Activation of proline biosynthesis is critical to maintain glutamate homeostasis during acute methamphetamine exposure

Methamphetamine (METH) is a highly addictive psychostimulant that causes long-lasting effects in the brain and increases the risk of developing neurodegenerative diseases. The cellular and molecular effects of METH in the brain are functionally linked to alterations in glutamate levels. Despite the well-documented effects of METH on glutamate neurotransmission, the underlying mechanism by which METH alters glutamate levels is not clearly understood. In this study, we report an essential role of proline biosynthesis in maintaining METH-induced glutamate homeostasis. We observed that acute METH exposure resulted in the induction of proline biosynthetic enzymes in both undifferentiated and differentiated neuronal cells. Proline level was also increased in these cells after METH exposure. Surprisingly, METH treatment did not increase glutamate levels nor caused neuronal excitotoxicity. However, METH exposure resulted in a significant upregulation of pyrroline-5-carboxylate synthase (P5CS), the key enzyme that catalyzes synthesis of proline from glutamate. Interestingly, depletion of P5CS by CRISPR/Cas9 resulted in a significant increase in glutamate levels upon METH exposure. METH exposure also increased glutamate levels in P5CS-deficient proline-auxotropic cells. Conversely, restoration of P5CS expression in P5CS-deficient cells abrogated the effect of METH on glutamate levels. Consistent with these findings, P5CS expression was significantly enhanced in the cortical brain region of mice administered with METH and in the slices of cortical brain tissues treated with METH. Collectively, these results uncover a key role of P5CS for the molecular effects of METH and highlight that excess glutamate can be sequestered for proline biosynthesis as a protective mechanism to maintain glutamate homeostasis during drug exposure.


Results
Acute METH exposure activates the proline synthetic pathway. METH exposure affects glutamate neurotransmission primarily by altering glutamate levels 12,13 . While, glutamine serves as the major source of glutamate [21][22][23] , proline metabolism is also closely linked to glutamate 25 (Fig. 1A). However, the functional link between proline and glutamate during drug exposure induced glutamate neurotransmission is poorly understood. Therefore, we examined the effects of METH exposure on proline metabolism using the SH-SY5Y neuronal cell model. First, we measured the levels of enzymes of the proline catabolic arm-POX and P5CDH-which sequentially convert proline to P5C and then to glutamate (Fig. 1A). We treated SH-SY5Y cells with METH in a dose dependent manner using concentrations of 1 mM and lower. Although higher concentrations of METH have been used in published studies 45,46 , we chose concentrations below 1 mM since higher levels of the drug are rarely achieved in METH associated disorders 47,48 . Lysates of METH-treated and control cells were analyzed Scientific Reports | (2021) 11:1422 | https://doi.org/10.1038/s41598-020-80917-7 www.nature.com/scientificreports/ by immunoblot to measure the levels of POX and P5CDH. These analysis revealed that SH-SY5Y cells endogenously express all the key metabolic enzymes of proline (Fig. 1B,D). Interestingly, METH exposure minimally altered the expression of POX and P5CDH when compared to the untreated controls ( Fig. 1B-D). Even at 1 mM concentration of METH, expression of these two enzymes were not significantly altered ( Fig. 1B-D), suggesting that acute METH exposure minimally alters the proline catabolic arm that converts proline to glutamate (Fig. 1A). We next measured the expression of enzymes of the proline biosynthetic arm (Fig. 1A), P5C synthase (P5CS) and P5C reductase (PYCR) that convert glutamate to proline by sequential catalysis 26 . Notably, the expression of P5CS was significantly increased in cells exposed to METH (Fig. 1E,F). Densitometry analysis illustrated a dosedependent increase in P5CS expression in cells treated with METH relative to untreated cells (Fig. 1F). Then, we measured the levels of PYCRs-PYCR1, and PYCR2, which are involved in biosynthesis of proline ( Fig. 1A) 27,29,49 . As presented in Fig. 1E,G,H, expression of the PYCR1 and PYCR2 was significantly increased in the METH treated cells. While the expression of PYCR1 was increased ~ 1.7-fold, the levels of PYCR2 was ~ sixfold higher in cells treated with 1 mM METH compared to the untreated control cells (Fig. 1E,G,H). Collectively, these results demonstrated that acute METH exposure activates the proline synthetic arm without altering the catabolic arm.
METH exposure upregulates key proline metabolic enzymes in differentiated neuronal cells and cortical brain slices. Our results in Fig. 1 showed upregulation of proline synthetic enzymes by METH exposure in undifferentiated SH-SY5Y cells. Even though the SH-SY5Y neuroblastoma cells show some characteristics of neurons, we have previously shown that these cells acquire a neuronal phenotype by differentiation utilizing ATRA 50 . Therefore, we used differentiated SH-SY5Y cells and treated them with METH in a dose dependent manner. 24 h post treatment cellular extracts were prepared to measure the expression of P5CS and Schematic depicting the metabolic link between glutamate and proline. Synthesis of proline begins with the conversion of glutamine to glutamate by the enzyme glutaminase (GLS). Glutamate is then converted by P5C synthase (P5CS) to glutamic-γ-semialdehyde that spontaneously cycles to Δ 1 -pyrroline-5-carboxylate (P5C). P5C is subsequently catalyzed to l-proline by P5C-reductase (PYCR). Conversely, catabolism of proline to glutamate is catalyzed in two consecutive steps by proline oxidase (POX) and P5C dehydrogenase (P5CDH). Glutamate is finally converted to glutamine by glutamine synthase (GS). (B-D) Effect of acute METH treatment on proline catabolic enzymes. SH-SY5Y cells were treated with increasing concentrations of METH for 24 h. Post-treatment cells were harvested and lysed. Cell lysates were subjected to immunoblot analyses. (B) Representative immunoblot of enzymes involved in catabolism of proline to glutamate (n = 3) (Fig. S1). (C,D) Densitometry analysis of POX and P5CDH, respectively normalized to β-actin. (E-H) Effect of acute METH treatment on proline metabolic enzymes. (E) Representative immunoblot of enzymes, P5CS, PYCR1 and PYCR2, which are essential for proline biosynthesis from glutamate (n = 3). (F-H) Densitometry analysis of P5CS, PYCR1, and PYCR2, respectively normalized to β-actin. Data presented in panels (C,D,F-H) are mean values of three independent experiments with error bars representing SEM. *p < 0.05, **p < 0.005 represent statistical comparison of untreated vs METH-treated cells. www.nature.com/scientificreports/ PYCR2, as representatives of proline synthetic enzymes. Immunoblot analysis showed that METH treatment resulted in a significant upregulation of both P5CS and PYCR2 when compared to untreated controls ( Fig. 2A-C). These observations are similar to the results obtained with the undifferentiated cells ( Fig. 1). To further probe the physiological relevance of these cell-based observations, we carried out METH exposure experiments using slices of cortical regions of rodent brain. Given the technical challenges associated with brain slice experiments, 200 µM METH was selected based on the significant upregulation of both P5CS and PYCR2 in neuronal cells at this concentration (Figs. 1E-H, 2A-C). Brain slices were incubated in METH or vehicle (saline) for a 6-h period. Then cortical enriched tissue was isolated and immunoblot analysis was carried out of the tissue lysates. Data from these studies illustrated that P5CS and PYCR2 levels were significantly upregulated in the cortical brain slices upon METH exposure ( Fig. 2D-F). Taken together, results in Figs. 1 and 2A-F establish that acute METH exposure upregulates the proline metabolic enzymes in neuronal cells and cortical brain slices. Next, we determined whether activation of proline synthetic arm is specific to METH exposure or a general mechanism utilized by other ALS. To test this, we first treated undifferentiated SHSY-5Y cells with increasing concentration of Amphetamine (AMPH). Immunoblot analysis of cellular lysates showed a minimal but nonsignificant increase in P5CS and PYCR2 protein expression by AMPH exposure when compared to untreated controls ( Fig. 2G-I). Interestingly, AMPH exposure of differentiated SH-SY5Y cells illustrated a modest increase in the expression of P5CS at 1000 µM METH. However, PYCR2 expression was minimally changed in the differentiated cells exposed to increasing concentrations of AMPH ( Fig. 2J-L). Finally, exposure of brain slices to AMPH (200 µM) showed a lack of significant increase in the expression of both P5CS and PYCR2 (Fig. 2L-O). Collectively, results described in Fig. 2 strongly suggested that acute METH exposure activates the proline catabolic arm, whereas AMPH exposure lacks such activation.
Acute METH exposure enhances proline levels but does not increase glutamate levels. Results in Figs. 1 and 2 demonstrated that METH exposure upregulated the expression of P5CS and PYCR2. Since these enzymes catalyze the synthesis of proline from glutamate, we measured intracellular proline levels in METH treated SH-SY5Y cells. We chose 200, 500 and 1000 μM of METH for this experiment since both P5CS and PYCR2 were significantly upregulated at these drug concentrations (Figs. 1, 2). To quantify proline levels, we utilized the acid-ninhydrin assay that specifically detects proline 51,52 . Results from this assay illustrate that the levels of intracellular proline were significantly elevated in METH treated cells when compared to the untreated control cells (Fig. 3A). Specifically, treatment of cells with 200 μM METH resulted in a ~ 100 μM increase in proline levels relative to the untreated cells. Similarly, treatment with 500 μM METH further increased the proline levels up to ~ 130 μM and 1000 μM METH enhanced proline levels up to ~ 155 μM. These results indicate that acute METH exposure significantly enhances proline biosynthesis in accordance with the induction of proline synthetic enzymes.
A number of studies have shown that METH exposure alters glutamate levels [53][54][55] . Therefore, we tested whether the increased proline levels were due to higher glutamate in METH-treated cells. We measured the levels of both intracellular and extracellular glutamate in SH-SY5Y cells. We treated these cells with 50-1000 µM METH for 24 h. Following that, we measured glutamate levels in the cellular lysates and supernatants by a colorimetric assay. As shown in Fig. 3B,C, treatment of SH-SY5Y cells with METH at concentrations of 0-1000 µM did not significantly increase either intracellular (Fig. 3B) or extracellular (Fig. 3C) glutamate levels. Surprisingly, there was a slight decrease in both intracellular and extracellular glutamate levels (~ 20-30%) in cells treated with METH at the concentrations 200-1000 µM. Collectively these results suggest that acute METH treatment at concentrations ≤ 1 mM minimally affects intracellular and extracellular glutamate levels.
A lack of increased glutamate in METH treated cells could be due to reduced glutamate synthesis from glutamine-the major source of glutamate in neurons (19,20). To test this hypothesis, we measured the levels of glutaminase (GLS) enzyme that converts glutamine to glutamate in neurons (22)(23)(24). Western blot analysis revealed that METH treatment did not reduce GLS levels in SH-SY5Y cells, (Fig. 3D,E). Interestingly, a modest increase in GLS expression was observed in METH treated cells (Fig. 2D,E). These observations suggested that  www.nature.com/scientificreports/ glutamate synthesis is not reduced in METH treated cells. Thus, a lack of increased glutamate during acute METH exposure was most likely not due to reduced synthesis from glutamine. Extracellular glutamate levels are also dependent on packaging of intracellular glutamate by the vesicular glutamate transporters: vGLUT1, vGLUT2, and vGLUT3 (22)(23)(24). Among these transporters, the functional characteristics of vGLUT1 and vGLUT2 for glutamate transport is similar 56 . Interestingly, vGLUT3 has been reported to be found in cholinergic neurons but not in glutamatergic cell populations 57 . Therefore, we measured the levels of vGLUT1 as the selected marker of glutamate release in METH treated cells. Surprisingly, western blot analysis revealed a dose-dependent decrease in vGLUT1 expression with increasing concentration of METH (Fig. 3D,F). The reduction in vGLUT1 expression suggested that the packaging of glutamate was most likely reduced in METH treated cells. Interestingly, the reduction in vGLUT1 expression is correlated with the decrease in extracellular glutamate levels in cells treated with METH at concentrations of 200-1000 µM (Fig. 3C). Finally, we also measured neuronal cytotoxicity by LDH assay in the supernatants of METH-treated cells. METH treatment did not cause any cytotoxicity even at 1 mM concentration when compared to untreated control cells (Fig. 3G). Collectively, these results demonstrate that acute METH exposure at concentrations below 1 mM significantly enhances proline levels but minimally increases extracellular glutamate and lack the ability to cause neuronal cytotoxicity.
P5CS is the key enzyme that regulates glutamate levels during acute METH exposure. Our results in Fig. 3A demonstrated that acute METH exposure activates proline biosynthesis. However, a lack of alterations in glutamate levels in these cells (Fig. 3B,C), led us to hypothesize that activation of proline biosynthesis may aid in regulating glutamate levels during acute METH exposure. To test this, we created SH-SY5Y cells www.nature.com/scientificreports/ that are deficient in proline biosynthesis from glutamate. We specifically depleted the P5CS enzyme since synthesis of proline from glutamate is regulated by P5CS (Fig. 1A) 26 and METH significantly increased P5CS levels (Figs. 1, 2). Even though PYCR2 was also upregulated by METH, blocking the reductase step required depletion of both PYCR1 and PYCR2 due to the functional redundancy of these two enzymes. Therefore, we generated P5CS knockout SH-SY5Y cells (P5CS-KO) using CRISPR/Cas9 genome-editing approach 58 . We employed a lentiviral vector-based expression of nuclease-active Cas9 and P5CS gRNAs and selected single clones of P5CS-KO. Western blot analysis confirmed lack of P5CS expression in the P5CS-KO cells when compared to the wild type cells (Fig. 4A). We then treated the P5CS-KO cells with METH in a dose-dependent manner for 24 h. Measurement of glutamate levels showed a significant increase in intracellular glutamate levels in the P5CS-KO cells upon METH treatment (Fig. 4B). For instance, METH exposure at 50 µM resulted in ~ 1.8-fold increase in intracellular glutamate levels in P5CS-KO cells compared to the untreated control. This increase was further enhanced to ~ 2.6-fold with 1 mM METH treatment P5CS-KO cells (Fig. 4B). Interestingly, acute METH exposure also significantly elevated extracellular glutamate levels in the P5CS-KO cells (Fig. 4C). These results are in contrast to the minimal effect of METH exposure on glutamate levels in the wild type cells that express P5CS (Fig. 3B,C). www.nature.com/scientificreports/ To further solidify the role of P5CS in regulating glutamate levels, we probed the expression of PYCR2 in the P5CS-KO cells, since METH treatment increased PYCR2 expression in wild type cells (Figs. 1, 2). Results presented in Fig. 4D,E clearly show that METH exposure minimally alters the expression of PYCR2 in the P5CS-KO cells. Moreover, in P5CS-KO cells treatment of METH minimally affected GLS expression (Fig. 4F,G), suggesting that depletion of P5CS does not affect glutamate synthesis from glutamine. Interestingly, a marginal increase in vGLUT1 expression was observed in the P5CS-KO cells upon METH treatment above 500 µM (Fig. 4H). This increase is consistent with the increased glutamate levels observed in the P5CS-KO cells after METH exposure. Collectively, these results establish that depletion of P5CS causes accumulation of glutamate and strongly suggest the involvement of proline biosynthetic arm in maintaining METH-induced glutamate homeostasis.  www.nature.com/scientificreports/ levels during acute METH exposure (Fig. 4). To further probe a functional link between proline synthetic arm and glutamate, we exploited the proline auxotrophic nature of Chinese Hamster Ovary (CHO) cells. CHO cells lack the proline biosynthetic arm that is needed to catalyze proline from glutamate 59,60 . Specifically, these cells do not express P5CS and PYCR2 as measured by western blot analysis (Fig. 5A). Therefore, to further strengthen the role of P5CS during glutamate homeostasis, CHO cells were treated with increasing doses of METH for 24 h. Then extracellular glutamate levels were measured in the culture supernatants of treated and untreated cells. As shown in Fig. 5C, a dose-dependent increase in glutamate levels was detected in the presence of increasing concentrations of METH. Next, we tested whether the higher levels of glutamate in METH-treated CHO cells is a consequence of lack of proline synthesis from glutamate. We exogenously expressed P5CS in the CHO cells by transfecting a P5CS expression construct. Post-transfection, western blot analysis of the cellular lysates of these cells confirmed the expression of P5CS compared to the lack of P5CS expression in wild type cells (Fig. 5B). Then, the P5CSexpressing CHO cells were exposed to increasing concentrations of METH for 24 h and extracellular glutamate levels were measured in the supernatants. Results from these assays show that P5CS expression significantly abrogates the increase in extracellular glutamate upon METH exposure as compared to the P5CS-lacking wild type cells (Fig. 5C). Taken together, these observations provide strong evidence that P5CS regulates glutamate levels during acute METH exposure.

METH exposure increases glutamate levels in proline-auxotrophic
Acute METH exposure activates proline biosynthetic pathway in the cortical regions of mice brain. Our studies using neuronal cell model and proline auxotrophic cells (Figs. 1, 2, 3, 4, 5) demonstrated a key role of proline biosynthetic pathway in regulating glutamate levels during acute METH exposure. To determine the in vivo significance of these observations, we investigated the effects of METH exposure on the proline synthetic pathway in a rodent model. Mice were administered with either METH (2 mg/kg IP) or saline under acute conditions for 24 h. Then, the cortical regions of the brain were isolated and total protein from these brain tissues were extracted and subjected to western blot analyses. We focused on the cortical regions, since www.nature.com/scientificreports/ glutamatergic neuronal projections originating in the prefrontal cortex (PFC) extend to the striatum, nucleus accumbens (NAc), ventral tegmental area (VTA) and substantia nigra of the midbrain 61,62 and METH has been shown to affect glutamate in the PFC of mice brain 63 . Data in Fig. 6A show that P5CS is constitutively expressed in the cortical region of mouse brain. Upon acute METH treatment of these mice, a significant increase in P5CS levels was detected when compared to the control animals, (Fig. 6A,B), suggesting upregulation of P5CS during METH exposure. Furthermore, protein levels of the PYCR2 were also increased in the cortical tissues of METH administered animals (Fig. 6C,D). However, the level of increase in PYCR2 was lower when compared to the significant increase in P5CS levels. These data are in accordance with the results obtained with the neuronal cell model (Figs. 1, 2). Collectively, these observations support the activation of the proline synthetic arm in the cortical region of mice brain during acute METH exposure.

Discussion
Glutamate is the key excitatory neurotransmitter in the brain [15][16][17] . During glutamatergic neurotransmission, glutamate is packaged in presynaptic neurons and released into the synapse [15][16][17] . The released glutamate binds to ionotropic and metabotropic receptors on postsynaptic neurons to propagate the incoming signals [15][16][17] . Subsequently, high-affinity transporters present on the glial cells and neurons rapidly sequester the glutamate to facilitate a continuous cycle of neurotransmitter activity 64 . Given the central role of glutamate in neurotransmission, it is involved in most aspects of normal brain function including cognition, memory, and learning 18 . Glutamate also plays a critical role in drug reward, reinforcement, and relapse [65][66][67][68] . Importantly, recent studies suggest that in individuals with substance use disorder, glutamate is associated with synaptic plasticity [65][66][67][68] . Specifically, METH has been reported to affect glutamate neurotransmission by altering glutamate levels 14 . However, the mechanism by which METH exposure affects glutamate homeostasis is poorly understood. METH has been shown to alter the extracellular concentrations of both glutamate and dopamine in the brain 13 . For example, chronic treatment with high doses of METH has been reported to increase glutamate efflux in the striatum 14 . Depletion of striatal dopamine levels with low dose METH has been shown to enhance glutamate release 69 . Similarly, a low dose of METH has been demonstrated to increase extracellular dopamine but not glutamate in PFC and striatum 63 . Additionally, repeated METH administrations were shown to enhance cortical glutamate efflux 63 . Interestingly, a single injection of a high dose of METH (30 mg/kg, i.p.), was reported to cause glutamate depletion 70 . These observations suggest that alterations in the levels of glutamate are dependent on the duration and dose of METH exposure. Accordingly, our results with a neuronal cell model demonstrated that acute METH exposure at concentrations below 1 mM minimally altered extracellular glutamate levels. It is noteworthy that majority of the published studies examined the effects of long-term exposure and relatively high doses of METH on glutamate 47,71 . Therefore, based on our results we speculate that a single dose of METH below 1 mM concentrations causes minimal changes in glutamate homeostasis.
Neuronal glutamate levels are regulated by a multi-prong mechanism that involve glutamate biosynthesis from glutamine and synaptic vesicle loading of intracellular glutamate. Glutamate synthesis is dependent on the glutamine-glutamate cycle between astrocytes and neurons 72 . In this metabolic cycle, GLS is primarily responsible for synthesizing glutamate from glutamine to continuously replenish the neurotransmitter pool (22)(23)(24). While, GLS regulates glutamate biosynthesis from glutamine (22)(23)(24), vGLUTs mediate the transport of glutamate into synaptic vesicles 73,74 and vGLUT expression determines the amount of glutamate packaged into vesicles 73,74 . Specifically, in glutamatergic neurons the functionally similar vGLUT1 and vGLUT2 are responsible for glutamate transport 56 . Interestingly, our results demonstrated that expression of GLS remains mostly unaffected in METH treated cells (Fig. 3). Moreover, with acute METH treatment, a minimal to marginal reduction in vGLUT-1 expression was detected (Fig. 3). These observations strongly suggest that acute METH exposure minimally alters ongoing glutamate biosynthesis and vesicular packaging, consistent with the minimal effect of METH on glutamate levels under acute exposure conditions. The glutamine-glutamate biosynthetic cycle is not stoichiometric, but is an open biochemical pathway wherein glutamate interfaces with several other metabolic pathways including carbohydrate and amino acid metabolism through the TCA cycle 70 . In this context, our results describing activation of proline synthetic arm by METH exposure highlights the novel role of proline biosynthesis during METH exposure. Glutamate can be channeled for proline synthesis through P5C. These biochemical reactions are sequentially catalyzed by P5CS and PYCR (Fig. 1A). Remarkably, our data showed that acute METH treatment resulted in a marked enhancement in both P5CS and PYCR2 expression (Figs. 1, 2). Concomitant increase in intracellular proline levels (Fig. 3A) indicated that induction of these two enzymes are most likely responsible for synthesizing proline from glutamate. Moreover, this increase was observed only with METH and not AMPH indicating the effect to be specifically mediated by METH. Previously, a genome-wide study, analyzed METH-induced mRNA expression in rodent brain to reveal the potential regulatory consequences in response to METH 75 . Pathways upregulated by METH included proline metabolism especially the proline synthetic enzyme P5CS (encoded by ALDH18A1) and PYCRs 75 . Similarly, a metabolomic study of brain tissues from METH affected animals uncovered neurochemical signatures related to the metabolism of amino acids including glutamate and proline 76 . Similar to METH, morphine is also known to affect glutamatergic neurotransmission 77,78 . In a proteomics analysis, PYCR2 was identified as one of the proteins that was induced by morphine in mouse hippocampal postsynaptic densityassociated proteins (HPSD) 79 . These studies indicate an interplay between glutamate and proline metabolism. However, the functional relevance of proline metabolism during METH exposure is unknown. Therefore, our studies showing a functional association between glutamate and proline metabolism during METH exposure are highly significant.
Our results provide strong evidence of glutamate efflux to proline biosynthesis. Specifically, a marked induction in the expression of P5CS was observed during acute METH exposure not only in the SH-SY5Y neuronal www.nature.com/scientificreports/ cells but also in brain cortical region of mice. Interestingly, disruption in proline biosynthesis by knocking out of P5CS dramatically increased both intracellular and extracellular glutamate with METH treatment. The role of P5CS was further confirmed using a proline auxotrophic cell line (CHO-K1) that lacks the proline biosynthetic machinery. A marked enhancement in glutamate levels was also observed in these cells with METH treatment and this effect was abrogated by overexpression of P5CS. Collectively, these results demonstrated sequestration of neuronal glutamate pool towards proline biosynthesis. A lack of neuronal cell death by acute METH exposure even at 1 mM concentrations supports the hypothesis that sequestration of glutamate to proline is a critical mechanism activated to avert the accumulation of excess glutamate and prevent excitotoxicity. Sequestration of glutamate to proline may have significant implications in the CNS. For instance, P5C formed during proline synthesis from glutamate can serve as the source for carbon exchange between the TCA and urea cycle 28 . Proline can also be deposited in collagen, the most abundant protein by weight in the human body 80 . Approximately, 25% of proline residues make up collagen, therefore, collagen has been suggested to be both a dump as well as a reservoir for proline 80 . This may be very relevant in the context of the extracellular matrix (ECM) of neurons. For example, excess glutamate can be converted to proline and deposited into the ECM as collagen without dramatically altering extracellular proline levels. High levels of proline have been shown to have excitotoxic properties and studies have shown that addition of proline to hippocampal slices decreased glutamate uptake causing glutamate to accumulate 81 . It is plausible that once the threshold for proline deposition is reached, especially with chronic METH, it may result in the accumulation of proline, which may also contribute to the glutamate excitotoxicity. Thus, the interplay between glutamate and proline may be an important regulator in METH-induced glutamate excitotoxicity.
Importantly, we envision that increased conversion of glutamate to proline offers a bioenergetic metabolic advantage. The bioenergetic cost of maintaining homeostatic levels of synaptic glutamate is expensive 82 . Glucose via glycolysis provides the main source of energy for maintaining glutamate homeostasis in the brain. However, for efficient glycolysis a sustained supply of electron acceptor nicotinamide adenine dinucleotide (NAD + ) is essential 83,84 . Proline biosynthesis has been shown to augment glycolysis by affecting the levels of NAD + , especially in cancer metabolism [85][86][87] . Therefore, our results demonstrating METH-induced proline biosynthesis may confer a metabolic bioenergetic advantage to neurons in response to increased neuronal stimulus by METH. In summary, our results identify a key role of P5CS during METH exposure and highlight that sequestering excess glutamate for proline biosynthesis is a key mechanism to maintain glutamate levels (Fig. 7).

Materials and methods
All the methods used in this study are in accordance with relevant guidelines and regulations. , and 1000 U/mL penicillin, and 100 mg/mL streptomycin (Gibco). CHO-K1 cells were grown and maintained in Ham's F-12K medium with 10% FBS, 2 mM Glutamine, 1000 U/mL penicillin, and 100 mg/mL streptomycin. SH-SY5Y were differentiated with ATRA as per published methods (48). All cells were maintained at 37 °C and 5% CO 2 before and during treatments. The cells were treated acutely for 24 h with METH in a dose dependent manner at physiologically relevant concentrations of 50-1000 μM.
Cloning. All three P5CS-targeting sgRNA sequences used in this study were from Doench et al. 88  Lentivirus transduction. HEK293T cells (5 × 10 5 ) seeded per well in 6-well plates and cultured overnight were transfected with plasmid DNAs using the PEI transfection reagent (Polysciences Inc.). For transfecting cells in each well, a 4:3:1 ratio of lentiviral transfer plasmid (described above), packaging plasmid psPAX2, and pseudotyping envelope plasmid pMD2.G, and the PEI at a DNA:PEI ratio of 1:3 were used. After overnight culturing, the culture medium was removed and the cells were replenished with fresh medium. Forty-eight hours post transfection, the lentivirus-containing culture medium was collected, centrifuged at 500×g for 5 min at room temperature to remove cell debris, and the supernatant was filtered through 0.45 μm filter. Unused virus stocks were stored at − 80 °C for later use.

P5CS CRISPR/Cas9 knockout in SH-SY5Y cells.
For generation of stable cell lines, SH-SY5Y cells were inoculated with lentivirus for 24-48 h, then replenished with fresh culture medium containing the selection drug puromycin at 1 μg/mL. Cells were replenished with fresh culture medium containing puromycin every 2 days until majority of the uninfected control cells were eliminated (in ~ 7-10 days). To generate single cell clones, cells were treated with 1 μg/mL DOX for 24 h and the GFP positive cells were sorted out by flow cytometry. The GFP-positive single cells were then seeded into a 96-well plate and allowed to expand. The cell clones that were depleted of the target protein as analyzed by western blot were used in functional experiments.
Lactate dehydrogenase (LDH) assay. To determine cytotoxicity, LDH release was measured with the LDH Cytotoxicity Assay Kit (Pierce) according to the manufacturer's protocol. 5 × 10 5 SH-SY5Y cells were plated into 6-well plates and maintained overnight before treatment, following which the cells were treated with METH at various concentrations for 24 h. After treatment, the cell supernatants were collected by centrifugation. Supernatants were assayed for LDH in triplicates as per the manufacturer's protocol. Absorbances were measured at 490 nm and 680 nm using a spectrophotometer (Biotek). The percentage of cytotoxicity was calculated based on the percentage difference compared with the LDH-positive control provided with the kit.
Glutamate assay. 5  www.nature.com/scientificreports/ orimetric Assay kit purchased from Cell Biolabs as per the instructions from the manufacturer. For measuring intracellular glutamate, the cells were lysed by sonication in sucrose buffer. After lysis, the cell extracts were centrifuged at 15,000×g for 5 min and the supernatants were subjected to glutamate assay using the kit. All the reactions were performed in triplicates. The glutamate concentrations in the samples were determined by plotting the data against the respective standard curve that was generated in parallel using a series of known glutamate concentrations ranging from 6.25 to 400 µM.
Measurement of intracellular proline levels. 5 × 10 5 SH-SY5Y cells were plated per well in 6-well plates overnight and treated with METH for 24 h. After treatment cells were harvested and lysed in cold PBS with 1% Triton X-100. The cellular debris were removed by centrifugation (10,000×g). The supernatants were transferred to a boiling water bath, and intracellular amino acids were extracted by boiling for 10 min. After centrifugation (5 min, 4 °C, 15,000×g), the supernatant was free of proteins, and intracellular proline was determined as described 51,52 . Briefly, 100 µL of the supernatant was incubated with 100 µl of acid-ninhydrin (0.25 g ninhydrin dissolved in 6 mL glacial acetic acid and 4 mL 6 M phosphoric acid) and 100 µL of glacial acetic acid for 1 h at 100 °C. The reaction was stopped by incubation on ice for 5 min, and the mixture was extracted with 200 µL toluene. The toluene phase was separated, and absorbance at 520 nm was used to determine the concentration of proline. All the reactions were performed in triplicates and a standard curve ranging in concentrations of 0.05-1 mM proline was generated for determining proline concentration of samples.
Western blot analyses. Cells were treated with various concentrations of METH for 24 h, after which cells were harvested and washed with PBS. Cell lysates were prepared as per our published method 50 and quantified according to standard BCA protein assay (Pierce, USA). Equal amounts of cell lysates (20 µg) were resolved on SDS-polyacrylamide gels and transferred to nitrocellulose membranes using a semi-dry blotter (Bio-Rad).
Membranes were blocked using Tris-buffered saline with 5% nonfat milk (pH 8.0; Sigma). Blots were then probed with the appropriate primary antibody in blocking buffer overnight at 4 °C. Incubation with secondary anti-mouse or anti-rabbit IgG antibodies conjugated to horseradish peroxidase (1:2000) was performed at room temperature for 1 h. All blots were washed in Tris-buffered saline with Tween 20 (pH 8.0; Sigma) and developed using the enhanced chemiluminescence (ECL) procedure (BioRad). Blots were routinely stripped by Restore Plus Stripping Buffer (Pierce) and reprobed with anti-actin monoclonal antibody to serve as loading controls. The density of the band was evaluated by Bio-Rad imaging-lab 4.0 software (https ://www.bio-rad.com/en-pl/ produ ct/image -lab-softw are?ID=KRE6P 5E8Z).

Transfection of P5CS expression construct in CHO-K1 cells. CHO-K1 cells (1 × 10 5 ) were seeded
per well in 6-well plates for 24 h in complete F12K media. Prior to transfection, the cells were pretreated with METH for 1 h and then transfected with pcDNA control vector or P5CS expression vector. Transfections were performed with JetPrime (Polyplus) according to the manufacturer's instructions. 24 h post transfection the cells were harvested by scraping. The cell supernatants and pellets were collected by centrifugation. Supernatants were assayed for glutamate and the cell pellets were analyzed by western blot as described before.
Animal studies. All protocols were approved by the University of Florida and Vanderbilt University Institutional Animal Care and Use Committee (IACUC) policies and adhered to the NIH guidelines. WT C57BL/6J male mice were obtained from Jackson Laboratories (Vanderbilt; Stock No. 000664) or from the University of Florida Animal Care Services and were maintained on a 12-h light/dark cycle with food and water available ad libitum in their home cages. The mice were injected with a single dose of METH at 2 mg/kg body weight (acute exposure). Control mice received saline injections. The animals were euthanized after 24 h and whole brain were extracted, rapidly rinsed with ice-cold PBS and immediately flash frozen. Punches from the cortical regions were taken for each sample and homogenized manually in a sucrose buffer system supplemented with protease inhibitors. Samples were spun down at high speed for 5 min, the supernatants were transferred into separate tubes and the protein concentration was estimated by BCA assay (Pierce). For brain slices studies 6-week-old male C57BL/6J mice were euthanized under isoflurane anesthesia after which 300-µm coronal brain slices containing the frontal cortex were prepared from whole brain tissue using a Leica Vibratome in oxygenated (95% O2; 5% CO2) ice-cold artificial cerebrospinal fluid (aCSF, in mM: 119 NaCl, 2.5 KCl, 1.3 MgCl2-6H2O, 2.5 CaCl2-2H2O, 1.0 NaH2PO4-H2O, 26.2 NaHCO3, and 11 glucose). Following a 30-min recovery period, slices were then incubated in either 200 µM amphetamine, 200 µM methamphetamine or vehicle (phosphate buffered saline, PBS). Following a 6-h incubation, cortical enriched tissue was isolated and flash frozen until processing for immunoblot analysis.
Statistical analyses. Data are expressed as mean ± SEM obtained from at least three independent experiments. Significance of differences between control and treated samples was determined by Student's t test. Values of p < 0.05 were considered to be statistically significant. www.nature.com/scientificreports/