Garlic augments the functional and nutritional behavior of Doenjang, a traditional Korean fermented soybean paste

Three different forms of garlic, namely, fresh garlic (2%, 6%, 10%), heat-dried (1%, 2%, 3%) and freeze-dried (1%, 2%, 3%), were supplemented in soybean paste to prepare Doenjang and further evaluated for functional, nutritional and safety aspects. Results showed a considerable antioxidant and anti-proliferative activity of garlic-supplemented Doenjang. As a measure of nutritive value, a high amount of total free amino acids, 4,290.73 mg/100 g–5,492.94 mg/100 g, was observed in prepared Doenjang. Among all preparations, 3% freeze-dried garlic-supplemented Doenjang proved the most effective against gastric adenocarcinoma and lung adenocarcinoma with 50% inhibition concentration of 7.66 ± 0.53 mg/mL and 7.82 ± 0.34 mg/mL, respectively. However 10% fresh-garlicsupplemented Doenjang (GGD-10) showed better activity against colorectal adenocarcinoma (HT29) cell line. Furthermore, GGD-10 effectively reduced colony formation and altered mitochondrial membrane potential of HT29 cells. Absence of pathogenic bacteria (Staphylococcus aureus, Salmonella species and Bacillus cereus) and aflatoxin was observed in Doenjang samples. In addition, nontoxic amount of anti-nutritional biogenic amines was observed in all the samples. The results collectively suggest that the addition of garlic in Doenjang can improve its nutritional and functional value, resulting in the protection of consumers from protein deficiencies and various stress conditions.


Results and Discussion
physiochemical analysis of the prepared Doenjang. Garlic is a storehouse of many important phytochemicals, mainly organosulfur compounds (allicin, alliin, S-allylcysteine or S-allyl mercaptocysteine), and has multiple health benefits 13 . Several clinical studies suggested its apparent role in hepatoprotection, immunomodulation, atherosclerosis, lipid metabolism, blood pressure maintenance and regulation of diabetes and cardiovascular diseases 13,14 . In the present study, nine different types of Doenjang were prepared using three forms of garlic (Table 1, Supplementary Fig. S1) and examined for its physicochemical analysis. Initially, pH was examined, which is considered as the most important factor of the fermented foods and vital to commence many biochemical activities. It not only provides a lucrative environment for many biochemical reactions but also acts as a barrier for many microbial contamination 6 . Results revealed the slight acidic pH in all Doenjang samples between 5.17 ± 0.04 and 6.12 ± 0.02 (Table 2). Highest acidic pH was observed in FGD-2 (5.17 ± 0.04), while the  Table 2. Physiochemical analysis and surface color value of Doenjang samples prepared with various form of garlic. Each value represents mean ± SD of three independent experiments. L* represents lightness in the scale 0 to 100 (black: 0, white: 100), a* represents redness (green: −60, red: 60) and b* represents yellowness (blue: −60, yellow: 60). Mean values with different lower case letters (a-f) in a column are significantly different (P < 0.05).
least was in the GGD-2 (6.12 ± 0.02). All the garlic-supplemented samples, except GGD-2 and HGD-1, have a slightly lower pH as compared to control. It is well established fact that acidic pH is an indication of microbial fermentation; hence, results insight that the decreased pH in garlic-supplemented samples is possibly because of good fermentation of soybean paste. The moderately acidic pH in the Doenjang samples attests it safe for consume, contrary to the excessively low pH which adversely affects health by causing acidosis. The results are consistent with previous findings, signifying the role of starter culture and alterations in processing conditions and ingredients on the pH of prepared Doenjang 6,15 . In addition to pH, salinity, ash and moisture content are important parameters to determine the fermented food quality. Sodium chloride is the most common food additive that not only acts as a food preservative but also improves taste, texture and shelf-life 16 . We observed salinity in the range of 12.97 ± 0.98%-14.36 ± 1.37% (Table 2). Minimum salinity was observed in HGD-2 (12.97 ± 0.98%), whereas maximum in GGD-2 (14.36 ± 1.37%). A negative correlation between salinity and increasing amount of fresh garlic (y = −0.06x + 14.44, r 2 = 0.91) was observed in the prepared Doenjang.
We further evaluated the moisture content, which is directly associated with texture, appearance, weight and shelf-life of foodstuff, along with a supportive role in microbial growth that accelerates the fermentation. Highest moisture content was observed in the control (62.07 ± 5.36%), whereas the lowest in HGD-1 (55.09 ± 3.22%) ( Table 2). Also, a positive correlation was observed in the moisture content and amount of fresh ground garlic added Doenjang (y = 0.15x + 59.03, r 2 = 0.97).
Next, we determined the ash content, which reflects the mineral content of the sample. The ash mainly consists of K, Na, Mg and Ca, along with trace amounts of Al, Fe, Cu, Mn and Zn. All these minerals have a significant role in human development and also act as cofactor of many proteins 17 and thus are very important to supply from the foods. We observed the ash content in the range of 13.27 ± 0.17%-14.21 ± 0.19% (Table 2). Maximum ash content was detected in HGD-1 (14.21 ± 0.19%), whereas the minimum content was in GGD-10 (13.27 ± 0.17%).
Food color is an important factor that determines its appearance and overall consumer acceptability. Surface color of all samples was determined using a colorimeter; minimum lightness (L*) value was reported for HGD-2 (35.30 ± 0.89) and the maximum for FGD-2 (39.54 ± 0.49), whereas the control group showed L* of 38.28 ± 0.65 (Table 2). In contrast, maximum and minimum redness (a*) values were noticed in GGD-2 (9.93 ± 0.81) and GGD-6 (5.32 ± 0.57) samples, respectively, and maximum and minimum yellowness (b*) values were noticed in GGD-2 (14.72 ± 1.44) and HGD-3 (6.79 ± 0.67), respectively. It is difficult to generalize the color change, however, variation of the color gives idea of the involvement of the diverse microbes that act on the complex biomolecules (starch and proteins) and degrade them into simple amino acids and sugars responsible for unique aroma, color and taste 6 . Results corroborated those of a recently published report on color change in Doenjang prepared using various starter cultures with varied ingredients 6 .
Determination of free amino acids. Total free amino acids in all prepared Doenjang were quantified, and the results are depicted in Table 3. An important role of amino acids in human metabolism as a building block of proteins, a stabilizer of nucleic acid and growth factor is well mentioned in the literature 18 . Besides the nutritive value, free amino acids are also responsible for color and taste of the food 6 . Owing to many benefits of amino acids, we quantified and characterized free amino acids in the prepared Doenjang and observed a high value between 4,290.73 mg/100 g and 5,492.94 mg/100 g (Table 3) (Supplementary Figs S2-S11). This high total free amino acid content is indicative of good fermentation and involvement of diverse microbial proteases, which is well supported by the results of previous reports, suggesting the importance of proteolysis during fermentation of soy products 6 .
We further characterized free amino acids as proteinogenic and non-proteinogenic and observed the highest proteinogenic amino acid content in FGD-3 (4,819.55 mg/100 g), followed by HGD-2 (4,535.26 mg/100 g) and the least in the control (3,713.24 mg/100 g). Furthermore, proteinogenic amino acids were regrouped into essential and non-essential amino acids. The outcome of the study suggests the higher essential amino acids in FGD-3 (2,125.19 mg/100 g), signifying its good health effect (Table 3). When compared with the control, FGD-3 showed 1.26-fold enhanced essential amino acid content. Among the all estimated essential amino acids, leucine was predominated with the highest value in FGD-3 (559.26 mg/100 g), followed by GGD-10 (556.96 mg/100 g), signifying 30.8% and 30.3% augmentation, respectively, as compared to control. Moreover, the overall high level of non-proteinogenic amino acids was observed in FGD-3 (673.39 mg/100 g), representing 1.16-fold higher values as compared to control (577.49 mg/100 g) ( Table 3).
Accumulating results suggest the indispensable biological role of non-proteinogenic amino acids, such as involvement of taurine with membrane stabilization, osmoregulation, modulation of Ca +2 signaling and essential of cardiovascular function, development and function of the central nervous system 19 . We detected taurine highest amount in FGD-3 (18.45 mg/100 g) and least in the FGD-2 (11.35 mg/100 g). In addition, we observed high level of β-aminoisobutyric acid which induces the expression of thermogenic genes, resulting in browning of white fat cells 20 , and is effective against obesity 21 . The maximum level of β-aminoisobutyric acid was detected in FGD-3 (248.37 mg/100 g), which was 1.3-fold higher than that in the control (187.83 mg/100 g) ( Table 3). Results collectively showed good amount of essential and non-proteinogenic amino acids in all the garlic-supplemented Doenjang and more precisely in FGD-3 and thus could be effective to overcome protein deficiency if consumed regularly.
We speculated that the deviation in the amount of total free amino acid in different samples is directly related with the variation of the raw material (different form and proportion of garlic) that leads fermentation through a diverse environment with different proteolytic events. The findings are supported by the previous report 6 representing the effect of various starter culture which leads to modulating the fermentation and diversification of free amino acids in Doenjang. However, unlike to this research, a lesser amount of proteinogenic amino acids was noticed in the various Doenjang samples 6 . Apart from nutritive value, amino acids are also responsible for the www.nature.com/scientificreports www.nature.com/scientificreports/ flavor to particular food material, therefore, we also categorized free amino acids based on their taste. Broadly amino acids can be classified into four major taste groups such as sweet taste (lysine, glycine, alanine, serine, threonine), umami taste (glutamic acid, aspartic acid, cysteine), bitter taste (isoleucine, leucine, methionine) and other taste (phenylalanine, valine, arginine, proline, tyrosine, histidine) 6 . We observed a high amount of total sweet taste amino acids in HGD-3 (1,311.25 mg/100 g), while least in GGD-2 (1,042.48 mg/100 g) ( Supplementary  Fig. S12). The highest amount for umami taste amino acids were detected in FGD-3 (1,339.43 mg/100 g), while least in the control (859.58 mg/100 g) ( Supplementary Fig. S12). Similarly, the maximum and minimum amount of bitter taste and other taste amino acids were detected in FGD-3 and control, respectively ( Supplementary  Fig. S12).
Estimation of phenolics, flavonoids and antioxidant potential of Doenjang. Diseases such as cancer, cataract, rheumatoid arthritis and neurodegenerative ailments such as Alzheimer's disease and Parkinson's disease are associated directly with the oxidative stress elevated by various reactive oxygen species (ROS) 22 . The cellular antioxidant machinery scavenges the ROS, although sometimes it is insufficient for scavenging it to the basal level particularly when large amounts of ROS are generated 22 . In such condition, exogenous antioxidants are required to scavenge the excessive ROS. As external antioxidant, herbs provides more promise due to its good efficiency and safety in nature unlike to synthetic compound. There are number of studies reporting the health beneficial potential of plants due to its high antioxidant activities. Phenolics and flavonoids are the important constituents of plants responsible for numerous biological activities including antioxidant activity. Moreover, a positive correlation of phenolics and flavonoids with antioxidant activities has been well documented 23 . A comparative total phenolics and flavonoid content in the Doenjang are depicted in Table 4, which revealed the higher phenolics content in the garlic-supplemented samples than in the control. The highest phenolics content was observed in the FGD-3 (35.14 ± 2.29 mg gallic acid equivalents (GAE)/g). Unlike phenolics, higher flavonoid content was observed in GGD-2 (4.25 ± 1.13 mg quercetin equivalents (QE)/g), followed by GGD-6 (4.17 ± 1.13 mg QE/g) and FGD-3 (4.14 ± 1.10 mg QE/g). Results are assisted with the number of reports presenting the diverse amount of phenolics and flavonoids in the soy fermented food 24,25 . In a recent study, variations in phenolic and flavonoid contents were reported in Doenjang formed by the addition of different plant materials 15 .

Doenjang samples
Control GGD-2 GGD-6 GGD-10 HGD-1 HGD-2 HGD-3 FGD-1 FGD-2 FGD-3  www.nature.com/scientificreports www.nature.com/scientificreports/ We further evaluated in vitro antioxidant activity by 2,2-diphenyl-1-picrylhydrazyl (DPPH), 2′-azino-bis (3-ethylbenzothiazoline-6-sulfonic acid) (ABTS), superoxide anion scavenging and ferric reducing antioxidant power (FRAP) assay. An antioxidant that vigorously donates hydrogen is considered as a potent scavenger of DPPH • with high antioxidant potential. We observed that HGD-2 was a strong scavenger of DPPH • with the least 50% inhibitory concentration (IC 50 ) value (20.39 ± 0.55 mg/mL) (Table 4). Further, the results of the ABTS •+ scavenging assay suggest that all the samples can effectively scavenge the cationic radical (ABTS •+ ). The best ABTS •+ scavenging activity was noted for FGD-3 with minimum IC 50 value (5.15 ± 0.67 mg/mL). Besides, all the Doenjang samples showed a great promise to scavenge superoxide anions by effectively reducing the oxidation of pyrogallol. However, the best activity was displayed by GGD-10 with a least IC 50 value (20.82 ± 0.83 mg/mL), followed by HGD-3 (23.11 ± 2.78 mg/mL). It has been well established that antioxidants can act as reductants by taking part in the redox reaction and thus have an effect on oxidants 26 . Therefore, we investigated the electron donating capacity of the Doenjang reflecting its reducing power using FRAP assay and observed a considerable activity of all the samples. Similar to DPPH assay, HGD-2 proved most efficient reducing agent, with 3.98 ± 0.20 mmol Fe +2 /g ferric reducing capacity (Table 4). Results are in good agreement with the published reports showing the antioxidant potential of Doenjang prepared by the defined starter culture 27 .
Anti-proliferative activity of Doenjang samples. Cancer is the most fatal disease increasing worldwide at a rapid pace. The reasons underlying the rapid progression include lifestyle, poor diet and obesity 28,29 . Therefore, the consumption of balanced hygienic food is critical as a preventive measure against cancer. A foodstuff that not only provides the essential nutrients but also has some anticancer property is much of interest as a preventive agents. Hence, we evaluated the anticancer activity of Doenjang against three different cancerous cell lines and observed a substantial effect (Table 5). FGD-3 showed the highest cytotoxicity against gastric adenocarcinoma (AGS) and human lung adenocarcinoma (A549) cell lines as evident by the lower IC 50 value (7.66 ± 0.53 mg/mL and 7.82 ± 0.34 mg/mL, respectively), which was 1.5 and 1.8-fold lower as compared to the control (Table 5). Interestingly GGD-10 showed the highest cytotoxicity against colorectal adenocarcinoma (HT29) with the least IC 50 value (10.98 ± 1.59 mg/mL). We believed that the anti-proliferative potential of the Doenjang is due to the presence of diverse phytochemicals in it. Importantly, phenolics and flavonoids are well known for imparting many biological activities including anticancer activity 30,31 . Results are in agreement with the published report suggesting the anticancer and antimetastatic properties of the Doenjang 32 . All garlic-supplemented Doenjang showed better activity than the control, indicating a critical role of garlic in anticancer activity, which also corroborates the results of a previous study demonstrating a potential role of garlic against various cancer cells 33 . Fresh garlic-supplemented sample (GGD-10) proved most effective against colon cancer cells (HT29) and thus selected as a representative to evaluate the clonogenic assay and the effect on mitochondrial membrane potential. Finding of the clonogenic assay revealed a significant (P < 0.05) dose-dependent reduction of HT29 cell proliferation and differentiation into colonies (Fig. 1a,b).
Mitochondria play an important role in the cancer cell survival and death. During the stress environment, mitochondrial membrane potential changes lead to opening of mitochondrial permeability transition pore and facilitate the release of cytochrome C which activates caspase cascade and ultimately induces cellular events responsible for cell death 34 . Results of mitochondrial membrane potential in response to GGD-10 are depicted in the Fig. 1c,d. GGD-10 has a significant (P < 0.05) effect on the mitochondrial membrane potential of HT29 evident by a dose-dependent reduced fluorescent intensity. Minimum fluorescent intensity (54.9 ± 5.6%) was observed in the cells treated with 7.5 mg/mL GGD-10 in contrast to 100% in control (without any treatment). The results collectively highlight the anti-proliferative potential of Doenjang, principally mediated by altering mitochondrial membrane potential and subsequently downstream events. In support to our findings, a study carried  www.nature.com/scientificreports www.nature.com/scientificreports/ out by Jeong et al. 35 revealed the role of Doenjang prepared by various starter culture on the up and down regulation of pro-apoptotic (Bax) and anti-apoptotic (Bcl2), respectively, in a murine model of colitis-associated colon carcinogenesis. In addition, several published reports have indicated the role of garlic in induction of apoptosis via various molecular mechanisms 33,36 . However, to the best of our knowledge, it is the first report representing the effect of garlic-supplemented Doenjang on alterations in mitochondrial membrane potential and ability on colony formation.
Quantification of biogenic amines. Safety of the food products is a primary concern before its consumption, thus evaluation of toxicants and fouling agents are the utmost steps to examine the food quality 6 . Biogenic amines are formed by microbial decarboxylation of amino acids, particularly in nutritionally poor environment 37 . The presence of excessive biogenic amines in food is an indirect indicator of microbial contamination 38 . In addition, it is also the precursor of carcinogens such as N-nitrosamine compounds 39 . The role of biogenic amines such as tyramine, 2-phenyethylamine and putrescine to induce migration, hypersensitivity, high blood pressure and  www.nature.com/scientificreports www.nature.com/scientificreports/ heart failure has been well cited in the literature 12 . The high amount of the biogenic amines such as tyramines and histamine in food are considered as the anti-nutritional elements 38 . Therefore, we determined the amount of biogenic amines such as agmatine, tryptamine, 2-phenylethylamine putrescine, cadaverine, histamine, tyramine, spermidine and spermine in Doenjang samples at the starting day and after two and four months of fermentation. The initial amount of biogenic amines in various samples was in the range of 25.36 mg/100 g-49.85 mg/100 g (Supplementary Table S1), which was decreased in most of the samples (other than HGD-2 and FGD-1) after two months of fermentation (Table 6). After two months of fermentation, the least amount of total biogenic amines was detected in FGD-3 (28.61 mg/100 g) and the highest in HGD-3 (48.22 mg/100 g) ( Table 6). A negative correlation between total biogenic amines and addition of freeze-dried garlic (y + . = . y x r 11 24 12 16, 098 2 ) was observed in Doenjang prepared by mixing various proportions of heat-dried garlic. Among the biogenic amines, histamine is considered as the most important with a noteworthy role in physiological and pathological conditions 40 . In food, histamine is considered as a serious toxicant leading to histamine poisoning (scombroid) with a symptom of illness, dizziness, headache, oral burning and sweating 41 . We observed the least value of histamine in the HGD-1 (6.34 mg/100 g) ( Table 6). In all garlic-supplemented Doenjang other than GGD-2 (20.28 mg/100 g), the histamine level was below the amount quantified in the control (18.78 mg/100 g) ( Table 6). In all samples, histamine content was considerably below its toxicity level of 500 mg/kg, at which it induces severe adverse effects on human health 12 . Interestingly, we did not observe any presence of agmatine in any of the prepared Doenjang. When compared with two month fermentation, level of biogenic amine was slightly increased in most of the garlic-supplemented Doenjang after four months of fermentation (Supplementary Table S2). However, the amount is below the level of toxicity recommended for the foods 12 . Our results are consistent with the previous published report exhibiting the biogenic amine level in the different Doenjang samples 39 , however, unlike to the report, we noticed the low level of biogenic amines, suggesting their safety to consume.

Determination of aflatoxin.
Aflatoxins are highly toxic secondary metabolites produced by fungal species and considered as the most potent food contaminants causing severe threat to human and animals 42 . Moreover, aflatoxin plays an active role in inducing cancer 43 . As the guidelines of Korea Food and Drug Administration, 10 ppb aflatoxin in food samples is considered hazardous 44 . More than 20 types of aflatoxin have been identified, the predominant among which is aflatoxin B1, which is activated in the liver by the action of cytochrome p450 and converted into aflatoxin B1-8, 9-epoxide, a potent renal carcinogen 42 . The main sources of aflatoxin poisoning are the intake of the contaminated food, therefore, a food free from aflatoxin proves its safety towards consumers. We determined aflatoxin level in Doenjang samples using Neogen Veratox kit, with 5 ppb as the minimum limit of quantification. The outcome of the study revealed an absence of aflatoxin in all the Doenjang samples and attests the safety of Doenjang to consume (Table 7).

Bacterial enumeration in Doenjang. Presence of microbes in the fermented food is very important to
decide the fate of the product. In the present study, consortia of microbes were evolved in the soybean paste that finally fermented into the Doenjang. Results showed the presence of high aerobic bacterial count between 6.22 ± 1.62 log CFU/g and 7.61 ± 1.82 log CFU/g (Table 7) in the fermented product. Interestingly, higher aerobic bacterial count was noticed in garlic-supplemented Doenjang than in the control. In all the prepared Doenjang, the highest aerobic bacterial count was observed in the GGD-2 (7.61 ± 1.82 log CFU/g), which was approximately 1.2-times higher than that in the control (6.22 ± 1.62 log CFU/g) ( Table 7). The high number of aerobic bacteria in Doenjang is an index of good fermentation, which leads to conversion of complex biomolecules into simple and vital phytochemicals responsible for imparting better metabolic profile to Doenjang. Results are in agreement with the previous observations regarding the involvement of various microbes in Doenjang formation 6,11 .
There is always being a high chance of pathogenic bacteria contamination in the complexed food material when stored for a long time. Common foodborne pathogens (Escherichia coli, Bacillus cereus, Salmonella species and Staphylococcus aureus) account for huge economic and health loss every year. The severity of Salmonella infection can be assumed by the fact that millions of cases of gastroenteritis are reported worldwide every year, primarily by the consumption of the contaminated food 45 . To access the pathogen-free preparation of Doenjang,   Table 6. Quantification of biogenic amines in Doenjang samples prepared with various combinations of garlic after 2 months of fermentation (mg/100 g).
www.nature.com/scientificreports www.nature.com/scientificreports/ we examined the Salmonella species and S. aureus in it and found the absence of these important pathogens. There are many possible reasons for the absence of these bacteria in Doenjang. One common reason is the acidic pH that acts as a barrier to the survival of these pathogens 6 . However, pathogens such as B. cereus can grow well in acidic environment and is considered one of the most serious pathogens in fermented food, where the level larger than 5.0 log CFU/g in food potentially causes food poisoning 46 . In our study, we observed 1.0 ± 0.0 log CFU/g in HGD-1, HGD-2 and FGD-3, which are well below the toxic level (5.0 log CFU/g). The remaining samples, including the control, were free from the B. cereus contamination ( Table 7). The finding suggests a pathogen-free preparation of Doenjang that could be used as a safe food.

Conclusion
Based on the findings, we conclude that garlic has a positive impact on Doenjang by enhancing its nutritional and functional properties. Garlic-supplemented Doenjang demonstrated an enhanced level of phenolics, flavonoid and antioxidant along with the substantial anticancer activity. A high level of total free amino acids in garlic-supplemented Doenjang makes it nutritionally more valuable and can be used as a source of functional food. Moreover, a lower level of antinutritional biogenic amines and absence of aflatoxin suggests that it is safe to consume. Although there are many reports of Doenjang preparation using different starter culture and supplementation of phytochemicals, to the best of our knowledge, it is the first comprehensive study suggesting Doenjang preparation augmented with garlic.

Materials and Methods
Chemicals and reagents. All chemicals, unless otherwise stated, were of the highest quality and were used as supplied. All the standard biogenic amines (putrescine dihydrochloride, agmatine sulfate, histamine dihydrochloride, 2-phenylethylamine, tyramine hydrochloride, spermine tetrahydrochloride, tryptamine hydrochloride, spermidine trihydrochloride and cadaverine dihydrochloride) were purchased from Sigma-Aldrich (St. Louis, MO, USA). High-performance liquid chromatography (HPLC) grade ammonium acetate and acetonitrile were purchased from Merck (Damstadt, Germany). Doenjang preparation. Soybean seeds were soaked for 12 h in water, followed by boiling for 1.5 h. The cooked seeds were ground to make an even soybean paste that were eventually transferred into stainless steel square plates (25 cm length × 20 cm width × 10 cm height) to make bricks for the formation of Meju. The soybean bricks were incubated at ambient temperature (~28 °C) for 14 days to allow growth of natural microflora (Meju). The soybean bricks (Meju) were divided into four different groups, (i) soybean bricks mixed with 2%, 6% and 10% (w/w) paste of fresh garlic, (ii) and (iii) soybean bricks mixed with 1%, 2% and 3% (w/w) heat-dried and freeze-dried garlic, respectively, and (iv) soybean bricks without any garlic mix (control) ( Table 1). Finally, 18% salt solution was added to each group and incubated at room temperature (~28 °C) for two months. After that liquid portion was separated, the solid part was further incubated for four months at ambient temperature (~28 °C) ( Supplementary Fig. S1). The experiment was performed in triplicate for each group. pH measurement. The pH of the Doenjang samples was determined using a pre-calibrated pH meter per the method described by Shukla et al. 39 . In brief, 10 g Doenjang samples were homogenized in 100 mL distilled water, followed by filtration with Whatman paper No. 2 (Advantec; Tokyo, Japan). Finally, the pH of the filtered samples was measured using a pH meter (Thermo Electron Corporation; Beverly, MA, USA).
Moisture, salt and ash content. The moisture content in the Doenjang samples was determined by a gravimetric method by placing 2 g sample into the preheated oven at 105 ± 2 °C until constant weight was achieved according to standard method of Association of Official Analytical Chemists 47 .
Salt content in the Doenjang samples was determined according to the method of the Association of Official Analytical Chemists 47 , and the results were expressed as percentage salinity.

3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) assay.
Anti-proliferative potential of Doenjang samples was evaluated by MTT assay against three different human cancer cell lines of AGS, A549 and HT29. All the three different cancer cell lines were seeded individually in separate 96-well cell culture plates at the density of 1 × 10 4 cells/well and incubated for 24 h. After incubation, cells were treated individually with each aqueous extracted Doenjang samples at different concentrations (0.0 mg/mL, 0.01 mg/mL, 0.1 mg/mL, 1.0 mg/mL and 10.0 mg/mL) and re-incubated for 24 h. After incubation, cells were treated with MTT solution (0.5 mg/mL) to produce dark blue color formazan, which was subsequently dissolved in the dimethyl sulfoxide. Finally, absorbance was measured at 540 nm in a microplate reader (Tecan; Mannedorf, Switzerland), and results were expressed as percent viable cells.
Clonogenic assay. For clonogenic assay, 8 × 10 2 HT29 cells/well were seeded in 12 well-culture plate, followed by the treatment of various concentrations (0.0 mg/mL, 2.5 mg/mL, 5 mg/mL and 7.5 mg/mL) of 3% freeze-dried garlic added Doenjang. Cells were allowed to grow for 7 days in a CO 2 incubator at 37 °C. Finally, media were aspirated and cells (clones) were washed with 0.1 M phosphate buffer saline (PBS) (pH 7.4) and stained with 0.5% crystal violet.
Mitochondrial membrane potential. Effect of Doenjang samples on the mitochondrial membrane potential of HT29 was evaluated by a cell permeable green fluorescent Rhodamine 123 dye 34  www.nature.com/scientificreports www.nature.com/scientificreports/ Determination of biogenic amines. Extraction of biogenic amines. Biogenic amines were extracted from Doenjang samples by the method adopted by Shukla et al. 39 . In brief, 5 g Doenjang samples were homogenized with 10 mL of 0.4 M perchloric acid, followed by centrifugation at 4,000 × g for 10 min at 4 °C. The supernatant was recovered, and the pellet was re-extracted twice with 0.4 M perchloric acid, followed by centrifugation. All the collected supernatants were polled, filtered through Whatman paper No. 1 and used for HPLC analysis after derivatization with dansyl chloride.
Derivatization of extracted biogenic amines. Derivatization of extracted biogenic amines was carried out by adopting the previously described method 39 . In brief, 1 mL of each extracted biogenic amine suspension was mixed with 200 µL of 2 M NaOH, 300 µL of NaHCO 3 (saturated solution) and 2 mL of 0.04 M dansyl chloride, followed by 45 min incubation at 40 °C. After the incubation, 100 µL of 25% NH 4 OH was added to stop the reaction, and removal of excess dansyl chloride was succeeded by the addition of acetonitrile to make the volume up to 5 mL. Finally, the mixture was centrifuged at 2,500 × g for 5 min and filtered through 0.2 µm syringe filter (Woongki Science Co.; Seoul, Korea). The filtered samples were stored at −25 °C until use for HPLC.
Separation and quantification of biogenic amines by HPLC. Separation and quantification of derivatized biogenic amines were carried out using HPLC analysis (Hewlett Packard series-1100, Agilent Technologies; Santa Clara, CA, USA) equipped with a UV-visible spectrophotometer as a detector. Separation was carried out by injecting 10 µL sample into a C 18 column (5 µm pore size, 4.6 mm internal diameter, 250 mm length) using 0.1 M ammonium acetate and acetonitrile as mobile phase with gradient programming up to 35 min. Separation was commenced at 30 °C with a constant mobile phase flow rate of 1 mL/min. Separated biogenic amines were analyzed at 254 nm. Biogenic amines in the samples were identified and quantified by comparing with the known standards.

Determination of aflatoxin.
Aflatoxin was extracted and quantified from Doenjang samples as per the instructions of the manufacturer (Neogen Corp.; Lansing, MI, USA), supplied with ELISA-based aflatoxin detection kit (Veratox for total aflatoxin). In brief, 10 g sample was homogenized with 50 mL of 70% methanol, followed by filtration with Whatman filter paper No. 2. The filtrate was directly used to detect aflatoxin using ELISA kit. After sufficient antigen (aflatoxin)-antibody interaction, chromogenic substrate was added, and then an absorbance of the colored product was measured at 650 nm using microplate reader (Infinite M200, Tecan; Mannedorf, Switzerland). Quantification of aflatoxin was made by Neogen's Veratox data reduction software by converting absorbance into concentration against reference standards. enumeration of aerobic bacteria in Doenjang samples. Presence of total aerobic bacteria was examined by homogenizing 10 g Doenjang sample in 90 mL sterilized saline (0.85% NaCl). The suspension was serially diluted up to 10 −7 , and then the100 µL of each dilution was spread over the plate count agar medium (Oxoid; Hampshire, England) and incubated at 30 °C for 24 h. In addition, the presence of common food pathogens (B. cereus, S. aureus and Salmonella species) was also determined.
Detection of Bacillus cereus. Ten gram sample was homogenized into 90 mL sterilized saline (0.85% NaCl) and serially diluted up to 10 −5 . One hundred microliter of each dilution was spread over mannitol-egg yolk-polymyxin agar plate (MYP) (Oxoid; Hampshire, England) and incubated at 30 °C for 24 h. Presence of B. cereus was evaluated by the appearance of characteristic pink colonies with transparent margins. Confirmatory test was carried out by streaking the positive isolates from MYP medium into nutrient agar plate, followed by identification using API 50 CHB kit (bioMerieux; Marcy I' Etoile, France). To exclude any rare possibility of the isolate as B. thuringiensis, a microscopic examination for crystal proteins have been done as per the method suggested by USFDA 54 .
Detection of Salmonella species. For detection of Salmonella species, 10 g sample was homogenized in 90 mL peptone water and incubated at 35 °C for 18 h. Further, 100 µL suspension was transferred into 10 mL of Rappaport Vassiliadis broth (Merck; Darmstadt, Germany) and incubated at 42 °C for 24 h. At last, 100 µL suspension was spread over MacConkey agar (Becton, Dickinson and Company; Sparks, MD, USA) plate and incubated at 35 °C for 24 h. Presence of Salmonella species were detected by the appearance of characteristic colorless colonies.
Detection of Staphylococcus aureus. Ten gram sample was homogenized in 90 mL tryptic soy broth (Becton, Dickinson and Company) with 10% NaCl and incubated at 35 °C for 16 h. Finally, 100 µL suspension was spreaded over Baird Parker agar plate (Oxoid; Hampshire, England) supplemented with egg yolk-tellurite emulsion (Oxoid). The plates were incubated at 35 °C for 24 h. Characteristic black color colonies were examined for the presence of S. aureus. statistical analysis. All experiments were carried out in triplicates. The results were expressed as mean ± standard deviation of three independent experiments. Multiple comparisons were performed using one-way analysis of variance, followed by Duncan test for post hoc analysis using SPSS16 software (SPSS Inc.; Chicago, IL, USA). P < 0.05 were considered statistically significant.

Data Availability
The authors declare that all the data supporting the finding of this study are available within the article and from the corresponding author on reasonable request.