Goolcharran C, Cleland JL, Keck R, Jones AJS and Borchardt RT Comparison of the Rates of Deamidation, Diketopiperazine Formation and Oxidation in Recombinant Human Vascular Endothelial Growth Factor and Model Peptides AAPS PharmSci 2000; 2 (1) article 5 (https://www.pharmsci.org/scientificjournals/pharmsci/journal/5.html).

Comparison of the Rates of Deamidation, Diketopiperazine Formation and Oxidation in Recombinant Human Vascular Endothelial Growth Factor and Model Peptides

Submitted: November 11, 1999; Accepted: March 1, 2000; Published: March 17, 2000

Chimanlall Goolcharran¹, Jeffrey L. Cleland², Rodney Keck², Andrew J.S. Jones² and Ronald T. Borchardt¹

¹Department of Pharmaceutical Chemistry, The University of Kansas, 2095 Constant Avenue, Lawrence, KS 66047

²Pharmaceutical Research and Development, Genentech, Inc., South San Francisco, CA 94080

Abstract

In this work, we examine the way in which stability information obtained from studies on small model peptides correlates with similar information acquired from a protein. The rates of deamidation, oxidation, and diketopiperazine reactions in model peptide systems were compared to those of recombinant human vascular endothelial growth factor (rhVEGF). The N-terminal residues of rhVEGF, a potent mitogen in angiogenesis, are susceptible to the aforementioned reactions. The degradation of the peptides L-Ala-L-Pro-L-Met (APM) and Gly-L-Gln-L-Asn-L-His-L-His (GQNHH), residues 1-3 and 8-12 of rhVEGF, respectively, and rhVEGF were examined at pH 5 and 8 at 37°C. Capillary electrophoresis and high-performance liquid chromatography (HPLC) stability-indicating assays were developed to monitor the degradation of the penta- and tripeptides, respectively. The degradation of rhVEGF was determined by tryptic mapping and quantified by RP-HPLC. The rates of degradation of both peptides and the protein followed apparent first-order kinetics and increased with increasing pH. The tripeptide APM underwent diketopiperazine formation (Ala-Pro-diketopiperazine) and oxidation of the Met residue, whereas the pentapeptide GQNHH degraded via the deamidation pathway. The results indicate that the rates of deamidation and oxidation of the protein are comparable to those observed in the model peptides at both pH values. However, the rate of the diketo-piperazine reaction was slower in the protein than in the model peptide, which may be the result of differences in the cis-trans equilibrium of the X-Pro peptide bonds in the 2 molecules.

Introduction

Vascular endothelial growth factor (VEGF) has been identified as a heparin-binding polypeptide mitogen that plays an important role in physiologic and pathologic angiogenesis. It is secreted by ischemic tissues and has a highly specific role in the maintenance and induction of growth of the vascular endothelial cells^1-4. Analysis of the human cDNA clone of VEGF revealed that the protein exists in 4 different isoforms having 121 (VEGF₁₂₁), 165 (VEGF₁₆₅), 189 (VEGF₁₈₉), and 206 (VEGF₂₀₆) amino acid residues⁵. The heterogeneity of VEGF was shown to arise from alternative exon splicing of a single VEGF gene⁵. VEGF₁₂₁ is a weakly acidic polypeptide that fails to bind to heparin⁶ , while VEGF₁₆₅, VEGF₁₈₉, and VEGF₂₀₆ are more basic and bind heparin with increasingly greater affinities⁷. Of these 4 species, VEGF₁₆₅ is the predominant molecular form produced in normal cells and tissues⁸. The other isoforms, VEGF₁₂₁ and VEGF₁₈₉, are detected in the majority of cells and tissues that expresses the VEGF gene, while VEGF₂₀₆ is a rare form that has been identified in human fetal liver only^5,7.

The native VEGF is a basic, heparin-binding, highly conserved homodimeric protein, with a molecular mass of 46 kDa^{1, 9}. The structure of the protein shows that it belongs to the cystine knot growth factor superfamily and is composed of an antiparallel homodimer, covalently linked by 2 disulfide bridges between Cys-51 and Cys-60¹⁰. Recombinant human vascular endothelial growth factor (rhVEGF) behaves in a manner similar to native VEGF in terms of its binding to heparin and its biological activity⁵. rhVEGF is a homodimeric protein consisting of 165 amino acids per monomer with a molecular weight of 38.3 kDa and a pI of 8.5. The protein consists of 2 domains, a receptor-binding domain (residues 1-110) and a heparin-binding domain (residues 111-165)^11,12.

It was observed that around neutral pH values, the N-terminal sequence of rhVEGF is susceptible to deamidation, diketopiperazine formation, and oxidation (J.L. Cleland, R. Keck, and A.J.S. Jones, unpublished data). The major degradation pathway is the deamidation of the Asn-10 residue. Diketopiperazine products and oxidation products (ie, Met-3 residue) were observed upon storage of the protein.

In this study, we were interested in examining the way in which stability information obtained from studies on small model peptides correlates with that of the in vitro chemical stability of a protein. In order to achieve this objective, the rates of deamidation, diketopiperazine formation and oxidation of rhVEGF and model peptides were compared. The model peptides we chose consisted of sequences corresponding to sites in the protein where these reactions occurred. The tripeptide L-Ala-L-Pro-L-Met (APM) and pentapeptide Gly-L-Gln-L-Asn-L-His-L-His (GQNHH), which constitute residues 1-3 and 8-12 of rhVEGF, respectively, were the model peptides used in this study.

Materials and Methods

Materials

The peptides APM and GQNHH were synthesized by California Peptide Research, Inc. (Napa, CA). Recombinant human vascular endothelial growth factor was produced by Genentech, Inc. (South San Francisco, CA). The protein bulk solution contains 2.4 mg/mL of rhVEGF, 104 mg/mL trehalose, 0.01% Tween-20, and 5 mM sodium succinate (pH 5.0). The protein bulk solution was concentrated using an Amicon Centriprep-30 concentrator (Beverly, MA) and lyophilized. Bovine lung aprotinin was purchased from Sigma Chemical Co. (St. Louis, MO), plasmin from Calbiochem-Novabiochem Corp. (La Jolla, CA), and TPCK-trypsin from Worthington Biochemical Corp. (Lakewood, NJ). High-purity urea was obtained from Pharmacia Biotech (Uppsala, Sweden) and low-peroxide Tween-20 from Karlshamms, USA (Columbus, OH). All other chemicals were of analytical grade and used as received. The water used in all studies was from a Millipore MILLI-Q^TM water system.

Separation of Receptor and Heparin Domainsof rhVEGF

The proteolytic enzyme plasmin (EC 3.4.21.7), which has a broad spectrum of activity, was used to separate the 2 domains of the protein. The technique used was similar to that employed by Keyt et al¹². Plasmin was added to the protein (0.5% w/w) in 50 mM phosphate buffer (PBS) at pH 7.4 and incubated at 25°C for 24 hours. The digest was stopped by adding to the mixture a 10-fold excess of aprotinin, a potent inhibitor, with respect to plasmin. A POROS^R HE-2 heparin column (1.5 x 10 cm) operated at room temperature was used to separate the fragments. The mobile phase consisted of 10 mM PBS, pH 7.0 (solvent A) and 3.0 M NaCl in 10 mM PBS, pH 7.0 (solvent B). Elution was accomplished using a flow rate of 2.0 mL/min and a linear gradient of 5% to 20% solvent B in 5 minutes (Figure 1). The fractions containing the receptor binding domain were collected and dialyzed against phosphate buffer (pH 7.0).

Tryptic Digest

The receptor domain samples were reduced and carboxymethylated prior to treatment with trypsin (EC 3.4.21.4). Denaturation of the protein was carried out in 8 M urea in 360 mM Tris buffer, pH 8.6 containing 2 mM ethylenediaminetetraacetic acid (EDTA). The disulfide bonds were reduced with 1 M dithiothreitol and alkylated with 1 M iodoacetic acid.

After buffer exchange into 100 mM Tris buffer (pH 8.3) containing 2 mM CaCl₂, trypsin (2% w/w) was added to the solution. The samples were then incubated at 37°C for 4 hours. At the appropriate time, the digestion was stopped by adding 85% phosphoric acid to the mixture.

HPLC Analysis

Analyses of the tryptic fragments of rhVEGF were performed using reversed-phase high-performance liquid chromatography (RP-HPLC). The system used consisted of a Hewlett-Packard 1090 liquid chromatograph system with a diode array detection and a Turbochrom 3 Chromatography Workstation data acquisition system (Perkin Elmer, Cupertino, CA). Separation was performed using a Phenomenox Jupiter 300 Å C-18 column (5 µm, 4.6 x 250 cm) operating at 40°C. The solvent systems consisted of 50 mM sodium phosphate, pH 3.5 (solvent A) and acetonitrile (solvent B). The tryptic fragments were separated using a linear gradient of 0 to 50% solvent B in 50 min. The flow rate was maintained at 1.0% mL/min and the detection wavelength was 214 nm. The fragments were collected from the analytical HPLC and analyzed by matrix-assisted laser desorption ionization mass spectroscopy (MALDI-MS).

Kinetic Measurements

Experiments were carried out in phosphate buffer at pH 5.0 and 8.0 at 37°C. A constant ionic strength of 0.5 M was maintained for each buffer by adding a calculated amount of NaCl. The buffers were prepared at the experimental temperature and the pH value for a given solution remained unchanged throughout the investigation.

Recombinant human vascular endothelial growth factor was dissolved in the appropriate buffer solution to yield an initial concentration of approximately 1 mM. Aliquots (750 µL) of the resulting solution were transferred to vials, sealed and stored at 37°C. At various times, a vial was removed, cooled to room temperature, and subjected to analysis.

The stability of the peptides was carried out under similar conditions as those of rhVEGF, in 100 mM phosphate buffer (I = 0.5) containing 0.01% Tween-20 at pH 5 and 8. The degradation of APM was monitored by HPLC with a Vydac C-18 column (4.5 x 250 mm, 300 Å), while the degradation of GQNHH was followed by a capillary electrophoresis stability-indicating assay¹³.

Results

Characterization of the Tryptic Fragments

There are 7 tryptic sites in the monomer of the receptor domain of rhVEGF, resulting in 8 peptide fragments (Table 1). The RP-HPLC separation of the fragments is shown in Figure 2. The fragments were identified by MALDI mass spectroscopy. The T7 and T8 fragments were not detected, being single amino acid and dipeptide species, respectively. Located within the T1 fragment are the residues susceptible to deamidation, diketopiperazine formation, and oxidation. As such, significant changes occur in the peak area of this fragment over time, with the appearance of new peaks.

Kinetics of Degradation of the T1 Tryptic Fragment

Deamidation of the Asn-10 residue, oxidation of Met-3, and the diketopiperazine formation reaction occur within the T1 fragment of the protein. Figure 3 shows a typical time course of the disappearance of the T1 fragment and the appearance of the degradation products. The apparent first-order rate constants for the appearance of the degradation products were obtained by a least-squares fitting procedure of the time course profiles at pH 5.0 and 8.0 to the following equations.

[A] = A₀ exp(-Kt) (1)

[B] = {(k₁ A₀) (exp(-k₅ t) - exp(-Kt))/(K - k₅)} (2)

[C] = {(k₂ A₀) (exp(-k₄ t) - exp(-Kt))/(K - k₄)} (3)

[D] = {(k₃ A₀) (exp(-k₆ t) - exp(-Kt))/(K - k₆)} (4)

These equations describe Scheme 1, where A, B, C, and D represent the T1, T1-deamidated, T1-oxidized, and des (Ala-Pro) T1 fragments, respectively. The constants k₁, k₂, and k₃ represent the rates of formation of the deamidated, oxidized, and truncated species, and k₄, k₅, and k₆ are the rates of disappearance of these species as they are generated (Scheme 1). The results of the least-squares procedure are summarized in Table 2. At both pH values, deamidation is the predominant degradation pathway and the most sensitive to changes in the pH of the solution. These data suggest that the protein undergoes deamidation at a relatively rapid rate. In a separate report, we have shown that the rate of deamidation of the Asn-10 residue is acid-base catalyzed by a His residue located on its C-terminal side¹³. The protein also undergoes oxidation faster than diketopiperazine formation, ie, approximately 1.3 times faster at the lower pH and 2.7 times faster at the higher pH value.

Degradation of AMP and GQNHH

The tripeptide APM can undergo diketopiperazine reaction and oxidation of the Met residue (Scheme 2. The disappearance of the initial peptide and the appearance of degradation products followed first-order kinetics. The apparent first-order rate constants were calculated from semi-logarithmic plots (Table 3). The diketopiperazine and oxidation reactions were sensitive to changes in the pH of the solution, occurring more rapidly at the higher pH value. The rate of diketopiperazine formation from the oxidized product, APM(O), was negligible at both pH 5.0 and 8.0. The rate constant for disappearance of APM at pH 8 is significantly greater than the sum of the rate constants for diketopiperazine formation and Ala-Pro-Met(O) formation, suggesting that an additional pathway(s) may be involved in the degradation of this peptide at this pH. Previously, we have shown that the rate of diketopiperazine formation is dependent on the ability of the X-Pro peptide bond to adopt a cis conformation¹⁴. In the cis conformation, the attacking N-terminal amino nitrogen atom is in proximity to the amide carbonyl between the second and third amino acids, facilitating ring closure ¹⁵. The cis-trans equilibrium of an X-Pro peptide bond has been shown to be dependent on the nature of the flanking amino acid and on the charge distribution around the bond^16-18.

The rates of degradation of GQNHH by the various pathways described above are summarized in Table 4. The rate of deamidation of the peptide followed pseudo-first-order kinetics and increased with increasing pH, which is consistent with literature reports¹⁹. At pH 5.0, the pentapeptide degraded exclusively to the Asp- and isoAsp-containing products in a 1:3.5 ratio. However, at pH 8.0, detectable levels of the Asn-His peptide bond hydrolysis products were observed. These hydrolytic products constitute less than 5 percent of the total degradation products and were not incorporated into the analysis of the data.

Discussion

Comparison of the Stability Information of the Model Peptides and rhVEGF

The rates of deamidation, diketopiperazine formation, and oxidation of rhVEGF and the model peptides are summarized in Table 4. These results indicate that the rates of deamidation and oxidation of the protein are comparable to those observed in the model peptides at both pH values. However, the diketopiperazine reaction occurs at a rate approximately 4 times slower in the protein than in the tripeptide.

The N-terminal residues of rhVEGF were shown by nuclear magnetic resonance spectroscopy to be disordered in solution²⁰. Thus, the labile Asn residue, which exists in a highly flexible region of the protein, is expected to have behavior similar to a model peptide having the same sequence. It was, therefore, not surprising that the rate of deamidation of the protein was comparable to the rate of deamidation of the model peptide used in this study. In addition, the mobile N-terminus of the protein is exposed to the solvent, so it was not unexpected that the rate of oxidation of the Met-3 residue was similar to that observed in the model peptide.

Conclusion

In the model peptide APM, the peptide bond preceding the Pro residue is free to rotate around its axis and to adapt the most favorable conformation. However, although the N-terminal residues of the protein are disordered, the electron cloud of the overall protein structure can influence the X-Pro bond. The charge distribution around the Ala-Pro peptide bond in the protein could result in a shift in the equilibrium of the cis-trans ratio of that bond and, consequently, alter the rate of the diketopiperazine reaction. Thus, predicting the rate of diketopiperazine reaction in proteins from model peptide data would be problematic, because one of the major factors in determining the rate of this reaction is the cis-trans equilibrium of the X-Pro peptide bond.

Acknowledgements

This work was supported by the United States Public Health Service (GM-088359, GM-54195). We would like to thank Drs. Richard Schowen and Teruna Saihaan for their advice and Iphigenia Koumenis for the MALDI analysis.

References

1. Ferrara N, Henzel WJ. Pituitary follicular cells secrete a novel heparin-binding growth factor specific for vascular endothelial cells. Biochem Biophys Res Commun.1989;161:851-859. [PUBMED]

2. Leung DW, Cachianes G, Kuang WJ, Goeddel DV, Ferrara N. Vascular endothelial growth factor is a secreted angiogenic mitogen.Science. 1989;246:1306-1309. [PUBMED]

3. Plouet J. Schilling J, Gospodarowicz D. Isolation and characterization of a newly identified endothelial cell mitogen produced by AtT-20 cells. EMBO J.1989;8:3801-3806. [PUBMED]

4. Gospodarowicz D, Abraham JA, Schilling J, Isolation and characterization of a vascular endothelial cell mitogen produced by pituitary-derived folliculo stellate cells. Proc Natl Acad Sci U S A.1989;86:7311-7315. [PUBMED]

5. Tischer E, Mitchell, R, Hartman T, Silva M, Gospodarowicz D, Fiddes JC, et al. The human gene for vascular endothelial growth factor. Multiple protein forms are encoded through alternative exon splicing. J Biol Chem. 1991;266:11947-11954. [PUBMED]

6. Cohen T, Gitay-Goren H, Sharon R, Shibuya M, Halaban R, Levi BZ, et al. VEGF121, a vascular endothelial growth factor (VEGF) isoform lacking heparin binding ability, requires cell-surface heparan sulfates for efficient binding to the VEGF receptors of human melanola cells. J Biol Chem. 1995;270:11322-11326. [PUBMED]

7. Houck KA, Ferrara N, Winer J, Cachianes G, Li B, Leung DW. The vascular endothelial growth factor family: identification of a fourth molecular species and characterization of alternative splicing of RNA.Mol Endrocrinol. 1991;5:1806-1814. [PUBMED]

8. Ferrara N, Houck KA, Jakeman LB, Winer J, Leung DW. The vascular endothelial growth factor family of polypeptides.J Cell Biochem. 1991;47:211-218. [PUBMED]

9. Pˆtgens AJG, Lubsen NH, van Altena MC, Vermeulen R, Bakker A, Shoenmakers J, et al. Covalent dimerization of vascular permeability factor/vascular endothelial growth factor is essential for its biological activity. Evidence from Cys to Ser mutations. J Biol Chem.1994;269:32879-32885. [PUBMED]

10. Muller YA, Li B, Christinger HW, Wells JA, Cunningham BC, De Vos AM. Vascular endothelial growth factor: crystal structure and functional mapping of the kinase domain receptor binding site. Proc Natl Acad Sci U S A. 1997;94:7192-7197. [PUBMED]

11. Houck KA, Leung DW, Rowland AM, Winer J, Ferrara N. Dual regulation of vascular endothelial growth factor bioavailability by genetic and proteolytic mechanisms. J Biol Chem.1992;267:26031-26037. [PUBMED]

12. Keyt BA, Berleau LT, Nguyen HV, Chen H, Heinsohn H, Vandlen R, et al. The carboxyl-terminal domain (111-165) of vascular endothelial growth is critical for its mitogenic potency. J Biol Chem.1996;271:7788-7795. [PUBMED]

13. Goolcharran C, Stauffer LL, Cleland JL, Borchardt RT. The effects of a histidine residue on the C-terminal side of an asparaginyl residue on the rate of deamidation using model pentapeptides. J Pharm Sci. (submitted).

14. Goolcharran C, Borchardt RT. Kinetics of diketopiperazine formation using model peptides. J Pharm Sci. 1998;87:283-288. [PUBMED]

15. Purdie J, Benoiton NL. Piperazinedione formation from esters of dipeptides containing glycine, alanine, and sarcosine: the kinetics in aqueous solution. J Chem Soc. 1973;13:1845-1852. [PUBMED]

16. Levitt M. Effect of proline residue on protein folding. J Mol Biol. 1981;145:251-263. [PUBMED]

17. Ramachandran GN, Mitra AK. An explanation for the rare occurrence of cis peptide units in proteins and polypeptides. J Mol Biol. 1976;107:85-92. [PUBMED]

18. Stein RL. Mechanism of enzymatic and nonenzymatic prolyl cis-trans isomerization. In: Anfinsen CB, Edsall JT, Richards FM, Eisenberg DS, eds. Advances in Protein Chemistry. New York: Academic Press; 1993:1-24.

19. Patel K, Borchardt RT. Chemical pathways of peptide degradation. II. Kinetics of deamidation of an asparaginyl residue in a model hexapeptide. Pharm Res. 1990;7:703-711. [PUBMED]

20. Christinger HW, Muller YA, Berleau LT, Keyt BA, Cunningham BC, Ferrara N, et al. Crystallization of the receptor binding domain of vascular endothelial growth factor. Proteins. 1996;26:353-357. [PUBMED]