On-Line Enhancement Technique for the Analysis of Nucleotides

Jul 29, 2006 - stationary, permitting potentially unlimited injection time, and hence unlimited sample enrichment power. The ability of PAEKI to maint...
0 downloads 0 Views 107KB Size
Download PDF
Anal. Chem. 2006, 78, 6608-6613

On-Line Enhancement Technique for the Analysis of Nucleotides Using Capillary Zone Electrophoresis/Mass Spectrometry Yong-Lai Feng and Jiping Zhu*

Chemistry Research Division, Health Canada, AL 0800C, EHC, Tunney’s Pasture, Ottawa, Ontario, Canada K1A 0L2

A new on-line capillary zone electrophoresis/mass spectrometry (CZE/MS), constant pressure-assisted electrokinetic injection (PAEKI), for the analysis of negatively charged nucleotides is reported. PAEKI uses an applied pressure to counterbalance the reverse electroosmotic flow in the capillary column during sample injection, while taking advantage of the field amplification in the sample medium. At balance, the running buffer in the column is stationary, permitting potentially unlimited injection time, and hence unlimited sample enrichment power. The ability of PAEKI to maintain a narrow sample zone over a long injection time seems to be a result of the formation of a high ion concentration band at the boundary of the two media due to rapid deceleration of the migrating ions at the boundary. The injected amount of analytes proved to be linearly proportional to both the field amplification factor, which is expressed as the ratio of resistivities of sample medium to running buffer, and the injection time, which extended up to 1200 s in CZE/MS and 3600 s in CZE/UV. For a 300-s on-line PAEKI injection in CZE/ MS, 3 orders of magnitude sample enhancement (5000fold enrichment) could be observed for the four single nucleotides without compromising separation efficiency and peak shape, and an achievement of detection limits between 0.04 and 0.07 ng/mL. With appropriate sample cleanup, PAEKI can be used in the analysis of single nucleotides in enzyme-digested DNA. Capillary electrophoresis (CE), a separation technique that includes a number of separation methods such as capillary zone electrophoresis (CZE) and micellar electrokinetic chromatography (MEKC), has emerged as a fast, low-cost, and powerful separation technique for both charged and neutral analytes.1-3 CE separation of single nucleotides over a wide range of pH values has also been reported.4-10 UV detection usually provides a typical detection limit in micromolar range. Therefore, a number of highly sensitive * Corresponding author. Fax: 613 946-3573. E-mail: [email protected]. (1) Iadarola, P.; Cetta, G.; Luisetti, M.; Annovazzi, L.; Casado, B.; Baraniuk, J.; Zanone, C.; Viglio, S. Electrophoresis 2005, 26, 752-766. (2) Issaq, H. J. Electrophoresis 2000, 21, 1921-1939. (3) Beale, S. C. Anal. Chem. 1998, 70, 279-300. (4) Geldart, S. E.; Brown, P. R. J. Chromatogr., A 1998, 828, 317-336. (5) Warnke, U.; Gysler, J.; Hofte, B.; Tjaden, U. R.; van der Greef, J.; Kloft, C.; Schunack, W.; Jaehde, U. Electrophoresis 2001, 22, 97-103. (6) Norwood, C. B.; Jackim, E.; Cheer, S. Anal. Biochem. 1993, 213, 194199.

6608 Analytical Chemistry, Vol. 78, No. 18, September 15, 2006

detection methods have been developed and used to increase the detection sensitivity of CE instruments. Among them, fluorescence and postlabeling provide excellent sensitivity but no structure information on the analytes.10,11 Besides, these detection methods also involve a complex derivatization procedure.10,11 Mass spectrometry (MS), on the other hand, emerged as a new powerful detection technique, providing not only high sensitivity but also structure information of the analytes.12-14 Despite the attempt to increase detection sensitivity, the main limitation of the CE technique lies with its extremely small injection volume, typically in the nanoliter range, 3 orders of magnitude smaller than the usual injection amount of highperformance liquid chromatography (HPLC). This limitation has greatly compromised the applications of the CE technique. In the past, several sampling enhancement techniques have been developed to address this issue. They include transient on-column isotachophoresis,15 which was well reviewed by Bocek,16 sample stacking,17,18 sweeping,19,20 and utilization of solid-phase packing at the head of the CE column.21,22 Among these, the sweeping technique in MEKC for neutral analytes using surfactants as pseudostationary phase19 and the sample stacking technique for charged analytes in CZE,18 especially for analysis of nucleotides with CZE/MS,13,14 are to date the most powerful sample enhancement techniques. The enhancement capacity of both techniques, (7) Barry, J. P.; Norwood, C.; Vouros, P. Anal. Chem. 1996, 68, 14321438. (8) Schmitz, O. J.; Worth, C. C. T.; Stach, D.; WieBler, M. Angew. Chem., Int. Ed. 2002, 41, 445-448. (9) Stach, D.; Schmitz, O. J.; Stilgenbauer, S.; Benner, A.; Dohner, H.; Wiessler, M.; Lyko, F. Nucleic Acids Res. 2003, 31, e2. (10) Cornelius, M.; Worth, C. G. C. T.; Kliem, H.-C.; Wiessler, M.; Schmeiser, H. H. Electrophoresis 2005, 26, 2591-2598. (11) Lan, Z. H.; Qian, X.; Giese, R. W. J. Chromatogr., A 1999, 831, 325330. (12) Aussenac, J.; Chassagne, D.; Claparols, C.; Charpentier, M.; Duteurtre, B.; Feuillat, M.; Charpentier, C. J. Chromatogr., A 2001, 907, 155-164. (13) Deforce, D. L. D.; Ryniers, F. P. K.; Van den Eeckhout, E. G. Anal. Chem. 1996, 68, 3575-3584. (14) Wolf, S. M.; Vouros, P. Anal. Chem. 1995, 67, 891-900. (15) Thompson, T. J.; Foret, F.; Vouros, P.; Karger, B. L. Anal. Chem. 1993, 65, 900. (16) Gebauer, P.; Bocek, P. Electrophoresis 2002, 23, 3858-3864. (17) Burgi, D. S.; Chien, R.-L. Anal. Chem. 1991, 63, 2042-2047. (18) Chien, R.-L.; Burgi, D. S. Anal. Chem. 1992, 64, 489A-496A. (19) Quirino, J. P.; Terabe, S. Science 1998, 282, 465-468. (20) Quirino, J. P.; Terabe, S. Anal. Chem. 2000, 72, 1023-1030. (21) Swartz, M. E.; Merion, M. J. Chromatogr. 1993, 632, 209-213. (22) Tomlinson, A. J.; Braddock, W. D.; Benson, L. M.; Oda, R. P.; Naylor, S. J. Chromatogr., B 1995, 669, 67-73. 10.1021/ac0608568 CCC: $33.50

© 2006 American Chemical Society Published on Web 07/29/2006

however, is restricted by the limitation in capillary column length employed for CE analysis. In this paper, we report the development of a new sampling enhancement approach, the pressure-assisted electrokinetic injection technique (PAEKI), which uses the time dimension instead of the space dimension to achieve one-step, on-line preconcentration of the charged analytes such as single nucleotides in CZE/ MS. The technique is based on the concept that the movement of running buffer in the capillary column due to electroosmotic flow (EOF) can be balanced during sample injection by an externally applied pressure in the opposite direction of the EOF under a given electric field, to achieve a stationary state of the running buffer inside the column. Integrated with field amplification techniques, PAEKI delivers a powerful, potentially unlimited sample enhancement for the charged analytes in CZE/MS analysis. EXPERIMENTAL SECTION Chemicals. Cytidine 5′-monophosphate disodium salt (CMP, 98+%), guanosine 5′-monophosphate disodium salt (GMP, 99+%), adenosine 5′-monophosphate sodium salt (AMP, 99+%), uridine 5′-monophosphate disodium salt sesquihydrate (UMP, 98%), tetrahydrofuran (99.9+%, HPLC grade, inhibitor-free), and acetic acid (99.7+%, ACS reagent) were purchased from Sigma-Aldrich (Sigma-Aldrich Co., St. Louis, MO). Tris(hydroxymethyl)aminomethane (Tris, 99.9+%), triethylamine (99.5%), ammonium acetate (99.99+%), ammonium carbonate (ACS reagent), ammonium hydroxide (28-30%, ACS reagent), hydrochloric acid (high-purity grade), ethylenediaminetetraacetic acid, disodium salt dehydrate (EDTA, 99+%), SAX cartridge (50 mg/1 mL), and magnesium chloride hexahydrate (minimum 99.0%, SigmaUltra) were obtained from Sigma-Aldrich Canada (Oakville, ON, Canada), sodium hydroxide solutions (0.1 and 1 N) were purchased from Agilent Technologies Inc. (Waldbronn, Germany). 2-Propanol and methanol, both 99.99+%, were purchased from EMD Chemicals Inc. (Gibbstown, NJ). Deionized (DI) water (18.3 Ω) was generated in-house in the laboratory using a Super-Q water generation system (Fisher Scientific, Ottawa, Canada). Preparation of Nucleotide Samples and Buffer Solutions. The stock solutions of 1 mg/mL of four single nucleotides were prepared by dissolving corresponding chemicals in DI water and stored in -20 °C. An ammonium carbonate buffer stock solution (0.2 M) for CZE analysis was prepared by weighing 1.561 g of ammonium carbonate, dissolving it in 100 mL of DI water, and adjusting the pH to 9.7 with 20% ammonium hydroxide solution. The working buffer solutions of ammonium carbonate were prepared by diluting the above buffer stock solutions to the required concentration in DI water, and the buffer solutions were again adjusted to pH 9.7 with 10% ammonium hydroxide solution. Tris-HCl stock buffer (1.0 M) was prepared by dissolving 6.057 g of Tris in 50 mL of DI water, and the buffer was adjusted to pH 7.5 with 1.0 M HCl. EDTA stock solution (0.5 M) was prepared by dissolving 18.612 g of EDTA in 100 mL of DI water (pH 8.0). MgCl2 stock solution (1.0 M) was prepared by dissolving 10.1655 g of magnesium chloride hexahydrate in 100 mL of DI water. A 10-mL aliquot of Tris-HCl stock buffer, 0.2 mL of EDTA stock solution, and 4 mL of MgCl2 stock solution were mixed and further diluted to 100 mL with DI water to obtain a 10× DNase I buffer,

which contained 100 mM Tris-HCl, 1.0 mM EDTA, and 40 mM MgCl2, pH 7.5. Spiking Samples. Samples of 50 ng each of CMP, AMP, GMP, and UMP were spiked into 1 mL of hydrolysis solution which contained 25 µg of DNase I, 30% 1× DNase I buffer, 20 mM ammonium bicarbonate, and 81 µg of NP1 as described by Ramsahoye.23 The well-mixed samples were cleaned up by SAX cartridge. The SAX cartridges were preconditioned with 3 mL of water (three times), 3 mL of methanol (three times), and 3 mL of 1:1 water/methanol (three times). After the sample was passed through the cartridge, the cartridge was washed twice with 2 mL of DI water. The nucleotides were eluted with 1 mL of 2% acetic acid from the cartridge. The eluate was freeze-dried and redissolved in 2 mL of DI water for PAEKI-CE/MS detection. CZE/MS Conditions. An Agilent capillary electrophoresis (3D-CE system) with a built-in diode-array detector (DAD) was linked to an Agilent 1100 series single quadrupole mass spectrometer (1100 series LC/MSD; Agilent Technologies Inc., Waldbronn, Germany), equipped with an Agilent electrospray ion source. Agilent CE ChemStation was used for the instrument control, data acquisition, and data analysis. The length for the bare fused-silica CE capillary column was 120 cm with an i.d. of 50 µm (Ploymicro Tech. LLC, Phoenix, AZ). The new capillary column was conditioned by flushing the column with a 1 N NaOH solution for 10 min, DI water for 5 min, and the running buffer for 10 min, respectively. The column cassette temperature was set at 20 °C. The optimal conditions of sample introduction under PAEKI were as follows: pressure of 50 mbar, injection voltage of -13 kV. Once the injection was completed, the voltage was switched to +30 kV for the CZE separation. The DAD wavelength was set at 258 nm. The electrospray ionization mass spectrometry was operating in both SIM and scan modes. For scan mode, the mass spectra were acquired over the mass range 300-400 Da. The ions monitored in MS were 322.1 for CMP, 323.1 for UMP, 362.1 for GMP, and 346.1 for AMP, respectively. The ESI capillary voltage of mass spectrometry was maintained at -4 kV, and the sampling orifice was set to ground potential for the negative ion mode. The fragmentor, step size, and gain were set at 65, 0.1, and 3.0, respectively. Nitrogen was used as the nebulizer and the drying gas, at a pressure of 25 psig and a flow rate of 8 L/min, respectively. The drying gas was heated at 300 °C. The sheath liquid consisted of 25% (v/v) 2-propanol and 1% aqueous triethylamine (pH 11) and was delivered at a rate of 2 µL/min by an 1100 series Agilent isopump. RESULTS AND DISCUSSION Balance of EOF. In CE, samples are usually introduced into the capillary column by hydrodynamic (also called pressure) injection or electrokinetic injection. While the mechanism of hydrodynamic injection is straightforward, the theory behind the electrokinetic injection is more complicated.24 It is generally believed that the amount of analytes injected under conventional electrokinetic condition can be expressed as in eq 1, where µepi is the electrophoretic mobility of the analyte in the sample solution, µEOF the mobility of bulk buffer due to EOF, A the cross sectional area of the capillary, Einj the electric field strength, which equals (23) Ramsahoye, B. H. Methods 2002, 27, 156-161. (24) Krivacsy, Z.; Gelencser, A.; Hlavay, J.; Kiss, G.; Sarvari, Z. J. Chromatogr., A 1999, 834, 21-44.

Analytical Chemistry, Vol. 78, No. 18, September 15, 2006

6609

Ni(t) ) (µepi - µEOF)EinjACinjt

(1)

to the injection voltage (Vinj) divided by the length of the capillary (L), Cinj the analyte concentration in the sample solution, and t the injection time. Negatively charged chemicals such as nucleotides studied in this investigation are eletrokinetically injected into the capillary under a negative voltage applied to the capillary. Under the negative voltage, the reverse EOF causes the bulk running buffer in the bare column to move toward the injection end. This usually undesired reverse EOF is used in the stacking technique, where the sample solution is injected into the capillary under hydrodynamic conditions and the sample solution then retreats toward the injection end by the EOF while analytes are being focused at the boundary layer of the sample solution and running buffer in the capillary.17,18 Instead of having to inject the sample solution and to remove it afterward, the PAEKI uses a real-time balancing approach to ensure that the EOF is compensated by an external pressure in the opposite direction and the running buffer in the capillary is in stationary state during the sample injection (µEOF ) 0 in eq 1), thereby eliminating the need for a long sample zone such as that in the stacking technique. The balance of EOF can be achieved by adjusting either the pressure under a given voltage or the voltage under a given pressure. Since the electrophoretic velocity (νepi) is proportional to the electric field strength (E) of the medium (νepi ) µepiE), it would be more desirable to have higher voltage during PAEKI injection with larger external pressure. Due to the limitation of the pressure capability (50 mbar maximum) of the commercial CE instrument used in this study, the balance of running buffer movement was achieved by adjusting the voltage applied to the column while the maximal external pressure of 50 mbar was maintained. When the electric voltage increased from -8 to -13 kV under the fixed external pressure of 50 mbar, the peaks of the analytes in the electropherogram became progressively sharper, and at the same time, the retention time of the peaks became longer (Figure 1). Such a phenomenon indicated that at lower voltage such as at -8 kV the EOF in the column was smaller than the flow generated by the external pressure (50 mbar), which results in the bulk solution in the column moving toward the detection end. That resulted in the creation of a sample zone in the column similar to that in the stacking technique. Because there was no water plug retreat and analyte focusing step in PAEKI, the peaks of the analytes were poor. With increasing voltage, the EOF gradually reduced the effects of external pressure and eventually reached the balance for the flow of running buffer. Reduced sample zone length narrowed the sample band and extended the effective column length, which resulted in a longer retention time with a better peak shape for the analytes. At -13 kV for this CZE/ MS system, a narrow sample band must have been created at the injection end of the column during PAEKI as this voltage resulted in the best peak shape. When the electric voltage was at -14 kV and beyond, the peak abundance was significantly reduced (Figure 1). At this point, the reverse EOF overpowered the external pressure, causing the sample zone to be pushed out of the capillary as shown in eq 1. In the worse case, the sample could not be injected into the capillary. 6610 Analytical Chemistry, Vol. 78, No. 18, September 15, 2006

Figure 1. Effect of injection voltage in the range of -8 to -14 kV. Injection pressure, 50 mbar. Injection time, 300 s. Running buffer, ammonium carbonate, 20 mM, pH 9.7. Column length, 120 cm, i.d. 50 µm; CZE operating condition, T ) 20 °C, V ) 30 kV. Analytes, 30 ng/mL each in water. MS conditions: capillary voltage, -4000 V, sheath liquid, 25% 2-propanol containing 1% TEA in water, delivered at a flow rate of 2 µL/min. Peak identification: 1, AMP; 2, CMP; 3, GMP; and 4, UMP.

Figure 2. Sample matrix effect in PAEKI technique. (A) 100 000 ng/mL AMP, CMP, GMP, and UMP in the running buffer, PAEKI injection at -13 kV, and 50 mbar of pressure for 300 s. (B) 10 000 ng/mL AMP, CMP, GMP, and UMP in the running buffer, PAEKI injection at -13 kV, and 50 mbar of pressure for 300 s. (C) 100 000 ng/mL AMP, CMP, GMP, and UMP in the running buffer, hydrodynamic injection at 50 mbar for 300 s. Running buffer and CZE/MS operating conditions were the same as in Figure 1.

Electric Field Amplification in PAEKI. Balance of EOF alone is not sufficient for the success of the enrichment of analytes in PAEKI. As shown in Figure 2, when analytes that were dissolved in the same running buffer injected under PAEKI, the analytes migrated simply from the sample vial to capillary without focusing at the inlet of the capillary. Although there was no movement of the running buffer under PAEKI, there was no focusing of analytes either. For example, for a 300-s PAEKI injection, a broad band of ∼5 min was observed (Figure 2A), very similar to the signal band in conventional pressure injection under the same conditions without voltage (Figure 2C). The broad peak

Figure 3. Relationship between injected amount of analytes and the ratio of resistivities. Analyte, 5 ng/mL CMP in water; PAEKI injection time, 180 s; running buffers were ammonium carbonate at various concentrations adjusted to pH 9.7; γ values were calculated from measured resistivity (F) values measured in sample medium and in various running buffers. Other CZE/MS conditions were the same as in Figure 1.

band was not attributed to the saturation of the capillary as even at 10 times less concentration; the same broadness of injected sample zone existed under PAEKI (Figure 2B). One of the possible contributors to form the narrow sample band in PAEKI could be the field amplification of the sample medium. In CZE, the electric field in sample medium (Esample) in the column is amplified by γ/[γx + (1 - x)] times, where x is a fraction of injected sample zone length to the total length of the capillary and γ is the ratio of resistivity (F) of the sample medium to the running bulk buffer in the column.18 Under optimal PAEKI conditions, the sample zone is at the tip of the injection end of the column. Therefore, theoretically x can be considered approaching zero. In such a case, Esample equals to γ times the electric field in column (Einj) and eq 1 can be deduced to eq 2 if combined with µEOF ) 0 as mentioned above.

Ni(t) ) γ µepiEinjACinjt

(2)

As shown in Figure 3, using CMP as a typical example, a linear relationship between the γ values and the injected amount, expressed by the increase of peak area, was indeed observed under the PAEKI conditions for the γ values between 500 and 3500. The resistivity values (inversion of the conductivity value) in sample medium and in various running buffer solutions with different buffer concentrations were experimentally measured. As the electric field strength can be amplified by γ times in the sample medium, the electrophoretic velocity (ν) of the analytes in the sample medium will be γ times of that in the running buffer. As shown in Figure 3, γ value up to 3500 could be easily achieved. Such a large γ value in the CZE system leads to the fast migration of analytes from sample medium onto the column and dramatic deceleration of the velocity of the analytes at the boundary, forming a concentrated sample zone between the sample medium and the running buffer as a result of sample stacking effect.17,18 The localized high ion concentration at the boundary may contribute to the maintenance of the narrow sample zone during

Figure 4. Relationship between injected amount of analytes and injection time. Analyte, 10 ng/mL CMP in water. PAEKI injection times varied. Buffer and CZE/MS conditions were same as in Figure 1. For CZE/UV, the UV detection window was opened at the position of 8.5 cm from the outlet end and the detection wavelength was set at 258 nm.

PAEKI. One possible reason is that the migration of ions crossing the boundary was slowed due to possible increase of localized viscosity (η) of the solution around the sample zone as the electrophoretic mobility is inversely proportional to the viscosity (µepi ∼ 1/η). The localized high viscosity may also well contribute to the prevention of diffusion of the concentrated analytes during the long injection time. Chien and Helmer have discussed mismatched flow profile in capillary filled with two buffer solutions of different resistivities such as that in stacking technique. The mismatched flow profile may cause the creation of laminar profile due to a local electroosmotic velocities flow resulting in potential broadening of the analyte band.25 The peak broadening was not observed in PAEKI, attributed to the fact that the analyte band is located at the injection tip and unlike the stacking technique and the length of sample solution x is negligible as discussed above. In such a case, the bulk electroosmotic velocity in the column can be considered the same as that of the running buffer, eliminating potential mismatch of the flow profile. It has also been observed in experiment that a linear relationship existed between injection time and the area counts of up to 1200 s in CZE/MS and 3600 s in CZE/UV (Figure 4). The ability of PAEKI to introduce samples during such a long injection time is unprecedented. To our best knowledge, it is very hard with other electrokinetic injection techniques in CZE to reach such a long sample injection time without compromising the peak shape or separation efficiency. The injection time in the conventional electrokinetic injection is usually limited to several seconds.18,26,27 Therefore, this several orders of magnitude increase of sample injection time could effectively increase the equivalent injection volume of CZE to a level comparable to the injection volume of HPLC. The possible reason for the shorter injection time of linearity under CZE/MS compared to that under CZE/UV is not (25) Chien, R. L.; Helmer, J. C. Anal. Chem. 1991, 63, 1354-1361. (26) Weston, A.; Brown, P. R. HPLC and CE Principles and Practice; Academic Press: San Diego, CA, 1997: p 134. (27) Chen, G.; Zhang, Y.; Yang, P. Anal. Sci. 2005, 21, 1161-1165.

Analytical Chemistry, Vol. 78, No. 18, September 15, 2006

6611

Figure 5. Typical PAEKI CZE/MS electropherogram. (A) 20 ng/mL of each analyte in water with PAEKI injection at -13 kV, 50 mbar for 300 s. (B) 100 000 ng/mL of each analyte in the running buffer with pressure injection at 50 mbar for 2 s. Other CZE/MS conditions were the same as in Figure 1. Peak identification: 1, AMP; 2, CMP; 3, GMP; and 4, UMP.

well understood. It is thought that it could be due to the infusion of the sheath liquid into the outlet end of capillary column, which could change the composition of the running buffer and hence the EOF in the capillary column. Performance of PAEKI CZE/MS. Figure 5 shows the sample enhancement of four nucleotides (AMP, CMP, GMP, UMP) in a 300-s PAEKI injection in CZE/MS as compared to the hydrodynamic injection (2 s). A roughly 5000-fold increase in sample enhancement was observed. The difference in relative peak intensity between the two injection methods could be attributed to the different electrophoretic mobility of the analytes under PAEKI. Since there is no sample enrichment in hydrodynamic injection, the increase reflects the real sample enhancement. The broad dynamic linear calibration range of 4 orders of magnitude from 0.2 to 100 ng/mL was achieved in PAEKI CZE/ MS (r2 ) 0.9990). The 4 orders of magnitude linear calibration range is very broad compared to other enrichment techniques, such as sample stacking.28 To demonstrate the reproducibility and recovery of the PAEKI technique, a series of experiments have been conducted in this study. Although longer injection time can be employed, 300-s injection time was chosen to evaluate the method performance. In other words, all reproducibility, detection limits, and other measurement performances were conducted based on 300-s PAEKI injection. Measurements of eight aqueous standards containing 20 ng/mL concentrations of four nucleotides showed a relative standard deviation of 10.3% for AMP, 8.4% for CMP, 7.7% for UMP, and 11.5% for GMP. The detection limits, defined as 3 times the signal-to-noise ratio in this study, was obtained under the same injection conditions. They were 0.05 ng/ mL for AMP, 0.04 ng/mL for CMP, 0.06 ng/mL for UMP, and 0.07 ng/mL for GMP. Like the sample stacking technique, one of the potential challenges of the PAEKI technique is its sensitivity toward the sample matrix as PAEKI performed best in the situations where large field amplification in sample medium exists. To test the (28) Chien, R. L. Anal. Chem. 1991, 63, 2866-2869.

6612 Analytical Chemistry, Vol. 78, No. 18, September 15, 2006

Figure 6. Typical chromatograph for the hydrolysate medium sample spiked with four nucleotides under PAEKI technique. Injection, -13 kV with 50 mbar for 300 s. A 50-ng sample of each four nucleotides was spiked in the hydrolysate medium and cleaned as described in the Experimental Section, and final water solution was 2 mL. Running buffer and CZE/MS conditions were the same as in Figure 1.

applicability of the PAEKI technique in real samples, four nucleotides were spiked into a DNA digestion cocktail solution including DNA hydrolysate and other chemicals such as TrisHCl, EDTA, and MgCl2. The sample was then cleaned up by using a SAX cartridge prior to CZE/MS analysis. Samples were loaded onto the instrument for analysis under a sequence run. Good peak shapes of all four analytes in the processed samples were observed on the electropherogram (Figure 6), and recovery from 89 to 104% was achieved for all four nucleotides. The ability to run a CZE/ MS sequence for the analysis also shows the on-line ability of PAEKI and the robustness of the PAEKI method. This is because once the optimal voltage for balancing the EOF is experimentally determined under a given external pressure, it will remain the same as long as the running buffer and column conditions are kept unchanged. CONCLUSION The PAEKI technique uses time dimension for the injection period, thereby providing potentially unlimited sample enhancement power. It is the first report that an electrokinetic injection time in CZE is no longer limited by the spatial restrictions. In combination with field amplification, PAEKI can achieve a tremendous increase in the injected amount of analytes in CZE. Compared to current stacking or stacking/sweeping techniques in CZE, PAEKI has not only the on-line sample enrichment feature for the CZE/MS system but also a much simpler procedure to operate. The sample zone at the tip of the capillary column and the low conductivity of the sample medium provided the same function in PAEKI as the water plug in the sample stacking. In addition to negatively charged analytes, PAEKI in principle can also be applied to positively charged analytes in CZE by reversing the external pressure and the polarity of the electric field. This sometimes however requires modifications to the existing instrument settings. With the ability to enrich either

positively or negatively charged analytes, PAEKI technique could open a whole new field in the preconcentration of charged analytes in CZE analysis.

technical assistance in the study. The study was funded by Canadian federal government under the Canadian Environmental Protection Act.

ACKNOWLEDGMENT

Received for review May 10, 2006. Accepted July 11, 2006.

We thank Xiaofeng Yang, a late employee of Health Canada, for the preparation of DNA and DNA adduct samples and other

AC0608568

Analytical Chemistry, Vol. 78, No. 18, September 15, 2006

6613