Anal. Chem. 2002, 74, 3772-3776
Analysis of Carbonic Anhydrase in Human Red Blood Cells Using Capillary Electrophoresis/ Electrospray Ionization-Mass Spectrometry Mehdi Moini,* Sandra M. Demars, and Hsiaoling Huang
Department of Chemistry and Biochemistry, The University of Texas at Austin, Austin, Texas 78733
Capillary electrophoresis/electrospray ionization-mass spectrometry (CE/ESI-MS) was applied to the analysis of human red blood cells (RBCs) using the split-flow technique for interfacing CE to MS. By using a long (∼125cm) and narrow (∼15-µm-i.d.) capillary, the four major proteins of the RBC, which are hemoglobin (Hb, r- and β-chains, 900 amol/chain), carbonic anhydrase I (CAI, ∼7 amol/cell), and carbonic anhydrase II (CAII, ∼0.8 amol/cell), were separated from each other and detected at low-attomole levels in one run and minimal sample preparation. Under these conditions, the detection limits for CAI and CAII in lysed RBCs were ∼20 and ∼44 amol, respectively. The ∼20-amol detection limit of CAI was confirmed by the CE/ESI-MS analysis of three intact RBCs that had been drawn into the capillary under a microscope. A shorter capillary (∼55 cm long) provided faster analysis time but did not separate CAII from the β-chain of hemoglobin, causing the CAII signal to be masked by the background chemical noise generated by the ∼1000× molar excess of the β-chain. Under this condition, the CAII detection limit increased to ∼500 amol. From three methods of sample introduction (injection of lysed blood, injection of intact cells under microscope, and injection of intact cells suspended in saline solution), injection of lysed blood provided the optimum sensitivity. It was found that a background electrolyte (BGE) containing 0.1% acetic acid in water worked best for the analysis of intact cells, while a BGE containing 0.1% acetic acid in water + acetonitrile (50/50 by volume) worked best for the analysis of lysed blood. Capillary electrophoresis (CE) has rapidly evolved as a preferred separation technique for the analysis of the chemical contents of single cells.1-6 A variety of detection systems have * Corresponding author. E-mail:
[email protected]. Tel: (512) 4717344. Fax: (512) 471-1420. (1) Swanek, F. D.; Ferris, S. S.; Ewing, A. G. Capillary electrophoresis for the analysis of single cells: electrochemical, mass spectrometric, and radiochemical detection. In Handbook of Capillary Electrophoresis; Landers, J. P., Ed.; CRC Press: Boca Ration, FL, 1997; Chaspter 17. (2) Kennedy, R. T.; Oates, M. D.; Cooper, B. R.; Nickerson, B.; Jorgenson, J. W. Science 1989, 246, 57. (3) Kennedy, R. T.; St. Claire, R. L.; White, J. G.; Jorgenson, J. W. Mikrochim. Acta 1987, 2, 37. (4) Kennedy, R. T.; Jorgenson, J. W. Anal. Chem. 1989, 61, 436. (5) Oates, M. D.; Cooper, B. R.; Jorgenson, J. W. Anal. Chem. 1990, 62, 1573.
3772 Analytical Chemistry, Vol. 74, No. 15, August 1, 2002
been employed as the CE detector for the analysis of single cells. These techniques can be divided into two general categories: nonmass spectrometric techniques and mass spectrometric (MS) techniques. Among the nonmass spectrometric techniques, electrochemical detection and laser-induced fluorescence (LIF)7-16 provide the highest sensitivity for single-cell analysis. Voltametry and wavelength-resolved fluorescence17,18 can also provide some structural information, but their chemical identification capability is limited when compared to mass spectrometric techniques. Mass spectrometric techniques provide accurate molecular weight (MW) information as a means of chemical identification, a feature that is especially useful when dealing with complex mixtures. Human red blood cells (RBCs) are among the most widely used cells for single-cell analysis using both mass spectrometric and nonmass spectrometric techniques.19-24 This is because, in addition to their facile accessibility, RBCs are nonnucleated (which simplifies analysis) and have diameters (∼7.2-7.9 µm)25 that can be observed under an optical microscope and are suitable for the capillary dimensions utilized. Moreover, the most abundant proteins of RBCs, the R- and β-chains of hemoglobin (∼900 amol/ cell), are well within the sensitivity range of today’s mass (6) Cooper, B. R.; Jankowski, J. A.; Leszyczyszyn, D. J.; Wrightman, R. M.; Jorgenson, J. W. Anal. Chem. 1992, 64, 691. (7) Ewing, A. G.; Wallingford, R. A.; Olefirowicz, T. M. Anal. Chem. 1989, 61, 292A. (8) Olefirowicz, T. M.; Ewing, A. G. Anal. Chem. 1990, 62, 1872. (9) Olefirowicz, T. M.; Ewing, A. G. Chemia 1991, 45, 106. (10) Hogan, B. L.; Yeung, E. S. Anal. Chem. 1992, 64, 2841. (11) Olefirowicz, T. M.; Ewing, A. G. J. Neurosci. Methods 1990, 34, 11. (12) Jorgenson, J. W.; Lukacs, K. D. Science 1983, 222, 266. (13) Gordon, M. J.; Huang, X.; Pentoney, S. L., Jr.; Zare, R. N. Science 1988, 242, 224. (14) Gilman, S. D.; Ewing, A. G. J. Capillary Electropho. 1995, 2, 1. (15) Yeung, E. S. Acc. Chem. Res. 1994, 27, 409. (16) Jankowski, J. A.; Tracht, S.; Sweedler, J. V. Trends Anal. Chem. 1995, 14, 170. (17) Cannon, D. M., Jr.; Winograd, N.; Ewing, A. G.; Annu. Rev. Biophys. Biomed. Struct. 2000, 29, 239-263. (18) Fuller, R. R.; Moroz, L. L.; Gillette, R.; Sweedler, J. V. Neuron 1998, 20, 173-181. (19) Cannon, D. M., Jr.; Winograd, N.; Ewing, A. G. Annu. Rev. Biophys. Biomol. Struct. 2000, 29, 239-263. (20) Chen, S.; Lillard, S. J. Anal. Chem. 2001, 73, 111-118. (21) Whittal, R. M.; Keller, B. O.; Li, L. Anal. Chem. 1998, 70, 5344-5347. (22) Hofstadler, S. A.; Severs, J. C.; Smith, R. D.; Swanek, F. D.; Ewing, A. G. Rapid Commun. Mass Spectrom. 1996, 10, 919. (23) Cao, P.; Moini, M. J. Am. Soc. Mass Spectrom. 1999, 10, 184-186. (24) Valaskovic, G. A.; Kelleher, N. L.; McLafferty, F. W. Science 1996, 273, 1199-1202. (25) Terzakis, J. A.; Santegada, E. Anal. Quant. Cytol. Histol. 2000, 22, 244246. 10.1021/ac020022z CCC: $22.00
© 2002 American Chemical Society Published on Web 06/22/2002
spectrometers. In its endogenous state, hemoglobin (Hb) exists as a tetramer, which consists of two dimers, each composed of an R-chain (average MW 15 126) and a β-chain (average MW 15 865).22 Each chain contains a heme group. Depending on the background electrolyte (BGE) used, the chains exist as either monomers, dimers, or tetramers, which can complicate analysis. Carbonic anhydrase I (CAI, average MW 28 780)26 and carbonic anhydrase II (CAII, average MW 29 156)27 are the next most abundant proteins, with the respective quantities of ∼7 and ∼0.8 amol in each adult RBC.28,29 In addition to the clear reasons for monitoring hemoglobin (e.g., for identification of variants), there is also a need for the identification of carbonic anhydrase isoforms. For example, a significant decrease in CAI levels could suggest hemolytic anemia or hyperthyroid Graves’ disease,30,31 CAII deficiency results in renal tubular acidosis,32 while osteoperosis, symmetrical cerebral calcification, and mental retardation have been noted in individuals with very low levels of CAII in their RBCs.33 Moreover, a decrease in the ratio of CAI/CAII is observed in individuals deficient in glucose-6-phosphate dehydrogenase. Therefore, it is important to develop sensitive analytical techniques for the detection and identifications of hemoglobin and carbonic anhydrase isoforms, preferably at single-cell levels. While detection of the R- and β-chains of hemoglobin in a single intact RBC has been accomplished by several researchers, detection of carbonic anhydrase in single cells has been more challenging. In both recent studies of a single intact RBC using CE/electrospray ionization (ESI)-MS, carbonic anhydrase was not detected.22,23 However, when a crude isolate from human blood was analyzed by CE/ESI-MS utilizing a Fourier transform ion cyclotron resonance (FTICR) mass spectrometer, a CA was detected with an average MW of 28 780.24 Although this work did not specify which CA isoforms was detected, the reported average molecular weight was consistent with the average molecular weight of CAI. The sample preparation procedure that was used isolated CA from blood by solvent extraction and ultrafiltration, and therefore, no other proteins were detected. In a more recent CE-LIF study of single human RBCs, only CAII was apparently detected.20 In this work, CE/ESI-MS was used to investigate the major proteins of human RBCs with emphasis on the separation and detection of all four major proteins of RBCs in one run without solvent extraction or ultrafiltration. EXPERIMENTAL SECTION Two fused-silica capillaries (55 cm long and 125 cm long) were utilized for these experiments, each with the dimensions of 15(26) Andersson, B.; Nayman, P. O.; Strid, L. Biochem. Biophys. Res. Commun. 1972, 48, 670-677. (27) Henderson, L. E.; Henriksson, D.; Nyman, P. O. Biochem. Biophys. Res. Commun. 1973, 52, 1388-1394. For a more recent CAII amino acid sequence, see the NCBI database at http://www.ncbi.nlm.nih.gov. (28) Lindskog, S. Pharmacol. Ther. 1997, 74, 1-20. (29) Aliakba, S.; Brown, P. R. Clin. Biochem. 1996, 29, 157-164. (30) Chiang, W.-L.; Chu, S.-C.; Lai, J.-C.; Yang, S.-F.; Chiou, H.-L.; Hsieh, Y.-S. Clin. Chim. Acta 2001, 314, 195-201. (31) Yoshida, K. Tohoku J. Exp. Med. 1996, 178, 345-356. (32) Nagai, R.; Kooh, S. W.; Balfe, J. W.; Fenton, T.; Halperin, M. L. Pediatr. Nephrol. 1997, 11, 633-636. (33) Aramaki, S.; Yoshida, I.; Yoshino, M.; Kondo, M.; Sato, Y.; Noda, K.; Jo, R.; Okue, A.; Sai, N.; Yamashita, F. J. Inherited Metab. Dis. 1993, 16, 982990.
µm inner diameter (i.d.) and 150-µm outer diameter (o.d.) (Polymicro Technologies, Phoenix, AZ). Both capillaries were derivatized using (aminopropyl)trimethoxysilane (APS).34 The CE/ ESI-MS interface utilized the split-flow design.35 The capillaries’ outlet tips were sharpened by etching with 40% hydrofluoric acid. Equipment and Chemicals. A P/ACE System MDQ CE instrument (Beckman Instruments, Fullerton, CA) in conjunction with a Finnigan LCQ MS (Finnigan, San Jose, CA) was used to collect all the CE/ESI-MS data. The mass spectrometer sensitivity was optimized for each BGE solution. The mass spectrometer scanned in the mass/charge (m/z) range of 1450-2000 with a scan time of 0.5-1 s. For all experiments, the CE inlet electrode was maintained at -30 kV (reverse polarity mode), and the electrospray voltages were adjusted to between 1.3 and 1.8 kV (according to the sharpness of the CE outlet tip and its distance from the MS inlet) to maintain a stable electrospray. To increase the MS sensitivity, the LCQ original heated capillary (0.4-mm i.d.) was replaced with an API II (Finnigan, San Jose, CA) heated capillary (0.5-mm i.d.). To maintain a pressure of 1.5 × 10-5 Torr in the ion trap analyzer housing, an additional mechanical pump (Edwards High Vacuum International, Crawley, Sussex, England) with a pumping capacity of 4.7 L/s was added to the original pumping system. In addition, because of the large inner diameter of the API II heated capillary, the temperature of the heated capillary was increased from 200 to 300 °C for optimum desolvation. The sensitivity of detection for carbonic anhydrase after instrument modification increased ∼5-fold. All chemicals were purchased from Sigma Chemical Co. (St. Louis, MO) and used without further purification. A solution containing human Hb, CAI, and CAII at the respective concentrations of 1.00, 0.01, and 0.01 mg/mL was analyzed as a standard to examine the relative migration order of the proteins and the mass accuracy of the instrument. The CE BGE solutions explored were 0.1% acetic acid solutions containing either 0, 25, 50, or 60% acetonitrile. However, among the BGE solutions containing the organic solvent, the BGE with 50% acetonitrile produced the optimum performance in terms of separation and electrospray stability. Therefore, the BGE used in all experiments presented here was a 0.1% acetic acid solution containing either 0 or 50% acetonitrile. Compared to the 0.1% acetic acid solution, the BGE containing 50% acetonitrile displayed a lower electroosmotic flow (EOF), which can cause ESI instability. However, sharp capillary tips and the lower surface tension of the acetonitrile-containing BGE provided stable ESI operation even at these low EOF levels. Methods and Techniques. The chemical contents of RBCs were analyzed using three different sampling methods: (1) A solution of lysed RBCs was injected into the capillary using the pressure injection mode. The solution was prepared by diluting 5 µL of fresh blood from a healthy adult male to 50 µL using saline solution, followed by centrifugation and removal of the supernatant, leaving only intact cells without the plasma. After adding 50 µL of water to lyse the cells, the lysed cells were then diluted either 5× or 200× with water to final dilutions of 50× and 2000×, respectively. (2) Intact RBCs (one or three cells) were drawn into the capillary inlet using suction and a 200× microscope for observation.23 (3) Intact RBCs suspended in a saline solution were (34) 28. Moseley, M. A.; Jorgenson, W. J.; Shabanowitz, J.; Hunt, D. F.; Tomer, K. B. J. Am. Soc. Mass Spectrom. 1992, 3, 289. (35) Moini, M. Anal. Chem. 2001, 73, 3497-3501.
Analytical Chemistry, Vol. 74, No. 15, August 1, 2002
3773
Figure 2. (A) Base peak electropherogram of the lysed blood at a CAII concentration of ∼500 amol/nL using the short column. Approximately 1 nL of lysed blood was injected. (B) Mass spectrum of the peak marked CAI & CAII of panel A. In panel B, I and II denote CAI and CAII, respectively. The 0.1% acetic acid solution containing 50% acetonitrile was employed for this experiment.
Figure 1. (A) Base peak electropherogram of the lysed blood at a CAII concentration of ∼88 amol/nL using the long column. Approximately 0.5 nL of lysed blood was injected. (B) Mass spectrum of the peak marked CAI of panel A. (C) Mass spectrum of the peak marked CAII of panel A. (D) Mass spectrum of the peak labeled Hb (β) of panel A. (E) Mass spectrum of the peak labeled Hb (R) of panel A. The 0.1% acetic acid solution containing 50% acetonitrile was used for this experiment. Unmarked peaks on panel A were not identified.
injected into the CE capillary using the pressure injection mode of the CE instrument. The solution of suspended intact RBCs was prepared in a manner almost identical to method 1, except that saline solution was used instead of water to maintain intact cells.36 To homogenize the solution, the sample vial was hand-shaken immediately before sampling. It should be noted that it was necessary to use fresh blood to separate the Hb chains; when a one-week-old blood sample was analyzed using the 0.1% acetic acid solution, only one peak with an average molecular weight of ∼65 000 was observed. RESULTS AND DISCUSSION Figure 1 panels A-E, respectively, show the base peak electropherogram of the lysed blood using the 125-cm-long capillary and the corresponding mass spectra of each peak of panel A. In this experiment, 0.5 nL of the 50× diluted lysed blood solution was injected, corresponding to the injection of ∼400 amol of CAI (Figure 1B) and ∼44 amol of CAII (Figure 1C). The measured average molecular weight of each compound was obtained from the deconvolution of the mass spectra of Figure 1 (36) Weinstein, R. S. In The Red Blood Cell, 2nd ed.; Surgenor, D. M-N., Ed.; Academic Press: New York, 1974; Chapter 5.
3774 Analytical Chemistry, Vol. 74, No. 15, August 1, 2002
panels B-E. The measured average molecular weights (in order of migration times) were 28 784, 29 159, 15 869, and 15 131, respectively. Since the LCQ mass accuracy at these quantitation levels had been experimentally determined to be 0.03%, it was concluded that the peaks were, respectively, CAI, CAII, the β-chain, and the R-chain. The average molecular weight obtained for CAI was consistent with the previous FTICR study in which a crude human blood extract was analyzed.24 In addition, the higher signal/noise ratio (S/N) of the mass spectrum of CAI versus CAII (Figure 1, panels B and C) was consistent with recent reports citing that CAI is ∼9× more abundant in red blood cells than CAII.29 The results, however, were inconsistent with a recent CELIF study in which apparently CAII was detected, although the presence of CAI was not discussed.20 The application of the long capillary was essential for the separation of all four major proteins of the RBC, as well as for the unambiguous identification of CAII. Using a shorter (55-cmlong) capillary, CAII was not detected. From Figure 1, and from the analysis of the standard solution using the acetic acid BGE both with and without 50% acetonitrile with the long capillary, it was observed that CAII always migrated after CAI and before the β-chain. It was therefore concluded that, in addition to its low quantity in RBCs, when using the shorter capillary the detection of CAII was also inhibited by its comigration with the β-chain, and therefore, its signal was masked by the chemical background noise associated with the ∼1000× molar excess of the β-chain using ESI. Under this condition, CAII was only detected at a very high blood concentration (∼500 amol of CAII/nL) where CAII was comigrated with CAI and showed up as a shoulder to an unseparated Hb peak (Figure 2, top panel). The further loss of separation efficiency at a high blood concentration was attributed to column overload because when 500 amol of CAII was injected, a large quantity of Hb (∼500 fmol of each chain) was injected as well. At this Hb concentration, the mass spectra of CAI and CAII were barely above the chemical background noise associated with Hb (Figure 2, lower panel). The high background chemical noise in the mass spectra of the major peaks was consistent with
Figure 3. (A) Base peak electropherogram generated from the CE/ ESI-MS analysis of an injection of ∼1 nL of the 2000× diluted lysed blood containing ∼20 amol of CAI. (B) Mass spectrum of the peak labeled CAI of panel A. (C) Mass spectrum of the peak labeled Hb (β) of panel A. (C) Mass spectrum of the peak labeled Hb (R) of panel A. The 0.1% acetic acid solution containing 50% acetonitrile was employed for this experiment.
previous studies of RBCs and with the analysis of other endogenous biological samples.22,23,37 The chemical noise was attributed to adduct formation (during the separation or ESI processes) of the major proteins (Hb) with solvents, salts, and other chemicals that are present in the endogenous samples or in the BGE.38 The use of selected ion monitoring (using the most prominent ions of CAI and CAII in the mass range studied) or the dynamic exclusion capability of the LCQ (excluding the prominent peaks of Hb) did not significantly improve the detection limit of CAII. The results imply that, at these quantitation levels, the background noise was due to reactions occurring outside the ion trap rather than processes occurring inside the ion trap (such as space charge effect) and that this was the dominant factor associated with the poor detection limit of CAII when its peak was comigrated with the β-chain. The mass resolution of the quadruple ion trap was not high enough to allow for the distinction between CAII peaks and the background species. Therefore, at low quantitation levels, when the CAII peak had comigrated with the peak of Hb, its signal was indistinguishable from the chemical background noise associated with Hb. Therefore, for the 15-µm-i.d., 55-cm-long capillary, the detection limit for CAII in RBCs was ∼500 amol. To obtain the detection limit for CAI in RBCs, a 2000× diluted lysed blood solution was analyzed using the short (55-cm) capillary. Compared to the long capillary, the main advantage of the shorter capillary was its shorter analysis time. In addition, at low blood concentrations, the separation efficiency of the short capillary was adequate for complete separation of CAI from Hb. Figure 3, panel A shows the base peak electropherogram obtained from the injection of 1 nL of the 2000× lysed blood solution containing ∼20 amol of CAI. Figure 3, panels B-D show the corresponding mass spectra of the electrophoretic peaks of Figure 3, panel A. As shown (Figure 3, panel B), CAI was detected with (37) Quenzar, T. L.; Emmett, M. R.; Hendrickson, C. L.; Kelly, P. H.; Marshall, A. G. Anal. Chem. 2001, 73, 1721-1725. (38) Wright, P. W.; Lister, A. S.; Dorsey, J. G. Anal. Chem. 1997, 699, 32513259.
Figure 4. (A) Base peak electropherogram of three intact RBCs that were drawn into the CE capillary and analyzed using CE/ESIMS. (B) Mass spectrum of the peak labeled CAI of panel A. (C) Mass spectrum of the peak labeled Hb (R,β) of panel A. Three RBCs contain ∼20 amol of CAI and ∼1.4 fmol of Hb. The 0.1% acetic acid solution was used for this experiment. Unlabeled peaks were unidentified. The peak marked with * represents the salt.
good sensitivity. Since 20 amol of CAI is approximately equivalent to the CAI content of three intact RBCs, three intact red blood cells were injected into the 55-cm capillary using methods 2 and 3 to confirm the detection limit. Figure 4, panel A shows the base peak electropherogram obtained from the injection (under microscope, method 2) of three intact RBCs. Panels B and C show the mass spectra of the peaks designated as CAI and Hb (R,β) of panel A, respectively. Poor S/N in panel B of Figure 4 (several charge states were barely above the background noise) indicate that the detection limit for CAI is ∼3 cells. Similar results were obtained when intact cells were injected using method 3. However, the S/N of the CAI mass spectrum obtained from the suspended cell solution (method 3) was even lower than when cells were injected under a microscope (method 2, Figure 4, panel B). The result implies that the cell density of the solution was below the theoretical value of 3 cells/nL. Cell precipitation or unavoidable cell adsorption to the vial wall may be responsible for the decrease in cell density. For intact cell analysis, best overall performance (separation and sensitivity) was achieved using the 0.1% acetic acid solution without any organic additives. However, when lysed blood was analyzed, the best overall performance was achieved using a 0.1% acetic acid solution containing 50% acetonitrile. Under these conditions, the presence of acetonitrile in the CE BGE reduced the EOF, 38 and consequently, the four major proteins of RBCs were all separated (Figure 1). Moreover, addition of acetonitrile also reduced the background chemical noise and adduct formation due to the incomplete desolvation of liquid droplets and, therefore, improved the S/N of CAI (Figure 3, panel B). The improvement of S/N due to the latter BGE was further confirmed by analyzing a standard mixture of CAI and CAII (0.1 mg/mL) using a 0.1% acetic acid solution and a 0.1% acetic acid solution containing 50% acetonitrile. It was observed that while the mass spectra of these compounds had similar intensity on the absolute scale, both noise level and adduct formation were significantly lower for the BGE containing 50% acetonitrile. Other possible factors for the higher Analytical Chemistry, Vol. 74, No. 15, August 1, 2002
3775
sensitivity of these proteins include better solubility and less adsorption to the capillary wall when the acetonitrile-containing BGE was used. CONCLUSION Human red blood cells were analyzed by CE/ESI-MS, and all four major proteins of the red blood cell were detected in one analysis with minimal sample preparation using a long capillary. Using both the short and the long capillary, CAI was separated from Hb chains. The detection limit for CAI was ∼20 amol. CAII, however, was only separated from CAI and Hb chains by using a long (125-cm) capillary. Under this condition, the detection limit of CAII in lysed blood was ∼44 amol. Using a short capillary, the detection limit of CAII increased to ∼500 amol. The increase in detection limit was because of comigration of CAII with the ∼1000× molar excess of the Hb β-chain. Under this condition, the background chemical noise associated with the ∼1000× molar excess of Hb β-chain masked the CAII signal at lower quantities (
