High Capacity Capillary Electrophoresis-Electrospray Ionization Mass

Oct 28, 2010 - Biomolecular Mass Spectrometry Unit, Department of Parasitology, ...... Albert Bondt , Yoann Rombouts , Dennis Blank , André M. Deelde...
0 downloads 0 Views 997KB Size
Download PDF
Anal. Chem. 2010, 82, 9476–9483

High Capacity Capillary Electrophoresis-Electrospray Ionization Mass Spectrometry: Coupling a Porous Sheathless Interface with Transient-Isotachophoresis Jean-Marc Busnel,*,†,‡ Bart Schoenmaker,† Rawi Ramautar,† Alegria Carrasco-Pancorbo,† Chitra Ratnayake,‡ Jerald S. Feitelson,‡ Jeff D. Chapman,‡ Andre´ M. Deelder,† and Oleg A. Mayboroda† Biomolecular Mass Spectrometry Unit, Department of Parasitology, Leiden University Medical Center, Leiden, The Netherlands, and Beckman Coulter, Inc., Brea, California 92822, United States A sheathless interface making use of a porous tip has been used for coupling capillary electrophoresis and electrospray ionization mass spectrometry. First, effective flow rates using the interface have been characterized. It was found that the interface is capable of generating a stable spray with flow rates ranging from below 10 nL/min to >340 nL/min, enabling its use in either the mass or concentration-sensitive region of the electrospray process. Subsequently, by analyzing peptide mixtures of increasing complexity, we have demonstrated that this platform provides exquisite sensitivity enabling the detection of very low amounts of materials with very high resolving power. Transient isotachophoresis (t-ITP) can also be integrated with this setup to increase the mass loading of the system while maintaining peak efficiency and resolution. Concentration limits of detection in the subnanomolar or nanomolar range can be achieved with or without t-ITP, respectively. The application of a vacuum at the inlet of the separation capillary further allowed the peak capacity of the system to be improved while also enhancing its efficiency. As a final step in this study, it was demonstrated that the intrinsic properties of the interface allows the use of coated noncharged surfaces so that very high peak capacities can be achieved. Since the hyphenation of capillary electrophoresis (CE) and electrospray ionization mass spectrometry (ESI-MS) was first reported in the late 1980s,1-3 numerous publications describing novel interface designs have been published. More than 20 years later, research in this field is still very active, reflecting the complexity of this task.4 Indeed, as compared to the coupling of liquid chromatography (LC) and ESI-MS, when hyphenating CE and ESI-MS, one must consider a number of unique features. First, CE is a miniaturized technique intrinsically operating at very low * Corresponding author phone:+31-71-526-93-86; fax: +31-71-526-69-07; e-mail: [email protected]. † Leiden University Medical Center. ‡ Beckman Coulter, Inc. (1) Smith, R. D.; Udseth, H. R. Nature 1988, 331, 639–640. (2) Smith, R. D.; Olivares, J. A.; Nguyen, N. T.; Udseth, H. R. Anal. Chem. 1988, 60, 436–441. (3) Lee, E. D.; Muck, W.; Henion, J. D.; Covey, T. R. J. Chromatogr. 1988, 458, 313–321. (4) Maxwell, J. E.; Chen, D. D. Y. Anal. Chim. Acta 2008, 627, 25–33.

9476

Analytical Chemistry, Vol. 82, No. 22, November 15, 2010

flow rates. Depending on the background electrolyte (BGE) composition, electric field, inner surface chemistry and internal diameter of the separation capillary, operating conditions may induce flow rates often well below 20 nL/min.5-7 For example, the use of noncharged (neutral) capillary coatings or bare fused silica capillaries at acidic pH usually results in a significant suppression of electroosmostic flow (EOF) where only negligible flow rates are obtained.8,9 Second, the conductivity of the BGEs used in CE are generally much higher than the mobile phases classically employed in reversed phase liquid chromatography (RPLC). Finally, CE is an electrokinetically driven separation technique and therefore requires the application of a voltage across the separation capillary that must remain stable during the separation in order to generate analyte peaks of very high efficiencies. Both sheath liquid and sheathless interfaces have been proven to enable the coupling of CE with ESI-MS.4 While the operating flow rates in CE are usually below 100 nL/min, sheath liquid interfaces deliver coaxially to the separation capillary a sheath liquid with a flow rate typically ranging from 1 to 5 µL/min. Within this range of flow rates, the MS behaves as a concentrationsensitive detector,10 with the higher flow rate and dilution by the makeup flow inducing a much lower sensitivity than sheathless designs.11,12 However, sheath liquid interfaces are more commonly used, as they are commercially available and have historically provided a more robust spray.13 The versatility of CE-based separation methodologies have generated increasing interest over the last several years in various research fields such as (5) Huhn, C.; Ramautar, R.; Wuhrer, M.; Somsen, G. W. Anal. Bioanal. Chem. 2010, 396, 297–314. (6) Valaskovic, G. A.; Kelleher, N. L.; McLafferty, F. W. Science 1996, 273, 1199–1202. (7) Wahl, J. H.; Goodlett, D. R.; Udseth, H. R.; Smith, R. D. Electrophoresis 1993, 14, 448–457. (8) Balaguer, E.; Demelbauer, U.; Pelzing, M.; Sanz-Nebot, V.; Barbosa, J.; Neususs, C. Electrophoresis 2006, 27, 2638–2650. (9) Busnel, J. M.; Josserand, J.; Lion, N.; Girault, H. H. Anal. Chem. 2009, 81, 3867–3872. (10) Kostiainen, R.; Buins, R. P. Rapid Commun. Mass Spectrom. 1994, 8, 549– 558. (11) Zamfir, A. D. J. Chromatogr., A 2007, 1193, 2–13. (12) Wahl, J. H.; Gale, D. C.; Smith, R.D. J. J. Chromatogr., A 1994, 659, 217– 222. (13) Mischak, H.; Coon, J. J.; Novak, J.; Weissinger, E. M.; Schantra, J. P.; Dominiczak, A. F. Mass Spectrom. Rev. 2009, 28, 703–724. 10.1021/ac102159d  2010 American Chemical Society Published on Web 10/28/2010

metabolomics,14-16 proteomics,17 glycoproteomics,18,19 neurochemistry,20 and analysis of intact proteins.21,22 In this report, a detailed analytical evaluation of a new sheathless sprayer is presented. This sprayer design is adapted from the work originally described by Moini et al.,23 and reported to have the performance properties of sheathless interfacing (high sensitivity while retaining separation efficiency), while significantly improving robustness. Important attributes from this design include a reduction in voltage required to drive the electrospray ionization (ESI), the separation of the ESI process from the byproduct of H20 electrolysis (bubbles), and the uniformity of the capillary ID with sharpened outer tip (reduced clogging). Essentially, this sheathless interfacing of CE and ESI-MS was achieved through the use of a separation capillary having the outlet etched with hydrofluoric acid (HF) to provide an outer terminus porous to the flow of small ions. Our purpose was to extensively characterize the operating flow rates of this interface and to assess its capabilities for peptide separation and MS detection. After having studied the spraying properties of the interface under infusion as a function of the flow rate, samples of varying complexity were studied. Tryptic digests of bovine serum albumin (BSA) or Escherichia coli (E. coli) were considered as samples and capillary zone electrophoresis (CZE) with or without the integration of transient isotachophoresis (t-ITP) was implemented. After demonstrating that this platform combines the best features of CE and ESI-MS to provide an exquisite sensitivity with the ability to detect analytes in the low attomole range, the potential application of a vacuum at the inlet of the separation capillary, when using bare fused silica capillaries at acidic pH, was evaluated to tune the peak capacity of the system. Finally, given the demonstrated compatibility of the interface with very low flow rates in the range of 10 nL/min and lower, the possibility to use noncharged neutral capillary coatings was evaluated. MATERIALS AND METHODS Chemicals. All chemicals used were of analytical reagent grade and obtained from Sigma-Aldrich (Zwijndrecht, The Netherlands). All buffers and sample solutions were prepared with water produced by an Alpha-Q Millipore system (Amsterdam, The Netherlands). Tryptic digest of bovine serum albumin was from Bruker Daltonics (Bremen, Germany). Angiotensin I and Angiotensin II were obtained from Beckman Coulter (Brea, CA). Escherichia coli Sample Preparation. A culture of E. coli strain K12 was grown in LB medium at 37 °C overnight at 120 rpm. Cells were harvested at an optical density of 0.6 at 600 nm, and were washed from the medium by triplicate centrifugation (14) Soga, T. Methods Mol. Biol. 2007, 358, 129–137. (15) Lapainis, T.; Rubhakin, S. S.; Sweedler, J. V. Anal. Chem. 2009, 81, 5858– 5864. (16) Ramautar, R.; Van der Plaas, A. A.; Nevedomskaya, E.; Derks, R. J.; Somsen, G. W.; De Jong, G. J.; Van Hilten, J. J.; Deelder, A. M.; Mayboroda, O. A. J. Proteome Res. 2009, 12, 5559–5567. (17) Simpson, D. C.; Smith, R. D. Electrophoresis 2005, 26, 1291–1305. (18) Zamfir, A. D.; Peter-Katalinic, J. P. Electrophoresis 2004, 25, 1949–1963. (19) Amon, S.; Zamfir, A. D.; Rizzi, A. Electrophoresis 2008, 29, 2485–2507. (20) Lapainis, T.; Sweedler, J. V. J. Chromatogr., A 2008, 1184, 144–158. (21) Neususs, C.; Pelzing, M. Methods Mol. Biol. 2009, 492, 201–213. (22) Haselberg, R.; De Jong, G. J.; Somsen, G. W. J. Chromatogr. A 2007, 1159, 81–109. (23) Moini, M. Anal. Chem. 2007, 79, 4241–4246.

and resuspension at 10 000g × 10 min at 4 °C in 10 mM HEPES/ NaOH buffer (pH 7.4) containing 0.3 M sucrose. A final (forth time) wash was done with the same HEPES buffer excluding sucrose. Aliquots of the pellet (∼3 mg) were lyophilized and stored at -80 °C in 1.5 mL tubes. After dissolving the E. coli pellet in deionized water containing 0.1% SDS, the protein concentration was determined by BCA (Micro BCA Protein Assay Kit, Pierce/ Thermo Cat. No. 23235) using BSA as a standard. Three milligrams of E. coli proteins were solubilized in three milliliters of 0.5 mol · L-1 TAEB containing 0.1% of RapiGest surfactant (Waters). Proteins were heat-denatured at 75 °C for 1 h and, after cooling, samples were digested overnight at 37 °C with trypsin (60 µg). To further cleave RapiGest in the sample, the pH was decreased below 2.0 with TFA, samples were incubated for one hour at room temperature and centrifuged at 10 000g for 10 min at 4 °C. The supernatant was finally collected, lyophilized, and stored at -80 °C. Capillary Electrophoresis. The CE experiments were carried out with a PA 800 plus capillary electrophoresis (CE) system from Beckman Coulter (Brea, CA) equipped with a temperature controlled autosampler and a power supply able to deliver up to 30 kV. Prototype fused-silica capillaries with a porous tip were obtained from Beckman Coulter (Brea, CA). Depending on the experiments, bare fused silica capillaries or noncharged capillaries were used. The neutral capillary coating is a proprietary product currently in development at Beckman Coulter. It is produced as a bilayer with the outer surface being comprised of a hydrophilic layer of polyacrylamide. Solutions of 10% acetic acid and ammonium acetate (pH 4 and various ionic strengths) were employed as BGE and leading electrolyte (LE), respectively. Injection volumes were calculated using the Poiseuille equation and a fluid viscosity of 1.04 cP. Viscosity Measurement. The viscosity of the respective BGEs was determined using the CE instrument with a previously reported method.24 Briefly, a 30 µm i.d. × 150 µm o.d. × 88.5 cm bare fused silica capillary was filled hydrodynamically with the BGE of interest. Next, a short plug of water was injected under 1 psi for 60 s. This zone was then mobilized hydrodynamically until its detection by the mass spectrometer. Viscosity (η) was calculated from the mobilization time applying the HagenPoiseuille law: η ) dc2∆Pt/(32L2)

(1)

Where dc is the internal capillary diameter, ∆P the mobilization pressure, t the mobilization time, and L the length of the capillary. CE-ESI-MS Interface. One end of the separation capillary was etched with hydrofluoric acid (HF) so that a porous section of approximately 3 cm was created. The etching procedure creates a capillary with an outer diameter at the etched portion of about 40 µm, but does not taper its inner part so that the prepared capillary presents a good ability to spray at low flow rates while reducing the potential for clogging. After the porous section is formed by etching, the etched portion of the capillary (outlet extremity) was inserted into a housing comprising a stainless steel (24) Bello, M. S.; Rezzonico, R.; Righetti, P. G. J. Chromatogr. 1994, 659, 199– 204.

Analytical Chemistry, Vol. 82, No. 22, November 15, 2010

9477

cylinder with a retractable outer head, which protects the porous tip from physical damage. The housing was then positioned in a bracket mounted on a custom stage (designed by Beckman Coulter, Brea, CA) to fit the mass spectrometer. This stage includes an x-y-z platform that allows the positioning of the porous tip in front of the mass spectrometer inlet. Typically, the tip was placed coaxially to the entrance of the MS instrument at a distance ranging from 2 to 5 mm, with the glass tip protruding from the stainless steel cylinder by approximately 5 mm. Before performing any CE-ESI-MS experiment, the stainless steel cylinder was filled with the BGE and grounded to close the electrical circuit of the CE. The position of the tip was subsequently optimized by following the MS response while hydrodynamically and/or electrophoretically introducing a test mixture. The porous sheathless CE-ESI interface is currently in development at Beckman Coulter, Inc. Beckman Coulter products discussed in this article are for Laboratory Use Only; not for use in diagnostic procedures. Mass Spectrometry. MS was performed using an ultrahighresolution-time-of-flight (UHR-TOF) maXis instrument from Bruker Daltonics (Bremen, Germany). Transfer parameters were optimized by direct infusion of a 2 µg/mL angiotensin I and angiotensin II peptide mixture. Spectra were collected at a data acquisition frequency of 2 Hz. CE-ESI-MS coupling was performed using the previously described interface. To allow the coupling, a nanoflow spray shield provided by Bruker (Bremen, Germany) was positioned on the inlet of the transfer glass capillary, the conventional ESI source removed and the specially designed stage positioned. The following spray conditions were typically used: ESI in positive ionization mode carried out with an ESI voltage ranging from -750 to -1500 V depending on the distance between the porous tip and the MS entrance; dry gas, 2 L/min; source temperature, 150 °C. RESULTS AND DISCUSSION Characterization of Interface Operating Flow Rates. As a preliminary step to perform sheathless CE-ESI-MS, the poroustip capillaries were first studied with regard to the range of flow rates in which they can generate a stable spray. Operating the ESI process at very low flow rates was shown to provide advantages, such as reduced ion suppression and improved sensitivity.25,26 Consequently, the purpose of this initial characterization step was to learn whether this sheathless CEESI-MS setup was compatible with such electrophoretic conditions and to what extent we could take advantage of the nanoflow behavior of the ESI process. To focus on the spraying properties, the characterization was done under hydrodynamic conditions. As a model peptide system, angiotensin I was dissolved in an aqueous BGE (10% acetic acid) and subsequently infused into the mass spectrometer at different flow rates using pressure applied at the capillary inlet. The flow rate generated at each pressure was calculated using the Poiseuille eq 2 and taking into account the experimentally determined BGE viscosity (1.04 cP) (25) Schmidt, A.; Karas, M.; Dulcks, T. J. Am. Soc. Mass Spectrom. 2003, 14, 492–500. (26) Kelly, R. T.; Page, J. S.; Zhao, R.; Qian, W. J.; Mottaz, H. M.; Tang, K. Q.; Smith, R. D. Anal. Chem. 2008, 80, 143–149.

9478

Analytical Chemistry, Vol. 82, No. 22, November 15, 2010

V ) (π/128)dc4((∆Pt)/(ηL))

(2)

After the capillary was filled with the sample to be infused, the inlet pressure was varied from 0.5 to 40 psi with pressure steps ranging from 0.5 to 10 psi with the MS spectra continuously recorded. Each pressure step was applied for two minutes and an averaged MS spectrum, over this duration, was used for subsequent data treatment. The full experiment was performed in triplicate. The peak intensities of the peptide under study are reported in Figure 1 as a function of the flow rate, taking into account both (M+H+)+ and (M+2H+)2+ ions. A stable spray was observed from the lowest to the highest pressure tested, which means that the interface can operate at flow rates ranging from approximately 4 nL/min to at least 336 nL/min. From Figure 1A, it appears that two regions can be distinguished. The first region corresponds to flow rates below ∼25 nL/min, where the signal intensity increases with the flow rate. The second region corresponds to flow rates above ∼25 nL/ min, where the peak intensities remain relatively constant while the flow rate is increased. Therefore, we propose that the described sheathless interface not only enables studies in the concentration-sensitive region of the ESI-process but also in the mass-sensitive region where the MS signal intensity is proportional to the number of molecules introduced into the mass spectrometer per unit time rather than unit volume (concentrationsensitive region). To further understand how the sensitivity of the detection process is related to the flow rate, the peak intensity was divided by the analyte quantity introduced into the MS for each two minute acquisition period, and the results subsequently reported as a function of the flow rate in Figure 1B. Figure 1B clearly demonstrates the dependence of sensitivity on the operating flow rate. Indeed, when the sensitivity obtained at a flow rate of 336 nL/min is compared to that at 12.6 nL/min, it appears that lowering the flow rate results in an improvement of the sensitivity by a factor of 18. Further examination of sensitivity levels measured within the mass-sensitive region of the ESI process revealed a slight decline in sensitivity when the lowest of the tested flow rates (4.2 nL/ min) is used. Considering that the ESI current was stable under these conditions, this observation could be explained by a lack of ion transmission at such a low flow rate. As compared to the sensitivity calculated at 8.4 nL/min, it corresponds to approximately a 15% decrease in sensitivity. The observations described by Figure 1A and B are consistent with previous literature where the detection sensitivity was significantly improved at very low flow rates.25 It appears that the porous capillary is compatible with a very wide range of electrophoretic conditions. Indeed, when the lowest (4.2 nL/min) and highest (336 nL/min) flow rates are considered, it can be calculated that the system is compatible with electroosmotic mobilities ranging from 2.9 to 233.10-5 · cm2 · V-1 · s-1. These values cover most of the conditions generally encountered in capillary zone electrophoresis (CZE), both with coated and uncoated capillaries used over a wide range of pH. Sheathless CZE-ESI-MS of a BSA Tryptic Digest. The results of the previous section have demonstrated that the porous sprayer can effectively function at ultralow flow rates. To further

Figure 1. Evolution of the peak intensity of Angiotensin I (a) and detection sensitivity as a function of the flow rate (b) Experimental conditions: capillary electrophoresis; bare fused silica capillary with a porous tip, total length 88.5 cm ×30 µm i.d. × 150 µm o.d.; Infused sample, Angiotensin I at 2 ng/mL in 10% acetic acid. Mass spectrometry; capillary voltage, -1350 V; detection range, 50-3000 m/z, other experimental conditions described in Materials and Methods. Analytical Chemistry, Vol. 82, No. 22, November 15, 2010

9479

Figure 2. Sheathless CZE-ESI-MS analysis of a BSA tryptic digest Experimental conditions: capillary electrophoresis; bare fused silica capillary with a porous tip, total length 88.5 cm ×30 µm i.d. × 150 µm o.d.; voltage, 30 kV; current, 4 µA; BGE, 10% acetic acid; sample, tryptic digest of BSA diluted in water (1 µM); Injection sequence, 1 psi 60 s sample, 1 psi 10 s BGE. Mass spectrometry; capillary voltage, -750 V; detection range, 50-3000 m/z, other experimental conditions as in Materials and Methods. (A) Base peak electropherogram. (B) Extracted ion electropherogram of m/z 536.7. (C) Extracted ion electropherogram of m/z 583.9. (D) Extracted ion electropherogram of m/z 722.30.

evaluate the performance of the sheathless porous capillary system, a tryptic digest of bovine serum albumin (BSA) was used. The experiments were performed using bare fused silica capillaries and a BGE consisting of 10% acetic acid (pH 2.2). These conditions are believed to generate a very low EOF due to protonation of most silanols on the capillary walls. Furthermore, a fully aqueous buffer of relatively high conductivity (0.22 S/m) was used and was not found to significantly affect the spray quality. A typical base peak electropherogram (BPE) corresponding to the CZEESI-MS analysis of a BSA tryptic digest is shown in Figure 2A. Three extracted ion electropherograms (EIE) are shown in Figure 2B-D corresponding to three arbitrarily selected tryptic peptides to further study the response of the MS. Experimental results in Figure 2 correspond to a 6.7 nL (1.1% of the total capillary volume) injection of a 1 µM solution of BSA tryptic digest, that is, only 6.7 fmol injected into the separation capillary. Still, even with such a low amount of material injected, most peptides yielded a very large signal-to-noise ratio with some of them even saturating the detector. Under these conditions, the FMF algorithm (Bruker Data Analysis) has been used to extract the masses of the peptides potentially present in the mixture. With search restricted only to compounds presenting a signal-to-noise ratio above 15, 83 features were attributed to BSA tryptic peptides (within 50 ppm mass accuracy and two missed cleavages allowed), corresponding to a 71% sequence coverage of the protein. Three replicate experiments were then performed with a BSA tryptic digest sample at 20 nM to assess the limits of detection (LOD) of the system when standard CZE experimental conditions (1.1% of the total capillary volume filled by the sample) are used. To calculate the LOD enabled by the system, three peptides were further chosen. Corresponding extracted ion electropherograms are shown in Figure 2. LODs were then calculated considering 9480

Analytical Chemistry, Vol. 82, No. 22, November 15, 2010

the signal-to-noise ratio (S/N). Supporting Information (SI) Table S-1 shows the average signal-to-noise ratio at the tested concentration (20 nM) for each of the three peptides and the subsequently calculated LODs taking into account a S/N of 3. With respect to SI Table S-1, the minimum peptide quantities needed for reliable detection were calculated for the corresponding sample volume (6.7 nL). As a result, the minimum quantity loads ranged from 9 to 55 amol. These very low values demonstrate the exquisite sensitivity of the platform. Use of Vacuum at the Inlet of the Separation Capillary to Tune Peak Capacity. As demonstrated in the SI, integrating a t-ITP step allowed a significant improvement in mass loading with a corresponding reduction in the limits of detection. In comparison to the results obtained without a preconcentration step, integrating a t-ITP step allowed to improve the LODs by a factor ranging from 10- to 86-fold, leading to subnanomolar concentration detection limits. Still, taking into account the very narrow time separation window achieved in SI Figure S-1, it appears that the separation power and consequently the peak capacity of the separation is modest as all the peptides are introduced into the mass spectrometer within a time window of only two minutes. Particularly, this very rapid separation is a serious drawback when complex samples are separated, especially considering the relatively slow MS/MS duty cycle of most mass spectrometers. As compared to the classical CZE separation of BSA (Figure 2), the resolution is significantly reduced in SI Figure S-1, which can be explained by the following two factors. First, under the experimental conditions reported, the sample plug represents almost one-third of the entire capillary length, with the stacking and the subsequent separation performed in the remaining two-thirds of the capillary. As a direct consequence, the length available for separation is significantly reduced and the separation power accordingly decreased. Second,

Figure 3. Use of vacuum at the capillary inlet to counter the electroosmotic pumping effect (Tryptic digest of BSA at 20 nM) Experimental conditions: Vacuum applied at the inlet was varied as follows. (A) 0 psi. (B) 0.5 psi. (C) 1 psi. (D) 1.5 psi. (E) 2 psi. Experimental conditions: Sample, BSA tryptic digest at 20 nM in 50 mM LE. Injection sequence, 15 psi 90 s sample, 1 psi 10 s BGE. Mass spectrometry; capillary voltage, -800 V; detection range, 50-3000 m/z, other experimental conditions as in Figure 2.

the sample matrix is at pH 4.0 while the BGE has a pH of 2.2. Since the experiments were carried out in bare fused silica capillaries, it can be proposed that the pH mismatch locally induces the partial deprotonation of the silanols in the capillary walls, which stimulates displacement of the entire capillary contents toward the detector through an EOF pump effect. To test the EOF pump hypothesis, we examined whether the application of a vacuum at the capillary inlet (which should counter the EOF pump) would in fact improve the peak capacity under stacking conditions. While the use of a hydrodynamic-driven flow in capillaries with internal diameters of 100, 75, or even 50 µm generally decreases separation efficiency due to the parabolic flow profile created, the magnitude of such an effect is expected to decrease with the internal capillary diameter.27 Indeed, when considering a parabolic flow profile hydrodynamically generated in a capillary, the difference between the velocities at its center and at another point of its cross-section increases with the distance between the two considered points.28 Since the capillaries used in this study have an internal diameter of only 30 µm, it was thus important to assess the possibility of tuning the separation resolution by making use of the vacuum capabilities of the PA 800 plus CE system. The experiments summarized in Figure 3 clearly illustrate the effect of applying vacuum at the capillary inlet. Indeed, when the vacuum applied at the capillary inlet was stepwise increased from (27) Tallarek, U.; Rapp, E.; Scheenen, T.; Bayer, E.; Van As, H. Anal. Chem. 2000, 72, 2292–2301. (28) Golay, M. J. E.; Atwood, J. G. J. Chromatogr. 1979, 186, 353–370.

0 to 2 psi, it can be seen in Figure 3 that the separation window and consequently the overall peak capacity is significantly enhanced. The experiment corresponding to Figure 3E was analyzed with the Bruker data analysis software and the compounds detected with a signal-to-noise ratio above 15 were considered for database search. Mascot search with these data returned 73 identified peptide matches, which corresponds to a 73% sequence coverage. The quantity of BSA tryptic digest loaded in this experiment was only 4 fmol. As can be seen in SI Figure S-3, a more precise characterization of the process is possible by extracting a single mass to charge ratio (m/z ) 722.3). Indeed, this marker demonstrates how beneficial the application of a vacuum can be when considering the separation resolution between two species having closely related electrophoretic mobilities. To further evaluate whether the application of a vacuum would negatively influence the separation efficiency, the peak width at half peak height of the signals reported in SI Figure S-3A were studied as a function of the vacuum applied (SI Figure S-3B). As can be seen, there is no significant trend, which suggests that the application of a vacuum does not hinder the achievement of high efficiencies., For example, when the peak efficiencies of the signals detected in the bottom panel of SI Figure S3-A were calculated, values of 660 000 plates (peak 1) and 670 000 plates (peak 2) were obtained. Similar to what was previously described using a sheath liquid interface,29 results generated with the porous sheathless CE-ESIMS platform are consistent with the fact that the ESI process might “pull” the contents of the capillary toward the mass Analytical Chemistry, Vol. 82, No. 22, November 15, 2010

9481

Figure 4. Base peak electropherogram corresponding to the analysis by sheathless CZE-ESI-MS of an E. coli tryptic digest in a noncharged neutral coated capillary. Experimental conditions: capillary electrophoresis; neutrally coated capillary with a porous tip, total length 95 cm ×30 µm i.d. × 150 µm o.d.; voltage, 30 kV; current, 3.7 µA; BGE, 10% acetic acid; sample, tryptic digest of E. coli at 0.5 mg/mL in 50 mM LE. Injection sequence, 5 psi 90 s sample, 1 psi 10 s BGE. Mass spectrometry; capillary voltage, -900 V; detection range, 50-3000 m/z, other experimental conditions as in the Materials and Methods.

spectrometer. As a result, a parabolic flow profile influencing the separation efficiency would be created inside the separation capillary, thereby hindering the achievement of very high theoretical plate values. On the basis of the results of SI Figure S-3, we propose that the application of a vacuum at the inlet of the separation capillary counterbalances the parabolic flow profile induced by the suction effect potentially created by the ESI process. If this interpretation is correct, it would mean, contrary to conventional thinking, that there may no need for significant EOF to couple CE and ESI-MS via a sheathless interface. Since the ESI process may itself impose suction on the contents of the separation capillary, and the porous sheathless interface is compatible with very low flow rates, the in-capillary flow generated by the ESI process might be sufficient to ensure a stable connection. In that case, it would be possible to envisage the use of noncharged neutral capillary coatings with highly suppressed EOF to achieve unprecedented levels of sensitivity and efficiency for the analysis of peptides and proteins by sheathless CE-ESIMS. Use of Noncharged Capillary Coatings in Sheathless CEESI-MS. It is often assumed that a significant EOF is required for the successful coupling of CE and ESI-MS using a sheathless interface. For this reason, numerous studies describing the use of acidic BGEs with sheathless CE-ESI-MS platforms have (29) Catai, J. R.; Sastre Torano, J.; De Jong, A. D.; Somsen, G. W. Electrophoresis 2006, 2091–2099.

9482

Analytical Chemistry, Vol. 82, No. 22, November 15, 2010

reported the use of positively charged capillary coatings so that a strong and stable EOF would be generated, whereas at acidic pH repelling the interaction of the net positively charged proteins/ peptides. In the view of the results of the previous section, it would appear that the intrinsic properties of the described interface might allow consideration of noncharged neutral capillary surfaces. If proven correct, this result would certainly represent a breakthrough in the field of CE-ESI-MS as it would open new perspectives in terms of peak capacity, preconcentration capabilities, and the potential of analyzing very complex samples under a large number of analytical conditions. As the final step in this study, we thus wanted to test the above hypothesis. To properly assess the potential peak capacity improvements allowed by the use of a noncharged neutral capillary surface, a tryptic digest of E. coli was used. In a preliminary set of experiments, different experimental conditions were tested. We assessed the possibility of using a neutral capillary surface for performing sheathless CE-ESI-MS without the additional application of any positive pressure at the separation capillary inlet Although such conditions permitted the detection of the migrating species, a lack of spray stability was observed. Consequently, the application of some positive pressure with application of an electric field was further tested. In this context, it was found that a pressure of only 0.5 psi was enough to provide a stable spray over the entire duration of the experiment. Taking into account the dimensions of the capillary (30 µm i.d. × 95 cm long), the viscosity

of the BGE (1.04 cP) and the Poiseuille equation, these conditions correspond to a bulk flow of approximately 4.5 nL · min-1. The corresponding experiment is shown in Figure 4. To further assess separation capabilities of the sheathless CE-ESI-MS system making use of a neutral capillary surface, the signals of 15 peptides with detection times ranging from 26 to 79 min were extracted from those detected in Figure 4. For each of the extracted peptides, the peak width at half peak height was then measured and an average value of 11 s was calculated. Further taking into account the ∼60 min separation window, the peak capacity was calculated to be 327. Such a high value clearly demonstrates the power of the described system as peak capacities of this magnitude have rarely been reported for monodimensional separations. CONCLUSION In this study, a porous sheathless CE-ESI-MS approach was extensively characterized on the basis of operating parameters and bioanalytical performances. We first demonstrated that the separation capillaries used for this coupling are compatible with ESI-MS within a very wide range of flow rates. As a consequence, depending on the electrophoretic conditions, the system can be operated either in the concentration-sensitive or mass-sensitive range of the ESI process, resulting in the achievement of exquisite sensitivities. We also demonstrated that a preconcentration technique such as t-ITP can easily be integrated to improve the

mass loading of the platform. It was shown that sample plugs as long as one-third of the separation capillary (>200 nL) can be used without significantly affecting the achievable resolution. Depending on whether t-ITP is used or not, the platform provides concentration limits of detection in the subnanomolar or low nanomolar range, respectively. Finally, it was demonstrated that the intrinsic properties of the interface even permit the use of noncharged neutral capillary coatings that exhibit only a minute quantity of EOF. With this approach we demonstrated that a very high peak capacity can be reached together with a very high sensitivity, ideal for the analysis of very complex samples only available in minute amounts. ACKNOWLEDGMENT We thank Rico Derks, Oleg Klychnikov, Ralf Schiewek, from the LUMC, Anna Lou, Jane Luo, Scott Mack, and John Hudson from Beckman Coulter, Inc. for valuable help and discussions. SUPPORTING INFORMATION AVAILABLE Additional information including two tables and three figures. This material is available free of charge via the Internet at http://pubs.acs.org. Received for review August 19, 2010. Accepted October 11, 2010. AC102159D

Analytical Chemistry, Vol. 82, No. 22, November 15, 2010

9483