisons at early ages (day 4) (Figure 3). Additionally, no significant differences were found in all 10 comparisons at later age (day 8) (Figure 3).

**Figure 2.** Mean number of prey killed per *C. attenuata* adult when fed 6, 12, 24, 36 and 48 adults of *B. impatiens* daily per predator adult (values are mean ± SE, single values were present when only one adult *C. attenuata* was left in the end).

**Figure 3.** Mean number of prey killed per *C. attenuata* adult when fed 6, 9, 15, 24 and 36 adults of *D. melanogaster* daily per predator adult (values are mean ± SE, single values were present when only one adult *C. attenuata* was left in the end).

#### *3.3. Preovipositional Period of C. attenuata Female Reared on Different Prey at Different Prey Densities*

There were no significant differences in preovipositional period in any of the treatments of *C. attenuata* females when fed 6, 9, 15, 24 or 36 adults of *D. melanogaster* prey (*F*4, 40 = 2.43, *p* = 0.064) (Figure 4). The preovipositional period of *C. attenuata* female was from 4.22 to 5.22 days when fed *D. melanogaster* prey. Similar, there were no significant differences in preovipositional period in any of the treatments of *C. attenuata* females when fed 6, 12, 24, 36 or 48 adults of *B. impatiens* (*F*4, 40 = 1.73, *p* = 0.162). The preovipositional period of *C. attenuata* female was from 3.89 to 4.67 days when fed *B. impatiens* prey. There were no significant differences in the preovipositional period for *C. attenuata* females when fed 6 adults of *D. melanogaster* prey or 6 adults of *B. impatiens* prey, or for *C. attenuata* females when fed 36 adults of *D. melanogaster* prey or 36 adults of *B. impatiens* prey (*t* = −0.686, *df* = 16, *p* = 0.5025; *t* = −1.715, *df* = 16, *p* = 0.1056, respectively) (Figure 4). The preovipositional periods were significantly longer for *C. attenuata* females fed 24 adults of *D. melanogaster* prey than of *C. attenuata* females fed 24 adults of *B. impatiens* prey. These differences, although statistically significant, were small (*t* = −2.132, *df* = 16, *p* = 0.0489) (Figure 4).

**Figure 4.** Preovipositional periods of *C. attenuata* female adult when fed 6, 9, 15, 24 and 36 adults of *D. melanogaster* and 6, 12, 24, 36 and 48 adults of *B. impatiens* daily per predator adult. Different letters above each bar indicate significant differences between prey densities using one-way ANOVA, Tukey's HSD test (*p* = 0.05 and n = 9).

#### *3.4. Total Fecundity of C. attenuata Female Reared on Different Prey at Different Prey Densities*

The total fecundity per female of *C. attenuata* fed *D. melanogaster* prey differed significantly with prey densities of 6 vs. 36 (*F*1, 16 = 17.38, *p* = 0.0007, Bonferroni-corrected *p* = 0.005), 9 vs. 36 (*F*1, 16 = 19.82, *p* = 0.0004, Bonferroni-corrected *p* = 0.005), 24 vs. 36 (*F*1, 16 = 18.90, *p* = 0.0005, Bonferroni-corrected *p* = 0.005) and the other 4 comparisons (6 vs. 9, 6 vs. 15, 6 vs. 24 and 15 vs. 36, all *p* < 0.0001). No significant differences were found in other 3 comparisons (Figure 5). The total fecundity per female of *C. attenuata* fed *B. impatiens* prey did not differ significantly with prey densities of 6 vs. 48 and 12 vs. 36 (*F*1, 16 = 0.17, *p* = 0.6829; *F*1, 16 = 9.54, *p* = 0.007, respectively, Bonferroni-corrected *p* = 0.005). However, the total fecundity per female of *C. attenuata* fed *B. impatiens* prey differed significantly between prey densities for the other 8 comparisons (all *p* < 0.0001) (Figure 5). Additionally, the total fecundity was much higher for females of *C. attenuata* fed *B. impatiens* adults than for those fed *D. melanogaster* prey at the same prey densities of 6, 24 and 36 (*t* = 31.971, *df* = 16, *p* < 0.0001; *t* = 36.609, *df* = 16, *p* < 0.0001; *t* = 32.954, *df* = 16, *p* < 0.0001, respectively).

**Figure 5.** Total fecundity per female of *C. attenuata* when fed 6, 9, 15, 24 and 36 adults of *D. melanogaster* and 6, 12, 24, 36 and 48 adults of *B. impatiens* daily per predator adult. Different letters above each bar indicate significant differences between prey densities (corrected *p* value for multiple testing by Bonferroni correction is 0.005 and n = 9).

#### *3.5. Proportion of Eggs Successfully Hatched in C. attenuata Reared on Different Prey at Different Prey Densities*

There was a significantly higher proportion of eggs that successfully hatched for *C. attenuata* fed 9 adults of *D. melanogaster* than for those fed 24 and 36 adults of *D. melanogaster* daily per predator adult (χ<sup>2</sup> = 8.37, *p* = 0.004; χ<sup>2</sup> = 20.56, *p* < 0.0001, respectively) (Figure 6). However, there were no significant differences in the proportion of eggs that successfully hatched between *C. attenuata* fed 9 adults of *D. melanogaster* and those fed 6 or 15 adults of *D. melanogaster* (χ<sup>2</sup> = 6.36, *p* = 0.012; χ<sup>2</sup> = 0.81, *p* = 0.368, respectively). There were no significant differences in proportion of eggs successfully hatched between *C. attenuata* fed 24 adults of *B. impatiens* and those fed 6, 12 or 36 adults of *B. impatiens* (χ<sup>2</sup> = 5.97, *p* = 0.015; χ<sup>2</sup> = 1.28, *p* = 0.257; χ<sup>2</sup> = 2.92, *p* = 0.087, respectively). The proportion of eggs that successfully hatched was significantly higher for *C. attenuata* fed 24 adults of *B. impatiens* than for those fed 48 adults of *B. impatiens* daily per predator adult (χ<sup>2</sup> = 6.87, *p* = 0.009). Additionally, the proportion of eggs that successfully hatched was much higher for *C. attenuata* adults fed *B. impatiens* adults than for those fed *D. melanogaster* adults at the same prey densities of 6, 24 and 36 (χ<sup>2</sup> = 15.14, *p* < 0.0001; χ<sup>2</sup> = 43.35, *p* < 0.0001; χ<sup>2</sup> = 43.26, *p* < 0.0001, respectively).

#### *3.6. Comparation of Body Weight and Body Length in C. attenuata, D. melanogaster and B. impatiens*

Adult females of *C. attenuata* were significantly longer than those of *D. melanogaster* and *B. impatiens* (*F*2, 87 = 1335.80, *p* < 0.0001) (Table 1). Adult females of *D. melanogaster* were significantly longer than those of *B. impatiens* (*F*2, 87 = 1335.80, *p* < 0.0001). Similar, adult males of *C. attenuata* were significantly longer than those of *D. melanogaster* and *B. impatiens* (*F*2, 87 = 2101.27, *p* < 0.0001) and the body length was significantly longer for adult males of *D. melanogaster* than for adult males of *B. impatiens* (*F*2, 87 = 2101.27, *p* < 0.0001).

**Table 1.** Body length (mm) and body weight (mg) of adult *Coenosia attenuata*, *Drosophila melanogaster* and *Bradysia impatiens* (n = 30).


Values are mean ± SE. Means in columns with the same letter are not significantly different at a 0.05 level of significant. <sup>a</sup> Adult body length measured < 24 h after adult emergence. <sup>b</sup> Adult body weight measured < 24 h after adult emergence.

Adult females of *C. attenuata* were significantly heavier than those of *D. melanogaster* and *B. impatiens* and there was a significant difference in the body weight of adult females between *D. melanogaster* and *B. impatiens* (*F*2, 87 = 2188.57, *p* < 0.0001). Similar, adult males of *C. attenuata* were significantly heavier than those of *D. melanogaster* and *B. impatiens* and there was a significant difference in the body weight of adult males between *D. melanogaster* and *B. impatiens* (*F*2, 87 = 1164.34, *p* < 0.0001) (Table 1).

#### **4. Discussion**

The flight of *C. attenuata* individuals was affected by environmental factors and was increased in response to increases in the number of prey flights [37]. Bonsignore (2016) found that predatory flights of adult *C. attenuata* comprised a small percentage (ca. 6%) of the total flights, with a predation success rate of 61% [37]. In our study, the mean proportion of damaged wings of *C. attenuata* females when fed 6, 12, 24 and 36 adults of *B. impatiens* daily per predator adult was increased in response to increases in the number of prey densities. However, the mean proportion of wing damage in *C. attenuata* females was lower for prey densities of 48 adults of *B. impatiens* than for prey densities of 36. The high density of 48 adults of *B. impatiens* probably increased the predation success rate and thereby decreased the mean proportion of damaged wings of *C. attenuata* female although the tiger-fly is regarded to have predation instinct [40,41]. Damaged wings of *C. attenuata* males fed on *B. impatiens* continued to increase with an increase in age of *C. attenuata*. However, the mean proportions of damaged wings in *C. attenuata* males were not consistent with those of females. Prey density did not cause significant effect on wing damage for *C. attenuata* males, which suggests prey density was not the only factor affecting wing damage in *C. attenuata* males. Being attacked by female *C. attenuata* and attempting to mate with female *C. attenuata* could also influence wing damage. Additionally, male adults required less prey compared to female adults, which means low prey densities could increase the predation success rate and thereby decreased the proportion of damaged wings of *C. attenuata* males. According to the damaged wings and longevity of *C. attenuata* adults, prey densities of 12 to 24 should be optimal density for mass rearing of adult *C. attenuata*.

Prey density of vinegar fly did not cause a significant effect on the mean proportion of damaged wings in both female and male adults of *C. attenuata*. It seems reasonable to conclude that the short lifespan of the tiger-fly was too short to manifest an effect of *D. melanogaster* prey density on wing damage of *C. attenuata* adults. *C. attenuata* adults fed *D. melanogaster* prey daily lost more wings compared to those fed *B. impatiens* prey at the same age for some prey density. This may be related to the increased difficulty in carrying heavier *D. melanogaster* adults than *B. impatiens* adults.

Bonsignore (2016) sorted adult *C. attenuata* flights into three groups, movement flights, territory defense flights and predatory flights in greenhouse [37]. However, we observed that there should be another type of flight, escape flights in cage. We hypothesize that adult *C. attenuata* want to escape from the cage when encountering high density of prey, resulting in less wing damage. Escaping from the environment with high prey density may be a self-protection response for adult *C. attenuata*.

We found that females lived longer than males, as reported by Kühne et al. (1997) [23]. However, these authors record 38 days and 33 days as the maximum female and male longevity, respectively, under laboratory conditions (25 ◦C and 50–60% RH) and an estimated longevity of eight weeks under greenhouse conditions. Predators fed *B. impatiens* adults in our study lived 46 days and 35 days as the maximum female and male longevity, respectively, possibly because they had a better food supply. However, in our study, predators fed adult *D. melanogaster* flies lived only 12 days and 10 days as the maximum female and male longevity, respectively. We speculate that *C. attenuata* adults were able to more easily capture lighter *B. impatiens* adults than heavier adult *D. melanogaster*. Additionally, adult *C. attenuata* were able to attack adult *B. impatiens* on the bottom of cage when they could only jump or crawl because of damaged wings. However, it is difficult for *C. attenuata* with damaged wings to capture adult *D. melanogaster*.

Female adult *C. attenuata* were found to exhibit a type I functional response to adult sciarid flies, which was conducted in glass vials 8 cm long and 8 cm in diameter at 25 ◦C at 60–80% RH, with a 16L:8D photoperiod. Sciarids were consumed in significantly different numbers at densities from 5 to 20 individuals (the number of killed flies changed from 2.90 to 8.4, respectively). However, increasing prey availability beyond 20 individuals resulted in no substantial increase in predation [30]. However, female adult *C. attenuata* were found to exhibit a type II functional response to adult *D. melanogaster* flies, which was conducted in Plexiglas cages with a dimension of 25 by 25 by 25 cm at 30 ◦C at 65 ± 5% RH, with a 12L:12D photoperiod. *D. melanogaster* flies were consumed in significantly different numbers at densities from 5 to 55 individuals (the number of attacked flies changed from 3.50 to 5.67, respectively) [42]. Kühne (2000) states that each adult *C. attenuata* needs either 1.5 adults of *D. melanogaster* or 6.9 adults of *B. impatiens* per day [25]. We did not analyze the functional response to adult *B. impatiens* or *D. melanogaster* flies because more than one factor affected functional response, such as intraspecific competition and predation. The number of killed prey in our study was more than those mentioned above, which was probably caused by intraspecific competition and predation instinct resulting from cage and space differences. The flight ability of adult *B. impatiens* is weak and often some of them stayed on the bottom of cage which made it more convenient for adult *C. attenuata* without flight ability, because of damaged wings, to catch the adult *B. impatiens*. In contrast, the flight ability of adult *D. melanogaster* is strong and it is more difficult for adult *C. attenuata* with weakened flight ability to catch adult *D. melanogaster*, although adult *C. attenuata* has been proved to be more efficient in information sampling and processing than adult *D. melanogaster* [43,44].

The preoviposition period of *C. attenuata* is approximately 4 days [23]. Our reports showed similar preoviposition periods when fed adults of *B. impatiens* with 3.89–4.67 days and adults of *D. melanogaster* with 4.22–5.22 days. Prey density, prey species and damaged wings did not cause negative effects on the preoviposition period of *C. attenuata*. Sanderson et al. (2009) found the tiger-flies laid more eggs with fungus gnat prey than shore fly prey [27]. We found that the tiger-flies laid much more eggs with fungus gnat prey than vinegar fly prey. However, the total fecundity per female of *C. attenuata* did not continue to increase with an increase in prey density. Shorter life span probably cause the lower fecundity for *C. attenuata* female when fed adult *D. melanogaster* compared to adult *B. impatiens*. Martins et al. (2015) presented an optimized method for mass rearing *C. attenuata* with fungus gnats and Drosophilids as prey, where the number of adults that emerged per parental pair ranged from 1.8 to 9.0 (= per pair progeny production, or the number of adult offspring that emerged in each cage divided by the number of parental pairs) [38]. In our study, the number of adult offspring that emerged ranged from 4.96 to 7.64 and 21.16 to 39.27 at least for per parental pair when fed adult *D. melanogaster* and *B. impatiens*, respectively according to survival rates of larvae, percentages of pupation and adult emergence in our previous reports [17,39]. The proportion of eggs that successfully hatched was much higher for *C. attenuata* adults fed *B. impatiens* adults than for those fed *D. melanogaster* adults at the same prey densities. Longer longevity in male *C. attenuata* and lighter body weight in *B. impatiens* prey correlated to increased proportion of eggs that successfully hatched in *C. attenuata.*

Predation by adult *C. attenuata* is rapid, and adults take off as soon as they observe their prey in flight, although they do not know the absolute size of the potential prey prior to the flight [45]. One important physical factor affecting predator responses is prey size [46]. Body length and body weight of adult *C. attenuata*, *D. melanogaster* and *B. impatiens* were analyzed in our study to better understand the complexity of predation. We report body length and body weight of adult *C. attenuata* from Tianjin to be similar to those reported by us previously [17,39] and to those reported in Uruguay where the *C. attenuata* flies measured approximately 2.5–5.00 mm in length [47]. Body weight of adult *D. melanogaster* measured in this study is similar to those reported by Chen et al. (2019) [48]. Body length of adult *B. impatiens* analyzed in this study is similar to those reported by Wilkinson and

Daugherty (1970) [49]. Obviously, it is more difficult for adult *C. attenuata* to catch and carry heavier adult *D. melanogaster* than lighter adult *B. impatiens*. Most importantly, it is more difficult for male adult *C. attenuata* to catch and carry female adult *D. melanogaster*, that are 74.15% weight of male adult of *C. attenuata* than to catch adult female *B. impatiens* that only weigh 28.57% of their weight.

In our study, we demonstrated that adult *B. impatiens* was an optimal prey in the mass rearing of adult *C. attenuata* although rearing drosophilids is quick, easy and not particularly expensive. In addition, we provide evidence that damage to wings of adult *C. attenuata* when fed adult *D. melanogaster* vs. *B. impatiens* is an important consideration for prey selection. We conclude a prey density of 12–24 adult fungus gnats daily per adult predator as optimal for mass rearing of adult *C. attenuata*. Rearing cost, nutritional difference, digestion efficiency, chemical, morphological and behavioral defense mechanisms of a prey will be explored in future studies.

#### **5. Conclusions**

We present the first report of wing damage for *C. attenuata* adults when reared on different prey. The results indicate that *C. attenuata* adults had higher fecundity, longer longevity and generally less wing damage when reared on *B. impatiens* compared to *D. melanogaster*. Lighter body weight and weaker flight ability in adult *B. impatiens* prey likely contributed to prolonged longevity and increased fecundity in adult *C. attenuata.* In this case, 12 to 24 adults of *B. impatiens* daily per predator were considered optimal prey density in the mass rearing of adult *C. attenuata* adult.

**Author Contributions:** Conceptualization, D.Z. and H.W.; methodology, D.Z., L.Z. and H.W.; software, W.X. and J.X.; validation, D.Z., T.A.C., L.Z., W.X., J.X. and H.W.; formal analysis, D.Z. and L.Z.; investigation, D.Z., M.W. and X.X.; resources, W.X. and H.W.; data curation, D.Z., M.W., X.X. and H.W.; writing—original draft preparation, D.Z.; writing—review and editing, T.A.C., L.Z., W.X., J.X. and H.W.; visualization, D.Z., T.A.C., L.Z., W.X., J.X. and H.W.; supervision, H.W.; project administration, D.Z., J.X. and X.X.; funding acquisition, D.Z., T.A.C., L.Z., W.X. and H.W. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by Science and Technology Innovation Foundation for Young Scientists of Tianjin Academy of Agricultural Sciences (No. 201911), National Key Research and Development Program of China (No. 2017YFD0201000, 2017YFD0201707) and the USDA Agricultural Research Service project Insect Biotechnology Products for Pest Control and Emerging Needs in Agriculture (No. 5070-22000-037-00-D).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** We thank Wan-Qi Xue (Shenyang Normal University, China) for helping in species identification.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses or interpretation of data; in the writing of the manuscript or in the decision to publish the results. 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.

#### **References**


### *Article* **Horizontal Honey-Bee Larvae Rearing Plates Can Increase the Deformation Rate of Newly Emerged Adult Honey Bees**

**Juyeong Kim 1, Kyongmi Chon 1,\*, Bo-Seon Kim 1, Jin-A Oh 1, Chang-Young Yoon 1, Hong-Hyun Park <sup>1</sup> and Yong-Soo Choi <sup>2</sup>**


**Simple Summary:** Rearing honey bee (*Apis mellifera*) larvae in vitro is an important method for studying bee larvae diseases or the toxicity of pesticides on bees. Laboratory experiments for bee larvae are usually performed by placing a rearing plate horizontally during all developmental stages. However, recent studies have demonstrated that a horizontal rearing environment can cause the deformation of emerged bees. Most studies adopted a vertical rearing method to reduce such deformation, but there is a lack of information on the emergence rates and deformation rates of bees reared on vertical or horizontal plates. Therefore, in this study, we examined the effect of placing the plates vertically and horizontally on newly emerged bees. There were no significant differences in larval mortality, pupal mortality, and adult emergence rates between horizontal and vertical rearing plates. However, the adult deformation rates of the horizontal plates were significantly higher than those of the vertical plates. In conclusion, we suggest that the vertical rearing method is more suitable when considering the deformation rate of the control group to verify the sublethal effects of pesticides on honey bees.

**Abstract:** Rearing honey bee larvae in vitro is an ideal method to study honey bee larval diseases or the toxicity of pesticides on honey bee larvae under standardized conditions. However, recent studies reported that a horizontal position may cause the deformation of emerged bees. Accordingly, the purpose of this study was to evaluate the emergence and deformation rates of honey bee (*Apis mellifera ligustica*) larvae reared in horizontal and vertical positions. The study was conducted under the same laboratory conditions with three experimental groups, non-capped or capped horizontal plates and capped vertical plates. However, our results demonstrated that the exhibited adult deformation rates of the horizontal plates were significantly higher (27.8% and 26.1%) than those of the vertical plates (11.9%). In particular, the most common symptoms were deformed wings and an abnormal abdomen in the horizontal plates. Additionally, adults reared on horizontal plates were substantially smaller (10.88 and 10.82 mm) than those on vertical plates (11.55 mm). Considering these conclusions, we suggest that a vertical rearing method is more suitable when considering the deformation rates of the control groups to verify the sublethal effects of pesticides on honey bees.

**Keywords:** *Apis mellifera*; deformation; emergence; honey bee; in vitro rearing; larvae

#### **1. Introduction**

Many studies have reported recent significant pollinator declines and increased honey bee (*Apis mellifera*) colony losses in many countries [1–5]. Several stressing factors, such as pathogens, climate change, parasites, habitat loss, lack of nutrition, pesticides, and diseases are considered to explain the decline and colony losses [6–9]. However, a single causative

**Citation:** Kim, J.; Chon, K.; Kim, B.-S.; Oh, J.-A; Yoon, C.-Y.; Park, H.-H.; Choi, Y.-S. Horizontal Honey-Bee Larvae Rearing Plates Can Increase the Deformation Rate of Newly Emerged Adult Honey Bees. *Insects* **2021**, *12*, 603. https://doi.org/ 10.3390/insects12070603

Academic Editors: Man P. Huynh, Kent S. Shelby and Thomas A. Coudron

Received: 6 June 2021 Accepted: 23 June 2021 Published: 1 July 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

stressor factor has not been conclusively identified, because of the complexity related to concurrent multiple stressors [4,10–12]. Among the several factors suggested, pesticides are regarded as one of the most crucial causes of adverse honey-bee health and colony declines [4,13–15]. Various studies have been conducted to determine the exposure effects of pesticides on adult honey bees, therefore, standard methods for investigating the effects of pesticides on adult bees have been well investigated both in vivo and in vitro [15,16]. Nevertheless, in contrast to controllable laboratory conditions, field experiments in hives are impacted by numerous uncontrollable factors such as season, colony genetic variation, climate, and resource availability [17–19]. Because of these uncontrolled variables, the in vitro procedure of rearing honey-bee larvae has been proposed to evaluate the toxicity of pesticides on honey-bee broods (larvae, pupae, and adults) [11]. Rearing larvae in vitro is a practical protocol to study larval pathogens, development, and caste differentiation in honey bees [20–23]. Nevertheless, available data in the publications concerning the lethal and sublethal effects on honey-bee larvae are rather poor compared to adult bees [11,24,25].

In 1933, the first informative report investigating the caste differentiation of queen and worker bees, as well as hand-feeding bee larvae with a diet containing royal jelly in vitro, was published [26]. The larval diet composition was improved and optimized by Rembold and Lackner [23], and Vandenberg and Shimanuki [27]. Vandenberg and Shimanuki [27] further developed methods of rearing one larva per cup and feeding them the correct quantity of diet daily. Wittmann and Engels [28] reported an in vitro rearing method as a risk assessment tool to study the toxicity of pesticides. Davis et al. [29] provided diets containing carbofuran and dimethoate to larvae reared in the laboratory, and Peng et al. [21] utilized rearing honey-bee larvae in vitro for assessing the toxicity of pesticides on honey bees. Most notably, seven laboratories across five different countries performed ring tests according to the improved in vitro methods to assess LD50 for acute toxicity of dimethoate in 2005 and 2008 [17]. The ring test participants achieved adult emergence rates greater than 80% in 43% of their control trials and greater than 90% in 17% of their control trials [12,17]. The Organization for Economic Co-operation and Development (OECD) guidelines for the honey-bee larvae toxicity tests under laboratory conditions were published in 2013 (single exposure) and 2016 (repeated exposure), based on the methods that were developed in ring tests [30,31].

Larval mortality and pupal mortality in the natural hive occurred at approximately 15% [32,33]. Therefore, the OECD guidelines specify that the total mortality during larval and pupal developmental stages should not exceed 15% in the controls, otherwise the study is considered invalid [30,31]. Namely, the mortalities of the control groups should be considered for validation of the test [32].

Accordingly, researchers have focused on increasing the survival rate of the larval stage, and there are several examples of these methods. They mainly improved the survival rate of larvae according to the composition of larval food, quality of royal jelly, larval age at grafting, rearing conditions (temperature and relative humidity) in the laboratory, or reducing contact between larvae and fecal materials using absorbents such as Kimwipes (filter paper) [12,32,34–36].

However, even if in vitro rearing protocols have been improved over the years, variable (inconsistent) survival rates in the controls of each laboratory have been reported continuously [12,17]. Zhu et al. [37] reported the larval mortalities of controls were approximately 17.5% at D6 after grafting. Additionally, low emergence rates of controls (≤50%) were noted in several experimental studies based upon Aupinel et al. [11,12]. Namely, this means that several research institutes have already performed a larval toxicity test, with control mortalities higher than the OECD guidelines [17,18,22,38]. The inconsistent results across different laboratories may reflect subtle differences in the brood sources and the laboratory conditions, or be related to the effects of the mechanical stress of grafting [12,19,39]. It is also a more common practice to set up the rearing using 48-well tissue plates, with the grafting cell cups placed horizontally, which is according to the OECD guideline (2016). However, some studies have mentioned that the horizontal rearing positions during the developmental stages may lead to the deformation of wings and abdomen (humpback) in emerged bees, because the larvae or pupae can then withstand abnormal vertical positions [32,40,41]. Riessberger-Gallé et al. [40] proposed a method to prevent these deformations, in which a 48-well tissue plate was sealed with a thin wax layer and set vertically as in natural beehives. However, presently, specific information on the emergence rates and deformation rates of the newly emerged honey bees when the rearing plates are placed vertically or horizontally has not yet been reported. Therefore, in this study, we aimed to evaluate the specific difference in the emergence rate and deformation rate of emerged bees when the rearing plates were positioned horizontally or vertically during the pupal developmental stage.

#### **2. Materials and Methods**

#### *2.1. Rearing Honey Bee Larvae In Vitro*

The honey-bee larvae (*Apis mellifera ligustica*) were randomly obtained from three healthy colonies at an apiary located at the National Institute of Agricultural Sciences in Korea (35◦49 47.0" N, 127◦02 26.0" E). The source colonies and bees were not treated against *Varroa destructor* for four weeks prior to the experiment. The honey-bee queens were caged on wax combs using queen excluders to lay eggs. The freshly laid eggs were confirmed the next day (24 h after the queens were caged), and after that, the combs containing the hatched first instar larvae (72 h after the queens laid the eggs) were delivered to the laboratory for grafting. Before grafting, the grafting cell cups (Nicotplast, Maisod, France) were disinfected with 70% ethanol and were then used, after UV sterilization for 30 min in a laminar-flow hood. The larval rearing procedure followed OECD No. 239 [31]. On day 1 (D1), 20 μL of larval diet A was loaded into the grafting cell cup, and healthy first instar larvae were transferred into the cell cups of 48-well tissue plates (SPL, Pocheon-si, Korea). During grafting, a clay pack (Caremate, Hwaseong-si, Korea) was preheated in the microwave and the pack was laid underneath 48-well tissue plates to minimize the temperature effect on the larvae [32]. After grafting, the 48-well tissue plates were placed horizontally in a sealed desiccator (Nalgene, Rochester, NY, USA) in a constant temperature incubator (DAIHAN Scientific Co., Wonju-si, Korea) maintained at 35 ◦C and 95 ± 5% relative humidity (RH) using a saturated solution of potassium sulfate (Junsei, Tokyo, Japan) during the larval stages (D1–D8). The pupal stages (D8–D15) were maintained at 80 ± 5% RH using a saturated solution of sodium chloride (Sigma–Aldrich, St. Louis, MO, USA). In the emergence stages (D15–D21), the 48-well tissue plates were transferred individually into emergence boxes with a 50% sucrose solution, and placed in the incubator to maintain 50% RH and 35 ◦C.

For larval feeding, d-glucose (Difco, Sparks, NV, USA) and d-fructose (Junsei, Japan) were added to water filtered with a 0.20 μm filter (Sartorius, Göttingen, Germany), and then yeast extract (Bacto, Sparks, NV, USA) was added and mixed. Lastly, the solution was mixed with royal jelly (Haechangol Honey Farm, Yeongwol, Korea) [12]. Following the OECD guidelines, the larval diets included a total of 160 μL of each standardized volume during the six days (excluding D2) of the larval stages, where the larval diet volume and components have been summarized in Table 1. Before feeding the larvae, diets were preheated in an incubator maintained at 35 ◦C.


**Table 1.** The volume and larval diet component percentages, according to the OECD guidelines [30,31].

#### *2.2. Experimental Design*

The non-capped and horizontally oriented groups (NHG) were placed in 48-well tissue plates without a wax layer and set horizontally, and the capped and horizontally oriented groups (CHG) were placed in 48-well tissue plates that were capped with the artificial wax layers and set horizontally. Meanwhile, the capped and vertically oriented groups (CVG) were placed in 48-well tissue plates that were capped with the artificial wax layers and set vertically. All groups were not treated with any chemical reagents. The experiments of each group were tested with 4 replicates (36 larvae per plate). Artificial wax layers were prepared by dissolving 4.0 g of pure beeswax. The size of the wax layers was 14 cm × 10 cm × 0.4 mm. In each grafting cell cup of the plate, small orifices were made to allow air exchange. Rearing plates were sealed with the perforated wax layers on D15 after grafting, particularly vertical plates (CVG) placed carefully upright so pupae were facing towards the opening.

#### *2.3. Mortality and Abnormal Symptoms*

Mortality and abnormal symptoms at each developmental stage were visually observed and recorded every day. After setting the survival rates of the larvae at 100% on D3, larval mortality, and abnormal symptoms were monitored as early death and melanizing death from D4 to D8 (larval stages) [12,30,31]. The larvae were considered as dead when the larval color became dark or they had no motion, and were removed daily from the test plates. On D7, no additional diet was fed, and on D8, the number of larvae with uneaten diets was recorded. From D8 to D21 (pupal stages), pupal mortality was assessed based on failed molt and failed adult molt [12,31,42]. From D16 to D21 (emergence stages), the number of newly emerged bees was observed and recorded daily. Individuals that died after emergence or fully developed bees that stayed in the cell without breaking the wax layer were considered to be emerged bees. All deformation symptoms and morphological characteristics (weight and length) of dead adult bees after emerging were observed and measured immediately. After all deformation symptoms of living emerged bees were observed on D21, the weight and length of the emerged bees were measured. The whole-body length of an adult bee was measured from the tip of its head to the tip of its abdomen. In particular, when a humpback was present in the bee, the total length of the body was measured as it was in the unstretched state. The symptoms of newly emerged adult bees were classified as surviving normal (SN), deformed wings (DW), deformed antennae (DA), and abnormal abdomen shape (AAS) (Table 2) [43]. The larval mortality, pupal mortality, adult emergence rate, and deformation rate were calculated for each group using the following formulae [31]:

Larval mortality = (the number of dead larvae from D3 to D8/the number of larvae on D3) × 100 Pupal mortality = (the number of dead pupae from D8 to D21/the number of pupae on D8) × 100 Adult emergence rate = (the number of emerged bees/the number of larvae on D3) × 100 Deformation rate = (the number of deformed bees/the number of emerged bees) × 100

**Table 2.** List and descriptions of mortality symptoms during developmental stage and deformation symptoms observed in emerged adult bees. This table is modified from Fine et al. [42] and Barbosa et al. [43].


#### *2.4. Statistical Analysis*

A statistical analysis of the data was carried out using the SPSS statistical software program (SPSS 20.0 Inc., Chicago, IL, USA). A Pearson's chi-square test was used to compare larval mortality, pupal mortality, and adult emergence rates among the three groups. A Fisher's exact test was used to assess differences between the total deformation rates of emerged bees in the three groups. The Kaplan–Meier log-rank test was used to compare the survival curves of each group. The mortality, emergence rates, and deformation rates were expressed as means ± SE. Means ± SE (standard error) were calculated for the four replicate values of each group. The emergence date, adult weight, and length were expressed as means ± SD (standard deviation). The dates of emergence, and the adult weight and length at D21, were tested by the one-way ANOVA and were determined using Tukey's HSD test to compare the values among the three groups. A *p*-value of <0.05 was considered as a statistically significant difference.

#### **3. Results**

#### *3.1. In Vitro Mortality, Adult Emergence Rates, and Survival*

Larval mortality and pupal mortality means ± SE were as follows: 4.9 ± 0.7% and 16.1 ± 2.8%, respectively, in the NHG; 7.7 ± 4.1% and 13.5 ± 3.4%, respectively, in the CHG; and 4.2 ± 2.7% and 15.3 ± 2.2%, respectively, in the CVG. The three groups satisfied the OECD test condition that the larval mortalities were less than 15% in the negative controls. On D21, total emergence rates were 79.9 ± 3.3% in the NHG, 79.9 ± 6.7% in the CHG, and 81.3 ± 0.9% in the CVG, which corresponds to the OECD test condition that the adult emergence rate should be ≥70% in the controls. No statistically significant differences were detected among the three groups with respect to larval mortality, pupal mortality, and adult emergence rates (chi-square test, *p* > 0.05, Table 3). The survival curves

of the honey-bee larvae have been illustrated in Figure 1, where no significant differences were found among the three groups in terms of their survival (Kaplan–Meier log-rank test, *p* > 0.05).

**Table 3.** Larval mortality, pupal mortality, and adult emergence rate in *Apis mellifera*. Values are means ± SE. No significant differences in mortalities and emergence rates were found among the three groups (chi-square test, *p* > 0.05).


<sup>1</sup> NHG: the non-capped and horizontally oriented groups; <sup>2</sup> CHG: the capped and horizontally oriented groups; <sup>3</sup> CVG: the capped and vertically oriented groups.

**Figure 1.** The survival curves of the honey-bee larvae of the horizontally oriented groups and the vertically oriented groups over 21 days, which were assessed by the Kaplan–Meier log-rank test (*p* > 0.05). NHG: the non-capped and horizontally oriented groups; CHG: the capped and horizontally oriented groups; CVG: the capped and vertically oriented groups.

#### *3.2. Emergence Rates by Time (Days)*

The emergence rates of each group were assessed according to the time (days). By D17, honey bees from the groups had not emerged, and then, on D18, worker bees began to emerge at 4.0% in the NHG, 0.7% in the CHG, and 1.4% in the CVG. On D19, more than half of the bees (75.7%) emerged in the NHG, and 59.0% in the CVG, compared to the 39.6% of bees that emerged in the CHG. Finally, on D21, 79.9% of bees emerged in the NHG, 79.9% of bees emerged in the CHG, and 81.3% of bees emerged in the CVG (Figure 2). The mean emergence date of the NHG was 19.00 ± 0.32 days, that of the CHG was 19.85 ± 0.93 days, and that of the CVG was 19.39 ± 0.74 days. The three groups demonstrated a statistically significant difference in emergence date (Tukey's HSD test, *p* < 0.05).

**Figure 2.** Emergence rates of each group from D15 to D21. NHG: the non-capped and horizontally oriented groups; CHG: the capped and horizontally oriented groups; CVG: the capped and vertically oriented groups.

#### *3.3. Deformation Rate of Newly Emerged Adult Bees*

The total deformation rates were 27.8 ± 7.7% in the NHG, 26.1 ± 6.9% in the CHG, and 11.9 ± 2.0% in the CVG. The deformation rates of NHG and CHG were significantly higher than those of CVG (Fisher's exact test, *p* < 0.05, Figure 3). Figure 4 illustrates examples of normal and deformed bees in the three groups. DW was shorter or had tangled ends compared to the normal wings. DA had curved ends compared to the normal antennae. Additionally, an AAS was more curved compared to the normal abdomen shape. Deformation symptoms were classified into 9 categories, including AAS, DSW, DTW, deformed with tangled wings asymmetrically (DTWA), and DA. In the NHG, the deformation rates were AAS (13.9%), DSW (0.9%), DTW (0.9%), DTWA (3.5%), AAS + DW (4.3%), DW + DA (0.9%), and AAS + DW + DA (2.6%). In the CHG, the deformation rates were AAS (4.3%), DSW (4.3%), DTW (3.5%), DTWA (6.1%), DA (0.9%), AAS + DW (4.3%), AAS + DA (1.7%), DW + DA (0.9%), and AAS + DW + DA (0.9%). In CVG, the deformation rates were AAS (3.4%), DSW (1.7%), DTW (1.7%), DTWA (3.4%), DA (0.9%), and DW + DA (0.9%) (Table 4).

**Figure 3.** Deformation rates of newly emerged adult bees. The different letters above the bars indicate significant differences among the three groups (Fisher's exact test, *p* < 0.05). NHG: the non-capped and horizontally oriented groups; CHG: the capped and horizontally oriented groups; CVG: the capped and vertically oriented groups.

**Figure 4.** The observed symptoms of newly emerged adult bees. (**A**) Surviving normal (SN); (**B**) Deformed with short wings (DSW); (**C**) Deformed with tangled wings (DTW); (**D**) Normal (above) and abnormal abdomen shape (AAS) (below); (**E**) Deformed antennae (DA). Scale bar = 1 mm.

**Table 4.** The overall percentage of observed symptoms of newly emerged adult bees. Each percentage in the following table is a calculated value of the ratio of the number of deformed bees to the total number of emerged bees.


<sup>1</sup> SN: survived and successfully eclosed as bees; AAS: abnormal abdomen shape; DSW: deformed with short wings; DTW: deformed with tangled wings; DTWA: deformed with tangled wings asymmetrically; DA: deformed antennae; DW: deformed wings; <sup>2</sup> NHG: the non-capped and horizontally oriented groups; <sup>3</sup> CHG: the capped and horizontally oriented groups; <sup>4</sup> CVG: the capped and vertically oriented groups.

The main deformation symptoms were simplified into three categories as follows: AAS, DW and DA. The percentage of SN, AAS, DW, and DA were 73.1%, 17.0%, 8.7%, and 1.3%, respectively, in the NHG; 73.0%, 7.7%, 16.8%, and 2.5%, respectively, in the CHG; and 88.0%, 3.4%, 7.3% and 1.3%, respectively, in the CVG (Figure 5).

**Figure 5.** Percentage of observed symptoms of emerged bees. SN: survived and successfully eclosed as bees; AAS: abnormal abdomen shape; DW: deformed wings; DA: deformed antennae.

#### *3.4. Body Weight and Length of Newly Emerged Adult Bees*

The mean weight and length of the emerged bees were 67.40 ± 12.24 mg and 10.88 ± 0.96 mm in the NHG, 69.77 ± 12.63 mg and 10.82 ± 0.92 mm in the CHG, and 71.95 ± 12.12 mg and 11.55 ± 1.00 mm in the CVG, respectively. Adult weights of CVG were significantly higher than those of NHG (*p* = 0.014). Additionally, the length of the emerged bees in the CVG was significantly larger than that of the NHG and CHG (one-way ANOVA, *F*(2, 344) = 21.072, *p* < 0.05, Table 5).

**Table 5.** The body weight and length of the newly emerged adult bees. Values are means ± SD. Means followed by the different letters across a row are significantly different (one-way ANOVA, *p* < 0.05).


<sup>1</sup> NHG: the non-capped and horizontally oriented groups; <sup>2</sup> CHG: the capped and horizontally oriented groups; <sup>3</sup> CVG: the capped and vertically oriented groups.

#### **4. Discussion**

Many studies have emphasized the importance of rearing honey-bee larvae in vitro for testing the toxicity of pesticides and, sequentially, experimental rearing methods have been systematically developed [12,18,32,44]. Natural or commercial hives consist of several honeybee combs with vertical structures for brood rearing and storing honey-bee products [45]. Conversely, as displayed in the OECD guidelines [30,31], many laboratories have generally performed experiments placing the rearing plates horizontally during all developmental stages (from larvae to adults). However, the horizontal rearing plates may induce deformations in emerged bees [32,41]. The purpose of this study was to analyze the effects of vertical and horizontal rearing plates on the emergence rates and deformation rates of newly emerged bees.

Honey bee larval defecation is usually on D7 after grafting [21]. Several recent studies suggest that the larvae should be transferred to a new clean plate, since larval mortality

may increase due to the defecation of the larvae [12,17,38]. However, transferring larvae may cause mechanical stress or contamination of the larvae [39]. Besides, when feeding the total diet of 160 μL, the larvae eat all the food provided, thus, it is not necessary to move the larvae to new cell cups or clean the grafting cell cups [32]. Brodschneider et al. [41] placed rearing plates capped with thin wax layers vertically on D11. As a result, the total mortality until emergence was 16.3% in the control groups; they also reported similar flight performance between reared bees in vitro and hives. Likewise, Krainer et al. [46] demonstrated a total mortality of 28.1% in the control groups. They sealed rearing plates with the wax layer and the plates were placed vertically on D12. In our experiments, there were no significant differences in larval mortality, pupal mortality, and adult emergence rates among the three groups. The total adult emergence rates were about 80% in the three groups. Thus, each group exhibited similar survival rates of bee larvae regardless of plate position (horizontal and vertical).

For the most part, the worker bees reared in vitro emerged on D17–D18 after grafting [20]. In our experiments, bees of each group began to emerge from D18. Bees in the NHG were the earliest to emerge among the three groups. These differences could be due to external stimulus by the newly emerged bees that roam and stimulate other non-emergent pupae in the cells of 48-well tissue plates. Namely, bees in CHG and CVG may emerge more slowly, since the wax layer of the plate could interrupt this external stimulus. In particular, the CHG was the slowest among the three groups. This may be because more force is needed to break through the wax layer against gravity on the horizontal plate than on the vertical plate.

Tehel et al. [47] inoculated the honey-bee pupa with deformed wing virus (DWV) and tested the relative effects of the genotype of DWV on the mortality and wing malformation of adult honey bees. They placed the plate vertically so that pupae were horizontal in the incubators and monitored the pupal development. They observed that 23% of emerged bees in the control had wing deformities. In our experiments, the deformations of adults included deformed wings, deformed antennae, and abnormal abdomen shape, and horizontally oriented groups (NHG and CHG) demonstrated higher deformation rates (27.8% and 26.1%) than the vertically oriented groups (11.9%). In particular, the NHG and CHG had more wing or abdominal deformations than the CVG. Although the cause of deformation in emerged bees is not clear, all adults with deformations were derived only from deformed pupae [43]. Additionally, the wings of honey bees are formed during the pupal development stage [48]. When the rearing plates were horizontal, the pupae in the cell cups hold a vertical state. Thus, the pupal body is pulled down by gravity, affecting the wing and abdominal development [41]. In this regard, deformed bees may already have external deformations from the pupal stage.

In other studies, the rearing plates were capped and set vertically on D11 after grafting when pupation started, but in our experiments, the plates were capped with a wax layer on D15 to observe the mortality during the pupal stages. For this reason, it is thought that the horizontal condition between D11 and D14 (before capping the plates with wax layer) had already affected the pupae, resulting in abdominal and wing deformations in the CVG. Mechanisms for explaining antennae deformations due to physical external deformation have not been described in other studies.

The average adult weight ranged from 67.40 to 71.95 mg in our experiments; similarly, Brodschneider et al. [41] measured 76.6 ± 11.6 mg in emerged bees of the control groups. The average adult weight of the CVG was significantly higher than that of the NGH. The emerged bees of the CVG were significantly larger by 0.7 mm than the NHG and CHG, and this difference appeared to be due to the abnormal shapes, such as the humpback and abdominal shrinkage that were observed in bees reared horizontally.

Barbosa et al. [43] reported that when azadirachtin and spinosad were treated on the stingless bee, *Melipona quadrifasciata*, deformed pupae and emerged bees with wing, antennae, and leg deformities occurred. Additionally, they reported that deformed bees had side effects regarding flight activity or olfactory activity. Therefore, the deformations

occurring in emerged bees have the potential to directly affect the activity of worker bees, and thus, these deformations can also be evaluated as sublethal effects [43]. In the present study, the mean adult deformation rates in CVG were approximately 12%. Thus, the vertical rearing method can be supported as a more appropriate method to verify the effects of pesticides on honey bees by considering the deformation rates in the control group. In future studies, the mechanisms of deformations in emerged honeybees that were identified here should be investigated.

#### **5. Conclusions**

Overall, there were no statistically significant differences in the emergence rates of adult bees between the horizontal and vertical plates, but the total deformation rates of the horizontal plates were significantly higher than those of the vertical plates. Our results are the first to discuss the emergence rates and deformation rates of honey bees concerning the position of plates in a laboratory. Considering these conclusions, the vertical rearing method with lower adult deformation rates appears to be more suitable, when considering the deformation rates of the control groups in order to verify the sublethal effects of pesticides on the bees. In the honey-bee larval toxicity test, according to the OECD guidelines, it is necessary to confirm the pupal mortality at the pupal stages (from D8 to D21). However, the vertical rearing plates must be capped with a wax layer on D15, so it may be difficult to check the pupal mortality after D15. Consequently, the rearing conditions and the position of rearing plates should be carefully considered depending on the purpose of the larval toxicity tests.

**Author Contributions:** Conceptualization, J.K. and K.C.; methodology, J.K. and K.C.; validation, J.K. and K.C.; formal analysis, J.K.; investigation, J.K., B.-S.K. and C.-Y.Y.; funding acquisition, K.C.; resources, Y.-S.C.; data curation, J.K.; writing—original draft preparation, J.K.; writing—review and editing, K.C.; visualization, J.K.; supervision, K.C., J.-A.O. and H.-H.P.; project administration, K.C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was supported by a grant from the Research Program for Agricultural Science & Technology Development (Project No. PJ015797), National Institute of Agricultural Science, Rural Development Administration, Korea. It was also partly supported by 2021 collaborative research program between the university and Rural Development Administration, Korea.

**Institutional Review Board Statement:** Not applicable.

**Acknowledgments:** The authors wish to thank the beekeepers of the Sericulture and Apiculture Division for their help.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design, conduct, investigation, analysis, or writing of the study.

#### **References**


### *Article* **Characterization of Thermal and Time Exposure to Improve Artificial Diet for Western Corn Rootworm Larvae**

**Man P. Huynh 1,2,\*, Adriano E. Pereira 1, Ryan W. Geisert 1, Michael G. Vella 3, Thomas A. Coudron 4, Kent S. Shelby <sup>4</sup> and Bruce E. Hibbard <sup>5</sup>**


**Simple Summary:** The western corn rootworm is a highly adaptive pest that has evaded nearly all management tactics developed to date. Antibiotics have been utilized in rootworm diets to mitigate bacterial contamination. However, antibiotic ingestion necessarily alters rootworm gut microbiota, clouding the outcome of diet toxicity bioassays used in determination of rootworm susceptibility to insecticides. Rapid heating, or pasteurization, is one of the most widely applied techniques to alleviate microbial contamination and could eliminate antibiotics from the diet. We characterized effects of temperatures and time intervals of thermal exposure on quality of rootworm diet by measuring larval weight, molting, and survival. Our results demonstrated non-linear effects of thermal exposure on the performance of diet, whereas no impacts were observed on the exposure intervals evaluated. These findings will guide the continued development of sterilized rootworm diets, facilitating mass production and provide insights into the design of diets for other insects.

**Abstract:** The western corn rootworm (WCR), *Diabrotica virgifera* LeConte, is the most serious pest of maize in the United States. In pursuit of developing a diet free of antibiotics for WCR, we characterized effects of thermal exposure (50–141 ◦C) and length of exposure on quality of WCRMO-2 diet measured by life history parameters of larvae (weight, molting, and survival) reared on WCRMO-2 diet. Our results indicated that temperatures had non-linear effects on performance of WCRMO-2 diet, and no impacts were observed on the length of time exposure. The optimum temperature of diet processing was 60 ◦C for a duration less than 30 min. A significant decline in development was observed in larvae reared on WCRMO-2 diet pretreated above 75 ◦C. Exposing WCRMO-2 diet to high temperatures (110–141 ◦C) even if constrained for brief duration (0.9–2.3 s) caused 2-fold reduction in larval weight and significant delays in larval molting but no difference in survival for 10 days compared with the control diet prepared at 65 ◦C for 10 min. These findings provide insights into the effects of thermal exposure in insect diet processing.

**Keywords:** *Diabrotica virgifera*; corn rootworm; WCRMO-2; diet processing; heating

#### **1. Introduction**

The western corn rootworm (WCR), *Diabrotica virgifera virgifera* LeConte (Coleoptera: Chrysomelidae), is the most serious pest of maize in the United States and some parts of Europe [1], causing 1 to 2 billion dollars (USD) in losses and control costs to U.S. maize growers each year [2]. Most damage associated with this species is the result of larval feeding on maize roots [3,4], though yield reduction of maize can result from adult

**Citation:** Huynh, M.P.; Pereira, A.E.; Geisert, R.W.; Vella, M.G.; Coudron, T.A.; Shelby, K.S.; Hibbard, B.E. Characterization of Thermal and Time Exposure to Improve Artificial Diet for Western Corn Rootworm Larvae. *Insects* **2021**, *12*, 783. https://doi.org/10.3390/ insects12090783

Academic Editor: Allen Carson Cohen

Received: 16 August 2021 Accepted: 29 August 2021 Published: 1 September 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

feeding on silks, pollen, kernels, and foliage of maize plants [5]. Management of WCR has been a challenge because this highly adaptive insect has evolved resistance to several management strategies, including chemical insecticides [6–9], transgenic maize hybrids expressing insecticidal crystalline toxins from *Bacillus thuringiensis* (Bt) Berliner [10–15], and cultural control techniques, such as crop rotation [16,17].

Given a history of developing resistance to nearly every management tactic utilized for managing WCR, a logical concern exists that WCR will possibly develop resistance to newer management tactics. To slow resistance development of this pest, the U.S. Environmental Protection Agency (EPA) has mandated monitoring resistance programs that involve annual collections of insect populations in regions of high adoption of the targeted trait followed by bioassays to determine potential reduction in susceptibility attributable to resistance development [18]. Diet assays, whereby insects are exposed to toxins in an artificial diet, that can be used in conjunction with on-plant assays to evaluate the susceptibility of WCR to insecticides are critical components of the resistance-monitoring programs [19–21].

An artificial diet capable of supporting WCR larval growth and development similar to those fed on maize roots would be greatly beneficial for research programs. Artificial diet development for *Diabrotica* spp. was initially conducted on the southern corn rootworm, *Diabrotica undecimpunctata howardi* Barber, as a model species [22,23] because diet work began prior to availability of the non-diapausing strain of WCR. An attempt to develop a diet for WCR rearing occurred in 2002 [24]. Later, Huynh et al. [25] developed an improved WCR diet (WCRMO-1) that was an optimization of the ingredients in the initial WCR diet [24]. The WCRMO-1 formulation is compatible with each of the four marketed Bt toxins targeting WCR [26]. However, both published WCR diets [24,25] require maize root powder, which is not available for purchase, thereby limiting the practical use of the diets. Recently, Huynh et al. [27] successfully developed a WCR diet without maize root powder (WCRMO-2) that supports performance of WCR larvae equal to or better than that of publicly available formulations [27,28]. The WCRMO-2 formulation was specifically designed to require only commercially available ingredients. This formulation supported approximately 97% of larvae for survival, molting, and increased larval weight gain after 10 days of feeding by 4-fold compared with the WCRMO-1 diet [27]. Both WCRMO-1 and WCRMO-2 diets have essentially zero microbial contamination [27], similar to the four proprietary diets previously used for WCR bioassays [26,28], through clean laboratory practices described previously [25]. The WCRMO-2 diet is now available commercially (WCRMO-2, Frontier Scientific Services, Newark, DE, USA).

Antibiotics are commonly used in insect diets for preventing bacterial contamination. Ingestion of antibiotics in insect diets has been reported to alter response of lepidopteran insects to Bt proteins due to its negative effects on insect gut microbiomes [29,30]. In fact, Paramasiva, Sharma, and Krishnayya [29] manipulated antibiotics to suppress gut microflora in the cotton bollworm (*Helicoverpa armigera* Hübner) and found that *H. armigera* larvae fed on larval diet with antibiotics had lower susceptibility to a commercial formulation of Bt and purified δ-endotoxins Cry1Ab and Cry1Ac compared with *H. armigera* larvae fed on larval diet without antibiotics. Additionally, pretreatment of *H. armigera* larvae with antibiotics to eliminate the gut microbes resulted in a decrease in larval mortality and an increase in the larval weight gain when *H. armigera* larvae were exposed to activated Cry1Ac, Bt formulation, and transgenic cotton [30]. In WCR, antibiotics have been reported to cause effects on the indigenous symbionts that contribute the development of resistance in WCR to crop rotation [31]. Differences in the abundance of multiple bacterial taxa (e.g., *Acinetobacter* sp., *Pseudomonas* sp., *Enterobacter* sp., *Lactococcus* sp.) in WCR gut microbiota were correlated with increased resistance to soybean-defense compounds. Gut microbiota of WCR derived from rotation-resistant populations and wild-type populations had differences in the abundance of *Klebsiella* sp., *Stenotrophomonas* sp., *Enterobacter* sp., and *Lactococcus* sp. [31]. The authors suppressed gut microbiota of rotation-resistant WCR by antibiotic treatments, finding that the resistance to the soybean-defensive compound of the resulting rotation-resistant WCR was reduced to a level similar to that of wild-type

WCR. Paddock et al. [32] reported that feeding on maize expressing a Bt toxin resulted in a shift in the gut microbiota in susceptible WCR larvae but led to no changes in the bacterial community within resistant insects. Currently, published diet formulations for *Diabrotica* spp. contain antibiotics (streptomycin and chlortetracycline) [22,24,27,33] that are widely used in other insect diets as the important antibacterial agents [34]. It is likely that these antibiotics have potential impacts on WCR gut microbiomes, though these effects of antibiotics on WCR larvae are not adequately characterized. The use of a diet free from antibiotics for WCR bioassays would likely provide more accurate phenotypic picture of test populations, as it relates to susceptibility to Bt toxins and other insecticide compounds. Therefore, development of a WCR diet free of antibiotics would facilitate resistance-monitoring programs as well as other research programs of this pest.

One of the most widely used techniques to alleviate microbial contamination in insect diets is thermal treatment as a means of diet preservation [34]. During diet processing, agar-based diets are often exposed to elevated temperatures ranging from mild blanching (50–100 ◦C) to high temperatures above the boiling point of water. These high-thermal treatments are typically manipulated with pressure, which can lead to the destruction of microbial contaminants. Flash sterilization is one of the heating techniques that utilizes high temperatures for a brief time of thermal exposure (less than 1 min) that sufficiently kills microbial contaminants in the diet [34] while minimizing the effects of overheating on heat-sensitive nutrients (e.g., vitamins, amino acids, or lipids) [35]. As a step toward development of a diet free of antibiotics for WCR, we investigated the effects of high temperatures (80–141 ◦C) for short time exposure (0.8–2.3 s) on quality of WCRMO-2 diet using a flash sterilization approach. Additionally, since the effects of using relatively low temperatures on the performance of insect diets have not been adequately explored, we further characterized the effects the mild thermal exposure (50–80 ◦C) and extended time of thermal exposure (5–30 min), which are commonly used in insect diet preparation, on the quality of WCRMO-2 diet. The quality of WCRMO-2 diet was assessed via the evaluation of life history parameters (weight, molt, and survival) of WCR larvae reared on the diets for 10 days.

#### **2. Materials and Methods**

#### *2.1. Insects and Egg Treatment*

Eggs of non-diapausing WCR populations were obtained from the USDA-ARS-Plant Genetics Research Unit (PGRU) laboratory in Columbia, MO. Egg plates containing WCR eggs and soil in Petri dishes were incubated at 25 ◦C in complete darkness until ~5% of eggs were hatched. Subsequently, the eggs were washed from soil with tap water and then were surface-sterilized using undiluted Lysol (Clean & Fresh Multi-Surface Cleaner, Reckitt Benckiser, Parsippany, NJ, USA) for 3 min and followed by 10% formaldehyde (HT501128, Sigma Aldrich, St. Louis, MO, USA) for 3 min, as described previously [24,36]. The eggs were dispensed onto coffee filter paper (Pure Brew, Rockline Industries, Sheboygan, WI, USA) placed inside a 16 oz. cup (11.7 × 7.62 × 9.6 cm, DM16R-0090, Solo Cup Company, Lake Forest, IL, USA) with a lid (LG8RB-0090, Solo Cup Company) using a 1-mL disposable pipette (13-711-9a, Fisher Scientific, Pittsburg, PA, USA). The eggs were then incubated at 25 ◦C in darkness. Larvae that hatched within 24 h were used for insect bioassay.

#### *2.2. Experimental Approach*

A series of experiments were performed to determine the effects of thermal and time exposure on the quality of WCRMO-2 diet. WCRMO-2 diet was made using WCRMO-2 dry mix (WCRMO-2, Frontier Scientific Services) at different temperatures and varying lengths of time of exposure to the different temperatures. The WCRMO-2 dry mix (Frontier Scientific Services) consists of ingredients, diet preservatives (antifungal and antibacterial agents), and agar that were published previously [27]. The quality of WCRMO-2 diet was evaluated by life history parameters of WCR larvae (weight, molt, and survival) reared on the diet treatments for 10 days.

Two-level factorial designs were constructed to determine the effects of two factors (temperature and time of thermal exposure) on the quality of WCRMO-2 diet. Two experiments were performed to evaluate high temperatures and brief time of thermal exposure (80–141 ◦C and 0.8–2.3 s) and mild temperatures and extended time of thermal exposure (50–80 ◦C and 5–30 min). The high temperature and the brief time of thermal exposure (141 ◦C for 2 s) was previously used for flash sterilization of a larval diet for *Trichoplusia ni* (Hubner) (Lepidoptera: Noctuidae) [37]. Since preliminary observations indicated that time of thermal exposure up to 10 min at 65 ◦C had no effects on the quality of WCRMO-2 diet, the extended time of thermal exposure up to 30 min was selected to further determine the effects of the longer duration of thermal exposure. The experimental designs for each experiment were generated with Design-Expert (Stat-Ease, Inc., Minneapolis, MN, USA). All designs consisted of 11 design points (diet treatments at different thermal exposure and time of exposure) with 5 model, 3 lack of fit, and 2 pure error degrees of freedom. In all experiments, WCRMO-2 diet made using the standard protocol according to the manufacturer's procedure, with a temperature of 65 ◦C and time of exposure for 10 min [27], was included as a control (Table 1).

**Table 1.** Diet treatments used in 2-factorial experiments to rear western corn rootworm larvae.


A flash sterilization system (Frontier Scientific Services) was utilized to produce WCRMO-2 diet at high thermal exposure (80–141 ◦C) and short time exposure (0.8–2.3 s). We observed consistent failures in the ability of the powdered agar to melt into solution during the brief time and rather high thermal exposure experienced in treatments with temperatures less than 88 ◦C (Table 1). This is likely due to the physical limitations of the agar (7060, Frontier Scientific Services) used to operate adequately at the mild temperatures tested. This resulted in a "soft-set" of the media, providing an inadequate matrix to support larval growth, leading to the death of the majority of larvae and increased bacterial contamination after 4 days post infestation. Low-melt agar (Frontier Scientific Services) was tested as a substitution of the agar used initially. However, the "soft-set" of the media was observed when the media were prepared with the low melt agar at the temperatures ≤ 65 ◦C for small windows of time (0.9–2.1 s). To further study the effects of mild thermal exposure (50–80 ◦C) on the quality of WCRMO-2 diet, an alternative approach involved completely melting the agar using a microwave (51101BZ, Hamilton Beach, Glen Allen, VA, USA) and cooling it to designed mild temperatures prior to adding it to the treated media was evaluated. The temperatures were then held for designed time exposure (5–30 min) using a hot plate (CimarecTM, Thermo Scientific, Waltham, MA, USA) (Table 1).

#### **3. Diet Preparation**

At high thermal exposure (80–141 ◦C), WCRMO-2 diet was made using a flash sterilization system (Frontier Scientific Services). Frontier's sterilizer consists of a circulating system of processing tubes that alternates between linear stretches of tubing and coiled spiraltherms that act as heat exchangers (Figure 1). To make 10 L of WCRMO-2 diet, 1.49 Kg of WCRMO-2 dry mix (WCRMO-2, Frontier Scientific Services) and 158 g of agar (7060, Frontier Scientific Services) were added to 9.26 L of cool tap water (~21 ◦C). The diet mixture was then pumped into the process lines from a tank and controlled via a metering pump that manages line pressure and flowrate of the product. The time of thermal exposure was set by manipulating the flowrate. The first heat exchanger was jacketed with hot oil set to a desired temperature. As the product flows through the coiled spiraltherm tubing, it rapidly gains thermal units from the hot oil jacketing the line. After departing the heated coil, the process line enters a series of linear switch-backs so that some excess heat is lost to atmosphere before reaching the cooling spiraltherm. At this point, the product enters a heat exchanger jacketed in cold (room temperature) water to rapidly lose thermal units and reach a desired dispense temperature. At the dispense temperature of 60 ◦C, 289.5 mL of 10% KOH (*w*/*v*) (F7633, Frontier Scientific Services) was added into the diet mixture to adjust the pH of the diet to 9, and the resulting diet solution was blended for 30 s to mix thoroughly. Subsequently, the diet solution was dispensed into a 96-well plate (3370, Corning Inc., Corning, NY, USA) using a repeater pipette (200 μL per well) in a biological cabinet (Forma 1800 Series Clean Bench, Thermo Scientific). The diet plate was then allowed to evaporate excess moisture for 10 min, stored in a refrigerator at 7 ◦C, and used for assays within 3 days.

**Figure 1.** Frontier Scientific's flash sterilization system used for the processing of insect media. From left to right: diet tank, pump, processing tubes, coiled spiraltherms, and control station.

At mild thermal exposure (50–80 ◦C), WCRMO-2 diet was made using WCRMO-2 dry mix (WCRMO-2, Frontier Scientific Services) according to the manufacturer's procedure [27]. To prepare 1 L of WCRMO-2 diet, agar (15.8 g) was added to 926 mL of purified water, and the solution was brought to a full boil using a microwave (Hamilton Beach) until agar was completely melted. The agar solution was then transferred to a blender placed in a biological safety cabinet (SG403, SterilGARD® III Advance cabinet, Sanford, ME, USA). When the agar solution cooled to designed temperatures (Table 1), 148.9 g of WCRMO-2 dry mix was added, and the mixture was blended for 10 s to mix thoroughly. Subsequently, 28.95 mL of 10% KOH (*w*/*v*) (P250, Fisher Scientific, Fair Lawn, NJ, USA) was added to increase the pH of the diet to 9, and the resulting diet solution was blended for 10 s to

mix thoroughly. The diet was poured into a 1-L glass beaker containing a stir-bar and placed on a stirring hot plate (CimarecTM, Thermo Scientific) set at the designed temperatures. The temperatures of diet solution were monitored using an infrared thermometer (IR002, Ryobi, Fuchu, Hiroshima, Japan) and held at the test temperatures for the designed times (5–30 min) using the hot plate. Subsequently, the diet solution was dispensed into a 96-well plate (3370, Corning Inc.) using a repeater pipette (200 μL per well), allowed to evaporate excess moisture for 10 min, stored in a refrigerator at 4 ◦C, and used for assays within 3 days.

#### *3.1. Insect Artificial Diet Bioassays*

The diet bioassays were conducted as described previously [25]. All materials used in the diet assays were surface-treated via exposure to UV light for 10 min in a biological cabinet (SterilGARD® III Advance cabinet). Each diet treatment, which is WCRMO-2 diet made at different temperatures and time of thermal exposure (Table 1), was randomly assigned to a 12-well row of the 96-well plate and replicated at least 3 times in different diet plates. Each well was infested with one WCR neonate (<24 h old) using a fine paintbrush. A sealing film (TSS-RTQ-100, Excel Scientific, Inc., Victorville, CA, USA) was used to cover the plate. For ventilation, a hole was made in the sealing film over each well using a number zero insect pin. The plates were kept in an incubator (Percival, Perry, IA, USA) at 25 ◦C in darkness for 10 days. Larval molting was recorded daily during the experiments, whereas larval weight, survival, and evidence of contamination were recorded at the end of the experiments. For larval dry weight, all live larvae in each treatment were pooled per replicate (12 possible) into 95% ethanol, dried in an oven (Binder GmbH, Tuttlingen, Germany) at 55 ◦C for 48 h, and weighed using a micro balance (MSU6.6S-000-DM, Sartorius Lab Instruments GmbH & Co. KG, Goettingen, Germany).

#### *3.2. Data Processing and Statistical Analyses*

Survival and molting data were calculated by dividing the number of live larvae and successful larval molt from 1st to 2nd instar and from 1st to 3rd instar per replicate, respectively, by the initial number of larvae infested and multiplying by 100 to obtain percentages. Weight per larva (mg) was determined by dividing the dry weight by the number of larvae that survived per replicate.

For the experiment with high temperatures (80–141 ◦C) and short time of thermal exposure, the diet treatments that resulted in a "soft-set" of the media were excluded in the analyses because the soft-set form led to the death of the majority of larvae after 4 days post infestation. Because of the exclusion of these treatments, the remaining data points were not adequate to generate the response surface models of the measured responses. The remaining data were analyzed as a completely randomized experiment. The data were analyzed with analysis of variance (ANOVA) using PROC MIXED in SAS 9.4 (SAS Institute, Cary, NC, USA). Diet was the fixed effect, and replication was the random variable. Differences between the remaining treatments were determined using Fisher's least significant difference (LSD) at *p* < 0.05. The percent variables (survival and molting) were arcsine square-root transformed prior to the analysis to meet assumptions of normality and homoscedasticity, whereas untransformed data were presented as mean ± SEM.

For the experiment with mild temperatures (50–80 ◦C) and extended time of thermal exposure, polynomial equations were generated to describe the impact of two factors (temperature and time of thermal exposure) on the measured responses (larval weight, molting, and survival). The best fit model for each measured response was selected from all possible models from linear to quartic polynomials generated with Design Expert® (Stat-Ease, Inc., Minneapolis, MN, USA). Model selection was based on several criteria, including low model *p*-value, lack of fit *p*-value, low standard deviation, high R-values, and a low PRESS value [38,39]. Once more than one satisfactory model was generated, adequacy tests were performed to further evaluate the selected model, as described previously [40].

#### **4. Results**

#### *4.1. Diet Quality with High Thermal Exposure and Short Time of Thermal Exposure*

Exposure to high temperatures (110–141 ◦C), even if constrained to a small window of time (0.9–2.3 s), had significant deleterious effects on the quality of WCRMO-2 diet for feeding WCR larvae. The WCRMO-2 diet exposed to the high temperatures for the brief time exposure resulted in larval weight significantly smaller compared to the control diet, WCRMO-2 diet made at a temperature of 65 ◦C for 10 min (*p* < 0.0001, *F6,13* = 15.88. Figure 2a). Average larval dry weight on the diet treatments ranged from 0.30 mg to 0.39 mg, while average dry weight on the control diet was 0.71 mg, an approximately 2-fold difference. There was no significant difference in dry weight on all diet treatments. Significant delays in larval molt to the 2nd instar were observed when they were reared on all diet treatments compared with the control diet (Figure 2b). There were significantly fewer 2nd instar larvae on the diet treatments than on the control diet after 5 days post infestation when larvae began to molt to 2nd instar. At day 5 post infestation, 20.6% of larvae had molted to 2nd instar on the control diet, whereas nearly 0% of 2nd instar larvae on all diet treatments (*p* < 0.0001, *F6,32* = 37.60). At day 7 post infestation, nearly 100% of the larvae had molted to 2nd instar on the control diet, significantly higher than that of all diet treatments (*p* < 0.0001, *F6,32* = 19.82). There was no significant difference in percent larvae molted to 2nd instar in the diet treatments (range from 39.3%–48.4%) at day 7 post infestation. No larvae that had molted to the 3rd instar were observed on the diet treatments by 10 days. Larval survivorship on all diet treatments ranged from 95.7% to 100%, which was not significantly different from survivorship on the control diet (*p* = 0.7764, *F6,13* = 0.53, Figure 2c).

**Figure 2.** Effects of high thermal exposure (110–141 ◦C) and short time intervals of thermal exposure (0.8–2.3 s): dry weight (**a**), percent successful completion of molt to 2nd instar (**b**), and percent survival (**c**). Western corn rootworm larvae were reared on WCRMO-2 diet for 10 days. Means with bars followed by different letters are significantly different (*p* < 0.05). Means ± SEM.

#### *4.2. Diet Quality with Mild Thermal Exposure and Extended Time of Thermal Exposure*

The two-level factorial experiment yielded significant response surface models for larval weight (*p* < 0.0015, *F3,7* = 16.27), molt to 2nd instar (*p* = 0.0410, *F3,7* = 4.76), molt to 3rd instar (*p* = 0.0100, *F3,7* = 8.46) and a marginally significant model for survival (*p* = 0.0531, *F3,7* = 4.23) (Table 2). Models for weight, molt to 3rd instar, and survival had insignificant lack of fit, whereas there was a significant lack of fit for molt to 2nd instar due to a very small value of pure error. The relationships between the two factors tested (mild thermal exposure and extended time of exposure) and the quality of WCRMO-2 diet were revealed in contour plots (Figure 3). The contour plots generated from the models of responses measured (weight, survival, and molting) displayed the performance of larvae when reared on the WCRMO-2 diet prepared at the mild temperatures for the extended duration. The magnitude of the response variables is coded in color and can be envisioned as perpendicular to the page, as indicated by labelled isobars.

**Table 2.** *p*-values, regression coefficients, and response surface-model fitting diagnostic statistics for western corn rootworm responses to a 2-factorial experiment at mild thermal exposure (50–80 ◦C) and extended time of thermal exposure (5–30 min). A: time of thermal exposure, B: temperature.


**Figure 3.** *Cont*.

**Figure 3.** Effects of mild thermal exposure (50–80 ◦C) for extended times of thermal exposure (5–30 min). Contour plots of dry weight (**a**), percent successful completion of molt to 2nd instar (**b**) and percent successful completion of molt to 3rd instar (**c**), and percent survival (**d**). Western corn rootworm larvae were reared on WCRMO-2 diet for 10 days. Color bars display the magnitude of the measured responses.

Models for weight, molt to 2nd instar, and molt to 3rd instar revealed that temperature had significant effects on these measured responses, while no significant effect on weight and molt was observed due to time of thermal exposure. *p* values of temperature were 0.0012, 0.0224, and 0.0051 in the models for weight, molt to 2nd instar, and molt to 3rd instar, respectively (Table 2). Non-linear effects of temperature on weight, molt to 2nd instar, and molt to 3rd instar were found (Figure 4). WCRMO-2 diet prepared a temperature of approximately 60 ◦C yielded the maximum larval weight and percent molt, whereas there were significantly negative effects on weight and molting when WCRMO-2 diet was produced at temperatures above 75 ◦C (Figures 3a–c and 4).

**Figure 4.** *Cont*.

**Figure 4.** Nonlinear effects of mild thermal exposure (50–80 ◦C) for 15 min on dry weight (**a**), percent successful completion of molt to 2nd instar (**b**), and percent successful completion of molt to 3rd instar (**c**). Western corn rootworm larvae reared on WCRMO-2 diet for 10 days.

A model for survival indicated marginally significant effects of time exposure (*p* = 0.0582) and the interaction between thermal exposure and time exposure (*p* = 0.0596). However, larval survivorship on all diet treatments was >95% (Figure 3d). A similar pattern was found in the experiment with high thermal exposure and short time of thermal exposure (Figure 2c). Consequently, survival was not considered as an important criterion for evaluation of the effects of the treatments.

#### *4.3. Contamination*

All experiments had minor contamination (<3%) except for diet treatments that experienced a soft-set of the media that were excluded from the analyses. No evidence for a relationship between contamination and two experimental factors (thermal exposure and time of thermal exposure) was determined. Similarly low contamination rates were observed previously [27].

#### **5. Discussion**

Published WCR diets require antibiotics as diet preservatives to alleviate bacterial contamination [24,25,27]. However, ingestion of antibiotics results in changes in WCR gut microbiota [31], thereby possibly interfering with the determination of the susceptibility of WCR to insecticide toxins using diet bioassays. The availability of a WCR diet free of antibiotics would facilitate research programs of this important pest. In diet processing, heating is required as one of the most important parts for activating gelling agents (e.g., agar) to stabilize diets and promote the form of suitable textures for insect feeding. This can be also used as an extremely effective means for destroying microbial contaminants derived from diet ingredients and preparation [34]. To further the goal of developing a diet free of antibiotics for WCR, we explored the effects of the thermal exposure from 50–141 ◦C for the short time (<3 s) and extended time (10–30 min) on the quality of WCRMO-2 diet based on life history parameters of WCR larvae fed on the treated diets. By using geometric and mathematical approaches, we further characterized the influence of both thermal exposure and time of thermal exposure, allowing determination of the optimum conditions (temperature and time) for making WCRMO-2 diet.

Our results indicated that the exposure to the high thermal exposure (110–141 ◦C) for brief intervals (0.9–2.3 s) caused detrimental effects on the performance of larvae on

WCRMO-2 diet, indicating a possible reduction in nutrients required for WCR growth and development due to the high thermal exposure even if constrained for the short duration. WCR larvae fed on the treated diets exhibited significant reductions in weight and molting compared to those fed on the control diet made at the mild temperature of 65 ◦C for 10 min, whereas no difference in larval survivorship was observed between the diet treatments and the control diet. Heating is known to provide many benefits, including destruction of microbial contaminants, activation of gelling agents, increasing protein digestibility, denaturation of digestive inhibitors and harmful enzymes (e.g., phenol oxidases, lipo-oxygenase), increasing flavor, and acceleration of desirable chemical reactions [34]. However, severe overheating can cause reduction of protein digestibility, nutrient destruction (e.g., ascorbic acid, unsaturated lipid), destruction of vitamins, and formation of complexes (sugar amino acid products) [41]. A similar pattern of negative effects of heating on the quality of insect diets was previously reported for a diet of *Chrysoperla carnea* Stephens (Neuroptera: Chrysopidae) [42]. Initially, Vanderzant [43] developed a liquid diet for *C. canea* larvae that contains soy and casein hydrolysates as the main protein sources along with fructose, vitamins (ascorbic acid, B-vitamins), and other diet ingredients. This formulation was successfully used for the production of 18 generations of *C. canea*, but it did not promote the growth of *C. canea* after autoclaving at 121 ◦C [42]. The author successfully developed an improved diet that can be autoclaved by removing ascorbic acid (a heat-sensitive ingredient), replacing fructose with sucrose, and adding yeast hydrolysate and casein. These substitutions effectively compensated the loss of heat-sensitive nutrients. Only minor difference in percent adult recovery from larvae between the improved diet with and without autoclaving, which was 39% and 34%, respectively, was observed [42]. However, there was approximately 2-fold reduction in the percentage of adult recovery from larvae when *C. canea* larvae were reared on the standard diet [43] compared with the heat sterilized diet [42]. Griffin et al. [44] compared percent adult recovery from larvae of the boll weevils, *Anthonomus grandis* Boheman (Coleoptera: Curculionidae), reared on artificial diets that were made high temperatures of 130 ◦C, 138 ◦C, 144 ◦C, and 151 ◦C for 30 s using a flash-type sterilizer. They found that no significant effects of these temperatures on the percentage of adult recovery from larvae, though the highest temperature of 151 ◦C yielded fewer and smaller weevils than other temperatures. Our ongoing efforts to develop a heat-sterilized diet for WCR facilitating fewer or no antibiotics focus on the identification of alternative ingredients (e.g., yeast hydrolysate, casein) that can be compensated for the loss of nutrition in WCRMO-2 diet due to the overheating.

With diet ingredients in WCRMO-2 formulation, we demonstrated that a temperature of 60 ◦C yielded the highest larval performance on WCRMO-2 diet, while a time exposure of less than 30 min did not have significant impacts on the quality of WCRMO-2 diet. The exposure of WCRMO-2 diet to temperatures below 55 ◦C or above 65 ◦C resulted in a reduction in larval weight and molting compared to WCRMO-2 diet made at the optimum temperature. The significant losses in nutrients needed for WCR larval growth and development were observed when the WCRMO-2 diet was heated over 75 ◦C. Although most insect diets are often made at temperatures of 65–70 ◦C [24] that are likely used to avoid the loss of nutrition of heat-sensitive ingredients (e.g., ascorbic acid, vitamins, lipids), no information on the effects of relative low temperatures (50–80 ◦C) on insect diets is available. This study adds to the limited number of studies characterizing the effects of thermal exposure and time of thermal exposure, especially at mild temperatures, on insect diets. Some insect diets that have been reported to be heat-tolerant are usually produced at temperatures of 121 ◦C for 15–20 min by autoclaving or 141 ◦C for a few seconds by flash sterilizing [37,44]. It is noteworthy that WCR is nearly monophagous on maize roots and can survive on a few grass species [45]. This pest may require specific nutrients, and their bioavailability in WCRMO-2 diet is significantly reduced when the diet is exposed to temperatures over 75 ◦C.

Diet assays determining susceptibility of corn rootworm larvae typically utilizes a 96-well microtiter plate format [46–48]. Typically, each well is filled with 200 μL of diet and overlaid with toxins and followed by an infestation of a single neonate larva and sealing the well. This assay involves a labor-intensive filling process. A high-throughput system designed to run and analyze assays on a large scale, factorially increasing the number of compounds screened in the case of discovery, or the number of assays completed in the case of insect resistance management (IRM) studies would greatly facilitate research programs of this important pest. Research with lepidopteran insects has shown the great value of high-throughput systems for mass production and diet bioassays. In fact, utilizing a flash sterilizer coupled to a form-fill seal machine allowed the mass production of 3 million corn earworm pupae, *Helicoverpa zea* (Boddie) (Lepidoptera: Noctuidae), annually [49]. More recently, the high-throughput system can produce 25,000 rearing units per hour containing artificial diet and eggs of *Trichoplusia ni* (Hubner) (Lepidoptera: Noctuidae) through the use of automation [37]. For these systems, in addition to a flash sterilizer and a form-fill seal machine, a heat-sterilized diet is one of the key components. WCR has proven to be one of the most challenging pests in North America [50]. Many management strategies have been developed for controlling WCR (crop rotation, soil insecticides, Bt maize), but this pest has evaded nearly all management tactics in recent years [10–17]. Recently, WCR has been reported to evolve resistance to the newest management technology, RNA interference (RNAi) [51]. Future research could aim to leverage automation and robotics technology to establish the high-throughput system for WCR that would accelerate discovery efforts related to novel insecticide compounds and their related products.

**Author Contributions:** M.P.H. contributed to writing the first draft of the manuscript. M.P.H., M.G.V., T.A.C. and B.E.H. conceptualized the study. M.P.H., M.G.V. and B.E.H. developed the methodology. M.P.H., A.E.P. and R.W.G. performed the experiments. M.P.H. collected data, performed the analyses, and provided the visualization. M.G.V., K.S.S. and B.E.H. provided materials required for the experiments. All authors have read and agreed to the published version of the manuscript.

**Funding:** Research for this study, in part, was funded by the Small Business Innovation Research (SBIR) Program at the National Institute for Food and Agriculture (NIFA) (grant number 2020-33610-31703).

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** The authors thank Julie Barry, Amanda Ernwall, Michelle Gregory at USDA-ARS (Columbia, Missouri), and David Tague at University of Missouri for technical assistance. Mention of trade names or commercial products is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the USDA. USDA is an equal opportunity provider and employer. This article reports the results of research only. Mention of a proprietary product does not constitute an endorsement or recommendation for its use by the USDA or the University of Missouri.

**Conflicts of Interest:** Michael Vella is an employee of Frontier Scientific Services, Newark, Delaware. All other authors declare no competing interests.

#### **References**


MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Insects* Editorial Office E-mail: insects@mdpi.com www.mdpi.com/journal/insects

MDPI St. Alban-Anlage 66 4052 Basel Switzerland

Tel: +41 61 683 77 34 Fax: +41 61 302 89 18

www.mdpi.com

ISBN 978-3-0365-2493-1