Efficient site-specific prokaryotic and eukaryotic incorporation of halotyrosine amino acids into proteins

Abstract

Post-translational modifications (PTMs) of protein tyrosine (Tyr) residues can serve as a molecular fingerprint of exposure to distinct oxidative pathways and are observed in abnormally high abundance in the majority of human inflammatory pathologies. Reactive oxidants generated during inflammation include hypohalous acids and nitric oxide (NO) derived oxidants, which oxidatively modify protein Tyr residues via halogenation or nitration, respectively, forming 3-chloroTyr, 3-bromoTyr, and 3-nitroTyr. Traditional methods to generate oxidized or halogenated proteins involve non-specific chemical reactions that result in complex protein mixtures, making it difficult to ascribe observed functional changes to a site-specific PTM or to generate antibodies sensitive to site-specific oxidative PTMs. To overcome these challenges, we generated a system to efficiently and site-specifically incorporate chloroTyr, bromoTyr, and iodoTyr, and to a lesser extent nitroTyr into proteins in both bacterial and eukaryotic expression systems, relying on a novel amber stop codon-suppressing mutant synthetase (haloTyrRS)/tRNA pair derived from the Methanosarcina barkeri pyrrolysine synthetase system. We used this system to study the effects of oxidation on HDL-associated protein paraoxonase 1 (PON1), an enzyme with important anti-atherosclerosis and antioxidant functions. PON1 forms a ternary complex with HDL and myeloperoxidase (MPO) in vivo. MPO oxidizes PON1 at tyrosine 71 (Tyr71), resulting in a loss of PON1 enzymatic function, but it is unclear the extent to which chlorination or nitration of Tyr71 contributes to this loss of activity. To better understand this biological process and to demonstrate the utility of our GCE system, we generated PON1 site-specifically modified at Tyr71 with chloroTyr and nitroTyr in E. coli and mammalian cells. We demonstrate that either chlorination or nitration of Tyr71 significantly reduces PON1 enzymatic activity. This tool for site-specific incorporation of halotyrosine will be critical to understanding how exposure of proteins to hypohalous acids at sites of inflammation alters protein function and cellular physiology. Further, it will serve as a powerful tool for generating antibodies able to recognize site-specific oxidative PTMs.

Graphical Abstract

INTRODUCTION

Immune cells produce reactive oxygen, halogen and nitrogen species that participate in host immune surveillance, resulting in tissue damage and oxidative modifications to proteins, DNA and lipids. Amino acid residues such as tyrosine, tryptophan, lysine, methionine, and cysteine are common targets for oxidative modifications including halogenation, hydroxylation, nitration, nitrosylation, carbamylation, oxidative cross-links and elevated states of sulfur oxidation.¹ In particular, post-translational modification (PTM) of tyrosine residues has emerged as biomarkers for oxidative damage that convey chemical information about the oxidative process involved in their generation and has been observed in over 50 human pathologies.^2–7 In neutrophils and eosinophils, myeloperoxidase (MPO) and eosinophil peroxidase, respectively, mediate the formation predominantly of hypochlorous and hypobromous acids.^8–12 In macrophages, activation of NADPH oxidase and inducible nitric oxide synthase (iNOS) results in the concurrent formation of superoxide and nitric oxide (NO), which together form peroxynitrite (ONOO⁻).^13,14 These reactive oxygen species (ROS) and reactive nitrogen species (RNS) are potent tyrosine oxidants, forming chloroTyr, bromoTyr, and nitroTyr under oxidative stress conditions.^{11,12,15–22} Site-specific tyrosine nitration and chlorination have been shown to affect protein-protein interactions with detrimental downstream effects on cellular function and have been established as important biomarkers of oxidative disease states.^23–31 However, the specific effects of tyrosine halogenation on protein function have not been well documented. It is proposed that these oxidative stress-induced post-translational modifications (ox-PTMs) alter protein function and protein-protein interactions.¹

Traditional approaches to generating proteins with ox-PTMs involves using non-specific chemical reactions in which native proteins are either mixed with reactive oxidant species like peroxynitrite or hypochlorous acid produced chemically or by enzymatic sources.^{26,28,32–36} These reactions generate proteins containing various oxidative modifications, including chloroTyr, bromoTyr and nitroTyr. Because these reactions are neither protein-, site-, nor residue-specific, heterogeneous populations of modified proteins are produced, making it impossible to accurately assess the functional consequences of a single ox-PTM.³⁷ One approach to overcome this problem is to utilize genetic code expansion (GCE) to incorporate noncanonical amino acids (ncAAs) into recombinant polypeptides, enabling detailed functional interrogation of site-specific incorporation of a modified amino acid corresponding to an ox-PTM identified as occurring in vivo.

Here we focus on generating a robust system for genetic incorporation of halotyrosine amino acids into proteins in E. coli and mammalian cells. Genetically incorporated halotyrosine amino acids (chloroTyr, bromoTyr, and iodoTyr) are useful as spectroscopic probes and pKa probes, as well as for protein stabilization and engineering of halogen bonds into proteins.^38,39 While a GCE system has been generated from the Methanocaldococcus jannaschii Tyr RS/tRNA pair to incorporate halotyrosine amino acids⁴⁰ this system is not orthogonal in eukaryotic cells. The incorporation of iodoTyr in Chinese hamster ovary cells⁴¹ has been demonstrated with an E. coli TyrRS/B. stearothermophilus Tyr tRNA pair, however, it has never been applied to chloroTyr or bromoTyr in human cells or E. coli. Here, we have developed an orthogonal synthetase/tRNA pair derived from the Methanosarcina barkeri pyrrolysine system that efficiently incorporates all three halotyrosine derivatives with high fidelity and efficiency, and is compatible with both bacterial and eukaryotic expression systems.

To demonstrate the utility of this system, we studied the effects of halogenation on HDL-associated protein paraoxonase 1 (PON1), an important cardioprotective and anti-inflammatory protein that is targeted for oxidative modification in vivo during atherosclerosis, a chronic inflammatory disease. PON1 is oxidized on numerous residues when isolated from atherosclerotic lesions, including Tyr71, a residue important for enzymatic activity.^30,42,43 Although mutagenesis with non-oxidizable amino acids has given clues as to which PON1 sites are important,^{26,30,31,44–46} the structural consequence on activity or the contributions of other sites globally hit by oxidation cannot adequately be accounted for using these methods. Using our GCE system we circumvent these confounding issues, generating pure homogeneously modified PON1 with chloroTyr and nitroTyr at Tyr71 in PON1 to verify the enzymatic activity of PON1 is significantly reduced with these modifications. This new tool for incorporation of halotyrosine, combined with our previously developed tools for nitroTyr,⁴⁷ will aid in furthering our understanding of how halogenation and nitration alter protein function and cellular pathways during periods of oxidative stress.

RESULTS

Synthetase Selection.

To identify an orthogonal Methanosarcina barkeri pyrrolysyl tRNA synthetase/tRNA_CUA pair (Mb-RS/tRNA) able to incorporate halotyrosine derivatives in response to an amber codon, we screened a library of Mb-RS variants in which five active-site residues were randomized to all 20 amino acids (L270, Y271, L274, N311, C313).⁴⁸ After a single round of positive selection in the presence of chloroTyr and a round of negative selection against canonical amino acids, 96 colonies were evaluated for their efficiency in expressing a TAG-codon interrupted sfGFP (sfGFP-150TAG) in the presence of nitroTyr and chloroTyr. In parallel, the ability of selected synthetases to discriminate against the canonical 20 amino acids was assessed by expressing the same TAG-interrupted sfGFP in the absence of ncAA. The top 18 performing clones, judged on their efficiency and fidelity, were further evaluated on a larger scale in the presence of 1 mM nitroTyr and chloroTyr, as well as bromoTyr (Figure 1B). Sequencing these 18 clones revealed 15 unique RS sequences (Supporting Table 1). While many of these RSs variants efficiently incorporated nitroTyr only one, the Mb-RS variant “C6”, showed remarkable permissivity for nitroTyr, chloroTyr, and bromoTyr. This Mb-RS variant named haloTyrRS was therefore chosen for further evaluation.

Figure 1.

Halotyrosine derivatives and initial screening of synthetase hits from library selection. (A) Structures of tyrosine and the tyrosine derivatives incorporated via genetic code expansion in this study. (B) Assessment of fluorescence from cultures expressing the sfGFP150TAG gene along with each of the 15 unique M. barkeri pyrrolysine synthetase variants identified from the selection process. Cultures were expressed in the absence of ncAA (blue) or the presence of 1 mM nitroTyr (red), chloroTyr (green) and bromoTyr (purple).

HaloTyrRS Synthetase Characterization.

The above expressions utilize a two-plasmid system in which the reporter (sfGFP-150TAG) and the amber suppressing Mb tRNA are co-expressed on the same p15a origin plasmid (pALS), while the cognate Mb RS is expressed on a pBR322 origin plasmid (pBK) (Supporting Figure 1). This system, though well-suited for library selections, is not convenient for conventional protein overexpression systems in which one’s protein of interest is expressed from a pBR322-origin plasmid (e.g. pET and pBAD vectors). To facilitate broader utility of these ncAA incorporation systems, promising synthetase variants are combined with their cognate amber suppressing tRNA into a single plasmid, pDule, with a p15a origin. We, therefore, moved the haloTyrRS gene from pBK to pDule and re-evaluated its efficiency, fidelity, and permissivity in DH10b cells with chloroTyr, bromoTyr, iodoTyr, and nitroTyr. Indeed, the Mb pDule-HaloTyrRS system enabled efficient ncAA-dependent suppression of sfGFP-150TAG in the presence of 1 mM haloTyr derivatives, modest suppression in the presence of 1 mM nitroTyr, and only trace suppression in the absence of ncAA (Figure 2A). Approximately 80 mg, 180 mg, 250 mg and 300 mg of sfGFP containing nitroTyr, chloroTyr, bromoTyr and iodoTyr at site 150, respectively, were purified to homogeneity per liter of media illustrating the permissivity of this system (Figure 2B). For comparison, WT-sfGFP yielded 550 mg per liter culture under similar conditions. While pDule-HaloTyrRS was able to incorporate nitroTyr, it does so at levels 2–4 fold less than our previously reported Mb-synthetases engineered specifically for nitroTyr.⁴⁷

Figure 2.

Assessment of ncAA incorporation using the pDule-haloTyrRS synthetase. (A) Culture fluorescence normalized to optical density at 600 nm for cells expressing wild-type sfGFP (blue) and sfGFP150TAG in the absence of ncAA (orange) or presence of 1 mM nitroTyr (gray), chloroTyr (yellow), bromoTyr (blue) and iodoTyr (green). Error bars represent standard deviations of three independent cultures. (B) Coomassie blue stained SDS-PAGE of sfGFP variants expressed and affinity-purified from cultures shown in panel A.

To confirm the haloTyr derivatives were accurately incorporated into recombinant proteins using pDule-HaloTyrRS, we measured the masses of purified sfGFP-WT, sfGFP-150-chloroTyr, sfGFP-150-bromoTyr, and sfGFP-150-iodoTyr using ESI-Q-Tof mass analysis (Figure 3). The observed and calculated molecular weights are reported in Table 1. Each of these variants had the expected masses associated with the incorporation of their respective ncAA into the peptide containing the amber codon. Interestingly, we observed several multiple higher molecular weight adducts of sfGFP-150-chloroTyr (Figure 3B) when expressed with 1 mM chloroTyr in the medium compared to the other ncAAs which warranted further characterization. The higher molecular weight adducts appeared as mass increases of 34 Da and were condition dependent.

Figure 3.

Verification of accurate incorporation of tyrosine derivatives using the haloTyrRS synthetase by mass spectrometry. (A-E) Deconvoluted mass spectra of purified sfGFP150 variants shown in Figure 2B. Variant identity is identified in the lower right corner of each panel. –Met: indicates the expected protein peak associated with loss of N-terminal methionine.

Table 1.

Expected and observed molecular masses of sfGFP variants.

sfGFP variant	Expected Mass (Da)	Observed Mass ± error (Da)
Wild-type	27830	27827 ±1
150-chloroTyr	27913	27910 ± 1
150-bromoTyr	27958	27955 ±1
150-iodoTyr	28005	28002 ±1
150-nitroTyr	27924	27920 ±1
Wild-type (HEK)	29142*	29144 ±1
150-chloroTyr (HEK)	29226*	29232 ±1
150-bromoTvr (HEK)	29270*	29272 ±1

^*These masses are consistent with the appended C-terminal V5 tag, demethionylation, and N-terminal acetylation of the sfGFP in HEK cells.

Global ChloroTyr Incorporation.

To investigate the source of the higher molecular weight adducts observed in sfGFP-150-chloroTyr, we expressed wild-type sfGFP in the presence of chloroTyr under similar conditions as above but without the pDule-HaloTyr machinery plasmid. Any incorporation of chloroTyr into WT-sfGFP under these conditions would result from natural E. coli translational components. Interestingly, mass spectrometry analyses revealed a series of mass increases of 34 ±1 Da in addition to the expected peak for WT-sfGFP (Supporting Figure 2). This confirmed that choroTyr can be globally incorporated by endogenous translational machinery at low levels, presumably via the natural tyrosine tRNA/RS pair in response to a Tyr codon based on the mass increases. To identify if global chloroTyr incorporation by E. coli is strain and conditionally dependent, DH10b and BL21ai cells expressing WT-sfGFP were cultured in 5 mL of media in 15 mL culture tubes (low aeration) and 50 mL of media in 250 mL flasks (high aeration) in the presence of 1 mM chloroTyr. sfGFP expressed in these cultures were purified and analyzed by mass spectrometry (Supporting Figure 3). Interestingly, protein expressed in 5 mL cultures with minimal aeration misincorporated chloroTyr more than the well-aerated 50 mL cultures. DH10b appeared to misincorporate chloroTyr more than BL21ai cells. We note that the native tyrosine synthetases from DH10b and BL21 cells differ at a single residue (Asp vs Glu at residue 402 in DH10b and BL21, respectively), however, it is not obvious how this minor alteration could explain the differences in misincorporation.

Effect of HaloTyr Concentration on Amber Suppression Efficiency.

From a practical standpoint, higher culture aeration and the use of BL21ai cells appear to be important to reduce global misincorporation of chloroTyr by the natural tyrosine tRNA/RS pair of E. coli. Typically during synthetase selection and evaluation, 1 mM ncAA is used in cultures to maximize amber suppression and so it was unclear whether the use of lower chloroTyr concentrations would compromise total protein production of sfGFP-150-chloroTyr. To test this, we assessed the effect that the concentration of chloro-, bromo- and iodoTyr had on amber suppression efficiency and therefore total sfGFP production by titrating increasing amounts of ncAA into expression media. Fitting these data to a simple dose-response curve allowed us to extrapolate the concentration of ncAA at which half-maximal protein production is observed, which we call the UP₅₀. We note that the UP₅₀ will depend on expression strain, growth condition, target protein, and expression levels of ncAA machinery components. However, because ncAAs are often expensive or can be toxic at high concentrations it is helpful to know the minimal amount of ncAA required to maximize protein production. These analyses revealed UP₅₀’s of 16 ± 3, 20 ± 3 and 39 ± 2 μM for chloro-, bromo- and iodoTyr amino acids, respectively (Figure 4). Since reducing the concentration of chloroTyr to 0.05 mM does not significantly compromise protein yield in BL21ai cells, chloroTyr concentrations were explored to verify global incorporation of chloroTyr is not a problem in BL21ai cells at these low concentrations. Mass analysis of expressed chloroTyr-GFP in BL21ai cells at 50 mL volumes with chloroTyr at concentrations of ranging from 0.05 mM to 2 mM did not show contamination of protein with globally incorporated chloroTyr (Supporting Figure 4). Taken together, these results demonstrate the pDule-HaloTyrRS machinery is a single, robust platform for the specific and efficient site-specific incorporation of halotyrosine derivatives into proteins.

Figure 4.

UP₅₀ determination for halotyrosine derivatives incorporated by the haloTyrRS. In-cell fluorescence was measured to determine relative amounts of sfGFP-150-haloTyr protein production in the presence of (A) chloroTyr, (B) bromoTyr and (C) iodoTyr. The red curves indicate the best-fit curve from which the indicated UP₅₀ values were derived.

Recombinant PON1 Protein Expression and Characterization.

To demonstrate the utility of the haloTyrRS incorporation system, we aimed to produce a biologically relevant protein known to be oxidatively halogenated or nitrated at tyrosine residues under physiologically relevant conditions. Here, we investigate the effects of chloroTyr and nitroTyr on the catalytic activity of PON1, an enzyme known to associate with HDL and hydrolyze thiolactones and xenobiotics. Previous work has shown the importance of Tyr71 for PON1 activity by site-specifically mutating Tyr71 to other natural amino acid residues, resulting in a significant loss of activity.³⁰ PON1 activity was also decreased after MPO-induced oxidative modification in which Tyr71, among other residues, was shown to be oxidized to nitroTyr or chloroTyr. However, it was unclear whether a modification to only Tyr71 was the key contributing factor to the overall decrease in PON1 activity. To evaluate modified PON1 we expressed and purified site-specifically incorporated chloroTyr and nitroTyr at residue Tyr71 of PON1 (Figure 5). All recombinant PON1 proteins were detected using a pan-PON1 antibody (Figure 5B), while a pan-nitroTyr antibody detected only recombinant PON1 when nitroTyr was incorporated into the protein (Figure 5C). Site-specific incorporation of nitro- and chloroTyr at position 71 of PON1 was further confirmed by tryptic digestion and liquid chromatography coupled to tandem mass spectrometry analysis (LC/MS/MS) (Supporting Figures 5 and 6).

Figure 5.

Assessment of PON1 purity and composition by SDS-PAGE and western blotting. Purified rPON1-Tyr71, rPON1–3chloroTyr71, and rPON1–3nitroTyr71 proteins had apparent predicted molecular weight (A). The proteins were also recognized by specific PON1 antibody (B), and the specific presence of nitroTyr in the protein was detected by western blot (C).

Functional Consequences of ox-PTMs on PON1 Catalytic Activity.

PON1-WT expressed and purified from E. coli showed similar enzymatic activity to previously reported recombinant PON1.³¹ However, when ox-PTM markers chloroTyr and nitroTyr were installed at position 71, significantly reduced enzymatic activity compared to wild-type was observed (by 71% and 88%, respectively, Figure 6). Crystallographic structures of PON1 have been determined^42,43 and consistent with those structures we have mapped Tyr71 of PON1 as part of the catalytic pocket via a photocrosslinkable cholesterol³⁰ in addition to other PON1 residues.³¹ Salient structural information at all of these PON1 positions points to an enriched aromatic and hydrophobic environment in which modification of the Tyr71 phenolic ring with a substituent (e.g., NO₂, Cl) could either sterically interfere with the active site cavity architecture or create unfavorable electrostatics due to the lowered pKa of the phenolic oxygen, thereby compromising the ability of Tyr71 to form a hydrogen bond with Asp183, a critical residue that modulates access to the PON1 active site.⁴³ Nevertheless, our data provide compelling evidence that oxidation at specifically residue 71 is a key contributing factor to the reduced activity of PON1 when oxidized by MPO.

Figure 6.

Paraoxonase activities of rPON1-Tyr71, rPON1-chloroTyr71, rPON1-nitroTyr71. The paraoxonase enzyme activities for purified WT rPON1 protein (WT), rPON1-nitroTyr71 (Y71NO₂), and rPON1-chloroTyr71 (Y71Cl) were assayed and shown as relative activity normalized to wild type protein activity. n=3±SD.

Eukaryotic expression of site-specifically halotyrosine containing proteins.

We next sought to determine if the haloTyrRS system could be used to express site-specific halotyrosine derivatives in eukaryotic cells. The effect of halotyrosine amino acids on the viability of HEK293T cells was tested by exposing cells to increasing concentrations of ncAA in the media. We found that none of these halogenated tyrosine derivatives inhibited cell viability up to 1 mM at 48 h after treatment (Supporting Figure 7).

We hypothesized that the Mb haloTyrRS/tRNA pair that was efficient at suppressing amber codons in E. coli could do so in mammalian cells. To test this, we cloned a human codon-optimized version of the haloTyrRS/tRNA in the pAcBac1 mammalian expression vector and introduced a single TAG codon in the sfGFP gene (Supporting Figure 8). When transfected into HEK293T cells, the haloTyrRS/tRNA pair induced sfGFP expression only in the presence of supplemented halotyrosine ncAA. A negligible read-through of the amber codon was detectable in the absence of halotyrosine (Figure 7). To estimate the efficiency of TAG suppression, transfected cells were analyzed by flow cytometry (Supporting Figure 9). Consistent with fluorescence microscopy, read-through was less than 1% and GFP positive populations increase linearly from 15% to almost 40% in the presence of 0.1 mM to 1 mM haloTyr. Suppression at TAG sites with haloTyr derivatives showed a comparable level of GFP expression (Figure 7, Supporting Figure 9). In order to correlate plasmids transfection efficiency with GCE machinery read-through efficiency we characterized the efficiency and fidelity of the haloTyrRS/tRNA pair using a reporter consisting of a TAG separated mCherry-GFP fusion protein (Supporting Figures 10 and 11).⁴⁹ All cells transfected with the reporter will express mCherry protein whereas only cells effectively suppressing the TAG stop codon will contain an mCherry-GFP fusion protein which can be quantified by 2D cell sorting (Supporting Figure 11). Transfection efficiency for all conditions was about 25% as assessed with the mCherry-GFP fusion reporter system (Supporting Figure 11B). Of cells displaying mCherry expression, TAG codon suppression efficiency increased linearly from 20% to 50% in response to increasing haloTyr amino acid concentrations from 0.1 mM to 1.0 mM (Supporting Figure 11C and 11D). This data shows that chloroTyr, bromoTyr, and iodoTyr can be site-specifically incorporated in mammalian cells using the haloTyrRS/tRNA pair with chloroTyr and bromoTyr being slightly more efficient than iodoTyr. To verify site-specific protein incorporation, WT-sfGFP, sfGFP-150-chloroTyr, sfGFP-150-bromoTyr was purified from HEK293T cells by metal affinity chromatography and characterized using ESI-Q-Tof mass analysis (Supporting Figure 12). The observed and calculated molecular weights are reported in Table 1. As detected with E. coli expression, each of these variants had the expected masses associated with the incorporation of their respective ncAA.

Figure 7.

Assessment of ncAA eukaryotic incorporation using the pAcBac1-haloTyrRS synthetase. The mean fluorescence intensity (MFI) of HEK293T cells transfected with the haloTyrRS/tRNA pair and sfGFP-150TAG in the presence or absence of 1 mM chloroTyr, 1 mM bromoTyr, 1 mM iodoTyr, or 1 mM bocLys was determined by flow cytometry. The WT PylRS was used in place of the haloTyrRS as a control. sfGFP-WT without the TAG mutation was used as a positive control. The MFI of the GFP positive cell population is shown. The significance of the difference in MFI was examined by one-way ANOVA with Dunnett post-test. *, P<0.05; **, P<0.01; #, not significant

Eukaryotic Expression of Human PON1 Containing Site-specific ChloroTyr.

Knowing that chloroTyr at position 71 of PON1 reduces enzymatic activity, we next tested whether we could express this site-specifically modified protein in HEK293T eukaryotic cells. To facilitate quantification of ncAA-PON, the signal peptide (the N-terminal 15 amino acid residues) was removed from the expression vector and EGFP was genetically fused to the C-terminus of human PON1. In this case, only if the TAG codon at position 71 was suppressed would fluorescence be observed. Consistent with sfGFP, there was negligible read-through for the PON1-TAG mutant in the absence of the ncAA (Figure 8). Specific incorporation was verified by in-gel fluorescence analysis of the crude cell lysates, which showed an unambiguous, chloroTyr-dependent fluorescent band with electrophoretic mobility consistent with that of the full-length WT-PON1-EGFP fusion protein (Figure 8A). The suppression efficiency was about 30-fold lower than WT-PON1, 4-fold lower than WT pyrrolysine RS/tRNA/Boc-Lys,⁵⁰ and 3-fold higher than another ox-PTM marker nitroTyr that was specifically incorporated to PON1 using A7-Mb RS/tRNA pair (Figure 8B).⁴⁷

Figure 8.

Incorporation of chloroTyr into paraoxonase 1 in mammalian cells. (A) HEK293T cells were transfected for 48 h with pAcBac1-PON1–71TAG-EGFP and pAcBac1-HaloTyrRS (CY), or pAcBac1-PON1-WT-EGFP (WT) using Lipofectamine2000 in the presence or absence of 1 mM chloroTyr. Crude cell lysates were analyzed by in-gel fluorescence assay. As controls, TAG suppression on PON1 was performed by A7-RS/nitroTyr (NY)⁴⁷, F4-RS/3-nitroPhe (NF)⁴⁷, or WT PylRS/bocLys (BocK)⁵⁰. The arrow indicates PON1 protein. BL, blank lane; CE, cell extract of untransfected control cells. (B) The relative band intensity was estimated by densitometry.

Western Blot Analysis.

Production of site-specific halogenated proteins should enable generation of specific and sensitive antibodies for the detection of halogenated proteins in biological systems. To our knowledge, only two commercial halotyrosine recognizing antibodies are available. First is an anti-chloroTyr antibody (Hycult Biotech, Plymouth Meeting, PA), for which there exist a few reports using it for co immunoprecipitation studies.^51,52 However its utility for western blotting has not been established.⁵³ The second is an anti-halotyrosine antibody (Santa Cruz Biotech) which was created using a bromoTyr containing peptide.⁵⁴ This antibody is advertised as western blot compatible but there are only limited reports of its use in the literature. Also available are a variety of anti-nitroTyr antibodies which were raised against protein chemically oxidized by peroxynitrite.

The ability to make site-specifically modified halotyrosine protein allowed us to evaluate the specificity and cross-reactivity of these commercial antibodies. We assessed the ability of each antibody to detect sfGFP chemically oxidized by HOCl, HOBr and ONOO⁻ alongside sfGFP site-specifically modified with chloro-, bromo- and nitroTyr (Supporting Figure 14). Interestingly, only the anti-nitroTyr antibody detected site-specifically modified sfGFP (Supporting Figure 14B, lanes 10–12). However, this antibody also detected all forms of chemically oxidized sfGFP even though nitroTyr is not expected to be present in HOCl and HOBr oxidized protein (Supporting Figure 14B, lanes 7–9). The anti-chloroTyr antibody did not detect any site-specific or chemically modified forms of sfGFP (Supporting Figure 14C), and the anti-halotyrosine weakly detected only HOBr oxidized sfGFP (Supporting Figure 14D, lane 8). No proteins from chemically oxidized cell extracts were detected by any of the antibodies. While we might expect that antibody binding to antigens be context or sequence dependent, these unsatisfactory results do highlight the need to develop better antibodies for detecting specifically halogenated or nitrated proteins.

DISCUSSION

In addition to enabling the study of ox-PTM biology and physiology, the ability to site-specifically halogenate tyrosines has other uses. For example, several GCE efforts have been reported to engineer halogen bonds into proteins for increased stability, such as lysozyme⁵⁵ and glutathione-S-transferase⁵⁶. Tyrosine halogenation has also been used to elucidate mechanisms of protein function⁵⁷ and to tune or alter protein function for specific analytical applications⁵⁸. In structural biology, protein specifically labeled with bromine or iodine gives rise to anomalous x-ray scattering, which can be used to facilitate de novo protein structure determination.^40,59 While important advancements in the field, these past systems rely on Methanocaldococcus jannaschii⁴⁰ and E. coli Tyr RS/tRNA pairs⁶⁰ that have a narrow range of utilities. Here, we report the development of a single robust genetic system capable of incorporating chloro-, bromo-and iodoTyr site-specifically and efficiently into proteins using a system compatible with bacterial and eukaryotic protein expression systems. This system also has the added benefit of being able to permissively incorporate nitroTyr too, albeit with lower efficiency than the Mb-RS/tRNA pair developed specifically for nitroTyr.⁴⁷ Our observations also highlight important practical considerations when incorporating chloroTyr into proteins, notably that this ncAA can be mis-incorporated into the proteome via endogenous E. coli synthetases. The use of BL21ai cells containing the efficient haloTyrRS/tRNA pair and low (0.1 mM) chloroTyr concentration during expression minimize these off-target effects and allow for site-specific incorporation into proteins. No off-target incorporation effects were detected in mammalian cells and low concentrations of supplemented ncAA was sufficient to efficiently express site-specific haloTyr containing protein in mammalian cells. Thus our new system combines three new advantages over previous systems: (i) compatibility with both eukaryotic and bacterial expression systems, (ii) the ability to incorporate all three biologically relevant halotyrosines efficiently and (iii) the ability to use low concentrations (0.1 mM) of chloroTyr to minimize proteome labeling by endogenous synthetases without compromising protein expression yields.

This haloTyrRS incorporation system should facilitate our ability to deconvolute specific effects associated with specific ox-PTMs by producing homogenously modified protein and by avoiding crude chemical-based oxidation systems such as hypochlorous, hypobromous acid and peroxynitrite treatment that produce a mixture of oxidation products. This clean production of halogenated antigens will enable generation of new antibodies for better detection of halotyrosine derivatives. We showed that commercially available antibodies against haloTyr, bromoTyr, and nitroTyr lacked specificity and sensitivity in detecting the ox-PTMs (Supporting Figure 14) illustrating the necessity for new antibodies. In situations where these antibodies do detect a putatively halogenated protein, this GCE-based tool will be particularly valuable for validating specific sites of tyrosine modification and discarding false-positives created by off-target reactivity with alternative oxidative modifications.

Proteomic analysis of HDL particles has identified over 50 different proteins associated with HDL⁶¹ and begat the idea that HDL-associated proteins may differ in subjects with cardiovascular disease versus healthy controls, raising the likelihood that alteration of HDL function may result from changes in HDL-associated proteins.^62–64 One such protein is the atheroprotective protein PON1, which has been linked to anti-inflammatory, antioxidant and lipid cargo-carrying activities of HDL and is found in roughly 10 – 12.5% of all HDL particles.^65–68 Our earlier studies identified ox-PTMs such as nitration and chlorination on apoA-I isolated from atherosclerotic lesions resulting in large part from the action of MPO. These PTMs resulted in loss of apoA-I/HDL-mediated atheroprotective functions,²⁸ which we and others mapped and identified to key sites playing important roles in cholesterol efflux and cholesterol ester formation.^{30,32,33,36,46,69–71} We showed that HDL, PON1, and MPO form a ternary complex that self-regulate each other’s activities.³⁰ MPO-mediated oxidation targets Tyr71 on PON1 amongst other residues both in vitro and in PON1 recovered from atherosclerotic lesions.³⁰ Importantly Tyr71 on PON1 was shown by crystallographic studies to undergo conformational change upon ligand binding and plays a key role in the hydrogen bond network near the PON1 active site needed for catalytic activity.⁴³ We wanted to strictly identify the contribution of Tyr71 oxidation on PON1 to its activity and decided to use ncAA incorporation strategies to achieve this goal. To demonstrate the utility of this newly generated bacterial and mammalian ncAA incorporation system we generated PON1 site-specifically modified at Tyrosine 71 with chloroTyr and nitroTyr by the haloTyrRS synthetase in E. coli and chloroTyr in mammalian cells. We found bacterially produced oxTyr71-PON1 species (both NO₂-Tyr71-PON1 and Cl-Tyr71-PON1) have significantly decreased enzymatic activities compared with the wild type recombinant PON1 lacking PTMs. The ability to generate site-specific nitroTyr or chloroTyr PON1 proteins will facilitate our ability to generate monoclonal antibodies that recognize dysfunctional forms of PON1. Further development and optimization of eukaryotic expression of the orthogonal halotyrosine machinery might be applicable to the simple model organism Danio rerio (zebrafish) to study cardiovascular disease effects caused by site-specific halo-PTMs of PON1 (or apoA-I), the target of two conserved proteins common to both zebrafish and man.

EXPERIMENTAL PROCEDURES

Antibodies and Reagents

Anti-His₆ antibody (Santa Cruz Biotechnology, Cat# SC-804). Anti-ChloroTyr antibody (Hycult Biotech, Cat# HP5002). Anti-Halotyrosine antibody, BTK-94C (Santa Cruz Biotechnology, Cat# SC-293152). Anti-NitroTyr antibody (Millipore (Upstate), Cat# 06–284; 1:2,000). Anti-PON1 antibody (Epitomics, 2835–1; 1:10,000). NitroTyr was purchased from Alfa Aesar (#A11018). ChloroTyr, bromoTyr, and iodoTyr were purchased from Ark Pharm Inc. (#AK-50086, AK-82171 and AK-86417, respectively). BocLys was purchased from Bachem (#4000211.0005).

Halotyrosine Synthetase Selection

The same plasmid library of aminoacyl-tRNA synthetases, randomized to all 20 natural amino acids at residues Leu270, Tyr271, Leu274, Asn311, and Cys313, and positive and negative selection plasmids used to identify nitroTyr specific aminoacyl-tRNA synthetases were to select for a haloTyr specific aminoacyl tRNA synthetases.⁴⁷

For positive selection, 100 μL of DH10B electrocompetent cells containing the positive selection plasmid, pREP-pylT, were electroporated with 1 μg of library plasmid. Cells were recovered for 1 hour in SOC media at 37 °C. Serial dilution of the combined recovery culture revealed ~10⁹ transformants, ensuring complete coverage of the library entering the selection process. Pooled recovery cultures were inoculated into 1 L of LB with 50 μg/mL kanamycin (Kan) and 25 μg/mL tetracycline (Tet) and grown for 18 hours at 250 rpm at 37 °C. Two milliliters of overnight grown pREP-pylT/D3-Library cells were used to inoculate 300 ml of LB. When the culture reached an OD₆₀₀ of 2.4, 200 μL aliquot of these cells were plated on eleven 15 cm LB-agar plates containing 25 μg/mL Kan, 12.5 μg/mL Tet, and 40 μg/mL chloramphenicol (Cm) and 1 mM ncAA, and grown overnight at 37 °C. Positive selections for nitroTyr, chloroTyr and 3-nitroPhe were performed in parallel. The plasmid DNAs of surviving members of the D3 library were collected and purified as described before⁴⁷ and kept as positive selection library.

Negative selection for the positive selection DNA library was performed using pYOBB2-pylT-containing DH10B cells as described previously⁴⁷ and the resulting DNA library was regarded as the first round of selection.

To evaluate the efficiency and fidelity of selected synthetases, pBK library DNA from the above described negative selection was transformed into cells containing the pALS plasmid. The pALS plasmid contains an arabinose-inducible sfGFP reporter with a TAG codon at residue 150 as well as a constitutively expressed Mb pyrrolysyl-tRNA_CUA. Transformed cells from each library were plated on autoinducing media containing 1% agar, 50 μg/mL Kan, 25 μg/mL Tet, and 1 mM ncAA. Plates were grown at 37 °C for 24 hours and then allowed to maturate at room temperature for an additional 24 hours. Green cells expressing a functional synthetase could be visibly identified due to TAG suppression and the production of full-length sfGFP. Mild toxicity from chloroTyr was apparent because of the retarded growth rates on plates containing chloroTyr at 1 mM.

Colonies expressing functional synthetases were selected from the plates and used to inoculate a 96-well plate containing 0.5 mL per well non-inducing media with 50 μg/mL Kan and 25 μg/mL Tet. 48 colonies were selected the 3-nitroPhe selection, 24 from the nitroTyr selection, and 24 colonies from the chloroTyr selection and allowed to grow at 37 °C with shaking (300 rpm). After 18 hours of growth, 10 μL of each culture was used to inoculate 0.5 ml of autoinduction media containing 50 μg/mL Kan, 25 μg/mL Tet. Four 96-well expressions (37°C, 300 rpm) were performed containing either nitroTyr, 3-nitroPhe, chloroTyr or no ncAA. sfGFP expression was measured after 36 hours of growth using a Synergy2 microplate reader (Biotek).

The top 18 performing ncAA-synthetases were expressed in 5 mL cultures of autoinduction media in duplicate. These 18 cultures were evaluated for efficiency and fidelity in the presence and absence of 1 mM ncAA; nitroTyr, 3-nitroPhe, chloroTyr, and bromoTyr. To mitigate the apparent toxicity from chloroTyr, this amino acid was added three hours after cultures were inoculated. Fluorescence measurements of the cultures were collected 48 hours as described above. The top-performing ncAA-synthetases were sequenced resulting in 15 unique clones (Supporting Table 1).

Construction of pDule1-haloTyrRS

The haloTyrRS synthetase gene was amplified from its respective pBK plasmid by standard PCR methods using the following primers: forward: 5’ GCTAAGGTACCTCGGGTTGTCAGC 3’, reverse: 5’ CGATACTCGAGAGCGGAATTAATTCG 3’. A double digest of amplified PCR product (1500 bp) was performed in parallel with plasmid pALS pylT (3400 bp) using KpnI and XhoI. The resulting fragments were purified, ligated and transformed using standard protocols. Rescued media was then plated on LB plates containing 25 ug/ml tetracycline. Plates were incubated at 37°C for 18 hours.

Colonies were selected into 5 ml of LB + 25 ug/ml tetracycline and grown to saturation. pDule2-HaloTyrRS was created to be functionally identical to that of pDule1-haloTyrRS except that it confers spectinomycin instead of tetracycline resistance. pDule2 was made by amplifying the spectinomycin resistance cassette from pCDFduet (Novagen) with flanking ClaI and BamHI restriction sites at the 5’ and 3’ end, respectively. This amplicon (1000 bp) and the pDule-haloTyrRS plasmid were digested with ClaI and BamHI (which removes the tetracycline resistance cassette of Mb-pDule1), gel-purified, ligated with T4 ligase, and transformed into DH10b cells as performed for pDule1 synthesis.

HaloTyrRS/tRNA Characterization

Electro-competent BL21ai cells were transformed with a pBAD vector (Invitrogen) containing only WT-sfGFP or co-transformed with a pBAD vector containing sfGFP-150TAG and the pDule vector containing the haloTyrRS. Auto-induction cultures (1.5 mL) were inoculated with 15 uL of a non-inducing starter cultures, split into three 0.5 mL cultures, and grown in the presence of either 0 mM, 0.01 mM, 0.05 mM, 0.1 mM, 0.25 mM, 0.5 mM, 0.75, and 1.0 mM of chloroTyr, bromoTyr, iodoTyr, and nitroTyr. Optical density and fluorescence measurements were measured at 24 and 48 hours.

Expression and Purification of sfGFP

Electro-competent E. coli cells (BL21ai or DH10b) were co-transformed with pBAD vectors containing genes for either WT-sfGFP or sfGFP-150TAG and pDule-HaloTyrRS via electroporation. Cells were rescued at 37°C for 1 hour in SOC and plated on LB plates (100 μg/mL ampicillin and 25 μg/mL tetracycline) and grown overnight at 37°C. Non-inducing media was inoculated with a single colony of the appropriate BL21ai cells and grown to saturation overnight at 37°C, shaking at 250 rpm. Auto-induction cultures (50 mL in a 250 mL baffled flask or 5 mL in 15 mL culture tube) were inoculated with 500 μL or 50 μL, respectively, of non-inducing starter cultures and grown in the presence of 1 mM chloroTyr, bromoTyr, iodoTyr, or nitroTyr at 37°C and shaking at 250 rpm. Cells were harvested after approximately 36 hours via centrifugation at 5000 rcf and stored at −80°C. Frozen cell pellets were resuspended in 10 mL of wash buffer (50 mM sodium phosphate, 300 mM NaCl, pH 7) and lysed via microfluidization at 18,000 psi to a final volume of approximately 35 mL. Lysates were clarified via centrifugation at 20,000 rcf for 30 min, which were then incubated with TALON Metal Affinity Resin (250 μL bed volume) for 1 hour at 4°C. The bound resin was washed with >30 column volumes of wash buffer and protein was eluted from the column with elution buffer (50 mM sodium phosphate, 300 NaCl, 150 mM imidazole, pH 7). Protein concentrations were assessed by Bradford assay.

Mass Spectrometry Analysis

Purified sfGFP protein expressed in E. coli was analyzed using an FT LTQ mass spectrometer as described earlier.⁴⁷

All PON1 mass spectrometry analyses were performed under Xcalibur data system using Proxeon Easy-nLC coupled to an LTQ Orbitrap Velos mass spectrometer (Thermo Scientific) with a nano-LC electrospray ionization source. Samples were first pressure loaded onto an IntegraFrit sample trap (ID 100 μm, length 2.5 cm, New Objective) and further separated by a PicoFrit column (stock# PF360-75-15-N-5, New Objective) packed in our laboratory (packing material: C18, 300 Å, 5 μm from Cobert Associates). Mobile phase: Buffer A: water with 0.1% formic acid; Buffer B: acetonitrile with 0.1% formic acid). The HPLC gradient started with 5% B and increased to 40% B over 60 min, then increased to 80% B over 5 min, held at 80% B for another 10 min. The flow rate was 0.3 μL/min. The applied spray voltage of the nano-LC electrospray ionization source is 2.5 kV. For MS2 data collection, one full scan by Orbitrap was followed by 10 data-dependent scans of the most intense ions by ion trap. Dynamic exclusion was also applied so that any ions that had been repeatedly scanned three times in 60 seconds would be excluded for further scan for 3 minutes.

Mass spectra were subject to SEQUEST (Thermo Scientific) and searched against a PON1.fasta database. Static modifications corresponding to NO₂,³⁵ Cl, and ³⁷Cl were applied to tyrosine. SEQUEST output files were filtered by XCorr (cross-correlation value, min XCorr = 1.8(+1), 2.5(+2), 3.5(+3) as described in the DTASelect). Other possible modifications were also applied during the SEQUEST search, including oxidation on methionine (+16), deamination of glutamine and asparagine (+1) and dehydration of serine and threonine (−18). All targeted peptides spectra were examined by manual inspection.

Site-directed Mutagenesis and Protein Expression of Recombinant Paraoxonase 1

Recombinant PON1 proteins used in this experiment were based on the G3C9 PON1 clone^30,31,72 with an M75K mutation to match the sequence of human paraoxonase 1 and a six-His tag appended at the C-terminus to facilitate protein purification. To incorporate the ncAAs (nitroTyr and chloroTyr) at Tyr71 position, the codon for Tyr71 was replaced with a TAG stop codon. The expression vector for the WT recombinant PON1 construct (G3C9 clone with an M75K mutation) was transformed into E. coli BL21(DE3) pLysS strain. The expression vectors for recombinant PON1–71TAG were co-transformed with pDule2-HaloTyrRS into BL21(DE3) pLysS. In this case recombinant PON1–71-chloroTyr and recombinant PON1–71-nitroTyr were produced using the same pDule2-HaloTyrRS plasmid. The purification procedure of these recombinant PON1 proteins is as described previously.^30,31

Paraoxonase Activity Assays for Recombinant PON1 Mutants

Recombinant PON1 wildtype and ncAA insertional mutants were pre-incubated with rHDLs (2 equivalence) in PON1 activity buffer (50 mM Tris, 50 mM NaCl and 1 mM CaCl₂, pH 8.0) at 37 °C for 30 min. Aliquots were taken and the paraoxonase activity was determined as previously described.^30,31 All samples were assayed in triplicate and the statistical significance was analyzed by unpaired t-test with Welch’s correction.

Construction of pAcBac for Expression of Halotyrosine-containing Proteins in Eukaryotic Cell Culture

The pAcBac1.tR4-MbPyl plasmid was a gift from Peter Schultz (Addgene plasmid # 50832).⁷³ The WT human codon-optimized Mb RS and the mCherry-TAG-EGFP-HA reporter plasmid were a kind gift from Jason Chin.⁴⁹ The construction of the haloTyr-RS expression vector was done in the same manner as our previous report⁴⁷ using the primers (Supporting Table 2).

Eukaryotic Cell Viability Assessment

The effect of chloroTyr and bromoTyr on the HEK293T cell viability was measured and analyzed using the CellTiter Glo assay kit (Promega) as previously described.⁴⁷

Transfection and Imaging

HEK293T cells were seeded in a 24-well plate (Greiner) and transfected for 24 h ~ 48 h with pAcBac1-haloTyrRS and pAcBac1-sfGFP-150TAG plasmid DNA in the presence of 0.1 mM chloroTyr, 0.1 mM bromoTyr, or vehicle using Lipofectamine 2000 (Thermo Fisher Scientific) according to the manufacturer’s protocol. PON1 expression was carried out in the presence of 1 mM chloroTyr to enhance the suppression efficiency. Fluorescence and phase contrast images were captured using the EVOS FL imaging system (Thermo Fisher) or the Keyence BZ-X710 all-in-one fluorescence microscope (Keyence Corporation, IL)

Flow Cytometry

GFP fluorescence of HEK293T cells was examined by the CytoFlex flow cytometer as described previously.⁴⁷

In-gel Fluorescence Assay and Densitometry

HEK293T crude cell lysates were prepared in non-reducing Laemmli buffer and resolved by SDS-PAGE. GFP fluorescence was recorded on Bio-Rad ChemiDoc MP imaging system using blue epi excitation and 530/28 nm emission filter. The band intensity was quantified by densitometric analysis using ImageLab 4.0 software (Bio-Rad).

Eukaryotic ncAA-Protein Mass Spectrometry

sfGFP proteins purified from HEK293T cells were analyzed by mass spectrometry as described previously.⁴⁷

Western Blot Analysis

Detection of recombinant PON1 and nitroTyr PTM was done by western blot using rabbit anti-PON1 antibody (EPITOMICS, 2835–1; 1:10,000) and rabbit anti-nitrotyrosine antibody (Millipore, 06–284; 1:2,000) respectively. The westerns were performed using chemiluminescence as described.^29–31

Purified sfGFP proteins containing 3-nitro-, 3-chloro- and bromoTyr at position 150 were expressed in E. coli and purified as described above and desalted into 50 mM sodium phosphate, 100 mM NaCl pH 7.0 using a PD-10 desalting column (GE Healthcare) prior to oxidation. To prepare soluble extracts, BL21ai cells containing an ApoA1 expression plasmid (pBAD-ApoA1) were expressed similarly for sfGFP and lysed by a microfluidizer also as described above. Soluble extracts were desalted into the 50 mM sodium phosphate, 100 mM NaCl pH 7.0 before exposure to oxidants.

HOBr and HOC1 oxidants were prepared and quantified as previously described.⁷⁴ Oxidation reactions were carried out in phosphate buffer pH 7.0 at room temperature for 20 min in 100 μL volume containing 0.25 mg/ml sfGFP and 0.2 mM oxidant. Reactions were quenched by the addition of 20 mM methionine. Approximately 2 μg of purified protein was loaded onto an SDS-PAGE gel. Cell extracts were oxidized similarly except that 2 mM oxidant was reacted with 50 μg of total cellular protein, of which 10 μg was loaded onto an SDS-PAGE gel. SDS-PAGE gels were transferred to a nitrocellulose membrane, which was blocked with 5% milk in TBST (25 mM Tris, 150 mM NaCl, 0.1 % [v/v] Tween 20) for 2 hours at room temperature. Anti-His₆, anti-chloroTyr, anti-nitroTyr, and anti-haloTyr antibodies were diluted 1:500, 1:50, 1:500 and 1:200 into 1% milk/TBST, respectively and incubated with membranes overnight at room temperature. Membranes were washed with TBST, incubated with IRDye® 800CW secondary antibodies (LiCOR, Lincoln NE) diluted 1:10,000 in 1% milk/TBST for 1 hour, washed with TBST, and imaged with a LiCOR imager.