Detecting Large Biomolecules from High-Salt Solutions by Fused

Apr 23, 2002 - Deliver Bioinformatics Services in Public Cloud: Challenges and Research Framework. Xi Sun , Liya Fan , Linlin Yan , Lei Kong , Yang Di...
0 downloads 7 Views 77KB Size
Anal. Chem. 2002, 74, 2465-2469

Detecting Large Biomolecules from High-Salt Solutions by Fused-Droplet Electrospray Ionization Mass Spectrometry Der-Yeou Chang, Chia-Cheng Lee, and Jentaie Shiea*

Downloaded by TUNGHAI UNIV on August 23, 2015 | http://pubs.acs.org Publication Date (Web): April 23, 2002 | doi: 10.1021/ac010788j

Department of Chemistry, National Sun Yat-Sen University, Kaohsiung 80424, Taiwan

A novel fused-droplet electrospray ionization (FD-ESI) source was developed to generate peptide and protein ions. The sample solution was first ultrasonically nebulized to form fine aerosols. The aerosols were then purged into a glass reaction chamber via nitrogen. Charged methanol droplets were continuously generated through electrospraying the acidic methanol solution from a capillary, which was located at the center of the reaction chamber. As the sample aerosols entered the reaction chamber, they fused with the charged methanol droplets from which electrospray proceeded continuously. The mass spectra of peptide and protein that FD-ESI-MS produced were practically identical to those that conventional ESI-MS produced. However, FD-ESI-MS resulted in an extremely high salt tolerance. Cytochrome c ions were detected in the solutions that contained 10% (w/w; 1.709 M) NaCl or 2.5% (425 mM) NaH2PO4. As with those obtained from the solution that lacked NaCl and NaH2PO4, the width of cytochrome c ion peaks remained nearly unchanged. Multiple channel electrospray ionization mass spectrometry (MC-ESI-MS) has been successfully connected to gas chromatography (GC) and flow pyrolyzer (FP) to detect volatile organic as well as small biological compounds.1-3 In MC-ESI-MS, the gaseous analytes that GC or FP elutes were directed into the central channel of a seven-channel electrospray ionization source via nitrogen. Concurrently, the surrounding six channels were electrosprayed with a methanol solution that contained 1% trifluoroacetic acid.2,3 The purpose of employing multiple electrosprayers is to ensure that there are sufficient amounts of charged methanol droplets for fusion to occur. Ion-molecular reactions (IMR) between the gaseous analyte (M) and protonated species (e.g., H+, (CH3OH)H+) produced protonated analyte ion (MH+) within the MC-ESI source.4-6 It is indicated that the charged droplets that the outlying electrosprayers generated absorbed the gaseous analyte, and IMR occurred within these droplets.2,3 (1) Wang, C. H.; Shiea, J. J. Mass Spectrom. 1997, 32, 247. (2) Lee, C. Y.; Shiea, J. Anal. Chem. 1998, 70, 2757. (3) Hong, C. M.; Tsai, F. C.; Shiea, J. Anal. Chem. 2000, 72, 1175. (4) Zhan, D.; Rosell, J.; Fenn, J. B. J. Am. Soc. Mass Spectrom. 1998, 9, 1241. (5) Fenn, J. B. Int. J. Mass Spectrom. 2000, 200, 459. (6) Amad, M. H.; Cech, N. B.; Jackson, G. S.; Enke, C. G. J. Mass Spectrom. 2000, 35, 784. 10.1021/ac010788j CCC: $22.00 Published on Web 04/23/2002

© 2002 American Chemical Society

Recently, MC-ESI-MS has been employed to successfully ionize and detect multiply charged peptide and protein ions.7 The mass spectra of these proteins were practically identical to those that conventional ESI-MS produced. In MC-ESI-MS, the sample solution was ultrasonically nebulized to form fine aerosols. The aerosols were then directed into the central channel of the MCESI source to fuse with the charged methanol droplets that were generated by electrospraying the acidic methanol solution from the outlying six electrosprayers. Detection of the multiply charged protein ions indicated that electrospray appeared to proceed continuously from the newly formed droplets. Separating the ionization and nebulization processes in the above fused-droplet electrospray ionization (FD-ESI) processes results in independent control of the sample preparation. For example, by simply dissolving the protein in deionized water, MC-ESI-MS continues to detect protein ion signals.7 However, MC-ESI source construction is somewhat tedious, because the exits of the six outlying elerctrosprayers must be situated exactly on the same plane;1-3 otherwise, the voltage applied to the neighboring electrosprayers will interfere with individual electrospray. Therefore, a modified FD-ESI source is presented herein to simplify ion source construction. As is generally known, the presence of moderate to large amounts of inorganic salts in a sample solution will decrease electrospray stability and reduce the analyte ion signals significantly.8,9 This is due to both ionization suppression as well as sodium-adduct ion formation.8,9 Therefore, in addition to conventional chromatographic techniques, desalting processes (such as conventional dialysis, solid-phase extraction, and multiple buffer exchange using membrane cartridges) are sometimes required to obtain a clean sample for conventional ESI-MS analysis.10-16 Recently, on-line microdialysis, which cleans protein and oligo(7) Shiea, J.; Chang, D. Y.; Lin, C. H.; Jiang, S. J. Anal. Chem. 2001, 73, 4983. (8) Tang, L.; Kebarle, P. Anal. Chem. 1991, 63, 2709. (9) Ikonomou, M. G.; Blades, A. T.; Kebarle, P. Anal. Chem. 1991, 63, 1989. (10) Wu, J. Q.; Pawliszyn, J. Anal. Chem. 1995, 67, 2010. (11) Leeds, J. M.; Graham, M. J.; Truong, L.; Cummins, L. L. Anal. Biochem. 1996, 235, 36. (12) Kanazzawa, H.; Konishi, Y.; Matsushima, Y.; Takahashi, T. J. Chromatogr. A. 1998, 797, 227. (13) Watkins, L. K.; Bondarenko, P. V.; Barbacci, D. C.; Song, S.; Cockrill, S. L.; Russell, D. H.; Macfarlane, R. D. J. Chromatogr. A 1999, 840, 183. (14) Liu, C.; Muddiman, D. C.; Tang, K.; Smith, R. D. J. Mass Spectrum. 1997, 32, 425. (15) Cheng, X.; Gale, D. C.; Udseth, H. R.; Smith, R. D. Anal. Chem. 1995, 67, 586. (16) Liu, C.; Verma, S. S. J. Chromatogr. A 1999, 835, 93.

Analytical Chemistry, Vol. 74, No. 11, June 1, 2002 2465

Downloaded by TUNGHAI UNIV on August 23, 2015 | http://pubs.acs.org Publication Date (Web): April 23, 2002 | doi: 10.1021/ac010788j

Figure 1. Schematic diagram of fused-droplet electrospray ionization mass spectrometry (FD-ESI-MS): A, ultrasonic nebulizer, B, Teflon tube; C, piezoelectric transducer; D, acrylic plate; E, threeway tee; F, glass reaction chamber; G, electrospray capillary in a Teflon tube.

nucleotide samples prior to ESI-MS analyses, has also been developed.17-19 Although on-line desalting approaches have been proved effective, they may produce memory effects and a clogging problem as a result of membrane usage. Our previous studies illustrated that via the MC-ESI-MS, only MH+ ions were detected within various pyrolytical products;3 however, rather than MH+, conventional ESI-MS detected MNa+ within these pyrolytical reactants. Although this difference is unexplained, it implies that during the ionization processes in MCESI-MS, sodium salts in the sample solution do not affect the MH+ signal. The objective of this work is then to examine the degree of salt tolerance within peptide and protein detection via the newly developed FD-ESI source. EXPERIMENTAL SECTION Instruments. Figure 1 schematically depicts the fused-droplet electrospray ionization source (FD-ESI), which is connected to a quadruple mass analyzer. The aqueous sample solution was pumped onto the surface of the piezoelectric transducer in an ultrasonic nebulizer (CETAC U-5000AT), which generated fine aerosols. The sample flow rate was 100 µL/min. The power of the piezoelectric transducer was increased gradually until fine aerosols were generated. Notably, this nebulizer is employed conventionally to generate aerosol in ICPMS;20-22 however, the heating chamber (or heat sink) and condenser, which dried the aerosol in the ICP/MS, were not employed. Via nitrogen gas, the sample aerosols were purged into a glass cylindrical reaction chamber (i.d. 2.5 cm, 5.5 cm long) through a sidearm. A Teflon tube (i.d. 0.9 cm, 50 cm long) was employed to connect the ultrasonic nebulizer to the glass reaction chamber. As well, the end of the glass chamber was positioned directly in front of the sampling skimmer of a quadruple mass spectrometer. (17) Liu, C.; Wu, Q.; Harms, A. C.; Smith, R. D. Anal. Chem. 1996, 68, 3295. (18) Liu, C.; Hofstadler, S. A.; Bresson, J. A.; Udseth, H. R.; Tsukuda, T.; Smith, R. D. Anal. Chem. 1998, 70, 1797. (19) Xiang, F.; Lin, Y.; Wen, J. A.; Matson, D. W.; Smith, R. D. Anal. Chem. 1999, 71, 1485. (20) Tarr, M. A.; Zhu, G.; Browner, R. F. Anal. Chem. 1993, 65, 1689. (21) Weber, A. P.; Keil, R.; Tobler, L.; Baltensperger, U. Anal. Chem. 1992, 64, 672. (22) Tarr, M. A.; Zhu, G.; Browner, R. F. J. Anal. At. Spectrom. 1992, 7, 813.

2466

Analytical Chemistry, Vol. 74, No. 11, June 1, 2002

The methanol solution, which contained 1% acetic acid, was electrosprayed continuously from a fused-silica capillary that was in the center of the glass reaction chamber. The distance between the exit of the electrospraying capillary and the sampling skimmer on a mass analyzer was ∼1 cm. The capillary was connected to a three-way tee, and a syringe pump delivered the methanol solution through the capillary. A stainless steel wire that was inserted into one arm of the three-way tee fed the high voltage (3500 V from a Glassman EH20R20 power supply) into the acidic methanol solution. The operator has to be cautious of electric shock when high voltage is applied onto the stainless steel wire. Mass Spectrometry. A PE Sciex API 1 mass spectrometer was employed to detect the positive ions that were generated within the source. Mass spectra were recorded at a scan rate of ∼2 s/scan. As well, the temperature of the electrospray interface chamber was maintained at 55 ( 1 °C. For comparison, the presenting mass spectra were obtained by averaging 10 continuous scans. Material. The peptide, protein, and solvents (HPLC grade) were purchased commercially (Sigma or Aldrich) and used without further purification. The concentration of all peptide and protein solutions used for conventional ESI-MS and FD-ESI-MS was 10-6 M (in pure water). Acidic methanol solution (1% acetic acid) was prepared for electrospray. RESULTS AND DISCUSSION The results of previous studies showed that multiply charged protein ions were produced by MC-ESI-MS. However, constructing the MC-ESI source is somewhat tedious and time-consuming. Herein, a modified ESI source was developed to ensure operation and construction ease (Figure 1). The modified fused droplet ESI (FD-ESI) source used only a capillary to generate electrospray (instead of six electrosprayers in the MC-ESI source). The electrosprayer was enclosed at the center of a cylindric glass reaction chamber with both sides opened. This is also different from the MC-ESI source, where the ionization was performed in an open space. Although the FD-ESI source uses one capillary to generate electrospray, multiply charged protein ions were still detected. This may be because the sample and the charged methanol droplets were confined in the glass reaction chamber, and the ionization efficiency was then enhanced. The flow rate of the purging nitrogen gas as well as the electrospray of the acidic methanol solution also affected the analyte ion signal. The optimum analyte ion signal was obtained at a nitrogen flow rate of 20 mL/min. The flow of the acidic methanol solution through the syringe pump was adjusted (from 30 µL/min to 0.22 µL/min), and the optimum ion signal was obtained at a flow rate of ∼2.6 µl/min. At a lower methanol flow rate, the analyte ion intensity decreased, and at higher one, the mass spectra degraded rapidly. The analyte ion signals depended strongly on relative positions of the sidearm (for sample aerosols to enter the reaction chamber) as well as the tip of the capillary electrosprayer. Thus, a reaction chamber that had a movable sidearm (1 cm increment for each movement) was constructed. The capillary was axially affixed to the center of the chamber. The sample aerosols were introduced into the reaction chamber through three positions: in front, parallel, or behind the capillary tip. The mass spectra of cytochrome c that were obtained at different positions were compared.

Downloaded by TUNGHAI UNIV on August 23, 2015 | http://pubs.acs.org Publication Date (Web): April 23, 2002 | doi: 10.1021/ac010788j

Figure 2. Positive ESI mass spectra of Lys-bradykinin (10-6 M) that (a) conventional ESI-MS and (b) FD-ESI-MS produced.

Notably, the sidearm in the position behind the capillary tip produced the strongest ion signals. The distance between the capillary electrosprayer and the sampling skimmer on the mass analyzer (0.8-1.3 cm) was then adjusted gradually until the best ion signal was obtained. Figure 1 depicts the experimental conditions that were determined in this manner, which were then employed for all of the analyses performed herein. The size of the reaction chamber also affected the analyte ion signals. It was discovered that to produce the analyte ion signals the internal diameter of the glass reaction chamber must exceed 2.5 cm. The charged methanol droplets that the electrosprayer generated fly to the nearest ground. That is, if the reaction chamber’s internal diameter is less than the distance between the capillary tip and the sampling skimmer on the mass analyzer, the charged methanol droplets fly toward the wall of the reaction chamber. When the internal diameter of the reaction chamber exceeded 2.5 cm, the charged methanol droplets flew toward the sampling skimmer and the mass analyzer. Figure 2 presents the ESI mass spectra of Lys-bradykinin that conventional ESI-MS (Figure 2a) and FD-ESI-MS (Figure 2b) produced, respectively. Within conventional ESI-MS, the analyte was dissolved in the acidic water (0.1% of acetic acid) and the deionized water was employed to dissolve the analyte in FD-ESIMS. Both techniques detected singly and doubly charged Lysbradykinin ions. Actually, the peak width of the peptide that FDESI-MS produced was practically identical to that conventional ESI-MS produced. However, the signal-to-noise ratio in the former was somewhat lower than it was in the latter. The peptide concentration employed in this analysis was 10-6 M. On the basis of the signal intensity of Lys-bradykinin in Figure 2b, meaningful Lys-bradykinin spectra should be possible at concentration lower than 10-6 M. To test the salt tolerance of FD-ESI-MS, Lys-bradykinin was dissolved in an aqueous solution that contained 5% NaCl (854 mM). Figure 3 depicts the mass spectra that conventional ESIMS (Figure 3a) and FD-ESI-MS (Figure 3b) produced, respec-

Figure 3. Positive ESI mass spectra of Lys-bradykinin (10-6 M with 5% NaCl (854 mM) in the sample solution) that (a) conventional ESI-MS and (b) FD-ESI-MS produced.

tively. Although the protonated analyte ion signal (MH22+) was observed in the conventional ESI mass spectrum (Figure 3a), it also contained extensive sodium adduct ions (e.g. MNaH2+ and MNa22+). However, within FD-ESI-MS, no sodium adduct ions were detected, and the mass spectrum (Figure 3b) remained practically identical to that without salt (Figure 2b). This implies that within FD-ESI-MS, a high NaCl concentration fails to affect the peak width of the analyte signal. There are at least two possible hypotheses that may explain this phenomenon. First, evaporation may decrease the aerosol’s diameter and increase salt concentration within the sample aerosols when the aerosols are purged through the Teflon tube. At a certain size, the sample aerosols may become oversaturated, thus forcing NaCl to precipitate out of the aerosol. To verify this possibility, electrospray was discontinued in the reaction chamber, and all sample aerosols that entered the reaction chamber were collected in a glass vial. A conventional ESI-MS was then employed to analyze the collected solution, which determined that interference of NaCl was still present and the mass spectrum was similar to that depicted in Figure 3a. Hence, the experimental results indicate that purging the sample aerosols through a Teflon tube does not remove NaCl. Another possible reason to avoid NaCl interference in FD-ESIMS is that following fusion, the NaCl is somehow excluded from the fused droplets. As is generally known, NaCl is extremely watersoluble, but it is insoluble in most organic solvents such as methanol and dichloromethane.23,24 As the aqueous sample droplet fused with the charged methanol droplet in the reaction chamber, NaCl could not dissolve in the charged methanol droplet. To investigate this possibility, a drop (∼0.05 mL) of saturated NaCl (23) Weast, R. C. CRC Handbook of Chemistry and Physics, 70th ed.; CRC Press: 1990, D-255. (24) Budavari, S. The Merck Index, 11th ed.; Merck & Co., Inc.: Rahway, NJ, 1989; pp 5868.

Analytical Chemistry, Vol. 74, No. 11, June 1, 2002

2467

Downloaded by TUNGHAI UNIV on August 23, 2015 | http://pubs.acs.org Publication Date (Web): April 23, 2002 | doi: 10.1021/ac010788j

Figure 4. Positive ESI mass spectra of cytochrome c (10-6M) that was dissolved in the aqueous solutions, which contained various amounts of NaCl (from 0 to 10% (1.709M)). The mass spectra were obtained by conventional ESI-MS (a-e) and FD-ESI-MS (f-j).

aqueous solution was added into 1 mL of methanol. The NaCl particle precipitated immediately. The same phenomenon could also occur when the sample aerosol (containing abundant NaCl) fused with the charged methanol solution. The experimental results confirmed that as the electrospray solution was changed from acidic methanol to acidic water, the NaCl-Lys-bradykinin adduct ions were produced, and the mass spectrum was again similar to that depicted in Figure 3a. The effects of the salt on the protein ion signal in FD-ESI-MS were also investigated. Figure 4 depicts the positive ESI mass spectra of cytochrome c that was dissolved in the aqueous solutions, which contained varying amounts of NaCl. The mass spectra were obtained by conventional ESI-MS (Figure 4a-e) and FD-ESI-MS (Figure 4f-j), respectively. Notably, within conventional ESI-MS, the mass spectra degraded with an increase of NaCl concentration. As the NaCl concentration exceeded 1.5% (255 mM), the NaCl cluster ions25 (e.g. m/z 782.6, (NaCl)13Na+ and m/z 1308.1, (NaCl)22Na+) predominated the mass spectra, and the cytochrome c ion signals became indistinguishable. The peak width of the cytochrome c ion peaks also increased concurrently (25) Juraschek, R.; Dulcks, T.; Karas, M. J. Am. Soc. Mass Spectrom. 1999, 10, 300.

2468 Analytical Chemistry, Vol. 74, No. 11, June 1, 2002

with the amount of NaCl that was dissolved in the sample solution, which was due to the formation of the protein-NaCl adduct ions. However, when FD-ESI-MS was employed to analyze these sample solutions, only protonated molecular ions were detected, and the width of the protein ion peaks remained nearly unchanged, even when the NaCl concentration was increased to 10% (1.709 M; Figure 4j). In fact, even within the sample solution that lacked NaCl, FD-ESI-MS produced narrower protein ion peaks than conventional ESI-MS did (a comparison of Figure 4a,f). As is generally known, within conventional ESI-MS, the sample might have contained trace amounts of NaCl impurities, and the ions with added sodium or protein-NaCl adduct ions increased the analyte ion peak width. However, caution was necessary when the 10% NaCl (1.709 M) sample solution was analyzed, because the signals of cytochrome c were weak, and the NaCl particles accumulated around the sampling skimmer and often blocked the ion entrance. Figure 5 depicts the ESI mass spectra of cytochrome c that was prepared in the aqueous solutions, which contained varying sodium phosphate concentrations. In conventional ESI-MS, as the sodium phosphate concentration neared 0.5% (85.5 mM), protein ion signals were barely detectable, and peaks that originated from

Downloaded by TUNGHAI UNIV on August 23, 2015 | http://pubs.acs.org Publication Date (Web): April 23, 2002 | doi: 10.1021/ac010788j

Figure 5. Positive ESI mass spectra of cytochrome c (10-6M) that was dissolved in the aqueous solutions, which contained various amounts of sodium phosphate (NaH2PO4) (from 0 to 2% (340 mM)). The mass spectra were obtained by conventional ESI-MS (a-e) and FD-ESI-MS (f-j).

the sodium phosphate predominated the mass spectrum (Figure 5a-e). However, in FD-ESI-MS, a mass spectrum with wellresolved protein ion peaks was still obtained (Figs. 5f-i). Finally, when the sodium phosphate concentration attained 2% (342 mM), cytochrome c ion peaks could still be detected (Figure 5j). The interferences of other organic salts such as TRIS and SDS (sodium dodecyl sulfate) with cytochrome c signals were also not observed when the sample solution was analyzed by FD-ESI-MS (data not shown). CONCLUSION A novel fused-droplet electrospray ionization (FD-ESI) source, which generated multiply charged peptide and protein ions was presented herein. Formation of these ions in FD-ESI-MS was performed in two steps. First, the sample solution was ultrasonically nebulized to produce fine aerosols. These aerosols were then fused with the charged methanol droplets that electrospray generated in a reaction chamber. Second, electrospray from the newly created droplets produced multiply charged protein ions. Because of their extremely low solubility in methanol, inorganic

salts within the sample solution were excluded from the newly created fused droplets. The mass spectra of the peptides and proteins produced in this manner were predominated by the protonated molecules, and inorganic salt interference was relatively undetectable. Thus, an extremely high salt tolerance (sodium chloride and sodium phosphate) was achieved for cytochrome c. However, the sample consumption in the present FD-ESI-MS exceeded that in a conventional ESI-MS, because a peristaltic rather than a syringe pump was employed herein to introduce the sample solution into the ultrasonic nebulizer. A FDESI source equipped with a nebulizer that reduces sample consumption is currently under development in our laboratory. ACKNOWLEDGMENT The authors would like to thank the National Science Council of Taiwan for financially supporting this research. Received for review July 12, 2001. Accepted March 8, 2002. AC010788J

Analytical Chemistry, Vol. 74, No. 11, June 1, 2002

2469