Post-translational modifications (PTMs) greatly expand the structures and functions of proteins in nature1,2. Although synthetic protein functionalization strategies allow mimicry of PTMs3,4, as well as formation of unnatural protein variants with diverse potential functions, including drug carrying5, tracking, imaging6 and partner crosslinking7, the range of functional groups that can be introduced remains limited. Here we describe the visible-light-driven installation of side chains at dehydroalanine residues in proteins through the formation of carbon-centred radicals that allow C–C bond formation in water. Control of the reaction redox allows site-selective modification with good conversions and reduced protein damage. In situ generation of boronic acid catechol ester derivatives generates RH2C• radicals that form the native (β-CH2–γ-CH2) linkage of natural residues and PTMs, whereas in situ potentiation of pyridylsulfonyl derivatives by Fe(ii) generates RF2C• radicals that form equivalent β-CH2–γ-CF2 linkages bearing difluoromethylene labels. These reactions are chemically tolerant and incorporate a wide range of functionalities (more than 50 unique residues/side chains) into diverse protein scaffolds and sites. Initiation can be applied chemoselectively in the presence of sensitive groups in the radical precursors, enabling installation of previously incompatible side chains. The resulting protein function and reactivity are used to install radical precursors for homolytic on-protein radical generation; to study enzyme function with natural, unnatural and CF2-labelled post-translationally modified protein substrates via simultaneous sensing of both chemo- and stereoselectivity; and to create generalized ‘alkylator proteins’ with a spectrum of heterolytic covalent-bond-forming activity (that is, reacting diversely with small molecules at one extreme or selectively with protein targets through good mimicry at the other). Post-translational access to such reactions and chemical groups on proteins could be useful in both revealing and creating protein function.
Subscribe to Journal
Get full journal access for 1 year
only $3.90 per issue
All prices are NET prices.
VAT will be added later in the checkout.
Tax calculation will be finalised during checkout.
Rent or Buy article
Get time limited or full article access on ReadCube.
All prices are NET prices.
Key raw MS data and primary numerical data for graphical plots have been deposited in the open-access depository ORA-data (https://doi.org/10.5287/bodleian:9ewjQ268q) and all raw data are available from the corresponding authors upon request. MS/MS raw data files have been uploaded to the PRIDE repository (accession number PXD019565, https://www.ebi.ac.uk/pride/archive). The following additional databases were used: MaxQuant contaminants database (https://www.maxquant.org) and Uniprot Human Database (https://www.uniprot.org/proteomes/UP000005640). Source data are provided with this paper.
Walsh, C. T., Garneau-Tsodikova, S. & Gatto, G. J., Jr. Protein posttranslational modifications: the chemistry of proteome diversifications. Angew. Chem. Int. Ed. 44, 7342–7372 (2005).
Deribe, Y. L., Pawson, T. & Dikic, I. Post-translational modifications in signal integration. Nat. Struct. Mol. Biol. 17, 666 (2010).
Howard, C. J., Yu, R. R., Gardner, M. L., Shimko, J. C. & Ottesen, J. J. Chemical and biological tools for the preparation of modified histone proteins. Top. Curr. Chem. 363, 193–226 (2015).
Yang, A., Cho, K. & Park, H.-S. Chemical biology approaches for studying posttranslational modifications. RNA Biol. 15, 427–440 (2018).
Ducry, L. & Stump, B. Antibody−drug conjugates: linking cytotoxic payloads to monoclonal antibodies. Bioconjug. Chem. 21, 5–13 (2010).
Hinner, M. J. & Johnsson, K. How to obtain labeled proteins and what to do with them. Curr. Opin. Biotechnol. 21, 766–776 (2010).
Leitner, A. et al. Probing native protein structures by chemical cross-linking, mass spectrometry, and bioinformatics. Mol. Cell. Proteomics 9, 1634 (2010).
Wang, L., Brock, A., Herberich, B. & Schultz, P. G. Expanding the genetic code of Escherichia coli. Science 292, 498 (2001).
Dumas, A., Lercher, L., Spicer, C. D. & Davis, B. G. Designing logical codon reassignment – expanding the chemistry in biology. Chem. Sci. 6, 50–69 (2015).
Chin, J. W. Expanding and reprogramming the genetic code. Nature 550, 53–60 (2017).
Klemes, Y., Etlinger, J. D. & Goldberg, A. L. Properties of abnormal proteins degraded rapidly in reticulocytes. Intracellular aggregation of the globin molecules prior to hydrolysis. J. Biol. Chem. 256, 8436–8444 (1981).
Chalker, J. M. & Davis, B. G. Chemical mutagenesis: selective post-expression interconversion of protein amino acid residues. Curr. Opin. Chem. Biol. 14, 781–789 (2010).
Wright, T. H., Vallée, M. R. J. & Davis, B. G. From chemical mutagenesis to post-expression mutagenesis: a 50 year Odyssey. Angew. Chem. Int. Ed. 55, 5896–5903 (2016).
Wright, T. H. et al. Posttranslational mutagenesis: a chemical strategy for exploring protein side-chain diversity. Science 354, aag1465 (2016).
Yang, A. et al. A chemical biology route to site-specific authentic protein modifications. Science 354, 623–626 (2016).
Tamura, T. & Hamachi, I. Chemistry for covalent modification of endogenous/native proteins: from test tubes to complex biological systems. J. Am. Chem. Soc. 141, 2782–2799 (2019).
Wright, T. H. & Davis, B. G. Post-translational mutagenesis for installation of natural and unnatural amino acid side chains into recombinant proteins. Nat. Protoc. 12, 2243–2250 (2017).
Sletten, E. M. & Bertozzi, C. R. Bioorthogonal chemistry: fishing for selectivity in a sea of functionality. Angew. Chem. Int. Ed. 48, 6974–6998 (2009).
Imiołek, M. et al. Selective radical trifluoromethylation of native residues in proteins. J. Am. Chem. Soc. 140, 1568–1571 (2018).
Isenegger, P. G. & Davis, B. G. Concepts of catalysis in site-selective protein modifications. J. Am. Chem. Soc. 141, 8005–8013 (2019).
Lim, R. K. V. & Lin, Q. Photoinducible bioorthogonal chemistry: a spatiotemporally controllable tool to visualize and perturb proteins in live cells. Acc. Chem. Res. 44, 828–839 (2011).
Twilton, J. et al. The merger of transition metal and photocatalysis. Nat. Rev. Chem. 1, 0052 (2017).
Prier, C. K., Rankic, D. A. & MacMillan, D. W. C. Visible light photoredox catalysis with transition metal complexes: applications in organic synthesis. Chem. Rev. 113, 5322–5363 (2013).
Bloom, S. et al. Decarboxylative alkylation for site-selective bioconjugation of native proteins via oxidation potentials. Nat. Chem. 10, 205 (2018).
Yu, Y. et al. Chemoselective peptide modification via photocatalytic tryptophan β-position conjugation. J. Am. Chem. Soc. 140, 6797–6800 (2018).
de Bruijn, A. D. & Roelfes, G. Chemical modification of dehydrated amino acids in natural antimicrobial peptides by photoredox catalysis. Chem. Eur. J. 24, 11314–11318 (2018).
Povie, G. et al. Catechols as sources of hydrogen atoms in radical deiodination and related reactions. Angew. Chem. Int. Ed. 55, 11221–11225 (2016).
Matsui, J. K., Lang, S. B., Heitz, D. R. & Molander, G. A. Photoredox-mediated routes to radicals: the value of catalytic radical generation in synthetic methods development. ACS Catal. 7, 2563–2575 (2017).
Robole, Z. M., Rahn, K. L., Lampkin, B. J., Anand, R. K. & VanVeller, B. Tuning the electrochemical redox potentials of catechol with boronic acid derivatives. J. Org. Chem. 84, 2346–2350 (2019).
Ghosh, T. et al. Mixed-ligand complexes of ruthenium(ii) containing new photoactive or electroactive ligands: synthesis, spectral characterization and DNA interactions. J. Biol. Inorg. Chem. 10, 496 (2005).
Dolbier, W. R. Structure, reactivity, and chemistry of fluoroalkyl radicals. Chem. Rev. 96, 1557–1584 (1996).
Zhang, L., Dolbier, W. R., Sheeller, B. & Ingold, K. U. Absolute rate constants of alkene addition reactions of a fluorinated radical in water. J. Am. Chem. Soc. 124, 6362–6366 (2002).
O’Hagan, D. Understanding organofluorine chemistry. An introduction to the C–F bond. Chem. Soc. Rev. 37, 308–319 (2008).
Lemos, A., Lemaire, C. & Luxen, A. Progress in difluoroalkylation of organic substrates by visible light photoredox catalysis. Adv. Synth. Catal. 361, 1500–1537 (2019).
Rong, J. et al. Radical fluoroalkylation of isocyanides with fluorinated sulfones by visible-light photoredox catalysis. Angew. Chem. Int. Ed. 55, 2743–2747 (2016).
Berlicki, L., Obojska, A., Forlani, G. & Kafarski, P. Design, synthesis, and activity of analogues of phosphinothricin as inhibitors of glutamine synthetase. J. Med. Chem. 48, 6340–6349 (2005).
Griller, D. & Ingold, K. U. Free-radical clocks. Acc. Chem. Res. 13, 317–323 (1980).
Chen, Y., Kamlet, A. S., Steinman, J. B. & Liu, D. R. A biomolecule-compatible visible-light-induced azide reduction from a DNA-encoded reaction-discovery system. Nat. Chem. 3, 146–153 (2011).
Huang, H. et al. Lysine benzoylation is a histone mark regulated by SIRT2. Nat. Commun. 9, 3374 (2018).
Dyer, P. N. et al. Reconstitution of nucleosome core particles from recombinant histones and DNA. Methods Enzymol. 375, 23–44 (2003).
Page, M. I. & Jencks, W. P. Entropic contributions to rate accelerations in enzymic and intramolecular reactions and the chelate effect. Proc. Natl Acad. Sci. USA 68, 1678–1683 (1971).
Krishnamurthy, V. M., Semetey, V., Bracher, P. J., Shen, N. & Whitesides, G. M. Dependence of effective molarity on linker length for an intramolecular protein−ligand system. J. Am. Chem. Soc. 129, 1312–1320 (2007).
Ng, S. S. et al. Crystal structures of histone demethylase JMJD2A reveal basis for substrate specificity. Nature 448, 87–91 (2007).
English, C. M., Adkins, M. W., Carson, J. J., Churchill, M. E. & Tyler, J. K. Structural basis for the histone chaperone activity of Asf1. Cell 127, 495–508 (2006).
Meeusen, J. W., Tomasiewicz, H., Nowakowski, A. & Petering, D. H. TSQ (6-methoxy-8-p-toluenesulfonamido-quinoline), a common fluorescent sensor for cellular zinc, images zinc proteins. Inorg. Chem. 50, 7563–7573 (2011).
Freedman, H. H. & Dubois, R. A. An improved Williamson ether synthesis using phase transfer catalysis. Tetrahedr. Lett. 16, 3251–3254 (1975).
Mandal, S. et al. A review on the advancement of ether synthesis from organic solvent to water. RSC Adv. 6, 69605–69614 (2016).
Levin, M., Stark, M. & Assaraf, Y. G. The JmjN domain as a dimerization interface and a targeted inhibitor of KDM4 demethylase activity. Oncotarget 9, 16861–16882 (2018).
Shin, S. & Janknecht, R. Diversity within the JMJD2 histone demethylase family. Biochem. Biophys. Res. Commun. 353, 973–977 (2007).
Karle, I. L. & Balaram, P. Structural characteristics of .alpha.-helical peptide molecules containing Aib residues. Biochemistry 29, 6747–6756 (1990).
Lonsdale, R. & Ward, R. A. Structure-based design of targeted covalent inhibitors. Chem. Soc. Rev. 47, 3816–3830 (2018).
Angerani, S. & Winssinger, N. Visible light photoredox catalysis using ruthenium complexes in chemical biology. Chem. Eur. J. 25, 6661–6672 (2019).
Yang, B. et al. Genetically introducing biochemically reactive amino acids dehydroalanine and dehydrobutyrine in proteins. J. Am. Chem. Soc. 141, 7698–7703 (2019).
Renaud, P., André-Joyaux, E., Kuzovlev, A. & Tappin, N. D. A general approach to deboronative radical chain reaction with pinacol alkylboronic esters. Angew. Chem. Int. Ed. 59, 13859 (2020).
Li, Q. et al. Developing covalent protein drugs via proximity-enabled reactive therapeutics. Cell 182, 85–97 (2020).
This research has received funding from the EPSRC (EP/V011359/1), UK Catalysis Hub (EPSRC Portfolio Grant EP/K014668/1; B.G.D., C.F.), the Swiss National Science Foundation (P2BSP2_178609; P.G.I.), BBSRC (BB/P026311/1; B.G.D., V.G., P.G.I.), Oxford-GSK-Crick Chemical Biology Centre for Doctoral Training Programme (EPSRC, GSK to G.R.) via the EPSRC Systems Approaches to Biomedical Science DTC (EP/R512333/1), Oxford Clarendon Scholarship (to B.J.), Rutherford Foundation (to T.H.W.), UCB (to B.J.B.), Brunei Government Scholarship (to A.W.J.P.) and EU H2020 under Grant Agreement 721902 (to O.A.). We thank S. Hester for experimental support in mass spectrometry, T. Mollner and M. Imiołek for providing small-molecule substrates, S. Faulkner for providing protein, the Chemistry Department workshop for construction of the photoreactor and C. am Ende, W. Stockdale and M. Moomersteeg for discussions.
A patent is being filed that might afford authors royalties were it to be licensed.
Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Extended data figures and tables
a, Retrosynthetic analysis highlights the chemoselective advantages of using carbon radical reagents paired with the radical acceptor Dha (right) over the typical heterolytic 2e− reagents (left) for the site-specific modification of proteins. b, c, Our previous work on radical addition to Dha14,17 (b) and the mechanism highlighting unwanted side reactions (red) (c). d, e, Summaries of previous works24,25 using photocatalysts for ‘on-peptide’ radical generation for site-selective peptide modification, highlighting their limitations and potential for undesired side reactions (red).
a, Oxidative-half potential (Eox) spectrum showing catalyst compatibility with protein-based chemistry for relevant catalysts found in the literature23,24 (bottom and top) and tested in this work (top). b, c, In situ formation of BACED reagents (for side chains 1, yellow highlight) advantageously allows Ruii-catalysed, low-Eox activation (compared to other derivatives) to RCH2• radicals, which then react with Dha in proteins to install side chains. Independent and mixed voltammetric responses of 1 mM catechol and 12 mM phenethylboronic acid on glassy carbon (GC) in PBS, pH 7.10 (inset). See also Extended Data Fig. 4 and Supplementary Discussion 2, 3 for more detailed electrochemical experiments. Intact protein LC–MS (bottom right; chromatogram and mass-to-charge ratio, m/z) shows homohomophenylalanine (1h) installation into histone H3 protein. d, Ruii-catalysed activation of pySOOF reagents to RCF2• radicals, which then react with Dha in proteins to install ‘zero-size’-labelled side chains. The added Feii drives an unprecedented efficiency (2–5 equiv. of precursor) by suppressing oxidation by Ruii* to imine (and hydrate), suggesting the key role of Feii as a reductant (readily available in biology) that quenches the α-carbon radical adduct generated during the reaction. Intact protein LC–MS shows that difluoroethylglycine (DfeGly, 2a) installation into histone H3 protein is successful with Feii (full conversion; top right, chromatogram and m/z) but not without iron (poor conversion to unwanted side products; bottom centre); see also Extended Data Figs. 5–7 for further details. For the full reaction scope of all side chains (types 1 and 2) edited into proteins, including those allowing previously inaccessible on-protein reactivity, see Extended Data Fig. 8 and Figs. 2, 3.
a, A 100% stacked bar chart (n = 1, with single data values represented by the y-axis span of the corresponding bars) showing the results of initial studies of the oxidation of benzyltrifluoroborate with different catalyst strengths and additives to achieve selective single addition. Catechol increased the reactivity for all catalysts, leading to the emergence of reactivity with the weakest Cat1 and almost complete conversion to the double-addition product with Cat3, whereas NaCNBH3 with Cat3 successfully quenched the α-carbon radical promoting single addition. b, Trends of increasing oxidative damage and decreasing reaction control with higher-*Eox catalysts with representative LC–MS ion series and spectra (see Supplementary Methods for details). c, A 100% stacked bar chart (n = 1, with single data values represented by the y-axis span of the corresponding bars) comparing Cat1 and Cat2 reaction conversions to the single-addition product and unwanted α-carbon radical quenching (via either double addition or catechol quenching) for different boronate substrates, arranged with higher-Eox primary boronate substrates on the left and increasingly stabilized radical precursors on the right. The trends suggest the utility or necessity of using the stronger Cat2 (*Eox = +0.99 V) for the primary boronate substrates and the increased efficiency when using the weaker Cat1 (*Eox = +0.77 V) for the more stabilized radical precursors (secondary and benzyl substrates), probably owing to a slow, controlled release of stable radicals, ensuring efficient addition to Dha instead of self-quenching or overalkylation. d, A light-intensity screen shows that increasing light intensity (450 nm blue LED, 0–50 W) allows high conversion efficiencies to the desired single-addition product with shorter reaction times and lower concentrations of Cat1 (Supplementary Table 5; n = 1; best-fit line overlayed). e, Reaction scheme and LC–MS spectra for the installation of iodonorleucine (Inl). Although Inl installation had poor initial conversion (~50%), the only other species present after the reaction was the starting material (Dha), allowing successive reactions to increase conversion efficiencies to ~75%. f, A screen of catechol derivatives finds that the naturally occurring catecholamines dopamine and l-DOPA could efficiently substitute for the role of catechol.
a, Reaction scheme screening different quinone derivatives for their influence on the oxidation of potassium phenethyltrifluoroborate and subsequent addition to Dha. b, Potential mechanism of catechol derivatives acting as redox mediators to bridge electron transfer between catalyst and substrate. c, LC–MS results of the quinone derivative screening (shown in a and b) ruled out the mechanism in b because only 1,2-diols showed substantial activity, with only catechol avoiding protein degradation. d, A potential mechanism of in situ catalyst modification with catechol, creating catecholo-Ru(bpy)2 (Cat6)30 was ruled out because it did not promote alkylation with or without the addition of exogenous catechol. e, In situ formation of a reactive boronic acid catechol ester is suggested. f, 1-Propylboronic acid catechol ester and 4-bromobutylboronic acid catechol ester were successfully added without the addition of exogenous catechol. This suggests that the formation of the catechol ester lowers the Eox value of the substrate to a range accessible by Cat1 (*Eox = +0.77 V). g, Voltammetric response of 1 mM catechol in the presence of increasing concentrations of 4-bromobutylboronic acid (black, 0 mM; brown, 3 mM; green, 6 mM; blue, 12 mM; red, 24 mM) at a glassy carbon macroelectrode in 50 mM phosphate buffer (pH 6) recorded at 100 mV s−1. h, Voltammetric response of catechol only (1 mM; black), 4-bromobutylboronic acid (12 mM; blue), and 4-bromobutylboronic acid in the presence of catechol (12 mM and 1 mM, respectively; red) recorded at 100 mV s−1. i, Simulated voltammetric response for the oxidation of 4-bromobutylboronic acid (12 mM) in the presence of catechol (1 mM) at 100 mV s−1, following the simplified mechanism outlined in e. The simulation highlights the importance of the oxidation of the boronic acid ester being catalytic and leading to the reformation of catechol (see Supplementary Discussion 2, 3). The rate of decomposition of the radical ester has been set at either 1 × 104 s−1 or 0 s−1 (blue and red, respectively). If the oxidation of the boronic acid ester is not catalytic, no peak is predicted to be voltammetrically observable at >1 V versus a saturated calomel reference electrode (SCE). j, Voltammetric response of 4-bromobutylboronic acid (12 mM) in the presence of catechol (1 mM) in 50 mM phosphate buffer (pH 6) as a function of scan rate (25−400 mV s−1). k, l, Voltammetric response of preformed 4-bromobutylboronic acid catechol ester (1 mM) (k) and catechol (1 mM) in 50 mM phosphate buffer (pH 6) (l) as a function of scan rate (25−400 mV s−1).
Extended Data Fig. 5 Initial experiments without iron using various hydride sources, and optimization study with sodium borohydride for pySOOF.
a, The formation of oxidative-derived side products H3-imine-DfeGly9 and H3-hemiaminal-DfeGly9 was detected in an initial, additive-free, light-driven protein-modification reaction with pySOOF and Ru(bpy)3Cl2. On the basis of this observation a mechanism was postulated, in which the on-protein radical intermediate is oxidized by the excited state of the photocatalyst. To avoid this reaction pathway, the use of hydride donors such as silanes, tertiary amines or borohydrides were assumed to favour the formation of DfeGly-modified histone at site 9. b, Inferior results were observed when 250 equiv. DIPEA, TTMS and catechol (100 equiv.) were used (with 200 equiv. pySOOF and 5 equiv. Ru(bpy)3) (Supplementary Tables 8, 11, 12). For catechol, excellent conversion efficiencies were observed; however, catechol-DfeGly-modified protein was detected as a major side product. Borohydrides such as Na(OMe)3BH (0.25 mg) and NEt4BH4 (0.25 mg) showed promising reactivity, but only the oxidation-derived side products were formed. Only with sodium borohydride was the desired DfeGly-modified histone formed in moderate conversion (Supplementary Table 8). c, Reduction of Dha to Ala by NaBH4 was identified as a potential limitation for the NaBH4-mediated reaction. To test this undesired pathway, a crude reaction mixture was buffer-exchanged to NaPi (100 mM, pH 9, 3 M GdnHCl) and incubated with a large excess of 2-mercaptoethanol. After 1 h at 37 °C, no formation of the corresponding thiol Michael-type protein adduct was detected, proving Ala formation. The 100% stacked bar graphs (n = 1, with single data values represented by the y-axis span of the corresponding bars) summarize the results of the optimization studies of the NaBH4-mediated photochemical reaction with different concentrations of NaBH4, photocatalyst and pySOOF and reaction times (Supplementary Tables 9, 10). Less than 500 equiv. NaBH4 resulted in the formation of undesired oxidative side products. Increasing the reagent concentration, photocatalyst loading or reaction times did not improve the reaction. d, Lower temperature increased the conversion, most probably by slowing down the reduction of Dha by NaBH4.
Extended Data Fig. 6 Optimization study of Fe(ii)-mediated protein modification reaction with pySOOF.
a, In the photochemical modification reaction with pySOOF, Fe(ii) likely acts as a reductive quencher for the photoredox cycle and as a single-electron reductant of the on-protein radical intermediate forming the enolate intermediate. However, side products, such as H3-imine-DfeGly9, H3-hemiaminal-DfeGly9 or H3-diDfeGly9, can be generated because of inefficient quenching of the on-protein radical intermediate. b, In an initial experiment with 200 equiv. pySOOF, 280 equiv. FeSO4, 5 equiv. Ru(bpy)3Cl2 and 66 μM histone H3-Dha9 with a protein concentration of 0.5 mg ml−1, 70% conversion to a mixture of H3-DfeGly9 and H3-diDfeGly (57:43) was observed. The formation of the mono-addition product was favoured at higher protein concentration (1 mg ml−1) and the conversion was increased to 92% (Supplementary Table 13). c, For the Fe(ii)-mediated reaction, various metallo- and organophotocatalysts with different *Ered values (−0.56 to −1.37 V) and radical precursors were tested, and Ru(bpy)3 and pySOOF were identified as the best combination (Supplementary Table 17). d, The 100% stacked bar charts (n = 1, single data values represented by the y-axis span of the corresponding bars) summarize the results of the optimization studies of the FeSO4-mediated photochemical reaction with different reaction times, concentrations of pySOOF and FeSO4, and catalytic amounts of Fe(ii) and photocatalyst, respectively (Supplementary Tables 10–15, 18, 20). Short reaction times and high efficiency with low concentrations of pySOOF were found. However, in cases with high levels of reactivity with <100 equiv. FeSO4, only oxidation-derived products were created, indicating the dual role of Fe(ii) for the single-electron reduction of the on-protein radical intermediate. e, With the optimized conditions in hand, good to excellent conversion efficiencies to histone H3-DfeGly9 were obtained at various protein concentrations (0.1–5 mg ml−1). SET, single-electron transfer.
a, Effect of substituents on the α centre of the created radical on reactivity, efficiency and selectivity (single versus double addition). No reactivity was observed for non- and mono-fluorinated precursors with a H, Me or Cl substituent. With an increase in the electron-withdrawing effect of the additional substituent, radical generation was observed and the resulting products were formed. Moreover, the highest levels of reactivity were observed for •CF2R. b, An equimolar mixture of three pySOOF substrates was applied to the protein-modification reaction to assess the reactivity order between •CF2H, •CF2Me and •CFHCONH2. During this competition experiment, the reactivity trend •CF2H > •CF2Me was identified, and no product formation was observed for •CFHCONH2. c, On the basis of the reactivity study of various mono- and difluoro-pySOOF reagents, a suitable radical precursor was designed to allow the efficient generation of •CF2R for the installation of difluorinated- amino acid residues or PTMs. d, With iodo-pySOOF, homolytic bond cleavage between iodine and the carbon centre was induced, and the pySOOF unit was installed on the Dha-tagged histone (Supplementary Tables 22–23). After further photoredox activation, a captodative-effect-stabilized on-protein radical was formed, which then allowed further on-protein chemical reactions by trapping the radical with various acceptors (Supplementary Tables 24–37). Moreover, the homolytic iodo-carbon bond cleavage inspired the design of bromo-difluoro carbonyl-based radical precursors for the installation of CF2Gln- or CF2Glu-derived amino acids. e, Mechanistic studies supported a radical mechanism, as no product formation was observed with 4-hydroxy-TEMPO (TEMPOL). Furthermore, a photocatalytic process was confirmed, as no conversions were detected without a photocatalyst or blue LED irradiation (Supplementary Table 16). On–off experiments showed only product formation during the irradiation period, excluding a radical chain mechanism. Finally, good reactivity (80% conversion) was observed when the sample was prepared under ambient atmosphere; however, levels of oxidative damage (Met oxidation) and oxidative-derived DfeGly-modified histones were also generated.
a, A comprehensive list of all side chains installed into proteins with the BACED reaction manifold. General conditions: protein 1 mg ml−1; 50 W, 450 nm light; 4 °C to RT; 100–1,500 equiv. BACED precursor reagent; 10 equiv. Cat1 or Cat2; 100 equiv. catechol; <6 ppm O2; pH 6.0 buffer (500 mM NH4OAc or PBS ± 3 M GdnHCl). b, A comprehensive list of all side chains installed into proteins with the pySOOF reaction manifold. General conditions: protein 1 mg ml−1; 50 W, 450 nm light; RT; 2–5 equiv. pySOOF precursor reagent; 0.4–4 equiv. Cat1; 50–100 equiv. FeSO4; <6 ppm O2; pH 6.0 buffer (500 mM NH4OAc or various other buffers).
The scalability of protein modification reaction with pySOOF was studied and high conversion efficiencies were observed for pySOOF-Lys, LysAc, Lys(Me)3, Met, Glu and DfeGly using 4–6 mg of Dha-tagged histone. After the photochemical reaction, the crude mixture was vortexed with either EDTA or DTT, followed by buffer exchange to deuterated buffer (NH4OAc, 250 mM, pH 7 in D2O with TFA as internal standard) using desalting columns. After purification, excellent yields were determined by checking the protein concentration by absorbance at 280 nm on a micro-volume spectrophotometer. Spectra were recorded on a Bruker NMR system (AV600) and all PTMs showed characteristic peaks with unique chemical shifts.
a, Milligram-scale chemical mutagenesis to generate the substrate histone eH3.1-CF2LysAc18, in order to investigate deacetylation by Sirt2 by comparing the 19F NMR analysis of the starting material and after enzymatic reaction. The desired difluoro-labelled histone was formed in excellent conversion, with low levels of Met oxidation. After the deacetylation, histone eH3.1-CF2LysAc18 and histone eH3.1-CF2Lys were detected by LC–MS, revealing that the fluorinated histone was accepted as substrate for Sirt2. b, c, 19F NMR spectra were acquired for ‘start’ and ‘product’ states, as well as when the solution containing start was added to the enzyme Sirt2 (reaction). The spectra of the start and product states consisted of an AB quartet (D) and an apparent singlet state (L). The spectrum of the reaction contained the singlet state of the product and the AB quartet of the start, suggesting that the singlet state almost completely reacted with Sirt2, whereas the species that gave rise to the AB quartet did not. To quantify this, the start and product spectra were simulated to determine the relevant resonance frequencies and coupling constants, and the ‘reaction’ spectrum was simulated by taking the species identified in start and product but with scaled intensities. d, The histone H3-DfeGly9 sample had its 19F NMR spectra recorded as an unfolded protein species in unfolding buffer (7 M GdnHCl, 10 mM Tris, 1 mM EDTA, 10 mM DTT, 1 mM benzamidine, pH 7.5, 50% D2O, 0.1 μl trifluoroethanol internal standard), folded species in Tris buffer (150 mM NaCl, 10 mM Tris, 1 mM EDTA, 2 mM βME, pH 7.5, 50% D2O, 2 mg ml−1 protein, 0.75 ml and 1 μl of 0.1% trifluoroethanol internal standard), reconstituted histone H3-H4 tetramer in refolding buffer (2.5 mg ml−1 tetramer, 1 ml buffer, 50% D2O, 0.1 μl trifluoroethanol internal standard) and reconstituted histone H2A-H2B-H3-H4 octamer in refolding buffer (1.6 mg ml−1 octamer, 0.5 ml buffer, 50% D2O, 0.1 μl trifluoroethanol internal standard). All the recorded 19F NMR spectra were compared and small changes in the chemical shifts were observed. Interestingly, the biggest changes were detected between the unfolded and folded histone H3 and after tetramer formation, potentially owing to changes in tumbling rate. LC–MS analysis of the octamer product confirmed the presence of all four histone types.
Extended Data Fig. 11 Effective molarity driven protein–protein crosslinking with electrophile-containing side chains.
a, A comparison of the connectivity of a modified lysine side chain (via methylation) and bromohomonorleucine (Bhn, 1u) used for crosslinking. b, Structure of histone eH3.1-mimicking peptide binding to KDM4A. Identified cysteines that undergo crosslinking (Cys306 and Cys234) are highlighted in orange and their distance to H3-K9 is highlighted. c, Workflow of the crosslinking reaction products and their identification by LC–MS/MS. Representative spectra are shown for each crosslinking site (top, eH3.1-Bhn4; middle, eH3.1-Bhn9; bottom, eH3.1-Bhn27) and captured cysteine residue. The bottom right spectrum shows an inter-histone crosslink via ether formation as found in the in-solution digest. d, Williamson C–O–C bond ether formation in an intermolecular fashion between H3 proteins (Bhn4 in one linked to hydroxyl in another) is driven by effective molarity, possibly suggesting a transient dimer model for KDM4A function. e, The histone eH3.1-Bhn9 alkylator protein was incubated with HeLa nuclear lysate to capture interaction partners via promixity-driven crosslinking. After enrichment via the HA tag (on histone eH3.1), an α-FLAG western blot reveals multiple higher molecular ‘weight’ (MW) bands corresponding to the mass of the histone plus that of the captured interaction partner. No higher MW bands were seen in conditions lacking Bhn.
About this article
Cite this article
Josephson, B., Fehl, C., Isenegger, P.G. et al. Light-driven post-translational installation of reactive protein side chains. Nature 585, 530–537 (2020). https://doi.org/10.1038/s41586-020-2733-7