Metabolic and transcriptome responses of RNAi-mediated AMPKα knockdown in Tribolium castaneum

Background The AMP-activated protein kinase (AMPK) is an intracellular fuel sensor for lipid and glucose metabolism. In addition to the short-term regulation of metabolic enzymes by phosphorylation, AMPK may also exert long-term effects on the transcription of downstream genes through the regulation of transcription factors and coactivators. In this study, RNA interference (RNAi) was conducted to investigate the effects of knockdown of TcAMPKα on lipid and carbohydrate metabolism in the red flour beetle, Tribolium castaneum, and the transcriptome profiles of dsTcAMPKα-injected and dsEGFP-injected beetles under normal conditions were compared by RNA-sequencing. Results RNAi-mediated suppression of TcAMPKα increased whole-body triglyceride (TG) level and the ratio between glucose and trehalose, as was confirmed by in vivo treatment with the AMPK-activating compound, 5-Aminoimidazole-4-carboxamide1-β-D-ribofuranoside (AICAR). A total of 1184 differentially expressed genes (DEGs) were identified between dsTcAMPKα-injected and dsEGFP-injected beetles. These include genes involved in lipid and carbohydrate metabolism as well as insulin/insulin-like growth factor signaling (IIS). Real-time quantitative polymerase chain reaction analysis confirmed the differential expression of selected genes. Interestingly, metabolism-related transcription factors such as sterol regulatory element-binding protein 1 (SREBP1) and carbohydrate response element-binding protein (ChREBP) were also significantly upregulated in dsTcAMPKα-injected beetles. Conclusions AMPK plays a critical role in the regulation of beetle metabolism. The findings of DEGs involved in lipid and carbohydrate metabolism provide valuable insight into the role of AMPK signaling in the transcriptional regulation of insect metabolism.


Background
The survival of all organisms depends on the maintenance of energy homeostasis. AMP-activated protein kinase (AMPK) is a cellular energy sensor conserved across all eukaryotic species [13]. As a serine/threonine protein kinase complex, AMPK consists of a catalytic subunit α and two regulatory subunits, β and γ, and is activated in response to energy stress by sensing increases in ADP/ ATP and AMP/ATP ratios, which leads to the activation of ATP-generating catabolic pathways including glycolysis and fatty acid oxidation and the inhibition of ATPconsuming anabolic pathways such as gluconeogenesis, fatty acid and protein synthesis [17]. While nucleotidedependent phosphorylation of Thr172 in the α subunit by liver kinase B1 (LKB1) is the principal event required for full activation of AMPK in mammalian cells [19,70], several studies have revealed the nucleotide-independent regulation of AMPK via the phosphorylation of Thr172 by calcium/calmodulin-dependent kinase kinase 2 in mammals (CAMKK2) [20,23,69].
Lipids and carbohydrates are major sources for energy storage and supply in cells. Under aerobic conditions, most energy is derived from fatty acids oxidation and the rest of energy is obtained from glucose oxidation. In contrast, glycolysis plays an important role in ATP production under anoxic condition [51]. Activation of AMPK occurs in response to stress circumstances such as starvation, exercise, hypoxia and ischemia, heat shock, and oxidative stress [9,11,18,45,47,49,67]. It is well known that AMPK can regulate lipid and carbohydrate homeostasis via direct phosphorylation of multiple downstream effectors. Acetyl-CoA carboxylase (ACC), the first rate-limiting enzyme in fatty acid synthesis, glycerol-3phosphate acyltransferases (GPAT), the rate-limiting enzyme in triglyceride (TG) synthesis, and 3-hydroxy-3methylglutaryl-CoA reductase (HMGR), the rate-limiting enzyme of cholesterol biosynthesis pathway, can be inhibited by AMPK through phosphorylation in rats [8,42,68]. Subsequently, AMPK was found to promote lipid absorption and release by directly phosphorylating lipases like hormone-sensitive lipase (HSL) and adipocytetriglyceride lipase (ATGL) in mice and Caenorhabditis elegans [1,44,64]. On the other hand, AMPK also stimulates glycolysis via the phosphorylation and activation of 6-phosphofructo-2-kinase/fructose-2,6-biphosphatase 3 (PFKFB3) and 6-phosphofructo-2-kinase (PFK2) in human tissues [37,38]. Additionally, AMPK activation was also shown to increase whole-body insulin sensitivity by phosphorylation of insulin receptor substrate-1 (IRS-1) in mouse [24] and insulin receptor (InR) in rodent muscle [10]. Given the functional attributes of AMPK in lipid and carbohydrate metabolism, AMPK is considered as an important therapeutic target for treating metabolic diseases including obesity and type 2 diabetes [43].
Although the role of AMPK in the regulation of cell metabolism is well studied in mammals, related research is still limited in insects. Notably, AMPK can regulate energy balance via modulation of transcriptional expression of metabolic enzymes in the long term, however, its downstream transcriptional pathways remains largely elusive [7]. Recently, we reported the transcriptional and post-translational activation of TcAMPKα by oxidative, heat and cold stresses in the red flour beetle, Tribolium castaneum [25]. In this study, RNAi was employed to determine the roles of TcAMPKα in lipid and carbohydrate metabolism. Comparison, annotation and classification of DEGs between dsTcAMPKα treatment and control groups were also conducted by high-throughput transcriptome sequencing to identify metabolism-related genes modulated by AMPK in T. castaneum.

Transcriptome sequence and reads mapping
The dsTcAMPKα and dsEGFP groups were analyzed by RNA-Seq (three independent biological replicates of each treatment). A mean of 23,570,938 clean reads were generated among six independent libraries (T01-T06) ( Table 1). Evaluation of clean data quality showed that the GC counts ranged from 42 to 45% and Q30 ratios were > 93%, indicating a high level of data quality. The alignment of clean reads to the reference genome database of T. castaneum showed that 83.50 and 77.15% reads of the dsEGFP and dsTcAMPKα groups were aligned on average, respectively (Table 2).
Correlation analysis was conducted with Pearson's Correlation Coefficient (R value) to evaluate the Fig. 1 The change of TG, glucose and trehalose levels in dsTcAMPKα-treated and AICAR-treated T. castaneum larvae. a-c: Relative TG, glucose and trehalose levels of T. castaneum in dsTcAMPKα group compared with dsEGFP group. d-f: Relative TG, glucose and trehalose levels of T. castaneum in AICAR group compared with IB group. (by Student's t-test, *p < 0.05, **p < 0.001) Fig. 2 The ratio between trehalose and glucose under treatment with dsTcAMPKα and AICAR compared with control groups. (by Student's t-test, *p < 0.05, **p < 0.001) biological repeatability based on expression values of each library [52]. The results showed that the Pearson's Correlation Coefficient between the three control samples (T01-T03) was 0.81, and that of dsTcAMPKα treatment samples (T04-T06) was 0.94 (Fig. 3a). Box plot analysis revealed that the three samples in each group had similar expression distributions of reads, while the control and treatment groups had significantly different expression distributions (Fig. 3b).
Further annotation of expressed unigenes revealed that a total of 14,095 unigenes out of 31,944 unique sequences were annotated and classified into at least one database of Non-redundant (Nr), EuKaryotic Orthologous Groups (KOG), Clusters of Orthologous Groups of proteins (COG), Kyoto Encyclopedia of Genes and Genomes (KEGG), Protein family (Pfam), Gene Ontology (GO) and Swiss-Prot databases ( Figure S1, Figure S2 and Table S2).

Changes in gene expression profiles
To identify the effects of knock-down of TcAMPKα on global unigene expression patterns of T. castaneum, DEGs between dsTcAMPKα and dsEGFP groups were identified based on their Fragments Per Kilobase of transcript sequence per Million base pairs sequenced (FPKM) values. 1184 DEGs were obtained including 349 upregulated and 835 downregulated unigenes (Fig. 4a, b and Table S3). The log 2 -fold variation range of DEGs was between − 6.07 and 3.75 (Fold change from − 67.18 to 13.45, P value from 1.27 × 10 − 53 to 0.04978).
DEGs were classified by searching against GO and KOG databases. GO term enrichments was used to further analyze physiological changes associated with DEGs in T. castaneum. GO enrichments revealed that the DEGs involved in cellular component category were enriched in cell part, cell, membrane and organelle ( Figure  S3). In molecular function category, most of the DEGs were enriched in binding and catalytic activity such as Retrovirus-related Pol polyprotein and ATPase inhibitor, while metabolic and cellular process were the most enriched subcategories in biological process ( Figure S3).
KOG database was used to annotate DEGs with specific physiological functions (Fig. 5, Table S4). Although the "General function prediction only" was the most abundant group among all subcategories in KOG database, a large number of DEGs were annotated into the metabolism classifications such as "Lipid transport and metabolism", "Carbohydrate transport and metabolism", "Amino acid transport and metabolism", "Energy production and conversion" and "Secondary metabolites biosynthesis, transport and catabolism" ( Table 3). Furthermore, most DEGs involved in protein translation were downregulated such as some ribosomal proteins (Table 3). Similarly, among the 42 DEGs in the posttranslational modification class, 35 DEGs were downregulated, including 3 heat shock proteins (Hsps), whereas the phosphatidylinositol 4,5-bisphosphate 3kinase (PIK3) and InR2 involved in the signal transduction mechanisms were upregulated.

Expression of genes involved in lipid metabolism, carbohydrate metabolism and insulin signaling
Besides the analysis of the entire gene set, we specifically checked for up-or downregulation of lipid and carbohydrate metabolism genes in our set of DEGs. The results showed that the insect adipose triacylglycerol lipase homologue, brummer (Accession no. TC011935), which is responsible for the first step of TG hydrolysis, was significantly downregulated with the expression log 2 -fold change of − 2.20 (Fold change: 0.22; ANOVA, df 1, 4 , F = 5.152, P value = 1.60 × 10 − 05 ), whereas two fatty acid synthetase genes (FAS1-2) (Accession no. TC015337 and TC015340) (Fold change: 2.03 and 3.22, ANOVA, df 1, 4 , F = 9.182 and 157.306, P value = 0.03878 and 0.00023) involved in fatty acid biosynthetic pathways, and the transcription factor ChREBP (Accession no. TC010471) (Fold change: 2.27, ANOVA, df 1, 4 , F = 96.147, P value = 0.00061), a key regulator of glucose and lipid metabolism and fat storage [62], were upregulated with the expression log 2 -fold change from 1.02 to 1.76 (Fold change from 2.03 to 3.39) (  data, the expression levels of these DEGs were determined using RT-qPCR ( Fig. 6 and Table 4). Gene expression levels validated by RT-qPCR showed the high consistency with transcriptome sequencing. To be as inclusive as possible, less strict screening criteria (Fold change |log 2 (treatment/control) | > 0.5) were applied, and additional lipid and carbohydrate metabolism-related genes with the expression changed in transcriptome were selected for further verification using RT-qPCR. The results showed that five genes involved in fatty acid and triglyceride biosynthetic pathways including three FAS (FAS3-5) (Accession no. TC011522, TC015400 and TC000238) (   (Table 4). Interestingly, SREBP1(Accession no. TC007163) (Fold change: 1.93; ANOVA, df 1, 4 , F = 9.490, P value = 0.03691), the master regulator of lipid homeostasis, and SREBP cleavageactivating protein (SCAP) (Accession no. TC013456) (Fold change: 1.75; ANOVA, df 1, 4 , F = 10.483, P value = 0.03174), a central regulator of lipogenesis that controls the activity of SREBP [53] were also significantly upregulated (Table 4).

Discussion
In insects, the energy for growth and development is mainly derived from the storage and utilization of lipids and carbohydrates in specific tissues such as fat body, midgut and oenocytes [39]. TG is the master form of lipids stored in fat body and plays an important role in energy storage and release [63]. Meanwhile, glucose (monosaccharide) and trehalose (disaccharide) provide energy through glycolysis [48]. Given that AMPK activates ATP-generating pathways and inhibits energyconsuming processes under conditions of low energy status [60], the attenuation of AMPK signaling in insects may disrupt the energy balance in vivo. In this study, we found that the TG and glucose levels in dsTcAMPKαinjected T. castaneum were significantly increased compared with dsEGFP-injected insects, while a decreased trehalose content was observed in dsTcAMPKα-injected insects. These results were further confirmed by in vivo AICAR treatment, which resulted in the decreased TG and glucose levels and increased trehalose content. Similarly, AICAR-induced AMPK activation resulted in significantly decreased TG level in lean and obese rodents, in vivo [6]. In human HepG2 cells, the kinase-inactive AMPKα increased lipid content and prevent the metformin from decreasing lipid accumulation [74]. However, lower TG levels were observed in D. melanogaster with reduced AMPK function during fed conditions [27]. On the other hand, while activation of AMPK triggered a  [72], and increased hepatic glucose production was observed in AMPKα2 knockout mice [3], knockdown of hypothalamic AMPK activity in male Sprague-Dawley rats led to a significant suppression of glucose production [73]. These results indicate the complexity of the role of AMPK in the regulation of metabolic processes.
As an energy sensor that regulates cellular metabolism, AMPK not only has acute effects on metabolic enzymes by direct phosphorylation, but also shows long term action to change the transcriptional levels of metabolic proteins and enzymes. For example, activation of AMPK in liver and adipocytes can downregulate FAS activity and inhibit lipogenesis [36]. Activation of AMPK by 4-Hydroxyderricin and xanthoangelol downregulates GPAT in 3 T3-L1 cells [75], an enzyme necessary for triacylglycerol synthesis [66]. Similarly, in C. elegans, AAK-2 can inhibit fat synthesis under stress condition by downregulating lipid synthesis-related genes such as Δ9 fatty acid desaturases which can produce monounsaturated fatty acids to constitute TGs [55,65]. To investigate the downstream transcriptional pathways of AMPK in T. castaneum, the transcriptomes of dsTcAMPKαinjected and dsEGFP-injected larvae under normal conditions were compared. Global gene expression profiles of dsTcAMPKα group were distinct from dsEGFP group with 1184 DEGs. Of particular interest is the upregulation of genes involved in fatty acid and TG biosynthetic pathways, such as FAS, ACC and GPAT, whereas the relative expression level of brummer involved in TG hydrolysis [14] was dramatically downregulated. We also observed the upregulation of several carbohydrate metabolism-related genes, such as TRE, and two key enzymes in gluconeogenesis (de novo synthesis of glucose), PC and PEPCK [16,26], in the dsTcAMPKα-injected insects. The transcriptional changes of genes encoding  metabolism enzymes might contribute to the increased TG and glucose levels and decreased trehalose content in dsEGFP-injected beetles. AMPK activation has been reported to cause a reduction in transcriptional activity of several metabolismrelated transcription factors. As a key transcription factor that regulates cellular lipogenesis in liver, skeletal muscle and adipose tissue, insulin-activated SREBP1 acts in synergy with glucose-senstive ChREBP, which mediates the response to dietary carbohydrates and is essential for regulating lipogenic gene expression [46]. AMPK-inhibited SREBP1 can block expression of some gluconeogenic and lipogenic genes, whereas SREBP1 overexpression can increase their transcription [12,31,54]. Activation of AMPK by metformin or an adenosine analogue suppresses the expression of SREBP1. In metformin-treated rats, hepatic expression of SREBP1 mRNAs and protein is reduced [76]. A recent study indicated that AMPK interacts with and directly phosphorylates SREBP1, suppresses SREBP1 cleavage and nuclear translocation, and represses SREBP1 target gene Fas expression in hepatocytes in response to metformin treatment, leading to reduced lipogenesis [34]. On the other hand, AMPK also phosphorylates Ser568 and reduce DNA binding capacity and promote nuclear exclusion of murine ChREBP [28,29]. Interestingly, the promotion of PEPCK expression is associated with increased expression of SREBP-1 and ChREBP in high free fatty acid (HFFA)-treated hepatocytes [32], and a recent study also revealed the role of ChREBP in gluconeogenesis [58]. In this study, SREBP1, ChREBP and SCAP were significantly upregulated in dsTcAMPKα-injected insects, which in turn might modulated the expression of genes involved in lipid and carbohydrate metabolism.
IIS pathway was involved in the regulation of glucose and lipid metabolism (Saltie and Kahn 2001) [50]. In addition to the regulation of lipid synthesis, studies reveal expanding roles for SREBP1 in controlling pathways for insulin resistance [34], in which the pathological process involves a series of cascades, including defective activation of IRS and PI3K [22,30,59]. Overexpression of SREBP1 decreased Irs-1 mRNA levels in a dose-dependent manner, and SREBP1 knockdown led to an upregulation of IRS-1 levels [34]. Further luciferase reporter assay confirmed that Irs-1 promoter activity was repressed by SREBP1 overexpression [34]. However, in liver with Nonalcoholic fatty liver disease (NAFLD), IRS-1 expression was enhanced and correlated positively with SREBP1 expression. In contrast, IRS-2 expression decreased by 50% and was not correlated with SREBP1 [31]. In sebocytes, insulin-like growth factor-1 (IGF-1) induces SREBP-1 expressions at both mRNA and protein levels in a PI3K-dependent manner, accompanied by an increase in the transcription of SREBP-1 target genes such as FAS [57]. Additionally, insulin-stimulated endogenous ChREBP expression was also observed in HepG2 and primary hamster hepatocytes (Sirek et al. 2009) [56]. In this study, we observed the upregulation of IIS-related DEGs like IRS1, InR2 and PI3K in dsTcAMPKα-injected beetles, which might result in the upregulation of SREBP1 and ChREBP. Interestingly, it has been reported that knockdown in expression of insulin like peptide 2 (ILP2) caused a decrease in TRE mRNA levels in T. castaneum [71], suggesting the positive control of TRE transcription by IIS signaling. Further study is needed to clarify the mechanism of upregulation of IIS signaling in dsTcAMPKα-injected insects.

Conclusion
This study confirmed that AMPK has an important role in the regulation of beetle metabolism. Specifically, our study showed that knockdown of AMPK causes alteration in expression levels of genes involved in lipid and carbohydrate metabolism as well as IIS signaling. Such investigations will help us understand the function of AMPK in transcriptional regulation of insect metabolism.

Double-strand RNA synthesis and injection
Gene specific primers (Table S1) with T7 promoter were designed to synthesize the dsRNAs targeting nucleotides 844-1285 (442 bp) of the ORF region of the TcAMPKα using TranscriptAid™ T7 High Yield Transcription Kit (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer's instructions. The synthesized dsRNAs were diluted in diethyl pyrocarbonate (DEPC)treated water with a concentration of 2 μg/μL, and about 200 ng of dsRNA in 200 nL IB was injected into 20-dayold larvae using a Nanoliter 2010 injector system (WPI, Sarasota, FL, USA) under a stereomicroscope. A total of 10 unsexed insect larvae were collected as a sample for RNA extraction at six-days post-injection. The dsEGFPinjected larvae (CK group) were set as controls in all injection experiments. Three biologically independent replicates were carried out with at least 100 insects (≥ 200 mg) in each replicate.

Triglyceride (TG) measurement
Total TG levels were determined using the liquid TG (GPO-PAP) method [4]. Briefly, each replicate with 10 injected larvae were homogenized in 270 μL of PBS (0.1 mol/L PH 7.4) and centrifuged at 2500 rpm for 10 min. The supernatant was collected, and the TG level was analyzed by using TG Assay Kit (catalogue no. A110-1, Jiancheng Bioengineering Institute, Nanjing, China) according to the manufacturer's instructions. Three independent biological replicates and four technical replicates were performed for every treatment.

Glucose and trehalose measurement
Total glucose levels were measured using glucose content assay kit (catalogue no. BC2500, Solarbio Science & Technology, Beijing, China) according to the manufacturer's instructions. Briefly, 10 larvae on day 6 after injection of dsTcAMPKα or dsEGFP were weighed and homogenized in 0.3 mL double distilled water. The supernatant was collected after centrifugation at 8000 g for 10 min and used to quantify the level of glucose at 505 nm with Multiskan FC microplate reader (Thermo Fisher Scientific, Waltham, MA, USA). Three independent biological replicates and 4 technical replicates were performed for every treatment.
Total trehalose level was measured using Trehalose content detection kit (catalogue no. A149-1-1, Jiancheng Bioengineering Institute, Nanjing, China) according to the manufacturer's instructions. Six days after injection of dsTcAMPKα or dsEGFP, 10 larvae were homogenized in 0.3 mL extraction solution and incubated at room temperature for 45 min. The supernatant was collected after centrifugation at 8000 g for 10 min, and used to quantify the level of trehalose at 620 nm with Multiskan FC microplate reader (Thermo Fisher Scientific, Waltham, MA, USA). Three independent biological replicates and four technical replicates were performed for every treatment.

AICAR treatment
To investigate the effects of AMPK activation on lipids and carbohydrates metabolism, 20-day-old larvae were injected with 200 nL of 1.6 mg/mL AICAR or IB as control. AMPK could be activated immediately by AICAR treatment among a couple of hours [40]. The total amounts of TG, glucose and glucose were determined at one-hour post-injection.

RNA extraction, library construction and sequencing
Total RNA was isolated from 10 larvae of dsTcAMPKα or dsEGFP group on the sixth day after injection using TRIzol Reagent (Invitrogen, USA), and digested by RNase-free DNase I (Takara, Dalian, China) to remove genomic DNA contaminants. To ensure the quality of the samples for transcriptome sequencing, concentration and integrity of RNA samples were checked using a Nanodrop ND-1000 spectrophotometer (NanoDrop Technologies, Wilmington, DE) and an Agilent Bioanalyzer 2100 (Agilent, Palo Alto, CA), respectively. The qualified RNA samples were used for mRNA preparation and cDNA library construction. Three independent biological replicates were performed for every treatment. cDNA libraries were sequenced as described by Meng et al. (2019) [41].

Sequence annotation
To ensure the accuracy of sequence alignment, raw reads were cleaned by removing adapter and primer sequences, reads with ambiguous nucleotides larger than 5% and lowquality reads. Clean reads were aligned to the reference genome of T. castaneum (Reference genome version: GCF_ 000002335.3_Tcas5.2_genomic.fna; Reference genome site: ftp://ftp.ncbi.nlm.nih.gov/genomes/all/GCF/000/002/335/ GCF_000002335.3_Tcas5.2/GCF_000002335.3_Tcas5.2_ genomic.fna.gz) [21] using HISAT2 (CCB, Johns Hopkins University, USA). Subsequently, the transcripts were subjected to blast against GO, KOG and KEGG databases. GO is an international standard classification system of gene function ( Figure S1 and Table S2). KOG database is based on the phylogenetic relationship of bacteria, algae and eukaryotes (Table S2). The KEGG database provides a powerful tool to discover the pathways in which genes are involved ( Figure S2). In addition, Nr, COG, Pfam, GO and Swiss-Prot databases were also used to annotate unigenes. The E-value threshold was set to 10 − 5 .

Differentially expressed genes analysis
Transcript abundances were measured by FPKM [61]. Differential expression analysis of dsTcAMPKα and dsEGFP groups were performed using DESeq2 software [2]. Unigene with False Discovery Rate (FDR) values < 0.05 and fold change |log 2 (treatment/control) | > 1 were set as the threshold for DEGs. GO and KOG database annotations were further analyzed to understand the functions of DEGs in T. castaneum.

Reverse transcription quantitative real-time polymerase chain reaction (RT-qPCR)
To validate the DEGs from the RNA-sequencing, RT-qPCR reactions were performed on the Bio-Rad CFX 96 Real-time PCR system using TB Green™ Premix Ex Taq™ (Takara, Dalian, China) and gene specific primers (Table  S1). The stably expressed gene encoding ribosomal protein S3 (rps3, GenBank: CB335975) was used as a reference gene [5]. PCR conditions were set as an initial incubation of 95°C for 30s, 40 cycles of 95°C for 5 s and 60°C for 30s, and a final melting curve analysis was performed. The mRNA levels were normalized to reference gene with the 2 -ΔΔCT method Livak and Schmittgen [35]. The means and standard errors (mean ± SE) for each time point were obtained from the average of at least three biologically independent sample sets.

Statistical analysis
Statistical analysis was performed using Graphpad Prism 6 (GraphPad Software Inc., San Diego, USA) by one-way analysis of variance, followed by Tukey's Honestly Significant Difference test. All data are presented as the mean ± SE.