Metal-catalyzed Oxidation of Histidine in Human Growth Hormone
MECHANISM, ISOTOPE EFFECTS, AND INHIBITION BY A MILD DENATURING ALCOHOL*

(Received for publication, October 23, 1996, and in revised form, January 9, 1997)

Fang Zhao Dagger , Elena Ghezzo-Schöneich Dagger , Gaby I. Aced Dagger , Jinyang Hong Dagger , Terry Milby § and Christian Schöneich Dagger

From the Dagger  Department of Pharmaceutical Chemistry, The University of Kansas, Lawrence, Kansas 66047 and the § Pharmaceutical R&D, Genentech, Inc., South San Francisco, California 94080

ABSTRACT
INTRODUCTION
EXPERIMENTAL PROCEDURES
RESULTS
DISCUSSION
FOOTNOTES
ACKNOWLEDGEMENTS
REFERENCES


ABSTRACT

Metal-catalyzed oxidation of proteins represents an important pathway of post-translational modification. We utilized human growth hormone (hGH), a protein with a well defined metal-binding site, to study the detailed mechanism of metal-catalyzed oxidation by ascorbate/Cu(II)/O2. Particularly His18 and His21 within the metal-binding site were oxidized, predominantly to 2-oxo-His with the incorporated oxygen originating from molecular oxygen, based on amino acid analysis, tryptic mapping, mass spectrometry, isotopic labeling, and 1H NMR. The anaerobic reduction of a hGH/Cu(II) mixture by ascorbate generated a hGH-Cu(I) complex with NMR spectral features different from those of native hGH and hGH/Cu(II). The anaerobic reaction of this hGH-Cu(I) complex with hydrogen peroxide resulted in the oxidation of His18 and His21, suggesting that a fraction of Cu(I) was bound at the metal-binding site of hGH. Site-specific oxidation of hGH required an intact metal-binding site and could largely (about 80%) be inhibited by the presence of >= 28% (v/v) 1-propanol which appears (i) to perturb the metal-binding site and (ii) to interact with a reactive oxygen species formed at the perturbed metal-binding site. The inhibition by 1-propanol-d7 (CD3CD2CD2OH) was significantly lower than that by 1-propanol-h7 with [residual hGH]1-propanol-h7/[residual hGH]1-propanol-d7 = 1.95 at 30% (v/v) 1-propanol, reflecting a kinetic isotope effect close to that for the reaction of a hydroxyl radical with Calpha -H/D bonds of methanol, suggesting the involvement of a hydroxyl radical-like species in the oxidation of His.


INTRODUCTION

In physiological compartments, provided with suitable reducing agents, transition metals are able to catalyze the activation of O2 to various reactive oxygen species (ROS)1 which readily attack biomolecules. The metal-catalyzed oxidation (MCO) of proteins plays an important role during oxidative stress and aging where modified proteins may accumulate in tissue (1-9). In addition, MCO also presents a stability problem for the biotechnological manufacturing of proteins (10, 11).

A unique feature often observed during MCO of proteins is that only a few amino acid residues within a certain domain of the given protein are modified (2, 12). Such site specificity is attributed to the generation of ROS at specific metal-binding sites, where the highly reactive ROS attack labile functional groups nearby rather than diffuse into the bulk medium. It has been shown that most of the enzymes labile to MCO require metals for activity (13). The oxidation of these enzymes is less sensitive to the inhibition by exogenous ROS scavengers than would be expected on the basis of known rate constants for the reactions of specific ROS with these scavengers. However, one problem associated with all these studies is that neither the oxidizing species nor the intermediary protein-transition metal complexes formed during MCO of proteins have been well characterized. This is most pertinent to copper-catalyzed oxidation (i.e. to the frequently used ascorbate/Cu(II)/O2 system) involving the Cu(II)/Cu(I) redox couple (2). Here, Cu(II) will most probably require a different geometry and different ligands within the protein ligand sphere than Cu(I). An initial reduction of Cu(II), bound to a metal-binding site, to Cu(I) may change the geometry of the protein-metal complex, and possibly even release Cu(I) from the metal-binding site. On the other hand, the metal may be retained when (i) the protein is flexible enough to adapt to the new geometrical requirements of the reduced metal, or (ii) re-oxidation of the metal by molecular oxygen or reduction products of molecular oxygen such as superoxide or hydrogen peroxide occurs faster than the release of the reduced metal from the metal-binding site.

In a recent example for electron transfer coupled ligand dynamics in the model complex Cu(I/II)(TTCN)2, we could show that oxidation of tetrahedral [Cu(I)(TTCN)2]+ potentially led to the loss of a ligand to yield [Cu(II)(TTCN)(H2O)3]2+ before geometrical reorganization around the Cu(II) atom allowed the addition of TTCN to give the final product, octahedral [Cu(II)(TTCN)2]2+.2 Analogous mechanisms may lead to kinetic complexities during protein oxidation which may be reflected in oxidation yields and kinetics as well as the nature of the oxidation products.

One amino acid particularly affected by MCO is histidine (His) (15-17). There is as yet no detailed mechanism available for His oxidation in proteins, except a crystallographic study which showed the feasibility of metal-catalyzed processes within the metal-binding domain of glutamine synthetase (18). However, such mechanisms and reaction stoichiometries are important to recognize, particularly in light of the fact that His oxidation products such as 2-oxo-His have been proposed as general markers for oxidative stress in vivo (19).

In this work, we report a detailed mechanistic investigation of the ascorbate/Cu(II)/O2-induced oxidation of His in a selected model protein, human growth hormone (hGH), with the particular objective to identify the nature of the oxidizing species and the structural requirements for MCO of hGH. Ascorbate/Cu(II)/O2 has not only been used extensively as a non-enzymatic model oxidizing system but is of physiological relevance to conditions of oxidative stress. hGH was chosen because of its well characterized metal-binding site which shows high affinity for divalent metal ions, such as Zn(II) and Co(II) (20). This site consists of His18 and His21 from helix I, and Glu174 from helix IV. The primary sequence of hGH is displayed in Fig. 1 (21).


Fig. 1. Primary structure of hGH and assignment of tryptic fragments. Two disulfide bonds are Cys53(-S-S-)Cys165 and Cys182(-S-S-)Cys189. hGH has four main alpha -helices composed of residues 9-34, 72-92, 106-128, and 155-184 (24). The metal-binding amino acids are His18, His21, and Glu174 (20), shown in bold letters.
[View Larger Version of this Image (16K GIF file)]



EXPERIMENTAL PROCEDURES

Materials

Recombinant hGH (lyophilized form) and its synthetic tryptic fragment T3 (Ala-His-Arg) were provided by Genentech, Inc. (South San Francisco, CA). The protein was exchanged into a desired buffer by ultrafiltration using Microcon-10 microconcentrators from Amicon, Inc. (Beverly, MA). Typically 10 mg of hGH were dissolved in 500 µl of 20 mM phosphate buffer, pH 7.4, the solution transferred into the microconcentrator and centrifuged for 20 min at 12,000 rpm in a tabletop centrifuge. Subsequently, the protein was washed three times with 250 µl of 20 mM phosphate buffer, pH 7.4. The protein was finally taken up in a total volume of 4.5 ml of 20 mM phosphate buffer, pH 7.4, and the concentration of hGH in the stock solution determined by UV absorption at 277 nm where A2770.1% = 0.82 cm-1 (22), [MWhGH = 22,125] (21). The Met14 sulfoxide derivative T2ox of the tryptic fragment T2 was obtained synthetically: first, an amount of about 1.5 nmol of authentic T2 was obtained by tryptic digestion of hGH (conditions see below), purification by reversed-phase HPLC (conditions see below), and lyophilization. The fragment was redissolved in 100 µl of neutral water and reacted for 12 h with 10 mM hydrogen peroxide. The only product of this reaction was T2ox which was purified by reversed-phase HPLC and characterized by MALDI-TOF mass spectrometry. Sequencing grade trypsin was obtained from Promega (Madison, WI); 1-propanol-d7 (CD3CD2CD2OH) and 18O-labeled water (95-98% atom pure 18O) were from Cambridge Isotope Laboratories, Inc. (Andover, MA). All other reagents were of the highest grade commercially available.

Reaction Conditions

Unless specified otherwise, the oxidation of hGH by the ascorbate/Cu(II)/O2 system was carried out under the following conditions (referred to as "standard conditions"): 200-µl solutions containing 18 µM hGH, 10 µM CuCl2, 100 µM ascorbate, and 10 mM sodium phosphate buffer, pH 7.4, were incubated at 25 °C in 2-ml vials. In this way sufficient oxygen was available in the headspace of the vial to replenish oxygen in the reaction solution consumed through the oxidation of ascorbate. The reagents were sequentially added in the following order: water, hGH (in phosphate buffer), CuCl2, and ascorbate. The reaction was started by the addition of ascorbate, although switching the order of CuCl2 and ascorbate did not affect the reaction kinetics or final yields (data not shown). All solutions were made with doubly distilled deionized water. The phosphate buffer stock solution was treated with Chelex 100 resin (Bio-Rad; 5 g/100 ml) to minimize metal contamination. Stock solutions of CuCl2 and ascorbate were freshly prepared prior to the reactions.

Atmospheric oxygen was excluded in some reactions. Since purging inert gas through the protein-containing solutions and even freeze-thaw cycles would cause foaming and denaturation of the protein, a special piece of glassware was designed to saturate the protein reaction mixture with N2. The glassware of an H-shape was made of two small top-end-open glass tubes joined in the middle. The reaction mixture was placed in one side of the device and H2O in the other (solution surfaces below the joint). The two openings were then fitted with rubber septa, and two hypodermic needles were used as the N2 gas inlet (needle immersed in H2O) and outlet (needle above the reaction solution), respectively. Before the inlet, N2 gas was purified from residual oxygen contamination by an OxiClear disposable gas purifier (LabClear, Oakland, CA). This whole setting enabled the efficient removal of O2 without protein denaturation or solvent evaporation. Other reagents, purged with N2 separately, were added through the rubber septa using a gas tight syringe.

Tryptic digestion of hGH samples was performed at an initial dose of trypsin: hGH = 1:50 (w/w) in 10 mM phosphate buffer, pH 7.4, at 37 °C. A second equivalent dose was added after 2 h, and the samples were incubated for another 4 h. The digest was then quenched by cooling to 4 °C.

Chromatographic Analysis

Intact hGH was monitored by weak anion exchange chromatography (AXC), reversed-phase HPLC (RP-HPLC), and size exclusion chromatography with UV detection at 214, 230, or 278 nm. Tryptic maps were analyzed by RP-HPLC and cation exchange chromatography (CXC). Table I summarizes the detailed information for the employed modes of chromatography.

Table I.

Information on chromatographic analyses of hGH


Method Column specifications Temp. Flow UV, lambda Mobile phase compositions

°C ml/min nm
AXCa TSK DEAE-5PW 7.5 × 75 mm 45 0.5 230 50 mM potassium phosphate, 4% ACN, pH 5.8 
RP-HPLC intact protein PLRP-S 4.6 × 150 mm 50 1.0 214 30 mM sodium phosphate, 44% ACN, pH 7.0 
SECa TSK G2000 SWXL 7.8 × 300 mm 25 1.0 214 Nondenaturing: 50 mM sodium phosphate, 0.15 M NaCl, pH 7.2 
278 Denaturing: 0.2 M sodium phosphate, 0.1% SDS, pH 6.8 
RP-HPLCa tryptic mapping Nucleosil C18 4.6 × 150 mm 40 0.5 214 Linear gradient: 5-60% ACN over 120 min, with constant 0.1% trifluoroacetic acid
CXC TSK SP-2SW 4.6 × 250 mm 25 1.0 214 Linear gradient: 50-500 mM sodium phosphate over 30 min

a Developed based on information from Ref. 21.

The quantitative analysis of ascorbate was performed by capillary electrophoresis equipped with an electrochemical detector. The capillary electrophoresis system employed a 50 µm x 70-cm fused silica capillary (Polymicrotechnologies; Phoenix, AZ), operating at 30 kV and 0.5 nA with 10 mM MES, pH 5.5, buffer as the mobile phase. The power supply was from Spellman High Voltage Corp. (Plainview, NY), and electrochemical detection was accomplished with an amperometric detector (model LC-4B; Bioanalytical Systems; West Lafayette, IN) employing a 33-µm carbon fiber electrode operating at 800 mV (versus Ag/AgCl). Prior to capillary electrophoresis with electrochemical detection analysis, the reaction samples were quenched by EDTA and filtered by Microcon-10 micro-concentrators to remove protein components.

NMR Analysis

The 500 MHz 1H NMR spectra were recorded on a Bruker AM-500 instrument. Typically 5 mg of hGH were exchanged into 400-800 µl of 10 mM sodium phosphate buffer in D2O, pD 7.4, by ultrafiltration using Microcon-10 microconcentrators (see "Materials") (the introduction of protons by 10 mM sodium phosphate into D2O can be considered negligible). When 1-propanol-d7 was used as a co-solvent (40%, v/v), it was added after the ultrafiltration. Before the addition of 1-propanol-d7, the pD was measured to be 7.4. The final concentrations of hGH in all NMR samples were 0.2-0.4 mM. The samples were then kept overnight at room temperature to ensure the exchange of amide protons with D2O before NMR analysis.

Mass Spectrometry

Molecular weights of the purified tryptic fragments of hGH were analyzed by electrospray ionization (ESI) and/or matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) mass spectrometry (MS). ESI-MS experiments were performed on an Autospec-Q tandem hybrid mass spectrometer (VG Analytical Ltd., Manchester, United Kingdom) equipped with an OPUS data system. MALDI-TOF mass spectra were obtained on a Hewlett Packard model G2025A which exhibited an accuracy of typically ± 1 atomic mass unit for peptides with molecular masses below 3 kDa. The matrices for MALDI-TOF measurements (supplied by Hewlett Packard, Atlanta, GA) were 3,5-dimethoxy-4-hydroxycinnamic acid for peptides of mass >1 kDa and alpha -cyano-4-hydroxycinnamic acid for peptides of mass <1 kDa.

Fluorescence and Circular Dichroism Spectroscopy

A Shimadzu spectrofluorometer (RF 5000U) was used to monitor the fluorescent behavior of the tryptophan residue after the oxidative modification of hGH (standard conditions). The excitation wavelength was set at 298 nm, and the emission spectrum was recorded between 310 and 400 nm. The presence of dityrosine, one potential oxidation product of tyrosine, was also examined by fluorescence spectroscopy with an excitation wavelength at 325 nm and emission spectra recorded between 400 and 500 nm (23). Circular dichroism (CD) measurements of hGH were performed on an AVIV model 60 DS spectrometer (Lakewood, NJ) essentially according to a published procedure (21).

Amino Acid Analysis

For the preparation of oxidized hGH for amino acid analysis, hGH (19 µM) was oxidized in 1 ml of solution (standard conditions). After a reaction time of 100 min, 100 µl of 0.1 mM EDTA were added, maintaining pH 7.4. The solution was divided into two aliquots of 550 µl, both of which were concentrated by centrifugation in a Microcon-10 microconcentrator and washed twice with 250 µl of 20 mM sodium phosphate buffer, pH 7.4. Both protein samples were then redissolved into 250 µl of 20 mM sodium phosphate buffer, pH 7.4, and two 100-µl aliquots of each sample lyophilized in small vials. Assuming an approximately 30% loss of material as a result of centrifugation in the microconcentrators, each aliquot contains about 3.3 nmol of protein. The protein samples were hydrolyzed with 6 M HCl at 110 °C for 20 h. Subsequent amino acid analysis was performed by Commonwealth Biotechnologies, Inc., Richmond, VA.


RESULTS

Product Characterization for the Oxidation of hGH by Ascorbate/Cu(II)/O2

Chemical Characterization

During the oxidation of 18 µM hGH by 10 µM CuCl2 and 100 µM ascorbate in air saturated 10 mM phosphate buffer at pH 7.4, AXC, pH 5.8, but not RP-HPLC revealed the rapid degradation of the hGH main peak paralleled by the appearance of two new more acidic peaks, implying the modification of some basic amino acid residues. Fig. 2 superimposes representative AXC chromatograms recorded immediately after the start (2 min) and at the end of a reaction (i.e. after complete consumption of ascorbate). The two major product peaks are indicated as peak 1 and peak 2. Each peak may contain more than one degradation product as suggested by the peak shoulders. The time profiles for their formation as well as for the consumption of hGH are shown in Fig. 3.


Fig. 2. Representative anion exchange chromatograms of hGH at 2 (reaction at initiation) and 150 (reaction at termination) minutes of exposure to ascorbate/Cu(II)/O2. Reaction conditions (standard reaction conditions): air-saturated aqueous solution containing 18 µM hGH, 10 µM CuCl2, 100 µM ascorbate, and 10 mM sodium phosphate buffer, pH 7.4, at 25 °C. AXC conditions are as described in Table I.
[View Larger Version of this Image (13K GIF file)]



Fig. 3. Time course of hGH oxidation by ascorbate/Cu(II)/O2 as monitored by AXC (standard reaction conditions, see legend of Fig. 2; chromatographic conditions, see Table I); x axis: reaction time, y axis: % peak area relative to peak area of initial unoxidized hGH. The solid lines through the points are included for clarity but are not computer fits.
[View Larger Version of this Image (15K GIF file)]


Prior to the characterization of peak 1 and peak 2 in Fig. 2, the chemical modifications of a complete reaction mixture were characterized by tryptic mapping (see Fig. 1 for the primary sequence of hGH and the assignment of the tryptic fragments). Most of the tryptic fragments are well separated by RP-HPLC; complimentary tryptic mapping by CXC was used to monitor the T3 fragment, which is highly positively charged and elutes closely to the solvent front during the RP-HPLC analysis. Representative tryptic maps of native and oxidized hGH are displayed in Fig. 4 for RP-HPLC and Fig. 5 for CXC. Peak assignment in the RP-HPLC chromatogram was achieved by fraction collection and off-line analysis by ESI or MALDI-TOF MS. The identities of the respective peaks together with the theoretical and measured molecular masses are displayed in Table II. The retention behavior of the T3 fragment during CXC was confirmed by comparison to a synthetic standard.


Fig. 4. Reversed-phase chromatographic analysis of tryptic maps of native and oxidized (standard conditions) hGH. Top, complete chromatograms. Bottom, expanded details for retention times of tR = 3-7 min and tR = 43-77 min of the top chromatograms. Peak assignment, mass analysis, and oxidation yields are summarized in Table II. RP-HPLC conditions are as described in Table I.
[View Larger Version of this Image (28K GIF file)]



Fig. 5. Representative chromatograms of tryptic maps analyzed by cation exchange chromatography of native and oxidized hGH. CXC conditions are as described in Table I.
[View Larger Version of this Image (14K GIF file)]


Table II.

Characterization and oxidation yields of the tryptic fragments of hGHa

Reaction conditions were the same as described in the caption of Fig. 2.


Peaka Tryptic fragment Amino acid sequence Mr theoreticalb Mr measuredc Lossd Gaine

%
1 T3 Ala17-Arg19 382.2 382.2f 63.5g
2 T3ox Ala17-Arg19 398.3f 63.5
3 T7 Glu65-Lys70 761.4 761.7 2.6
4 T14 Gln141-Lys145 625.3 625.4  -0.8
5 T14ph Gly141-Lys145 608.7 608.3  -4.6
6 T12 Leu128-Arg134 772.4 772.7 0
7 T10c1i Ser95-Asn99 536.3 536.3  -2.4
8 T13 Thr135-Lys140 692.4 692.5 0
9 T20-SS-T21 Ile179-Arg183 + Ser184-Phe191 1399.6 1399.6 9.7
10 T15 Phe146-Lys158 1488.7 1488.7 0.1
11 T2ox Leu9-Arg16 995.0 9.7
12 T8 Ser71-Arg77 843.5 843.5  -3.4
13 T17-18-19 Lys168-Arg178 1380.6 1380.7 12.6
14 T18-19 Asp169-Arg178 1252.3 1252.7 8.5
15 T2 Lys9-Arg16 978.5 978.6 10.1
16 T1 Phe1-Arg8 929.5 929.6 3.7
17 T10c2i Ser100-Lys115 1742.9 1742.8 17.4j
18 T11 Asp116-Arg127 1360.7 1360.7 3.6
19 T4ox1 Leu20-Lys38 2356.1f 2.0
20 T4ox2 Leu20-Lys38 2391.4f 2.8
21 T4 Leu20-Lys38 2341.1 2341.4 49.1
22 T4ox3 Leu20-Lys38 2357.4 24.9
23 T10 Ser98-Lys115 2261.1 2261.4  -2.7
24 T6-SS-T16 Tyr42-Arg64 + Asn159-Arg167 3760.8 3762.2 5.2
25 T9 Ile78-Arg94 2054.2 2054.2 2.7

a Peak numbers are assigned in the order of the retention time as shown in Fig. 4.
b Theoretical values of Mr are calculated based on the most abundant monoisotopic species.
c These Mr values were obtained from ESI mass spectrometry unless specified otherwise. The values are reported as [M + 1]+ - 1.
d Loss of the tryptic fragments was calculated based on oxidation-resistant fragments T12 and T13 as internal standards.
e The % gain of the oxidation products was calculated based on the extinction coefficient of the parent peptide.
f These masses were determined by MALDI-TOF MS. The values are reported as [M + 1]+ - 1.
g The quantitative yield of T3 was obtained from tryptic mapping by CXC.
h This peptide derives from the pyroglutamate formation at the N terminus of T14.
i These two peptides were from chymotryptic-like cleavage of T10.
j This loss is not paralleled by any loss of T10 (peak 23) and, therefore, probably not related to any chemical modification during hGH oxidation.

A comparison of the chromatograms in Figs. 4 and 5 reveals that a number of fragments were modified by the ascorbate/Cu(II)/O2 system, but T3 and T4 constituted the predominant targets. Table II summarizes the loss of each fragment and the yields of oxidation products. Two oxidation resistant fragments, T12 and T13, were chosen as the internal standards. Each oxidation product was quantified by comparing its peak area with the original peak area of its parent fragment in the unmodified protein, assuming similar response factors.

Tryptic mapping by RP-HPLC in conjunction with off-line ESI-MS revealed several product peaks originating from T4 (labeled T4ox1 to T4ox3) with the total area of these peaks accounting for as much as 60% of the loss of T4. The major product (T4ox3) had a molecular weight corresponding to MWT4 + 16, implying the incorporation of an oxygen atom into T4. One minor product had MWT4 + 15 (T4ox1) which, considering the accuracy of our MALDI-TOF MS measurements of ± 1 atomic mass unit, might in fact correspond to MWT4 + 16. One additional minor product (T4ox2) corresponded to MWT4 + 50. 

Although T3 was well separated from the other fragments by CXC (which enabled the quantification of its loss), no new product peak was observed, suggesting that the products were less positively charged and co-eluted with the other earlier eluting fragments (Fig. 5). By RP-HPLC (Fig. 4), T3 (peak number 1) was barely resolved from the solvent front even with initially low concentrations of acetonitrile. The MALDI-TOF mass spectrum of peak number 1 suggested that it contained T3 as an exclusive compound. During the oxidation reaction, the loss of peak number 1 was accompanied by the emergence of a later eluting peak (peak number 2 in Fig. 4). MALDI-TOF MS analysis of peak number 2 revealed the presence of one major component with MWT3 + 16 and two minor components with MWT3 + 38 and MWT3 + 54. Peak number 2 is thus assigned as T3ox, the product of oxygen incorporation into T3. The two minor components likely represent the potassium adducts of T3 (MWT3 + 38) and T3ox (MWT3 + 54), respectively. Based on the peak areas of peak number 1 and peak number 2 in Fig. 4 it appears that T3 is nearly quantitatively converted into T3ox.

In addition to oxidation products from T3 and T4, RP-HPLC also resolved a product peak, number 11, which had identical retention time and molecular weight as the synthetic standard of the Met14 sulfoxide variant of T2. This conversion is nearly quantitative, i.e. 10.1% loss of T2 compares to 9.7% gain of T2ox.

The fragments T17-T18-T19 and T18-T19 suffered an approximately 10% loss which is, however, not accompanied by any detectable major product. The potential oxidation sites on these fragments are Met170 and/or Glu174. The Glu174 residue is the third component of the metal-binding site in addition to the two His residues (20). Met170, although buried in the hydrophobic core, is also located close to the metal-binding site according to the crystal structure of hGH (24). Our results (see below) will provide evidence for the formation of hydroxyl radical-like species at the metal-binding site of the protein. Both Glu (25) and Met (26-30) are susceptible to the attack by hydroxyl radicals, yielding a variety of non-selective reaction products. We note that Met170 could theoretically be oxidized to Met sulfoxide by two-electron oxidants such as hydrogen peroxide, but no corresponding sulfoxide variants of T18-T19 or T17-T18-T19 were experimentally observed. The observed loss of T10c2 (chymotryptic cleavage of T10; Ser100-Lys115) is considered an experimental artifact since both fragments T10c1 and T10 did not show any loss induced by metal-catalyzed oxidation.

The two major oxidation peaks (peaks 1 and 2) of hGH, resolved by AXC (Fig. 2), were purified for individual analysis by tryptic mapping. Peak 2 contained all the original fragments except T3, suggesting that this peak represented a rather homogeneous fraction of a hGH variant only modified at T3. Peak 1 suffered significant loss of T4 (about 65%) in addition to a minor loss of T3 (about 13%).

It appears that the oxidation of hGH by the ascorbate/Cu(II)/O2 system is rather selective to amino acid residues located on fragments T3 and T4. The behavior of the oxidized protein and peptide species during AXC and CXC suggests that the modifications occur at basic amino acid residues. Both T3 and T4 contain histidine residues, His18 and His21, respectively. Theoretically those are susceptible to oxidation/hydroxylation to yield 2-oxo-histidine (2-oxo-His), characterized by a gain of 16 atomic mass units in molecular weight (19, 31-33). Amino acid analysis of oxidized hGH revealed that only His was modified to a significant extent with a loss of 1.4 ± 0.2 His residues per protein, corresponding to 47 ± 7% of the total His residues. This is consistent with the observed loss of T3 and T4. Notably His18 and His21 are components of the metal-binding site of hGH (20). A third His residue, His151, is located next to the known deamidation site of the protein and solvent accessible (21). However, the corresponding tryptic fragment, T15, was not significantly modified by ascorbate/Cu(II)/O2 (Fig. 4 and Table II).

We employed 1H NMR to confirm the degradation of His by virtue of the characteristic chemical shifts of the imidazole C-2 protons (34, 35). As shown in Fig. 6a, the imidazole C-2 protons of the three His residues in native hGH gave distinct chemical shifts at 7.5-8.2 ppm at pD 7.4, comparable to the published values (34, 35). Characteristic resonances were also observed at delta  = 6.47 ppm (resonance II), representing Tyr28 or Tyr160 (35), and at delta  = 6.1 ppm (resonance I), representing a presently non-assigned Tyr and/or Phe residue (34, 35). The oxidative modification by the ascorbate/Cu(II)/O2 system led to a significant decrease of signal intensity only for the C-2 protons of His18 and His21, but not for that of His151 (Fig. 6b). Thus, the NMR results corroborate the findings from both the amino acid analysis and tryptic mapping. We note that the oxidation of hGH also yields a new weak resonance at 6.2 ppm. Earlier we had shown that 2-oxo-His within the peptide Gly-(2-oxo-His)-Gly-Met-Gly-Gly-Gly displays a resonance at around 6.3 ppm, assigned to the C-4 proton of the 2-oxo-His residue (36).


Fig. 6. 500 MHz 1H NMR spectra of: a, 200 µM hGH in D2O, pD 7.4, 10 mM sodium phosphate; b, 200 µM oxidized hGH in D2O, pD 7.4, 10 mM sodium phosphate; a 2-ml sample containing 18 µM hGH was oxidized under standard conditions, the protein separated from oxidized ascorbate by ultrafiltration using Microcon-10 microconcentrators (Amicon), lyophilized, and redissolved in a small volume (150 µl) of D2O, containing 10 mM sodium phosphate, pD 7.4; c, N2-saturated aqueous (D2O) solution of 200 µM hGH, 95 µM CuCl2, 50 µM ascorbate, and 10 mM sodium phosphate, pD 7.4, after complete consumption of reduced ascorbate; d, 400 µM hGH and 200 µM CuCl2 in D2O, 10 mM sodium phosphate, pD 7.4; e, 400 µM hGH and 400 µM ZnCl2 in D2O, 10 mM sodium phosphate; f, 200 µM hGH in 60:40 D2O/CD3CD2CD2OH (v/v), 10 mM sodium phosphate, pD 7.4 (measured before addition of CD3CD2CD2OH); g, 200 µM hGH in D2O, pD 8.5, 10 mM sodium phosphate.
[View Larger Version of this Image (16K GIF file)]


Thus, on the basis of NMR, amino acid analysis, tryptic mapping, and mass spectrometry, we conclude that 2-oxo-His at positions 18 and 21 constitutes a major oxidation product of hGH modified by ascorbate/Cu(II)/O2. Minor yields of other hydroxylated/oxidized products are observed, such as T4ox1 (MWT4 + 15) and T4ox2 (MWT4 + 50), where T4ox2 most likely corresponds to 4,5-dihydroxy-2-oxo-His (37), but its low yield precluded further characterization by NMR.

Structural Characterization

Oxidation of hGH by the ascorbate/Cu(II)/O2 system did not significantly affect the higher order structure of hGH. Nondenaturing size-exclusion chromatography revealed about 23% loss of the hGH monomer paralleled by some dimer formation, while denaturing size-exclusion chromatography indicated only a 4% loss in monomer content. hGH has a single tryptophan in helix II, Trp86, which constrains the distance between helices II and IV through a hydrogen bond with Asp169 on helix IV (24). The fluorescence property of Trp86 is a good indicator of the integrity of the protein tertiary structure (38, 39). With excitation at 298 nm, the emission spectrum of hGH was recorded between 310 and 400 nm. No significant change of the spectrum was observed for oxidized hGH. Fluorescence spectroscopic analysis also revealed no formation of dityrosine during the oxidation (lambda ex = 325 nm, lambda em = 400-500 nm). Far-UV circular dichroism measurements showed only minimal loss of overall alpha -helical content of the oxidized hGH. In addition, the NMR spectra of the native and oxidized hGH (Fig. 6, a and b) were largely identical besides the changes of the intensity of the histidine-related signals, confirming that the integrity of the higher order structure of hGH was maintained after the oxidation by ascorbate/Cu(II)/O2.

The Source of Oxygen in 2-Oxo-His

Although 2-oxo-His has been documented as a product from metal-catalyzed oxidation of His (19, 31-33, 36), its mechanism of formation as well as the source of oxygen has never been characterized in detail. In particular the source of oxygen is critical: incorporation of oxygen from the solvent water would indicate an intermediary imidazole cation or radical cation, generated by electron transfer to any type of oxidant, followed by reaction with aqueous OH-; in contrast, the incorporation of oxygen from molecular oxygen of air would indicate the involvement of 1O2, HO· (formed via three consecutive reduction steps of O2), or possibly peroxyl radicals formed by the addition of O2 to any type of carbon-centered radical of the His residue. Therefore, our oxidation experiment of hGH by ascorbate/Cu(II)/16O2 was carried out in 18O-labeled water, H218O (95-98% atom pure 18O), followed by tryptic digest in the same solvent. By RP-HPLC separation, four fractions containing T3, T4, and their 2-oxo-His variants (T3ox and T4ox3) were collected and lyophilized immediately. MALDI-TOF and ESI-MS analysis revealed a gain of 4 atomic mass units for purified fractions of T3 and T4, and the fractions of T3ox and T4ox3 exhibited molecular weights of MWT3 + 20 and MWT4 + 20, respectively. The mass differences between the products and their parent peptides were still 16 atomic mass units, indicating that the oxygen incorporation into 2-oxo-His resulted from an interaction with a species derived from molecular oxygen. In control experiments we verified that there was no oxygen exchange between 2-oxo-His and the solvent during HPLC purification and subsequent lyophilization. This was done by incubation of 2-oxo-His(16O) of T3 and T4 in H218O/acetonitrile mixtures of compositions and pH identical to the eluting mobile phase during RP-HPLC analysis. The observed increase in molecular weight of 4 atomic mass units for all fragments obtained by tryptic digest in H218O versus H216O is likely the result of oxygen exchange between solvent and a tetrahedral intermediate formed during tryptic hydrolysis.

The Importance of an Intact Metal-binding Site and Site Specificity of Oxidation

The importance of an intact metal-binding site and site specificity for the oxidation of His18 and His21 was further demonstrated by experiments with (i) individual tryptic fragments containing only His18 (T3) or His21 (T4), and (ii) with different concentrations of Cu(II) specifically selected to obtain conditions of [Cu(II)] < [hGH] and [Cu(II)] > [hGH]. Both fragments T3 (synthetic) and T4 (purified by RP-HPLC after tryptic digestion of native hGH) were resistant to oxidation under conditions identical to those applied to hGH (i.e. 18 µM peptide, 10 µM Cu(II), 100 µM ascorbate in air-saturated 10 mM phosphate buffer, pH 7.4).

At constant concentrations of ascorbate and hGH, the oxidation yields were sensitive to different concentrations of Cu(II), [Cu(II)], summarized in Table III. An increase of [Cu(II)] from 2 to 10 µM, i.e. under conditions of [Cu(II)] < [hGH], resulted in a significant increase of the consumption of hGH, expressed as Delta hGH, paralleled by an increase of the yields of oxidation products, quantified as peaks 1 and 2 by AXC (see Fig. 2). However, conditions of [Cu(II)] > [hGH] (i.e. at [Cu(II)] = 25 and 50 µM) resulted in a slight decrease of hGH consumption, paralleled by a slight decrease of the product peaks 1 and 2, as compared with [Cu(II)] = 10 µM. The inspection of the respective ratios peak 1/Delta hGH and peak 2/Delta hGH reveals a bell-shaped characteristic for peak 1/Delta hGH with low values at 2 and 50 µM Cu(II), whereas peak 2/Delta hGH shows a steady decrease with increasing [Cu(II)]. These differences are likely the result of different kinetics for the oxidation of His18, represented by peak 2, and His21, the major oxidized amino acid of peak 1. It appears that His18 represents an early target of hGH, followed by His21 and, eventually, some other minor targets such as Met14.

Table III.

Oxidation of hGH in the presence of various initial concentrations of Cu(II)

Conditions: air-saturated aqueous solution containing 100 µM ascorbate, 18 µM hGH, 10 mM sodium phosphate buffer, pH 7.4.


Cu(II) [hGH]lefta  Delta hGHa,b Peak 1a,c Peak 2a,c  <FR><NU>Peak 1</NU><DE>&Dgr;<SUB><UP>hGH</UP></SUB></DE></FR>  <FR><NU><UP>Peak</UP> 2</NU><DE>&Dgr;<SUB><UP>hGH</UP></SUB></DE></FR>

µM %
2 88.6  ± 0.8 11.4 1.3  ± 0.4 6.4  ± 0.8 0.11 0.56
5 54.6  ± 4.8 45.4 9.6  ± 1.8 20.9  ± 0.3 0.21 0.46
10 5.0  ± 1.8 95.0 17.8  ± 0.2 35.7  ± 1.2 0.19 0.38
25 6.5  ± 1.6 93.5 15.0  ± 0.7 31.6  ± 0.4 0.16 0.34
50 13.1  ± 1.6 86.9 13.0  ± 1.0 27.5  ± 1.0 0.15 0.32

a % Related to peak area of initial concentration of native hGH.
b Delta hGH = [hGH]initial - [hGH]left.
c Refers to peaks 1 and 2 obtained by AXC (Fig. 2).

The oxidation of hGH (18 µM hGH, 10 µM Cu(II), 100 µM ascorbate in air-saturated 10 mM phosphate buffer, pH 7.4) was completely inhibited by the addition of 15 µM EDTA. However, Zn(II) (added as ZnCl2) was unable to inhibit the oxidation of hGH until the ratio of [Zn(II)]:[Cu(II)] was higher than 80:1. This result suggests a strong binding of Cu(II) to the metal-binding site of hGH and is particularly interesting with regard to the fact that Zn(II) regulates storage and biological activity of hGH under physiological conditions (20, 40).

Characterization of Hydrogen Peroxide and Cu(I) as Important Intermediates

Hydrogen Peroxide

The oxidation of hGH was inhibited by >50 units/ml native catalase (18 µM hGH, 10 µM Cu(II), 100 µM ascorbate in air-saturated 10 mM phosphate buffer, pH 7.4) but not by heat-inactivated catalase, indicating that freely diffusable hydrogen peroxide plays an important role during the reaction. Cu,Zn-superoxide dismutase (at 200 units/ml) had no effect on the final yields of hGH oxidation but slowed down the reaction kinetics.

Cu(I)

Three vials (I-III) with each containing an N2-saturated solution of 18 µM hGH and 15 µM Cu(II) in 10 mM phosphate buffer, pH 7.4, were incubated under N2 for 20 min with 7.5 µM ascorbate, i.e. with a ratio of [ascorbate]:[Cu(II)] = 1:2 to promote a stoichiometric formation of Cu(I). Subsequently, to vial I a final concentration of 100 µM EDTA (separately saturated with N2) was added under N2, vial II was opened and exposed to air, and to vial III a final concentration of 100 µM hydrogen peroxide (H2O2) (separately saturated with N2) was added under N2. Analysis of the reaction mixture of vial I by AXC revealed that hGH had not been oxidized during the 20-min incubation (before the addition of EDTA), whereas analysis by capillary electrophoresis with electrochemical detection revealed that the reduced form of ascorbate had been completely consumed. Analysis of the reaction mixtures of vials II and III revealed that hGH had been oxidized, with higher yields in vial III, where the product stoichiometries were identical to those obtained during the oxidation of hGH by the standard system in air-saturated solutions, i.e. showed predominant oxidation of T3 and T4. In a separate experiment we added a final concentration of 100 µM H2O2 aerobically to a solution containing 18 µM hGH, 10 µM Cu(II), and 10 mM phosphate buffer, pH 7.4. Here, no oxidation of hGH was observed over a time of 150 min. These results imply the anaerobic reduction of Cu(II) to Cu(I) via net reactions 1 and 2, where AH-, Abardot , and A denote ascorbate, the ascorbyl radical anion, and dehydroascorbic acid, respectively.
<UP>Cu</UP>(<UP>II</UP>)+<UP>AH<SUP>−</SUP> </UP>⇌<UP> Cu</UP>(<UP>I</UP>)+<UP>A&cjs1138; </UP>+<UP>H<SUP>+</SUP></UP> (Eq. 1)
<UP>Cu</UP>(<UP>II</UP>)+<UP>A&cjs1138; </UP>→<UP>Cu</UP>(<UP>I</UP>)+<UP>A</UP> (Eq. 2)
Subsequently, upon exposure to oxygen or H2O2, Cu(I) promotes the formation of an oxidizing species ultimately responsible for the oxidation of His in T3 and T4, respectively (see "Discussion"). Since aqueous Cu(I) is thermodynamically unstable with respect to disproportionation (reaction 3) (41, 42) we expect that all Cu(I), present after the complete anaerobic consumption of ascorbate, is ligated by the protein.
2 <UP>Cu</UP>(<UP>I</UP>)→<UP>Cu</UP>(<UP>O</UP>)+<UP>Cu</UP>(<UP>II</UP>) (Eq. 3)
Spectroscopic experiments were undertaken to test this hypothesis. An N2-saturated aqueous (D2O) solution containing 10 mM phosphate buffer, pD 7.4, 200 µM hGH, and 95 µM CuCl2 was reacted with 50 µM ascorbate and subsequently transferred under N2 in a gas tight syringe into an NMR tube (separately saturated with N2). This mixture essentially contained hGH, hGH-Cu(I), and oxidized ascorbate (see above). The mixture was allowed to stand for 24 h (to promote H/D exchange) before the NMR spectrum, displayed in Fig. 6c, was recorded. The comparison of Fig. 6, a and c, reveals that neither the three His resonances nor resonances I and II were affected by the presence of Cu(I). However, two new sets of resonances appear around the signal for the C-2 proton of His151, labeled III and IV, respectively. Furthermore, several broad signals in the region of delta  >=  8.4 ppm show higher intensities as compared with the NMR spectrum of native hGH (Fig. 6a). All resonances persist over a period of 3 days. The resonances III and IV may represent aromatic protons whereas the resonances at delta  >=  8.4 ppm likely represent amide protons more resistant to H/D exchange. Several control experiments were run. (i) We ensured that the anaerobic reaction of 50 µM ascorbate with 95 µM Cu(II) in the absence of hGH did not produce any signal between 6.0 and 9.0 ppm (data not shown). (ii) The 1H NMR spectrum of an aerobic aqueous (D2O) solution containing 10 mM phosphate buffer, pD 7.4, 400 µM hGH, and 200 µM Cu(II) (Fig. 6d) shows several significant differences as compared with the spectra of the hGH-Cu(I) complex (Fig. 6c) and native hGH (Fig. 6a). The resonances of His18 and His21 at delta  = 8.29 ppm and delta  = 7.58 ppm, present for native hGH and hGH-Cu(I), are essentially absent in the presence of Cu(II). Instead of a sharp resonance for His151, Fig. 6d displays a relatively weak, broad, and noisy signal around delta  = 7.8-7.9 ppm. Small resonances on both sides of the large peak at delta  = 6.8 ppm, present for native hGH and hGH-Cu(I), are nearly absent in the presence of Cu(II). The intensity of the resonance at delta  = 7.4 ppm (Fig. 6, a and c) is significantly reduced in the presence of Cu(II). (iii) The 1H NMR spectrum of an aerobic aqueous (D2O) solution containing 10 mM phosphate buffer, pD 7.4. 400 µM hGH and 400 µM Zn(II) is displayed in Fig. 6e. Some features of the spectrum are similar to those observed in the presence of Cu(II), namely the absence of the resonances for His18 and His21 at delta  = 8.29 ppm and delta  = 7.58 ppm, respectively, the absence of the small resonances on both sides of the peak at delta  = 7.4 ppm, and the broad, although more intense, resonance around delta  = 7.9 ppm. Binding of Zn(II) to the metal-binding site of hGH is well established (Kd approx  1 × 10-6 M) (20). The similar effects of Zn(II) and Cu(II) on the 1H NMR spectrum of hGH may indicate that also a fraction of Cu(II) binds to the metal-binding site. However, the physical rationale for the disappearance of the resonances of His18 and His21 may be different for both metals: first, the geometry of the Zn(II) and the Cu(II) complex may be different resulting in different ligand fields. Second, due to the paramagnetic nature of Cu(II), the disappearance of the His resonances is likely caused by extensive peak broadening. We note, however, that in the presence of Cu(II) particularly the resonances at delta  <=  6.8 ppm (Fig. 6d) are clearly visible and their peak shape comparable to those of native hGH (Fig. 6a), indicating that peak broadening by the paramagnetic Cu(II) seems to be restricted to amino acid side chains directly involved in the binding of Cu(II) and not related to an overall field effect. This feature may be taken as an additional evidence for the binding of Cu(II) to the metal-binding site of hGH. The NMR results can be summarized as follows. The incubation of hGH with Cu(II) results in spectral changes some of which are similar to those induced by the binding of Zn(II) to the metal-binding site, suggesting that a fraction of Cu(II) also binds to the metal-binding site of hGH. The spectrum of the hGH-Cu(II) system is distinctly different from the spectrum of the hGH-Cu(I) system which, nevertheless, differs from the spectrum of native hGH. The fact that some of the amide resonances of the hGH-Cu(I) complex appear to be more resistant to H/D exchange than those of native hGH suggests a more rigid structure of hGH-Cu(I). This could be rationalized by the binding of Cu(I) to the metal-binding site of hGH, supported by the observation that the exposure of the hGH-Cu(I) to O2 or hydrogen peroxide (anaerobic) resulted in the site-specific oxidation of hGH at the His residues of T3 and T4, i.e. within the metal-binding site.

The Effect of 1-Propanol

It was reported that 1-propanol induces a small change of the tertiary structure of hGH to a molten globular state (39). When we exposed hGH to the ascorbate/Cu(II)/O2 system in mixtures of (i) methanol and water, and (ii) 1-propanol and water, there was no effect of methanol up to a content of 40% (v/v). However, the extent of oxidation followed a sigmoidal behavior with no significant change between 0 and 20% 1-propanol (v/v), but about 80% inhibition at 28% 1-propanol (Fig. 7).


Fig. 7. Effect of the content of 1-propanol and methanol on hGH oxidation by ascorbate/Cu(II)/O2 (standard reaction conditions; see legend to Fig. 2).
[View Larger Version of this Image (17K GIF file)]


This significant change in inhibition within a narrow range of 1-propanol content suggests the involvement of a structural transition of the protein. Indeed, the sigmoidal inhibition curve of hGH oxidation approximately paralleled the variation of protein tertiary structure characterized by fluorescence spectroscopy (39). On the other hand, the lack of any effect of comparable concentrations of methanol shows that it was not merely the presence of an alcoholic functional group in the solvent which caused the observed protection. Fig. 6f shows the 1H NMR spectrum of native hGH in D2O in the presence of 40% 1-propanol-d7 (CD3CD2CD2OH) (v/v; pD = 7.4, measured before the addition of 1-propanol-d7). The three resonances in the range delta  = 7.6-7.8 ppm were assigned to the three His residues in the following way. A sample of hGH was exposed to ascorbate/Cu(II)/O2 in the absence of 1-propanol to oxidize His18 and His21. Oxidized hGH is characterized through the lack of the characteristic resonances of His18 and His21 in the NMR spectrum (Fig. 6b). Subsequently, the sample containing oxidized hGH was lyophilized and redissolved in aqueous (D2O) solution, pD 7.4, containing 40% (v/v) 1-propanol-d7. The NMR spectrum of this sample was identical to Fig. 6f except that it did not show the two resonances in the region delta  = 7.60-7.65 ppm. Therefore, in Fig. 6f the two resonances around delta  = 7.60-7.65 ppm can be assigned to His18 and His21 in some order, whereas the resonance around 7.8 ppm can be assigned to His151, as labeled. In particular the resonance of His18 is significantly shifted in the presence of 40% (v/v) 1-propanol-d7 as compared with native hGH in the absence of 1-propanol. A nearly similar upfield shift (to delta  = 7.64 ppm) of the His18 resonance is achieved in the absence of 1-propanol by an increase of the pD to pD 8.5, shown in the NMR spectrum in Fig. 6g where the broad peak around delta  = 7.65 ppm contains the resonances of both His18 and His151 (35). At pD 8.5 all His residues of hGH present their imidazole side chains in the neutral form (35) and the salt bridge between His18 of helix I and Asp171 of helix IV is lost, resulting in a change of the relative positions of helix I and helix IV (35). By analogy this suggests that 1-propanol perturbs the interactions of helix I and helix IV through the weakening of hydrophobic interactions and, therefore, perturbs the metal-binding site. This hypothesis derives further support from the 1H NMR spectra of Fig. 6, a, f, and g, in the 6.0-6.6 ppm region. Native hGH shows resonances I and II in the absence of 1-propanol at pD 7.4 (Fig. 6a) and pD 8.5 (Fig. 6g), whereas both resonances are absent in the presence of 40% 1-propanol at pD 7.4 (Fig. 6f). Both Tyr28 and Tyr160, represented by resonance II, are located in the contact area of helices I and IV, respectively (35). Resonance I represents a presently unidentified Tyr or Phe residue (34, 35) several of which are located in the contact area of helices I and IV such as, e.g. Phe25, Phe31, and Phe166 (35).

At pH 8.5 in the absence of 1-propanol (18 µM hGH, 10 µM Cu(II), 100 µM ascorbate in air-saturated 10 mM phosphate buffer), hGH was oxidized to an extent comparable to that at pH 7.4 in the absence of 1-propanol (data not shown). Thus, the mere lack of a salt bridge between His18 and Asp171 (see above) is not sufficient for an inhibition of oxidation, suggesting that the protection exerted by 1-propanol requires (i) additional structural modifications of the metal-binding site and/or (ii) other additional parameters. If 1-propanol perturbs the metal-binding site by weakening hydrophobic interactions between helix I and helix IV, 1-propanol has to insert between the helices where it may locate close to the perturbed metal-binding site. A close proximity would enable 1-propanol to scavenge ROS generated at the metal. To test this possibility we performed the oxidation of hGH (standard conditions) in the additional presence of (i) 30% (v/v) 1-propanol-h7 or (ii) 30% (v/v) 1-propanol-d7 (CD3CD2CD2OH). A content of 30% (v/v) 1-propanol-h7 exerted maximum protection with about 77 ± 7% native hGH left after oxidation (Fig. 7). In contrast, the inhibition by 30% (v/v) 1-propanol-d7 was significantly less efficient, with only 39 ± 5% native hGH left after oxidation (data not shown). This product isotope effect, expressed as [residual hGH]1-propanol-h7/[residual hGH]1-propanol-d7 = 1.95, likely reflects a kinetic isotope effect with 1-propanol-d7 being a less efficient scavenger of ROS. Thus, part of the inhibition by 1-propanol may be due to the scavenging of ROS, generated within a perturbed metal-binding site, and the product isotope effect may serve as a basis for mechanistic considerations.


DISCUSSION

The Potential Mechanism of Oxidation

The exposure of hGH to ascorbate/Cu(II)/O2 leads to a rather selective oxidation of His18 and His21 which are part of the metal-binding site of hGH. Hydrogen peroxide represents an important intermediate which appears to exist freely diffusable in solution based on the fact that added catalase inhibited the oxidation of hGH. Mechanistically, the formation of H2O2 can be formulated by Equations 4-8 where the ligand system Lx can represent any potential ligand in our experimental system, i.e. ascorbate, phosphate, protein, or water, or combinations of these. Evidence for Equations 5-8 and the existence of peroxodicopper (II) complexes (trans-µ or µ-eta 2:eta 2), here represented as LxCu(II)(O22-)Cu(II)Lx, has been provided (43).
<UP>AH<SUP>−</SUP></UP>+2 <UP>L<SUB>x</SUB>Cu</UP>(<UP>II</UP>)→<UP>A</UP>+<UP>H<SUP>+</SUP></UP>+2 <UP>L<SUB>x</SUB>Cu</UP>(<UP>I</UP>) (Eq. 4)
<UP>L<SUB>x</SUB>Cu</UP>(<UP>I</UP>)+<UP>O</UP><SUB>2</SUB>→[<UP>L<SUB>x</SUB>Cu</UP>(<UP>II</UP>)(<UP>O &cjs1138;<SUB>2</SUB> </UP>)] (Eq. 5)
[<UP>L<SUB>x</SUB>Cu</UP>(<UP>II</UP>)[<UP>O&cjs1138;<SUB>2</SUB> </UP>]<UP> ⇌ L<SUB>x</SUB>Cu</UP>(<UP>II</UP>)+<UP>O &cjs1138;<SUB>2</SUB> </UP> (Eq. 6)
[<UP>L<SUB>x</SUB>Cu</UP>(<UP>II</UP>)<UP>O &cjs1138;<SUB>2</SUB> </UP>)]+<UP>L<SUB>x</SUB>Cu</UP>(<UP>I</UP>)→<UP>L<SUB>x</SUB>Cu</UP>(<UP>II</UP>)(<UP>O</UP><SUP><UP>2−</UP></SUP><SUB><UP>2</UP></SUB>)<UP>Cu</UP>(<UP>II</UP>)<UP>L<SUB>x</SUB></UP> (Eq. 7)
<UP>L<SUB>x</SUB>Cu</UP>(<UP>II</UP>)(<UP>O</UP><SUP><UP>2−</UP></SUP><SUB><UP>2</UP></SUB>)<UP>Cu</UP>(<UP>II</UP>)<UP>L<SUB>x</SUB></UP>+2 <UP>H<SUP>+</SUP></UP>→2 <UP>L<SUB>x</SUB>Cu</UP>(<UP>II</UP>)+<UP>H<SUB>2</SUB>O<SUB>2</SUB></UP> (Eq. 8)
The site-specific oxidation of His in hGH requires the initial presence of ascorbate and the interaction of H2O2 with Cu(I) chelated by the metal-binding site of hGH, in the following referred to as [hGH-Cu(I)]mbs. This is supported by the following experimental observations. (i) The anaerobic incubation of a hGH-Cu(I) system with 100 µM H2O2 resulted in an efficient oxidation of His18 and His21 within the metal-binding site of hGH. (ii) The aerobic reaction of 100 µM H2O2, representative for maximum hydrogen peroxide concentrations achievable through the oxidation of 100 µM ascorbate (our standard conditions), with a hGH/Cu(II) system did not lead to the oxidation of hGH, although generally, H2O2 can reduce Cu(II) to Cu(I) (Equation 9), e.g. within Cu,Zn-superoxide dismutase (44).
<UP>L<SUB>x</SUB>Cu</UP>(<UP>II</UP>)+<UP>H<SUB>2</SUB>O<SUB>2</SUB></UP>→<UP>L<SUB>x</SUB>Cu</UP>(<UP>I</UP>)+2 <UP>H<SUP>+</SUP></UP>+<UP>O&cjs1138;<SUB>2</SUB></UP> (Eq. 9)
At first we shall present a potential rationale for the lack of hGH oxidation by Cu(II)/H2O2. Our NMR experiments suggest that a fraction of Cu(II) may bind to the metal-binding site of hGH. However, the geometry of the hGH-Cu(II) complex of the metal-binding site could sterically inhibit the reaction with H2O2. Some Cu(II) may also bind to surface-exposed sites of the protein, referred to as [hGH-Cu(II)]surf. In fact, Fig. 6d reveals that not only the resonances of His18 and His21 but also that of His151 is perturbed in the presence of Cu(II). Here, [hGH-Cu(II)]surf will be more solvent accessible and more reactive but any reaction of [hGH-Cu(II)]surf with H2O2 may have little effect on the protein, in particular on His18 and His21. We can exclude a competition in favor of oxygen between oxygen and H2O2 for Cu(I), generated through Equation 9, since the bolus addition of 100 µM H2O2 to a hGH/Cu(II) system corresponds to the maximum possible H2O2 concentration of the standard oxidizing system where oxidation was observed. We would expect also that ascorbate reacts faster with [hGH-Cu(II)]surf instead of [hGH-Cu(II)]mbs. However, [hGH-Cu(II)]surf may not persist in the presence of ascorbate due to Equations 10 or 11 (for simplicity, here the copper ligand in Equations 10 and 11 was chosen as AH- but may be a combination of several ligands; see above). In fact, in particular, any [(AH-)nCu(I)] resulting from Equation 11 may either serve to reduce [hGH-Cu(II)]mbs to [hGH-Cu(I)]mbs or create [hGH-Cu(I)]mbs through transport of Cu(I) into the metal-binding site of hGH.
[<UP>hGH-Cu</UP>(<UP>II</UP>)]<SUB><UP>surf</UP></SUB>+n <UP>AH<SUP>−</SUP> </UP>⇌<UP> hGH</UP>+[(<UP>AH</UP><SUP><UP>−</UP></SUP>)<SUB>n</SUB><UP>Cu</UP>(<UP>II</UP>)] (Eq. 10)
[<UP>hGH-Cu</UP>(<UP>II</UP>)]<SUB><UP>surf</UP></SUB>+(n+1)<UP>AH<SUP>−</SUP></UP>→
<UP>hGH</UP>+[(<UP>AH</UP><SUP><UP>−</UP></SUP>)<SUB>n</SUB><UP>Cu</UP>(<UP>I</UP>)]+<UP>A&cjs1138; + H<SUP>+</SUP></UP> (Eq. 11)
We note that ascorbate will not stay bound at the Cu(I) center within the metal-binding site, i.e. at [hGH-Cu(I)]mbs. This conclusion is based on the fact that no reduced ascorbate was detectable in anaerobic solutions of hGH-Cu(I) subsequent to the addition of EDTA (see above). By 1H NMR spectroscopy we have obtained some evidence for the existence of an hGH-Cu(I) complex although the spectroscopic data alone do not allow us to identify this complex as [hGH-Cu(I)]mbs. This may, however, be concluded from the fact that the anaerobic reaction of a hGH-Cu(I) system with hydrogen peroxide resulted in the specific oxidation of His18 and His21 at the metal-binding site, i.e. likely required the existence of [hGH-Cu(I)]mbs. Further support is derived from the fact that an intact metal-binding site is necessary for the oxidation of His18 and His21. This fact also rationalizes why maximum levels of hGH decomposition were obtained under conditions of [Cu(II)] < [native hGH]. Here, the steady-state concentrations of LxCu(I), independent of whether ligated by the protein or other ligands, cannot exceed the concentration of native hGH until at least native hGH has been consumed to a level of [native hGH] < [Cu(II)]. Thus, at the initial stage of the reaction all Cu(I) has the theoretical chance of incorporation into the metal-binding site, promoting site-specific oxidation. On the other hand, under conditions of [Cu(II)] > [native hGH] there is the possibility that the steady-state levels of LxCu(I) exceed [native hGH] and promote oxidation reactions outside the metal-binding site, most probably leading to an accelerated "nonproductive" decomposition of ascorbate.

There was a marked reduction of hGH oxidation in the presence of additional >= 28% (v/v) 1-propanol, i.e. under conditions where hGH undergoes a structural transition to a molten globular state. The 1H NMR spectrum of hGH recorded in aqueous solution containing 40% (v/v) 1-propanol suggests that the presence of 1-propanol affects the interactions between helix I and helix IV of hGH which construct the metal-binding site. However, even at a high content of 1-propanol there is a residual level (about 23%) of metal-catalyzed oxidation of hGH, suggesting that even the molten globular state of hGH provides a chelating entity sufficient for some binding of the metal. This observation leads to the question whether the basis of the inhibitory role of 1-propanol is only the perturbation of the metal-binding site or the combination of such perturbation and additional factors. The experimental product isotope effect of [residual hGH]1-propanol-h7/[residual hGH]1-propanol-d7 = 1.95 likely reflects a kinetic isotope effect where an oxidizing species reacts less efficiently with the C-D bond of 1-propanol-d7 as compared with the C-H bond of 1-propanol-h7. Hence, there is a more efficient protection by 1-propanol-h7. The fact that a product isotope effect is measurable may indicate that the perturbation by 1-propanol is through insertion between the hydrophobic sites of helix I and helix IV bringing 1-propanol sufficiently close to the yet perturbed metal-binding site to scavenge reactive oxygen species. Three experimental facts would support an assignment that this reactive species is either a hydroxyl radical or a metal-bound equivalent, a complexed hydroxyl radical: (i) the anaerobic reaction of hydrogen peroxide with a hGH-Cu(I) complex, most likely [hGH-Cu(I)]mbs, is expected to proceed according to Equations 12 and 13 (44).
[<UP>hGH-Cu</UP>(<UP>I</UP>)]<SUB><UP>mbs</UP></SUB>+<UP>H<SUB>2</SUB>O<SUB>2</SUB></UP>→[<UP>hGH-Cu</UP>(<UP>II</UP>)<UP>/HO<SUP>−</SUP>/HO</UP><SUP><UP>&z.ccirf;</UP></SUP>]<SUB><UP>mbs</UP></SUB> (Eq. 12)
[<UP>hGH-Cu</UP>(<UP>I</UP>)]<SUB><UP>mbs</UP></SUB>+<UP>H<SUB>2</SUB>O<SUB>2</SUB></UP>→[<UP>hGH-Cu</UP>(<UP>II</UP>)<UP>-OH/HO<SUP>−</SUP>/</UP>]<SUB><UP>mbs</UP></SUB> (Eq. 13)
(ii) The most reactive C-H bonds of 1-propanol are the respective Calpha -H bonds (45). Pulse radiolytic studies have shown that the primary kinetic isotope effect for the reaction of the hydroxyl radical with the Calpha -H and the Calpha -D bonds of methanol-h3 and methanol-d3, respectively, is on the order of kH/kD = 2.25 (46), i.e. close to our experimentally obtained product isotope effect for 1-propanol. (iii) The oxygen incorporated into the product, 2-oxo-His, originated from molecular oxygen. On the basis of Equations 4-8, 12, and 13 the oxygen of the hydroxyl radical species will originate from molecular oxygen. A simple mechanism for the direct incorporation of the hydroxyl radical oxygen into 2-oxo-His is displayed in the Scheme 1 (reactions 14-16).


Scheme 1. Formation of 2-oxo-His.
[View Larger Version of this Image (13K GIF file)]


We note that the formation of 2-oxo-His via the displayed pathway requires the net addition of HO· at C-2 of the imidazole ring whereas ESR investigations on the addition of radiation chemically produced freely diffusable hydroxyl radicals to imidazole derivatives not complexed to transition metals have provided evidence for the addition of HO· at C-4 and C-5 (47). A preferred net addition of HO· to C-2 of His during the metal-catalyzed oxidation of hGH could be the result of metal-binding, the geometry of the metal-binding site, and the fact that a complexed rather than a free hydroxyl radical is the actual reactive oxygen species. In our tryptic map of hGH we identified low quantities of other oxidation products of His (T4ox1 and T4ox2) most likely representing products of the net addition of HO· at other positions of the imidazole ring such as C-4 and C-5. In this regard we note that the generation of free and/or complexed hydroxyl radicals has also been proposed for the reaction of hydrogen peroxide with Cu(I) superoxide dismutase (44, 48, 49) where 2-oxo-His was a major product of the modification of His118 (32) located at the metal-binding site.

Biological Significance

The binding of hGH to the hGH receptor is largely Zn2+-independent whereas binding to the human prolactin (hPRL) receptor requires Zn2+ and an intact metal-binding site of hGH (14). When each of the three amino acids involved in metal binding, i.e. His18, His21, and Glu174, was individually mutated to Ala, there was a 3,000-12,000-fold decrease of affinity of the respective hGH mutants to a recombinant protein containing the hGH-binding sequence of the hPRL receptor (hPRL binding protein, hPRLbp) (14). In contrast, these hGH mutants showed less significant change of binding to a recombinant protein containing the hGH-binding sequence of the hGH receptor (hGH-binding protein) (14). Thus, it appears that the chemical modification of a metal-binding side chain to a non-metal-binding side chain of only one amino acid of the metal-binding site is sufficient for the change of the affinity to hPRL-binding protein. We can reasonably assume that the 2-oxo tautomer of 2-oxo-His has a significantly lower affinity for metals as compared with a native imidazole ring due to the lack of the pyridine nitrogen. Therefore, we expect that the modification of His to 2-oxo-His in hGH may result in a decrease of the affinity to hPRLbp but not to hGH-binding protein, but this remains to be demonstrated in future experiments.


FOOTNOTES

*   This work was supported by the Association for International Cancer Research (AICR), National Institutes of Health Grant PO1AG12993, and a New Investigator Award from Eli Lilly & Company (to Ch. S.).The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
   To whom correspondence should be addressed. Tel: 913-864-4880; Fax: 913-864-5736; E-mail: schoneich{at}smissman.hbc.ukans.edu.
1   The abbreviations used are: ROS, reactive oxygen species; AXC, anion exchange chromatography; CXC, cation exchange chromatography; ESI-MS, electrospray ionization mass spectrometry; hGH, human growth hormone; MALDI-TOF MS, matrix-assisted laser desorption ionization-time of flight mass spectrometry; MCO, metal-catalyzed oxidation; TTCN, 1,4,7-trithiacyclononane; hGH, human growth hormone; HPLC, high performance liquid chromatography; RP-HPLC, reverse phase-HPLC; MES, 4-morpholineethanesulfonic acid.
2   Sanaullah, Hungerbühler, H., Schöneich, Ch., Morton, M., Vander Velde, D. G., Wilson, G. S., Asmus, K.-D., and Glass, R. S. (1997) J. Am. Chem. Soc., in press.

ACKNOWLEDGEMENTS

We thank Drs. Martha Morton and David Vander Velde for NMR analysis, Dr. Todd D. Williams for the assistance in product identification by mass spectroscopy, and Dr. James Q. Oeswein and Brian L. Miller for helpful discussions. We also thank the reviewer for valuable suggestions on the manuscript.


REFERENCES

  1. Halliwell, B., and Gutteridge, J. M. C. (1989) Free Radicals in Biology and Medicine, 2nd Ed., Clarendon Press, Oxford
  2. Stadtman, E. R. (1990) Free Rad. Biol. Med. 9, 315-325 [CrossRef][Medline] [Order article via Infotrieve]
  3. Stadtman, E. R. (1992) Science 257, 1220-1224 [Medline] [Order article via Infotrieve]
  4. Smith, C. D., Carney, J. M., Tatsumo, T., Stadtman, E. R., Floyd, R. A., and Markesberry, W. A. (1992) N. Y. Acad. Sci. 663, 110-119
  5. Sohal, R. S., and Dubey, A. (1994) Free Rad. Biol. Med. 16, 621-626 [CrossRef][Medline] [Order article via Infotrieve]
  6. Sohal, R. S., Sohal, B. H., and Orr, W. C. (1995) Free Rad. Biol. Med. 19, 499-504 [CrossRef][Medline] [Order article via Infotrieve]
  7. Benzi, G., and Moretti, A. (1995) Free Rad. Biol. Med. 19, 77-101 [CrossRef][Medline] [Order article via Infotrieve]
  8. Cabiscol, E., and Levine, R. L. (1995) J. Biol. Chem. 270, 14742-14747 [Abstract/Free Full Text]
  9. Sohal, R. S., and Weindruch, R. (1996) Science 273, 59-63 [Abstract]
  10. Manning, M. C., Patel, K., and Borchardt, R. T. (1989) Pharm. Res. 6, 903-918 [CrossRef][Medline] [Order article via Infotrieve]
  11. Cleland, J. L., Powell, M. F., and Shire, S. J. (1993) Crit. Rev. Ther. Drug Carrier Syst. 10, 307-377 [Medline] [Order article via Infotrieve]
  12. Stadtman, E. R. (1993) Annu. Rev. Biochem. 62, 797-821 [CrossRef][Medline] [Order article via Infotrieve]
  13. Stadtman, E. R., and Oliver, C. N. (1991) J. Biol. Chem. 266, 2005-2008 [Free Full Text]
  14. Cunningham, B. C., and Wells, J. A. (1991) Proc. Natl. Acad. Sci. U. S. A. 88, 3407-3411 [Abstract]
  15. Levine, R. L. (1983) J. Biol. Chem. 258, 11823-11827 [Abstract/Free Full Text]
  16. Levine, R. L. (1983) J. Biol. Chem. 258, 11828-11833 [Abstract/Free Full Text]
  17. Farber, J. M., and Levine, R. L. (1986) J. Biol. Chem. 261, 4574-4578 [Abstract/Free Full Text]
  18. Liaw, S.-H., Villafranca, J. J., and Eisenberg, D. (1993) Biochemistry 32, 7999-8003 [Medline] [Order article via Infotrieve]
  19. Uchida, K., and Kawakishi, S. (1993) FEBS Lett. 332, 208-210 [CrossRef][Medline] [Order article via Infotrieve]
  20. Cunningham, B. C., Mulkerrin, M. G., and Wells, J. A. (1991) Science 253, 545-548 [Medline] [Order article via Infotrieve]
  21. Pearlman, R., and Bewley, T. A. (1993) in Stability and Characterization of Protein and Peptide Drugs: Case Histories (Wang, Y. J., and Pearlman, R., eds), pp. 1-58, Plenum Press, New York
  22. USP Pharmacopeial Forum (1989) The United States Pharmacopeia (1990), 22nd Ed., p. 1246, Mack Printing Co, Easton, PA
  23. Giulivi, C., and Davies, K. J. A. (1994) Methods Enzymol. 233, 363-371 [Medline] [Order article via Infotrieve]
  24. De Vos, A. M., Ultsch, M., and Kossiakoff, A. A. (1992) Science 255, 306-312 [Medline] [Order article via Infotrieve]
  25. Davies, M. J., Fu, S., and Dean, R. T. (1995) Biochem. J. 305, 643-649 [Medline] [Order article via Infotrieve]
  26. Hiller, K.-O., Masloch, B., Göbl, M., and Asmus, K.-D. (1981) J. Am. Chem. Soc. 103, 2734-2743
  27. Bobrowski, K., and Holcman, J. (1987) Int. J. Radiat. Biol. 52, 139-144
  28. Bobrowski, K., and Holcman, J. (1989) J. Phys. Chem. 93, 6381-6387
  29. Bobrowski, K., Schöneich, C., Holcman, J., and Asmus, K.-D. (1991) J. Chem. Soc. Perkin Trans. 2, 353-362
  30. Schöneich, Ch, Zhao, F., Madden, K. P., and Bobrowski, K. (1994) J. Am. Chem. Soc. 116, 4641-4652
  31. Uchida, K., and Kawakishi, S. (1986) Biochem. Biophys. Res. Commun. 138, 659-665 [Medline] [Order article via Infotrieve]
  32. Uchida, K., and Kawakishi, S. (1994) J. Biol. Chem. 269, 2405-2410 [Abstract/Free Full Text]
  33. Li, S., Nguyen, T., Schöneich, Ch, and Borchardt, R. T. (1995) Biochemistry 34, 5762-5772 [Medline] [Order article via Infotrieve]
  34. Turner, C., Cary, P. D., Grego, B., Hearn, M. T. W., and Chapman, G. E. (1983) Biochem. J. 213, 107-113 [Medline] [Order article via Infotrieve]
  35. Abildgaard, F., Jørgensen, A. M. M., Led, J. J., Christensen, T., Jensen, E. B., Junker, F., and Dalbøge, H. (1992) Biochemistry 31, 8587-8596 [Medline] [Order article via Infotrieve]
  36. Zhao, F., Yang, J., and Schöneich, Ch (1996) Pharm. Res. 13, 931-938 [Medline] [Order article via Infotrieve]
  37. Tomita, M., Irie, M., and Ukita, T. (1969) Biochemistry 8, 5149-5160 [Medline] [Order article via Infotrieve]
  38. Mulkerrin, M. G., and Cunningham, B. C. (1993) in Protein Folding: in Vivo and in Vitro; ACS Symposium Series 526 (Cleland, J. L., ed), pp. 240-253, American Chemical Society, Washington, D. C.
  39. Wicar, S., Mulkerrin, M. G., Bathory, G., Khundkar, L. H., and Karger, B. L. (1994) Anal. Chem. 66, 3908-3915 [Medline] [Order article via Infotrieve]
  40. Willam, S., Ultsch, M., De Vos, A. M., and Kossiakoff, A. A. (1994) Nature 372, 478-481 [CrossRef][Medline] [Order article via Infotrieve]
  41. Shaw, K., and Espenson, J. H. (1968) Inorg. Chem. 7, 1619-1622
  42. Ershov, B. G., Janata, E., Michaelis, M., and Henglein, A. (1991) J. Phys. Chem. 95, 8996-8999
  43. Kitajima, N., and Moro-oka, Y. (1994) Chem. Rev. 94, 737-757
  44. Cabelli, D. E., Allen, D., Bielski, B. H. J., and Holcman, J. (1989) J. Biol. Chem. 264, 9967-9971 [Abstract/Free Full Text]
  45. Asmus, K. D., Möckel, H., and Henglein, A. (1973) J. Phys. Chem. 77, 1218-1221
  46. Buxton, G. V., Greenstock, C. L., Helman, W. P., and Ross, A. B. (1988) J. Phys. Chem. Ref. Data 17, 513-886
  47. Samuni, A., and Neta, P. (1973) J. Phys. Chem. 77, 1629-1635
  48. Sato, K., Akaike, T., Kohno, M., Ando, M., and Maeda, H. (1992) J. Biol. Chem. 267, 25371-25377 [Abstract/Free Full Text]
  49. Yim, M. B., Chock, P. B., and Stadtman, E. R. (1993) J. Biol. Chem. 268, 4099-4105 [Abstract/Free Full Text]

©1997 by The American Society for Biochemistry and Molecular Biology, Inc.