Anal. Chem. 2007, 79, 2714-2721
18
O Labeling Method for Identification and Quantification of Succinimide in Proteins
Gang Xiao, Pavel V. Bondarenko, Jaby Jacob, Grace C. Chu, and Dirk Chelius*
Department of Pharmaceutics, Amgen, Inc., Thousand Oaks, California 91320
We have developed a new method for identification and quantification of succinimide in proteins. The method utilizes 18O water to monitor succinimide hydrolysis. 18Olabeled isoaspartic acid and aspartic acid peptides were produced by hydrolysis of a succinimide-containing protein in 18O water (H218O) followed by tryptic digestion in regular water (H216O). The peptides that had 18O incorporated were 2 Da heavier than their 16O native counterparts. The mass difference was detected and quantified by electrospray time-of-flight mass spectrometry. The amount of 18O incorporation into the isoaspartic acid- and aspartic acid-containing peptides was used to quantify the amount of succinimide present in the native sample. The method was applied to analyze a degraded recombinant monoclonal antibody, which exhibited the accumulation of succinimide after storage in mildly acidic buffers at elevated temperatures for a few weeks. We unambiguously identified amino acid residue 30 located in the antibody light chain as the site of aspartic acid isomerization. At this site, there were 20% isoaspartic acid and 80% aspartic acid detected by peptide mapping in the degraded sample (8 weeks, 45 °C, pH 5.0). Hydrolysis in 18O water showed that 80% of the isoaspartic acid and 6% of the aspartic acid had 18O incorporated. The only explanation of 18O incorporation was the presence of succinimide in the sample. Together, a total of 21% (0.8 × 20% isoaspartic acid + 0.06 × 80% aspartic acid) of aspartic acid residue 30 was found to be present in the form of succinimide in this degraded sample. As a control, the same sample, analyzed using regular 16O water did not show any incorporation of 18O water. By monitoring the amount of 18O-labeled isoaspartic acid and aspartic acid over time under both denaturing and native conditions at pH 8.2, we found that, at denaturing conditions, succinimide at light chain residue 30 hydrolyzed very rapidly (in less than 5 s), but slower (succinimide halflife of ∼ 6 h) under native conditions. We also found that, under denaturing conditions, succinimide hydrolyzed at an isoaspartic acid/aspartic acid ratio of 3.5:1, but hydrolyzed almost exclusively to aspartic acid under native conditions. This finding indicates that protein structure plays an important role in the kinetics of succinimide hydrolysis as well as in the generation of the hydrolysis products isoaspartic acid and aspartic acid.
* To whom correspondence should be addressed. Phone: (805) 447-7532. Fax: (805) 447-3401. E-mail:
[email protected].
2714 Analytical Chemistry, Vol. 79, No. 7, April 1, 2007
The isomerization of aspartic acid (Asp) to isoaspartic acid (βaspartate, isoAsp) is a spontaneous reaction that can alter protein structure and function.1-7 Asp residues can form a cyclic imide intermediate (Figure 1), which spontaneously hydrolyzes in peptides to a mixture of isoAsp/Asp typically at a ratio of 3:1.8,9 To complicate matters further, the isomerization is accompanied by partial racemization of the newly formed isoAsp and Asp residues, which occurs preferentially at the stage of succinimide intermediate.10 The effects of pH and temperature on the formation of isoAsp have been reported in great detail, showing that low pH and high temperature accelerate the isomerization reaction. Unlike the isomerization reaction of asparagine (Asn) residues, the solvent (e.g., dielectric constants) does not seem to affect the rate of isomerization of Asp.11 The effect of different amino acids Cterminal of aspartic acid on succinimide formation and isomerization has been reported, indicating that glycine and serine residues are most favorable for succinimide formation for both Asp and Asn residues.10,12-14 Several methods for the identification of isoAsp have been reported in the literature. Typically, Edman sequencing is used to detect and identify isoAsp residues.15-17 In this approach, the inability of isoAsp to react with phenylisocyanate to cleave the (1) Manning, M. C.; Patel, K.; Borchardt, R. T. Pharm. Res. 1989, 6, 903-918. (2) Ahern, T.; Manning, M. C. In Stability of protein pharmaceuticals: Part A, Chemical and Physical Pathways of Protein Degradation; Plenum: New York, 1992. (3) Cleland, J. L.; Powell, M. F.; Shire, S. J. Crit. Rev. Ther. Drug Carrier Syst. 1993, 10, 307-377. (4) Pearlman, R.; Wang, Y. J. In Formulation, Characterization, and Stability of Protein Drugs: Case Hstory; Plenum: New York, 1996; Vol. 9. (5) Liu, D. T. Trends Biotechnol. 1992, 10, 364-369. (6) Aswad, D. W. In Deamidation and Isoaspartate Formation in Peptide and Proteins; CRC Press: Boca Raton, FL, 1995. (7) Aswald, D. W.; Paranandi, M. V.; Schurter, B. T. J. Pharm. Biomed. Anal. 2000, 21, 1129-1136. (8) Patel, K.; Borchardt, R. T. Pharm. Res. 1990, 7, 787-793. (9) Oliyai, C.; Borchardt, R. T. Pharm. Res. 1993, 10, 95-102. (10) Geiger, T.; Clarke, S. J. Biol. Chem. 1987, 262, 785-94. (11) Brennan, T. V.; Clarke, S. Protein Sci. 1993, 2, 331-338. (12) Stephenson, R. C.; Clarke, S. J. Biol. Chem. 1989, 264, 6164-6170. (13) Paranandi, M. V.; Guzzetta, A. W.; Hancock, W. S.; Aswad D. W. J. Biol. Chem. 1994, 269, 243-253. (14) Kosky, A. A.; Razzaq, U. O.; Treuheit, M. J.; Brems, D. N. Protein Sci. 1999, 8, 2519-2523. (15) Zhang, W.; Czupryn, J. M.; Boyle, P. T. Jr.; Amari, J. Pharm. Res. 2002, 19, 1223-1231. (16) Sadakane, Y.; Yamazaki, T.; Nakagomi, K.; Akizawa, T.; Fujii, N.; Tanimura, T.; Kaneda, M.; Hatanaka, Y. J. Pharm. Biomed. Anal. 2003, 30, 18251833. (17) Di Donato, A.; Ciardiello, M. A.; de Nigris, M.; Piccoli, R.; Mazzarella, L.; D’Alessio, G. J. Biol. Chem. 1993, 268, 4745-4751. 10.1021/ac0617870 CCC: $37.00
© 2007 American Chemical Society Published on Web 02/22/2007
Figure 1. Scheme of the isomerization of Asp residues and the incorporation of 18O to isoAsp and Asp during succinimide hydrolysis used in this study.
peptide bond between the N-terminal amino acid and the next amino acid is used for identification. Edman sequencing always stops at isoAsp residues. Another approach for the identification is the use of protein-L-isoaspartyl methyltransferase, an enzyme, which catalyzes the transfer of the methyl group of S-adenosylL-methionine to isoAsp sites. After methylation, the difference in mass can be easily detected by tryptic peptide mapping and mass spectrometric analysis.18 More recently, Asp-N digestion was used in combination with Edman sequencing to identify isoAsp formation. Zhang et al. showed that Asp-N does not cleave at the N-terminus of isoAsp.15 The identification of the succinimide intermediate in proteins is challenging because it hydrolyzes rapidly under neutral to basic pH conditions. The same conditions are typically used for enzymatic digestion for the identification of the modified peptides, which makes the identification of succinimide nearly impossible. One method for the identification has been reported in the literature using cation-exchange chromatography to separate the succinimide intermediate from its native counterpart, followed by enzymatic digestion and identification of the hydrolysis products isoAsp and Asp in the succinimide-containing fraction.19,20 However, those identification techniques are indirect and can easily miss succinimide if the succinimide-containing molecules cannot be separated by the ion-exchange chromatography or if the succinimide intermediate does not hydrolyze to isoAsp. Here we present a new approach for the identification and quantification of succinimide and its hydrolysis products by (18) Johnson, B. A.; Shirokawa, J. M.; Hancock, W. S.; Spellman, M. W.; Basa, L. J.; Aswad, D. W. J. Biol. Chem. 1989, 264, 14262-14271. (19) Hui, J. O.; Chow, D. T.; Markell, D.; Robinson, J. H.; Katta, V.; Nixon, L.; Chang, B. S.; Rohde, M. F.; Haniu, M. Arch. Biochem. Biophys. 1998, 358, 377-384. (20) Harris, R. J.; Kabakoff, B.; Macchi, F. D.; Shen, F. J.; Kwong, M.; Andya, J. D.; Shire, S. J.; Bjork, N.; Totpal, K.; Chen, A. B. J. Chromatogr., B: Biomed. Sci. Appl. 2001, 75, 233-245.
utilizing heavy-isotope 18O water. This method was applied to analyze a recombinant monoclonal antibody, which exhibited the accumulation of succinimide after storage in mildly acidic buffers at elevated temperatures. 18O water-exchange experiments together with 13C NMR have been reported to study the pH-induced succinimide hydrolysis of small peptides.21 The incorporation of 18O has also been widely used for quantification in proteomics research because the chromatographic and mass spectrometric properties of 18Ocontaining peptides do not change compared to those that contain 16O.22 In this study, we produced 18O-labeled isoAsp- and Aspcontaining peptides from a recombinant monoclonal antibody by hydrolysis of a succinimide in 18O water (H218O) at pH 8.2 followed by tryptic digestion in regular water (H216O). The analysis by electrospray time-of-flight mass spectrometry showed that the peptide that contained isoAsp or Asp with 18O incorporated was 2 Da heavier than its 16O native counterpart. The amount of 18O incorporation into the isoAsp and Asp peptides was used to quantify the amount of succinimide present in the native sample. We also studied the rate of succinimide hydrolysis of this recombinant antibody under denaturing and native conditions at pH 8.2, demonstrating that this new method can be used for kinetic studies under different conditions. EXPERIMENTAL PROCEDURES Materials. The recombinant human monoclonal antibody analyzed in this study was expressed at Amgen and purified using standard manufacturing procedures. It has been stored in mildly acidic buffers (pH 5.0) at 45 °C for 8 or 12 weeks before analysis. The control sample was stored frozen at -70 °C in the same (21) Xie, M.; Velde, D. V.; Morton, M.; Borchardt, R. T.; Schowen, R. L. J. Am. Chem. Soc. 1996, 118, 8955-8956. (22) Stewart, I. I; Thomson, T.; Figeys, D. Rapid Commun. Mass Spectrom. 2001, 15, 2456 - 2465.
Analytical Chemistry, Vol. 79, No. 7, April 1, 2007
2715
buffer. 18O water was purchased from Cambridge Isotope Laboratories, Inc (97% enriched, lot no. WP-05-36). Guanidine hydrochloride and ammonium bicarbonate were from Mallinckrodt (St. Louis, MO). Dithiothreitol (DTT) was from MP Biomedicals (Irvine, CA), and iodoacetic acid (IAA) was from Sigma (St. Louis MO). 18O Incorporation during Succinimide Hydrolysis at Different Time Points. A 33-µL sample of the recombinant monoclonal antibody at 30 mg/mL was taken and dried using a Speed Vac centrifuge to remove 16O water. The sample was resuspended in 18O water and dried by a Speed Vac centrifuge again to ensure the residual 16O water has been removed. The sample was then resolubilized in 0.5 mL of buffer that was prepared in 18O water (97% enriched) and contained 100 mM NH4HCO3 and 8 M guanidine hydrochloride pH 8.2. No guanidine hydrochloride was added to study the hydrolysis under native conditions. The resuspended sample was incubated at 37 °C. Aliquots of 50 µL were taken from the sample at different time points as indicated in the text and were immediately frozen in methanol/dry ice to stop the succinimide hydrolysis reaction. After 18O water was removed from all the frozen aliquots of the sample using a Speed Vac centrifuge, these aliquots were ready for reduction, alkylation, and tryptic digestion in regular water for peptide mapping analysis. Reduction, Alkylation, and Tryptic Digestion. Following 18O incorporation, the dried aliquots of the sample were suspended in 0.5 mL of buffer containing 7.5 M guanidine hydrochloride, 0.25 M Tris-HCl, and 1 mM EDTA at pH 7.5. A 5-µL sample of 0.5 M DTT solution was added followed by incubation at 37 °C for 30 min. All the aliquots were then cooled to room temperature. After 13 µL of 0.5 M IAA solution was added, the aliquots were placed at room temperature for 40 min in the darkness. All the samples were then buffer exchanged to 1 mL of 0.1 M Tris-HCl (pH 7.5) by NAP-5 column (Amersham BioSciences, Uppsala, Sweden) following the manufacturer’s protocol. Tryptic digestion was performed for 4 h at 37 °C using an enzyme/protein weight ratio of 1:50. The same amount of trypsin was added a second time after 2 h of incubation. The digests were frozen at s20 °C and analyzed together. Liquid Chromatography Tandem Mass Spectrometry (LC/ MS/MS) Analysis of Tryptic Peptides. The method for the analysis of the tryptic peptides has been described before.23 Briefly, the tryptic peptides were separated on a Polaris C18 ether column (250 × 2.0 mm Varian, Torrance, CA) using a linear gradient from 0 to 65% B over 195 min. Solvent A was 0.1% TFA in water, and solvent B was 0.089% TFA, 90% acetonitrile (Baker, Phillipsburg, NJ) in water. Before sample injection, the column was equilibrated with 100% solvent A. The column temperature was maintained at 50 °C. The flow rate was 0.2 mL/min, and 20 µg of protein digest was injected onto the column for analysis. The HPLC was directly coupled to a Finnigan LCQ Deca ion trap mass spectrometer (Thermo Electron, San Jose, CA) equipped with an electrospray ionization source. The spray voltage was 4.5 kV, and the capillary temperature was 250 °C. The fragmentation mass spectra were obtained using ion trap collision energies of 35%. Peptides were identified automatically by two different computer programs. SEQUEST algorithm of BioWorks version
3.1 (Thermo Finnigan) was used to correlate the experimental tandem mass spectra against theoretical tandem mass spectra from a database. Software written in-house was also employed to correlate the experimental tandem mass spectra against theoretical tandem mass spectra generated from the known antibody amino acid sequence for peptide identification.24,25 Liquid Chromatography Electrospray Time-of-Flight Mass Spectrometry (LC-ESI-TOF) Analysis of Tryptic Peptides. Agilent 1100 HPLC (Agilent, Palo Alto, CA) was coupled to a Waters LCT Premier equipped with an ESI source. A Jupiter C18 column (250 × 1.0 mm, from Phenomenex, Torrance, CA) with 5-µm nominal diameter and 300-A pore size resin was used for analysis. The solvents, gradient and column temperature were the same as those in the LC/MS/MS analysis. The flow rate was 50 µL/min, and a total of 5 µg of protein digest was injected onto the column for analysis. The analysis on the LCT Premier was carried out in V mode with instrument resolution at 5000. The capillary and sample cone voltages were set at 2600 and 125 V, and the desolvation and source temperatures were set at 350 and 80 °C, respectively. All other voltages were optimized to provide maximal signal intensity for peptide analysis in V mode. All raw data were processed by Waters MassLynx 4.0 software. Fluorescence Spectroscopy. Photon Technology International fluorometer QM6 model (Birmingham NJ) equipped with a LPS-220B light source unit, MD-5020 motor drive unit, SC-500 shutter control unit, and a BryteBox detector unit was used for fluorescence spectroscopy. The data were analyzed using FeliX 32 fluorescence analysis software. The excitation wavelength was 280 nm and the emission wavelength was set to 357 nm to monitor changes in tryptophan fluorescence. Unfolding is initiated by mixing the folded antibody with denaturing buffer (8 M guanidine hydrochloride, 100 mM NH4HCO3, pH 8.2), which is continuously stirred with a magnetic stirrer during the measurement. The denaturing buffer is equilibrated at the appropriate temperatures (37 or 10 °C) in the fluorometer sample holder before starting the measurement and adding the folded protein. The final protein concentration used in these measurements is 1 mg/mL.
(23) Chelius, D.; Rehder, D. S.; Bondarenko, P. V. Anal. Chem. 2005, 77, 60046011.
(24) Zhang, Z. Anal. Chem. 2004, 76, 3908-3922. (25) Zhang, Z. Anal. Chem. 2004, 76, 6374-6383.
2716
Analytical Chemistry, Vol. 79, No. 7, April 1, 2007
RESULTS AND DISCUSSION Method for Identification of Succinimide in Proteins. Liquid chromatography mass spectrometry of the tryptic digests of a recombinant monoclonal antibody was used to monitor chemical degradation during storage at elevated temperatures. Figure 2 shows the detection of a new peak eluting at 97.5 min in the tryptic peptide map of a degraded sample (45 °C, pH 5.0, 8 weeks) in comparison with a control sample stored at -70 °C. This new peak had the same molecular weight and very similar tandem mass spectra (data not shown) as the peptide ASQSVD30SNLAWYR eluting at 100 min. Therefore, based on this result and previously reported data stating that isoAsp-containing peptides tend to elute earlier than Asp-containing peptides, we predicted isomerization of Asp-30 to isoAsp to be the cause for the formation of this new peak.23 No other modification of the protein was detected during peptide mapping. Cation-exchange chromatography of this degraded antibody showed an increase of a post main peak species compared to the
Figure 2. Reversed-phase chromatograms with UV absorbance at 215 nm of the tryptic peptides of the recombinant monoclonal antibody stored at 45 °C for 8 weeks (dotted back line) and stored at -70 °C (back line). A new peak eluting at 97.5 min was detected and identified as the isoform of peptide ASQSVD30SNLAWYR that eluted at 100 min. Mass spectrometry analysis of this isoform showed that it had isoAsp at residue 30 (see details in the text).
Figure 3. Tandem mass spectra for the isoAsp peptide ASQSVD30SNLAWYR (A) with 18O incorporation and (B) without 18O incorporation. Data were obtained by collision-induced dissociation of the (M + 2H)2+ precursor ions (m/z 749) in the LCQ Deca ion trap. 18O incorporation into isoAsp is revealed by the mass spectra that shows y8-ions (containing isoAsp) in (A) are 2 Da heavier than the corresponding y-ions in (B). The y7-ions (not containing isoAsp) have the same m/z in both (A) and (B).
control sample (data not shown). The conversion of Asp to succinimide changes the charge and introduced a basic postpeak variant in the cation-exchange chromatogram.19,20 Since conversion of Asp to isoAsp does not change the charge state of the molecule, we surmised that succinimide was present in the degraded sample. Succinimide hydrolyzes rapidly under neutral and alkaline conditions to form isoAsp and Asp, making it almost impossible to
detect succinimide by tryptic peptide mapping, which requires a pH above 7.0 for a long period of time (4 h). Efforts to perform reduction and alkylation using tris(2-carboxyethyl)phosphine hydrochloride and 1-cyano-4-dimethylaminopyridinium tetrafluoroborate followed by pepsin digestion at pH 3 to avoid hydrolysis of succinimide failed in detecting the succinimide intermediate. Analytical Chemistry, Vol. 79, No. 7, April 1, 2007
2717
To overcome this difficulty, we developed a method for the identification and quantification of succinimide utilizing 18O water to monitor the succinimide hydrolysis. The rationale is to incorporate 18O into the succinimide hydrolysis products isoAsp and Asp during succinimide hydrolysis in 18O water (Figure 1). Then, after different durations of incubation, the protein is buffer exchanged into regular 16O-containing buffer for tryptic digestion and peptide mapping analysis. The tryptic peptides that have 18O incorporated into the Asp or isoAsp or both are 2 Da heavier than the 16O-containing counterparts, which can be detected and measured by LC-ESI-TOF. The amount of 18O incorporation into isoAsp and Asp can be quantified and directly correlated to the amount of succinimide that is present in the sample. The degraded recombinant monoclonal antibody was denatured, reduced, and alkylated in the presence of 18O water as described in Experimental Procedures. Figure 3 shows MS/MS spectra of peptide ASQSVD30SNLAWYR identifying the presence of succinimide at residue 30 because of the incorporation of the 18O into isoAsp-30. The y-ions (y8, y10) containing residue 30 were 2 Da heavier than the corresponding y-ions of the peptides from the sample prepared in 16O buffer, while y-ions (y7, y3, y4) not containing residue 30 had the same m/z value. The b-ion series of both the 18O and 16O peptides also had the same results, showing that ions containing Asp30 residue had an increase of 2 Da while ions that did not contain Asp-30 did not (Figure 3). The 2 Da increase of y8 but not of y7 was proof that 18O was incorporated into residue 30 but not into other residues (Figure 3). The incorporation of 18O water can only be explained by the presence of succinimide at this position because 18O can be only incorporated into isoAsp via hydrolysis of succinimide. Quantification of Succinimide and Its Hydrolyzed Product isoAsp and Asp. The incorporation of 18O into the isoAspcontaining peptide could also be detected using ESI-TOF mass spectrometry. ESI-TOF mass spectrometers operate at a much higher resolution (5000 in V-mode) than the ion trap mass spectrometers used to generate MS/MS data, allowing the baseline separation of the isotopes of the peptides. LC-ESI-TOF was used to acquire high-resolution mass spectral data to separate the isotope peaks of the peptides. Figure 4 shows the isotope peak distribution of the doubly charged ion of peptides ASQSVD30SNLAWYR containing isoAsp (Figure 4A) and the Asp (Figure 4B) at residue 30. Since the 16O- and the 18O-containing peptides differ by only 2 Da in molecular mass and 1 m/z for the doubly charged peptide ions, the natural isotope distribution of the 16O peptides partially overlapped in the mass spectrum with the isotope distribution of the 18O peptides. The first peak (peak 0, black trace in Figure 4) of the 18O isotope is overlapped with the third peak (peak 2, red trace in Figure 4) of the 16O isotopes. Therefore, the first and second isotope peaks (shown in Figure 4 black line) measured by LC-ESI-TOF represent the first (peak 0, red trace) and second peak (peak 1 in red) of the 16O isotopecontaining peptide. Other measured isotope peaks shown (black line) in Figure 4 are contributed by both the 16O- and the 18O isotope-containing peptides. The intensity for each 16O isotope peak was calculated based on the theoretical isotope distribution of the elements (H, C, O, S, and N) for this peptide using the intensity of the first peak (1H, 12C, 16O, 14N, 32S) for normalization (see Table 1). The theoretical intensity for each 16O isotope peak 2718 Analytical Chemistry, Vol. 79, No. 7, April 1, 2007
Figure 4. Mass spectra quantifying 18O incorporation into the isoAsp peptide eluting at 97.5 min (A) and Asp peptide eluting at 100 min (B) obtained by a high-resolution ESI-TOF. Both spectra show the isotope peak intensity distribution for the doubly charged peptide ions. Peaks plotted by the dotted red trace represent the theoretical 16O isotope distribution, and peaks plotted by the solid trace represent the observed isotope peak intensity distribution. The 18O isotope peak intensity was calculated by subtracting theoretical 16O isotope intensity from the observed isotope peak intensity. For example, the intensity of the first 18O isotope peak equals the intensity of the third observed isotope peak (peak 0, black trace) minus the third theoretical 16O isotope peak (peak 2, red trace).
distribution is also shown in the dotted red line in Figure 4. The 18O isotope peak intensity can be calculated by subtracting the normalized theoretical intensity of each 16O isotope peak from the intensity of the measure isotope peak (black trace Figure 4). The ratio of 18O/16O-containing peptides can be calculated by dividing the calculated intensity of the first 18O isotope (peak 0, black trace) with the first normalized 16O isotope (peak 0, red trace), the second calculated 18O isotope (peak 1, black trace) with the second normalized 16O isotope (peak 1, red trace), the third calculated 18O isotope (peak 2, black trace) with the third normalized 16O isotope (peak 2, red trace), and so forth. In our experiment, the intensities of the first three isotopes were used for the calculation because of limited intensities of higher order isotopes. Alternatively the ratio could also be calculated by summing up the intensities of the first three 18O peptide isotopes and dividing this value by the sum of the intensities of the first three normalized 16O peptide isotopes. The earlier has the advantage that three ratios were calculated instead of just one, which enabled the calculation of the average ratio and the standard deviation of the three measurements, allowing an estimate of the error introduced
Table 1. (A) Calculation of 18O Incorporation into isoAsp of the Peptide ASQSVisoD30SNLAWYR of a Degraded Antibody Sample Stored at 45 °C at pH 5.0 for 8 Weeks isotope peaks of the IsoAsp peptide measured isotope peak intensity theoretical isotope distributiona theoretical 16O peptide isotope peak intensities calculated 18O labeled peptide isotope peak intensities ratio of 18O peptides to 16O peptidesb % of 18O peptidesc
peak 0
peak 1
peak 2
peak 3
528 100 528 1979 3.75 79
448 82 433 1508 3.48 78
2180 38 201 678 3.38 77
1577 13 69 229 3.35 77
peak 4
peak 5
694 3 16
235 1 6
total C isotope peak intensity 5662 1251 4396
(B) Calculation of 18O Incorporation into Asp of the Peptide ASQSVD30SNLAWY of a Degraded Antibody Sample Stored at 45 °C at pH 5.0 for 8 Weeks isotope peaks of the Asp peptide
peak 0
peak 1
peak 2
peak 3
measured isotope peak intensities theoretical isotope distributiona theoretical 16O peptide isotope peak intensities calculated 18O labeled peptide isotope peak intensities ratio of 18O peptides to 16O peptides d % of 18O peptidese
9832 100 9832 670 0.068 6.37
8590 82 8062 358 0.042 4.00
4406 38 3736 248 0.066 6.22
1636 13 1278
peak 4
peak 5
694 3 16
235 1 6
total isotope peak intensities 25006 23202 1276
(C) Calculation of the ratio of Theoretical and Measured Isotope Intensities of the isoAsp Peptides of a Degraded Antibody Sample Stored at 45 °C at pH 5.0 for 8 Weeksf isotope peaks of the IsoAsp peptide peak 0 peak 1 peak 2 peak 3 theoretical isotope distributiona normalized isotope peak intensities measured isotope peak intensities ratio of measured to normalized theoretical intensitiesg
100 2336 2336 1.00
82 1826 1916 1.05
38 864 888 1.03
13 271 304 1.12
a The theoretical isotope distribution was calculated using IsoPro software. b The average ratio is 3.5 and CV is 5%. c The average percentages are 78% and CV is 5%. Considering 18O water was 97% enriched, 18O isoAsp peptides accounts for 80% of the total isoAsp peptides. d The average ratio is 0.059 and CV is 4%. e The average percentages are 5.6% and CV is 4%. Considering 18O water was 97% enriched, the 18O Asp peptides accounts for 6% of the total Asp peptides. f The instrument error is represented by the ratios of theoretical and measured 13C isotope intensity. g The average ratio is 1.05 and standard deviation is 5%
by the instrumentation. In this sample, the ratio of 18O-containing peptide to 16O-containing peptides was calculated to be 3.5 to 1. This means that ∼80% ((3.5/(3.5 + 1)) of the isoAsp peptide contained 18O and therefore originated from succinimide in the sample (Table 1A). The percentage of 18O-containing Asp peptide was calculated to be 6% (Table 1B), which also originated from succinimide. The degraded sample (37 °C, pH 5.0, 8 weeks) used in this experiment showed ∼20% isoAsp at residue 30 based on the peak area UV chromatogram (Figure 2). Combining the data shown above (Table 1A, B), we calculated the presence of 21% succinimide (80% 18O isoAsp peptides × 20% total isoAsp peptides + 6% 18O Asp peptides × 80% total Asp peptides) at residue 30 in the degraded antibody sample, but only 4% isoAsp (20% 16O isoAsp peptide × 20% isoAsp peptide). Although we detected a total of 20% isoAsp by peptide mapping, the actual amount of isoAsp in the sample was only 4%. This difference can be explained by the rapid hydrolysis of the 21% succinimide present in the sample during the peptide mapping procedure. The error introduced by the instrumentation was also calculated by comparing the theoretical isotope distribution and the measured isotope distribution obtained from an experiment using regular water (Table 1C). The theoretic isotope distribution matched the observed isotope distribution very well within an error of less than 5% (Table 1C). Kinetics of Succinimide Hydrolysis. A succinimide hydrolysis kinetics study of this recombinant antibody under denaturing and native conditions was carried out by monitoring the amount of 18O incorporation into isoAsp and Asp over time. Hydrolysis of
Figure 5. Succinimide hydrolysis kinetics study of the recombinant antibody under denaturing conditions (8 M guanidine hydrochloride, 0.1 M NH4HCO3, pH 8.2) by monitoring the increase of 18O incorporation into isoAsp over time. The percentage of the amount of 18O incorporation was calculated by dividing the of 18O isoAsp peptide isotope intensity by the total 18O and 16O isoAsp peptide isotope intensity. The results showed that succinimide hydrolysis was completed in less than 5 s.
succinimide at light-chain residue 30 at denaturing conditions (8 M GuHCl, pH 8.2, 37 °C) was completed in less than 5 s (Figure 5). To compare the fast rate of the hydrolysis with the rate of unfolding of the protein under the same denaturing and alkaline conditions, we monitored the unfolding of the antibody by fluorescence spectroscopy using the change in fluorescence of the tryptophan residues over time as shown in Figure 6. The antibody unfolds very fast (within a few seconds), indicating that upon unfolding, the succinimide hydrolyzes immediately to isoAsp and Asp. The ratio of the isoAsp peptide 18O isotope intensity to Analytical Chemistry, Vol. 79, No. 7, April 1, 2007
2719
Figure 6. Kinetic fluorescence traces comparing unfolding of the antibody in 8 M guanidine hydrochloride, 100 mM NH4HCO3, pH 8.2 at 10 °C (gray trace) and 37 °C (black trace). The excitation wavelength was 280 nm, and the monitored emission wavelength was 357 nm. The change in fluorescence as the molecule unfolds was monitored over time. The folded protein sample was added to the denaturing buffer (8 M guanidine hydrochloride) a few seconds after data acquisition started, which was reflected in a sharp increase in fluorescence signal.
the Asp peptide 18O isotope intensity (4396, and 1276, respectively; Table 1A, B) represents the ratio of succinimide hydrolysis to isoAsp and Asp under denaturing conditions and was calculated to be ∼3.5 to 1. Our data are in good agreement with those reported previously that unstructured small peptides containing succinimide hydrolyzed to isoAsp and Asp at a ratio of ∼3:1.8,9 Under native conditions, the rate of succinimide hydrolysis was much slower compared to denaturing conditions with a half-life of succinimide of ∼6 h (Figure 7). Note that the sample used for this experiment was different from the sample used to study hydrolysis under denaturing condition. This sample was degraded for a longer period of time (pH 5.0, 45 °C, 12 weeks) and therefore showed an increased amount of isoAsp (∼28%) by peptide mapping compared to the degraded sample used to study the hydrolysis under denaturing conditions (pH 5.0, 45 °C 8 weeks, 20% isoAsp). The amount of succinimide in this sample was calculated to be ∼27% compared to the 21% of the previous sample. The amount of succinimide present in the sample did not affect the rate of its hydrolysis. Interestingly, under the native conditions, almost all 18O was found to be incorporated into Asp peptides. Contrary to succinimide hydrolysis under denaturing conditions, which resulted in an isoAsp/Asp ratio of 3.5:1, this result suggests that succinimide hydrolyzed almost exclusively to Asp. Therefore, the structure of the antibody must play an important role in succinimide hydrolysis. Figure 7B shows the decrease of the relative amount of isoAsp peptides over time during succinimide hydrolysis of the recombinant antibody under native conditions (0.1M NH4HCO3 at pH 8.2). There was very little 18O incorporation into the isoAsp peptide (data not shown). However, the decreasing amount of isoAsp peptide over time correlates very well with the increase in 18O incorporation to the Asp peptide, indicating that the succinimide present in the sample almost exclusively forms Asp and not isoAsp under native conditions at neutral and alkaline pH values. The significance of this phenomenon is that isomerization of Asp can be repaired through a nonenzymatic mechanism under conditions 2720 Analytical Chemistry, Vol. 79, No. 7, April 1, 2007
Figure 7. (A) Succinimide hydrolysis kinetics of the recombinant antibody under native conditions (0.1 M NH4HCO3, pH 8.2) by monitoring the increase of 18O incorporation into Asp peptides over time. The incorporation of 18O was calculated by dividing the 18O Asp peptide isotope intensity by the total 18O and 16O Asp peptide isotope intensity. The data were fitted to a first-order exponential association using GraphPad Prism software, resulting in a half-life of succinimide of ∼6 h with a rate constant k ) 0.002 min-1. (B) Decrease of the relative amount of isoAsp peptides over time during succinimide hydrolysis of the recombinant antibody under native conditions (0.1 M NH4HCO3) at pH 8.2. The amount was calculated based on the total ion intensities of the mass spectral data, which compared well with calculation based on the UV signal. The data were fitted to a first-order exponential decay using GraphPad Prism software resulting in a half-life of ∼6 h with a rate constant k ) 0.002 min-1 for the hydrolysis of succinimide.
where succinimide almost completely converts back to the native Asp amino acid. CONCLUSION We have developed a method for identification and characterization of succinimide in proteins. The method allows the unambiguous identification of succinimide in proteins and allows monitoring the hydrolysis of succinimide under different conditions for kinetic studies. The method utilizes 18O water for the hydrolysis of succinimide, and the resulting 18O-labeled isoAsp acid residues and Asp residues are identified and quantified using high-resolution ESI-TOF mass spectrometry of the proteolytic peptides. The method was used for the identification and characterization of succinimide in a recombinant monoclonal antibody. MS/MS data of 18O-labeled peptide were used to unambiguously identify succinimide at amino acid residue 30 of the light chain. Additionally, the method was used to study the kinetics of succinimide hydrolysis under native and denaturing conditions. Our results indicate that hydrolysis under denaturing conditions is very fast (less than 5 s) but much slower under native conditions
with a half-live of ∼6 h at pH 8.2. The isoAsp/Asp ratio was measured to be ∼3.5:1 under denaturing conditions, which is in agreement with values reported in the literature for small unstructured peptides (7, 8). However, hydrolysis under native conditions resulted almost exclusively in the formation of Asp and not isoAsp, suggesting that the structure of the protein affects not only the rate of succinimide hydrolysis but also its isomerization pathway.
ACKNOWLEDGMENT We thank Gerd Kleemann for his valuable suggestions and comments during the preparation of the manuscript. Received for review September 21, 2006. Accepted December 21, 2006. AC0617870
Analytical Chemistry, Vol. 79, No. 7, April 1, 2007
2721
