|Home | About | Journals | Submit | Contact Us | Français|
This study is the first on combined HPLC and MALDI-TOF MS analysis of phenolic acids. The analyses were carried out for phenolic acid mixtures and showed a unique, individual co-crystalline pattern for each phenolic acid. HPLC could distinguish phenolic acids and MALDI-TOF MS provided comparable mass (m/z) profiles for the samples. This combined study proved to be rapid in the accurate identification and structural analysis of phenolic acids with different masses.
Phenolic acids seem to be universally distributed in the plant kingdom, essential for the growth and reproduction of plants, and are produced as a response to defense against pathogens. The importance of antioxidant activities of phenolic compounds and their possible usage in processed foods as a natural antioxidant has received attention in recent years. These compounds are diverse in structure but are characterized by hydroxylated aromatic rings (e.g., flavan-3-ols) and polymerized into larger molecules.1
Due to the abundance of different classes of phenolic acids and their diverse chemical properties, a variety of separation and identification methods have been developed using thin layer chromatography (TLC), high-performance liquid chromatography (HPLC), and gas liquid chromatography (GLC). In general, these methods are useful for identification but do not provide information on specific chemical composition. Analytical techniques have been coupled to mass spectrometers (MS), creating gas chromatography MS (GC-MS), liquid chromatography MS (LC-MS), inductively coupled argon plasma MS (ICP-MS), supercritical fluid MS (SCF-MS), nuclear magnetic resonance MS (NMR-MS), and infrared MS (IR-MS). The main advantage of MS is that it not only provides information on the molecular mass of a compound of interest, but can also generate structurally significant information. Soft ionization techniques—initially fast atom bombardment (FAB)2 and plasma desorption (PD)3, followed by the development of electrospray ionization (ESI)4—have been used for phenolic acid analyses. A limitation of the mobile phase for ESI in LC-MS analysis is that it accepts only volatile salts or solvents.
Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) is an efficient tool for large biomolecule analysis.5 MALDI-TOF MS has several advantages over other methodologies, including speed of analysis, high sensitivity, and wide applicability with a good tolerance toward contaminants, and ability to analyze complex mixtures.6 Recently, MALDI-TOF MS has been extensively used in food analysis for oligomeric polyphenol and anthocyanin detection.7,8 This article presents MALDI-TOF MS of individual phenolic acids from the mixtures of phenolic compounds, without the aid of matrix.
Stock solutions of ferulic acid, caffeic acid, 3,5-dihy-droxybenzoic acid, and vanillic acid (Figure 1) (Sigma, St. Louis, MO) were prepared in 1 mL 80% aqueous MeOH (methanol) solution at a concentration of 100 ppm. Phenolic acid mixtures were analyzed by HPLC (Agilent 1100 series) coupled with a UV-vis diode array detector (DAD, G1315B) using a ZORBAX-Eclipse XDB-C18 column (4.6 × 150 mm, particle size 5 μm) with a polar mobile phase comprised of 32:68 methanol:1 mM aqueous trif-luoroacetic acid (TFA) at a flow rate of 1 mL min−1 for 20 minutes and monitored at 310 and 254 nm. Before injection, the samples were filtered on Millex-SR 0.22-μm filters (Millipore, Billerica, MA). Selected peaks of the HPLC chromatogram were collected using a fraction collector (Gilson, Gambetta, France) coupled with HPLC, each fraction was concentrated (100 μL) by Speed-Vac, and then 50 μL of acetonitrile was added to each sample vial. Two microliters of sample was employed to a 100-well stainless steel MALDI sample plate.
Prior to MALDI analysis, the co-crystalline pattern of each phenolic acid on MALDI sample plate was captured directly from video monitor coupled with MALDI MS using a digital camera. MALDI analysis was performed on an Applied Biosystem Voyager-DE PRO MALDI TOF mass spectrometer with a nitrogen laser (337 nm) operated in an accelerating voltage (20 kV). Each spectrum was collected in the positive ion linear mode as average of 100 laser shots of predetermined or random positions across a spot. The data were externally calibrated using angiotensin and ACTH (Applied Biosystems, Foster City, CA). Reproducibility of each spectrum was checked 20 times from duplicate prepared sample.
Liquid chromatography is routinely used for the separation of individual phenolic compounds from complex mixtures, based on their different affinities for a resin-packed column. Changing the pH and/or ionic strength of the solution helps to allow the compounds of interest to elute in a sequential manner. Scalbert et al.9 described the chromatographic separation of cyanidin and delphinidin from methanolic extracts of oak heartwood using a C-18 Novapak column. Tamagnone et al.10 used HPLC coupled to a diode array detector to evaluate phenolic compounds of leaf, using a C-18 column with a flow rate of 1 mL/min. The solvents consisted of 10% (v/v) methanol in water with 1 mM TFA (solvent A), and 80% (v/v) methanol in water with 1 mM TFA (solvent B). The solvent gradient was 90% solvent A and 10% solvent B to 100% solvent B for 60 minutes. Morreel et al.11 also used a C-18 column with 1% aqueous triethylammonium acetate (TEAA) as solvent A, and a 25:75 mixture of methanol and acetonitrile containing 1% TEAA as solvent B, at a flow rate of 0.3 mL/min with a gradient of 100% solvent A to 100% solvent B in 40 min. We have earlier reported on caffeic acid and rosmarinic acid separation using HPLC performed with an RP C18 Techsphere (Macclesfield, UK) 50 ODS column (4.6 × 250 mm) and a mobile phase of 40:60 methanol:aqueous 1 mM TFA at a flow rate of 1 mL min−1.12 Figure 2 clearly indicates a good separation of phenolic acids by this HPLC system. The identification of the compounds is based on a combination of the retention time and either a UV-vis spectrum or a mass spectrum. Large searchable databases exist to aid with the identification of compounds based on mass spectral data.
For MALDI-TOF MS analysis, each fraction was spotted onto a MALDI stainless steel plate and, interestingly, each fraction showed characteristics unique to the co-crystallization pattern on the MALDI plate (Figure3). We used acetonitrile in MALDI sample preparation to allow for better co-crystallization, which is probably a consequence of acetonitrile’s ability to form analyte-doped crystals on the target.13 The co-crystallization pattern might be useful to generate a rapid idea of phenolic compound identity, in comparison to the reference compound.
A positive ion linear mode MALDI-TOF spectrum of coeluted HPLC fraction 1 presents a series of masses (m/z 637, 659, 842, and 996) (Figure 4). The difference between two intense fragment ions was m/z 154 (m/z 996–m/z 842), corresponding to the mass of 3,5-dihydroxybenzoic acid (DHB). The ion at m/z 637.57 [4M + Na+] might be the tetramer of DHB in addition to sodium ion. MALDI-TOF MS of polyflavons tends to favor an association with sodium [M + Na+] and potassium [M + K+] ions over the formation of a protonated molecular ion [M + H+].14 The analyte was applied directly to MALDI-TOF from HPLC without addition of sodium salt, assuming that naturally occurring Na+ are abundant in the sample. Another observed value was m/z 659 (4M + 2Na+), predicted to be the tetramer of DHB with multiple sodium (Na+) ions. The difference between two (m/z 659 and m/z 842) ions was m/z 183 (154 + 29), representing one DHB monomer adduct with C2H5+ (m/z 29) ions. MALDI-TOF MS analysis of fraction 2 represents a distinct oligomeric series of phenolic acid units. The observed masses of fraction 2 were m/z 604, 685, 798, and 835 (Figure 5). The possible explanation of observed masses m/z 604 and m/z 798 corresponds to trimer and tetramer of ferulic acid, respectively, in addition with one sodium ion (3M + Na+). Evidence comes from the mass difference of two distinct ions (m/z 604 and m/z 798) equal to the mass of ferulic acid m/z 194. The observed m/z 685 (m/z 604 + C2H9O3+) might be equivalent to the addition of a C2H9O3+ ion with trimeric ferulic acid.
A similar pattern was also observed in the case of fraction 3. The caffeic acid trimer m/z 563 (3M + Na+) and tetramer at m/z 743 (4M + Na+) were observed. Each monomer was m/z 180, equal to the exact mass of caffeic acid. Major intense ions between m/z 563 and m/z 743 were m/z 611 (m/z 563 + CH4O2+) and m/z 655 (m/z 611 + C2H4O+) (Figure 6). However, in the case of fraction 4 the highest intense peak was observed at m/z 673 (Figure 7), corresponding to the tetramer of vanillic acid without any addition to metal cation, indicating vanillic acid has less affinity to metal ions compared with other tested phenolic acids. Other observed ions were at m/z 781 and m/z 933, and the difference of these two ions (m/z 152) might be vanillic acid (m/z 168) with loss of one hydroxyl group. Another loss of C2H4O2+ (m/z 60) ion from a monomer of vanillic acid was observed between the first two major ions (m/z 108).
Mass distribution provides evidence for a particular phenolic acid identification from existing mass spectral databases. Moreover, the data have provided insight into the processes involved in phenolic acid co-crystallization that might be useful for partial identification. The results indicate that characterization of HPLC-fractionated phenolic acids is effective by MALDI-TOF mass fingerprinting.
The authors gratefully acknowledge the Central Research Facility, Indian Institute of Technology Kharagpur, India, for providing the MALDI-TOF MS facility.