Nanosecond Laser-Induced Photochemical Oxidation Method for

Oxidation Method for Protein Surface Mapping with. Mass Spectrometry ... National University of Singapore, 14 Science Drive 4, Singapore 117543. We ha...
13 downloads 0 Views 329KB Size
Anal. Chem. 2005, 77, 5814-5822

Nanosecond Laser-Induced Photochemical Oxidation Method for Protein Surface Mapping with Mass Spectrometry Thin Thin Aye,† Teck Yew Low,† and Siu Kwan Sze*,†,‡

Genome Institute of Singapore, 60 Biopolis Street, Singapore 138672, and Department of Biological Sciences, National University of Singapore, 14 Science Drive 4, Singapore 117543

We have developed an ultrafast pulse method for protein surface footprinting by laser-induced protein surface oxidations. This method makes use of a pulse UV laser that produces, in nanoseconds, a high concentration of hydroxyl (OH) free radicals by photodissociation of a hydrogen peroxide (H2O2) solution. The OH radicals oxidize amino acid residues located on the protein surface to produce stable covalent modifications. The oxidized protein is then analyzed by mass spectrometry to map the oxidized amino acid residues. Ubiquitin and apomyoglobin were used as model proteins in this study. Our results show that a single laser pulse can produce extensive protein surface oxidations. We found that monooxidized ubiquitins were more susceptible to further oxidations by subsequent laser irradiation, as compared to nonoxidized ones. This is due to the conformational changes of proteins by oxidation that increases the solvent-accessible surface area. Therefore, it is crucial to perform this experiment with a single pulse of laser so as to avoid oxidation of proteins after conformation of the protein changes. Subsequently, to obtain a high frequency and coverage of the oxidation sites while keeping the number of laser shots to one, we further optimized the laser power and concentration of hydrogen peroxide as well as the concentration of protein. This ultrafast OH radical generation method allows for rapid and accurate detection of surface residues, enabling mapping of the solvent-accessible regions of a protein in its native state. Mass spectrometry (MS) is one of the most widely used analytical techniques in proteomics. In addition to being used for protein identification, it is also gaining popularity for deciphering protein structure, particularly for probing the solvent-accessible surfaces of proteins.1-3 Portions of a protein that are solventaccessible are available to interact with ligands or other proteins. Interactions at these sites can induce conformational changes that * Corresponding author: Phone: 065-64788111. Fax: 065-64789060. E-mail: [email protected]. † Genome Institute of Singapore. ‡ National University of Singapore. (1) Hori, R.; Baichoo, N. Protein-Protein Interact. 2002, 285-311. (2) Kolb, A.; Belyaeva, T.; Savery, N. DNA-Protein Interact. 2000, 161-174, 164-plates. (3) Nagai, H. Seibutsu Butsuri 1998, 38, 116-118.

5814 Analytical Chemistry, Vol. 77, No. 18, September 15, 2005

can modify the solvent-accessible surfaces of proteins. Current methods for protein structure determination include hydrogendeuterium exchange4-8 and chemical modifications9 with chemicals such as methyl bromide10 followed by MS analysis. However, hydrogen-deuterium exchange is hampered by complex backexchange kinetics. As for the latter, the reactivity and universality of such chemical probes to all amino acids are not yet well-defined. Recently, the hydroxyl (OH) free radical has been chosen for labeling the surface residues of proteins due to its numerous advantageous attributes.11 First, OH radicals are highly reactive12 and are able to oxidize a variety of amino acid side chains.2,13 In addition, the rate of oxidation of the side chains of amino acids by OH radicals is much higher than oxidation-induced backbone cleavages.2,13,14 Therefore, proteins can readily be oxidized without much backbone fragmentation. Currently, OH radicals used for such studies are usually generated by Fenton chemistry,15 UV irradiation of hydrogen peroxide,11 or by radiolysis of water with a high-energy X-ray synchrotron beam.16-19 Each of these three methods has its own shortcomings. Apart from the long incubation time required, iron salts and EDTA used in the Fenton reaction may distort the native conformation of a protein.20 As for UV irradiation, a time frame of (4) Akashi, S.; Naito, Y.; Takio, K. Anal. Chem. 1999, 71, 4974-4980. (5) Busenlehner, L. S.; Armstrong, R. N. Arch. Biochem. Biophys. 2005, 433, 34-46. (6) Kipping, M.; Schierhorn, A. J. Mass Spectrom. 2003, 38, 271-276. (7) Wu, Q.; Bulleigh, K.; Van Stipdonk, M. J. 39th Midwest Regional Meeting of the American Chemical Society 2004, MID04-053. (8) Zhu, M. M.; Rempel, D. L.; Gross, M. L. J. Am. Soc. Mass Spectrom. 2004, 15, 388-397. (9) Hanai, R. Tanpakushitsu Kakusan Koso 1996, 41, 929-933. (10) Scaloni, A.; Ferranti, P.; De Simon, G.; Mamone, G.; Sannolo, N.; Malorni, A. FEBS Lett. 1999, 452, 190-194. (11) Sharp, J. S.; Becker, J. M.; Hettich, R. L. Anal. Chem. 2004, 76, 672-683. (12) Loizos, N. Methods Mol. Cell. Biol. 2004, 261, 199-211. (13) Xu, G.; Takamoto, K.; Chance, M. R. Anal. Chem. 2003, 75, 6995-7007. (14) Xu, G.; Chance, M. R. Anal. Chem. 2004, 76, 1213-1221. (15) Sharp, J. S.; Becker, J. M.; Hettich, R. L. Anal. Biochem. 2003, 313, 216225. (16) Dhavan, G. M.; Chance, M. R.; Brenowitz, M. Kinet. Anal. Macromol. 2003, 75-86. (17) Guan, J.-Q.; Almo, S. C.; Reisler, E.; Chance, M. R. Biochemistry 2003, 42, 11992-12000. (18) Kiselar, J. G.; Janmey, P. A.; Almo, S. C.; Chance, M. R. Mol. Cell. Proteomics 2003, 2, 1120-1132. (19) Kiselar, J. G.; Janmey, P. A.; Almo, S. C.; Chance, M. R. Proc. Natl. Acad. Sci. U.S.A. 2003, 100, 3942-3947. (20) Heyduk, T.; Baichoo, N.; Heyduk, E. Met. Ions Biol. Syst. 2001, 38, 255287. 10.1021/ac050353m CCC: $30.25

© 2005 American Chemical Society Published on Web 08/12/2005

up to minutes in scale is usually required. Finally, although a highenergy X-ray synchrotron beam is capable of pulse-labeling proteins in milliseconds without the presence of chemicals,15 it is out of reach to most researchers. To probe the solvent-accessible surface of a protein in its native, biological state, OH radicals must react with the protein in its native conformation in a suitable buffer solution. Oxidation by OH radical is a random process dependent on the solvent-accessible surface as well as the chemical properties of the exposed amino acids.13 Upon oxidation, the three-dimensional structure of a protein may change to a certain extent, depending on the sites of oxidation.21 As a result, any further oxidations after this conformational change can result in erroneous allocation of surfaceaccessible residues. The time frame for protein folding/unfolding kinetics ranges from microseconds to milliseconds.22 It is, therefore, advantageous to pulse-label the surface amino acid residues below this time frame. However, the duration of treatment for current oxidation methods usually exceeds this time frame by at least an order of magnitude. In this paper, we report a novel nanosecond laser pulse-labeling technique for protein surface footprinting. In this method, OH radicals are formed in 3-5 ns by a pulse Nd:YAG laser. The halflife of the well-studied OH radicals was reported as nanoseconds to microseconds, depending on the composition of the solution.23-25 Therefore, with a single pulse of laser, formation of OH radicals, as well as their subsequent oxidation and quenching, is complete before the conformation of the protein changes. Ubiquitin and apomyoglobin were used as model proteins in this work because their detailed structural data are available in the Protein Data Bank. Furthermore, apomyoglobin has been widely used for surface footprinting studies,15,26,27 thus allowing us to validate our method. EXPERIMENTAL SECTION Sample Preparations. Ubiquitin, apomyoglobin and other chemicals were purchased from Sigma Chemicals (St. Louis, MO). Lyophilized proteins were reconstituted at different concentrations from 20 to 80 µM in a 10 mM phosphate buffer (pH 7.0), and H2O2 was added to the sample to a final concentration from 0.3 to 1% just before laser irradiation. The proteins were then oxidized with OH radicals, generated by exposing H2O2 to a pulse Nd: YAG laser (Polaris II-30 Hz, New Wave Research, Fremont, CA) operating at 266 nm and 30 Hz. A 20-µL portion of sample solution in a microfuge tube was held horizontally at a distance of 10 cm from the laser source, and its opening was properly aligned with the laser beam. Immediately after 1 to 100 laser shots, the oxidation was quenched by freezing the sample in liquid nitrogen, and subsequently, lyophilizing the sample in a vacuum chamber at 10-3 Torr for 30 min. Solvent and residual H2O2 were removed by sublimation. (21) Nabuchi, Y.; Asoh, Y.; Takayama, M. J. Am. Soc. Mass Spectrom. 2004, 15, 1556-1564. (22) Dai, S. Y.; Gardner, M. W.; Fitzgerald, M. C. Anal. Chem. 2005, 77, 693697. (23) Boveris, A. Medician (Buenos Aires) 1998, 58, 350-356. (24) Fessenden, R. W.; Verma, N. C. Biophys. J. 1978, 24, 413-430. (25) LaVerne, J. A. Radiat. Res. 2000, 153, 196-200. (26) Chance, M. R. Biochem. Biophys. Res. Commun. 2001, 287, 614-621. (27) Tanaka, N.; Ikeda, C.; Kanaori, K.; Hiraga, K.; Konno, T.; Kunugi, S. Prog. Biotechnol. 2002, 19, 47-54.

Trypsin Digestion of Oxidized Protein. The lyophilized oxidized proteins were then reconstituted to 20 µM in a 100 mM ammonium bicarbonate (NH4HCO3) buffer containing 50 mM dithiothreitol (DTT, Bio-Rad Laboratories, Hercules, CA). This solution was incubated at 60 °C for 1 h, and iodoacetamide (SigmaAldrich, St. Louis, MO) in 100 mM of NH4HCO3 was subsequently added to a final concentration of 100 mM before being incubated for 30 min in the dark at room temperature. Sequencing grade modified trypsin (Promega, Madison, WI) was then added in a 100:1 ratio. The mixture was incubated for 16 h in a 37 °C shaker incubator. The digestion was quenched by acidifying the sample with 5% of TFA (Fluka Chemicals, St. Louis, MO). Finally, the tryptic digests were kept at -80 °C prior to LC/MS/MS and FTMS analysis. FT-MS Analysis of Oxidized Proteins. (A) A Fourier transform ion cyclotron resonance mass spectrometer (FT-ICRMS) (APEX IV 94QE, Bruker Daltonics, Billerica, MA) was used to monitor the level of protein oxidation and oxidative backbone cleavages. The lyophilized, oxidized protein was redissolved in ultrahigh purity water (J. T. Baker, Phillipsburg, NJ) and cleaned by a C-4 zip-tip (Millipore, Billerica, MA) according to the manufacturer’s protocols before being loaded directly into a nanospary emitter (Proxeon Biosystems, Denmark). The nanospray voltage was set at 900 V. Ions were generated with a homebuilt static nanospray ionization source and accumulated in a hexapole ion trap before being transferred into the ICR cell by gate trapping. An external mass calibration was performed with ubiquitin at 20 µM before analysis of the oxidized sample. (B) LC/FT-MS was performed to detect oxidized, tryptic peptides originating from the samples. A 5-pmol portion of tryptic digests was separated by HPLC (LC-Packing, Sunnyvale, CA) using a 75-µm-i.d. C-18 column (LP-Packing, Sunnyvale, CA). The gradient was set at 200 nL/min and ramped from 5% ACN to 60% in 40 min, then to 80% ACN in 45 min and kept at 80% ACN for 3 min and ramped back to 5% ACN. The eluant was ionized by online nanospray at 1800 V into FT-MS with an Apollo nanospray source (Bruker Daltonics, Billerica, MA). The LC/FT-MS data were analyzed by a program written in-house and validated by manual interpretation of the spectra. This program searches for oxidized peptide mass peaks with reference to calculated masses of tryptic peptides that harbor all possible oxidations. The identification of the peptides was achieved through matching the experimental accurate monoisotopic masses measured by the FTMS with calculated monoisotopic masses of selected proteins. LC/MS/MS Analysis of Tryptic Peptides. All LC/MS/MS experiments were performed with a Surveyor HPLC system (Thermo Finnigan, San Jose, CA) interfaced to an LCQ-Deca XP Plus ion trap mass spectrometer (Thermo Finnigan, San Jose, CA) equipped with a nanospray source (Thermo Finnigan, San Jose, CA). A 20-pmol portion of protein digest was injected, subsequently concentrated, and desalted with an online peptide trap (MicromBioresources, Auburn, CA). Chromatographic separation was carried out with a home-packed nanobored C18 column (75µm i.d × 10 cm × 300 Å × 5-µm particles), operating at a flow rate of 200 nL/min. The LCQ was operated in a data-dependent mode by performing MS/MS scans for the 3 of the most intense peaks from each MS scan. Dynamic exclusion was inactivated. MS/MS data were analyzed by TurboSEQUEST using a database Analytical Chemistry, Vol. 77, No. 18, September 15, 2005

5815

Figure 1. UV absorption spectrum of H2O2 showing that the absorption maximum was about 250 nm. This is very close to the wavelength, that is, 266 nm, which we applied for photochemical oxidation of solvent-accessible amino acid residues.

consisting of the proteins being studied with a differential modification of +16 Da set up for each of the 20 amino acids. Possible oxidation sites were confirmed manually by checking the mass spectra. As a comparison, GETAREA 1.128 was used to calculate the solvent-accessible surface of the NMR structure of ubiquitin29 and X-ray crystal structure of apomyoglobin.30 Circular Dichroism Studies. CD studies were performed with a Jasco J-810 spectropolarimeter (Tokyo, Japan) with a continuous scanning mode from 190 to 260 nm. Five spectra were accumulated for each sample with a scanning speed of 50 nm/ min and 1-nm bandwidth. RESULTS AND DISCUSSION Laser-Induced Oxidation of Proteins. A H2O2 molecule readily undergoes endothermic dissociation (+49 kcal/mol) to form two molecules of OH free radicals which are potent oxidizing agents due to the unpaired electrons of the oxygen atom.31

H2O2 + hv f 2•OH The scanning spectrogram acquired for H2O2 at different wavelengths shows that H2O2 absorbs UV wavelengths from 300 nm to the beginning of the Rydberg transition region at 124 nm,32 with an absorption maximum at 250 nm (Figure 1). Therefore, a pulsed Nd:YAG laser operating at 4 harmonic (266 nm) would provide an ideal UV source for this experiment. Though OH radicals have an extremely short half-life,23-25 it is just as important to effectively quench the oxidation of proteins (28) Fraczkiewicz, R.; Braun, W. J. Comput. Chem. 1998, 19, 319-333. (29) Cornilescu, G.; Marquardt, J. L.; Ottiger, M.; Bax, A. J. Am. Chem. Soc. 1998, 120, 6836. (30) Evans, S. V.; Brayer, G. D. J. Mol. Biol. 1990, 213, 223-230. (31) Ondrey, G.; van Veen, N.; Bersohn, R. J. Chem. Phys. 1983, 78, 37323737. (32) Karl-Heinz, G.; Stefan, K.; Franz, J. C. J. Chem. Phys. 1986, 85, 4463-4479.

5816

Analytical Chemistry, Vol. 77, No. 18, September 15, 2005

by residual H2O2 after laser irradiation, because H2O2 can also contribute to cumulative oxidation of proteins in the relatively long storage and processing time before analysis.15 We tried several methods to quench the oxidation by H2O2. These methods include immediate quenching by adding DTT (100 mM), Tris (1 M),15 or ammonium bicarbonate (1 M),33 followed by snap-freezing in liquid nitrogen. Direct snap-freezing was found to be the most effective method. This method was performed by immediately snap-freezing the sample after laser irradiation in liquid nitrogen and subsequent lyophilization to remove the residual H2O2. However, even in the untreated sample, oxidation of about 2% of the ubiquitin was consistently observed. This phenomenon was also observed for other proteins, such as apomyoglobin. This basal level of oxidation can be attributed to the oxidations of methionine residue or other reactive amino acids in the protein during the nanospray ionization in air.34 Dynamics of Oxidation. Both Figure 3 and Table 1 show the results acquired from FT-MS analysis of ubiquitin ions and its oxidized counterparts carrying 10 charges, after irradiation with varying shots of laser. The ion intensities were integrated over the isotopic cluster, and total ions were normalized to 100. The ionization efficiency and detection sensitivity of nonoxidized and oxidized proteins were assumed to be identical. To enable an appropriate comparison between photochemical oxidations obtained from a UV source,15 H2O2 and protein concentrations were fixed at 0.3% and 20 µM, respectively, in this study. With a single shot of laser, 64% of the ubiquitin was converted to monooxidized product in 0.3% H2O2 (Figure 2B). A negligible amount of multiple oxidized ubiquitin was generated. However, with 100 laser shots, the ubiquitin was oxidized at multiple sites, with up to 18 oxides per protein (Figure 2C). With (33) Stowell, J. P.; Jensen, J. N. Water Res. 1991, 25, 83-90. (34) Maleknia, S. D.; Chance, M. R.; Downard, K. M. Rapid Commun. Mass Spectrom. 1999, 13, 2352-2358.

Figure 2. FT-MS spectra of ubiquitin ion and its oxidized counterparts carrying 10 charges after different numbers of shots of laser: (A) no laser shot, (B) one laser shot, (C) 100 laser shots, and (D) one laser shot of 8 M urea denatured ubiquitin.

reference to Table 1 and Figure 3, it is noteworthy that with an increasing number of laser shots, the proportion of monooxidized species diminished sharply while the multiply oxidized species increased. In contrast, the proportion of the nonoxidized species did not decrease as markedly. For instance, when the number of laser shots was increased from 1 to 10, the nonoxidized species remained relatively consistent (30.15 to 32.58%), while the monooxidized species dropped from 64.22 to 46.77%. When the number of laser shots was further increased from 1 to 100, the nonoxidized species dropped 1.7-fold (30.15 to 18.18%), while the monooxidized species dropped 3.3-fold (64.22 to 19.71%). This

observation indicates that the monooxidized ubiquitin is more susceptible to further oxidations. Because oxidation of proteins by OH radicals is dependent on the solvent-accessible surface as well as the chemical properties of the exposed amino acids, it is highly probable that this susceptibility is due to the changes in the solvent-accessible surface and, thus, the solvent-exposed amino acid residues, both of which occur as a result of changes in the protein conformation. On the other hand, denatured ubiquitin (8 M urea) treated by a single laser shot contains a higher proportion of multiply oxidized ubiquitin, as compared to native ubiquitins treated with Analytical Chemistry, Vol. 77, No. 18, September 15, 2005

5817

Figure 3. Normalized intensities of oxidized ubiquitin ions carrying 10+ charges after irradiation by different numbers of laser shots. Molecules carrying the same number of oxides are linked. It is noteworthy that with an increasing number of laser shots, the proportion of monooxidized species diminishes sharply while the multiply oxidized species increase. In contrast, the proportion of nonoxidized species does not decrease as markedly. Table 1. Oxidation Percentage of Ubiquitin under Different Numbers of Laser Shotsa no. of laser shots no. of oxidations

0

1

10

50

100

1 (in urea)

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

96.87 2.32 0.65 0.16 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

30.15 64.22 2.06 1.42 1.25 0.90 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

32.82 46.77 7.63 4.05 2.00 3.73 1.52 0.81 0.48 0.48 0.12 0.07 0.00 0.00 0.00 0.00 0.00 0.00 0.00

25.58 34.07 10.80 7.53 5.07 5.53 3.29 2.38 1.66 1.17 0.94 0.68 0.56 0.46 0.28 0.00 0.00 0.00 0.00

18.18 19.71 10.51 9.24 7.93 7.36 5.68 4.44 3.47 2.86 2.21 1.75 1.57 1.19 1.12 0.88 0.67 0.55 0.38

61.47 23.41 5.72 2.95 1.84 1.84 1.54 1.23 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00

a FT-MS analysis of ubiquitin ions and its oxidized counterparts carrying 10+ charges, after irradiation with varying shots of laser. The ion intensities were integrated over the isotopic cluster and total ions were normalized to 100.

a single laser shot (Figure 2D and Table 1). A small amount of oxidative backbone cleavage products were also observed (data not shown). This observation can be attributed to the larger solvent-exposed surface areas of denatured ubiquitin, because they have a more open 3-dimensional structure that allows the exposure of its backbone for oxidation-induced fragmentation. Taken together, this suggests that the oxidized ubiquitins had undergone some conformational changes that are similar to denaturation by 8 M urea. Indeed, far UV (190-260 nm) spectra obtained from circular dichroism (CD) studies show that the conformation of ubiquitin started to change even after 1 laser shot (Figure 4B). With 10 and 100 laser shots, the conformational changes became more evident. The corresponding control samples were irradiated with 1, 10, and 100 laser shots, respectively, 5818

Analytical Chemistry, Vol. 77, No. 18, September 15, 2005

without H2O2. The CD spectra for these controls (Figure 4A) did not change significantly as compared to untreated ubiquitins, suggesting that structural changes of treated ubiquitins were due to oxidation and not laser irradiation. Subsequently, a similar set of CD experiments was repeated for apomyoglobin (Figure 4C). These experiments show that the conformation of apomyoglobin in 0.3% H2O2 changed after being irradiated by a single shot of laser. For laser operating at 30 Hz for 1, 10, and 100 shots, the protein was exposed to OH radicals for a few microseconds, 330 ms, and 3.3 s, respectively. Even exposing the sample to two laser shots would have taken 66 ms. Therefore, it is desirable to perform the laser-induced oxidation in a single shot with high laser intensity, and hence, all subsequent experiments for protein solvent-accessible surface foot-printing were performed with a single laser shot. Collectively, this set of experiments show that (1) monooxidized ubiquitins are more susceptible to oxidations, as compared to nonoxidized ubiquitins. This phenomenon is comparable to ubiquitins denatured by 8 M urea. (2) This susceptibility is due to the change in the 3-D conformation of ubiquitins. (3) The change in conformation is due to oxidation itself. (4) The conformation of a protein can be changed by oxidation, even with a single shot of laser, which takes only nanoseconds to microseconds, and any presence of OH radicals after this would result in nonspecific oxidation of surface amino acid residues. (5) Therefore, it is desirable to perform the laser-induced oxidation in a single shot with high laser intensity. Optimization Studies. To obtain high frequency and coverage of the oxidation sites and partly to compensate for the low concentration of OH radicals resulting from only a single laser shot, we optimized the laser intensity, H2O2 concentration, and protein concentration. First, we varied the H2O2 concentration from 0.3 to 0.6 and 1% while the concentration of apomyoglobin was fixed at 20 µM and the laser intensity was set to ∼2 mJ. Our results (Figure 5) show that with increasing H2O2 concentration, more oxidations per protein were observed. For example, when 1% H2O2 was used, the unoxidized species was lost, and up to 13 oxidations were observed per protein. In contrast, unoxidized apomyoglobin was still present when 0.3% H2O2 was used, and only up to eight oxidations per protein was observed. The influence of laser intensity was analyzed by altering it from 0.2 to 2 mJ while H2O2 concentration and protein concentration were kept at 0.3 and 20 µM, respectively. Generally, with the increase in laser intensity, protein peaks shifted to higher oxidation state, although the effect of laser intensity on apomyoglobin oxidation is not as obvious as changes in H2O2 concentration (Figure 6). When the same experiments were repeated with 0.6 and 1% H2O2, similar results were obtained. This can be attributed to the concentration of H2O2, which is a limiting factor in such an experiment. Finally, with a laser intensity of 2 mJ and 0.3% H2O2, we varied the protein concentration from 20 to 80 µM. FT-MS analysis of oxidized apomyoglobin is shown in Figure 7. Figure 7B shows the oxidation of 80 µM apomyoglobin performed with a single laser shot. The normalized ion intensities are 9.6, 50.2, 34.5, 3.9, and 1.8% for 0, 1, 2, 3, and 4 oxidations, respectively. More than 90% of the protein was oxidized to different extents with a single laser shot. This demonstrates that apomyoglobin has a larger

Figure 4. Far-UV CD spectra for ubiquitin and apomyoglobin: (A) control experiment in which ubiquitin without H2O2 was treated with different numbers of laser shots, (B) the same set of experiments repeated with 0.3% H2O2. Results obtained show that oxidation produced with a single laser shot changes the conformation of ubiquitin, even at the level of secondary structure. (C) The same findings were also found in the experiments repeated with apomyoglobin.

Analytical Chemistry, Vol. 77, No. 18, September 15, 2005

5819

Table 2. Number of Oxidations Detected by LC/FT-MS and Reactive Amino Acid Residues Identified by LC/MS/MS for Ubiquitin oxidation sites of solvent- accessible meas mass detected by oxidation by surface area peptide (Da) by a a a by NMR (A2) residues LC/FT-MS LC/FT-MS LC/MS/MS

Figure 5. Normalized intensities of oxidized apomyoglobin ions carrying 18+ charges after irradiation by a single laser shots. It is found that with increasing H2O2 concentration, the proportion of monooxidized species diminishes while the number of multiply oxidized species increases.

1-6

780.4201

M+O

12-27

1802.9145

M+O

34-42

1070.5102

M + 2O

43-47 48-63 64-72

699.3591 1794.8742 1082.6081

M+O M+O M+O

Table 3. Number of Oxidations Detected by LC/FT-MS and Reactive Amino Acid Residues Identified by LC/MS/MS for Apomyoglobin site of meas mass oxidation solvent-accessible oxidation (Da) by detected by detected by surface area LC/FT-MSa LC/FT-MaS LC/MS/MSa by NMR (A2)

1-16

1830.8891

M+O

17-31

1637.8371 1653.8321 1286.6506 1302.6455 699.3591 1366.6286 1521.9242 1537.9191 1248.6620

M + 1O M + 2O M + 1O M + 2O M+O M+O M + 1O M + 2O M + 2O

1517.6568 1533.6517 763.4228 646.3286 665.3020

M + 1O M + 2O M+O M+O M+O

32-42 43-47 51-62 64-77 88-98

Figure 6. FT-MS spectra of apomyoglobin ion and its oxidized counterparts carrying 18+ charges after irradiation of different laser powers from ∼0.2 to 2 mJ. 119-133

5820

Analytical Chemistry, Vol. 77, No. 18, September 15, 2005

23.57 69.87 39.63 65.70 70.83 61.76 32.28 67.15 36.52

a Six oxidized peptides were found in LC/FT-MS analysis, whereas three oxidized peptides were detected in LC/MS/MS analysis. The solvent-accessible surface areas were calculated using GETAREA 1.127 using NMR structure of bovine ubiquitin.28

peptide residues

solvent-exposed surface area than ubiquitin. About 85% of the apomyoglobin was mono- or dioxidized. Figure 7C shows the result for 20 µM apomyoglobin oxidized with a single laser shot. We found up to 10 oxidations per protein without unoxidized species of apomyoglobin. Collectively, these experiments show that H2O2 concentration has greater effects on the frequency and coverage of the oxidized apomyoglobin using a single laser shot. However, when protein concentration changes, we need to adjust these parameters accordingly so as to obtain the optimum level of frequency and coverage of the oxidized residues. Analysis of Oxidation Sites. LC/FT-MS and LC/MS/MS analysis were performed on the trypsin-digested ubiquitin and apomyoglobin to locate the oxidation sites. The results are shown in Tables 2 and 3, respectively. Although 70% of the ubiquitin and 90% of the apomyoglobin were oxidized with a single shot of the laser, the oxidation sites were randomly distributed to different surface amino acid residues. As a result, the concentration of any

M1 Q2 T12 P19 S20 N25 I36 P37 P38

134-139 140-145 148-153

S3 E6 W7 Q8 L11 K16 Q26 E27

58.7 47.5 19.0 83.7 42.6 47.8 19.3 54.1

M55 H64 T66 P88 L89 S92 H93 H96 M131 T132

7.3 33.0 35.7 61.6 48.6 23.9 58.5 55.9 0.2 52.3

a Eleven peptides are detected as oxidized peptides in LC/FT-MS analysis, although six oxidized peptides are observed in LC/MS/MS. The solvent-accessible surface areas were calculated using GETAREA 1.127 using NMR structure of horse heart myoglobin.29

unique oxidized peptides was considerably low. A sensitive technique was therefore critical for detecting these low-abundance peptides for successful mapping of the solvent-accessible surface. LC/FT-MS is a highly sensitive technique that is capable of detecting multiple oxidized isoforms of oxidized peptides. The lowabundance peptides were separated and concentrated in the C-18 column before the nanospray ionization. This separation minimized ion suppression from other high-abundance peptides. These oxidized peptides were identified by accurate mass measurement using our in-house software. The LC/FT-MS chromatogram

Figure 7. FT-MS spectra of apomyoglobin: (A) nanospray untreated apomyoglobin, (B) oxidized 80-µm apomyoglobin by a single laser, shot and (C) oxidized 20-µm apomyoglobin with a single laser shot.

showed that modifications of the same peptide at different amino acid residues eluted out of the C-18 column at slightly different times as a cluster of peaks (data not shown). To locate the site of oxidation to a specific amino acid, datadependent MS/MS was performed with a LC-ion trap MS instrument, which generated more than 2000 spectra in a 60-min gradient. Mono-, di- and tri-oxides (16, 32, and 48 Da) were included as variable modifications for all 20 amino acids when TurboSequest was used for data analysis. To enable a faster search for the combination of multiple possible oxidation products, an in-house program was developed to screen the MS/MS spectra before performing database queries. This program screened for MS/MS spectra with parent ions having masses that correspond to the calculated mass of oxidized tryptic peptides, with a mass tolerance of (1.4 Da. This reduced the total number of spectra to analyze from 2000 to 350. As shown in Table 3, LC/MS/MS analysis of trypsin-digested apomyoglobin achieved 99.3% sequence coverage. Only residues 154-155 were not detected. Figure 8 shows a MS/MS spectrum of an oxidized peptide GLSDGEWQQVLNVWGK from apomyoglobin. The mass of this doubly charged peptide (m/z 924.83) was 16 units more than expected, as compared to the nonoxidized

peptide (m/z 908.5). This indicates the presence of two oxidations. As shown in the spectrum, for the y ion series, a mass shift of 16 units was observed for the y10 fragment ion. However, this mass shift was not observed for ion series from y9 to y3. This indicates that this oxidation site is located at Trp7. Meanwhile, ion series y11-y13 contained 32 units more than their corresponding nonoxidized product ions, indicating an extra site of oxidation other than Trp7. On the other hand, for b ion series, the b6 ion was observed to have a mass increase of 16 units, which was absent in the b5 ion. This shows that the extra oxidation occurs at Glu6. In addition, each ion from the ion series b7-b15 contained mass shifts of 32 units. This further confirms the assignment. For residues 17-31, a mass shift of 16 units was detected, implying two possible oxidation sites for this doubly charged peptide, VEADIAGHGQEVLIR. For y ion series, y4 ions, y5 + O ions and y6 + 2O were detected, implying that both Glu26 and Glu27 were oxidized. This finding was confirmed by the presence of the b14 and b11 ions, both harboring two oxidations. According to the structure of apomyoglobin obtained with X-ray crystallography, both Gln26 and Glu27 have a high solvent-accessible side chain area. Furthermore, they are also located on the outer surface of the protein. Applying the same logic, oxidation was also Analytical Chemistry, Vol. 77, No. 18, September 15, 2005

5821

Figure 8. MS/MS spectrum of an oxidized peptide obtained from apomyoglobin with sequence GLSDGEWQQVLNVWGK. The oxidation sites are located at Trp7 and Glu6 after analysis with TurboSequest.

confined to Met55 in residues 51-62; His64 as well as Thr66 in residues 64-77; and finally, Met131 and Thr132 in residues 119-133. Except for methionine, all of the oxidized residues identified in these experiments are highly solvent-accessible, according to the calculated side chain, solvent-accessible surface area of NMR structure. Using accurate mass measurement, we found that LC/FT-MS is capable of detecting more oxidized peptides that were not identified by LC/MS/MS, although mapping of oxidation sites could not be performed. As for LC/MS/MS, the lower rate of detection might be due to the low abundance of the oxidized peptides, because insufficient fragment ions were generated in the CAD process for detection and spectral assignment. Therefore, a combination of both methods was used. CONCLUSION Our results indicate that an extremely short reaction time is crucial for generating OH radicals for mapping protein solventaccessible surface. To probe the native protein structure, the

5822

Analytical Chemistry, Vol. 77, No. 18, September 15, 2005

oxidation of protein surface residues has to be completed before the modified protein change conformation. Thus, laser-induced dissociation of hydrogen peroxide can provide a convenient source of ultrashort pulse of high-intensity OH radicals for this purpose. ACKNOWLEDGMENT We thank the Singapore Agency for Science Technology and Research (A*Star) for its generous financial support. We are also grateful to Edison Liu, Edwin Cheung, Henry Mok, Vellaichamy Adaikkalam, Hua Lin, Liu Jining, and Pradeep Gopalakrishnan for useful discussions and to Chan Siew Leong and Jaran Jainhuknan for valuable technical assistance.

Received for review February 28, 2005. Accepted July 8, 2005. AC050353M