Potential of three probiotic lactobacilli in transforming star fruit juice into functional beverages

Abstract The star fruit is popularly cultivated and consumed in Southeast Asia due to its high antioxidant capacity and various nutrients. In this study, three commercial probiotic strains (Lactobacillus helveticus L10, Lactobacillus paracasei L26, and Lactobacillus rhamnosus HN001) were evaluated in star fruit juice fermentation and all strains grew well with the final cell counts of around 108 CFU/ml. The star fruit juice fermented by L. rhamnosus produced the highest amount of lactic acid, resulting in a significant lower pH (4.41) than that of L. helveticus (4.76) and L. paracasei (4.71). Most of aldehydes and esters endogenous in star fruit juice decreased to low or undetectable levels, while ketones, alcohols, and fatty acids were produced at varying levels that could impart different aroma notes to the beverages. Therefore, the selection of appropriate probiotics can be an alternative way to develop new functional beverages from star fruit juice with specific aroma notes.

L. casei could promote cellular cholesterol reduction (Lye, 2010), and the important roles of L. rhamnosus GG and L. casei in prevention and treatment of pediatric diarrhea have also been well studied (Nixon, Cunningham, Cohen, & Crain, 2012;Wanke & Szajewska, 2014). Furthermore, the beneficial effects of lactobacilli on oral health (e.g., the reduction in dental caries incidences and salivary mutan formation) are also documented (Campus et al., 2014).
Probiotics are mostly found in yoghurt and fermented milks, because they are known to be excellent carriers for probiotics due to their good buffering capacity. However, consumers who suffer from lactose intolerance may not be able to enjoy the benefits of probiotic dairy products (Hertzler, Dennis, Jackson Karry, Bhriain, & Suarez, 2013). Therefore, nondairy probiotic beverages such as probiotic fruit juices would serve as an alternative for such consumers.
Of late, Lee, Boo, and Liu (2013) and Lu, Putra, and Liu (2018) have reported the successful probiotic fermentation (using L. acidophilus and L. casei) in coconut water and durian pulp, respectively.
The probiotic fermentation contributed unique flavor profiles to these fruit juices, which further raises interest in studying such fruit juices. However, the relatively low pH of fruit juices (<4) would pose a challenge to the growth of probiotics in such acidic media (Nagpal, Kumar, & Kumar, 2012). To enhance the survival and sustainability of probiotic lactobacilli in acidic juices, adjustment of pH of the media and selection of a robust strain could be conducted (Sheehan, Ross, & Fitzgerald, 2007). Therefore, the aim of this study was to evaluate the growth and metabolism of three probiotic lactobacilli (Lactobacillus helveticus L10, Lactobacillus paracasei L26, and L. rhamnosus HN001) at 30°C (prevalent in tropical countries). Ultimately, the intent is to develop a novel probiotic fermented functional star fruit juice beverage.

| Star fruit juice preparation
Star fruits were purchased from a local supermarket in Singapore.
Skin and seeds were removed from the pericarp before juicing in a blender. The crude juice was then centrifuged and filtered using a muslin cloth to remove the suspended solids. The initial total soluble solids content (°Brix) and pH were 7.09 and 3.58, respectively. The pH of the star fruit juice was adjusted to 5.9 (1 mol/L NaOH) to enable growth of lactobacilli. The star fruit juice was then filter-sterilized by sequentially passing through a 0.65μm and 0.45μm polyethersulfone filter membrane aseptically.

| Probiotic strains and preculture preparation
Three probiotic strains including L. helveticus (formerly acidophilus) L10 and L. paracasei L26 (both from Lallemand,Montreal,Canada) and L. rhamnosus HN001 (DuPont-Danisco, Singapore) were used in this study. The freeze-dried pure cultures were propagated in respective MRS broth at 37°C for 48 hr. The pure cultures were then stored at −80°C before use.
The precultures of probiotic strains were prepared separately by inoculating 10% (v/v) of the respective pure cultures into sterile star fruit juice. This was then followed by incubation at 37°C for 48 hr to achieve the cell forming unit (CFU) at least 10 7 per ml.

| Fermentation of lactobacillus strains in start fruit juice
Triplicate fermentations were conducted by inoculating 1% (v/v) precultures of each probiotic strain into 250 ml of sterile star fruit juice in 500-ml conical flasks. The fermentation was then incubated at 30°C for 8 days. Samples were taken at Days 0, 1, 2, 4, 6, and 8 for chemical and microbiological analyses under aseptic condition.

| Analytical determinations
The pH was measured using a pH meter (Metrohm, Herisau, Switzerland), and °Brix was determined by a refractometer (ATAGO, Yushima, Japan), respectively. The viable cell counts of Lactobacillus strains were determined by plating on MRS agar (62 g/L; Oxoid, Basingstoke, UK) and incubated at 37°C for 48 hr before plate counting.
High-performance liquid chromatography (HPLC) (Shimadzu, Kyoto, Japan) was used for the determination of sugars and organic acids. The separation and detection of sugars were performed using a Zorbax carbohydrate column (Agilent, Santa Clara, CA USA) with a low temperature evaporative light scattering (ELSD-LT) detector (40°C,350 kPa N 2 ). An isocratic flow rate of 1.4 ml/min was used for mobile phase consisting of acetonitrile-water mixture (80:20 v/v).
The organic acids were analyzed using a Supelcogel C-610H column (300 × 7.8 mm, Supelco, Sigma-Aldrich, Barcelona, Spain) connected to a SPD-M20A photodiode array detector at 210 nm (Shimadzu, Kyoto, Japan). The column was eluted with sulfuric acid (0.1% v/v) as the mobile phase at 40°C and 0.4 ml/min flow rate. Calibration curves were established for all analyzed compounds with R 2 > 0.99.
Prior to injection, samples were centrifuged at 20,379 g for 15 min at 4°C, followed by filtration using a 0.20μm regenerated cellulose filter membrane (Sartorius Stedim Biotech, Gottingen, Germany).
Headspace solid-phase microextraction (SPME) sampling was combined with gas chromatography (GC)-mass spectrophotometer (MS) and flame ionization detector (FID) for qualitative analysis of the volatiles as described by Lee, Ong, Yu, Curran, and Liu (2010).
The star fruit juice was adjusted to pH 2.5 with 1 mol/L HCl, and 5 ml of the sample was transferred to a 20-ml glass headspace vial sealed with a polytetrafluoroethylene septum. The extraction of volatiles was performed by a SPME autosampler (CTC, Combi Pal, Switzerland) using a carboxen/polydimethylsiloxane fiber (85 μm film thickness, Supelco, Sigma-Aldrich, Barcelona, Spain). Sample was subjected to 250 rpm agitation at 60°C for 45 min. This was followed by thermal desorption of the SPME fiber at 250°C in the injection port of an Agilent 7890A gas chromatograph coupled to an Agilent 5975C triple-axis MS and FID (Santa Clara, CA, USA). Separation of volatiles was carried out in an oven temperature programmed from 50°C where t represents the retention time of interest compounds in min, n is the number of carbon atoms of the n-alkane eluting before the compound; whereas t n and t n+1 are the retention time of the alkanes eluting before and after the interest compound, respectively.

| Statistical analysis
All analyses were carried out based on the data from the triplicate fermentations. One-way analysis of variance (ANOVA) and Scheffe's test were performed using SPSS 19.0 (Statistical Program for Social Sciences, SPSS Corporation, Chicago, IL), and significant difference was evaluated at the 95% confidence interval. Principal component analysis (PCA) was performed using software MATLAB R2008a (MathWorks, Natick, MA, USA) to analyze the distribution of aroma profiles of star fruit juice and star fruit juice beverages fermented with different probiotic strains.

| Growth of three probiotic strains
The growth of three lactobacilli strains in star fruit juice is shown in Figure 1. Lactobacillus paracasei exhibited a longer lag phase (4 days), whereas L. helveticus and L. rhamnosus increased rapidly after 2 days of the lag phase ( Figure 1). Lactobacillus rhamnosus increased to 6.23 × 10 7 CFU/ml on Day 4, while the other two strains were able to reach similar cell counts on Day 6 ( Figure 1). After which, L. rhamnosus and L. paracasei entered the stationary phase on Day 4 and Day 6, respectively. On the other hand, although L. helveticus started from a lower cell count (8.50 × 10 3 CFU/ml) compared with the other two strains (~10 5 CFU/ml), it was able to increase to 7.30 × 10 7 CFU/ ml on Day 6 and continued to grow to 2.07 × 10 8 CFU/ml on Day 8 (Table 1).
The growth patterns of the three probiotic stains used in this study were not consistent with that observed by Lee et al. (2013), who showed that L. paracasei and L. helveticus increased to ~10 8 CFU/ml within 2 days in coconut water without experiencing a lag phase. The lag phase in this study could be ascribed to the suboptimal fermentation temperature (30°C) and differences in nutrients. This agreed with the findings of Mousavi, Mousavi, Razavi, Emam-Djomeh, and Kiani (2011), where also reported a lag phase of L. paracasei and L. acidophilus in pomegranate juice fermentation at 30°C. However, our results were in contrast to some other studies, where lactobacilli could grow rapidly in fruit and vegetable juices at 30°C without going through the lag phase (Wang, Ng, Su, Tzeng, & Shyu, 2009). This could infer that other factors such as growth inhibitors and nutrients availability in the media may also affect the growth of probiotic strains (Siragusa et al., 2014). Besides, L. helveticus reached a final cell count of 1.6-fold to 2.0fold higher than that of L. paracasei and L. rhamnosus, respectively (Table 1). This indicated that L. helveticus could be a better candidate for star fruit juice fermentation.

| Changes of °Brix and pH
The changes in °Brix and pH served as indicators to monitor the fermentation progress in star fruit juice. All three probiotic strains resulted in slight decreases in °Brix from 7.09 to around 6.93-6.98 growth, production of lactic acid, and consumption of sugars during fermentation, especially by L. rhamnosus relative to the other two lactobacilli ( Figure 1, Table 1). Figure 3 shows sugar utilization by all three probiotic strains. Sucrose was totally depleted, and fructose decreased to 20.5-22.8 g/L on day 4 ( Figure 3). Nevertheless, glucose remained unchanged. Our results were in line with findings of Lee et al. (2013). The decrease in sucrose could be ascribed to the acid and/or enzymatic hydrolysis.

| Changes in sugars
However, no increase in glucose and fructose was observed during this period despite the decomposition of sucrose, indicating that Lactobacillus strains utilized glucose and fructose as their energy sources (Srinivas, Mital, & Garg, 1990) in counterbalance to the formation of glucose and fructose from sucrose hydrolysis.

| Changes in organic acids
The changes in organic acids in star fruit juice fermentation are shown in Table 1. The slight decrease in citric acid in all fermentations could be ascribed to the citrate fermentation pathway via citrate lyase (Hugenholtz, 1993;Mortera, Pudlik, Magni, Alarcón, & Lolkema, 2013), resulting in the formation of acetic acid and flavor compounds (diacetyl and acetoin) as shown in Table 1 and   Table 2, respectively. The star fruit juice fermented with L. helveticus (0.28 g/L) produced significantly higher level of acetic acid than that of L. paracasei (0.04 g/L) and L. rhamnosus (0.03 g/L) (Table 1), possibly due to metabolism of some amino acids such as serine and alanine.
Malic acid was the most abundant organic acid in fresh star fruit juice (Table 1). It was significantly reduced from 3.5 g/L to around 1.9-2.0 g/L in all fermentations (Table 1). This could be largely attributed to malolactic reaction by decarboxylation of malic acid to lactate (Schümann et al., 2013). In fact, most lactobacilli could decarboxylate malic acid directly into lactic acid by a single malolactic enzyme (Hutkins, 2007).
Lactic acid was the major acid produced during fermentation (Table 1). L. rhamnosus produced significantly higher level of lactic acid (4.4 g/L) than that of L. helveticus (3.43 g/L) and L. paracasei (3.70 g/L) (Table 1) in correlation with the pattern of sugar consumption ( Figure 3) and pH reduction ( Figure 2b). As mentioned earlier, malic acid could be one of the major sources for the accumulation of lactic acid. However, the major pathway for lactic acid production in this study should be from the transformation of a hexose into two pyruvic acids through the Embden-Meyerhof pathway, followed by the reduction in pyruvic acid into lactic acid by NAD + dependent dehydrogenases (Lengeler, Drews, & Schlegel, 2009), as all the lactobacilli used are homofermentative.
Similar but trace amounts of α-ketoglutaric acid (0.06 g/L) were produced in all star fruit juices fermented by different probiotic strains. α-Ketoglutaric acid could be formed from the catabolism of glutamic acid (Thage et al., 2004). Oxalic acid remained stable during fermentation (Table 1). This indicated that probiotics used in this study would not be able to degrade the oxalic acid in star fruit juice fermentation at 30°C. Oxalic acid is undesirable due to its ability to form salts of oxalic acid that may cause kidney stones.

| Changes in volatile profiles
Volatiles in star fruit juice before and after fermentation including acids, alcohols, aldehydes, esters, ketones, and terpenes are summarized in Table 2. The different probiotic strains resulted in drastic variations of the volatiles in star fruit juice beverages ( Table 2).
On the other hand, the aldehydes including benzaldehyde and tolualdehyde that were perceived as nutty and almond-like aroma notes were increased after fermentation, with higher amounts produced by L. helveticus and L. paracasei (Table 2). These compounds may be derived from the aromatic amino acids such as phenylalanine via the aminotransferase reaction (van Kranenburg et al., 2002).
The second most abundant volatiles in fresh star fruit juice were esters (methyl and ethyl esters, acetate esters), contributing to 22.99% of total peak area ( Table 2). All endogenous esters except for methyl benzoate were significantly degraded to trace or undetectable levels after fermentation (Table 2). Lactobacillus helveticus showed the highest ester degradation compared with the other two strains (Table 2). It is interesting to note that the shortchain esters (methyl butanoate, ethyl butanoate, n-hexyl acetate, and methyl heptanoate) were degraded more drastically compared to the long-chain esters (e.g., methyl salicylate and methyl anthranilate) ( Table 2). Our results agreed with the findings of Bintsis, Vafopoulou-Mastrojiannaki, Litopoulou-Tzanetaki, and Robinson (2003), in which most Lactobacillus strains, especially L. acidophilus, exhibited high esterase activities, which were involved in the breakdown of short-chain fatty acid esters.

TA B L E 2 (Continued)
Ketones were the largest volatile group produced in all fermentations with the increase of RPA from 2.20% to 30.72%-40.84% (Table 2). Diacetyl and acetoin were the major ketones that were produced with the highest production in star fruit juice fermented with L. rhamnosus ( Table 2). The production of diacetyl and acetoin by the probiotic lactobacilli was well documented (Benito de Cárdenas, Ledesma, Pesce de Ruiz Holgado, & Oliver, 1985;Liu, Holland, & Crow, 2003). These two buttery aroma compounds could be derived from the serine catabolism (Liu et al., 2003) or from citric acid (Hugenholtz, 1993). On the other hand, L. helveticus was found to be a good producer of 2-nonanone (contributing fruity and musty odor) compared with the other two probiotic strains (Table 2). This was in line with the findings in probiotic fermented coconut water (Lee et al., 2013).
The increases in fresh, sweet green like C 6 alcohols such as 1-hexanol and (E)-2-hexen-1-ol may be due to the reduction in corresponding C 6 aldehydes, as a reflection of the Lactobacillus in maintaining the redox balance (Budinich et al., 2011). In addition, these C 6 alcohols could also be produced by hydrolyzing the hexenyl and hexanyl esters during fermentation as discussed above.
Isoamyl alcohol could be derived from leucine via amino acid metabolism and is commonly found in foods fermented by Lactobacillus (Thage et al., 2004). Linalool, which gives rise to the citrus and floral aroma in star fruit, was increased in fermented juice by L. paracasei and L. rhamnosus but not L. helveticus (Table 2). This observation was in agreement with the fermentation of probiotic coconut water, in which L. helveticus did not produce linalool after fermentation (Lee et al., 2013). On the other hand, the production of 1-octen-3-ol and 2-ethylhexanol could be derived from the oxidation of linoleic and linolenic acids (Broadbent et al., 2004). These two compounds could contribute to the mushroom-like and sweet fruity-like aroma notes to the star fruit juice beverages, respectively.
Volatile fatty acids (VFAs) were another important volatile group produced after star fruit juice fermentation (Table 2). These acids were mostly derived from the hydrolysis of esters or from sugars, organic acids, and amino acids. The increase in hexanoic acid and (E)-2-hexenoic acid corresponded to the decrease in 1-hexanal and (E)-2-hexanal (Table 2), indicating the 6-carbon aldehydes could be oxidized into their corresponding volatile acids by the Lactobacillus.
The higher production of acetic, hexanoic, and (E)-2-hexenoic acids in star fruit juice fermented with L. helveticus could be explained by its higher hydrolytic activity of the corresponding esters.

| CON CLUS IONS
The potential of three different probiotic lactobacilli to ferment star fruit juice was evaluated, and the results showed that all three lactobacilli were able to grow well with final cell counts of around 10 8 CFU/ml. The highest level of lactic acid was produced by L. rhamnosus, resulting in the significantly lower pH of star fruit juice beverage than the juices fermented with L. helveticus and L. paracasei.
Endogenous volatile compounds in star fruit juice were degraded to low or undetectable levels, while new volatile compounds including ketones, alcohols, and fatty acids were produced by different probiotic strains at varying levels, contributing flavor complexity to the beverage. Therefore, the findings suggest that probiotic strains can be used to develop novel nondairy functional star fruit juice beverage with different flavor notes.

ACK N OWLED G EM ENT
The authors would like to thank the Food Science and Technology Programme of the National University of Singapore for providing the research facilities.

CO N FLI C T S O F I NTE R E S T S
All authors declare that they do not have conflicts of interests.