In Situ Click Reaction Coupled with Quantitative Proteomics for

Subscriber access provided by UNIVERSITY OF TOLEDO LIBRARIES

Article

In-situ Click Reaction Coupled with Quantitative Proteomics for Identifying Protein Targets of Catechol Estrogens Huei-Chen Liang, Yi-Chen Liu, Hsin Chen, Ming Chun Ku, QuynhTrang Do, Chih-Yen Wang, Shun-Fen Tzeng, and Shu-Hui Chen J. Proteome Res., Just Accepted Manuscript • DOI: 10.1021/acs.jproteome.8b00021 • Publication Date (Web): 13 Jun 2018 Downloaded from http://pubs.acs.org on June 18, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

In-situ Click Reaction Coupled with Quantitative Proteomics for Identifying Protein Targets of Catechol Estrogens Huei-Chen Liang,1 Yi-Chen Liu,1 Hsin Chen,1 Ming-Chun Ku,1 Quynh-Trang Do,1 Chih-Yen

Wang,2 Shun-Fen Tzeng2 and Shu-Hui Chen1* Department of Chemistry,1 Department of Life Sciences,2 National Cheng Kung University, Tainan, Taiwan

*[email protected], TEL:886-62757575 ext.65339

Abstract Catechol estrogens (CEs) are metabolic electrophiles that actively undergo covalent interaction with cellular proteins, influencing molecular function. There is no feasible method to identify their binders in a living system. Herein, we developed a click chemistry-based approach using ethinylestradiol (EE2) as the precursor probe coupled with quantitative proteomics to identify protein targets of CEs and classify their binding strengths. Using in-situ metabolic conversion and click reaction in liver microsomes, CEs-protein complex was captured by the probe, digested by trypsin, stable isotope labeled via reductive amination, and analyzed by liquid chromatography-mass spectrometry (LC-MS). A total of 334 liver proteins were repeatedly identified (n ≥ 2); 274 identified proteins were classified as strong binders based on precursor mass mapping. The binding strength was further scaled by D/H ratio (activity probe/solvent): 259 strong binders had D/H > 5.25; 46 weak binders had 5.25 > D/H > 1; 5 non-specific binders (keratins) had D/H < 1. These results were confirmed using spiked covalent control (strong binder) and noncovalent control (weak binder), as well as in vitro testing of cytochrome c (D/H = 5.9) which showed covalent conjugation with CEs. Many identified strong binders, such as glutathione transferase, catechol-O-methyl transferase, superoxide dismutase, catalase, glutathione peroxidase, and cytochrome c, are involved in cellular redox processes or detoxification activities. CE conjugation was shown to suppress the superoxide oxidase activity of cytochrome c, suggesting that CEs modification may alter the redox action of cellular proteins. Due to structural similarity and inert alkyne group, EE2 probe is very likely to capture protein targets of CEs in general. Thus, this strategy can be adopted to explore the biological impact of CEs modification in living systems. Keywords: catechol estrogens, click chemistry, electrophiles, LC/MS, dimethyl labeling, metabolites, proteomics, redox modification, post translational modification, oxidative stress

1

ACS Paragon Plus Environment

Journal of Proteome Research 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Introduction Redox-dependent covalent modification of proteins, mediated by exogenously or endogenously generated electrophiles, may induce cell signaling responses and influence the metabolic or gene regulation.1-6 The most well-known sources of electrophilic modifiers of cellular proteins are the unsaturated fatty acids of membranes and lipoproteins such as malondialdehyde, 4-oxononenal, and 4-hydroxynonenal (HNE).2,6 Elucidating action mechanism of these or other unknown electrophiles remains an important area in need of further research. Click chemistry that generated biotin-tagged protein adducts for proteomic analysis has proven useful for interrogating biochemical targets of HNE in the complex environment of the cell.7 Except HNE, however, methods for identifying potential targets of other electrophiles are largely lacking. Catechol estrogens (CEs), major metabolites of estrogens, estrone (E1), estradiol (E2), and hydroxyestriol (E3), are another important active set of electrophiles that form stable covalent conjugation with endogenous proteins recently identified from human serum.8 CEs are formed by aromatic hydroxylation (OH) of primary estrogens at either the C-2 (2OHE1/E2/E3) or C-4 (4OHE1/E2/E3) position.9 Apart from the genotoxic and cytotoxic effects of CEs due to the formation of DNA adducts,10-11 CEs may also participate in the mechanism of estrogen action on other tissues, such as the brain and pituitary gland, which contain high concentrations of CE biosynthesis enzymes (e.g. cytochrome P450).9 The redox activity of CEs was shown to be greater than that of catechol alone.12 Although some studies reported that catechol derivatives formed covalent bonds with proteins when catalyzed by enzymes or reactive oxygen species (ROS)13,14 recent studies demonstrated that CEs were able to form stable covalent conjugates with proteins under ambient conditions without enzyme or ROS, whereas catechol was unable to do so.12 Moreover, CE-conjugated endogenous proteins were detected in human serum8 yet covalent protein conjugation was not 2


Page 2 of 31

Page 3 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


detected for either parent estrogen or catechol (our unreported results). These findings indicate that the potency of CEs differs from that of either of its functional moieties alone and could play a significant role in CE-dependent protein interaction and signaling. By using the more rudimentary method of radio-isotope labeling, it was initially shown that conjugation of CEs with cellular proteins, such as β-actin, could alter its molecular function.15-18 However, radio-isotope labeling cannot confirm whether covalent conjugation occurs nor help reveal any unknown actions. Due to rapid reaction to achieve covalent ligation with proteins prior to probe metabolism and clearance, click chemistry19 has been used to develop ex-vivo methods to identify protein targets of small molecules by fluorescence or mass spectrometry (MS).20-21 Click chemistry is also a bioorthogonal chemistry that can occur inside of living systems without interfering with native biochemical processes. So far, the most efficient click chemistry for HNE targets is the 1,3-dipolar cycloaddition between azide (or alkyne) labeled probes to conjugate alkyne (or azide) labeled reporter tags.7,22 The attachment of either an azide or alkyne tag to HNE is a subtle change so that HNE, azido-HNE, and alkynylHNE exhibit comparable cytotoxicity and abilities to stimulate gene expression. However, in addition to direct binders that form covalent binding with active electrophiles, indirect binders in cells are also captured via non-covalent or non-specific binding and detected by fluorescence or MS. Because all the captured interacting proteins are digested into peptides and identified by bottom-up proteomics technique, peptides containing covalently bound reporter tags are in low abundant and hard to identify by MS. Several methods such as stringent washings to remove non-specific binders or fluoro proteomics23 to enrich HNE-modified peptides after digestion were reported to improve the identification of covalent binders. It is, however, still quite challenging to identify covalent vs non-covalent binders in a large scale analyses using MS. On the other hand, the conjugated motif of a protein may fail 3



Page 4 of 31

to be recognized and digested by specific enzymes. Its corresponding peptides thus fail to be identified by bottom-up proteomics technique. Individual identified proteins were validated as HNE targets based on other methods such as the concentration-response to HNE analogues and the specificity of their modification.2,7 In this study, we developed a click reaction/MS-based strategy to identify cellular protein targets of CEs. In contrast to HNE which was directly introduced into the investigated system, the CE metabolites were produced in situ by the enzymes present in the microsomes from an endogenous estrogen precursor probe, ethynyl estradiol (EE2). This ensures that the pertinent biological activity which produces the reactive species from the initial estrogen leaves the reactive species at a biologically relevant condition and better mimics the biosynthetic pathway of active electrophiles in a living system. Furthermore, we intended to improve the current method in classifying different binding strengths with the identified proteins by combining click chemistry, stable isotope dimethyl labeling-based comparative quantification,24 reporter tag mapping, and internal controls. The strategy and design of the method as well as the biological implications of our findings are discussed in detail.

Experimental sections Materials.

Iodoacetamide

(IAM),

dithiothreitol

(DTT),

formic

acid

(FA),

4-hydroxyestradiol (4OHE2), 4-hydroxy 17β-estrone (4OHE1), 17α-ethinylestradiol (EE2), 4-hydroxy 17α-ethinylestradiol (4OHEE2), 3-azido-1-propanamine, nicotinamide adenine dinucleotide phosphate- hydrogen (NADPH), [tris(3-hydroxypropyltriazolylmethyl)amine) (THPTA, 50 mM in PBS), H2- or D2-formaldehyde (4% in water), sodium cyanoborohydride, purified cytochrome c from bovine heart, and purified human serum albumin (HSA, ≥ 95.0% purity by agarose gel electrophoresis) were purchased from Sigma-Aldrich (St. Louis, MO, USA). Trichloroacetic acid, L-glycine, sodium dodecyl sulfate (SDS), ammonium bicarbonate, 4


Page 5 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


and

acetonitrile

(ACN),

2-(biotinamido]ethylamido)-3,3-dithiodipropionic

acid,

N-hydroxysuccinimide ester, dimethyl formamide (DMF), guanidine-HCl, PBS buffer composed of 0.2g KH2PO4 (1.8mM), 8.01g NaCl (137mM), 0.2g KCl (2.7mM), and 1.15g Na2HPO4 (10mM) at pH 7.4 were purchased from JT Baker (Center Valley, PA, USA). Glu-C and xanthine-xanthine oxidase were purchased from Roche Life Sciences (Indianapolis, IN, USA). Trypsin was purchased from Promega BioSciences (San Luis Obispo, CA, USA). Therapeutic insulin (Humulin R) and teriparatide (Forteo) were from Eli Lilly and Company (Indianapolis, IN, USA). Synthesis of biotin azide bi-functional linker (Bi-L). 165.2 µL of biotin disulfide N-hydroxysuccinimide ester (6.89 µg/µL in DMF)) was mixed with 8.5 µL of 3-azido-1-propanamine and the mixture was incubated at 4 °C for 3 hours. The yielded bi-functional linker (Bi-L) was dried by vacuum, re-dissolved in DI-water, and purified by reversed phase liquid chromatography using a C18 column (4.6 mm i.d. × 10 cm, 3µm C18, Vydac 218TP, Grace Davison Discovery Sciences, Columbia, Maryland, USA). The purified Bi-L was dried by vacuum and stored at -80°C until use. The structure of the synthesized probe was confirmed by NMR (AVANCE 600MHz NMR, BRUKER, Germany 400MHz, Bruker): 1H-NMR (MEOD, 500 MHz) δ

4.49 (dd, 1H, J= 7.5, 5.0 Hz), 4.31 (dd, 1H, J=

8.0, 4.5 Hz), 3.36 (t, 2H, J = 7.0 Hz), 3.27 (t, 3H, J = 7.0 Hz), 3.19-3.23 (m, 1H), 2.94-2.97 (m, 4H), 2.92 (d, 1H, J = 5.0 Hz), 2.71 (d, 1H, J = 13.0 Hz), 2.60 (t, 4H, J = 7.5 Hz), 2.21 (td, 2H, J = 7.5, 2.5 Hz), 1.71-1.79 (m, 3H), 1.56-1.70 (m, 3H), 1.40-1.48 (m, 2H) shown in Figure S1A; FTIR (Perkin Elmer FT-IR Spectrometer Spectrum RX I (RX-1)) shown in Figure S1B; and ESI-MS (G2-S, Waters, MA, USA): MH+ 561.2145 Da shown in Figure S1C. The EE2 precursor probe was conjugated to Bi-L via CuAAC click reaction, captured by streptavidin beads, and thereafter eluted by disulfide cleavage using dithiothreitol (DTT) and alkylated by iodoacetamide (IAM). The structure of the linker-conjugated precursor probe (EE2-Bi-L) and the eluted precursor probe (CR2p) 5



containing the reporter tag (CR2) were confirmed by ESI-MS: MH+ of 857.3787 Da for EE2-Bi-L and 542.2897 Da for CR2p (Figure S2).

Liver microsome preparation. Adult Sprague-Dawley (SD) rats were anesthetized and perfused by saline. Liver tissues were removed and hepatic microsome samples were prepared by the following method. Hepatic tissues were chopped into small pieces and homogenized in cold 100 mL KCl (150 mM)-sucrose (250 mM) buffer (pH 7.5) on ice. The liquid contents (approximately 80%) were collected, whereas the solid contents were homogenized again with additional 40 mL KCl-sucrose buffer. All the contents were pooled and centrifuged at 11,000 x g for 22 min at 4oC, and then the supernatant was centrifuged at 11,000 x g for 70 min. The pellet was re-suspended in 100 mL solution containing 80% 0.1 M phosphate and 20% glycerol Buffer (pH 7.5). The samples were stored at -80oC until use. Animal surgery and care were provided in accordance with the Laboratory Animal Welfare Act and NIH Guide for the care and use of laboratory animals, and were approved by the Institutional Animal Care and Use Committee of National Cheng Kung University, Tainan, Taiwan. Metabolic conversion and click chemistry. Liver microsome sample containing 2 mg total protein (determined by Lowry assay) was spiked with 200 µg of HSA and 200 µg of Forteo, and then divided into two equal volume. One was added with EE2 at a final concentration of 1.6 mM and the other one was added with DI-water. Each samples were further added with NADPH (30 mM in PBS) and PBS buffer till a final volume of 100 µL and then both incubated at 37 oC for 24 h. The remaining EE2 and other reagents were removed using a 3 kDa cut-off filter to yield a final volume of 100 µL. For click chemistry, each vial was added with Bi-L (1.08 µmol), CuSO4 (8 mM), THPTA (0.5 mM), and sodium ascorbate (10 mM). Using insulin as the model compound, the optimal reaction time was estimated to be around 4 h at room temperature. The yield of coupling was estimated to be around 40% (Figure S3). Excess reagents were removed using a 3 kDa cut-off membrane. 6


Page 6 of 31

Page 7 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Affinity purification. A volume of 1200 µL of streptavidin bead solution was activated by 12 mL of the phosphate buffer composed of sodium phosphate (0.02 M) and sodium chloride (0.015 M). Each microsomal solution (EE2 or DI-water) was added with 1.2 mL of the activated bead and incubated at the room temperature for 1.5 h. The remaining reagent was removed using a 3 kDa cut-off filter and washed with 20x volume of PBS buffer. For elution, each vial was added with 300 µL DTT (100 mM in 6M guanidine-HCl) and incubated at 95 °C for 15 mins. Subsequently, the solution was added with 33 µL IAM (1 M) and incubated in the dark at room temperature for 30 mins. After the reaction, the solution was dialyzed against Di-water by100-500 Da membrane at 4 °C for 12 hr or overnight to remove remaining reagents and dried by vacuum. Trypsin digestion and dimethyl labeling. Each dried solution was re-dissolved with 50 µL ammonia bicarbonate buffer (50 mM, pH 8.0) and then digested with 2 µL trypsin (1 µg/µL) at room temperature for 18 hr. After digestion, the solution treated with the EE2 probe was labeled with D2-formaldehyde, and the other treated with DI-water was labeled with H2-formaldehyde, respectively, following the procedure previously described.23 Briefly, the tryptic peptide solution was dissolved in 100 mM sodium acetate buffer (pH 5-6) and then added with 5µL of H2- or D2-formaldehyde (4% in water) and 5µL of freshly prepared sodium cyanoborohydride (600 mM). The mixture was vortexed for 10 min and then added with 5µL ammonium hydroxide (7% in water) to quench the remaining formaldehyde. Two solutions were then mixed at 1:1 v/v ratio and desalted by a C18 cartridge (Sep-Pak C18, Waters, MA, USA). The labelled and dried digest was dissolved in 50% ACN solvent and injected (100 µg total protein) onto a hydrophilic interaction chromatography (HILIC) column (2.1 mm i.d. × 15 cm, 3.5µm, 100Å, SeQuant® ZIC®-HILIC PEEK, Darmstadt, Germany). Peptides were eluted over 50 min using the following gradient: 0-5min, 90% ACN; 5-10 min, 90 to 60% ACN; 10-40 min, 60 to 35% ACN; 40-45 min, 35 to 90% ACN; and 45-50 min, 90% ACN. 7



Fractions were collected at every 2 min at a flow rate of 0.1 mL/min and combined into a total of 8 fractions. Each fraction was dried, and re-dissolved in 0.1% FA for nanoLC-MS analysis. NanoLC-MS/MS Analyses. For peptide analyses, 0.5 µL of the re-dissolved HILIC fraction was injected into a nano-UPLC system (ACQUITY UPLC, Waters, Corporation, Milford, MA, USA) equipped with a trapping-column (180 µm × 20 mm, 5 µm C18, Waters Corporation, Milford, MA, USA) and eluted initially over 5 min for sample loading using 5% ACN in 0.1% FA at a flow rate of 5 µL/min. This was followed by a 70-min gradient using mobile phase A: 0.1% FA and mobile phase B:ACN in 0.1% FA at a flow rate of 300 nL/min: 0-5 min, 5% mobile phase B; 5-40 min, 5 to 35% mobile phase B; 40-45 min, 35 to 90% mobile phase B; 45-50 min, 90% mobile phase B; 50-55 min, 90 to 5% mobile phase B; and 55-70 min, 5% mobile phase B. MS data were acquired using the data-dependent mode where one full scan with m/z 300-2000 in the LTQ-Orbitrap XL mass spectrometer (Thermo Fisher Scientific, San Jose, CA, USA) (R = 60,000 at m/z 400) at a scan rate of 30 ms/scan was followed by the five most intense peaks for fragmentation with a normalized collision energy value of 35% in the LTQ. A repeat duration of 30 s was applied to exclude the same m/z ions from being re-selected for fragmentation. Database search and quantification. All MS/MS spectra acquired from each HILIC fraction were converted to peak lists (in msm format) using an in-house Mascot Distiller program (version 2.3, Matrix Science Ltd., London, UK) with default settings for Orbitrap to search against a database composed of SwissProt 20110921 protein database for Rattus (532146sequences; 188719038 residues) and the control proteins, HSA and Forteo. For a standard search, a mass tolerance of ± 10 ppm of precursor ions and ± 0.8 Da of product ions, two allowable miss-cleavage, and a Mascot probability score (P 0.8 for correlation. The quantified peptide ratios from each fraction were merged. The peptide ratios corresponding to the same sequence but generated under different parameter, including retention time, charge state, and fraction number were combined and a Q-test was performed to exclude outliers (95% confidence). The protein ratio was calculated based on the peptide ratio of different sequences assigned to the same protein. The output of the quantification results was exported in the comma separated values (CSV) format that allows users to easily inspect and manipulate the data. The experiment was repeated for three times. The average of protein ratio and standard deviation were calculated for proteins which were identified from at least two repeated experiments.

Superoxide oxidase activity of cytochrome c. 200 µL of cytochrome c (1 µg/µL in PBS buffer) was incubated with 10 µL of 4OHE1 (5 µg/µL in ACN) at 37⁰C for 72 hrs. The 9



reaction was terminated by cooling and the remaining reagent was removed by a 3-kDa cut-off membrane. The solutions before and after incubation were injected onto an HPLC column (1.0 mm i.d. × 10 cm, 1.7µm C18, BEH130, Waters Corporation, Milford, MA, USA) coupled online to a Q-TOF instrument (ACQUITY UPLC and Xevo G2 QTof, Waters Corporation, Milford, MA, USA), and eluted over 27 min using the following gradient at a flow rate of 0.5 mL/min: 0-3 min, 5% methanol in 0.1% FA; 3-6 min, 5 to 30% methanol in 0.1% FA; 6-12 min, 30% methanol in 0.1% FA; 12-21 min, 30 to 95% methanol in 0.1% FA; 21-24 min, 95% methanol in 0.1% FA; 24-27 min, 95 to 5% methanol in 0.1% FA. Three fractions containing different percentages of the conjugated form were collected, dried, and re-dissolved in PBS buffer. The concentration of total cytochrome c in each collected fraction was adjusted to be equal (10 µM in PBS) based on the measurement of Lowry assay. The superoxide oxidase activity of each fractions was measured as the rate of cytochrome c reduction by monitoring the change in absorbance difference between 550 nm and 540 nm (A550-A540 nm) using UV instrument (SpectraMax M2e Multi-Mode Microplate Reader, Molecular Devices, CA, USA). Each fraction (12 µL) was added with 1 µL of xanthine-xanthine oxidase (hypoxanthine 5mM, xanthine oxidase 0.1 Unit, NAD+ 5mM) and the absorbance was monitored immediately at every 8 sec till 10 min. The reduction rate was calculated from the slope of the linear regression based on the data points of the first minute.23

Results and discussion Characterization of the workflow. As shown in Figure 1, CEs (using 4OHE2 as an example) are metabolites of estradiol (E2) catalyzed by P450 enzymes and form conjugates with proteins through covalent binding with specific sites of cysteine, lysine, or histidine residues on a protein.8 The entire analysis workflow involving multiple reaction steps for identification of the CE targets was depicted in Figure 2. In this reaction scheme, we were 10


Page 10 of 31

Page 11 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


able to make direct use of EE2, an analog of E2 and also an endogenous estrogen with an alkyne functional group, as the precursor probe and Bi-L, a cleavable biotin azide synthesized in house, as the capture probe to enable the copper-catalyzed azide-alkyne cycloaddition (CuAAC).19 EE2 is one of the major endogenous estrogen in humans and an orally active estrogen. One of the most important issues was whether the precursor probe was converted in situ by endogenous cytochrome P450, and whether the conjugation site of the converted CE probe differed from CE metabolites of the parent estrogen (estradiol) which did not contain the alkyne tag. Human serum albumin (HSA) and insulin, which spontaneously form covalent conjugation with CEs,12 were used as the model proteins to test the efficiency of enzymatic conversion and protein conjugation. As confirmed by MS analysis via dansyl chloride derivatization,24 the EE2 precursor was converted to hydroxylated EE2 (2/4OHEE2) by liver microsome P450 (Figure S4). This was further confirmed by the detection of the reporter tag (CR2)-conjugated tryptic peptide, NECCR2FLQHKDDNPNLPR (m/z=842.74, z=3), of HSA from both samples treated with the commercial 4OHEE2 (Figure S5A) and with the EE2 precursor probe (Figure S5B). For insulin, as shown in Supplement Figure S6, the conjugated CR2 tag was found to locate on C7 of the A chain (Figure S6A), as well as C7 (Figure S6B), H10, and K29 (Figure S6C) of the B chain. These CR2 tag conjugation sites on insulin were the same as those conjugated by the metabolite (4OHE2) from the parent estradiol,12 regardless of the alkyne tag. This indicates the EE2 precursor probe was suitable for probing the action of the estrogen family. However, the fragmentation efficiency was much poorer for the same peptide conjugated by CR2 (Figure S6) compared to that conjugated by CEs,12 possibly due to the large CR2 tag which made energy dispersed to different vibrational modes. Following EE2 precursor probe treatment, eluted samples were labeled with the D (heavy) form of formaldehyde, while those treated with buffer were labeled with the H (light) 11



Page 12 of 31

form. Both samples were spiked with equal amounts of covalent (HSA) and non-covalent (Forteo) controls. Forteo was previously shown to remain unconjugated when co-incubated with CEs.12 Based on extracted ion chromatograms of two closely eluted peptides from the two respective controls, as shown in Figure 3, the extracted light form (sample treated with buffer) mainly contained the peptide (S*VSEIQLMHNLGK*, m/z = 756, z = 2) for the non-covalent control; whereas the extracted heavy form (sample treated with the EE2 probe) mainly contained the peptide (E*TCFAEEGK*K*, m/z = 647.8677, z = 2) for the covalent control.

Apparently, affinity purification was shown to capture the covalent control, HSA,

with much greater affinity than the non-covalent control, Forteo, in the EE2 probe treated sample labeled with D formaldehyde (Figure 3). Identification of interacting proteins in rat liver microsomes. The full analysis workflow (Figure 2) was repeated three times and the collected raw data subjected to an automatic database search against the Rattus proteome database. More than 600 proteins were initially identified, 334 of which (including HSA and Forteo) were repeatedly (n ≥ 2) identified (Table S1) and quantified (Table S2). Liver microsomes are vesicle-like organelles that re-form from pieces of the endoplasmic reticulum and ribosomes when they are broken-up. Our data for the 334 repeatedly identified proteins were consistent with this fact based on functional annotation using DAVID (http://david.abcc.ncifcrf.gov/):

the identified

proteins were mainly localized in cytosolic small ribosomal subunits (p = 8.6E-18), the endoplasmic reticulum lumen (p = 2.2E-12), and the mitochondrial inner membrane (p = 2.6E-4). Moreover, as depicted in Figure S5A, more than 20% of the identified proteins were ribosomal proteins and over 50% were soluble enzymes. Liver microsomes also contain high concentrations of cytochrome P450 enzymes which catalyze biosynthesis and metabolism of estrogens. Whereas, most proteins were identified by non-conjugated tryptic peptides. Few conjugated peptides were identified. This is possibly due to low abundance, mis (or poor) 12


Page 13 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


digestion, or poor fragmentation efficiency since most conjugated tryptic peptides were relatively large. The conjugated CR2 reporter tag may block trypsin recognition motif and result in un-cleaved or mis-cleaved large peptides. For example, as shown in Figure 4A, the CR2-conjugated peptide, HPEAKCR2RMPCAEDYLSVVLNQLCVLHEK (1290.6416, 3+) derived from the spiked HSA, was generated by two miscleavages (trypsin) and with a molecular weight as large as 3868.9031 Da. Moreover, as shown in Figure 4B, another large CR2-conjugated CR2

tryptic

peptide,

INNGGCQDLCLLTH

QGHCR2VNCSCRGGRILQEDFTCR (1743.4652, 3+) derived from protein Lrp1 in

microsomes, was also generated by two miscleavages (trypsin). Using collision induced dissociation, only 4 fragments were generated and they were all mapped with the theoretical b/y ion predicted based on the sequence of this precursor ion. Although further enrichment at the peptide level or multiple enzyme digestions may help to identify more conjugated peptides, sample preparation will become much more complicated. Classification of binding strength. As depicted in Figure 5, we proposed a two-step search algorithm to further classify non-covalent and covalent binders with the identified proteins. For the second stage search, a smaller dataset was constructed using the identified (> 600) proteins including controls, against which the raw data were searched again using the CR2 tag (∆m = 555.2515 Da or 553.2359 Da for dehydrogenated form12) as an additional variable modification. The use of smaller dataset helps to reduce mis-matches since the chance of identification of conjugated peptides derived from un-identified proteins (by non-conjugated peptides) is low without enrichment. Due to poor fragmentation efficiency of CR2-conjugated peptides as a result of collision-induced dissociation (CID), we did not use any cutoff for the peptide score but still kept 10 ppm as the threshold for precursor mass accuracy. Proteomics techniques repeatedly (n≥2) identified and quantified 334 interacting proteins of which 274 were mapped with one or more CR2-tagged peptides in the second 13



search of triplicate experiments. Proteins mapped by CR2-tagged peptides are likely to be covalent binders even although the peptide score was low. Moreover, the covalent control (HSA) was mapped with CR2 tag (Figue 4A) but the non-covalent control (Forteo) was not mapped with any CR2 tag. Thus, proteins with or without CR2-tagged peptides were classified as strong or weak binders, respectively. As shown in Figure 6, for example, cytochrome c was identified by two different tryptic peptides based on initial database search, ADLIAYLK (m/z = 485.8248, z = 2, D/H = 3.3) (Figure 6A) and GITWGEDTLMEYLENPK (m/z = 687.3557, z = 3, D/H = 8.3) (Figure 6B) with an average D/H ratio around 5.9. As shown in the inset of Figure 6B, the later peptide with one miscleavage, NKGITWGEDTLMEYLENPK+CR2 (m/z = 705.8318, z = 4), was further mapped with CR2 tag during the second search. In-vitro testing was performed to confirm whether cytochrome c is a strong binder of CEs. As shown in Figure S8, cytochrome c (purified from bovine heart) readily formed conjugation with CEs upon co-incubation (Figure S8A and S8B). About 20-60% of the co-incubated cytochrome c solution (Figure S8C) was one or two CEs-conjugated form. Moreover, as shown in Figure S9, Lys 55 on this peptide backbone, NKCEGITWGEETLMEYLENPK (m/z = 846.0839, z = 3), except D (rattus)→E (bovine) mutation on AA62 was further identified to be one of the conjugation sites. We further employed stable isotope dimethyl labeling-based comparative quantification to differentiate the binding strength of CEs to each protein. As shown in Figure 7 and Table S2, the D/H ratios of all identified proteins ranged from > 100 to 0.4 of which only the five lowest D/H values were < 1. It is noticeable that the six lowest D/H values ( 5.25) were mapped with one or more CR2 tags. In addition, 46 of the remaining 61 proteins (1 < D/H < 5.25) were not mapped with any CR2 tag. Moreover, the spiked covalent control, HSA, had a D/H value (8.7) among the top 259 strong binders, while the non-covalent control, Forteo, had a D/H value (4.8) between 1 and 5.25, which was dominated by weak binders. Generally speaking, results obtained from the CR2 tag mapping well correlated with that obtained from the D/H scale. Combining both results, the identified proteins were thus classified into strong binders, weak binders, as well as non-specific binders (Table S2). However, we did not intend to make clear cuts for the classification. For example, fifteen strong binders had D/H values in-distinguishable (1 < D/H < 5.25) to those of weak binders. Moreover, we were unsure whether the non-covalent control, Forteo, exhibited specific or non-specific interactions with CEs since the concentration of the spiked controls were relatively high compared to those of endogenous proteins. The classification was not intended to provide identification for site-specific modification either since the peptide score was low. Nevertheless, our classification can provide information about the binding strength that may reduce the work load for tedious target validations.2,7 Direct binders. Many reported CE direct binders were classified as strong binders by our method. As depicted in Figure S5B, the percentage of enzymes and transporters is higher for strong binders. By contrast, the percentage of ribosomal, structural, and DNA/RNA binding proteins is higher for weak binders. It is notable that over 50% of the identified enzymes were strong binders of CEs. Liver is a site for CE biosynthesis and biotransformation and many enzymes involved in these processes are their direct binders. CE-metabolic enzymes, such as catechol-O-methyl transferase (COMT), which converts CEs to 2- and 4-methoxy metabolites,17,27 and glutathione S-transferase (GST), which conjugates CE-Qs to GSH,28 were identified here as strong binders of CEs (Table S2). The main function of these 15



enzymes is detoxification, so as to protect cells from cytotoxicity, by converting toxic CEs into non-toxic forms. These identified direct binders were classified as strong binders of CEs here, providing additional justification for our method. In addition to these detoxification enzymes, the three main cellular antioxidant enzymes for ROS, such as catalase, glutathione peroxidase, and superoxide dismutase, were also identified and classified as strong binders of CEs. These antioxidant enzymes are also likely to be covalent binders since CE conjugation can be catalyzed by ROS, which oxidizes CEs to CE-Qs. Our data indicate that CEs tend to form covalent bonds with strong binders that catalyze their own metabolism or detoxification of ROS, thereby affecting their detoxification activity. This is somehow consistent with strong correlations between antioxidant activity and estrogen action which were believed to depend on their basic chemical properties instead of genomic or hormonal effects.29 Effect on superoxide oxidase activity of cytochrome c. In order to test whether covalent conjugation of CEs with stronger binders changes their biological function, superoxide oxidase activity of cytochrome c, a strong binder, was investigated. Co-incubation of 4OHE1 with cytochrome c (bovine heart) resulted in about 20-60% of the protein was conjugated (Figure S8). Three fractions were collected following elution of non-incubated (F1) and incubated cytochrome c (F2 and F3) using reversed-phase liquid chromatography (Figure S6) and each fraction was adjusted to contain equal concentration of total protein. As shown (Fig. S8C), F1 contained 100% non-conjugated cytochrome c; F2 contained ~20% conjugated cytochrome c; and F3 contained ~60% conjugated cytochrome c. As shown in Figure 8, superoxide oxidase activity was measured during the first 35 second of the reaction (Figure 8A) and found to decrease (F3 < F2 < F1) as the percentage of CE-conjugation increased (Figure 8B). Similar results were also obtained using 4OHE2 for conjugation (data not shown).

As with the three main cellular antioxidant, soluble cytochrome c is also a well-known antioxidant21,22 that oxidizes superoxide back to O2. Cytochrome c was also herein identified to be strong binders which are likely to form covalent conjugation with CEs. Moreover, our data showed that CE conjugation decreased the peroxidase or anti-oxidant activity of 16


Page 16 of 31

Page 17 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


cytochrome c by removing the superoxide generated by xanthine oxidase (Figure 8). Thus, the detoxification or antioxidant capacity of other enzymes may also be altered via covalent conjugation of CEs. On the other hand, in mitochondria, the heme group of cytochrome c accepts electrons from the bc1 complex and transfer electrons to cytochrome c oxidase. Notably, these electron carriers, namely cytochrome c, bc1 complex, and cytochrome c oxidase, were all identified to be strong binders (Table S2). It remains to be investigated whether CE conjugation will also affect electron transport processes and cell apoptosis.

Conclusions We developed a click reaction/MS-based strategy to systematically identify a large scale of cellular protein targets of CEs. The method directly used a major endogenous estrogen, EE2, which contains an alkyne functional group as the precursor probe and a cleavable biotin hydrazide linker to enable click chemistry. This allows CE probes to be synthesized in-situ by endogenous enzymes and captured by their interacting proteins in the early stage before probe metabolism and clearance. Moreover, by combining reporter tag mapping and stable isotope dimethyl labeling, this method is able to classify strong (likely to be covalent), weak (likely to be non-covalent), and non-specific binders with the identified proteins. Although most of the endogenous catechol estrogens don’t contain alkyne group like EE2 probe, the azide and alkyne functional groups are largely inert towards biological molecules and aqueous environments. In contrast, due to structural similarity, we believed the identified protein targets by EE2 probe are very likely to be targets of CEs in general. Our data indicate that CEs tend to form covalent bonds with direct binders that catalyze their own metabolism or detoxification of ROS, thereby affecting their detoxification activity. In light of the hydrophobic characteristics of CEs, coupled with their high reactivity towards cysteine (disulfide linkages), lysine, and histidine residues, CEs modification may alter the molecular structure and function of a variety of proteins. We herein highlight a new area of metabolism that requires further research. Our method can be widely 17



Page 18 of 31

applied to study cellular targets of CEs in any biological systems. By using suitable electrophilic probes, this approach can be adopted to develop methods for identifying and classifying protein targets of other active electrophiles.

The

following

supporting

information

is

available

free

of charge

at

ACS

website http://pubs.acs.org Figure S1 NMR, FTIR, and ESI-MS data of Bi-L; Figure S2 ESI-MS of (A) the linker bound precursor probe and (B) its eluted form; Figure S3 Coupling yield of click chemistry. Figure S4 Metabolic conversion of EE2 to hydroxylated EE2. Figure S5. Conjugation of the precursor probe to HSA via microsomal P450 conversion. Figure S6. Target sites of the probe on insulin via click chemistry. Figure S7 Functional annotation for identified proteins. Figure S8 In-vitro testing for spontaneous conjugation of cytochrome c with CEs. Figure. S9. Covalent cytochrome c modification confirmed in vitro by co-incubation. Table S1 Identification of CEs Target Proteins; Table S2 D/H ratios of identified target proteins; Table S3 Peptides mapped with CR2 tags.

Acknowledgements: The authors would like to thank Mr. Ming-Feng Chen of Instrument Development Center, National Cheng Kung University for his technical support in using orbitrap mass spectrometry.

Funding Sources: This work was supported by Ministry of Science and Technology, Taiwan, Republic of China (MOST 105-2113-M-006-014-MY3).

18


Page 19 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


References 1. Rudolph, T.K., Freeman, B.A. Transduction of redox signaling by electrophile-protein reactions. Science Signaling 2009, 2(90), re7. 2. Jacobs, A.T., Marnett, L.J. Systems analysis of protein modification and cellular responses induced by electrophile stress. Acc. Chem. Res. 2010, 43(5), 673-683. 3. Li, X., Gianoulis, T.A., Yip, K.Y., Gerstein, M., Snyder, M. Extensive in vivo metabolite-protein interactions revealed by large-scale systematic analyses. Cell 2010, 143, 639–650. 4. Link, H., Kochanowski, K., Sauer, U. Systematic identification of allosteric protein-metabolite interactions that control enzyme activity in vivo. Nat. Biotechnol.2013, 31, 357-362. 5. Liebler, D. C. Protein damage by reactive electrophiles: targets and consequences. Chem. Res. Toxicol. 2008, 21, 117–128. 6. Gallego, O., Betts, M.J., Gvozdenovic‐Jeremic, J., Maeda, K., Matetzki, C., Aguilar‐ Gurrieri, C., Beltran‐Alvarez, P., Bonn, S., Fernández‐Tornero, C., Jensen, L.J., Kuhn, M., Trott, J., Rybin, V., Müller, C.W., Bork, P., Kaksonen, M., Russell, R.B., Gavin, A.-C. A systematic screen for protein-lipid interactions in Saccharomyces cerevisiae. Mol. Syst. Biol. 2010, 6:430. 7. Vila, A.; Tallman, K. A.; Jacobs, A. T.; Liebler, D. C.; Porter, N. A.; Marnett, L. J. Identification of protein targets of 4-hydroxynonenal using click chemistry for ex vivo biotinylation of azido and alkynyl derivatives Chem. Res. Toxicol. 2008, 21, 432-444. 8. Fang, C.M., Ku, M.C., Chang, C.K., Liang, H.C., Wang, T.F., Wu, C.H., Chen, S.H. Identification of endogenous site-specific covalent binding of catechol estrogens to serum proteins in human blood. Toxicol. Sci. 2015, 148(2), 433–442. 9. Maclusky, N. J., Naftolin, F., Krey, L. C., Franks, S. The catechol estrogens. J. Steroid. Biochem. 1980, 55, 111-124. 10. Cavalieri, E., Frenkel, K., Liehr, J.G., Rogan, E., and Roy, D. Estrogens as endogenous genotoxic agents--DNA adducts and mutations. J. Natl. Cancer Inst. Monogr. 2000, 27, 75–93. 11. Cavalieri, E., Chakravarti, D., Guttenplan, J., Hart, E., Ingle, J., Jankowiak, R., Muti, P., Rogan, E., Russo, J., Santen, R., et al.

Catechol estrogen quinones as initiators of breast

and other human cancers: implications for biomarkers of susceptibility and cancer prevention. Biochim. Biophys. Acta 2006, 1766, 63–78. 12. Ku, M.C., Fang, C.M., Cheng, J.T., Liang, H.C., Wang, T.F., Wu, C.H., Chen, C.C., Tai, J.H., Chen, S.H. Site-specific covalent modifications of human insulin by catechol estrogens: Reactivity and induced structural and functional changes. Sci. Rep. 2016, 6, 28804; doi: 10.1038/srep28804. 13. Bednarek-Tupikowska, Antioxidant properties of estrogens. G. Ginekol Pol. 2002, 73(1), 19



61-67. 14. Nicolis, S., Monzani, E., Pezzella, A., Ascenzi, P., Sbardella, D., Casella, L. Neuroglobin modification by reactive quinone species. Chem. Res. Toxicol. 2013, 26, 1821-1831. 15. Epe, B., Hegler, J., Metzler, M. Site-specific covalent binding of stilbene-type and steroidal estrogens to tubulin following metabolic activation in vitro. Carcinogenesis 1987, 8, 1271-1275. 16. Epe, B., Harttig, U., Stopper, H., Metzler, M. Covalent binding of reactive estrogen metabolites to microtubular protein as a possible mechanism of aneuploidy induction and neoplastic cell transformation Environ. Health Perspect. 1990, 88, 123-127. 17. Haaf, H., Li, S.A., Li, J.J. Covalent binding of estrogen metabolites to hamster liver microsomal proteins: inhibition by ascorbic acid and catechol-o-methyltransferase. Carcinogenesis 1987, 8, 209-215. 18. Markides, C., Liehr, J. Specific binding of 4-hydroxyestradiol to mouse uterine protein: evidence of a physiological role for 4-hydroxyestradiol. J. Endocrinol. 2005, 185, 235-242. 19. Kolb, H.C., Finn, M.G., Sharpless, K.B. Click chemistry: diverse chemical function from a few good reactions. Angew. Chem. International Edition 2001, 40 (11), 2004–2021. 20. Rix, U., Superti-Furga, G. Target profiling of small molecules by chemical proteomics. Nat. Chem. Biol. 2009, 5, 616-624. 21. Wright, M.H., Sieber, S.A. Chemical proteomics approaches for identifying the cellular targets of natural products. Nat. Prod. Rep. 2016, 33, 681-708. 22. Rostovtsev, V. V.; Green, L. G.; Fokin, V. V.; Sharpless, K. B. A stepwise huisgen cycloaddition process: copper(i)-catalyzed regioselective Ligations of azides and terminal alkynes. Angew. Chem., Int. Ed. 2002, 41, 2596–2599. 23. Yuan, W.; Zhang, Y.; Xiong, Y.; Tao, T.; Wang, Y.; Yao, J.; Zhang, L.; Yan, G.; Bao, H.; Lu, H. Highly selective and large scale mass spectrometric analysis of 4-hydroxynonenal modification via fluorous derivatization and fluorous solid-phase extraction. Anal. Chem. 2017, 89, 3093-3100. 24. Hsu, J.L., Huang, S.Y., Chow, N.H., Chen, S.H. Stable-isotope dimethyl labeling for quantitative proteomics. Anal. Chem. 2003, 75, 6843-6852. 25. Huang, H.J., Chiang, P.H., Chen, S.H. Quantitative analysis of estrogens and estrogen metabolites in endogenous MCF-7 breast cancer cells by liquid chromatography–tandem mass spectrometry. J. Chromatogr. B 2011, 879, 1748-1756. 26. Hodge, K., Have, Cleaning up the masses: exclusion lists to reduce contamination with HPLC-MS/MS. J Proteomics. 2013, 88, 92–103. 27. Dawling, S., Roodi, N., Mernaugh, R.L., Wang, X., Parl, F.F. Catechol-o-methyltransferase (COMT)-mediated metabolism of catechol estrogens: comparison of wild-type and variant COMT isoforms. Cancer Res. 2001, 61, 6716–6722. 20


Page 20 of 31

Page 21 of 31 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


28. Yao, J., Chang, M., Li, Y., Pisha, E., Liu, X., Yao, D. Inhibition of cellular enzymes by equine catechol estrogens in human breast cancer cells: specificity for glutathione S-transferase P1-1. Chem Res Toxicol. 2002, 15(7), 935-42. 29. Moosmann, B., Behl, C. The antioxidant neuroprotective effects of estrogens and phenolic compounds are independent from their estrogenic properties. Proc. Natl. Acad. Sci. USA 1999, 96, 8867–8872.

Figure legends Figure 1. Biosynthesis and conjugation of CEs (using 4OHE2 as an example) with proteins via covalent binding with specific AA (Cys, Lys, or His) residues.

Figure 2. Workflow of quantitative chemical proteomics for identifying CEs-interacting proteins. The reporter tag, CR2 (555.2515 Da), represents EE2 plus the cleaved linker moiety as indicated.

Figure 3. Affinity purification of the controls. Extracted ion chromatograms of the light form (upper) and the heavy form (lower) of the precursor ions, *ETCFAEEGK*K* derived from the covalent control (HSA), and S*VSEIQLMHNLGK derived from the non-covalent control (Forteo). The light form was labeled for the sample treated with DI water and the heavy form was labeled for the ample treated with EE2 precursor probe before mixing.

Figure

4. CR2

HPEAK

Tandem

MS

(MS2)

spectra

of

the

CR2-conjugated

peptides:

(A)

RMPCAEDYLSVVLNQLCVLHEK (1290.6416, 3+) derived from the spiked

HSA and (B) INNGGCQDLCLLTH CR2QGHCR2VNCSCRGGRILQEDFTCR (1743.4652, 3+) derived from protein Lrp1 in microsomes. D form dimethyl labeled (D), M oxidized (oxi), NQ deamidated (DE), and carbamidomethyl (IAM) labeled sites were indicated.

Figure 5. Algorithm for identification classification of covalent binders, non-covalent binders, and non-specific binders based on two-step Mascot database search and quantification.

Figure 6. Identification and classification of cytochrome c as covalent binder of CEs. CID-MS2 spectra of the peptides derived from cytochrome c (A) ADDLIAYLKD (m/z = 485.8248, z = 2) and (B) GDITWGEDTLMEYLENPKD (m/z =687.3550, z = 3). Their 21



Page 22 of 31

isotopic pair profile used for quantification was shown in the inset. The isotopic profile of the miscleaved

peptide

which

was

mapped

with

CR2

reporter

tag,

NKGITWGEDTLMEYLENPK+CR2 (m/z=705.8352, z=4), was also shown in the inset of (B). D form dimethyl labeled (D) sites were indicated.

Figure 7. Distribution of all D/H ratios of the 334 identified proteins.

Figure 8. Superoxide oxidase activity of the native and CE-modified cytochrome c. (A) Rate of cytochrome c reduction measured as the change in absorbance (A550–A540) over time using a xanthine-xanthine oxidase system. (B) The averaged initial rate based on the fitted linear slope of data acquired during the first 35 second of four repeated measurements (n=4). The error bar indicates the standard deviation and * indicates significant changes (p

In Situ Click Reaction Coupled with Quantitative Proteomics for

Recommend Documents