https://orcid.org/0000-0003-0456-8544
Abstract
This study evaluated the effects of autochthonous starters on the quality characteristics of budu. The effects of individual or combined inoculation of Virgibacillus halodenitrificans PS21 and Staphylococcus simulans PMRS35 were compared to those of non-inoculation during 150 days of budu production. The starters directly affected characteristics of budu with browner colour and enhanced proteolysis and lipolysis as indicated by degree of hydrolysis and free fatty acid contents. The combined starter inoculation resulted in a high aspartic acid, glutamic acid and lysine levels. Additionally, key desirable volatiles, including 2-methylbutanal, 3-methylbutanal, benzaldehyde and 2-ethyl furan, were dominant in the starter-inoculated budu. In a sensorial analysis, the highest scores for appearance, colour, taste and flavour were assigned for inoculated budu with combined starters. Therefore, the addition of starters (V. halodenitrificans PS21 and S. simulans PMRS35) in budu production improved the quality characteristics and reduced the fermentation period.
Introduction
Budu is a high salt fermented fish sauce that is largely consumed as a food condiment or flavouring additive in many Thai cuisines. This product is brown or dark brown in colour with a salty and distinctive flavour (Mohamed et al., 2012). In Thailand, budu is made from small marine fish (mostly from Stolephorus spp.) mixed with salt (25–30%, w/w) and then naturally fermented under anaerobic conditions for 12–18 months (Kanjan & Sakpetch, 2020). During fermentation of budu, protein and lipid of raw material are hydrolysed into peptide, amino acids and free fatty acids by the action of enzymes from fish and halophilic bacteria (Lee et al., 2014). Therefore, the unique flavour of budu is to facilitate the proteolysis and lipolysis process. However, it is widely known that endogenous enzyme from fish always loses its activity under high salt environment. Additionally, the spontaneous fermentation occurring under uncontrolled environment may lead to the unstable of the final product or the loss of special flavour (Kim et al., 2016).
Currently, fermented fish sauces are produced with starter cultures to improve the flavour profiles and product quality (Anihouvi et al., 2012). The strains used as starter cultures should be selected from indigenous microorganisms because of their better adaptability (Lee et al., 2014). As previously reported, Bacillus sp., Lentibacillus sp., Halomonas sp., Virgibacillus sp., Staphylococcus sp., Micrococcus sp. and Tetragenococcus sp. are commonly found in fish sauce (Namwong et al., 2005). Most of them exhibited protease activity in high salt condition. Yongsawatdigul et al. (2007) isolated Virgibacillus sp. from fish sauce and applied these strains as starter cultures because they could produce various hydrolytic enzymes. In the same way, Udomsil et al. (2017) employed halophilic T. halophilus as starter cultures to improve flavour and speed of fermentation of Thai fish sauce.
Not only Virgibacillus and T. halophilus species, but also coagulase-negative staphylococci (CNS) were detected as predominant groups of bacteria in fermented fish products (Jeong et al., 2014; Udomsil et al., 2017). These staphylococci species could promote desirable volatile compounds via lipid and protein hydrolysis. Several studies employed coagulase-negative staphylococci in the dry fermented sausage to improve flavour and aroma of the products (Cachaldora et al., 2013). Jeong et al. (2014) reported that the key desirable aldehydes were higher in the sample inoculated with Staphylococcus equorum. In addition, Staphylococcus sp. CMC5-3-1 and CMS5-7-5 can reduce the faecal note contributed by dimethyl trisulfide and dimethyl disulphide in fish sauce (Udomsil et al., 2015). Hence, the addition of autochthonous starter on managing proteolysis, lipolysis and flavour generation during fermentation of budu is to be increased.
Therefore, the purpose of this study was to investigate the use of autochthonous Virgibacillus halodenitrificans PS21 and Staphylococcus simulans PMRS35 as a starter culture to improve the product quality and to accelerate the fermentation of budu.
Materials and methods
Preparation of starter cultures
Virgibacillus halodenitrificans PS21 and S. simulans PMRS35 previously isolated from naturally fermented budu were used as starter cultures. Virgibacillus halodenitrificans PS21 and S. simulans PMRS35 were thereafter cultivated in tryptic soy broth (TSB) and de Man Rogosa and Sharp (MRS), respectively, containing 10% NaCl and incubated at 35 °C for 3 days. Subsequently, the resting cells were collected by centrifugation at 8000 g for 15 min at 4 °C. After washed twice with sterilised phosphate buffer (PBS), the bacterial density of each strain was adjusted to 107 CFU mL−1.
Budu processing and sampling
Anchovy (1 kg) was mixed with 25% sea salt and 10% (v/w) of either V. halodenitrificans PS21 or S. simulans PMRS35 inoculum. The samples were coded as PS21 or PMRS35, respectively. For the combined starters, 5% of V. halodenitrificans PS21 and 5% of S. simulans PMRS35 were added to anchovy containing 25% sea salt. This combined starter inoculation was assigned as PS21+PMRS35. The control was prepared but did not receive any starter cultures. The samples were packed in earthen jars and then fermented at room temperature for 150 days. Sampling of 15 g was performed on days 0, 15, 30, 60, 90, 120 and 150 for microbial analysis.
Growth of proteolytic bacteria and staphylococci in budu samples
The samples (10 g) were mixed with 90 mL of sterile saline solution and then homogenised by stomacher for 2 min. The mixture of 0.1 mL with 10−1 to 10−5 dilutions was spread onto the skim milk agar for proteolytic bacteria and MRS agar with added 0.5% (w/v) CaCO3 for staphylococci, respectively. After incubation at 35 °C for 5–7 days, the number of viable cell counts was calculated and expressed as log10 CFU g−1.
Analysis of degree of hydrolysis (DH)
The samples fermented for 150 days were determined for degree of hydrolysis. An aliquot 1 g of each sample was homogenised with 9 mL of 5 % (w/v) SDS. Each mixture was heated at 85 °C for 30 min. After being centrifuged at 10 000 g for 10 min, the supernatant was analysed and DH was calculated and expressed as described by Benjakul & Morrissey (1997).
Analysis of amino acid profiles
Each sample (1 g) was homogenised using distilled water at a ratio of 1:10. Subsequently, the diluted samples (2 mL) were hydrolysed with 2 mL of 12 N HCl containing 1% (w/v) of phenol at 110 °C for 24 h using an autoclave. The acid was eliminated, and the precipitates were dissolved in deionised water and filtered through membrane filter (0.22 µm). Total amino acid was quantified by an amino acid analyser (Buckinghamshire, UK). The contents of amino acids were calculated and expressed as mg per 100 g sample (Yongsawatdigul et al., 2007).
Analysis of free fatty acids (FFAs) content
The amount of FFA in the samples was examined by titration in accordance with the method of Truong et al. (2016), with minor modification. The sample (2 g) was suspended in a solvent mixture of 25 mL diethyl ether, 25 mL ethanol (70%) and 1 mL of 1% phenolphthalein solution. After being homogenised at 12 000 g for 1 min, the solution was then slowly titrated using 0.1 N NaOH until the pink colour appeared. The FFA was calculated as described by Truong et al. (2016).
Analysis of Thiobarbituric acid reactive substances (TBARS) value
The sample (10 g) was homogenised with 50 mL of distilled water for 2 min. The homogenate was washed with 47.5 mL of distilled water in distillation flask. Subsequently, 2.5 mL of 4 N HCl was added to bring the pH to 1.5. After being distilled about 50 mL in 10 min, an aliquot of 5.0 mL of distillate was heated together with 5 mL TBA reagent for 35 min. The sample was collected by centrifugation, and the absorbance was measured at 532 nm. TBARS value was calculated and expressed as described by Pongsetkul et al. (2018).
Analysis of colour and browning intensity
The colour of samples was measured by a colorimeter (Hunter Lab, Reston, VA, USA) and expressed as L* (lightness), a* (redness/greenness) and b* (yellowness/blueness). Meanwhile, the browning was determined spectrophotometrically. The samples were diluted with deionised water, and the appropriate dilution was determined at 420 nm by UV-1601 spectrometer (Shimadzu, Kyoto, Japan).
Analysis of volatile compounds
After a fermenting period of 150 days, volatile compounds existed in the samples were analysed using solid-phase micro-extraction (SPME) gas chromatography–mass spectrometry (GC-MS). The sample (2 g) was extracted for volatile compounds as described by Pongsetkul et al. (2015). After extraction, the sample was allowed to absorb into the SPME fibre at 60 °C for 1 h before desorbed in the GC injector port. GC-MS analysis was performed on the Thermo Scientific TRACE GC Ultra Gas Chromatograph together with the ISQ Single Quadrupole Mass Spectrometer (Thermo Scientific Inc., New Jersey, USA). The retention time and fragmentation pattern were compared with the literatures and the Wiley 275.L data library of the GC-MS system.
Sensory evaluation
A group of thirty panellists, who consume budu regularly, was selected for the sensory evaluation. All samples (approximately 10 mL in 15 mL glass cups) were coded with three-digit random numbers and divided into four groups. Each group was randomly served in white paper plate. The panellists were instructed to use drinking water and cucumber for rinsing their mouths between different samples. Researchers then instructed the panellists to give linking scores for five attributes: overall liking, appearance, colour, flavour and taste using the 9-point hedonic scale.
Statistical analysis
The SPSS software version 17 (SPSS for window, SPSS Inc., Chicago, IL, USA) was used for analysed data. Statistical significance was evaluated using Duncan's multiple range and accepted at P < 0.05.
Results and discussion
Survival of the inoculated bacterial starters during fermentation
The populations of proteolytic bacteria and staphylococci in samples are shown in Fig. 1. The counts of proteolytic bacteria in all inoculated samples were higher than that in the control throughout the fermentation, which may be due to the high initial bacterial starter cultures (inoculated population 106 CFU g−1). Counts of proteolytic bacteria for inoculated samples increased at day 15, and then gradually declined until the end of fermentation (day 150), but for control sample, proteolytic bacteria slightly increased at 30 days and then decreased until the end of fermentation (Fig. 1a). The population of proteolytic bacteria in samples inoculated with a starter culture that included V. halodenitrificans PS21 increased faster than those in the control and in sample inoculated only with S. simulans PMRS35. More specifically, after 150 days, the proteolytic bacteria had a population of 5.33 and 5.58 log CFU g−1 in the PS21 and PMRS35+PS21 samples, while those of the PMRS35 and control samples were 4.42 and 2.02 log CFU g−1, respectively. The results indicated that V. halodenitrificans PS21 could be used as a starter culture for budu production under high salt condition. Yongsawatdigul et al. (2007) also reported that Virgibacillus sp. SK37 remained in the system up to 8 months.
Changes of proteolytic bacteria on the skim milk agar containing 10% NaCl (a), staphylococci on MRS containing 10% NaCl with added 0.5% CaCO3 (b) of budu samples fermented by various starter cultures inoculation at 35 °C. Results are mean values of five replicates ± SD.
The population of staphylococci appeared higher in PMRS35 and PS21+PMRS35 than in PS21 sample and the control, in the initial stage of fermentation, probably due to the inoculation of S. simulans PMRS35 (Fig. 1b). As fermentation progressed, staphylococci levels of PMRS35 and PS21+PMRS35 samples slowly increased at day 15 and gradually decreased towards the end of fermentation, while for the PS21 and control sample staphylococci levels slowly decreased throughout fermentation. The final staphylococci counts for the PMRS35 and PS21+PMRS35 were 4.13 and 4.70 log CFU g−1. Meanwhile, the counts were 2.2 and 2.3 log CFU g−1 for the control and PS21 samples, respectively. The decrease in staphylococci counts may be due to the high NaCl concentration (25%, w/v) and low water activity (Udomsil et al., 2015).
Degree of hydrolysis (DH)
The DH of samples inoculated without and with different bacterial strains is shown in Fig. 2a. DH is equivalent to the percentage of free amino groups obtained by the cleavage of peptide linkages (Benjakul & Morrissey, 1997). Among all samples, control had the lowest DH (40.81%) after 150 days of fermentation. All samples which were inoculated with starter culture showed higher DH than the control sample (P < 0.05), but there was no significant deference in PS21 and combined cultures (PS21+PMRS35) (P > 0.05). The differences in DH between control and inoculated samples generally resulted from the extracellular proteases from S. simulans PMRS35 and V. halodenitrificans PS21 used as starter culture, which contributed to the different aroma and taste of the finished products.
Degree of hydrolysis (a), free fatty acid content (b), and TBARS value (c) of budu samples fermented by various starter cultures. budu inoculated with single culture, Virgibacillus halodenitrificans PS21 (PS21) or Staphylococcus simulans PMRS35 (PMRS35), combined cultures (PS21+PMRS35) and the control without starter cultures. Data represent mean ± SD. Different superscripts indicate significant differences (P < 0.05).
FFAs content
Free fatty acid contents of samples inoculated without and with different bacterial strains are shown in Fig. 2b. FFAs are important substrates for lipid oxidation, which produce volatile compounds and contribute to the flavour quality of fermented fish products (Xu et al., 2018). The control sample had the lowest FFA content (5.50 g per 100 g lipid sample) (P < 0.05). Compared with the control samples, there was a higher FFA content in all inoculated samples, especially in PS21+PMRS35 (11.68 g per 100 g lipid sample), followed by PS21 (9.51 g per 100 g lipid sample) and PMRS35 (8.75 g per 100 g lipid sample), indicating that the inoculation with starter cultures had the ability to breakdown lipids. Thus, the inoculation of V. halodenitrificans PS21 and S. simulans PMRS35 could enhance lipolysis of budu. Generally, the specific flavours such as oily, fatty and tallow are originated from FFA breakdown.
TBARS analysis
TBARS values of samples inoculated without and with different bacterial strains are shown in Fig. 2c. Generally, TBARS values were used as indicators for testing of lipid oxidation in many fermented foods, which had impact on sensory characteristics. The control sample showed the lowest TBARS value (2.32 mg MDA per kg dry weight sample). The inoculated samples possessed higher TBARS value (3.3–3.8 mg MDA per kg dry weight sample), compared with control (P < 0.05), correlating with FFA contents (Fig. 2b). The results obtained indicated that the use of starter cultures might be facilitated the degradation of lipid hydroperoxides into malondialdehyde. In addition, fish muscle is rich in polyunsaturated FAs and is consequently prone to oxidative reactions during fermentation. The maximum TBARS value of the high-quality fermented fish products was 8 mg MDA per kg sample (Fan et al., 2008).
Total amino acid profiles
Total amino acid contents of samples inoculated without and with different bacterial strains are depicted in Fig. 3. As expected, the total amino acids in all of the inoculated samples were obviously higher than that in the control sample, especially in samples that added V. halodenitrificans PS21 (PS21+PMRS35 and PS21). These results indicated that V. halodenitrificans PS21 play a more crucial role than S. simulans PMRS35 in the amino acid production of the final product. As previously observed, V. halodenitrificans PS21 appeared to hydrolyse protein to a greater extent than S. simulans PMRS35 as indicated by higher degree of hydrolysis (Fig. 2a). The dominant amino acids in PS21 and PS21+PMRS35 samples were aspartic acid (2556.21 mg per 100 g for PS21+PMRS35), glutamic acid (1367.34 mg per 100 g for PS21+PMRS35) and lysine (1386.21 mg per 100 g for PS21+PMRS35), followed by proline, alanine and glycine, respectively. According to Yongsawatdigul et al. (2007), histidine, lysine, glutamic acid and aspartic acid were dominant in Thai fish sauce. Glutamic acid and alanine mainly contribute to good flavour characteristic of fermented fish-chilli paste (Hu et al., 2020). Meanwhile, proline contributed to sweetness, which could have a positive contribution to the overall taste characteristics of sea urchin roe (Park et al., 2001). Additionally, the inoculated samples with V. halodenitrificans PS21 also showed higher branched-chain amino acids (leucine and valine), when compared with the control (Fig. 3). It serves as a precursor for aldehyde compounds in fish sauce via amino acid catabolism pathway (Smit et al., 2009).
![Profiles of amino acids identified from budu samples fermented by various starter cultures. Data are given as the mean values ± SD. [Colour figure can be viewed at wileyonlinelibrary.com]](https://oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/ijfst/56/8/10.1111_ijfs.15035/1/m_ijfs15035-fig-0003-m.jpeg?Expires=1750319432&Signature=AjIA-zSvPOj61kzqe2NbTM-RhVfaPXER4mrNSKKbcesIkxfGXE~yfJpOFF88NYnep1sHtTeNsJVQUPVANPlOgd1R8HiXcVbx9slAR2as99CqtXcpmSaj3TZIXUaqO0ke175spW~6TZF4ybZRe~sgoQP8H80WZsJXsOyLg8wd1mMgZJfQh~2Wh~RtGV-UMV5ShIE~r0NZWVCYJkG8jKQKDoudrcR1yRrABSsgbpqtyu5DowwtIT-eOWup4Pz4iTg3V44EfN-qiyyly6HW-IiZ8sUsLsDThbCYqj3ESQZ92acHXAPXpSl2cmrGIwLffHpkZQ9U3oIVi1otc-QtIO7n4g__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
Profiles of amino acids identified from budu samples fermented by various starter cultures. Data are given as the mean values ± SD. [Colour figure can be viewed at wileyonlinelibrary.com]
Colour value and browning intensity
The samples inoculated without and with different bacterial strains had different colour characteristics and browning intensity (A420) as presented in Table 1. The colour is one of the parameters determining the overall consumer satisfaction, which brown is a preferable colour of budu (Mohamed et al., 2012). As shown in Table 1, browning intensity was higher in all inoculated samples, compared with the control sample (P < 0.05). Among all inoculated samples, PS21+PMRS35 had the highest browning intensity (P < 0.05). This was essentially correlated with the lower L*- and b*-values, along with the higher a*-value of the inoculated samples (Table 1). The result indicated that the combined inoculation of V. halodenitrificans PS21 and S. simulans PMRS35 could accelerate the development of browning intensity. Protease secreted by these starter cultures produced higher content of free amino acids and peptides (Fig. 3), which were subsequently functioned as reactants for Maillard reaction (Pongsetkul et al., 2018).
Colour and browning intensity (A420) of budu samples fermented by various starter cultures after 150 days fermentation
| Samples | Colour | Browning intensity (A420) | ||
|---|---|---|---|---|
| L* | a* | b* | ||
| Control | 40.88 ± 0.24a | 4.52 ± 0.17d | 17.18 ± 0.14a | 0.202 ± 0.01c |
| PMRS35 | 38.36 ± 0.29b | 4.65 ± 0.04c | 16.60 ± 0.13b | 0.225 ± 0.02b |
| PS21 | 35.69 ± 0.32c | 5.22 ± 0.12b | 14.69 ± 0.10c | 0.241 ± 0.01a |
| PS21+PMRS35 | 35.61 ± 0.13c | 5.73 ± 0.10a | 14.71 ± 0.14c | 0.251 ± 0.01a |
| Samples | Colour | Browning intensity (A420) | ||
|---|---|---|---|---|
| L* | a* | b* | ||
| Control | 40.88 ± 0.24a | 4.52 ± 0.17d | 17.18 ± 0.14a | 0.202 ± 0.01c |
| PMRS35 | 38.36 ± 0.29b | 4.65 ± 0.04c | 16.60 ± 0.13b | 0.225 ± 0.02b |
| PS21 | 35.69 ± 0.32c | 5.22 ± 0.12b | 14.69 ± 0.10c | 0.241 ± 0.01a |
| PS21+PMRS35 | 35.61 ± 0.13c | 5.73 ± 0.10a | 14.71 ± 0.14c | 0.251 ± 0.01a |
Data represent mean ± SD. Different superscripts in the same column indicate significant differences (P < 0.05).
Colour and browning intensity (A420) of budu samples fermented by various starter cultures after 150 days fermentation
| Samples | Colour | Browning intensity (A420) | ||
|---|---|---|---|---|
| L* | a* | b* | ||
| Control | 40.88 ± 0.24a | 4.52 ± 0.17d | 17.18 ± 0.14a | 0.202 ± 0.01c |
| PMRS35 | 38.36 ± 0.29b | 4.65 ± 0.04c | 16.60 ± 0.13b | 0.225 ± 0.02b |
| PS21 | 35.69 ± 0.32c | 5.22 ± 0.12b | 14.69 ± 0.10c | 0.241 ± 0.01a |
| PS21+PMRS35 | 35.61 ± 0.13c | 5.73 ± 0.10a | 14.71 ± 0.14c | 0.251 ± 0.01a |
| Samples | Colour | Browning intensity (A420) | ||
|---|---|---|---|---|
| L* | a* | b* | ||
| Control | 40.88 ± 0.24a | 4.52 ± 0.17d | 17.18 ± 0.14a | 0.202 ± 0.01c |
| PMRS35 | 38.36 ± 0.29b | 4.65 ± 0.04c | 16.60 ± 0.13b | 0.225 ± 0.02b |
| PS21 | 35.69 ± 0.32c | 5.22 ± 0.12b | 14.69 ± 0.10c | 0.241 ± 0.01a |
| PS21+PMRS35 | 35.61 ± 0.13c | 5.73 ± 0.10a | 14.71 ± 0.14c | 0.251 ± 0.01a |
Data represent mean ± SD. Different superscripts in the same column indicate significant differences (P < 0.05).
Volatile flavour compounds
Generally, the flavour is a very important factor affecting the quality characteristic of budu (Mohamed et al., 2012). Figure 4 presented the relative contents of volatile flavour compounds in each sample obtained by SPME-GC-MS. A total of fifty-three volatile components were identified at the end of the fermentation and divided into eight groups, namely alcohols, aldehydes, ketones, hydrocarbons, acids, furans, N-containing compound and S-containing compound. Among all compounds, aldehyde is the most prominent volatile compound found in budu samples followed by hydrocarbons and alcohols, respectively.
![Heatmap of the volatile profiles present in budu samples throughout fermentation. The colour scale in each column represents the intensity of volatile profiles fermented by various starter cultures. [Colour figure can be viewed at wileyonlinelibrary.com]](https://oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/ijfst/56/8/10.1111_ijfs.15035/1/m_ijfs15035-fig-0004-m.jpeg?Expires=1750319432&Signature=rZIoOL25xx9mCJv3yrFdy8DXoq34h68jOKXI5cdpWQU83Z5kPycj1vkzrTGLNRKnsLgHH~ldwEycvM6WOlJj0e1XJKwKCNpxXNQ902duS6ePOl0FjLJJ6hRR8q1XuTwHXjzi6jmmjZ8pp-iueQ6gU9dK1VS~x1urKoEIol5IVdJEpmnnBSVBiTmi~eqTBnFncZPiGPIBXZeR641U5YRDtPFLem2j3fkMTePwl~0ykQiMfVaEa1-jDmMb5npbfEHb5j1IkUSGZlkTkym~ldTEbmULHm~LoZ0Vthlwd9zRP3SMhiR0knm3mOxSipNjbf2KwtWcxB2DrD661wxE2LYEpg__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
Heatmap of the volatile profiles present in budu samples throughout fermentation. The colour scale in each column represents the intensity of volatile profiles fermented by various starter cultures. [Colour figure can be viewed at wileyonlinelibrary.com]
Aldehydes and ketones are important group of compounds creating the unique flavours of fermented fish products because of their low odour thresholds (Zeng et al., 2017) and derived from the lipid oxidation during fermentation. The higher abundance of aldehydes and ketones were attained in the inoculated samples for PS21 and PS21+PMRS35, compared with control sample (Fig. 4). This correlated well with the higher lipolysis and lipid oxidation found in the samples inoculated with starter culture, as evidenced by the higher FFA content and lipid oxidation products (Fig. 2b and c). The major aldehyde contents when inoculated with starter cultures included 2- and 3-methylbutanal, hexanal, n-heptanal and benzaldehyde. Among them, 2-methylbutanal and 3-methylbutanal, generated through Strecker degradation or microbial metabolism, contributed to aroma-active compounds in budu (Mohamed et al., 2012). Hexanal, which was responsible for a fresh grass odour (Gao et al., 2016), was typically resulted from linoleic acid oxidation (Gao et al., 2016) while benzaldehyde, which was responsible for a pleasant almond, nutty and fruity aroma in finish products, contributed to the aroma of fish sauce (Smit et al., 2009).
Twelve ketones were identified. Only 1-penten-3-one was identified among all samples inoculated with starter cultures but not found in the control (Fig. 4). Meanwhile, 2-propanone, 2-nonanone and 3-undecen-2-one were the major ketones found in all inoculated samples (Fig. 4), which was aligned with the findings of Yongsawatdigul et al. (2007) in fish sauce. In these samples, 2-propanone had the highest abundance in volatile compound, indicating the ability of selected starter cultures for generating this compound in budu. Ketones contributed to the cheesy note in budu odour (Mohamed et al., 2012), and generally originated by the Maillard reaction or from microbial enzymatic actions on lipids or amino acids (Takeungwongtrakul & Benjakul, 2013). However, ketones had little effect on flavour due to their high odour threshold.
Figure 4 showed that the amount of various alcohols was higher in all inoculated samples than in the control, including 1-penten-3-ol, 1-octen-3-ol, 1, 5-octadien-3-ol, and benzene ethanol. Those alcohol compounds might be the products from the degradation of lipid oxidation, which induced by the starter culture added. Isoamyl alcohol was noticeable in all inoculated samples but not detected in the control. Among all alcohol compounds, 1-octen-3-ol, which is unsaturated alcohols, appeared to have the most crucial impact on fermented fish aroma because of its low odour threshold value (Varlet & Fernandez, 2010). This alcohol compound is rapidly produced from arachidonic acid oxidation and usually found in fish sauce, fish miso, and soy sauce (Pham et al., 2008). Furthermore, 1-penten-3-ol and 1,5-octadien-3-ol, which created a meaty, burnt, grassy and green odour, were found at high abundance in most samples, but due to their high detection threshold values they may not have a significant impact (Gao et al., 2019).
The hydrocarbons were identified as the second most abundant volatile compound in most samples, especially in the inoculated samples (Fig. 4). Pentadecane, heptadecane and tetradecane were found in high amounts for PMRS35 sample, while nonane, 2-propenyl- cyclohexane, 3, 5-octadiene existed in low levels in all samples. Hydrocarbons were mainly attained from lipids during the autoxidation of long-chain fatty acids (Lu et al., 2011). Hydrocarbons' aroma contributions to overall sensory quality might be minimal because of their high sensory threshold values even though their flavour notes are typically desirable (Chen et al., 2017).
The most prominent furan compounds in all inoculated samples were 2-ethyl furan, of which the highest content was found in PS21 followed by PS21+PMRS35, PMRS35 and the lowest content was in the control sample (Fig. 4). Other furans found at high amount in the samples were 2-pentyl furan and trans-2-(2-Pentenyl) furan. In general, furans are found in dehydrated or fermented carbohydrate condensates or derived from Amadori rearrangement pathways (Taylor & Mottram, 1990). 2-ethyl furan contributed to a rubber and pungent smells, whereas 2-pentyl furan contributed to a beany, licorice-like and grassy taste. Mohamed et al. (2012) reported that 2-ethyl furan and 2-pentyl furan positively contributed to the overall flavour of budu because of their low odour threshold values. Moreover, trans-2-(2-pentenyl) furan also produced the same odour as 2-pentyl furan.
In addition, the amount of volatile fatty acids obtained in all inoculated samples were higher than those found in the control (Fig. 4). Those included heptanoic acid, octanoic acid, tetradecanoic acid and pentanoic acid. Two volatile fatty acids were negligible in control sample, including formic acid and butanoic acid. Those volatile fatty acids were also detected in PMRS35, which was not present in PS21 and PS21+PMRS35. According to Giri et al. (2010), volatile acids were generated by fatty acid oxidation and amino acid metabolism induced by hetero-fermentative lactic acid bacteria (LAB) and staphylococci. Volatile acid was one of the desirable characteristics of the product (Michihata et al., 2002).
The sulphur-containing compound obtained in budu samples was dimethyl disulphide. The samples of PS21 and PMRS35 had lower contents of dimethyl disulphide, compared with the control (Fig. 4). Dimethyl disulphide was not found in the PS21+PMRS35 with added combined cultures. Sulphur-containing compounds are products of methionine catabolism via transamination (Yvon & Rijnen, 2001), considering to be potent odorants due to their low threshold values. Dimethyl disulphide contributed to a faecal note which is an undesirable odour in fermented fish products. Therefore, samples inoculated by our starter cultures could eliminate the undesirable odour caused by dimethyl disulphide and enhance desirable odour characteristics of budu.
The only major nitrogen-containing compound in budu samples was 2, 6-dimethylpyrazine, which contained in low levels in all inoculated samples (Fig. 4). Meanwhile, nitrogen-containing compounds were not identified in the control sample. Fukami et al. (2004) showed that the inoculation of Staphylococcus sp. R4Nu increased 2, 6-dimethylpyrazine content of fish sauce. These compounds, which contributed to preferable odour in many fermented foods, are produced from Maillard reactions during fermentation (Mohamed et al., 2012). Although existed in low amounts, 2, 6-dimethylpyrazine could contribute to peanut, roasted, and coffee aromas in miso (Giri et al., 2010).
Sensory assessment
The median scores of the sensorial analysis of samples inoculated without and with different bacterial strains are presented in Fig. 5. Among all samples, PS21+PMRS35 showed the highest liking score in every sensory characteristic which includes taste, colour, appearance, flavour and overall. However, the overall liking scores of samples of PS21+PMRS35 and sample of PS21 were similar. The control sample was given the lowest score which did not differ from overall liking score of samples of PMRS35. The differences in sensorial analysis among samples were possibly influenced by differences in amino acid composition (Fig. 3) and volatile compounds (Fig. 4), which hydrolysed by the combined starter. Gao et al. (2019) reported that using mixed starter cultures improve sensory characteristic of fermented fish paste when compared with single culture. The sample of PS21+PMRS35, with the highest liking score, contained the highest umami amino acids, especially glutamic acid, aspartic acid and lysine. Yongsawatdigul et al. (2007) observed that taste of fish sauce was enhanced by glutamic acid and lysine, influencing sweetness and umami, respectively.
![Sensory evaluation of budu samples fermented by various starter cultures after 150 days fermentation. Results are mean values of three replicates ± SD. [Colour figure can be viewed at wileyonlinelibrary.com]](https://oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/ijfst/56/8/10.1111_ijfs.15035/1/m_ijfs15035-fig-0005-m.jpeg?Expires=1750319432&Signature=zi1gdzn1e4fggNIDVD85Rh~QBiRqC5vEASfQLVN-XYDLFTRSC8UJcrj0ZQolgAB0Lav7vhrBb-p5V~2W0bdlJpzugL0fqG1TL54huQ9agbcvHPpN7PrtjjmO0vOuqZDJK3QZW04Zy3nxWlhlEUJ0rHyKc8M--9JZY7gTtEeZlBz7e4s1SEKl2dilKyrbpeVy-g5oOlwBGZ5zn6axsvGPsOv07B7p5YoQZ39vgTOtvibhmG0clt~dYPuggVxWN5qa-KY9LKlpYRlrja5nj-VIPhreG5Fli7eSQZkJ2fGBOIaY949KQbKKHzAyk9HjWeCND~oGlX4LZrz0mfJT~F13pw__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
Sensory evaluation of budu samples fermented by various starter cultures after 150 days fermentation. Results are mean values of three replicates ± SD. [Colour figure can be viewed at wileyonlinelibrary.com]
Conclusions
This study suggested the advantages of using starter cultures of V. halodenitrificans PS21 and S. simulans PMRS35 in high salt fermented fish sauce (budu) production. Compared to the sample without starter cultures, the one inoculated with combined starters favoured development of budu characteristics, including colour development, lipid oxidation, total amino acid contents and volatile compounds in shorter time. Consequently, combined inoculation could be potential strains applied to accelerate budu fermentation with overall sensory characteristics.
Acknowledgments
The work was supported by the Halal Institute of Prince of Songkla University (Grant No. SAT03H61).
Conflict of interest
The authors declare no conflict of interest.
Author contribution
Pochanart Kanjan: Conceptualization (lead); Formal analysis (lead); Funding acquisition (lead); Investigation (lead); Methodology (lead); Project administration (lead); Supervision (lead); Writing-original draft (lead); Writing-review & editing (lead). Phat Sakpetch: Data curation (equal); Formal analysis (equal); Methodology (equal); Software (equal); Writing-original draft (equal). Payap Masniyom: Funding acquisition (supporting). Tipparat Hongpattarakere: Conceptualization (supporting); Methodology (supporting); Writing-original draft (supporting).
Ethical approval
Ethics approval was not required for this research.
Peer review
The peer review history for this article is available at https://publons.com/publon/10.1111/ijfs.15035.
Data availability statement
Research data are not shared.
References
This paper reported that 2-methylbutanal, 3-methylbutanal, dimethyl disulfide, 3-(methylthio)-propanal, 3-methylbutanoic acid and benzaldehye is the aroma-active compounds in commercial budu. It helped to support our statement that the inoculation of autochthonous starters could enhance flavor characteristic of budu product in shorter time.
This paper confirmed that the addition of combined starter cultures improved quality in terms of volatile compounds, glutamic acid content and overall acceptability. It helped to support our experiment design and result analyses.
This article studied the effect of autochthonous starter cultures on the chemical and microbial properties of Thai fish sauce. It provided important technique support for our budu products.
This article reported the physicochemical properties and sensory evaluation of Thai fish sauce produced using proteinases and bacterial starter cultures. It was cited for providing support for the parameter determination and result analyses in the volatile flavor compounds.
