To assess the ability of human intestinally derived strains of Lactobacillus and Bifidobacterium to produce γ-aminobutyric acid (GABA).
To assess the ability of human intestinally derived strains of Lactobacillus and Bifidobacterium to produce γ-aminobutyric acid (GABA).
Strains of Lactobacillus and Bifidobacterium were grown in medium containing monosodium glutamate (MSG). Growth of the bacteria and conversion of MSG to GABA were measured. Of 91 intestinally derived bacteria assessed, one Lactobacillus strain and four strains of Bifidobacterium produced GABA. Lactobacillus brevis DPC6108 was the most efficient of the strains tested, converting up to 100% of MSG to GABA. The ability of the cultured intestinal strains to produce GABA was investigated using a simple pH-controlled anaerobic faeces-based fermentation, supplemented with 30 mg ml−1 MSG. The addition of Lact. brevis DPC6108 to a faeces-based fermentation significantly increased the GABA concentration (P < 0·001), supporting the notion that this biosynthesis could occur in vivo.
The production of GABA by bifidobacteria exhibited considerable interspecies variation. Lactobacillus brevis and Bifidobacterium dentium were the most efficient GABA producers among the range of strains tested. The addition of Lact. brevis DPC6108 to the culturable gut microbiota increased the GABA concentration in fermented faecal slurry at physiological pH.
Identification of optimal MSG conversion to GABA by particular cultured elements of the commensal intestinal microbiota and the demonstration that this can occur under simulated in vivo conditions offer new prospects for microbiota modulation to promote health.
γ-Aminobutyric acid (GABA) is a major inhibitory neurotransmitter of the vertebrate central nervous system (Roberts and Frankel 1950) and is the main inhibitory neurotransmitter in the brain, found in relatively large amounts, regulating many physiological and psychological processes, with dysfunctions in the GABA system implicated in anxiety and depression (Cryan and Kaupmann 2005; Schousboe and Waagepetersen 2007). It is involved in the regulation of cardiovascular conditions such as blood pressure and heart rate, and plays a role in the sensation of pain and anxiety (Mody et al. 1994). A number of further potential health benefits of GABA have been described, including control of growth hormone secretion (Volpi et al. 1997), protective effect against glycerol-induced acute renal failure in rats (Kim et al. 2004) and anti-proliferative activity (a reduction of the induced migratory activity of SW480 colon carcinoma cells (Joseph et al. 2002)).
Interest in the potential role of GABA as an anti-hypertensive dietary component has recently increased in Japan. This was due in particular to the high sodium intake in the diet in that country, where the daily intake of salt was estimated to be 11·2 g day−1 in 2003 (Yamakoshi et al. 2007). Interestingly, foods enriched with GABA have been defined as ‘foods for specified health use’ (FOSHU) in Japan (Sanders 1998). This led to increased research into the development of fermented products containing GABA. Inoue et al. (2003) developed a fermented milk product containing GABA, which was shown to have a blood pressure-lowering effect in spontaneously hypertensive rats (Hayakawa et al. 2004) and in mildly hypertensive people (Inoue et al. 2003). Recent years have also witnessed an increase in the development of additional GABA-enriched fermented dairy, soybean, kimchi and juice products, which could be used as potential GABA-delivery vehicles (Park and Oh 2007; Seok et al. 2008; Chang et al. 2009; Kim et al. 2009).
In parallel with the heightened interest in the possible physiological effects following the consumption of GABA-enriched fermented food, there was a concomitant surge in the isolation of novel fermented food-derived or dairy-starter cultures with the ability to produce GABA (Siragusa et al. 2007; Hiraga et al. 2008; Komatsuzaki et al. 2008; Li et al. 2008). Lactic acid bacteria (LAB) from food sources have been shown to have the ability to produce GABA (Ueno et al. 1997; Nomura et al. 1998, 1999a,b; Siragusa et al. 2007). GABA is produced primarily from the irreversible α-decarboxylation of L-glutamate by the enzyme glutamate decarboxylase (GAD), a pyridoxal 5′-phosphate (PLP)-dependant enzyme. It has been found in animals, higher plants (Ueno 2000) and bacteria, where it plays a role in acid resistance (Cotter et al. 2001). Many studies have reported the presence of a gad gene in LAB (Nomura et al. 1998, 1999a; Siragusa et al. 2007; Hiraga et al. 2008; Komatsuzaki et al. 2008) and indeed, GABA has been produced by cheese starters during cheese ripening (Nomura et al. 1998; Siragusa et al. 2007). It is clear that one approach to increasing GABA levels in humans is by consuming GABA-enriched food products. However, another potential way of increasing GABA levels in the gut could be by harnessing the production ability of the intestinal microbiota or ingested probiotic bacteria that could use dietary monosodium glutamate (MSG) to generate GABA (Lyte 2011).
The purpose of this study was therefore to screen human intestinally derived bifidobacteria and lactobacilli for the ability to produce GABA, as a first step in investigating the latter strategy.
The human-derived strains used in this study were maintained in the Teagasc Moorepark Food Research Centre culture collection. Strains of lactobacilli and bifidobacteria were cultured in MRS broth (Difco, Detroit, MI, USA) supplemented with 0·05% (w/v) l-cysteine-hydrochloride (mMRS) (98% pure; Sigma Chemical Co., St Louis, MO, USA) under anaerobic (anaerobic jars with Anaerocult®A gas packs; Merck, Darmstadt, Germany) conditions at 37°C. When solid medium was required, 1·5% (w/v) agar (Oxoid, Hampshire, UK) was added to the mMRS medium. Standard cultures were prepared by the inoculation of 10 ml mMRS broth with 10 µl of a frozen stock (−80°C) followed by incubation at 37°C for 16–24 h. Strains were then subcultured in 10 ml mMRS broth for 16–24 h at 37°C prior to inoculation into the fermentation vessel.
Prior to the examination of the strains for GABA production, each strain was subcultured in mMRS. Strains were inoculated [1% (v/v)] anaerobically in mMRS broth supplemented with 30 mg ml−1 MSG (Sigma) at 37°C for 72 h. One ml of the culture broth was centrifuged at 14 000 g for 1 min and the level of GABA in the culture supernatant was determined by analysing the free amino acid content as described below.
The growth of Lact. brevis DPC6108 was monitored in mMRS supplemented with MSG (30 mg ml−1). The strain was inoculated from a fresh overnight culture to a final cell density of 106 CFU ml−1 and the culture was grown anaerobically at 37°C. Bacterial growth [cell counts on mMRS agar, following serial dilution in maximum recovery diluent (MRD), at 37°C for 48 h] and conversion of MSG to GABA (determined by analysing the free amino acid content as described below) were monitored in triplicate at regular time intervals over 55 h.
The faecal fermentation medium was prepared as previously described (Fooks and Gibson 2003), with a single modification which involved the addition of 30 mg ml−1 MSG. The medium was allowed to cool overnight at room temperature following sterilization and 160 ml was transferred to a fermentation vessel and flushed with nitrogen for 30 min prior to inoculation. Fresh faecal samples were obtained from three healthy adults, who had not been prescribed antibiotics in the previous 3 months. However, levels of MSG intake in their diet in the days before sampling were unknown. Fresh faecal slurry [20% (w/v)] was prepared in anaerobic MRD containing 0·05% (w/v) cysteine. The anaerobic fermentation medium was inoculated with 40 ml of fresh faecal slurry in each fermentation vessel. To determine the effect of spiking the fermentation medium with GABA-producing human-derived bacteria, Lact. brevis DPC6108 was grown anaerobically at 37°C in mMRS supplemented with 30 mg ml−1 MSG for 48 h. Cells were harvested by centrifugation at 4000 g for 10 min, washed once with anaerobic MRD containing 0·05% (w/v) cysteine and inoculated into the faecal fermentation slurry to give an initial number of c. 108 cells ml−1 in the fermentation vessel. Fermentations were conducted at 37°C under nitrogen and maintained at pH 6·8 for 24 h. Samples were withdrawn at 0, 4, 9, 21 and 24 h for GABA analysis. Faecal fermentations were carried out in duplicate.
Samples were deproteinized by mixing equal volumes of 24% (w/v) trichloroacetic acid (TCA) and sample. They were then allowed to stand for 10 min before centrifuging at 14 000 g (Microcentaur; MSE, London, UK) for 10 min. Supernatants were removed and diluted with 0·2 mol l−1 sodium citrate buffer, pH 2·2 to yield c. 250 nmol of each amino acid residue. Samples were then diluted 1 in 2 with the internal standard, norleucine, to give a final concentration of 125 nm ml−1. Amino acids were quantified using a Jeol JLC-500/V amino acid analyzer (Jeol Ltd, Garden City, Herts, UK) fitted with a Jeol Na+ high-performance cation exchange column.
Results in the text, tables and figures are presented as mean per group ± standard error of the mean (SEM). To assess whether differences between individuals were significant, data were analysed using one-way analysis of variance (anova) followed by post hoc Tukey's multiple comparison tests using GraphPad Prism ver. 4·0 for Windows (GraphPad Software, San Diego, CA, USA). Results were considered significant as follows: ***P < 0·001.
In this study, 91 human-derived lactobacilli and bifidobacteria strains, obtained from a number of previous studies and maintained in the Moorepark Food Research Centre culture collection, were assessed for the ability to generate GABA from MSG. The assessed strains belonged to the following species Lact. brevis (1), Lactobacillus casei (7), Lactobacillus gasseri (2), Lactobacillus reuteri (1), Lactobacillus rhamnosus (5), Lactobacillus ruminis (1), Lactobacillus salivarius (4), Bifidobacterium adolescentis (4), Bifidobacterium angulatum (1), Bifidobacterium bifidum (6), Bifidobacterium breve (16), Bifidobacterium catenulatum (2), Bif. dentium (2), Bifidobacterium gallicum (1), Bifidobacterium longum subsp. infantis (4), Bifidobacterium longum subsp. longum (27), Bifidobacterium lactis (4), Bifidobacterium scardovi (1), Bifidobacterium subtile (1) and Bifidobacterium thermophilum (1). Four human-derived Bifidobacterium and one Lactobacillus GABA-producing strains were thus identified as GABA producers. These were Bif. dentium DPC6333, Bif. dentium NCFB2243, Bif. infantis UCC35624, Bif. adolescentis DPC6044 and Lact. brevis DPC6108. The details of strains with the ability to produce GABA from MSG are shown in Table 1. The conversion abilities of these strains varied when grown in different concentrations of MSG. This screening programme indicated that there was also interspecies variation in the ability of bifidobacteria to produce GABA. Lactobacillus brevis DPC6108 was the most efficient of the strains tested. Lactobacillus brevis DPC6108 was capable of 100% conversion of 10 and 20 mg ml−1 MSG to 11·03 ± 0·18 and 20·47 ± 0·07 mg ml−1 GABA respectively, in mMRS. Increasing the MSG concentrations to 30, 40 or 50 mg ml−1 resulted in decreased percentage conversions to GABA of 94·4% (28·02 ± 0·13 mg ml−1), 75·98% (30·39 ± 0·32 mg ml−1) and 64·64% (32·32 ± 0·82 mg ml−1), respectively, by Lact. brevis DPC6108 (Table 1). The four remaining GABA-producing strains, Bif. dentium DPC6333, Bif. dentium NCFB2243, Bif. infantis UCC35624 and Bif. adolescentis DPC6044, also converted MSG to GABA (Table 1). Conversion efficiencies of MSG to GABA ranged from 22% (2·2 ± 0·25 mg ml−1) to 60·9% (6·09 ± 0·15 mg ml−1) and 15·86% (3·17 ± 0·23 mg ml−1) to 61·6% (12·32 ± 0·40 mg ml−1) in 10 and 20 mg ml−1 MSG in mMRS, respectively, by the 4 bifidobacterial strains. Similar to Lact. brevis DPC6108, increasing the MSG concentrations to 30, 40 or 50 mg ml−1 resulted in decreased percentage conversions of MSG to GABA (Table 1). All of the remaining 86 Bifidobacterium and Lactobacillus strains assayed in this study grew in mMRS containing 30 mg ml−1 MSG, but did not convert MSG to GABA at any significant level.
|Species||Strain source||GABA (mg ml−1) converted from increasing concentrations of MSG|
|10 (mg ml−1) MSG||20 (mg ml−1) MSG||30 (mg ml−1) MSG||40 (mg ml−1) MSG||50 (mg ml−1) MSG|
|Lactobacillus brevis DPC6108||Infant faeces||11.01 ± 0.18||20.47 ± 0.07||28.02 ± 0.13||30.39 ± 0.32||32.32 ± 0.82|
|Bifidobacterium adolescentis DPC6044||Infant faeces||2.2 ± 0.25||3.17 ± 0.23||2.76 ± 0.21||1.24 ± 0.06||3.07 ± 0.15|
|Bifidobacterium dentium DPC6333||Infant faeces||5.25 ± 0.19||7.69 ± 0.43||8.62 ± 0.50||5.54 ± 0.16||6.16 ± 0.27|
|Bifidobacterium dentium NFBC2243||Dental carries||6.09 ± 0.15||12.32 ± 0.40||12.48 ± 0.35||5.68 ± 0.02||8.63 ± 0.40|
|Bifidobacterium infantis UCC35624||Ileal-caecal region||3.46 ± 0.03||3.26 ± 0.08||2.63 ± 0.13||5.68 ± 0.08||2.04 ± 0.07|
Lactobacillus brevis DPC6108 was the most efficient strain tested for the conversion of MSG to GABA. The growth of Lact. brevis DPC6108 in mMRS with 30 mg ml−1 was monitored over time (Fig. 1). Production of GABA was associated with the stationary phase of growth of the culture. The maximum cell count was achieved at 28 h and maximum conversion of MSG to GABA was at 55 h (Fig. 1).
The effect of 30 mg ml−1 MSG and c. 108 live Lact. brevis DPC6108 on the culturable gut microbiota was investigated using a simple pH-controlled anaerobic faeces-based fermentation (Fooks and Gibson 2003). Stool samples were obtained from three adult subjects, and faecal fermentations were performed in duplicate with each sample. Samples were withdrawn at 0, 4, 9, 21 and 24 h for GABA analysis. GABA was not detected in the faecal fermentation using the sample collected from subject number 1 at 0, 4, 9 or 21 h. It was however detected at 24 h, at a concentration of 0·56 μg ml−1 of faecal fermentation (Fig. 2). GABA was detected in a faecal-based fermentation using a sample collected from subject number 2. The GABA concentration reached a maximum at 4 h and subsequently decreased at 9, 21 and 24 h, with the concentrations of 5·82, 5·03, 1·75 and 0·94 μg ml−1, respectively (Fig. 2). GABA was also detected in a faecal-based fermentation using a sample collected from subject number 3. The maximum GABA concentration of 1·83 μg ml−1 was recorded at 0 h and the GABA concentration subsequently decreased at 4, 9, 21 and 24 h, with the concentrations of 0·39, 0, 0 and 0 μg ml−1, respectively (Fig. 2). Lactobacillus brevis DPC6108 was added to a faecal-based fermentation to examine the effect of adding a potentially probiotic c. 108 GABA-producing strain on a faecal fermentation at physiological pH. The faecal sample used to make the faecal slurry was obtained from subject 3. The GABA concentration increased to 66·25 μg ml−1 after 4 h (P < 0·001) and reached a maximum at 9 h of 70·72 μg ml−1 (P < 0·001) (Fig 2). Similar to other faecal-based fermentations, the GABA concentration decreased at 21 and 24 h to 4·18 and 0·47 μg ml−1, respectively (Fig. 2).
Bifidobacteria and lactobacilli are reported to produce a wide range of metabolites, which may act as a basis for probiotic function, including vitamins (Deguchi et al. 1985; Conly and Stein 1992; Hou et al. 2000; Crittenden et al. 2003), bacteriocins (Corr et al. 2007), exopolysaccharides (Cerning 1995; Hosono et al. 1997; Sreekumar and Hosono 1998) and conjugated linoleic acid (Ogawa et al. 2001; Oh et al. 2003; Coakley et al. 2003, 2006; Barrett et al. 2007; Rosberg-Cody et al. 2011). A further metabolite produced by intestinal lactobacilli and bifidobacteria that may also be linked to desirable host effects is GABA. It has been suggested that microbially produced GABA in the gut may have an effect on the brain–gut–microbiome axis (Bienenstock et al. 2010), an emerging concept in health said to be crucial for maintaining homoeostasis (Bonaz and Sabate 2009; Cryan and O'Mahony 2011). This study was undertaken to assess the ability of intestinally derived strains of lactobacilli and bifidobacteria to produce GABA from MSG. It would appear that the prevalence of GABA-producing lactobacilli and bifidobacteria in the human gastrointestinal tract (GIT) is not as widespread as it is among food-derived LAB, where GABA-producing LAB have been isolated from tempeh, fruit juices and fermented dairy and soy products (Higuchi et al. 1997; Nomura et al. 1998; Hou et al. 2000; Aoki et al. 2003; Inoue et al. 2003; Siragusa et al. 2007; Chang et al. 2009; Kim et al. 2009; Lim et al. 2009). Five of 91 human-derived lactobacilli and bifidobacteria strains had the ability to bioconvert MSG to GABA. In our study, Lact. brevis and Bif. dentium were the most efficient GABA producers. Lactobacillus brevis strains isolated from food sources have previously been reported to produce GABA (Ueno 2000; Park and Oh 2005; Hiraga et al. 2008; Kim et al. 2009); however, in this case, the Lact. brevis DPC6108 and the four bifidobacteria were all isolated from the human GIT.
In Listeria monocytogenes, GAD activity is critical for survival in acidic conditions and allows the producing bacterium to overcome the low pH stresses of fermented foods, gastric juice, volatile fatty acids in the GIT and low pH of the macrophage phagosome (Cotter and Hill 2003). Perhaps, the GAD enzyme will offer the same protection to Lact. brevis following exposure to low pH stress in the GIT, if for example the producing strain was orally administered as a probiotic. While the GAD system in different bacteria vary in molecular details and kinetics, generally glutamate uptake by a specific transporter is followed by the removal of an intracellular proton during glutamate decarboxylation. The resulting GABA is exported from the cell via an antiporter, increasing the value of the cytoplasmic pH, because of the removal of hydrogen ions, and also slightly increasing the extracellular pH because of the exchange of extracellular glutamate for GABA (Cotter and Hill 2003).
In addition to producing GABA in synthetic laboratory media, Lact. brevis DPC6108 was capable of producing GABA in faecal fermentations in the presence of culturable gut-derived bacteria. The levels produced in the pH-controlled faecal fermentations were far lower than in non-pH-controlled screening synthetic media, most likely due to the high pH (pH 6·8) of the faecal fermentations. The culturable bacteria from the three subjects were able to produce GABA at low levels; however, the level was significantly increased upon inclusion of Lact. brevis DPC6108 into the faecal fermentation (P < 0·001). It is clear that while some intestinally derived bacteria are capable of producing GABA, it is also utilized by intestinal bacteria. While it is apparent that bacteria of the gut microbiota can both produce and utilize GABA, in vivo studies are required to ascertain whether the host would benefit from the microbially produced GABA or if the GABA would be utilized by members of the gut microbiota.
A criterion of probiotic bacteria is that they confer beneficial effects on the host. It has been suggested that the in vivo action of GABA-producing intestinal bacteria, acting on dietary glutamate, could have potential health benefits (Mills et al. 2009; Lyte 2011). It is tempting to suggest that because of the beneficial health effects of GABA, the ability of strains of bifidobacteria and lactobacilli, natural inhabitants of the intestine, to convert MSG to GABA, may be considered as a novel probiotic trait. In addition, the discovery of lactobacilli and bifidobacteria with the ability to synthesize GABA may offer new opportunities in the design of improved health-promoting functional foods, with the benefits of enriched GABA and probiotic bacteria.
The technical assistance of Paula O'Connor is gratefully acknowledged. The authors are supported, in part, by Science Foundation Ireland, through The Alimentary Pharmabiotic Centre (APC), and the Irish Government under the National Development Plan 2000–2006.