Amino Acids Regulate Expression of Antizyme-1 to Modulate Ornithine Decarboxylase Activity

Background: Amino acids regulate ornithine decarboxylase (ODC) activity through antizyme1 (AZ1).


INTRODUCTION
The polyamines, spermidine and spermine and their diamine precursor putrescine, are required for the proliferation of eukaryotic cells and the growth of tissues (1,2,3). In addition, polyamines are involved in the regulation of a variety of other cellular functions such as transmembrane ion flux, apoptosis and migration (4,5,6). The importance of polyamines to cellular function has been demonstrated in both normal and cancer cells from a variety of tissues (7,8). The intracellular levels of polyamines are closely controlled and primarily depend on the activity of ornithine decarboxylase (ODC; EC 4.1.17), which catalyzes the first rate-limiting step in polyamine synthesis. ODC activity itself is highly regulated and varies in response to many different stimuli in different tissues and cells. In general, activity increases dramatically and transiently in response to trophic hormones, growth factors, other proliferative stimuli and damage (1,9,10,11). ODC activity has one of the shortest half-lives of any mammalian enzyme (12,13). The intracellular concentration of polyamines is controlled at several steps, including their uptake and their biosynthesis. The latter is predominantly achieved by controlling ODC activity by a mechanism involving ODC antizyme (AZ) (14,15). AZ was first identified in mammals and exists in several isoforms (AZ1, AZ2, AZ3, AZ4) (16)(17)(18). AZ1 prevents the formation of active ODC homodimers by forming ODC/AZ heterodimers (19,20). Furthermore, AZ1 binding mediates ODC degradation by the 26S proteosome (20)(21)(22). AZ levels increase with the elevation in intracellular concentration of polyamines. The induction of AZ by polyamines thus, constitutes a feedback control in polyamine homeostasis. AZ expression is regulated by polyamines in a unique manner during translation of the AZ mRNA, which involves a programmed +1 frameshift, resulting in the expression of the functional full-length AZ protein (18,23). AZ binds to and accelerates the degradation of ODC (24), and other proteins shown to regulate proliferation and cell death such as Aurora-A, cyclinD1, and Smad1 (25)(26)(27)(28).
Among normal dietary constituents, certain amino acids markedly increase ODC activity and mucosal growth when administered intragastrically (29,30). The effects of amino acids on ODC activity have been shown to occur in a wide variety of tissues and cells. Fausto (1) found that some amino acids could increase ODC activity in regenerating liver. Chen and Canellakis (13) and Chen and Chen (31) examined the ability of amino acids to induce ODC activity in cells incubated in a simple saltglucose solution (EBSS), thus, eliminating stimulation from growth factors present in serum or other amino acids and nutrients normally found in incubation medium. They showed that ODC activity was increased in neuroblastoma cells by asparagine (ASN) and glutamine. Additional studies by the same group (32)(33)(34) found that, although ASN was the most effective, other amino acids transported by one of the Na +dependent systems could also induce ODC activity. While induction of ODC activity was dependent on Na + , it was independent of metabolism or the ability of the amino acid to be incorporated into protein. However, other amino acids like lysine (LYS) and valine (VAL) were unable to induce ODC in EBSS. Of particular interest was the finding, in a variety of cell types, that the induction of activity by hormones and growth factors required the presence of ASN or a similar amino acid (32,33).
In previous work, our laboratory reported that ASN but not EGF could stimulate ODC activity in intestinal epithelial cells incubated in EBSS. EGF elicited a potentiated response in the presence of ASN. Although induction of ODC activity did not involve the synthesis of new ODC protein, it did require both transcription and translation (35). These data indicate that amino acids either induced or repressed a specific regulator of ODC enzyme activity.
Recent studies have shown that the mammalian target of rapamycin (mTOR) regulates the translation machinery by controlling the phosphorylation of both 4E-BP1 and p70-S6 ribosomal protein kinase (p70S6K) via mTORC1 and mTORC2 (36,37). GLN has been shown to be essential for the metabolism of intestinal mucosa and for proliferation of colon cancer cells (38). In most studies, cellular deprivation of a single essential amino acid or branched chain amino acid decreased mTOR signaling and inhibited protein translation. However, the effect of ASN and GLN on mTOR signaling is unknown.
The significance of the current study is the finding of a unique mode of regulation of AZ1 expression. Generally, amino acid starvation inhibits protein synthesis, and the addition of amino acids restores it. However, AZ1 synthesis occurred in the absence of amino acids, and its synthesis was prevented by the addition of specific amino acids. This makes sense since AZ inhibits ODC and in the absence of amino acids ODC activity is unable to stimulate or support growth. Therefore, this study provides an important insight into the complexity of amino acid signaling and, thereby, its role in the regulation of various cellular processes. We show that ASN stimulated ODC activity by preventing the expression of AZ1 without a change in the level of ODC protein. The disappearance of antizyme is essential for EGF to increase the activity of the enzyme. ASN significantly decreased the levels of AZ1 mRNA and protein. Although, amino acids normally regulate protein translation via the mTOR pathway, AZ1 synthesis occurred independently of mTORC1 but depended on mTORC2 signaling. ASNmediated inhibition of AZ1 synthesis involved the dephosphorylation of 4EBP1 by mTORC2. This is the first demonstration of the mechanism whereby ASN and related amino acids regulate ODC activity.

Plasmids
Three AZ1 SiRNA oligonucleotide sequences were designed and cloned into a plasmid vector (pcDNA6.2-GW/EmGFP-MiR) and confirmed by sequencing using appropriate primer pairs. Selected clones for the vector (MiR-LacZ-EGFP) and AZ1 (MiR-LacZ-AZ1-EGFP) were used to prepare plasmid DNA for the transfection of IEC-6 cells using EndoFree Plasmid Maxi kit from QIAGEN. and Mg 2+ . They were counted and plated (day 0) at 6.25 104 cells/cm2 in DMEM plus 5% dialyzed FBS, 10 µg insulin, and 50 µg gentamicin sulfate/ml. The cells were fed on day 2. On day 3, the medium was removed, the cells were washed once with Hanks' balanced salt solution, and the medium was replaced with serum-free DMEM with 10 µg insulin and 50 µg gentamicin sulfate/ml. On day 4, the medium was removed, the cells were washed once with Hanks' balanced salt solution, and the medium was replaced with Earle's balanced salt solution (EBSS) containing whatever additions were required by the experiment.
ODC assay: The activity of the enzyme ODC was assayed with a radiometric technique in which the amount of 14 CO 2 liberated from DL-[ 1-14 C] ornithine was estimated (39). Briefly, after experimental treatment, the dishes were placed on ice; the monolayers were washed three times with cold Dulbecco's PBS (DPBS); and 0.5 ml of 1 mM Tris buffer (pH 7.4) containing 1 mM EDTA, 0.05 mM pyridoxal phosphate, and 5 mM dithiothreitol was added. The cells were frozen at -80°C until assayed. At the time of assay, the cells were thawed on ice, scraped into Microfuge tubes, sonicated, and centrifuged at 12,000 g at 4°C for 10 min. The ODC activity of an aliquot of supernatant was incubated in a stoppered tube in the presence of 6.8 nmol of [ 14 C] ornithine for 15 min at 37°C. The 14 CO 2 liberated by the decarboxylation of ornithine was trapped on a piece of filter paper impregnated with 20 µl of 2 N NaOH, which was suspended in a center well above the reaction mixture. The reaction was stopped by the addition of TCA to a final concentration of 5%. The 14 CO 2 trapped in the filter paper was measured by liquid scintillation spectroscopy. Aliquots of the supernatant were assayed for total protein with the method described by Bradford (40). Enzymatic activity is expressed as picomoles of CO 2 per hour per milligram of protein.

RNA extraction, reverse transcription, and real-time PCR:
Serum starved cells incubated in EBSS with or without ASN for 4h were washed twice with DPBS followed by addition of TRIzol reagent for the extraction of total RNA following the manufacturer's (TRIzol) instructions. The yield and quality of total RNA was measured by absorbance at 260/280 nm. One µg total RNA and 0.5 µg of the random primers were used for reverse transcription with M-MLV RT following the manufacturer's instructions. The resulting cDNA was diluted to 100 µl with DEPC treated water and used as a template for realtime PCR. Briefly, PCR primers were designed with a melting temperature (Tm) of 59-61 °C. Amplicon size was 50-150 bases. Forward and reverse primers spanning exonexon junctions were selected to avoid amplification of genome sequences (AZ1: Forward primer -5'-ACTGCTTCGCCAGAGAGAA, Reverse primer-ATCTTCAGGGGT GGGTGAG-3'; ACTIN: Forward primer-5'-GCTGAAAAGATGACCCAGATCA, Reverse primer-CACAGCCTGGATGGCTACGT-3'. PCR reactions were performed using an ABI Prism 7700 Sequence Detection System (Applied Biosystems).
The reaction contained 25µl SYBR reagent, 5 µl diluted cDNA, and 300-900 nM primers in 50 µl volume. The thermal cycling conditions involved an initial denaturing step at 95 °C for 15 sec and an anneal-extension step at 60 °C for 1 min. Quantitative values were obtained from the threshold cycle value (Ct), which is the point where a significant increase of fluorescence is first detected. The transcript number of rat ß-actin was quantified as an internal RNA control, and each sample was normalized on the basis of its β-actin content.
The relative gene expression level of each sample was then normalized to the EBSS (control). Final results, expressed as N-fold difference in gene expression relative to ß-actin and the control group, termed N, were calculated as: SiRNA Transfection: AZ1 and AZ2 specific siRNA (20-25 nucleotides) derived from the Rattus norvegicus (Santa Cruz Biotechnology) were used for the transient transfection of IEC-6 cells. FuGENE-HD transfection reagent was mixed with 10 µM SiRNA (3:5 µl) in serum-free medium to a total volume of 100 µl and incubated for 45 min at room temperature. IEC-6 cells at a relatively early passage were grown to 60% confluence in 6-well plates. For transfection, the cell monolayer was rinsed with serumfree medium, and the siRNA/FuGENE-HD mixture was added drop by drop onto the cell monolayer and incubated for 24 h followed by incubation in serum containing medium for an additional 24h at 37 °C. Transiently transfected cells were serum starved for 24 h and rinsed with Hank's balanced salt solution and further incubated with serum free medium or EBSS for 3h. Cells were washed with DPBS and lysed with MPER containing protease and phosphatase inhibitors to determine the expression of AZ1. AZ1 SiRNA sequence was also cloned in MiR-LacZ-EGFP vector as per manufacturer's instruction for the stable transfection of cells to knock down AZ1. Empty vector (MiR-LacZ-EGFP) or AZ1 (MiR-AZ1-EGFP) plasmid DNA (10 µg) was transfected as described above. The success of transfection was ascertained by monitoring GFP expression, and stable clones were selected using blasticidin. Isolated clones were characterized by determining the expression of AZ1and induction of ODC activity by ASN. 35 S-methionine/cysteine incorporation and AZ1 synthesis: Confluent serum starved IEC-6 cells grown in 6 well plates (triplicate for each treatment) or T75 flasks were incubated in EBSS with or without ASN in the presence of 10 µci/ml 35 S-Methionine/Cystine for 2 and 4h. 6-well plates were washed 5 times and cells were lysed using MPER containing protease inhibitors. The supernatant was used to determine the protein concentration by the BCA method. An equal amount of protein from each sample was precipitated using TCA and resuspended in SDS sample buffer at a concentration of 1 µg/µl. Incorporation of 35 S-Methionine/Cystine was measured by liquid scintillation spectroscopy. Equal amounts of protein (30 µg) were resolved on 12% SDS PAGE and analyzed by western blotting for AZ1 expression. Cell extracts prepared from T-75 flasks (900 µg protein) were incubated with protein-AG plus agarose beads (25 µl) for 2h at 4 0 C on a rocker followed by centrifugation. Precleared cell extracts were incubated with AZ1 antibody (3 µg) followed by addition of protein-AG plus agarose beads (25 µl) and processed as mentioned above. Agarose beads were washed twice, resuspended in SDS sample buffer and resolved on 12% SDS PAGE. Immunoprecipitated AZ1 transferred on PVDF membrane was sprayed with Enhance reagent and exposed to X-ray film overnight at -70 0 C. Cap binding assay: Cells treated as indicated were washed with DPBS and lysed in binding buffer (150 mM NaCl, 50 mM Tris pH 7.5, 50 mM NaF, 10 mM Na pyrophosphate, 1 mM EDTA, containing phosphatase and protease inhibitor cocktails. Protein (500 µg) from the lysates in 500 µl of cap binding buffer was added to 50 µl pre-washed 7-methyl GTP-sepharose beads (illustra TM from GE Healthcare UK Limited, Buckinghamshire, UK) and Slurry was incubated on a rocker for 2 hr at RT. Samples were washed twice with cap binding buffer, followed by the addition of sample buffer, boiled, and equal amounts were resolved by SDS-PAGE and transferred onto PVDF membrane. Blots were probed for indicated antibodies.
Western blot analysis: After experimental treatments, IEC-6 cells were washed twice with DPBS. The DPBS was removed, 500 µl of MPER containing protease and phosphatase inhibitors was added, and the plates were frozen overnight and scraped. The cell lysate was centrifuged at 10,000 rpm for 10 min. The supernatant was used to determine the protein concentration by the BCA method. Total cell protein (30 µg) was separated on 10% SDS-PAGE and transferred to a PVDF membrane for Western blotting with a specific primary antibody and appropriate secondary antibody labeled with horseradish peroxidase (HRP). Immune complexes were detected by chemiluminescence system. Statistics: All data are expressed as means +/-SE. Densitometric analysis of western blots from three different experiments was performed. Analysis of variance and appropriate post-hoc testing determined the significance of the differences between means. Values of p < 0.05 were considered significant.

Expression of AZ1, AZin, and ODC-
Based on the known steps in the regulation of ODC activity, we hypothesized that the effects of ASN on ODC activity might be due to the inhibition of AZ or stimulation of antizyme inhibitor (AZin) expression. IEC-6 cells grown to confluence and serum starved for 24 hours were used throughout the investigation. Serum starved cells were incubated in EBSS with or without additives for the indicated period. When IEC-6 cells were incubated in EBSS for 3h, a 20-kDa isoform of antizyme-1 (AZ1) was detected (Fig.1A) using a specific antibody generated against Rat AZ1 protein (41). The appearance of AZ1 was prevented by ASN. To confirm the involvement of AZ1, IEC-6 cells transiently transfected with AZ1 and AZ2 specific SiRNA were exposed to serum free DMEM or EBSS for 3h, and the levels of AZ1 protein were determined by western blotting. Results in figures 1B and C show that AZ1 specific SiRNA significantly decreased basal AZ1 as well as the expression of AZ1 induced during the incubation with EBSS. However, AZ2 SiRNA failed to affect the expression of AZ1. Since AZ2 SiRNA failed to alter the expression of AZ1 despite sharing high amino acid homology with the AZ1, it indicates that AZ SiRNAs had no off-target effects. Furthermore, transfection of plasmid based AZ1 SiRNA decreased AZ1 expression in EBSS and significantly increased ODC activity compared to that induced in vector-transfected cells in response to ASN ( Fig. 1D and E) confirming our SiRNA results shown in figure 1B. These results indicate that EBSS induced the expression of AZ1.
We determined whether the regulation of AZ1 expression by ASN is unique to IEC-6 cells or is a more general phenomenon by measuring the ODC activity and AZ1 levels in NIH3T3 cells. Addition of ASN to EBSS significantly increased ODC activity in both cell types, although the increase was significantly higher in NIH3T3 cells (Fig. 1F). The basal level of AZ1 in EBSS was lower in NIH3T3 cells compared to that observed in IEC-6 cells (Fig. 1F). The addition of ASN significantly decreased AZ1 expression in both cell types (Fig. 1F). Therefore, despite differences in levels of ODC activity and protein, ASN had a similar effect in both cell lines. Addition of 10 mM ASN in EBSS decreased AZ1 protein (Fig. 1A). The effects of EGF and ASN+EGF on AZ1 expression were more or less similar to those of EBSS and EBSS+ASN respectively. ODC and AZin proteins were unchanged in response to EBSS, ASN, EGF, or ASN+EGF (Fig. 1A).
Thus, ASN appears to regulate AZ1 expression, and EGF may have a secondary role in ODC activation.

Effect of ASN on expression of AZ1 and
ODC activity-In order to gain insight into the mode of action of amino acids, we examined the effectiveness of ASN in a short term and extended time course study. In the presence of ASN, ODC activity was detected within 60 minutes and continued to increase for at least 180 minutes, which correlated with the absence of AZ1 (Figs. 2A and B). Furthermore, the levels of AZ1 protein continue to increase in EBSS through 6 hours ( Fig. 2C and D). Although, ASN prevented the expression of almost all AZ1, a significant amount of AZ1 was detected after 4h incubation compared to 0h (Fig. 2C, lanes 8, 9 and D). ASN-induced ODC activity increased rapidly for 4h and then began to decline, indicating a regulatory role of AZ1 in the activation of ODC by ASN. Furthermore, it is apparent that the very low levels of AZ1 expression at 4h and 6 h efficiently reduced ODC activity (Fig. 2E). The levels of ODC protein remained unchanged throughout the EBSS and ASN treatment (Fig. 2C, upper panel).
Thus, AZ1 appears to be a high affinity modulator of ODC activity. To test this notion, cells were incubated in EBSS for 2, 30, and 60 minutes to allow accumulation of low levels of AZ1 and were exposed to EBSS or ASN for 3h. ODC activity in response to ASN was significantly decreased in cells preincubated with EBSS for 30 and 60 minutes (Fig. 3A). These observations were confirmed by measuring the ODC activity of ASN-stimulated cell extract combined with the extract of cells treated with EBSS for 3h and 6h. EBSS cell extract significantly decreased ASN-induced ODC activity (Fig. 3B). These data indicate that ODC activity largely depends on the level of AZ1 protein, and that its expression is regulated by ASN in IEC-6 cells.
Polyamines are known to stimulate AZ expression by a frame shifting mechanism (41)(42)(43)(44). We used putrescine to augment AZ1 expression in EBSS. Figure  3C shows that putrescine (10 µM) significantly increased AZ1 compared to that seen in EBSS (Fig. 3C). AZ1 immunoprecipitated from extracts of cells treated with EBSS and EBSS plus putrescine significantly inhibited ODC activity induced by ASN (Fig. 3C, upper panel and D). The coomassie brilliant blue stained gel shows the level of immunoprecipitated AZ1, which was higher in EBSS plus putrescine compared to EBSS alone (Fig. 3D, right panel). The extent of ODC inhibition correlated with the level of AZ1 (Fig. 3D). These results confirmed the role of AZ1 in the regulation of ODC activity.
To identify the mechanism by which ASN regulates AZ1 expression, we allowed AZ1 to accumulate in EBSS treated cells for 3h. We then incubated cells for an additional 3h in the presence of ASN and studied the disappearance of AZ1 ( Fig. 4A and B). The prolonged incubation in EBSS (total 6h) significantly increased AZ1 protein levels. Addition of ASN after 3h incubation in EBSS nearly eliminated the accumulated AZ1 within 1h (Fig. 4A, lane 7 and B). Thus, the effectiveness of ASN in eliminating AZ1 appears to depend on the level of preexisting AZ1 protein. Prolonging the incubation in EBSS beyond 6h caused AZ1 protein to decrease spontaneously within 1.5h (Fig.  4C, lanes 3 and 4 and D). Addition of ASN after 6h incubation in EBSS further decreased AZ1 (Fig. 4C, lane 7 and D). The decrease in AZ1 in EBSS was about 45%, and ASN accelerated it to about 80% (Fig.  4B, lanes 4 and 7).
We determined the stability of AZ1 in EBSS for 4h to avoid the spontaneous degradation phase (6h), and then added CHX (cycloheximide), ASN, or ASN+CHX and monitored AZ1 levels over the course of 120 mins (Fig. 5A). In EBSS, the level of AZ1 increased for 120 min, after which the addition of CHX caused a rapid decrease resulting in the loss of about 50% of AZ1 protein within 75 minutes, which was similar to the decrease observed with EBSS+ASN. Addition of CHX to the EBSS+ASN group did not significantly affect the decrease in AZ1 compared to EBSS+ASN alone (Fig. 5B). ASN significantly decreased the expression of AZ1 mRNA as measured by the real time polymerase chain reaction assay (RT-PCR) (Fig. 5C). Inhibition of transcription by ACT-D failed to prevent AZ1 synthesis in EBSS (Fig. 5D) indicating that ongoing transcription may not control the synthesis of AZ1.
Actinomycin-D completely prevented the expression of p21Waf1/Cip1 protein within 1h and β-actin after 3h indicating an effective inhibition of transcription (data not shown). Therefore, we used coomassie brilliant blue staining to confirm equal loading of the samples.
In order to confirm the role of protein synthesis in the expression of AZ1, we measured 35 S-methionine/cysteine incorporation into proteins synthesized in EBSS with and without ASN for 2 and 4h. First, we found that the addition of methionine/cysteine in EBSS had no effect on AZ1 synthesis (data not shown). We then analyzed equal amounts of protein for the incorporation of 35 S-methionine/cysteine by scintillation spectroscopy. Figure 6A shows that ASN significantly decreased the incorporation of radioactive amino acids into newly synthesized proteins compared to that seen in EBSS. TCA precipitated proteins from 35 S-labelled samples were resolved by SDS-PAGE, transferred to a PVDF membrane, and analyzed by western blotting. The membrane was washed, dried, sprayed with the Enhance reagent and exposed to X-ray film as described in the methods. Results in figure 6B show that in the presence of ASN there was no detectable AZ1 expression at 2h, and only a slight increase at 4h (Fig. 6B, middle panel). A band from the superimposed western blot and 35 S-labelled film shows decreased labeling of the band corresponding to AZ1, which began to appear at 4h in the presence of ASN. However, increasing intensity of the band was noted at 2 and 4h in the presence of EBSS (Fig. 6B, upper panel). We further confirmed these observations by immunoprecipitation of AZ1 from the extracts. The newly synthesized AZ1 increased in a time dependent manner in EBSS, while it was reduced in the presence of ASN (Fig. 6C). A band marked by the asterisk showing the pattern identical to AZ1 might be a degradation product generated during storage and immunoprecipitation. It is known that AZ1 is highly unstable and requires a high concentration of salt and/or triton-X-100 for stabilization (45). We had not attempted to stabilize AZ1 to avoid effects on immunoprecipitation. These results support the conclusion that ASN modulated AZ1 levels by regulating its synthesis.
Effect of amino acids on the expression of AZ1-Previous studies including our own showed that like ASN, glutamine (GLN) and α-amino iso-butyric acid (AIB) increased and lysine (LYS) and valine (VAL) failed to increase ODC activity (35). These results led us to conclude that system A & N type amino acids are inducers, while other amino acids do not induce ODC activity. ASN, GLN, and AIB prevented AZ1 expression (Fig. 7A). Interestingly, VAL and LYS, noninducers of ODC activity, increased AZ1 expression, although the effect of VAL was weaker compared to LYS (Fig. 7B). Both LYS and VAL increased AZ1 protein in a time dependent manner (Fig. 7B). VAL significantly increased AZ1 levels compared to EBSS within 2h, and LYS induced AZ1 within 1h and caused strong increases at 2 and 3 h (Fig. 7C). These data suggest that LYS and VAL inhibit ODC activity via AZ1.
LYS, VAL, and ORN significantly decreased (40%) ODC activity stimulated by ASN, which correlated with levels of AZ1 ( Fig. 8A and B). VAL and ORN added together almost completely inhibited ODC activity. However, ODC protein levels were unchanged in response to the addition of these amino acids to EBSS (Fig. 8B, upper panel). Thus, it is clear from these results that LYS, VAL, and ORN oppose the effect of ASN in the clearance of AZ1. The growth medium, DMEM, contains amino acids including GLN (3.9 mM), LYS (1.0 mM), and VAL (0.8 mM) shown to have opposing effects on AZ1 expression. Therefore, the amino acid composition of culture medium should influence ODC activity. Addition of 10mM ASN to DMEM in the absence of amino acids increased ODC activity 3-fold compared to that observed in DMEM containing amino acids (Fig. 8C). Cells exposed to DMEM containing amino acids and EBSS for 3h showed significantly higher levels of AZ1 compared to that in DMEM (Fig. 8D). These results indicate that the amino acid composition of growth medium may affect ODC induction by regulating AZ1 levels. Effect of amino acids on mTOR and AZ1-Nutrient deprivation decreases the rate of global protein synthesis by regulating the levels or activities of the components of protein translation machinery. Eukaryotic initiation factor 4E binding protein (4EBP1) plays a crucial role in nutrient sensing mechanisms in eukaryotic systems (46). 4EBP1 binds eukaryotic initiation factor 4E (eIF4E) and prevents its association with eIF4G and, thereby, prevents translation. 4EBP1 is inactivated by phosphorylation, allowing eIF4E to initiate translation (46). The 7-methyl GTP-Sepharose binding assay showed slightly lower binding of eIF4E in cells incubated in DMEM without (AA-) amino acids compared to that seen in DMEM with amino acids (AA+) and addition of ASN to the DMEM (AA-) restored the binding of eIF4E to m7-GTP cap (Fig. 9A). Furthermore, m7-GTP cap bound 4EBP1 levels were undetectable compared to that seen in the DMEM without amino acids (AA-) and ASN added DMEM (AA-). Interestingly, the levels of phosphorylated 4EBP1 were higher in DMEM (AA-) compared to that seen in DMEM (AA+) and DMEM (AA-) to which ASN was added (Fig. 9A). However, the increase in p-4EBP1 in the absence of amino acids is slight in comparison to the increase in 4EBP1 itself. It is also possible that the p-4EBP1 is hypophosphorylated and may not dissociate from eIF4E. These results indicate that the absence of amino acids inhibits the translation of mRNA containing 5'-m7-GTP cap and that addition of ASN failed to reverse the inhibition of translation initiation. Furthermore, the phosphorylation of 4EBP1 was decreased in the DMEM (AA-) compared to DMEM (AA+). However, the levels of total 4EBP1 protein decreased in a time dependent fashion in DMEM (AA-). Addition of ASN to DMEM (AA-) prevented the rapid decrease in the 4EBP1 protein seen in DMEM (AA-). Overall the total amounts of 4EBP1 were higher in the DMEM (AA-) to which ASN was added (Fig. 9B). The increase in unphosphorylated 4EBP1 and inhibition of AZ1 expression in response to ASN suggested the involvement of the mTOR pathway. The activation of mTOR signaling was higher in the medium containing amino acids compared to that seen in the medium lacking amino acids. ASN further decreased active mTOR protein compared to AA+ and AA-conditions (Fig. 9C). Immunoprecipitation of mTOR further confirmed the phosphorylation state of mTOR in these conditions. In order to ascertain the role of mTOR on AZ1 expression, we determined the effect of rapamycin on AZ1 expression and p70-S6K activity in serum starved cells incubated for 4h in DMEM containing a full set of amino acids and DMEM lacking amino acids (Fig. 9D). Rapamycin indeed inhibited mTOR phosphorylation and p70S6K activity as evident by the complete inhibition of ribosomal protein-S6 phosphorylation in the cells incubated with amino acid containing medium. Interestingly, rapamycin prevented the phosphorylation of ribosomal protein S6 in both the presence and absence of amino acids and the activity of p70S6K decreased to a lower extent in the absence of amino acids. Furthermore, rapamycin failed to alter the levels of AZ1 protein in medium without amino acids (Fig. 9D). Since AZ1 synthesis is undetectable in DMEM containing amino acids, the effect of rapamycin on AZ1 expression in this condition can't be predicted. Addition of Asn, Leu, or Met to DMEM (AA-) failed to restore p70S6K activity to the levels seen in the presence of full set of amino acids (Fig. 9E). Rapamycin treatment for 4h had no effect on the expression of AZ1 in the presence or absence of ASN in the DMEM without amino acids (Fig. 9F). However, prolonging the treatment with rapamycin for 24 h induced AZ1 expression in DMEM (AA+) comparable to that seen in DMEM (AA-) (Fig. 9G). These results suggest that the mTORC1 inhibition induces AZ1 synthesis. Since rapamycin-independent mTOR signaling is mediated by mTORC2, we determined the involvement of mTORC2 in AZ1 expression. PP242, an inhibitor of mTOR having a more pronounced effect on mTORC2, inhibited AKT-Ser473, a downstream target of mTORC2, ribosomal protein S6 phosphorylation, and decreased AZ1 expression (Fig. 10A). Since putrescine induces AZ1 translation by ribosomal frameshifting, we determined the effect of rapamycin and PP242 on putrescine induced AZ1 expression. Putrescine increased AKT-Ser473 phosphorylation and AZ1 levels, which were not altered by rapamycin, suggesting the activation of mTORC2 but not mTORC1 (Fig. 10B). However, PP242 decreased both the basal-and putrescineinduced AZ1 levels (Fig. 10B). Furthermore, PP242 completely inhibited AKT-Ser473 and 4EBP1 phosphorylation with concomitant accumulation of total 4EBP1 protein. These findings indicate a role for mTORC2 in the regulation of AZ1 translation. Thus, ASN appears to inhibit AZ1 by decreasing its frameshift-induced translation via mTORC2. ASN decreased putrescine induced AZ1 in a time dependent fashion correlating with a decrease in AKT-Ser473 and ribosomal protein S6 phosphorylation (Fig. 10C). These results suggest that mTORC2 regulates AZ1 synthesis, and that ASN decreases mTORC2 activity in the absence of amino acids.

Discussion
Polyamines are essential for normal growth and differentiation, but they also play an important role in the development of cancer. The overexpression of ODC in NIH3T3 cells causes neoplastic transformation (47). Furthermore, sustained activation of ODC is associated with neoplastic transformation caused by oncogenic Ras (48), Src (49), activated RhoA (50), and forced expression of eIF4E (51). Antizymes are crucial, negative modulators of intracellular polyamine levels. Antizyme-1 interacts with ODC leading to ubiquitin-independent degradation by the 26S proteosome. A single molecule of AZ-1 can cause the degradation of several ODC molecules. Furthermore, AZ-1 also regulates the entry and export of polyamines across the plasma membrane, and thereby, maintains intracellular polyamine levels. Several studies have shown that overexpression of AZ-1 prevents tumor growth (52), inhibits cell proliferation and causes cell cycle arrest analogous to polyamine depletion by DFMO, a suicide inhibitor of ODC activity. An ODC like protein, called AZin, increases the activity of ODC by binding to AZ. Serum and phorbol ester increase AZin expression earlier than ODC activity and promote growth and tumor formation respectively (53). Thus, AZ and AZin play crucial roles in maintaining the homeostatic balance of polyamines by regulating ODC activity and, thereby the normal growth of cells.
Since A & N type amino acids are required for the induction of ODC activity by growth factors (35), we postulated that ASN might affect ODC activity by regulating the expression of antizyme or antizyme-inhibitor proteins. One of the most interesting findings in this study was the observation that amino acid starvation increased the expression of AZ1 in IEC-6 cells (Fig. 1). ODC induction correlated with a decrease in AZ1 in cells incubated with ASN or ASN+EGF. However, the levels of AZin remained unchanged in response to ASN, EGF, or ASN/EGF compared to EBSS indicating that ASN induces ODC activity by regulating AZ1 (Fig. 1A). As mentioned earlier, AZ targets ODC protein for degradation, thus, constant ODC protein levels indicate that AZ1 may not control ODC protein degradation. However, the degradation of ODC by AZ1 can't be ruled out, since the rates of degradation of ODC may be similar among the treatment groups. AZ1 accumulation over 6h incubation in EBSS did not result in the degradation of ODC protein (Fig. 2C and D). Since AZ1 shares 50% homology with AZ2, we speculated that AZ1 specific antibody might recognize AZ2. However, transfection of cells by AZ1 or AZ2 SiRNA clearly demonstrated that AZ1, induced in response to amino acid starvation, is a product of AZ1 mRNA and excluded the involvement of AZ2. Furthermore, anti-AZ1 antibody immunoprecipitated AZ1 from the extracts of cells exposed for 4h to EBSS or EBSS plus putrescine (Fig. 3D). These results suggest that AZ1 inhibits ODC activity by binding to ODC and preventing the formation of active ODC dimers. A time dependent increase in AZ1 in response to amino acid starvation (EBSS) was prevented by ASN leading to elevated ODC activity, indicating a role for ASN in either preventing the expression or increasing the turnover of AZ1 (Figs. 2 and 4).
In an extended time course study, a sharp decline in ASN-induced ODC activity after 4h correlated with low but detectable levels of AZ1 (Fig. 2D, lanes 8, 9 and 2E). A decrease in intracellular ASN due to metabolic conversion might be responsible for the increase in AZ1. The activity of the system A amino acid transporter increases in response to prolonged amino acid starvation. This mechanism, called adaptive regulation was described in mesenchymal cells of avian origin (54), but it has also been found in many other models (55). A & N type amino acids induce ODC activity in various cell types (32)(33)(34), and our results show that amino acid starvation induces AZ1 in both IEC-6 and NIH3T3 cells. Since ASN prevents AZ1 induction (Fig. 1F), this suggests a common mechanism regulating AZ1 expression by amino acids in mammalian cells. The undetectable levels of AZ1 in the presence of ASN at earlier time points (30 min), and the onset of ODC induction after 60 minutes ( Fig. 2A and B), suggest that an undetectable amount of AZ1 protein might be sufficient for the inhibition of ODC activity. Pre-incubation of cells in EBSS for 30 and 60 minutes significantly decreased ASN-induced ODC activity (Fig.  3A), and in vitro inhibition of ASN-induced ODC activity by extracts of cells incubated for 3h in EBSS (Fig. 3B), support the notion that low levels of AZ1 can inhibit ODC activity. This was further confirmed by the inhibition of ODC activity in vitro by immunoprecipitated AZ1 from extracts of cells incubated in EBSS or EBSS plus putrescine (Fig. 3D). The effects of putrescine on AZ1 expression in EBSS support previous reports showing that putrescine stimulates AZ1 expression by a +1 frameshifting mechanism (41)(42)(43). However, it is unclear whether ASN prevents the synthesis or the accumulation of AZ1 protein. The levels of AZ1 decrease after 6h incubation in the absence of amino acids (EBSS) (Fig. 4C and D). The decline in AZ1 may be due to the lack of amino acids in the EBSS resulting in a decrease in the rate of protein synthesis. ASN addition along with EBSS or its addition after 3h incubation in EBSS completely eliminated AZ1 levels within 1h. However, ASN addition after 6h in EBSS required 1.5 h to significantly decrease AZ1 protein. These results clearly show that the effectiveness of ASN in decreasing AZ1 depends on the level of the protein.
Earlier reports showed that A&N type amino acids including ASN, GLN, and AIB (a nonmetabolizable analogue of ASN) induced ODC activity in EBSS while LYS, VAL, and other amino acids failed to do so (35). Thus, amino acids unable to activate ODC were considered non-inducer amino acids. Our results confirmed that ODC inducing amino acids (GLN and AIB) prevent AZ1 accumulation similar to ASN (Fig. 7A), and that amino acids unable to induce ODC activity (LYS, VAL, and ORN) fail to decrease AZ1 protein (Figs. 7 and 8). Because, AZ1 increases in EBSS in a time dependent manner, it is also possible that LYS, VAL, and ORN may not modulate AZ1 expression but may stabilize AZ1 or its interaction with ODC or AZin. However, a time course study clearly showed that both LYS and VAL increased AZ1 compared to EBSS (Fig. 7B). Furthermore, the increases in AZ1 began earlier in LYS and VAL compared to EBSS, indicating an active role for these amino acids in suppressing ODC activity. These observations are supported by the fact that ASN-induced ODC activity significantly decreases in the presence of either LYS or VAL (Fig. 8A). Although, VAL induces AZ1 to a lesser extent compared to LYS, its suppression of ASNinduced ODC activity is similar to that of LYS. It has been shown that VAL decreases intracellular ornithine (ORN), thereby, depleting ODC substrate (56). The decrease in ASN-induced ODC activity in the presence of VAL (Fig. 8) could be due to the depletion of ORN. In that case, the addition of ORN should reverse the VAL suppression of ASN-induced ODC activity. Surprisingly, addition of ORN along with VAL significantly inhibits ASN-induced ODC activity, and ORN itself significantly decreases ASN-induced ODC activity by increasing the levels of AZ1 similar to LYS. Furthermore, VAL and ORN had an additive inhibitory effect on ODC activity. Although, ASN significantly decreased AZ1 induced by these amino acids, ODC induction was significantly lower, affirming that almost complete elimination of AZ1 is required for the full induction of ODC activity. Our results confirm that the amino acids ASN and GLN repress, and LYS, VAL, and ORN induce AZ1 expression and show that these amino acids play an important role in the regulation of ODC activity. These results also suggest that the amino acid composition of a culture medium is an important factor in the regulation of polyamine levels during the normal growth of cells. ASN (10 mM) induces ODC activity in amino acid containing medium (DMEM) and in DMEM without amino acids. However, ASN induced significantly higher levels of ODC activity in DMEM alone compared to DMEM containing other amino acids (Fig. 8  D). The decrease in ASN-induced ODC activity in amino acid containing medium can be attributed to the presence of AZ1 inducing amino acids like VAL and LYS. Furthermore, comparison of AZ1 expression in DMEM and EBSS showed that amino acid starvation specifically induces AZ1. In addition, putrescine increased AZ1 several fold over that observed in EBSS (Fig. 3C).
Several features of antizyme induction suggest regulation at the level of translation. Antizyme synthesis is inhibited by cycloheximide but not by actinomycin D (16,57). While only a minute amount of antizyme is present in mammalian tissues (about 1-2 ppm of total soluble protein), its mRNA level is relatively high and not further elevated by polyamines (23). Addition of ASN along with EBSS or following the accumulation of AZ1 (prolonged incubation in EBSS) significantly decreased AZ1, indicating that ASN prevented the translation of or augmented the degradation of AZ1 and, thus, allowed the efficient homodimerization of ODC monomers. If ASN modulates AZ1 degradation, inhibition of protein synthesis should have an additive effect leading to rapid elimination of AZ1 in the presence of ASN. Inhibition of protein synthesis by CHX alone or in combination with ASN decreased AZ1 levels at the same rate as ASN alone, suggesting that both are acting via the same mechanism ( Fig. 5A and B). Incorporation of 35 S-methionine/cysteine confirmed that the observed increase in AZ1 in EBSS is indeed due to synthesis of AZ1. ASN caused a generalized decrease in ongoing protein synthesis including AZ1 (Fig. 6). While ASN significantly decreased AZ1 mRNA (Fig. 5C), the failure of actinomycin-D (ACT-D) to prevent the accumulation of AZ1 indicates that translation of preexisting AZ1 mRNA was sufficient to sustain the level of AZ1 (Fig. 5  D). Thus, it is evident from the above results that ASN decreases AZ1 by primarily regulating its translation.
Several studies have demonstrated that limiting amino acids, inhibition of protein synthesis, and serum starvation block cell cycle progression (58). However, more recent studies indicate that amino acids can act as signaling molecules, suggesting that amino acids control the cell cycle by modulating the expression of specific genes independently of effects on global protein synthesis. Induction of p21 and p27 expression by amino acid deprivation in HepG2 human hepatoma cells involved mRNA stabilization (59). Nakajo et al., 2005 showed that GLN was required for the growth of rat intestinal epithelial cells and that it inhibited mTOR activity induced in response to arginine and proposed a differential role for individual amino acids rather than more general effects on gut homeostasis (38).
Generally, amino acid limitation results in dephosphorylation of p70-S6 kinase (p70S6K) and eukaryotic initiation factor 4E-binding protein (4EBP), whereas addition of amino acids leads to rapid TORC1/2-dependent phosphorylation of these molecules (60,61). eIF4E plays a key role in the regulation of mRNA translation and cellular physiology. Overexpression of eIF4E can lead to cell transformation, unregulated cell growth, and enhanced translation of mRNA containing 5'untranslated regions (62). Binding of 4EBP1 to eIF4E prevents the interaction with other partners and blocks cap-dependent mRNA translation initiation. 4EBP1 undergoes multiple phosphorylations, which disrupt its ability to bind eIF4E, leading to the release of eIF4E, which then binds eIF4G and initiates protein translation. The phosphorylation of 4E-BP1 is strongly modulated by the phosphoinositide 3-kinase (PI3K)/Akt pathway and by nutrients/amino acids. In addition to mTOR, other factors that have been implicated in the regulation of 4E-BP1 phosphorylation/activation states include the kinases pim-2 (66), PKCδ, c-Abl (67) and phosphatases such as serine/threonine protein phosphatase 2A (PP2A) (68)(69)(70). Thus, a complex system of signaling molecules, comprised of scaffold proteins, kinases, and phosphatases, appears to control the phosphorylation/activity of 4E-BP1. Eukaryotic initiation factor 4E (eIF4E) binds the 5'-cap structure of mRNA, forms complexes with other initiation factors and facilitates the binding of the 40S subunit to mRNA and the identification of the AUG start codon. Thus, amino acid deprivation should mimic the effect of rapamycin, a specific inhibitor of mTOR.
The absence of all the constituent amino acids in DMEM decreased mTOR signaling as evident by the decreased phosphorylation of p70S6K and it's down stream substrate, ribosomal protein S6 and increased binding of 4EBP1 to the mRNA cap (Fig. 9). It is evident from the mTOR activity in DMEM (AA+) and DMEM (AA-) that the activation of mTOR signaling requires the presence of other amino acids besides asparagine. However, AZ1 synthesis is inhibited in DMEM (AA+). Thus, AZ1 synthesis occurs on a background of low mTOR activity. Short-term inhibition (4h) of mTOR activity by rapamycin failed to induce AZ1 synthesis in DMEM (AA+), and rapamycin had no effect on AZ1 induced in response to amino acid starvation (AA-). However, prolonged inhibition of mTOR by rapamycin treatment for 24 h induced AZ1 expression in DMEM (AA+) thus, mimicking the absence of amino acids (Fig.  9G). The addition of single amino acids (Asn, Leu or Met) to DMEM (AA-) did not restore the phosphorylation of ribosomal protein S6 to levels observed in DMEM (AA+), suggesting that a full complement of amino acids is necessary to activate mTOR.
These results clearly established that the decreased mTOR activity in response to amino acid starvation inhibits cap-dependent mRNA translation. However, increased AZ1 synthesis under these conditions suggests a cap-independent mRNA translation. The 5' pseudoknot structure in the AZ1 mRNA, and a distinct +1 ribosomal frameshifting mechanism are involved in decoding AZ1 mRNA further suggesting the involvement of cap-independent translation. The fact that polyamines induce the frameshifting (17,18) suggests that AZ1 synthesis might be regulated by an additional translational control (Fig. 10). Spermine, for example, has been shown to regulate synthesis of mitochondrial proteins by stabilizing a small subunit of the ribosome and fMet-tRNA complex (71). Furthermore, recently Choo and Blenis, 2009 (72) reported that rapamycin inhibited S6K in various cell lines. However, prolonging the treatment resulted in the hyper phosphorylation of 4EBP1. This led to rapamycin-resistant capdependent translation. These results suggest that additional mechanisms regulating the phosphorylation state of 4EBP1 might be involved. We have also observed 4EBP1 phosphorylation with a low mTOR background (AA-, Fig. 9).
The mTOR pathway forms two functionally distinct protein complexes, mTORC1 and mTORC2. Decreased mTORC1 activity in the absence of amino acids might lead to an increase in mTORC2 activity in order to sustain translation of stress response proteins. AZ1 is induced by stress (19) and in addition to regulating ODC activity, is involved in the regulation of cyclin D activity (73) and polyamine transport (23,24,44), suggesting involvement of the mTORC2 pathway in the regulation. Since rapamycin does not inhibit AKT-Ser473 phosphorylation, a downstream target of mTORC2, it allowed us to differentiate between the contributions of these complexes in the regulation of AZ1. Inhibition of mTORC2 by PP242, as judged by the decrease in AKT-Ser473 phosphorylation, decreased the levels of AZ1 (Fig. 10). Although, polyamines regulate AZ1 translation by ribosomal frameshifting, AZ1 translation induced by putrescine also increased AKT-Ser473 phosphorylation, suggesting the involvement of mTORC2 in AZ1 translation (Fig. 10). ASN decreased putrescine-induced AKT-Ser473 phosphorylation with a concomitant decrease in AZ1 expression. PP242 completely inhibited 4EBP1 phosphorylation and caused the accumulation of 4EBP1. Analogous to PP242, ASN also caused the accumulation of 4EBP1 and decreased AZ1. These results suggest that ASN decreased transcription of AZ1 mRNA and prevented its translation by directly inhibiting mTORC2 activity, and/or the ribosomal frameshifting.
In the presence of ASN, or a stimulating amino acid, the normal signaling pathways initiated by EGF result in the potentiation of ODC activity compared to that elicited by ASN acting alone. Several complex 5'-UTR mRNAs involved in cell growth, cell cycle progression, and angiogenesis are selectively translated via eIF4E including ODC (74,75). The amount of eIF4E is much lower compared to other translation factors and is activated by mitogenic stimuli (75). The translation of mRNA containing complex 5'-UTR (pseudoknot structure), IRES, and 5'-TOP also occurs independent of classical 5'-capdependent mechanisms. In confluent serum starved IEC-6 cells incubated in EBSS, addition of ASN or EGF failed to stimulate ODC synthesis (35) suggesting that eIF4E may be necessary but not sufficient to support the translation of ODC mRNA. Although, eIF4E stimulates translation of multiple mRNAs, the availability of translatable mRNA in confluent serum starved cells may be a limiting factor. In fact, decreased AZ1 mRNA in the presence of ASN resulted in the decreased availability of the translatable pool of AZ1 mRNA (Fig.  5C). This decreased mRNA levels could possibly contribute to decreased AZ1 synthesis. Figure 11A shows that the AZ1 is decoded from two open reading frames (ORFs). Translation is initiated at a start codon for ORF1 and subsequent +1 translational frameshifting at the end of ORF1 (stop codon, UGA) is necessary for the synthesis of functional AZ1. Putrescine increases the efficiency of the obligatory +1 frameshifting (17,18). Amino acid deprivation induced AZ1 synthesis. ASN prevented AZ1 synthesis in response to amino acid deprivation and putrescine. Our previous studies and present results suggest a scheme (Fig. 11B) showing that the decreased mTOR activity in the absence of amino acids inhibits cap-dependent translation of mRNA. However, translation of AZ1 mRNA containing complex 5'-UTR (pseudo knot structure) occurs in a capindependent fashion. Furthermore, polyamine-induced mTORC2 activity stimulated AZ1 translation. Analogous to PP242, ASN inhibits mTORC2 and prevents AZ1 mRNA translation. Decreased AZ1 expression allows the homodimerization of ODC monomers leading to the formation of active ODC.     Confluent serum starved cells were exposed to EBSS for indicated time period in the presence and absence of actinomycin-D (ACT-D) and cell extracts were analyzed by western blot to determine the levels of AZ1. A Representative western blot from three observations is shown.
Membranes stained with coomassie brilliant blue to ensure equal loading are shown (Loading). FIGURE 6. Confluent serum starved cells were exposed to EBSS in the presence and absence of ASN for the indicated time in the presence of 10µCi/ml 35 S-methionine/cysteine. Cell extracts were prepared as described in the methods. Equal amounts of protein were precipitated by TCA, and 10 µg protein from each sample was used to quantify incorporation of 35 S-Methionine/Cystine by liquid scintillation spectroscopy (A). Equal amounts of protein (30 µg) were resolved on 12% SDS PAGE and analyzed by western blotting for AZ1 and actin (B). Equal amounts (900 µg) of protein from the above samples were subjected to immunoprecipitation using AZ1 antibody and resolved on SDS-PAGE followed by transfer to PVDF membrane. Then membrane was dried, sprayed with the enhance reagent and exposed to X-ray film as described in the methods (C). Values are means ± SE; n = 3 cultures. *, & #, p < 0.05 compared with corresponding control values. Representative western blots from three observations are shown.  Representative western blots from three observations are shown. FIGURE 9. (A) Confluent serum starved cells were exposed to DMEM without amino acids (AA-) in the presence or absence of ASN and DMEM with amino acids for 4h. Protein (500 µg) from cell extracts was subjected to m7-GTP pull down assay and analyzed by western blot to determine the levels of p-eIF4E, p4EBP1 and total 4EBP1. Protein (15 µg) from whole cell extract were also loaded on the gel. Representative western blots from three observations are shown. (B) Confluent serum starved cells were exposed to DMEM without amino acids (AA-) in the presence or absence of ASN for the indicated time, and total and phosphorylated 4EBP1 were determined by western blot analysis. (C) Confluent serum starved cells were exposed to EBSS in the presence and absence of ASN for 4h, and cell extracts were analyzed by western blot to determine phospho-and total-mTOR, 4EBP1 and AZ1; DMEM served as a positive control (AA+). mTOR immunoprecipitated from 500 µg protein was probed with pSer2448-mTOR antibody. (D) Confluent serum starved cells were exposed to DMEM containing amino acids AA+) or without amino acids (AA-) in the presence or absence of rapamycin (20 nM) for 4h, and cell extracts were analyzed by western blot to determine the levels of phospho-and total-mTOR, phospho-p70S6K, phospho-ribosomal protein S6, AZ1, and Actin. (E) Confluent serum starved cells were exposed to EBSS in the presence and absence of ASN, leucine (Leu), and methionine (Met) for 3h and cell extracts were analyzed by western blot to determine the levels of phosphoribosomal protein S6 and Actin. DMEM served as a positive control (AA+). (F) Confluent serum starved cells were exposed to DMEM without amino acids (AA-) in the presence or absence of ASN followed by addition of rapamycin (20 nM) to appropriate groups for 4h, and cell extracts were analyzed by western blot to determine the levels of AZ1 and Actin. Representative western blots from three observations are shown. (G) Confluent cells were exposed to 20 nM rapamycin (Rapa) during serum starvation (24 h) followed by incubation in DMEM (AA+) in the presence or absence of rapamycin or DMEM (AA-) for 4h, and cell extracts were analyzed by western blot to determine the levels of AZ1 and Actin. Representative western blots from three observations are shown.