Capillary electrophoresis-electrospray mass spectrometry for the

Baczynskyj, and George E. Bronson. Anal. Chem. , 1992, 64 (17), ... Charles C. Liu , Jyy S. Huang , David L. J. Tyrrell , Norman J. Dovichi. ELECTROPH...
0 downloads 0 Views 653KB Size
Download PDF
Anal. Chem. 1992, 64, 1864-1870

1884

Capillary Electrophoresis-Electrospray Mass Spectrometry for the Analysis of Recombinant Bovine and Porcine Somatotropins Kiyoshi Tsuji' Control Biotechnology, Pharmaceutical Control Diuision, T h e Upjohn Company, Kalamazoo, Michigan 49001

Lubomir Baczynskyj and George E. Bronson Physical & Analytical Chemistry, Upjohn Laboratories, T h e Upjohn Company, Kalamazoo, Michigan 49001

A Beckman P/ACE 2050 high-performance caplllary eiectrophoresls (HPCE) Instrument has been Interfaced with a Vestec electrospray lonlzatlon (ESI) mass spectrometer for the analysls of recomblnant proteins. Peak resolutlon Is not compromlsed by coupllng HPCE to an ESI mass spectrometer. Recombinant bovine and porclne somatotropins (rbSt and rpSt) were used as model proteins. The standard curve of the caplllary zone electrophoresk(CZE) method wtth UV detectlon for the determlnatlon of rpSt Is linear In the range of 7-300 fmol with theoretlcai plates of approximately 410 000 m-l. The relative standard devlatlon for the rpSt peak mlgratlon tlme Is less than 1%. The muttiplytharged Ion clusters obtalned In the CZE-ESI mass spectrum for a sample of rpSt ranged from m/r 1363.2 (the cluster with 16 charges) to 1982.5 (the cluster wtth 11 charges). The average molecular welghts of 21 812.6 and 21 798.3 for a sample of rbSt and rpSt determlned In thls study were nearly identical to the theoretical values of 21 812.0 and 21 797.9, respectlvely. Detectlon llmH of the CZE-ESI mass spectrometer is approxlmately 100 fmol. The CZE method separated monoand dldeamldated specles and monoacetylated compounds while the ESI mass spectrometer detected an analogue and a truncated homologue of rpSt comlgratlng with the major peak. The presence of mono- and dloxldlzed homologues was also detected In the major peak of some rbSt and rpSt samples. These data clearly Indicatedthat, lndlvldually, both CZE and ESI mass spectrometric methods could not detect all Impuritles. Coupllng of the HPCE instrument and the ESI mass spectrometer enhances analytical capabilities of both tools for rapid characterlzatlon of recomblnant proteins.

INTRODUCTION Due to its superb peak separation capability, highperformance capillary electrophoresis (HPCE) has become an accepted research technique. HPCE is especially suited for protein/peptide analysis because of its low (- 1O-pL) sample requirements. However, because of the small sample volumes (nanoliters) that can be injected into a fused-silica capillary column, characterization of peaks is extremely difficult. Collecting the peak fractions for identification132 can be too time consuming. Tsuji employed a sodium dodec(1)Guzman, N. A.; Trebilock, M. A.; Advis, J. P. The use of a concentration step to collect urinary components separated by capillary electrophoreses and further characterization of collected analytes by mass spectrometry. J. Liq. Chromatogr. 1991, 15, 997-1015. (2) Fujimoto, C.; Muramatsu, Y.; Suzuki, M.; Jinno, K. Capillary electrophoretic separation of amino acids: fraction collection. J.High Resolut. Chromatogr. 1991, 14, 178-180. 0003-2700/92/0364-1864$03.00/0

yl sulfate-polyacrylamide gel (SDS-PAG) filled capillary column to aid in the separation and characterization of recombinant proteins.3 Since the relative standard deviation of the gel-filled capillary system for the determination of molecular weights was fl-2 76,molecular weights of two proteins differing by a single amino acid could not be determined by this technique. Recent advances in electrospray ionization (ESI) mass spectrometry enabled determination of molecular weights of proteins exceeding 100 OOO with greater accuracy than attainable by the SDS-PAGEte~hn010gy.4~5 In a series of papers, Smith et al. described a liquid sheath interface697 and detailed its application for CZE-ESI mass spectrometric determination of peptides and proteins.g10 Henion et al. developed a liquid junction interface for coupling CZE-ESI mass s p e ~ t r o m e t r y . ~However, ~-~~ no interface to couple HPCE with an ESI mass spectrometer is commercially available. Such a device has been built in-house to interface a Beckman HPCE instrument with aVestec ESI mass spectrometer. The purpose of this paper is to examine the applicability of capillary zone electrophoresis (CZE) coupled with an ESI mass spectrometer to aid in rapid characterization of recombinant biotechnology derived proteins. (3) Tsuji, K. High-performance capillary electrophoresis of proteins. Sodium dodecyl sulfate-polyacrylamide gel-filled capillary column for the determination of recombinant biotechnology-derived proteins. J. Chromatogr. 1991,550, 823-830. (4) Feng, R.;Konishi, Y.; Bell, A. W. High Accuracy Molecular Weight Determination and Variation Characterization of Proteins up to 80 ku Mass Spectrom. 1991, 2, by Ionspray Mass Spectrometry. J. Am. SOC. 387-40 1. ( 5 ) Covey, T. R.; Bonner, R. F.; Shushan, B. I.; Henion, J. The determination of proteins, oligonucleotide and peptide molecular weights by ion-spray mass spectrometry. Rapid Comm. Mass Spectrom. 1988, 2, 249-256. (6) Smith, R. D.; Barinaga, C. J.; Udseth, H. R. Improved Electrospray Ionization Interface for Capillary Zone Electrophoresis-Mass Spectrometry. Anal. Chem. 1988,60, 1948-1952. (7)Smith, R. D.;Udseth, H. R.; Barinaga, C. J.; Edmonds, C. G. Instrumentation for high-performance capillary electrophoresis-mass spectrometry. J. Chromatogr. 1991,559, 197-208. (8) Loo, J. A.; Udseth, H. R.; Smith, H. D. Peptide and protein analysis by electrospray ionization-mass spectrometry and capillary electrophoresis-mass spectrometry. Anal. Biochem. 1989, 179, 404-412. (9) Smith, R.D.;Loo, J. A.; Edmonds, C. G.; Barinags, C. J.; Udseth, H. R. New Developments in Biochemical Mass Spectrometry. Anal. Chem. 1990,62, 882-899. (10) Loo, J. A.; Ogorzalck Loo, R. R.; Light, K. J.; Edmonds, C. G.; Smith. R. D. Multiulv Charged Neeative Ions bv Electrosurav Ionization of Polypeptides and-Protehs. Anal. Chem. 1992, 64, ai-88. (11)Lee, E. D.;Mueck, W.; Henion, J. D. On-line capillary zone electrophoresis-ion spray tandem mass spectrometry for the determination of dynorphins. J. Chromatogr. 1988, 458, 313-321. (12) Lee, E. D.;Mueck, W.; Henion, J. D. Liquid junction coupling for capillary zone electrophoresis/ion spray mass spectrometry. Biomed.Enuiron. Mass Spectrom. 1989, 18, 844-850. (13) Lee, E. D.;Kans, 0.; Covey, T. R.; Henion, J. D. Liquid junction coupling for capillary zone electrophoresis/ion spray spectrometry. U S . Patent 4,994,165, Feb 19, 1991.

0 1992 American Chemical Society

ANALYTICAL CHEMISTRY, VOL. 64, NO. 17, SEPTEMBER 1, 1992 Teflon Tubing 11\16' 0 D ) for make-up buffer

1865

Plastic (Delrin) support tubing

UE Capillary Column Stainless Steel '1'

UE Capillary

SO micrometer I d 140 micrometer 0 D

Stainless Steel Tubing 1/7 000' I d iii6- o D

Flgure 1. Schematlc diagram of the interface built in-house to couple a Beckman P/ACE 2050 high-performance capillary electrophoresis system to a Vestec Model 201A single-quadrupole mass spectrometer with an electrospray ionization source.

EXPERIMENTAL SECTION HPCE Instrumentation. A P/ACE System 2050 highperformance capillary electrophoresis (HPCE)instrument (Beckman, Palo Alto, CA) was used throughout the study. Each HPCE run involved a 5-5 nitrogen (0.5-psi)pressure injection of asample into a bare fused-silica capillary column. The compounds migrating in the column were monitored at 214 nm. The temperature in a column cartridge was maintained at 25 "C with a circulating coolant, and an electrophoretic run was conducted at a constant voltage of 20 kV (350 V/cm). The area under the protein peak was integrated by means of a program, developed in-house, residing in a VAX mainframe computer. Capillary Columns for CZE Analysis. A roll of fused-silica capillary tubing (50-pm i.d., 375-pm 0.d.) was obtained from Polymicro Technologies (Phoenix, AZ). Capillaries were cut to a few centimeters longer than the final column length of 57 cm and a window was created at 7 cm from the outlet end by removing the polyimide coating with a polyimide stripper (Model S200, Polymicro). The capillary was mounted in the column cartridge (Beckman) and cut to an effective column length of 50 cm. The capillary column was treated with 0.1 N aqueous NaOH for 5 min. After rinsing with water for 5 min, the column was filled with an electrolyte buffer for 5 min. Just prior to each sample analysis, the column was rinsed with the electrolyte buffer for 5 min. Reagents. Reagents used for the electrophoretic analysis were made of analytical reagent grade chemicals obtained from J. T. Baker Chemical Co. (Phillipsburg,NJ), Mallinckrodt,Inc. (Paris, KY), Burdick & Jackson (Baxter, Muskegon, MI), Sigma Chemical Co. (St.Louis, MO), and/or Aldrich Chemical Co., Inc. (Milwaukee, WI). Stock CZE buffer solutions were prepared by mixing 100 mM ammonium hydroxide and 100 mM acetic acid to a pH of approximately 9.0 and another at pH 10.0. A stock 100 mM ammonium biborate solution was also prepared by adjusting the pH to about 9.7 with 100mMammonium hydroxide. For the analysis of rpSt, the electrolyte buffer was prepared by mixing 3.0 mL of HzO, 5.0 mL of the stock buffer solution (pH 9.0)and 2 mL of acetonitrile. An electrolyte buffer for the analysis of rbSt was prepared by mixing 2.0 mL of acetonitrile, 6.0 mL of water, and 2.0 mL of stock ammonium biborate solution. Samples of recombinant bovine and porcine somatotropins (rbSt and rpSt) used in this experiment were manufactured by The Upjohn Co. (Kalamazoo, MI). The rpSt molecule consists of a single polypeptide containing 191 amino acid residues with two intrachain disulfide bridges. The total theoretical average molecular weight and the isoelectric point ( P I ) of the rpSt are 21 797.88 and 7.0, respectively. Separation of Impurities in Recombinant Somatotropins. The recombinant rbSt and rpSt were analyzed at a concentration of about 1 mg/mL in the native form. The sample was injected onto the column for 5 s (ca. 6 nL) under nitrogen pressure (0.5 psi) and electrophoreticallyseparated for 8 min under a constant voltage of 20 kV (350 V/cm, ca. 9.5 wA).

CZE-ESI Mass Spectrometry. CZE-ESI Interface. A schematic diagram of the CZE-ESI mass spectrometer interface is presented in Figure 1. The Vestecelectrospray probe assembly was modified by replacing the original stainless steel tubing with a 20-cm-long stainless steel tubing of 180-pm i.d. and 1.6-mm 0.d. (l/TW-in.i.d., l/Is-in. 0.d.). A zerodead-volume stainless steel "T" connector was attached to the stem of the stainless steel tubing for introduction of the makeup solution. At the tip of the probe/tubing, a 2-cm-long 27-gauge stainless steel needle (200pm i.d., 400-pm 0.d.) was connected. A 28-gauge stainless steel needle may also be used as the spray needle. The fused-silica capillary column of 50-pm i.d., 140-pm o.d., and approximately 1.2-1.5 m was threaded through the stainless steel probe/needle assembly. The optimum peak resolution was obtained when the capillary column was allowed to protrude approximately 200-300 pm from the tip of the nozzle. HPCE Instrument. For the CZE-ESI mass spectrometric analysis, the Beckman P/ACE 2050 HPCE instrument was modified by installing a switch to free float the ground. Capillary Column. A 50-wm4.d. and 140-pm-0.d.fused-silica capillary column approximately 1.2-1.5 m long (Polymicro) was used. A window was created at approximately 20 cm from the inlet end of the column. After inserting the capillary column into several sections of an another fused-silica capillary tubing (200-wmid., 375-wm 0.d.) to give mechanical strength at strategic places, the combined capillary column/tubing was mounted in the column cartridge (Beckman). The capillary column was treated with 0.1 N aqueous NaOH for 15 min. After rinsing with water for 15 min, the column was filled with an electrolyte buffer for 15 min. Just prior to each sample analysis, the column was rinsed with the electrolyte buffer for 15 min. Electrolyte B u f f e r . For the analysis of rpSt the electrolyte buffer was prepared by mixing 6.0 mL of HzO, 2.0 mL of the stock buffer solution (pH 9.0),and 2.0 mL of acetonitrile. A makeup solution containing 5% acetic acid in 50% methanol was pumped at a flow rate of 1-3 pL/min. The electrolyte buffer used for the analysis of rbSt was prepared by mixing 5.5 mL of water, 2.0 mL of acetonitrile, 2.0 mL of the stock buffer solution (pH 10.0), and 0.5 mL of the stock ammonium biborate solution. A makeup solution containing 1%acetic acid in 95% methanol was gravity fed at a flow rate of approximately 5-10 pL/min for the analysis of rbSt. CZE-ESI Mass Spectrometric Analysis. Approximately 1.5 pmol of rbSt or rpSt sample was injected into the capillary column and electrophoretically analyzed for approximately 40 min under a constant voltage of 23 kV (167 V/cm) to compensate for the ESI voltage of 3 kV. CZE-ESI Mass Spectrometric Operation. Peaks migrating in the capillary column were identified by use of a Vestec quadrupole mass spectrometer (Model 201A, Vestec Corp., Houston, TX) outfitted with an electrospray ionization (ESI) source. A

1868

ANALYTICAL CHEMISTRY, VOL. 64, NO. 17, SEPTEMBER 1, 1992

CZE-ESI interface was designed and built in-house. The basic design of the interface is described in detail in the Results and Discussion. To obtain a consistent and stable spray in the ionization chamber, a sample flow rate of approximately 1-10 pL/min was found optimum for the Vestec ESI source. Since the electroosmotic flow of the CZE is a low nanoliter per minute level, a makeup solution is required to maintain an effectivespray. The makeup solution was 1?6 acetic acid in methanol-water (955) solution or 5 5% aceticacid in methanol-water (1:l)solution;both contained approximately 3 pg/mL bradykinin. The makeup solution was pumped using a syringe pump (Model 341B, Sage Instruments) or gravity fed at a flow rate of approximately 1-10 pL/min. Bradykinin was included in the makeup solution to assist in fine tuning of the ESI mass spectrometer. In the Vestec electrospray source, the electrical field was generated by the application of 2-3 kV to the electrospray needle. Conditions for the ESI mass spectrometer were as follows: block temperature, 255 "C; spray chamber temperature, 44 "C; lens temperature, 54 "C; spray voltage, 2.60 kV; spray current, 0.240 PA; nozzle voltage, 160 V. The electroapraymass spectra (massrange 1000-2000Da) were scanned every 2 s and data acquired with the Vector Two data system (Teknivent Corp., Maryland Heights, MO). Three or four scans corresponding to a given CZE peak were averaged and transmitted to the Harris 800 dual CPU computer (Fort Lauderdale, FL) for further analysis. The electrospray mass spectra were processed using programs developed in-house. In one approach, centroids of the individual clusterswere determined and the number of charges for each cluster calculated as described previ~usly.~ The average molecular weights were then computed. In the other approach, the electrospray mass spectra were deconvoluted as described by Fenn and co-workers.14 Molecular weights of rbSt, rpSt, and their homologues and analogueswere determined from the multiply-charged ion clusters detected in the ESI mass spectra by both m e t h ~ d s . ~ J ~

A

0

5

10

15

Minutes

Figure 2. Separationof minor components in a sample of rbSt by CZE as monitored at 214 nm. Conditions: 187 Vkm; 13 pA; column temperature, 25 OC; migrationdlstance, 100 cm; fused-silica capillary, 50-pm 1.d.; electrolyte buffer, 20% acetonitrlie in 20 mM ammonium biborate, pH 9.7. Peaks A and B contain mono- and dldeamido derivatives, respectively.

Table I. Precision of Peak Migration by the CZE Determination of rpSt peak migration peak migration run no. time (min) run no. time (min) 1 4.88 7 4.82 2 4.85 8 4.81 3 4.84 av 4.839 4 4.84 76 RSD 0.43 5 4.84 6 4.83

RESULTS AND DISCUSSION Analysis of Recombinant Proteins by Capillary Electrophoresis with UV Detection. A typical electropherogram indicating separation of the major and minor compounds in an rbSt sample is presented in Figure 2. A similar electropherogram was also obtained for the analysis of rpSt. By analogy of the CZE studies of the bovine and human growth hormones17J8 and the peak migration time of the deamido species, the minor peaks A and B were tentatively identified to contain mono- and dideamido forms, respectively. (See action on CZE-ESI Mass Spectrometric Characterization for further discussion). When this part of the electropherogram was expanded, peaks A and B appeared to contain approximately five and two compounds, respectively. Identification of each of these minor compounds would require improvementa in the peak resolution. For quantitative analysis of rpSt, the standard curve was linear ( r > 0.991) for protein concentrations from 7-300 fmol. From the data a linear regression equation resulted in Y = 156.4X + 217 50, where Y represents the peak area and X represents the rpSt concentration in pg/mL. The relative (14) Mann, M.; Meng, C. K.; Fenn, J. B. Interpreting Mass Spectra of Multiply Charged Ions. Anal. Chem. 1989,61, 1702-1708. (15) Bergman, T.; Agerberth, B.; Joernvall, H. Direct analysis of peptides and amino acids from capillary electrophoresis. FEBS Lett. 1991, 283, 100-103. (16) Tran, A. D.; Huynth, 0.T.;Park, S.; Ryall, R. R.; Lisi, P. J.; Lane, P. A. Separation of carbohydrate-mediated microheterogenity of recombinant human erythropoietin by free solution capillary electrophoresis: effects of pH and buffer type. J. Chromatogr. 1991, 542, 459-471. (17) Grossman, P. D.; Colburn, J. C.; Lauer, H. H.; Nielsen, R. G.; Riggin, R. M.; Sittampalm, G.S.; Richard, E. C. Application of FreeSolution Capillary Electrophoresis to the Analytical Scale Separation of Proteins and Peptides. Anal. Chem. 1989,61, 1186-1194. (18)Secchi, C.;Biondi, P. A.;Negri, A.;Borroni,R.; Ronchi, S. Detection of Desamido Forms of Purified Bovine Growth Hormone. Int. J.Peptide Protein Res. 1986,28, 298-306.

standard deviation (RSD) of the peak migration time was less than 1%(Table I). Depending upon the volume and concentration of rpSt injected, theoretical plates of the major rpSt peak ranged from 149 OOO to 415 OOO m-1. The highest column performance was obtained using a 5-s (3-nL) injection of a solution containing 500-1000 pg/mL rpSt, which corresponds to 75150-fmol injection. Due to column overloading, a decrease in column performance was noted when over 300 fmol of rpSt was injected. Also, injection of less than 15 fmol of protein caused loss of theoretical plates, due perhaps to nonspecific adsorption of the protein on the column surface. An increase in the injection volume by 3 (158) also decreased the column performance. The detection limit of the capillary electrophoretic method using UV detection for the analysis of rbSt and rpSt samples is approximately 5 fmol. At this concentration, a signal to noise ratio of 3 to 1 was observed. CZE-ESI Mass Spectrometry. Although the Vestec electrospray (ESI) mass spectrometer is technically an atmospheric pressure mass spectrometer, it requires a probe of approximately 30 cm long to reach the ion spray chamber. We examined several different interface designs, including that of Henion;13 the HPCE-ESI interface described here is similar in concept to that of Smith et al.6 No operational advantage or improvement in peak resolution was noted when an all nonmetallic interface, similar to that of Smith et al.,7 was constructed and examined. The height differential between an electrolyte buffer level in the HPCE instrument and the ESI spray nozzle is approximately 12-15 cm. Although Smith et aL7maintained the differential of about 1.5 cm, we experienced no adverse effect on peak resolution at our height differential.

ANALYTICAL CHEMISTRY, VOL. 64, NO. 17, SEPTEMBER 1, 1992

1867

Intensity 2200000

2 150000

2 100000

2050000

2000000

Ilh

1950000

A 190000c

185000C

180000~

175000( 0.5

5:o

Id. 0

.

15.0

18.0

T i . * Flguro 3. CZE-ESI mass spectrometertotal Ion current electropherogram(A)and a conventlonalCZE UV electropherogram(B) of rpSt. Conditions: fused-silica capillary column: 50-pm 1.d. by (A) 120 cm and (E)50 cm long; electrolyte buffer, 20% acetonitrile In 20 mM ammonium acetate, pH 9.0.

In order to obtain a reasonable peak resolution and to maintain a stable, fine spray for ESI mass spectrometric analysis, the internal diameter of the spray needle must be as narrow as possible and the fused-silica capillary column be slightly protruding out of the spray nozzle. These requirements dictate the use of a thin-walled capillary column. The frail capillary column required mechanical reinforcement at strategic places in the column cartridge to prevent breakage. This was accomplished by insertion of the fused-silicacapillary column into pieces of an another, larger diameter fused-silica capillary tubing prior to installation in the column cartridge (Beckman). CZE-ESI Mass Spectrometric Operation. ESI mass spectrometry requires use of a volatile electrolyte buffer which has some electric resistance. Therefore, the followingvolatile buffers were examined: ammonium acetate, ammonium formate, ammonium carbonate, ammonium trifluoroacetate, triethylamine with formic acid, and triethylamine with ammonium hydroxide. Ammonium acetate buffer at pH 9.0gave the best column performance for the analysis of rpSt. Acetonitrile (20% of solution) was added to improve the peak selectivity.15J6 For an optimum peak resolution of rbSt impurities, inclusion of 5 mM ammonium biborate was required in an electrolyte buffer of 20% acetonitrile and 20 mM ammonium acetate buffer at pH 10.0. Composition of electrolyte buffer dictates the selection of the makeup solution to minimize undue electric current feedback to the ESI power supply. The makeup solution composed of methanol-wateracetic acid (47.547.55.0)allows the use of an electrolyte buffer containing 20% acetonitrile in 20 mM ammoniumacetate buffer (pH 9.0). However, high methanol containing makeup solution (methanol-water-

acetic acid, 95:4:1)is required to operate the CZE with an electrolyte buffer containing 5 mM ammonium biborate in 20% acetonitrile and 20 mM ammonium acetate (pH 10.0). The presence of 5 mM ammonium biborate significantly improves the resolution of impurity peaks in rbSt samples. A CZE-ESI mass spectrometer total ion current electropherogram and a conventional CZE UV electropherogram of the same sample of rpSt are presented in Figure 3A,B, respectively. It is evident from these electropherograms that the peak resolution is not compromised by coupling an HPCE to an ESI mass spectrometer via the interface described. Differences of peak migration time between the mass spectrometric tracing (Figure 3A) and the HPCE electropherogram (Figure 3B) can be attributed to the differences in length of the capillary columns, 1.2 m vs 50 cm. From the amount of rpSt sample (1.5pmol) injected onto the capillary column and the heights of the impurity peaks, the sensitivity of the CZE-ESI mass spectrometer was estimated to be approximately 100 fmol. CZE-ESI Mass SpectrometricCharacterization. The CZE-ESI mass spectrum of the major rpSt peak is shown in Figure 4. The multiply-charged ion clusters ranging from m / z 1363.2(the cluster with 16 charges) to 1982.5 (the cluster with 11 charges) were detected. The average molecular weight of 21 798.3 f 3.6 for rpSt determined in this study was close to the theoretical value of 21 797.9. CZE-ESI mass spectrometry was also used to characterize recombinant bovine somatotropin (rbSt). The average molecular weight of rbSt, 21 812.6f 4.3,determined by the CZE-ESImaes spectrometer also agreed well with that of the theoretical value of 21 812.0. Deconvolution of the ESI mass spectrum detected compounds with average molecular weights of 21 581 (M - 218)

ANALYTICAL CHEMISTRY, VOL. 64, NO. 17, SEPTEMBER 1, 1992

1868

m 60 0

V

c a 50

m >

.-

1678.

Y

40

0 0

e

1817.3

30

20

1558.1

1

1454.5

1982.5

,

.

10

;

0

1100

,

,

,

.

l

.

1200

.

.

.

,

.

1300

.

.

.

l

.

1400

.

.

.

,

.

.

.

,

1500

l

,

1600

'

,

,

.

l

,

.

,

1700

.

l

,

1800

,

,

.

l

.

1900

.

,

.

~

2000

nfz Flguro 4. CZE-ESI mass spectrum of the major rpSt peak showing multiply-charged ion clusters (mlz). Conditions: electrolyte buffer, 20% acetonltrile in 20 mM ammonium acetate, pH 9.0; makeup solution, 5 % acetic acld in 50% methanol; calibrant, bradykinin 3 pg/mL In makeup solution; flow rate of the makeup soiutlon, 1-3 pL/min; block temperature, 255 O C ; spray chamber temperature, 44 O C ; lens temperature, 54 O C ; spray voltage, 2.60 kV; spray current, 0.240 pA; nozzle voltage 160 V.

B

I I I

I I I

I I 5

r

21194

I

1

I

1

I

22389

Flguro 5. Distrlbutlon of the average molecular weights deconvoiuted from the CZE-ESI mass spectrum (refer to Figure 3) Indicating the presence of three compounds comlgrating wlth the major rpSt peak (refer to Figure 1): (A) M - 218; (B) rpSt (M = 21 798); (C) M 4- 42.

and 21 841 (M + 42) comigrating with the major rpSt peak (Figure 5). The compound corresponding to M - 218,peak A, was tentatively identified as a lower molecular weight homologue with the loss of ALA-PHEfrom either the C- or N-terminus. M + 42,peak C, is possibly a monoacetylated derivative of rpSt. Detection of monoacetylated rpSt in the

major rpSt peak was perplexing since the mass to charge ratio of acetylated rpSt is significantly different from the parent rpSt. Thus, CZE should have separated these compounds. Indeed, when an enriched sample of acetylated rbSt and the parent rbSt were analyzed by CZE under conditions identical to those of rpSt, these two compounds were clearly

ANALYTICAL CHEMISTRY, VOL. 64, NO. 17, SEPTEMBER 1, 1992

1880

B

i I

1

22471

21182

Flgure 6. Dlstrlbution of the average molecular welghts deconvoluted from the CZE-ESI mass spectrum of a rbSt sample (refer to Figure 1) lndlcatlng the presence of a comlgrating monooxkilzed derivative (M 16) In the major rbSt peak: (A) rbSt (M = 21 813); (b) M 16.

+

+

separated. The peak migration time of rbSt and acetylated rbSt were 3.18 and 4.2 min, respectively. Although the makeup solution contains acetic acid for protonation of protein, acetylation of rpSt during the ESI mass spectrometric analysis has been ruled out. This is because no compound with the mass of M + 42 was detected when rbSt as well as another sample of rpSt were analyzed under the identical CZE-ESI mass spectrometric conditions. Thus, the compound with the mass M + 42 (f4)may be an analogue of rpSt differing by one amino acid, possibly substituted during posttranslational processes. The ability of CZE-ESI mass spectrometry to distinguish proteins differing by one amino acid was examined. Out of five analogues of rpSt studied, the CZE clearly differentiated two analogues by differences in the peak migration time. However, the remaining three analogues comigrated with the parent rpSt peak. The average molecular weights of these comigrating analogues determined by CZE-ESI mass spectrometry were 21 732.7 f 6.3,21 772.6 f 4.7,and 21 824.9 f 3.4. These values are all within 3 mass units of the theoretical values. Proteins are known to be posttranslationally modified by deamidation. The major sites of deamidation of rbSt were reported to occur to ASN at positions 13 and 148 and GLN at position 140.l8 Chemical formation of isoaspartate from ASN a t position 99 in both rbSt and rpSt was reported to occur at neutral or alkaline pH c0ndition.1~ The CZE separation of the deamido homologues (refer to Figure 2) may be due to an increased net negative charge by the replacement of the amide group with a carboxyl group. Identification of the mono- and dideamidated compounds by (19) Violand, B. N.;Schlitter, M. R.; Toren, P. C.; Siegel, N. R. Formation of Isoaapartate 99 in Bovine and Porcine Somatotropins. J. Protein Chem. 1990,9,109-117.

the ESI mass spectrometer is beyond the capability (f4.0 units) of our ESI mass spectrometer. Also the CZE separation is not specific to the deamidated position on the molecule. For the specific identification of the deamidated site, CZEMS and/or HPLC-MS analysis of a trypsin-treated sample would be required. rbSt and rpSt may also degrade by oxidation of the thioether of MET. Oxidation of MET a t position 5 of rbSt was reported to be the most active followed in decreasing order by positions 149,124,and 179.*O Since oxidation of MET does not alter the total net charge of rbSt or rpSt, CZE is incapable of separating these homologues. However, the ESI mass spectrometer was able to detect the presence of monooxidized compounds, differing by the mass of 16,in a sample of rbSt enriched with the monooxidized homologue (Figure 6). Dioxidized compound was similarly detected in an another sample.

CONCLUSION The CZE-ESI mass spectrometric data obtained in this study clearly demonstrated that high theoretical plates, 400 000-450 OOO m-l, of capillary electrophoresis alone does not guarantee the purity of the peak and that it takes another analytical tool, e.g. mass spectrometry, to determine its composition. It must be stressed that the CZE-ESI mass spectrometric identification is tentative and that the identity of the compounds must be confirmed by other analytical (20) Cascone, 0.; Biscoglio, M. J.; Bonino, de J.; Santome, J. A. Oxidation of Methionine Residues in Bovine Growth Hormone by Chloramine-T. Int. J. Peptide Protein Res. 1980,16, 299-305.

1870

ANALYTICAL CHEMISTRY, VOL. 64, NO. 17, SEPTEMBER 1, 1992

techniques, e.g. amino acid sequencing, tryptic digest-mass spectrometry, etc. Nevertheless, the results obtained in this study clearly electrophoretic and demonstrate that ESI mass spectrometry complement each other for separation and rapid characterization of the recombinant proteins and closely related derivatives. We are confident that CZE-ESI mass spectrometry will play an indispensable role not only in protein chemistry but also in many analytical applications.

ACKNOWLEDGMENT We sincerely thank Prof. Jack D. Henion of Cornel1 University, Ithaca, NY,for CoUrtesl’ presented to us during our visit to his laboratory. The technical assistance of R. J. Little of Control Biotechnology and J. F. Locker of Laboratory Instrument Support is appreciated. RECEIVED for review January 6, 1992. Accepted May 6, 1992. Registry No. Somatotropin, 9002-72-6.