Extent of the Oxidative Side Reactions to Peptides and Proteins

Sep 1, 2016 - Siheng Li , Lin Wang , Fei Yu , Zhiling Zhu , Dema Shobaki , Haoqing Chen , Mu Wang , Jun Wang , Guoting Qin , Uriel J. Erasquin , Li Re...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/bc

Extent of the Oxidative Side Reactions to Peptides and Proteins During the CuAAC Reaction Siheng Li,† Honghao Cai,† Jilin He,† Haoqing Chen,† Srujana Lam,† Tao Cai,† Zhiling Zhu,† Steven J. Bark,‡ and Chengzhi Cai*,† †

Department of Chemistry and ‡Department of Biology and Biochemistry, University of Houston, 4800 Calhoun Road, Houston, Texas 77204, United States S Supporting Information *

ABSTRACT: The copper-catalyzed azide−alkyne cycloaddition (CuAAC) reaction is a powerful tool for bioconjugation of biomolecules, particularly proteins and peptides. The major drawback limiting the use of the CuAAC reaction in biological systems is the copper-mediated formation of reactive oxygen species (ROS), leading to the oxidative degradation of proteins or peptides. From the studies on a limited number of proteins and peptides, it is known that, in general, the copper mediated oxidative damage is associated with the copper coordination environment and solvent accessibility. However, there is a lack of data to help estimate the extent of copper-mediated oxidation on a wide range of proteins and peptides. To begin to address this need, we quantitatively measured the degree of copper-mediated oxidation on libraries of 1200 tetrapeptides and a model protein (bovine serum albumin, BSA) using liquid chromatography mass spectrometry (LC-MS). The collected data will be useful to researchers planning to use the CuAAC reaction for bioconjugaton on peptides or proteins.



INTRODUCTION The copper-catalyzed azide−alkyne cycloaddition (CuAAC) reaction1−3 has been extensively used in the field of pharmaceutics and chemical biology as an efficient method for covalent modification of biologically active or proactive molecules.4 Examples include the site-specific PEGylation of protein-based therapeutics,5 protein/DNA-template assisted fragment-based drug discovery,6 and chemoselective biomolecule tagging for affinity purification or fluorescence detection.7 Compared to other types of copper-free click reactions, the CuAAC reaction has a unique advantage in that the resultant small triazole linkage is similar to a peptide linkage in size and polarity,8 thus causing the least perturbation to the biological functions of the conjugate.8,9 In addition, as compared to the less available or more costly handles for copper-free click reactions, such as cyclooctyne, trans-cyclooctene, and tetrazine derivatives, the small azido and alkynyl handles are more readily incorporated into biomacromolecules by chemical methods or biosynthetic pathways.10,11 Despite the broad utility of the CuAAC reaction for bioconjugation, there is a major drawback originating from the concomitant CuI-mediated reactive-oxygen species (ROS) generation in many systems.6,12,13 Specifically, in the presence of molecular oxygen, the catalytically active CuI species is rapidly oxidized, requiring a reducing agent, most commonly sodium ascorbate (NaAsc), for CuI regeneration.4,14 This redox cycle continuously generates ROS until the reducing agent is © 2016 American Chemical Society

consumed. Excessive ROS may degrade delicate biomolecules to varying extents,15 including the oxidative modification of certain amino acid residues, particularly histidine, cysteine, methionine, tryptophan, and tyrosine.16 The oxidative stress induced by the copper catalyst is also a major source of cytotoxicity which limits the use of the copper catalyst in living cells.17 Copper-catalyzed generation of unregulated ROS and the subsequent oxidative modification/degradation of peptides/ proteins has been an active field of research, particularly in the study of neurodegenerative diseases.18 Studies in this field have focused on a few peptides and proteins, such as amyloid-β,19−22 α-synuclein,23,24 and huntingtin.25 The results have shown that the extent of copper-mediated oxidation depends on the local CuI and CuII coordination environment and the oxidative susceptibility of amino acid residues.19,26 CuI ligands have been used to greatly enhance the catalytic activity and stabilize the CuI oxidation state in the CuAAC reaction. For the latter, NHC ligands virtually prevent CuI oxidation in air.27 Hence, NHC ligands hold great potential to eliminate the oxidative side reactions seen during CuAAC reaction,28 although certain CuI−NHC complexes can be deactivated by free α-amino acids and glutathione.27 To date, Received: May 28, 2016 Revised: July 31, 2016 Published: September 1, 2016 2315

DOI: 10.1021/acs.bioconjchem.6b00267 Bioconjugate Chem. 2016, 27, 2315−2322

Article

Bioconjugate Chemistry

Figure 1. Structural formulas of tris(triazolylmethyl)amine-based ligands used to accelerate the CuAAC reaction and reduce the oxidative degradation of the biomolecules.

Figure 2. Oxidative loss of each peptide in 3 tetrapeptide libraries L1 (a), L2 (b), and L3 (c) under CuAAC reaction conditions measured by LCMS. The libraries consist of 400 tetrapeptides listed in the table with DL (L1, (a)), HL (L2, (b)) and LL (L3, (c)) as the 2 internal residues and the combination of 2 of the 20 amino acids (column and row headings) at the termini.

the most widely used ligands for the CuAAC reaction are based on tris(triazolylmethyl)amine derivatives.29 The water-soluble analogues,30−32 such as THPTA (1), OEG-TTAA (2), and BTTAA (3), significantly accelerated the CuAAC reaction (Figure 1).33 The complexation of these ligands also largely increased the CuII/CuI redox potential.29 Despite the use of tris(triazolylmethyl)amine ligands significantly reducing oxidative side reactions during the CuAAC reaction, we and others have shown that, even in the presence of such ligands, oxidation still occurs to cysteine residues in proteins5 and free histidine.30 Moreover, the resultant carbonyl group formation may lead to protein cross-linking.34 To reduce protein oxidation, excess ligands were used for sacrificially scavenging ROS,30,34 but a high ligand/copper ratio generally compromised the catalytic activity. The above studies focused on a limited number of peptides and proteins. A large-scale evaluation of copper-mediated oxidative damage in peptides and proteins has not been reported. Such studies performed under optimized CuAAC reaction conditions would be valuable to the field of protein

and peptide bioconjugation using the CuAAC reaction. Herein, we report a study performed with libraries of 1200 tetrapeptides and a model protein (bovine serum albumin, BSA). Besides evaluating the roles of peptide sequence and the local copper coordination environments on copper-mediated oxidative damage during various CuAAC reaction conditions, we also reported the use of ligands with oligo(ethylene glycol) (OEG) side chains as ROS scavengers to reduce peptide/protein oxidative degradation.



RESULTS AND DISCUSSION

Copper-Mediated Oxidative Damage in Tetrapeptide Analogs. In this study, we designed a simplified system consisting of three tetrapeptide libraries constructed by solidphase combinatorial synthesis. All copper-binding motifs within the tetrapeptide were capable of coordinating with the same copper atom, thus enabling us to assess how the neighboring amino acid compositions affected the copper-mediated peptide oxidative damage. During the solid-phase synthesis of the tetrapeptide library, the first and the fourth residues were varied 2316

DOI: 10.1021/acs.bioconjchem.6b00267 Bioconjugate Chem. 2016, 27, 2315−2322

Article

Bioconjugate Chemistry

Figure 3. (a) Extent of peptide 4 oxidative degradation after incubation with copper catalyst for 10 min in air (A, B, and C) and in anaerobic environment (D). (b) Extent of peptide 4 oxidative degradation during CuAAC reaction of various reaction time and sodium ascorbate concentration. (c) Extent of ligand 2 oxidative degradation during CuAAC reaction with various reaction time and sodium ascorbate concentrations.

with the 20 natural amino acids by the “split and mix” method. The two internal amino acids residues were Asp-Leu (L1), HisLeu (L2), and Leu-Leu (L3) in each library, respectively. The selection of the two internal amino acids residues took consideration of their difference in copper affinity: the carboxylate anion in aspartic acid stabilizes the CuII complex, histidine has a high affinity to both CuI and CuII, and leucine is inert to copper. To enhance the hydrophobicity of the tetrapeptides for sufficient retention on C18 column during liquid chromatography mass spectrometry (LC-MS) analysis, the resin-bound peptides were end-capped with butyric acid. After cleavage from the solid support, three 400 peptidecontaining libraries were obtained. To study the copper-mediated oxidative damage of each individual peptide in the libraries during the CuAAC reaction, the peptides were mixed with ligand-free CuAAC reagents in air for 1 h. The mixture was analyzed by LC-MS and each individual peptide was resolved by tandem MS (MS/MS). An internal standard peptide (MRFA) was spiked into the peptide mixture for quantifying the peptide oxidative loss. We noted that this method was not capable of differentiating the leucineisoleucine isobaric isomers and the sequence isomers of reversing the first and fourth residues. For an overview of the data, Figure 2a−c shows the range (represented by different colors) of oxidative loss for each peptide in all three libraries. The measured percentage of oxidative loss for each peptide is shown in Figure S3. All peptides suffered various degrees of oxidative loss. Among all the tetrapeptides, histidine-containing peptides were the most vulnerable in copper-mediated oxidation, as shown by the results in library L3. In particular, less than 40% of cysteinecontaining peptides in library L3 were substantially oxidized, while all histidine-containing peptides in library L3 were oxidized to a large extent. Greater oxidation in library L1 than in library L2 indicated that the presence of the CuIIcoordinating carboxylate group in aspartic acid promoted more oxidation than the CuI- and CuII-coordinating histidine. Additionally, in libraries L1 and L2, the presence of (additional) carboxylate groups (in aspartic acid and glutamic acid) promoted the oxidation of peptides with hydrophobic amino acids. Increased oxidation also occurred to methionine-,

tryptophan-, and tyrosine-containing peptides in libraries L1 and L2. Reducing Peptide Oxidation by Using CuI Ligand. Among all the peptides in the mixture, peptide 4 was among the most vulnerable for ROS damage, as it was oxidized completely in the CuAAC reaction mixture containing 100 μM CuSO4 and 5 mM sodium ascorbate for 1 h. The oxidation product was the singly oxidized peptide 4 (+ 16 Da), corresponding to the oxidation of the histidine residue to 2oxo-histidine (Figure S6). Therefore, peptide 4 was used to study the oxidation under common CuAAC reaction conditions in the presence of ligand 2 or 3, in situ generated CuI from CuSO4, and excess sodium ascorbate in air. As a positive control to prevent oxidation, the reaction was performed in an anaerobic environment using the 2-CuI (1/1) complex prepared from Cu(I) (MeCN) 4 PF 6 (Figure S5). The concentration of peptide 4 in the reaction mixture was determined by HPLC. As shown in Figure 3a, over 90% peptide 4 was oxidized in the mixture of CuSO4/sodium ascorbate without a ligand (Figure 3a, A). Both ligand 2 and ligand 3 largely reduced the oxidative damage to 75% of the peptide 4 remained intact after the ligand 2-assisted CuAAC reaction. On the contrary, the remaining 4 was