Fmoc-O-allyl-D-tyrosine

The synthon for O-thiophosphotyrosine, Fmoc-Tyr[PS(OBzl)2]-OH (1c), was prepared in 63% yield from Fmoc-Tyr-OH by first transient protection as the tBuMe2Si-ester and phosphinylation with (BzlO)2PNiPr2/tetrazole followed by oxidation of P(III) to P(V) with S8 in CS2. Building block 1c was incorporated in the Fmoc solid-phase synthesis of two O-thiophosphotyrosine-containing peptides H-Thr-Glu-Pro-Gln-Tyr(PS)-Gln-Pro-Gly-Glu-OH (2) and H-Thr-Arg-Asp-Ile-Tyr(PS)-Glu-Thr-Asp-Phe-Phe-Arg-Lys-OH (3), corresponding to sequences of the p60src (523-531) protein and an insulin receptor (IR) (1142-1153) analogue, respectively. An alternative approach of synthesis, the global phosphorylation of a resin-bound peptide, also proved useful. Thus, the free tyrosyl side-chain containing-peptide IR (1142-1153) on support was phosphinylated with the above phosphoramidite reagent followed by oxidation with either S8/CS2 or tetraethylthiuram disulfide/CH3CN solutions. Deprotection and peptide-resin cleavage was performed with a TFA/thiophenol (H2O) mixture. Crude peptides 2 and 3 were stable to the acidolytic deprotection. Preparative RP(C18)HPLC was initially performed using 0.1% TFA(aq)/EtOH solvents. However, analyses of fractions resulting from the purification step indicated significant decomposition of thiophosphopeptide in solution. Stability measurements both as a function of time and pH, further confirmed this initial finding. Purifications performed at intermediate pH using a triethylammonium acetate (pH 7.5)/CH3CN solvent system overcame this problem.

We introduce solid-phase syntheses of H- and methylphosphonopeptides, giving access for the first time to a new class of mimics for o-phosphoamino acids. The model peptides H-GlyGlyXaaAla-OH (Xaa = Ser, Thr) were synthesized on a solid-phase using Fmoc/tBu strategy and HBTU/HOBt activation by incorporation of hydroxyl-protected serine and threonine. As selectively cleavable hydroxyl-protecting groups we used triphenylmethyl and tert-butyldimethylsilyl for both amino acids, as described in the literature. All peptides were phosphitilated with O, O-di-tert-butyl-N,N-diethylphosphoramidite and yielded H-phosphonopeptides after trifluoroacetic acid cleavage. Alternatively we phosphonylated the peptides with O-tert-butyl-N,N-diethyl-P-methylphosphonamidite, which was synthesized by a two-step one-pot procedure starting from commercially available chemicals. All H- and methylphosphonopeptides were obtained in high purities and yields, as shown by reversed-phase high-performance liquid chromatography and anion-exchange chromatography. The phosphonopeptides were characterized by 1H and 31P NMR. We confirmed their molecular masses by electrospray mass spectrometry and analyzed their fragmentation schemes, which seemed to be characteristic for each class of analogues. The H-phosphonopeptides lost phosphonic acid (H3PO3, 82 mass units) and the methylphosphonopeptides lost methylphosphonic acid (MeH2PO3, 96 mass units). Both H- and methylphosphonopeptides represent a new and simply accessible class of mimics for phosphopeptides. Compared with the corresponding phosphopeptides all phosphonopeptides were synthesized in higher yields and purities (> 80%).

Many proteins are regulated by reversible O-glycosylation and O-phosphorylation. Whereas O-glycosylation of hydroxy-L-proline is common and well investigated, phosphorylation has not been proved so far in vivo, but this post-translational modification is entirely possible. As a first step to identify this phosphoamino acid, we describe both the syntheses of peptides phosphorylated at 4-hydroxy-L-proline and the 1H and 31P NMR parameters of these phosphopeptides. The model peptides were synthesized on solid-phase using Fmoc-strategy. Both natural isomers of 4-hydroxy-L-proline (containing the hydroxyl group in either the cis or trans position) were introduced without side-chain protection. All peptides were globally phosphorylated with O,O'-tert-butyl-N,N-diethylphosphoramidite on the solid phase and cleaved with trifluoroacetic acid. Additionally, we synthesized two classes of phosphonopeptides that mimic phosphopeptides, namely H- and methylphosphonopeptides. The NMR data were based on the model peptide Gly-Gly-Hyp-Ala, which is regarded as a typical random-coil sequence. The NMR parameters showed a significant influence of the phosphate group on the cis-trans isomerization of the Gly-Hyp bond, which may reflect a possible regulation of proteins by changing their local conformations. The 1H and 31P NMR parameters differed for each isomer, and were distinct from the parameters of phosphorylated serine, threonine and tyrosine. These known shifts can be used to identify both cis- and trans-O-phospho-4-hydroxy-L-proline in vivo.