Buruli ulcer (BU) infections caused by Mycobacterium ulcerans
require a rapid diagnosis and initiation of antibiotic therapy at an early stage in order to prevent the devastating consequences of advanced ulceration. The pathogenesis of M. ulcerans
is dependent on the activity of mycolactone, a plasmid-encoded polyketide-derived macrolide toxin [1
]. Mycolactone is responsible for the cytotoxicity [5
], immunosuppression [7
], and painless ulcer formation [8
] characteristics of BU disease. The toxin has been detected in the blood and tissue of mice [9
] and patients [12
] prior to ulcer formation, making mycolactone a promising biomarker for BU diagnostics. However, the unique biochemistry of this lipid-like toxin has hampered research in the host-pathogen biology of this disease, and consequently, the development of a simple point-of-care diagnostic tool. There are at least five known structural variants of mycolactone produced by closely-related mycobacteria, all of which contain a highly conserved 12-membered lactone ring and upper side chain, as well as a variable unsaturated fatty acyl lower side chain [1
]. Of these, mycolactone A/B is the most toxic form [19
]. Even minor structural alterations to the lower (termed southern side chain) can significantly reduce the biological activity of the toxin [20
], and this includes labeling with dye [20
] and biotin [26
] for use in laboratory experiments and studies.
Purification of natural mycolactone in the form of lipid extractions from bacterial culture has been performed by chloroform–methanol (2:1 v
) on liquid broth cultures [3
] or ethanol on bacterial colonies scraped directly from agar plates [28
]. Further purification can then be done using thin-layer chromatography (TLC), whereby the UV-active toxin can be separated from other lipids in the extract [3
]. However, it is known that mycolactone purified from M. ulcerans
(herein referred to as natural mycolactone) is often contaminated by unknown compounds [17
]. Total chemical synthesis of mycolactone has provided a means for obtaining robust, homogenous, well-defined samples with high reproducibility [17
]. However, total synthesis of mycolactone is an extremely complicated process and requires substantial synthetic expertise, thus limiting the global supply of this crucial material [18
There is evidence of variability in assay methods, handling conditions, and purification strategies for this lipophilic toxin. Lipids stick to plastic surfaces, aggregate in aqueous solutions, and are sequestered by other lipid-containing molecules [34
]. Previous work from our team and others have shown that amphiphilic biomarkers of infection are carried by host lipoproteins in the aqueous milieu of blood, and this sequestration is critical for the transport, toxicity, and pathogenicity of the organism, while influencing immune induction mechanisms as well [34
]. It has been suggested that the lipid-like structure of mycolactone associates with other lipophilic molecules, such as lipid membranes [19
], and we hypothesize that lipoproteins constitute a major category of lipophilic host moieties with which mycolactone may interact. We propose that generating antibodies to mycolactone is not only challenging because of its low antigenicity [19
], but also owing to the physical interaction of mycolactone with other lipidic molecules in physiological solutions. These interactions must either be accounted for or disrupted in any detection platform in order to improve sensitivity in physiological conditions.
Several investigators have explored the conditions for optimal storage and handling of mycolactone. Marion et al. provided a thorough characterization of mycolactone handling and storage conditions, showing that the toxin is sensitive to photodegradation (50% reduction in toxicity after only 7.5 min of exposure to UV light) and stable at high temperatures (100 °C for 6 h) [40
]. Toxin degradation was most profound under UV-irradiation (254–365 nm), and to a lesser extent in sunlight and artificial light. The toxin was found to be stable under red light. Rearrangement in mycolactone chemical structure as a result of photodegradation was shown to significantly reduce cytotoxicity [41
]. Storage in glass tubes slowed the photodegradation slightly when compared to quartz tubes, while storage in amber tubes showed partial protection from UV light, and complete protection from photodegradation by natural light [40
]. Mycolactone stored in ethanol was also shown to be more stable than the toxin stored in acetone, while storage in aqueous solutions was not examined. Similarly, storage of toxin-containing samples in ethanol proved crucial to the detection of mycolactone in mouse footpads by TLC [10
]. Whereas these studies shed some light on optimal conditions for the storage and characterization of mycolactone, a systematic study of the impact of various solvents, storage conditions, and measurement techniques on the detection of mycolactone in laboratory and physiologically relevant matrices does not exist. In this manuscript, we present a systematic assessment of the impact of storage conditions (plastic, glass, and others), solvents (ethanol, water, media, and others), and approach (UV-Vis, mass spectrometry (MS), and TLC) on the measurement of mycolactone A/B (synthetic vs. natural), and expand the findings to assess the interaction of the lipophilic toxin with carrier moieties in host samples. Beyond highlighting the most suitable conditions for laboratory testing of mycolactone A/B, we hope to document the differential association and presentation of the toxin in serum, which is required for the better understanding of host-pathogen biology, as well as the development of diagnostic approaches.
Mycolactone produced by Mycobacterium ulcerans is a major diagnostic and therapeutic target for Buruli ulcer. In this study, we characterized the toxin from a biochemical perspective, appreciating its amphiphilicity and speculating on the consequent impact and interactions in aqueous milieu. Expanding on work from other investigators, we defined preferred conditions for the storage, purification, and handling of the toxin, highlighting the discrepancy between various detection methods in their ability to accurately measure mycolactone concentration. CMC measurements indicate that natural mycolactone is likely to behave very differently from modified (e.g., biotinylated) toxin, and that these variabilities should be considered in the design of experiments. Ethanol or a surfactant like SDS can be used to solubilize the toxin in aqueous media in order to enhance its detection by UV-Vis. On the other hand, the storage container material (glass, plastic, or siliconized plastic) proved to have little impact on mycolactone recovery under the conditions tested.
Most significantly, our results demonstrate that lipophilic mycolactone associates with host lipoproteins in human serum and may be present in this sequestered form in physiologically relevant samples. This observation is also supported by the fact that we are unable to measure the toxin in media without serum or lipoproteins. These results help to explain the difficulty in generating antibodies against mycolactone as it is most likely presented in a sequestered form with other lipophilic molecules, including lipoproteins, and other host-derived assemblies [19
]. Amphiphiles are extremely common in biological systems. Most conserved bacterial signatures, those routinely recognized by the host innate immune system (i.e., lipopolysaccharides, lipoteichoic acids, lipoarabinomannans), are amphiphilic in nature. Previous work from our group has shown that these molecules associate with host lipoproteins, which not only impacts their availability for detection but also directly affects the induction of innate immune signaling [34
]. Yet, little consideration is often given to the biochemistry and physiological presentation of these moieties when measurements are taken. These associations may be used to develop schemes for the direct detection of mycolactone in serum, using lipoproteins as a means for capturing and concentrating the toxin, a strategy developed by our team termed lipoprotein capture [35
]. Further characterization of the interactions between the toxin and host lipoproteins, such as HDL and LDL, will be crucial in order to achieve the level of sensitivity required in a clinical diagnostic tool.
These studies indicate the significance of amphiphilic biomolecules on host-pathogen interactions and associated pathogenesis, and the need for further study of mycolactone specifically. Incorporating the biochemical properties of the molecule in question is critical to an accurate understanding of pathogenesis, and for the development of physiologically relevant diagnostic approaches. In this manuscript, we present new insights into the characterization of amphiphilic mycolactone. Our findings indicate the need for more reliable standards and purification methods, and specifically call for broader efforts to synthesize the toxin for consistency in research and development.
4. Materials and Methods
4.1. Materials and Reagents
Synthetic mycolactone A/B was a kind gift of Professor Yoshito Kishi (Harvard University, Cambridge, MA, USA) and biotinylated mycolactone was a kind gift of Dr. Caroline Demangel (Pasteur Institute, Paris, France). Natural mycolactone was obtained from M. ulcerans
bacterial culture (strains 1615 or 1059). Mycolactone concentrations were quantified by UV-Vis spectrophotometry as described previously (λmax
= 362 nm; log ε = 4.29) [29
]. Purified human high- and low-density lipoproteins (HDL and LDL, respectively) were purchased from Bio-Rad (catalog no. 5685-2004 and 5685-3204), resuspended in filter-sterilized nanopure water, and stored at 4 °C until use.
4.2. Natural Mycolactone Purification
strains 1615 [50
] and 1059 [51
] were kindly provided by Dr. Pamela Small, University of Tennessee. M. ulcerans
1059AL is an autoluminescent derivative of strain 1059 [52
]. Bacterial strains were grown on selective Middlebrook 7H11 agar. Colonies from 2 or 3 confluent plates were scraped into a pre-weighed Eppendorf tube and weighed. The material was then put in a foil-wrapped 125 mL glass flask with 10 mL absolute EtOH and stirred for 2 h at room temperature. Using a Pasteur pipette, the dissolved or suspended material was transferred to a pair of glass centrifuge tubes, the tops were crimped, and the tubes were spun at ~3000 rpm for 20 min at 4 °C. The ethanol phase was collected into a 25 or 50 mL round bottom glass flask and dried under nitrogen with heat set between 50–60 °C. The residue was resuspended in 200–300 µL of absolute EtOH and stored at −20 °C. Thin-layer chromatography (TLC) was performed as previously described [2
]. Briefly, ethanol extracts from M. ulcerans
were applied to a silica TLC plate and after drying, mobilized using chloroform:methanol:water (90:10:1 v
) as a solvent system. Mycolactone-containing lipids were extracted as a light-yellow band following visualization under UV light, resuspended in absolute ethanol, and stored in amber glass vials at −20 °C until use.
4.3. Electron Spray Ionization Mass Spectrometry (MS)
Ethanol extracts were diluted in acetonitrile and directly perfused into an electrospray ionization source on a Waters Synapt G2S mass spectrometer. The MS conditions were initially optimized to erythromycin, as well as a synthetic mycolactone standard before applying them to the ethanol extracts. The MS conditions were: flow rate 50 µL/min over a 2 min period; capillary voltage 2.98 kV; cone voltage 75 V; desolvation temperature 550 °C; desolvation gas 945 L/h; source temperature 150 °C; acquisition range 100–2000 m/z. Mycolactone A/B was detected by the presence of the more abundant sodium adduct [M + Na]+
(m/z 765.5); the protonated molecular ion [M + H]+
(m/z 743.5), and the dehydrated protonated molecular ion [M + H − H2
(m/z 725.5) as described previously [2
4.4. Cytotoxicity Assays
L929 mouse fibroblasts (ATCC CCL-1) were purchased from the American type culture collection and maintained in Eagle’s minimum essential medium (Corning, 10009CV) with 10% horse serum (Sigma-Aldrich, H1270-500ML) in tissue culture flasks and incubated in 5% CO2 at 37 °C. In this manuscript, Eagle’s minimum essential medium containing 10% horse serum is referred to as “EMEM”, while media without serum is called “EMEM no serum”. Cells were passaged at least 3 times prior to use in cytotoxicity experiments. 24-well tissue culture plates were seeded with 5 × 104 cells per well and allowed to adhere overnight. Medium was discarded and replaced with 500 µL of fresh EMEM containing the desired concentration of mycolactone. Mycolactone dilutions in EMEM were performed in glass test tubes protected from light. After 48 h of incubation, cytotoxicity was measured using the resazurin-based reagent PrestoBlue™ (A13262, Invitrogen) by adding 50 µL of reagent and incubating at 37 °C for 60 min. Measurements were taken using a SpectraMax Gemini EM plate reader (Molecular Devices, Sunnyvale, CA, USA) with excitation at 560 nm and emission at 590 nm. Relative fluorescence values plotted are the mean ± standard deviation (n = 4 replicate wells).
4.5. Mycolactone Quantification by UV-Vis
Synthetic mycolactone A/B stocks at 0.2 mg/mL or 0.1 mg/mL in acetonitrile were diluted 1:20 or 1:10 (respectively) in the desired test media condition in a glass vial and then subdivided into a total of 3 glass vials (National Scientific, C5000-995), 3 plastic tubes (Fisher Scientific, 05-408-137), and 3 siliconized low-retention plastic tubes (Fisher Scientific, 02-681-320) for each test condition (technical triplicates). Samples were then incubated at room temperature protected from light. At various time points ranging from 10 min to 24 h, quantification of mycolactone by UV-Vis absorption at 362 nm was performed using a NanoDrop™ ND-1000 spectrophotometer (Thermo Scientific). Measurements were taken using 2 µL of each sample. Blanks prior to each measurement were performed using the appropriate media conditions in the absence of mycolactone. In the case of HDL and LDL, absorbance of lipoprotein (at the indicated concentration) in the absence of mycolactone was subtracted from the value given for each lipoprotein sample containing mycolactone [Abs362 with mycolactone − Abs362 no mycolactone]. Experiments were performed at least twice. Values plotted are the mean absorbance values at 362 nm ± standard deviation.
4.6. Critical Micelle Concentration (CMC)
The CMC of mycolactone A/B was determined using the detergent critical micelle concentration (CMC) assay kit (ProFoldin, CMC1000) as per the manufacturer’s instructions. Briefly, synthetic (1 mg/mL stock in ethyl acetate) or biotinylated mycolactone (1.125 mg/mL stock in ethanol) was diluted in filter-sterilized nanopure water in 2-fold dilutions ranging from 60–0.03 µM, including a “no mycolactone” negative control in 50 µL volumes in duplicate. Then 50 µL of 1X CMC dye (also diluted in water) was added to each sample, bringing the final mycolactone concentration to 1X at 30–0.015 µM. The samples were incubated at room temperature and protected from light for 30 min. The fluorescence was measured using a SpectraMax Gemini EM plate reader (Molecular Devices, Sunnyvale, CA, USA) with excitation at 360 nm and emission at 465 nm. When in monomeric form, there will be little to no change in fluorescent signal as compared to the negative control. Once the CMC is reached, aggregation will occur, and dye will be incorporated within the micelle resulting in an increased fluorescent emission at 465 nm. Thus, the measured fluorescence values for each mycolactone concentration were divided by the value of the negative control signal to yield normalized fluorescence values (n = 2). In monomeric form, the normalized fluorescent value will equal 1. Therefore, a normalized value greater than 1 indicates micelle formation, which was identified by testing for statistical significance (Student’s t-test, P < 0.05) pairwise between samples of 2-fold increasing concentrations of mycolactone. The CMC was defined as the 2-fold concentration range in which a statistically significant increase in normalized fluorescence units first appears. Experiments were performed twice.
4.7. Statistical Analysis
Absorbance values are plotted as means ± standard deviation. Data were analyzed using Students t-test to determine statistical significance between individual groups. A significance level (P) of less than 0.05 was considered statistically significant (*** P < 0.001, ** P < 0.01, or * P < 0.05).