Two 2[5H]-Furanones as Possible Signaling Molecules in Lactobacillus helveticus

Abstract

Two 2[5H]-furanones, in association with medium-chain fatty acids, were released in whey by Lactobacillus helveticus exposed to oxidative and heat stresses. This species plays an important role in cheese technology, particularly for Swiss-type cheeses and Grana cheese. Moreover, it significantly contributes to cheese ripening by means of an early autolysis and the release of enzymes during processing. Experimental evidence of the involvement of the two 2[5H]-furanones, detected by a gas chromatography-mass spectrometry/solid-phase microextraction technique, in the autolysis phenomenon has been obtained. Zymograms performed by using renaturing sodium dodecyl sulfate-polyacrylamide gels were used to detect the bioactivity of the supernatants containing the two furanones on fresh cells of the same strain. In addition to bands corresponding to known autolysins, new autolysins were detected concomitant with the exposure of Lactobacillus helveticus to the supernatants, which can be regarded as conditioned media (CM), and to a commercial furanone, 5-ethyl-3-hydroxy-4-methyl-2[5H]-furanone (HEMFi), having spectral data similar to those of the newly described 2[5H]-furanones. Morphological changes were observed when fresh cells were exposed to CM containing the two 2[5H]-furanones and HEMFi. The two furanones produced by Lactobacillus helveticus, which met a number of criteria to be included in cell-cell signaling molecules, have a presumptive molecular mass lower than those of already known 3[2H]-furanones having an autolytic activity and being produced by gram-negative bacteria. Moreover, they present a different chemical structure with respect to the furanones already identified as products of Lactococcus lactis subsp. cremoris or to those identified in some cheeses with Lactobacillus helveticus as a starter culture.

Lactobacillus helveticus is a homofermentative thermophilic lactic acid bacterium that is widely used in the manufacture of cooked cheeses, and it is the dominant species during whey fermentation for Parmigiano Reggiano and Grana Padano cheese production where high temperatures (53°C) and low pH (3.3) are reached (7). Moreover, L. helveticus is largely used for other long-ripened cheeses such as Emmental and Provolone. The occurrence of appreciable amounts of furanones such as 2,5-dimethyl-4-hydroxy-3[2H]-furanone (DMHF) and 4-hydroxy-5-methyl-3[2H]-furanone (MHF), which are regarded as important key odorants, has been identified in cheeses whose production is associated with L. helveticus (15, 32, 37, 42). Moreover, it has been reported that the growth of Lactobacillus helveticus and Lactobacillus delbrueckii in aqueous suspensions of whey powder resulted in the formation of appreciable amounts of DMHF (38). DMHF and MHF have been detected in a wide range of foods fermented by fungi and lactic acid bacteria such as soya sauce, shoyu, miso, and beer (19, 20, 34, 46, 49).

Furanones are naturally occurring compounds which are associated with a variety of diverse biological phenomena. The high specificity and high sensitivity of the responses that they generate in some organisms are consistent with a key function in living cells. In fact, on the basis of the available evidence, a biological role in the signaling mechanisms between individual organisms has been proposed for furanones (46). The chemical properties of these compounds are ideal for the signaling mechanisms. In fact, several furanones are water and/or fat soluble or volatile depending on the substituents on the central ring (46). This last feature is also a prerequisite for air-mediated transmission. The extracellular production of DMHF, MHF, and a molecule called autoinducer-2 (AI-2), identified as furanosyl borate diester, has been reported for a variety of gram-negative bacteria including Vibrio fischeri, Escherichia coli, Salmonella enterica serovar Typhimurium, Helicobacter pylori, Mannheimia haemolytica, Pasteurella multocida, and Pasteurella trehalosi (13, 23, 30, 50, 55, 56). The latter molecule is able to induce bioluminescence in Vibrio harveyi. AI-2 is chemically formed from 4,5-dihydro-2,3-pentadione that is in turn generated by the action of LuxS on S-ribosylhomocysteine (43). On the other hand, MHF and DMHF have been shown to possess low AI-2 activity in Vibrio harveyi (43, 56). Halogenated furanones produced by the macroalga Delisea pulchra inhibit quorum sensing in some gram-negative species and have been proposed as potential antibacterial coatings on biomaterials (2, 11, 24).

Since MHF and DMHF are both known to be crucial for cheese flavor, particularly for Parmigiano Reggiano, the mechanisms regulating their release by starter or nonstarter microorganisms are potentially of significant biotechnological importance. The generation of these molecules by L. helveticus and/or other lactic acid bacteria during the Parmigiano Reggiano cheesemaking process suggests that their accumulation could be associated to the cell stress response. In fact, natural whey starter preparation, curd cooking, and ripening phases can be regarded, due to process conditions (temperature, pH, or water activity), as inhospitable environments (8). Likewise, in E. coli, the regulation of the AI-2 furanone seems to be a channeling condition of stress (10) and starvation towards the quorum-sensing circuit, presumably through the GroESL chaperonin complex.

The aim of this investigation was to establish whether L. helveticus releases MHF, DMHF, or other furanones as a response to exposure to chemicophysical stress. Moreover, the possible roles of these molecules in signaling mechanisms and in autolysis induction were examined. In fact, an important technological feature of this species is the early cell autolysis resulting in the release of peptidases and enzymes involved in cheese ripening (52, 53, 54).

The effects of sublethal osmotic, acid, and oxidative stress combinations simulating the different conditions occurring during Parmigiano Reggiano production were assessed. In particular, the quantitative release of furanones or other molecules as well as the dynamics of their release during exposure of L. helveticus CNBL 1156 to individual stresses were evaluated on the basis of gas chromatography (GC)-mass spectrometry (MS)/solid-phase microextraction (SPME). This method does not require extended sample preparation and is able to quantify, in addition to polar and nonpolar molecules of low volatility, furanones occurring in complex food systems (40). The bioactivity on fresh cells of the supernatants of the stress-exposed cells was evaluated in comparison with the bioactivities of pure commercial furanones including 2-ethyl-4-hydroxy-5-methyl-3(2H)-furanone (HEMF) and 3-hydroxy-5-ethyl-4-methyl-2[5H]-furanone (HEMFi).

MATERIALS AND METHODS

Strains and culture conditions.

Lactobacillus helveticus CNBL 1156, obtained from the collection of the Instituto di Microbiologia, Università Cattolica del Sacro Cuore, Piacenza (Italy), was grown in MRS broth under anaerobic conditions at 45°C (Anaerocult A; Merck, Darmstadt, Germany). The medium used for stress exposure was whey obtained from a preparation of Parmigiano Reggiano cheese and sterilized by filtration as previously reported (18). The viability assessment was carried out by anaerobic plate counting on MRS broth. Vibrio harveyi strain BAA 1117, obtained from the ATCC, was routinely grown in Marine broth (Difco, MA) at 26°C under conditions of aerobiosis.

Experimental design.

Filter-sterilized whey was inoculated with a whey culture of L. helveticus CNBL 1156 grown overnight and incubated at 45°C in a fermenter (Chemap). The pH was maintained at 5.6 by the addition of NH₃. After 4 h, the cells were collected and inoculated to obtain a cell density of 8 ± 0.3 log CFU/ml in acidified whey samples with H₂O₂ and NaCl added according to the conditions of the experimental design (Table 1). The levels of pH, NaCl, and H₂O₂ were modulated in order to generate conditions that occur during Parmigiano Reggiano production. The pH was modified at the defined values of 3.2, 3.8, 4.4, 5.0, and 5.6 by adding lactic acid to the whey cultures, while NaCl was added to final concentrations of 0.1, 0.3, 0.5, 0.7, and 0.9 M. Oxidative stress was achieved by adding H₂O₂ to reach 0.3, 1.5, 2.6, 3.8, and 5 mM concentrations. The oxidative stress was included because H₂O₂ and peroxidase are both present in fresh whey (14). Moreover, in order to evaluate the effect of the temperature as an individual factor, the cells were inoculated in whey brought to 45 (control), 48, and 53°C. Cell viability was verified by plate counting on MRS agar plates (Difco).

TABLE 1.

Levels of sublethal factors used for the experimental design

Combination	H2O2 level (mM)	NaCl level (M)	pH
1	1.5	0.3	3.8
2	3.8	0.3	3.8
3	1.5	0.7	3.8
4	3.8	0.7	3.8
5	1.5	0.3	5.0
6	3.8	0.3	5.0
7	1.5	0.7	5.0
8	3.8	0.7	5.0
9	2.6	0.5	4.4
10	2.6	0.5	4.4
11	0.3	0.5	4.4
12	5.0	0.5	4.4
13	2.6	0.1	4.4
14	2.6	0.9	4.4
15	2.6	0.5	3.2
16	2.6	0.5	5.6
17	2.6	0.5	4.4

After 100 min and 44 h of incubation at 45°C, the cell suspensions were centrifuged (2,500 × g for 10 min at 4°C), and the supernatants were analyzed with GC-MS/SPME.

The supernatants of runs 12, 13, and 14 were also used as conditioned media (CM) (CM12, CM13, and CM14, respectively) to suspend fresh cells (8 log CFU/ml) of L. helveticus in order to analyze their morphological changes by means of scanning electron microscopy (SEM).

Detection of the dynamics of furanone excretion in whey under individual oxidative or osmotic stress.

A culture of L. helveticus CNBL 1156 grown overnight and propagated at 45°C in whey under anaerobic conditions was recovered by centrifugation and resuspended to attain a cell concentration of 8 log CFU/ml in fresh filter-sterilized whey (FS) (control), 5 mM H₂O₂ (FSO), and 0.5 M NaCl (FSN). Three repetitions of each sample were prepared. After 10, 100, 300, 480, 1,440, and 1,860 min at 45°C, aliquots of the cell cultures were aseptically withdrawn and centrifuged. The cell-free supernatants collected after 10, 100, 300, 480, and 1,440 min were analyzed both with GC-MS/SPME and, for their bioactivity, with sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and bioluminescence assay.

Evaluation of the bioactivity of the CM collected during incubation of cells in FS, FSN, and FSO.

A culture of L. helveticus CNBL 1156 was grown overnight (optical density at 600 nm of 1.2) in whey medium under conditions of anaerobiosis at 45°C, and the cells were recovered by centrifugation (2,500 × g for 10 min at 4°C). The cells were resuspended for 100 min at a concentration of 8 ± 0.4 log CFU/ml in the various CM collected at different times during incubation in whey, and NaCl or H₂O₂ was added as described above.

After 100 min of incubation at 45°C, the samples for renaturing polyacrylamide gel electrophoresis were drawn. To analyze cell wall hydrolase activity, 2 ml of the cell suspension in each CM was harvested and centrifuged (2,500 × g for 10 min at 4°C), and the cells were resuspended in 40 μl of SDS-PAGE sample buffer. Cell extracts were heated for 3 min at 100°C, placed on ice for 5 min, and loaded onto an SDS-12% (wt/vol) polyacrylamide gel containing 0.2% (wt/vol) lyophilized Micrococcus lysodeikticus cells (Sigma, St. Louis, Mo.) (6). The protein concentration was determined using a protein assay (Bio-Rad, Hercules, CA), and serum albumin (Bio-Rad) solution was used as a standard. After electrophoresis, proteins were renatured to detect lytic activity according to a previously reported method (26), with some modifications: the gels were incubated with gentle shaking in 0.05 M Tris-HCl buffer (pH 6.8) containing 1% (vol/vol) Triton X-100 overnight at 42°C. The renaturation of the protein in the samples and visualization of clear zones in the gel with methylene blue staining were performed as previously described (3). Equivalence of loading between lanes was assessed by Coomassie blue staining of SDS-PAGE gels run in parallel. Densitometric analysis of zymograms was achieved by using Image J 1.3 software, provided by Wayne Rasband (National Institutes of Health, Bethesda, Md.).

Bioluminescence assay with Vibrio harveyi.

V. harveyi luminescence bioassay was performed essentially as previously described (51). For the determination of V. harveyi BAA1117 bioluminescence, a bacterial culture of V. harveyi grown overnight was diluted 1:5,000 into fresh Marine broth (MB) medium. The CM collected after 100 min from the suspension of L. helveticus cells exposed to 5 mM H₂O₂ was added to a final concentration of 10% (vol/vol). The culture was incubated at 26°C, shaken at 250 rpm in a rotary shaker, and assayed at intervals. Positive controls contained 10% (vol/vol) cell-free CM from V. harveyi BAA 1117, while negative controls contained 10% (vol/vol) sterile whey. The level of bioluminescence was determined with a systemSURE luminometer (Nova Biomedical, Waltham, MA) and was expressed as relative light units.

GC-MS/SPME analysis of volatile compounds of the whey and CM.

A divinylbenzene-carboxen-polydimethylsiloxane-coated fiber (65 μm) and a manual SPME holder (Supelco Inc., Bellefonte, PA) were used in this study after preconditioning according to the manufacturer's instruction manual. Before each headspace sampling, the fiber was exposed to the GC inlet for 5 min for thermal desorption at 250°C in a blank run. Five milliliters of the sample was placed into 10-ml vials, and the vials were sealed. The samples were then equilibrated for 15 min at 70°C. The SPME fiber was exposed to each sample for 40 min, and finally, the fiber was inserted into the injection port of the GC for 5 min of sample desorption.

GC-MS analyses were carried out using an Agilent 6890 gas chromatograph (Agilent Technologies, Palo Alto, CA) coupled to an Agilent 5970 mass selective detector operating in electron impact mode (ionization voltage, 70 eV). A Chrompack CP-Wax 52 CB capillary column (50-m length, 0.32-mm internal diameter) was used (Chrompack, Middelburg, The Netherlands). The temperature program was 50°C for 2 min and then programmed at 1°C/min to 65°C and finally at 5°C/min to 220°C, which was maintained for 22 min. Injector, interface, and ion source temperatures were 250, 250, and 230°C, respectively. Injections were performed in splitless mode, and helium (1 ml/min) was used as the carrier gas. Compounds were identified by use of available mass spectra databases (NIST/EPA/NIH version 1998 and Wiley version 1996) as well as by MS data in the literature (4). The quantification of medium-chain fatty acids (FAs) was performed by means of calibration curves obtained using pure standards (Sigma-Aldrich, Milan, Italy). For the quantification of furanones, due to the unavailability of commercial standards, HEMFi was used to prepare the calibration curve using five known concentrations ranging between 0.15 and 2.5 mg/liter.

Determination of sugars in whey and in the various CM.

Lactose, galactose, and glucose levels in the various CM and in the whey were determined by high-performance liquid chromatography as previously described (33).

SEM.

SEM observations of control cells and cells after exposure to the various conditions were performed as previously described (35).

Statistical analysis.

The quantitative gas chromatographic data for the experimental design were analyzed in order to obtain polynomial equations to describe the individual and interactive effects of pH, NaCl, and H₂O₂ on furanones and FA release. A software package (Statistica 6.0; Statsoft) was used. The variables with a significance lower than 95% (P > 0.05) were not included in the final models. The data regarding the experimental design and the dynamics of furanone excretion are the means of three repetitions prepared with the same whey.

RESULTS

GC-MS/SPME analysis of L. helveticus supernatants after exposure to various stress combinations.

The effects of the exposure of L. helveticus (8 ± 0.3 log CFU/ml) to different temperatures (45°C, 48°C, and 53°C) and to the various stress combinations in whey of the experimental design at 45°C were evaluated after 100 min on the basis of both the loss of viability and metabolite release in comparison with those of the control. Only the combinations characterized by a pH of 3.8 and high H₂O₂ and NaCl concentrations gave rise to a viability decrease of between 1 and 2 log CFU/ml. However, cell recovery was observed after incubation (data not shown).

The SPME analyses of the three repetitions of the various combinations after 100 min of stress exposure showed the release of three medium-chain fatty acids, hexanoic acid, octanoic acid, and decanoic acid, identified on the basis of the comparison of their mass spectra and retention times with those of pure standards (Table 2).

TABLE 2.

Amount of volatile compounds produced by L. helveticus after 100 min of exposure to stress combinations of the experimental design at 45°C and in whey at 45°C, 48°C, and 53°C

Runa	Amt of compound (mg/liter) ± SD
	Furanone A	Furanone B	Hexanoic acid	Octanoic acid	Decanoic acid
1	1.21 ± 0.13	0.54 ± 0.08	0.79 ± 0.11	3.27 ± 0.39	5.97 ± 0.66
2	0.71 ± 0.09	0.29 ± 0.04	0.60 ± 0.08	2.6 ± 0.31	7.02 ± 0.77
3	1.33 ± 0.16	0.60 ± 0.08	0.74 ± 0.10	0.31 ± 0.04	6.89 ± 0.76
4	0.19 ± 0.04	NDb	0.66 ± 0.09	2.28 ± 0.27	1.86 ± 0.20
5	1.10 ± 0.12	0.43 ± 0.06	0.27 ± 0.04	1.09 ± 0.13	2.03 ± 0.22
6	0.91 ± 0.11	0.27 ± 0.04	0.51 ± 0.07	1.86 ± 0.22	1.41 ± 0.16
7	1.54 ± 0.17	0.53 ± 0.07	0.32 ± 0.04	1.68 ± 0.20	2.13 ± 0.23
8	1.17 ± 0.12	0.44 ± 0.06	0.51 ± 0.07	2.23 ± 0.27	1.94 ± 0.21
9	0.90 ± 0.11	0.37 ± 0.05	0.51 ± 0.07	2.35 ± 0.28	1.41 ± 0.16
11	0.23 ± 0.04	ND	0.46 ± 0.06	12.33 ± 1.48	13.48 ± 1.41
12	1.85 ± 0.19	0.76 ± 0.11	0.42 ± 0.05	2.84 ± 0.34	2.55 ± 0.28
13	1.70 ± 0.18	0.82 ± 0.11	0.34 ± 0.04	2.25 ± 0.27	2.44 ± 0.27
14	1.36 ± 0.15	0.51 ± 0.07	0.50 ± 0.07	2.37 ± 0.28	1.92 ± 0.21
15	1.48 ± 0.16	0.78 ± 0.11	0.46 ± 0.06	3.33 ± 0.41	3.02 ± 0.33
16	1.77 ± 0.18	0.81 ± 0.12	0.56 ± 0.07	2.70 ± 0.32	2.89 ± 0.31
C45	ND	ND	ND	ND	ND
C48	0.30	0.2	ND	ND	ND
C53	0.60	ND	ND	ND	ND

^aC45, C48, and C53 indicate whey control at 45°C, 48°C, and 53°C, respectively.

^bND, not detectable.

Moreover, two peaks with retention times of 45.4 min and 45.9 min, hereafter called furanone A and furanone B, respectively, were detected. These peaks were present in traces in the controls. The spectral data for compound A, with the masses of the characteristic ions in kilodaltons and their intensities (in relative abundance) in parentheses, are as follows: 41 (39), 57 (80), 67 (28), 69 (25), 79 (18), 85 (8), 97 (100), 99 (73), 109 (40), 123 (7), 143 (9), while for compound B, they were 43 (28), 57 (47), 69 (11), 83 (17), 97 (100), 111 (10), 123 (16), 137 (10), 151 (4), 165 (4), and 180 (10). On the basis of the comparison of these spectral data with data from literature concerning MS fragmentation patterns (4), it was possible to assume that both molecules are 3-hydroxy-2[5H]-furanones with an α-hydroxy-γ-lactone configuration. The chemical structures shown in Fig. 1 can be proposed for furanone A and furanone B.

FIG. 1.

Proposed molecular structure of furanone A and furanone B.

The electron impact MS fragmentation patterns of furanones A and B showed that these molecules were characterized by different compositions and configurations of substituent R. However, the definitive identification of the two substituents (R) is still in progress. On the basis of the analysis of the mass spectral data, molecular weights of about 143 and 180 can be proposed for furanone A and furanone B, respectively.

The main daughter ions of molecules A and B do not correspond to those previously reported for the DMHF, MHF, AI-2, or other acylhomoserine lactones (AHLs) (56). In fact, these latter furanones, on the basis of their fragmentation patterns, are 3[2H]-furanones, while compounds A and B appear to be 2[5H]-furanones on the basis of their fragmentation patterns.

The increase of the incubation temperature up to 53°C induced a significant increase in furanone A only (Table 2).

The data relative to the concentration of the two furanones and the medium-chain FAs (means of three repetitions) after 100 min of exposure to the stress conditions modulated according to the experimental design, excluding those relative to temperature effects, were analyzed in order to obtain polynomial equations describing the effects of the independent variables as individual or quadratic terms and of their interactive effect on the concentration of the molecules detected by GC-MS/SPME. A significant relationship was obtained only between furanone B concentration and the individual and quadratic terms of H₂O₂ according to the following equation: furanone B = 0.354(H₂O₂) − 0.058(H₂O₂)² (regression coefficient = 0.897; F value = 30.808; standard error = 0.24).

According to the reported model, only the H₂O₂ concentration significantly affected B molecule accumulation. The positive sign of the individual factor and the negative sign of the quadratic term of the H₂O₂ concentration indicate that the production of furanone B increased with H₂O₂ up to a certain level, after which it decreased. Both furanone A and furanone B concentrations significantly decreased or did not occur when H₂O₂ was at the lowest level of the experimental design, as in run 11. In this condition, the octanoic and decanoic acid levels were comparatively elevated. When the cells were incubated under the various conditions for 44 h, the concentration of furanones tended to decrease. On the other hand, after this incubation time, the presence of compounds A and B was also detected in the control (C45), where they attained 0.62 and 0.19 mg/liter, respectively. Also, at 48 and 53°C, their values did not exceed values of 0.4 and 0.22 mg/liter, respectively. After the same incubation time, hexanoic and decanoic acids attained levels of about 0.21 and 2.55 mg/liter, respectively, in the control.

To validate the model obtained, combinations 3, 11, 12, 13, and 14 have been replicated. The concentrations of furanones obtained presented differences lower than 10% with respect to the predicted values.

In order to establish whether the variability within whey samples from different batches was able to induce differences in the production of furanones A and B, combination 12 of the experimental design, characterized by the highest level of furanone A and furanone B, was prepared using three different batches of whey. The mean values and standard deviations of furanones A and B were 1.76 ± 0.12 mg/liter and 0.79 ± 0.07 mg/liter, respectively.

Dynamics of excretion of furanones A and B and relationship with the bioactivity of CM.

Furanone A was released immediately (10 min) after the transfer of fresh cells to whey containing 5 mM H₂O₂ (Fig. 2a). The maximum extent of furanone A excretion was attained in the first 100 min at 45°C, after which its concentration decreased. A minor accumulation of furanone A was observed in the control (FS) and in the presence of 0.5 M NaCl (FSN). Furanone B presented the same behavior in the first 100 min and continued to increase over time in whey containing 5 mM H₂O₂ (FSO). The dynamics of furanone B release in the FSN and FS showed negligible differences (Fig. 2b).

FIG. 2.

(a) Evolution of furanone A released by L. helveticus exposed to osmotic (FSN) and oxidative (FSO) stress in the CM collected over time. ○, FS; □, FSN; ▴, FSO. (b) Evolution of furanone B produced by L. helveticus exposed to osmotic (FSN) and oxidative (FSO) stress in the CM collected over time. ○, FS; □, FSN; ▴, FSO.

In order to assess the relationship between furanone A and furanone B content and CM bioactivity, cell-free CM collected after 10, 100, 300, 480, 1,440, and 1,860 min were analyzed with the zymogram technique performed by using a renaturating SDS-PAGE gel containing M. lysodeikticus cells as a substrate.

The zymogram revealed degradation halos corresponding to lytic activities, which were evaluated as relative units (RU). In particular, a translucent zone having an apparent molecular mass of 31 kDa was observed in all the samples. Moreover, two bands with an apparent molecular mass between 43 and 45 kDa were present when the cells were exposed to CM of cells exposed to oxidative stress (FSO). Only the band at 43 kDa occurred both in the control (FS) and when the cells were suspended in CM of cells subjected to osmotic stress (FSN) (data not shown).

In Fig. 3, the summation of the RU values for the various bands detected in the CM under the three conditions tested (FS, FSO, and FSN) are plotted in relation to the exposure time. A comparison of the evolution of the bioactivity of the CM with data shown in Fig. 2a and b suggests that furanone A can be involved in the bioactivity of the CM more than furanone B. In fact, furanone A secretion was earlier and was related to the bioactivity data under both H₂O₂ and NaCl stress conditions. In the control, levels of both furanone A and furanone B increased over time in parallel with the RU value and lytic activity.

FIG. 3.

Autolytic activity of L. helveticus exposed to osmotic (FSN) and oxidative (FSO) stress in the CM collected over time. ○, FS; □, FSN; ▴, FSO.

Effects of H₂O₂ concentration and commercial furanones on autolytic activity of L. helveticus.

In order to evaluate the relationship between the autolytic activities and the H₂O₂ concentration of the medium and to evaluate the effects of commercial furanones or the medium-chain FAs corresponding to those occurring in CM, L. helveticus cells were transferred from a late-exponential-phase culture to fresh whey (condition I) (control), fresh whey containing 0.3 mM, 2.6 mM, and 5 mM H₂O₂ (conditions II, III, and IV, respectively), and fresh whey added with commercial furanones such as HEMF (7 μM) and HEMFi (7 μM) and medium-chain FAs (4 μM) as well as CM of conditions II, III, and IV. HEMFi was chosen due to its spectral similarity with furanones A and B. The biological effects were evaluated on the basis of cell viability, and the zymogram technique was performed by using a renaturing SDS-PAGE gel containing M. lysodeikticus cells as a substrate.

In Table 3, the RU values of the various bands having autolytic activity as well as of L. helveticus viability after exposure to the above-described conditions for 100 min are shown. The zymogram revealed translucent bands corresponding to the lytic activities whose numbers and intensities depended on the conditions used. The first band, with a molecular mass of about 31 kDa, was present in the control and under all the conditions tested except when the cells were exposed to 2.6 mM H₂O₂ and 5 mM H₂O₂. When CM II and III were put into contact with fresh cells, a band of 43 kDa appeared. Moreover, a band of about 45 kDa, whose intensity decreased with H₂O₂ concentration, was observed when the cells were put into contact (100 min) with CM II and III. In fact, the exposure to CM II, III, and IV gave rise to a significant decrease in viability. A band corresponding to a molecular mass of about 63 kDa was observed when cells were exposed to the commercial 2[5H]-furanone HEMFi, but the same band did not occur in presence of the 3[2H]-furanone HEMF. The exposure to medium-chain fatty acids reduced the autolysin occurrence and caused a viability decrease of about 1.4 log CFU/ml. It has previously been reported that autolysin A, with a molecular mass of 41 kDa, was associated with viable cells, but as soon as the cells died, autolysin A disappeared (52). On the other hand, the band of 63 kDa that was associated with the exposure to HEMFi decreased with H₂O₂ concentration apparently regardless of the cell viability level of the cell suspension. Commercial 2[3H]-furanones such as HEMF did not display any additional effect on autolytic activity with respect to the control.

TABLE 3.

Intensities of the lytic bands of L. helveticus CNBL 1156 after 100 min of exposure to oxidative stress and CM of cells exposed to oxidative stress and to commercial furanones or mixtures of fatty acids

Condition	Intensity (RU) ± SD					Viability (log CFU/ml) ± SD
	31 kDa	43 kDa	45 kDa	63 kDa	Total
Whey (control) (I)	82.7 ± 4.9	NDa	ND	ND	82.7	7.6 ± 0.3
Whey H2O2, 0.3mM (II)	43.7 ± 2.6	ND	ND	ND	43.7	7.4 ± 0.3
Whey H2O2, 2.6mM (III)	ND	ND	ND	ND	ND	7.4 ± 0.3
Whey H2O2, 5mM (IV)	ND	ND	ND	ND	ND	6.8 ± 0.3
HEMF, 7mM	39.8 ± 2.4	ND	ND	ND	39.8	6.8 ± 0.3
HEMFi, 7mM	74.8 ± 4.5	ND	ND	101.7 ± 6.1	176.5	6.6 ± 0.3
CM I	35.2 ± 1.6	ND	ND	ND	35.2	7.1 ± 0.3
CM II	41.1 ± 2.5	54.9 ± 3.3	31.1 ± 1.9	ND	127.1	4.3 ± 0.1
CM III	77.0 ± 4.6	54.0 ± 3.2	20.9 ± 1.2	ND	151.9	4.5 ± 0.1
CM IV	79.1 ± 4.7	0	ND	ND	79.1	4.4 ± 0.1
FAs, 4mM	53.5 ± 3.2	ND	ND	ND	53.5	6.2 ± 0.2

^aND, not detectable.

Evaluation of the bioactivity of the CM on V. harveyi.

To determine whether L. helveticus produces an AI-2-like signaling molecule under the conditions tested, V. harveyi BAA 1117 was inoculated into MB medium containing 10% CM whey from L. helveticus CNBL 1156. Under these conditions, the level of V. harveyi luminescence was maintained for 6 h. No relevant differences in luminescence were detected among V. harveyi cells inoculated with CM whey of L. helveticus CNBL 1156 and the negative control. These data suggest that L. helveticus does not produce AI-2-like molecules under the conditions taken into consideration.

Morphological changes of cells associated with exposure to CM and commercial furanones.

Fresh L. helveticus cells were suspended (to obtain a cell concentration of about 8 log CFU/ml) in CM of combinations 12, 13, and 14 of the experimental design containing different levels of furanone A, furanone B, and FAs (Table 3) and in whey with the commercial furanone HEMFi (7 μM) or HEMF (7 μM) added. After 100 min of exposure at 45°C, the possible morphological cell changes were analyzed with SEM. Autolysis phenomena were observed with the three CM, as shown in Fig. 4a, relative to CM12. In particular, irregular lesions and cell debris were observed. Moreover, the cells presented a heterogenous length (1.5 to 5.0 μm). A total of 250 cells in five micrographs per condition were individually measured. The percentage of cells with a length of ≤1.5 μm was 38% when fresh cells were exposed to CM12 (Fig. 4b). This percentage was 35 and 33% when the cells were exposed to the CM13 and CM14, respectively. On the other hand, the proportion of cells with a length of ≤1.5 μm was about 7% in the control (cells suspended in fresh whey) (Fig. 4c). This morphological modification was presumably not associated with nutrient limitation. In fact, the detection of glucose, lactose, and galactose in the different CM of the experimental design collected after 100 min revealed that the galactose level ranged from 0.3 g/liter (combinations 6, 12, 13, and 14) to 0.85 g/liter (combination 11), the glucose level ranged from 0.5 g/liter (combinations 6 and 9) to 0.9 g/liter (combinations 3, 4, and 5), and the lactose level ranged from 35 g/liter (combinations 1, 2, and 6) to 55 g/liter (combinations 5 and 7). In the whey control incubated at 45°C, the galactose, glucose, and lactose concentrations were 0.6, 0.7, and 40 g/liter, respectively.

FIG. 4.

(a) SEM micrograph of fresh cells of L. helveticus exposed for 100 min to cell-free CM12 of the experimental design showing cell autolysis and debris. (b) SEM micrograph of fresh cells of L. helveticus exposed for 100 min to cell-free CM12 of the experimental design showing short cells. (c) SEM micrograph of fresh cells of L. helveticus exposed for 100 min to whey control. (d) Fresh cells of L. helveticus exposed for 100 min to commercial HEMFi (7 μM) in whey.

Forty-one percent of very short cells (length, ≤1.5 μm) were observed when fresh cells were exposed for 100 min to 7 μm HEMFi in whey. Moreover, this 2[5H]-furanone as well as HEMF (7 μm) gave rise to anomalous cell formation, as shown in Fig. 4d, relative to HEMF. These morphological anomalies presented different shapes and occurred in about 8% of cells. Also, the exposure to hexanoic, octanoic, and decanoic acids (4 μM) gave rise to the occurrence of shorter cells (data not shown).

DISCUSSION

Besides autoinducers of the ATP-binding cassette (ABC) transporter for secretion, the most common mechanism of quorum sensing in gram-positive bacteria consists of a peptide and a two-component system for sensing the autoinducer concentration (12). However, LuxS homologues associated with AI-2 synthase of gram-negative bacteria have also been reported for the genome sequences of Lactobacillus acidophilus (1), Lactobacillus plantarum (25), Lactobacillus johnsonii (39), and Bifidobacterium longum (44). Moreover, unidentified molecules able to stimulate the quorum-sensing system and regulate the expression of the luciferase operon in Vibrio harveyi have been detected in cell-free culture fluids of Lactobacillus rhamnosus and Lactobacillus casei using the bioassay for bioluminescence induction (9). However, under the conditions tested in this work, L. helveticus CNBL 1156 did not secrete signaling molecules able to promote LuxS synthesis in Vibrio harveyi for bioluminescence production in whey. In fact, all the molecules with AI-2 activity, whose configuration has been identified, are 3[2H]-furanones (55) with 4,5-dihydroxy-2,3-pentadione as a precursor. Moreover, furanones A and B presented different configurations with respect to the furanones already identified as products of Lactococcus lactis subsp. cremoris (21) or to those identified in some matured cheeses with L. helveticus as a starter culture (15, 37). In fact, the furanones whose presence has been reported in cheeses are 3[2H]-furanones.

Although the complete identification of 2[5H]-furanones A and B released by L. helveticus, as well as their precursors and biosynthesis, requires further investigation, their origin from 4,5-dihydroxy-2,3-pentadione does not seem realistic from a chemical point of view. The dependence of furanones A and B on oxidative stress, the contemporaneous release of medium-chain FAs, and the previously reported formation of epoxides of linoleic acid in L. helveticus following oxidative stress (18) suggest that epoxidated or hydroxylated C₁₈ chain membrane FAs may be precursors of furanones A and B. It is known that in many organisms, reactive oxygen species such as H₂O₂ and the superoxide ion are produced under many physiological conditions. They can oxidize unsaturated acyl chains integrated into membrane phospholipids. 2[5H]-furanones can subsequently be produced throughout a sequence of shortening by β-oxidation and lactonization reactions. Peroxidation reactions are in fact the first step in the generation of plant signal compounds such as jasmonic acid (48).

Compared with the many reports about the isolation of furanones from microorganisms and their potential applications (2, 11, 24), there are only a few reports regarding natural furanones substituted at positions 2 and 5. In particular, the synthesis of four 2[5H]-furanones with antibiotic activity against Pseudomonas aeruginosa has been reported previously for Streptomyces antibioticus (5). Some of these compounds proved to be active in the quorum-sensing system of Chromobacterium violaceum (16). Moreover, five 2[5H]-furanones, the aporpinones, were reported to be secondary metabolites of Aspergillus terreus (45) and Aporpium caryae (27). A metabolite of Pseudomonas aureofaciens, 3-(1-hexenyl)-5-methyl-2[5H]-furanone, is the first antifungal 2[5H]-furanone whose release in a bacterium has been reported, and it has been described as a biocontrol agent of fungal plant pathogens (36). While the biosynthesis of the 2[3]-furanones and of AHLs has been deeply studied (57), the biosynthesis mechanisms of the above-mentioned 2[5H]-furanones and their precursors are almost unknown. However, some bioactive 2[5H]-furanones endowed with pharmacological properties have been obtained through chemical synthesis (16). These synthetic compounds were able to simulate AHL activity and induce violacein formation in Chromobacterium violaceum (31).

Compounds A and B, which are volatile and have a presumptive molecular mass lower than that of the above-described antibiotics (M_w, 143 to 180), meet a number of criteria previously proposed (55) for inclusion of a metabolite in the cell-to-cell signal molecules. In fact, they were released following sudden stress exposure or during the stationary phase in the control, and they were then metabolized or degraded. Likewise, it has been previously reported that AI-2 and MHF are taken up and utilized during late stages by Pseudomonas aeruginosa (56). Moreover, the exposure of fresh cells to the various CM containing furanones A and B as well as medium-chain FAs triggered or was associated with morphological changes in cells and autolysin production. Similar effects were also observed when cells were exposed to a pure commercial HEMFi furanone with similar chemical characteristics. In particular, HEMFi induced the production of an additional autolysin band (with an apparent molecular mass of 63 kDa). The significant increase of the proportion of short anomalous cells suggests interference with cell wall synthesis or cell division and provides further indirect evidence of the role of the specific chemical configurations of furanones A and B.

The occurrence of these molecules as late metabolites in control whey also indicates that, as previously observed for Saccharomyces cerevisiae (29) and Candida albicans (22), L. helveticus can co-opt its own metabolites to promote cell differentiation.

It may also be questioned whether the medium-chain free FAs released play a crucial role in cell differentiation as suggested by the morphological changes observed. The FAs released by cells exposed to chemicophysical stresses have been associated with an interrupted biosynthesis of FAs in yeast cultures or to phospholipid unsaturated FA peroxidation following oxidative stress and subsequent degradation (17). It has been proposed that in Salmonella enterica serovar Typhimurium, the degradation of long-chain FAs through β-oxidation would generate acetyl coenzyme A to feed the tricarboxylic acid cycle for energy production during starvation (47). However, on the basis of the sugar content and its changes during L. helveticus incubation, whey cannot be regarded as a limiting medium. Moreover, it has been suggested that in Myxococcus xanthus, mixtures of straight and branched FAs have an autocide activity and are involved in a cell-cell communication system (41). In Myxococcus xanthus, the extracellular accumulation of FAs was accompanied by the lysis of a large proportion of the cell population. However, the exposure of fresh cells of L. helveticus to a mixture of FAs did not induce autolysin production.

The effect of exogenous furanones on cell morphology suggests that the morphological changes, and particularly the appearance of short cells, can be associated with the phenomena involved in the programmed death of a population induced by pheromones (28). In fact, L. helveticus cells respond in concert to stress, releasing furanones and FAs. These molecules can promote autolysis or agglutination and morphology changes that give rise to shorter or anomalous cells.

Preliminary research on other lactic acid bacterial species such as Lactobacillus helveticus, Lactobacillus plantarum, and Lactobacillus sanfranciscensis in hydrolyzed wheat flour (M. E. Guerzoni, personal communication) evidenced the occurrence of two molecules with the same mass spectral data of furanone A and furanone B when the cells attained a level of 8 log CFU/ml or when the cells were exposed to osmotic and oxidative stresses. Final proof of the role of the 2[5H]-furanones requires the use of chemically defined molecules. In this phase of the research, only small amounts of furanones A and B were released, which did not allow either a determination of the absolute configuration or a more detailed investigation of the full biological properties. Therefore, given the growing interest in the antimicrobial activities of furanones of diverse origins and their biological effects observed at low concentrations, the identification of the precursors of the 2[5H]-furanones and of the genes controlling their synthesis in L. helveticus is of recognizable importance.