Lactobacillus reuteri ATCC 55730 obtained from BioGaia, Stockholm, Sweden, was used as a reference strain and the test strain, L. reuteri LR47 was selected from our previous study (30). The bacterial indicator strains (obtained from an in-house lab culture collection) Escherichia coli K12, Listeria monocytogenes ATCC 13932, Enterococcus faecalis NCDC 135, Serratia marcescens ATCC 13880, Salmonella Typhi NCDC 113, Yersinia enterocolitica ATCC 130715, Klebsiella pneumoniae NCDC 138 and Pediococcus acidilactici LB42 were used for the antibacterial activity testing of L. reuteri LR47 supernatant. All the bacterial cultures (excluding P. acidilactici LB42) were grown in brain heart infusion broth (BHI, HiMedia, Mumbai, India), while P. acidilactici LB42 and L. reuteri strains were grown in de Man, Rogosa and Sharpe broth (MRS; HiMedia, Mumbai, India) at 37 °C for 24-48 h. The bacterial strains were maintained as 15% glycerol stock cultures at -20 °C in BHI and MRS broths. The organisms were propagated twice before their use in the experiments.

Cell-free supernatants (CFS) of L. reuteri LR47 were prepared according to the method described by Lüthi-Peng et al. (31). A single colony with typical morphology (small pinpointed and creamy white colony) from the MRS agar plates was picked and inoculated into MRS broth and incubated anaerobically (gas pack system, HiMedia) at 37 °C for 20 h. After incubation, the culture was centrifuged (Precision Biotech Instruments Pvt. Ltd., Delhi, India) at 4000×g and 4 °C for 10 min. The supernatant was filter-sterilised through a 0.22-μm syringe filter (Millipore, New Delhi, India) to remove any remaining cells. The cell pellet was used for secondary fermentation to prepare a suspension of MRS supplemented with glycerol (360 mM). The supernatant obtained from secondary fermentation of MRS with glycerol was designated MRS-Gly, while the supernatant derived from MRS broth was designated MRSB. Finally, the pH values of the above supernatants were determined using a portable pH meter (Orion Star™ A111; Thermo Fisher Scientific, Waltham, MA, USA). The pH of the supernatants was adjusted from pH=4.6 to 6.5 (in order to eliminate the cause of microbial death due to acidic pH) using 0.1 M sodium hydroxide at room temperature prior to testing their antimicrobial activity.

Antibacterial activity of L. reuteri LR47 CFS was previously tested in our study (30) using agar well diffusion method (32). In this study, we used the microtiter plate method to test the antibacterial activity of the CFS. An aliquot of 20 μL of CFS was added, in triplicate, into 180 μL of BHI broth containing 10⁶ CFU/mL of each pathogen, and incubated at 37 °C for 18 h. Negative control contained pathogen cultures only, while positive control contained commercial antibiotics (γ(ampicillin)=100 mg/mL as stock solution; Sigma-Aldrich, St. Louis, MO, USA) instead of CFS. The inhibition of indicator strains was quantitatively determined by measuring the absorbance (A) (Spectrostar Nano BMG Labtech, Ortenberg, Germany) at 620 nm, in comparison with the control. The percentage of inhibition was calculated as follows:

PCR was conducted using genomic DNA of L. reuteri LR47 to detect the presence of dhaB (part of dha gene operon) and pduC (gene coding for large fragment of gdh protein) genes, responsible for conversion of glycerol into reuterin, using DHAB1 (5’-AACTACGATAACATGTTTGC-‘3), DHAB7 (5’-CCTTCTTCTTCAATTCCGGCA-‘3) and pduCF (5’-CCTGAAGTAAAYCGCATCTT-‘3), pduCR (5’-GAAACYATTTCAGTTTATGG-‘3) primer pair as per the method described by Versalovic et al. (33) and Walter (34), respectively. The resulting PCR products were separated and visualized using standard agarose gel electrophoresis.

CFSs (MRS-Gly) of three L. reuteri strains (L. reuteri LR47, L. reuteri LR11 and L. reuteri LR 49) were assayed to quantitatively determine glycerol utilisation and the resulting concentration of reuterin present in the CFS. Reuterin and glycerol concentrations were determined, in triplicate, using a colourimetric method adapted from Spagnola (35). Briefly, an aliquot of 330 µL of CFS (MRS-Gly) from each bacterium was mixed with 225 µL of 10 mM tryptophan (dissolved in 0.05 M HCl) and 150 µL of 95% ethanol. Each mixture was incubated at 40 °C for 50 min and the absorbance was measured at A_{560 nm} using spectrophotometer microplate reader (Spectrostar Nano BMG Labtech, Ortenberg, Germany). The reuterin concentration was calculated by comparing the absorbance of the samples with a standard curve of acrolein, assuming that 1 mol of reuterin was dehydrated to 1 mol of acrolein that reacted with tryptophan in the presence of HCl.

The activity of L. reuteri LR47 CFS was quantified by minimum inhibitory concentration (MIC) assay using a microtiter plate against E. coli K12 as described by Chung et al. (36). Overnight culture of E. coli K12 was harvested, washed twice with phosphate buffer (pH=7.2, 50 mM), suspended in the same buffer, and diluted to an A_{620 nm} of 0.2 measured using a spectrophotometer (Spectrostar Nano BMG Labtech). This suspension was diluted 1:10, which corresponds to about 10⁶ CFU/mL. The diluted suspension (0.1 mL) was used to inoculate 0.2 mL of serial dilutions of CFS in Müller-Hinton medium (HiMedia). The microtiter plate was incubated at 37 °C for 12 h and then the growth of indicator strain E. coli K12 was examined. The unit of the CFS activity was defined as the reciprocal of the highest dilution that did not permit visible growth of the indicator strain and expressed in arbitrary unit (AU). Accordingly, the activity of the CFS was calculated to be 1600 AU/mL.

Pediocin was prepared by inoculating active cultures of pediocin-producing Pediococcus pentosaceus 34 (37) in 100-mL aliquots of MRS broth and incubated at 37 °C for 24 h. CFS was prepared by centrifugation of the culture in a refrigerated centrifuge (Precision Biotech Instruments Pvt. Ltd.) at 4000×g for 10 min. The supernatant was filter-sterilized by passing through 0.22-µm pore size membrane filter after neutralization. Pediocin activity was quantified by the MIC method against Pediococcus acidilactici LB42 as described by Kaur et al. (38) using microtiter plates. Nisin was procured from Sigma-Aldrich. Stock solutions of nisin (1000 IU/mL) were prepared in 0.02 M HCl and stored at -40 °C. The nisin stock solution was diluted prior to use in experiments.

Raw milk samples of Bos taurus (Sahiwal breed) were obtained from the Experimental Dairy plant of ICAR-NDRI, Karnal, India. All samples were collected in the morning. After the collection of samples, it was immediately transported to the laboratory in a tightly closed container placed in an ice box. At the start of each experiment, a volume of 100 mL of raw milk was aliquoted in 14 sterile screw-cap glass bottles as per experimental design. Each sample was marked with the type of antimicrobial treatment to be given. Antimicrobial compounds in different combinations and different activities (Table 1) were added concomitantly to the respective sample bottles. Milk sample without the addition of antimicrobials served as a control. All the untreated and treated milk samples were incubated at 37 °C for 9 h and parameters were measured after 3, 6 and 9 h of experiment.

The analysed chemical parameters of the raw milk samples were pH and titratable acidity (TA). The pH was measured using a digital pH meter (Orion Star™ A111; Thermo Fisher Scientific) and the TA was determined by titration method (39) at different time intervals (0, 3, 6 and 9 h). Each sample was measured in triplicate. Total plate count (TPC) and coliform counts were also tested simultaneously by the sample dilution pour plate method. Briefly, tenfold serial dilutions of samples were made up to volume fraction of 10^-7 in normal saline solutions (0.85%). A volume of 1 mL of milk sample was diluted in a series of normal saline solutions (9 mL of saline down to 10^-8) and 1-mL aliquots of milk in the saline solution from appropriate volume fraction dilutions: 10^-1, 10^-3 or 
10^-5 for coliform count, and 10^-3, 10^-5 or 10^-7 for TPC were transferred to Petri dishes. Samples were plated in triplicates, using pour plate technique. Nutrient agar medium (HiMedia) was used for the TPC, while violet red bile agar (HiMedia) was used for counting coliforms incubated at 37 °C. After incubation, plates containing 30 to 300 colonies were selected for screening and results were expressed as log CFU/mL.

The data obtained in this study are presented as mean value±standard deviation (S.D.) and analyzed statistically with the GraphPad Prism software (40). Microbiological counts were converted to log CFU/mL and the statistical significance among different antimicrobials and their combination in milk system was compared by two-way analysis of variance (ANOVA) followed by Tukey’s test and considered significant at p≤0.01.

Many wild-type variants of Lactobacillus reuteri have the ability to produce reuterin from anaerobic conversion of glycerol (22). Reuterin is a potent and a wide-spectrum antimicrobial agent that suppresses various kinds of bacteria, fungi, protozoa, etc. (21). However, many inherent genetic factors as well as other external factors influence reuterin production (41). Hence, efforts are mainly focused on bioprospect high reuterin-producing L. reuteri strains from different niches. We selected L. reuteri strain LR47 from our previous study (30), which had shown highest antimicrobial activity during the initial screening and antimicrobial activity, which was further tested against eight different bacterial indicators that included important food pathogens. Both, CFS before (pH=4.5) and after the adjustment of the pH value (pH=6.5) derived from MRSB and MRS-Gly were used to test the antimicrobial activity. Antimicrobial activity was observed in both MRS-Gly and MRSB (Fig. 1); however, inhibitory effect of MRSB is very low or insignificant at pH=6.5 (Fig. 1a). L. reuteri, being a LAB, secretes organic acids such as lactic acid, acetic acid, etc. as primary fermentation products and many others as secondary metabolites (15) that have been reported to produce strong antimicrobial agents, particularly in a relatively low pH environment, which may have contributed to the antimicrobial activity observed in the MRSB. However, antimicrobial activity could not be replicated at pH=6.5. On the other hand, CFS (pH_adjusted=6.5) from MRS-Gly exhibited strong antimicrobial activity against the majority of indicator strains. It was able to strongly inhibit both Gram-negative and Gram-positive bacterial indicators. The activity observed in this case can be attributed to reuterin, since it is used as an antimicrobial agent against a wide range of bacterial targets at different pH values including neutral pH (42), contrary to organic acids and bacteriocins that are effective only against a limited number of bacterial targets and often lose their potential at an altered pH. However, reuterin did not have any significant inhibitory effect against E. faecalis and P. acidilactici. This is not surprising as it has been reported that reuterin affects lactic acid bacteria to a lesser degree, and a relatively higher concentration of reuterin is required to inhibit the growth of LAB (21).

Few bacterial strains have the metabolic capacity to utilize glycerol and produce reuterin. Among them are L. reuteri strains, which possess all the necessary genes for bioconversion of glycerol into reuterin. Glycerol dehydratase and diol dehydratase are two discrete isofunctional enzymes, which catalyse the reaction of glycerol dehydration to reuterin by two independent pathways (41, 43). The structural genes of the glycerol dehydratase (dhaBCE) of L. reuteri are part of the dha regulon and its expression is induced under anaerobic conditions in the presence of glycerol. On the other hand, diol dehydratase genes of L. reuteri are part of pdu operon (pduCDE) which also contains genes for glycerol utilization. Glycerol, a second substrate of pduCDE, is transformed into an intermediate compound (reuterin), which can be further metabolized to 1,3-propanediol or 3-hydroxypropionate. Earlier, Schaefer et al. (44) demonstrated the elimination of reuterin production by disrupting/deleting the pduC gene in L. reuteri strains ATCC 23272 and confirming the role of pduC gene in reuterin production. Therefore, we also attempted to check the presence of either one or both genes in L. reuteri LR47 by PCR assay. The PCR results showed a positive amplification of both genes, viz. pduC (105 bp) and dha (900 bp) in L. reuteri LR47 (data not shown), which confirmed that this strain possessed these genes and molecular set-up to drive the conversion of glycerol to reuterin.

To confirm and augment the genetic screening results, we also performed the colourimetric estimation of reuterin production in CFS of L. reuteri LR47. Colourimetric estimation of reuterin was performed to determine the levels of reuterin produced among different strains of L. reuteri (L. reuteri LR47, L. reuteri LR49 and L. reuteri LR11) (30). Initial and final concentrations of reuterin were determined and compared within and among different strains used in the study. Also, the amount of utilised glycerol was determined and compared with the respective supernatant of these three L. reuteri strains. It is evident from Fig. 2 that reuterin production in L. reuteri LR47 was significantly higher than in the other two L. reuteri strains. The different levels of reuterin production (12.58, 13.65 and 21.08 mg/mL) could also correlate well with the differences in the glycerol utilization (18.06 20.77 and 24.00 mg/mL) of L. reuteri LR49, LR11 and LR47, respectively. Other research groups have reported reuterin production in the range of 45.7 to 298 mM (30, 41, 42). However, most of these studies have used different L. reuteri strains and conditions, and also E. coli system to stimulate reuterin production. In contrast, we carried out a direct assay in MRS-Gly broth for reuterin production and obtained quite significant reuterin concentrations without any external stimulation. This process can be particularly useful in in situ biopreservation of food where antimicrobial production is required. Nevertheless, the E. coli system could also be tested here to achieve a still higher reuterin concentrations in L. reuteri LR47 strain.

Based on the literature survey on reuterin and bacteriocins (11-13, 17, 24, 26, 29), current work was undertaken to investigate the effect of different concentrations of reuterin, nisin and pediocin individually or in different combinations on the shelf-life of raw milk. Literature search indicates that there are hardly any studies conducted on the preservation of raw milk using reuterin and bacteriocins. Thus, we proceeded to experiment and verify if different combinations of these antimicrobial compounds would be effective in controlling the initial microbial counts without compromising the physiochemical properties of raw milk. The selected concentrations of reuterin, nisin and pediocin were based on the earlier studies (11, 17, 29) and total plate count (TPC), coliform count, pH and titratable acidity (TA) were measured to monitor microbial progression at 0, 3, 6 and 9 h of incubation at 37 °C. Raw milk samples were significantly influenced by the type of antimicrobial combination in the treatments and the duration of incubation at 37 °C (shown in Fig. 3 and Fig. 4, Table S1 and Table S2).

In the first subgroup, reuterin, nisin and pediocin were added singly to each raw milk sample and physicochemical parameters were measured. At the beginning of the incubation period (0 h), the overall TPC and coliform count was 4.6 and 4.1 log CFU/mL, respectively, in all raw milk samples. However, after the initial 3-hour incubation, there was a significant increase in the TPC and coliform count in the control sample. The addition of reuterin, nisin and pediocin showed an initial bacteriostatic effect, having respective TPC and coliform count of 2.4, 1.6 and 0.7 log CFU/mL and 1.1, 1.5 and 0.4 log CFU/mL (Fig. 3a and Fig. 4a), which was lower than control sample count (Table S1 and Table S2). After further 3-hour incubation (i.e. total 6 h of incubation), the TPC and coliform count continued to increase in the control sample, as well as in the treated samples, but they remained lower than in the control. After 9 h of incubation, the antibacterial action of the three added biopreservatives seemed to have been exhausted and very low bacterial inhibition was visible in the treated samples. Earlier, EI-Ziney and Debevere (45) and Arqués et al. (24), who studied the efficacy of reuterin in UHT skimmed milk at 7 or 37 °C, had shown that reuterin had a bacteriostatic effect against individual Gram-negative bacteria. However, it was also observed that reuterin enacted strong inhibitory action against other pathogens and inhibition of up to 4 log units could be attained, but at a lower temperature of 7 °C. Arqués et al. (11) in another study showed that when added singly, nisin exhibited bactericidal effect against S. aureus at 37 °C during 4 h of incubation. However, the pathogen regained its growth after reaching bacterial counts similar to that in control milk after 24 h. Hence, it is quite clear that individual antimicrobial metabolites have limited capacity against bacterial targets in the milk system. Inhibitory effect can be more suppressed when mixed microflora in ambient conditions are used, like in this study. Hence, the different combinations of reuterin, nisin and pediocin were further tested to achieve better results.

In the second subgroup, a combination of two antimicrobial compounds was used. A greater initial inhibition of TPC and coliform count was observed than when these antimicrobial compounds were applied individually in raw milk system. After 3 h, the raw milk treated with 150 AU/mL reuterin and 100 IU/mL nisin (R1+N1), and 150 AU/mL reuterin and 2185 AU/mL pediocin (R1+P1) had TPC and coliform counts of 3.2 and 2.2 log CFU/mL and 2.6, 1.5 log CFU/mL respectively, which was lower than the control sample (Fig. 3a and Fig. 4a). However, at the end of 6 h of incubation, the initial inhibitory effect of the combination started diminishing and there was a steady increase in the TPC and coliform count in the samples treated with both combinations. Thereafter, the growth resumed in the two samples after 9 h of incubation and the TPC and coliform bacteria reached counts that were comparable with those in the control sample. The initial reduction in the bacterial counts was synergistic because the combined effect of two treatments was greater than the sum of individual inhibition treatments. However, some researchers have reported that the combined effect of reuterin with nisin could not improve the antimicrobial efficacy of reuterin against Gram-negative bacteria in UHT skimmed milk at 4 or 8 °C, although they also reported that reuterin in combination with nisin had a synergistic effect against L. monocytogenes in milk at 37 °C (25). Arqués et al. (26) also reported a synergistic effect of reuterin (8 AU/mL) with nisin (100 IU/mL) against L. monocytogenes and S. aureus in milk at 37 °C. The reuterin activity against L. monocytogenes has also been reported to be enhanced by the addition of 3% salt (45). Hence, the results of the current study confirm the synergistic action of reuterin and nisin against the raw milk microflora. In addition, a combination of nisin and pediocin PA-1 also showed a strong synergistic bactericidal activity against milk pathogens (46). However, the dual combination was not effective enough to obtain a sufficient level of shelf-life extension of raw milk at 37 °C, hence, triple combinations were further tested.

In the third subgroup with the combination of three biopreservatives at different concentrations, the treatment with 150 AU/mL reuterin, 100 IU/mL nisin and 600 AU/mL pediocin (R1+N1+P2) showed the best results and had a significant effect on the TPC and coliform count. After 3 h of incubation, the TPC and coliform count were 3.9 and 2.9 log CFU/mL respectively, which is lower than in the control milk (Fig. 3b and Fig. 4b). The efficacy of 150 AU/mL reuterin, 100 IU/mL nisin and 600 AU/mL pediocin (R1+N1+P2) treatment was maintained even after 6 h of incubation and very minimal increase in the TPC and coliform count was observed. The effect, however, seemed to fade thereafter and increase in both TPC and coliform count could be observed (Fig. 3b and Fig. 4b). As anticipated, the triple combination of antimicrobial compounds showed the best effect, which could be attributed to a synergistic action among the used antimicrobials. Arqués et al. (11) had shown earlier that a similar triple combination consisting of three biopreservatives reuterin, nisin and lactoperoxidase system (LPS) in curdled milk had a very strong synergistic bactericidal effect against L. monocytogenes with only 0.3 log CFU/mL bacterial count remaining in the milk samples compared to control (8.3 log CFU/mL). In compliance with our obtained results, they also reported that the combination of three biopreservatives comprising both reuterin and nisin along with LPS system exhibited superior action to single or a combination of two biopreservatives in milk system.

Efficacy of biopreservatives used in triple combinations is evident in the pH and acidity of the treated raw milk samples. Among different combinations of biopreservatives, only the treatment with R1+N1+P2 combination was able to maintain TA of 0.22% lactic acid (Fig. 5) and pH=6.43 (Fig. 6) after 6 h of incubation, which was within the commercially acceptable range of pH=6.22-6.77 and TA of 0.16-0.25% lactic acid (47), while after all the other treatments higher pH and TA values exceeded their commercially acceptable limits (Fig. 5 and Fig. 6). Under the prevailing tropical conditions, the major bacterial spoilage encountered in raw milk is curdling, which is mainly due to the increase in titratable acidity and decrease of pH. The commercial dairy industry also uses TA values to judge the freshness of raw milk as a basis for its overall acceptability for further processing. The rise in TA values is a direct indicator of microbiological activity in milk, which is more rapid at ambient temperatures (48). Apart from various other microbial groups, the LAB groups and coliforms are the most active in the initial increase in the TA of raw milk. Bacteriocins used in the study such as nisin and pediocin are derived from LAB groups and are active against closely related species (15). Hence, bacteriocins used in the study may have played an important role in controlling the LAB groups. However, bacteriocins are not very effective against Gram-negative bacteria such as coliforms, Pseudomonas and other similar bacteria that are highly prevalent in raw milk. Hence, the reuterin system may have worked in parallel to inhibit and suppress the growth of these groups of bacteria in the raw milk and extend the shelf-life of raw milk for a period of 6 h at 37 °C (Table S3 and Table S4). The reuterin system works by inducing oxidative stress in target cells (44), while both bacteriocins typically act by forming pores in the bacterial cell membrane. This multitarget system could also explain the increased efficacy and synergistic action of the triple combination of biopreservatives in raw milk system. Thus, our results show that reuterin could be successfully combined with nisin and pediocin at specific concentrations and used to inhibit complex microflora that is prevalent in raw milk and maintain it for at least 6 h under ambient temperature conditions.