THE JOURNAL OF BIOLOGICAL CHEMISTRY (0 1992 by The American Society for Biochemistry and Molecular Biology, Inc

Oxidative Fragmentation of Collagen and by Cu(II)/H202 CONVERSION OF PROLINE RESIDUE TO 2-PYRROLIDONE*

Val. 267, No. 33, Issue of November 25, pp. 23646-231351,1992 Printed in U.S.A.

Prolyl Peptide

(Received for publication, July 13, 1992)

Yoji Kato, Koji Uchida, and Shunro KawakishiS . .

Nagoya 464-01, Japan-

Oxidative degradation of collagen and the model pep- tides by Cu(II)/H202 has been studied. The depolymer- ization of collagen was predominantly observed by use of gel filtration chromatography. Polyproline was used as a model for collagen, and the oxidative modification was examined by amino acid analysis. Glutamic acid and y-aminobutyric acid were identified in the hydrol- ysates of oxidized polyproline. The formation of glu- tamic acid was reduced by treatment with NaBH4. The model peptide, (Pro-Pro-Gly)l,, was also degraded by Cu(II)/H202, and a new N-terminal glycine was gener- ated in proportion to the reaction time. Hydroxyl rad- ical scavengers show only partial inhibition of the degradation of (Pro-Pro-Gly)lo. In order to estimate the fragmentation mechanism, we used N-tert-butox- ycarbonyl (Boc)-L-prolylglycine as a model for colla- gen and (Pro-Pro-Gly)lo. The degradation products were isolated and characterized. Then N-tert-Boc-2- pyrrolidone, which provides y-aminobutyric acid by acid hydrolysis, was identified. The formation of a 2- pyrrolidone compound from oxidized Boc-L-prolylgly- cine is direct evidence for the scission of the peptide bond. The time-dependent formation of N-tert-Boc-2- pyrrolidone and liberation of glycine from N-tert-Boc- L-prolylglycine exposed to Cu(II)/H202 was observed. These results suggest that the cleavage of the peptide bond (Pro-Gly) was caused by oxidation of the proline residue, which led to the formation of the 2-pyrroli- done compound. We confirmed that proline oxidation leads to the fragmentation of proteins, accompanied by the formation of a 2-pyrrolidone structure.

Active oxygen is thought to be one of the major factors in aging and many human diseases. These reactive species (uiz. ’ OH, ferry1 ion) can attack biomolecules such as nucleic acids, lipids, and proteins. Many studies of protein oxidation with active oxygens have been done. It has been reported that the aggregation and fragmentation of proteins occur with modi- fications of amino acids when proteins are exposed to oxygen radicals. Amici et al. (1) reported the selective damage of His, Lys, Arg, and Pro in a protein and amino acid homopolymers by metal-catalyzed oxidation. We have investigated the selec-

* The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

$ T o whom correspondence should be addressed Laboratory of Chemistry of Plant Products, Department of Food Science & Tech- nology, Nagoya University, Furo-cho, Chikusa-ku, Nagoya 464-01, Japan. Tel.: 052-781-5111 (ext. 6323); Fax: 052-781-4447.

tive modification of histidine by Cu(II)/ascorbate (2) and tryptophan by Fe(II)/EDTA/ascorbate (3) in the bovine serum albumin. It has been considered by Wolff et al. (4) that proline oxidation contributes to the fragmentation of a protein because of the formation of a pyroglutamyl structure. Schues- sler et al. (5) proposed that bovine serum albumin exposed to y-rays was cleaved at the position of proline residues. How- ever the relationship between proline oxidation and fragmen- tation is unknown.

Collagen is a major structural protein and the main com- ponent of the extracellular matrix in the human body. Colla- gen is characterized by a triple helical structure and a repeat- ing sequence of Gly-X-Y (X and Y are often proline and hydroxyproline, respectively). Age-related changes in collagen are thought to be associated with increased cross-linking between collagen molecules. Kano et al. (6) reported the aggregation of collagen treated by Cu(II)/ascorbate, whereas many workers reported that collagen is degraded by active oxygen species (7-10).

We have previously found that collagen is degraded by metal-catalyzed oxidation, especially Cu(II)H202, and signif- icant loss of the proline residue in collagen occurs (11). The oxidation of collagen and its model peptides by Cu(1I)/H2O2 was investigated in detail. The fragmentation of collagen molecules was observed predominantly. The model peptide was cleaved at the proline position, and then a 2-pyrrolidone residue was newly generated. We confirmed that the oxidation of proline residues leads to the cleavage of the peptide bond.

EXPERIMENTAL PROCEDURES

Materials-Collagen IV (from human placenta), catalase (from bovine liver), polyproline ( M , lo3 to 3 X lo4), and L-prolylglycine were purchased from Sigma. Superoxide dismutase (from beef eryth- rocytes) was purchased from ICN Biomedicals, Inc. tert-Butyl ,944, 6-dimethylpyrimidin-2-yl) thiolcarbonate and 3,4-dehydro-~-proline were obtained from Aldrich. (Pro-Pro-Gly),o.9H20 was purchased from Peptide Institute, Inc. Dansyl chloride, sodium borohydride (NaBH,), dimethyl sulfoxide (Me2SO), histidine, EDTA, 2-pyrroli- done and di-tert-butyl dicarbonate were obtained from Wako Pure Chemical Industries Ltd. (Osaka). All other reagents were of the highest grade commercially available.

Preparation of tert-Butoxycarbonyl (Boc)’ Deriuatiues-N-tert-Boc- L-prolylglycine was prepared by tert-butoxycarbonylation of L-pro- lylglycine with tert-butyl S-(4, 6-dimethylpyrimid~n-2-yl) thiolcar- honate according to the method of Nagasawa et al. (14). After tert- butoxycarbonylation, the derivative was crystallized with ethyl ace- tate. The N-tert-Boc-L-prolylglycine was identified by amino acid analysis, FAB-MS, and ‘H NMR. The tert-butoxycarbonylation of 2-

’ The abbreviations used are: Boc, tert-butoxycarbonyl; Z, benzy- loxycarbonyl; GABA, y-aminobutyric acid; dansyl, 5-dimethylami- nonaphthalene-1-sulfonyl; HPLC, high performance liquid chroma- tography; FAB-MS, fast atom bombardment-mass spectroscopy.

From the Laboratory of Chemistry of Plant Products, Department of Food Science & Technology, Nagoya University,

23646

Oxidative Fragmentation of Collagen and Prolyl Peptide

pyrrolidone with di-tert-butyl dicarbonate was performed as described by Grehn et al. (15). After the reaction, the solution was concentrated, and further purification was performed by reversed-phase HPLC on a Develosil ODS-10 (20 X 250 mm). The derivative was eluted with a solution of 0.1% trifluoroacetic acid/methanol (1:l) a t a flow rate of 5.0 ml/min. The synthetic product was identified as N-tert-Boc-2- pyrrolidone by amino acid analysis, 'H NMR, and "'C NMR.

Oxidation of Protein and Polypeptide-A reaction mixture (1.0 ml) containing 0.5 mg of collagen, 0.5 mg of polyproline, or 0.1 mg of (Pr~-Pro-Gly)~,, and 0.05 mM CuSO, in 0.1 M phosphate buffer (pH 7.4) was incubated in air a t 37 "C. The reaction was started by the addition of 100 pI of 50 mM hydrogen peroxide, which was freshly prepared. After incubation, the reaction was stopped by adding EDTA to a final concentration of 0.1 mM.

Oxidation of Boc-L-prolylglycine and Isolation of the Oxidation Products-The reaction mixture (1 ml) containing 1 mM BOC-L- prolylglycine and 0.05 mM CuSO, in 0.1 M phosphate buffer (pH 7.4) was incubated at 37 "C. The reaction was started and stopped as described above.

For the isolation of oxidation products, 500 ml of the reaction mixture was used. The solution was incubated at 37 "C for 24 h in the presence of 5 mM hydrogen peroxide and 0.05 mM CuS04, and then the reaction was stopped by addition of EDTA. The reaction mixture (500 ml) was freeze-dried, extracted with methanol to remove a large amount of inorganic salts, and then evaporat,ed in uacuo. The extract was dissolved in water/methanol solution and submitted to HPLC connected to a Develosil ODS-10 column (20 X 250 mm). The products were eluted with a solution of 0.1% trifluoroacetic acid/ methanol (3:2) a t a flow rate of 5 ml/min, and each peak was isolated.

Gel Filtration-An aliquot of an oxidized collagen solution was used for gel filtration, which was performed on a TSK-gel G3000SWxI. (7.8 X 300 mm) (Tosoh Co.). The column was equilibrated in 0.1 M phosphate buffer (pH 7.0), containing 0.1 M NaC1, and eluted by this buffer solution at a flow rate of 0.8 ml/min. The elution was monitored by UV absorbance a t 215 nm.

Sodium Dodecyl Sulfate-Polyacrylamido Gel Electrophoresis-The polyacrylamido gel electrophoresis in the presence of sodium dodecyl sulfate was performed as described before ( l l ) , according to the method of Laemmli (12), using a vertical slab gel.

The Reduction of Natiue and Oxidized Polyproline by Treatment with NaBH,-The reaction was stopped by adding EDTA. The so- lution was divided into two portions; one was used for reduction and the other (control) was not treated. NaBH4 was added until the pH of the reaction mixture was adjusted to 9.0. Then the solution was held at room temperature for 1 h. The reduced solution was acidified by adding acetic acid in order to decompose the residual NaBH,. The reduced and untreated solutions were hydrolyzed as described in amino acid analysis.

Amino Acid Analysis-Amino acid analysis was performed with a JEOL JLC-300 amino acid analyzer. The samples were prepared as follows. After incubation, the solutions were freeze-dried and then hydrolyzed with 6 N HCl in uacuo for 24 h a t 110 "C. The hydrolysates were concentrated, dissolved in citrate buffer (pH 2.2), and then submitted for amino acid analysis.

HPLC Determination of (Pro-Pro-Gly)lo-(Pro-Pro-Gly)lo was de- termined by reversed-phase HPLC on a Develosil ODS-5 column (4.6 X 250 mm) (Nomura Chemical Co., Ltd.). An aliquot of a solution was applied to the column equilibrated in 0.1% trifluoroacetic acid. The peptide was eluted with a linear gradient of acetonitrile (0.75%/ min) at a flow rate of 0.8 ml/min, and the elution was monitored by absorbance at 215 nm. The residual substrate was determined from the chromatographic peak height.

N-terminal Analysis of Native and Oxidized (Pro-Pro-GILy)lo by the Dansyl Method-The sample (0.8 ml) was passed through a CIR cartridge (SEP-PAK"). The cartridge was washed with 2 ml of dis- tilled water in order to remove the inorganic salts. Then the sample was eluted with 5 ml of acetonitrile, and the eluent was collected and concentrated in uacuo. The sample was dissolved with 1 ml of distilled water and then filtered (0.45 pm). The eluent was freeze-dried and then used for derivatizat.ion. The derivatization of the N-terminal amino acids was performed according to the method of Tapuhi et al. (13). The sample was dissolved in 50 pl of 40 mM LiCO,, which had been adjusted to pH 9.5 by HCI, 25 pl of dansyl chloride (1.5 mg/l ml of acetonitrile) was added, and the sample was sealed in a container and then mixed. The solution was held a t room temperature for 30 min in the dark. The reaction was stopped by addition of 5 p1 of 2% monomethylamine hydrochloride and stored for 5 min. The solution was evaporated and hydrolyzed with 6 N HCI a t 110 "C for 24 h in

23647

uucuo. Then the sample was evaporated in uacuo and extracted with 100 pl of ethyl acetate, which had been saturated with water, in order to remove excess dansyl-OH. The extract was evaporated, dissolved with 50 pl of water/methanol(2:1), and submitted for reversed-phase HPLC on a Develosil ODS-5 column (4.6 X 250 mm). The derivatives were eluted with a linear gradient of a two-solvent system a t a flow rate of 0.8 ml/min. Solvent A (50 mM acetate/sodium acetate buffer (pH 4.18)/tetrahydrofuran (19:l)) and solvent B (acetonitrile/tetra- hydrofuran (9:l)) were used for the gradient. The gradient employed was as follows: isocratic at 100% A for 3 min, 100 to 60% A in 60 min, 60 to 0% A in 2 min, 0 to 100% A in 3 min. The detection of the derivatives was performed by fluorescence detector TOY0 SODA FS8000 (excitation 340 nm, emission 530 nm). The identification of the derivatives was performed by comparison with a dansyl derivative of authentic proline or glycine. The peak area was determined by use of a Shimadzu Chromatopac CR-3A.

Determination of Boc-L-prolylglycine, Boc-2-pyrrolidone, and Gly- cine in the Reaction. Mixture-Boc-L-prolylglycine and Boc-2-pyrrol- idone were determined by reversed-phase HPLC on a Develosil ODS- 5 column (4.6 X 250 mm). An aliquot of the reaction mixture was applied to the column equilibrated with 0.1% trifluoroacetic acid/ methanol (5:4). The peptide and the products were eluted a t a flow rate of 0.8 ml/min, the elution being monitored by absorbance a t 215 nm. These amounts were determined by comparison with synthetic Boc-L-prolylglycine or Boc-2-pyrrolidone.

The liberation of glycine from oxidized Boc-L-prolylglycine was determined as follows. The reaction mixture (1.1 ml), after the reaction had been stopped by adding EDTA, was acidified by addition of 1 N HC1 (100 p l ) and applied to a cation-exchange column (10 X 50 mm), which had a 4.0-ml cation resin bed of Dowex 50W X 8 (H+), 100-200 mesh. The acidified solution was placed onto the column, and the resin was washed with distilled water until the pH of the effluent was neutral. Then glycine was eluted from the column with a 20-ml volume of 7 N NH,OH eluent. The eluate was collected and concentrated. The sample was dissolved in citrate buffer (pH 2.2) and applied to an amino acid analyzer, JLC-300.

RESULTS

Degradation of Collagen by Cu(II)/H202-Collagen solution (0.5 mg/ml) was incubated with Cu(JI)/H2O2 at 37 "C. The time-dependent reaction of the protein with the Cu(II)/H202 system was examined by HPLC on a TSK-gel G3000SW~,. column. As shown in Fig. 1, the high molecular weight parts of the collagens decreased, and then the lower molecular fragments increased, which was proportional to the incubation time. The formation of an aggregated collagen could not be determined because the elution volume of collagen is close to the void volume of the column. The fragmentation of collagen

I

I 1 Ohr

L

I 4 hr

I 8 hr

24 hr

0 5 10 15 0 5 10 15

Retention Time (min)

FIG. 1. Time-dependent degradation of collagen oxidized by Cu(1I)/H2O2. Collagen a t a concentration of 0.5 mg/ml was incubated with 50 p M Cu(I1) and 5 mM H20, for the times indicated in the chromatograms. The time-dependent changes in collagen were ex- amined by HPLC on TSK-gel G3000SWxL (7.8 X 300 mm) as de- scribed under "Experimental Procedures."

23648 Oxidative Fragmentation of Collagen and Prolyl Peptide

by Cu(II)/H202 was observed predominantly in our system. To access the contribution of oxidizing species to collagen

oxidation, we examined the effects of inhibitors, such as hydroxyl radical scavengers, chelating agents for metal ions, and antioxidant enzymes on the degradation of collagen using SDS-polyacrylamido gel electrophoresis. As shown in Fig. 2, catalase, EDTA, or histidine inhibited the degradation. A hydroxyl radical scavenger, mannitol, failed to inhibit the degradation of the collagen. Interestingly, Me2S0, which is also known as a hydroxyl radical scavenger, inhibited in part. A site-specific mechanism for the degradation of collagen was assumed, since mannitol had no effect on the oxidation.

Oxidation of Polyproline-We recently reported that proline residues in collagen molecules were decreased by metal-cata- lyzed oxidation systems (11). We used a polyproline, as a model of proline residues in a collagen molecule, and the oxidation of proline was investigated. The polyproline was exposed to Cu(II)/H202 at 37 “C for 24 h. The oxidized solu- tion was divided into two portions; one was reduced with NaBH4, and the other (control) was not treated (see “Exper- imental Procedures”). The two samples were hydrolyzed with 6 N HC1 in uucuo and then analyzed by a JLC-300 amino acid analyzer. As shown in Fig. 3A, the unreduced sample gener- ated considerable amounts of glutamic acid and unknown product X in the acid hydrolysates. X was very close to the elution position of hydroxyproline. The peak of X disappeared on treatment with NaBH,. X may be dehydroproline, which

catalase +

FIG. 2. Effects of inhibitors on oxidation of collagen with Cu(II)/H202. Collagen a t a concentration of 0.5 mg/ml was incubated with 50 pM Cu(I1) and 5 mM H202 for 30 min in the presence or absence of various active oxygen scavengers. The polyacrylamido gel electrophoresis in the presence of SDS was performed as described (11). The new band between the major collagen band and the dye front in lane 3 derived from catalase itself, which was added as an inhibitor (arrow). Lane 1, control; lane 2, Cu(II)/H202, 30 min; lane 3, Cu(II)/H202/catalase (500 units/ml), 30 min; lane 4, Cu(II)/H202/ EDTA (0.1 mM), 30 min; lane 5, Cu(II)/H202/mannitol (10 mM), 30 min; lane 6, Cu(I1)/H2O2/histidine (10 mM), 30 min; lane 7, Cu(II)/ H2O2/Me2SO (10 mM), 30 min.

FIG. 3. Chromatograms of amino acid analysis of polyproline oxidized by Cu(II)/HzO2. Polyproline at a con- centration of 0.5 mg/ml was incubated with 50 pM Cu(I1) and 5 mM H2O2 for 24 h. The reduced sample was prepared with treatment by NaBH, as described under “Experimental Procedures.” A, hydrolysates of nonreduced sample; B, hydrolysates of reduced sample.

e I) U

has been detected in a solution of an oxidized proline deriva- tive (16), because of the same retention time as authentic dehydroproline. However, the structural identification of X was not done. In addition, the peak of X completely disap- peared after dialysis of the oxidized sample (data not shown). This result suggests that the precursor of X is too labile (maybe hydroperoxide) or the formation of X by acid hydrol- ysis is due to contamination (maybe hydrogen peroxide).

Glutamic acid decreased on treatment with NaBH, (Fig. 3B). This result indicates that one of the precursors of glu- tamic acid would be y-glutamyl semialdehyde. Both y-glu- tamyl semialdehyde and pyroglutamyl residue were consid- ered by Amici et al. (1) as oxidation products of proline residues. After the reduction, two new products were detected at the position of 11 and 20 min by the amino acid analyzer. We considered that these two peaks, which were produced by treatment with NaBH4, may be 5-hydroxy-2-aminovaleric acid and 5-chloroaminovaleric acid, respectively, on the basis of the eluted positions as shown by Amici et al. (1). However, large parts of the precursors are probably a pyroglutamyl residue, because most precursors were not reduced with NaBH,.

We also detected a considerable amount of y-aminobutyric acid (GABA) in the hydrolysates of oxidized polyproline. The generation of GABA was possibly derived from the 2-pyrrol- idone residue, which is converted to GABA by acid hydrolysis. The formation of 2-pyrrolidone suggests the fragmentation of the peptide bond on the basis of its structure.

Oxidation of Collagen-like Peptide by Cu(II)/H2O2”In an effort to clarify the degradation mechanism, a collagen-like peptide, (Pr~-Pro-Gly)~o, was treated with Cu(II)/H202. The degradation was determined by reversed-phase HPLC and amino acid analysis of the reaction mixture. As shown in Fig. 4, the time-dependent loss of (Pro-Pro-Gly),, was observed in the presence of H202 and Cu(I1) but not in the presence of H202 or Cu(I1) only. During the incubation with Cu(II)/H202, proline in the peptide decreased progressively (Fig. 5). The time-dependent formation of glutamic acid and GABA was observed in the hydrolysates of oxidized peptide.

To investigate the degradation mechanism and its oxidizing species in the Cu(II)/H202 system, the residue of (Pro-Pro- Gly),, by incubation with Cu(II)/H202 was determined in the absence or presence of various radical scavengers and antiox- idant proteins. As shown in Fig. 6, hydroxyl radical scavengers mannitol and ethanol show only partial inhibition (20-25%), whereas Me2S0 inhibited degradation somewhat (about 50%). Diazabicyclo[2.2.2]octane and sodium azide, which are known to be IO2 scavengers, also inhibited the degradation slightly (20-30%). Histidine, EDTA, and diethylenetriaminepentaa- cetic acid exhibited considerable inhibition. These inhibitory effects on the degradation were probably based on the ability

_i B (reduced)

0 3 6 0 12 15 18 21 24 27 30 0 3 6 9 12 15 18 2l 24 27 30

Retention lime (min) Uuention Tim (rnin)

Oxidative Fragmentation of Collagen and Prolyl Peptide 23649

0.12

0.10

0.08

0.06

0.04

0.02

0.00

Time (hr)

FIG. 4. Time-dependent disappearance of (Pro-Pro-Gly)lo during incubation with Cu(II)/H202. (Pro-Pro-Gly),” a t a concen- tration of 0.1 mg/ml was incubated with 50 p~ Cu(I1) and 5 mM Hz02 for the times indicated on the abscissa. The concentration of (Pro- P r ~ - G l y ) ~ , , was determined by reversed-phase HPLC as described under “Experimental Procedures.”

600 R Proline

300

Glycine

I t P I

0 2 4 6 8 1 0 0 2 4 6 8 1 0

Time (hr) Time (hr)

FIG. 5. Time-dependent changes in amino acid composition of (Pro-Pro-Gly),, during incubation with Cu(II)/H202. (Pro- Pro-GIy)lll at a concentration of 0.1 mg/ml was incubated with 50 p~ Cu(I1) and 5 mM H202 for the times indicated on the abscissa. The changes in amino acid composition were determined as indicated under “Experimental Procedures.” Left panel, changes in proline and glycine. Right panel, changes in glutamic acid and 7-aminobutyric acid.

Mannitol(l0 mM)

Histidine(l0 mu) Elhanol(10 mM)

DABCO(l0 mM) Sodium azide(l0 mMI

MeSO(10 mM)

EDTA(O.1 mM)

DTPA(O.1 mM1

SOD(100 uniUml1

Denatured SOD

C a t a l a r q w unitlml)

Denatured catalase

-20 0 20 40 60 80 100

inhibition ?A)

FIG. 6. Effects of inhibitors on oxidation of (Pro-Pro-Gly),,, with Cu(II)/H202. (Pro-Pro-Gly)lo a t a concentration of 0.1 mg/ml was incubated with 50 p~ Cu(I1) and 5 mM H,Oz for 90 min in the absence or presence of various active oxygen scavengers at the con- centrations indicated. The inhibition was calculated on the basis of the residual (Pro-Pro-Gly),, determined by reversed-phase HPLC as described under “Experimental Procedures.” DABCO, diazabicy- clo[2.2.2]octane; DTPA, diethylenetriaminepentaacetic acid; SOD, superoxide dismutase.

t o chelate metal ion. Contribution of hydrogen peroxide to the oxidation of (Pro-Pro-Gly)lo was confirmed, since catalase inhibited the degradation completely. Presumably, the super- oxide anion radical did not participate in the oxidation of the peptide because superoxide dismutase had no effect on the degradation. The effects of the concentration of mannitol on the degradation were investigated in detail (Fig. 7). Concen-

Concentration of mannitol (mM)

FIG. 7. Effects of concentration of mannitol on oxidation with Cu(II)/H202. (Pro-Pro-Gly)lo a t a concentration of 0.1 mg/ml was incubated with 50 p~ Cu(I1) and 5 mM H202 for 90 min in the presence of mannitol (0-100 mM). The inhibition was calculated on the basis of the residual (Pro-Pro-Gly)lo determined by reversed- phase HPLC as described under “Experimental Procedures.”

0 0.5 1 2 3

Time (hr)

FIG. 8. Time-dependent changes in N-terminal amino acids of (Pro-Pro-Gly),, during incubation with Cu(II)/H202. (Pro- Pro-Gly),, a t a concentration of 0.1 mg/ml was incubated with 50 p~ Cu(I1) and 5 mM H202 for the times indicated on the abscissa. The determination of N-terminal amino acids was performed by the dansyl method as described under “Experimental Procedures.” B, proline (%I; U, glycine (%).

tration-dependent inhibition was observed proportionally from 0 to 10 mM. However, further addition of mannitol to the solution had little effect on the degradation. The partial protection of peptide in the degradation by hydroxyl radical scavengers and the perfect protection of peptide in the deg- radation by metal ion chelators suggest that a site-specific reaction was a major factor in the degradation. Stadtman et al. (17) have suggested that a “caged” process occurs in the Fe(II)/H202 system in the presence of amino acid and bicar- bonate.

To estimate the degradation of (Pro-Pro-Gly),o, we inves- tigated the N-terminal analysis of native and oxidized (Pro- Pro-Gly)lo by the dansyl method. Although the native N- terminal of (Pro-Pro-Gly),, is proline, a time-dependent ap- pearance of N-terminal glycine was observed (Fig. 8). This result suggests that cleavage of the peptide bond in (Pro-Pro- Gly)lo occurred to a considerable extent during incubation with the Cu(II)/H202 system.

Isolation and Characterization of Products Derived from Oxidized Boc-L-prolylglycine-To determine the precursors of the amino acids newly detected in the acid hydrolysates of oxidized polyproline and (Pro-Pro-Gly),o, the oxidation of Boc-L-prolylglycine was investigated. Fig. 9 shows the HPLC profile of the reaction mixture after 24 h of incubation with Cu(II)/H202. The reaction of Boc-L-prolylglycine with Cu(II)/ H 2 0 n provided mainly two products (BPG-1 and BPG-2). These two products were isolated and characterized. The

23650 Oxidative Fragmentation of Collagen and Prolyl Peptide

u 0 5 10 15

Retention Time (rnin)

FIG. 9. HPLC profile of Boc-L-prolylglycine and oxidation products with Cu(II)/H202. Boc-L-prolylglycine at a concentration of 1 mM was incubated with 50 PM Cu(I1) and 5 mM HzO, for 24 h. An aliquot of the solution was submitted for reversed-phase HPLC on a Develosil ODs-5 column (4.6 X 250 mm) as described under “Experimental Procedures.”

HnN7COOH 0

Nn.t-Boc-L-prolyIglycine ~.t-Boc-2.pyrrolidone Glycine

(BPG-2)

.OHIO2

+ ~ ~ ~ - c n o n

Nn-t-Boc-L-pyroglutamylglycine

(BPG-1)

FIG. 10. Proposed structures for oxidation products of Boc- L-prolylglycine with Cu(II)/H202. Structural data of BPG-1 and BPG-2 are as follows. BPG-1: ‘H NMR, 6(ppm) 1.32-1.55 (9H, tert-

NCH); BPG-1 was converted to both glutamic acid and glycine by acid hydrolysis; FAB-MS 287 (M + 1). BPG-2: ‘H NMR, b(ppm) 1.43-1.50 (9H, tert-butyl), 1.93-2.09 (2H, CH,), 2.46-2.56 (2H, CH,), 3.70-3.81 (2H, CH,); BPG-2 was converted to GABA by acid hydrol- ysis.

proposed structures of the isolated products are shown in Fig. 10.

Identification of BPG-1 has been performed by amino acid analysis, ’H NMR, and FAB-MS. BPG-1 was converted to glutamic acid and glycine by acid hydrolysis. The ‘H NMR spectrum of BPG-1 revealed the disappearance of the C-5 proton of proline, suggesting the presence of a pyroglutamyl structure. The FAB-MS spectrum of BPG-1 gave a peak of the (M + H)* ion, m/z 287. Recently, we reported the for- mation of pyroglutamyl compounds from Boc-L-prolylglycine and Boc-L-proline exposed to ultraviolet light (18). Conse- quently, BPG-1 was identified as N-tert-Boc-L-pyroglutamyl- glycine. The production of a pyroglutamyl compound indi- cates the oxidative modification of prolyl residues.

BPG-2 provided GABA by hydrolysis. We recently identi- fied Boc-2-pyrrolidone as an oxidation product of BOC-L-

butyl), 1.98-2.74 (4H, CHzCHz), 3.81-4.11 (2H, CHz), 4.80-4.85 ( lH,

prolylglycine by ultraviolet irradiation (18) and N-benzylox- ycarbonyl (Z)-2-pyrrolidone as an oxidation product of Z- proline during incubation with H202 in the presence of cop- per(I1) ion (19). Authentic 2-pyrrolidone compound is con- verted to GABA by acid hydrolysis. Then we presumed the presence of a 2-pyrrolidone structure in BPG-2. The ‘H NMR spectrum of the product was similar to those of Z-2-pyrroli- done and authentic 2-pyrrolidone and completely agreed with that of synthetic Boc-2-pyrrolidone, which was prepared by tert-butoxycarbonylation of 2-pyrrolidone with di-tert-dicar- bonate. Therefore, the structure of BPG-2 was estimated to be N-tert-Boc-2-pyrrolidone. In addition, the (M + H)’ ion of BPG-2 was undetected clearly in the FAB-MS spectrum. The result possibly derived from the nature of Boc-2-pyrrol- idone itself because the (M + H)’ ion of synthetic Boc-2- pyrrolidone was also undetected in the FAB-MS spectrum.

The formation of Boc-2-pyrrolidone from oxidized BOC-L- prolylglycine suggests that the scission of the peptide bond has occurred at the position of the prolyl residues in the peptide, followed by the liberation of glycine (Fig. 10). To be sure of this hypothesis, the formation of Boc-2-pyrrolidone and glycine in the reaction mixture was determined. The time-dependent loss of Boc-L-prolylglycine and the formation of Boc-2-pyrrolidone and glycine in the reaction mixture were observed (Fig. 11). These results suggest that the cleavage of the prolylglycine bond causes the production of Boc-2-pyrrol- idone and glycine. The formation of glycine was less than the 2-pyrrolidone formation. This indicates that not all of the formation of Boc-2-pyrrolidone leads to the liberation of glycine. This result suggests that the N terminus of glycine was blocked or the structure of glycine was destroyed at least in part. Anyway, the formation of Boc-2-pyrrolidone is derived from proline oxidation, which leads to the cleavage of the peptide bond.

DISCUSSION

The depolymerization of collagen was observed primarily by use of gel filtration chromatography, following oxidation with Cu(II)/H202 (Fig. 1). The results are in agreement with those of Curran et al. (7) who reported the degradation of collagen oxidized by Fenton’s reagent or ozone. We have previously investigated the oxidation of collagen by metal- catalyzed oxidation systems (Cu(II)/ascorbate, Fe(II)/EDTA/ ascorbate, Cu(II)/H202, and Fe(II)/EDTA/H20J (11). We have determined that the major band of native collagen dis- appeared using SDS-PAGE, and a significant loss of proline in the amino acid composition of collagen was the most predominant (11).

0 8 16 24 0 8 16 24

Time (hr) Time (hr)

FIG. 11. Time-dependent changes in Boc-L-prolylglycine, Boc-2-pyrrolidone, and glycine during incubation with Cu(II)/H202. Boc-L-prolylglycine at a concentration of 1 mM was incubated with 50 PM Cu(I1) and 5 mM HZ02 for the times indicated on the abscissa. These compounds were determined as described under “Experimental Procedures.”

Oxidative Fragmentation of Collagen and Prolyl Peptide 23651

In this experiment, we used polyproline as a model for collagen, and the oxidation products were investigated by amino acid analysis. Glutamic acid, GABA, and unknown product X were identified in acid hydrolysates of polyproline oxidized by Cu(II)/H202 (Fig. 3). The production of glutamic acid in acid hydrolysates of polyproline exposed to Cu(II)/ H20e has been reported previously by Cooper et al. (20). Amici et al. (1) have already supposed that the generation of glutamic acid by a metal-catalyzed oxidation system is derived from y- glutamyl semialdehyde and pyroglutamyl residue. These com- pounds provide glutamic acid by acid hydrolysis. The presence of a y-glutamyl semialdehyde structure in a solution of oxi- dized polyproline was confirmed by the fact that the peak area of glutamic acid decreased on treatment with NaBH4, which can reduce an aldehyde structure to alcohol. However, a considerable amount of unreducible precursor exists after the reduction. We assume that most of the precursor, which converts to glutamic acid, is a pyroglutamyl residue.

The model peptide, (Pro-Pro-Gly)lo, is rapidly degraded by Cu(II)/H202 (Fig. 4). The time-dependent increase in glu- tamic acid and GABA in accordance with the loss of proline in the hydrolysates of the peptide was observed (Fig. 5 ) . The degradation was partly inhibited by the addition of mannitol, ethanol, and Me2S0, which are known to be hydroxyl radical scavengers (Fig. 6). This result is in agreement with that of collagen (Fig. 2). The inhibitory effect of mannitol increased in proportion to the concentration from 0 to 10 mM (Fig. 7). However, further addition of mannitol had little effect on the degradation. This result suggests that a random 'OH attack such as ionizing radiation partially contributes to the degra- dation. In addition, the oxidative degradation of (Pro-Pro- Gly),,, by ionizing radiation was completely inhibited by hy- droxyl radical scavengers such as mannitol or ethanol.* A peptide(protein)-Cu-H202 complex or metal-oxo[Cu(III)-0] complex presumably participates in the degradation.

We previously isolated Z-2-pyrrolidone from the oxidation products of Z-proline (19). On the basis of the structure, the 2-pyrrolidone compound is probably generated by H-abstrac- tion of the a-carbon of proline, accompanied by the peptide bond scission. In this experiment, we isolated Boc-2-pyrroli- done from Boc-L-prolylglycine treated with Cu(II)/H202 (Fig. 10). The generation of Boc-2-pyrrolidone was accompanied by the liberation of glycine (Fig. 11). This result suggests that the appearance of a new N-terminal glycine from (Pro-Pro- Gly)lo exposed to Cu(II)/H202 was due to the oxidation of the proline residue, accompanied by the formation of a 2-pyrrol- idone compound (Fig. 8). The appearance of a considerable amount of glycine in the N-terminal sequence analysis of oxidized collagen by Cu(II)/H202 (11) or by the xanthine oxidase-hypoxanthine system has been reported (9). The frag- mentation of collagen presumably depended upon the oxida- tive degradation of the proline residue, which is one of the major components of collagen, because many workers have suggested that proline oxidation leads to the scission of the peptide bond. Wolff et al. (4) have supposed that a bond cleavage occurs preferentially at the proline residue via a pyroglutamyl structure and that a new N-terminal glutamate residue is formed by spontaneous hydrolysis. However, in our experiment, no N-terminal glutamic acid was detected from

Y. Kato, K. Uchida , and S. Kawakishi, unpublished result.

the degradation products of (Pro-Pro-Gly)lo. In addition, Boc- L-pyroglutamylglycine isolated from oxidized Boc-L-prolyl- glycine was relatively stable. Schuessler et al. ( 5 ) have also assumed that a possible target for peptide bond scission of bovine serum albumin by radiation-induced oxidation may be the amino acyl-proline bond. Bateman et al. (21) have re- ported that thyrotropin-releasing hormone derivatives were cleaved at the position of proline in the presence of copper, ascorbate, and oxygen, and then a peptide amide was formed. We isolated Boc-2-pyrrolidone derived from Boc-L-prolylgly- cine exposed to Cu(II)/H202. The result indicates that peptide scission at the proline residue by metal-catalyzed oxidation, accompanied by the formation of a 2-pyrrolidone residue, is one of the mechanisms for the fragmentation of collagen. The degradation of collagen by various oxidation systems has been reported (7-11). This degradation of collagen may be derived from the oxidation of the proline residue. We have shown that the fragmentation of collagen by ultraviolet irradiation probably arose from the oxidation of the proline residue, accompanied by the formation of a 2-pyrrolidone residue (18). Though many workers have reported the fragmentation of proteins by various oxidation systems (2, 3, 5, 22-25), the mechanisms for these fragmentations have been unknown in detail. These fragmentations may be derived from oxidation of a proline residue. Recently, Cremo et al. (26) have reported that vanadate catalyzes the photocleavage of adenylate ki- nase, and its fragment has a 2-pyrrolidone structure as the C terminus, which is at the position of proline 17 in the native adenylate kinase. This fact supports our speculation on the fragmentation mechanism of proteins.

REFERENCES 1. Amici, A,, Levine, R. J., Tsai, L., and Stadtman, E. R. (1989) J. Biol. Chem.

3. Uchida, K., Enomoto, N., Itakura, K., and Kawakishi, S. (1989) Agric. Bid. 2. Uchida, K., and Kawakishi, S. (1988) Agric. Biol. Chem. 52, 1529-1535

4. Wolff, S. P, Garner, A,, and Dean, R. T. (1986) Trends Biochem. Sci. 1,

6. Kano, Y., Sakano, Y., and Fujimoto, D. (1987) J. Biochem. 102,839-842 5. Schuessler, H., and Schilling, K. (1984) Int. J. Radiat. Bid. 45 , 267-281

7. Curran, S. F., Amoruso, M. A,, Goldstein, B. D., and Berg, R. A. (1984)

8. Bowes, J. E., and Moss, J. A. (1962) Radiat. Res. 1 6 , 211-223 9. Monboisse, J. C., Braquet, P., Randoux, A,, and Borel, J. P. (1983) Biochem.

10. Monhoisse, J. C., GardieAlbert, M., Randoux, A,, Borel, J. P., and Ferradi,

11. Uchida, K., Kato, Y., and Kawakishi, S. (1992) J. Agric. Food Chem. 40,

12. Laemmli, U. K. (1970) Nature 227,680-685 13. Tapuhi, Y., Schmidt, D. E., Linder, W., and Karger, B. L. (1981) Anal.

Biochem. 115, 123-129 14. Nagasawa, T., Kurosawa, K., and Narita, K. (1973) Bull. Chem. Soc. Jpn.

46, 1269-1272 15. Grehn, L., Gunnarsson, K., and Ragnarsson, U. (1986) Acta Chem. Scand.

16. Poston, J. M. (1987) Fed. Proc. 4 6 , 1979 (abstr.) Ser. B Org. Chem. Biochem. 40, 745-750

17. Stadtman, E. R., and Berlett, B. S. (1991) J. Biol. Chem. 266,17201-17211 18. Kato, Y., Uchida, K., and Kawakishi, S. (1992) J. Agric. Food Chem. 40,

19. Uchida, K., Kato, Y., and Kawakishi, S. (1990) Biochem. Biophys. Res.

20. Cooper, B., Creeth, J. M., and Donald, A. S. R. (1985) Biochem. J. 228 ,

264,3341-3346

Chem. 53, 3285-3292

27-31

FEES Lett. 176, 155-160

Pharmacol. 32,53-58

C. (1988) Biochim. Biophys. Acta 965,29-35

9-12

9-12

Commun. 169,265-271

fil.5-fi9.6 21. Bateman, R. C., Youngblood, W. W., Busby, W. H., and Kizer, J. S. (1985)

22. Hunt, J. V., Simpson, J. A., and Dean, R. T. (1988) Biochem. J . 250,87-

"_ "_ J. Bid. Chem. 260 , 9088-9091

49 23. Wikf, S. P., and Dean, R. T. (1986) Biochem. J. 234,399-403 24. Marx, G.. and Chevion. M. (1985) Biochem. J. 236.397-400 25. Fong,.L. G., Parthasarathy, S., Witztum, J. L., and'steinherg, D. (1987) J.

26. Cremo, C. R., Loo, J. A,, Edmonds, C. G., and Hatleid, K. M. (1992) Lipid Res. 28, 1466-1477

Biochemistry 31,491-497