Mass Spectrometry of Flavonoids

A. Mass Spectrometry of Flavonoids

Flavonoid herbs have been increasingly studied by mass spectrometry (MS) since the introduction of the thermospray (TSP), ESI, and APCI interfaces, which allow direct coupling of MS with HPLC. Being characterized by ‘‘soft’’ ionization, these techniques permit the analysis of flavonoids in their native form without derivatization [21]. TSP-MS was used, at first, to analyze flavonoids in different plant extracts, such as Arnica montana, Gentianaceae species, Ginkgo biloba, Calendula officinalis, and Hypericum perforatum [11]. Unfortunately, TSP-MS fails in the case of thermolabile compounds, such as the flavonol-glycosides.

These compounds undergo fragmentation and yield mainly the aglycone fragment [A+H]+, with the molecular ion [M+H]+ present in very low quantity.

This was one reason to switch to the ESI and APCI interfaces, which involve a low level of fragmentation. Both ESI and APCI produce mainly molecular ions, and they are particularly suitable for detecting intact molecular species present in herbal extracts. For further structural information, these interfaces may be coupled to an ITMS analyzer to promote mass fragmentation, and this arrangement provides data helpful in identifying the flavonoids of interest.

1. Electrospray Ionization Mass Spectrometry

ESI-MS produces ions as a result of the application of a potential to a flowing liquid, which causes the liquid to charge and spray. Electrospray forms very small droplets of solvent containing the analytes. Usually, the solvent is removed by heat and multicharged ions are produced. As previously stated, ESI has the advantage over TSP of producing low fragmentation of flavonoid derivatives.

To elucidate the difference between TSP and ESI, the following example is illustrative. TSP-MS of kaempferol-3-O-rutinoside (molecular weight [MW] 594 da) mainly produces kaempferol fragment (m/z 286) resulting from the loss of rutinose ([M-rutinose+H]+), and the kaempferol-rhamnoside fragment (m/z 433, [M-glucose + H]+). The molecular ion (m/z 595 = [M+H]+) is present in very low quantity. Conversely, ESI-MS of the same glycoside predominantly yields the sodium and potassium adducts of the molecular ions: ([M+Na]+), m/z 617; ([M+K]+), m/z 633. Fragmentation ions are almost absent.

For this reason, ESI-MS is particularly suitable for direct analysis of samples without preliminary chromatographic separation. As a result, specific fingerprints of complex natural mixtures are easily and rapidly obtained. This information on the overall components is particularly valuable for herbal medicines, because they are in toto regarded as the active principle rather than single constituents.

Closely related to ESI is APCI in that the source operates at nearatmospheric pressure. APCI produces almost molecular ions with very little fragmentation, and it provides fingerprints of herbs [22].

2. Ion Trap Mass Spectrometry

MS spectra with fragmentation of molecules require collision-activated dissociation (CAD) and triple quadrupole analyzers. In these instruments, the analysis is performed as follows: the first quadrupole selects the interesting ion (parent ion), the second produces the fragments from the isolated ion, and the third quadrupole analyzes the fragmentation products (daughter ion spectrum). These steps (ion isolation, fragmentation, and analysis) can be repeated by addition of n quadrupole devices (multisector mass spectrometer) to allow multiple MS/MS experiments (MSn) to be performed.

As an alternative, MSn analysis can be carried out in the same physical space by means of ITMS. This approach involves using combinations of direct and rf-field pulse sequences on trapped ions in a helium reagent gas atmosphere.

Besides being simple, ITMS offers significant advantages in terms of sensitivity over a triple quadrupole and it may play an important role in flavonoid analysis.

To exemplify, ITMS is capable of isolating the ions m/z 271, 301, 353, 447, and 609 from the negative ESI-MS spectrum of naringenin, quercetin, chlorogenic acid, quercitrin, and rutin and fragmenting them to produce an ‘‘ion map,’’ which shows both isolated ions (parent m/z axis) and their fragments (product m/z axis). In practice, rutin and quercitrin are identified from their molecular ions (m/z 609 and 447) and from the same fragment (the aglycone quercetin, m/z 300).

Similarly, the identity of chlorogenic acid is given by the molecular ion (m/z 354) and the fragment m/z 191, which represents quinic acid.

The techniques described allow three different analytical approaches: infusion, direct injection, and injection after a separation step. The infusion is the simplest method of sample introduction (by means of a syringe) into the mass spectrometer. High sample volumes (50–150 AL) at flow rates (3–10 AL/ min) are required, and these conditions facilitate the structural investigation of the analytes subjected to a continuous infusion into the spectrometer.

In the second approach, the sample solution is injected by means of a HPLC injector directly into the mass spectrometer, without using any chromatographic column. Direct injection involves low sample volumes (1–10 AL) and has the advantage over the infusion method that no cleaning is needed after each analysis. In addition, the analysis times are very short (1–2 min), thereby permitting rapid screening of many samples. Furthermore, the direct injection approach allows minimization of the ion suppression effects due to the matrix by adding different concentrations of flavonoid standards to the herbal sample, and it may be considered for semiquantitative and rapid screening of herbal extracts.

The third approach involves coupling of a separative system (usually HPLC) with the mass spectrometer. This procedure simultaneously provides chromatographic, ultraviolet, and mass spectrometric data, and this range of information may be very helpful when assessing the identity of principles present in herbs. Further, HPLC coupled to MS (LC-MS) permits discrimination of compounds with the same molecular masses and exclusion of interference from the herbal matrix. Therefore, this approach remains the method of choice for quantitative analyses.

3. Sample Preparation

The herb is usually extracted with methanol or aqueous methanol at room temperature or at 40–50⁰C, depending on the stability of its components. The resulting crude extract may be purified to remove undesired constituents, such as lipids, chlorophyll, sugars, organic acids, and salts. In the case of commercial extracts, which are normally enriched in specific compounds, this step may be eliminated. Similarly, the purification may not be necessary in the case of LCMS, since the analytes of interest are separated from the interfering compounds during the chromatographic elution. By contrast, purification of the sample is recommended in the infusion and direct injection approaches. Indeed, the presence in the herbal matrix of different molecular species at concentrations ranging from 1 to 10 mM can cause ion suppression: i.e., the MS analyzer fails to detect the ions. In some circumstances the matrix effect may be reduced by diluting the sample and/or lowering the flow rate. These expedients appear to be successfull when highly sensitive and salt-compatible MS instruments are used.

4. Alkali Adducts

In positive ESI-MS, some molecular species can form adducts with alkali cations (sodium and potassium). In particular, potassium adducts are typical of raw herbal samples, because vegetable matrices are rich in potassium salts.

Alkali adduct formation may be diminished by desalting the samples through solid phase extraction (SPE). Diluting the sample solutions is a simple way to replace potassium ions with sodium ions. The latter are the most common in commercial extracts of herbs.

Not all flavonoids are capable of yielding alkali adducts. Thus, flavonol-3-O-glycosides generate sodium or potassium adducts. By contrast, these adducts are not obtained from flavonol-4V-O-glycosides and flavone glycosides.

So the spectrum of rutin (quercetin-3-O-rutinoside, MW 610 da) is characterized by the presence of the sodium adduct (m/z 633); the molecular ion ([MH]+, m/z 611) is almost absent. Also, the abundance of the aglycone residue (m/z 302) is low, indicating that the removal of the glucose residue is hindered. The same behavior can be observed for other flavonol-3-O-glycosides, as described for

Ginkgo biloba and St. John’s wort extracts (see Secs. II.A and II.B).

Conversely, in the case of spiraeoside (quercetin-4V-O-glucoside) no adduct is formed. Its spectrum mainly presents the molecular ion ([MH]+ = 465) and a relevant fragmentation occurs. Likewise, the flavone-glycoside (lacking the 3-OH group) rhoifolin (apigenin-7-O-neohesperidoside, MW 578) produces the m/z values 579 ([MH]+) and 270, corresponding to molecular and aglycone residue ions, respectively. This finding may suggest that the presence of the hydroxyl group at position 3 may be important for adduct formation.

However, this hypothesis is not appropriate, since flavonol aglycones, such as quercetin, kaempferol, and isorhamnetin, do not produce sodium adducts. What seems crucial is the 3-O-glycosylation, which may favor the formation of a crown-embedded stable cation.

Isoflavones form molecular ion adducts. Likely, the isoflavone ring at position 3 plays the same role as the sugar moiety in 3-O-glycosyl flavonols. In fact, the ESI-MS spectrum of the isoflavone biochanin A (MW 284 da) shows as main ion the sodium adduct ([M+Na]+, 307 m/z).

Alkali adducts are also formed with other flavonoid classes (see Secs. II.D. and II.G).