Use of a Polybrene Capillary Coating in Capillary Electrophoresis for

Jul 1, 1997 - An ion trap storage/reflectron time-of-flight (IT/reTOF) mass spectrometer, described in previous work,24,25was used for mass analysis...
3 downloads 4 Views 148KB Size
Anal. Chem. 1997, 69, 2451-2456

Use of a Polybrene Capillary Coating in Capillary Electrophoresis for Rapid Analysis of Hemoglobin Variants with On-Line Detection via an Ion Trap Storage/Reflectron Time-of-Flight Mass Spectrometer Michael X. Li, Lin Liu, Jing-Tao Wu, and David M. Lubman*

Department of Chemistry, The University of Michigan, Ann Arbor, Michigan 48109-1055

A polybrene capillary coating in capillary electrophoresis (CE) has been used for rapid analysis of hemoglobin variant digests. The use of the polybrene capillary coating has allowed sufficient separation to resolve the large number of digest products formed upon tryptic digestion of the whole protein, so that prior separation of the hemoglobin r and β chains is not required. The resolution of the digest peaks obtained by CE is sufficient so that even single amino acid substitutions can easily be detected using UV absorption detection. The digest is further analyzed by capillary electrophoresis separation with on-line detection using electrospray ionization interfaced to the ion trap storage/reflectron time of flight device (CE/ESI-IT/reTOF), where a comparison of the total ion electropherograms and mass spectra of the mutant and normal hemoglobins can detect the presence of a mutation site. The CE separation and mass analysis can be accomplished in typically 10-15 min. The unique capability of the CE/ESI-IT/reTOF system for detection of fast separations with narrow peaks that may be under 1 s fwhm is demonstrated. The speed of this system is essential for resolution of the large number of peaks that are separated in a short time duration using CE separations. The detection and identification of hemoglobin (Hb) variants has become an important area in clinical practice.1 Most states require the screening of every newborn for the presence of hemoglobin variants so that a method for rapid, widespread testing is needed.2 More than 700+ 3 structural variants of hemoglobin are known to date, some of which may be responsible for severe diseases such as sickle cell anemia, which is directly related to the structure of the Hb, while others are clinically silent.4 Many of these variants are single amino acid substitutions in either the R or β chain, although double mutation, chain elongation, deletion, and other mutations have been observed.5 The detection of these (1) Ingram, V. M. Nature 1956, 178, 792-4. (2) Huisman, T. H. J. The Hemoglobinopathies; Churchill Livingstone: Edinburgh, New York, 1986. (3) International Hemoglobin Information Center List. Hemoglobin 1994, 18, 77-161. (4) Shackleton, C. H. L.; Witkowska, H. E. Anal. Chem. 1996, 68, 29A-33A. (5) Prome´, D.; De´on, C.; Prome´, J.; Wajcman, H.; Galacteros, F.; Blouquit, Y. J. Am. Soc. Mass Spectrom. 1996, 7, 163-7. S0003-2700(97)00076-0 CCC: $14.00

© 1997 American Chemical Society

abnormal hemoglobins has relied on gel electrophoresis6 and isoelectrofocusing,7 which are labor intensive and slow methods. Recent work has seen the use of reversed-phase high-performance liquid chromatography (RP-HPLC) of either Hb or its tryptic digests to identify Hb variants in less than 1 h.8 However, HPLC does not often have sufficient resolution to separate all peptides in Hb tryptic digests, and overlapping or unresolved peaks may make identification difficult. In more recent work,9,10 HPLC has been coupled to mass spectrometry with electrospray ionization (ESI/MS) as a means to identify the digestion fragments resulting from Hb variants. The mass spectrum has been used to identify the mutation fragments based on mass difference and ultimately by MS/MS experiments to pinpoint the mutation sites. In the tryptic digest of normal Hb, there are 29 digestion products though typically 25-35 products are produced for most of the variants. The resulting HPLC separation of these digests is such that coeluting of the normal and mutant fragments may make identification difficult in an ESI/MS experiment, especially if the mass difference between the normal and mutant forms is 1 Da or less. Many of the mutations involve substitution of only one or two amino acids so that the mass difference is often small.11 In such cases, generally the R and β chains of hemoglobin are first separated and then digested separately to simplify the separation. The basic HPLC/MS procedure though is still time-consuming for large scale screening and the resolution often limited for many substitutions. The development of capillary electrophoresis12,13 with its high sensitivity and resolving power has provided a rapid method for the separation of complex biological mixtures. Several groups have used capillary electrophoresis (CE) for separation of tryptic digests of various Hb variants with UV detection14,15 or off-line (6) Schneider, R. G. Clin. Chem. 1974, 20, 1111-5. (7) Basset, P.; Beuzard, Y.; Garel, M. C.; Rosa, J. Blood 1978, 51, 971-82. (8) Congote, L. F.; Kendall, A. G. Anal. Biochem. 1982, 123, 124-32. (9) Witkowska, H. E.; Bitsch, F.; Shackleton, C. H. L. Hemoglobin 1993, 17, 227-42. (10) Davis, M. T.; Lee, T. D.; Ronk, M.; Hefta, S. A. Anal. Biochem. 1995, 224, 235-44. (11) Wada, Y.; Matsuo, T.; Sakurai, T. Mass Spectrom. Rev. 1989, 8, 379-434. (12) Jorgenson, J. W.; Lukacs, K. D. Science 1983, 222, 266-72. (13) Gordon, M. J.; Huang, X.; Pentoney, Jr., S. L.; Zare, R. N. Science 1988, 242, 224-8. (14) Ross, G. A.; Lorkin, P.; Perrett, D. J. Chromatogr. 1993, 636, 69-79. (15) Castagnola, M.; Messana, I.; Cassiano, L.; Rabino, R.; Rossetti, D. V.; Giardina, B. Electrophoresis 1995, 16, 1492-8.

Analytical Chemistry, Vol. 69, No. 13, July 1, 1997 2451

detection by mass spectrometry,16 while Lee et al.17 employed online capillary isoelectric focusing and mass spectrometry for Hb variant analysis. An important advantage of CE is that separation can be accomplished in under 10 min. The resolution of CE separations depends upon the fragment electrophoretic mobilities that are determined by their charges and masses. Thus, mutants with small mass differences of even 1 amu or less can often be separated by this method, and the use of on-line CE/MS would be invaluable for detecting mutations in such digests.18-20 In recent work, we have reported on analysis of globin chain digests in variant hemoglobins using on-line CE/MS with amine-coated fused-silica capillaries where the CE anodic end was directly applied as the microelectrospray needle.21 This work demonstrated that in principle by using CE combined with mass analysis a rapid and sensitive method was feasible for screening and detection of abnormal Hb’s. The main limitation in this work was the necessity for preseparation of the R- and β-globin chains to reduce the number of digestion peaks to simplify the mass spectrum. The major obstacle in analysis of the whole Hb protein digests is the large number of product peaks formed upon enzymatic digestion. The (3-aminopropyl)trimethoxysilane (APS)coated capillary used in previous work resulted in elution of the samples too quickly to allow sufficient separation among the large number of peptides.21 In the present work, we report on the use of a polybrene capillary coating with on-line CE/MS detection that is capable of resolving and identifying the large number of digest peptides from both Hb and variant Hb’s. This method provides improved sensitivity in detection without loss of resolution over a broad mass range for detection of low-intensity signals in these CE separations, where a sample of 10-50 fmol of hemoglobin was typically used in these experiments and separated and identified in 10-15 min. EXPERIMENTAL SECTION Preparation of Hb Proteins and Digests. Hb proteins were processed as reported in previous work.21 Initially, hemolysates were prepared from crude blood samples by washing erythrocytes three times with an equal volume of normal saline buffer. Washed red cell samples (0.1 mL) were centrifuged to remove any additional contaminants, and the supernatant was aspirated. The cells were lysed by diluting with 0.5 mL of distilled water and centrifuged at 8000g for 15 min to isolate the cell membrane and other solid materials. The supernatant was then added to a cold acid acetone (2%) solution to precipitate the globin and to remove the heme group, while the tube was shaken constantly by using a vortex machine (Model G-560E, Scientific Industries, Inc., Bohemia, NY). After centrifugation at 3000g for 10 min, the supernatant was removed and acetone was added to wash the precipitate. The precipitate was washed twice before it was vacuum dried. The proteins were redissolved in deionized water to a concentration of 1 µg/µL for further use. Subsequently, an aliquot of 10 µg of both the normal and variant Hbs was digested (16) Ferranti, P.; Malorni, A.; Pucci, P.; Fanali, S.; Nardi, A.; Ossicini, L. Anal. Biochem. 1991, 194, 1-8. (17) Tang Q.; Harrata, K.; Lee, C. S. Anal. Chem. 1996, 68, 2482-7. (18) Figeys, D.; Vanoostveen, I.; Ducert, A.; Aebersold, R. Anal. Chem. 1996, 68, 1822-8. (19) Smith, R. D.; Wahl, J. H.; Goodlett, D. R.; Hofstadler, S. A. Anal. Chem. 1993, 65, 574A-84A. (20) Cai, J.; Henion, J. D. J. Chromatogr. 1995, 703, 667-92. (21) Li, M. X.; Wu, J.-T.; Liu, L.; Lubman, D. M. Rapid Commun. Mass Spectrom. 1997, 11, 99-108.

2452

Analytical Chemistry, Vol. 69, No. 13, July 1, 1997

separately by TPCK treated-trypsin with a substrate to enzyme ratio of 50:1 in NH4HCO3 buffer solution at pH 8.4 for 24 h at 37 °C. The digests were vacuum-dried and reconstituted in the CE buffer. Separation. Capillary electrophoresis was performed on an in-house constructed setup using a (30 kV high-voltage power supply (Model CZE 1000R, Spellman High Voltage Electronics Corp., Plainview, NY). In order to minimize the adsorption of peptides to the capillary wall, the inner walls of the capillaries were coated with positively charged layers to prevent the peptides from adhering to the wall. It was previously reported21 that the APS-coated capillary can be used with CE/MS to separate tryptic digest products of Hb and thus identify the mutation sites in Hb β chains. However, the APS coating strongly repels the peptides and has a large EOF flow and subsequently a rapid elution time of typically 4 min where the separation period of the protein digests may only be 1-2 min in a 50 cm long capillary with a ∼-300 V/cm electrical field. In the case of a whole Hb protein digest analysis, under these conditions complete separation would be difficult to achieve since one would need to separate ∼10 eluents/min. In order to determine mutation sites from Hb digests, it is necessary to apply a separation method that can slow down the separation speed compared to that of the APS capillary and expand the difference among the peak elution times of various fragments. The separation selectivity needs to be improved while resolution is retained. The coating technique used to accomplish this goal is similar to that described by Thibault et al.22,23 with some modifications in the equilibration conditions and regeneration environment. The advantages of this procedure are the simplicity and convenience compared to the APS coating. In preparing an APS coating the capillary was often easily plugged by APS, if care was not properly taken when flushing with the coating solution. Since ethylene glycol and polybrene do not readily freeze, the coating layer is very reproducible, durable, and easily prepared. The buffer used in all of the experiments was an aqueous solution containing 100 mM formic acid and 5 mM ammonium acetate with a pH of ∼3.0, depending on the different samples separated. The capillary was conditioned with the CE buffer for 1 h prior to use and reactivated after several runs. All of the CE separations were performed in the negative mode with an electrical field strength of ∼-300 V/cm. The separation process was initially monitored by a variable-wavelength UV detector (Model SC100, Thermo Separation Products, Fremont, CA) with the wavelength set at 200 nm. Samples were injected using the electrokinetic method, which loads approximately 10-50 fmol of protein digests for each separation. The sample loading, CE separation, and data acquisition of UV signals were remotely controlled and performed by a personal computer (Model P5-66, Gateway 2000 Inc., N. Sioux City, SD), where an ADC-16 analog board embedded in the computer was used to collect the UV signal at a frequency of 7 Hz and custom software controlled all procedures. On-Line Sheathless CE/MS. An ion trap storage/reflectron time-of-flight (IT/reTOF) mass spectrometer, described in previous work,24,25 was used for mass analysis. A quadrupole ion trap (22) Kelly, J. F.; Locke, S. J.; Ramaley, L.; Thibault, P. J. Chromatogr. 1996, 720, 409-27. (23) Bateman, K. P.; White, R. L.; Thibault, P. Rapid Commun. Mass Spectrom. 1997, 11, 307-15. (24) Michael, S. M.; Chien, B. M.; Lubman, D. M. Anal. Chem. 1993, 65, 261420.

storage device (Model C-1251, R. M. Jordan Co., Grass Valley, CA) was used to store ESI-produced ions which were subsequently pulsed out to a reflectron TOF device for mass analysis (Model D1450). Typically, externally produced ESI ions were trapped and stored for a period of 0.25 s by an applied rf field (1100 V, 1.1 MHz) on the ring electrode using a He buffer gas to enhance trapping, before being ejected to the TOF device for mass detection. The ions stored in the trap were ejected to the TOF tube by applying a dc pulse (400 V) on one end cap and detected by a 25 mm triple microchannel plate detector (Model C-2501, R. M. Jordan Co.). The pressures inside the ion trap main chamber and the TOF tube were 1 × 10-5 and 2 × 10-7 Torr, respectively, while the approximate pressure in the trap was ∼10-3 Torr. In order to improve both the sensitivity and resolution of the system, a sheathless microelectrospray interface was constructed in this work, which has been reported previously.26 In this setup, one capillary end was directly applied as the electrospray needle. First, the capillary tip was etched from the initial size of 105 µm o.d. to a fine diameter by concentrated hydrofluoric acid after 0.5 cm of polyimide coating was burned. Then, the outer surface of this tip was silver coated using the electroless plating process27 described in detail in previous work. The coated tip was inserted into a 2 cm long stainless steel (ss) tube (125 µm i.d., 250 µm o.d.) connected to the electrospray power supply, to provide an electrical contact for both electrospray and capillary electrophoresis. The power supply of the capillary electrophoresis was set at ∼-12 kV, while the electrospray needle was at a voltage of ∼3.5 kV, so that the total separation voltage was ∼-15 kV. A heated (140 °C) 20 cm long ss (500 µm i.d.) tube was used as the counter electrode and the transit pathway for the electrospray-generated ions into the mass spectrometer. The liquid flow inside the coated capillary, which was determined by both the eletroosmotic and electrophoretic flows, was sufficient to form a stable electrospray background under this sheathless condition. The mass signals were collected, using a 250 MHz transient recorder (Model 9846, Precision Instruments Inc., Knoxville, TN) embedded in a P5-66 personal computer (Gateway Inc.). This data acquisition system could acquire data as fast as 25 Hz for over 0.5 h, using custom software developed in our laboratory.28 A time-of-flight range of 0-150 µs was generally used as the mass acquisition period, which corresponds to a m/z range of 0-1500. The mass calibration was conducted using three standard peptides, bradykinin, angiotensin III, and methionine enkephalin-ArgPhe, according to the following equation, (m/z)1/2 ) aT + b, where T is the flight time of various ions. Materials and Chemicals. Ammonium bicarbonate, ammonium acetate, formic acid, (3-aminopropyl)trimethoxysilane, polybrene, ethylene glycol, angiotensin III, bradykinin, and methionine enkephalin-Arg-Phe were purchased from Sigma (St. Louis, MO) and were used without further purification. TPCK treated-trypsin was obtained from Promega (Madison, WI). The water was filtered and deionized prior to use by a Milli-Q water purification system (Millipore Corp., Bedford, MA). HbA was obtained from Sigma Co., while all other crude Hb samples were (25) Qian, M. G.; Lubman, D. M. Anal. Chem. 1995, 67, 234A-42A. (26) Wu, J.-T.; Qian, M. G.; Li, M. X.; Liu, L.; Lubman, D. M. Anal. Chem. 1996, 68, 3388. (27) Mallory, G. O.; Hajdu, J. B. Electroless Plating: Fundamentals and Applications, American Electroplaters and Surface Finishers Society: Orlando, FL, 1990. (28) Qian, M. G.; Wu, J.-T.; Parus, S.; Lubman, D. M. Rapid Commun. Mass Spectrom. 1996, 10, 1209-14.

Figure 1. HbA digest: (A) CE electrophoregram with UV absorption detection and (B) CE/MS TIE, and inset mass spectra of (1) βT1 fragment at m/z of 477.2 and (2) βT3 fragment at m/z of 658.3. CE conditions: VCE ) -15 000 V; capillary, polybrene coated, 60 cm, 40 µm i.d., 105 µm o.d.; buffer, 100 mM formic acid and 5 mM ammonium acetate, pH ) 2.9; UV detection, 200 nm. CE/MS conditions: VCE ) -12 000 V, VESI ) 3500 V; mass acquisition speed, 4 Hz; other conditions are the same as those in the CE separation.

obtained as a gift from the World Laboratory (Ann Arbor, MI). Fused-silica capillaries were purchased from Polymicro Technologies (Phoenix, AZ). RESULTS AND DISCUSSION The aim of this work is to develop a rapid method of screening hemoglobin variants. As shown in Figure 1, this is accomplished by tryptic digestion of Hb with high-resolution CE separation of the products followed by mass spectrometric identification. The key problem is that the total digestion may result in as many as 30 peaks so that the peaks may be closely spaced and difficult to separate. The ability to separate the peaks is critical in order to observe the abnormal Hb digestion products that result from tryptic digestion where a mass difference of only 1 Da may need to be detected. In Figures 1-4 are shown the CE separations of the tryptic digests of hemoglobin variants as detected by UV absorption and mass spectrometry. As expected, the electropherograms obtained are similar, but there are clearly differences in the profiles. The use of the polybrene capillary coating with the proper choice of buffer composition allowed separation of a majority of the digest peaks so that abnormal peaks could be readily observed and subsequently identified by mass spectrometry. The buffer compositions in particular were adjusted so that optimal separation of peaks in the time interval where the majority of the peaks eluted could be achieved. The polybrene capillary coating was very Analytical Chemistry, Vol. 69, No. 13, July 1, 1997

2453

efficient in slowing down the EOF relative to the APS coating used in previous work,21 allowing a slower separation time which resulted in an improved resolution of the narrow peaks produced in CE. A direct comparison using an APS-coated capillary to separate these digests resulted in total ion electropherograms (TIEs) in which many of the peaks were poorly separated and overlapped with one another. In addition, it had been reported29 that the addition of a small amount of organic modifier into the aqueous buffer solution resulted in enhanced separation and in particular slowed down the peptide migration speed in the capillary. In the present work, a small amount of acetonitrile was shown to delay elution times but no improvement in the separation was observed. Therefore, in all of the separations performed, no organic modifier was used. Hemoglobin A. In panels A and B of Figure 1 are shown the tryptic digests of normal adult hemoglobin separated by CE with UV absorption and mass spectrometric detection, respectively. Normal adult hemoglobin contains two main chains, R and β, which provide 29 peaks under complete tryptic digestion. The separation of the HbA digest was accomplished using a polybrenecoated capillary, 100 mM formic acid, and 5 mM ammonium acetate buffer at pH 2.9. In Figure 1A, the separation of at least 25 fragments was completed in 15 min where most peaks were detected before 10 min. The late components appear to be larger fragments or even incompletely digested protein. Most of the peptides can be separated down to the baseline, where the separation efficiency was above 5.0 × 105 theoretical plates. The resolution attained with the polybrene coating was significantly higher than that achieved with the APS coating used in previous work.21 The resolution required to separate all the digest products without preseparation of the R and β chains was not possible using the APS coating. In comparison, the polybrene coating provided sufficient control over the EOF to provide the resolution needed to separate this complex peptide mixture and also prevented the peptides from sticking to the capillary wall. In Figure 1B is shown the TIE of the separation of the HbA digest using the on-line CE/ESI-IT/reTOF/MS system. The TIE reproduced the CE electropherogram relatively well except for several peaks at later elution times in the electropherogram that did not appear in the TIE. The main reason for the lack of appearance of these products in the TIE is the limited detection range (0-1500 Da) used in these experiments and also the relatively low rf trapping voltage. It should be noted that RT12, RT13, and βT12 are absent in both the TIE and UV absorption spectra as they are insoluble and precipitate out during the digestion.14 It is also possible that the trap efficiency is much lower for the low-mass peptides such as RT8 and thus reduces the detection efficiency. In addition, the high resolution and narrow peak widths produced by CE require rapid detection speed by the mass spectrometer detector in order to retain the CE resolution and peak shapes. The nonscanning nature of the IT/ reTOF/MS provides a >99% duty cycle in these experiments, but the ability to retain the narrow peaks from CE depends upon the data system and the mass spectral acquisition speed. In these experiments, generally a mass spectrometer acquisition speed of 4 Hz was used as the optimal compromise between sensitivity and the resolution of the CE separation. In the electropherogram, the narrowest peak width is ∼0.8 s fwhm so that at 4 Hz mass (29) Weinmann, W.; Maier, C.; Baumeister, K.; Przybylski, M.; Parker, C. E.; Tomer, K. B. J. Chromatogr., A 1994, 664, 271-5.

2454 Analytical Chemistry, Vol. 69, No. 13, July 1, 1997

Table 1. Comparison of the Observed and Calculated m/z Values for Tryptic Digests of HbA in CE/MS m/z peptide βT1a βT2 βT3b βT4 βT5 βT6,7 βT9 βT10,11 βT13 βT14 βT15 RT1 RT2 RT3 RT4 RT5 RT6 RT7 RT8 RT9 RT10,11 RT14

sequence no.

calcd

obsd

accuracy (%)

1-8 9-17 18-30 31-40 41-59 60-65 67-82 83-104 121-132 133-144 145-146 1-7 8-11 12-16 17-31 32-40 41-56 57-60 61-61 62-90 91-99 140-141

952.1 932.1 1314.4 1274.5 2059.3 656.8 1669.9 2547.8 1378.5 1149.4 318.3 728.4 460.3 531.3 1528.7 1070.5 1834.0 397.4 146.1 2997.3 1105.3 337.2

952.4 932.3 1314.1 1274.4 2059.5 657.1 1671.1 2546.3 1378.0 1149.6 318.1 728.8 460.4 531.3 1529.0 1070.4 1834.1 397.6 nd 2998.4 1105.9 338.1

0.03 0.02 -0.02 -0.01 0.01 0.03 0.01 -0.05 -0.02 0.01 -0.01 0.05 0.02 0.00 0.02 -0.01 0.01 0.03 0.04 0.05 0.26

a Calculated m/z values: (HbS) The substitution of glutamic acid by valine at β6 position causes the resulting fragment to be found at m/z 922.1. (HbC) The substitution of glutamic acid by lysine at β6 position causes two additional fragments at m/z of 275.1 (β7-8) and 694.0 (β1-6). b Calculated m/z values: (HbE) The substitution of glutamic acid by lysine at β26 position results in an additional cleavage point under tryptic digestion. The resulting two fragments have m/z values of 916.0 (β18-26) and 415.5 (β27-30), where the total of the two digest products is 1331.5.

spectral acquisition rate there are at least five or six points to define the peak. In Figure 1, it is shown that the CE separation was basically retained in the TIE. There was still some loss of the CE resolution in the TIE compared to that achieved in the electropherogram. As shown in previous work,21,30-32 this loss of resolution is not due to the speed of the mass spectral data acquisition but, rather, due to the electrospray source. In particular, the electrostatic pressure and evaporation at the tip in the sheathless electrospray induce a parabolic movement of the solvent toward the anodic end even without applying CE voltage and subsequently lowers the resolution. This effect depends heavily on the particular buffers and solvent used in the CE separation. It should be noted that the elution time for peaks in the TIE was longer than that for the electropherogram. This is due to the detection window in the UV setup being located 6 cm from the capillary end where the peptides traveled a longer distance to reach the ESI source for mass spectrometric detection. All of the tryptic digest peptides obtained through CE/MS are listed in Table 1. HbS Heterozygous (Sickle Cell Anemia). In Panels A and B of Figure 2 are shown the CE electropherogram and TIE of the tryptic digests of HbS, respectively. The digests were separated under conditions similar to those in HbA (Figure 1), except that a slightly higher pH value for the buffer solution (pH (30) Valaskovic, G. A.; McLafferty, F. W. Rapid Commun. Mass Spectrom. 1996, 10, 825-8. (31) Wilm, M. S.; Mann, M. Int. J. Mass Spectrom. Ion Processes 1994, 136, 16780. (32) Valaskovic, G. A.; Kelleher, N. L.; McLafferty, F. W. Science 1996, 273, 1199-202.

Figure 2. HbS digest: (A) CE electrophoregram with UV absorption detection and (B) CE/MS TIE and inset mass spectrum of variant βT1 fragment at m/z of 462.1. The only difference in conditions compared to Figure 1 is that the buffer pH is 3.0.

3.0) was used to optimize the separation. There was one significant additional peak in the TIE of HbS compared to HbA as marked by an arrow in Figure 2. The mutation in HbS represented by this peak corresponds to a substitution of a glutamic acid group by a valine at the β6 position, where the variation of the charge changed the elution time under the electrical field for this specific fragment. This was confirmed by the mass spectrum corresponding to this additional peak in the TIE as shown in the inset of Figure 2B. The mass spectrum showed that the HbS βT1 fragment (922.1 Da) was detected as a doubly charged ion at m/z 462.1. In Figure 1b, there is a corresponding mass spectrum inset where the HbA βT1 peptide (952.1 Da) was detected as a doubly charged ion at m/z 477.2. The mass shift between these two doubly charged ions is 15 Da, which corresponds to a mass difference of 30 for singly charged ions. This mass shift was as expected for the substitution of a glutamic acid by a valine group. HbC (Homogenous). The mutation site for HbC is the same as HbS, but with a different substitution, where the glutamic acid group was replaced by a lysine. This substitution resulted in an additional tryptic digestion position for HbC, where two fragments of 694 and 275.1 Da can be generated by the successful cleavage. As observed in Figure 3, the electropherogram and TIE of the HbC digest is similar to that of HbA, but differences are clearly visible. The peak corresponding to the βT1 fragment in HbA was barely observed, while two additional peaks were produced in the HbC electropherogram, which were marked as peak 1 and peak 2. The digest peaks observed in the TIE in Figure 3B matched the electropherogram quite well. The βT1 fragment was an extremely small peak in both the TIE and the mass spectrum. In the inset figures are shown the mass spectra of the βT1 fragment

Figure 3. HbC digest: (A) CE electrophoregram with UV absorption detection and (B) CE/MS TIE and inset mass spectra of the corresponding fragments in HbC βT1 (1) HbC β1-6 fragment at m/z of 348.2 and (2) HbC β7-8 fragment at m/z of 276.3. These two fragment ions result from the tryptic digestion at the lysine residue at β6 position in HbC. All conditions are the same as in Figure 2.

peptides corresponding to peaks 1 and 2 in HbC, which were produced when the β6 position was cleaved by trypsin. Peak 1 in the electropherogram was of very low intensity in both the TIE and corresponding mass spectrum. This was probably due to the low molecular weight of this specific fragment, which was 276.3 Da, and the trapping RF used in this experiment, which was optimized for higher mass. Peak 2 was detected by the IT/reTOF mass spectrometer at m/z 348.2 for the doubly charged ions, which corresponds to a molecular ion of 696.4 Da. The disappearance of the normal βT1 fragment and the emergence of the new peaks provided the evidence for the existence of the mutation. It is also noticeable that the intensity of βT14 and other fragments in the range of 6-8 min of the HbC TIE is relatively low compared to that in HbA, although the reason for this is not known. HbE (Heterozygous). The HbE variant results from the substitution of the β26 glutamic acid by a lysine group. In this case, enzymatic digestion results in two additional peaks when the lysine position is cleaved by trypsin. The result is demonstrated in Figure 4A by the difference in the electropherogram of the HbE digest compared to that of the HbA digest in Figure 1A. Several peaks that were initially suspected of being the mutant fragments are marked by arrows. The corresponding TIE for the HbE digest shown in Figure 4B confirmed that two additional peaks resulted from this variant digest. For HbA, the βT3 peptide was observed at a m/z of 658.3 for the doubly-charged ion, while the two extra ions were detected for HbE at a m/z of 416.5 and 458.5, respectively, where the latter was a doubly charged ion. The three mass spectra of these peptides are shown as insets in Analytical Chemistry, Vol. 69, No. 13, July 1, 1997

2455

resolving and identifying the large number of digest peptides from Hb and variant Hb’s. The use of the polybrene coating provides sufficient separation to resolve the large number of digest products formed upon tryptic digestion of the whole protein. The resolution of the CE method is such that even single amino acid substitutions can easily be detected using UV absorption detection. The digest is further analyzed on-line using electrospray ionization followed by detection using the IT/reTOF/MS, where a comparison of the TIEs and mass spectra of the mutant and normal hemoglobins can detect the presence of a mutant site. The entire analysis, CE separation and mass analysis, can be accomplished in typically 10-15 min. The use of the IT/reTOF/MS as a nonscanning mass spectrometer capable of detecting the narrow peaks from fast CE separation is demonstrated. The speed of this mass spectrometer system is essential for resolution of the large number of peaks that are separated in a short time duration by CE and in particular for detecting small changes in the tryptic digestion that may result from point mutations.

ACKNOWLEDGMENT

Figure 4. HbE digest: (A) CE electrophoregram with UV absorption detection and (B) CE/MS TIE and inset mass spectra of the corresponding fragments in HbE βT3. (1) HbE β27-30 fragment at m/z of 416.5 and (2) HbE β18-26 fragment m/z of 458.5. These two fragment ions result from the tryptic digestion at the lysine residue at β26 position in HbE. All conditions are the same as in Figure 2.

We gratefully acknowledge support of this work by the National Institutes of Health under Grant R01GM49500 and partial support of this work from the Army/ERDEC under Grant DAAD05-95-9-3517 and the National Science Foundation under Grant BIR-9513878. M.X.L. thanks Dr. Anthony Killeen of the Clinical Diagnostics Laboratory at the University of Michigan Medical Center for providing the patient hemoglobin samples.

Figures 1B and 4B, respectively. The mass values obtained matched those predicted theoretically with an average deviation of 0.05%.

Received for review January 22, 1997. Accepted April 15, 1997.X

CONCLUSION In conclusion, we have demonstrated that the use of a polybrene capillary coating with on-line CE/MS is capable of

AC970076M

2456 Analytical Chemistry, Vol. 69, No. 13, July 1, 1997

X

Abstract published in Advance ACS Abstracts, May 15, 1997.