mrgshhhhhhgmasa k16e t


FIG. 1. Recombinant histidine-tagged ER-DBD of 99 amino acids and two synthetic peptides [K16-Q50 (Pepl) and H52-K88 (Pep2)] prepared for these studies. Amino acid residues at the trypsin cleavage sites that give peptides that identify oxidation of the individual fingers are numbered.

g [h52 nd_rkcyevgmmk88 [ggirkdrrgg


FIG. 1. Recombinant histidine-tagged ER-DBD of 99 amino acids and two synthetic peptides [K16-Q50 (Pepl) and H52-K88 (Pep2)] prepared for these studies. Amino acid residues at the trypsin cleavage sites that give peptides that identify oxidation of the individual fingers are numbered.

carboxymethylate the thiol groups, each of which should increase the molecular mass by 58 Da. A mixture of products was obtained, the most abundant of which corresponded to the alkylation of eight and nine cysteines. A minor degree of over-reaction gave other, weaker peaks for the addition of 10 and 11 carboxymethyl groups. As anticipated, ER-DBD that had been oxidized with hydrogen peroxide and then treated with IAA proved to be resistant to carboxymethylation. Thus it was confirmed that ER-DBD reduced with DTT had the anticipated number of free thiols.

Oxidation of Estrogen Receptor DNA-Binding Domain

Initial experiments on the oxidation of recombinant ER-DBD were carried out with either 5 mM hydrogen peroxide or 5 or 25 mM diamide for periods extending from 25 min to overnight. Diamide [(CH3)2NCON = NCON(CH3)2] is a reagent that specifically oxidizes cysteine to cystine.36 The extent and nature of the oxidation were evaluated by HPLC-MS of the intact protein without digestion to separate

36 N. S. Kosower, E. M. Kosower, andB. Wertheim, Biochem. Biophys. Res. Commun. 37,593 (1969).

peptides. ER-DBD contains nine cysteines and five methionines. Treatment of ER-DBD with hydrogen peroxide induced two different types of oxidation: (1) the loss of one or more pairs of hydrogen atoms as disulfide bonds were formed, and (2) the addition of multiple oxygen atoms, almost certainly to form methionine sulfoxide. Peroxide treatment at 5 mM for 25 min removed only four hydrogens and added zero, one, or two oxygens, whereas overnight treatment removed six hydrogens and added one to five oxygens. After oxidation for 25 min, further treatment with 50 mM DTT reduced the disulfide bonds but did not reduce the methionine sulfoxides. Treatment with 5 mM diamide for 25 min partially oxidized ER-DBD by the removal of an average of four hydrogen atoms, whereas 25 mM diamide removed eight hydrogens but also gave rise to a new species corresponding to the addition of a diamide molecule. Treatment with DTT reduced the cystines back to cysteine and removed the diamide adduct. Unlike peroxide, diamide induced little methionine sulfoxide formation.

Subsequent titrations with oxidizing agents were followed by digestion with trypsin so that differential oxidation of different regions of ER could be monitored. Initially the resulting peptides were found to be prone to oxidation during digestion and chromatography, probably because of dissolved air, even when the digestion time was limited to 3 min. To minimize unwanted oxidation, dissolved air was removed from all solutions by vacuum filtration immediately before their use and all solutions were maintained under a positive pressure of helium. Oxidation was carried out by incubating 10 /¿M reduced ER-DBD (previously stored in 100 \xM DTT) in 11.5 mM ammonium acetate, pH 7.4, and either 60 fiM zinc(II) sulfate or 100 fiM Na2EDTA with various concentrations of either hydrogen peroxide or diamide. The oxidation was allowed to proceed for 15 min before analysis by rapid tryptic digestion. To achieve faster enzymatic digestion of ER-DBD protein than would be obtained in solution, the protein was injected onto a Porozyme packed immobilized trypsin column (PE Biosystems, Framingham, MA) at a flow rate of 50 fj, 1 min"1. The digestion buffer was 70 mM ammonium bicarbonate-5% (v/v) acetonitrile, pH 8.0, from which dissolved air had been removed by vacuum filtration followed by helium purging and continuous exposure to helium at 3 lb in"2. The pH was checked and adjusted with acetic acid or ammonium hydroxide before each analysis. The enzyme column was connected in series via a Rheodyne (Rohnert Park, CA) switching valve to a Vydac (Hesperia, CA) 150 x 1 mm C|g microbore column. After 3 min the valve was switched to allow analysis of the resulting peptides by HPLC-MS at a flow rate of 50 ¡i\ min"1. The enzyme column was rinsed after each analysis with at least 10 ml of digest buffer. For this HPLC analysis, solvent A was 0.1% (v/v) formic acid and solvent B was 5 :2 (v/v) ethanol-n-propanol with 0.05% (v/v) formic acid; both solvents were vacuum filtered and purged with helium and then continuously exposed to helium at 3 lb in"2. After holding the HPLC at 5% (v/v) solvent B for 2 min a linear gradient was run from 5 to 60% solvent B in 30 min. The eluted peptides were passed through the

UV flow cell with detection at 215 nm and the flow was split to transfer ~10% to the ESI source.

Calculation of Degree of Oxidation

Oxidation of cysteine residues in a zinc finger results in the formation of disulfide bonds with the loss of two protons for each disulfide formed. For cysteine-containing peptides in a tryptic digest, the degree of peptide oxidation can be calculated by determining the amount of oxidized peptide present (Mr lower than that of the reduced peptide by 2n, where n = the number of disulfides), compared with the amount of reduced peptide present, based on the heights of the peaks in the mass spectrum. If the peptides are not separated chromatographically, the isotopic patterns for the reduced and oxidized forms will overlap. For a peptide containing two cysteines the peak height of the reduced form can be corrected by subtracting the isotopic contribution of the third isotopic peak of the oxidized form. Because ER-DBD has no trypsin cleavage sites in zinc finger 1, the entire four-cysteine structure was observed intact, containing none, one, or two disulfide bonds. Therefore, calculation of the amount of singly oxidized peptide (one disulfide bond) was corrected by subtracting the contribution of the third isotopic peak of the doubly oxidized peptide (two disulfide bonds); the peak height of the reduced peptide was corrected for the contribution of the fifth isotope of the doubly oxidized peptide and the third isotope of the singly oxidized peptide. As a result of incomplete digestion, zinc finger 1 was actually represented by two overlapping peptides corresponding to amino acids 17-47 and 20-42. The results from these two peptides were averaged to give the percent oxidation of zinc finger 1. Baseline levels of oxidation observed in the absence of any added oxidant were subtracted from the data.

The target peptides were identified from an analysis of the multiply charged ions in the HPLC-MS spectra, from which the relative abundances of the reduced and oxidized forms were calculated. For zinc finger 2 of ER-DBD there were four possible trypsin cleavage sites (Fig. 1). Again the digestion was incomplete, giving rise to additional peptides. The data from two peptides spanning amino acids 48-67 and 48-69 were averaged to give the percent oxidation of the first half (ZF2ci_2)-Although two peptides were also observed representing the second half of zinc finger 2, that is, amino acids 71-77 and 72-77, the peak for amino acids 71-77 was usually weak and in a more noisy region of the spectrum. Thus, only the amino acid 72-77 peptide fragment was used in the calculation of the percent oxidation of the second half of finger 2 (ZF2c3_4). Taking account of the isotope corrections, the amount of oxidation occurring under different conditions was calculated for different concentrations of oxidant. This revealed only low levels of oxidation of finger 1 under conditions that caused extensive oxidation of finger 2; furthermore, the oxidation of finger 2 was asymmetric, with a much higher level of disulfide formation between cysteines 3 and 4 (Cys3 and Cys4) than between cysteines 1 and 2 (Cysl and Cys2). For example, 2 mM peroxide was sufficient to oxidize 100% of ZF2Cys3-4, whereas 7.5 mM was not sufficient to fully oxidize any of the other pairs of cysteines. Similar trends were observed with 0-500 ¡xM diamide in the presence of zinc ions. Interestingly, the removal of zinc by EDTA increased the sensitivity of finger 1 to oxidation such that both fingers were oxidized to the same degree. The observation that finger 1 is more resistant to oxidation than finger 2 only when zinc ions are present and not in the absence of zinc is consistent with structural studies by nuclear magnetic resonance (NMR), showing that zinc in finger 2 is less well coordinated than in finger 1,37

If the tetrahedral arrangement of the cysteine residues was maintained during oxidation, within a single finger there would be equal cross-linking of peptides by disulfide formation between Cysl-Cys2/Cys3-Cys4, Cysl-Cys3/Cys2-Cys4, and Cysl-Cys4/Cys2-Cys3. Statistically, a tryptic digest of finger 2 should reveal twice the abundance of the cross-linked peptides, such as peptides 48-67/72-77 compared with the separate oxidized peptides 48-67 and 72-77. However, the signal for the cross-linked peptide was of much lower intensity (~10%) than the separate peptides, suggesting that oxidation disrupts the geometry and expels the Zn2+ ion from this finger before the disulfide bonds are formed. It was not possible to make a direct comparison with the behavior of finger 1 as this was not cleaved by trypsin, but HPLC separation of the fully oxidized species gave three peaks of approximately equal abundance, which are likely attributable to the three different disulfide-bonded isomers. This would indicate that the finger structure was maintained more strongly during oxidation of finger 1 than finger 2.

Methionine Oxidation: Absence of Protection Against Thiol Oxidation

It has been proposed that methionine residues at strategic sites in proteins might act as scavengers for ROS and thereby protect cysteine thiols from oxidation.38 Oxidation of methionine to methionine sulfoxide was monitored in the same experiment as a function of oxidant concentration by observing the formation of peaks corresponding to the addition of one oxygen atom, that is, 16 Da. Three methionine residues are located in or adjacent to zinc finger 2 as follows: Met-56, immediately adjacent to the first cysteine of finger 2, and Met-86 and Met-87 situated just after the C-terminal helix of this finger. A peptide spanning Ser-48 to

37 A. Wilkstrom, H. Berglund, C. Hambraeus, S. van den Berg, and T. Hard, J. Mol. Biol. 290, 96 (1999).

38 R. L. Levine, L. Mosoni, B. S. Berlett, and E. R. Stadtman, Proc. Natl. Acad. Sci. U.S.A. 93, 15036 (1996).

Lys-67 revealed a background level of ~3-4% oxidation for Met-56, which did not increase at all on exposure to up to 7.5 mM hydrogen peroxide or 500 mM diamide, even though these amounts were sufficient to completely disrupt and oxidize finger 2. Furthermore, the other methionine residues (Met-86 and Met-87), situated just outside the C terminus of the finger 2 helix, showed limited oxidation totaling less than 10% above baseline on exposure to up to 7.5 mM peroxide and no significant oxidation on exposure to diamide. Thus, contrary to the suggestion that methionine might protect against oxidation, the zinc finger cysteines are more susceptible to oxidation than any of the three methionine residues contained in the ER-DBD.

Stoichiometiy of Zinc Binding to Estrogen Receptor DNA-Binding Domain Monitored by Electrospray Ionization-Mass Spectrometry

Zinc binding by recombinant ER-DBD was monitored by ESI-MS in a weak buffer at pH 7.4, conditions designed to allow observation of intact noncovalent complexes and reveal the stoichiometry of any complex formed between the protein and zinc ions. The strongest peak representing approximately two-thirds of total protein molecules was free of zinc, whereas the remaining third added a single Zn2+ cation. There was no observable addition of two Zn2+ ions. No zinc had been added at any stage in the expression or purification, and therefore it was concluded that the protein had scavenged the zinc from the cells in which it was expressed. The presence of less than the stoichiometric amount of zinc was consistent with a circular dichroic (CD) spectrum showing a lower than normal degree of a- helicity compared with fully zinc-loaded ER-DBD, which shifted to higher helix content on addition of zinc. The addition to the buffer of 2 mol of Zn2+ per mole of reduced ER-DBD showed strong formation of the complex ER-DBD ■ Zn2+ and a lesser amount of ER-DBD • 2Zn2+. The CD spectrum confirmed that this amount of Zn2+ was sufficient to restore the a-helicity to its normal level. Exposure of ER-DBD to 5 mol of Zn2+ per mole of protein caused the ER-DBD ■ 2Zn2+ peak to become the most intense in the mass spectrum. Weak binding of a third Zn2+ ion was also seen to occur, possibly by binding to the histidine tag, or simply as a nonspecific effect of ESI-MS. In contrast, ER-DBD oxidized with either peroxide or diamide could bind only low levels of Zn2+ in a nonspecific manner, again perhaps involving the histidine tag.

The precision of the ESI-MS mass measurements for ER-DBD • 2Zn2+ was sufficient to establish that the binding of two Zn2+ ions involved the elimination of four protons from ER-DBD, that is, on average each Zn2+ ion added only 63.5 Da to the mass, even though the average atomic mass of zinc is 65.5 Da. Four cysteine residues participate in stabilizing each of the two zinc fingers, and thus it can be concluded that each finger involves two thiolate anions forming ionic bonds and two thiols forming coordinate bonds. A similar observation by ESI-MS was also reported for a Cys^His zinc finger.39

Cooperative Action of Two Zinc Fingers in Stabilizing Structure

To probe the behavior of each zinc finger of ER-DBD in isolation, both were synthesized as separate peptides (Pepl and Pep2; Fig. 1) and probed by EMS A, ESI-MS, and CD. Although we determined by EMSA that neither peptide was able to bind to ERE, ESI-MS showed that with a peptide: Zn2+ molar ratio of 1:1, each peptide strongly bound a single Zn2+ ion. The effect of Zn2+ on the secondary structures of these peptides was monitored by CD spectroscopy and compared with the behavior of ER-DBD. We had previously established that salt at near-physiological concentration was necessary for ER-DBD to maintain its structure and therefore the same was assumed to be true of the individual peptides. Buffers containing 100 mM sodium fluoride were used for all CD experiments, rather than sodium chloride, as the chloride ion absorbs UV radiation below 200 nm. The CD spectrum of a typical random coil structure should show a minimum at ~195 nm and relatively little signal at ~222 nm, whereas a-helical proteins generally show a maximum at ~ 195 nm and two minima at 208 and 222 nm that correspond to n-n* and n-n* electronic transitions, respectively. In the absence of Zn2+ both synthetic peptides showed minima strongly blue-shifted from the «-helix value to ~198 nm, and weak signals at 222 nm. This was indicative of low a-helix content, which was confirmed by secondary structural analysis, giving only 5% a-helix for Pepl and 2% for Pep2. Furthermore, there were only weak increases in a-helical structure on the addition of one molar equivalent of Zn2+, particularly for Pep2, for which the a-helix increased to only 6% despite the relatively strong Zn2+-binding properties of both peptides demonstrated by ESI-MS. For neither peptide was the a-helical content comparable to the 23% determined for ER-DBD itself.

Further Experimental Considerations

Quantitative studies such as those described here involve a comparison between peptides that contained cysteines that are either oxidized or reduced. It was found that if peptides containing free cysteines are kept in solution for any length of time, such a procedure may be problematical as dissolved air during trypsin digestion or HPLC purification may cause additional oxidation. An alternative is to alkylate the free thiols with a reagent such as IAA, iodoacetamide, or ¿V-ethylmaleimide before any processing or manipulation, which will prevent further reaction. A number of authors have adopted a strategy of alkylating free thiols with one reagent and then reducing the oxidized cysteines and alkylating with a second agent, giving

39 D. Fabris, J. Zaia, Y, Hathout, and C. Fenselau, J. Am. Chem. Soc. 118, 12242 (1996).

two groups of differently alkylated cysteines that can be distinguished by their characteristic masses. An affinity tag may be attached to the first alkylating agent to selectively purify cysteine-containing peptides from a tryptic digest.40 However, reduction and alkylation would destroy other disulfide-bonded species, such as the .S'-nitrosylglutathione adducts that we have observed as intermediates from oxidation by S-nitrosoglutathione. Nevertheless, in studying a zinc finger protein extracted from a biological material such as a human tumor, the risk of unwanted oxidation is much greater and an initial alkylation step should be regarded as mandatory. It is likely that a successful purification of a specific target protein will involve an immunoaffinity purification; therefore it is important to establish that alkylation does not diminish the affinity of the antibody for the protein.

The products of other oxidation processes may be sought, and it should be noted that although disulfide formation is the common product of cysteine oxidation, at least five other more extensively oxidized products could potentially be formed, although not all occur in biological samples. Xu and Wilcox have described three products of oxidation of a single cysteine formed by introduction of one, two, or three oxygen atoms: cysteine sulfenic acid, sulfinic acid, and sulfonic acid (cysteic acid), respectively. Also, a disulfide can add one oxygen to give thiosulfinate or two oxygens to give thiosulfonate.41 Other oxidative processes can proceed through nitric oxide and .S'-nitrosoglutathione, and these may play a role in the regulation of gene expression.42 Consequently it may be relevant to search for corresponding cysteine derivatives in the zinc fingers of other transcription factors.


We thank Drs. J. Meza and R. M. Whittal for valuable contributions. Mass spectrometry was carried out in the UCSF Mass Spectrometry Facility, supported by NIH NCRR 01614. This work was also supported in part by NIH Grant CA71468 as well as by Hazel P. Munroe (Buck Institute) and Janet Landfear (Mt. Zion Health Systems) memorial funds.

40 T.-Y. Yen, R. K. Joshi, H. Yan, N. O. L. Seto, M. M. Palcic, and B. A Macher, J. Mass Spectrom. 35,990 (2000).

41 H. E. Marshall, K. Merchant, and J. S. Stamler, FASEB J. 14, 1889 (2000).

42 Y. Xu and D. E. Wilcox, J. Am. Chem. Soc. 120, 7375 (1998).

0 0

Post a comment