1355
Anal. Chem. 1993, 65, 1355-1359
Assessment of Metals in Reconstituted Metallothioneins by Electrospray Mass Spectrometry Xiaolan Yu,’Marek Wojciechowski,+and Catherine Fenselau Department of Chemistry and Biochemistry, University of Maryland Baltimore County, Baltimore, Maryland 21228
A method has been developed that combines electrospray ionization mass spectrometry with pH control to provide analysis of metals in native or reconstituted metallothioneins. These metalloproteins cooperatively bind seven divalent metal ions, most commonly Zn2+and Cd2+. Since the protein is denatured and metal ions are lost below pH 3, the pH of the electrospraysolution is critical to successful results. The metal-free apoprotein was detected with its most abundant ions in a charge state of 6+, while the folded metallothionein-metal complexes were observed with lower charge states. The retention of seven metals in the molecular ions detected is consistent with the hypothesis that metallothionein retains its conformation in the gas phase. This mass spectrometric technique can be used to determine rapidly and accurately how many and what cations are incorporatedper moleculeof protein. Information about molar distributionsand estimatesof relative abundances of various complexes in the sample can be acquired in a single measurement.
INTRODUCTION Metalloproteins are involved in many biochemical processes. The metals they bind and the number of metal ions bound are often crucial to maintaining three-dimensional structure and proper functioning. Although there have been applications of anodic stripping voltammetry techniques,’ metal contents in metalloproteins are usually analyzed by atomic absorption spectrophotometry (AAS)or inductively coupled plasma (ICP) atomic emission spectroscopy. Each element of interest has to be analyzed individually in AAS, and it is usually time consuming. ICP atomic emission spectroscopy allows multielement analysis at the same time. It has been reported that ICP-MS offers better detection limits than the above two techniques.2 Unfortunately, none of these techniques has the capability to analyze metal and protein at the same time in a complex in ita native form. Separate protein quantification has to be carried out in order to establish the net molar ratio of metals to protein, which also increases the actual amount of protein sample needed for complete analysis. Several years ago the analysis of metals in a preformed dimer of a zinc-finger peptide was made using plasma desorption mass spectrometry.3 While the present study was underway, a series of papers appeared reporting Present address: Enzyme Technology, 710 West Main St., Durham, NC 27701. (1)Nurnberg, H. W.In Analytical Techniques for Heavy Metals in Biological Fluids; Facchetti, S., Ed.; Elsevier: New York, 1983;pp 209232. ~.~ (2)Mason, A. 2.;Storms, S. D.; Jenkins, K. D. Anal. Biochem. 1990, 186, 187-201. (3)Frankel, A. D.;Chen, L.; Cotter, R. J.; Pabo, C. 0.Proc. Natl. Acad. Sci. U.S.A. 1988,85,6297-6300. +
0003-2700/93/0365-1355$04.00/0
laser desorption of metal-peptide complexesand electrospray ionization mass spectrometric characterization of metalpeptide complexes formed in situ.4-9 Metallothionein (MT) is a unique metalloprotein which still awaita functional characterization after 35 years of investigation.lOJ1The mammalian form of metallothionein (MW 6000-7000) has -61 amino acid residues, of which 20 are cysteines. It naturally binds up to 7 divalent transition d10 metal ions, most often Cd2+,Zn2+,and Cu2+. There is no disulfide bond in the native form of metallothionein and all 20 Cys are considered to be coordinated to the 7 metal ions through metal-thiolate bonds. Structural studies by NMR12J3 and X-ray crystallogra~hyl~ have established a two-cluster model for the three-dimensional structure of fully coordinated metallothionein. The amino terminal half of the protein (cluster B)carries 3 metal ions and the carboxyl terminal half of the protein (cluster A) carries 4 metal ions. Each metal ion is coordinated to 4 Cys residues in a tetrahedral configuration, and 8 Cys are ‘bridging” sites in which each is bound to 2 metal ions. Removal of metals from and collapse of the three-dimensional structure of metallothionein are effected by exposure to chelating agents or by acidification.l”l1 Zn ions are reported to be lost completely at pH 3 while Cd ions are lost at pH 2. Ongoing biochemical studies of the function(s) of metallothionein require that the protein be maintained in ita native state, fully metalated and folded. NMR and other studies of structure and function require reconstitution of the complex with selected populations of metal ions. In support of such studies, mass spectrometric methods for analysis of the number and kinds of metal ions in native or reconstituted metalloproteins would appear to offer real advantages in speed, complexity, and sample requirements. However, it is imperative that these methods accommodate metalloprotein samples without denaturation and demetalation. In the present work we combine the established facts about met(4)Allen, M. H.; Hutchen, T. W. Rapid Commun. Mass Spectrom. 1992,6,308-312. (5)Hutchens, T. W.; Nelson, R. W.; Allen, M. H.; Li, C. M.; Yip, T.-T. Biol. Mass Spectrom. 1992,21,151-159. (6) Hutchens, T. W.; Nelson, R. W.; Yip, T.-T.J.Mol. Recognit. 1991, 4, 151-153. (7)Hutchens, T. W.; Nelson, R. W.; Yip, T.-T.FEBS Lett. 1992,296, 99-102. (8)Nelson, R. W.; Hutchens, T. W. Rapid Commun. Mass Spectrom. 1992,6,4-8. (9) Hutchena,T. W.; Nelson, R. W.; Li, C. M.; Yip,T.-T, J.Chromatog. 1992,604,125-132. (10)Kagi, J. H. R.; Schaffer, A. Biochemistry 1988,27, 8509-8515. (11)Methods in Enzvmoloav: Riodan. J. F.. Vallee. B. L.. Eds.: Acadamic Press: New Ybrk, l%l; Vol. 205. (12)Otvos, J. D.;Armitage, I. M. Proc. Natl. Acad. Sci. U S A . 1980, 77,7094-7098. (13)Messerle, B.A.; Schaffer, A.; Vasak, M.; Kagi, J. H. R.; Wuthrich, K. J. Mol. Biol. 1990,214,765-779. (14)Robbins,A.H.;McRee,D.E.;Williamson,M.;Collett,S.A.;Xuong, N. H.; Furey, W. F.; Wong, B. C.; Stout, C. D. J. Mol. Biol. 1991,221, 1269-1293. (15)Kagi, J. H.R.; Vallee, B. L. J.Biol. Chem. 1961,236,2435-2442. (16)Rupp, H.; Weser, U. Biochim. Biophys. Acta 1978,533,209-226. (17)Li, T.-Y.; Kraker, A. J.; Shaw, C. F., III; Petering, D. H. h o c . Natl. Acad. Sci. U.S.A. 1980,77,6334-6338. ’
0 1993 Amerlcan Chemical Society
1356
ANALYTICAL CHEMISTRY, VOL. 05, NO. 10, MAY 15, 1993
allothionein with recent advances in understanding the role of pH,lS21 solvent wrnpOBition,l~zebwt.,U* and, other conditions2E@ in electrospray ionhatioil to demonstate that electrospray maas spectrometry can be used to determine rapidly and accurately how many and what cations are incorporated per molecule of protein. Information about molecular distributions and good estimates of molar ratios of various complexes in the sample can be acquired in a single measurement, instead of simply the average molar ratio of metal to protein.
6+ 1021.6
I
5f 1226.0
EXPERIMENTAL SECTION Materials. Rabbit liver metallothionein 2 (MT-2) was purchased from Sigma (St. Lo&, MO). T r i k o a o e t i c acid (TFA) and tris(hydmxymethy1)minminomethane (nb)were also purchased from Sigma. CdC12 yaa purchased from Fisher Scientifii(Fair Lawn, NJ)and Zn$O4.7HzOwas from J.T. Baker from (Phillipsburg,NJ). HPLC-grade acetonitde was purJ. T. Baker, and water was deionized to 18 Ma by Ionpure (Landover, MD). Protein Purification. To obtain purified metal-free apoMT2a, the protein was demetalatd f i t in 0.1 % TFA&O, followed by gel fiitration sing a Sephadex G-25M column (Pharmacia PD-10, Uppsala, Sweden). After lyophilization, the protein was further pMied on II reversed-phase HPLC column (Aquapore RP-300,4.6 X 250 mm, Applied Bioeypzem, San Jose, CA), using 0.1% TFA in water (solvent A) and acetonitrile (solvent E). Gradient csnditioPewere gs follawa: solvent B from 10%to 19% in 3minand then from19% to 25% in 13min, usingtwo Shimadzu (Kyoto, Japan) SPD-6A pumps. Flow rate was 1mL/min, and the protein was monitored at 215 nm by W detector. The homogeneity of the purified apoMT-2a was later checked by electrospray ionization mass spektrometry. ProteinReconrtitution. The metal-free apoMT-2awas then reconstituted in 0.5 MTTis/HClbuffer at pH 8.6 in an Ar-purged glovebag, following procedures published by Vasak and coworkers.2' Briefly, w 500 pg of apoMT-2a was dissolved in 1 mL of 0.1 M HCl, with corresponding amounta of CdCL or ZnSO, so that the f i i molar ratio of metal to protein is 7.61. The pH of the mixture was then raised stepwise to 8.6 by addig 0.5 M Tris base, and the reconstitution solution was kept at room temperature for -1 h before the metalloprotein was isolated and concentrated from the reaction mixture by ultrafiltration, using aYM membrane (Centricon-3centriconcentrator,Amicon, Beverly, MA). Uncoordinated metal ions are removed in this process. The final concentration of the reconstituted MT-2a was -2 pglpL, in a volume of 250 pL. Partially reconstituted MT-2a (C&-MT-2a) was acquired from Cd,-MT-2a by adjusting the pH to 3.5 and isolating the complex from a Sephadex G-25M column (Pharmacia PD-10) in pH 3.5 acetic acid. Atomic Absorption. Metal content was determined by graphitefurnaceatomicabsorption spectrophotometry (GFAAS), using a Varian (Mulgrave,Australia) AA-1475spectrophotometer end a Varian GTA-95 graphite tube atomizer. Protein concentration was quantified in 0.1 M HCI at 220 nm, using the UV (18) Chowdhury,S.K.;Katta,V.;Chait,B.T. J.Am.Chem. Soc. 1990, 112,9012-9013. (19) Loo, J. A.; Loo, R. R. 0.;Udseth, H.R.; Edmonds, C. G.; Smith, R. D. Rapid Commun. Mass Spectrom. 1991,5,101-105. (20)Loo, J. A.; Loo, R. R.0.;Light,K. J.; Edmonds, C. G.; Smith, R. D. Anal. Chem. 1992,64,81-88. (21) Kelly, M. A.; Vestling, M. M.; Fenselau, C. C.; Smith, P. B. Org. Maas Spectrom. 1992,27,1143-1147. (22) Loo, J. A.; Udseth, H. R.; Smith,R. D.Biomed. Enuiron. Mas8 Spectrom. 1988,17,411-414. (23) Le Blanc, J. C. Y.; Beuchemin, D.; Siu, K. W. M.; Guevremont, R.; Berman, S. S. Org. Mass Spectrom. lai,26,831-839. (24) Allen, M. H.; Vestal, M. L. J . Am. he. Maua Spectrom. 1992,3, 1926. _ _ _. (25) Smith, R. D.;Loo, J. A.; Edmonds,C. G.; Barinaga, C. J.; Udeeth, H.R. AM^. Chem. 1990,62,882-889. (26) Loo, J. A.; Edmonds, C. G.; Udseth, H.R.; Smith, R. D. Anal. Chem. 1990,62,693-698. (27) Vasak, M. In Methods in Enzymology; Rordan, J. F., Vnllee, B. L.,Ede.; Acadamic Press: New York, 1991; Vol. 205, pp 452-458.
..
I., Iz
1533.6
Flguro 1. ESI ma80 spectrum of apometalkthknehr (rabbit her MT2a and motal-free fmm). (Mditbm am dewitmd in the
a',
ExpeJrimentelsectlon. absorption of the apoprotein (c2m = 48 200 M-l em-').= ElectrosprayIonization Maw Spectrometry. Electmpray
~ on a Vestec ionization (EN)mass spectrometry w a performed (Houston, TX)electrospray source fitted to a Hewlett-Packard (PaloAlto,CA) 5988A quadrupole mass spectrometer, a P h r w r (Duarte,CA) high-energy dynode detector, and a Technivent (St. Lo@, MO) data system. Samples were introduceid by a Harvard [South Natick, MA) syringe pump at a flow rate of 3 pL/min. The stainleas steel needle was held at -2.2 kV,and the source temperature was at 240 "C. For the analysis of apoMT2a samples, lyophilized MT was dissolved in diluted acetic acid at pH 4 with final cancentration at 10 p M . Reconstituted MT sample wae mixed with the diluted acetic acid solution at pH 4 (30pL of MT + 70 6 of acetic acid) m d then electrosprayed. Ammonia/water at pH 10 was also used for the analysis of the reconstituted MT samples. The molecular weighta reported and calculated in this study are averages.
-
RESULTS AND DISCUSSION Analysis of Metallothionein by Electrospray Ionization Mass Spectrometry (ESMS). When the purified apoMT-2a sample was dissolved in diluted acetic acid at pH 4 with a final concentration of -10 NM,excellent electrospray signale could be generated. A typical spectrumis shown in Figure 1. The molecular weight of the apsMT-2a observed from ESMS is 6126.1 f 0.9. The average molecular weight calculated from the published sequence is 6126.2. A second isoform of MT-2a was also present in the sample, with an observed molecular mass of 30 Da higher than that of the main imform. This is consistent with the A/T heterogeneity reported by Hunziker.28 Sequences for both isoforms have been determined by H u i k e r and they are shown here in Figure.2, along with the schematically shown diagram of the metal-protein clusters determined by NMR structure studies.m The minor peaks observed to the left of the intact protein peaks are fragment ions resulting from loes of N-acetylated Met and Asp residues from the amino terminal of the protein. The fragmentation is attributed to the electrospray process, since the intensities of these peaks vary with their charge states (they are more intense at higher charge states). It was observed that formation of these ions from the apoprotein was enhanced at higher repeller voltages (data not shown). This fragmentation waa greatly reduced in spectra of reconstituted metallothioneins. (28)BuRler, R. H. 0.; Kagi, J. H.R. Erperientia Suppl. 1979, 34, 211-220. (29)Hunziker,P.E.InMethodsin Enzymology;Riordm,&F.,Vallee, B. L.,Ede.; Acadamic Press: New York, 1991; Vol. 205, pp 421-426. (30) Braun, W.; Wagner, G.; Worgotter, E.;V d , M.;Kagi, J. H.R.; Wuthrich, K. J. Mol. Biol. 1!86,187, 125-129.
ANALYTICAL CHEMISTRY, VOL. 65, NO. 10, MAY 15, 1993
IV
I11
I1
1957
4+
1643.8
MT-2a : Ac-MDPN MT-2a':
T
v
VTI
I
VI
MT-2a: Flgwe 2. Amlno acM sequences of rabbit liver metallothloneln 2a and 2a'. Also shown Is the metal-cystelne bondlng as Indicated by lines between each metal ion (I-VII) and cysteine (C) (adapted from refs 29 and 30).
q Q
604
4+ 1725.6
I
.I*
Flgure 4. ESI mass spectrum of fully reconstltuted metallothionein (Zn7-MT-2a). Conditions are descrlbed In the Experimental Section.
Table 1. Metal Content of Reconstituted Metallothionein
protein
Flguro 3. ESI mass spectrum of fully
reconsthted metallothionein (Cd,-MT-Pa). Condkbns are described in the Experlmental Section.
Metallothionein complexes were reconstituted between purified apoMT-2a and Cd2+or Zn2+ions and separated from the small excess of metal ions before analysis by mass spectrometry. The stability of metallothionein complexes is known to be dependent on pH, with complete demetelation of Zn2+ occuring below pH 3 and of Cd2+ below pH 2.159'6 In order to retain the metal ions and the integrity of the reconstituted complex during the electrospray process, special attention was given to the pH of the solution. The spectrum shown in Figure 3 was obtained from a solution of the reconstituted Cd complex in aqueous acetic acid at pH 4. The peak at mlz 1725.6 corresponds to a protein species of C d r MT-2a carrying four extra protons at charge state 4+. An average molecular weight of 6897.6f0.9 was obtained with three determinations over the period of 2 weeks, which compares well with the expected molecular weight for a neutral species of 6898.0 (apoMT + 7Cd - 14H). We conclude that the metalloprotein was completely reconstituted by our procedure. By analogy with conclusions based on X-ray crystallographic and NMR studies that MT is folded when it is fully metalated and that it must be metalated to be folded, we suggest that this metalloprotein retains conformation in the gas phase that is simiiar to that in solution. Fragmentation in the electrosprayprocess was not observed from these folded metallothioneins. Similar results were also obtained with the Zn.rMT-2a sample,where the only charge state observed was 4+, as shown in Figure 4. The observed average molecular weight was 6571.2 f 2.0 and the calculated one is 6568.8 (apoMT + 7Zn - 14H). Aliquota of the same Cd.rMT-2a and Zn-MT-2a samples were also analyzed by graphite furnace atomic absorption spectrophotometryin combination with protein analysis. The results were consistent, as shown in Table I. In another experiment, reconstituted CdrMT-2a was exposed to pH 3.5 acetic acid during gel filtration. The mass spectrum obtained is shown in Figure 5. The most abundant peak at mlz 1314.3 was assigned to a Cdd-MT-2a species in
metal expected
CdrMT
Cd
7
CdrMT Zn7-MT
Zn Zn
0 7
ZnrMT Cdd-MT
Cd Cd
0
Cdd-MT
Zn
0
2
metal contenta atomic absb ESIc 7.0 f 0.2 7.4 f 1.2 (75% Cd7-MT, 16% Cde-MT, 10% Cdg-MT) 0.Of 0.2 0 7.1 f 0.2 7.6 f 1.2 (62% Zn,-MT, 24% Zna-MT, 15% Zng-MT) 0.Of 0.2 0 3.7 f 0.3 3.8 f 0.8 (8% Cdo-MT, 83% Cdd-MT, 6% CdB-MT, 3% Cds-MT) nad 0
4
Metal content is expressed as molar ratio (moles of metal per mole of protein). Molar ratio was an average of two determinations. Uncertainties are calculated from five measurements in each determination. Average molar ratios were obtained from the sum of each species with relative abundance calculated from peak heights. Uncertainties are estimated to be - 5 % . Not analyzed.
*
5 + , 4 Cd 1314.3
6+, 4 Cd 1095.0 I
et4 1000
1050
1100
1150
ll00
1150
1100
1110
1400
1410
.I*
F b r e 5. ESI mass spectrum of partially reconstituted metallothionein (CdrMT-2a). Conditions are described In the Experimental Section. F1 and Fo are fragments with one and two residues lost from the N-termlnal of the protein, respectively.
the sample, with a charge of 5+. On the basis of the NMRdetermined metal-MT cluster structure (schematicallyshown in Figure 2),the retention of 4 Cd2+in cluster A should interact with the free SH groups of 11Cya. The observed molecular weight was 6566.5 f 1.0, which compares well with the calculated average molecular weight for neutral C&-MT-aa: apoMT + 4Cd - 8H = 6566.8. It has been observed that the
1358
ANALYTICAL CHEMISTRY, VOL. 65, NO. 10, MAY 15, 1993
reconstitution of metallothionein from apoMT and free metal ions proceeded preferentially via the initial formation of cluster A, and when a metal-chelating reagent was added to the protein solution, cluster B decomposed first.31 We suggest that the protein behaves in a similar way when it is exposed to acid so that, at pH 3.5, the 3 Cd2+in cluster B were lost during the gel filtration process and the remaining 4 Cd2+ were all in cluster A. Besides the most abundant peak of Cd4-MT-2a at miz 1314.3 with its charge state at 5+, peaks indicating the presence of complexes with0 CdL+,5 Cd2+,and 6 Cd2+bonded to MT-2a can be observed, as shown in Figure 5. No intermediate-state molecular complexes containing 1,2, and/ or 3 CdL' were detected. This result is consistent with other metallothionein studies. Earlier reports indicate that metal ions are bound to the carboxyl terminal half of the apoMT in a cooperative way to form cluster A, and no intermediate forms of MT-metal complexes have been reported between the metal-free apoMT and the Me4-MT." On the other hand, the less stable cluster B is reported to be more flexible and cluster formation less c o o p e r a t i ~ e , ~consistent *,~~ with our observations of Cds-MT-2a and Cd6-MT-2a complexes. It is likely that, although the three-metal cluster (cluster B) fell apart when the Cdy-MT-2a was treated with pH 3.5 acetic acid, there still was some Cd2+tetrahedrally bound to the amino-terminal domain of the protein, which showed up in small abundances as Cdj-MT-2a and Cde-MT-2a. To further confirm the results from electrospray ionization mass spectrometry, an aliquot of the same sample was also analyzed by graphite furnace atomic absorption spectrophotometry. The results are listed in Table I. The average molar ratio of Cd to apoMT-2a was determined by GFAAS to be 3.7 f 0.3. Intensities of all peaks in all charge states assigned to metalated and apoMT (including fragmentation peaks) were summed to estimate the molar ratios of various species by ESMS. Molar ratios were found to be 83% , 8% , 6 % , and 3 5% , for species with 4,0,5, and 6 Cd2+bonded, respectively. An average molar ratio of 3.8 f 0.8 Cd per MT molecule was then calculated for the sample, which compares well with the GFAAS result. Charge States in Electrosprayed Metallothionein Ions. Examination of Figures 1 and 3-5 reveals that, in the case of metallothionein, the charge-state distribution shifts toward lower values as each domain becomes incorporated with metals. As metallothionein undergoes cluster formation from metal-free apoMT (most relaxed form) to a partially reconstituted (cluster A formation) and then to the fully reconstituted (both clusters A and B) form, the most abundant charge state observed in ESMS decreases from 6+ to 5+ and finally to 4+. This is consistent with a number of studies that indicate that protein conformation in solution influences the overall charge state observed in electrospray mass spectra. 16-21 One paper in a recent series by Allen and Hutchens4reports no change in the overall charge envelope during the process of peptide-Zn complexation. One explanation given by the authors was that 2H+ + 2" instead of 4 H+ were replaced by each Zn2+during the peptide-metal interaction, so that there would be no net change with regard to the protein charge state in solution. For the case of metallothionein, the coordination of 7 Cd?+or 7 Zn2+ions into the protein could result in a net increase of 6 negative charges on the protein (7 M2+replaces 20 H+).33The fact that a net decrease of 2 (31) Nielson, K. R.; Winge, D. R. J . Biol. Chem. 1983, 258, 1306313069. (32) Bernhard. W.R.; Vasak, M.; Kaai, J. H. R. Biochemistrv 1986.25. 1975-1980. (33) Kojima, Y.; Berger, C.; Vallee, B. L.; Kagi, J. H. R. Proc. N d i . Acad. SCL.V.S.A, 1976, 73, 3413-3417.
1CCI-
1713 8 I
1725 6
P \
n i
Figure 8. ESI mass spectrum of a commercial rabbit liver metallothionein 2 (MT-2). Conditions are described in the Experimental Section.
in the charge state (now at 4+) is observed in positive ion electrospray indicates that it is not the total number of chargeable sites on the protein that determines the overall charge distribution in ESMS, but rather the accessable sites as determined by both the primary sequence and the tertiary structure. Furthermore, when ammonialwater (pH 10) was used as the electrospray solvent, essentially the same results were obtained as with pH 4 acetic acid; i.e., Cd;-MT-2a was observed as a peak at mlz 1725.5 at charge state 4+ (data not shown). This clearly indicates, as discussed in more detail by Kelly et a1.,*1 that charge-state distribution of an electrospray mass spectrum is not determined simply by the net charge predicted for the protein at certain pH conditions. Comparison of ESMS with Graphite Furnace Atomic Absorption Spectrophotometry. The traditional method for trace metal analysis, namely, GFAAS used in conjunction with protein analysis, provides the average molar ratio of metal to protein. In this report, we demonstrate that ESMS can provide more information about metalloprotein composition in a single rapid analysis. As shown in Figure 5 and Table I, a variety of different metal-protein complexes can be detected in a single sample and distinguished by their different molecular weights. Detection of the coexistence of metal-free MT, Cds-MT, and C&-MT with Cd4-MT provides important information about the population of protein-metal complexes, in addition to the average molar ratio of 3.7. In this case, the good agreement between GFAAS and ESMS of the overall molar ratio supports the idea that the relative abundances of the various complexes in the samples can be estimated from the abundances of respective ions in the ESI spectrum. To test the method further, a sample of MT-2 was analyzed directly as it was received from the commercial supplier. Again, as shown in Figure 6 and Table 11, more information was obtained from ESMS measurements. It is clear that the commercial sample contains at least three different complexes. One has 7 Cd2+ (indicated by the peak at mlz 1725.6), one has 6 Cd2+and 1Zn2+(indicated by the peak at mlz 1713.8),and one has 5 Cd2+and 2 Zn2+ (peak at mlz 1702.2). In addition, with an estimated deviation of 5 % , these three peaks have relative abundances of 36 % ,39% ,and 25 % ,respectively. As can be seen in Figure 6, there are also some other peaks in the ESI spectrum, which indicate the presence of other protein molecules (molecular weights not shown).
ANALYTICAL CHEMISTRY, VOL. 65, NO. 10, MAY 15, 1993
Table 11. Metal Content of a Commercial Metallothioneins labelb atomic abs ESI Cd Zn
4.2 0.9
6.4 f 0 . 2 1.9 f 0.2
6.1 0.9
* 0.gC 0.2
molar distrib (re1 abundance, % ) 7 Cd, 0 Zn (36) 6 Cd, 1 Zn (39) 5 Cd, 2 Zn (25) Rabbit liver metallothionein 2 (MT-2) was from Sigma (Lot No. 90H9550). Molar ratios were determined by inductively coupled plasma atomic emission. Uncertainties are estimated from peak height measurements.
*
Table 111. Comparison of Atomic Absorption and Electrospray Mass Spectrometry as Methods for Determination of Molar Ratios of Metals in Protein Comulexes
anal. sensitivity (nmol) anal. time (min) result
atomic abs and protein quantitation 4 60 av molar ratio
ES 1 5 molar distrib of various complexes
A comparison of the two methods is listed in Table 111, with regard to sensitivity, minimal time needed to carry out a complete analysis, and results provided by each technique. In the case of metallothionein analysis, it was found that, with essentiallythe same requirements for amount of sample, ESMS is able to provide more information about the protein more rapidly. While only picomoles of metal-freeMT sample
1958
was needed for ESMS analysis, lower sensitivitywas observed for MT samples containing metal. It is probable that the sample requirements reported here will be improved for metalloprotein analyses in the future. CONCLUSIONS
We have shown that metallothionein complexes in which 7 divalent metals are coordinated by 20 sulfhydryl groups can be analyzed by electrospray mass spectrometry if care is taken to use pH values that maintain the native cooperative binding. The ability of electrospray to generate ions from solutions across a broad pH range makes it compatible with this approach, and observation of intact complexes is consistent with the hypothesis that metallothionein retains its conformation in the gas phase. This mass spectrometric technique can determine not only how many and what cations are bound per protein molecule but also the molecular distributions and, in some cases, the relative abundance8 of various complexes in the protein sample. It is faster and more sensitive than atomic absorption spectrophotometry coupled with protein analysis. ACKNOWLEDGMENT
We thank Michele Kelly for instruction in the use of the electrospray ionization mass spectrometer. We also thank Dr. Martha Vestling and Dr. Takashi Ishida for many helpful discussions. The work was supported by a grant from the U.S.Public Health Service National Institutes of Health, GM-21248.
RECEIVED for review July 15, 1992. Accepted January 11, 1993.
