Characterization by Tandem Mass Spectrometry of Biologically Active Compounds Produced by Bacillus Strains

*Correspondence:
Bure C . CGFB, 146 rue Leo Saignat , 33076 Bordeaux cedex, France.
Phone/Fax. +33 (0)5 57 57 56 23/16 84.

Citation:
Tosco A, Chobelet A, Bathanya K, Schmitter JM, Urdaci MC, Buré C. Characterization by Tandem Mass Spectrometry of Biologically Active Compounds Produced by Bacillus Strains. J Appl Bioanal 1(1), 19-25 (2015).

Open-access and Copyright:
©2015 Tosco A et al. This article is an open access article distributed under the terms of the Creative Commons Attribution License (CC-BY) which permits any use, distribution, and reproduction in any medium, provided the original author(s) and source are credited.

Funding/Manuscript writing assistance:
The authors have no financial support or funding to report and also declare that no writing assistance was utilized in the production of this article.

Competing interest:
The authors have declared that no competing interest exist.

Article history: Received: 2 December 2014, Revised 1 January 2015, Accepted 6 January 2015

Abstract

A methodology relying on LC-ESI-MS/MS has been applied to the characterization of biologically active compounds produced by Bacillus strains grown in liquid culture media. This approach enabled off-line evaluation of biological activities and complementary MALDI-MS/MS analysis. In this way, major limitations of direct MALDI analysis such as spectral discrimination and lack of information on low molecular weight species were circumvented. Fragmentation patterns combined with accurate mass determination led to reliable structure assignments for amicoumacins and major lipopeptide series: surfactins and fengycins. In this latter series, LC-ESI-MS/MS correlated with activity tests confirmed the biological activities of fengycins, whereas linear forms of fengycins were inactive.

Keywords

lipopeptides, MS, MS/MS, liquid chromatography, bacillus strains.

Introduction

Among Gram-positive bacteria, Bacillus strains are known to produce numerous compounds that display a large spectrum of biological activities [1]. Owing to their production of antimicrobial substances, Bacillus strains can act as biological control agents for plant diseases [2] and some of these substances have an effect on plant pathogens [3]. Furthermore, these bacteria have beneficial effects on human health that are directly connected to their production of anti-microbial substances [4] and probiotics based on Bacillus bacteria contribute to the prevention and treatment of infectious diseases [5] and gastrointestinal microbial disorders [6]. 
Many compounds produced by Bacillus strains are characterized by an alkyl chain of variable length linked to a cyclic peptide, and thus can be classified as lipopeptides. Their amino acid composition and fatty acid chain length are variable, as well as the type of ring closure, thus leading to the occurrence of numerous isoforms and isobaric species (Figure 1). Surfactins are powerful biosurfactants with exceptional emulsifying and foaming properties [7].  This family encompasses structural variants, but all members are heptapeptides interlinked with a β-hydroxy fatty acid to form a cyclic lactone ring structure. Fengycins, which are also called plipastatins, are lipodecapeptides with an internal lactone ring in the peptidic moiety linked to a β-hydroxy fatty acid chain that can be saturated or insaturated. Fengycins are less haemolytic than surfactins but retain a strong fungitoxic activity [7].
Other compounds synthesized by Bacillus strains were listed as antibiotic compounds. In 1974, Berdy reported 167 antibiotics produced by members of the Bacillus genus [8]. Since then, many new antibiotics have been isolated from strains of the Bacillus genus and have found applications in the pharmacology, veterinary and food industries [9]. These compounds include low molecular weight species (below 500 Da) such as amicoumacins (Figure 1) that have antibacterial, anti-inflammatory and anti-ulcer activity [10].
Several analytical methods have been developed for the detection of lipopeptides, such as capillary electrophoresis (CE), thin layer chromatography (TLC), liquid chromatography coupled to an UV detector (LC-UV) or also immunological detection by competitive ELISA assays for the quantification and localization of lipopeptides on the surface of plant tissues. However, none of these methods provides enough information for a safe identification of individual lipopepides, whereas mass spectrometry turned out to be a more informative methodology for the characterization of lipopeptide structures.
Matrix Assisted Laser Desorption Ionization – Mass Spectrometry (MALDI-MS) has been widely used for the screening of bacterial strains [11-13], whereas electrospray ionization (ESI) has been less used [14-16]. MALDI-MS is simple, fast and sensitive, but three major drawbacks limit its application to the characterization of biologically active substances produced by bacteria, namely spectral discrimination, loss of information for small molecules with molecular weights below 700 Da and dominance of cationic adducts among molecular species. The first two points largely explain the fact that the full panel of biological activities displayed by a given strain is rarely reflected by the compounds revealed by MALDI-MS fingerprinting. For the third point, the occurrence of multiple ionic species for a single molecule complicates full scan spectra  and is directly related to the composition of culture media; furthermore, MS/MS of [M+Na]⁺ and [M+K]⁺ precursor ions most often leads to a low content of structure information than hinders structure characterization. 
An analytical method based on LC-ESI-MS/MS (liquid chromatography coupled to tandem mass spectrometry) could have the double advantage of the separation of the different families of lipopeptides and of mass measurement combined to fragment ion spectra for structure identification.
Therefore, we designed a more reliable methodology for rapid characterization of biologically active compounds produced by Bacillus strains that associates measurement of biological activities and structure determination.

Materials and methods

Reference compounds
Surfactin A (mixture of compounds with different alkyl chain lengths), and Glu-fibrinopeptide were obtained from Sigma-Aldrich (St. Louis, MO, USA).

Culture media and strains
Bacillus subtilis strains were grown in Difco sporulation medium and Landy medium for 18 and 70 h at 30°C, at 200 rpm, with an estimation of 109 bacteria/mL for the Difco and 108 bacteria/mL for the Landy medium. Culture broths (10 mL) were centrifuged at 9,500 rpm for 15 min at 4°C. Bacterial cells left in the supernatant were removed by filtration (0.45µm membrane, Millipore, Molsheim, France). Supernatant were stored at -20°C.

Extraction 
Bacteria grown in liquid medium (10 mL cultures) were harvested by centrifugation and resuspended in 200 µL of sterile physiological water. A 100 µL aliquot of the suspension was extracted with 400 µL of a 1/1 solution (v/v) of acetonitrile/isopropanol containing 0.1% formic acid. Solvent was removed under vacuum (SpeedVac, Savant, Holbrook, NY, USA) prior to re-dissolution and LC/MS/MS analysis. 

Solubilisation of extracts
Lipopeptides tend to remain adsorbed on the wall of Eppendorf tubes. Optimal re-dissolution of bacterial extracts was obtained with 50 µL of 1/1/2 (v/v/v) acetonitrile/isopropanol/H₂O solution containing 0.1% formic acid, followed by treatment in an ultrasonic bath (two 30 s period, with 30 s pause time).

LC-ESI-/MS and MS/MS analysis

A C4 column (3.9 x 150 mm, 5 micrometer particles, 100 Å pore size; Waters, St Quentin en Yvelines, France) was used for reverse phase liquid chromatography (RPLC) at 1 mL/min flow rate with a linear gradient from 0% to 100% B in 40 min (eluent A: 0.1% formic acid; eluent B: 80% acetonitrile, 20% isopropanol with 0.1% formic acid). The injection volume was 100 µL of bacterial extract diluted 1/4 in 0.1% aqueous formic acid. Using a post-column split, 90 % of the effluent was collected (1 min fractions) for further characterization and 10% was analyzed by ESI-MS/MS (LCQ Advantage, ThermoFisher, San Jose, CA, USA) in the positive ion mode. MS/MS spectra were acquired in data dependant scan mode with dynamic exclusion for 1 min after 3 scans. High flow rate settings were used for the ion source: 270°C capillary temperature; voltages were 10 V for the capillary and 4.5 kV for the spray. The analyses were achieved in triplicate.

MALDI-MS/MS analysis

A MALDI Q-ToF Premier instrument (Waters, Manchester, UK) fitted with a nitrogen laser (337 nm) was used for accurate mass and MS/MS analysis of fractions collected from the C4 column. The matrix was alpha-cyano-4-hydroxycinnamic acid (3.6 mg/mL in 50% acetonitrile containing 0.1% TFA). Glu-Fibrinopeptide was used for single point mass correction (Lockmass). Mass accuracy for pseudo-molecular ions was better than 5 ppm.

Measurement of biological activities

Solvent extracts were separated by reversed phase liquid chromatography on a micropreparative scale. This step was followed by the measurement of biological activities on collected fractions. This methodology is suited to the characterization of active compounds produced by Bacillus strains in liquid culture media. Antifungal activities were tested against Botrytis cinerea and Fusarium oxysporum by the agar diffusion method [17]. Fungi were deposited in the centre of Petri dishes containing MALT-Agar solidified medium. Once fungi had started development, a 20 µL aliquot of the solutions to be tested was deposited on the surface of the agar medium. Petri dishes were incubated at 25°C in an aerobic atmosphere for 48-72 h. Growth inhibition was determined by measurement of the halo diameters around the deposit.

Results and discussion

Extraction of lipopeptides
A Bacillus subtilis strain producing a large panel of biologically active compounds including low molecular weight amicoumacines was used for the selection of the best extraction solvent. C14 surfactin was added as a standard for quantitative evaluation of the extraction process, while using a constant volume of bacterial suspension and extraction solvent. The best extraction solvent composition was 1/1 acetonitrile/isopropanol (v/v) containing 0.1% formic acid.

Figures and Tables

[Click to enlarge]

Limitations of fast screening by MALDI 
Direct analysis by MALDI mass spectrometry is widely used for bacterial typing because it gives a rapid overview of the different species present in a sample. However, selective desorption-ionisation effects prevent its application to the exhaustive identification of active substances produced by a bacterial strain [11,18]. Isobaric species cannot be differentiated by MALDI-MS and the presence of different adducts (Na⁺ and K⁺) in lipopeptides [11-12] complicates considerably the identification of these lipopeptides.  Furthermore, low molecular weight compounds such as amicoumacins are hardly detectable by MALDI-MS in a mass range where clusters of matrix ions are dominant. Considering these limitations, we have designed an analytical strategy relying on micro-preparative liquid chromatography combined to ESI-MS/MS analysis. Collected fractions were assayed for antifungal activity, and used for accurate mass determination and MALDI-MS/MS with a Q-ToF instrument when structure assignments were ambiguous. 

Chromatographic conditions
As lipopeptides of the surfactin series are strongly retained by most C18 grafted supports, a C4 support with 0.1 nm porosity was preferred. A clear separation of surfactins and fengycins was observed, with enough selectivity for separating surfactins and fengycins having various fatty acid chain lengths (Figure 2). LC/MS analysis of various Bacillus strains provided evidence for the variability of compositions in the amicoumacin, fengycin and surfactin series (Figure 2a-c, frames A, F and S). LC/MS analysis also highlighted the influence of culture conditions on the relative abundance of compounds produced by Bacillus strains. For the same strain, a higher concentration of fengycins and surfactins was obtained for the Landy medium (Figure 2a), compared to Difco sporulation medium (Figure 2d). Similarly, a higher concentration of fengycins and surfactins was obtained for longer culture times (data not shown). As shown for instance for surfactins, compositions according to the alkyl chain length also vary from one strain to the other, and from one culture condition to the other (Figure 2, frame S). 

Handling of collected fractions
Fractions collected (900 µL) from the C4 column were dried under vacuum and redissolved in 50 µL of solvent (concentration factor was 18). Then, 20 µL of this solution was used to measure biological activities. The use of C14 surfactin as a standard for a quantitative evaluation revealed that a simple addition of solvent such as 1/1 acetonitrile/water was insufficient to bring lipopeptides back into solution, most of the material being lost on the wall of Eppendorf tubes used for fraction collection. At best, a 70 ± 5 % re-dissolution yield was obtained with 1/1/2 (v/v/v) acetonitrile/isopropanol/water containing 0.1% formic acid. This solvent composition was also suited for biological activity assays without further sample handling.

Identification of Amicoumacins
Amicoumacin A is characterized by an amidated terminal group whereas amicoumacin B bears a carboxylate end group (Figure 1). These compounds were never detected by direct MALDI-MS analysis, but were easily characterized by LC-ESI-MS/MS. Amicoumacins A and B had the same retention time on the C4 column (5.3 min, Figure 2) and the same fragment ions in MS/MS. They were thus differentiated by the sole m/z value of the pseudomolecular ion (m/z 424.2 and m/z 425.2 for Amicoumacin A and B, respectively). In this study, only Amicoumacin B was found in strain 2.

Identification of surfactins LC-ESI-MS/MS analysis resolved most spectral discrimination problems encountered with direct MALDI-MS analysis and particularly provided a clear separation between surfactins and other lipopeptides (Figure 2). Identification within the surfactin series by means of LC-ESI-MS/MS relies on fragment ions that characterize the residues of the peptidic cycle (m/z 441, 554, 667 and 685) as well as the length of the β-hydroxy fatty acid chain bound to it (m/z 469 and 582) (Figure 3). Surfactins can be separated by RPLC on the C4 column according to their fatty acid chain length: C12, C13, C14, C15, C16 and C17 surfactins were eluted at 17.2, 17.8, 18.5, 19.2, 20.1 and 20.8 min, respectively (Figure 2). However, surfactins A and C (Leu/Ile substitution) cannot be distinguished by this low energy MS/MS approach. Only surfactins A and/or C were found in the investigated Bacillus strains.

Identification of cyclic and linear fengycins
The presence of ornithine in the structure of fengycins leads to abundant doubly charged ions in ESI mode (Figure 4a-b). Ala/Val isoforms, corresponding to fengycins A and B, were easily distinguished by ESI and MALDI-MS/MS by their fragment ions (Figure 5a) [14 15], and by their LC retention times (for instance C16 fengycin A eluted at 13.1 min, whereas C16 fengycin B eluted at 13.7 min) (Figure 2). In Bacillus strain 2, we detected a series of compounds eluted from the C4 column close to fengycins, but with molecular masses in excess of 18 amu (Figure 4c). Accurate mass measurements allowed attributing this mass difference to a water molecule, suggesting that these compounds were linear forms of fengycins. Sequence ions obtained by MS/MS confirmed the absence of the Tyr-Ile bond, and located Ile at the C-terminus of the linear peptide chain (Figures 4e and 5b).  The C16-C18 linear fengycins A and a C17 linear fengycin B were eluted before their cyclic counterparts (from 11.7 to 12.9 min, Figure 2). Activity tests run on a fraction containing only the linear forms of fengycins did not reveal any antifungal activity (data not shown). Furthermore, we noticed that lesser amounts of linear fengycin forms were found with longer culture times. These observations are in agreement with the fact that cyclisation is the last step of fengycin synthesis, and is an absolute requirement for the biological activity of these compounds [19].

Contrary to MALDI-MS, the use of LC-ESI-MS/MS has the real advantage to separate the different species of lipopeptides, such as surfactins, iturins or fengycins [14-16], even if some of them are isobaric species [15-16]. Moreover, the low amount of adducts does not complicate mass spectrum thanks to the previous chromatographic separation and can be useful if the sodiated ion is used to confirm MS/MS fragmentation [16]. RPLC with C18 columns is usually used to separate lipopeptides [15-16]. In our case, RPLC with a C4 column coupled to tandem mass spectrometry permitted to separate different lipopetides, surfactins and fengycins, but also amicoumacines and linear fengycins.

Conclusions

The analytical methodology that has been set up is adapted to the characterization of the structure and evaluation of the activity of compounds produced by Bacillus. It allowed separating the surfactin series from the cyclic and linear fengycin series. Notably, the fact that linear forms of fengycins are devoid of antifungal activity confirmed that cyclisation is essential for the activity of fengycins. 

Acknowledgement

This work is dedicated to the memory of Anne Chobelet. The authors gratefully acknowledge Luc Négroni and Anne-Marie Elie for their help, and the Conseil Régional d’Aquitaine for financial support.

References

[1]. Falardeau J, Wise C, Novitsky L et al. Ecological and mechanistic insights into the direct and indirect antimicrobial properties of Bacillus subtilis lipopeptides on plant pathogens. J. Chem. Ecol. 39 (7), 869-878 (2013). [CrossRef]
[2]. Urdaci MC, Pinchuk I. Antimicrobial activity of Bacillus probiotics. In: Bacterial spore formers – Probiotics and emerging applications. Horizon Bioscience, Norfolk, U.K, 171-182 (2004).
[3]. Wise C, Falardeau J, Hagberg I et al. Cellular Lipid Composition Affects Sensitivity of Plant Pathogens to Fengycin, an Antifungal Compound Produced by Bacillus subtilis Strain CU12. Phytopathology. 104 (10), 1036-1041 (2014). [CrossRef]
[4] Ray P, Sanchez C, O’Sullivan DJ et al. Classification of a bacterial isolate, from pozol, exhibiting antimicrobial activity against several gram-positive and gram-negative bacteria, yeasts, and molds. J.  Food Prot. 63, 1123-1132 (2000). [CrossRef]
[5]. Iwashita MK, Nakandakare IB, Terhune JS et al. Dietary supplementation with Bacillus subtilis, Saccharomyces cerevisiae and Aspergillus oryzae enhance immunity and disease resistance against Aeromonas hydrophila and Streptococcus iniae infection in juvenile tilapia Oreochromis niloticus. Fish Shellfish Immunol. 43 (1) 60-66 (2014). [CrossRef]
[6]. Sorokulova IB, Kirik DL, Pinchuk IV. Probiotics against Campylobacter pathogens. J. Travel Med. 4, 167-170 (1997). [CrossRef]
[7]. Ongena M, Jacques P. Bacillus lipopeptides: versatile weapons for plant disease biocontrol. Trends  Microbiol. 16 (3), 115-125 (2008). [CrossRef]
[8]. Berdy J. Recent developments of antibiotics research and classification of antibiotics according to chemical structure.  Adv.  App.  Microbiol. 18, 309-406 (1974). [CrossRef]
[9]. Kwon JW, Kim SD. Characterization of an antibiotic produced by Bacillus subtilis JW-1 that suppresses Ralstonia solanacearum. J. Microbiol. Biotechnol. 24 (1), 13-18 (2014). [CrossRef]
[10]. Pinchuk IV, Bressollier P, Sorokulova IB et al. Amicoumacin antibiotic production and genetic diversity of Bacillus subtilis strains isolated from different habitats. Res.  Microbiol. 153, 269-276 (2002). [CrossRef]
[11]. Cheng F, Tang C, Yang H et al. Characterization of a blend-biosurfactant of glycolipid and lipopeptide produced by Bacillus subtilis TU2 isolated from underground oil-extraction wastewater. J. Microbiol. Biotechnol. 23 (3), 390-396 (2013). [CrossRef]
[12]. Kim PI, Ryu J, Kim YH et al. Production of biosurfactant lipopeptides Iturin A, fengycin and surfactin A from Bacillus subtilis CMB32 for control of Colletotrichum gloeosporioides. J. Microbiol. Biotechnol. 20 (1), 138-145 (2010).
[13]. Pennanec X, Dufour A, Haras D et al. A quick and easy method to identify bacteria by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry. Rapid Commun. Mass Spectrom.  24, 384-392 (2010). [CrossRef]
[14]. Chen H, Wang L, Su CX et al. Isolation and characterization of lipopeptide antibiotics produced by Bacillus subtilis. Lett. Appl. Microbiol. 47 (3), 180-186 (2008). [CrossRef]
[15]. Li Xy, Mao ZC, Wang YH et al. ESI LC-MS and MS/MS Characterization of Antifungal Cyclic Lipopeptides Produced by Bacillus subtilis XF-1. J. Mol. Microbiol. Biotechnol. 22, 83-93 (2012). [CrossRef]
[16]. Pecci Y, Rivardo F, Martinotti MG et al. LC/ESI-MS/MS characterisation of lipopeptide biosurfactants produced by the Bacillus licheniformis V9T14 strain. J. Mass Spectrom. 45, 772-778 (2010). [CrossRef]
[17] Kimura H, Sashihara T, Matsusaki H et al. Novel bacteriocin of Pediococcus sp. ISK-1 isolated from well-aged bed of fermented rice bran. Annals of New York Academy of Sciences. 864, 345–348 (1998). [CrossRef]
[18]. Liu H, Du Z, Wang J et al. Universal sample preparation method characterization of bacteria by matrix-assisted laser desorption ionization-time of flight mass spectrometry. Appl. Environ. Microbiol. 73 (6), 1899-1907 (2007). [CrossRef]
[19]. Samel SA, Wagner B, Marathiel MA et al. The thioesterase domain of the fengycin biosynthesis cluster: a structural base for the macrocyclization of a non-ribosomal lipopeptide. J. Mol. Biol. 359 (4), 876-889 (2006). [CrossRef]

All site content, except where otherwise noted, is licensed under a Creative Commons Attribution 4.0 License

---