Chronic stress decreases GS activity through Tyr-nitration in the mPFC

In CIS-induced depression mice (stressed, STR, Fig. 1a), CORT and ROS/RNS were increased in the plasma (Fig. 1b and c; t = 2.773, df = 11, P = 0.0181 and t = 3.163, df = 11, P = 0.009). There was also an increase in ROS/RNS (Fig. 1d, t = 4.861, df = 12, P < 0.001) and a decrease in GS activity (t = 3.273, df = 12, P = 0.007) without a change in GS expression in the mPFC in the STR group compared with those in the control (CTL) group (Fig. 1e and f). An immunoprecipitation study showed greater Tyr-nitration of GS in the STR group than in the CTL group (Fig. 1g and h, t = 2.611, df = 10, P = 0.0260).

Fig. 1: Tyrosine nitration (N-Y) of glutamine synthetase (GS) is increased in the medial prefrontal cortex (mPFC) in the stress group (STR) compared with the control group (CTL) by chronic immobilization stress (CIS).

a Experimental scheme for CIS (upper panel). Changes in body weight (lower left) and food intake (lower right) during the experiment. b Plasma corticosterone level. c, d Reactive oxygen species (ROS)/reactive nitrogen species (RNS) levels in plasma and the mPFC, respectively. e, f Activity and expression levels of GS. g nitrotyrosine-GS level in the mPFC. h Representative Western blot images for GS, N-Y-GS, and α-tubulin. Data are presented as the mean±SEM. *P < 0.05, **P < 0.01, ***P < 0.001 (t-test).

Tyr has antidepressive and anti-anxiety properties

We examined the denitration of GS via Tyr using rhGS and mouse mPFC lysate (Fig. 2a and b). Tyr protected GS activity against PN-induced Tyr-nitration of GS in a dose-dependent manner. Mice were subjected to CIS and behavior tests with or without a Tyr-supplemented diet (Fig. 2c). A bodyweight difference was only found between the CTL and STR groups, and no difference was found between the normal diet (N) and Tyr-supplemented diet groups. Interestingly, mice fed a Tyr-supplemented diet (STR-Y mice) showed fewer depressive symptoms including anxiety-related behaviors, helplessness, or anhedonic behaviors than did STR-N mice (Fig. 2d–g; d, F(3,24) = 3.446, P = 0.033; t = 2.190, df = 12, P = 0.0490; f, F(3,24) = 4.224, P = 0.016; g, F(3,24) = 11.11, P < 0.001). Plasma CORT and ROS/RNS levels were significantly decreased in STR-Y mice compared with those in STR-N mice (Fig. 2h and i; h, F(3,24) = 16.49, P < 0.001; i, F(3,24) = 11.26, P < 0.001), and ROS/RNS levels in the PFC were also decreased in STR-Y mice compared with those in STR-N mice (Fig. 2j, F(3,24) = 14.65, P < 0.001). The activity, expression, and Tyr-nitration of GS were analyzed simultaneously. Reduced GS activity was increased by Tyr treatment (F(3,24) = 15.27, P < 0.001) with no change in GS expression (Fig. 2k and l). Tyr-nitration of GS was increased in STR mice compared with that in CTL mice but was decreased by Tyr supplementation (F(3,24) = 6.482, P = 0.0023) (Fig. 2m and n). Because GS activity directly affects Glu, Gln, and GABA in the mPFC, we analyzed their amounts in both the PFC and plasma (Fig. 2o–s). CIS decreased Glu and Gln in the mPFC (Glu, F(3,24) = 2.696, P = 0.0685; t = 2.463, df = 12, P = 0.0299; Gln, F(3,24) = 2.689, P = 0.0690) but did not affect their plasma levels. Tyr supplementation reversed Glu, Gln, and Tyr, but not GABA, to CTL levels. To evaluate the recovery of glutamatergic signaling with Tyr supplementation, sEPSCs were measured using glutamatergic neuron-specific labeled transgenic mice (Vglut2-Cre::CRISPR-CAS9) (Fig. 2t). We found that the antidepressive and anti-anxiety effect of Tyr was closely related to the increment in glutamatergic neurotransmission (Fig. 2t–w; u, F(2,33) = 5.048, P = 0.0122; w, F(2,33) = 4.307, P = 0.0213; t = 2.686, df = 24, P = 0.0129).

Fig. 2: Tyrosine (Y) shows antidepressive effects in chronic immobilization stress (CIS)-induced depressive mice via activation of glutamine synthetase (GS) in the medial prefrontal cortex (mPFC).

a and b Denitrative effect of Y on human recombinant GS and mouse mPFC lysate. c Scheme for Y diet supplementation, CIS, and behavioral tests. Changes in body weight and food intake during the experiment among groups (normal diet control group: CTL-N, Y diet control group: CTL-Y, normal diet stress group: STR-N, Y diet stress group: STR-Y, n = 7 per group). d–g Behavioral test results: open field test (d and e), tail suspension test (f), and sucrose preference test (g). h Plasma corticosterone level. i and j Reactive oxygen species (ROS)/reactive nitrogen species (RNS) levels in plasma and the mPFC, respectively. k and l Activity and expression levels of GS. m Nitrotyrosine-GS (N-Y-GS) level in the mPFC. n Representative Western blot images for GS, N-Y-GS, and α-tubulin. o–r Glutamate (Glu), glutamine (Gln), Tyr, and γ-aminobutyric acid (GABA) levels in the mPFC. s Amino acid (Glu, Gln, Tyr, and GABA) levels in plasma. t Traces of spontaneous excitatory postsynaptic currents (sEPSCs) in glutamatergic neurons of the mPFC. The changes of frequency (u), amplitude (v), and cumulative amplitude (w) among groups. Data are presented as the mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001 (multiple comparisons test), and #P < 0.05 (individual comparison test) vs. CTL-N or STR-N groups.

Tyr-Gln (YQ) and Gln-Tyr (QY) have antidepressive and denitration effects on GS

To examine the bioavailability of YQ and QY, we administered mice to each dipeptide-supplemented diet and examined whether the dipeptides had an antidepressive effect in a CIS-induced depression model (Figs. 3 and S1). Before the in vivo experiment, we performed an in vitro test of the denitration effect of the dipeptides on GS compared with that on PN using rhGS and mouse brain lysate (Figs. 3a–b and S1a–b), similar to our approach with Tyr. We found that a higher concentration could be applied and more denitration of GS was achieved, indicating the expanded clinical availability of dipeptides compared with that of the single amino acid Tyr. When YQ- or QY-supplemented diets were tested in our well-established CIS regimen, YQ showed antidepressive effects in behavioral tests (Fig. 3d–g; g, F(3,23) = 7.926, P < 0.001) and QY showed similar mean differences in each behavior test (Fig. S1d–g; d, F(3,25) = 7.864, P < 0.001; f, F(3,25) = 4.430, P = 0.013; g, F(3,25) = 3.560, P = 0.029). The CORT levels in plasma and ROS/RNS levels in both plasma and the mPFC were also reduced by YQ or QY treatment (Figs. 3h–j and S1h–j; h, F(3,23) = 4.138, P = 0.0175; i, F(3,23) = 5.756, P < 0.001; j, F(3,23) = 9.056, P < 0.001; Fig. S1h, F(3,25) = 11.06, P < 0.001; S1i, F(3,25) = 8.375, P < 0.001; S1j, F(3,25) = 14.15, P < 0.001). As expected, YQ or QY increased GS activity compared with that in the STR-N group without GS expression changes (Figs. 3k and l,  S1k and l; Fig. 3k, F(3,23) = 13.64, P < 0.001; Fig. S1k, F(3,25) = 10.95, P < 0.001). An increase in the Tyr-nitration of GS was observed in STR-N mice, but both YQ and QY remarkably reduced the Tyr-nitration of GS to the level observed in CTL-N mice (Figs. 3m and n, S1m and n; Fig. 3m, F(3,23) = 5.803, P = 0.004; Fig. S1m, F(3,25) = 6.055, P = 0.003). Tyr-Gln increased Glu and Gln levels in the mPFC (Fig. 3o and p; o, F(3,23) = 3.573, P = 0.029, p, F(3,23) = 5.093, P = 0.008), but QY did not affect these amino acid levels (Fig. S1o and p). No changes were found in plasma Glu, Gln, Tyr, or GABA levels, similar to the effects observed after Tyr treatment (Fig. 2s). Although CIS evoked hypoactive glutamatergic neurotransmission, YQ recovered sEPSCs and the cumulative amplitude to levels observed in CTL mice (Fig. 3t–w; u, F(2,33) = 5.048, P = 0.0122; w, F(2,33) = 4.307, P = 0.0213; t = 2.686, df = 24, P = 0.0129).

Fig. 3: Tyr-Gln (YQ) shows antidepressive effects in chronic immobilization stress (CIS)-induced depressive mice via activation of glutamine synthetase (GS) in the medial prefrontal cortex (mPFC).

a and b Denitrative effect of YQ on human recombinant GS and mouse mPFC lysate. c Scheme for YQ diet supplementation, CIS, and behavioral tests. Changes in body weight and food intake during the experiment among groups (normal diet control group: CTL-N, YQ diet control group: CTL-YQ, normal diet stress group: STR-N, YQ diet stress group: STR-YQ, n = 7 per group). d–g Behavioral test results: open field test (d and e), tail suspension test (f), and sucrose preference test (g). h Plasma corticosterone level. i and j Reactive oxygen species (ROS)/reactive nitrogen species (RNS) levels in plasma and the mPFC, respectively. k and l Activity and expression levels of GS. m Nitrotyrosine-GS (N-Y-GS) level in the mPFC. n Representative Western blot images for GS, N-Y-GS, and α-tubulin. o–r Glutamate (Glu), glutamine (Gln), Tyr, and γ-aminobutyric acid (GABA) levels in the mPFC. s Amino acid (Glu, Gln, Tyr, and GABA) levels in plasma. t Traces of spontaneous excitatory postsynaptic currents (sEPSCs) in glutamatergic neurons of the mPFC. Changes in frequency (u), amplitude (v), and cumulative amplitude (w) among groups. Data are presented as the mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001 (multiple comparisons test), and #P < 0.05 (individual comparison test) vs. CTL-N or STR-N groups.

Both pre- and post-supplementation with YQ produce antidepressive and denitration effects

To evaluate whether YQ could be used to prevent depression and treat depression after onset, we provided a diet containing YQ before and after CIS (Fig. 4a and p). Both pre- and post-supplementation of YQ showed antidepressive effects on CIS-induced depressive behaviors. The STR-YQ group, which was pre-supplied with YQ, showed an increased duration in the center in the OFT, an increased duration in the open arms + center in the EPM, and a decreased duration in the closed arms compared with those in the STR-N group (Fig. 4b–g; b, F(2,12) = 5.631, P = 0.019; d, F(2,12) = 12.86, P = 0.001; f, F(2,12) = 8.438, P = 0.005; g, F(2,12) = 7.868, P = 0.007). In the FST, the STR-YQ group also showed a decrease in immobile duration compared with that in the STR-N group (Fig. 4h; F(2,12) = 5.355, P = 0.022). The CORT and ROS/RNS levels in the plasma and mPFC were lower in the STR-YQ group than those in the STR-N group (Fig. 4i–k; i, F(2,12) = 17.29, P < 0.001; j, F(2,12) = 9.177, P = 0.004; k, F(2,12) = 23.76, P < 0.001). GS activity was restored in the STR-YQ group without a change in the GS expression level (Fig. 4l and m; l, F(2,12) = 67.08, P < 0.001), and Tyr-nitration of GS was decreased in the STR-YQ group compared with that in the STR-N group (Fig. 4n and o; n, F(2,12) = 9.303, P = 0.004).

Fig. 4: Tyr-Gln (YQ) of both pre- (a-o) and post-supplementation (p-aa) shows antidepressive effects in chronic immobilization stress (CIS)-induced depressive mice.

a and p Experiment scheme of YQ supplementation including CIS and behavior tests and change of body weight and daily food intake during experiment (normal diet control group: CTL, normal diet stress group: STR-N, YQ diet stress group: STR-YQ, n = 5 for pre-supplementation and n = 7 for post-supplementation per group). b, c, and q, r Open field test. d–g Elevated plus maze. h Forced swimming test. i and u Corticosterone level in plasma. j and v Reactive oxygen species (ROS)/reactive nitrogen species (RNS) level in plasma. k and w ROS/RNS level in mPFC. l and x Activity of glutamine synthetase (GS) in mPFC. m and y Relative GS expression level in mPFC. n and z Relative nitrotyrosine-GS (N-Y-GS) level in mPFC. o and aa Representative Western blot images for GS, N-Y-GS, and α-tubulin. s Tail suspension test. t Sucrose preference test. Data are presented as the mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001 (multiple comparisons test) vs. CTL or STR-N groups.

The STR-YQ group, which was post-supplied with YQ, showed a decreased immobile duration in the TST and an increased sucrose preference compared with those in the STR-N group (Fig. 4s–t; s, F(2,18) = 11.95, P < 0.001; t, F(2,18) = 5.676, P = 0.012). The CORT and ROS/RNS levels in the plasma and mPFC were lower in the STR-YQ group than those in the STR-N group (Fig. 4u–w; u, F(2,18) = 8.673, P = 0.003; v, F(2,18) = 4.082, P = 0.035; w, F(2,18) = 10.50, P < 0.001). GS activity was restored in the STR-YQ group without a change in the expression level (Fig. 4x and y; x, F(2,18) = 20.23, P < 0.001), and Tyr-nitration of GS was also significantly decreased in the STR-YQ group (Fig. 4z and aa; z, F(2,18) = 3.837, P = 0.041).

Tyr and YQ reduce kainic acid-induced epileptic seizures and decrease Tyr-nitration of GS in the hippocampus

We determined whether Tyr and/or YQ induce GS denitration and reduce seizures caused by kainic acid (KA) (Fig. 5). Tyr and YQ were provided to mice via their diet or intraperitoneal (i.p.) injection (Fig. 5a). Diet supplementation of 5×Y/3×YQ and i.p. administration of YQ (100 mg/kg) decreased seizure levels compared with those in the N group (Fig. 5b and i; b, F(3,44) = 3.939, P = 0.0142; i, F(2,33) = 4.966, P = 0.013). KA markedly increased IBA-1 in the CA3 hippocampal region in the KA-N group, which was decreased in the 3×Y, 5×Y, and 3×YQ diet–supplemented groups (Fig. 5c and j; c, F(4,5) = 19.95, P = 0.003; j, F(3,4) = 9.288, P = 0.028). ROS/RNS were also increased in hippocampal tissue after KA injection; however, diet supplementation of 3×Y or 5×Y or YQ i.p., 3×YQ, or i.p. injection of YQ reduced ROS/RNS levels (Fig. 5d and k; d, F(4,9) = 33.42, P < 0.001; k, F(3,21) = 14.52, P < 0.001). GS activity was reduced by KA, but Tyr or YQ treatment increased GS activity without a change in its expression (Fig. 5e–g, and l–n; e, F(4,10) = 5.717, P = 0.012; i, F(3,31) = 3.278, P = 0.034, t = 2.358, df = 19, P = 0.029, KA-N vs. KA YQ i.p. / t = 2.155, df = 16, P = 0.047, KA-N vs. KA 3xYQ). GS nitration was also increased by KA but was decreased by Tyr or YQ treatment (Fig. 5f, h, m, and o; o, F(3,18) = 4.707, P = 0.014, t = 2.618, df = 11, P = 0.024, KA-N vs. KA YQ i.p. / t = 2.450, df = 7, P = 0.044, KA-N vs. KA 3xYQ).

Fig. 5: Tyr-Gln (YQ) attenuates kainic acid (KA)–induced seizures and the decrement in glutamine synthetase (GS) activity in the hippocampus.

a Scheme for Y/YQ diet supplementation and YQ/KA i.p. injections (n = 10 per group). b and i Changes in seizure levels after KA and different treatments (normal diet without KA treatment group: CTL, normal diet with KA treatment group: N, 1×Y diet with KA treatment group: 1×Y, 3×Y diet with KA treatment group; 3×Y, 5×Y diet with KA treatment group: 5×Y, YQ i.p. administration with KA treatment group: YQ i.p., 3×YQ diet with KA treatment group: 3×YQ). c and j Representative images and relative densities of IBA-1 expression in the CA3 region of the hippocampus (n = 4 per group). d and k Reactive oxygen species (ROS)/reactive nitrogen species (RNS) levels. e and l GS activity changes in the hippocampus. f and m Representative Western blot results for N-Y-GS, GS, and α-tubulin. Changes in GS expression (g and n) and N-Y-GS levels (h and o) among groups. Data are presented as the mean ± SEM. *P < 0.05, **P < 0.01, ***P < 0.001 vs. CTL and #P < 0.05, ##P < 0.01, ###P < 0.001 vs. N (one-way ANOVA with multiple comparisons test).

Tyr and YQ alleviate blood ammonia levels and liver dysfunction in hyperammonemia models

We determined whether Tyr or YQ has a beneficial effect by regulating blood ammonia via GS activation in the liver in two hyperammonemia mouse models, azoxymethane (AOM)-induced liver failure as an acute model and bile duct ligation (BDL)-induced liver failure as a chronic model. In the AOM model, blood ammonia and Tyr-nitration levels in the liver were increased by AOM, but these increments were attenuated by Tyr (100 mg/kg) or YQ (200 mg/kg) treatment (Fig. 6b, d, f, and h; b, F(2,13) = 17.42, P < 0.001; d, F(2,7) = 4.109, P = 0.041; f, F(2,10) = 54.15, P < 0.001; h, F(2,7) = 11.08, P = 0.007). The elevated plasma ALT level was reduced by Tyr or YQ treatment, implying that Tyr and YQ at least partly protect liver function against AOM toxicity (Fig. 6c and g; c, F(2,10) = 170.4, P < 0.001; g, F(2,10) = 12.26, P = 0.002). The decrease in the blood ammonia level may have been due to the maintenance of GS activity by YQ treatment (Fig. 6i. F(2,6) = 16.41, P = 0.004). Similar to AOM-induced liver failure, BDL-induced liver dysfunction was evidenced by increased blood ammonia and plasma ALT and alkaline phosphatase levels, but these changes were alleviated by Tyr or YQ treatment (Fig. 6k, l, m, p, q, and r; k, F(3,42) = 8.448, P < 0.001; l, F(3,47) = 37.74, P < 0.001; m, F(3,43) = 64.22, P < 0.001; p, F(3,17) = 5.806, P = 0.006; q, F(3,20) = 21.93, P < 0.001; r, F(3,16) = 9.749, P < 0.001). Additionally, there was a large increment in Tyr-nitration of liver proteins and a decrement of GS activity by BDL. Chronic oral administration of Tyr (100 mg/kg) or YQ (200 mg/kg) after BDL increased Tyr-denitration and GS activity (Fig. 6n, o, s, and t; n, F(3,10) = 6.510, P = 0.01; o, F(3,8) = 5.099, P = 0.029; s, F(3,7) = 7.101, P = 0.016; t, F(3,16) = 9.749, P < 0.001).

Fig. 6: Tyr and Tyr-Gln (YQ) reduce blood ammonia levels and liver injury in azoxymethane (AOM) and bile duct ligation (BDL)–induced hyperammonemia mouse models.

a and j Schemes for Y or YQ administration during AOM and BDL-induced hyperammonemia model, respectively. b, f, k, and p Changes in blood ammonia level among groups (normal control: CTL; vehicle administration with AOM, sham, or BDL treatment: V; Y administration with AOM, sham, or BDL treatment: Y; YQ administration with AOM, sham, or BDL treatment: YQ). c, g, l, and q Plasma alanine aminotransferase (ALT). d, h, n, and s Changes in nitrotyrosine-containing protein levels in the liver. Nitrotyrosine-containing protein levels were normalized to those of β-actin. e, i, o, and t GS activity changes in the liver. m and r Plasma alkaline phosphatase (ALP). Data are presented as the mean ± SEM. *P < 0.05, **P < 0.01, ***P< 0.001 vs. CTL or V groups (multiple comparisons test).