Anal. Chem. 2001, 73, 4154-4161
Observation of the Iron-Sulfur Cluster in Escherichia coli Biotin Synthase by Nanoflow Electrospray Mass Spectrometry H. Herna´ndez,† K. S. Hewitson,‡ P. Roach,§ N. M. Shaw,| J. E. Baldwin,‡ and C. V. Robinson*,†
Oxford Centre for Molecular Sciences, Oxford University, New Chemistry Laboratory, South Parks Road, Oxford, OX1 3QT, U.K., Dyson Perrins Laboratory, Oxford University, South Parks Road, Oxford, OX1 3QY, U.K., Department of Chemistry, University of Southampton, Highfield, Southampton, SO17 1BJ, U.K., and Biotechnology Research, Lonza A.G., CH-3930 Visp, Switzerland
Biotin synthase from Escherichia coli was analyzed by nanoflow electrospray ionization mass spectrometry. From solution conditions in which the protein is in its native state, a distribution of monomeric, dimeric, and tetrameric species was observed. The distribution of these species was sensitive to changes in ionic strength: in the positive ion spectrum, biotin synthase at low ionic strength (pH 7.0-8.5) yielded less than 10% dimer. The masses of the monomeric species were consistent with the presence of a [2Fe-2S] cluster with a mass difference of 175.3 Da from the apomonomer with one disulfide bond. Despite the molecular mass of the noncovalent dimer (77 kDa), it was possible to observe a dimeric species containing one iron-sulfur cluster in both positive and negative ion spectra. Additionally, observation of a series of charge states assigned to the apodimer indicated that binding of the iron-sulfur cluster was not required to maintain the dimer. Binding of Cu2+ to biotin synthase was also observed; in the presence of excess chelating agent, free metals were removed and the iron-sulfur cluster remained intact. Evidence for the coordination of the iron-sulfur cluster in biotin synthase was obtained in a tandem mass spectrometry experiment. A single charge state containing the cluster at m/z 2416.9 was isolated, and collision-induced dissociation resulted in sequential loss of sulfur and retention of Fe3+. Electrospray ionization (ESI) mass spectrometry is increasingly being applied to the characterization of metalloproteins.1,2 The strength of the technique is the ability to determine the stoichiometry of metal-protein complexes based on mass measurement of individual species. A number of iron-sulfur proteins have been studied by ESI mass spectrometry, including rubredoxins,3-6 * Corresponding author: (e-mail)
[email protected]; (fax) (44) 01865 275948. † Oxford Centre for Molecular Sciences, Oxford University. ‡ Dyson Perrins Laboratory, Oxford University. § University of Southampton. | Lonza A.G. (1) Loo, J. A. Mass Spectrom. Rev. 1997, 16, 1-23. (2) Yu, X.; Wojciechowski, M.; Fenselau, C. Anal. Chem. 1993, 65, 5, 13551359.
4154 Analytical Chemistry, Vol. 73, No. 17, September 1, 2001
ferredoxins,5,7-10 and iron hydrogenase.11 The intact iron-sulfur clusters in these small metalloproteins (generally 90% protein as judged by SDSPAGE analysis were combined and concentrated to ∼50 mg/mL before being stored at -80 °C. Native Gel Electrophoresis. Native gel (20.0%) electrophoresis was performed using standard techniques with a Bio-Rad Mini Protean II kit. Iron Measurement. Iron content was measured by the colorimetric method of Fish.34 After release of the bound iron from the protein using HCl and KMnO4, an Fe-ferrozine complex was formed and the absorbance at 562 nm used for measurement of iron. Buffer Exchange and Sample Preparation for Mass Spectrometry. Aliquots (5 µL) of purified biotin synthase were buffer exchanged into water or ammonium acetate using Micro Biospin-6 (33) Sambrook, J.; Fritsch, E. F. and Maniatis, T. Molecular Cloning: A Laboratory Manual, Cold Spring Habor Laboratory, Cold Spring Habor, NY, 1989. (34) Fish, W. W. Methods Enzymol. 1988, 158, 357-364.
Analytical Chemistry, Vol. 73, No. 17, September 1, 2001
4155
Figure 2. Comparison of positive and negative ion nanoflow ESI mass spectra of biotin synthase in ammonium acetate showing the distribution of oligomeric species. (a) Positive ion ESI spectrum from 10 µM biotin synthase in 20 mM ammonium acetate, pH 7.0; cone 130 V; extractor 0 V; indicated analyzer pressure 1.0 × 10-5 mbar. (b) Negative ion ESI spectrum from 10 µM biotin synthase in 10 mM ammonium acetate, pH 7.0; cone 150 V; extractor 0 V; indicated analyzer pressure 1.0 × 10-5 mbar.
chromatography columns (Bio-Rad Laboratories, Hercules CA). Samples were diluted to 50 µL in either water or ammonium acetate prior to loading on the column. The pH of ammonium acetate solutions and water was adjusted using acetic acid and ammonium hydroxide. The buffer was exchanged twice, and the resulting samples were stored at -80 °C. The final concentrations of the samples after buffer exchange were determined using Bradford reagent35 with bovine serum albumin as the standard. Nanoflow ESI Mass Spectrometry. All data were acquired using a quadrupole time-of-flight (Q-ToF) mass spectrometer (Micromass UK Ltd., Altrincham, U.K.). Cesium iodide in water was used to calibrate the instrument in positive and negative ion modes, and denatured myoglobin was used to check the calibration. The acquisition range was 50-10 000 m/z (manual pusher set at 180 µs) with an acquisition step of 5 s. Positive and negative ion spectra were acquired in continuum mode using the same tuning parameters (in magnitude, but reversed in polarity). Samples were loaded into borosilicate capillaries, 1.0 mm o.d. × 0.5 mm i.d. (Clark Electromedical Instruments, Reading, U.K.), which were drawn down to a fine taper and coated with gold inhouse. The capillary tip was cut manually under a stereomicroscope to give the required diameter and flow. A nitrogen backing gas line was used to initiate and maintain a flow from the capillary. Nitrogen at room temperature was also used as drying gas, and the ESI source was not heated. The cone/extractor voltages and analyzer pressure required to maintain the bound cluster were in the following ranges: 130150 V cone, 0-20 V extractor, and 7.5 × 10-6 to 1.3 × 10-5 mbar analyzer pressure read-back. Collisional cooling was achieved
through adjustment of the source rotary pump isolation valve to give the required pressure in the source and transfer hexapole region. The exact source conditions are given in the figure legends. MS/MS spectra were acquired with quadrupole mass resolution set to give a peak width at half-height of ∼1 m/z for parent ion selection. Argon was used as the collision gas. All spectra were processed using MassLynx software (Micromass UK Ltd., Altrincham, U.K.), and masses are given as the experimental mass without correction for the charge on the cluster. [In calculating the expected mass difference between apobiotin synthase and the cluster-bound form, the formal charge on the cluster was included as the deconvolution algorithm in mass spectrometry software assumes that all charge is derived from protons.36 The experimental molecular mass for both negative and positive ion spectra of the holoform will differ by -1.008n from the sum of the neutral apoprotein and iron-sulfur cluster masses, where n is the number and polarity of formal charge on the ironsulfur cluster and 1.008 is the average mass of hydrogen. The expected experimental mass for the holo form with a [2Fe-2S]2+ cluster is given by aposeq - 2H + 2Fe + 2S, where aposeq is the molecular mass of apobiotin synthase calculated from the sequence with eight free cysteines (SWISS-PROT P12996).]
(35) Bradford, M. M. Anal. Biochem. 1976, 72, 248-254.
(36) Edmonds, C. G.; Smith, R. D. Methods Enzymol. 1990, 193, 412-433.
4156 Analytical Chemistry, Vol. 73, No. 17, September 1, 2001
RESULTS The Oligomeric Nature of Biotin Synthase. The positive and negative ion ESI mass spectra of 10 µM biotin synthase in ammonium acetate are shown in Figure 2. Four distinct chargestate distributions were observed in both ionization modes. The relative intensities of the charge-state distributions from monomeric species were strongly dependent on the source conditions
Table 1. Calculated and Experimental Molecular Masses for Apo- and Holobiotin Synthase Monomer and Dimer calculated values (Da) mol massa
∆mb
apo apo -2H apo -4H apo + [2Fe+2S]2+
38 517.04 38 515.03 38 513.01 38 690.85
2.0 4.0 173.8
apo + Cu2+
38 578.57
61.5
(apomonomer)2
77 034.08
apomonomer + holo monomer (apomonomer)2 + EDTA
77 207.89
173.8
77 326.33
292.2
apomonomer + holomonomer + EDTA (apomonomer)2 + 2EDTA
77 500.14
466.0
77 618.58
584.4
apomonomer + holomonomer + 2EDTA
77 792.39
758.0
biotin synthase
experimental values (Da) mol mass ( SDa,c
∆m ( SDd
38 515.0 ( 1.3 38 511.1 ( 0.4 38 689.5 ( 1.9 38 687.1 ( 0.6 38 576.8 ( 0.6 38 573.5 ( 1.0 77 050.7 ( 11.9 77 029.4 ( 2.7 77 224.0 ( 7.0 77 206.3 ( 5.8 77 347.5 ( 14.9 77 353.8 ( 3.5 77 512.8 ( 2.2 77 502.5 ( 5.4 77 633.3 ( 6.5 77 637.4 ( 11.0 77 788.9 ( 2.1
mass error (%) 0.005 0.015
174.5 ( 1.6 176.0 ( 0.5 61.8 ( 1.3 62.4 ( 0.6 173.3 ( 7.2 176.5 ( 3.7 296.8 ( 10.8 324.0 ( 5.2 462.1 ( 11.4 472.8 ( 3.1 582.6 ( 9.6 602.7 ( 0.48 760.5 ( 1.6
0.003 0.010 0.005 0.013 0.022 0.006 0.021 0.002 0.027 0.035 0.016 0.003 0.019 0.024 0.004
a As reported by MassLynx software. b Mass difference from apo calculated mass. c Calculated from at least three charge states; positive ion data for the monomer from eight replicates, negative ion data in italics from three replicates; data for the dimer from three replicates, negative ion data in italics. d Average mass difference and standard deviation from apomonomer or dimer from individual spectra.
and collisional cooling.37 The major series of charges states at m/z 4000-5000 have an average mass that corresponds to the homodimer (Table 1). A similar distribution of dimer and monomer was observed from a 10-fold dilution of the biotin synthase solutionsthe fact that the dimer remains the dominant species even after dilution indicates that the dimer is specific. At concentrations above 5 µM, a tetramer was also observed (154 259.0 ( 117.5 Da experimental mass from seven replicates). This is consistent with the observation of a tetramer during purification of biotin synthase with an approximate mass of 160 kDa.23 Binding of the Iron-Sulfur Cluster to the Monomer and Dimer. In the positive ion spectrum, the higher charge states (+20 to +14) assigned to the apomonomer were present with few adducts. For the charge states +13 to +11, two additional series of ions were present with experimental mass differences from the apomonomer of 61.8 and 174.5 Da, respectively (Figure 3a). The series corresponding in mass to +61.8 Da from the apomonomer (m/z 3215.7 in Figure 3a) was the most intense species present at higher charge states and was also observed in the negative ion spectrum (m/z 2142.1, Figure 3b). The mass difference from the apomonomer was consistent with copper binding; to investigate this, biotin synthase was treated with ethylenediaminetetraacetic acid (EDTA) and buffer exchanged to remove the chelating agent. Separate titrations of the protein with solutions of Cu2+ and Fe3+ (over the range 0.3 -1.3 molar excess) were monitored in positive ion mode: an increase in the intensity of the ion at m/z 3215.7 (+12 charge state) was observed for Cu2+ (data not shown). Binding of Fe3+ or Fe2+ (at 1.3 molar excess) was not observed. The second series at +174.5 Da from the apomonomer in Figure 3a (m/z 3225.1) is assigned to the monomer with a bound iron-sulfur cluster, [2Fe-2S]2+. This species was usually only observed at lower charge states in positive ion mode (+13 to +11) but over all of the charge states (-24 to -11) in negative ion (37) Krutchinsky, A. N.; Chernushevich, I. V.; Spicer, V. L.; Ens, W.; Standing, K. G. J. Am. Soc. Mass Spectrom. 1998, 9, 569-579.
mode (for example, Figure 3b, -18 charge state). In both positive and negative ion modes, the iron-sulfur cluster was dissociated by increasing the cone or extractor voltages by 10-20 V or reducing the amount of collisional cooling. A summary of the experimentally determined masses for the monomers is included in Table 1. Examination of the peaks assigned to the monomer after addition of an excess of EDTA (Figure 3, upper traces) showed that the iron-sulfur cluster remained bound to biotin synthase (even in a 65 molar excess of chelating agent). The copper adduct at +62 Da and other minor adducts were removed by EDTA yielding the apo form. Given the relative increase in intensity of apobiotin synthase in the presence of an excess of EDTA, it is apparent from the intensity of the iron-sulfur cluster species in Figure 3 that the bound cluster is quite stable under these solution conditions. This was also found to be the case for the 77-kDa dimer as, in the presence of EDTA, a well-resolved species was assigned to the dimer with one bound cluster (Table 1 and Figure 4). The presence of the chelating agent brought about a significant improvement in separation of the dimer components without disrupting the binding of the iron-sulfur cluster. With this resolution, it was possible to confirm that a significant amount of dimer with two bound clusters was not present as anticipated from the yield of holobiotin synthase (usually ∼50%) from the aerobic purification procedure.30 Effect of Ionic Strength. In positive ion mode, after equilibration in 20 mM ammonium acetate at pH 5.0, 6.0, 7.0, and 8.0, the proportion of monomer to dimer was unchanged. Variation of the ionic strength over the range 5-100 mM at pH 7.0 also did not affect the relative intensities of monomer to dimer. However, a significant effect on the monomer/dimer distribution in both positive and negative ion modes was observed on reducing the ionic strength by exchanging the buffer from ammonium acetate to water. Over the range pH 7.0-8.5, the positive ion spectra obtained from aqueous solution in the absence of buffer showed a large shift toward the monomer, with very little dimer present Analytical Chemistry, Vol. 73, No. 17, September 1, 2001
4157
Figure 3. Expansion of the monomer charge states from the spectra in Figure 2 showing the bound iron-sulfur cluster in the presence (upper traces) and absence (lower traces) of a 30-fold molar excess of EDTA: (a) positive ion, +12 charge state; (b) negative ion, -18 charge state.
(∼10%, based on the ratio of peak heights over monomer and dimer charge states, Figure 5a). The same source conditions (in magnitude) were also used to obtain a negative ion spectrum (Figure 5b) in which the shift toward monomers was less marked compared to positive ion mode. For both ion modes (Figure 5c and d), the iron-sulfur cluster was observed although, in common with the spectra obtained from the buffered solutions, the higher negative ion charge states (-21 to -10) retained the cluster whereas in positive ion mode the cluster was generally only present at the lower charge states (+15 to +10). It is interesting to note that although the distribution of monomeric and dimeric species was sensitive to ionic strength, the proportion of cluster is largely independent of solution conditionssthe largest effect appears to be the ionization polarity. Given that the spectra shown in Figure 5 were acquired by using source conditions similar to those presented in Figure 2, it is apparent that the dimer interactions, in the gas phase at least, are not maintained to the same extent at low ionic strength. This conclusion was further supported by the results of native PAGE, which showed that the 4158 Analytical Chemistry, Vol. 73, No. 17, September 1, 2001
Figure 4. Expansion of a single dimer charge state transformed onto a mass scale obtained from biotin synthase in an excess of EDTA. Solution and source conditions for positive and negative ion modes were as described in Figure 2. Centroid m/z values have been converted to mass (Da). Peaks labeled D are the apodimer: (a) positive ion, +19 charge state; (b) negative ion, -19 charge state.
dimer was present in both the water sample at pH 8.5 and ammonium acetate solution at pH 7.0 (data not shown). Taken together, these data strongly suggest that the monomer observed in the spectra resulted from dissociation of the dimer (during either ionization or in the gas phase) due to a reduction in dimer interactions on decreasing the ionic strength. MS/MS of the Iron-Sulfur Cluster. To investigate further the stability of the iron-sulfur cluster, an individual charge state assigned to the iron-sulfur cluster (m/z 2416.9, -16 charge state) was isolated and dissociated in MS/MS mode. At collision cell voltages in the range 20-40 V, two well-resolved product ions were observed corresponding to mass differences of 35.2 and 67.2 Da from the precursor ion (Figure 6). No other product ions were detected. Further increasing the collision energy to 50-60 V led to peak broadening and loss of signal intensity. The most likely explanation for the product ions observed is loss of one and two sulfide ions. In assigning these product ions, the molecular mass of the cluster ion was calculated as 38 688.5 Da for the parent ion, assuming that the -16 charge state resulted from loss of 18 protons as the cluster carries a charge of +2. Loss of S2- would presumably involve loss of two additional protons to maintain the
Figure 5. Positive and negative ion ESI mass spectra of biotin synthase in aqueous solution at low ionic strength. The spectra were acquired using the same source conditions (cone 150 V, extractor 0 V, indicated analyzer pressure 1.0 × 10-5 mbar). Dimers are indicated by an asterix and the +12 and -18 monomer charge states by MCxx. (a) Positive ion ESI mass spectrum from 6 µM biotin synthase in water, pH 8.5. (b) Negative ion ESI mass spectrum from 30 µM biotin synthase in water, pH 8.5. The monomer/dimer distribution at a concentration of 6 µM was unchanged although the peak widths were greater. Insets: Expansion of the monomer charge states for positive and negative ion spectra showing retention of the iron-sulfur cluster in water: (a) +12 monomer charge state; (b) -18 monomer charge state.
Figure 6. Negative ion ESI MS/MS spectra showing loss of sulfide ions from the -16 charge state of the monomer. The spectra were obtained from 8 µM biotin synthase in water, pH 8.0 with source conditions: cone 150 V, extractor 0 V. Collisional cooling and argon collision gas were used (indicated analyzer pressure 3.1 × 10-5 mbar)
-16 charge state observed in the spectra, yielding a calculated m/z for the product ion of 2414.8. Another S2-/2H+ loss would generate an ion at m/z 2412.6. These calculated m/z values due to loss of sulfide ions from the cluster are in good agreement with the two product ions observed by MS/MS and provide further evidence for the composition of the cluster. DISCUSSION Effect of Ionization Polarity and Ionic Strength on Dimer Interactions. The observation of biotin synthase predominantly
as a dimer, in the mass spectra acquired from ammonium acetate at pH 7.0, is evidence that biotin synthase is dimeric under these solution conditions and is in agreement with native gel electrophoresis (data not shown) and gel filtration data.22 The observation of resolved peaks, corresponding in mass to the apodimer, in the positive and negative ion spectra acquired in the presence of EDTA demonstrates that the iron-sulfur cluster in the [2Fe-2S]2+ form is not essential for formation of the dimer. Moreover, the shift in the monomer/dimer distribution, which was apparent on decreasing the ionic strength of the sample solution, and the differences between positive and negative ion spectra obtained at low ionic strength both suggest that electrostatic interactions between monomer subunits may be important in maintaining the dimer. It is anticipated that some charge separation occurs in the nanoflow capillary such that ions of a polarity opposite to the capillary voltage may be drawn toward the capillary walls and discharged.38-40 If counterions are important in maintaining the dimer, a switch in ionization polarity might be expected to alter the monomer/dimer distribution as the movement of counterions in the capillary is reversed. This appeared to be the case as it was noted that if the ionization polarity was changed from positive to negative, the first few negative ion spectra showed a decrease in the proportion of dimer compared to subsequent spectra. Given that the dimer was maintained to a greater extent at low ionic strength in negative ion mode, this implies that negatively charged counterions are more important than positively charged ions in (38) Ikonomou, M. G.; Blades, A. T.; Kebarle, P. Anal. Chem. 1990, 62, 957967. (39) Ikonomou, M. G.; Blades, A. T.; Kebarle, P. Anal. Chem. 1991, 63, 19891998. (40) Cole, R. B. J. Mass Spectrom. 2000, 35, 763-772.
Analytical Chemistry, Vol. 73, No. 17, September 1, 2001
4159
maintaining the dimer in the gas phase. Factors other than electrophoretic processes, such as electrochemically induced pH changes, may also influence the monomer/dimer distribution. The major difference between the positive and negative ion spectra recorded for biotin synthase in either water or ammonium acetate was the presence of the cluster at higher charge states in the negative ion spectra. Higher charge states can be indicative of a more unfolded conformation:41 loss of the cluster from a partially folded conformation would be expected to occur more readily than loss from the native conformation, which appears to be consistent with the positive ion spectra. Since the cluster is positively charged, it would be expected to bind more tightly to a negatively charged protein even if the protein was partially unfolded. This could explain the presence of the cluster over all monomer charge states in the negative ion spectra. However, the MS/MS data (see below) suggest that the specific protein-cluster interactions are maintained at higher charge states in negative ion. This would imply that unfolding has not occurred to such an extent that the interaction between the negatively charged protein and positively charged cluster is simply electrostatic. Binding and Dissociation of the Iron-Sulfur Cluster. In MS mode, dissociation of the cluster from the monomer yielded the apomonomer. No intermediates were detected, either due to their absence or their coincidence with the copper-bound species. The observation of copper binding in preference to iron was unexpected and probably due to a low level of copper ions in solution from the reagents employed during purification and sample preparation. A coordination number of four with a distorted trigonal-pyramidal geometry is usual for blue copper proteins;42 the copper binding observed in biotin synthase may occur within the cluster binding site through the four cluster ligands. Further evidence that suggests that the copper binding is specific is the absence of a copper adduct in the presence of the bound ironsulfur cluster. The spectra obtained with EDTA present in solution show that binding of the copper is a weaker interaction compared to the iron-sulfur cluster under these conditions as the copper was completely removed by the chelating agent whereas the cluster was observed intact. The +62-Da species, attributed to copper by titration and chelation with EDTA, is however also consistent with the addition of two sulfur atoms. Observation of persulfide species in the ESI mass spectra of iron-sulfur proteins has been reported.14 In the case of biotin synthase, it is not possible from the results obtained to determine whether any contribution to the +62-Da species is due to persulfide formation. However, complete removal of this species in the presence of EDTA would strongly suggest that persulfide species were not present. It is possible that intramolecular disulfide bond formation had occurred between two of the eight cysteinyl residues in the protein monomer as the experimental mass for the apomonomer differed from the calculated value by -2 Da in positive ion mode. Loss of an iron-sulfur cluster has been shown by mass spectrometry to lead to disulfide bond formation across the cysteinyl ligands in the cluster binding site in smaller proteins such as rubredoxin (41) Chowdhury, S. K.; Katta, V.; Chait, B. T. J. Am Chem. Soc. 1990, 112, 90129013. (42) Guss, J. M.; Merritt, E. A.; Phizackerley, R. P.; Freeman, H. C. J. Mol. Biol. 1996, 262, 686-705.
4160
Analytical Chemistry, Vol. 73, No. 17, September 1, 2001
from Clostridium pasteurianum and ferredoxins from Pyrococcus furiosus.14 Given that samples were prepared in nonreducing conditions and that the yield of holobiotin synthase from the purification procedure was ∼50%, some apo form was expected in solution. Without the iron-sulfur cluster, oxidation of the cysteinyl ligands in solution is probable. If so, two apo forms of biotin synthase would be present in the mass spectrasone with a disulfide bond formed in solution and one with free thiols due to in-source dissociation of the cluster. The mass difference of 2 Da would not be resolved given the mass of the protein, resolution of the instrument, and presence of salt adducts. However, the average mass of 38 515.0 ( 0.7 Da suggests that the majority of the apo form observed in the mass spectra contains one disulfide bond. Additionally, the experimental mass difference (calculated as an average from individual spectra) between the apo form and the bound iron-sulfur cluster from positive and negative ion spectra was 175.3 Da. This value is more consistent with the calculated difference of 175.8 Da for an apomonomer with two oxidized cysteines as opposed to 173.8 Da for the fully reduced protein (Table 1), strongly implying that a proportion of the apo form contains a single disulfide bond. In contrast to the MS data, the MS/MS spectra showed that it was possible to observe two intermediates due to loss of the sulfide ions from the cluster. In addition to the mass difference in MS mode, these product ions are further evidence for the ironsulfur cluster. Loss of iron or the intact cluster was not observed over the range of collision cell voltages used, which indicates that the interaction between Fe3+ and the protein ligands is stronger than the interaction between the cluster ions under these conditions. The results demonstrate the potential advantages of MS/ MS over MS mode; from comparison of the MS/MS and MS spectra, it is apparent that ions resulting from partial dissociation of the cluster (through loss of sulfides) in MS mode would not be resolved from salt adducts on the adjacent copper-bound monomer peak. It is also possible that loss of sulfides did not occur in MS mode due to differences between the energy of insource collisions and dissociation in the collision cell. It has been shown that a 34S-labeled cluster in biotin synthase can exchange with 32S in solution,26 which is consistent with the MS/MS data presented here. Moreover, observation of labile sulfides by MS/ MS supports the current mechanism proposed for biotin synthase activity where the sulfur atom inserted into dethiobiotin is derived from the cluster itself.26-28 CONCLUSIONS The results shown here demonstrate the feasibility of resolving an iron-sulfur cluster within a 77-kDa noncovalent protein dimer. This was possible despite the added complexity of nonspecific binding of other metal ions and adducts. Furthermore, the observation of the cluster allowed an examination of the stability of the protein-protein and protein-cluster interactions over a wide range of solution and mass spectrometry conditions. More generally, these results add to the growing body of information that underpins the application of mass spectrometry to metalloproteins where the oligomeric nature of the protein as well as the stoichiometry of metal binding can be determined solely from mass measurements.
ACKNOWLEDGMENT This is a contribution from the Oxford Centre for Molecular Sciences, which is supported jointly by the BBSRC, EPSRC, and MRC. K.S.H. is supported by a BBSRC studentship and funding from Lonza U.K. The research of C.V.R. and P.R. is supported by the Royal Society through University Research Fellowships. The
authors thank the other members of the mass spectrometry group (C.V.R.) for useful discussion. Received for review March 5, 2001. Accepted June 15, 2001. AC0102664
Analytical Chemistry, Vol. 73, No. 17, September 1, 2001
4161
