- Email: [email protected]
Two-dimensional capillary electrophoresis using tangentially connected capillaries
Journal of Chromatography A, 1154 (2007) 454–459
Two-dimensional capillary electrophoresis using tangentially connected capillaries Eskil Sahlin ∗ Department of Chemistry, G¨oteborg University, 412 96 G¨oteborg, Sweden Received 26 January 2007; received in revised form 29 March 2007; accepted 2 April 2007 Available online 6 April 2007
Abstract A novel type of fused silica capillary system is described where channels with circular cross-sections are tangentially in contact with each other and connected through a small opening at the contact area. Since the channels are not crossing each other in the same plane, the capillaries can easily be filled with different solutions, i.e. different solutions will be in contact with each other at the contact point. The system has been used to perform different types of two-dimensional separations and the complete system is fully automated where a high voltage switch is used to control the location of the high voltage in the system. Using two model compounds it is demonstrated that a type of two-dimensional separation can be performed using capillary zone electrophoresis at two different pH values. It is also shown that a compound with acid/base properties can be concentrated using a dynamic pH junction mechanism when transferred from the first separation to the second separation. In addition, the system has been used to perform a comprehensive two-dimensional capillary electrophoresis separation of tryptic digest of bovine serum albumin using capillary zone electrophoresis followed by micellar electrokinetic chromatography. © 2007 Elsevier B.V. All rights reserved. Keywords: Capillary electrophoresis; Tangentially connected capillaries; Two-dimensional separation
1. Introduction In order to increase the knowledge of life and our environment, it is essential to be able to identify and quantify compounds in clinical and environmental samples. However, these samples can be extremely complex when considering low concentrations. Body fluids, cells and some environmental samples such as urban dust are important examples of samples that contain or can contain a vast number of compounds in a large concentration range. For instance, in a human cell there are in the order of 104 proteins which are expressed in a particular type of cell at any time at different levels of concentration with a dynamic range of five-fold [1]. Tryptic digests of cells could contain as much as a million peptides [1]. Hence, in order to analyze or to perform studies on such samples, analytical techniques with extremely high selectivity are needed and a separation step is
∗ Present address: Chemistry and Materials Technology, SP Technical Research Institute of Sweden, Brinellgatan 4, 501 15 Bor˚as, Sweden. Tel.: +46 10 516 5265; fax: +46 33 12 37 49. E-mail address: [email protected].
0021-9673/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2007.04.001
often included. However, even for the most efficient separation techniques, efficiencies and peak capacities are not high enough for many complex samples [1–3] and two-dimensional separation systems have therefore been developed. Proteins are commonly separated using two-dimensional polyacrylamide gel electrophoresis (2D-PAGE) [1,4,5]. In addition, several on-line two-dimensional separation systems (often LC–LC and LC–CE) have been described in the literature [2–4,6–8]. Most of them utilize liquid chromatography in the first separation step, however, this separation is not easily stopped and is therefore often kept running while the second separation takes place. Hence, the second separation must be relatively fast when the two-dimensional system is operated comprehensively. This problem is avoided when capillary electrophoresis is used in both separation steps, though a fast second separation step is still desirable in order to have reasonable total analysis times when operated comprehensively. Two-dimensional capillary electrophoresis systems have been described several times in the literature using systems based on either fused silica capillaries or microfabricated microfluidic systems [3,7–18], and put into practice especially by the Dovichi group [9,11,14,15,17] and the Ramsey group [10,12,13]. An important issue with
E. Sahlin / J. Chromatogr. A 1154 (2007) 454–459
two-dimensional capillary electrophoresis systems is the design of the capillary–capillary interface where analytes are transferred from the first to the second separation. The Dovichi group has constructed a system with two fused silica capillaries aligned in series and where separation media to the second capillary enters the system at a small gap between the capillaries [9,11,14,15,17]. Two-dimensional separations have also been performed in microfabricated systems where the different channels are crossing each other in the same plane [10,12,13,18]. Transport of analytes and solution through the crossing is performed by controlling the voltage in several reservoirs simultaneously. An additional capillary–capillary interface includes an etched porous junction [16]. Another issue to consider is the choice of separation modes in the first and the second dimension for various analytes. Two-dimensional separations of proteins with high orthogonality have been demonstrated using capillary sieving electrophoresis–micellar electrokinetic chromatography [9,14,15,17], submicellar capillary electrophoresis–capillary zone electrophoresis (at two different pH) [11], capillary isoelectric focusing–capillary zone electrophoresis [16] and capillary gel electrophoresis–micellar electrokinetic chromatography [18]. Peptides have been separated using micellar electrokinetic chromatography–capillary zone electrophoresis [12,13], and (open-channel) electrochromatography–capillary zone electrophoresis [10]. In this paper, a novel low dead-volume capillary–capillary interface is fabricated without microfabrication technology that allows straightforward filling of the systems with different solutions in the two different capillaries. The capillary system is characterized and used to perform two-dimensional capillary electrophoresis separations in a comprehensive mode. Several other possible applications of the system are also discussed briefly. 2. Materials and methods 2.1. Materials and tools Fused silica capillaries (75 m i.d., 360 m o.d.) and polyether ether ketone (PEEK) tubing (0.8 mm i.d.) were purchased from Upchurch Scientific (Oak Harbor, WA, USA), and heat shrink/melt tubing with an inner diameter of 1.65 mm was obtained from Cole Parmer (Vernon Hills, IL, USA). Tungsten wire with a diameter of 50 m and platinum rods with a diameter of 1.0 mm were obtained from Goodfellow (Huntingdon, UK). Heating of the heat/shrink melt tubing was obtained with a heat gun with a variable temperature range of 60–590 ◦ C. The tungsten wires were held in place during heating with a tapered pair of tweezers. 2.2. Instrumentation The capillary electrophoresis system is shown in Fig. 1. The system consisted of four reservoirs (marked A and numbered 1–4) that were connected to each other using a capillary connection (marked B) with four arms. The capillary connection is further described below. Two of the arms consisted
455
Fig. 1. A four armed capillary electrophoresis system: (A) reservoirs, (B) capillary connection, (C) fused silica capillaries (25 cm × 75 m i.d.), (D) fused silica capillaries (5 cm × 75 m i.d.), (E) PEEK tubing (15 cm × 0.8 mm i.d.) and (F) detector. The four arms are joined together in a FEP channel connection (marked B) where the two channels are tangentially in contact with each other. The detector is located 6.5 cm from the channel connection.
of 25 cm × 75 m i.d. fused silica capillaries (marked C) and ended in reservoirs 1 and 4. The other two arms consisted of 5 cm × 75 m i.d. fused silica capillaries (marked D) connected to 15 cm × 0.8 mm i.d. PEEK tubing (marked E) and ended in reservoirs 2 and 3. A UV/vis absorbance detector (Spectra 100 UV/vis detector, Thermo Separation Products, Fremont, CA, USA) (marked F) was used for detection and was located 6.5 cm from the capillary connection. The system was housed in a Plexiglas box with holders for the reservoirs and capillaries. The reservoirs consisted of plastic tubes that were filled with 1.30 ml of separation media. Reservoirs 1 and 3 could be sealed and pressurized using nitrogen gas (2 bar) facilitating filling and flushing of system with separation media. Platinum electrodes (1.0 mm diameter) were inserted into the reservoirs and connected to a high voltage power supply (Model 2462, Bertan Associates, Hauppauge, NY, USA). A high voltage switching relay (Model G2 (powered by 12 V dc), Gigavac, Santa Barbara, CA, USA) was used to change the location of the high voltage in the systems (i.e. two reservoirs were always floating). The high voltage switching relay and the high voltage power supply were controlled with timers (Model H5CX, Omron Electronics, Schaumburg, IL, USA) so that the system could be operated without manual assistance when the sample had been injected. The timers and the switching relay were powered by an isolated dc power source consisting of a rechargeable 12 V battery. Detection was performed at 210 or 225 nm and the signal was filtered by setting the rise time of the detector to 0.1 or 1.0 s. Data were collected at 10, 30 or 50 Hz using a multifunction I/O board (PCI-6034E from National Instruments, Austin, TX, USA) controlled with LabVIEW 6.0 (National Instruments).
456
E. Sahlin / J. Chromatogr. A 1154 (2007) 454–459
Fig. 2. (a) Initial alignment of: (A) heat/shrink melt tubing, (B) fused silica capillaries (25 cm × 75 m i.d.), (C) fused silica capillaries (5 cm × 75 m i.d.) and (D) tungsten wires. (b) Schematic drawing of the tangentially connected channels.
2.3. Chemicals 2-Chlorobenzoic acid (98%), trypsin (T-1005), bovine serum albumin (A-7906), 2-(N-morpholino)ethanesulfonic acid (MES) (>99.5%), 2-(cyclohexylamino)ethanesulfonic acid (CHES) (>99%), and sodium dodecyl sulfate were purchased from Sigma–Aldrich, mesityl oxide (4-methyl-3-penten-2-one) (90%) was purchased from Lancaster (Morecambe, U.K.), and m-nitrophenol, boric acid and sodium hydroxide (analytical reagent grade) were obtained from Merck. The buffer systems were prepared from the corresponding acid followed by addition of concentrated sodium hydroxide to the aimed pH. All solutions were prepared using Milli-Q water from a Millipore system (Billerica, MA, USA). 2.4. Trypsin digestion of protein A solution of bovine serum albumin (10 mg/ml) in 100 mM boric acid buffer (pH 8.40) was thermally denatured in a water bath at 80 ◦ C for 60 min. Then, a solution of trypsin (1.0 mg/ml) in 100 mM boric acid buffer (pH 8.40) was added at a 50:1 ratio (protein/trypsin) by weight and the mixture was kept at 37 ◦ C in a water bath for 24 h with gentle stirring. The solution was then divided into smaller fractions and stored frozen (−20 ◦ C). Prior to use, the sample was diluted five times in 100 mM boric acid buffer (pH 8.40). 2.5. Separation procedure If not otherwise stated, sample injection and initial separation in the first capillary were performed by applying +10 kV to reservoir 1 with reservoir 2 grounded. Transfer of a fraction from the first to the second separation capillary was performed by applying +10 kV to reservoir 1 with reservoir 4 grounded. The subsequent second separation was then performed by changing the location of the high voltage from reservoir 1 to reservoir 3. The system was operated in a comprehensive mode by repeating a cycle consisting of transfer of analytes from the first capillary to the second capillary and a subsequent separation in the second capillary. 3. Results and discussion 3.1. Fabrication of capillary connections and capillary systems The capillary connections were made from heat/shrink melt tubing that consists of an outer layer of poly(tetrafluoroethylene)
(PTFE) layer and an inner layer of fluorinated ethylene propylene (FEP). When heated (>350 ◦ C) the outer PTFE layer shrinks and the inner FEP layer melts filling all empty space inside the tubing. The connections were fabricated by aligning a piece of heat/shrink melt tubing, two fused silica capillaries (25 cm × 75 m i.d.), two fused silica capillaries (5 cm × 75 m i.d.) and two tungsten wires (50 m diameter) as shown in Fig. 2a. This was achieved by threading a long (25 cm) and a short (5 cm) fused silica capillary on each wire with a small gap between the two capillaries. The two tungsten wires were then aligned inside the heat/shrink melt tubing so that the two wires crossed each other at the gap between the capillaries (see Fig. 2a). The tubing was then heated with a heat gun to achieve simultaneous shrinking and melting of the tubing. In order to achieve contact between the two wires a tapered pair of tweezers was used to press the tubing together at the point where the wires were crossing each other during the heating process. After cooling to room temperature, the pair of tweezers was removed, and the tungsten wires were removed by pulling them out. In this way, a channel connection consisting of two channels with circular cross-sections that were tangentially in contact with each other was created (see Fig. 2b). The use of heat/shrink melt tubing to create other type of simple microchannel structures without microfabrication technology has been described before [19]. The fused silica capillaries and the FEP connection part were glued onto a Plexiglas plate in order to stabilize the construction. In order to facilitate fluid connection of the two short fused silica capillaries with the reservoirs these capillaries were connected to PEEK tubing with a relatively large i.d. (0.8 mm) (see Fig. 1A). PEEK tubing was chosen since it is flexible and available with large i.d.’s. In this way, the high voltage will mainly be applied over the two long fused silica capillaries where the first and the second separation take place, respectively. Note that the two separation capillaries can be filled with different separation media by simply filling reservoirs 1 and 3 with different media and by applying a gas pressure to the two reservoirs. 3.2. Operation and characterization of the system When running the system comprehensively it is essential that the solution in the first capillary is kept motionless while the second separation is going on. Movement of species can occur either by migration in an electric field where the electric field originates from leakage of current in the high voltage switch, or by movement of the solution due to pressure differences in the system. Movement of solution in the first capillary while the
E. Sahlin / J. Chromatogr. A 1154 (2007) 454–459
Fig. 3. Separation of 100 M mesityl oxide (A) and 100 M 2-chlorobenzoate (B) in 20 mM CHES, pH 9.30 using a separation voltage of +10 kV applied between reservoirs 1 and 4. (a–c) After an initial separation for 30 s the separation was paused for 30 min while: (a and c) +10 kV was applied between reservoirs 3 and 4, (b) no voltage was applied anywhere in the system; (d) the separation was not paused. (a and b) All reservoirs were at the same height; (c and d) reservoir 1 was located 12 mm below reservoirs 2, 3 and 4. Injection was performed by applying +10 kV between reservoirs 1 and 2 for 2 s and detection was performed at 225 nm.
second separation was running was studied by filling the system including all the reservoirs with 20 mM CHES, pH 9.30. Injection of a sample containing 100 M of mesityl oxide and 2-chlorobenzoate was performed by applying +10 kV to reservoir 1 with reservoir 2 grounded for 2 s. A separation was then performed by applying +10 kV to reservoir 1 with reservoir 4 grounded. However, after 30 s this separation was paused for 30 min. During this time, +10 kV was applied to reservoir 3 with reservoir 4 grounded in one experiment, or no high voltage was applied to the system in another experiment. The corresponding electropherograms obtained when the original separation was continued are shown in Fig. 3a and b, respectively. The differences in migration times between the two species are given in the figure together with the migration times for mesityl oxide (neutral). Clearly, both species (mesityl oxide and 2-chlorobenzoate) have moved forward during the 30 min period when +10 kV was applied between reservoirs 3 and 4. Since the difference in migration times between the two species is smaller in the presence of a high voltage between reservoirs 3 and 4, it can be concluded that the two species did not separate from each other during the movement. Furthermore, the ratio between the differences in migration times of mesityl oxide and 2-chlorobenzoate
457
(2.16/1.27 = 1.70) equals the ratio between the migration times of mesityl oxide (3.59/2.08 = 1.73). Hence, the transport of the two species in the first capillary is due to a pressure difference in the system generated by the voltage applied between reservoirs 3 and 4. The pressure difference is a result of different capillary materials (PEEK and fused silica) present in the system. This pressure difference was eliminated by lowering reservoir 1 a distance of 12 mm relative to the other three reservoirs. In this way a hydrostatic pressure was introduced that counteracted the pressure difference originating from the second separation. When repeating the experiment where +10 kV was applied between reservoirs 3 and 4 during the 30 min break, no movement of solutes had occurred as demonstrated in Fig. 3c. Fig. 3d shows the electropherogram obtained when the separation was not paused at all. The above experiment was repeated with five different capillary systems and in all five systems the described pressure difference was present and it was found necessary to lower reservoir 1 relative to the other reservoirs 12, 12, 8, 9 and 12 mm, respectively. This further demonstrates that the capillary systems can be fabricated with an acceptable reproducibility. In order to study the recovery of analyte transfer between the two separation capillaries, two different types of experiments were performed six times in 20 mM CHES, pH 9.30. In both types of experiments, a sample containing 100 M mesityl oxide and 2-chlorobenzoate was injected at +10 kV for 30 s. In one of the experiments, the system was operated in a comprehensive mode using a voltage of +10 kV, with a transferring time of 5.0 s, and allowing the subsequent separation to continue for 80 s. In the other experiment, a single separation was performed by applying +10 kV to reservoir 1 with reservoir 4 grounded. In the comprehensive mode experiment, the two analytes appeared as peaks in eight or nine subsequent electropherograms and a total peak area for each analyte could be obtained by adding the peak areas in these electropherograms. In order to compare these areas with the corresponding areas obtained in the single separation experiment, it is necessary to correct for the different electric field strengths used in the two types of experiments. The comparison of the mean values from both types of separation experiments is given in Table 1 where the corrected mean peak area (mean peak area × electric field strength) is given for each analyte for both types of experiments together with the corresponding corrected standard deviations (standard deviation × electric field strength). By comparing the corrected mean peak areas for each analyte it can be concluded that there is
Table 1 Recovery of capillary-to-capillary analyte transfer Mesityl oxide
Corrected mean peak areaa (a.u.) Corrected standard deviationb (a.u.)
2-Chlorobenzoate
Comprehensive mode experiment (n = 6)
Single separation experiment (n = 6)
Comprehensive mode experiment (n = 6)
Single separation experiment (n = 6)
18.79 0.735
19.46 0.320
11.92 0.651
12.86 0.279
a.u., arbitrary unit. a Corrected mean peak area = mean peak area × electric field strength. b Corrected standard deviation = standard deviation × electric field strength.
458
E. Sahlin / J. Chromatogr. A 1154 (2007) 454–459
no significant difference in the amount of analyte that has passed the detector in the two types of experiments (t = 0.68 for mesityl oxide, and t = 1.09 for 2-chlorobenzoate, where the critical value of t = 2.23 for both compounds at P = 0.05). Hence, running the system in a comprehensive mode does not result in any loss of analytes. 3.3. Two-dimensional capillary zone electrophoresis separation using two different pH By filling the channels with different solutions, it is possible to perform capillary electrophoresis separations under different conditions in the two channels. Reservoirs 1 and 2 were filled with 20 mM MES, pH 6.10, and reservoirs 3 and 4, were filled with 50 mM CHES, pH 9.30. A sample containing 100 M mesityl oxide and m-nitrophenol in 20 mM MES, pH 6.10, was injected into the system for 5 s. Following a short separation of the sample components, in the first capillary (by applying +10 kV to reservoir 1 with reservoir 2 grounded), the system was operated comprehensively with a capillary-to-capillary transfer time of 2 s and with the second separation going on for 80 s. m-Nitrophenol has pKa equal to 8.34 [20], i.e. m-nitrophenol is neutral at pH 6.10 and carries a net charge of −0.90 at pH 9.30. Thus, in the first separation media both analytes are neutral and are not separated from each other, while in the second separation media mesityl oxide is neutral and m-nitrophenol carries a charge of −0.90. The corresponding 2D electropherogram is shown in Fig. 4. Clearly, mesityl oxide (neutral) and m-nitrophenol (negatively charged) are separated at the pH used in the second capillary. In addition, m-nitrophenol has been concentrated when transferred from the first capillary to the second capillary due to a dynamic pH junction effect [21,22] further demonstrating that the compound has experienced a change in pH. This also demonstrates how a post-separation focusing step can be incorporated into a capillary electrophoresis separation in order to enhance the sensitivity and to improve the resolution.
Fig. 4. Two-dimensional separation of 500 M mesityl oxide (A) and 50 M m-nitrophenol (B) using zone electrophoresis at two different pH. The first separation was performed at pH 6.10 (20 mM MES), and the second separation was performed at pH 9.30 (50 mM CHES). Electrokinetical injection and initial separation were performed for 5 and 95 s, respectively, by applying +10 kV between reservoirs 1 and 2. A cycle consisting of +10 kV between reservoirs 1 and 4 for 2 s (transfer of analyte) followed by +10 kV between reservoirs 3 and 4 for 80 s (separation in the second dimension) was then repeatedly executed. Detection was performed at 225 nm.
The above experiment also demonstrates a seldom-used way of performing a type of two-dimensional separation although it has been applied successfully to protein separations [11]. However, zone electrophoresis separations at different pH are not fully orthogonal to each other. 3.4. Comprehensive two-dimensional capillary electrophoresis of tryptic digest of a protein Capillary zone electrophoresis and micellar electrokinetic chromatography are two separations mechanisms with high orthogonality [12]. Here, a two-dimensional capillary electrophoresis system, consisting of capillary zone electrophoresis (in 100 mM borate buffer, pH 8.40) in the first separation and micellar electrokinetic chromatography (in 20 mM sodium dodecyl sulfate, 100 mM borate buffer, pH 8.40) in the second separation, was used to separate a tryptic digest of bovine serum albumin (1.7 mg/ml). The resulting two-dimensional electro-
Fig. 5. Two-dimensional capillary electrophoresis separation of tryptic digest of bovine serum albumin (1.7 mg/ml). The first separation (zone electrophoresis) was performed in 100 mM borate buffer, pH 8.40, and the second separation (micellar electrokinetic chromatography) was performed in 20 mM sodium dodecyl sulfate, 100 mM borate buffer, pH 8.40. Electrokinetic injection and initial separation were performed for 3 and 85 s, respectively, by applying +10 kV between reservoirs 1 and 2. A cycle consisting of applying +10 kV between reservoirs 1 and 4 for 5 s (transfer of analyte) followed by applying +10 kV between reservoirs 3 and 4 for 90 s (separation in the second dimension) was then repeatedly executed. Detection was performed at 210 nm. Migration time repeatability for peaks marked with (A), (B) and (C) is given in the text.
E. Sahlin / J. Chromatogr. A 1154 (2007) 454–459
459
Table 2 Repeatability data for peaks marked A, B and C in Fig. 5 in seven consecutive separations Peak
Fraction in which peak mainly appeared
Mean migration time in second separation (n = 7) (min)
Relative standard deviation of migration time in second separation (n = 7) (%)
A B C
7, 6, 6, 6, 6, 7, 6 11, 11, 10, 10, 11, 12, 11 20, 20, 20, 20, 21, 23, 21
0.65 0.80 0.95
4.0 3.7 1.7
pherogram is shown in Fig. 5. The two-dimensional separation was completed in 55 min. Though not as impressive as some other two-dimensional separations with laser induced fluorescence detection [12,15] the electropherogram shows numerous resolved peaks due to the presence of a second separation dimension and the overall resolution is greatly improved. With more sensitive detection techniques even more peaks will appear. Peaks marked with (A), (B) and (C) (in the 6th, 10th and 20th fraction) were used to study the migration time repeatability and the resulting data are given in Table 2. Clearly, the system could be operated with an acceptable repeatability. 4. Summary and conclusions The described capillary system has a large potential to be used in different types of capillary electrophoresis applications. Here, it has been demonstrated that the system can be used to perform comprehensive two-dimensional capillary electrophoresis separations. The maximum peak capacity, PC2D , is given by [2,6,23,24]: PC2D = PC1 × PC2
(1)
where PC1 and PC2 are the peak capacities in the individual separations, i.e. a two-dimensional system is superior to a onedimensional system when analyzing complex samples. Another advantage with two-dimensional systems is the reduced probability of component overlap and improved peak identification capabilities since the exact position of a compound in a twodimensional electropherogram is dependent on two different separation mechanisms. The use of fused silica capillaries in the system is attractive for several reasons. Different types of coated fused silica capillaries are commercially available and can be used to eliminate or reduce surface interactions between solutes and capillary wall surfaces. Fused silica capillaries can also easily be connected to mass spectrometers using commercially available interfaces. In addition, the interfaces and the capillary systems can be fabricated reproducibly with simple equipment without microfabrication technology. The described system will also be a versatile tool for manipulation of individual compounds present in small complex samples. Possible manipulations comprise concentrating an isolated compound, chemical modification (post-separation derivatization), binding studies (by performing affinity capillary electrophoresis in the second separation), separation of enantiomeric isomers isolated from a complex sample, or clean-up
of media (for instance by decreasing the electrolyte or surfactant concentration) prior to nanospray mass spectrometry detection. These applications will be explored in the near future. Acknowledgement The Swedish Research Council is acknowledged for financial support. References [1] H.J. Issaq, Electrophoresis 22 (2001) 3629. [2] T.F. Hooker, D.J. Jefferey, J.W. Jorgenson, in: M.G. Khaledi (Ed.), High Performance Capillary Electrophoresis, Wiley, New York, 1998, p. 581. [3] H.J. Issaq, K.C. Chan, G.M. Janini, T.P. Conrads, T.D. Veenstra, J. Chromatogr. B 817 (2005) 35. [4] H.J. Issaq, T.P. Conrads, G.M. Janini, T.D. Veenstra, Electrophoresis 23 (2002) 3048. [5] K.S. Lilley, A. Razzaq, P. Dupree, Curr. Opin. Chem. Biol. 6 (2002) 46. [6] C.R. Evans, J.W. Jorgenson, Anal. Bioanal. Chem. 378 (2004) 1952. [7] T. Stroink, M.C. Ortiz, A. Bult, H. Lingeman, G.J. de Jong, W.J.M. Underberg, J. Chromatogr. B 817 (2005) 49. [8] J.W. Cooper, Y.J. Wang, C.S. Lee, Electrophoresis 25 (2004) 3913. [9] S. Hu, D.A. Michels, M.A. Fazal, C. Ratisoontorn, M.L. Cunningham, N.J. Dovichi, Anal. Chem. 76 (2004) 4044. [10] N. Gottschlich, S.C. Jacobson, C.T. Culbertson, J.M. Ramsey, Anal. Chem. 73 (2001) 2669. [11] D.A. Michels, S. Hu, R.M. Schoenherr, M.J. Eggertson, N.J. Dovichi, Mol. Cell. Proteomics 1 (2002) 69. [12] J.D. Ramsey, S.C. Jacobson, C.T. Culbertson, J.M. Ramsey, Anal. Chem. 75 (2003) 3758. [13] R.D. Rocklin, R.S. Ramsey, J.M. Ramsey, Anal. Chem. 72 (2000) 5244. [14] M.A. Fazal, V.R. Palmer, N.J. Dovichi, J. Chromatogr. A 1130 (2006) 182. [15] J.R. Kraly, M.R. Jones, D.G. Gomez, J.A. Dickerson, M.M. Harwood, M. Eggertson, T.G. Paulson, C.A. Sanchez, R. Odze, Z.D. Feng, B.J. Reid, N.J. Dovichi, Anal. Chem. 78 (2006) 5977. [16] H.C. Liu, L.H. Zhang, G.J. Zhu, W.B. Zhang, Y.K. Zhang, Anal. Chem. 76 (2004) 6506. [17] D.A. Michels, S. Hu, K.A. Dambrowitz, M.J. Eggertson, K. Lauterbach, N.J. Dovichi, Electrophoresis 25 (2004) 3098. [18] H. Shadpour, S.A. Soper, Anal. Chem. 78 (2006) 3519. [19] E. Sahlin, A.T. Beisler, S.J. Woltman, S.G. Weber, Anal. Chem. 74 (2002) 4566. [20] S. Budavari (Ed.), The Merck Index, 13th ed., Merck & Co., Whitehouse Station, NJ, 2001. [21] J.B. Kim, P. Britz-McKibbin, T. Hirokawa, S. Terabe, Anal. Chem. 75 (2003) 3986. [22] C.H. Lin, T. Kaneta, Electrophoresis 25 (2004) 4058. [23] J.C. Giddings, Anal. Chem. 56 (1984) 1258. [24] J.C. Giddings, J. High Resolut. Chromatogr. Chromatogr. Commun. 10 (1987) 319.
Two-dimensional capillary electrophoresis using tangentially connected capillaries Eskil Sahlin ∗ Department of Chemistry, G¨oteborg University, 412 96 G¨oteborg, Sweden Received 26 January 2007; received in revised form 29 March 2007; accepted 2 April 2007 Available online 6 April 2007
Abstract A novel type of fused silica capillary system is described where channels with circular cross-sections are tangentially in contact with each other and connected through a small opening at the contact area. Since the channels are not crossing each other in the same plane, the capillaries can easily be filled with different solutions, i.e. different solutions will be in contact with each other at the contact point. The system has been used to perform different types of two-dimensional separations and the complete system is fully automated where a high voltage switch is used to control the location of the high voltage in the system. Using two model compounds it is demonstrated that a type of two-dimensional separation can be performed using capillary zone electrophoresis at two different pH values. It is also shown that a compound with acid/base properties can be concentrated using a dynamic pH junction mechanism when transferred from the first separation to the second separation. In addition, the system has been used to perform a comprehensive two-dimensional capillary electrophoresis separation of tryptic digest of bovine serum albumin using capillary zone electrophoresis followed by micellar electrokinetic chromatography. © 2007 Elsevier B.V. All rights reserved. Keywords: Capillary electrophoresis; Tangentially connected capillaries; Two-dimensional separation
1. Introduction In order to increase the knowledge of life and our environment, it is essential to be able to identify and quantify compounds in clinical and environmental samples. However, these samples can be extremely complex when considering low concentrations. Body fluids, cells and some environmental samples such as urban dust are important examples of samples that contain or can contain a vast number of compounds in a large concentration range. For instance, in a human cell there are in the order of 104 proteins which are expressed in a particular type of cell at any time at different levels of concentration with a dynamic range of five-fold [1]. Tryptic digests of cells could contain as much as a million peptides [1]. Hence, in order to analyze or to perform studies on such samples, analytical techniques with extremely high selectivity are needed and a separation step is
∗ Present address: Chemistry and Materials Technology, SP Technical Research Institute of Sweden, Brinellgatan 4, 501 15 Bor˚as, Sweden. Tel.: +46 10 516 5265; fax: +46 33 12 37 49. E-mail address: [email protected].
0021-9673/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2007.04.001
often included. However, even for the most efficient separation techniques, efficiencies and peak capacities are not high enough for many complex samples [1–3] and two-dimensional separation systems have therefore been developed. Proteins are commonly separated using two-dimensional polyacrylamide gel electrophoresis (2D-PAGE) [1,4,5]. In addition, several on-line two-dimensional separation systems (often LC–LC and LC–CE) have been described in the literature [2–4,6–8]. Most of them utilize liquid chromatography in the first separation step, however, this separation is not easily stopped and is therefore often kept running while the second separation takes place. Hence, the second separation must be relatively fast when the two-dimensional system is operated comprehensively. This problem is avoided when capillary electrophoresis is used in both separation steps, though a fast second separation step is still desirable in order to have reasonable total analysis times when operated comprehensively. Two-dimensional capillary electrophoresis systems have been described several times in the literature using systems based on either fused silica capillaries or microfabricated microfluidic systems [3,7–18], and put into practice especially by the Dovichi group [9,11,14,15,17] and the Ramsey group [10,12,13]. An important issue with
E. Sahlin / J. Chromatogr. A 1154 (2007) 454–459
two-dimensional capillary electrophoresis systems is the design of the capillary–capillary interface where analytes are transferred from the first to the second separation. The Dovichi group has constructed a system with two fused silica capillaries aligned in series and where separation media to the second capillary enters the system at a small gap between the capillaries [9,11,14,15,17]. Two-dimensional separations have also been performed in microfabricated systems where the different channels are crossing each other in the same plane [10,12,13,18]. Transport of analytes and solution through the crossing is performed by controlling the voltage in several reservoirs simultaneously. An additional capillary–capillary interface includes an etched porous junction [16]. Another issue to consider is the choice of separation modes in the first and the second dimension for various analytes. Two-dimensional separations of proteins with high orthogonality have been demonstrated using capillary sieving electrophoresis–micellar electrokinetic chromatography [9,14,15,17], submicellar capillary electrophoresis–capillary zone electrophoresis (at two different pH) [11], capillary isoelectric focusing–capillary zone electrophoresis [16] and capillary gel electrophoresis–micellar electrokinetic chromatography [18]. Peptides have been separated using micellar electrokinetic chromatography–capillary zone electrophoresis [12,13], and (open-channel) electrochromatography–capillary zone electrophoresis [10]. In this paper, a novel low dead-volume capillary–capillary interface is fabricated without microfabrication technology that allows straightforward filling of the systems with different solutions in the two different capillaries. The capillary system is characterized and used to perform two-dimensional capillary electrophoresis separations in a comprehensive mode. Several other possible applications of the system are also discussed briefly. 2. Materials and methods 2.1. Materials and tools Fused silica capillaries (75 m i.d., 360 m o.d.) and polyether ether ketone (PEEK) tubing (0.8 mm i.d.) were purchased from Upchurch Scientific (Oak Harbor, WA, USA), and heat shrink/melt tubing with an inner diameter of 1.65 mm was obtained from Cole Parmer (Vernon Hills, IL, USA). Tungsten wire with a diameter of 50 m and platinum rods with a diameter of 1.0 mm were obtained from Goodfellow (Huntingdon, UK). Heating of the heat/shrink melt tubing was obtained with a heat gun with a variable temperature range of 60–590 ◦ C. The tungsten wires were held in place during heating with a tapered pair of tweezers. 2.2. Instrumentation The capillary electrophoresis system is shown in Fig. 1. The system consisted of four reservoirs (marked A and numbered 1–4) that were connected to each other using a capillary connection (marked B) with four arms. The capillary connection is further described below. Two of the arms consisted
455
Fig. 1. A four armed capillary electrophoresis system: (A) reservoirs, (B) capillary connection, (C) fused silica capillaries (25 cm × 75 m i.d.), (D) fused silica capillaries (5 cm × 75 m i.d.), (E) PEEK tubing (15 cm × 0.8 mm i.d.) and (F) detector. The four arms are joined together in a FEP channel connection (marked B) where the two channels are tangentially in contact with each other. The detector is located 6.5 cm from the channel connection.
of 25 cm × 75 m i.d. fused silica capillaries (marked C) and ended in reservoirs 1 and 4. The other two arms consisted of 5 cm × 75 m i.d. fused silica capillaries (marked D) connected to 15 cm × 0.8 mm i.d. PEEK tubing (marked E) and ended in reservoirs 2 and 3. A UV/vis absorbance detector (Spectra 100 UV/vis detector, Thermo Separation Products, Fremont, CA, USA) (marked F) was used for detection and was located 6.5 cm from the capillary connection. The system was housed in a Plexiglas box with holders for the reservoirs and capillaries. The reservoirs consisted of plastic tubes that were filled with 1.30 ml of separation media. Reservoirs 1 and 3 could be sealed and pressurized using nitrogen gas (2 bar) facilitating filling and flushing of system with separation media. Platinum electrodes (1.0 mm diameter) were inserted into the reservoirs and connected to a high voltage power supply (Model 2462, Bertan Associates, Hauppauge, NY, USA). A high voltage switching relay (Model G2 (powered by 12 V dc), Gigavac, Santa Barbara, CA, USA) was used to change the location of the high voltage in the systems (i.e. two reservoirs were always floating). The high voltage switching relay and the high voltage power supply were controlled with timers (Model H5CX, Omron Electronics, Schaumburg, IL, USA) so that the system could be operated without manual assistance when the sample had been injected. The timers and the switching relay were powered by an isolated dc power source consisting of a rechargeable 12 V battery. Detection was performed at 210 or 225 nm and the signal was filtered by setting the rise time of the detector to 0.1 or 1.0 s. Data were collected at 10, 30 or 50 Hz using a multifunction I/O board (PCI-6034E from National Instruments, Austin, TX, USA) controlled with LabVIEW 6.0 (National Instruments).
456
E. Sahlin / J. Chromatogr. A 1154 (2007) 454–459
Fig. 2. (a) Initial alignment of: (A) heat/shrink melt tubing, (B) fused silica capillaries (25 cm × 75 m i.d.), (C) fused silica capillaries (5 cm × 75 m i.d.) and (D) tungsten wires. (b) Schematic drawing of the tangentially connected channels.
2.3. Chemicals 2-Chlorobenzoic acid (98%), trypsin (T-1005), bovine serum albumin (A-7906), 2-(N-morpholino)ethanesulfonic acid (MES) (>99.5%), 2-(cyclohexylamino)ethanesulfonic acid (CHES) (>99%), and sodium dodecyl sulfate were purchased from Sigma–Aldrich, mesityl oxide (4-methyl-3-penten-2-one) (90%) was purchased from Lancaster (Morecambe, U.K.), and m-nitrophenol, boric acid and sodium hydroxide (analytical reagent grade) were obtained from Merck. The buffer systems were prepared from the corresponding acid followed by addition of concentrated sodium hydroxide to the aimed pH. All solutions were prepared using Milli-Q water from a Millipore system (Billerica, MA, USA). 2.4. Trypsin digestion of protein A solution of bovine serum albumin (10 mg/ml) in 100 mM boric acid buffer (pH 8.40) was thermally denatured in a water bath at 80 ◦ C for 60 min. Then, a solution of trypsin (1.0 mg/ml) in 100 mM boric acid buffer (pH 8.40) was added at a 50:1 ratio (protein/trypsin) by weight and the mixture was kept at 37 ◦ C in a water bath for 24 h with gentle stirring. The solution was then divided into smaller fractions and stored frozen (−20 ◦ C). Prior to use, the sample was diluted five times in 100 mM boric acid buffer (pH 8.40). 2.5. Separation procedure If not otherwise stated, sample injection and initial separation in the first capillary were performed by applying +10 kV to reservoir 1 with reservoir 2 grounded. Transfer of a fraction from the first to the second separation capillary was performed by applying +10 kV to reservoir 1 with reservoir 4 grounded. The subsequent second separation was then performed by changing the location of the high voltage from reservoir 1 to reservoir 3. The system was operated in a comprehensive mode by repeating a cycle consisting of transfer of analytes from the first capillary to the second capillary and a subsequent separation in the second capillary. 3. Results and discussion 3.1. Fabrication of capillary connections and capillary systems The capillary connections were made from heat/shrink melt tubing that consists of an outer layer of poly(tetrafluoroethylene)
(PTFE) layer and an inner layer of fluorinated ethylene propylene (FEP). When heated (>350 ◦ C) the outer PTFE layer shrinks and the inner FEP layer melts filling all empty space inside the tubing. The connections were fabricated by aligning a piece of heat/shrink melt tubing, two fused silica capillaries (25 cm × 75 m i.d.), two fused silica capillaries (5 cm × 75 m i.d.) and two tungsten wires (50 m diameter) as shown in Fig. 2a. This was achieved by threading a long (25 cm) and a short (5 cm) fused silica capillary on each wire with a small gap between the two capillaries. The two tungsten wires were then aligned inside the heat/shrink melt tubing so that the two wires crossed each other at the gap between the capillaries (see Fig. 2a). The tubing was then heated with a heat gun to achieve simultaneous shrinking and melting of the tubing. In order to achieve contact between the two wires a tapered pair of tweezers was used to press the tubing together at the point where the wires were crossing each other during the heating process. After cooling to room temperature, the pair of tweezers was removed, and the tungsten wires were removed by pulling them out. In this way, a channel connection consisting of two channels with circular cross-sections that were tangentially in contact with each other was created (see Fig. 2b). The use of heat/shrink melt tubing to create other type of simple microchannel structures without microfabrication technology has been described before [19]. The fused silica capillaries and the FEP connection part were glued onto a Plexiglas plate in order to stabilize the construction. In order to facilitate fluid connection of the two short fused silica capillaries with the reservoirs these capillaries were connected to PEEK tubing with a relatively large i.d. (0.8 mm) (see Fig. 1A). PEEK tubing was chosen since it is flexible and available with large i.d.’s. In this way, the high voltage will mainly be applied over the two long fused silica capillaries where the first and the second separation take place, respectively. Note that the two separation capillaries can be filled with different separation media by simply filling reservoirs 1 and 3 with different media and by applying a gas pressure to the two reservoirs. 3.2. Operation and characterization of the system When running the system comprehensively it is essential that the solution in the first capillary is kept motionless while the second separation is going on. Movement of species can occur either by migration in an electric field where the electric field originates from leakage of current in the high voltage switch, or by movement of the solution due to pressure differences in the system. Movement of solution in the first capillary while the
E. Sahlin / J. Chromatogr. A 1154 (2007) 454–459
Fig. 3. Separation of 100 M mesityl oxide (A) and 100 M 2-chlorobenzoate (B) in 20 mM CHES, pH 9.30 using a separation voltage of +10 kV applied between reservoirs 1 and 4. (a–c) After an initial separation for 30 s the separation was paused for 30 min while: (a and c) +10 kV was applied between reservoirs 3 and 4, (b) no voltage was applied anywhere in the system; (d) the separation was not paused. (a and b) All reservoirs were at the same height; (c and d) reservoir 1 was located 12 mm below reservoirs 2, 3 and 4. Injection was performed by applying +10 kV between reservoirs 1 and 2 for 2 s and detection was performed at 225 nm.
second separation was running was studied by filling the system including all the reservoirs with 20 mM CHES, pH 9.30. Injection of a sample containing 100 M of mesityl oxide and 2-chlorobenzoate was performed by applying +10 kV to reservoir 1 with reservoir 2 grounded for 2 s. A separation was then performed by applying +10 kV to reservoir 1 with reservoir 4 grounded. However, after 30 s this separation was paused for 30 min. During this time, +10 kV was applied to reservoir 3 with reservoir 4 grounded in one experiment, or no high voltage was applied to the system in another experiment. The corresponding electropherograms obtained when the original separation was continued are shown in Fig. 3a and b, respectively. The differences in migration times between the two species are given in the figure together with the migration times for mesityl oxide (neutral). Clearly, both species (mesityl oxide and 2-chlorobenzoate) have moved forward during the 30 min period when +10 kV was applied between reservoirs 3 and 4. Since the difference in migration times between the two species is smaller in the presence of a high voltage between reservoirs 3 and 4, it can be concluded that the two species did not separate from each other during the movement. Furthermore, the ratio between the differences in migration times of mesityl oxide and 2-chlorobenzoate
457
(2.16/1.27 = 1.70) equals the ratio between the migration times of mesityl oxide (3.59/2.08 = 1.73). Hence, the transport of the two species in the first capillary is due to a pressure difference in the system generated by the voltage applied between reservoirs 3 and 4. The pressure difference is a result of different capillary materials (PEEK and fused silica) present in the system. This pressure difference was eliminated by lowering reservoir 1 a distance of 12 mm relative to the other three reservoirs. In this way a hydrostatic pressure was introduced that counteracted the pressure difference originating from the second separation. When repeating the experiment where +10 kV was applied between reservoirs 3 and 4 during the 30 min break, no movement of solutes had occurred as demonstrated in Fig. 3c. Fig. 3d shows the electropherogram obtained when the separation was not paused at all. The above experiment was repeated with five different capillary systems and in all five systems the described pressure difference was present and it was found necessary to lower reservoir 1 relative to the other reservoirs 12, 12, 8, 9 and 12 mm, respectively. This further demonstrates that the capillary systems can be fabricated with an acceptable reproducibility. In order to study the recovery of analyte transfer between the two separation capillaries, two different types of experiments were performed six times in 20 mM CHES, pH 9.30. In both types of experiments, a sample containing 100 M mesityl oxide and 2-chlorobenzoate was injected at +10 kV for 30 s. In one of the experiments, the system was operated in a comprehensive mode using a voltage of +10 kV, with a transferring time of 5.0 s, and allowing the subsequent separation to continue for 80 s. In the other experiment, a single separation was performed by applying +10 kV to reservoir 1 with reservoir 4 grounded. In the comprehensive mode experiment, the two analytes appeared as peaks in eight or nine subsequent electropherograms and a total peak area for each analyte could be obtained by adding the peak areas in these electropherograms. In order to compare these areas with the corresponding areas obtained in the single separation experiment, it is necessary to correct for the different electric field strengths used in the two types of experiments. The comparison of the mean values from both types of separation experiments is given in Table 1 where the corrected mean peak area (mean peak area × electric field strength) is given for each analyte for both types of experiments together with the corresponding corrected standard deviations (standard deviation × electric field strength). By comparing the corrected mean peak areas for each analyte it can be concluded that there is
Table 1 Recovery of capillary-to-capillary analyte transfer Mesityl oxide
Corrected mean peak areaa (a.u.) Corrected standard deviationb (a.u.)
2-Chlorobenzoate
Comprehensive mode experiment (n = 6)
Single separation experiment (n = 6)
Comprehensive mode experiment (n = 6)
Single separation experiment (n = 6)
18.79 0.735
19.46 0.320
11.92 0.651
12.86 0.279
a.u., arbitrary unit. a Corrected mean peak area = mean peak area × electric field strength. b Corrected standard deviation = standard deviation × electric field strength.
458
E. Sahlin / J. Chromatogr. A 1154 (2007) 454–459
no significant difference in the amount of analyte that has passed the detector in the two types of experiments (t = 0.68 for mesityl oxide, and t = 1.09 for 2-chlorobenzoate, where the critical value of t = 2.23 for both compounds at P = 0.05). Hence, running the system in a comprehensive mode does not result in any loss of analytes. 3.3. Two-dimensional capillary zone electrophoresis separation using two different pH By filling the channels with different solutions, it is possible to perform capillary electrophoresis separations under different conditions in the two channels. Reservoirs 1 and 2 were filled with 20 mM MES, pH 6.10, and reservoirs 3 and 4, were filled with 50 mM CHES, pH 9.30. A sample containing 100 M mesityl oxide and m-nitrophenol in 20 mM MES, pH 6.10, was injected into the system for 5 s. Following a short separation of the sample components, in the first capillary (by applying +10 kV to reservoir 1 with reservoir 2 grounded), the system was operated comprehensively with a capillary-to-capillary transfer time of 2 s and with the second separation going on for 80 s. m-Nitrophenol has pKa equal to 8.34 [20], i.e. m-nitrophenol is neutral at pH 6.10 and carries a net charge of −0.90 at pH 9.30. Thus, in the first separation media both analytes are neutral and are not separated from each other, while in the second separation media mesityl oxide is neutral and m-nitrophenol carries a charge of −0.90. The corresponding 2D electropherogram is shown in Fig. 4. Clearly, mesityl oxide (neutral) and m-nitrophenol (negatively charged) are separated at the pH used in the second capillary. In addition, m-nitrophenol has been concentrated when transferred from the first capillary to the second capillary due to a dynamic pH junction effect [21,22] further demonstrating that the compound has experienced a change in pH. This also demonstrates how a post-separation focusing step can be incorporated into a capillary electrophoresis separation in order to enhance the sensitivity and to improve the resolution.
Fig. 4. Two-dimensional separation of 500 M mesityl oxide (A) and 50 M m-nitrophenol (B) using zone electrophoresis at two different pH. The first separation was performed at pH 6.10 (20 mM MES), and the second separation was performed at pH 9.30 (50 mM CHES). Electrokinetical injection and initial separation were performed for 5 and 95 s, respectively, by applying +10 kV between reservoirs 1 and 2. A cycle consisting of +10 kV between reservoirs 1 and 4 for 2 s (transfer of analyte) followed by +10 kV between reservoirs 3 and 4 for 80 s (separation in the second dimension) was then repeatedly executed. Detection was performed at 225 nm.
The above experiment also demonstrates a seldom-used way of performing a type of two-dimensional separation although it has been applied successfully to protein separations [11]. However, zone electrophoresis separations at different pH are not fully orthogonal to each other. 3.4. Comprehensive two-dimensional capillary electrophoresis of tryptic digest of a protein Capillary zone electrophoresis and micellar electrokinetic chromatography are two separations mechanisms with high orthogonality [12]. Here, a two-dimensional capillary electrophoresis system, consisting of capillary zone electrophoresis (in 100 mM borate buffer, pH 8.40) in the first separation and micellar electrokinetic chromatography (in 20 mM sodium dodecyl sulfate, 100 mM borate buffer, pH 8.40) in the second separation, was used to separate a tryptic digest of bovine serum albumin (1.7 mg/ml). The resulting two-dimensional electro-
Fig. 5. Two-dimensional capillary electrophoresis separation of tryptic digest of bovine serum albumin (1.7 mg/ml). The first separation (zone electrophoresis) was performed in 100 mM borate buffer, pH 8.40, and the second separation (micellar electrokinetic chromatography) was performed in 20 mM sodium dodecyl sulfate, 100 mM borate buffer, pH 8.40. Electrokinetic injection and initial separation were performed for 3 and 85 s, respectively, by applying +10 kV between reservoirs 1 and 2. A cycle consisting of applying +10 kV between reservoirs 1 and 4 for 5 s (transfer of analyte) followed by applying +10 kV between reservoirs 3 and 4 for 90 s (separation in the second dimension) was then repeatedly executed. Detection was performed at 210 nm. Migration time repeatability for peaks marked with (A), (B) and (C) is given in the text.
E. Sahlin / J. Chromatogr. A 1154 (2007) 454–459
459
Table 2 Repeatability data for peaks marked A, B and C in Fig. 5 in seven consecutive separations Peak
Fraction in which peak mainly appeared
Mean migration time in second separation (n = 7) (min)
Relative standard deviation of migration time in second separation (n = 7) (%)
A B C
7, 6, 6, 6, 6, 7, 6 11, 11, 10, 10, 11, 12, 11 20, 20, 20, 20, 21, 23, 21
0.65 0.80 0.95
4.0 3.7 1.7
pherogram is shown in Fig. 5. The two-dimensional separation was completed in 55 min. Though not as impressive as some other two-dimensional separations with laser induced fluorescence detection [12,15] the electropherogram shows numerous resolved peaks due to the presence of a second separation dimension and the overall resolution is greatly improved. With more sensitive detection techniques even more peaks will appear. Peaks marked with (A), (B) and (C) (in the 6th, 10th and 20th fraction) were used to study the migration time repeatability and the resulting data are given in Table 2. Clearly, the system could be operated with an acceptable repeatability. 4. Summary and conclusions The described capillary system has a large potential to be used in different types of capillary electrophoresis applications. Here, it has been demonstrated that the system can be used to perform comprehensive two-dimensional capillary electrophoresis separations. The maximum peak capacity, PC2D , is given by [2,6,23,24]: PC2D = PC1 × PC2
(1)
where PC1 and PC2 are the peak capacities in the individual separations, i.e. a two-dimensional system is superior to a onedimensional system when analyzing complex samples. Another advantage with two-dimensional systems is the reduced probability of component overlap and improved peak identification capabilities since the exact position of a compound in a twodimensional electropherogram is dependent on two different separation mechanisms. The use of fused silica capillaries in the system is attractive for several reasons. Different types of coated fused silica capillaries are commercially available and can be used to eliminate or reduce surface interactions between solutes and capillary wall surfaces. Fused silica capillaries can also easily be connected to mass spectrometers using commercially available interfaces. In addition, the interfaces and the capillary systems can be fabricated reproducibly with simple equipment without microfabrication technology. The described system will also be a versatile tool for manipulation of individual compounds present in small complex samples. Possible manipulations comprise concentrating an isolated compound, chemical modification (post-separation derivatization), binding studies (by performing affinity capillary electrophoresis in the second separation), separation of enantiomeric isomers isolated from a complex sample, or clean-up
of media (for instance by decreasing the electrolyte or surfactant concentration) prior to nanospray mass spectrometry detection. These applications will be explored in the near future. Acknowledgement The Swedish Research Council is acknowledged for financial support. References [1] H.J. Issaq, Electrophoresis 22 (2001) 3629. [2] T.F. Hooker, D.J. Jefferey, J.W. Jorgenson, in: M.G. Khaledi (Ed.), High Performance Capillary Electrophoresis, Wiley, New York, 1998, p. 581. [3] H.J. Issaq, K.C. Chan, G.M. Janini, T.P. Conrads, T.D. Veenstra, J. Chromatogr. B 817 (2005) 35. [4] H.J. Issaq, T.P. Conrads, G.M. Janini, T.D. Veenstra, Electrophoresis 23 (2002) 3048. [5] K.S. Lilley, A. Razzaq, P. Dupree, Curr. Opin. Chem. Biol. 6 (2002) 46. [6] C.R. Evans, J.W. Jorgenson, Anal. Bioanal. Chem. 378 (2004) 1952. [7] T. Stroink, M.C. Ortiz, A. Bult, H. Lingeman, G.J. de Jong, W.J.M. Underberg, J. Chromatogr. B 817 (2005) 49. [8] J.W. Cooper, Y.J. Wang, C.S. Lee, Electrophoresis 25 (2004) 3913. [9] S. Hu, D.A. Michels, M.A. Fazal, C. Ratisoontorn, M.L. Cunningham, N.J. Dovichi, Anal. Chem. 76 (2004) 4044. [10] N. Gottschlich, S.C. Jacobson, C.T. Culbertson, J.M. Ramsey, Anal. Chem. 73 (2001) 2669. [11] D.A. Michels, S. Hu, R.M. Schoenherr, M.J. Eggertson, N.J. Dovichi, Mol. Cell. Proteomics 1 (2002) 69. [12] J.D. Ramsey, S.C. Jacobson, C.T. Culbertson, J.M. Ramsey, Anal. Chem. 75 (2003) 3758. [13] R.D. Rocklin, R.S. Ramsey, J.M. Ramsey, Anal. Chem. 72 (2000) 5244. [14] M.A. Fazal, V.R. Palmer, N.J. Dovichi, J. Chromatogr. A 1130 (2006) 182. [15] J.R. Kraly, M.R. Jones, D.G. Gomez, J.A. Dickerson, M.M. Harwood, M. Eggertson, T.G. Paulson, C.A. Sanchez, R. Odze, Z.D. Feng, B.J. Reid, N.J. Dovichi, Anal. Chem. 78 (2006) 5977. [16] H.C. Liu, L.H. Zhang, G.J. Zhu, W.B. Zhang, Y.K. Zhang, Anal. Chem. 76 (2004) 6506. [17] D.A. Michels, S. Hu, K.A. Dambrowitz, M.J. Eggertson, K. Lauterbach, N.J. Dovichi, Electrophoresis 25 (2004) 3098. [18] H. Shadpour, S.A. Soper, Anal. Chem. 78 (2006) 3519. [19] E. Sahlin, A.T. Beisler, S.J. Woltman, S.G. Weber, Anal. Chem. 74 (2002) 4566. [20] S. Budavari (Ed.), The Merck Index, 13th ed., Merck & Co., Whitehouse Station, NJ, 2001. [21] J.B. Kim, P. Britz-McKibbin, T. Hirokawa, S. Terabe, Anal. Chem. 75 (2003) 3986. [22] C.H. Lin, T. Kaneta, Electrophoresis 25 (2004) 4058. [23] J.C. Giddings, Anal. Chem. 56 (1984) 1258. [24] J.C. Giddings, J. High Resolut. Chromatogr. Chromatogr. Commun. 10 (1987) 319.