Microbiological Research 166 (2011) 1—13 www.elsevier.de/micres Properties of different lactic acid bacteria isolated from Apis mellifera L. bee-gut M. Carina Audisioa,b,c,, Marı́a J. Torresa,b, Daniela C. Sabatéa,b, Carolina Ibargurena,b, Marı́a C. Apellab,d a Instituto de Investigaciones para la Industria Quı́mica (INIQUI-CONICET), A4402FDC- Salta, Salta, Argentina Consejo Nacional de Investigaciones Cientı́ficas y Técnicas (CONICET), Argentina c Facultad de Ingenierı́a, Universidad Nacional de Salta; Salta, Argentina d Centro de Referencia para Lactobacilos (CERELA-CONICET); 4000-San Miguel de Tucumán, Tucumán, Argentina. b Received 11 November 2009; received in revised form 30 December 2009; accepted 3 January 2010 KEYWORDS Lactobacillus johnsonii; Apis mellifera L.; Lactic acid bacteria; Bee-gut microbiota Summary Eight strains belonging to Lactobacillus spp. and five to Enterococcus spp. were isolated from the gut of worker Apis mellifera L. bees. Studies based on 16S rRNA sequencing revealed that AJ5, IG9, A15 and CRL1647 strains had a 99% identity with Lactobacillus johnsonii, while SM21 showed a 99% similarity with Enterococcus faecium. L. johnsonii CRL1647, AJ5 and IG9 were high lactic acid producers (values were between 177 and 275 mM), and in vitro they inhibited different human food-borne pathogens and Paenibacillus larvae, the American foulbrood agent. This bacterium was the most sensitive to the lactic acid effect being inhibited by 44 mM of this metabolite. L. johnsonii CRL1647, AJ5 and IG9 also presented important surface properties. These cells showed between 77% and 93% of auto-aggregation. The preliminary study of the chemical nature of the aggregating factors revealed that the molecules involved in the surface of each L. johnsonii strain were quite complex; and something of a peptidic nature was mainly involved. E. faecium SM21 produced bacteriocin-like compounds with anti-Listeria effects. Furthermore, a band close to 6.0–7.5 kDA was detected by SDSPAGE studies, and the entA, B and P structural genes were amplified by PCR reactions. For the first time, bee-gut associated L. johnsonii and E. faecium strains have been isolated, identified, cultivated and some of their functional properties reported. & 2010 Elsevier GmbH. All rights reserved. Corresponding autor at: INIQUI (Instituto de Investigaciones para la Industria Quı́mica), Universidad Nacional de Salta (UNSa), Food Microbiology, Av. Bolivia 5150, A4402FDC-Salta, Argentina. Tel./fax: þ54 387 4251006. E-mail address: [email protected] (M. Carina Audisio). Introduction The microbiota associated to the honeybee Apis mellifera mellifera L. is complex and far from being fully understood, or even known. It has been 0944-5013/$ - see front matter & 2010 Elsevier GmbH. All rights reserved. doi:10.1016/j.micres.2010.01.003 2 described as being mainly composed of yeasts, Grampositive bacteria (such as Lactobacillus rigidus apis, L. constellatus, Bacillus spp., Streptococcus and Clostridium), and Gram-negative or Gram variable bacteria (Achromobacter, Citrobacter, Enterobacter, Erwinia, Escherichia coli, Flavobacterium, Klebsiella, Proteus and Pseudomonas) (Rousseau et al., 1969; Gilliam and Morton, 1978; Gilliam and Prest, 1987; Rada et al., 1997; Mohr and Tebbe, 2006). Gilliam (1997) reported that these bacteria were likely to be endemic in the alimentary tract of adult bees and were neither dependent on seasonal nor food factors. Glinski and Jarosz (1995), also suggests that this bees’ normal microbiota comes from pollen consumption or through contact with older bees. The presence of intestinal enterobacteria in hibernating honeybees was also extensively discussed by Lyapunov et al. (2008). Interestingly, only five studies have been carried out regarding the presence of lactic acid bacteria associated with honeybee microbiota, and their isolation and identification have yielded inconclusive results (Rada et al. 1997; Gilliam, 1997; Máchová et al., 1997; Kačániova et al. 2004, Olofsson and Vásquez 2008). In some cases, bacteria considered as probable members of Lactobacillus group, produced catalase and reduced nitrate (Gilliam, 1997). These properties are negative for this genus (Hammes et al. 1992). Kačániova et al. (2004) could not detect Lactobacillus spp. at all. Thus, only these scientific articles report the presence of culturable bee-gut Lactobacillus, Bifidobacterium or Enterococcus strains associated with the intestinal tract of the honeybee. Lactic acid bacteria are important inhabitants of the intestinal tract of man and other mammalian and vertebrate animals. Lactobacillus and Enterococcus are members of this family and are also present in food and fermentation processes. These microorganisms disclose interesting properties not only for the food industry but also for health (Naidu et al., 1999). The antimicrobial potential of these bacteria includes, among others, the synthesis of compounds such as lactic acid, short-chain-volatile-fatty acids, and bacteriocinlike molecules (Jack et al. 1995; Wilson et al. 2005). Antagonistic studies are generally directed towards food-spoilage and/or pathogenic microorganisms related to the host or product from which the lactic acid bacteria were isolated. Different microorganisms affect the honeybee and the most serious bacterial agent may be Paenibacillus larvae, a sporulated Gram-positive bacillus that causes the American foulbrood disease in larvae (Williams, 2000). An indiscriminate antibiotic use against this bee pathogen has brought about not M. Carina Audisio et al. only chemoresistant strains, but also honey contamination (Evans, 2003). The aim of this work was to isolate, characterize, cultivate, preserve and pheno and genotypically identify lactic acid bacteria associated to bee gut in order to improve the understanding of this microbial environment. Different properties were also analyzed such as their surface characteristics and in vitro antimicrobial spectra against the bacterial bee pathogen, and in a broader sense, against human foodborne pathogen bacteria. Materials and methods Honeybee intestinal tract sampling Intestine samples of summer adult worker bees from different hives of El Galpón (Salta, Argentina) were used to obtain homogenates in physiological solution (0.85% w/w NaCl). The whole intestinal tracts (esophagus to rectum) were analyzed and aseptically handled to avoid contamination with the external surface of the bee body. Culture media and growth conditions for lactic acid bacteria isolation Known aliquots of the different homogenates were sown on the Streptococcus selective medium (MSS agar) devised in our laboratory (tryptone 1.4%; glucose 0.5%; sodium citrate 0.1%; NaCl 0.4%; sodium azide 0.022%; meat peptone 0.5%; sodium sulphite 0.022%; L-cysteine 0.022%; pH 6.5; Audisio et al., 2005), on MRS agar (Britania, Argentina) and on brain-heart infusion agar (BHI, Britania, Argentina). The plates were incubated at 37 1C for 24–48 h except for the MRS plates which were incubated at 37 1C for 48–72 h under microaerophilic conditions. In all cases, the incubation period was extended until visible colonies appeared. White, small, round colonies from the MSS and BHI agar were recovered and placed into BHI broth and incubated at 37 1C. Colonies with similar characteristics from MRS agar were transferred to MRS broth. Lactic acid bacteria phenotypic characterization Morphological observations were carried out using Gram staining. Carbohydrate fermentation patterns (APICH50, Biomérieux) and phenotypical identification according to basic biochemical tests were carried out for Lactobacillus (Hammes et al., 1992). Enterococcus strains were also characterized Properties of different lactic acid bacteria according to typical biochemical tests for this genus (Devriese et al., 1992). Antimicrobial spectrum analyses Antimicrobial substance synthesis by the isolated lactic acid bacteria was analyzed in the cell-free supernatants (CFS), obtained from cultures of the lactobacilli (strains AJ5, A15, IG9 and CRL1647) and enterococci (strains SM21 and CA12) in MRS and BHI broths, respectively. After 24 h of incubation at 37 1C and under the appropiate atmosphere for each strain, the CFS were recovered by centrifuging (10,000g, 10 1C, 15 min), filter-sterilized (0.22 mm) and kept at 4 1C until used. Listeria monocytogenes 99/287 was provided by Instituto ‘‘Dr. Carlos Malbrán’’ (Buenos Aires, Argentina). Bacillus cereus C1 was supplied by Dr. Morea (ISPA, Bari, Italy). Escherichia coli 0157:H7 was supplied by Dr. Raya (CERELA, Argentina). Staphylococcus aureus 29213 from ATCC (American Type Culture Collection). Six different P. larvae strains (I, II, lll, IV, Azul and C) were provided by Dr. Terzolo and Engineer Borracci (INTA-Balcarce, Argentina). P larvae 7 and 35 were from Unidad de Bacteriologı́a of CIDEFI (La Plata, Buenos Aires – Argentina). P. larvae strains were activated in MPYGP (Dingman and Stahly, 1983) agar at 37 1C for 72 h in a microaerophilic atmosphere. The other indicator strains were grown in BHI broth at 37 1C for 24 h. All strains were kept at 20 1C in MRS or in BHI broth plus glycerol (20%). The different CFS were split off in two fractions: one was kept unmodified and the other was adjusted to pH 6.0 with NaOH 2.0 N; both aliquots were filter-sterilized (0.22 mm pore size cellulose acetate filter) and kept at 4 1C until used. Different lawns of the human pathogen strains were obtained on BHI agar (1.5%) for the agar diffusion assay. P. larvae cell suspensions were prepared by recovering the colonies grown on MPYGP agar, with a sterile cotton swab soaked in buffered-peptone water. They were resuspended in 2 mL of MPYGP broth (ca. 1  107 CFU/mL) and 300 mL were inoculated in 10 mL of molten MPYGP agar. Twenty five microliters of each CFS were placed in wells (5 mm diameter) cut in agar plates previously sown with the indicator strains. The plates were incubated at 30 1C for 12–24 h and examined for inhibition halos. For the microplaque technique, 100 mL of each CFS fractions were placed into 96 wells microplaque and 10 mL of the different indicator strain cell suspensions (final concentration varies from 106 to 108 CFU/mL) were added. Then, the microtitulation plaques were incubated at 37 1C for 2, 4 and 24 h under the optimal conditions for the tested 3 bacteria. Lack of turbidity in the wells sown with the pathogen indicated its inhibition by the sample under analysis. Peptone–water (0.1%) dilutions were made and viable indicator cell number was determined by plate counting. In the particular case of P. larvae, aliquots of 5 ml of those suspensions that presented slight turbidity were sown on MPYGP agar and the plates were incubated in order to determine P. larvae residual viability. All assays were carried out in triplicate. Phylogenetic analyses of selected bacteria Lactobacilli CRL1647, A15, AJ5 and IG9 and the Enterococcus faecium SM21 were selected for genotypical characterization and further analyses. Their total genomic DNA were extracted from overnight MRS or BHI cultures by alkaline lysis (Pospiech and Neumann 1995). Two oligonucleotides, based on the report of Daffonchio et al. (1998), were used to determine 16S rDNA of the isolate: S-D-BACT-1494-A-S-20 (GTCGTAACAAGGTAGCCGTA) and L-D-BACT-0035-A-A-15 (CAAGGCATCCACCGT). PCR amplification was performed in a 25 mL reaction mix containing 2.5 mL 10X STR reaction buffer (Promega), 20 ng total DNA, 0.5 mM of each primer and 1 U of Taq DNA polymerase (Promega). Amplification consisted of an initial denaturation step at 94 1C for 5 min, followed by 35 cycles at 94 1C for 1 min, 50 1C for 2 min and 72 1C for 2 min. A final extension step consisting of 7 min at 72 1C was included (Sabaté et al., 2009). The PCR amplification fragments were resolved by agarose 0.8% gel electrophoresis at 75 V for 1 h 30 min. The gel was stained with ethidium bromide, and the bands were visualized under UV illumination at 254 nm. The 16S rDNA amplified with PCR was purified and the DNA sequencing carried out by Macrogen Services. Sequences were compared and aligned with sequences from the GenBank database by using the BLAST program of the National Center for Biotechnology Information (NCBI; http://www. ncbi.nlm.nih.gov) network server. Phylogenetic trees, based in the 16S-rRNA gen sequences (1,339 bp) of the isolate and culturable lactic acid bacteria were constructed using the Tree Builder program of the Ribosomal Database Project II (http://rdp.cme.msu. edu) (Cole et al. 2009). The neighbour-joining tree was bootstrapped 1000 times. Lactobacillus acidophilus (acc. Number 780913) was used as an outgroup. Physicochemical characterization of the antagonistic compounds The concentration of all organic acids was determined by HPLC. Prior to the assay, CFS from 4 MRS (lactobacilli) and BHI (enterococci) cultures were deproteinized. In brief, 1 mL of sample was added to 2 mL of a solution in 1.8% Ba(OH)2, adding 2 mL of a solution of 2% ZnSO4. The mixture was vortex shaken, allowed to settle for 10 min and centrifuged at 3500g for 5 min. Each supernatant was filter sterilized (0.45 mm) before the HPLC analyses. The sample amount injected was 20 mL. The column temperature was 55 1C and the flow rate of the H2SO4 10 mM mobile phase was 0.6 mL/min. Detection was carried out by determining the refraction index using a 2142 LKB Differential Refractometer. The chromatography column (Rezec Organic Acid, Phenomenex) had a diameter of 7.8 mm and a length of 300 mm. Peak profiles, integration, and quantification were obtained with a CR601 Shimadzu chromatopac integrator (Shimadzu Corporation, Analytical Instrument Division, Kyoto, Japan). All tests were carried out in duplicate. Bacteriocin-like substances production Bacteriocin-like substances production was determined by the well-diffusion assay with L. monocytogenes 99/287 as the indicator strain. Trypsin, proteinase K, pronase E, a-chymotrypsin and pepsin (final concentration of 1 mg/L in phosphate buffer pH 7.00, 0.05 M), were assayed to test the proteinaceous nature of the inhibitors. Catalase, a-amylase and lipase effects (final concentration of 1 mg/L in phosphate buffer pH 7.00, 0.05 M), were also analyzed. Heat resistance was studied after CFSs were heated to 70 1C for 30 min, 100 1C for 15 min and 121 1C for 15 min (in an autoclave) while another aliquot without thermal treatment was used as control. Then, they were cooled and tested for antimicrobial activity. To determine bacteriocin molecular size, the peptides present in the CFS of E. faecium SM21 were separated by Tricine-SDS-PAGE as described by Schäger (2006). Polyacrylamide concentration on the separating gel was 16% (w/v). Electrophoresis was conducted at a constant voltage of 50 V for 1 h, 75 V for 2 h and 150 V for 1 h. A low rainbow molecular weight marker with sizes ranging from 3.5 to 38 kDa (Amersham Pharmacia, Germany) was used. The gel was fixed with an isopropanol (20% v/v) – acetic acid (10% v/v) solution, exhaustively washed with distilled water, and overlaid on BHI agar inoculated with L. monocytogenes 99/287. The presence of enterocin-encoding genes was studied by PCR amplification with primers for the following well-known enterococcal bacteriocins: enterocin A; enterocin B; enterocin P; enterocin 31 and enterocin L50A/B (du Toit et al., 2000). M. Carina Audisio et al. Total genomic DNA from strains was used. PCR products were analyzed by electrophoresis (85 V for 1 h 5 min) on 2% (w/v) agarose gels in 1  TAE buffer. Gels were stained GelRed (Invitrogen) and observed under UV light. The sizes of the amplified fragments were determined using 100-bp DNA Ladder (Invitrogen) as a molecular weight marker. As a positive control, the entA, B and P bacteriocingene sequenced of E. faecium CRL1385 were used (accession numbers to the GeneBank database GQ369790, GQ369791 and FJ57726, respectively). Lactobacillus cell surface properties When AJ5, IG9 and CRL1647 cells were grown in an MRS (Britania, Argentina) medium, discernible cell grouping appeared to the naked-eye, which adhered to the tube walls or sedimented to the bottom of the tube, due to this, their aggregation properties were studied. These properties were analyzed according to Handley et al. (1987) with some modifications. Lactobacilli cells were washed twice with PBS (NaCl 0.8; KH2PO4, 0.034; K2HPO4, 0.121, pH 7.4), resuspended in the same buffer and adjusted to OD600=0.6070.05. Cell suspensions (2 mL) were centrifuged at 2000g for 15 min and optic density (OD) at 600 nm was measured, without agitation, at room temperature every hour for 4 h. The percentage of autoaggregation was expressed as: A%=(A0At)/A0  100, where A0 represents the absorbance (A600) at 0 h and At represents the absorbance (A600) taken at 1, 2 and 3 hours. Preliminary physic-chemical studies of molecules involve in the lactobacilli surface interactions In order to approach the chemical nature of the molecules involved in the surface adsorption phenomena, L. johnsonii cells were put in contact with solutions of either 1 mg/mL of trypsin and pepsin at 37 1C for 1 h, lipase and sodium mperiodate at 25 1C for 1 h, and heated to 100 1C for 15 min. All enzymes were from Sigma and were prepared in their optimal buffer. Following this, the aggregation assay was performed. Statistical analyses All analyses were carried out according to the Tukey test and the results considered significant at the Po0.05 level. Assays were carried out in duplicate. Properties of different lactic acid bacteria 5 Results Antimicrobial spectrum Screening and phenotypic characterization of lactic acid bacteria associated to bee-gut The agar-diffusion assay was used to ‘‘quick screen’’ the antimicrobial spectra of the 13 lactic bacteria strains isolated from the bee gut. It was observed that only Lactobacillus AJ5, IG9, A15 and CRL1647 were able to inhibit S. aureus ATCC29213, B. cereus C1, E. coli O157:H7, L. monocytogenes and P. larvae; but strain A15 failed to inhibit two of the eight P. larvae strains tested. While, E. faecium SM21 only exhibited anti-L. monocytogenes activity. The lactobacilli antagonistic effects disappeared in all cases after adjusting CFS pH to 6.0, but the coccus kept its antilisteria activity under this new condition (see Table 2). Different types of colonies developed on the surface of BHI agar plate after 1 day of incubation at 37 1C, and after 48 or 72 h on the surface of MRS agar. Thirteen microorganisms were purely isolated and showed biochemical characteristic of lactic acid bacteria: small, round, opaque and white colonies, non-sporulated, cocci (5) or rods (8) Gram-positive, non-motile cells, catalase negative and nitrate negative. Classic biochemical tests revealed that the eight bacilli were homofermentative and belonged to Lactobacillus genus. The strains CRL1647 and IG9 were characterized as acidophilus and crispatus, respectively; but the species of the other strains could not be determined and they were designated as Lactobacillus spp. The cocci were grouped into the Enterococcus genus. Basic biochemical tests revealed that, at least, SM21, CA12 and SM7 can be considered as E. faecium (Table 1). Phylogenetic identification and analyses of selected isolates Five strains were selected due to their antimicrobial spectra and their species-specific identifications were derived using a 16S-rRNA sequence analysis. Comparison of the near complete 16S rRNA gene sequence of the bee-gut associated CRL1647, AJ5, IG9 and A15 Table 1. Characterization of the most representative lactic acid bacteria associated with summer adult worker bee gut samples. Isolates Characters Tentative identification IG9, CRL1647, AJ5, A15 C22, SM16, SM18, CA22 Gþ, long rods, non motile, non sporulated, nitrate (), Catalase (), arginine (þ), esculin (þ), growth at 37 1C (þ), grwoth at 45 1C (þ), glucose (þ) without gas production Lactobacillus spp. IG9, CRL1647 AJ5, A15 Gluconate (), esculin (þ), melezitose (), sacarose (þ), Glycerol (), fructose (þ), mannose (þ), cellobiose (þ), Rhamnose ()a, sorbitol (þ), manitol (þ), Darabinose (þ) Salicin (wþ), trehalose (þ), xylitol (), L-fucose (þ), Gentibiose (þ), maltose (þ), arbutine (þ), tagatose (þ), Raffinose (þ), amygdaline (þ), melibiose ()a, lactose ()a L. acidophilus CRL1647 L. crispatus IG9 Lactobacillus spp. A15 Lactobacillus spp. AJ5 CA12, SM21, SM7, G(þ), cocci in short chains, immobile, non sporulated, nitrate () Catalase (), growth at 15, 37 and 451C (þ), growth at pH 9.6 (þ); Growth at 6.5% NaCl (þ), Voges Proskauer (), arginine (þ), Esculin (þ), glucose (þ) Enterococcus spp. Cellobiose (þ), melibiose (þ), melezitose (), sacarose (), Starch (þ), glycerol (þ), L-arabinose (þ), ribose (þ), fructose (þ), Galactose (þ), manitol (þ), D-mannose (þ), rhamnose (þ), Sorbitol (þ), lactose (þ), salicin (þ), trehalose (þ), xylitol (), D-arabinose (þ), gentibiose (þ), maltose (þ), arbutine(þ), Amygdaline (þ), gluconate (þ) Enterococcus faecium CA10, CA11 CA12, SM21, SM7 þ, positive; wþ, weak positive; , negative. (þ) for AJ5 and A15 strains. a 6 Table 2. M. Carina Audisio et al. Property profile of the selected lactic acid bacteria isolated from the bee-gut of Apis mellifera L. Phylogenetic identification Genbank accession code Lactic acid synthesis (mM) Bacteriocin production Surface property Inhibitory spectrum: P. larvae L. monocytogenes S. aureus ATCC29213 B. cereus C1 E. coli O157:H7 CRL1647 AJ5 IG9 A15 SM21 L. johnsonii EU428007 17777 – þ L. johnsonii EU428008 27578 – þ L. johnsonii EU780913 18575 – þ L. johnsonii EU428009 10678 – – E. faecium EU428012 3475 þ – 8/8 3/3 þ þ þ 8/8 3/3 þ þ þ 8/8 3/3 þ þ þ 6/8 3/3 þ þ þ 0/8 3/3 – – – Figure 1. Phylogenetic tree, based in the 16S-rRNA gen sequences (1339 bp) of the four L. johnsonii strains culturable lactic acid bacteria from bee-gut, was constructed using the Tree Builder program of the Ribosomal Database Project II (http://rdp.cme.msu.edu) (Cole et al., 2009). strains with the corresponding Lactobacillus sequences from the GenBank database showed all strains lay in the evolutionary clade of Lactobacillus. The four strains had a 99% identity with known Lactobacillus johnsonii, and their sequences were introduced into the GenBank database (Table 2). A phylogenetic tree of these lactobacilli was built with the neighbourjoining algorithms, using different sequences from the L. acidophilus group with, at least, 95% identity as shown in Figure 1. L. johnsonii NCC533, a type strain with its complete genome known, was also included in this analysis. It can be seen that the bee-gut associated strains are closely related to L. johnsonii. A similar analysis was done with SM21 strain confirmed as an E. faecium with a 99% identity with E. faecium strain DSM20477 (=ATCC19434), a type strain, and other known and reported sequences (see Table 2 and Figure 2). Lactic acid synthesis and antagonistic effect analyses The lactic acid production by L. johnsonii CRL1647, IG9, AJ5, A15 and E. faecium SM21 was determined by HPLC. Interestingly, even though the pH values of the lactobacilli cultures were similar with an average value of 3.870.5; the lactic acid concentration profile was characteristic of each strain. L. johnsonii AJ5 synthesized the largest amount of this metabolite followed by IG9, CRL1647 and A15 (Table 2). While, E. faecium SM21, also a homofermentative lactic acid bacteria, acidified BHI broth at pH 4.5 and produced the lowest concentration of lactic acid. BHI broth was selected to culture Enterococcus spp. because it was determined in the laboratory that MRS broth was not as efficient for cocci growth. E. coli O157:H7, S. aureus ATCC29213, B. cereus C1 and the three L. monocytogenes assayed showed the same behaviour against lactic acid; all of them were inhibited by 177 mM of this acid (Table 2). But, 8875 mM had no impact on cell viability, independently of the detection method, i.e. welldiffusion agar or direct contact in microplaques. In particular, Figure 3A demonstrates the effect of 177 mM lactic acid synthesized by L. johnsonii CRL1647 on L. monocytogenes 99/287 nonproliferative cells viability. Properties of different lactic acid bacteria 7 7.5 9.5 7 9 6.5 8.5 log CFU/mL log CFU/mL Figure 2. Phylogenetic tree, based in the 16S-rRNA gen sequences (1339 bp) of the E. faecium SM21 culturable lactic acid bacterium from bee-gut, was constructed using the Tree Builder program of the Ribosomal Database Project II (http://rdp.cme.msu.edu) (Cole et al., 2009). 6 5.5 8 7.5 5 7 4.5 6.5 6 4 0 1 Time (h) 2 0 1 2 Time (h) Figure 3. Listeria monocytogenes 99/287 inhibition by effect of: (A) lactic acid produced by L. johnsonii CRL1647 (~ control; m 177 mM lactic acid at pH 6.0; ’ 177 mM lactic acid at pH 3.9) and (B) bacteriocin-like compounds produced by E. faecium SM21 (m control; ’ crude CFS at pH 6.0; ~ CFS at pH 6.0 and trypsin treated). The vegetative cells of P. larvae were the most sensitive to lactic acid action; no colonies were detected on MPYGP agar after a 2 h contact period with 44 mM of lactic acid. These results were confirmed by the microplaque technique when different aliquots of cultures, from wells of the microplaque with slight turbidity, showed a viable plate count number three orders less in the log of CFU/mL than the control (Figure 4). Bacteriocin-like compounds synthesis E. faecium SM7, SM21 and CA12 produced compounds with strong anti-L. monocytogenes activity. The antagonistic effect had no significant difference when the filtered cell-free supernatants were adjusted to pH 6.0 with NaOH. Also lipase, catalase and a-amylase did not affect the antiListeria activity, but disappeared after a treatment 8 M. Carina Audisio et al. 8 1 3 log CFU/mL 7 2 6 5 4 0 2 Time (h) 4 Figure 4. Paenibacillus larvae IV response in the presence of different CFS from L. johnsonii CRL1647. (A) By the welldifussion assay (1: pH 3.9 and 177 mM lactic acid; 2: pH 5 and 177 mM lactic acid, and 3: pH 5 and 44 mM lactic acid) and (B) by the microplaque technique ((~) control of P. larvae IV cells, (m) P. larvae IV cells in contact with 177 mM lactic acid; and (&) in contact with 44 mM lactic acid). with pepsin, a-chymotrypsin, trypsin, pronase E and proteinase K, confirming their peptidic nature. The compounds were also highly thermostable keeping their anti-listeria effect after autoclave sterilization (121 1C for 15 min). A bactericidal effect of E. faecium SM21 cellfree supernatant on L. monocytogenes 99/287 non-proliferative cells viability was observed (Figure 3B). Also, it was confirmed that the antagonistic effect was mainly due to bacteriocinlike compounds because when the CFS was treated with trypsin, its antilisteria effect disappeared (Figure 3B). In particular, the analysis of E. faecium SM21 antimicrobial peptides by Tricine-SDS-PAGE, and further antimicrobial assays against L. monocytogenes 99/287, showed a broad band with a molecular weight of approximately 6.0–7.5 kDa (Figure 5A). The analyses of different structural enterocin genes in the DNA of E. faecium SM21 by PCR reactions was also done and revealed that the entA, B and P genes were amplified (Figure 4B). The size of the fragments of entA (126 bp) and entB (168 bp) were identical to those obtained for E. faecium CRL1385, the strain used as the positive control. The specific PCR fragment of 120 bp, corresponding to the entP gene, was also amplified. Moreover, no specific PCR fragments were detected with the set of primers for bacteriocins L50A/B and 31 (Figure 5B). L. johnsonii surface properties The presence of cells adhered to the tube wall were observed only in three of the fourteen lactobacilli isolated from bee-gut, as shown in Figure 6A for L. johnsonii CRL1647. A similar situation was observed for L. johnsonii AJ5 and IG9 cultures. After 3 h, without agitation, a reduction in the OD (600 nm) values were determined from 24-hour-old cells of IG9 and AJ5 (93.2% and 83.4%, respectively) resuspended in PBS. While, L. johnsonii CRL1647 showed 77.3% of autoaggregation. By scanning microscopy analyses, lactobacilli cells were aligned like bricks in a wall and some random coils were also observed (Figure 6B). Flocculation was not suspected as this phenomena is more common in yeast than in bacteria cells (Table 3). The preliminary study of the chemical nature of the autoaggregating factors, revealed the molecules involved on the surface of each L. johnsonii strain were quite complex. All being sensitive to trypsin, sodium m-periodate and highly thermostable. But, only those of AJ5 were affected by pepsin. Also lipase action was more pronounced on IG9 cells autoaggregation (Table 4). Discussion The scientific works where the presence of lactic acid bacteria associated to honeybee-gut is re- Properties of different lactic acid bacteria 9 38 KDa 31 KDa 24 KDa 17 KDa 12 KDa 8.5 KDa 3.5 KDa 1 2 1 2 3 4 5 6 Figure 5. (A) Tricine-SDS-PAGE gel showing the zone of growth inhibition of L. monocytogenes 99/287 by the band of the CFS of E. faecium SM21. Lane 1: molecular mass marker; lane 2: CFS of E. faecium SM21; (B) Amplification of structural enterocin genes from DNA of E. faecium SM21. Lane 1: 100 bp Ladder (invitrogen), lane 2: ent A (126 bp), lane 3: ent B (162 bp), lane 4: ent P (120 bp), lane 5: ent bac 31 and lane 6: ent L50. Figure 6. Auto-aggregating phenomena of L. johnsonii CRL1647 cells. (A) macroscopic observation, (B) scanning electron microscopy. Table 3. Autoaggregation percentage of the different L. johnsonii cultures. Strain CRL1647 AJ5 4 IG9 Time (h) 0 1 2 3 33.170.5 36.670.6 31.671.1 36.871.2 49.670.8 81.771.1 59.170.9 71.470.4 92.170.8 77.370.5 83.470.7 93.271.1 ported are limited. They were carried out with different purposes. For example, Mohr and Tebbe (2006) revealed the presence of Lactobacillus in bee-gut only by a genetic analysis, but no isolations were carried out and the strains detected were uncultured. Mrázek et al. (2008) also analyzed the diversity of insect intestinal microbiota, without any strain isolation, and only reported the possible presence of Bifidabacterium sp. in honeyebees. Rada et al. (1997) were able to isolate Lactobacillus sp., Enterococcus sp. and some Bifidobacterium sp. when they were studying the microbiota in the honeybee digestive tract. However, they only characterized a Bifidobacterium asteroides by biochemical tests and analyzed the sensitivity of these bacteria to veterinary drugs. In this work, Lactobacillus and Enterococcus strains were isolated from the gut of the summer adult worker Apis mellifera L. bees of the Argentinean Northwest region. The Streptococcus 10 Table 4. M. Carina Audisio et al. Autoaggregation percentagea of the different L. johnsonii strains after different treatments. Treatment Control Trypsin Pepsin Lipase Sodium m-periodate Boiling temperature for 15 min a Strains CRL1647 IG9 AJ5 77.370.5 30.871.2 75.170.5 61.670.8 59.370.4 75.671.1 93.270.7 59.470.3 90.371.1 67.371.3 70.570.7 92.470.6 83.471.1 45.370.5 65.370.8 83.170.4 67.771.2 82.870.7 Each point corresponds to an average of at least four readings. selective medium (MSS), modified by Audisio et al. (2005), was selective only for Enterococcus isolation. However Lactobacillus isolation was somewhat difficult. If MRS agar was prepared adjusting the pH close to 5.0 with acetic acid, as indicated by the manufacturer, scant colony development was observed and sometimes, there was no colony at all. On the other hand, if MRS agar was prepared without an acidic final pH, i.e. equal to 6.5, a higher number of typical lactic acid bacteria colonies were detected, including many others. Thus, to isolate lactobacilli from the bee-gut environment, the samples were seeded both on MRS at pH 6.5 and on MRS at pH 5.0, and small, round, opaque and white typical lactic acid bacteria colonies were picked up from both media. The response of classic biochemical tests performed for Enterococcus (Devriese et al., 1992) allowed us to correctly identify some of the enterococci strains, confirmed by a phylogenetic analysis for the SM21 strain. The isolation of these microorganisms from bee-gut was an important step as there are only four scientific papers reporting their presence in the honeybee intestinal environment (Rousseau et al., 1969; Snowdon and Cliver, 1996; Rada et al., 1997; Kačániova et al., 2004) and with this information a clearer understanding of the bee-gut microbiota composition can be achieved. Although there were differences between the phenotypical characterization of strains CRL 1647, IG9, AJ5, A15 and their genotypical profile, they were not surprising, as authors had already reported similar situations (Berger et al., 2007). Other articles have shown that there are six Lactobacillus strains of distinct homology groups (A1, A2, A3, A4, B1 and B2) that respond phenotypically to the characteristics of the acidophilus species, but a study with molecular biology tools revealed that the species may be crispatus, gasseri, johnsonii, gallinarum, amylovorus or acidophilus (Johnson et al., 1980; Pot et al. 1993). All the lactobacilli preselected in this work were homo- fermentative, i.e., they do not produce gas from glucose. These results agree with those reported by Olofsson and Vásquez (2008) who isolated and characterized many homofermentative lactobacilli from the honeybee stomach. They observed that lactobacilli and bifidobacteria were predominant within the lactic microbiota associated to the honeybee stomach, but failed to report the presence of L. johnsonii or Streptococcus or Enterococcus in this bee environment. However, in this work it was determined, by phylogenetic analyses, that the homofermentative lactobacilli associated to bee-gut were closely related to L. johnsonii. In particular, the IG9 strain showed a 97% (ca. 400 bp) identity with the 50 end of L. johnsonii NCC533, a type strain with its complete genome known (Boekhorst et al., 2004). A high in vitro susceptibility of P. larvae vegetative cells to a low pH, unreported so far for this bacterium, was observed in this work. The three L. johnsonii strains selected were able to inhibit the vegetative cells of the American foulbrood agent by lactic acid synthesis and 44 mM of this compound were necessary. E. faecium SM21, cultured in BHI broth, reached a pH close to 5.0 and produced 34 mM of lactic acid, but this concentration was not enough to affect P. larvae viability. Among the pre-selected lactobacilli and enterococci, the final pH values for the CFS used in the antagonistic tests were close to 4.0 or 4.5 in most cases, but organic acid concentration was variable and not all lactic acid bacteria selected, particularly enterococci, inhibited E. coli O157:H7, S. aureus ATCC29213, B. cereus C1 or P. larvae, by acidity. Although lactic acid production is an essential characteristic of the bacteria studied here, the amount of this metabolite that is synthesized depends on each particular strain. Susceptibility to the antagonistic activity of lactic acid also varies among the pathogenic microorganisms (Ogawa et al., 2001; Wilson et al., 2005). So, the present in vitro study revealed that certain Lactobacillus species, such as L. johnsonii exert Properties of different lactic acid bacteria growth inhibition on the indicator microorganisms assayed. To the authors’ knowledge, this would be the first report on P. larvae growth inhibition due to acidity generated by lactobacilli isolated from bee gut. In vitro inhibition of P. larvae by Bacillus sp. bacteria has been previously reported (Evans and Armstrong, 2005; Alippi and Reynaldi, 2006; Sabaté et al., 2009). However, Bacillus sp. is a genus normally related to the honeybee environment (Gilliam, 1987, 1997; Snowdon and Cliver, 1996). The selected enterococci were remarkably active against L. monocytogenes. This inhibition was not due to lactic acidity but to a synthesis of bacteriocin-like compounds by four E. faecium of the nine isolated. These compounds were thermostable and showed both a proteinaceous nature and a bactericidal effect against L. monocytogenes 99/287; all these characteristics agree with those found for different enterocins produced by E. faecium isolated from other sources (FoulquiéMoreno et al., 2006). It has been informed that the structural gene of enterocin A is widely distributed among E. faecium strains, whereas that of enterocin B always occurs in the presence of enterocin A; besides, the high-frequency and variability of enterocin structural genes among enterococci of different origin by means of simple PCR reactions, without purification or sequentiation, was also reported (Poeta et al., 2007). This trend was confirmed in this paper and this would be the first case where these genes have been detected in enterococci associated with the bee digestive tract. Bacterial aggregation/surface property at bacterial cell envelope level can be determined and quantified by the reduction of optic density values at different times and should not be confused/ misunderstood with the flocculation phenomena (Handley et al., 1987; Kos et al., 2003). Flocculation is a phenomena more common/current for yeasts than bacteria. Furthermore, yeast flocculation is a non-sexual and reversible cell aggregation in which cells adhere to each other to form flocs (Zhao and Bai 2009). In this work, an important auto-aggregation was observed in three different L. johnsonii strains isolated from bee-gut. The assays carried out to determine the chemical nature of the factor involved in the lactobacilli auto-aggregation revealed in the cell envelope, different compounds for each strain; a lipopetide for IG9 and a glicopeptide for CRL1647 and AJ5. The involvement of proteinaceous bacterial surface compounds in lactobacilli adhesion has been analyzed and characterized for strains from human, fish and dairy product origins (Ventura et al., 2002; Ying et al., 2007). However, no reports exist on the 11 study of this property in Lactobacillus with regards to insects, furthermore no analysis has been performed on honeybee-derived Lactobacillus until now. Further studies are needed to determine which molecules are present in the autoaggregation phenomena. Even though this work analyzes human foodborne pathogens, it should be pointed out that the honeybee or its products are not E. coli O157:H7 or L. monocytogenes hosts or vectors. The origin of a strain was not considered to be an obstacle to ascertain its full antimicrobial potential; thus, the findings may open the way for different applications other than apiculture practices. Conclusions Lactobacillus and Enterococcus strains associated with adult worker bees were isolated and characterized to enhance the knowledge of microorganisms found in the intestinal environment of this insect. The presence of L. johnsonii has now been reported for the first time in the bee gut. Furthermore, this is the first report on P. larvae growth inhibition due to acidity generated by lactobacilli isolated from the bee gut ecosystem. Acknowledgements This paper was supported by Agencia Nacional de Promoción Cientı́fica y Técnica (ANPCyT, Grant PICTR890/06), CIUNSa (PI No 1725) and Consejo Nacional de Investigaciones Cientı́ficas y Técnicas (CONICET). M.C. Audisio and M.C. Apella are members of the Research Career of CONICET. The authors thank Dr. S.A. Cuozzo and Dr. R.R. Raya for the excellent assistance with the phylogenetic analyses of the isolated strains. References Alippi AM, Reynaldi FJ. Inhibition of the growth of Paenibacillus larvae the causal agent of American Foulbrood of honey bees by selected strains of aerobic spore forming bacteria isolated from apiarian sources. J Invertebr Pathol 2006;91:141–6. Audisio MC, Terzolo HR, Apella MC. Bacteriocin from honeybee beebread Enterococcus avium, active against Listeria monocytogenes. Appl Environ Microbiol 2005;71:3373–5. Berger BR, Pridmore D, Barretto C, Delmas-Julien F, Schreiber K, Arigoni F, et al. Similarity and differences in the Lactobacillus acidophilus group identified 12 by polyphasic analysis and comparative genomics. J Bacteriol 2007;189:1311–21. Boekhorst J, Siezen RJ, Zwahlen MC, Vilanova D, Pridmore RD, Mercenier A, et al. The complete genomes of Lactobacillus plantarum and Lactobacillus johnsonii reveal extensive differences in chromosome organization and gene content. Microbiol 2004;150: 3601–11. Cole JR, Wang Q, Cardenas E, Fish J, Chai B, Farris RJ, et al. The ribosomal database project: improved alignments and new tools for rRNA analysis. Nucl Acids Res 2009;37:D141–5. Daffonchio D, Borin S, Frova G, Manachini P, Sorlini C. PCR fingerprinting of whole genomes: the spacers between the 16S and 23S rRNA genes and of intergenic tRNA gene regions reveals a different intraespecific genomic variability of Bacillus cereus and Bacillus licheniformis. Int J Syst Bacteriol 1998;48:107–16. Devriese LA, Collins MD, Wirth R. The genus Enterococcus. In: Balows A, Trüper HG, Dworkin M, Harder W, Heinz Schleifer K, editors. The Prokaryotes, vol. II. New York: Springer-Verlag; 1992. p. 1465–81. Dingman DW, Stahly DP. Medium promoting sporulation of Bacillus larvae and metabolism of medium components. Appl Environ Microbiol 1983;46:860–9. du Toit M, Franz CM, Dicks LM, Holzapfel WH. Preliminary characterization of bacteriocins produced by Enterococcus faecium and Enterococcus faecalis isolated from pig faeces. J Appl Microbiol 2000;88: 482–94. Evans JD. Diverse origins of tetracycline resistance in the honey bee bacterial pathogen Paenibacillus larvae. J Invertebr Pathol 2003;83:46–50. Evans JD, Armstrong TM. Inhibition of the American foulbrood bacterium, Paenibacillus larvae larvae, by bacteria isolated from honey bees. J Apicult Res 2005;44:168–71. Foulquié-Moreno MR, Sarantinopoulos P, Tsakalidou E, De Vuyst L. The role and application of enterococci in food and health. Int J Food Microbiol 2006;1:1–24. Gilliam M, Morton HL. Bacteria belonging to the genus Bacillus isolated from honey bees, Apis mellifera, fed 2,4-D and antibiotics. Apidologie 1978;9:213–22. Gilliam M, Prest DB. Microbiology of feces of the larval honey bee, Apis mellifera. J Invert Pathol 1987;49: 70–5. Gilliam M. Identification and roles of non-pathogenic microflora associated with honey bees. FEMS Microbiol Lett 1997;155:1–10. Glinski Z, Jarosz J. Mechanical and biochemical defences of honey bees. Bee World 1995;76:110–8. Hammes WP, Weiss N, Holzapfel W. The genera Lactobacillus and Carnobacterium. In: Balows A, Trüper HG, Dworkin M, Harder W, Heinz Schleifer K, editors. The Prokaryotes, vol. II. New York: Springer-Verlag; 1992. p. 1535–94. Handley PS, Harty DWS, Wyatt JE, Brown CR, Doran JP, Gibbs ACC. A comparison of the adhesion, coaggregation and cell-surface hydrophobicity properties of fibrilar and fimbriate strains of Streptococcus salivarius. J Gen Microbiol 1987;133:3207–17. M. Carina Audisio et al. Jack R, Tagg H, Ray B. Bacteriocins of Gram-positive bacteria. Microbiol Rev 1995;59:171–200. Johnson JL, Phelps CF, Cummins CS, London J, Gasser F. Taxonomy of Lactobacillus acidophilus group. Int J Syst Bacteriol 1980;30:53–68. Kačániova M, Chlebo R, Kopernichý M, Trakocvická A. Microflora of the honeybee gastrointestinal tract. Folia Microbiol 2004;49:169–71. Kos B, Suskovic J, Vukovic S, Simpraga M, Frece J, Matosic S. Adhesion and aggregation ability of probiotic strain Lactobacillus acidophilus M92. J Appl Microbiol 2003; 94:981–7. Lyapunov Y, Kuzyaev RZ, Khismatullin RG, Bezgodova OA. Intestinal enterobacteria of the hibernating Apis mellifera mellifera L. bees. Microbiology 2008;77: 373–9. Máchová M, Rada V, Huk J, Smékal Fr. Development of probiotics for bees. Apiacta 1997;4:99–111. Mohr KI, Tebbe C. Diversity and phylotype consistency of bacteria in the guts of three bee species (Apoidea) at an oilseed rape field. Environ Microbiol 2006;8: 258–72. Mrázek J, Štrosová L, Fliegerova K, Kott T, Kopečný J. Diversity of insect intestinal microflora. Folia Microbiol 2008;53:229–33. Naidu AS, Bidlack WR, Clemens RA. Probiotic spectra of lactic acid bacteria (LAB). Crit Rev Food Sci Nutr 1999;39:13–26. Ogawa M, Shimizu K, Nomoto K, Tanaka R, Hamabata T, Yamasaki S, et al. Inhibition of in vitro growth of Shiga toxin-producing Escherichia coli O157:H7 by probiotic Lactobacillus strains due to production of lactic acid. Int J Food Microbiol 2001;68:135–40. Olofsson TC, Vásquez A. Detection and identification of a novel lactic acid bacterial flora within the honey stomach of the honeybee Apis mellifera. Curr Microbiol 2008;57:356–63. Poeta P, Costa D, Rojo-Bezares B, Zarazaga M, Klibi N, Rodrigues C, et al. Detection of antimicrobial activities and bacteriocin structural genes in faecal enterococci of wild animals. Microbiol Res 2007;162: 257–63. Pospiech A, Neumann B. A versatile quick-prep of genomic DNA from Gram-positive bacteria. Trends Genet 1995;11:217–8. Pot B, Hertel C, Ludwig W, Descheemaeker P, Kersters K, Schleifer KH. Identification and classification of Lactobacillus acidophilus, L. gasseri and L. johnsonii strains by SDS-PAGE and rRNA-targeted oligonucleotide probe hybridization. J Gen Microbiol 1993;139: 513–7. Rada V, Máchová M, Huk J, Marounek M, Duskova B. Microflora in the honeybee digestive tract: count, characteristics and sensitivity to veterinary drugs. Apidologie 1997;28:357–65. Rousseau M, Tysset C, Durand C. Presence of streptococci of the Lancefield D group in healthy working bees (Alpis mellifica L.). Interpretation of their presence in alimentary bacteriology. Bull Acad Vet Fr 1969;42:173–86. Sabaté DC, Carrillo L, Audisio MC. Inhibition of Paenibacillus larvae and Ascosphaera apis by Bacillus subtilis Properties of different lactic acid bacteria isolated from honeybee gut and honey samples. Res Microbiol 2009;160:193–9. Schäger H. Tricine-SDS-PAGE. Nature protocols 2006;1:16–22. Snowdon JA, Cliver DO. Microorganisms in honey. Int J Food Microbiol 1996;31:1–26. Ventura M, Jankovic I, Walker DC, Pridmore RC, Zink R. Identification and characterization of novel surface proteins in Lactobacillus johnsonii and Lactobacillus gasseri. Appl Environ Microbiol 2002;68:6172–81. Ying C, Lei W, Jiazhong L, Zhantao S, Liguo A. Identification and purification of a novel adhesion-associated protein in a new strain of Lactobacillus, L15, from 13 flounder (Paralichthys olivaceus). Vet Microbiol 2007; 122:116–22. Williams DL. Review. A veterinary approach to the European honey bee (Apis mellifera). Vet J 2000;160: 61–73. Wilson AR, Sigee D, Epton HA. Anti-bacterial activity of Lactobacillus plantarum strain SK1 against Listeria monocytogenes is due to lactic acid production. J Appl Microbiol 2005;99:1516–22. Zhao XQ, Bai FW. Yeast flocculation: new story in fuel ethanol production. Biotechnol Adv 2009;27: 849–56.