COOH
4. Separation- Separation-Detection
4.3. High-
Occurrence and Analysis of Phenolic Compounds 77
content due to the fact that these two aglycones are present in the urine as glucuronides.
An improved derivatization procedure, by Stalikas et al., pro-posed an in-vial derivatization–extraction method for the GC–
MS analysis of methylated flavonoids and phenolic acids in various herbs (38). Derivatization takes place under basic con-ditions so that the hydroxyl groups of the analytes are depro-tonated. The anionic nucleophiles are transferred to the organic phase as ion-pairs using a phase-transfer catalyst and are next sub-jected to reaction with methyl iodide (66). Polymer-bound tri-n-butylmethylphosphonium chloride proved to be the best cata-lyst. In the selective-ion monitoring mode, good separation was attained (Fig. 5.4) and the limits of detection of the phenolics in the extracts were 4–40 ng/mL.
Fig. 5.4. The GC–MS(SIM) chromatogram of aMentha spicata fortified extract after derivatization with methyl iodide and phase-transfer catalysis. Peak assignment: (1)p-hydroxy benzoic acid, (2) trans-cinnamic acid, (3) homovanillic acid, (4) vanillic acid, (5) 2-hydroxy cinnamic acid, (6) 4-hydroxy cinnamic acid, (7). syringic acid, (8) ferulic acid, (9) naringenin, (10) galangin, (11) kaempferol, (12) luteolin, and I.S. internal standard (Reprinted with permission from Stalikas et al.
(38)).
78 Stalikas
length and usually with an internal diameter of 3.9–4.6 mm are employed. Particle sizes are in the range of 3–10 m. Narrow-bore columns (internal diameter 2 mm) are also recommended for special HPLC–MS applications (67).
The quality of sorbent (purity, end-capping) and the particle size do not have a very significant influence on the peak symmetry of the analyzed phenolic acids but the chromatographic resolu-tion and the efficiency of the column are better for columns with very good free silanol group covering, end-capping, or embed-ding. For the analysis of phenolic compounds in beer with LC, separation conditions were optimized as regards 11 different sta-tionary phases (all C18 bonded silicas) for a standard mixture of several flavone aglycones and glycosides (68). Four columns qual-ified as the most appropriate. In the same context, different RP columns of conventional dimensions were applied for the anal-ysis of flavonoid glycosides. According to other authors, a col-umn which is well end-capped should be preferred to counter-parts lacking end-capping because residual silanol groups appear to impair the separation of flavonoid glycosides (69).
A fast RP-HPLC method for the simultaneous separation of 11 flavonoid aglycones was developed using a monolithic continuous-bed column (Chromolith Performance C18 end-capped, from Merck). The method was successfully applied to the analysis of these compounds in complex natural samples such as propolis and Ginkgo biloba (70).
Recently, improved chromatographic performance was achieved by the introduction of ultra-performance liquid chro-matography (UPLC) capitalizing on basic chromatographic prin-ciples to perform separations using columns packed with smaller particles and/or at higher flow rates. Four flavonoids were deter-mined in the flower of Trollius ledibouri from different sources (71). The analysis was performed on an AcQuity UPLC BEH C18 column using gradient elution with a mobile phase of 0.1%
acetic acid and acetonitrile, over 20 min.
Stereochemistry in the studied field is rarely an issue in the recent literature. A systematic work on the enantiomeric separa-tion of flavanones and the diastereomeric separasepara-tion of flavanone glycosides has been reported by Ficarra et al., who utilized four chiral liquid chromatographic columns based on polysaccharide derivatives (72).
As regards the elution conditions, both isocratic and gradi-ent elution are routinely applied for the analysis of phenolic com-pounds. The choice depends on the number and type of ana-lytes and the nature of matrix of concern. Methanol, acetonitrile, and tetrahydrofuran are the most commonly used organic modi-fiers, as well as acidified aqueous solvents such as aqueous acetic acid, formic acid, phosphoric acid, or trifluoroacetic acid. In some cases, acetonitrile leads to better resolution in a shorter analysis
Occurrence and Analysis of Phenolic Compounds 79
time than methanol and, generally, acetonitrile gives sharper peak shapes, resulting in a higher plate number. Occasionally, tetrahy-drofuran and 2-propanol as less polar solvents with their high elu-tion strength have been used.
The greatest alteration observed in the mobile phases was the type of acid used as modifier to minimize peak tailing. Pheno-lic acids have pKa values of about 4 while flavonoids presenting several ionizable hydroxyl groups have pKa values relatively close to each other but certainly greater than 4. Therefore, the recom-mended pH range for the HPLC assay is 2–4. Dalluge et al. found that a deactivated C18column in combination with trifluoroacetic acid as the acidic modifier of the mobile phase, greatly improves peak shape and reproducibility of retention times of catechins in tea (73). Besides, the use of aqueous buffers (citrate, phosphate, acetate) instead of the addition of acid is recommended at con-centrations from 5 to 50 mM.
Free phenolic acids (chlorogenic, protocatechuic, p-hydroxybenzoic, caffeic, vanillic, syringic, p-coumaric, and ferulic) could be separated in medicinal plants or pharma-ceutical preparations using a simple isocratic mobile phase (methanol–water–acetic acid) (74). When phenolic acids of different chemical structures and different polarities have to be analyzed simultaneously, gradient elution is indispensable. As a general observation, phenolic acids are eluted from RP columns according to decreasing polarities. The loss of polar hydroxy groups and the presence of the methoxy groups or ethylenic side chains could decrease the polarity and increase the retention time.
Some phenolic acids could be present in natural plants as geo-metric isomers. The greatest number of phenolic acids occurs in nature as trans-isomers, but on exposure to UV radiation or day-light they are gradually transformed to cis-isomers, which elute usually, before trans-isomers. Their simultaneous separation is usually possible using RP stationary phases including an optically active molecule in the mobile phase (75).
Finally, as far as detection means is concerned, phenolics are commonly detected using ultraviolet-visible (UV-VIS), photodi-ode array (DAD), and fluorescence detectors. Because every phe-nol exhibits a higher or lower absorption of UV or UV-VIS light, given the intrinsic existence of conjugated double and aromatic bonds, UV detection is the ideal method to localize a phenol in the effluent of a column. Most of the benzoic acid derivatives have absorption maxima at 246–262 nm, with a shoulder at 290–
315 nm, except for gallic and syringic acids with absorption max-ima at 271 and 275 nm, respectively. The cinnamic acids absorb in two ranges: 225–235 and 290–330 nm. Detection at 280 nm is the most generally used wavelength for the simultaneous sepa-ration of mixtures of phenolic acids.
80 Stalikas
All flavonoid aglycones contain at least one aromatic ring and, consequently, efficiently absorb UV light. The first maximum, which is localized in the 240–285 nm range, is due to the A-ring and the second maximum in the 300–550 nm range is attributed to the substitution pattern and conjugation of the C-ring. Rea-sonably, UV detection became the preferred tool in LC-based analyses and, even today, LC with multiple-wavelength or DAD is the prevalent tool in studies dealing with, e.g., screening, quan-tification, and provisional sub-group classification.
For a comprehensive HPLC-UV method of flavonol deter-mination in human urine and plasma, readers can refer to the publication of Hollman (76).
Fluorescence detection in phenolic acid analysis is scant. The nature of the functional groups and their substitution pattern determine whether a particular flavonoid is fluorescent or not.
In case that fluorescence detection is feasibile in combination with UV, it offers the possibility to discriminate between fluores-cent and non-fluoresfluores-cent co-eluting compounds (77). To extend the application range of fluorescence detection, derivatization has been used. For example, quercetin, kaempferol, and morin, with their 3-OH, 4-keto substituents, can form complexes with metal cations, some of which are highly fluorescent (78, 79).
Other systems used for the detection of phenolics encompass electrochemical, mass spectrometric, and nuclear magnetic reso-nance (NMR) detectors.
Electrochemical detection is very sensitive for the compounds that can be oxidized or reduced at low-voltage potentials. Pheno-lic acids in food and human plasma extracts are routinely detected by HPLC-electrochemical coulometric detection +600 mV (80).
A multi-channel coulometric detection system, being compati-ble with gradient elution, may serve as a highly sensitive way for detecting phenolic acids and flavonoids in a wide range of sam-ples well as it can be applicable to the overall characterization of antioxidants (81).
For many years, liquid chromatography–mass spectrometry (MS) systems have been applied for the detection and identifica-tion of flavonoid glycosides in plants extracts and various biologi-cal fluids. In some cases, HPLC with different sensitivity detectors (UV, electrochemical, fluorescence) and HPLC–MS are simulta-neously used for the identification and quantification of phenolics in natural plants and related food products.
Electrospray ionization MS has been employed for structural confirmation of phenolics in plums, peaches, grapeseeds, soy-food, cocoa, olive oil, and walnut leaves (82) in human urine and plasma (83) with a polymeric chromatographic column. Conven-tional RP columns were coupled to DAD detector and a magnetic sector-type MS equipped with an electrospray ionization source was applied to the analysis of flavonoid glycosides in Crataegus
Occurrence and Analysis of Phenolic Compounds 81
extract (69). It was demonstrated that UV spectra and first-order electrospray ionization mass spectra allowed a fast characteriza-tion of flavonoids even if reference compounds are not at hand or available.
Electrospray ionization and atmospheric pressure chemical ionization interfaces were both used combined with HPLC for the quantitative analysis of flavonoids and their metabolites in biofluids (84). The detection was carried out in the positive or negative ion mode; data could be collected in multiple-reaction monitoring, multiple-reaction monitoring, and selective-ion monitoring mode.
Detailed investigations on the identification of flavonoid metabolites after the consumption of onions were published by Mullen et al. (85). The analysis was performed by LC-ion-trap MS and 23 metabolites of quercetin were identified, which are illustrated in HPLC traces of Fig. 5.5. The differ-ent classes of metabolites that were detected can be summa-rized as follows: quercetin monoglucuronides, quercetin diglu-curonides, methylquercetin monogludiglu-curonides, methylquercetin diglucuronides, quercetin, quercetin sulfates, quercetin curonide sulfates, quercetin glucosides, quercetin glucoside glu-curonides, and quercetin glucoside sulfates.
Unquestionably, NMR is the technique that generates more information for unambiguous identification of a molecule. In a
Fig. 5.5. Gradient reversed-phase HPLC with detection at 365 nm of quercetin metabolites in (a) a plasma extract and (b) urine obtained from a human volunteer after the consumption of fried red onions. Separation was carried out using a 250× 4.6 mm i.d. 4 m Synergi Max-RP column eluted with a 60 min gradient of 5–40% acetonitrile in 1% formic acid, at a flow rate of 1 mL/min and maintained at 40◦C. Peaks 1–23 belong to the compound classes reported in the text (Reprinted with permission from Mullen et al. (85)).
82 Stalikas
recent study on the flavonoid constituents of a red clover extract, stopped-flow LC-NMR and stand-alone NMR were used to iden-tify structural isomers that could not be distinguished on the basis of MS/MS information (86). A recently developed cryoflow NMR probe exhibits detectability about fourfold better than with conventional probes or, alternatively, the scan time is 16-fold shorter for the same amount of sample. The probe has been applied for the analysis of an oregano extract where five flavonoids were identified using an LC–UV–solid-phase extraction–NMR–
MS setup (87).
Other less common means of detection, coupled to LC, have been through refractive index and evaporative light scattering detection. The latter offers freedom from some of the limitations of spectroscopic detection because it is not limited to compounds that contain UV-absorbing chromophores and it is immune to mobile-phase variations and gradient baseline shifting. Although being less sensitive than those previously described, both of the detectors have been successfully used: the HPLC-refractive index system in the quantification of (3,3,4,4,5,7-hexahydroxyflavan) in unripe banana pulp (88) and the HPLC-evaporative light scattering detection in the determination of Radix Astragali flavonoids (89).
4.4. Capillary