Ribonucleoside labeling with Os(VI): A methodological approach to evaluation of RNA methylation by HPLC-ICP-MS

Abstract

Covalent modifications of nucleobases are thought to play an important role in regulating the functions of DNA and various cellular RNA types. Perhaps the best characterized is DNA methylation on cytosine (methyl tag attached to carbon 5 position) and such modification has also been detected in stable and long-lived RNA molecules. In this work, we propose a novel procedure enabling very sensitive quantification of methylcytidine and other ribonucleosides, based on reversed phase liquid chromatography with inductively coupled plasma mass spectrometry (ICP-MS) detection. The procedure relies on labeling ribose residues with osmium, by formation of a ternary complex between cis-diol ribose groups, hexavalent osmium (K₂OsO₂(OH)₄) and tetramethylethylenediamine (TEMED). The derivatization reaction was carried out with 50 : 1 molar excess of Os to ribonucleoside, pH 4, for 2 h at room temperature. The structures of Os-labeled cytidine and methylcytidine were confirmed by electrospray ionization mass spectrometry. The separation of Os-labeled cytidine (C), uridine (U), 5-methylcytidine (5mC) and guanosine (G) was achieved on C18 column (Gemini, 150 × 3 mm, 5 μm) with isocratic elution (0.05% triethylamine + 6 mmol L⁻¹ ammonium acetate, pH 4.4: methanol (85 : 15)) and a total flow rate 0.6 mL min⁻¹. The column effluent was on-line introduced to ICP-MS (a model 7500 ce, Agilent Technologies) for specific detection at ¹⁸⁹Os. Calibration was performed within the concentration range 0–200 nmol L⁻¹ of each ribonucleoside and the analytical figures of merit were evaluated. For 100 μL injection, the detection limits for C, U, 5mC, G were 24, 38, 21 and 28 pmol L⁻¹, respectively. While introducing Os(VI)-TEMED to the column, it eluted in the dead volume and the detection limit for osmium was 20 pmol L⁻¹. The results obtained in this work might be helpful in the analysis of RNA digests, providing quantitative data on the ribonucleoside composition and RNA methylation (measured as the percentage of methylated cytidines with respect to total RNA cytidines).

---

Introduction

Regulation of gene expression depends on the two components, namely genetic and epigenetic. Genetic component relies on the DNA sequence, which decides on the stability of transcriptional and messenger RNA. Otherwise, changes of DNA sequence (mutations) would be heritable somatically and through the germ line. The second, epigenetic component of gene expression consists of alterations in chromatin structure without changing DNA sequence.^1,2 Primary molecular mechanisms underlying epigenetics are covalent modifications of DNA and histones.^3,4 In particular, DNA methylation that occurs at carbon 5 position in cytosine pyrimidine ring, is perhaps one of the best characterized epigenetic events.⁵ The aberrant methylation patterns play a prominent role in carcinogenesis in a manner that gene-specific hypermethylation generally is involved in deactivating tumor suppressor genes and global DNA hypomethylation leads to activation of genes important for cancer development.^6–8 Therefore, the analysis of actual DNA methylation status is gaining strength in the fields of epigenetics, cancer risk assessment, diagnosis and therapy monitoring.^9,10

Within the context of this work it is important that dsRNA-derived species can direct changes in the actual methylation of DNA regions with which they share sequence identity.^11,12 The mechanism by which RNA signals are translated into DNA methylation imprints is currently unknown, however the participation of specific cytosine methyltransferases in this process has been implicated.^8,13 Furthermore, in different experimental systems, methylation of cytosine in non-coding small RNA molecules has been reported,^14,15 yet its functional characterization is still hindered, in part due to the lack of molecular methods for RNA methylation analysis.¹³

Regarding techniques used in molecular biology, sodium bisulfite treatment widely used for DNA analysis, has been recently applied for detection of methylated cytidines in RNA molecules.¹³ In this procedure, cytosine bases are converted to uracil, while methylated cytosines resist conversion and can be then identified by microarrays or deep sequencing procedures.

On the other hand, about 30 years ago, osmium tetroxide was introduced to the field of nucleic acid research, based on its ability for binding to nucleobases pyrimidine in the presence of tertiary nitrogen donor ligands.¹⁶ Later, it was demonstrated that, depending on the reaction conditions, osmium can be also attached to ribose groups in ribonucleosides.^17,18 In particular, while potassium osmate(VI) and 2,2′-bipyridyl were used, the tag was attached to the cis-diol group with formation of oxo-osmium(VI) bipyridyl sugar ester.¹⁸ The electroactivity of osmium labels, the feasibility of electroanalytical techniques and fluorimetric detection for DNA/RNA probing have been extensively studied.^19–26

The original idea of this work was to introduce osmium as external elemental tag for labeling ribose groups in 5-methylcytidine and in ribonucleosides in order to achieve their quantification at extremely low concentration levels by high performance liquid chromatography with inductively coupled plasma mass spectrometry. The analytical performance of the procedure developed indicates its potential utility in the evaluation of RNA methylation in sub-microgram samples. Furthermore, the results obtained confirm the utility of ICP-MS as attractive, complementary analytical tool in epigenetic studies.²⁷

Experimental

Reagents and solutions

All chemicals were of analytical reagent grade. Deionized water (18.2 MΩ cm, Labconco, USA) and HPLC-grade methanol (Fisher Scientific, Pittsburgh, USA) were used throughout.

The standard solutions of 5-methylcytidine, cytidine, guanosine and uridine (10 μmol L⁻¹) were prepared by dissolving respective Sigma reagents in deionized water. Working solutions were prepared daily by appropriate dilution. The stock solution of potassium osmate (15 mg L⁻¹) was obtained by dissolving Aldrich reagent in 1 mol L⁻¹ hydrochloric acid. N,N,N′,N′-tetramethylethylenediamine (TEMED) was from Sigma, the solution 5 mg L⁻¹ was prepared.

The following Sigma reagents were also used: hydrochloric acid, phosphoric acid, potassium hydroxide, triethylamine, ammonium acetate.

Apparatus

An Agilent series 1200 liquid chromatographic system equipped with a quaternary pump, a well plate autosampler, a column oven, a diode array detector, and a ChemStation (Agilent Technologies, Palo Alto, CA, USA) was used in the preliminary experiments. The chromatographic column was a Gemini C18 (150 × 3 mm, 5 μm) from Phenomenex. For specific detection on osmium, the column effluent was on-line introduced to inductively coupled plasma-mass spectrometry system via the short-length Teflon tubing.

A model 7500 ce ICP-MS (Agilent Technologies, Tokyo, Japan) was used with a MiraMist Teflon® nebulizer. A Peltier-cooled chamber was operated at 2 °C. Tuning procedure was performed daily using diluted Agilent solution (Li, Y, Tl, Ce, 1 μg L⁻¹ each). The chromatographic and ICP-MS instrumental operating conditions are given in Table 1.

Table 1 HPLC-ICP-MS instrument operating conditions

Reversed phase liquid chromatography separation
Column	Gemini C18 (150 × 3 mm, 5 μm)
Mobile phases	A: 6 mmol L^−1 ammonium acetate, pH 4.4, 0.05% triethylamine; B: methanol
Elution	Isocratic 85% A, 15% B
Temperature	Ambient
Flow	0.6 mL min^−1
Injection volume	100 μL
ICP-MS detection
Forward power	1500 W
Nebulizer gas flow	0.9 mL min^−1
Make-up gas	0.15 mL min^−1
Nebulizer	MiraMist Teflon®
Spray chamber	Peltier cooled chamber, 2 °C
Sample and skimmer cones	Nickel
Sample depth	10 mm
Quadrupole bias	−3 V
Octopole bias	−6 V
Channels monitored	^189Os
Acquisition mode	Time-resolved analysis
Dwell time	100 ms

---

A 100 μL syringe, running at 18 μL h⁻¹, was connected to an electrospray ionization/ion trap mass spectrometer (Agilent HPLC-Chip/Trap XCT Ultra, Agilent Technologies, Santa Clara, CA, USA) with the nanospray interface, operating in a positive ion mode with the following parameters: nanospray needle voltage: 2000 V; drying temperature: 325 °C; drying gas: 3.00 L min⁻¹; maximum accumulation time: 300 ms. The mass spectra were recorded across the range 100–1000 m/z in ultra scan mode. The instrument was externally calibrated using a calibration mixture (Agilent Technologies, Santa Clara, CA, USA) containing masses 118.09, 322.05, 622.03, 922.01, 1521.97, 2121.93 and 2721.89. All MSⁿ experiments were performed with an isolation width of 1.0 mass unit and the fragmentation amplitude was 0.6 (arbitrary unit). Ten parent ions were selected and isolated by the instrument, producing MS² spectra by collision induced dissociation with He gas.

Nucleoside labeling with osmium

For calibration, 100 μL of TEMED (5 mg L⁻¹) were mixed with 100 μl of K₂OsO₂(OH)₄ (15 mg L⁻¹), 100 μL of the nucleoside mixed standard solution (0–200 nmol L⁻¹ each) and 700 μL of phosphate-acetate buffer (10 mmol L⁻¹, pH 4.0) were added. The mixture was left for 2 h at room temperature, centrifuged (10 000 g 10 min) and 100 μL of the supernatant introduced to the column.

For ESI-MS analysis, the labeled cytidine and 5-methylcytidine were obtained as described above, but using 100 times higher concentrations of nucleosides, osmate and TEMED. The column fractions corresponding to elution of derivatized C and 5mC were collected.

Results and discussion

The original idea of this study was to label nucleosides ribose groups with osmium and to explore the feasibility of such an approach for very sensitive quantification of 5-methylcytidine and ribonucleosides by reversed phase liquid chromatography separation coupled to ICP-MS detection. It was assumed that osmium would be a good candidate for an external elemental tag, because of its relatively high ionization efficiency (about 80%), no isobaric interferences expected at isotope ¹⁸⁹Os and low detection limit (in the order of 0.1 ng L⁻¹). Furthermore, osmium is practically absent in biological samples and the reaction of Os(VI) with RNA has been studied extensively for the purposes of electrochemical analyses.^23,28 Chromatographic separation of labeled cytidine and 5-methylcytidine and their quantification based on ICP-MS detection at osmium would enable to us evaluate the percentage of methylated cytosines with respect to total cytosines, which in the application to the real-world micro samples could be useful in assessing RNA methylation.

As mentioned in the introduction, labeling reaction involves the formation of ternary complex between cis-diol ribose group, Os(VI) and tertiary nitrogen donor ligand. In particular, Os(VI) is attached to ribose through formation of diester structure, which is stabilized by a nitrogen donor ligand.¹⁶ In this work, tetramethylethylenediamine (TEMED) was selected as a stabilizing ligand, owing to its higher polarity and hence, lower requirement for organic modifier in reversed phase HPLC-ICP-MS hyphenation, as compared to 2,2′-bipyridine, commonly used in the previous studies.²² The reaction equation is presented schematically in Fig. 1. The reaction conditions were studied systematically, by varying pH, molar ratio Os(VI)-TEMED : ribonucleoside, temperature and reaction time. At this stage, UV detection at 270 nm was employed²⁹ instead of the much more expensive ICP-MS. Relatively low sensitivity of spectrophotometric detection was not a limiting factor during method development. The experiments were carried out using standard solutions of 5-methylcytidine, cytidine, guanosine and uridine (10 μmol L⁻¹ each). For each nucleoside, the labeling reaction was carried out at pH 2, 4, 7 and 9 (phosphate–acetate, or phosphate buffer) and the mixture was injected to HPLC-UV system. In Fig. 2a the values of peak height obtained for individual ribonucleoside are plotted against pH value. As can be observed, the highest chromatographic signals were obtained at pH 4 and this pH was used in further development. Similarly, the effect of molar excess of osmium with respect to ribonucleosides was studied. The results obtained are shown in Fig. 2b, indicating that the molar ratio 50 : 1 of Os(VI)-TEMED to ribonucleoside should be used. No significant differences between chromatographic signals were observed, while performing the reaction at room temperature, 40 °C and 60 °C, so it was decided to keep the reaction mixture at room temperature. Reaction time was set for 2 h, since for longer times no further increase of the chromatographic signals (peak height) was observed. Worth mentioning, in earlier studies the attachment of Os(VI)-2,2′-bipyridine to ribose was achieved at pH 6.5–7.5, with 2 : 1 molar excess of osmium, at room temperature (1.5 h).²³ Finally, the obtained compounds were stable for at least two weeks at room temperature, as verified by chromatographic analysis of individual Os-labeled standards carried out daily during this period of time (relative standard deviation evaluated for the individual peak height measurements did not exceed 8%).

Fig. 1 Labeling 5-methylcytidine with Os(VI)-reaction scheme.

Fig. 2 Effect of (a) pH and (b) molar ratio Os(VI)-TEMED: ribonucleoside on the analytical signals of labeled cytidine (·–◆–·), uridine (⋯■⋯), 5-methylcytidine (–▲–) and guanosine (–●–) obtained by HPLC-UV (270 nm). Chromatographic conditions given in Table 1, UV detection at 270 nm, peak height measurement mode.

Chromatographic separation of the labeled C, 5mC, U and G was studied using reversed phase analytical column and ICP-MS-compatible mobile phases (diluted aqueous phase and low percentage of organic modifier). The preliminary experiments were carried out with UV detection in order to avoid exposure of MS analyzer and detector to excessive amounts of Os. In Fig. 3a, a typical chromatogram of mixed standard solution is presented; the baseline resolution of four compounds of interest was achieved in 14 min, using isocratic elution with relatively low concentration of acetate buffer and low percentage of methanol (6 mmol L⁻¹ ammonium acetate, pH 4.4 + 0.05% triethylamine : methanol, 85 : 15). It was examined by ICP-MS detection that the bulk peaks in the region of dead volume (Fig. 3a, UV detection) corresponded to the elution of osmium species. Furthermore, while injecting Os(VI)-TEMED solution, similar early-eluting peaks were observed with UV detection and ICP-MS detection confirmed the presence of Os, which indicates that these chromatographic peaks corresponded to the excess of labeling reagent. In order to protect mass analyzer and detector, it was decided to start MS data acquisition at 2.5 min of chromatographic run.

Fig. 3 Typical reversed phase chromatograms of mixed standard solution (C: cytidine, U: uridine, 5mC-5-methylcytidine, G: guanosine) after labeling ribonucleosides with Os(VI): (a) 10 μmol L⁻¹ of each compound, UV detection at 270 nm, (b) 5.0 nmol L⁻¹ and 50 nmol L⁻¹ of each compound, ICP-MS detection at ¹⁸⁹Os, data acquisition started at 2.5 min of the chromatographic run.

Once established the separation conditions, structural characterization of Os-labeled cytidine and 5-methylcytidine was carried out by electrospray ionization mass spectrometry. The reaction mixture of individual ribonucleoside was introduced onto the column and the fraction was collected from 6.5 min to 7.0 min of chromatographic run for C and from 8.7 min to 9.2 min for 5mC. Mass spectra obtained for the two compounds are shown in Fig. 4a. For cytidine, molecular ion corresponded to m/z 582.1, for 5-methylcytidine to m/z 596.1. As shown in Fig. 4b and c, the two clusters exhibited isotopic patterns in good agreement with those calculated theoretically for the structures of the two compounds. The second ion cluster in the ESI-MS spectra was identical for Os-labeled C and for 5mC (m/z 471.1), indicating partial fragmentation of the molecular ions in a way that methyl group attached to pyridine (in 5mC) is lost and osmium atom preserved. The isotopic pattern of this ion cluster (Fig. 4b) was in agreement with that theoretically calculated (Fig. 4c) for the fragmentation ion proposed (Fig. 5). The collision-induced mass spectra (ESI-MS/MS) of two molecular ions and the proposed fragmentation patterns are presented in Fig. 5. Taking into account that molecular ion of cytidine was taken for ¹⁹⁰Os (m/z 580.1) and that of 5-methylcytidine for ¹⁹²Os (m/z 596.0), the identical spectra were obtained for the two compounds (Fig. 5a), which indicates that methyl group attached to pyridine ring in 5mC is always lost during fragmentation. The fragmentation patterns proposed are presented in Fig. 5b. The difference between m/z 580.1 and 469.1 for cytidine (or between 596.0 and 471.1 for methylated cytidine) points to the loss of cytosine (or 5-methylcytosine) and hydrogen atom, both of them bound to carbon 1 position in ribose. It was assumed that the fragmentation ions m/z 423.1 and 425.1 were formed by opening the pyridine ring on double bonds, loss of –CH–CH(NH₂)– or –C(CH₃)–C(NH₂)– group (for C and 5mC respectively) and separation of TEMED (Fig. 5b, dashed line). The formation of m/z 381.1 for C (or 383.1 for 5mC) involves detachment of TEMED and suggests the loss of –CH–CH(NH₂)=N–C(O)– group in the case of C and –C(CH₃)–C(NH₂)=N(O)– in the case of 5mC (Fig. 5b dotted line). Overall, the above results enable to confirm the structures of Os-labeled ribonucleosides, according to the reaction scheme presented in Fig. 1.

Fig. 4 Structural characterization of Os-labeled cytosine (C–Os(VI)-TEMED) and 5-methylcytosine (5mC–Os(V)-TEMED): (a) ESI-MS spectra; (b) isotopic pattern of ion clusters; (C) theoretical isotopic pattern of molecular ions and of fragmentation ion (m/z 471.1, structure proposed in Fig. 5).

Fig. 5 Structural characterization of Os-labeled cytosine and 5-methylcytosine: (a) ESI-MS/MS spectra of m/z 580.1 (¹⁹⁰Os) for (C–Os(VI)-TEMED) and m/z 596.0 (¹⁹²Os) for (5mC–Os(V)–TEMED); (b) proposed fragmentation of molecular ions.

Finally, calibration was performed using two detection systems. For ICP-MS detection, the mixed standard solutions contained 0, 5, 25, 50, 100, 150 and 200 nmol L⁻¹ of each ribonucleoside were used, while for UV detection, the concentration levels were 0, 5, 10, 25, 50 and 100 μmol L⁻¹. In Fig. 3b, the ICP-MS chromatograms of two calibration solutions are presented at 5.0 nmol L⁻¹ and 50 nmol L⁻¹ concentration levels of each compound. In addition to labeled nucleosides, the elution of other, minor Os-containing compounds can be observed. At this stage of our study, these retained species of osmium remain unidentified, however it seems possible that they could correspond to lower oxidation states of Os, originated from the excess of the derivatizing reagent (Os(VI)-TEMED). The analytical figures of merit were evaluated based on peak height measurements for UV and peak area integration for ICP-MS detection. These parameters are presented in Table 2. The values of detection limits for UV detection were in agreement with those reported in the previous work for non-labeled nucleosides.²⁹ More importantly, it can be observed in Table 2 that the ICP-MS detection made it possible to substantially lower the detection limits provided by UV detection. Furthermore, when calibration solutions injected into HPLC-ICP-MS system contained Os(VI)-TEMED (0, 5, 25, 50, 100, 150 and 200 nmol L⁻¹), the detection limit of 20 pmol L⁻¹ was obtained. The coefficients of variation at concentration level 5 nmol L⁻¹ did not exceed 1.2% and those evaluated at concentration level 100 nmol L⁻¹ were lower than 0.8% (n = 5, within day precision). The sensitivities, evaluated as slope values of linear regression functions and the values of DL for Os(VI)-TEMED (slope 20078, DL 20 pmol L⁻¹), for Os-labeled C (slope 19037, DL 24 pmol L⁻¹) and Os-labeled 5mC (slope 18 392, DL 21 pmol L⁻¹) were essentially identical, which proves that ICP-MS detection on ¹⁸⁹Os was compound-independent. The agreement between calibration slopes and detection limits obtained for Os(VI)-TEMED standard and for Os-labeled C and 5mC also indicates practically 100% yield of nucleoside labeling with osmium, Since the HPLC-ICP-MS procedure proposed here provides baseline separation and quantification of low concentration levels of nucleosides labeled on ribose with Os(VI), we consider this procedure potentially useful for the evaluation of RNA methylation, defined as the ratio between methylated cytidine and total cytidines in RNA or RNA fragments with particular interest for the analysis of micro samples.

Table 2 Analytical parameters evaluated. For ICP-MS detection, peak area measurement mode and calibration range 0–200 nmol L⁻¹ were used for each ribonucleoside, while for UV detection, peak height measurement mode and calibration range 0–100 μml L⁻¹ were used. T_ret: retention time, DL: based on six standard deviations of baseline measured in the elution region of each analyte

Nucleoside	T ret ± SD, min (n = 5)	ICP-MS detection, ^189Os		UV detection, 270 nm
		R ^2	DL, pmol L^−1	R ^2	DL, nmol L^−1
Cytidine	3.9 ± 0.1	0.9999	24	0.9996	1.07
Uridine	5.6 ± 0.1	0.9998	38	0.9997	1.58
5-Methylcytidine	6.4 ± 0.1	0.9998	21	0.9998	1.32
Guanosine	10.1 ± 0.2	0.9997	28	0.9998	0.98

---

Conclusions

In this work, a novel procedure is proposed for the quantitative analysis of ribonucleosides, based on labeling ribose residues with Os(VI), reversed phase separation and ICP-MS detection. The extremely low detection limits obtained (20–38 pmol L⁻¹) would fit the requirements existing in biological studies, in which small amounts of RNA are available. The potential utility of our approach for the evaluation of RNA methylation, defined as the percentage of methylated cytidines with respect to total amount of cytidines in RNA or RNA fragments, has been demonstrated. Since analytical signals of labeled C and 5mC (areas of the chromatographic peaks) depend on the amount of Os in the eluting compound and is independent on the compound structure, the evaluation of RNA methylation based on peak area measurements becomes possible, without necessity of calibration. The proposed procedure might be helpful in studies focusing functional characterization of methylation of the non-coding RNA in relevance to epigenetic regulation of gene expression. To demonstrate the feasibility of the procedure in such studies, in future development, it will be applied for the analysis of real-world biological micro samples.

Acknowledgements

The authors thank Professor Joseph A. Caruso for discussion of the results and for valuable suggestions. The financial support from CONACYT (Mexico), project 49405 is gratefully acknowledged.

References

1. K. Gronbaek, C. Hohter and P. A. Jones, APMIS, 2007, 115, 1039–1059.
2. F. Mohn and D. Schubeler, Trends Genet., 2009, 25, 129–136.
3. J. E. Sutherland and M. Costa, Ann. N. Y. Acad. Sci., 2003, 983, 151–160.
4. J. Mellor, P. Dudek and D. Clynes, Curr. Opin. Genet. Dev., 2008, 18, 116–122.
5. G. P. Delcuve, M. Rastegar and J. R. Davie, J. Cell. Physiol., 2009, 219, 243–250.
6. A. Dueñas_Gonzalez, M. Lizano, M. Candelaria, L. Cetina, C. Arce and E. Cervera, Molecular Cancer, 2005, 38–62.
7. A. H. Zhang, H. H. Bin, X. L. Pan and X. G. Xi, J. Toxicol. Environ. Health, Part A, 2007, 70, 970–975.
8. S. L. Reamon-Buettner, V. Mutschler and J. Borlak, Mutat. Res., Rev. in Mutat. Res., 2008, 659, 158–165.
9. H. Fieghl and K. Elmasry, Dis. Markers, 2007, 23, 89–96.
10. E. S. McKenna and C. W. Roberts, Cell Cycle, 2009, 8, 23–26.
11. O. Mathieu and J. Bender, J. Cell Sci., 2004, 117, 4881–4888.
12. M. Svedruzic Zeljiko, Curr. Med. Chem., 2008, 15, 92–106.
13. M. Schaefer, T. Pollex, K. Hanna and F. Lyko, Nucleic Acids Res., 2009, 37, e12/1–e12/10.
14. M. G. Goll, F. Kirpekar, K. A. Maggert, J. A. Yoder, C. L. Hsieh, X. Zhang, K. G. Golic, S. E. Jacobsen and T. H. Bestor, Science, 2006, 311(5759), 395–398.
15. F. F. Costa, Gene, 2008, 410, 9–17.
16. C. H. Chang, M. Beer and L. G. Marzilli, Biochemistry, 1977, 16, 33–38.
17. E. Palecek, Methods Enzymol., 1992, 212, 305–318.
18. F. B. Daniel and E. J. Behrman, Biochemistry, 1976, 15, 565–568.
19. L. Havran, M. Fojta and E. Palecek, Bioelectrochemistry, 2004, 63, 239–243.
20. E. Palecek, Methods Enzymol., 1992, 212, 139–155.
21. F. Jelen, P. Karlovsky, E. Makaturova, P. Pecinka and E. Palecek, Gen. Physiol. Biophys., 1991, 10, 461–473.
22. E. Palecek, P. Boublikova and K. Nejedly, Biophys. Chem., 1989, 34, 63–68.
23. M. Trefulka, V. Ostatna, L. Havran, M. Fojta and E. Palecek, Electroanalysis, 2007, 19, 1281–1287.
24. E. Palecek, Electroanalysis, 2009, 21, 239–251.
25. M. Fojta, F. Jelen, L. Havran and E. Palecek, Curr. Anal. Chem., 2008, 4, 250–262.
26. K. Tanaka, K. Tainaka, T. Kamei and A. Okamoto, J. Am. Chem. Soc., 2007, 129, 5612–5620.
27. K. Wrobel, K. Wrobel and J. A. Caruso, Anal. Bioanal. Chem., 2009, 393, 481–486.
28. E. Palecek, S. Billova, L. Havran, R. Kizek, A. Miculkova and F. Jelen, Talanta, 2002, 56, 919–930.
29. A. Alcazar Magana, K. Wrobel, Y. Alvarado Caudillo, S. Zaina, G. Lund and K. Wrobel, Anal. Biochem., 2008, 374, 378–385.

---