Chiral speciation of Marfey's derivatized DL-selenomethionine using capillary electrophoresis with UV and ICP-MS detection

Abstract

The chiral speciation of DL-selenomethionine by capillary electrophoresis (CE) with UV absorbance and inductively coupled plasma mass spectrometry (ICP-MS) is described. Chiral derivatization of DL-selenomethionine with 1-fluoro-2,4-dinitrophenyl-5-L-alanine amide (Marfey's reagent) forms diastereomers that are separated with CE in reversed polarity. Three CE buffer systems were evaluated using UV absorbance detection. A 30 mM ammonium phosphate buffer at pH 3.3 was chosen for good resolution of the D- and L-forms in under 14 min. Limits of detection for D- and L-selenomethionine were 250 ppb (ng mL⁻¹) using CE-UV, and 50 ppb using CE-ICP-MS. The migration time reproducibility of both detection limits was comparable (∼2%). This method was applied to the qualitative chiral identification of selenomethionine in selenized yeast digested with proteinase K (enzymatic hydrolysis).

---

Introduction

Many forms of selenium are of biological and toxicological importance, including the selenate and selenite ions, elemental selenium and metal selenides. In addition, some organisms may replace the sulfur atom in amino acids with a selenium atom, making seleno amino acids. The chemical and biochemical pathways of such a selenium cycle are not fully understood.¹ In humans, selenium deficiency is associated with muscle disease and several types of cancer and also selenium is an essential trace element. However, at high concentrations, selenium is toxic, causing problems such as dermatitis, fatigue and hair loss.² There is a small gap between the beneficial and toxic doses of selenium. Human nutritional studies have shown that appropriate doses of supplemental selenium enhance cellular defense against oxidative damage and may prevent certain types of cancers.^3,4

Dietary selenium supplements often take the form of selenate or selenite, although better health effects might be obtained by ingesting methylated selenium species.⁴ Selenized yeast is commonly used for selenium nutritional supplements. Such yeasts are complex samples with many amino acids, including seleno amino acids, that contain an asymmetric α-carbon and thus DL-enantiomers. However, usually only L-forms of amino acids are incorporated into proteins in living organisms. In addition, D-forms or racemic mixtures of amino acids may have harmful effects on humans when consumed.⁵ For this reason, improved analytical techniques that can differentiate between D- and L-enantiomers have become an active area of research, especially in the pharmaceutical industry.

High-performance liquid chromatography (HPLC), coupled to ICP-MS detection, may be used for the speciation of DL-seleno amino acid enantiomers using chiral stationary phases, such as β-cyclodextrin,⁶ crown ethers⁷ and, most recently, antibiotics.⁸ Gas chromatography (GC) has also been studied for enantiomeric separation of Se-methionine with ICP-MS detection.⁹ However, chiral columns are expensive, and amino acids as well as small peptides may potentially co-elute, making peak identification difficult. As a result, chiral separations are increasingly performed by capillary electrophoresis (CE). The advantages of CE are well known, but the high separation efficiency, low cost and rapid analysis times are the most significant for this application. Reviews of the state-of-the-art methodology in chiral separations by CE have been published recently.^10,11

CE offers the ability to separate the components of complex mixtures in short analysis times. CE coupled to ICP-MS allows sensitive elemental speciation detection.^12,13 Previous studies have shown chiral separation of seleno amino acids with vancomycin as a chiral selector using CE with UV detection.^5,14,15 Also, a study with cyclodextrin-modified micellar electrokinetic chromatography (MEKC) of seleno amino acid derivatives using UV detection has been published.¹⁶ Seleno amino acids have been separated using CE and detected by ICP-MS.^17–19 However, there are no chiral speciation studies of seleno amino acids by CE-ICP-MS.

Chiral derivatization of amino acids is a common strategy for enantiomeric speciation. Current HPLC-ICP-MS studies²⁰ in our laboratory use L-Marfey's reagent (1-fluoro-2,4-dinitophenyl-5-L-alanine amide)²¹ to react with the amine group of seleno amino acids forming diastereomers. This reagent has been used for the analysis of enantiomeric forms of amino acids using CE-UV²² and for the separation of DL-dopa using MEKC.²³ This paper investigates the application of Marfey's derivatization of DL-selenomethionine using CE with UV and ICP-MS detection. Different buffer systems are evaluated, analytical figures of merit are presented, and application to the analysis of DL-selenomethionine in selenized yeast is shown.

Experimental

Instrumentation

The ICP-MS instrument used was an Agilent 7500s (Agilent Technologies, Tokyo, Japan) equipped with a grounded “ShieldTorch” system. ICP-MS operating conditions and data acquisition settings are shown in Table 1. The nebulizer gas and ICP-MS lenses were tuned by introducing a solution of Se (10 ppb) via the self-aspirating CE interface. The ion lenses and quadrupole settings were set for highest Se signals, but also for lowest background at ⁷⁸Se (23.6%) and ⁸²Se (9.2% abundant). Typical signals using the self-aspirating micronebulizer, with the interface in place, were 1.5 × 10⁴ counts s⁻¹ with 1 × 10³ counts s⁻¹ background (≈15 signal∶background ratio) for ⁸²Se. The ⁷⁸Se signal was ∼5 × 10⁴ counts s⁻¹ with 1.5 × 10⁴ counts s⁻¹ background (≈3 signal∶background). Both ⁸²Se and ⁷⁸Se were monitored during each CE run to verify selenium isotope patterns in electropherogram peaks.

Table 1 ICP-MS and data acquisition parameters and CE operating conditions for chiral separation of DL-selenomethionine

ICP parameters—
Plasma rf power	1300 W with Shield Torch system
Sampling depth	4.5 mm
Nebulizer	MicroMist AR30-1-F02
Nebulizer gas flow rate	1.10 L min^−1
Spray chamber	Scott-type double pass (glass), 2 °C
Isotopes monitored	^78Se, ^82Se (integration time/mass = 0.3 s)
CE conditions—
Buffer and pH	30 mM NH4PO3, adjusted with 30 mM H3PO4 to pH 3.3
Voltage, temperature	−25 kV (70 µA), 23 °C
Capillary	75 µm inner diameter, 365 µm outer diameter, 50 cm long
UV detection	214 nm

---

The CE instrument was a Waters Quanta 4000 capillary ion analysis system (Waters Corporation, Milford, MA, USA). The UV absorbance data were acquired using a laboratory-built analog/digital converter and stored as text files. Data processing was performed using software from Grams32 v5.1 with Chromatography Pack (Galactic Industries, Salem, NH, USA). CE-ICP-MS data and peak areas were calculated using “ICP-MS Chromatographic Software” (Agilent Technologies, Tokyo, Japan).

CE conditions and interface

The CE-ICP-MS schematic of the interface is shown in Fig. 1. The interface consists of tubing, finger-tight fittings and a four-way tee union, all made of PEEK material purchased from Upchurch Scientific (Oak Harbor, WA, USA). A copper wire fitted with alligator clips provided the electrical connection from the grounded electrode in the CE instrument. All fittings in the CE-ICP-MS interface were airtight. CE separation conditions were optimized by performing indirect UV analysis using new capillaries, which were conditioned by purging with 0.1 M HCl for 5 min, 0.1 M NaOH for 5 min and fresh CE electrolyte for 5 min. When interfaced to ICP-MS, capillary conditioning was performed by hand using Luer lock syringes (Becton Dickinson, Franklin Lakes, NJ, USA) adapted to the capillary via a PEEK sleeve and fittings. When interfaced to the ICP-MS, the heights of both ends of the CE capillary were equal to prevent siphoning effects in the capillary (induced laminar flow). The make-up electrolyte in the CE-ICP-MS studies matched the CE separation buffer. The CE final conditions are summarized in Table 1.

Fig. 1 CE-ICP-MS interface schematic (not to scale).

Chemicals and standards

Selenomethionine (98% purity), and selenocystine (98%) were obtained from Aldrich (Milwaukee, WI, USA). Sodium bicarbonate, ammonium phosphate, phosphoric acid, hydrochloric acid, acetic acid, sodium hydroxide and acetone (all analytical-reagent grade) were from Fisher (Fairlawn, NJ, USA). Tetraethylammonium tetrafluoroborate (TEA/TFB) was from Fluka (Ronkondoma, NY, USA). Sodium selenate, sodium selenite, proteinase K and L-Marfey's reagent (1-fluoro-2,4-dinitophenyl-5-L-alanine amide) were purchased from Sigma (St. Louis, MO, USA). SelenoExcell^™ selenized yeast (1.25 mg g⁻¹ Se dry weight) was obtained from Cypress Systems, Inc. (Fresno, CA, USA). Milli-Q+ deionized water (18 MΩ cm) was used for all solution preparation and dilutions.

Sample preparation and derivatization

A 1000 µg mL⁻¹ (ppm) solution of DL-selenomethionine was prepared by dissolution in 0.3 M NaHCO₃. For yeast samples, 0.2 g of selenized yeast was hydrolyzed using 5 mL of water mixed with 0.02 g of Proteinase K in a polypropylene container. The samples were shaken at 250 rev min⁻¹ in the dark for 20 h at 30 °C and centrifuged at 3000 g for 30 min. The supernatant solution was removed and passed through a 0.25 µm filter. For Marfey's derivatization, 200 µL of a 1% w/v solution of Marfey's reagent dissolved in acetone were added to 150 µL of selenomethionine standard, or a yeast sample as prepared above, in a clean 3 mL glass vial, and the solution was reacted for 2 h at 40 °C. After cooling, the mixture was neutralized by adding ∼20 µL of 2 M HCl. The final solution was diluted in water. All solutions were stored in darkness.

Results and discussion

Selection of CE run buffer

Three CE buffer systems were evaluated for separation of DL-selenomethionine by CE-UV. The goal was to achieve adequate resolution of the DL-selenomethionine for application to complex mixtures that could contain many amino acids and small peptides such as selenized yeast. An acetic acid–sodium acetate buffer (pH 5.3) with an applied voltage of +15 kV [Fig. 2(A)] gave baseline resolution of both diastereomers within 6 min. In this buffer system, normal CE polarity was used to produce the electroosmotic force (EOF) toward the detector, so that the analytes would migrate against the EOF. However, the magnitude of EOF is greater than the electrophoretic mobility of the analytes and carries them towards the detector. The separation is then based on the combination of the EOF effect and the inherent mobility of the analytes. Therefore, larger and higher mass analytes move more slowly toward the detector than small, lower mass ions. The major disadvantage of using this buffered system is the poor reproducibility and resolution achievable, therefore other electrolyte solutions were investigated.

Fig. 2 Three buffer systems investigated for chiral separation of Marfey's derivatized DL-selenomethionine: (A), acetic acid; (B), TEA/TFB; and (C), phosphate buffer. See Table 2 for details. 1, Excess Marfey's reagent; 2, L-selenomethionine; 3, D-selenomethionine; and 4, impurity in DL-selenomethionine.

A second approach involved the use of tetraalkylammonium salts such as tetraethylammonium tetrafluoroborate (TEA/TFB) and ammonium phosphate. TEA/TFB is stable, non-toxic, and provides favorable electrolytic conductivity.²⁴ Using this buffer, a reverse EOF condition is created where the negatively charged silanol groups on the capillary wall are covered with ammonium ions. The negative buffer ions, as well as the negatively charged analytes, migrate toward the anode (at the detection end of the capillary) and the separation is based on the electrophoretic mobility (size and charge) of the individual analytes.^25,26 An example of the TEA/TFB separation is shown in Fig. 2(B). The migration times of the diastereomers are higher and baseline resolution is achieved. A 30 mM ammonium phosphate buffer (pH 3.3) with an applied voltage of −25 kV gave the best separation of the DL-enantiomers [Fig. 2(C)]. These conditions were chosen for further studies including the application to selenized yeast digested by hydrolysis with the enzyme proteinase K.

Optimization of pH

Previous studies have used a 50 mM phosphate buffer at pH 3.3 for the analysis of sulfur-containing amino acids derivatized by Marfey's reagent,²² but the effect of pH was re-evaluated by us for the selenomethionine diastereomers. The data were acquired using UV absorbance detection at 214 nm. At pH 3.0 the migration times are much slower for each of the diastereomers, because the EOF is slower and increases the resolution and migration time. At pH 3.3, the EOF is higher in the direction of the detector and the analyte ions migrate faster. At higher pH values (3.9 and 4.2), the resolution decreases due to the higher EOF, and the resolution and migration times are smaller. This information was useful for choosing the conditions for customizing the resolution for analysis of selenized yeast. Also, a 30 mM phosphate buffer (pH 3.3) was investigated and gave good separation of the enantiomers in much faster migration times; this was used for rapid detection of DL-selenomethionine using CE-ICP-MS. The lower concentration of phosphate in the buffer is beneficial for ICP-MS detection because of lower background signals for mass 82.

CE-ICP-MS of DL-selenomethionine

An equivalent length of CE capillary was used to directly compare CE migration times observed using UV and ICP-MS detection.^27,28 Also, the make-up electrolyte level, or height, was optimized by the following procedure. If CE-ICP-MS migration times were shorter than those found with UV detection, then the make-up electrolyte height was raised, which applied hydrostatic pressure to the end of the capillary to neutralize nebulizer suction. The optimum make-up electrolyte height for all experiments was 1.0 cm below the end of the CE capillary.

The capillary position inside the nebulizer was adjusted to produce the narrowest peak shapes. The position was adjusted by loosening the fitting/sleeve that seals the capillary into the interface. This releases the seal around the CE capillary and allows manipulation of the position in the back of the nebulizer. The capillary position may be observed through the glass walls of the nebulizer. The optimum position was found with the end of the capillary positioned ∼1.0 mm from the inner nebulizer capillary opening. Inserting the capillary further into the nebulizer must be avoided to prevent isolation of the nebulizer suction to the CE capillary. This condition excludes the sheath or make-up gas flow. Also, when the end of the CE capillary contacts the nebulizer, small chips of CE capillary or its polyimide coating might break off, clogging the nebulizer. To help prevent this, all capillaries used were ground to a tapered tip (preventing small chips) to provide an optimum fit in the back of the nebulizer.

Fig. 3(A) shows a CE-ICP-MS electropherogram corresponding to the derivatization of a DL-selenomethionine standard. The concentration of selenium is 1 µg ml⁻¹ per enantiomer, but the derivatization yields are about 65%.²¹ The complete separation of the derivatized diastereomers is shown in under 14 min. This is the first reported chiral speciation of selenium-containing amino acids by CE-ICP-MS of which we are aware. Another advantage of ICP-MS detection is that the inorganic forms of selenium (Se⁶⁺ and Se⁴⁺) may also be detected using the same conditions as shown in Fig. 3(B).

Fig. 3 CE-ICP-MS electropherograms of standards: (A) Marfey's derivatized DL-selenomethionine and (B) 1 ppm mixture of selenate and selenite. Conditions: 30 mM (NH₄)₂HPO₄, pH 3.3 and −25 kV.

The analytical figures of merit for the analysis of a Marfey's derivatized DL-selenomethionine standard are shown in Table 2 for both methods of detection. Limits of detection for D- and L-selenomethionine were 240 and 260 ppb (ng mL⁻¹), respectively, using CE-UV, and 48 and 52 ppb as selenium using CE-ICP-MS. A direct comparison of LOD between the two methods can be made by measuring amino acid concentration in ppb. Using CE-ICP-MS, the limits of detection are 130 ppb for L-selenomethionine and 120 ppb for D-selenomethionine, an improvement by a factor of two compared to the UV detection data. The migration time reproducibility for both detection methods was comparable (∼2%). The linearity of the calibration plots (R² ≥ 0.99) was acceptable at the low concentrations analyzed in the range from 0.25 to 10 ppm. These results show the good sensitivity of the CE-ICP-MS method, compared to CE-UV detection, with similar reproducibility and linearity, although in this case the difference between the two techniques is not as striking due to the excellent UV chromophore. However, ICP-MS detection also allows mass selectivity and isotope information.

Table 2 Figures of merit for CE-UV and CE-ICP-MS analysis of Marfey's derivatized DL-selenomethionine (T_m = CE migration time)

	L-Selenomethionine	D-Selenomethionine
UV absorbance detection (214 nm)—
T m (%RSD) (n = 6)	1.8	1.9
LOD (ppb)	240	260
R ^2	0.992	0.994
ICP-MS detection—
T m (%RSD) (n = 4)	2.0	2.1
LOD (ppb as Se) {as amino acid}	48 {130}	52 {120}
R ^2	0.990	0.993

---

Analysis of selenized yeast by CE-UV and CE-ICP-MS

Selenized yeast was hydrolyzed using proteinase K and evaluated by CE-UV (see Experimental for sample preparation). Fig. 4(A) shows the unspiked sample and Fig. 4(B) shows the same sample spiked with 500 ppb of DL-selenomethionine. Several peaks are present, which should be Marfey's derivatized amino acids, but no information is available for selenium content from UV detection. For yeast samples, only L-forms of amino acids are expected, though a small degree of racemization could occur during the sample preparation and derivatization steps. However this was not observed in the analysis of pure L-selenomethionine (results not shown) suggesting that, if racemization occurs, it cannot be detected by these CE-UV or CE-ICP-MS methods. Previous studies using HPLC-UV²⁰ have demonstrated that these compounds are Marfey's derivatized amino acids. The sample spiked with a racemic mixture of derivatized selenomethionine shown in Fig. 4(B) shows the same peaks, but also a larger one at 20 min and a new reproducible peak at 24 min that is D-selenomethionine. Note that the migration times are longer than those shown in Fig. 2 because the concentration of the phosphate running buffer was increased to 50 mM instead of 30 mM, which has the effect of increasing migration times of the analytes and thus the resolution of this complex mixture. A 30 mM phosphate buffer was used in CE-ICP-MS experiments because the selectivity of ICP-MS detection gave simple electropherograms compared to the UV absorbance data, and the increased CE resolution provided by the 50 mM phosphate buffer was not required. Also, the lower concentration buffer gave lower ICP-MS background signals for the ⁸²Se isotope compared to the 50 mM buffer, and allowed lower detection limits to be achieved.

Fig. 4 CE-UV electropherograms of selenized yeast digested with Proteinase K: (A) unspiked and (B) spiked with 500 ppb DL-selenomethionine standard.

Fig. 5 shows the same selenized yeast sample analyzed by CE-ICP-MS. The broken line corresponds to the sample and the solid line corresponds to the sample diluted 1∶2 in H₂O and spiked with 500 ppb of DL-selenomethionine standard. A small peak corresponding to the migration time of L-selenomethionine is observed in the unspiked sample, which agrees with UV results. Identification of the L-enantiomer was accomplished by derivatizing a pure L-selenomethionine standard under the same conditions (not shown). Also, two larger peaks eluted after the DL-selenomethionine peak that did not appear in the UV electropherogram. These peaks are unidentified, but they contain selenium as confirmed by the isotopic ratios for ⁷⁸Se and ⁸²Se. However, they are not Marfey's-derivatized, as they do not appear in the UV electropherograms. A study by Tran et al.²² has shown that dipeptides and tripeptides elute after the amino acids, and it is probable that these peptides are present in the selenized yeast sample, but could not be confirmed by a qualitative technique or by the analysis of peptide standards, as such standards are not available. It would be possible to use this technique to identify such peptides for future studies, but qualitative studies are better suited for such purposes. Therefore, studies are being performed on electrospray MS for identification of these species.

Fig. 5 Overlaid CE-ICP-MS electropherograms of selenized yeast sample, as shown in Fig. 4: 1, selenite; 2, L-selenomethionine; and 3, D-selenomethionine. Conditions: 30 mM (NH₄)₂HPO₄, pH 3.3 and −25 kV.

Conclusions

The use of chiral derivatization by Marfey's reagent is effective for the separation of DL-selenomethionine by CE. A 30 mM ammonium phosphate buffer at pH 3.3 (−25 kV) allows the baseline resolution of these compounds within a 14 min electropherogram (half of the time required for the HPLC analysis). The same conditions also allow the separation of the inorganic forms of selenium in less than 14 min. The limits of detection were close to 50 ppb and reproducibility between injections was about 2% for CE-ICP-MS. Spiking the selenized yeast samples with a derivatized DL-selenomethionine standard confirmed the presence of L-selenomethionine. Other selenium species that were observed using CE-ICP-MS, but not observed with CE-UV, are possibly small peptides containing selenium.

Acknowledgements

The authors kindly acknowledge Dr. Clayton B'Hymer for bringing the Marfey's derivatization technique to their attention, Agilent Technologies for instrument cooperation, and NIEHS grant no. ES04908 for funding J. Day during doctoral studies.

References

1. C. Ip and H. E. Ganther, in Selenium in Biology and Human Health, ed. R. F. Burk, Springer-Verlag, New York, 1994, pp. 171–180.
2. A. Pineau, in Selenium in Medicine and Biology, ed., J. Nève and A. Favier, Walter de Gruyter, New York, 1988, pp. 345–249.
3. L. C. Clark, G. F. Combs, B. W. Turnbull, E. H. Slate, D. K. Chalker, J. Chow, L. S. Davis, R. A. Glover, G. F. Graham, E. G. Gross, A. Krongrad, J. L. Lesher, H. K. Park, B. B. Sanders, C. L. Smith and J. R. Taylor, J. Am. Med. Assoc., 1996, 276, 1957.
4. C. Redman, J. A. Scott, A. T. Baines, J. L. Basye, L. C. Clark, C. Calley, D. Roe, C. M. Payne and M. A. Nelson, Cancer Lett., 1998, 125, 103.
5. K. L. Sutton, R. M. C. Sutton, A. M. Stalcup and J. A. Caruso, Analyst, 2000, 125, 231.
6. S. P. Mendez, E. B. Gonzalez, M. L. F. Sanchez and A. Sanz Medel, J. Anal. At. Spectrom., 1998, 13, 893.
7. K. L. Sutton, C. A. P. de Leon, K. L. Ackley, R. M. C. Sutton, A. M. Stalcup and J. A. Caruso, Analyst, 2000, 125, 281.
8. S. P. Mendez, E. B. Gonzalez and A. Sanz Medel, J. Anal. At. Spectrom., 2000, 15, 1109.
9. S. P. Mendez, M. M. Bayon, E. B. Gonzalez and A. Sanz Medel, J. Anal. At. Spectrom., 1999, 14, 1333.
10. R. Vespalec and P. Bocek, Chem. Rev., 2000, 100, 3715.
11. H. J. Issaq and K. C. Chan, Electrophoresis, 1995, 16, 467.
12. A. R. Timerbaev, Talanta, 2000, 52, 573.
13. J. W. Olesik, in Elemental Speciation: New Approaches for Trace Element Analysis, ed. J. A. Caruso, K. L. Sutton and K. L. Ackley, Elsevier, New York, 2000, pp. 151–211.
14. R. Vespalec, H. Corstjens, H. A. H. Billiet, J. Frank and K. Luyben, Anal. Chem., 1995, 67, 3223.
15. R. Vespalec, H. A. H. Billiet, J. Frank and K. Luyben, J. High Res. Chromatogr., 1996, 19, 137.
16. A. P. Mendez, E. B. Gonzalez and A. Sanz Medel, Anal. Chim. Acta, 2000, 416, 1.
17. B. Michalke, J. Chromatogr. A, 1995, 716, 323.
18. B. Michalke and P. Schramel, Biol. Trace Elem. Res., 1997, 59, 45.
19. L. Bendahl, B. Gammelgaard, O. Jons, O. Farver and S. H. Hansen, J. Anal. At. Spectrom., 2001, 16, 38.
20. M. M. Bayon, C. B. B'Hymer, C. Ponce de Leon and J. A. Caruso, J. Anal. At. Spectrom., 2001, 16, 945.
21. P. Marfey, Carlsberg Res. Commun., 1984, 49, 591.
22. A. D. Tran, T. Blanc and E. J. Leopold, J. Chromatogr., 1990, 516, 241.
23. G. Wu and M. Furlanut, Pharmacol. Res., 1997, 35, 553.
24. E. G. Yanes, S. R. Gratz and A. M. Stalcup, Analyst, 2000, 125, 1919.
25. M. P. Harrold, M. J. Wojtusik, J. Riviello and P. Henson, J. Chromatogr., 1993, 640, 463.
26. C. Quang and M. Khaledi, Anal. Chem., 1993, 65, 3354.
27. J. A. Day, K. L. Sutton, R. S. Soman and J. A. Caruso, Analyst, 2000, 125, 819.
28. J. A. Day, J. A. Caruso, J. S. Becker and H. J. Dietze, J. Anal. At. Spectrom., 2000, 15, 1343.

---