Detection of Bacteriocins by Matrix-Assisted Laser Desorption/Ionization Time-of-Flight Mass Spectrometry

Abstract

The use of matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) for the detection of bacteriocins was investigated. A 30-s water wash of the sample on the MALDI-TOF MS probe was effective in removing contaminants of the analyte. This method was used for rapid detection of nisin, pediocin, brochocin A and B, and enterocin A and B from culture supernatants and for detection of enterocin B throughout its purification.

Bacteriocin-producing lactic acid bacteria and purified bacteriocins have the potential for use as biopreservatives to extend storage life or to enhance the safety of foods. Bacteriocins are proteinaceous compounds produced by lactic acid bacteria that exhibit activity against closely related bacteria, and they may also be active against species beyond the same ecological niche (17).

Currently, researchers rely on bioassays such as deferred inhibition and agar diffusion tests for the detection of bacteriocin production and the determination of bacteriocins in foods (20). Such methods are indirect because they rely on a sensitive indicator organism that varies among laboratories. In addition, the results are expressed in arbitrary units (AU), which vary with experimental conditions (i.e., pH, temperature, nutrients, and choice of indicator) (6, 13).

A sensitive, rapid detection method for bacteriocins could be a useful method to track purification procedures, to detect bacteriocin production in experiments involving genetic manipulation, and to detect bacteriocins in foods (6, 20). Detecting the bacteriocin by searching for a compound with the appropriate molecular mass is one method of confirming the presence of bacteriocins in cultures or food products. Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) appears to have potential as one such method.

MALDI-TOF MS was first recognized in 1988 as a method of examining large molecules (16). With rapid advances in this technology, MALDI-TOF MS is becoming an essential tool in the analysis of biopolymers, including peptides and proteins. Generally, MALDI-TOF MS involves ion formation using a pulsed UV laser beam to deposit energy into a cocrystallized matrix-analyte sample and ejection of the sample into a vacuum for desorption analysis. The major advantages of MALDI-TOF MS over other mass spectrometric techniques are (i) its ease of use, (ii) its picomole-to-femtomole sensitivity (22), (iii) its high-mass (>200,000-Da) range and ability to give molecular mass values with a reported accuracy of ±0.1% or better (4, 5, 7, 22), and (iv) its relative tolerance to contaminants in the sample (5, 24, 25, 27). These advantages make MALDI-TOF MS technology suitable for the analysis and characterization of bacteriocins.

MALDI MS is usually used in conjunction with TOF mass analyzers that give a defined time of ion generation and are compatible with the pulsed lasers of MALDI MS (7). The TOF of the ion through the drift region is proportional to the mass/charge (m/z) ratio of the ion, which allows for the determination of mass. MALDI-TOF MS systems have also greatly advanced with the implementation of delayed extraction (3, 7, 14, 23, 25). Delayed extraction has been reported to significantly improve the resolving power, detection sensitivity, signal-to-noise (S/N) ratios, and mass measurement accuracy of MALDI-TOF MS (7, 14, 23, 25, 27).

Presently, MALDI-TOF MS is recognized as a valuable tool for molecular mass measurements and is showing promise as a quantitation tool. MALDI-TOF MS has been used routinely for the analysis of synthetic peptides and proteins. However, the analysis of peptides and proteins extracted from biological sources has been hampered by the presence of contaminants, such as salts, glycerol, and detergents (1, 4, 5, 18, 26, 27). These contaminants may suppress the peptide and protein signals completely. Sodium dodecyl sulfate and salts have been shown to interfere with the signal obtained for bovine serum albumin and cytochrome c (1, 24). Wang et al. (24) reported that differences in salt content among samples might result in detection of certain peptides and proteins over others, yielding different mass spectral patterns and consequently poor spectrum reproducibility.

In bacteriocin purification, the presence of detergents such as Tween in All-Purpose Tween (APT) and de Man, Rogosa and Sharp (MRS) medium, both commonly used for the growth of bacteriocin-producing organisms, may interfere with the analysis. Tween 80 is an essential medium component for bacteriocin production and detection (12, 15). It is likely that partial purification of bacteriocin preparations will be necessary to obtain a clean sample for analysis by MALDI-TOF MS; however, purification methods are often time-consuming, result in the loss of biological material, and may even introduce more contaminants that are incompatible with MALDI-TOF MS (26).

Researchers have investigated the use of activated synthetic membranes as an alternative to stainless steel in MALDI-TOF MS analysis. By spotting a sample onto a membrane such as polyethylene, contaminants can be washed away, leaving the peptides and proteins intact at the surface for analysis (4, 5, 26). This process results in equal or greater sensitivity and mass resolution for all samples compared to those desorbed from stainless steel. Synthetic membranes are particularly suitable for high-mass (i.e., >30,000-Da) molecules because severe ion suppression is typically observed in the analysis of high-mass mixtures (4, 5). Other researchers have used or suggested further processing of the sample on the probe, such as a wash with cold water to remove the contaminants (18, 24, 27).

In this study, we developed a MALDI-TOF MS method for the detection of bacteriocins in culture supernatant and used the method to determine the fate of a bacteriocin throughout a purification procedure.

Bacterial strains and media.

The bacteriocin-producing lactic acid bacteria used in this study are listed in Table 1. Enterococcus faecium CTC 492 was obtained from M. Hugas (Institut de Recera i Technologia Agroalimentàries, Girona, Spain), and E. faecium BFE 900 was isolated from black olives by Franz et al. (8). Carnobacterium divergens LV13, obtained from B. G. Shaw (Institute of Food Research, Langford, Bristol, United Kingdom), and Lactobacillus sake ATCC 20017 were used as sensitive indicator organisms against the producer strains. Frozen stock cultures were maintained at −70°C in Bacto APT broth (Difco Laboratories, Detroit, Mich.) supplemented with 20% glycerol (vol/vol). Prior to experimental use, Lactococcus lactis ATCC 11454, Brochothrix campestris ATCC 43754, Pediococcus acidilactici PAC-1.0, E. faecium CTC 492 and BFE 900, and C. divergens LV13 cultures were subcultured twice and grown overnight in APT broth. L. sake ATCC 20017 was subcultured twice and grown overnight in lactobacillus MRS broth (Difco) prior to use. Solid agar medium was prepared by adding 1.5% (wt/vol) granulated agar (Difco) to either APT or MRS media. Soft APT and MRS agar were prepared with 0.75% agar (wt/vol).

TABLE 1.

Bacteriocin-producing organisms used in this study and characteristics of their bacteriocins

Producer strain	Bacteriocin	No. of amino acid residues	Molecular mass (Da)	Reference
L. lactis subsp. lactis ATCC 11454	Nisin	34	3,354	10
B. campestris ATCC 43754	Brochocin A	59	5,242	19
B. campestris ATCC 43754	Brochocin B	43	3,943	19
P. acidilactici PAC-1.0	Pediocin PA-1	44	4,629	11
E. faecium CTC 492	Enterocin A	47	4,829	2
E. faecium BFE 900	Enterocin B	53	5,463	9

Preparation of culture supernatant.

L. lactis ATCC 11454, P. acidilactici PAC-1.0, B. campestris ATCC 43754, and E. faecium CTC 492 and BFE 900 were grown for 18 h at 30°C in 10 ml of APT broth. One-milliliter aliquots of each grown culture were pipetted into Eppendorf tubes and boiled for 1 min. Cells were removed by centrifugation at 10,000 × g for 10 min at 4°C. The supernatant fluid was collected and stored at 4°C until used for analysis, within 1 day.

Purification of enterocin B.

Enterocin B was purified by the method described by Franz et al. (9). E. faecium BFE 900 was grown for 18 h in 1 liter of APT broth supplemented with 3% glucose (vol/vol). The culture was heated for 30 min at 75°C, and cells were removed by centrifugation (8,000 × g; 30 min). The supernatant was collected and loaded onto an Amberlite XAD-8 column (150 by 75 mm; BDH Chemical Ltd., Poole, United Kingdom). The column was washed with 1 liter of 0.1% trifluoroacetic acid (TFA) and then with 750 ml of 30% ethanol in 0.1% TFA. The active fraction, as determined by a spot-on-lawn assay, was eluted with 60% ethanol in 0.1% TFA and concentrated to approximately 75 ml by a rotary evaporation. The concentrated fraction was adjusted to pH 4.5 with 20 mM sodium acetate buffer (SAB; pH 5). The fraction was loaded onto an SP Sepharose Fast Flow cation exchange column (110 by 13 mm; Pharmacia Biotech, Baie D’Urfe, Quebec, Canada), and the column was sequentially washed with 100 ml of SAB, 60 ml of 100 mM sodium chloride in SAB, and 60 ml of 500 mM sodium chloride in SAB, which eluted the bacteriocin. This active fraction was loaded onto a Sep Pak C₁₈ reverse-phase column (Waters Ltd., Mississauga, Ontario, Canada). The column was washed with 10 ml of milli-Q water and then with 10 ml of 30% ethanol. This bacteriocin was eluted with a final wash of 10 ml of 95% ethanol. This active fraction was freeze-dried overnight and resuspended in 0.1% TFA. At each step during the purification, aliquots of the supernatant, XAD-8 column, cation-exchange column, and resuspended freeze-dried protein fractions were collected and stored at 4°C for further analysis.

Bacteriocin activity assay.

The culture supernatants were assayed for activity against C. divergens LV13 by the spot-on-lawn technique with APT agar. The plates were incubated for 18 h at 30°C. Fractions from the enterocin B purification were assayed by the same method using MRS agar and L. sake ATCC 20017 as the indicator strain. Activity was measured by taking the reciprocal of the highest dilution that exhibited a clear zone of inhibition and was expressed as AU per milliliter.

MALDI-TOF MS.

All mass spectra were acquired on a linear MALDI-TOF mass spectrometer equipped with delayed extraction technology (Proflex III; Bruker, Billerica, Mass.) with a 125-cm flight tube. All spectra were acquired in positive ion linear mode with a nitrogen laser (λ = 337 nm) for desorption/ionization of the samples and an acceleration voltage of 20 kV. The spectra are representative of 60 consecutive laser shots. Angiotensin II (MH⁺ = 1,046.542; Sigma Chemical Co., St. Louis, Mo.) and insulin bovine (MH⁺ = 5,734.557; Sigma) were used as the calibrants for external mass calibration. The instrument was calibrated for each sample preparation method by using the conditions described below.

Sample preparation.

The use of synthetic membranes and washing the probe with water were examined as methods of removing sample contaminants to provide effective MALDI-TOF MS analysis. Polyethylene membranes (Fisher Scientific, Fair Lawn, N.J.) were prepared by the method described by Worral et al. (26). The membrane was saturated with methanol, air-dried, and fixed to the stainless steel probe with double-sided tape. Culture supernatant (0.5 μl) was spotted on the membrane and allowed to dry. The membrane was washed three times with 20 μl of 70% methanol in water and was air-dried between each set of washes. A saturated solution of sinapinic acid (0.5 μl; Sigma) was spotted on the sample. The supernatant of L. lactis ATCC 11454 was used to determine the most effective washing method of the sample directly on the probe (on-target washing). Supernatant samples (0.5 μl) were placed on a stainless steel MALDI probe, and the probe was allowed to air dry. The probe was dipped into water and held static for 0, 10, 30, or 60 s. The excess water was shaken off, and the probe was air-dried. When dry, 0.5 μl of a saturated solution of sinapinic acid in 0.1% TFA–acetonitrile (2:1) was added to the sample spot and allowed to dry before analysis.

Detection of bacteriocins by bioassay.

All supernatants and fractions tested had activity against the appropriate indicator strains (Table 2). The concentration of bacteriocins in the supernatants varied from 1,600 to 6,400 AU/ml when assayed against C. divergens LV13. Throughout the purification procedure for enterocin B, the relative AU per milliliter in each of the fractions increased.

TABLE 2.

Concentrations of bacteriocins in prepared culture supernatants and fractions collected during purification of enterocin B

Source	Producer strain or sample	Indicator organism	AU/ml
Culture supernatant	L. lactis ATCC 11454	C. divergens LV13	1,600
	B. campestris ATCC 43754	C. divergens LV13	6,400
	P. acidilactici PAC-1.0	C. divergens LV13	1,600
	E. faecium CTC 492	C. divergens LV13	3,200
	E. faecium BFE 900	L. sake ATCC 20017	6,400
Enterocin B purification	Supernatant	L. sake ATCC 20017	6,400
	XAD-8 column	L. sake ATCC 20017	12,800
	Cation-exchange column	L. sake ATCC 20017	25,600
	Resuspended freeze-dried protein	L. sake ATCC 20017	204,800

Detection of bacteriocins by MALDI-TOF MS.

Attempts with membranes to adsorb bacteriocins for MALDI-TOF MS analysis were unsuccessful (data not shown). However, the on-target washing method proved to be useful for the detection of bacteriocins by MALDI-TOF MS. Figure 1A shows the spectrum of the prepared culture supernatant from L. lactis ATCC 11454, which should contain nisin. The natural contaminants in the prepared culture supernatant were present in sufficient concentration to affect the quality of the protein signal. A 10-s wash with milli-Q water removed a portion of the contaminants (Fig. 1B); however, a 30-s wash (Fig. 1C) was the most effective in washing away the contaminants, resulting in a peptide signal of greater intensity, better resolution, and less noise. The 60-s wash (Fig. 1D) was also effective in removing the contaminants, but it appeared to degrade the sample signal. Despite poor signal intensity and resolution for some of the samples in Fig. 1, the m/z ratios for all samples were similar to the mass expected for nisin (MH⁺ = 3,354 ± 0.1%) with signals ranging from 3,355 to 3,359 Da. Similar results were obtained when enterocin B was washed for different times (data not shown). The 30-s wash was used as the method for sample preparation when attempting to detect other bacteriocins by MALDI-TOF MS.

FIG. 1.

Comparison of mass spectra obtained from a crude bacteriocin preparation of nisin with either no water rinse (A), a 10-s water rinse (B), a 30-s water rinse (C), or a 60-s water rinse (D).

MALDI-TOF MS was able to detect brochocin A and B, pediocin PA-1, enterocin A, and enterocin B in cell supernatants of the producer organisms. Figure 2A shows the spectra obtained for the culture supernatant of B. campestris ATCC 43754, which produces brochocin A (5,242 Da) and brochocin B (3,943 Da) (19). The peak in the 2,920-Da range has not been identified. Figure 2B shows the MALDI-TOF MS spectra obtained from the supernatant of P. acidilactici PAC-1.0, which produces pediocin PA-1 (4,629 Da) (11). Figures 2C and D are the mass spectra obtained from the supernatants of E. faecium CTC 492 and E. faecium BFE 900, respectively. E. faecium CTC 492 produces enterocin A, a 47-amino-acid bacteriocin with a molecular mass of 4,829 Da (2). The MALDI mass spectrum confirms that E. faecium CTC 492 produced enterocin A, but it also shows another peptide at approximately 5,479 Da, which corresponds to the mass of enterocin B (5,463 Da) (9, 21). E. faecium CTC 492 also appears to be producing another uncharacterized substance, with a molecular mass of approximately 5,800 Da. The peak at approximately 5,479 Da in Fig. 2D confirms that enterocin B is being produced by E. faecium BFE 900.

FIG. 2.

Mass spectra obtained from crude bacteriocin preparations from B. campestris ATCC 43754 (A), P. acidilactici PAC-1.0 (B), E. faecium CTC 492 (C), and E. faecium BFE 900 (D).

MALDI-TOF MS was also effective in determining the presence of bacteriocins in the active fractions obtained during the purification of enterocin B (Fig. 3). Figure 3A is the MALDI spectrum for the cell supernatant with a small peak at the appropriate mass range for enterocin B. The MALDI-TOF MS spectrum for the fraction collected after purification on the XAD-8 column (Fig. 3B) also confirmed the presence of enterocin B. Figure 3C shows that enterocin A and B were present in the fraction eluted from the cation exchange column. Figure 3D is the MALDI-TOF MS spectrum of the sample after resuspension of the freeze-dried bacteriocin. Throughout the purification, as the bacteriocin concentration increased, the arbitrary intensity of the peaks increased. However, there does appear to be a slight discrepancy in the mass among the samples. Discrepancies between the MALDI-TOF MS mass measurements shown in this figure and the reported masses given in Table 1 could be the result of many factors. For example, the mass of enterocin B determined by MALDI-TOF MS (5,479 Da) was 16 Da greater than that reported in the literature. This difference is likely due to oxidation of the peptide. Wang et al. (24) attributed mass discrepancy to difficulty in accurately determining the peak centroid due to the low resolving capabilities of MALDI-TOF MS, as well as the use of external versus internal calibration. External calibration is the method of choice when speed and sample consumption rather than mass accuracy are of interest. A lower level of mass accuracy is obtained with external calibration because slight changes in laser power and sample preparation may cause differences in the desorption/ionization process. However, an advantage of MALDI-TOF MS is that the results are the average of many individual laser pulses (23). Therefore, the combination of a large number of ions and good calibration should alleviate concern regarding the accuracy of the mass. In this study, it was shown that MALDI-TOF MS should not be relied upon for an accurate measurement of molecular mass; however, it is capable of providing reproducible spectra for the detection of bacteriocins within the expected mass range.

FIG. 3.

Mass spectra obtained from fractions of an enterocin B purification supernatant (A), from an XAD-8 column (B), from a cation-exchange column (C), and of resuspended freeze-dried protein (D).

MALDI-TOF MS is the first reported mass spectrometric technique to be used to detect bacteriocins in the cell-free supernatants of a culture. This is largely due to the ability to purify samples on target. In our studies, a sterile water wash was chosen as the method for removing contaminants. Other researchers have suggested the use of polyethylene or polypropylene membranes for use as activated membranes to which the peptide or protein binds (4, 5, 26). However, Joosten and Nuñez (15) report that Tween 80 prevents adsorption of the bacteriocins nisin and enterocin on polypropylene surfaces. This may explain the loss of peptide signal found when we studied the use of polyethylene membranes with various rinses as a sample surface for MALDI-TOF MS analysis.

The presence of contaminants in the culture supernatant greatly suppresses the signal of the peptide. Blackledge and Alexander (4) suggested that the contaminants prevent effective crystallization of the matrix and are desorbed with more efficiency than the peptide. This, in turn, suppresses the peptide signal. Amado et al. (1) indicated that the loss of protein signal may be explained by partial precipitation of surfactant-protein ionic pairs during sample preparation. They also suggested that the signal degradation found with high concentrations of surfactants may be the result of surfactants coating the matrix crystals, thus diminishing energy transfer and desorption/ionization efficiency. The results presented here show that a 30-s water rinse in sample preparation removes the majority of the contaminants, resulting in a better S/N ratio, better peak resolution, and better signal intensity of the sample peak.

It has also been shown that MALDI-TOF MS can be used to identify components of various samples throughout the bacteriocin purification process and has potential for future use in the detection of bacteriocins in genetic and food experiments. Of particular interest is the potential of MALDI-TOF MS as a quantification tool. Bouksaïm et al. (6) reported that to use bacteriocins as food preservatives, it is important to understand the relationship between activity and exact quantity or real concentration of bacteriocin in a food system. However, in order to quantify bacteriocins by MALDI-TOF MS, a pure sample would be needed. The method proposed in this paper for detection would not be very reproducible for quantification given the evidence that spectral changes occur with washing (Fig. 1). The use of MALDI-TOF MS as a quantitation tool is currently being examined for pure nisin. Purification methods for other bacteriocins are being developed, and these will be examined by MALDI-TOF MS.

Analysis by MALDI-TOF MS is just beginning to be recognized, and more work is needed in obtaining higher-resolution spectra, better sensitivity, better sample preparation, and faster data analysis (1, 7, 23, 26). The results presented here show that MALDI-TOF MS is a rapid and sensitive detection method for bacteriocins. Its ability to generate mass spectra from the supernatant makes it particularly attractive for use in industry and commercial application. Another major advantage of MALDI-TOF MS is its ability to screen supernatant and purification samples for bacteriocin production. This process takes minimal time (minutes) compared to the overnight incubation of traditional bioassays. Further research will involve the use of MALDI-TOF MS to detect bacteriocins in multicomponent food systems and the examination of the interaction between the bacteriocins and the different food components.