Chemical Preparation of an Isotopically Enriched Superoxide

For the isotopic enrichment of SOD, Cu and Zn metal (CK Gas Products, Hampshire, UK) with a certified isotopic abundance of 63Cu 0.8% and 65Cu 99.2% o...
0 downloads 0 Views 347KB Size
Download PDF
Anal. Chem. 2007, 79, 8381-8390

Chemical Preparation of an Isotopically Enriched Superoxide Dismutase and Its Characterization as a Standard for Species-Specific Isotope Dilution Analysis Christian L. Deitrich,† Andrea Raab,† Barbara Pioselli,‡ Jane E. Thomas-Oates,‡ and Jo 1 rg Feldmann*,†

College of Physical Science, Department of Chemistry, University of Aberdeen, Meston Walk, Aberdeen, UK AB24 3UE, and Department of Chemistry, University of York, Heslington, York, UK YO10 5DD

The development of methods to analyze accurately and precisely individual metalloproteins is of increasing importance. Here we describe for the first time the chemical preparation and characterization of an isotopically enriched metalloenzyme containing two different metal isotopes. Its evaluation as a standard in species-specific isotope dilution analysis by HPLC coupled to inductively coupled plasma mass spectrometry is carefully evaluated. Our model enzyme bovine superoxide dismutase (SOD) contains both Cu and Zn and is remarkably stable at high temperatures and even under denaturing conditions. The enzyme’s metal cofactors were removed under a range of different conditions and replaced with isotopically enriched 65Cu and 68Zn. Depending on the conditions, various isotopic ratios differing from the natural Cu and Zn abundances were obtained for the reconstituted enzyme. Both the wild type and isotopically enriched enzyme had the same migration pattern on native 1D-PAGE. Using an enzyme activity test, we showed that the incorporated 65Cu was bound to the right SOD-binding site, since the measured activity correlated directly with the amount of Cu incorporated. Mixing the native and the isotopically enriched enzyme standard with free enriched 65Cu and 68Zn or a metal chelator did not result in any exchange or loss of the metals from the enzyme at neutral pH. This verifies the stability of the enzyme metal center under the chosen conditions. The isotopically enriched enzyme standard was spiked into a wild type SOD solution to evaluate its use for species-specific isotope dilution experiments. To our knowledge, this is the first report of the chemical preparation of a metalloenzyme containing two different isotopically enriched metals. We provide evidence that the incorporated isotopically enriched metals are bound to the right binding site of SOD using an specific enzymatic activity assay. Proteins are pivotal in cellular processes. More than 30% of all proteins contain a metal and ∼40% of those metalloproteins need * To whom correspondence should be addressed. Phone: +44 (0) 1224 272911. Fax: +44 (0) 1224 2721. E-mail: [email protected]. † University of Aberdeen. ‡ University of York. 10.1021/ac071397t CCC: $37.00 Published on Web 09/25/2007

© 2007 American Chemical Society

the metal as a cofactor to perform specific functions.1 The detection and quantification of metalloproteins is of great importance for clinical diagnosis, since alterations in their formation or expression can be an indicator of disease. Hyphenated techniques are commonly used for trace elemental speciation. This involves the coupling of a high-resolution separation technique such as HPLC, GC, or CE, with a highly sensitive detector. Inductively coupled plasma mass spectrometry (ICPMS) is nowadays the first choice of an elemental detector.2-4 However, in biomolecules such as metalloproteins, metals form coordination complexes that can be labile under certain conditions, allowing the metal to be released and the protein to elute or migrate as its apo form.5 Dissociation and transformation or irreversible binding of proteins to the column material during the analytical procedure must be carefully considered. Furthermore, quality control is hampered by the lack of suitable standards and reference materials for biological species.6 Therefore, quantification of metalloproteins is still far from routine. The development of new methods for analyzing accurately and precisely individual metalloproteins is of increasing importance. Methods based on using isotopically labeled or enriched protein standards are therefore required to obtain a sufficient degree of accuracy and precision. Stable isotope labeling is widely used for quantitative proteomic analysis.7 Labeled proteins are usually obtained by in vivo metabolic labeling of amino acids in cell culture (SILAC) or by introducing a stable isotope tag into a particular amino acid residue via an isotope-coded affinity tag or onto specific functional groups such as primary amines (iTRAQ). Stable proteins, or more frequently their tryptic peptides, that are isotopically labeled are almost identical to the unlabeled protein/ peptide but can be distinguished by their mass, using molecular MS. However, these techniques based on ESI-MS or MALDI-MS are for determination of relative rather than absolute protein (1) Wilson, C. J.; Apiyo, D.; Wittung-Stafshede, P. Q. Rev. Biophys. 2004, 37, 1-30. (2) Heumann, K. G. Anal. Bioanal. Chem. 2002, 373, 132-139. (3) Michalke, B. TrAC, Trends Anal. Chem. 2002, 21, 142-153. (4) Prange, A.; Schaumlo¨ffel, D.; Bra¨tter, P.; Richarz, A-N.; Wolf, C. Fresenius J. Anal. Chem. 2001, 371, 764-774. (5) Szpunar, J. Analyst 2005, 130, 442-465. (6) Michalke, B. Ecotoxicol. Environ. Saf. 2003, 56, 122-139. (7) Monteoliva, L.; Albar, J. P. Brief Funct. Genomic Proteomic 2004, 3, 220239.

Analytical Chemistry, Vol. 79, No. 21, November 1, 2007 8381

abundances.8 Online isotope dilution mass spectrometry (IDMS) has recently been developed as a superior tool in elemental speciation to perform quantitative analysis of trace elements in proteins.9-11 It provides a method for validating current speciation methods, by highlighting systematic errors in analytical approaches, as recently reviewed by Sanz-Medel and co-workers.12 For the determination of elements in biomolecules of unknown identity using a species-unspecific isotope dilution approach, an isotopically enriched spike of the element to be determined is continuously added postcolumn (the spike can be in a chemical form different from that of the analyte). The eluting trace elements bound to the biomolecule can then be quantified accurately using ICPMS.12 When the stoichiometry of metals in a metalloprotein is known and an identical isotopically enriched species is available, it is possible to perform species-specific isotope dilution mass spectrometry (SS-IDMS). In this case, the protein concentration can be determined via its elemental concentration. This is also possible using the species-unspecific approach if the stoichiometry of the metal-protein complex is known. However, the speciesspecific spiking approach is superior because losses of the analyte after spiking can be corrected for. SS-IDMS of metalloproteins, especially where noncovalent protein-metal interactions are involved, requires carefully chosen and maintained conditions in order that no loss of the metal center or redistribution between free metal isotopes and the isotopes of the spike and sample protein occur. An important benefit of SS-IDMS is that species transformation during sample preparation and chromatography can be controlled and even quantified.13 SS-IDMS is, thus, a useful tool for evaluating analyte stability and behavior during the analytical procedure. Only a few examples have been published of the production of isotopically enriched proteins for elemental SS-IDMS, since the application and production of those standards is still a challenging task.14-16 The Cu-containing metalloproteins rusticyanin14 and plastocyanin15 were successfully isotopically enriched and characterized from their nonmetalated apo forms. Recently, SanzMedel and co-workers reported the use of SS-IDMS for the quantification of transferrin isoforms.16 An isotopically enriched transferrin spike saturated with 57Fe was synthesized, characterized, and used to quantify individual transferrin isoforms in serum samples. These studies showed the feasibility of SS-IDMS for metallobiomolecules. It is noteworthy that in these studies it was not shown whether the enriched metals were bound to the correct binding motif of those proteins. The transfer of 67Zn from (8) De la Calle Guntin ˜as, M. B.; Bordin, G.; Rodriguez, A. R. Anal. Bioanal. Chem. 2004, 378, 383-387. (9) Schaumlo ¨ffel, D.; Prange, A.; Marx, G.; Heumann, K. G.; Bra¨tter, P. Anal. Bioanal. Chem. 2002, 372, 155-163. (10) Ferrarello, C. N.; Ruiz Encinar, J.; Centineo, G.; Garcia, Alonso, J. I.; Fernandez, de la Campa, M. R.; Sanz-Medel, A. J. Anal. At. Spectrom. 2002, 17, 1024-1029. (11) Goenaga Infante, H.; Van Campenhout, K.; Schaumlo ¨ffel, D.; Blust, R.; Adams, C. F. Analyst 2003, 128, 651-657. (12) Rodriguez-Gonzalez, P.; Marchante-Gayon, J. M.; Garcia, Alonso, J. I.; SanzMedel, A. Spectroc. Acta, Part B 2005, 60, 151-207. (13) Rodriguez-Gonzalez, P.; Ruiz Encinar, J.; Garcia, Alonso, J. I.; Sanz-Medel, A. J. Anal. At. Spectrom. 2004, 19, 685-691. (14) Harrington, C. F.; Vidler, D. S.; Watts, M. J.; Hall, J. F. Anal. Chem. 2005, 77, 4034-4041. (15) Hann, S.; Obinger, C.; Stingeder, G.; Paumann, M.; Furtmu ¨ ller P. G.; Koellensperger, G. J. Anal. At. Spectrom. 2006, 21, 1224-1231. (16) Del Castillo Busto, M. E.; Montes-Bayon, M.; Sanz-Medel, A. Anal. Chem. 2006, 78, 8218-8226.

8382

Analytical Chemistry, Vol. 79, No. 21, November 1, 2007

isotopically enriched metallothionein to the enzyme carbonic anhydrase (CA) was described by Mason and co-workers.17 Here the authors used an activity assay to show that the transfer and the incorporation of 67Zn into CA activated the apometalloenzyme. The isotopic enrichment of a metalloenzyme offers the possibility of performing a metal-dependent enzyme assay, which can be used to indicate the correct incorporation of enriched metals into those enzymes. Here, we demonstrate an SS-IDMS approach using, for the first time, a metalloenzyme containing two different metal cofactors. The aim of this study was the chemical preparation of the isotopically enriched metalloenzyme superoxide dismutase (SOD) and the evaluation of the correct metal binding to the enzyme using an activity assay. In addition, its use as a standard for SS-IDMS was evaluated, in order to assess whether IDMS analysis is possible or whether redistribution reactions take place during sample preparation. Cu/Zn-SOD is a homodimeric enzyme (32 kDa) and was first discovered by McCord and Friedovic.18 Within each subunit, the active site hosts the binuclear metal center, comprising a closely spaced Cu(II) and Zn(II) ion pair. The enzyme catalyzes the disproportionation of superoxide via its Cu ion redox cycle [Cu(II)/Cu(I)], protecting the organism from oxidative stress, while the neighboring Zn ion plays a structural role.19 The specific coordination structure of SOD and neighboring residues are responsible for the tight binding of cofactors and the enzyme activity.20 An enzyme assay is available, which depends on the Cu content of the enzyme and therefore can be used to verify the correct incorporation of Cu into the binding motif of the enzyme.21 The involvement of the enzyme in disease development and its possible role as a biomarker made it an interesting candidate for our SS-IDMS studies.22-23 We report the full characterization of the isotopically enriched enzyme by ICPMS, enzymatic assays, and native PAGE. The stability of the enzyme was tested, and evaluation of our isotopically enriched SOD as a standard in SSIDMS was performed by spiking it into a solution of known SOD concentration. EXPERIMENTAL SECTION Materials and Reagents. Analar grade chemicals or higher were used in this work unless otherwise stated. All solutions and dilutions were made with high-purity deionized water (>18 MΩ, Elga, UHQ). Reagents include tris(hydroxymethyl)methylamine (Tris), ammonium acetate, ethylenediaminetetraacedic acid disodium salt (EDTA), potassium hydroxide, pyrogallol (99.5% minimum assay), and calcium chloride (all from BDH, Dorset, UK), diethylenetriaminepentaacetic acid (DTPA; Sigma-Aldrich, Dorset, UK), hydrochloric acid (37%), and nitric acid (∼69.5%; Fluka, Buchs, Switzerland). Cu/Zn-SOD from bovine liver (EC 1.15.1.1, (17) Mason, A. Z.; Moeller, R.; Thrippleton, K.; Lloyd, D. Anal. Biochem. 2007, 369, 87-104. (18) McCord, J. M.; Fridovich, I. J. Biol. Chem. 1969, 244, 6049-6055. (19) Bordo, D.; Pesce, A.; Bolognesi, M.; Stroppolo, M. E.; Falconi, M.; Desideri, A. In Handbook of Metalloproteins, Volume 2; Messerschmidt A., Huber, R., Poulos, T., Wieghardt, K., Eds.; John Wiley: New York, 2002; pp 12841300. (20) Tainer, J. A.; Getzoff, E. D.; Beem, K. M.; Richardson, J. S.; Richardson, D. C. J. Mol. Biol. 1982, 160, 181-217. (21) Marklund, S. L.; Marklund, G. Eur. J. Biochem. 1974, 47, 469-474. (22) Goto, J. J.; Butler Gralla, E.; Valentine, J. S.; Cabelli, D. E. J. Biol. Chem. 1998, 273, 30104-30109. (23) Lin, C. T.; Lee, T. L.; Duan, K. J.; Su, J. C. Zool. Stud. 2001, 40, 84-90.

∼80% biuret) was purchased from Sigma-Aldrich. The standards used to size-calibrate the size exclusion column included glutathione (0.31 kDa), cyanocobalamine (1.36 kDa), metallothionein (6.10 kDa, rat liver), myoglobin (17.6 kDa, horse skeletal muscle), pepsin (35.0 kDa, porcine stomach mucosa), albumin (66 kDa, bovine serum), and phospholipase D (97.4 kDa, cabbage) all purchased from Sigma. Stock solutions were prepared in TrisHCl buffer pH 7.4 and stored in acid-cleaned plastic vials in the freezer at -20 °C prior to use. High-purity solutions of 1000 mg L-1 Cu and Zn (BDH), with natural isotopic ratios of 65Cu/63Cu 0.446 and 68Zn/64Zn 0.387,24 were used as calibration stocks for total metal and speciation analysis. Standards were made up in 1% (v/v) aqueous nitric acid for quantifying totals, chromatographic recovery, and extraction efficiencies. For speciation experiments, external calibration was performed by on-column injection of standards prepared in Tris-HCl buffer pH 7.4. Every standard included an amount of EDTA equimolar to its total Cu and Zn concentration. For the isotopic enrichment of SOD, Cu and Zn metal (CK Gas Products, Hampshire, UK) with a certified isotopic abundance of 63Cu 0.8% and 65Cu 99.2% or 68Zn 99.0%, 64Zn 0.38%, and 66Zn 0.33%25 were used. Solutions of isotopically enriched Cu or Zn were made up by dissolving the corresponding metal in concentrated nitric acid and diluting them to a working solution of 10 mg L-1. Preparation of Isotopically Enriched Superoxide Dismutase. Commercially available wild type SOD was dissolved in either 20 mM ammonium acetate pH 7.4 or 20 mM Tris-HCl buffer pH 7.4. These stock solutions were then transferred into 3-mL dialysis membranes with a 5000 molecular weight cutoff (MWCO) (Spectra/Pro CE Flot-A-Lyzer, Spectrum Europe B.V.). Enzyme solutions were dialyzed against 20 mL of buffer for 24 h at room temperature to remove low molecular weight impurities in order to obtain a purified stock solution. Scheme 1 gives an overview of the entire preparation process. The purity of the SOD stock solutions was tested using a total protein and a SOD enzyme activity assay. In addition HPLC with UV detection (280 nm) did not show any additional peak in the dialyzed protein solutions. Metal concentrations and isotopic abundances of the SOD stock solutions were checked by ICPMS using standard solutions with natural isotopic abundances. The isotopically enriched SOD was then prepared in two steps. First, the apoenzyme was prepared by lowering the pH of the protein solutions below 4 using 1 M HCl. This was followed by extensive dialysis against a 20 mM Tris or acetate solution held at the same pH with and without EDTA (Scheme 1). The outer solution was changed at various time points and together with samples of the inner enzyme solution submitted to total metal analysis. The isotopically enriched enzyme was prepared by remetalating the resulting apoenzyme solutions. Apoenzyme solutions were therefore diluted in 20 mM buffer, and the pH was then increased above 5.0 with 1 M KOH. For incorporation of the enriched metals into the apoenzyme, an excess of 65Cu or 68Zn was added. The metals were either added sequentially at different pH values or together at pH 7.4 (Scheme 1). The enzyme solutions were then allowed to stir for at least 24 h at room temperature. A final 24-h dialysis at room temperature in either Tris-HCl or acetate buffer at pH 7.4 was performed (24) Rosman, K. J. R.; Taylor, P. D. P. Pure Appl. Chem. 1998, 70, 217-235. (25) CK Gas Products, UK. Certificate of analysis, order no. 11767, Lot no. 10;2004.

Scheme 1. Production of Isotopically Enriched SOD from the Commercially Available Wild Type Enzyme

to remove any unbound 65Cu and 68Zn. The remetalated enzyme solutions were stored at -20 °C prior to their characterization. Enzyme Characterization by SEC-ICPMS. Characterization of the wild type, apo, and isotopically enriched SOD was carried out using SEC-ICPMS. The degree of 65Cu and 68Zn incorporation into the enzyme as well as recovery and isotopic abundances were evaluated. A Tricorn (Superdex75 10/300 GL, Amersham Biosciences, Bucks, UK) high-performance size exclusion column (SEC), 10 × 300-310-mm bed, 13-µm particle size, was coupled to the nebulizer of an ICPMS quadrupole instrument (SpectroAnalytical Chemistry, Vol. 79, No. 21, November 1, 2007

8383

Table 1. Instrumental Parameters Used for SEC-ICPMS Experiments column flow rate injection volume mobile phase UV detection

SEC Parameters Tricom 10/300 GL Superdex 75 0.8 and 1 mL min-1 100 µL 20 mM Tris-HCl/2 mM CaCl2 pH 7.40 294 (nm)

instrument nebulizer spray chamber gas flow rates

ICPMS Parameters SpectroMass 2000 concentric Meinhard type C cyclonic spray champer cooled at 4 °C coolant 18; auxiliary 1.5; nebulizer 1.1 (mL min-1)

rf power torch cones collision cell gas flow rates isotopes monitored dwell times

1.35 (kW) standard quarz torch 2.5 mm i.d. injector Ni sampler and skimmer none 63Cu, 65Cu, 64Zn, 66Zn, 68Zn, 115In 200 ms for all isotopes

Mass2000, Spectro Analytical, Kleve, Germany), with PEEK tubing. Sample injection was achieved via a metal-free injection valve (model 9125, Rheodyne, Rohnert Park, CA) equipped with a 100-µL PEEK sample loop and a PEEK loop filler port (model 9012, Rheodyne). The mobile phase contained 20 mM Tris-HCl pH 7.4 buffer and was supplied with an HPLC pump (model 2150, LKB Bromma, Uppsala, Sweden) at a flow rate of 1.0 mL min-1. The eluent was freshly prepared with the addition of 2.0 mM calcium chloride to inhibit nonspecific metal binding to the SEC column.26 The SEC column was size-calibrated using different macromolecular standards (concentration 100 mg L-1 as protein). The retention times of the size calibrants were monitored via UV detection (280 nm, 1100 series, Agilent). To monitor the stability of the plasma during the chromatographic run, indium solution (10 µg L-1 in 1% (v/v) aqueous nitric acid) was added postcolumn via a T-piece at a flow rate of 0.15 mL min-1. The ICPMS was optimized daily using a solution of 10 µg L-1 Cu and In in 1% aqueous nitric acid. Conditions and instrumental parameters are shown in Table 1. The following isotopes were measured: 63Cu, 65Cu (spike),64Zn, 66Zn, 68Zn (spike), and 115In (internal standard). Metal concentrations and isotopic abundances of the wild type SOD were determined using on-column calibration with standards of natural isotopic abundance. The concentration and isotopic abundances of the isotopically enriched SOD were determined using reverse isotope dilution by measuring a standard with a natural isotopic abundance of Cu and Zn. All measurements were corrected for mass bias by measuring a standard with natural isotopic abundance. In addition, SOD concentrations were also determined via the specific enzymatic assay. Characterization of the Isotopically Enriched Enzyme by Activity Assay and Native 1D-PAGE. For the further characterization of the isotopically enriched enzyme, an enzyme activity assay was performed. SOD inhibits the autoxidation of pyrogallol, which can be followed spectrophotometically at 420 nm.21 Standards in the range of 0-2 units of SOD mL-1 were prepared in 50 mM Tris-HCl pH 8.2. The samples were diluted in the same buffer so that their activity was within the calibration range. For (26) Inagaki, K.; Mikuriya, N.; Morita, S.; Haraguchi, H.; Nakahara, Y.; Hattori, M.; Kinosita, T.; Saito, H. Analyst 2000, 125, 197-203.

8384 Analytical Chemistry, Vol. 79, No. 21, November 1, 2007

Agilent 7500c concentric Meinhard type C scott double pass cooled at 4 °C coolant 15; auxiliary 0.9; nebulizer 1.25 (mL min-1) 1.58 (kW) standard quartz torch 1.5-mm-i.d. injector Ni sampler and skimmer H2 2.5 (mL min-1) 63Cu, 65Cu, 64Zn, 66Zn, 68Zn, 115In 50 ms for all isotopes

measurement, standards and samples were mixed with 50 mM Tris-HCl pH 8.2 containing 10 mM DTPA, and a fixed volume of 100 µL of 0.2 mM pyrogallol in 10 mM HCl was added. The samples were allowed to stand for 2 min, and the change in absorption of all solutions was then determined over 4 min using a UV-vis spectrophotometer (Lambda 25, Perkin-Elmer). The activity assay was performed on the wild type SOD solutions, apo, and the isotopically enriched enzyme solutions directly from the stock solutions or from fractions collected after HPLC. An additional total protein assay was performed in order to confirm the purity of the SOD solutions, which is possible in combination with the enzyme activity assay. Total protein concentration was assayed using the Bradford assay. Protein samples were diluted to the required concentration of 1-10 µg mL-1 in Tris or acetate buffer. The samples were then mixed 1:1 with Bradford reagent (Sigma-Aldrich) containing Brilliant Blue G dye. Samples were left 20 min at room temperature for incubation, and the absorbance was measured at 595 nm. Total protein concentration was calculated by comparing the values against a standard curve produced using bovine serum albumin standards in the same buffer. The isotopically enriched and wild type enzymes were also characterized by gel electrophoresis under native conditions. The discontinuous native Tris-glycine buffer system was used. All reagents were purchased from Sigma-Aldrich and were used without further purification. The solution of acrylamide/bisacrylamide (Protogel) was purchased from National Diagnostics (Yorkshire, UK). The 12% T% resolving gels, 4% T% stacking gels were hand-cast in Invitrogen (Paisley, UK) disposable plastic cassettes for casting gels 1 mm thick, according to the native Tris-glycine discontinuous system (Invitrogen) technical notes. Aliquots of the samples were mixed with Tris-glycine native sample buffer. Electrophoresis was run in an X-Cell SureLock mini system (Invitrogen) at a constant voltage of 150 V until the dye front reached the bottom of the gel. Runs were performed at room temperature. Gels were immediately stained using the rapid silver stain procedure based on the Amersham Biosciences PlusOne Silver Stain Kit.

Figure 1. Demetalation of wild type SOD under different conditions, triangles, Cu; squares, Zn. Black lines, demetalation at pH 3.8 using EDTA (experiment A). Gray lines, demetalation at pH 2.7 (experiment F). n ) 10 replicates.

Evaluation of the Isotopically Enriched Metalloenzyme as a Standard for SS-IDMS. For testing the stability of the isotopically enriched SOD, an aliquot was spiked either with 0.25 mL of a 10 mM EDTA solution or with an excess of free 65Cu, 68Zn (10 mg L-1 solution in 20 mM Tris buffer). Solutions were made up to a final volume of 2 mL with 20 mM Tris buffer. For the evaluation of the isotopically enriched SOD as a standard in SS-IDMS, the isotopically enriched enzyme was mixed with purified wild type SOD in a 1:1 ratio. These results obtained were compared with those obtained from external calibration and the enzyme activity assay. For SEC-ICPMS of those solutions, a 7500c series ICPMS (Agilent Technologies) equipped with a collision cell for removing interferences with Cu and Zn isotopes was used. The reaction gas was hydrogen. For instrumental parameters and HPLC conditions, see Table 1. The instrument was optimized daily using a solution of 7Li, 89Y, and 205Tl (10 µg L-1 in 1% v/v aqueous nitric acid). An HPLC system (1100 series, Agilent Technologies) including a degasser, autosampler, and UV detector was used for delivering the mobile phase. A 10 µg L-1 In solution in 1% (v/v) aqueous nitric acid was continuously added postcolumn at 0.2 mL min-1 as an internal standard. All connections were made from PEEK or Teflon. The detection was focused on the Cu and Zn isotopes (Table 1). RESULTS AND DISCUSSION Preparation of Isotopically Enriched SOD. On the basis of the work of McCord and Fridovich,18 who demetalated SOD at pH 3.8 using 1 mM EDTA, we demetalated the enzyme under a range of different pH conditions with and without the addition of EDTA (experiments A and B or C-H, Scheme 1). Remetalation of the apoenzyme was then performed by either the combined addition of 65Cu and 68Zn at pH 7.4 or by sequential addition of 65Cu and 68Zn at a range of different pHs (Scheme 1). All experiments were performed using the same solution of isotopically enriched 65Cu and 68Zn. Monitoring total Cu and Zn levels during the SOD demetalation procedure allowed us to calculate the efficiency of the demetalation process under different conditions. Figure 1 shows the time-dependent demetalation of SOD for experiments A and F. After 24-h dialysis at pH 3.8 with EDTA (experiment A), removal of 77.5 ( 12.2% of the total Cu from the enzyme had been achieved. At pH 2.7 without EDTA (experiment

F), 94.2 ( 5.4% of the Cu was removed from SOD after 24 h. Under the same conditions, 76.8 ( 6.8 and 94.1 ( 5.2%, respectively, of the total Zn had been removed. The results show that the use of lower pH has a greater effect on demetalating the enzyme than does the addition of EDTA at higher pH. Demetalation efficiency was considered to be sufficient, since remetalation was performed with solutions highly enriched in 65Cu and 68Zn. Therefore, SOD species can be produced in which the isotopic abundances of the remetalated enzyme are clearly distinguishably from those of the wild type. SEC-ICPMS of the remetalated enzyme solutions from experiments A and B (both demetalated at pH 3.8 in the presence of EDTA) did not result in any detectable Cu or Zn SOD signal. An explanation is the strong nonspecific binding of EDTA to SOD reported elsewere.27 Thus, EDTA was probably not removed from the enzyme solution during dialysis and scavenged the free 65Cu and 68Zn added in the remetalation step. This meant that the 65Cu and 68Zn bound to EDTA could not be transferred to the binding site of apo-SOD. Hence, further demetalation experiments were carried out without the addition of EDTA, but using lower pHs (experiments C-G Scheme 1). Characterization of the Enzyme by SEC-ICPMS. The approach used for the characterization of isotopically enriched SOD is based on SEC-ICPMS since it has been established that this is a gentle method for the analysis of metalloproteins.3,28,29 The method can be performed under physiological conditions preserving the protein-metal complex, it does not require organic solvents, and the conditions are compatible with ICPMS. Experiment C was designed to study the effect of demetalation at low pH without the addition of EDTA. Remetalation followed, with a sequential incubation of first Zn then Cu for 24 h at pH 7.4 (Scheme 1). Figure 2 shows the Cu and Zn chromatograms of wild type and isotopically enriched SOD from experiment C. The Cu and Zn signals for the remetalated enzyme both have an altered isotopic abundance when compared with those of the wild type. The retention time for the metal signals in the remetalated enzyme corresponds to the retention time of the wild type SOD. Table 2 gives a summary of the incorporation efficiency for both the 65Cu and 68Zn isotopes, the isotopic abundances and ratios, and the total metal recovery of the isotopically enriched enzyme obtained from experiment C. In addition to experiment C, further experimental conditions (D-H) were tested to obtain an isotopically enriched enzyme standard characterized by a superior incorporation of 65Cu and 68Zn (Table 2). In every case, the purity and content of the remetalated enzyme solutions were determined using a combination of total protein and activity assay and HPLCUV-ICPMS. The metal concentration and isotopic abundances of the wild type stock solutions and the isotopically enriched SOD solutions were determined as described in the Experimental Section using standards with natural-abundance Cu and Zn. All results were corrected for mass bias. The extent of enrichment of the enzyme with 65Cu and 68Zn can be improved by increasing the treatment time or lowering the pH during the demetalation process. This can be clearly observed as an increased incorpora(27) Valentine, J. S.; Pontoliano, M. W.; McDonnell, P. J.; Burger, A R.; Lippard, S. J. Proc. Natl. Acad. Sci. U.S.A. 1979, 76, 4245-4249. (28) Wang, J.; Houk, R. S.; Dreessen, D.; Wiederin, D. R. J. Am Chem. Soc. 1998, 120, 5793-5799. (29) Boulyga, S. F.; Loreti, V.; Bettmer, B.; Heumann, K. G. Anal. Bioanal. Chem. 2004, 380, 198-203.

Analytical Chemistry, Vol. 79, No. 21, November 1, 2007

8385

Figure 2. (a) Copper SEC-ICPMS chromatogram of wild type SOD: black 63Cu; gray 65Cu. (b) Copper SEC-ICPMS chromatogram of isotopically enriched SOD from experiment C: black 63Cu; gray 65Cu. (c) Zinc SEC-ICPMS chromatogram of wild type SOD: black 64Zn; gray 68Zn. (d) Zinc SEC-ICPMS chromatogram of isotopically enriched SOD from experiment C: black 64Zn; gray 68Zn. All chromatograms were corrected with respect to the signal for the internal standard. Table 2. Characteristic Values of the Isotopically Enriched SOD for Various Treatmentsa isotopic abundance

experiment stating material

incubation efficiency (%) no incubation

65Cu

(%)

32.8 ( 0.1

68Zn

(%)

21.5 ( 0.3

A and B C

E F G H

total SODmetal recovery (%)

enzyme activity (%)

63Cu/65Cu

66Zn/68Zn

0.52 ( 0.01

0.74 ( 0.02

100

100

Cu ) 67.8 ( 5.3 Zn ) 92.1 ( 14.5 Cu ) 53.7 ( 4.5 Zn ) 86.1 ( 9.9 Cu ) 39.7 ( 2.2 Zn ) 71.1 ( 7.9 Cu ) 33.0 ( 2.6 Zn ) 99.0 ( 9.4 Cu ) 43.9 ( 2.2 Zn ) 106.7 ( 9.2 Cu ) 137.4 ( 6.0 Zn ) 31.4 ( 1.3

70.2 ( 2.7

no detected Cu,Zn-SOD signal ) 12.6 ( 0.4 ) 43.2 ( 1.0 ) 17.7 ( 2.1 68Zn ) 66.4 ( 3.2 65Cu ) 27.8 ( 1.1 68Zn ) 67.6 ( 2.9 65Cu ) 71.2 ( 2.7 68Zn ) 100.0 ( 4.2 65Cu ) 82.8 ( 2.7 68Zn ) 94.5 ( 0.5 65Cu ) 91.7 ( 1.3 68Zn ) 52.6 ( 3.9 65Cu 68Zn

D

isotopic ratio

65Cu

42.2 ( 0.5

39.5 ( 0.1

1.44 ( 0.03

0.60 ( 0.01

46.1 ( 0.3

74.4 ( 0.5

1.26 ( 0.02

0.14 ( 0.001

54.1 ( 0.3

75.2 ( 0.9

0.96 ( 0.01

0.13 ( 0.01

81.1 ( 0.2

89.6 ( 0.2

0.25 ( 0.004

0.07 ( 0.002

89.5 ( 0.4

93.4 ( 0.3

0.13 ( 0.005

0.03 ( 0.003

95.2 ( 0.5

58.2 ( 1.3

0.05 ( 0.01

0.25 ( 0.03

54.8 ( 5.8 35.2 ( 7.4 39.4 ( 3.9 42.6 ( 1.4 57.6 ( 8.8

a Values for the isotopically enriched enzymes were compared with the wild type SOD stock solutions, which were used as a starting material and were always measured together with the isotopically enriched enzyme solutions. All measurements were carried out in triplicate. The activity is given as a percentage of the total activity of the enzyme stock solution and not as normally stated in units of enzyme. Letters correspond to the different experimental conditions used in the production process of the isotopically enriched enzyme (Scheme 1).

tion of the enriched isotopes into the remetalated enzyme. A spike suitable for use in SS-IDMS should be highly enriched in the less naturally abundant isotope, so that less spike is needed for the analysis in order to achieve the ideal ratio of unity between the spike and the sample isotopes.30,31 However, in our case, this is however only achieved for Zn, which is in accordance with the literature. In a study aimed at investigating the pH dependence of metal ion binding to the native Zn site of SOD, it was reported that Zn was more easily removed from SOD than Cu at a pH of 3.6.32 (30) Hearn, R.; Evans, P.; Sargent, M. J. Anal. At. Spectrom. 2005, 20, 10191023. (31) Hill, S. J.; Pitts, L. J.; Fisher, A. S. TrAC, Trends Anal. Chem. 2000, 19, 120-126.

8386

Analytical Chemistry, Vol. 79, No. 21, November 1, 2007

The recovery of Zn after its reincorporation into the enzyme is effectively 100% for most of the experiments. This suggests that incubation with Zn prior to Cu is a suitable approach for the remetalation step. However, Cu recovery after reincorporation into SOD decreases as the pH is lowered or treatment time is increased in the demetalation step. This is in contrast with the literature, where demetalation performed at pH 3.8 using EDTA gave 80% reincorporation of Cu after remetalation of the protein.18 In the demetalation step, we use a working pH, which is significantly lower than the proteins’ own isoelectric point of 4.95. Therefore, it is possible that alteration in the conformation of the Cu binding (32) Pantoliano, M. W.; Valentine, J. S.; Mammone, R. J.; Scholler, D. M. J. Am. Chem. Soc. 1982, 104, 1717-1723.

site during the demetalation process, inhibiting further uptake of free 65Cu ions, may occur. The use of such pH conditions might not only have affected the metal binding affinity of the enzyme but also the conformational equilibrium between different enzyme structures (monomeric/dimeric forms). These depend on the presence of intrasubunit disulfide bridges, prosthetic metals, and strong hydrophobic interactions between subunits, which might have been affected by the low working pH and the resulting overall charge on the protein during demetalation.33 This is supported by the fact that when isotopically enriched SOD was spiked with additional free 65Cu no further increase in the Cu reincorporation was observed after SEC separation. These results suggest that the enzyme conformation or the metal binding site might have been altered during the demetalation process; hence, the enzyme is incapable of binding further Cu in the Cu-binding motif. However, to obtain such structural information, techniques such as EXFAS or XANES are necessary, which are capable of characterizing the neighboring environment of the metal cofactors. Experiment H represents an exception. Here 65Cu was added before 68Zn in the remetalation step. A low total Zn but a high Cu reincorporation was obtained for the enriched enzyme (Table 2). This is an indication that free Cu ions nonspecifically bind or migrate to the Zn binding site during remetalation. In general, we can exclude the migration of Zn into the Cu binding site during remetalation since reincorporation of Zn was always effectively 100% and the conditions chosen should inhibit the migration of Zn into the Cu motif.34 Characterization of the Isotopically Enriched Protein by Activity Assay and Native 1D-PAGE. When preparing a standard for isotope dilution analysis, an enriched species with the same behavior and properties as the species under investigation is desired. Therefore, further investigations of the isotopically enriched enzyme standard are necessary. The isotopically enriched enzyme standards were therefore submitted to an activity assay. Since SOD is an enzyme, an activity assay can give specific information about whether metals are correctly incorporated into the enzyme, since its function is metal dependent. Metalloproteins that have no enzyme activity require advanced spectroscopic techniques such as XANES or EXFAS to determine whether the metals have been correctly incorporated into the protein. This study thus shows novelty since we have evaluated the correct metal binding to the enzyme through an activity assay in contrast to previous SS-IDMS studies.14-16 Activities measured for the isotopically enriched enzyme are shown in Table 2. Enzyme activity is given as a percentage of the activity of the wild type enzyme solution used as the starting material in each experiment. In general, longer incubation at lower pH during demetalation decreases the activity of the enriched enzyme standard obtained after metal reincorporation. Enzyme activities correlate well with the total Cu values (Figure 3). The graph shows a good correlation (r2 ) 0.978) with a slope of nearly 1. This indicates that the Cu incorporated into our isotopically enriched enzyme standards is specifically bound to the Cu motif because only those Cu ions are responsible for the enzyme activity.35 While the enzyme (33) Arnesano, F.; Banci, L.; Bertini, I.; Martinelli, M.; Furukawa, Y.; O’, Halloran, T. V. J. Bio. Chem. 2004, 279, 47998-48003. (34) Goto, J. J.; Zhu, H. N.; Sanchez, R. J.; Nersissian, A.; Butler, Gralla, E.; Valentine, J. S. J. Biol. Chem. 2000, 275, 1007-1014. (35) Forman, H. J.; Fridovich, I. J. Biol. Chem. 1973, 248, 2645-2649.

Figure 3. Relationship between recovered total protein Cu and measured enzyme activity for isotopically enriched SOD solutions from different experiments (error n ) 3 replicates). The capital letters correspond to the different experimental conditions from Scheme 1.

activities for all other enriched 65Cu, 68Zn-SODs correlated well with the Cu recovery (Figure 3), only the isotopically enriched SOD from experiment H failed to follow this trend. This confirmed the nonspecific incorporation of Cu into the enzyme since only the Cu site and not the Zn site has been reported to be involved in the enzyme activity.22 All activities measured for the isotopically enriched enzymes were higher than the activities of the demetalated apo-SOD fractions. Activities of the apoenzymes were in general below 15% with an average of 11.7 ( 4.5%. For experiment C, the apoenzyme fraction was still 50% active when compared to the wild type enzyme. However, this corresponds well with the degree of demetalation, which was not as efficient as in the other experiments. Generally, the activity of the enzyme is decreased during the demetalation process where Cu is removed from the enzyme. During remetalation, the enzyme activity is then increased but could not be totally restored since the Cu reincorporation into the remetalated enzyme was not 100%. The results are in contrast with the findings of Forman and Fridovich, who reported a total restoration of the enzyme activity after remetalation.35 The measurements are evidence that the standards contain only fully active Cu/Zn-SOD with some inactive copperfree apo-SOD (E/Zn-SOD). However, only further biochemical characterization via structural and functional experiments can confirm this. These findings highlight a general problem for all metalloproteins. A metal can be bound nonspecifically to a protein and would therefore not represent the native state of the protein, so that the criteria for species specificity are not fulfilled. The incorporation of the metal into the right binding site of a protein cannot be detected simply by applying mass spectrometric techniques alone. It rather requires spectroscopic approaches such as an enzymatic activity assay, as used in our study, or the use of a combination of techniques such as mass spectrometry and NMR as, for example, reported on the study of Zn incorporation into methallothioneins.36 Additionally, the activity assay can be used as a means to determine the concentration and purity of the SOD enzyme solutions. For the purified wild type SOD stock solutions, the results of the total protein concentration determined by the (36) Blindauer, C. A.; Polfer, N. C.; Keiper, S. E.; Harrison, M. D.; Robinson, N. J.; Langridge-Smith, P. R. R.; Sadler, P. J. J. Am. Chem. Soc. 2003, 125, 3226-3327.

Analytical Chemistry, Vol. 79, No. 21, November 1, 2007

8387

Figure 4. Native 1-D PAGE of SOD at different stages of the deand remetalation process from experiment G. Lane a, untreated wild type SOD solution in buffer pH 7.4. Lane b, wild type SOD dialyzed for 24 h against buffer pH 7.4. Lane c, SOD dialyzed for 24h at pH 2.7. Lane d, same as lane c but enzyme transferred into buffer pH 7.4. Lane e, isotopically enriched SOD in buffer pH 7.4. Lane SOD, control native enzyme without any treatment. Lane SODd, SOD denatured at 70 °C with 1.5% SDS. D is the native dimeric form of SOD (32 kDa), 2(M) is the partially folded protein, and 2(U) is the dissociated monomer completely unfolded (16 kDa).

Bradford assay were compared to the results of the enzyme activity assay. This confirmed the purity of the SOD solutions in terms of their protein content. Purity was sufficient with an average 98.8 ( 5.7% of the total protein concentration showing SOD activity. Furthermore, the SOD concentrations determined from the activity assay correlated well with the concentrations obtained via metal determination using ICPMS, as shown below. This confirms that the enzyme was pure in terms of its metal content. For testing whether the isotopically enriched enzyme standard was still in the dimeric form and did not change its conformation during the preparation process, native 1D-PAGE was performed. Native 1D-PAGE preserves the enzyme structure and therefore is a good method to compare the wild type with the isotopically enriched enzyme. Conditions were optimized so that the native dimeric (D), partially folded 2(M) and monomeric 2(U) form of the enzyme could be separated as reported elsewhere.37 Figure 4 shows the native gel of SOD at different stages of the de- and remetalation process (experiment G). It can be seen that the remetalated isotopically enriched SOD has the same migration pattern in the native gel as the untreated wild type SOD and is mainly present in its dimeric form. This is an indication that the isotopically enriched enzyme has the same size, shape, and overall charge as the native protein. In addition, only one protein band was observed, which indicates there is one dominant form of the enzyme present. Although native PAGE is not strictly designed to evaluate protein purity, it supports the fact that the enzyme solutions appeared pure in terms of their protein content. However, structural differences due to the entire preparation process cannot be excluded. Native 1D-PAGE might not be sensitive enough to detect slight changes to the relative structural (37) Manning, M.; Colo´n, W. Biochemistry 2004, 43, 11248-11254.

8388

Analytical Chemistry, Vol. 79, No. 21, November 1, 2007

stability of the different conformations present in equilibrium in solution. It is worth noting the clear differences in the migration between the demetalated enzymes kept at different pHs. While at low pH, the demetalated SOD is mainly present in its partially unfolded form; while at physiological pH, the enzyme seems to be in its dimeric form. This confirms that the metals are not absolutely necessary for the enzyme to maintain its dimeric structure. Stability of the Isotopically Enriched Enzyme Standard under SEC-ICPMS Conditions. An important issue when producing an isotopically enriched protein for use as a standard for SS-IDMS is that the protein-metal complex stays stable during the analytical procedure. Ideally, this means there should be no change in the isotopic metal abundance of either the spike or the sample protein. In practice, it is essential that the enriched isotopes in the spike protein do not undergo rapid exchange reactions either by loss of metals from the protein or by scavenging of free natural-abundance metals. A possible loss of metals from the protein spike with the potential of reacquiring free metals could lead to a change in the equilibrium between the spike and the analyte isotopes. Isotope dilution would not correct for this because the isotopic content of the enriched protein would be altered. However, when sample and spike protein are the same and a possible loss of metals during the speciation process is not reversible then SS-IDMS can be used to quantify the overall loss of the protein-metal complex during the analytical procedure since both the sample and spike protein are affected the same way.13 Isotopically enriched protein standards are useful tools in SS-IDMS since they not only provide an improvement in absolute protein quantification but are also probes for identifying exchange of metals with other ligands, which is otherwise a difficult endeavor. By spiking both the wild type and remetalated isotopically enriched enzyme with an aliquot of a strong chelator and free inorganic Cu and Zn with different isotopic abundances, it was possible to evaluate the stability of the enzyme metal center. In addition, it is possible to establish whether Cu and Zn redistribution occurs under normal chromatographic conditions. Both wild type and isotopically enriched SOD metalloenzyme solutions were spiked with excess EDTA (at pH 7.4, rather than at low pH as during demetalation) and with a solution containing free naturalabundance and isotopically enriched Cu and Zn. This was to represent a possible sample matrix and to determine whether this had any effect on the elution characteristics or the isotopic content of the enzyme solutions. The results presented in Figure 5 show that neither the addition of EDTA nor the addition of free Cu and Zn alters the elution profile of the wild type and isotopically enriched enzymes. The signal areas and the corresponding isotopic abundances did not change during the procedure. The chromatographic recoveries for native SOD after spiking with an EDTA solution were 98.1 ( 1.3% for Cu and 97.7 ( 1.7% for Zn. Recoveries of 95.0 ( 1.2% for Cu and 94.9 ( 1.2% for Zn were achieved when the enzyme was spiked with isotopically enriched solutions of 65Cu and 68Zn (n ) 3 replicates). The isotopically enriched enzyme solution (experiment G) was analyzed after treatment with EDTA, and the recoveries were 98.2 ( 0.6% for Cu and 103.9 ( 3.4% for Zn. Addition of a solution of naturalabundance Cu and Zn resulted in a recovery of 100.3 ( 0.8% for

Figure 5. SEC-ICPMS chromatograms of unspiked and spiked wild type and isotopically enriched SOD. (a) 65Cu and 68Zn SEC-ICPMS chromatograms of wild type unspiked SOD; black line and SOD spiked with free 65Cu/68Zn; gray line. (b) 65Cu and 68Zn SEC-ICPMS chromatograms of wild type unspiked SOD; black line, and SOD spiked with 10 mM EDTA gray line. (c) 65Cu and 68Zn chromatograms of unspiked isotopically enriched SOD; black line, isotopically enriched SOD spiked with free 63Cu/64Zn; dark gray line and isotopically enriched SOD spiked with 10 mM EDTA light gray line. Spiking showed no effects, and chromatograms in (a-c) are superimposed upon each other. Elution rates were at 1 mL/min (a) and at 0.8 mL/ min (b, c). The Zn chromatograms have been upraised for clarity.

Cu and 100.5 ( 1.2% for Zn (n ) 3 replicates). Values were calculated from the unspiked wild type and isotopically enriched SOD stock solutions used as starting materials. These results indicate no loss or reacquisition of enzyme-bound metal on spiking with EDTA or free metals during sample handling. For further proof of the enzyme’s integrity during chromatography, activities of the enzyme stock solutions and the enzyme fractions eluted from SEC were measured and compared. The activities measured for both the wild type enzyme and the isotopically enriched enzyme were the same before and after the chromatographic separation. For the wild type enzyme, an activity of 5233 ( 489 U was measured in the collected fraction after chromatography, corresponding to 99.3 ( 29.7% of the original activity (n ) 3 replicates). For the isotopically enriched enzyme

fraction, an activity of 5380 ( 301 U, corresponding to a recovery of 99.0 ( 6.7% of the original activity after chromatography, was measured (n ) 3 replicates). This indicates that no change to the enzymes during the chromatographic run takes place. It should be noted that in this case the native and the isotopically enriched enzymes were not from the same batch, so that the similarity in measured activities is coincidental. When apo-SOD was spiked with a natural-abundance inorganic Cu and Zn solution (chromatogram not shown), it was clearly seen that the enzyme can take up any available Cu and Zn present in the solution in which the enzyme is being analyzed. That is especially important in the case of column-bound metal, which can lead to species-species transfer. This observation is expected and highlights the fact that isotopically enriched enzymes, which are used as spikes for SS-IDMS, must be present in a completely metal-saturated form and that the conditions must not lead to the exchange or loss of those metals.14 However, metal uptake by the apoenzyme is only possible if the metal binding site of the enzyme is still in its original native form and has not been altered during sample preparation or analysis. The results indicate that both the wild type enzyme and the isotopically enriched enzyme are stable under the SEC conditions, and therefore, SS-IDMS using isotopically enriched SOD is possible. Further spiking experiments were carried out as seen below to further evaluate the potential of isotopically enriched SOD in SS-IDMS. Evaluation of Isotopically Enriched SOD in SS-IDMS with a Single Wild Type SOD Standard. The enriched enzyme solution was mixed (1:1, v/v) with an aliquot of a wild type, purified SOD stock solution. Since all variables were known or measurable, the concentration of the wild type SOD in the solution can be quantified using SS-IDMS. The results could then be compared with those from an external calibration or an enzyme activity assay to evaluate the potential of this new quantification method for SOD. Figure 6 shows the elution profiles of wild type and isotopically enriched SOD and a mixture of the two. As expected, the peak profile of the mixture represents the sum of the other two profiles. Analysis of the wild type SOD stock solution with external calibration gave a total Cu concentration of 1.54 ( 0.073 mg L-1. Metal concentrations can be converted into protein concentrations since the stoichiometry of the protein-metal complex is known for both metals, and has a ratio of 2:1 metal per protein.19 The molarity of the metal present in the protein is calculated using, in the case of wild type SOD, the measured natural abundance of the metals, which is then converted into the protein molarity from its metal/protein ratio. The Cu concentrations obtained for wild type SOD with external calibration therefore correspond to 12.1-12.2 µmol L-1 enzyme (n ) 2). For Zn, a concentration of 1.46 ( 0.072 mg L-1 corresponding to 11.211.7 µmol L-1 enzyme (n ) 2) was determined. In contrast, with SS-IDMS, a total Cu concentration of 1.68 ( 0.066 mg L-1, corresponding to 13.16 ( 0.52 µmol L-1 enzyme (n ) 3 replicates), and a Zn concentration of 1.34 ( 0.04 mg L-1 corresponding to 10.2 ( 0.35 µmol L-1 enzyme, were obtained. For comparison, the activity assay gave a enzyme concentration of 12.8 ( 1.20 µmol L-1 (n ) 3 replicates), which corresponds to a Cu concentration of 1.63 ( 0.15 mg L-1 and a Zn concentration of 1.68 ( 0.16 mg L-1. The Cu results from the external calibration and the activity assay are in good agreement with the results obtained for the Analytical Chemistry, Vol. 79, No. 21, November 1, 2007

8389

an activity assay. Future work will focus on the determination of SOD in biological samples using this SS-IDMS approach. For this, multidimensional separation mechanisms will be necessary to achieve complete separation of metal-containing species. The establishment of the stability of the analyte in a real sample matrix containing free metals and other chelating agents under extraction and separation conditions has to be carefully considered. This is currently difficult because of the lack of certified reference materials in this area.

Figure 6. (a) 65/63Cu SEC-ICPMS isotopic ratio chromatogram of wild type SOD, 1; isotopically enriched SOD, 2; and a mixture (1:1, v/v) 3. (b) 68/64Zn SEC-ICPMS isotopic ratio chromatogram of wild type SOD, 1; isotopically enriched SOD, 2; and a mixture (1:1, v/v) 3. Vertical lines indicate elution positions of the molecular weight standards used for column calibration (kDa).

SS-IDMS and show no significant difference. The recovery of 109 ( 3.5% for the SS-IDMS experiment compared to that of external calibration indicates eventual loss of the sample during the chromatographic run, which cannot be corrected for when external calibration is used. This is supported by a mass balance calculation, which showed that only 71.8-93.3% of the total CuSOD could be recovered after SEC. In contrast to Cu, the results for Zn from the three methods show significant differences. The lowest Zn-SOD concentration was determined using the SS-IDMS method while the activity assay gave the highest Zn concentrations. An explanation is a possible blank problem during the Zn determination. Excess Zn can bind nonspecifically to SOD and therefore influence the results obtained for both IDMS and external calibration. In addition, Zn has only structural importance for the enzyme, and therefore, the activity assay is only indirectly related to its Zn content. The results for Cu show that it is possible to use an SOD isotopically enriched in 65Cu for the quantification of wild type SOD using IDMS. This is supported by the fact that we can be sure that Cu is correctly incorporated into our enzyme by use of

8390 Analytical Chemistry, Vol. 79, No. 21, November 1, 2007

CONCLUSIONS To our knowledge, this is the first report showing the production of an isotopically enriched metalloenzyme, containing two different isotopically enriched metal cofactors, which can be used as a standard in species-specific IDMS. The study shows that it is possible to chemically produce an isotopically enriched SOD with various isotopic abundances of the enriched 65Cu and 68Zn. The stability of the SOD standard under speciation conditions was tested, and no exchange of the 65Cu and 68Zn was observed after addition of excess Cu, Zn, or a chelating agent. Using an activity assay, we have shown that isotopically enriched Cu used for labeling the enzyme can be incorporated into the correct binding motif. Furthermore, an enzyme sample of known SOD concentration was spiked with an isotopically enriched enzyme standard. The results showed that the Cu values obtained from the SS-IDMS approach were in good agreement with those obtained by external calibration and using the activity assay. This application shows the potential of using isotopically enriched SOD for the accurate quantification of that enzyme in real samples. General limitations of this approach are that the native enzyme activity was not completely restored in the isotopically enriched enzyme. In addition, denaturation of the protein might have taken place during the demetalation procedure especially at low pH. It has to be mentioned that despite all the advantages of SS-IDMS, a huge effort is necessary to produce and characterize an isotopically enriched protein spike, such as described here. Therefore, this approach might only be sensible in practice for the assay of essential proteins or biomarkers. ACKNOWLEDGMENT The authors acknowledge the UK Engineering and Physical Research Council for support (EPRSC GRS98689/01 to J.F. and GR/S98696/01 to J.E.T.-O.) and the Department of Chemistry at the University of Aberdeen for the award of a Ph.D. studentship. J.E.T.-O. gratefully acknowledges funding from the Analytical Chemistry Trust Fund, the RSC Analytical Division, and EPSRC. Received for review July 2, 2007. Accepted August 7, 2007. AC071397T