|Home | About | Journals | Submit | Contact Us | Français|
Access to phosphoproteins with stoichiometric and site-specific phosphorylation status is key for understanding the role of protein phosphorylation. Here we report an efficient method to generate pure, active phosphotyrosine-containing proteins by genetically encoding a stable phosphotyrosine analogue that is convertible into native phosphotyrosine. We demonstrate its general compatibility with proteins of various sizes, phosphotyrosine site and function, and reveal a possible negative regulation role of tyrosine phosphorylation in ubiquitination.
Protein phosphorylation is a major post-translational modification (PTM) that plays a pivotal role for signal transduction and in regulating cellular events1,2. In particular, phosphorylation of tyrosine is involved in many processes including cell proliferation, cell cycle progression, metabolic homeostasis, transcriptional activation, neural transmission, differentiation, development, and aging1,3. Conversely, dysfunction of tyrosine phosphorylation results in various diseases, most prominently in cancer4–6. Limited access to site-specifically phosphorylated proteins hampers the investigation of this major PTM in disease-relevant proteins7.
Many methods have been developed to study protein phosphorylation. A generally useful approach is to substitute glutamate or aspartate for phosphoserine or phosphothreonine, despite the charge difference between carboxylate and phosphate mono-ester. A mimic of phosphotyrosine (pTyr) however is challenging since there are no negatively charged aromatic residues in the 20 amino acids. Proteins can be phosphorylated with kinases, but in vitro kinase phosphorylation has limited site-specificity and often results in sub-stoichiometric phosphorylation8. To overcome this limitation, chemical approaches, including native chemical ligation, semisynthetic and cell-free methods, have been developed to introduce pTyr site selectively, but these methods are technically challenging and cannot be generally applied to all proteins9–13. Through the expansion of the genetic code14, a chemical analogue of pTyr has been incorporated into proteins in E. coli15, yet the analogue cannot faithfully mimic the phosphate group and its negative charges. Direct incorporation of phosphorylated amino acids in live cells also encounters multiple challenges16–20: the phosphorylated amino acid enters cells poorly due to its negative charge, is unstable inside cells, and is incompatible with the elongation factor Tu (EF-Tu) for translation. Despite feeding excessive amount of phosphorylated amino acids to cells, engineering EF-Tu, and reducing cellular phosphatase activity, these challenges still result in low incorporation efficiency and other amino acids at the phosphorylation site. To date, there is no general, efficient system available to prepare proteins with native pTyr introduced site-specifically.
Here we developed an efficient and easily accessible method for the production of pure tyrosine-phosphorylated proteins. Our strategy circumvents the aforementioned critical problems: A charge-neutral and stable phosphotyrosine analogue was genetically incorporated into proteins at the target phosphorylation site through the expansion of the genetic code, which was subsequently converted into a native pTyr by a facile pH shift (Fig. 1a). Using this approach we prepared multiple proteins with pTyr site-specifically introduced at the phosphorylation site, with good yields and without disrupting protein activity. We also investigated the effect of tyrosine phosphorylation on ubiquitin (Ub) structure and function, revealing a possible negative regulatory role of Ub phosphorylation.
Because the phosphate group of pTyr is labile to cellular phosphatases and its negative charge limits cell permeability, we genetically encoded the pTyr analogue 1 containing a phosphoramidate group (Fig. 1a), which is stable, neutral, and can be cleaved under acidic conditions to generate the native pTyr21. We synthesized unnatural amino acid (Uaa) 1 in 82% yield and high purity without the need for column chromatography (Supplementary Results, Supplementary Note 1). To optimize a cleavage condition suitable for proteins, we used low concentrations of HCl to remove the protecting group in 1. Already, at 0.04 M HCl (pH ~2), the protecting group was cleaved completely within 36 h at room temperature to generate pTyr (Supplementary Fig. 1).
To genetically encode Uaa 1 in E. coli, we evolved the Methanosarcina mazei /PylRS pair22 to be specific for this Uaa. A mutant library of PylRS was generated with residues 302, 309, 322, 346, 348, 401, 417, and 419 mutated as previously described23,24. From this library we identified a clone showing phenotypic dependence on Uaa 1 (Supplementary Fig. 2), which harbored a mutant (named MmNpYRS) containing Ser302, Met309, Leu322, Ala346, Gly348, Val401, Thr417, and Gly419.
To investigate the efficiency and fidelity of MmNpYRS to incorporate 1, we expressed in E. coli a gene for Xenopus calmodulin (CaM) that contained an amber codon TAG at site Met76 and a C-terminal His×6 tag, together with the /MmNpYRS. CaM consists mainly of α-helices and has low molecular weight (18 kDa) to facilitate MS characterization with high accuracy. Full-length CaM was obtained in good yield (ca. 1.0 mg L−1) in the presence of Uaa 1 (1 mM), but was undetectable in its absence (Supplementary Fig. 3). Analysis of the expressed CaM by electrospray ionization ion trap mass spectrometry (ESI-MS) confirmed incorporation of 1 into CaM (Fig. 1b). An observed peak at 18,103.8 Da corresponds to intact CaM containing Uaa 1 at site 76 (expected [M+H]+ = 18,104.0 Da); a second peak corresponds to the Uaa 1-containing CaM lacking the initiator Met (expected [M-Met+H]+ = 17972.8 Da, measured 17972.7 Da). No peaks were observed corresponding to CaM containing a natural amino acid at the TAG site. In addition, the high fidelity of MmNpYRS was also confirmed by incorporating 1 into another protein, myoglobin (Supplementary Fig. 4). In contrast to the need of an engineered EF-Tu for incorporating phosphoserine16, incorporation of 1 into proteins in E. coli suggests that WT E. coli EF-Tu is compatible with Uaa .
To convert 1 into pTyr within proteins, we treated CaM solution (0.6 mg mL−1) with HCl (final conc. 0.4 M, pH~1) for 48 h at 4 °C. The sample was lyophilized to remove the acid and then dissolved in water. ESI-MS measurements clearly demonstrated the removal of the protecting group and formation of pTyr within CaM: A peak at 18,049.2 Da (Fig. 1c) corresponds to intact CaM with pTyr at site 76 (pYCaM, expected [M+H]+ = 18,049.9 Da). A second peak corresponds to the pTyr-containing CaM lacking the initiator Met (expected [M–Met+H]+ = 17,918.7 Da, measured 17,918.2 Da).
We next introduced pTyr into green fluorescent protein (GFP), which has a different secondary structure β-sheet from the largely helical CaM. A GFP gene containing a TAG codon at Tyr182 and a C-terminal His×6 tag was expressed with the /MmNpYRS in E. coli. The mutant GFP was purified, yielding 1.25 mg L−1, corresponding to 31% of WT GFP purified under the same conditions. HCl (final conc. 0.4 M, pH~1) was added to a diluted (0.1 mg mL−1) solution of the mutant GFP; the mixture was incubated at 4 °C for 16 h, and then readjusted to pH 7.5 with NaOH. Western-blot analysis using anti-His×6 antibody showed that full-length GFP was expressed only in the presence of Uaa 1 (Supplementary Fig. 5), confirming specific incorporation of Uaa 1 by the /MmNpYRS. The GFP band upshifted after HCl treatment, suggesting that deprotection generated pTyr whose negative charge slowed phosphoprotein movement in SDS-PAGE. The pTyr-specific antibody detected a band in the HCl-treated GFP only, at the position co-localizing with the up-shifted GFP band in the anti-His×6 blot, indicating the conversion of Uaa 1 into pTyr. The pTyr conversion efficiency was ~90% based on band intensities. Before and after acidic cleavage of Uaa 1, the GFP proteins were both green fluorescent (Supplementary Fig. 6), indicating that Uaa 1 and conversion into pTyr did not prevent GFP from folding into the correct fluorescent structure. To maximize deprotection of Uaa 1, exposure time to HCl may depend on protein concentration and structure. While we found 16 h sufficient for GFP and others (vide infra), CaM needed 48 h.
We next generated pTyr-containing ubiquitin (Ub) and investigated the impact of tyrosine phosphorylation on Ub conformation and function. While serine phosphorylation has been extensively studied for Ub, phosphorylation of Tyr59, the only tyrosine residue of WT Ub, has been exclusively observed in cancerous tissue though its biological relevance remains unknown25,26. In WT Ub, the hydroxyl of Tyr59 hydrogen bonds with the backbone amide of Glu51, forming a loop that is disrupted by mutation of Tyr5927. To provide direct evidence whether phosphorylation of Tyr59 alters the Tyr59–Glu51 loop in Ub and affects Ub function, we incorporated Uaa 1 into Ub at site 59 and generated phosphotyrosine by acid cleavage followed by lyophilization. SDS-PAGE and MS confirmed the successful preparation of Tyr59-phosphorylated Ub (Supplementary Fig. 7). ESI-MS analysis of acid-treated Ub showed a peak at 9468.1774 Da, corresponding to intact Ub with pTyr at site 59 lacking the initial methionine (expected [M+H–Met]+ = 9468.6856 Da); trypsin digest and MS/MS sequencing of this protein confirmed that pTyr was incorporated at the TAG site 59.
To examine a potential conformational change induced by pTyr59, we measured the HSQC NMR spectra of WT Ub and the Tyr59-phosphorylated Ub (pTyrUb). The signals of Tyr59 and Glu51 as well as other relevant amino acids in the loop shifted dramatically in pTyrUb compared to WT Ub (Fig. 2a). Most notably, the Glu51 signal showed a similar large shift as observed previously in the Y59F Ub mutant27. In addition, most of the assigned and shifted residues were located in or near the Y59–E51 loop (Fig. 2b). To exclude possible shifts brought by the acid treatment, we also subjected the WT Ub to the same treatment and detected no difference in the HSQC NMR spectra before and after treatment (Supplementary Fig. 8). These NMR data thus indicate that the Y59–E51 loop was indeed altered after phosphorylation of Tyr59.
To investigate whether pTyr59 would interfere with thioester formation between Ub and E2 conjugating enzyme, we studied the conjugation of pTyrUb to the E2 enzyme UBE2D3. In contrast to Ser65-phosphorylated Ub, which exhibits enhanced Ub–E2 conjugation18,26, we found that pTyrUb dramatically decreased the Ub–E2 conjugation (Fig. 2c). To exclude any potential interference from the acid treatment, we subjected WT Ub to the same treatment as the Uaa 1-incorporated Ub, and found that the treated WT Ub still conjugated to UBE2D3 (Supplementary Fig. 9). Therefore, Tyr59 phosphorylation altered Ub conformation and decreased its ability to conjugate with the E2 enzyme UBE2D3, suggesting that Tyr59 phosphorylation on Ub could play a negative regulatory role in the ubiquitination process.
In summary, we have developed a new method to prepare phosphorylated proteins with phosphotyrosine site-specifically introduced in high quality and good yields. Combining genetic incorporation with facile pH conversion, this method enables phosphotyrosine to be introduced in different secondary structures of various proteins with broad compatibility of protein type, size, and phosphorylation site. The method does, however, require the target protein to withstand low acid treatment and to refold into the functional state upon acid removal. As the tRNAPyl/MmNpYRS was derived from the tRNAPyl/PylRS, which is orthogonal in prokaryotic and eukaryotic cells, we expect that the tRNAPyl/MmNpYRS can also be used in eukaryotic cells for preparing proteins difficult to express in bacteria. The technical simplicity of this approach should allow general adoption, affording a robust method to facilitate the investigation of tyrosine phosphorylation in biology and diseases.
(S)-2-(((9H-fluoren-9-yl)methoxy)carbonylamino)-3-(4(bis(dimethylamino)phosphoryloxy) phenyl)propanoic acid (Fmoc-Tyr(PO(NMe2)2) was obtained from EMD Millipore. All other chemicals were obtained from commercial sources and used without further purification. Mass spectrometric analysis of proteins was performed by Jadebio (San Diego) using methods previously described28,29. LC-MS measurement of Uaa 1 was performed on AGILENT 1200 series LC system in combination with an Advion Expression CMS mass spectrometer. Oligonucleotides and primers were obtained from IDT. His-tag antibody (HRP) was from Invitrogen (MA-1–21315-HRP), Phospho-Tyrosine Mouse mAb (P-Tyr-100) was from Cell Signaling Technology (9411S) and IgG mouse secondary antibodies were from Santa Cruz Biotechnology (sc-2314). Ub-antibody (MAB1510) was from Millipore. All antibodies were used in a dilution 1:2000. NMR spectra of Uaa 1 were recorded on Bruker 300 MHz. 1H-15N HSQC spectra for 15N labeled WT and pTyr59 ubiquitin were acquired on a Bruker 800 MHz spectrometer and processed with Bruker Topspin. Samples were suspended at a concentration of 100 μM in 20 mM Na-phosphate (pH 6.0) 10% D2O/90% H2O buffer in a total volume of 300 μL. For all experiments, sample temperature was set to 300.7 K. Spectra and assignments were carried out in CcpNMR analysis by comparison to a standard assigned by HNCACB experiments. Chemical shift differences between WT and pTyr59 were calculated by taking the root-mean-square of differences for a given residue.
To a solution of 1 (1 mg mL−1) in water was added hydrochloric acid to reach a final concentration of 0.04 M HCl (pH~2). The reaction mixture was incubated at 25°C and the reaction was followed over time by successive LC-MS measurements at 1 h, 18 h and 36 h. After 36 h, ~95% of pTyr 2 had been formed confirming that even very low hydrochloric acid concentrations allowed cleaving the phosphoramidate group in 1.
The library CRIZ previously designed for bulky amino acids was electroporated into DH10βT1 competent cells harboring pREP selection plasmid. Selection was carried out on GMM-L plates containing 12.5 μg mL−1 tetracycline, 50 μg mL−1 kanamycin, 100 μg mL−1 chloramphenicol, and 1 mM Uaa 1 as previously described23,24. Single green colonies were picked and re-streaked on LB/tetracycline/kanamycine and on LB/tetracycline/kanamycine/chloramphenicol. One clone (B1) was sensitive to chloramphenicol selection in the absence of the Uaa on the LB/T/K/Cm plate, but was viable and nonfluorescent on the LB/T/K plate. This clone was characterized in detail and named as MmNpYRS.
To genetically incorporate Uaa 1 into proteins (calmodulin, myoglobin, GFP, ubiquitin) in E. coli, BL21 cells were transformed with corresponding plasmids (pTak-CaM76TAG, pTak-Myo4TAG, pTak-GFP182TAG, or pTak-Ub59TAG with pBK-MmNpYRS). All the pTak plasmids22 have a His-tag at the C-terminus of the target protein. For the E2 charging assay, N-terminal His-tagged Ub was expressed. One colony was picked and grown overnight in 5 mL 2xYT supplemented with 30 μg mL−1 chloramphenicol and 50 μg mL−1 kanamycin at 37 °C. This starter culture was used to inoculate 100 mL of 2xYT containing antibiotics. When OD600 reached 0.5, 1 mM of Uaa 1 was added, and cells were induced for protein expression by adding 0.5 μM IPTG. After 16 h, cells were lysed and sonicated in 5 mL lysis buffer (50 mM TrisHCl, pH 8.0, 500 mM NaCl, 20 mM imidazole pH 8.0, 1% (v/v) Tween 20, 10% (v/v) glycerol and 0.5 mg mL−1 lysozyme). Lysed cells were centrifuged for 30 min at 14,000 g, and clarified supernatant was passed through a 0.1 mL column of Ni2+-NTA agarose resin (Qiagen). The column was washed with 10 column volumes of wash buffer (lysis buffer without Tween 20 and lysozyme). Protein was eluted with 400 μL of elution buffer (wash buffer containing 250 mM imidazole, pH 8.0). The sample was concentrated using a Microcon Ultracel YM-10 (CaM, myoglobin, GFP) and YM-3 (ubiquitin) spin column (Millipore) yielding concentrations of 1.0 mg L−1 of CaM, 1.25 mg mL−1 myoglobin, 1.25 mg mL−1 GFP, and 1.75 mg L−1 ubiquitin. SDS-PAGE was 15% polyacrylamide, and His-tag or pTyr specific antibodies were used for immunoblotting with a dilution of 1:2000.
To a diluted protein solution (0.1–1.0 mg mL−1) in Tris buffer (50 mM Tris, 150 mM NaCl, pH 7.5) was added HCl (4 M) to reach a final HCl concentration of 0.4 M and a pH ~1–2, respectively. Protein solutions of myoglobin, GFP, ubiquitin were incubated at 4°C for 16 h. Protein solution of CaM was incubated at 4°C for 48 h. For Western blot analysis, the pH of the samples was readjusted to pH 7.5 using NaOH (0.5 M). Subsequently, samples were loaded onto 15% gel for SDS-PAGE, blotted, and visualized using His-tag or pTyr specific antibodies. Alternatively, HCl was removed by lyophilization of the reaction mixture. For recording MS spectra of CaM, HCl was removed by lyophilization and samples were re-dissolved in water. In addition, either C-terminally or N-terminally His-tagged Ub were freeze dried after HCl treatment and re-dissolved in water.
To prepare 15N labeled ubiquitin with pTyr at position 59 in E. coli, BL21 cells were transformed with plasmids pTak-Ub59TAG and pBK-MmNpYRS. One colony was picked and grown overnight in 5 mL 2xYT supplemented with 30 μg mL−1 chloramphenicol and 50 μg mL−1 kanamycin at 37 °C. This starter culture was used to inoculate 1400 mL of 2xYT containing antibiotics. When OD600 reached 0.6, cells were centrifuged for 5 min at 5,000 g and resuspended in 700 mL M9 media with antibiotics, 1 mM Uaa 1 and induced for protein expression by adding 0.5 μM IPTG. After 6 h, cells were lysed and sonicated in 10 mL lysis buffer (50 mM TrisHCl, pH 8.0, 500 mM NaCl, 20 mM imidazole pH 8.0, 1% (v/v) Tween 20, 10% (v/v) glycerol and 0.5 mg mL−1 lysozyme). Lysed cells were centrifuged for 50 min at 14,000 g, and clarified supernatant was passed through a 1.4 mL column of Ni2+-NTA agarose resin (Qiagen). The column was washed with 10 column volumes of wash buffer (lysis buffer without Tween 20 and lysozyme). Protein was eluted with 1400 μL of elution buffer (wash buffer containing 250 mM imidazole, pH 8.0). The sample was concentrated using a Microcon Ultracel YM-3 spin column (Millipore). After diluting to a concentration of 0.5 mg mL−1 in Tris-buffer (50 mM Tris, 150 mM NaCl, pH 7.5), the ubiquitin was treated with HCl (0.4 N) o/n, freeze dried and the buffer was exchanged to phosphate buffer (pH 5.8) yielding 1 mg mL−1 labeled ubiquitin pTyr59. Labeled WT ubiquitin was prepared accordingly. In brief, plasmid pTakWTUb was electroporated into BL21 cells and one colony was picked and grown overnight in 5 mL 2xYT supplemented with 30 μg mL−1 chloramphenicol at 37 °C. This starter culture was used to inoculate 600 mL of 2xYT containing antibiotic. When OD600 reached 0.6, cells were centrifuged for 5 min at 5,000 g and resuspended in 300 mL M9 media with antibiotic and induced for protein expression by adding 0.5 μM IPTG. After 6 h, cells were lysed and sonicated in 10 mL lysis buffer (50 mM TrisHCl, pH 8.0, 500 mM NaCl, 20 mM imidazole pH 8.0, 1% (v/v) Tween 20, 10% (v/v) glycerol and 0.5 mg mL−1 lysozyme). Lysed cells were centrifuged for 50 min at 14,000 g, and clarified supernatant was passed through a 0.6 mL column of Ni2+-NTA agarose resin (Qiagen). The column was washed with 10 column volumes of wash buffer (lysis buffer without Tween 20 and lysozyme). Protein was eluted with 600 μL of elution buffer (wash buffer containing 250 mM imidazole, pH 8.0) yielding 15 mg mL−1 WT ubiquitin after buffer was exchanged to phosphate buffer (pH 5.8). As a control, we also treated the WT ubiquitin with hydrochloric acid (0.4 N) for 16 h, freeze dried the sample, re-dissolved in buffer and measured HSQC NMR.
To genetically incorporate Uaa 1 into ubiquitin in E. coli, BL21 cells were transformed with plasmids pTak-Ub59TAG_N-His and pBK-MmNpYRS. One colony was picked and grown overnight in 5 mL 2xYT supplemented with 30 μg mL−1 chloramphenicol and 50 μg mL−1 kanamycin at 37 °C. This starter culture was used to inoculate 200 mL of 2xYT containing antibiotics. When OD600 reached 0.5, 1 mM of Uaa 1 was added, and cells were induced for protein expression by adding 0.5 μM IPTG. After 3 h, cells were lysed and sonicated in 10 mL lysis buffer (50 mM TrisHCl, pH 8.0, 500 mM NaCl, 20 mM imidazole pH 8.0, 1% (v/v) Tween 20, 10% (v/v) glycerol and 0.5 mg mL−1 lysozyme). Lysed cells were centrifuged for 50 min at 14,000 g, and clarified supernatant was passed through a 0.2 mL column of Ni2+-NTA agarose resin (Qiagen). The column was washed with 10 column volumes of wash buffer (lysis buffer without Tween 20 and lysozyme). Protein was eluted with 400 μL of elution buffer (wash buffer containing 250 mM imidazole, pH 8.0). The sample was concentrated using a Microcon Ultracel YM-3 spin column (Millipore) yielding 1.0 mg L−1 of Ub. Buffer was exchanged to Tris-buffer (50 mM Tris, 150 mM NaCl, pH 7.5) before HCl treatment. Trypsin digest and MS/MS sequencing confirmed the pTyr at appropriate site.
Human E1 (UbE1) was obtained from Fisher Scientific (part number E304050). UBE2D3 (from addgene 15784) was prepared by standard cloning procedures with His-tag. Reaction mixtures in Tris-buffer (50 mM Tris, 150 mM NaCl, pH 7.5) contained UBE1 (0.5 μM), UBE2D3 (2 μM), MgCl2 (10 mM), ATP (10 mM) and the pTyrUb (10 μM) or WTUb (10 μM) at a total volume of 40 μL. As negative control, the same reaction was carried out without ATP. The mixture was incubated at 37 °C for 20 h. For SDS-PAGE, to 8 μL of each reaction 2 μL of non-reducing Lämmli buffer (Bio-Rad) was added, loaded onto a 15% SDS gel and stained with coomassie. For Western blot 1 μL of each reaction mixture was diluted with 5 μL Tris and 1.5 μL of non-reducing Lämmli buffer and loaded on 15% SDS gel, transferred to a membrane and visualized with His-tag (HRP) antibody. Subsequently, the membrane was stripped under mild conditions and visualized with anti-Ub (mouse) and IGg-mouse antibody in 1:2000 dilution. In parallel, sample were treated at 100 °C with reducing Laemmli buffer containing 2-mercaptoethanol, loaded onto SDS gel, transferred and immunoblotted with His-tag (HRP) antibody in 1:2000 dilution. In addition, control experiments were conducted to exclude any interference of the E2 charging assay with acid treatment. WT ubiquitin in Tris-buffer (50 mM Tris, 150 mM NaCl, pH 7.5) was diluted to 0.6 mg mL−1 and treated with hydrochloric acid (0.4 N) at 4 °C for 16 h. The sample was freeze fried and reconstituted with water. Western blot immunoblotted with His-tag (HRP) antibody shows that the Ub-E2 thioester is formed.
We thank Mark Kelly (UCSF NMR Core) for help with NMR measurements and helpful discussions. L.W. acknowledges the support of NIH (R01GM118384).
Author Contributions: C.H. conducted experiments and characterized data, A.W. assigned data for NMR, B.Y. prepared UBE2D3, S.L., T.H., K.M.S. provided helpful discussions, L.W. conceived and directed the project, and C.H. and L.W. wrote the manuscript with inputs from S.L., T.H., and K.M.S.
Competing Financial Interests Statement: The authors declare no competing financial interests.
Data availability. All data generated or analyzed during this study are included in this published article (and its supplementary information files).