Development of Convergent Hybrid Phase Ligation for Efficient and Convenient Total Synthesis of Proteins

Abstract

Simple and efficient total synthesis of homogeneous and chemically modified protein samples remains a significant challenge. Here, we report development of a convergent hybrid phase native chemical ligation (CHP-NCL) strategy for facile preparation of proteins. In this strategy, proteins are split into ~100-residue blocks, and each block is assembled on solid support from synthetically accessible peptide fragments before ligated together into full-length protein in solution. With the new method, we increase the yield of CENP-A synthesis by 2.5-fold compared to the previous hybrid phase ligation approach. We further extend the new strategy to the total chemical synthesis of 212-residue linker histone H1.2 in unmodified, phosphorylated, and citrullinated forms, each from eight peptide segments with only one single purification. We demonstrate that fully synthetic H1.2 replicates the binding interactions of linker histones to intact mononucleosomes, as a proxy for the essential function of linker histones in the formation and regulation of higher order chromatin structure.

Graphical Abstract

Introduction:

The ability to prepare homogeneously modified protein samples is essential for structural and functional studies of protein post-translational modifications (PTMs)¹, but the technical challenge has imposed a barrier to the number and types of systems that can be studied. In the past two decades, many approaches have been explored to improve the efficiency of chemical protein synthesis^2–6, but they all share inherent limitations in practice. One interesting new approach is automated flow chemistry, which has achieved initial success in synthesis of small proteins try^7,8; it remains to be seen if it will fulfill its promise in wider application.

Over that same timeframe, native chemical ligation (NCL)^5,9, combined with solid phase peptide synthesis (SPPS)¹⁰, and desulfurization^11,12 has emerged as the dominant strategy for total protein synthesis, since it is traceless and recovers the native peptide bond^2–4,6,13. Moreover, total chemical protein synthesis provides full control of protein composition, and thus can overcome inherent limitations of biological synthesis, for example generating naturally absent mirror-image D-proteins which have potential research and therapeutical applications^14,15. However, due to the size limitation of peptides prepared by SPPS¹⁰, multiple rounds of NCLs are often required to achieve synthesis of full-length proteins. While the ligation reaction itself is typically efficient, intermediate purification steps are unavoidable for either sequential or convergent solution phase strategies, which often leads to significant yield losses that compound over each round. One-pot ligation^16,17, which only requires one single final purification step, can overcome the compounding losses, but has an inherent limit in the number of segments that can be assembled^17–21.

Theoretically, carrying out reactions on the solid phase (solid phase NCL)^22,23 should overcome these limitations. While the concept of solid phase NCL was demonstrated more than two decades ago, it has only achieved limited success for preparation of relatively small proteins (<100–150 residues)^22–27. The reason for this size barrier is unclear, although non-specific interactions between polypeptide chains and/or the solid support, leading to aggregation, have been proposed. But the problem is exemplified by a study from our own laboratory; when preparing the 139-residue CENP-A protein by sequential solid phase ligation, we observed full yield (by weight) for solid phase assembly of a 70-residue block prepared by solid phase ligation, but almost no product recovery at 105 residues or for the full-length protein²⁴. For that particular use case, we developed a hybrid phase ligation strategy in which the last two ligation steps were carried out in solution; while the final product was obtained, the yield was reduced by purification during the solution-phase steps (Figure 1A). Raibaut et al. compared hybrid phase ligation and fully solid phase ligation for preparation of a 94-residue protein; within that size regime, they found that including a solution phase step reduced overall yield²⁵, also consistent with a loss of efficiency with increasing numbers of solution phase ligation steps.

Figure 1.

Development of Convergent Hybrid Phase Native Chemical Ligation (CHP-NCL). A. Sequential solid phase ligation and hybrid phase ligation gives no or low yield for preparation of CENP-A protein. B. Two activation pathways of 3,4-diaminobenozic acid linker into thioester precursors: on-resin with protected peptide or in-solution with unprotected peptide.

A logical extension of the hybrid approach would be convergent hybrid phase native chemical ligation (CHP-NCL). Here, rather than following up a solid phase NCL assembly with sequential solution ligation, ligation “blocks” could be independently assembled on the solid phase, requiring only one solution phase step. This strategy requires a chemical handle that is unreactive throughout the chemistry required for SP-NCL, yet can be activated for the subsequent ligation step. We recognized that these properties could be introduced through the crypto-thioester 3,4-diaminobenzoic acid (Dbz), which was originally reported for use in generating peptide thioesters by Fmoc-SPPS²⁸ and subsequently found to undergo conversion to thioester with unprotected polypeptides²⁹ (Figure 1B), to enable following solution phase ligation.

Here, we demonstrate the feasibility of the new strategy in two protein systems. We first extend our prior work with a 2.5-fold increase in yield of the 139-residue CENP-A synthesis protein from 5 peptide segments assembled in two blocks. We also accomplished the total chemical synthesis of the 212-residue linker histone H1.2 from 8 synthetic fragments in two blocks, with only one single final purification.

Materials and Methods:

All chemicals mentioned in this manuscript were used following their standard safety data sheets (SDS) and all procedures were carried out in designated chemical use areas and with appropriate PPE, in accordance with standard laboratory safety protocols and guidelines of The Ohio State University Environmental Health & Safety.

Materials

Rink Amide MBHA resin LL (100–200 mesh, 0.3–0.4 mmol/g loading) was purchased from Novabiochem. PL-PEGA resin (300–500 μm, 0.2 mmol/g) was purchased from Varian. Amino PEGA resin (150–300 μm, 0.34 mmol/g) was from Novabiochem. N,N-Dimethylformamide C₃H₇NO (DMF), dichloromethane CH₂Cl₂ (DCM), acetonitrile CH₃CN (ACN), and diethyl ether (C₂H₅)₂O were purchased from Sigma Aldrich or Fisher Scientific. N-methyl pyrrolidone C₅H₉NO (NMP) was purchased from AGTC Bioproducts or Fisher Scientific. Piperidine C₅H₁₁N, 4-nitrophenyl chloroformate ClCO₂C₆H₄NO₂ (4-NPCF), 4-mercaptophenyl acetic acid HSC₆H₄CH₂CO₂H (MPAA), C₉H₁₅O₆P Tris(2-carboxylmethyl)phosphine (TCEP), N,N-diisopropylethylamine C₈H₁₉N (DIEA), Phenylsilane C₆H₈Si, and Tetrakis(triphenylphosphine)palladium(0) (Pd(PPh₃)₄) were purchased from Sigma Aldrich. Fmoc or Boc protected amino acids were purchased from AAPPTec and Novabiochem. Fmoc-6-Aminohexanoic acid (Fmoc-Ahx-OH), Fmoc-OSu, Fmoc-L-norleucine (Fmoc-Nle-OH), and 4-dimethylaminopyridine C₇H₁₀N₂ (DMAP) were purchased from Novabiochem. (2-(7-aza-1H-benzotriazole-1-yl)-1,1,3,3-tetramethyluronium. Oxyma-Pure was purchased from Novabiochem. (2-(7-aza-1H-benzotriazole-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate C₁₀H₁₅F₆N₆OP (HATU), 2-(6-chloro-1H-benzotriazole-1-yl)-1,1,3,3-tetramethylaminium hexafluorophosphate C₁₁H₁₅ClF₆N₅OP (HCTU), and 1-hydroxy-6-chlorobenzotriazole C₆H₄ClN₃O (6-Cl-HOBt) were purchased from AAPPTec. Acetic anhydride and sodium 2-sulfanylethane sulfonate C₂H₅NaO₃S₂ (MESNA) was purchased from Sigma-Aldrich. N,N’-diisopropylcarbodiimide C₇H₁₄N₂ (DIC) was purchased from Chem-Implex International. VA-044-US C₁₂H₂₂N₆·2HCl was purchased from Wako Chemicals. Ultra-pure guanidine-HCl CH₆ClN₃ (GnHCl) was purchased from MP Biomedicals. 3,4-diaminobenzoic acid (Dbz) was purchased from Sigma-Aldrich. 3-[(9-Fluorenylmethoxycarbonyl)amino]-4-(methylamino)benzoic acid (Fmoc-MeDbz) was purchased from Peptide Institute. INC. Allyl chloroformate C₄H₅ClO₂ was purchased from Acros Organics. Triisopropylsilane C₉H₂₂Si (TIS) was purchased from Sigma-Aldrich. α-Cyano-4-hydroxycinnamic acid C₁₀H₇NO₃ (HCCA) was purchased from Bruker Daltonics.

Synthesis of monoFmoc-Dbz

3-Fmoc-Dbz-OH was prepared as described previously³⁰. In brief, 3,4-Diaminobenzoic acid (1 g, 6.5 mmol) was resuspended in 125 mL 1:1 CH₃CN:NaHCO₃. Reaction was initiated by the dropwise addition of Fmoc-OSu (2.4 g, 7.1 mmol) in 15 mL 1:1 CH₃CN:NaHCO₃ and proceeded for 2 h. HCl was added to a final pH of 1.0, and the mixture was filtered. Filtrate was dissolved in 4 mL DMSO, precipitated with acidified reaction buffer, washed extensively, and dried under vacuum to yield a light gray product.

Peptides Synthesis and Purification

Peptides were synthesized using the standard Fmoc-N-α protection strategies either manually (for short sequences) or on an automated AAPPTec APEX 396 or CEM Liberty Lite synthesizer. Peptides synthesized on an AAPPTec APEX 396 automated synthesizer proceeded as following for each coupling cycle unless otherwise specified: 5 min Fmoc deprotection with 20% piperidine for two times, 30min coupling with HCTU/DIEA/Fmoc-AA (4/8.8/4.4), 5min capping with capping solution (300 mM 6-Cl-HOBt and 300 mM Acetic anhydride in 1:9 DCM: DMF). The peptides synthesized on a CEM Liberty Lite synthesizer were carried out with microwave assistance. For standard syntheses when MeDbz was used as a linker, microwave-heated syntheses followed the following procedure for every coupling cycle: 90 seconds Fmoc deprotection with 10% piperazine or 20% piperidine at 88 °C, 180 seconds amino acid/DIC/Oxyma (1:1:1, 5 equivalents) coupling at 88 °C, 120 seconds capping with 10 equivalents of DIEA/acetic anhydride (optional). When Dbz(Alloc) was used as a linker with microwave, the deprotection conditions were modified to be carried out at 25 °C for 2× 5 min and other conditions were not changed. Fmoc-MeDbz or Fmoc-Dbz was manually coupled to rink amide MBHA resin with HCTU activation (amino acid/HCTU/DIEA=2.2/2/4.4 equivalents to resin loading). Alloc protection (overnight treatment of the resin with 250 mM allyl chloroformate in DCM with 1 eq DIEA to resin loading) was followed to prepare Fmoc-Dbz(Alloc) resin. Amino acid coupling directly onto 4-Alloc-Dbz resin was accomplished using HATU activation (amino acid/HATU/DIEA=16.5/15/33) with 1 hour coupling time³⁰, followed by acetyl capping using a fresh solution containing 15% acetic anhydride and 15% DIEA in DMF. Amino acid following the HMBA linker was double-coupled as the symmetric anhydride: 10 equivalents Fmoc-AA-OH were dissolved in DCM. 5 equivalents of DIC were added and incubated on ice for 30 minutes. The filtrate and 0.1 equivalent Dc MAP were added to DMF-swollen resin. The reaction was allowed to proceed for 1 hour. The removal of Alloc was accomplished on resin by treatment with 0.35 eq Pd (PPh₃)₄ and 20 eq Phenyl silane in DCM for 45 minutes. If not specified, Dbz was converted into Nbz by treating the resin with 50mM NPCF in DCM for 30 minutes, followed by 0.5 M DIEA in DMF for 15 minutes as reported²⁴. MeDbz is converted into MeNbz by treating the resin with 0.25M 4-NPCF in DCM for 2h, followed by 1M DIEA treatment in DMF for 1h as described previously³¹. Other conditions will be specified when needed. All peptides were cleaved in 95:2.5:2.5 TFA:TIS:water for 2 hours where otherwise specified. TFA was reduced with a stream of nitrogen, and peptides were precipitated and washed with cold Et₂O, then resuspended in water/acetonitrile and lyophilized before analysis and purification. Analytical reverse phase HPLC (RP-HPLC) was run on a Shimadzu or Waters instrument using an analytical column (Supelco C18 15 cm × 4.6 mm × 5 μm, flow rate 0.9 mL/min). Preparative RP-HPLC was run on a Waters instrument using a semi-preparative column (Supelco C18 25 cm × 10 mm × 10 μm, flow rate 5 mL/min), or a preparative column (Supelco C18 25 cm × 21.2 mm × 10 μm, flow rate 18 mL/min). Mini-preparative RP-HPLC for full-length linker histone H1.2 was run on a Shimadzu instrument using an analytical column (Discovery BIO Wide Pore C5 10 cm × 4.6 mm × 5 μm). Buffer A was 0.1 % TFA in water, and Buffer B was 1:9 water:acetonitrile, 0.1% TFA. Peptide masses were confirmed by MALDI-TOF-MS (Bruker Daltonics Microflex) and analyzed using flexAnalysis 3.3 software. α-Cyano-4-hydroxycinnamic acid (HCCA) was used for the matrix and Peptide Calibration Standard II (Bruker) or Protein Calibration Standard I (Bruker) was used for calibration.

Fmoc Deprotection of Base Resin

200 μL of the base resin Fmoc-Thz-Ala-Ahx-Tyr-Lys-Gly-Rink-PL PEGA, swelled in methanol was measured in a 1.2 mL bed volume Bio-Spin Chromatography column (Bio-Rad). Based on the theoretical loading of 0.0025 mmol/mL, the scale was 0.5 μmol. The resin was treated with three repetitions of 20% Piperidine in NMP for 5 minutes. The resin was then washed with 15 column volumes of DMF, followed with 5 columns volumes of methanol, and then with 15 column volumes of water. The resin was nutated in water for 10 minutes. The resin was then flow-washed with 9 column volumes of Wash Buffer, and nutated in Wash Buffer for 5 minutes. The flow-wash/nutation step was repeated 3 more times.

Thz Deprotection on Solid-Phase

Conversion of thiazolidine to cysteine was typically carried out in Thz Ring Opening Buffer (RO buffer). Resin was treated with 2–3 column volumes of buffer and nutated for 2 hours. The column was drained, and the cycle repeated for a total of 3–4 cycles. The resin was flow-washed with 15 column volumes of wash buffer, then nutated in wash buffer for 5 minutes and drained. This wash cycle was repeated x2. Finally wash buffer with 10mM TCEP was added, nutated for 10 minutes, drained, and flow-washed with 5 column volumes wash buffer, then nutated with ligation buffer for 5 min and drained prior to ligation.

Thioester resin activation

Diglycolic acid PEGA resin was activated with Thiophenol/DIC/DMAP in DMF (for 1mL reaction mixture 48μL DIC 30μL thiophenol, 10μL 0.2M DMAP were added). The activation was carried out twice, each for 30min. The activated resin was washed with DMF thoroughly and then with ethanol followed by water in a flow-wash/nutation manner. Flow-wash with 3 column volumes Wash Buffer, followed by 2–3 column volumes ligation buffer, and drain before addition of peptide in Ligation Buffer.

Solid-Phase Native Chemical Ligation

For a typical ligation cycle, after thorough washing, ~2–5 equivalents peptide was dissolved in Ligation Buffer, and reaction allowed to proceed with nutation for 4 hours. Ligation was assessed via microcleavage and MALDI-TOF MS, supplemented by RP-HPLC analysis of the peptide content of the supernatant liquid. If complete, resin was drained, and flow washed with 15 column volumes of Wash Buffer. An additional Cys wash step was optional; 50 mM cysteine in ligation buffer was added to resin and nutated for 30 min at room temperature, followed by thorough washing.

Solid-Phase Desulfurization

2–3 column volumes of desulfurization buffer (3 M Guanidine, 0.1 M phosphate, 400mM TCEP, 150 mM MESNA, pH 7.4) were added to the resin. The solution was thoroughly sparged with argon. Free radical initiator VA-044 was added (to 20 mM final concentration), and the mixture was heated to 42 °C in a water bath. Reaction proceeded until complete as assessed by MALDI-TOF MS; typically, 4 hours. If reaction was incomplete after 4 hours, resin was drained, washed, and the cycle repeated.

C-terminal segment product release: NaOH Cleavage at the HMBA linker

C-terminal ligated segment was released from the resin by treatment with 0.01 M NaOH for 30 minutes. In a typical reaction, 500μL of 0.01M NaOH was added and nutated for 30 minutes at room temperature; pH was checked to ensure the reaction was carried out at pH 10.0. The column was drained, and 500μL of 0.01 M HCl was added and drained to neutralize the NaOH. 3x additional column volumes were added and drained into the cleavage. The resin was rinsed with water (stored separately), and finally extracted with TFA (stored separately) to ensure maximal product recovery.

Solution Phase Ligation and Desulfurization

Typically, 583 μg H1.2-(Ser₁-Gly₉₉)-Dbz-x’ (quantified by SDS-PAGE, 50 nmol) was dissolved in 50 μL acidified guanidine buffer (6M guanidine, 0.1M phosphate, pH 3–4) and prechilled in ice/salt bath (−15 °C). 5 μL 0.4M NaNO₂ (2 μmol, 40 equivalents) was added and incubated in ice/salt bath for 20min. 2.0mg H1.2-(Cys₁₀₀-Lys₂₁₂)-OH (200 nmol, 4 equivalents) was dissolved in 50 μL MPAA buffer (6M guanidine, 0.1M phosphate, 0.1M MPAA, pH 7) and added to the above NaNO₂-activated peptide solution. The pH was adjusted to ~6.5. The reaction process was monitored by SDS-PAGE. The ligation was generally complete within the several hours. The ligation mixture was dialyzed against guanidine buffer (6M guanidine, 0.1M phosphate, pH 7) three times to remove all MPAA. Desulfurization was carried out under the conditions same as solid-phase described above.

Preparation of core histones and histone octamer

Core histones and histone octamer were produced as previously described³² with some modifications. Briefly, human H2A(K119C), H2B, H3(C110A), and H4 were expressed in Rosetta or BL21 (DE3) pLysS cells. Core histones were purified in denaturing conditions using size exclusion chromatography followed by cation exchange chromatography. Histone octamer was formed by resuspending lyophilized core histones in unfolding buffer (20 mM Tris-HCl pH 7.5, 7 M guanidinium, 10 mM DTT) at 5 mg/mL and mixing in a ratio of H2A:H2B:H3:H4 of 1.1:1.1:1:1. Histones were refolded, and octamer produced by performing double dialysis into refolding buffer (10 mM Tris-HCl pH 7.5, 1 mM EDTA, 2 M NaCl, 5 mM BME). Histone octamer was labeled with Cy5 maleimide (GE Healthcare) as described previously³³ with addition of 10 fold molar excess dye dissolved in DMF at 22 mM reacted at room temperature on a spinning rotisserie for 1 hour then overnight at 4 °C. Labeling was quenched with 10 mM DTT and octamer was purified with a Superdex 200 size exclusion column (GE Healthcare) to remove free histones, heterodimer, and excess dye.

Preparation of nucleosomes

Nucleosomes were prepared, as done previously^32,34, by mixing DNA (147 base pair Widom 601 positioning sequence³⁵ – see supplemental for preparation details) and octamer in a molar ratio of 1.25:1 (DNA:octamer) in 0.5x TE pH 8.0, 2 M NaCl, 1 mM benzamidine-HCl in 50–100 μL and reconstituted by salt double dialysis into 0.5x TE pH 8.0, 1 mM benzamidine-HCl. Nucleosomes were added to the top of a 5%−30% w/v sucrose gradient and purified with an Optima L-90K Ultracentrifuge (Beckman Coulter) with a SW41 rotor spinning at 41,000 rpm for 22 hours at 4 °C. Sucrose gradients were fractionated into 0.4 mL fractions and those containing correctly positioned nucleosomes were concentrated and buffer exchanged into 0.5x TE pH 8.0.

H1 concentration determination

The concentration of H1 used for in vitro assays was determined with Coomassie stained SDS-PAGE gels and a modified Lowry assay³⁶ (Pierce). For Coomassie stained gels, a range of H1.0 (NEB) of known concentration, determined by absorbance at 280 nm by the manufacturer, was loaded onto the gel with the unknown H1. The intensity of stained bands was determined by ImageJ and a standard curve was made with the H1.0 and used to quantify unknown H1 concentrations. Separately, concentration was determined with a modified Lowry assay following the manufacturer’s instructions using H1.0 to produce a standard curve and calculate the unknown H1 concentration.

Electrophoretic mobility shift assay

EMSAs were performed with 1 nM nucleosomes incubated with a range of H1 (0–300 nM) in 10 mM Tris-HCl pH 8.0, 130 mM NaCl, 10% Glycerol, 0.005% TWEEN20 in a 20 uL volume at room temperature for 20 min. Reactions were ran on a 4% polyacrylamide, 0.3x TTE, 10% glycerol gel at 4 °C for 2 hours and imaged on a Typhoon imager (GE Healthcare).

Results and Discussion

New strategy increased the yield of CENP-A synthesis by 2.5 fold

Previously, we prepared CENP-A from five peptide fragments with three assembled on solid phase and extra two rounds of solution phase ligation²⁴. The solution phase ligation and following purification significantly lowered the overall yield to ~7%. With the new strategy, we first tested its feasibility in the preparation of CENP-A from five peptide fragments with splitting sites as previouly reported. We divided the five peptides into two groups, and each group of peptides were assembled on solid support to form a polypeptide block in high yield and purity (Figure 2A). The N-block consists of two peptides CpA1 (residue 1–33) and CpA2 (residue 34–69) and the C-block consists of three peptides CpA3 (residue 70–96), CpA2 (residue 97–114), and CpA5 (residue 115–139). Assembly of CpA345 on solid phase was as described previously with a 4-(hydroxymethyl) benzoic acid (HMBA) linker for protection of terminal carboxylic acid (Figure 1A and 2A)^24,37. A Dbz moiety was equipped in the C-terminal of CpA2 peptide to enable the new strategy and side chain cysteine is installed for capture of the peptide by thioester resin (Figure 2A). Both blocks assembled on solid support were obtained in high yield and purity (Figure 2B and 2C), which can be used directly in the final solution phase ligation. With this new strategy, we achieved 18% overall yield in the preparation of CENP-A, which is 2.5-fold high than that of hybrid phase ligation (Figure 3).

Figure 2.

CHP-NCL improves the yield of CENP-A synthesis by 2.5-fold with single purification. A. Scheme of convergent hybrid phase native chemical ligation for preparation of CENP-A; the linker indicated as a black box is the same structure as in Figure 1A. Solid phase assembly of CpA345. C. Solid phase assembly of CpA12.

Figure 3.

Preparation of CENP-A by CHP-NCL. A. Solution phase ligation of CpA12 and CpA345 followed by desulfurization and RP-HPLC purification to give the final full-length synthetic CENP-A protein. B. MALDI-TOF MS characterization of the desulfurization. C. MALDI-TOF MS of the purified CENP-A.

After successful preparation of CENP-A in high yield with the new strategy, we seek to extend its application to the synthesis of 212-residue linker histone H1.2³⁸ constructing from more peptide fragments. In eukaryotic cells, the linker histone H1s play important roles in chromatin compaction and dynamics^39,40. In the past decades, many different post-translational modifications (PTMs) have been identified³⁹ but the molecular mechanism of these modifications largely remain to be explored partially due to the inability to prepare homogeneously modified protein samples. We anticipate that the capability to synthesize linker histone H1s with site-specific PTMs will shed new lights in related structural and functional studies.

Side reactions on Dbz and its derivatives during peptide thioester synthesis

When successfully preparing all required peptides with Dbz or its variants by Fmoc-SPPS, we found some useful notes on the choice of Dbz and its derivatives for peptide thioester preparation. While Dbz is a useful epimerization-free thioester precusor in Fmoc-SPPS²⁸, branched side products have been observed on plain Dbz in the synthesis of Gly-rich peptides, leading to the development of derivatives Dbz(Alloc) and MeDbz^30,31, which make capping step possible and eliminate over-acylated products. Intriguingly, during the synthesis of the H1.2 peptides, we found that various problems exist in the application of Dbz and its variants (full discussion in ESI). We found that N-methyldiaminobenzoic acid (MeDbz) is compatible with microwave peptide synthesis, but it adds complexity to the characterization of the peptide-MeDbz intermediate due to formation of the benzimidazole derivative under acidic cleavage or HPLC conditions⁴¹ (it does not affect the final results). We also observed reduced yield of the expected N-acylurea products resulted from incomplete MeNbz conversion for some peptides with the recommended conditions as previously reported and figured out that increased temperature could drive the conversion into completion. The N-(alloc)diaminobenzoic acid derivative is compatible with microwave synthesis only if piperidine treatment for Fmoc removal is carried out at room temperature while heated deprotection cycles resulted in premature N-acylurea formation and peptide release from the resin⁴². However, we observe minimal product loss (~1% loss per coupling cycle) when carrying out coupling steps using microwave heating. In the review process of this manuscript, another derivative Dbz(Boc)⁴³ has been reported to overcome the Nbz formation in high temperature Fmoc-SPPS which might be an alternative to prepare the N-terminal anchor peptide.

Assembly of linker histone H1.2 blocks on solid support

With all H1.2 peptides in hand, we proceeded to assess the solid phase ligation for the two major protein blocks. To generate the N-terminal block (18), we first prepared thioester resin (2). Rink amide PEGA resin was treated with diglycolic anhydride. This was further treated with thiophenol in presence of DIC and DMAP to obtain resin thiophenol thioester. Anchor peptide H1-D (14) was added, and native chemical ligation afforded the linked product. A key step was a capping step in which the resin was treated with excess of cysteine in ligation buffer to deactivate any remaining thioester, to avoid potential cross-reactivity in subsequent ligation steps.

Treatment with methoxylamine afforded the “ring-opened” Cys product suitable for use in subsequent ligation cycles (15–18). For documentation of these in itial syntheses, ligation progress was also monitored by RP-HPLC (Figure 4B) and SDS-PAGE (see ESI); however, these analyses consume significantly more material than mass spectrometric analysis, and are less sensitive than MALDI-TOF MS for detection of unreacted material. For preparative purposes, we recommend eliminating these steps and restricting analysis to MALDI-TOF MS. In all, we carried out 4 replicated assembly of the N-terminal segment 18desulf to assess the robustness of the approach; one with unmodified segments, and three with citrullinated Arg53 to generate 18desulf-R53cit.

Figure 4.

Preparation of linker histone H1.2 with CHP-NCL strategy. A. Ligation scheme of H1.2 through CHP-NCL. B. Solid phase assembly of each peptide block monitored by RP-HPLC and MALDI-TOF MS.

Typically, desulfurization is carried out as the last step in the convergent ligation scheme. However, desulfurization might be necessary prior to the full-length protein (for example in the preparation of proteins with all cysteines in C-terminus as ligation sites) in some cases and we noted that the thiols in the N-terminal segment 18 are not required for subsequent ligation steps. We therefore explored desulfurization on the solid phase to further validate our approach. Across the four preparations, yields after four ligation/desulfurization cycles ranged from 50–72% by weight, in sufficient purity to use directly in solution phase ligations.

After validating the efficiency of preparation of N-terminal segment on solid phase, we followed the protocols previously developed for the synthesis of H4 and CENP-A²⁴ for preparation of the C-terminal block. PEGA-PL resin was used for solid support. Fmoc-Thz-Ala-Ahx-Tyr-Lys-Gly handle sequence (sufficient for detection by RP-HPLC and MALDI-TOF MS) was introduced on the rink amide PEGA resin. Immediately prior to the first ligation, the Fmoc was removed by treatment with 20% piperidine in DMF. Treatment with methoxylamine then afforded the N-terminal cysteine, to enable the attachment of H1-(Thz₁₈₉-Lys₂₁₂)-HMBA-RG-MeNbz/Nbz through native chemical ligation. After ligation, resin was washed, and the cycle was repeated for the remaining rounds of SP-NCL. To assess the robustness of our approach, we repeated this synthesis 3 times for unmodified peptide segments. Across the three syntheses, the yields ranged from 77–81% by weight, and the crude product was typically sufficiently pure to be used without purification for the final solution phase ligation (Figure 2B and ESI).

In the preparation of C-terminal segment with Ser172ph as described above, we assessed an alternative rapid palladium-based protocol for deprotection of the thiazolidine⁴⁴. Given the lower purity observed for this product (details in ESI), we recommend the methoxyamine protocol on the solid phase. However, the crude product (73% yield) was still of sufficient purity to use without further purification.

Solution-phase ligation of blocks into full-length linker histone H1.2

With both large protein segments in hand, we proceeded to the final solution phase ligation. The N-terminal peptide 18desulf is activated by treatment of sodium nitrite to form the triazole intermediate, for conversion to thioester in situ for the following ligation. In initial ligation tests, we carried out thioester conversion of the N-terminal followed by addition of the C-terminal segment in ligation buffer (6M Gdn, 0.1M phosphate, 50mM MPAA). However, with ligation carried out at pH 7.3, we found significant hydrolysis of the triazole in competition with thioester formation to reduce our overall yield. The hydrolysis problem has been reported in the case of hydrazide activation and solved by the use of alternative thiol catalyst⁴⁵. Here we reasoned that linker histone peptide is super basic and the pH might play a significant role. We demonstrated the hypothesis using the model peptide [H1.2(Lys₈₉-Gly₉₉)-Dbz-Gly-NH₂], and found minimal hydrolysis over 6 hours at pH 6. We also assessed whether the process could be simplified by pre-mixing 18desulf with 24 prior to activation^29,46, rather than sequential addition of 24 to converted 25. We found no significant differences in ligation kinetics or yield for this considerably simpler experimental protocol. We therefore converged to ligation via activation of the mixed components, followed by addition of ligation buffer to reach pH ~6.5 for ligation over 4 hours at room temperature (Figure 5A).

Figure 5.

Total chemical synthesis of modified linker histone H1.2 with CHP-NCL strategy. A. Final solution phase ligation of citrullinated H1.2 N- and C-block peptides monitored by RP-HPLC. B. SDS-PAGE monitoring of final solution phase ligation. C. Purified H1.2-R53cit protein checked by SDS-PAGE. D. MALDI-TOF MS result of purified H1.2-R53cit. E. Binding of modified or unmodified synthetic linker histone H1.2, and commercially available recombinant H1.0 to mono-nucleosome checked by EMSA test. Nuc: nucleosome, Agg: aggregates.

The ligation product 25 was dialyzed extensively to remove MPAA, which can quench phosphine-mediated desulfurization⁴⁷. The final full-length protein was purified by RP-HPLC after desulfurization. Although the ligation reaction proceeds to near completion (Fig 5A and 5B), purification yields are typically poor for histone proteins – which is one driving factor for the development of minimal purification approaches such as solid phase NCL. A 11% isolated yield (assessed by in-gel quantitation against a commercial H1.0 standard) was obtained for the full-length H1.2-R53cit after ligation, desulfurization, and RP-HPLC purification. All ligations were characterized by SDS-PAGE (Figure 5B and ESI). The purity of the synthetic protein was confirmed by SDS-PAGE and MALDI-TOF MS (Figure 5C and 5D). The protein identity was further confirmed by high resolution mass spectrometry (HR-MS) in all cases (details in ESI).

Synthetic linker histone H1.2 retains its function binding to nucleosome

The primary native function of linker histones, including variants H1.0 and H1.2, is to interact with nucleosomes to form a chromatosome, which facilitates chromatin compaction and may alter nucleosome dynamics and local protein interactions^48–51. Recent structural determination provides a consensus structure for the folded core of the H1.0 protein bound to the nucleosome⁵², although the majority of the H1 sequence remains relatively unstructured. To assess the functionality of our synthetic proteins, we carried out electrophoretic mobility shift assays (EMSA) to compare the binding of commercial unmodified H1.0 (NEB), our synthetic unmodified H1.2, H1.2-R53cit, and H1.2-S172ph to purified mono-nucleosomes (Figure 5E). In each case, we observe that the mononucleosome band begins to shift at 5–10 nM with similar changes in electrophoretic mobility indicating that the unmodified and modified synthetic H1.2 form chromatosomes similarly to recombinant H1.0. At higher concentrations (≥30 nM), the nucleosomes do not migrate into the gel, which is consistent with the formation of larger aggregates mediated by H1. EMSA is not expected to be sufficiently sensitive to detect quantitative differences in H1.2 function due to individual modifications. H1 binds nucleosomes at picomolar affinities⁵³ and has a very slow dissociation rate that is great than 10 minutes⁵⁴. So, in this EMSA assay, H1 binds stoichiometrically; however, these results clearly demonstrate that the synthetic proteins carry out their expected functions, and that the modifications do not reduce the ability of H1 to form a chromatosome. These syntheses therefore enable future studies to determine how H1 modifications, individually or in combination, may impact nucleosome dynamics or macromolecular interactions, for example transcription factor binding to DNA sites within the linker histone-bound nucleosome. More investigation on the function of these synthetic proteins with site-specific modifications has been published somewhere else⁵⁵.

Conclusion

In summary, we have developed a practical method to allow the benefits of solid phase ligation to be extended to the total synthesis of large proteins by exploiting diaminobenzoic acid (peptide o-aminoanilides) as a cryptic thioester. This allows the modular assembly of ~100-residue protein segments from small, synthetically accessible peptides which can be further ligated into a big protein. We improved the yield of CENP-A synthesis by 2.5-fold with the new strategy. We also documented its extension to the preparation of modified or unmodified linker histone H1.2 in a reasonable yield. While we have demonstrated this new strategy for proteins assembled from two “blocks” with only one purification step, we hypothesize that the method could logically be extended to one-pot, sequential, or convergent ligation of 3–5 “blocks” for the facile assembly of very large proteins. We also anticipate that this strategy may find wide applications for researchers from other fields when combined with many developed protein synthesis approaches, such as expressed protein ligation and enzymatic peptide ligation⁶.

Data Availability Statement

The data supporting the findings of this study can be found in the article and supporting materials. Any additional related data are available from the corresponding authors on request.