Screening of lactic acid bacteria strains isolated from Iranian traditional dairy products for GABA production and optimization by response surface methodology

A total of 50 lactic acid bacteria (LAB) isolates from Iranian traditional dairy products (Motal and Lighvan cheeses, and artisanal yogurt) were screened for gamma-aminobutyric acid (GABA) production. Firstly, a rapid colorimetric test was performed to evaluate the glutamate decarboxylase (GAD) activity among the LAB isolates examined. Thin layer chromatography (TLC) was then performed on selected strains to identify isolates with high/moderate GABA producing capacity, and a GABase micro-titer plate assay was employed to quantify GABA. Finally, two Lactococcus (Lac.) lactis strains were selected for GABA production optimization via Response Surface Methodology (RSM) following Central Composite Design (CCD). Forty-one out of the 50 isolates showed GAD activity according to the colorimetric assay. Eight isolates displayed strong GAD activity, while nine showed no activity; low to moderate GAD activity was scored for all other isolates. GABA production was confirmed by TLC in all isolates with high GAD activity and in four selected among isoaltes with moderate activity. Among the Lactococcus strains tested, Lac. lactis 311 and Lac. lactis 491 were the strongest GABA producers with amounts of 3.3 and 1.26 mM, respectively. These two strains were subjected to GABA production optimization applying RSM and CCD on three key variables: Monosodium glutamate concentration (MSG) (between 25 and 150 mM), incubation temperature (between 25 and 37 °C), and pH (between 4.0 and 5.0). Optimal conditions for GABA production by Lac. lactis 311 and Lac. lactis 491 of temperature, pH and MSG concentration were, respectively, 35.4 and 30 °C, pH 4.5 and 4.6, and MSG concentration of 89 and 147.4 mM, respectively. Under the above conditions, the amount of GABA produced by Lac. lactis 311 and Lac. lactis 491 was 0.395 and 0.179 mg/mL, respectively. These strains and the optimal culture conditions determined in this study could be used for the biotechnological production of GABA or applied in food fermentations for the development of naturally GABA-enriched foods.


LAB strains and growth conditions. A total of 50 LAB isolates previously isolated from Iranian tradi-
tional dairy products (Motal cheese 33 , Lighvan cheeses 31,34 , and Iranian yogurt 35 ), identified using molecular techniques and 16S rRNA gene sequencing, were screened for GABA production ( Table 1). All examined isolates, including lactobacilli (Lactobacillus, Lactiplantibacillus, and Levilactobacillus) (27 isolates), Lactococcus (20 isolates), and Streptococcus (3 isolates), were kept frozen in MRS (lactobacilli) or M17 (lactococci and streptococci) broth with 15% glycerol at -80 °C. Pre-cultures were made by adding 20 µL of stock culture to 10 mL fresh medium. Lactobacilli strains were grown in MRS medium, while M17 medium was used for lactococci and Streptococcus (St.) thermophilus strains. Based on the types, incubations were done at 30 or 37ºC for 24-48 h under aerobic or anaerobic conditions. Qualification screening of LAB isolates. Determination of GAD activity. Glutamic acid decarboxylase (GAD) assay is a rapid colorimetric test used to assess the GABA-producing potential among LAB isolates. Sample preparation for GAD activity determination was done as described by Lacroix et al. 36 . An overnight grown culture of each isolate (5 mL) was washed with 0.9% (w/v) NaCl solution and centrifuged at 5000 × g (20 min, 25 ºC). Cells were suspended in 500 µL of GAD reagent solution (pH = 4), which consists of L-glutamic acid (1 g), Triton X-100 (300 µL), NaCl (90 g), and bromocresol green (0.05 g) in 1 L deionized water. After anaerobic in-  www.nature.com/scientificreports/ cubation for 4 h at 37ºC, GAD activity was visually examined by color change. The development of a greenish or blueish color was considered as low or high GAD activity, respectively. However, no color change (yellow) was considered with no activity 36 .
Thin layer chromatography (TLC). To identify isolates with high and moderate GABA producing ability, a TLC assay was performed as described by Lee et al. 37 . Briefly, strains were inoculated in their respective broth medium supplemented with 0.5% monosodium glutamate (MSG) and incubated at their optimal temperature for 24 h. To test the presence of GABA, supernatants were collected by centrifugation (8000 × g, 4ºC, 5 min) and then 2 µL of each filtered supernatant was spotted on a TLC plate (Silica gel 60 F254). TLC was conducted using n-butanol:acetic acid:water ratio of 5:2:2 (v/v/v). The silica gel plate was subsequently developed by spraying with a 2% ninhydrin solution and heating at 105ºC for 5 min.
DNA extraction and strain typing by rep-PCR. Total genomic DNA was extracted from GABAproducing isolates using a commercial kit (High Pure PCR Template Preparation Kit; Roche, Basel, Switzerland), following the manufacturer's instructions. To assess the interspecies diversity between isolates, repetitive extragenic profiling PCR (rep-PCR) was done using BoxA2R (5´-ACG TGG TTT GAA GAG ATT TTCG-3´) as reported by Koeuth et al. 38 . In brief, reactions were performed in 25 µL volume containing 2 µL purified DNA, 12.5 µL Red master mix (Ampliqon, Odense, Denmark), 9.3 µL Nuclease-free water, and 1.2 µL of the primer. Thermal conditions were as follows: an initial step of denaturation at 95 °C for 5 min, followed by 35  GABA determination. Quantitative measurment of GABA in broth medium was done by a GABase microtiter plate assay as described by Tsukatani et al. 39 . Bacterial isolates were grown in MRS broth medium containing 0.5% MSG for 2 days at their optimal growth temperature. The reactive mixture contained 750 mM sodium sulfate, 10 mM dithiothreitol, 1.4 mM NADP + , 2.0 mM α-ketoglutarate in Tris-HCl buffer (80 mM, pH 9.0) and GABA-T (Gamma-aminobutyrate glutamate aminotransferase) (Sigma-Aldrich). The mixture was added to each well (in a 96-well microtiter plate) in a final volume of 90 µL. The initial absorbance was read at 340 nm with a microplate reader (BioTek, ELX 808) before adding a 10 µL sample of a standard GABA solution. The substrate of NADP + turns to NADPH in the presence of GABA and α-ketoglutarate after incubation at 30-37ºC for 1 h.
To calculate the GABA content formed by each isolate, the difference between initial and final absorbances at 340 nm was compared to standard curves of calibrated GABA solutions.
Optimization of GABA production using response surface methodology (RSM). Optimization of GABA production in MRS medium and studying the interaction of key parameters were done using response surface methodology (RSM). A Central Composite Design (CCD) of RSM was applied to optimize the 3 effective variables: MSG concentration (between 25 and 150 mM), incubation temperature (from 25 to 37 °C), and pH (4.0-5.0). Varying these three independent parameters, a total of 20 experiments for each isolate were designed and performed. The statistical software Design Expert 10.0.7.0 was used for data analysis 12,40 . Laboratory-scale fermentations were set up to investigate optimal parameters for GABA production by selected isolates. Bacterial isolates were inoculated into falcon tubes containing 50 mL MRS broth medium with different concentrations of MSG with initial pH adjustment and incubated at different temperatures for 48 h to reach optimum GABA production. After 48 h of incubation, samples were withdrawn from the each falcon tube to determine GABA content.

Statistical analysis.
To analyze the results related to the potential of GABA production by examined bacteria, Duncan's one-way analysis of variance (ANOVA) was applied using the SPSS software (ver. 16), while to optimize the conditions of GABA production, RSM with Design Expert 10.0.7.0 software was used.

Results
Screening of GAD activity. Total number of LAB isolates and the results of their previous identification 31, [33][34][35] are presented in Table 1. These isolates include members of lactobacilli (of the genera Lactobacillus, Lactiplantibacillus, and Levilactobacillus) (27 isolates), Lactococcus (20 isolates), and Streptococcus (3 isolates). The GAD activity of the 50 isolates from Iranian traditional dairy products are also summarized in Table 1. The cultures of the isolates Levilactobacillus (Lb.) brevis M2, M4 and M12; Streptococcus (St.) thermophilus 43, and Lac. lactis 261, 311, 454, and 491 developed a strong blue color, consistent with a high GAD activity. No color change was observed for the isolates M3, M6, M7, M8, M9, M13, M15, M17, and M19; thus these are considered not to have GAD activity. Other isolates with greenish-blue, green and greenish-yellow color were considered as having medium, low and very low GAD activity, respectively. Fourteen isolates, including all high and some with moderate GAD activity were selected for subsequent experiments. GAD assay results revealed that most of the GAD-positive isolates were isolated from cheese (Lighvan and Motal); whereas, only one isolate from a traditional yogurt exhibited high GAD activity. www.nature.com/scientificreports/ Genetic analysis of isolates. Lactobacilli, Streptococcus and Lactococcus species were targeted by rep-PCR to evaluate the intramolecular diversity among the isolates, as well as to assess whether isolates with GAD activity belonging to the same species were replicates or independent strains. Further, typing of the strains could serve for selecting different strains for GABA production optimization. The different patterns of similarity were clustered using the SM coefficient by the UPGMA method. Given the reproducibility of the assay around 94 (Fig. 1), isolates sharing a percentage of similarity > 94% were considered to belong to the same strain. Figure 2  a also showed very high homology. The relation between the intra-species diversity with GABA production in these species showed that GABA production was significantly different between isolates with high similarity and comparable between isolates with low similarity. Unlike other examined strains, there was no significant difference in GABA production in St. thermophilus isolates with high and low similarity.

Confirmation of GABA by thin layer chromatography.
Isolates that showed no GAD activity by the colorimetric method were excluded from further analyses. Based on the results of GAD activity, we selected 14 isolates corresponding to the species St. thermophilus (three), Lb. delbrueckii (four), Lac. lactis (four), Lb. brevis (three) for further studies. The GABA production was confirmed by observing red spots on the TLC plates (Fig. 3). The spot mobility from the culture supernatant of the 14 isolates was consistent with GABA, as a retention factor (Rf) value (0.77 cm) of the sample matched with the GABA standard. Thus, all 14 selected isolates were considered as GABA producers. GABA's of strains show the same retention factor (Rf) value as the GABA standard (= 0.77 cm). The retention factor was defined as the ratio of the distance traveled by the center of a spot to the distance traveled by the solvent front.
Quantitative GABA production measurement. GABA production of selected strains was quantified using a GABase microtiter plate enzymatic determination. As shown in Table 1, GABA production varied widely ranging from 0.01 to 0.36 mg/L (equivalent to 0.12 and 3.52 mM). All Lb. brevis isolates from Motal cheese  Tables 2 and 3, respectively. The interaction effect of these factors on GABA production concentration by Lac. lactis 311 is described by the following equation: where R is the concentration of GABA (mg/mL) and A, B and C are coded values of the independent variables, viz., MSG, temperature and pH, respectively. www.nature.com/scientificreports/ www.nature.com/scientificreports/ Three-dimensional surface plots were generated to estimate the effect of the combinations of the independent variables on the GABA production using Design Expert software. The 3D plot in Fig. 4A shows the interaction effect of MSG and temperature, keeping pH at its central level 4.5. According to the 3D plot, predictably, the GABA concentration increased with increasing MSG concentration and the increase of temperature. Nevertheless, in the low MSG concentration, increasing GABA concentration was more sharply compared to a higher concentration. The maximum concentration (3.83 mM) was observed with MSG in concentration of 89 mg/mL and temperature of 35.4 °C. Figure 4B presents the 3D plot showing the dependency of GABA concentration on MSG concentration and pH, keeping the third factor (temperature) at its central level of 31 °C. The graph predicted that the highest GABA concentration by relevant strain at 35.4 °C was observed with the mid pH (4.5) and MSG concentration (89 mg/ml). However, GABA concentration at lower pH remained approximately constant by increasing MSG concentration. The 3D curve showing the interaction effect of temperature and pH has shown in Fig. 4C. The low GABA concentration was observed with the lowest temperature and lowest pH. The maximum GABA concentration (3.83 mM) was observed with pH (4.5) and temperature (35.4 °C).
Equation 2 explains the interaction effect of the independent variables on GABA concentration by Lac. lactis with strain code 491: Figure 5A showed that with increasing the concentration of monosodium glutamate and increasing the temperature, the amount of GABA formed increased. This trend is in agreement with the results reported elsewhere by other authors 41,42 .
An increase in the amount of GABA was observed with increasing MSG and pH to 89 mg/ml and 4.59, respectively, at a constant temperature (Fig. 5B). When the concentration of monosodium glutamate was kept www.nature.com/scientificreports/ constant, the GABA level increased sharply with increasing temperature, but increased pH led to the production of low amounts of GABA (Fig. 5C). The incubation temperature is another important factor that may affect GABA production. In addition to the stability and biological activity of the enzyme, temperature also affects the thermodynamic equilibrium of the reaction.
Selection of optimal treatment and validation of model. The results of GABA production by two examined strains showed the optimal conditions for GABA production by Lac. lactis 311 and Lac. lactis 491 had a temperature of 35.4 and 30 °C, a pH of 4.5, 4.6 and a concentration of 89 and 147.4 mM of MSG, respectively. In the above conditions, the amount of GABA production by these two strains was 3.83 and 1.73 mM, respectively. Examination of the results in Fig. 6 shows that the points with a good approximation are on the straight line and the values obtained experimentally are in a good agreement with the values predicted by the model and showed the accuracy of the equation model.

Discussion
Regarding the Motal cheese isolates in this study, the results are consistent with previously reported levels by Ebadi Nezhad et al. (2020) 44 . GABA production begins in the exponential growth phase and increases near to the stationary phase due to an increased in GAD activity, which agrees with GAD being an intracellular enzyme produced in response to acidic conditions 45 . The nutrient sources required for the GABA production by bacteria include the primary sources of carbon, nitrogen (including GABA substrate, glutamic acid) and minerals 46 .
To avoid the use of replicates from the optimization studies, a strain-level typing scheme by rep-PCR was employed. Knowing intra-species diversity also helps to know how diverse the extend of GABA production is in www.nature.com/scientificreports/ closely-related and unrelated strains of the same species 10 . As for many other phenotypic traits, large phenotypic variations have occasionally been reported for genetically closely-related LAB strains 47 . Typing of LAB isolates from Koozeh and Lighvan cheese by rep-PCR with BOXA2R primer and typing of GABA-producing LAB isolates with rep-PCR by (GTG) 5 primer has been previously reported by others 10,31 . As in this work, high intraspecies diversity was reported in the referred studies. Presumptive detection of GABA production by LAB species by TLC, followed by its confirmative identification and quantification by High-performance liquid chromatography (HPLC) has been previously reported to be a successful strategy by Zhang et al. 48 . These authors reported that the mobility spot of the culture supernatant of BC114 strain was basically consistent GABA standard and could be preliminarily identified as GABA within a negligible margin of error. The measured sample from the fermentation broth (a complex matrix), may also result in incomplete agreement with pure GABA standards 48 . The strain under investigation, Lb. plantarum BC114, showed the highest GABA production as measured by HPLC, which reached a value of 1.52 ± 0.07 g/L 48 . Tanamool et al. 49 also used the TLC method to screen GABA-producing lactic acid bacteria strains. These authors found that among 44 LAB, the isolates L10-11 clearly produced the highest GABA based on the TLC results 49 . In our results, 14 isolates among 50 LAB were considered as GABA producers. Valenzuela et al. 25 showed the ability of Lb. brevis to produce GABA is consistent with other reported levels for Lb. brevis LMG6906 (0.29 g/L). Lb. brevis CECT8183 isolated from Spanish cheese by Diana et al. 17 produced 0.1 g/L of GABA. Lb. brevis BJ20 produced 0.002 g/L in a fermented sea tangle solution, which is a popular traditional marine food in Korea based on brown seaweed 50 . Lb. brevis PM17 was among the most potent studied isolates by Franciosi et al. 9 . By contrast, no significant GABA production was detected for any St. thermophilus strain isolated from yogurt products, compared to lactobacilli and Lactococcus species in study. According to the literature, large numbers of strains from these genera from various fermented foods are indicated as GABA-producing LAB 13 . There are many publications reporting variable GABA production levels by different LAB species. Among other factors, production could be affected by the GABA detection methods. For instance, Lb. delbrueckii PR1 (63 mg/kg) and Lac. lactis PU1 showed the highest GABA concentrations 9 . The results of Ly et al. 10 showed Lb. futsaii, Lb. namurensis and Lb. plantarum to produce high amounts of GABA in the range of 1.7-2 g/L. Hwang et al. 51 have examined GABA production by two Lac. lactis strains in MRS containing 5% MSG at 30 °C. Maximum GABA production was observed after 40 h (1.37 g/L). Redruello et al. 52 studied the effect of 6 Lac. lactis GABA-producing strains in cheese making. GABA accumulated at concentrations up to 0.457 g/Kg in cheese. Galli et al. 53 revealed GABA content of their fermented milks by two mix starters (Lac. lactis and Lb. rhamnonus or Lb. paracasei) 0.185 and 0.319 g/L, respectively by adding 249 mg/L MSG. Santos-Espinosa et al. 54 demonstrated the highest GABA concentration (0.086 g/L) in fermented milk with Lac. lactis L-571 at 37 °C with 3 g/L of glutamate substrate. In a study conducted by Tajabadi et al. 55 the optimum conditions for maximum GABA production by Lb. plantarum Taj-Apis362 were an initial glutamic acid concentration of 497.97 mM, culture temperature of 36 °C, initial pH of 5.31 and incubation time of 60 h, which produced 0.74 g/L of GABA.
GABA-producing ability is linked to the activity of GAD enzyme in LAB 56 . The source of LAB might affect GAD activity and GABA production capacity, as it was speculated that acidified foods could probably be the natural niche of GABA producers 10 . Li et al. 57 identified and characterized the GABA-producing lactic acid bacteria and claimed that most LAB strains are able to produce the highest amount of GABA in the range of pH 4 to 5 at 30 to 50 °C and in the presence of glutamic acid. For this reason, this temperature and pH range was used in their study 57 . GABA production occurs under acidic conditions. LAB metabolism mainly leads to the production of organic acids such as lactic and acetic acid. This means that they have to frequently face acidic stress and therefore, they have developed different mechanisms to cope with acidic conditions such as the GAD pathway. Laroute et al. 58 claimed the activation of this pathway occurred after changing the pH to 4.6 and during www.nature.com/scientificreports/ the stationary phase. Their results are consistent with our results which optimum pH for GABA production in two examined isolates is 4.5 and 4.6. As reported by Laroute et al. 58 a dose of 152 mM glutamate, increased the production of GABA by Lac. lactis NCDO 2118. This dosage of glutamate is similarly optimum dose for GABA production in our examined isolate (Lac. lactis 491). The findings of our work showed Lac. lactis 311 produced more GABA in lower concentration than Lac. lactis 491. The probable reason that lower concentrations of MSG can increase the efficiency of GABA production in comparison to higher concentrations of it, is that large amounts of MSG increase the osmotic pressure of cells and disrupt bacterial metabolism, leading to a decrease in GABA efficiency by bacteria 59 . Although the optimal concentrations of MSG are different for various microorganisms in GABA production, some researchers have proven that excessive MSG could inhibit cell growth and minimize GABA production 60,61 . Lu et al. 62 showed that the highest concentration of GABA production by Lac. lactis isolated from kimchi was at 34° C (3.68 g/L). These results are similar to our results about Lac. lactis 311. However, they stated that the optimal pH of GABA production was in the range of 7-8 62 , which was not in compliance with the results obtained in our study.

Conclusions
In the present study, firstly, LAB isolates were screened for GABA production according to a qualitative method based on a colorimetric assay. Significant variability in GABA production was encountered among the different species and strains tested. GABA-producing strains were then subjected to a TLC analysis to detect strains with high GABA producing ability. Response surface methodology was finally used to optimize conditions for maximum GABA production in two Lac. lactis strains, taking into account the variables affecting the response. The selected strains and the culture conditions producing the highest amounts of this bioactive compound could be used to develop GABA-enriched functional foods or in other biotechnological applications.