The Tribolium castaneum cell line TcA: a new tool kit for cell biology

The red flour beetle, Tribolium castaneum, is an agriculturally important insect pest that has been widely used as a model organism. Recently, an adherent cell line (BCIRL-TcA-CLG1 or TcA) was developed from late pupae of the red flour beetle. Next generation transcriptome sequencing of TcA cells demonstrated expression of a wide variety of genes associated with specialized functions in chitin metabolism, immune responses and cellular and systemic RNAi pathways. Accordingly, we evaluated the sensitivity of TcA cells to dsRNA to initiate an RNAi response. TcA cells were highly sensitive to minute amounts of dsRNA, with a minimum effective dose of 100 pg/mL resulting in significant suppression of gene expression. We have also developed a plasmid containing two TcA-specific promoters, the promoter from the 40S ribosomal protein subunit (TC006550) and a bi-directional heat shock promoter (TcHS70) from the intergenic space between heat shock proteins 68a and b. These promoters have been employed to provide high levels of either constitutive (TC006550) or inducible (TcHS70) gene expression of the reporter proteins. Our results show that the TcA cell line, with its sensitivity to RNAi and functional TcA-specific promoters, is an invaluable resource for studying basic molecular and physiological questions.

to be significantly enriched in the TcA Up5k (RPKM.0.4686) group; 124 of which were found in the Up5k group of all four RNAseq databases (O in Fig. 2b). The enriched GO-terms for both the Up5k and Down0 had the highest similarity to that of the elytra RNAseq data, with the abdominal carcass and whole pupae comprising the second and third similar collection of enriched GO-terms. The similarity of the TcA enriched GO-terms with those of the elytra suggest that the TcA cell line may be of an epidermal origin (Fig. 2a,  b, c).
The enriched GO-terms commonly found in all four RNAseq for the Up5k (O in Fig. 2b) generally included housekeeping genes, including those for ribosomal proteins (GO:0005840) and subunits (GO:0015934 and 0015935), as well as mitochondrial respiratory components (GO:0005739, 0005740, 0005743, 0005746, 0005747, 0005753, and 0005759). The commonly enriched GO-terms in Down0 groups of all four RNAseq datasets included genes involved in sensory neural function (Fig. 2c, GO:0007600, 000718, 0007606, 0038023, 0004930, 0004888). Large numbers of enriched GO-terms for Up5k were shared between TcA, abdominal carcass, and elytra (N in Fig. 2b) or between TcA and elytra (G), while the similarity was relatively low between TcA and whole pupae (F). Among the GOterms in the category N (common to TcA, abdominal carcass, and elytra in Fig. 2b), it is noteworthy to mention that many of the GOterms are related to intracellular vesicle formation and activity (GO:0006897, 0006909, 0006911, and 0016192), as it was also the case with category G (common in TcA and elytra; GO:0030133, 0006888, and 0048193). In category A (uniquely enriched only in TcA), it is interesting to note that GO-terms related to negative regulators of cell death (GO:0043066 and 0043069) are among those in the Up5K group and may be partly responsible for the establishment of the immortalized, stable cell line.
We used immunohistochemistry to confirm the presence of proteins from genes that were highly expressed in the tanscriptome. An antibody specific for a signaling peptide, calcitonin-like diuretic hormone 8 , showed small cytoplasmic vesicles in every cell (Fig 3a), indicating a uniformity of expression among the TcA cells. However, an antibody specific for chitin deacetylase stained vesicles in only a subpopulation of TcA cells (Fig. 3b), suggesting some variability in gene expression between cells.
Although a number of genes involved in cuticle metabolism were found to be highly expressed, supporting the idea of TcA as an epidermal cell type, there were a large number of chitin metabolic genes that were captured in Down0. Indeed, category A, enriched only in TcA with no detectable expression (Down0. Fig. 2c), included many GO-terms for chitin catabolism and chitinase activities. Therefore, it seems TcA expresses a specific subset of cuticle proteins and chitinases at significant levels while it lacks expression of many other genes involved in cuticle metabolism, perhaps because the TcA cell line is a clonal descendent from a specific subtype of epidermal cells. It would be of interest in the future to determine if the TcA line can be used as a model as an insect epidermal cell type with the activity for chitin synthesis.
In addition to the genes involved in chitin metabolism discussed above, we also found that numerous genes putatively involved in insect immunity 9 are expressed in the TcA cell line. These are listed in Table S3 and entire expression data is in data set 1.
Expression of genes encoding candidate RNAi machinery. RNAi is very effective in T. castaneum during all life stages, and has been widely used to study candidate gene function 10,11 . Previous studies have established the basic components of the RNAi pathway that are found in the T. castaneum genome 4,12 . Based on the information from these studies, we examined the expression of genes encoding core  RNAi machinery (Table 1) and genes implicated in systemic RNAi (Table 2) in the TcA cell line. All genes encoding components of the cellular RNAi machinery were expressed in the TcA cell line (Table 1) including the three dicer family genes (dicer 1, dicer 2, and drosha), the genes for their associated dsRNA binding proteins (loquacious, R2D2 and C3PO, and pasha), all 5 argonaute genes (Ago 1, 2A, 2B, 3, and piwi), and the exonuclease, snipper. In addition, many of the candidate genes thought to be associated with the systemic RNAi response are expressed in the TcA cell line ( Table 2). The apparent intact nature of the RNAi response system in the TcA cell line vastly increases the potential for this cell line to be used in highly informative investigations of cellular and molecular physiology.   TcA-specific promoters for reporter gene expressions. Heterologous expression of a native, chimeric, or fluorescently-tagged protein in a cell line is a commonly used and extraordinarily useful tool. Use of the TcA for expression of target proteins depends on efficient promoters that constitutively or conditionally promote the expression of the exogenous gene. Our initial trials with promoters that are commonly used in insect cell lines were not successful at expressing either the fluorescent reporter protein, Tomato, or enhanced green fluorescent protein, EGFP. In a test of the common promoters frequently used in insect cell systems, the Actin 5 promoter of D. melanogaster, the CMV promoter in the pcDNA3.11 plasmid, the OpIE1 and OpIE2 promoters in the pIZT vector (Invitrogen), and the alpha tubulin promoter of T. castaneum 13 were all ineffective at producing detectable levels of Tomato up to 3 days after transfection of the plasmid into TcA cells. Therefore, we expanded our efforts to test endogenous Tribolium promoters. We chose candidate promoters based on the expression levels of genes in the TcA (,.5 in the RPKM), defined small sizes of intergenic regions at the 59 end of the genes, and a lack of large introns. The putative promoters for three genes encoding ribosomal protein subunits, TC014362, TC000476, and TC006550, and one heat shock protein 68 (hereafter named as TcHS70 by Drosophila homology) were identified and selected for testing. Two promoters, TC006550 encoding a 40S ribosomal protein subunit and TcHS70, were found to be active for production of the reporter protein.
The TcHS70 contained multiple putative heat shock sequence elements ( Fig. 3a and b). Initially, we tested the TcHS70 activity after a heat shock in the endogenous cell. Immediately after a heat shock of TcA cells at 40uC for 30 min, total RNA was extracted to investigate the transcriptional responses of the hsp genes. We found ,300 to 500-fold increases in mRNA levels for each hsp68a, hsp68b, and another highly similar hsp TC010172 after normalization to rps3 transcript levels and comparison to the no heat shock control (Fig. 4c). We subsequently used the 1910bp intergenic region between the coding genes for hsp68a and hsp68b as the TcHS70 promoter ( Fig. 4a and 5a). This sequence included the putative TATA boxes, transcription start sites located in each end, and numerous heat shock sequence elements (Fig. 4b).
We constructed a plasmid using the pUC18 vector as a backbone, containing reporter and antibiotic proteins: the TC006550 promoter driving expression of an EGFP-Zeocin resistance fusion protein and the bidirectional TcHS70 promoter driving expression of secretory nanoluciferase (Promega) and photo switchable Orange 14 (Fig. 5a). Transient transfection of this construct was followed by heat shocks at 38uC for 10, 20, and 40 min. and at 40uC for 10 or 20 min. The naouluciferase activity in the cell culture media was significantly increased 5 hr after heat shock except in the case of 38uC for 10 min. Application of a second heat shock further increased naoluciferase activity compared to the first heatshock (Fig. 5b). The changes in nanoluciferase activity over time indicated that the min- imum expression and maturation time for nanoluciferase activity is ,3 hr after heatshock, and it reaches a plateau at ,8 hr following heat shock at 38uC for 40 min (Fig. 5b and c). The cells with constitutive expression of EGFP were also found to have inducible expression of psOrange following heat shock, although induction of visible levels of psOrange appears to require a longer time for maturation than that of nanoluciferase. Occasionally, we found autofluorescence in the TcA that is similar to the EGFP wavelength. This autofluorescence was limited to large vacuolar structures appearing only in a small subpopulation of the cells (Fig. 5d).
In conclusion, based on expression of the reporter proteins, the TcHS70 is a highly inducible bidirectional promoter. Further, the TC006550 promoter also provides a moderate level of constitutive reporter expression. The plasmid containing those promoters will be a valuable tool for expression of heterologous genes.
RNAi in the TcA cell line. Expression of many of the components of the RNAi machinery led us to investigate the sensitivity of this cell line to RNAi. Accordingly, we selected a relatively highly expressed chitin protein, CP6, as our target for silencing, and used the common control target, vermilion, as a negative control. dsRNA was synthesized (231 bp for CP6, 400 bp for vermilion) and mixed directly with culture media at various concentrations (1 fg/mL to 1 mg/mL, vermilion was mixed at 100 ng/mL) prior to applying to the TcA cells. qPCR quantification showed that the TcA cell line is highly sensitive to RNAi of CP6. Treatment with as little as 1 ng/mL resulted in .99% suppression of CP6 mRNA expression (Fig. 6a). Suppression of CP6 mRNA decreased in a dose-dependent manner as dsRNA concentration decreased below 1 ng/mL, with no significant reduction in CP6 expression occurring with 1 pg/mL dsRNA. The sensitivity to dsRNA of TcA cells is similar to that of the Tc81 cell line as measured with dsRNA targeting the met gene 7 , and more than 10,000 times greater than that for the S2 Drosophila cell line 15 .
We also tested our CP6 dsRNA in vivo by injecting dsRNA (2 or 200 ng/beetle) into T. castaneum early instar larvae to ascertain the effects of CP6 RNAi in vivo. No obvious phenotype was associated with silencing of CP6 in any developmental stage (late instar larvae, pupae, or adult), although the beetles injected with vermilion dsRNA showed the characteristic white eyes associated with this treatment.
Analysis of CP6 mRNA showed that both 2 and 200 ng treatments per individual resulted in .99% suppression of expression at the organismal levels without noticeable phenotype in the observation until the early adult stage.
We also subsequently tested the efficiency of RNAi to silence the expression of heterologously expressed proteins, transiently transfected for heat shock inducible nanoluciferase (Fig. 6b). Silencing of nanoluciferase activity induced by heat shock was measured following treatment with dsNluc. Interestingly, significant suppression of nanoluciferase activity required a minimum of 1 ng/mL (,50% suppression, Fig. 6b), which is more than 10-fold greater than that required for suppression of CP6 expression. We suspect that this difference is likely caused by the rapid induction of extremely high levels of expression of the nanoluciferae following heat shock treatment.
High sensitivity of the TcA cells to dsRNA in the media requires efficient mechanisms for cellular uptake of primary dsRNA followed by dsRNA processing into small interference RNA (siRNA) which forms complexes with the target messenger RNA. Our results show that TcA expresses a large number of candidate molecular components of the RNAi pathway (Table 1) and shows high sensitivity to dsRNA, provideing an excellent opportunity to explore novel, highly efficient RNAi mechanisms in the TcA cell.

Methods
Cell culture. TcA, an adherent cell line, was derived from BCIRL-TcA-CLG1 that was originally established from the late pupal T. castaneum 6 . TcA cells were grown in ExCell 420 media (Sigma, St. Lois, MO.) supplemented with 10% insect cell culturetested fetal bovine serum (Atlanta Biologicals, Lawrenceville, GA) and 4.5 mg/ml of bovine insulin at 27uC and 60% relative humidity, although bovine insulin was found to be not necessary for normal cell growth in the later study (Park et al, unpublished data).
Transcriptome sequencing and analyses. Total RNA was harvested from confluent cells using a commercial kit (Qiagen, Valencia, CA.). RNA quality was tested using an Agilent Bioanalyzer prior to submission for sequencing. Illumina-based (GXII platform) rRNA-depleted transcriptome sequencing of a single RNA sample was carried out by Cofactor Genomics (St. Louis, MO.). Obtained sequences were single direction, 50 nt in the length, and filtered for the sequences with a q-score higher than 30 and longer than 30 nt by using the FastX toolkit implemented in DNA Galaxy 16 . Assembly of the sequences to the reference genome Tcas3.0 was made in DNA Nexus (https://classic.dnanexus.com/home). The RPKM (read per kilobase per million www.nature.com/scientificreports SCIENTIFIC REPORTS | 4 : 6840 | DOI: 10.1038/srep06840 mapped reads) was used for identifying the top 5,000 highly expressed genes (Up5k) and the genes with undetectable transcript levels (Down0) (Fig. 1).
The transcriptome of TcA was evaluated by comparisons of enriched GO-terms to other tissue specific RNAseq data. Direct comparisons of RNAseq data for expression levels of each gene by RPKM in different sets of data were not possible because of varying sample preparation procedures and sequencing approaches in each RNA library. For example, normalization of our sequence data to another library introduced largely skewed raw RPKM values, and was thus not appropriate to use in conventional methods for comparing expression levels. As an alternative approach, the lists of genes for the Up5k and Down0 were used for a GO-term enrichment test against the official gene set (Tcas 3.0 OGS 17 ,) in Blast2GO 18 at the False Discovery Rate (FDR) of 0.05 (Fig. 2). Enriched GO-terms of each library representing RNAseq data from elytra of late pupal stage (elytra, the data provided by Dr. S. Muthukrishnan), abdominal carcass excluding the gut and the central nervous system in the early adult stage (abdominal carcass, data provided by Dr. T. Phillips), and whole pupal stage (whole pupae, data provided by Dr. G. Butcher) were compared for each Up5k and Down0.
Immunohistochemistry. Anti-DH37 (rabbit, 15500), originally raised against the diuretic hormone 37 of Tenebrio molitor 8 , and anti-CDA1 (rabbit, 15500) against T. castaneum chitin deacetylase 1 19 were used to label their respective targets in fixed TcA cells grown on a petri dish. A goat anti-rabbit IgG antibody conjugated with Alexa Fluor 488 (Life technologies) was used as the secondary antibody. Additionally, we used phalloidin conjugated with Alexa Fluor 594 (Life technologies) to stain actin and 49,69-diamino-2-phenylindole (DAPI 300 nM, Sigma) to stain cell nuclei (Fig. 3). Images were captured on an LSM 700 confocal microscope under a 40X objective (Zeiss).
Plasmid construct. pIZTV5-His (Invitrogen), the plasmid originally developed for lepidopteran insects, was used for constructions of other variants. Plasmids we constructed were for testing putative promoters in the TcA cell line for expression of fluorescent protein dTomato 20 as the reporter. The promoters tested were: the CMV promoter from the pcDNA3.11 plasmid (Invitrogen), the OpIE1 and OpIE2 promoters from the pIZT vector, the Tribolium alpha tubulin promoter 13 , and the Tribolium heat shock promoter (TcHS70, in this study). The dTomato was amplified using the primers dTomatoF and dTomatoR (Table S1) and cloned into a pGEMT easy vector. Then, dTomato was introduced into the pIZTV5-His vector by using a NotI restriction site (pIZT-dTomato). Subsequently, the promoters were cloned into the pGEMT easy vector and then inserted into pIZT-dTomato replacing the OpIE2 promoter by digestion with the corresponding restriction enzymes and ligation  Figure S1). Among these constructs, the plasmid containing the Tribolium heat shock promoter was further modified to add an ampicillin selection marker driven by a bla promoter. The OpIE1 promoter upstream of the GFP-Zeocin chimeric gene in pIZTV5-His was replaced by the putative promoters for three genes encoding ribosomal protein subunits, TC014362, TC000476, or TC006550. Subsequently, the constitutively active promoter TC006550 (upstream ,1.2 kb) was used for the expression of GFP-Zeocin.
After we tested promoter efficiencies by using the plasmids described above, we constructed the plasmid that contains several reporter genes in Figs. 4, 5, and S1. The plasmid contains a bla promoter with ampicillin resistance, a TC006550 promoter for GFP::Zeocin resistance (Invitrogen) with SV40 39 UTR, a TcHSP70 bidirectional promoter for secretory nanoluciferase (from pGL3 of Promega) with a BGH 39UTR, and psOrange 14 with OpIE2 39UTR (Invitrogen). The primers used in the construction of the plasmid are listed in Table S1 and the plasmid map including the major restriction sites that were used for plasmid construction is shown in Figure S1.
PCR was conducted using the high fidelity DNA polymerase PrimeStar (Takara). All the restriction enzymes, the T4 DNA ligase as well as the competent cells, DH5a TM high efficiency, were purchased from New England Biolabs.
RNA interference. dsRNA was synthesized using the Ampliscribe T7 Flash Transcription Kit (Epicentre Biotechnologies, Madison, WI) from cDNA templates using gene-specific primers (Table S1) with a T7 nucleotide sequence attached to the 59 end of the primer sequences of CP6 (cuticle protein 6, or also named as CPR71), vermilion, and nanoluciferase. The dsRNA was purified using a commercial kit (MEGAClear, Life Technologies) and run on a 2% agarose gel to confirm the band size. dsRNA was quantified using a Nanodrop Spectrophotometer (Thermo Scientific, Wilmington, DE) and diluted to the desired concentration with nucleasefree water.
For CP6 RNAi experiments in cell culture, TcA cells were seeded onto 24-well plates (5 3 10 6 cells/mL) and left to grow for at least 24 h. RNAi was achieved in TcA cells by adding dsRNA for CP6 or vermilion directly to the culture media. RNA was harvested after 96 h using TRIzol reagent (Life Technologies), and cDNA was synthesized using the ImProm-II TM Reverse Transcription System (Promega). cDNA samples were then diluted 20-fold and CP6 expression was quantified by qPCR utilizing comparative Ct 21 on a Biorad thermal cycler (Biorad) with ribosomal protein S3 (rpS3) serving as the internal loading control. Gene-specific primers for CP6 and rpS3 are listed in Table S1.The thermal cycling conditions were as follows: 95uC for 2 min, 40 cycles of 95uC for 15 sec, 60uC for 15 sec and 72uC for 30 sec. Melting curve analysis (55-95uC) and gel electrophoresis were applied to all reactions to ensure consistency, specificity, and amplicon size of the amplified product. All CP6 Ct values were first normalized to rpS3 in the same sample, and then secondarily to the vermilion treatment in order to produce a relative expression value. These experiments were replicated three times and the results are presented as the mean 6 standard error.
For in vivo RNAi experiments of CP6, early last instar larvae were microinjected with dsCP6 (200 ng or 2 ng/individual) or vermilion (200 ng/individual) dsRNAs. Two weeks post-injection, adults were visually examined for any defects in development. Four individuals from each treatment group were pooled for total RNA isolation. cDNA was synthesized from total RNA (1 mg/rxn) using the ImProm-II TM Reverse Transcription System (Promega). RNAi efficiency was determined by qPCR as described above.
For RNAi of nanoluciferase, TcA cells were seeded onto a 96-well plate (5 3 10 6 cells/mL) and left to grow for at least 24 h. The TcA cells were transfected with the expression plasmid (Fig. 5a), using FuGENE HD transfection reagent (Promega). Subsequently, various concentrations of dsRNA for nanoluciferase were directly added to the culture media for final dsRNA concentrations ranging from 100 fg/mL to 100 ng/mL. About 48 h after the transfection, the cells were subject to heat shock (HS) at 38uC for 40 min. Approximately 18 h after the initial HS, the media was changed to new media and one more HS was made at 38uC for 40 min. Six hours after the HS, secretory nanoluciferase activity in the cell culture media was measured using the Nano-Glo luciferase assay system (Promega). dsVermilion (100 ng/mL) was used as a negative control.