Anal. Chem. 2007, 79, 8323-8326
Monitoring the Interactions of Tocopherol Homologues with Reversed-Phase Stationary HPLC Phases by 1H Suspended-State Saturation Transfer Difference High-Resolution/Magic Angle Spinning NMR Spectroscopy Siri Schauff, Volker Friebolin, Marc David Grynbaum, Christoph Meyer, and Klaus Albert*
Institut fu¨r Organische Chemie, University of Tu¨bingen, Auf der Morgenstelle 18, 72076 Tu¨bingen, Germany
The separation process in reversed-phase high-performance liquid chromatography employing C18 phases is mainly due to hydrophobic interactions. The separation of tocopherol isomers, exhibited by the C30 phases, however, is additionally driven by shape selectivity. This phenomenon is investigated by suspended-state nuclear magnetic resonance spectroscopy using the saturation transfer difference technique, which was originally introduced to study protein-ligand interactions. The interaction strength between β-/γ-tocopherol and three different stationary phases was estimated qualitatively. The nuclear magnetic resonance data are compared to chromatographic data, and a similar mode of interaction between the analytes and the stationary phases is elucidated. Vitamin E is an important antioxidant and radical scavenger, which showed anticarcinogenic effects in several studies,1-3 and is found in most vegetable oils and nuts. The tocopherols constitute a major part in this vitamin and consist of four homologues, namely, R-, β-, γ-, and δ-tocopherol, showing different bioavailabilities. Nowadays, chromatographic separations in high-performance liquid chromatography (HPLC) are performed on a large variety of silica-based reversed-phase stationary phases. Alkyl chain modified silica phases, such as C8 and C18, are most commonly used to cope with usual separation problems. The majority of separations are based on hydrophobic interactions,4,5 and thus, analytes with minor structural differences, e.g., β- and γ-tocopherol, cannot be separated on these reversed-phase stationary phases. Both homologues contain the same amount of methyl groups at the aromatic ring but have different configurations. * To whom correspondence should be adressed: E-mail: klaus.albert@ uni-tuebingen.de. Tel: +49 70712975335. (1) Biesalski, H. K.; Schrezenmeier, J.; Weber, P.; Weiβ, H. E. Georg Thieme: Stuttgart, 1997. (2) Pryor, W. A.; Bowman, B. A.; Russel, R. M. Present Knowledge in Nutrition; International Life Sciences Institute Press: Washington, DC, 2001. (3) Kristensen, D.; Hansen, E.; Arndal, A.; Trinderup, R. A.; Skibsted, L. H. Int. Dairy J. 2001, 11, 837. (4) Meyer, V. Praxis der Hochleistungs-Flu ¨ ssigchromatographie. 8. Auflage; Salle+Sauerla¨nder: Frankfurt,1999. (5) Unger, K. K.; Weber, E. A. A Guide to Practical HPLC; GIT-Verlag: Darmstadt, 1999. 10.1021/ac071069t CCC: $37.00 Published on Web 10/05/2007
© 2007 American Chemical Society
However, preceding studies unveiled the suitability of polymeric C30 phases to separate all four tocopherol homologues.6 This effect was referred to as “shape selectivity” and is associated with the formation of domains with different alkyl chain mobilities on the surface of the material.7,8 The saturation transfer difference (STD) technique, a special suspended-state 1H high-resolution/magic angle spinning nuclear magnetic resonance (HR/MAS NMR) experiment, was developed by Mayer et al. In Mayer's study, it is generally used to investigate receptor-ligand interactions taking place at the active sites of proteins. In this case, various types of analyte molecules are present in the mixture, from which just one is selectively interacting with the protein.9,10 In our study, it is shown that suspended-state 1H STD HR/ MAS NMR spectroscopy can be used to study the retention mechanism, interaction strength, and sites of the two tocopherol isomers (β and γ) upon separation on three reversed phases. The results from NMR spectroscopy are compared to the chromatographic HPLC data. Saturation Transfer Difference NMR. Macromolecules (e.g., the stationary phases) are selectively saturated by irradiation with a train of Gaussian-shaped pulses (Figure 1), and a complete saturation of the polymer is obtained due to efficient spin diffusion.11 Then the saturation is transferred to analyte molecules, closely interacting with the polymer, which dissociate in a time scale, where the degree of saturation depends on the interaction time. The dissociated analyte memorizes the saturation, and thus, the 1H NMR peak intensity of molecules that were in contact with the polymer is reduced. However, nonsaturated analytes will most likely exceed the saturated one, leading to differentiation problems (6) Krucker, M.; Lienau, A.; Putzbach, K.; Grynbaum, M. D.; Schuler, P.; Albert, K. Anal. Chem. 2004, 76, 2623-2628. (7) Strohschein, S.; Pursch, M.; Lubda, D.; Albert, K. Anal. Chem. 1998, 70, 13-18. (8) Dachtler, M.; Glaser, T.; Ha¨ndel, H.; Lacker, T.; Tseng, L. H.; Albert, K. In Encyclopedia of Separation Science II; Wilson, I. D., Adlard, E. R., Cooke, M., Poole, C. F., Eds.; Academic Press: London; 2000. (9) Mayer, M.; Meyer, B. Angew. Chem.1999, 111, 1902-1906; Angew. Chem., Int. Ed. 1999, 38, 1784-1788. (10) Meyer, B.; Peters, T. Angew. Chem. 2003, 115, 890-918; Angew. Chem., Int. Ed. 2003, 42, 864-890. (11) Raitza, M.; Wegmann, J.; Bachmann, S.; Albert, K. Angew. Chem. 2000, 112, 3629-3632.
Analytical Chemistry, Vol. 79, No. 21, November 1, 2007 8323
Figure 1. Pulse sequence of the STD NMR technique.
in the spectra. In order to obtain spectra only showing the saturated analyte molecules, a difference experiment is performed: The spectrum of a nonsaturated system is substracted from a saturated system by an appropriate phase cycle after each scan. This can be achieved by the acquisition of one spectrum with on-resonant and one spectrum with off-resonant irradiation.9,10 EXPERIMENTAL SECTION NMR. The 1H suspended-state STD HR/MAS NMR experiments were performed in 4-mm HR/MAS double-bearing rotors, made of ZrO2 on a Bruker ARX 400-MHz spectrometer, at a spinning frequency of 4 kHz. The spectrometer was equipped with a 2H lock setup, which was set to the resonance frequency of methanol-d4. The 1H pulse length was 9.5 µs, and a total of 4k transients were recorded at room temperature. For each STD experiment, 5.5 mg of the corresponding stationary phase was combined with 68 µL of a 0.1 M solution of the corresponding tocopherol in 100% methanol-d4. For means of reference, the experiment was also carried out on the pure tocopherol solutions in absence of stationary phase. Materials. The tocopherol isomers were obtained from CalBiochem, USA. Methanole-d4 with a purity of 99.8% was obtained from Cambridge Isotope Laboratories. The C30 and C18 stationary phases were kindly provided by Bischoff Chromatography, Leonberg. The polymeric C30 phase and the monomeric C18 phase had a particle size of 3 µm with a pore size of 120 Å. The synthesis of the polyethylene-co-acrylic acid phase (PEAA) is published elsewhere.12 HPLC. The tocopherols were separated on a C30 column with an isocratic mobile-phase mixture of 96% methanol and 4% water at a flow rate of 1 mL/min.6 They were also separated on a PEAA capillary column, using an isocratic mixture of 90% methanol-d4 and 10% D2O at a flow rate of 5 µL/min at ambient temperatures.13 RESULTS AND DISCUSSION The interaction behavior of β- and γ-tocopherol toward a monomeric C18, a polymeric C30, and a PEAA phase12 are compared using 1H suspended-state STD HR/MAS NMR spectroscopy. The structures of the homologues are given in Figure 2, the structures of the stationary phases in Figure 3. For comparison, HPLC separations show the different separation efficencies of the reversed phases. The chromatograms on C18 and C30 were obtained from analytical HPLC,6 whereas the chromatogram on PEAA was obtained from capillary-HPLC. However, the retention order is still comparable. Identical retention order (δ-γ-β-R) is observed for all phases (Figure 4). Separation of β- and γ-tocopherol is achieved on the (12) Meyer, C.; Skogsberg, U.; Welsch, N.; Albert, K. Anal. Bioanal. Chem. 2005, 382, 679-690. (13) Grynbaum, M. D.; Meyer, C.; Putzbach, K.; Rehbein, J.; Albert, K. J. Chromatogr., A 2007, 1156, 80-86.
8324 Analytical Chemistry, Vol. 79, No. 21, November 1, 2007
Figure 2. Structures of (i) β- and (ii) γ-tocopherol with a ) H and c, c′ ) CH3.
Figure 3. Structures of (a) C30 phase, (b) PEAA phase, and (c) C18 phase.
Figure 4. HPLC chromatograms of the tocopherols on (a) C30 phase, (b) PEAA phase, and (c) C18 phase.
C30 phase, whereas only a slight separation is achieved on the PEAA phase. The homologues cannot be separated on the C18 phase at all. The structural differences of the tocopherol homologues are based upon the amount and sterical order of the methyl groups
1H
Figure 5.
1H
NMR spectra of (a) β- and (b) γ-tocopherol.
Figure 6. Bar diagram reflecting the measured peak integrals of peaks a and c from 1H suspended-state STD HR/MAS NMR spectra. The solid bars represent the intensity of signal a, the hatched bars peak c, c′.
at the aromatic ring. A difference in hydrophobicity of R- and δ-tocopherol is present according to the higher number of aromatic methyl groups of R-tocopherol. Therefore, R-tocopherol reveals stronger interactions to the stationary phase, leading to its longer retention time in HPLC. Due to the structural similarity at the aromatic ring of β- and γ-tocopherol, they are assumed not to be separated at all. However, since the sterical order is not alike, they are most probable to be separated due to shape selectivity.
HR/MAS NMR spectra of β- and γ-tocopherol solutions in methanol-d4 indicate the structural difference of both homologues. They differ in the chemical shift of their alicyclic CH2 group (b), which is given in Figure 5. Opposite shift behavior is observed for the aromatic proton signals (a) of γ-tocopherol and β-tocopherol, respectively. 1H suspended-state STD HR/MAS NMR spectra of β- and γ-tocopherol solutions were recorded in the presence of each stationary phase, respectively. The peak intensities of signals a and c were integrated, in reference to signal b. Since both homologues only differ in the aromatic region, the alkyl chain signal intensities can be neglected. For convenience, the results are graphically shown in a bar diagram in Figure 6. For comparison and as a proof of principle, the fist two bars represent the signal intensities of pure tocopherol solutions, which are equal. However, addition of the C30 phase causes higher intensities for signals a and c in the case of β-tocopherol in comparison to γ-tocopherol. A similar but smaller effect can be observed for the PEAA phase. Predictably, the tocopherols show similar peak integrals in the presence of the C18 phase. Deviations from the expected intensity behavior can be seen in the case of peak c, which is most probably due to stronger hydrophobic interactions of the methyl groups toward the stationary phases, exceeding the “minor” shape-selective interactions. The obtained 1H STD HR/MAS NMR spectra reveal the most intense peaks for parts of analyte molecules that show the strongest interaction with the stationary phase. The aromatic region of the tocopherol homologues interacts with both, the C30 and PEAA phase. The strongest effect and interaction difference is observed for the C30 phase and to a somewhat smaller extent for the PEAA phase. No interaction difference for the tocopherol homologues can be seen in the presence of the C18 phase, though an overall signal increase is present, which represents an equal interaction of the tocopherols toward the stationary phase. However, this interaction seems to be much weaker in the presence of the C18 phase than of both C30 and PEAA phases. These results match the information obtained from HPLC. In order to explain the retention behavior of both tocopherol homologues on a C30 phase, the structural difference has to be drawn: the methyl groups of β-tocopherol are arranged on opposite sides of the aromatic ring, whereas they are located on one side of the aromatic ring for γ-tocopherol. On behalf of this,
Figure 7. Schematic picture of (a) γ-tocopherol and (b) β-tocopherol penetrating between the C30 alkyl chains.
Analytical Chemistry, Vol. 79, No. 21, November 1, 2007
8325
we suggest that β-tocopherol is able to penetrate deeply into the C30 chains, whereas γ-tocopherol might penetrate in a more askew way, thus showing weaker interaction and therefore eluting earlier (Figure 7). CONCLUSIONS It was shown that 1H suspended-state STD HR/MAS NMR spectroscopy is suitable to reveal information on the interaction strength and sites between analyte molecules and chromatographic sorbents as a model of the chromatographic process. In case of a chromatographic separation, a rather dynamic equilibrium takes place between the analyte molecules and the stationary phase, whereas a more static equilibrium is present in an NMR rotor. However, since both processes rely on an exchange process, the STD technique is capable to monitor the interaction.
8326
Analytical Chemistry, Vol. 79, No. 21, November 1, 2007
In this study, the interaction strength of hydrophobic as well as shape-selective interactions could be evaluated in the system of β- and γ-tocopherol interacting with reversed-phase stationary phases. In particular, the shape selectivity of a polymeric C30 phase was monitored. ACKNOWLEDGMENT The authors acknowledge the helpful discussions with Berthold Maier and Paul Schuler on suspended-state NMR spectroscopy. Financial support was provided by the DFG, Graduiertenkolleg “Chemie in Interphasen”. Received for review May 23, 2007. Accepted August 23, 2007. AC071069T
