Synthesis and Evaluation of Non‐Hydrolyzable Phospho‐Lysine Peptide Mimics

Abstract The intrinsic lability of the phosphoramidate P−N bond in phosphorylated histidine (pHis), arginine (pHis) and lysine (pLys) residues is a significant challenge for the investigation of these post‐translational modifications (PTMs), which gained attention rather recently. While stable mimics of pHis and pArg have contributed to study protein substrate interactions or to generate antibodies for enrichment as well as detection, no such analogue has been reported yet for pLys. This work reports the synthesis and evaluation of two pLys mimics, a phosphonate and a phosphate derivative, which can easily be incorporated into peptides using standard fluorenyl‐methyloxycarbonyl‐ (Fmoc‐)based solid‐phase peptide synthesis (SPPS). In order to compare the biophysical properties of natural pLys with our synthetic mimics, the pK a values of pLys and analogues were determined in titration experiments applying nuclear magnetic resonance (NMR) spectroscopy in small model peptides. These results were used to compute electrostatic potential (ESP) surfaces obtained after molecular geometry optimization. These findings indicate the potential of the designed non‐hydrolyzable, phosphonate‐based mimic for pLys in various proteomic approaches.

Side products occurred during SPPS using building block 3 and corresponding yields. While the desired peptide 12 was obtained, a considerable P-N bond hydrolysis was observed, as well as another side product, in which the detected mass would correspond to a Tyr addition and subsequent acetylation at the lysine side-chain.

Reagents and Solvents
Reagents and solvents were, unless stated otherwise, commercially available as reagent grade and did not require further purification. Chemicals were purchased either from Sigma-

Peptide Synthesis
Peptides were prepared using the Fmoc solid phase strategy by manual peptide synthesis in   Peptides containing the hCys-derivative 1 were treated for 105 min with the standard cleavage cocktail, before a solution of 60 μL EDT and 300 μL TMSBr was added and the mixture shaken for further 15 min. The resin was filtered off, the TFA filtrate collected in a 10-fold excess of deep-frozen Et2O and let sit for precipitation in the freezer. After at least 15 min, the mixture was centrifuged, the solution decanted, the precipitate dried under nitrogen and re-dissolved in ACN/H2O for UPLC analysis and preparative HPLC.

TLC analysis
The thin layer chromatography (TLC) was performed on silica gel plates with fluorescence indicator F254 (Merck KGaA). Detection was performed at 254 or 366 nm. Compounds without any chromophore were stained with any of staining solutions such as potassium permanganate solution, ninhydrin or vanillin reagent.

Purification
Peptidic substrates were purified by preparative, semi-preparative or analytical HPLC, performed either on a Gilson PLC 2020 system (Gilson Inc., Middleton, Wisconsin, USA), a Shimadzu  [2,3] were performed using a gradient version of the experiment in which a WATERGATE [4] water suppression had been implemented. 2,048 ( 1 H) • 64 ( 31 P) complex points were acquired, with acquisition times tH,max = 204.8 ms and tP,max = 6.4 ms (i.e. a spectral window of 10,000 Hz was used in each dimension) and 16 scans. NMR data were processed and spectra viewed using topspin 3.2 (Bruker Biospin).

Phosphatase Activity Assay
Phosphatase activities were determined on a SAFIRE 2

Electrostatic Potential Maps
Electrostatic potential (ESP) maps were calculated relying on density functional theory (DFT) optimized molecular structures. The exchange correlation functional B3LYP [6][7][8] together with the polarized triple zeta basis set def2-TZVPP [9,10] on all atoms was employed and subsequent harmonic vibrational frequency analysis was carried out in order to check if a real local minimum on the potential energy surface was found. From the created electron densities, the electrostatic potential energy (in a.u.) was calculated and mapped on the respective density plot for a contour value of 0.01. The Turbomole program package V7.0.2 [11][12][13] was used for all DFT electronic structure and geometry optimizations. ESP evaluations and visualizations were done with Molden 5.9. [14,15]

Vinylphosphonic dichloride (crude)
Vinylphosphonic dichloride was synthesized as described before with minor adjustments. Briefly, in a

Dibenzyl vinylphosphonate (4)
In a heated flask and in Ar atmosphere, 50 mL dry THF were pre-cooled.

N6-(bis(2,2,2-trichloroethoxy)phosphoryl)-L-lysine, H2N-Lys(NPO(OTc)2)-OH
N-and C-terminal deprotection was achieved with hydrogenation. 600 mg (850 µmol) 11 were weighed into a Schlenk flask, dissolved to a 75 mM solution with AcOH:TFA:MeOH (5:5:90, v/v/v, 11 mL) and kept under vacuum for 5 min. In inert atmosphere, 96 mg Pd on activated charcoal (10% Pd, 113 mg per mmol starting material) were added and the gas exchanged with H2. The mixture was stirred for 45 min in H2 atmosphere until UPLC analysis indicated full conversion. The catalyst was filtered off and washed with MeOH. Solvents were evaporated by bubbling N2 through the solution and residual acid removed by lyophilization. The crude product was applied in the next step without further purification.

1-(2-nitrophenyl)ethanol (14)
In a 500 mL round-bottom flask 10 g (60.6 mmol) 2'-nitroacteophenone were dissolved in 100 mL MeOH:dioxane (3:2, v/v) and cooled with an ice bath. Under vigorous stirring, 2.5 eq. sodium borohydride (151.4mmol, 5.7 g) were added portion wise over 90 min. The resulting mixture was equipped with septum and balloon and left to warm to r.t. ovn while stirring. After 16 h residual NaBH4 was quenched by the addition of 50 mL of acetone and the solvents evaporated under reduced pressure.
The residue was diluted with H2O and EE and the layers separated. The organic layer was washed twice with water, the combined aqueous layers were washed once with EE, eventually, the combined organic layers were washed with brine, dried over