English

• Lysine methylation is a classic posttranslational modification (PTM) and regulates many cellular activities. For example, histone H3K4me3 promotes gene transcription while histone H3K27me3 represses gene transcription [1–4]. Reader proteins are an important class of proteins that bind site-specific modified proteins for downstream biological consequences [2,5–7]. SGF29 is a reader of H3K4me3 in Spt-Ada-Gcn5-acetyltransferase (SAGA) complex that contains acetyltransferase GCN5. Therefore, the reader recruits SAGA complex to local chromatin and then stimulates hyperacetylation for gene transcription [8,9]. EED is a reader of H3K27me3 that brings polycomb repressive complex 2 (PRC2) to local chromatin, that elevates H3K27 methylation by methyltransferase EZH2 [3,6,10]. Although there are diverse reader domains with different foldings, the methyllysine binding strategy is generally the same by aromatic cages with 2–4 aromatic residues for hydrophobic interactions and π-cation interactions [1,11,12].

  Although significance of methyllysine readers has been demonstrated, there are still some biological important methyllysine sites without known readers [13–15]. Therefore, development of chemical probes that selectively crosslink readers is highly desired in chemical biology studies. Photoreactive groups such as diazirine and p-benzoyl-l-phenylalanine (BPA) have been widely used, but the placement of such groups requires intensive optimization for best crosslinking [16–19]. Also, the highly reactive intermediates from ultraviolet (UV) irradiation may cause non-specific crosslinking that may interfere with identification of reader proteins [20].

  We recently reported a methyllysine reader-selective crosslinking method by sulfonium-containing peptides. NleS⁺me2 (norleucine-ε-dimethylsulfonium) as a Kme2 (dimethyllysine) mimic binds the methyllysine reader to form sulfonium-indole electron donor-acceptor (EDA) complex in the binding pocket. Upon UV irradiation, single-electron transfer (SET) enables selective crosslinking between Nle and Trp. This method offered great opportunities to identify new readers from cell samples [21].

  Tyrosine is another important residue for the binding activity of methyllysine reader and there are some readers without tryptophan in aromatic cages such as SGF29 and EED (Fig. 1a). Although there are many tyrosine-selective bioconjugation methods such as aryl diazonium and sulfur fluoride exchange (SuFEx), they lack site selectivity and are very challenging to crosslink tyrosine inside the pocket due to less exposure (Fig. 1b) [22,23]. In this study, we developed a tyrosine site-selective crosslinking method by sulfonium probes. Selective single-electron transfer between proximate tyrosine and sulfonium enables selective crosslinking to methyllysine readers with broad applications (Fig. 1b).

  Figure 1

  

  Figure 1.  Overview of strategies of tyrosine bioconjugation. (a) Crystal structures of diverse tyrosine-containing methyllysine reader domains complexed with their ligands. The binding pockets contain conserved tyrosine residues. PDB, protein data bank. (b) Approaches of crosslinking to tyrosine residues. Current strategies of tyrosine bioconjugation are with poor site-selectivity and much less reactive to tyrosine inside pockets due to low accessibility. In this work, dimethylsulfonium as a dimethyllysine mimic enables selective photo crosslinking to tyrosine in the binding pocket of readers under UV irradiation.

  DownLoad: Full-Size Img PowerPoint

  In our previous study of sulfonium-tryptophan crosslinking, a UV-B lamp with a 305 nm longpass filter was applied to ensure only tryptophan can be excited among 20 natural amino acid residues [21]. Since phenol group of tyrosine is an electron rich aromatic ring, we proposed that it could also play as an electron donor that is similar to indole for site-selective crosslinking to NleS⁺me2 peptides under UV irradiation.

  CBX1 is a well-known reader of H3K9 methylation that drives formation of heterochromatin [24,25]. The aromatic cage of chromodomain contains three key aromatic residues, Tyr21, Trp42 and Phe45 (Fig. 2a), and we thus prepared recombinant CBX1 (W42Y) mutant to explore Tyr crosslinking by H3K9NleS⁺me2 peptide (Fig. 2b). We next measured binding affinity and isothermal titration calorimetry (ITC) data demonstrated that the W42Y mutant is still active to bind H3K9me2 and H3K9NleS⁺me2 peptides, but at lower level than wildtype CBX1 (Fig. 2c and Fig. S2a in Supporting information). This data also indicates the significance of W42 for methyllysine binding affinity of CBX1.

  Figure 2

  

  Figure 2.  H3K9NleS⁺me2 peptide binds to CBX1(W42Y) and displays selective photo crosslinking to the tyrosine in the binding pocket under 302 nm UV irradiation or basic conditions. (a) The key residues in CBX1 aromatic cage for H3K9me3 binding. (b) Structure of histone H3K9NleS⁺me2 peptide. (c) ITC analysis of CBX1(W42Y) with H3K9me2 and H3K9NleS⁺me2 at pH 7.4. Experiments were conducted under the conditions of 25 µmol/L protein concentration and 750 µmol/L peptide concentration using a MicroCal PEAQ-ITC Automated instrument at 25 ℃. (d) Ultraviolet emission spectra of UV lamp without or with a longpass filter, called 302 nm and 313 nm, respectively. (e) Photo crosslinking between CBX1(W42Y) and H3K9NleS⁺me2 under different UV conditions. CBX1(W42Y) (3 µmol/L) was incubated with H3K9NleS⁺me2 (100 µmol/L) on ice for 15 min in the buffer of pH 7.4 and then crosslinked under 313 nm UV and 302 nm UV for an additional 15 min. The analytical yield of peptidyl conjugate H3K9Nle-CBX1(W42Y) increased from 17.1% to 62.5% when the wavelength of UV source changed from 313 nm to 302 nm. (f) Photo crosslinking between CBX1(Y21F, W42F) and H3K9NleS⁺me2 under 302 nm UV. CBX1(Y21F, W42F) (3 µmol/L) crosslinked H3K9NleS⁺me2 peptide (100 µmol/L) in the buffer of pH 7.4 under 302 nm UV for 15 min. No crosslinking product was detected. (g) Deprotonation of tyrosine residue in the basic condition. (h) Absorption spectra of tyrosine amino acid and ARTESYKA peptide. The absorption peaks red shifted and the absorbance rose as increase of pH value. (i) ITC measurements of CBX1(W42Y) and H3K9NleS⁺me2 at different pH. H3K9NleS⁺me2 (750 µmol/L) was titrated into CBX1(W42Y) (25 µmol/L) at 25 ℃ in the buffers of different pH. (j) Photo crosslinking between CBX1(W42Y) and H3K9NleS⁺me2 at different pH. CBX1(W42Y) (3 µmol/L) was incubated with H3K9NleS⁺me2 (100 µmol/L) on ice for 15 min in the buffers of different pH and then crosslinked under 313 nm UV for an additional 15 min. The crosslinking efficiency of CBX1(W42Y) was promoted by increase of pH value. Methyl-CBX1(W42Y): methyl conjugate, the product through SET along with generation of H3K9NleSme due to alternative homolysis of C-S bond. (k) Photo crosslinking between CBX1(Y21F, W42F) and H3K9NleS⁺me2 in the basic condition of pH 10.1. CBX1(Y21F, W42F) (3 µmol/L) crosslinked H3K9NleS⁺me2 peptide (100 µmol/L) in the basic condition of pH 10.1 under 313 nm UV for 15 min. No crosslinking product was detected.

  DownLoad: Full-Size Img PowerPoint

  Since the binding affinity indicated proximate interactions between tyrosine and sulfonium, we moved forward to explore potential SET and crosslinking activities. Because tyrosine absorbs UV with λ_max around 280 nm, we set two UV conditions either with or without 305 nm longpass filter [26]. Emission spectra demonstrated the lamp covers partial tyrosine absorbance area but the longpass filter minimizes excitation to tyrosine. Due to the distinct emission λ_max, we call them 302 nm UV or 313 nm UV in the following text (Fig. 2d).

  Mixtures of CBX1(W42Y) and H3K9NleS⁺me2 peptide were placed under different UV wavelengths for 15 min irradiation, and the resulting mixtures were analyzed by LC-MS. Peptidyl conjugate of CBX1(W42Y) was detected in both conditions as a mass peak that matches calculated H3K9Nle-CBX1(W42Y) due to crosslinking with a leave of dimethyl sulfide. Analytical yield of the peptidyl conjugate under 302 nm UV was much higher than 313 nm UV due to better tyrosine excitation (Fig. 2e). To confirm the crosslinking selectivity to tyrosine inside methyllysine binding pocket, we prepared CBX1(W42F) and CBX1(Y21F, W42F) mutants. The CBX1(W42F) mutant with one Tyr and two Phe in the aromatic cage was less reactive and the CBX1(Y21F, W42F) mutant with only Phe in the aromatic cage was inactive (Fig. 2f and Fig. S2b in Supporting information). Next, we compared crosslinking under 302 nm UV or 313 nm UV for wildtype CBX1, and the reaction yields were the same (Fig. S2b). We think 313 nm UV offered sufficient excitation for tryptophan excitation so that 302 nm UV did not improve the reaction. Nonetheless, the initial results demonstrated the crosslinking activity between sulfonium and excited tyrosine by UV in the absorption range of tyrosine.

  Tyrosine is a phenol-containing amino acid so that the side chain is a weak acid and can be deprotonated to phenoxide in basic conditions (Fig. 2g). Due to the negative charge with one more delocalized electron, a lower gap between highest occupied molecular orbital (HOMO) and lowest unoccupied molecular orbital (LUMO) with a longer wavelength absorbance is expected. To understand the correlation between pH and UV absorbance, we measured the absorption spectra of tyrosine amino acid and a tyrosine-containing peptide in various buffers. First, the range of absorption wavelength changed from 254–288 nm to 262–313 nm with λ_max shifting from 274 nm to 290 nm. In addition, the absorbance of phenoxide at 290 nm was about 2-fold of phenol at 274 nm, that indicated an improved extinction coefficient (Fig. 2h). We thus think basic conditions may promote photo crosslinking between tyrosine and sulfonium.

  Negatively charged phenoxide may promote the interaction with positively charged sulfonium in the binding pocket. However, the high pH may also affect protein folding to result in loss of binding activities. To evaluate the overall effect of phenoxide, we carried out ITC experiments to measure the binding between sulfonium peptide and CBX1(W42Y) mutant containing different forms of tyrosine, that are from phenol form to phenoxide form based on the pK_a value. The data demonstrated similar affinity with a slight decrease as pH increased (Figs. 2c and i). Since phenoxide-containing pocket bound the sulfonium peptide, we thus screened photo crosslinking reactions in basic buffers. The data demonstrated higher pH increased the yield of peptidyl conjugate as more phenoxide formed in binding pockets (Fig. 2j and Fig. S3b in Supporting information). On the contrary, the crosslinking of wildtype CBX1 did not benefit from high pH. We think tryptophan was much more reactive than tyrosine and contributed to the majority of crosslinking products so that phenoxide of Y21 in the binding pocket did not enhance overall crosslinking (Figs. S3a and b in Supporting information). To further confirm the site-selectivity at basic environment, we applied CBX1(Y21F, W42F) mutant, and no crosslinking product was observed at pH 10.1 (Fig. 2k). Since pH 11.0 did not further improve crosslinking compared to pH 10.1, we selected pH 9.2 and pH 10.1 for further characterization.

  In our previous study, sulfonium-mediated tryptophan crosslinking is binding-dependent because SET requires the formation of EDA complex [21]. To further validate the tyrosine crosslinking reaction is also binding-dependent, we carried out crosslinking reactions in the presence of unfolding agent or binding competition peptides. Guanidine (4 mol/L) resulted in the loss of conjugate, so correct folding is necessary for crosslinking. H3K9me3 peptide as the CBX1 preferential ligand dramatically reduced the crosslinking products while unmodified H3 peptide did not (Fig. S4a in Supporting information). Therefore, binding to sulfonium peptide is a key step for tyrosine crosslinking (Fig. 3a). To further confirm high pH does not cause non-specific crosslinking between sulfonium and phenoxide, we mixed DYGGGSGGGSMK peptide with the H3K9NleS⁺me2 peptide and no crosslinking product was observed by high performance liquid chromatography (HPLC) analysis (Fig. S4b in Supporting information).

  Figure 3

  

  Figure 3.  Characterization of the molecular mechanism of photo crosslinking between CBX1(W42Y) and H3K9NleS⁺me2. (a) Scheme of proposed binding-based crosslinking mechanism. In the hypothesis, photo crosslinking happens inside a protein-ligand (PL) complex which is formed through the binding of CBX1(W42Y) and H3K9NleS⁺me2. (b) Kinetic analysis of photo crosslinking to compare effects made by UV wavelength and pH. CBX1(W42Y) (5 µmol/L) was incubated with H3K9NleS⁺me2 peptide at various concentrations ranging from 12.5 µmol/L to 400 µmol/L and then crosslinked for 15 min under UV irradiation. Average values and errors (±s.e.m.) were calculated from three independent experiments (n = 3).

  DownLoad: Full-Size Img PowerPoint

  We further characterized crosslinking kinetics by alternation of UV source and pH. A group of crosslinking mixtures with different sulfonium peptide concentrations were analyzed. Initial reaction rates calculated from linear range were plotted against peptide concentrations, and Michaelis–Menten equation was applied to calculate apparent K_m and V_max (Fig. 3b). Firstly, in the condition of 313 nm UV at pH 7.4, the reaction conversion was so low that V₀ could not be obtained (Fig. 2e). With enhanced crosslinking by alternative UV source and pH of the reaction buffer, kinetics were analyzable. Secondly, the apparent K_m in all conditions was close and around 40 µmol/L, that matched to similar binding affinity data from ITC. This further validated the binding-dependent crosslinking mechanism. Thirdly, although 302 nm UV and basic environment resulted in higher V_max, it was still much lower than V_max of wildtype CBX1 (1.1 µmol L⁻¹ min⁻¹) [21]. It could explain why wildtype CBX1 crosslinking could not benefit from 302 nm UV and basic conditions.

  With the data above, we proposed that the mechanisms of sulfonium-mediated tyrosine crosslinking and sulfonium-mediated tryptophan crosslinking in reader binding pockets are similar. Due to the binding affinity, sulfonium interacts with tyrosine to form EDA complex. Upon UV irradiation, SET happens between phenol/phenoxide and sulfonium to form neutral sulfur radical. After homolysis of C-S bond and leaving of dimethyl sulfide, the alkyl radical was rebound to phenol for conjugation (Fig. S4c in Supporting information). To further confirm the mechanism, the crosslinking reaction mixture was treated with toluene and the organic phase extract was analyzed by gas chromatography-mass spectrometry (GC–MS). Dimethyl sulfide was unambiguously detected by comparison with the standard sample (Fig. S5 in Supporting information).

  After investigation of crosslinking between tyrosine and sulfonium using CBX1 mutant as a model, we next moved to methyllysine readers of which binding pockets with tyrosine but without tryptophan. SGF29 (reader of H3K4me3 with tandem Tudor domain), UHRF1 (reader of H3K9me3 with tandem Tudor domain) and EED (reader of H3K27me3 with WD40 domain) were applied to crosslink with corresponding sulfonium peptides and Western blot analysis clearly showed the crosslinking activities (Figs. 1a and 4a). SHH1 (reader of H3K9 methylation with SAWADEE domain) and ATRX (reader of H3K9 methylation with ADD domain) were also crosslinked by corresponding sulfonium peptides (Fig. 1a, Figs. S1 and S6a in Supporting information) [6,9,12,27,28]. Generally, reactions at pH 9.2 exhibited best crosslinking. We think this pH is at a balance among phenoxide formation, protein folding and binding affinity.

  Figure 4

  

  Figure 4.  NleS⁺me2 peptide probes enable selective crosslinking to proximate tyrosine widely. (a) NleS⁺me2 peptide probes were active to crosslink their specific methyllysine readers and the crosslinking was promoted in basic conditions. (b) SGF29, UHRF1 and EED were selectively crosslinked by the corresponding sulfonium probes. (c) The crystal structure of NanoBiT. (d) Western blot analysis of photo crosslinking between Flag–SmBiT79(E166Met⁺me) and His₆-LgBiT(W11Y). (e) Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blot analysis of photo crosslinking between Flag-SmBiT79(E166Met⁺me) and His₆-LgBiT(W11Y) in HeLa cell lysate.

  DownLoad: Full-Size Img PowerPoint

  Because sulfonium-mediated crosslinking is binding-dependent, we would expect high crosslinking selectivity among different readers. SGF29 as a reader of H3K4 methylation, was selectively crosslinked by H3K4NleS⁺me2 peptide other than H3K9NleS⁺me2 and H3K27NleS⁺me2 peptides. Similar selectivity results were obtained from UHRF1, EED and SHH1 (Fig. 4b and Fig. S6b in Supporting information). However, the specificity of crosslinking by K9 and K4 sulfonium peptides were not excellent to ATRX (Fig. S6b). We think the unique ATRX binding pocket, that contains two acidic residues but not an aromatic cage, may contribute non-specific interactions between tyrosine and H3K4NleS⁺me2. Therefore, we could conclude that sulfonium-mediated tyrosine crosslinking is highly binding-dependent, so that reader-selective crosslinking is applicable in complicated protein samples.

  So far, sulfonium-tyrosine crosslinking was only applied in methyllysine binding pockets. Based on the reaction mechanism, we proposed that such reaction could be applied to proteins without an aromatic cage. The proximate interactions between tyrosine and sulfonium could be driven by protein-ligand interactions. Therefore, we selected a pair of protein and ligand for following investigation. NanoBiT is a split protein and the subunits LgBiT and SmBiT have high affinity at nmol/L level. Based on the reported 3D structure, there is a close interaction between LgBiT-W11 and SmBiT-E166 (Fig. 4c) [29]. We hence proposed a specific crosslinking between LgBiT-W11Y and SmBiT-E166Met⁺me from our previous successful experience of LgBiT-W11 and SmBiT-E166Met⁺me [21]. Therefore, we prepared the recombinant His₆-LgBiT mutant and Flag-SmBiT79(E166Met⁺me) peptide for crosslinking. The desired conjugate was observed by Western blot (Fig. 4d). Next, we conducted the crosslinking in the presence of HeLa cell lysate and the specific conjugate was detected by Western blot. Although the cell lysate offered a very complicated environment, the crosslinking selectivity was very high with only one non-specific band (Fig. 4e). We thus conclude that sulfonium-tyrosine crosslinking could be bioorthogonal that is not limited to aromatic cages of methyllysine readers.

  Tyrosine residues play important roles in protein folding, ligand binding and catalysis [30]. Although a number of chemical methods have been developed for tyrosine-selective crosslinking, the site-selective strategy is still very limited. Here we developed a site-selective tyrosine photo crosslinking strategy based on proximate interactions between sulfonium and tyrosine. In order to excite tyrosine for single-electron transfer to sulfonium probes, we demonstrated 302 nm UV source and high pH enable efficient tyrosine crosslinking. The reactivity is reader binding-dependent and thus exhibits great site-selectivity. Since it is applicable to methyllysine readers that do not contain tryptophan in binding pockets, we expect most methyllysine readers are capable of crosslinking by sulfonium probes. In addition, we think this strategy can be widely used to other proximity-based tyrosine selective crosslinking by sulfonium beyond aromatic cages. Although sulfonium-based tools are promising in chemical biology research, there are still some limitations for improvement. For example, UV-B is likely to damage protein and cell samples. Therefore, we always limit the irradiation time to less side effects. We may avoid harmful UV by incorporation of photocatalyst for visible light range in the future. In addition, at some point, sulfonium peptides cannot mimic intact proteins. Consequently, new synthetic method that enables sulfonium protein probes and nucleosome probes is our future goal.

  Declaration of competing interest

  The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

  CRediT authorship contribution statement

  Yingxiao Gao: Writing – review & editing, Formal analysis, Data curation. Feng Feng: Data curation. Ting Luo: Data curation. Yusong Han: Data curation. Mingxuan Wu: Writing – original draft, Supervision, Funding acquisition, Conceptualization.

  Acknowledgments

  We thank the support from National Natural Science Foundation of China (No. 22161132006), Key R&D Program of Zhejiang (No. 2024SSYS0036), and Westlake University Startup. We acknowledge the Instrumentation and Service Center for Molecular Sciences for the instrument support. We also thank the Mass Spectrometry and Metabolomics Core Facility of the Biomedical Research Core Facilities at Westlake University. We thank the Instrumentation and Service Center for Physical Sciences for supporting the ITC measurement. We thank Xin Li from Instrumentation and Service Center for Molecular Sciences at Westlake University for the kind assistance in the measurement and analysis of gas chromatography-mass spectrometry.

  Supplementary materials

  Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2024.110756.

  
  
• 1. [1]

     M. Luo, Chem. Rev. 118 (2018) 6656–6705. doi: 10.1021/acs.chemrev.8b00008

  2. [2]

     K.P. Bhat, H. Ü. Kaniskan, J. Jin, O. Gozani, Nat. Rev. Drug Discov. 20 (2021) 265–286. doi: 10.1038/s41573-020-00108-x

  3. [3]

     W. Qi, K. Zhao, J. Gu, et al., Nat. Chem. Biol. 13 (2017) 381–388. doi: 10.1038/nchembio.2304

  4. [4]

     H.T. Li, S. Ilin, W.K. Wang, et al., Nature 442 (2006) 91–95. doi: 10.1038/nature04802

  5. [5]

     T. Wagner, D. Robaa, W. Sippl, M. Jung, ChemMedChem 9 (2014) 466–483. doi: 10.1002/cmdc.201300422

  6. [6]

     R. Margueron, N. Justin, K. Ohno, et al., Nature 461 (2009) 762–767. doi: 10.1038/nature08398

  7. [7]

     A.P. Chen, Y. Zhong, Y.X. Liu, et al., Chin. Chem. Lett. 34 (2023) 107923. doi: 10.1016/j.cclet.2022.107923

  8. [8]

     L.Y. Huang, D.W. Hsu, C.J. Pears, Nucleic Acids Res. 49 (2021) 3781–3795. doi: 10.1093/nar/gkab154

  9. [9]

     C.B. Bian, C. Xu, J.B. Ruan, et al., EMBO J. 30 (2011) 2829–2842. doi: 10.1038/emboj.2011.193

  10. [10]

      S. Ha, J.C. Zhu, H. Xiang, G.S. Luo, Chin. Chem. Lett. 35 (2024) 109192. doi: 10.1016/j.cclet.2023.109192

  11. [11]

      C.R. Travis, K.M. Kean, K.I. Albanese, et al., J. Am. Chem. Soc. 146 (2024) 3086–3093. doi: 10.1021/jacs.3c10031

  12. [12]

      K. Arita, S. Isogai, T. Oda, et al., Proc. Natl. Acad. Sci. U. S. A. 109 (2012) 12950–12955. doi: 10.1073/pnas.1203701109

  13. [13]

      X.J. Cao, B.A. Garcia, Curr. Protoc. Protein Sci. 86 (2016) 24.8.1-24.8.19. doi: 10.1002/cpps.16

  14. [14]

      K.K. Biggar, S.S.C. Li, Nat. Rev. Mol. Cell Biol. 16 (2014) 5–17. doi: 10.1038/nrm3915

  15. [15]

      S.M. Carlson, K.E. Moore, E.M. Green, G.M. Martín, O. Gozani, Nat. Protoc. 9 (2013) 37–50. doi: 10.1038/nprot.2013.164

  16. [16]

      A.J. Burton, M. Haugbro, L.A. Gates, et al., Nat. Chem. 12 (2020) 520–527. doi: 10.1038/s41557-020-0474-8

  17. [17]

      T.P. Yang, Z. Liu, X.D. Li, Chem. Sci. 6 (2015) 1011–1017. doi: 10.1039/C4SC02328E

  18. [18]

      X. Li, E.A. Foley, K.R. Molloy, et al., J. Am. Chem. Soc. 134 (2012) 1982–1985. doi: 10.1021/ja210528v

  19. [19]

      K. Zou, Y.H. Xiao, J.Y. Yang, M.X. Wu, Chin. Chem. Lett. 35 (2024) 109497. doi: 10.1016/j.cclet.2024.109497

  20. [20]

      H. Park, J.Y. Koo, Y.V.V. Srikanth, et al., Chem. Commun. 52 (2016) 5828–5831. doi: 10.1039/C6CC01426G

  21. [21]

      F. Feng, Y.X. Gao, Q. Zhao, et al., Nat. Chem. 16 (2024) 1267–1277. doi: 10.1038/s41557-024-01577-y

  22. [22]

      P.A. Szijj, K.A. Kostadinova, R.J. Spears, V. Chudasama, Org. Biomol. Chem. 18 (2020) 9018–9028. doi: 10.1039/d0ob01912g

  23. [23]

      E.J. Choi, D. Jung, J.S. Kim, Y. Lee, B.M. Kim, Chem. Eur. J. 24 (2018) 10948–10952. doi: 10.1002/chem.201802380

  24. [24]

      K.I. Albanese, M.W. Krone, C.J. Petell, et al., ACS Chem. Biol. 15 (2020) 103–111. doi: 10.1021/acschembio.9b00651

  25. [25]

      S. Arora, W.S. Horne, K. Islam, J. Am. Chem. Soc. 141 (2019) 15466–15470. doi: 10.1021/jacs.9b05725

  26. [26]

      S. Del Galdo, G. Mancini, I. Daidone, et al., J. Comput. Chem. 39 (2018) 1747–1756. doi: 10.1002/jcc.25351

  27. [27]

      J.A. Law, J.M. Du, C.J. Hale, et al., Nature 498 (2013) 385–389. doi: 10.1038/nature12178

  28. [28]

      S. Iwase, B. Xiang, S. Ghosh, et al., Nat. Struct. Mol. Biol. 18 (2011) 769–776. doi: 10.1038/nsmb.2062

  29. [29]

      A.S. Dixon, M.K. Schwinn, M.P. Hall, et al., ACS Chem. Biol. 11 (2015) 400–408. doi: 10.1021/acschembio.5b00753

  30. [30]

      Y. Yang, M. Liang, R. Wang, C. He, Chin. Chem. Lett. 34 (2023) 107806. doi: 10.1016/j.cclet.2022.107806

• 

  Figure 1  Overview of strategies of tyrosine bioconjugation. (a) Crystal structures of diverse tyrosine-containing methyllysine reader domains complexed with their ligands. The binding pockets contain conserved tyrosine residues. PDB, protein data bank. (b) Approaches of crosslinking to tyrosine residues. Current strategies of tyrosine bioconjugation are with poor site-selectivity and much less reactive to tyrosine inside pockets due to low accessibility. In this work, dimethylsulfonium as a dimethyllysine mimic enables selective photo crosslinking to tyrosine in the binding pocket of readers under UV irradiation.

  下载: 全尺寸图片 幻灯片
  

  Figure 2  H3K9NleS⁺me2 peptide binds to CBX1(W42Y) and displays selective photo crosslinking to the tyrosine in the binding pocket under 302 nm UV irradiation or basic conditions. (a) The key residues in CBX1 aromatic cage for H3K9me3 binding. (b) Structure of histone H3K9NleS⁺me2 peptide. (c) ITC analysis of CBX1(W42Y) with H3K9me2 and H3K9NleS⁺me2 at pH 7.4. Experiments were conducted under the conditions of 25 µmol/L protein concentration and 750 µmol/L peptide concentration using a MicroCal PEAQ-ITC Automated instrument at 25 ℃. (d) Ultraviolet emission spectra of UV lamp without or with a longpass filter, called 302 nm and 313 nm, respectively. (e) Photo crosslinking between CBX1(W42Y) and H3K9NleS⁺me2 under different UV conditions. CBX1(W42Y) (3 µmol/L) was incubated with H3K9NleS⁺me2 (100 µmol/L) on ice for 15 min in the buffer of pH 7.4 and then crosslinked under 313 nm UV and 302 nm UV for an additional 15 min. The analytical yield of peptidyl conjugate H3K9Nle-CBX1(W42Y) increased from 17.1% to 62.5% when the wavelength of UV source changed from 313 nm to 302 nm. (f) Photo crosslinking between CBX1(Y21F, W42F) and H3K9NleS⁺me2 under 302 nm UV. CBX1(Y21F, W42F) (3 µmol/L) crosslinked H3K9NleS⁺me2 peptide (100 µmol/L) in the buffer of pH 7.4 under 302 nm UV for 15 min. No crosslinking product was detected. (g) Deprotonation of tyrosine residue in the basic condition. (h) Absorption spectra of tyrosine amino acid and ARTESYKA peptide. The absorption peaks red shifted and the absorbance rose as increase of pH value. (i) ITC measurements of CBX1(W42Y) and H3K9NleS⁺me2 at different pH. H3K9NleS⁺me2 (750 µmol/L) was titrated into CBX1(W42Y) (25 µmol/L) at 25 ℃ in the buffers of different pH. (j) Photo crosslinking between CBX1(W42Y) and H3K9NleS⁺me2 at different pH. CBX1(W42Y) (3 µmol/L) was incubated with H3K9NleS⁺me2 (100 µmol/L) on ice for 15 min in the buffers of different pH and then crosslinked under 313 nm UV for an additional 15 min. The crosslinking efficiency of CBX1(W42Y) was promoted by increase of pH value. Methyl-CBX1(W42Y): methyl conjugate, the product through SET along with generation of H3K9NleSme due to alternative homolysis of C-S bond. (k) Photo crosslinking between CBX1(Y21F, W42F) and H3K9NleS⁺me2 in the basic condition of pH 10.1. CBX1(Y21F, W42F) (3 µmol/L) crosslinked H3K9NleS⁺me2 peptide (100 µmol/L) in the basic condition of pH 10.1 under 313 nm UV for 15 min. No crosslinking product was detected.

  下载: 全尺寸图片 幻灯片
  

  Figure 3  Characterization of the molecular mechanism of photo crosslinking between CBX1(W42Y) and H3K9NleS⁺me2. (a) Scheme of proposed binding-based crosslinking mechanism. In the hypothesis, photo crosslinking happens inside a protein-ligand (PL) complex which is formed through the binding of CBX1(W42Y) and H3K9NleS⁺me2. (b) Kinetic analysis of photo crosslinking to compare effects made by UV wavelength and pH. CBX1(W42Y) (5 µmol/L) was incubated with H3K9NleS⁺me2 peptide at various concentrations ranging from 12.5 µmol/L to 400 µmol/L and then crosslinked for 15 min under UV irradiation. Average values and errors (±s.e.m.) were calculated from three independent experiments (n = 3).

  下载: 全尺寸图片 幻灯片
  

  Figure 4  NleS⁺me2 peptide probes enable selective crosslinking to proximate tyrosine widely. (a) NleS⁺me2 peptide probes were active to crosslink their specific methyllysine readers and the crosslinking was promoted in basic conditions. (b) SGF29, UHRF1 and EED were selectively crosslinked by the corresponding sulfonium probes. (c) The crystal structure of NanoBiT. (d) Western blot analysis of photo crosslinking between Flag–SmBiT79(E166Met⁺me) and His₆-LgBiT(W11Y). (e) Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and Western blot analysis of photo crosslinking between Flag-SmBiT79(E166Met⁺me) and His₆-LgBiT(W11Y) in HeLa cell lysate.

  下载: 全尺寸图片 幻灯片
•