Glutamic acid decarboxylase (GAD, EC 4.1.1.15) catalyzes the irreversible α-decaboxylation of glutamate to produce GABA. GAD uses PLP as coenzyme. Therefore, from the theoretical point of view, an addition of PLP to medium may be a workable method to increase GAD activity to enhance synthetic capacity of GABA. Previous studies had shown that the addition of PLP to medium could effectively increase the GABA production of LAB [¹⁵,¹⁹]. In our present study, however, the addition of PLP neither affected the cell growth (Figure 1A) nor increased GABA content (Figure 1B) in the fermentation. The possible reason was that NCL912 cells could synthesize sufficient PLP for themselves.

Figure 2 shows that the considerable variation in the yields of both bacterial growth and GABA production under different fermentation temperatures. The bacterial growth increased with the increase of temperature and peaked at 35°C, then decreased with the increase of temperature. For GABA production, a trend similar to the bacterial growth was observed. It was clear that high cell density was required for effective synthesis of GABA. On the other hand, GABA concentration at 30°C was almost the same to that at 35°C but biomass at 30°C was less than that at 35°C. In addition, NCL912 could not produce GABA at 45°C while the strain could grow under this temperature. These data indicated that appropriate temperature was beneficial to produce GABA, and excessively high temperature was unfavorable to the GABA production. The above results indicated that high efficient conversion glutamate to GABA needed not only high cell density but also appropriate temperature. By comprehensive consideration of the above data, 32°C was selected for the following tests.

It was reported that the GABA biosynthesis in LAB was strictly pH regulated [¹⁵,¹⁹,³⁶]. To investigate the effect of different pH levels on the production of biomass and GABA of NCL912 in the course of fermentation, the initial pHs of the media were adjusted to 3, 4, 5 and 6, respectively. During the fermentation process, the pHs were adjusted to corresponding initial values immediately after sampling, respectively. As shown in Figure 3, pH has a significant effect on the production of biomass and GABA. The strain could hardly grow at pH3.0 and almost no GABA produced. A highest yield of GABA was obtained at pH5.0 in the fermentation course even if the biomass was less than that of pH6.0 after 24 h. The optimal pH value for the GABA production was 5.0 that accorded with the previous reports about the optimal pH values for maintaining the activity of LAB GADs were in the range of 4.0 to 5.0 [¹⁴,^37-39]. The higher or lower pH may lead to the partial loss of the GAD activity.

pH 5.0 was also the optimal pH for the cell growth, though the biomass decreased drastically after 24 h. However, the biomass kept increasing during the fermentation course when pH values were maintained at 4.0 or 6.0. It might be the combined inhibitory effect of high concentration of GABA and H₂SO₄. For pH5.0, more GABA was produced than pH4.0 and 6.0. The decarboxylation of glutamate to GABA catalyzed by GAD takes the following general form [¹,⁸]:

Decarboxylation of glutamate occurred in LAB results in the stoichiometric release of the end product GABA and the consumption of a proton. The net effect of this reaction is to increase the alkalinity of the cytoplasm and environment. For pH5.0, more H₂SO₄ was therefore supplemented into the fermentation broth in order to offset pH increase arising from the decarboxylation.

As shown in Figure 4A, a moderate addition of glutamate to the medium resulted in an increase in biomass. LAB can metabolize sugars to produce large amount of low weight molecule organic acids, resulting in an acidic environment that was hard for bacteria growth. Some LAB can employ GAD system for maintaining neutral cytoplasmic pH when the external pH drops because the decarboxylation of glutamate within the LAB cell consumes an intracellular proton. In Figure 4C, the pH value of the culture without glutamate decreased to 3.4 after 24 h of fermentation. pH values of the cultures supplemented with glutamate, however, are generally above 5.0. It implied that GAD system of Lb. brevis NCL912 acted under low pH and resulted in an increase of pH in medium with glutamate, and protected cell survival from acidic condition. The role of GAD conferring acid resistance to microbial cells was further verified in the present study. On the other hand, the cell growth and biomass decreased with the increase of glutamate concentration at the given levels (0.25, 0.5, 0.75 and 1.0 M). It was apparent that extra high concentration of glutamate was harmful to the strain growth. An addition of 0.25-0.5 M of glutamate was suitable for the strain NCL912 growth and production of GABA.

In the fed-batch process, the initial concentration of glutamate in medium was 400 mM, for the cell growth was strongly inhibited when glutamate exceeded 500 mM. The time courses of GABA production, residual glutamate, DCW, and residual glucose were tested (Figure 5). The cell growth occurred immediately after inoculation and biomass rapidly increased at the first 12 h. And then the biomass dramatically decreased due to the combined inhibitory effect of high concentration of GABA, glutamate and H₂SO₄. The concentrations of GABA and glutamate were 381.6 mM and 484.6 mM at 12 h, respectively. In addition, about 100 ml of 10 N H₂SO₄ was supplemented into the fermentor in order to offset pH increase arising from the glutamate addition and the decarboxylation. Such harsh conditions were certainly detrimental to the cells and therefore resulted in a sharp decline in biomass after 12 h. The cell growth was almost completely inhibited after 36 h. Without question, compared to feeding strategies maintaining low level glutamate, the current utilized feeding strategy strongly inhibited the cell growth. This feeding strategy, however, was simple and energy-saving. The most important was that the current fermentation method still exhibited powerful capacity of synthesis of GABA (reaching 1095.63 ± 61.03 mM at the end of the fermentations). The biosynthetic kinetics indicate that the GABA concentration increased rapidly with the fermentation time from 0-36 h, increased slowly from 36-60 h, and kept constant after 60 h. The GABA concentrations at 36, 48 and 60 h were 931.50 ± 29.65, 1005.81 ± 47.88 and 1075.05 ± 82.72 mM, respectively, which were obviously higher than that (345.83 mM) [³⁵] before the optimization. Based on a comprehensive consideration of the GABA concentration, energy conservation and fermentation period, 48 h of fermentation was recommended in the future practical production. The volume of the fermentation broth was increased to about 3.75 L due to the inoculation and feed. Residual glutamate and glucose were 134.45 ± 24.22 mM and15.28 ± 0.51 g L^-1 at 48 h. Total 738.24 g glutamate (plus the glutamate in the seed medium) was added into the fermentation medium, in which the converted glutamate was 705.81 ± 33.60 g according to the generated GABA mol number, and the residual glutamate was 94.35 ± 17.00 g. This good balance of glutamate added, converted and residual showed that the added glutamate did not participate in other metabolisms. Complete conversion of substrates was beneficial to save material and purify the end product from culture broth. In further practical applications, an addition of 112.28 g glutamate at 24 h, and 35 g L^-1 of glucose in the medium are suggested.

GABA-producing Lb. brevis NCL912 was isolated by our laboratory from paocai, a Chinese traditional fermented vegetable [¹¹]. In a previous study, we optimized the fermentation medium for production of GABA by Lb. brevis NCL912, and this medium was theoretically composed of (g L^-1) [³⁵]: glucose, 55.25; soya peptone, 30.25; MnSO₄·4H₂O, 0.0061; and Tween-80 1.38 mL L^-1. To facilitate the calculation and preparation of the medium in practice, we adjusted the medium components to (g L^-1): glucose, 50; soya peptone, 25; MnSO₄·4H₂O, 0.01; and Tween 80, 2 mL L^-1. Unless otherwise emphasized, the initial sodium L-glutamate concentration in the medium was 500 mM. The seed medium was composed of (g L^-1): glucose, 50; soya peptone, 25; MnSO₄·4H₂O, 0.01; L-glutamate, 150 mM; and Tween 80, 2 mL L^-1. Nitrogen sources, glutamate and the other compositions were autoclaved separately at 121°C for 20 min and mixed together prior inoculation. Lb. brevis NCL912 was cultured in the seed medium at 32°C for 10 h till the absorbance value at 600 nm (A₆₀₀) between 4.0 and 6.0 and then used for seed culture inoculation. Culture condition optimization was conducted in 250-mL Erlenmeyer flasks that contained 100 mL of the medium. The flasks were inoculated with 10% (v/v) seed culture grown to early exponential phase, and incubated under static conditions in an incubator. The fed-batch fermentation was carried out in a 5-L fermentor (Labo-controller MDL-8C; B. E. Marubishi, Tokyo, Japan) under following conditions: medium volume 3 L, inoculum size 10% (v/v), temperature 32°C, pH 5.0, agitation speed 100 rpm, and fermentation time 84 h. The pH was kept constant at 5.0 with addition of 10 N H₂SO₄. 280.70 g and 224.56 g glutamate were supplemented into the bioreactor at 12 h and 24 h, respectively. Each flask fermentation was performed in two replicates and the fed-batch fermentation was performed in three replicates.

Glutamate and GABA concentrations in the culture broths were determined by pre-staining paper chromatography [⁴⁰]. Cell growth was monitored by measuring A₆₀₀ on a TU-1901 UV-vis spectrophotometer (Beijing Purkinje General Instrument, China). Dry cell weight (DCW) was calculated by a consistent calibration curve of DCW versus A₆₀₀. Glucose concentration was determined by 3, 5-dinitrosalicylic acid method [⁴¹]. Each sample was analyzed in duplicate and the mean values were calculated.