Household food waste samples (mainly containing rice, vegetables, and fruits) were collected from Thanjavur city, Tamil Nadu, India, and serially diluted up to 10^-7. One milliliter of sample from each dilution was spread-plated onto nutrient agar plates, which were incubated at 37°C for 48 h. Morphologically diverse colonies were selected and stored at 4°C for further studies.

The following methods were adopted to screen potential biosurfactant-producing strains: (i) drop-collapse test, by adding mineral oil to 96-well microtiter plates (Rodrigues et al., 2006); (ii) oil displacement activity, by adding weathered crude oil (Anandaraj and Thivakaran, 2010); (iii) emulsification activity, by adding kerosene and an equal volume of cell-free supernatant (Deepak and Jayapradha, 2015). The most efficient biosurfactant-producing strain was selected for further studies.

The morphological and biochemical characterization of the selected biosurfactant-producing strain was performed according to Bergey’s Manual of Determinative Bacteriology, 9th edition (Holt et al., 1994). Molecular identification was performed by 16S rRNA gene sequencing. Wizard^® genomic DNA purification kit (Promega, Madison, WI, USA) was used for genomic DNA extraction. PCR amplification and sequencing of the 16S rRNA gene were carried out using universal primers (5′-GAGTTTGATCCTGGCTCAG-3′; 5′-AGAAAGGAGGTGATCCAGCC-3′). Sequences were compared to those of other 16S RNAs obtained from GenBank using the BLAST program.

The screened isolate was inoculated in 25 mL of sterile nutrient broth and incubated at 37°C for 24 h at 300 rpm. Biosurfactant production was carried out according to the method described earlier (Giri et al., 2016c). Briefly, 10 mL of inoculum was transferred to a 1-L conical flask containing 500 mL of previously described medium (Joshi et al., 2013) and incubated at 37°C, pH 7.0, and 300 rpm for 72 h. After adjusting the pH to 7.0, cultures were centrifuged at 11300 × g at 4°C for 20 min to obtain cell-free supernatants. The supernatant was precipitated by adding 6 N HCl and pH was adjusted to 2.0. The collected precipitate (11300 × g at 4°C for 20 min) was dissolved in 5 mL of distilled water and pH was adjusted to 7.0 with 1 N NaOH. The biosurfactant solution was dialysed against demineralised water at 4°C using a Cellu-Sep© membrane (Seguin, USA) for 48 h. The partially purified biosurfactant was freeze-dried and stored at -20°C until use.

Initial identification of biosurfactant was performed using the phosphate (Okpokwasili and Ibiene, 2006) and biuret (Feigner et al., 1995) tests. Next, the partially purified biosurfactant was analyzed using a gas chromatography–mass spectrometry (GC–MS) system (Shimadzu QP2010 Plus, Japan) equipped with a capillary column (DB-5 MS; 0.25-mm film thickness, 0.25 mm i.d., 30 m length) (Giri et al., 2016c). “Briefly, helium was used as the carrier gas with a flow rate of 1.0 mL min^-1. The injection port temperature was 250°C. The column oven temperature was held at 80°C for 2 min and then the temperature was increased at every 10°C min^-1 up to 250°C, after which the temperature was finally raised at a rate of 5°C min^-1 to 280°C; this temperature was held for 10 min. Electron impact spectra were acquired in positive ionization mode between m/z 40 and 450″ (Giri et al., 2016c).

The (cetyltrimethylammonium bromide) CTAB agar test (Siegmund and Wagner, 1991) was used to identify the type of surfactant, and the orcinol test (Chandrasekharan and Bemiller, 1980) was performed for the detection of reducing sugars. To estimate protein concentration, a Lowry test (Lowry et al., 1951) was performed. In addition, a phosphate test was performed to check for the presence of a phosphate moiety (Okpokwasili and Ibiene, 2006).

Vegetables such as potato, carrot, radish, and ginger were collected from the local market of Thanjavur, Tamil Nadu. Vegetables were washed thoroughly with running tap water for 30–40 min and then washed with double distilled water. The skins were removed, cut into small pieces, immersed in a 1:1 HCl solution for 10 min, and then washed with double deionised water (Anjum et al., 2016).

Cadmium chloride stock solution was prepared in milli-Q water at a concentration of 1000 mg L^-1. A violet color developed when diphenylcarbazide was added and the absorbance was then determined at 540 nm using a UV-VIS spectrometer and a standard curve was plotted.

Vegetables were exposed to cadmium chloride at concentrations of 0.3, 0.4, and 0.5 mg mL^-1 for 30 min and diphenylcarbazide was added to develop the violet color (Anjum et al., 2016). The absorption of the Cd ion concentration was determined using a UV-VIS spectrophotometer. Few of the vegetables from the same stock were treated with biosurfactant; after the absorption with cadmium chloride, the change in concentration was again determined by measuring the OD at 540 nm. The cadmium ion removal percentage due to bio-adsorption was calculated as: percent Cd removal = (C_i – Co/C_i) × 100, where C_i = initial concentration of cadmium (mg L^-1) and Co = final concentration of cadmium (mg L^-1).

Biofilm inhibition assay was performed following the previously described method of Anjum et al. (2016) with slight modification. Overnight culture of Escherichia coli was prepared in 96-well plates containing 100 μL LB medium per well (37°C, 200 rpm). After 20 h, a 96-pin replicator (Boekel Scientific) was applied to inoculate 96-well test plates containing 150 μL of relevant pre-heated medium. To avoid the evaporation of medium, test plates were transferred to large plastic bags and incubated at 37°C for 48 h without shaking. After 48 h, biofilm was formed in a 96-well microtiter plate. The supernatant was discarded and the wells were washed with PBS to remove non-adherent cells. Then, the biofilm was treated with biosurfactant along with the control. The plates were incubated at 37°C for 2 h and 4 h. Subsequently, the plates were washed twice to remove non-adherent cells. A 100-μL solution of 0.4% crystal violet was added, incubated for 30 min, and washed with PBS to remove the extra stain. The biofilm was air dried for 5 min, then absolute ethanol was added to solubilize the crystal violet. The optical density was determined using a spectrophotometer (Perkin-Elmer, USA) at 590 nm.

Analysis of variance (ANOVA) tests were used to analyze the data. A Tukey’s test was used to analyze differences between the treatments. The OriginPro software (version 8; OriginLab Corporation, Northampton, MA, USA) was used for statistical analysis. The significance level fixed at P < 0.05. Results were expressed as the mean ± SEM.

Eighteen different bacterial colonies were selected from nutrient agar plates, of which eight isolates exhibited positive haemolysis activity. Among them, strain VS16 exhibited the highest haemolytic activity in terms of zone of clearance (2.97 cm radius) on blood agar. Emulsion formation is the stable interaction between the hydrophobic and hydrophilic phases. Among the isolates, VS16 had highest emulsification index of 49.03% during the tropophase and hence was selected for further studies. Results of the drop collapse and oil displacement tests confirmed that VS16 is highly capable of producing biosurfactant. Hence, VS16 was used for further studies.

Strain VS16 consisted of gram-positive, rod shaped bacteria that were arranged as chains. The results of biochemical tests indicated that the isolate was similar to Bacillus spp. Phenotypic identification was confirmed by 16S rRNA gene sequence analysis. BLAST search analysis¹ of the obtained 16S rRNA sequence (629 bp) showed high identity scores when compared to sequences from the Bacillus genus. Strain VS16 had the highest nucleotide sequence similarity value (99%) when compared to that of B. licheniformis (GenBank accession numbers: KJ526872.1, KF737353.1, AY842876.1, and AY859479.1) (Phylogenetic data not shown). The nucleotide sequence of this strain has been submitted to NCBI GenBank under the accession number KX943303.

The result of CTAB showed no coloration of colonies or of CTAB agar, which indicates that the biosurfactant was not a glycolipid. Based on the Orcinol test, the expected red-orange coloration was not observed, indicating the absence of reducing sugars in the biosurfactant. The partially purified biosurfactant tested positive using Lowry’s test, thus indicating the presence of a lipopeptide moiety. Moreover, a phosphate assay confirmed its phospholipid type. These results demonstrated that the B. licheniformis VS16-derived biosurfactant could be a type of phospho-lipopeptide biosurfactant.

Gas chromatography–mass spectrometry analysis (Figure 1 and Table 2) revealed the presence of 11 lipid-based compounds with biosurfactant properties. Major peaks at retention times of 16.15, 17.75, 21.44, and 21.87 min were confirmed as eicosanoic acid, heptacosane, tert-hexadecanoic acid, and hexadecanoic acid, with molecular weights of 326, 380, 258, and 270 Da, respectively. Similarly, heneicosane, octadecanoic acid and methyl ester, myristoleic acid, and pentadecanoic acid were identified at the retention times of 19.25, 19.45, 20.93, and 21.22 min with molecular weight (Da) of 296, 298, 226, and 242, respectively. Peaks at 15.06, 23.30, and 25.84 min were confirmed as 9-hexadecenoic acid, methyl ester, myristic acid, methyl ester, and mannosamine with molecular weights of 254, 232, and 215.6, respectively. The highest percentage of peak area was for hexadecanoic acid and the lowest percentage peak (2.91%) was identified for myristoleic acid at a retention time of 20.93 min.

The highest LA and ACP activities (P < 0.05) were recorded 14 days in S220 group (Table 3). However, both LA and ACP activities slightly declined at 21 days, when compared to those recorded at 14 days, except in the S55 group.

Significantly higher PA was recorded in the S220 and S330 groups than in the control group at all time points (Table 4), with highest in the S220 group at 14 days. IgM levels were higher in the treated groups (Table 4). Fish in the S220 group exhibited significantly higher IgM levels at all time points, with a highest in the S220 group at 7 days (Table 4).

Administration (i.p.) of biosurfactant enhanced digestive enzyme activities in fish (Table 5). Amylase activity was significantly higher in the S220–S330 groups at 14 and 21 dpa than in the controls, and was highest in the S220 group at 14 dpa. Protease activity was significantly higher in the S220 group at 14 and 21 dpa with the highest levels at 14 dpa. Lipase activity was higher (P < 0.05) in the S220 and S330 groups at 14 and 21 dpa (Table 5); however, highest lipase activity was recorded in the S330 group at 7 dpa.

The expression of cytokine genes in the head kidney of L. rohita was altered by biosurfactant administration (i.p.) as shown in Figures 2–4.

Fish administered biosurfactant (S110–S330) had significantly lower TNF-α levels at 14 dpa (Figure 2) when compared to those in the controls, with the lowest levels observed in the S220 group. At 7 dpa, lower (P < 0.05) TNF-α expression was recorded in the S330 group. The expression of IL-1β was lower in the treatment groups and differences were significant in the S220 and S330 groups at 7 and 14 dpa (Figure 2).

Administration of biosurfactant to fish provoked the expression of anti-inflammatory cytokines IL-10 and TGF-β during the whole experimental period (Figure 3). IL-10 expression was higher (P < 0.05) in the S110–S330 groups at 14 dpa, with the highest levels in the S220 group. TGF-β mRNA expression was significantly higher in the S220 and S330 groups at 7 and 14 dpa. However, both IL-10 and TGF-β showed a declining tendency at 21 dpa.

The expression of NF-κB p65 showed a tendency to decline in the treated groups at all time points (Figure 4). NF-κB p65 expression was significantly lower in the S110–S330 groups at 14 dpa. At 7 and 21 dpa, differences were significant only for the S330 and S110 groups, respectively.

Fish treated with biosurfactant tended to exhibit lower IKKβ expression than those in the control groups at all time points, but the differences were significant only for the S220 group at 14 dpa (Figure 4). IκB-α expression was significantly higher in the S110–S330 groups at 14 dpa (Figure 4) than in the controls, with the highest expression in the S220 group. At 21 dpa, significantly higher IκB-α expression was recorded in the S220 group.

The results of the challenge study revealed that fish administered biosurfactant exhibited better protection against A. hydrophila infection. Fish administered 220 μg mL^-1 of biosurfactant (S220) exhibited the highest RPS (72.72%), which was followed by fish in the S330 (57.56%), S110 (39.39%), and S55 (12.11%) groups. No mortality was observed in the negative control group. The lowest survival (8.33%) was recorded in the positive control group.

Bacillus licheniformis VS16-derived biosurfactant was effective in inhibiting biofilm formation, and in removing heavy metals from vegetables (Table 6). The biofilm inhibition study showed that the biosurfactant was effective in suppressing biofilm formation by up to 54% (Table 6A). Further, the biosurfactant had the potential to remove cadmium from carrots, radishes, ginger, and potatoes (Table 6B). Biosurfactant removed 60.98% of Cd from Cd-contaminated ginger.

Biosurfactant production by B. licheniformis VS16 was confirmed through several simple yet powerful screening methods such as emulsification activity, drop collapse test, and oil displacement tests. The composition of biosurfactant was characterized through GC–MS analysis (Figure 1 and Table 2). The percentages of peak area were 13.62, 14.3, 18.16, and 21.76% for eicosanoic acid, heptacosane, tert-hexadecanethiol, and hexadecanoic acid, respectively. The highest percentage of peak area was available for hexadecanoic acid and the lowest percentage peak (2.91%) was identified for myristoleic acid at a retention time of 20.93 min. The major abundance of methyl esters of hexadecanoic acid and eicosanoic acid indicated that C₁₆ and C₂₀ were the main fatty acid components of the purified biosurfactant. The results revealed that the biosurfactant was a mixture of 11 different compounds (C₆–C₂₇) consisting of long chain β-hydroxy fatty acid, hexadecanethiol, and heptacosane. Results obtained in this study were consistent with those of Yakimov et al. (1995), wherein lipopeptide surfactant from B. licheniformis BAS50 had a mixture of 14 β-hydroxy fatty acids ranging in size from C₁₂ to C₁₇. The result was also similar to that of another study by Rajeswari et al. (2016), wherein GC–MS data for a biosurfactant from S. hominis showed that it was a mixture of 17 different β-hydroxy fatty acids (C₈–C₂₇).

The present study demonstrated that B. licheniformis-derived biosurfactant was able to modulate immune responses in L. rohita. Lysozyme has substantial bactericidal activity against both gram-positive and gram-negative bacteria in the presence of complement (Giri et al., 2015). In fish, complement alarms the host immune system to the existence of potential pathogens and assists in their subsequent clearance (Holland and Lambris, 2002). The highest (P < 0.05) LA and ACP activities were recorded at 14 days in the S220 group. However, both activities declined slightly at 21 days, when compared to those obtained at 14 days. Similarly, dietary supplementation with B. licheniformis SY-52-derived secondary metabolites was previously shown to enhance LA in C. carpio (Chen et al., 2015). Dietary administration of extracellular proteins from Aeromonas veronii BA-1 and Flavobacterium sasangense BA-3 significantly enhanced lysozyme and complement C3 activities in C. carpio (Chi et al., 2014). L. rohita administered cellular components of probiotic bacteria had significantly higher LA and ACP activities (Giri et al., 2015). In addition, phospholipid biosurfactants isolated from S. hominis increased the LA in O. mossambicus (Rajeswari et al., 2016).

‘PA is an important tool and primordial defense mechanism of the non-specific immune system’ (Chen et al., 2015). In the present study, PA was significantly higher in the S220 and S330 groups than in the control group at all time points (Table 4). Earlier studies revealed that bacterial metabolites can increase PA in fish (Chen et al., 2015; Giri et al., 2016c). In the present study, IgM levels were higher in the biosurfactant treated groups (Table 4). Fish in the S220 group exhibited higher (P < 0.05) IgM levels at all time points (Table 4). IgM is the predominant antibody type in fish. It is used to identify and neutralize foreign objects such as bacteria and viruses (Giri et al., 2014). Similarly, cellular components of the probiotics Kocuria SM1 and Rhodococcus SM2 increased IgM levels in Oncorhynchus mykiss (Sharifuzzaman et al., 2011). Earlier studies demonstrated that probiotic supplementation increased IgM levels in fish at an initial time point; however, at later time points, these levels declined (Sun et al., 2010; Giri et al., 2013, 2014). In this study, IgM levels increased initially and thereafter they decreased gradually, suggesting that IgM increases is a temporary phenomenon attributable to immunostimulants. Further, LA, PA, ACP, and IgM activities were lower in the S330 group than in the S220 group, suggesting that administration of overdoses of bacterial metabolites could result in the suppression of immune responses in fish (Giri et al., 2016c).

Administration (i.p.) of biosurfactant enhanced digestive enzyme activities in fish (Table 5). Amylase, protease, and lipase activities were higher in the treated groups. The digestion processes in aquatic animals can be enhanced by the addition of certain microorganisms or their by-products, which can boost the production of extracellular enzymes such as proteases and lipases, and/or can have the intended ability of supplying necessary growth factors as fatty acids and vitamins, among others (Ibrahem, 2015). Earlier studies reported that probiotic bacteria could enhance the secretion of digestive enzymes in fish (Essa et al., 2010; Mohapatra et al., 2012; Sankar et al., 2016). Therefore, increased production of digestive enzymes could improve digestion, which could in turn increase the growth performance of the fish.

“Cytokines originate from macrophages, lymphocytes, granulocytes, dendritic cells, mast cells, and epithelial cells and include ILs, TNFs, TGFs, IFNs, and chemokines” (Savan and Sakai, 2006). Pro-inflammatory cytokines IL-1β and TNF-α are mainly produced by monocytes and macrophages and regulate numerous aspects of the immune response (Giri et al., 2015). In the present study, administration of biosurfactant (S220 and S330) significantly lowered TNF-α and IL-β expression in the head-kidney of fish at 14 dpa (Figure 1). In a similar study, we previously showed that administration of biosurfactant could down-regulate the expression of TNF-α and IL-1β in L. rohita (Giri et al., 2016c). However, biosurfactant (S110–S330) significantly upregulated the expression of anti-inflammatory cytokines (IL-10 and TGF-β) at 14 dpa, and maximum expression was recorded in the S220 group (Figure 2). This is consistent with increased IL-10 and TGF-β expression observed in L. rohita immunized with biosurfactant (Giri et al., 2016c). In addition, IL-10 expression was up-regulated in rainbow trout cell cultures stimulated with lipopolysaccharide (Fierro-Castro et al., 2013) and in C. carpio administered bacterial secondary metabolites (Chen et al., 2015). The observed inverse relationships between the expression of pro- and anti-inflammatory cytokines in the present study is consistent with earlier studies wherein fish were administrated immunostimulants (Feng et al., 2016; Giri et al., 2016c). IL-10 is produced mainly by monocytes, macrophages, B cells, T cells, and dendritic cells. It inhibits the expression of many pro-inflammatory cytokines (De Waal Malefyt et al., 1991). “TGF-β inhibits B and T cell proliferation and differentiation, antagonizes pro-inflammatory cytokines such as IL-1β, TNF-α, and IFN-γ, and blocks the expression of IL-1β and IL-2 receptors” (Li and Flavell, 2008). These results suggest that the administration of 220 μg mL^-1 of biosurfactant attenuates pro-inflammatory responses in fish.

NF-κB, a critical transcription factor that is involved in the inflammatory response, controls the expression of various compounds such as IL-6, IL-1β, TNF-α, iNOS, and COX-2 (Cho et al., 2013). In this study, mRNA expression of NF-κB p65 significantly declined during the entire duration of the trial (Figure 3) and the lowest expression was recorded in the S220 group at 14 dpa. Inhibition of NF-κB was previously shown to down-regulate the expression of IL-1β and IL-8 in mouse monocytes (Li et al., 2008). Further, IκB-α expression was higher in the biosurfactant-treated groups (Figure 3). However, IKK-β mRNA was lower in the treated groups and this difference was significantly lower only in the S220 group at 14 dpa (Figure 3). Up-regulation of IκB-α, an NF-κB binding protein, was shown to lead to down-regulation of pro-inflammatory cytokines in mouse muscle fibers (Hnia et al., 2008). Further, IL-10 can attenuate pro-inflammatory cytokine production through the suppression of NF-κB activation by sustaining IκB protein expression (Chon et al., 2010). Therefore, our results indicated that biosurfactant might act by down-regulating the expression of IKK-β to attenuate the degradation of IκB-α, thereby inhibiting NF-κB p65 nuclear translocation, to suppress the expression of pro-inflammatory cytokines (Giri et al., 2016c). However, the underlying mechanism requires further investigation.

In the present study, carp treated with (i.p.) biosurfactant showed increased protection against A. hydrophila. Fish administered 220 μg mL^-1 of biosurfactant (S220) exhibited the highest RPS (72.72%). In a recent study, we demonstrated that L. rohita treated with (i.p.) with 200 μg mL^-1 of B. subtilis-derived biosurfactant were most protected against A. hydrophila infection (Giri et al., 2016c). Dietary administration of secondary metabolites isolated from B. licheniformis XY-52 enhanced the survival of C. carpio against A. hydrophila challenge (Chen et al., 2015). O. mossambicus administered S. hominis-derived phospholipid biosurfactant (200 mg kg^-1) were most protected against A. hydrophila challenge (Rajeswari et al., 2016). The improved immune parameters (LA, ACP, PA, and IgM) and digestive enzyme activities, and higher expression of anti-inflammatory cytokines in fish treated with 220 μg mL^-1 of biosurfactant might be linked with the enhanced resistance to A. hydrophila, which could result in higher post-challenge survival. The groups injected with 55, 110, or 330 μg mL^-1 of biosurfactant exhibited lower post-challenge survival, which might be associated with the obtained results regarding immune responses, digestive enzyme activities, and immune gene expression. These results indicate that B. licheniformis VS16-derived biosurfactant has immunomodulatory effects on carp.

The biosurfactant was effective in suppressing biofilm formation (Table 6A) by up to 54%. Controlling the adherence of microorganisms to the food contact surface is critical for providing safe and healthy food products to consumers (Anjum et al., 2016). Therefore, the use of biosurfactants in inhibiting biofilm formation could be an important tool for the food industry. Glycolipid biosurfactant (30 μg glycolipid mL^-1) derived from Brevibacterium casei was shown to significantly inhibit biofilm production by Vibrio spp., E. coli, and Pseudomonas spp. of both mixed culture and individual strains (Kiran et al., 2010a). Nocardiopsis sp. MSA13A-derived biosurfactant (300 μg mL^-1) significantly disrupted biofilm formation in Vibrio alginolyticus (Kiran et al., 2014). Biosurfactant isolated from Lysinibacillus fusiformis S9 exhibited notable inhibition of biofilm formation by pathogenic E. coli or Streptococcus mutans (Pradhan et al., 2014). They reported that at a concentration of 40 μg mL^-1 the biosurfactant efficiently inhibited biofilm formation by E. coli and S. mutans. Further, Falagas and Makris (2009) hypothesized that ‘probiotics or their products (biosurfactants) could be applied to patient care equipment such as tubes or catheters with the aim of decreasing colonization by nosocomial pathogens.’

The intake of heavy metals though food or any other means is highly detrimental to human health. The uptake of heavy metals by crops is often associated with the plant species, growth phase, soil type, metal variety, soil conditions, environment, and weather. We found that biosurfactant has the potential to remove cadmium from carrots, radishes, ginger, and potatoes (Table 6B). Recently, Anjum et al. (2016) demonstrated that Bacillus sp. METCC 5877-derived biosurfactant removes considerable amounts of heavy metals from vegetables such as potatoes, garlic, radishes, and onions. In addition, rhamnolipids (glycolipids) removed nickel and cadmium from soils with efficiencies of 80–100%, when tested in a controlled environment, and this removal efficiency was 20–80%, when using field samples (Wang and Mullingan, 2004). Therefore, the controlled and cautious use of these exciting surface active molecules will definitely assist in the removal of the toxic environmental pollutants and could help provide a clean environment.