Activation of the TOR Signalling Pathway by Glutamine Regulates Insect Fecundity

The target of rapamycin (TOR) positively controls cell growth in response to nutrients such as amino acids. However, research on the specific nutrients sensed by TOR is limited. Glutamine (Gln), a particularly important amino acid involved in metabolism in organisms, is synthesised and catalysed exclusively by glutamine synthetase (GS), and our previous studies have shown that Gln may regulate fecundity in vivo levels of the brown planthopper (BPH) Nilaparvata lugens. Until now, it has remained unclear whether Gln activates or inhibits the TOR signalling pathway. Here, we performed the combined analyses of iTRAQ (isobaric tags for relative and absolute quantification) and DGE (tag-based digital gene expression) data in N.lugens at the protein and transcript levels after GS RNAi, and we found that 52 pathways overlap, including the TOR pathway. We further experimentally demonstrate that Gln activates the TOR pathway by promoting the serine/threonine protein kinase AKT and inhibiting the 5′AMP-activated protein kinase AMPK phosphorylation activity in the pest. Furthermore, TOR regulates the fecundity of N. lugens probably by mediating vitellogenin (Vg) expression. This work is the first report that Gln activates the TOR pathway in vivo.

serine/threonine protein kinase AKT and inhibits the 5′ AMP-activated protein kinase AMPK phosphorylation activity. Furthermore, TOR regulates the fecundity of N. lugens probably by mediating vitellogenin (Vg) expression. This study is the first report to show that Gln activates the TOR pathway in vivo.

Results
Global changes at the protein and transcript levels after GS RNAi. To ensure that GS knockdown was effective in the 2 samples, the mRNA expression of the GS gene was detected by qRT-PCR (Supplementary Figure 1). Afterwardsthe iTRAQ Mass-Tagging mass spectrometric results from the two repetitions of the GS-RNAi and GFP-RNAi experiments were analysed using the ProteinPilot TM software 2.0.1 (Applied Biosystems, USA). N. lugens lacks a reference genome; therefore, the annotation of its proteins primarily depends on its EST databases and other insect databases 17 . This analysis produced a quantitative proteome of 958 proteins for the GS knockdown experiment ( Table 1). The GS knockdown experiment resulted in 124 proteins that were statistically significant responders (p < 0.01), which is approximately 12.94% of the identified quantitative proteome. Among these proteins, 49 and 75 proteins were up-and down-regulated, respectively (Fig. 1A, Supplementary Table 1). Regarding the down-regulated proteins, we found the RNAi target protein GS (NLU012724.1) and fecundity-related protein vitellogenin (Vg) (NLU019204.1) (Fig. 1B), indicating that the RNAi was effective. Several TOR pathway proteins of interest, including two key upstream activators of Rheb (NLU021861.1) and 3′ -phosphoinositide-dependent kinase (PDK1) (NLU017432.3), were also down-regulated.
The differentially expressed genes following RNAi treatment were analysed using the Digital Gene Expression (DGE) method. The results revealed 293 genes with significantly differential expression between the GS-RNAi and GFP-RNAi libraries, including 151 up-and 142 down-regulated genes (Fig. 1C). In the down-regulated genes, we also found GS and Vg in addition to several TOR pathway genes such as TOR (NLU007095.1), Rheb and S6 K (NLU007225.2). According to the GO classification, most of the differentially expressed genes correlated to biological processes, including the reproductive process and various developmental processes (Fig. 1D).
With the iTRAQ and DGE high-quality data sets, we performed an integrated analysis of the proteome and transcriptome. The overall analysis of the robustly regulated genes at the protein and mRNA levels found 5 genes that overlap in the up-regulated group and 7 genes in the down-regulated group (Fig. 1E, Supplementary Table 2). Pathway enrichment analysis for the proteins and transcripts was also performed using the KEGG database. In addition to the "unable to classify" group, the protein metabolism, carbohydrate metabolism and amino acid metabolism groups were ranked in the top 3 most enriched pathway ontology groups (Fig. 1F). Fifty-two overlapping pathways were found between iTRAQ and DGE (Fig. 1G), including the TOR signalling pathway.
Glutamine activates the TOR pathway in vivo. To validate the above result suggesting that GS knockdown influences the TOR signalling pathway, several experiments were conducted. After studying the feature of GS (Supplementary Figure 2), we depleted GS for 48 h and detected the transcript change of the TOR pathway-associated genes. The RNAi was effective because the GS mRNA and protein levels as well as enzymatic activity decreased ( Fig. 2A,B). Three TOR pathway-associated genes, Rheb, TOR and S6 K, had significantly lower expression levels after GS RNAi compared to the control (dsGFP) ( Fig. 2A), indicating that a positive relationship exists between GS and TOR.
Secondly, we detected the phosphorylation of S6 K, a molecular marker of TOR pathway activation 18 . After injecting 250 ng dsTOR or 10 mM rapamycin into the BPH adults, the phosphorylation of S6 K decreased (Supplementary Figure 3). Similar results were observed with the injection of dsGS or L-methionine sulfoximine (MSX), a GS specific inhibitor 8 (Fig. 3A, lane 2 and 5). When Gln (the GS specific product) was added into dsGS or MSX, S6 K phosphorylation increased (Fig. 3A, lane 3 and 6). When feeding the BPH nymphs with the artificial diet (with Gln) 19 , S6 K phosphorylation increased compared to the diet without Gln (Fig. 3B, lane 2 and 4). When dsGS was added into the diet, S6 K phosphorylation also decreased (Fig. 3B, lane 1). These results show that Gln activates the TOR pathway via the phosphorylation of S6 K in vivo. In addition, the Vg mRNA and protein levels were down-regulated after the injection of dsGS and MSX (Fig. 2C), indicating that GS regulates BPH fecundity 20 .

TOR regulates fecundity in BPH.
After the injection of dsTOR into the BPH adults, the TOR mRNA level significantly decreased by 59.66%, 68.66% and 66.68% at 24 h, 48 h and 72 h, respectively, compared to the animals injected with dsGFP (Fig. 4A). The expression level of the Vg gene was also significantly down-regulated by 47.66%, 52.74% and 59.68% after dsTOR injection. When the BPH adults   were sampled 48 h after injectedion of rapamycin or dsTOR, significant reductions in the Vg protein level was observed (Fig. 4B, Supplementary Figure 3). As a result, emerged brachypterous BPH adults had fewer offspring (53.22 per pair, a 71.21% decrease). Most dsTOR-treated adults had less than 100 offspring; however, 200-300 nymphs were observed for the majority of the dsGFP-treated pairs (Fig. 4C). These results suggest that TOR regulates the fecundity of N. lugens probably by mediating Vg expression.

Gln promotes AKT and inhibits AMPK.
To further study the molecular mechanism of the activation of the TOR pathway via Gln, we chose two key upstream genes of the TOR pathway, AKT and AMPK. After silencing GS through RNAi, both the enzyme activity and phosphorylation of AKT decreased. When Gln was added, the enzyme activity and phosphorylation of AKT were up-regulated (Fig. 5A,B), suggesting that Gln promoted AKT. However, the opposite trend was observed when AMPK was measured (Fig. 5C,D), indicating that Gln inhibited AMPK. When AKT was knocked down using RNAi, the TOR mRNA level significantly decreased (Fig. 5E). More importantly, GS activity decreased by 60.65% (Fig. 5F); however, its mRNA level did not significantly change (Fig. 5E), implying that AKT has a positive feedback mechanism that acts on GS.
In conclusion, GS regulates the content of Gln, which activates the TOR signalling pathway by promoting AKT and inhibiting AMPK. Furthermore, TOR regulates the fecundity of N. lugens probably by mediating Vg expression (Fig. 5G).

Discussion
Targeting a gene via RNAi is often used to study its function.However, the reports on the global changes of mRNA and protein profile that occur after RNAi targeting of a specific gene have been limited 21,22 . For example, after knock down of ISWI in Drosophila SL2 cells, Bonaldi et al. studied the global changes of transcriptome and proteome using microarray and SILAC-MS technologies, and found that there is a very little correlation between mRNA and protein level 23 . In cultured male-like tissue cells of D. melanogaster, the expression level of many genes on the X chromosome decreased after RNAi targeting MSL complex; however, the gene expression on autosomes remained largely unchanged 24 . In our study, the combinative analyses of iTRAQ proteomic and DGE transcriptomic data after GS RNAi revealed 52 overlapping pathways, including the TOR pathway (Fig. 1). We identified 12 genes in which 5 were up-regulated and 7 were down-regulated at both mRNA and protein level (Supplementary Table 2). Similar to the previous report 23 , the fold of change in mRNA and protein levels was different. Recently, accumulated evidences showed that the TOR pathway regulates protein synthesis and cell proliferation through integrating multiple upstream signals 25 . In N. lugens, the knockdown of TOR leads to decreased fecundity that is correlated with lower Vg expression (Fig. 4), implying that TOR regulates its fecundity probably by mediating Vg expression. However, TOR could also affect fecundity by other mechanisms. While in A. aegypti, TOR pathway is essential for the activation of Vg expression 26,27 , TOR-dependent activation of S6 kinase is a crucial step in the transduction of nutritional information during egg development of mosquitoes 28 . Moreover, in Drosophila, TOR regulates ovary size 29 , and is a determinant of follicle number by promoting cell survival 30 . In Apis mellifera, TOR governs diphenic development 31 . Both phenotypic and mechanistic studies are desired in order to explain roles of TOR in regulating fecundity of N. lugens.
TOR is a well known nutrient sensor in many organisms 8,32 . In multiple types of cancer cells, a positive correlation between Gln and TOR was observed [12][13][14] . Consistent with this, we found that Gln activates the TOR pathway via phosphorylating S6 K (Fig. 3). However, a different result was reported when Ba/F3 cell line was used 15 . Possible explanation is that the effect of Gln on TOR signalling may be cell type dependent 15 . Another possibility concerns a threshold effect of Gln concentration in cells. Since all experiments of this study were conducted in vivo, the findings may be more physiologically relevant.
Interestingly, we also found that Gln apparently promotes AKT activity, while inhibiting AMPK activity in N. lugens (Fig. 5G). This finding provides evidence that Gln regulates TOR not by a single pathway but by a network composed of growth (insulin) and energy pathways. However, it is still not clear how Gln stimulates AKT phosphorylation/activation. In mammalian systems, Gln stimulates synthesis and secretion of insulin-like growth factor 2 (IGF2) as well as IGF2-dependent AKT phosphorylation in beta cell 33 . In addition, Gln, Leu and Pro increase AKT phosphorylation in HepG2 liver cells 34 , presumably, through the inhibition of NF-κ B. Also, how Gln inhibits AMPK is an open question. It is thus interesting and important to understand the mechanism underlying Gln-mediated AKT activation and AMPK inhibition. Western Blotting. Western blotting analyses were performed using standard techniques. The proteins were separated on a 12% SDS-PAGE gel and transferred to PVDF membranes (0.4 μ m, Millipore), and the membranes were immunoblotted with AKT, phospho-S473 AKT, phospho-pT183+ T172 AMPK (1:2000, Abcam, UK), phospho-T389 S6 K (1:1000, CST, USA), GS, Vg and S6 K (1:5000) serum (prepared by our lab). IgG goat anti-mouse and goat anti-rabbit antibodies conjugated with HRP were used for secondary antibodies (1:5000, Abcam, UK), and the membranes were visualised by ECL (enhanced chemiluminescence).

Methods
Quantitative Real-Time PCR Analysis. The primers used for real-time PCR are listed in Table 1.
The synthesised first-strand cDNA was amplified by PCR in 10 μ L reaction mixtures using a Light Cycler 480 system (Roche, USA), and the β-actin gene was used as an internal standard 37 . After the amplifications, a melting curve analysis was performed in triplicate, and the results were averaged. The quantitative variation was calculated using three independent biological samples by a relative quantitative method (2 −ΔΔCT ) 38 .
RNA Interference and Sampling. The dsRNA of a target gene was produced using the T7 RiboMAX ™ Express RNAi System (Promega, USA). After synthesis, the dsGS (439 bp), dsTOR (457 bp), dsAKT (447 bp) and dsGFP (688 bp) were quantified by ultramicro-spectrophotometers (NanoDrop 2000, Thermo) and were maintained at − 80 °C until use. The sequence was verified by sequencing (Invitrogen Company, Shanghai, China). To deliver dsRNA into the body of BPHs, the dsRNA was added to the artificial diet 39 or directly injected 40 . Before injection, the dsRNA and phenol red solution were mixed for observation. The purified dsRNA or ddH 2 O were slowly injected using 3.5 Drummond needles and the NARISHIGE IM-31 (Nikon, JAPAN).
AKT and AMPK Activity Quantitative Test. AKT and AMPK quantitative enzymatic activity test kits (Genmed Scientifics, Inc.) were used to conduct to detect the activity. According to the manufacturer's protocol steps, the absorbance was read at 340 nm. Finally, the AKT and AMPK active quantitative indexes were calculated using separate formulas 41 . Glutamine Synthetase Activity Assay. Glutamine synthetase (GS) was assayed using the method with a modification 42  FeCl 3 , 0.2 M TCA and 0.6 M HCl) and placed at room temperature after a moment. After centrifugation, the absorbance of the supernatant was read at 540 nm against a reagent blank. Statistical analysis. The results are expressed as the means± S.E.M. SPSS 13.0 software was used to perform t-tests to identify significant differences at a 95% or 99% confidence level.