Semi-synthetic or synthetic antibiotics are extensively used in animal farming to treat bacterial diseases and to promote growth (Lin et al. Reference Lin, Hunkapiller, Layton, Chang and Robbins2013). The benefit of using such antibiotics as growth promoters in avian production lies in their ability to improve growth by enhancing feed conversion efficiency and to control pathogens. However, the emergence of resistant strains of bacteria after exposure to antibiotics is an important consequence and is considered to be the biggest challenge of the 21st century because it compromises antimicrobial efficiency (Marshall & Levy, Reference Marshall and Levy2011). In recent years, European countries have banned the use of antibiotics as growth promoters. Alternatives to the administration of conventional antibiotics include food supplementation with probiotics (Kaur et al. Reference Kaur, Kuhad, Garg and Chopra2009) or essential oils (Brenes & Roura, Reference Brenes and Roura2010), vaccination (Dahiya et al. Reference Dahiya, Wilkie, Van Kessel and Drew2006), and bacteriophage therapy (Wong et al. Reference Wong, Sieo, Tan, Abdullah, Hair-Bejo, Abu and Ho2014). As well, antimicrobial peptides (AMP) can be seen as potential alternatives to conventional antibiotics (Joerger, Reference Joerger2003). A study by Stern et al. (Reference Stern, Svetoch, Eruslanov, Kovalev, Volodina, Perelygin, Mitsevich, Mitsevich and Levchuk2005) showed that bacteriocin-based treatment can reduce the colonisation and intestinal level of Campylobacter jejuni in chicken.
Antimicrobial peptides are natural bactericidal substances found in a wide variety of organisms (Hancock & Chapple, Reference Hancock and Chapple1999). They are encrypted within endogenous precursor proteins and are released by proteolysis. It has been well documented that peptides with antimicrobial activity can also be generated from food proteins following ingestion and gastrointestinal digestion or during fermentation. Dairy products are interesting sources of AMP (Sibel Akalin, Reference Sibel Akalin2014; Théolier, Reference Théolier, Fliss, Jean and Hammami2014), and specific peptides have been identified from hydrolysates of the proteins casein (Lahov & Regelson, Reference Lahov and Regelson1996), lactoglobulin (Pellegrini et al. Reference Pellegrini, Dettling, Thomas and Hunziker2001), lactalbumin (Pellegrini et al. Reference Pellegrini, Thomas, Bramaz, Hunziker and Von Fellenberg1999), and lactoferrin (Bellamy et al. Reference Bellamy, Takase, Yamauchi, Wakabayashi, Kawase and Tomita1992b) following hydrolysis by digestive enzymes. Moreover, some of the antimicrobial activity against Escherichia coli and Salmonella enterica var. Typhimurium is associated with proteolysis of milk fat globule membranes (MFGM) (Clare et al. Reference Clare, Zheng, Hassan, Swaisgood and Catignani2008) that are released from fat globules during churning for butter production. The best known and most active AMP from dairy products is lactoferricin B (LfcinB) [f(17–41), molecular weight 3126 Da, isolectric point (pI) 11·84], a cationic peptide produced from hydrolysis of lactoferrin by pepsin (Bellamy et al. Reference Bellamy, Takase, Wakabayashi, Kawase and Tomita1992a, Reference Bellamy, Takase, Yamauchi, Wakabayashi, Kawase and Tomitab).
The preparation of highly purified AMP is a challenge, and yields are usually very low. To be used as alternative to antibiotics, it is essential to produce, at low cost, large amounts of active, partially purified preparations from inexpensive food substrates using large-scale, commercially viable processes. Buttermilk, the by-product of butter production, is essentially skim milk containing large amounts of MFGMs (Dewettinck et al. Reference Dewettinck, Rombaut, Thienpont, Le, Messens and Van Camp2008). This currently under-exploited milk product is interesting as a crude source of AMP to optimise dairy AMP recovery. Moreover, milk-derived AMP can be considered GRAS (generally recognised as safe) because they are naturally produced during the digestive process. Indeed, some milk-derived AMP, such as Lfcin, can be naturally detected and may be potentially active within the gut (Kuwata et al. Reference Kuwata, Yamauchi, Teraguchi, Ushida, Shimokawa, Toida and Hayasawa2001; Troost et al. Reference Troost, Steijns, Saris and Brummer2001).
The objective of this study was to develop a rapid and efficient method to prepare and fractionate peptides to produce peptide extracts rich in AMP and then to test their antimicrobial activity against specific enteric avian bacterial strains, namely, Salmonella enterica var. Enteritidis, Salmonella enterica var. Typhimurium, Escherichia coli O78:H80, and Escherichia coli TK3 O1:K1, as well as against a Gram-positive strain, Staphylococcus aureus. The method of production of active peptide extracts was based on ultrafiltration (UF) followed by ionic membrane adsorption chromatography (MAC) of buttermilk hydrolysate, with lactoferrin hydrolysate also being fractionated as a control for the process.
Materials and methods
Production of peptide extracts
Buttermilk powder (Agropur, Longueuil, QC, Canada) and lactoferrin powder (Vitalarmor LF 1401; Euro Proteins Inc., Wapakoneta, OH, USA) were dissolved in water at 5% (w/v) protein, as determined by the Kjeldahl nitrogen analysis method (Lynch & Barbano, Reference Lynch and Barbano1999) and dialysed against water for 24 h at 4 °C (molecular weight cut-off [MWCO] 12–14 kDa). The protein solutions were digested with pepsin from porcine gastric mucosa (enzyme/substrate ratio 1/20; Sigma-Aldrich, Oakville, ON, Canada) for 6 h at pH 2·5 and at 37 °C as described by Bellamy et al. (Reference Bellamy, Takase, Yamauchi, Wakabayashi, Kawase and Tomita1992b). After the digestion, the solutions were centrifuged at 7000 rpm for 30 min, and the supernatant was subjected to UF on a Kvick Start column (MWCO 10 kDa; GE Healthcare, Baie-d'Urfé, QC, Canada). Permeates corresponding to total peptide extracts (TPE) were kept frozen until use. The kinetic of hydrolysis was evaluated by the OPA (o-phthalaldehyde) spectrophotometric assay (Church et al. Reference Church, Swaisgood, Porter and Catignani1983).
Fractionation of peptides
The TPE were fractionated on the basis of the peptides' charge through a cation-exchange adsorptive membrane (Mustang S capsule, CLM05MSTGSP1, 10-ml bed volume; Pall Corporation, St. Laurent, QC, Canada) followed by an anion-exchange adsorptive membrane (Mustang Q capsule, CLM05MSTGQP1, 10-ml bed volume; Pall Corporation). Before use, the capsules were cleaned with 20 volumes of NaOH 1 N and 10 volumes of NaCl 1 N, and then equilibrated with ammonium acetate (NH4OAc) 1 m, pH 5 for cationic MAC or pH 9 for anionic MAC. Finally, the capsules were rinsed with water until the conductivity was below 2 mΩ/cm. After use, the capsules were regenerated with NaOH, neutralised with NH4OAc 1 m (pH 7), rinsed with water, and kept in 20% ethanol. To prepare the cationic peptide extracts (CPE), the pH of TPE were adjusted to pH 5·0 with NH4OH 1 m and then diluted with water to bring conductivity to below 2 mΩ/cm. The solutions were filtered on 0·22-μm membrane filters (EMD Millipore, Taunton, MA, USA) and applied on the Mustang S capsule at a flow rate of 50 ml/min. The capsule was washed with about 350 to 400 ml of NH4OAc 10 mm (pH 5·0) until the absorbance at 280 nm was less than 1. Adsorbed biomaterial corresponding to the CPE was eluted from the capsule with NaCl 1 m in NH4OAc 50 mm (pH 9·0) and kept frozen until use. The unbound eluate from the Mustang S chromatography was adjusted to pH 9·0 with NH4OH 1 m, diluted with approximately 3 volumes of water (conductivity < 2 mΩ/cm), and applied on the Mustang Q anionic capsule at a flow rate of 50 ml/min. After being washed with 350 to 400 ml of NH4OAc 10 mm (pH 9·0), anionic peptide extracts (APEs) were eluted with NaCl 1 m and NH4OAc 50 mm (pH 5·0). Samples that did not specifically bind to any capsule were considered the non-adsorbed peptide fraction. Peptide contents were quantified using a Pierce BCA Protein Assay Kit. Unless stated, the chemicals were from Fisher Scientific (Nepean, ON, Canada).
Analytical and preparative RP-HPLC
Peptide samples (0·15 to 0·5 mg/ml in 0·1% trifluororacetic acid [TFA]) were injected on a C18 RP-HPLC (reversed-phase high-performance liquid chromatography) column (Aeris PEPTIDE 3·6 µm XB-C18, 250 × 4·6 mm; Phenomenex, Torrance, CA, USA) at a flow rate of 1 ml/min and at 25 °C and then eluted with a linear gradient of 0–50% acetonitrile – 0·1% TFA for 60 min and detected at 214 nm. Spiking experiments were conducted by adding synthetic LfcinB (GenScript USA Inc., Piscataway, NJ, USA) as reference peptide in desalted lactoferrin TPE and CPE before C18 analysis. To enable in vitro antimicrobial tests, the peptide extracts were desalted by RP-HPLC. The peptide extracts (250 mg) were solubilised in 0·1% TFA and injected onto a C12 RP-HPLC semi-preparative column (Jupiter 10 µm proteo 90A; Phenomenex) at a flow rate of 3 ml/min, eluted with 80% ethanol in water and then freeze-dried.
Bacterial strains and antimicrobial assay
Four avian clinical isolates supplied by the Chair in Poultry Research and the EcL Laboratory of the Faculté de médecine vétérinaire, University of Montreal were used for antimicrobial tests. Selected pathogens were associated with poultry diseases and were Salmonella enterica var. Enteritidis (SHY-04-1540), Salmonella enterica var. Typhimurium (DT104 #4393), Escherichia coli O78:H80, and Escherichia coli TK3 O1:K1. A Gram-positive strain, Staphylococcus aureus (ATCC 25923), was also tested. The bacterial cultures were maintained in 20% glycerol at −80 °C. The inocula were thawed, transferred at 1% (v/v) into tryptic soy broth (Difco; Fisher Scientific, Nepean, ON, Canada), and grown at 37 °C for 18 h. The bacterial suspensions were centrifuged (2000 g , 5 min), washed with 0·85% (w/v) NaCl, and suspended at 106 CFU/ml in PYG medium (10 g/l bacto-peptone, 10 g/l glucose, 0·5 g/l yeast extract). Microdilution assays were carried out in a 96-well microplate (Corning Inc., Canton, NY, USA). Briefly, 0·1 ml of serial dilution from filter-sterilised (0·22 µm; EMD Millipore) peptide extracts (final concentrations of 0, 0·3, 0·6, 1·3, 2·5, and 5 mg/ml) in PYG were in contact with bacterial strains (0·025 ml, final bacterial concentration of approximately 0·5–1 × 106 CFU/ml in PYG) for 2 h at 37 °C. At the end of the treatment, 0·125 ml of twice-concentrated brain-heart infusion broth was added to the wells containing bacterial suspensions, which initiated growth of the bacteria that had survived the bactericidal effect of the peptide extracts. The bacterial counts were determined based on the time required for a sample grown in liquid media to reach threshold turbidity, as described by Brewster (Reference Brewster2003). The growth curve was recorded using a PowerWave x microplate reader (BioTek Instruments, Winooski, VT, USA). The lag time needed to reach 0·1 OD600nm was determined and compared to the corresponding lag time from the standard curve established with various initial bacterial counts to evaluate bacterial viability. Within each test day, the antimicrobial activity of LfcinB, at concentrations from 0·003 to 0·05 mg/ml, was analysed as the control to ascertain the quality and reproducibility of the tests.
Statistical analyses
Analysis of variance (SAS software package; SAS Institute Inc., Cary, NC, USA) was performed on the data (five bacterial strains, six different concentrations of peptide extracts, six different extracts from two different preparations of peptides). Antimicrobial analysis was done in duplicate on test days, and three or four replicates were carried out per experimental condition. The alpha threshold was adjusted using the Benjamini–Hochberg method; P < 0·0005 was considering significant. The initial bacterial population was used as a co-factor to standardise all analyses.
Result
Production of peptide extracts
According to the kinetics of protein hydrolysis, estimated by the OPA spectrophotometric assay, hydrolysis by pepsin reached a plateau for all samples between 3 and 6 h. Following hydrolysis, UF (MWCO 10 kDa) was used to remove enzyme. On a protein basis, 45–50% of the starting proteinaceous materials were recovered in the UF permeate as TPE. The TPE were sequentially fractionated based on their net charge using, in sequence, cationic (Mustang S) and anionic (Mustang Q) MAC. The yields of recovery of the buttermilk CPE and APE were 6·5 ± 1·8% (n = 5) and 7·0 ± 3·5% (n = 4) of the TPE, respectively. The yield of recovery of the lactoferrin CPE was higher, with a value of 15·4 ± 3·8% (n = 2). The maximum capacity was 240 mg for a 10-ml Mustang S capsule (mean obtained with two different capsules), indicating that up to 3·5 g of buttermilk TPE could be applied per run, taking into account the yield of recovery of buttermilk CPE.
The RP-HPLC profiles of the lactoferrin TPE and CPE on Mustang S are presented in Fig. 1. At least 14 peaks were detected in the lactoferrin CPE. The spiking experiment using commercial LfcinB made it possible to locate Lfcin within the lactoferrin CPE at 35·54 min (Fig. 1b insert). The Lfcin contents in the lactoferrin TPE and CPE were about 2·5 and 3·3%, respectively, based on the relative absorbance at 214 nm on the chromatogram, the area under the peak. It is also estimated that the cationic extract accounted for 41% of the lactoferrin TPE. Taking into account the amino acid sequence of Lfcin (25 aa), covering 3·5% of lactoferrin, the fact that 41% of the TPE adsorbed on Mustang S, and that about 3·3% of the peptides were eluted as Lfcin, it was estimated that the yield of recovery of Lfcin was 37·6%.
RP-HPLC elution profile of (a) lactoferrin total peptide extract and (b) lactoferrin cationic peptide extract. Insert: lactoferrin cationic extract spiked with lactoferricin B (LfcinB) f(17–41) [lactoferricin cationic peptide extract without LfcinB (- - -) and with LfcinB f(17–41) (____) and LfcinB f(17–41) (….)]. Linear gradient from 5–50% acetonitrile containing 0·1% trifluoroacetic acid on a C18 column at 20 °C 60 min.
The RP-HPLC analyses of the buttermilk TPE, CPE, and APE (Fig. 2) revealed that the fractionation efficiency and elution pattern of the CPE greatly differed from those of the APE. The buttermilk CPE tended to be eluted more slowly than the buttermilk APE. The main CPE-specific peaks were between 26 and 29 min and at 19·5 min. The amount of Lfcin in the buttermilk CPE cannot be estimated, given that lactoferrin is a minor milk protein and its concentration in milk reaches only 140 mg/l (Arnould et al. Reference Arnould, Soyeurt, Gengler, Colinet, Georges, Bertozzi, Portetelle and Renaville2009).
RP-HPLC elution profile of (a) buttermilk total peptide extract, (b) buttermilk cationic peptide extract, and (c) buttermilk anionic peptide extract. Linear gradient from 5–50% acetonitrile containing 0·1% trifluoroacetic acid on a C18 column at 20 °C, 60 min.
Antimicrobial activity of buttermilk peptide extracts
The antimicrobial activities of buttermilk TPE are presented in Table 1. The bacterial counts of E. coli O78:H80 significantly declined by 0·8 log CFU/ml after 2 h in the presence of the TPE at 0·6 mg/ml, and the counts gradually decreased thereafter to reach a viability loss of 1·9 log CFU/ml at 5 mg/ml. The other strains were more resistant to the antimicrobial activity of the TPE: viability decreased by less than 1·3 log CFU/ml at 5 mg/ml for E. coli TK3 O1:K1, Salm. Enteritidis, and Salm. Typhimurium and Staph. Aureus was not significantly affected by the TPE at 5 mg/ml. Enrichment with cationic peptides by MAC greatly improved the specific antimicrobial activity (Table 2). For instance, the buttermilk CPE at 5 mg/ml decreased Staph. aureus viability by 1·5 log CFU/ml, although the TPE did not significantly affect viability. Moreover, it took about eight times less buttermilk CPE than buttermilk TPE to decrease the bacterial viability of the other strains by 1 log CFU/ml from the initial bacterial content. Thereafter, the viability gradually decreased with increased CPE concentration, to reach 1·7 to 2·9 log CFU/ml. The buttermilk APE and the non-adsorbed peptide fractions did not significantly decrease the bacterial viability of the strains under study at any concentrations tested (data not shown).
Bacterial viability (log CFU/ml) of four avian strains and for Staphylococcus aureus in contact with increasing concentrations of buttermilk total peptide extract
*For a given strain, values that are followed by different superscript letters are significantly different (P < 0·0005)
Bacterial viability (log CFU/ml) of four avian strains and for Staphylococcus aureus in contact with increasing concentrations of buttermilk cationic peptide extract
*For a given strain, values that are followed by different superscript letters are significantly different (P < 0·0005)
Antimicrobial activity of lactoferrin peptide extracts
The antimicrobial activity of the lactoferrin TPE and CPE against the five selected pathogens is presented in Tables 3 and 4. Three strains (E. coli O78:H80, Salm. Enteritidis, and Staph. aureus) were significantly sensitive to the action of the lactoferrin TPE (Table 3); bacterial viability losses of more than 3 log CFU/ml were reached in the presence of peptides at 1·3 mg/ml. Both E. coli TK3 O1:K1 and Salm. Typhimurium were more resistant, with bacterial counts decreasing by 3·9 and 2·5 log CFU/ml at 5 mg/ml, respectively. Lactoferrin CPE (Table 4) was the most active peptide extract. The concentration of the lactoferrin CPE can be lowered by two to more than four times, depending on the strain, to obtain a killing effect similar to the one obtained with the TPE. The antimicrobial activity of the synthetic LfcinB also varied with the strains, with bacterial losses ranging from 1·5 ± 0·36 log CFU/ml for Salm. Enteritidis and Staph. aureus to 2·7 ± 0·36 log CFU/ml for E. coli O78:H80 at 0·025 mg/ml, to more than 4 log CFU/ml at 0·05 mg/ml.
Bacterial viability (log CFU/ml) of four avian strains and for Staphylococcus aureus in contact with increasing concentrations of lactoferrin total peptide extract
*For a given strain, values that are followed by different superscript letters are significantly different (P < 0·0005)
Bacterial viability (log CFU/ml) of four avian strains and for Staphylococcus aureus in contact with increasing concentrations of lactoferrin cationic peptide extract
*For a given strain, values that are followed by different superscript letters are significantly different (P < 0·0005)
Discussion
Several studies already showed that potent AMP can be generated from skim milk (review by Sibel Akalın, Reference Sibel Akalin2014; Théolier et al. Reference Théolier, Fliss, Jean and Hammami2014) and from membrane-associated MFGM proteins (Clare et al. Reference Clare, Zheng, Hassan, Swaisgood and Catignani2008). Buttermilk, a by-product of the butter-making process, was selected as the starting material for the production of peptide extracts because it is essentially skim milk containing a large amount of MFGMs (Dewettinck et al. Reference Dewettinck, Rombaut, Thienpont, Le, Messens and Van Camp2008). Lactoferrin, extensively studied for its bactericidal activity, was used here as a control to validate the method of preparation. The purpose of this study was to propose an easy, efficient, and scalable method to increase the specific antimicrobial activity of peptide extracts and to test their activity against four avian pathogens as well as Staph. aureus. The peptide extracts from buttermilk and lactoferrin were produced according to the protocol described. Pepsin-treated protein preparations were subjected to UF to obtain low-molecular-weight peptide fragments, namely the TPE. In this study, UF replaced heat treatment (80 °C) and high-speed centrifugation, proposed by Bellamy et al. (Reference Bellamy, Takase, Yamauchi, Wakabayashi, Kawase and Tomita1992b) to stop enzymatic activity and precipitate denatured proteins. This change was made because UF is a more convenient method for industrial applications. On a protein basis, about 45–50% of the starting material was recovered as buttermilk TPE. The TPE were fractionated on the basis of their charge using MAC, which is an interesting alternative to conventional resin-based chromatography for the large-scale production of biomolecules because MAC is simple and fast, and thus high flow rates can be used. Moreover, robust ready-to-use modules are available in various sizes, allowing scale-up for industrial production. This approach was previously applied in the dairy sector to fractionate whey proteins (Teeters et al. Reference Teeters, Root and Lightfoot2002; Plate et al. Reference Plate, Beutel, Buchholz, Demmer, Fischer-Frühholz, Reif, Ulber and Scheper2006; Kreub & Kulozik, Reference Kreub and Kulozik2009; Voswinkel & Kulozik, Reference Voswinkel and Kulozik2014) and to separate the glycosylated and aglycosylated populations of milk caseinomacropeptide (Robitaille, Reference Robitaille2013). The fractionation of lactoferrin TPE has also been conducted in this study essentially to validate the peptide production method using Mustang S capsules and to evaluate the antimicrobial efficiency of the resulting CPE. The results showed that cationic peptides, including LfcinB, can bind to the cationic membrane and were eluted with a high salt concentration. The amount of buttermilk CPE, 6% of the TPE, was about two times lower than the amount of lactoferrin CPE produced from the lactoferrin TPE. This difference was expected, given that lactoferrin has a high pI of 8·7, with about 35% cationic peptides (pI > 9) as estimated from the work of Chan & Li-Chan (Reference Chan and Li-Chan2007). The mean pI of buttermilk proteins is below 6. Moreover, it should be emphasised that a desalting step after MAC was added to the procedure because AMP are known to be affected in vitro by monovalent and divalent ions (Bellamy et al. Reference Bellamy, Takase, Wakabayashi, Kawase and Tomita1992a); this step would not be necessary if the peptide extracts are to be used as food supplements. Without a desalting step, a higher yield would be expected.
Almost all antimicrobial activity detected with buttermilk samples were recovered into CPE, indicating that antimicrobial activity was associated mainly with cationic peptides; neither the APE nor the unbound fraction was active at a concentration of 5 mg/ml. This was expected, given that the vast majority of antimicrobial peptides are cationic (Brogden, Reference Brogden2005). In contrast to the bacterial survival of the lactoferrin-CPE-sensitive strain, for which a dramatic viability decrease occurred within a short range of concentrations, the antimicrobial effects of the buttermilk CPE increased gradually until the maximum tested concentration, 5 mg/ml, was reached. This can be explained by the peptide solubility at high concentrations (Agyei & Danquah, Reference Agyei and Danquah2011), that likely compromised antimicrobial activity.
The lactoferrin TPE was more efficient than the buttermilk TPE and CPE, given that the bacterial counts were close to 0 for three of the five strains at concentrations between 1·3 and 5 mg/ml. In the literature, lactoferrin TPE is bactericidal at concentrations that varied from study to study. Bellamy et al. (Reference Bellamy, Takase, Wakabayashi, Kawase and Tomita1992a) reported antimicrobial activity at 0·1 mg/ml. Concentrations from 0·25 to 0·75 mg/ml (Ellison & Giehl, Reference Ellison and Giehl1991; Dionysius & Milne, Reference Dionysius and Milne1997; Shin et al. Reference Shin, Yamauchi, Teraguchi, Hayasawa, Tomita, Otsuka and Yamazaki1998; Branen & Davidson, Reference Branen and Davidson2000) and up to 2 mg/ml (Murdock & Matthews, Reference Murdock and Matthews2002) can also be found. This variability was due to various factors: the preparation of the extract, the species and/or strains used, the growth phase of bacteria, the culture media and the method used for the assay. In the present study, the antimicrobial assay was modified from the assay described by Bellamy et al. (Reference Bellamy, Takase, Yamauchi, Wakabayashi, Kawase and Tomita1992b). As PYG is a pour media that hardly support growth for some strains, the bacterial suspension was diluted two times with twice-concentrated blood-heart infusion broth after the 2 h treatment, in order to enrich the medium and inactivate antimicrobial activity. However, this dilution in a rich media probably allowed some recovery of bacterial viability. Consequently, the antibacterial efficiency was underestimated in this study compared to the others. Incidentally, the antimicrobial activity for Lfcin, the reference AMP, recorded in this study was lower than the activity obtained by Bellamy et al. (Reference Bellamy, Takase, Wakabayashi, Kawase and Tomita1992a) for E. coli species, confirming that our approach underestimates bacterial viability.
The lactoferrin CPE was the most active extract, killing four strains at more than 1·3 mg/ml, and Salm. Typhimurium viability reached only 2·88 log CFU/ml with CPE at 5 mg/ml. Furthermore, the concentration of TPE had to be two to more than four times higher than the concentration of CPE to achieve similar bactericidal effects. All the tested avian pathogens were sensitive to buttermilk CPE, with the bacterial counts decreasing by at least 1·6 log CFU/ml at 5 mg/ml. Like the effects of the lactoferrin extracts, the effects of the buttermilk CPE varied with species and strains. Bacteria can use several defence mechanisms against antimicrobial peptides (Joo & Otto, Reference Joo and Otto2015), including the production of secreted proteases, the modification of lipopolysaccharide lipid A and membranes to reduce their negative surface charge, and the secretion of negatively charged exopolysaccharides, such as colonic acid produced by E. coli and Salm. Typhimurium (Foschiatti et al. Reference Foschiatti, Cescutti, Tossi and Rizzo2009).
Conclusion
Membrane adsorption chromatography using Mustang capsules was effective for the fractionation of peptides from buttermilk and lactoferrin hydrolysates, with the antimicrobial activity in CPE concentrated using two steps that can be easily scaled up at the industrial level. All avian pathogens tested in this study were sensitive to the buttermilk CPE at various levels. As noted by Sibel Akalın (Reference Sibel Akalin2014) for isracidin, another milk AMP, the antimicrobial activity could be low in vitro but very efficient in animal model because AMP are associated with other activities such as immunostimulation and antiadhesion in support of nonspecific resistance to infections (Hernández-Ledesma et al. Reference Hernández-Ledesma, García-Nebot, Fernández-Tomé, Amigo and Recio2014). It would be important to verify antimicrobial activity in avian models before concluding on the effectiveness of such peptide extracts. Those results are very promising for the development of functional ingredients from buttermilk cationic extracts for use in poultry feed as an alternative or a complement to antibiotics.
This study was supported by Agriculture and Agri-Food Canada's research programme.
