A general preparation of protected phosphoamino acids

Article abstract of DOI:10.1021/ol203332j

Fmoc-O-benzyl-l-phosphoserine is an important building block in the synthesis of Forigerimod, a phosphopeptide being investigated for Systemic Lupus Erythematosus (SLE). An efficient one-pot process was developed using inexpensive, readily available starting materials. This general procedure was used to prepare a variety of protected phosphoamino acids.

Full text of DOI:10.1021/ol203332j

ORGANIC
LETTERS
2012
Vol. 14, No. 5
1206–1209
A General Preparation of Protected
Phosphoamino Acids
Daniel E. Petrillo,* Dale R. Mowrey, Shawn P. Allwein, and Roger P. Bakale
Chemical Process Research and Development, Cephalon Inc., 383 Phoenixville Pike,
Malvern, Pennsylvania 19355, United States
dpetrill@cephalon.com
Received December 13, 2011
ABSTRACT
Fmoc-O-benzyl-L-phosphoserine is an important building block in the synthesis of Forigerimod, a phosphopeptide being investigated
for Systemic Lupus Erythematosus (SLE). An efficient one-pot process was developed using inexpensive, readily available starting materials.
This general procedure was used to prepare a variety of protected phosphoamino acids.
Protein phosphorylation and dephosphorylation are
ubiquitous reactions involved in many important cell
processes, including cell signaling.1 Accordingly, the pre-
paration of phosphopeptides has received much attention
in the chemical literature.2 Phosphopeptides can be pre-
pared either by phosphorylation of an existing peptide
(global phosphorylation strategy) or by the use of indivi-
dual phosphoamino acid building blocks in standard pep-
tide synthesis (prephosphorylation strategy). As withother
amino acids, the choice of the side chainprotecting group is
critical in order to minimize side reactions.
Fmoc strategy. It contains two serine residues, only one
of which is phosphorylated, making prephosphorylation
an attractive option.
Fmoc-O-benzyl-phospho-L-serine (1) (Figure 2) is a
commonly employed phosphoamino acid building block4
used in the synthesis of forigerimod. It was introduced in
1994 by Wakamiya and co-workers for use in solid phase
synthesis using the Fmoc strategy.3a The use of monoben-
zyl protection is necessary to avoid elimination of the
phosphate group to form dehydroalanine under the basic
Fmoc deprotection conditions.4a In addition, the single
benzyl group can be removed under the same conditions as
other acid-labile side-chain protecting groups.
Compound 1 has been used in the synthesis of several
biologically important peptides, including a portion of
heat shock protein 27,4c and the C-terminus of the c-fos
protein.4d It is commercially available from several ven-
dors; however it is expensive,5 compared to other Fmoc-
protected amino acids, and of inconsistent quality.
Although several syntheses of 1 (Figure 2) have been
reported in the literature, all are multistep, requiring
protection and deprotection of the carboxylate moiety,
and use expensive and/or hazardous reagents.4 For exam-
ple, all published preparations involve the use of dialkyl
N,N-phosphoramidite reagents, introduced by Beaucage
and Caruthers for the synthesis of oligonucleotides.6
Forigerimod (Figure 1) is a spiceosomal 21 amino acid
phosphopeptide currently being investigated for the treat-
ment of Systemic Lupus Erythematosus (SLE).3 It is
synthesized via solid phase peptide synthesis using the
(1) Cohen, P. Eur. J. Biochem. 2001, 268, 5001.
(2) (a) Arendt, A.; Hargrave, P. A. In Peptide Synthesis Protocols;
Pennington, M. W., Dunn, B. M., Eds.; Humana Press Inc.: Totowa, NJ, 1994;
Vol. 35, Chapter 9, p 187. (b) Perich, J. W. In Methods in Enzymology;
Fields, G. B., Ed.; Elsevier: 1997; Vol. 289, Chapter 12, p 245. (c) Attard,
T. J.; O’Brien-Simpson, N.; Reynolds, E. C. Int. J. Pept. Res. Ther. 2007,
13, 447.
(3) (a) Monneaux, F.; Lozano, J. M.; Patarroyo, M. E.; Briand, J. P.;
Muller, S. Eur. J. Immunol. 2003, 33, 287. (b) Muller, S.; Monneaux, F.;
Schall, N.; Rashkov, R. K.; Oparanov, B. A.; Wiesel, P.; Geiger, J.-M.;
Zimmer, R. Arthritis Rheum. 2008, 58, 3873.
(4) (a) Wakamiya, T.; Saruta, K.; Yasuoka, J.; Kusumoto, S. Chem.
Lett. 1994, 1099. (b) Wakamiya, T.; Nishida, T.; Togashi, R.; Saruta, K.;
Yasuoka, J.; Kusumoto, S. Bull. Chem. Soc. Jpn. 1996, 69, 465. (c)
Wakamiya, T.; Togashi, R.; Nishida, T.; Saruta, K.; Yasuoka, J.;
Kusumoto, S.; Aimoto, S.; Kumagaye, K. Y.; Nakajima, K.; Nagata,
K. Bioorg. Med. Chem. 1997, 5, 135. (d) Luo, S.; Li, Y.; Chen, Z.; Abe,
H.; Cui, L.; Nakanishi, H.; Qin, X.; Zhao, Y. Lett. Pept. Sci. 2003, 10, 57.
(5) 1 is available from Sigma-Aldrich for $252/g (2012À2014
catalog).
r
10.1021/ol203332j
Published on Web 02/22/2012
2012 American Chemical Society

After an evaluation of this and other known synthetic
routes4 to 1, we desired a more scalable, step-economical
synthesis for the production of forigerimod. We initially
envisioned the use of either POCl3 or PCl3 as an inexpen-
sive, highly reactive source of phosphorus that would not
require an activating agent. We also wished to access the
monoprotected amino acid directly.
Scheme 1. Wakamiya Synthesis of 14b
Initial reactions employing Fmoc-L-serine, POCl3, and
benzyl alcohol led to complex mixtures, presumably due to
activation of the carboxylate group by POCl3. We then
focused our attention on PCl3 and a report by Prashad and
co-workers, inwhicha cyclicmixedphosphiteintermediate
was oxidized to a phosphodiester (see Scheme 2).8
Figure 1. Forigerimod (Arg-Ile-His-Met-Val-Tyr-Ser-Lys-Arg-
Sep-Gly-Lys-Pro-Arg-Gly-Tyr-Ala-Phe-Ile-Glu-Tyr).
Scheme 2. One-Pot Preparation of Phosphodiester8
Although such reagents are often readily available, they
require the use of an activator, such as 1-H-tetrazole,
which is no longer available due to its explosive nature.
Although other heterocycles have been shown to promote
phosphitylation, many are expensive or require synthesis.7
In addition, most syntheses using phosphoramidites target
a doubly protected phosphoamino acid as the penultimate
intermediate, which requires an additional deprotection
step to form the monobenzyl-protected species.
The original synthesis4a,b by Wakamiya and co-workers
is shown in Scheme 1. In their synthesis, a dialkyl N,
N-phosphoramidite (prepared in two steps) is reacted with
a phenacyl-protected serine derivative in the presence of
3 equiv of 1-H-tetrazole. The resulting phosphite inter-
mediate is oxidized with m-CPBA. Finally the phenacyl
group and the trichloroethyl (TCE) group are removed via
reduction with zinc in acetic acid.
As shown in Scheme 3, treatment of PCl3 (1.1 equiv)
with BnOH (1.0 equiv) in the presence of 2,4,6-collidine
(6.5 equiv) led to the presumed benzyl dichlorophosphite
intermediate, as determined by 31P NMR analysis (177 ppm
in CDCl3). Addition of Fmoc-L-serine as either the anion or
the neutral species led to the presumed cyclic phosphite
intermediate, as detected by 31P NMR analysis (134 ppm in
CDCl3). Hydrolysis led to the acyclic phosphite, detected by
31P NMR analysis and reversed-phase LC-MS. Oxidation
of this species with Br2 (1.3 equiv) led to 1, albeit with
numerous other impurities.
Using these conditions as a starting point, the process
was rapidly optimized. Using 1.3 equiv of PCl3 and 1.5
equiv of BnOH led to a 62% HPLC assay yield of 1, based
on an external standard. Using a greater excess of both
reagents led to lower yields.
Figure 2. Fmoc-O-benzyl-phospho-L-serine.
(7) Maung, J.; Mallari, J. P.; Choy, C. J.; Berkman, C. E. Tetrahedron
Lett. 2004, 45, 6497.
(8) Prashad, M.; Amedio, J. C.; Ciszewski, L.; Lee, G.; Villa, C.;
(6) (a) Beaucage, S. L.; Caruthers, M. H. Tetrahedron Lett. 1981, 22,
1859. (b) Bannwarth, W.; Trzeciak, A. Helv. Chim. Acta 1987, 70, 175.
(c) Perich, J. W.; Johns, R. B. Tetrahedron Lett. 1987, 28, 101.
Chen, K.-M.; Prasad, K.; Repic, O. Org. Process Res. Dev. 2002, 6, 773.
Org. Lett., Vol. 14, No. 5, 2012
1207

Table 1. Effect of Oxidant on the Yield of 1
Scheme 3. One-Pot Preparation of 1
entry
oxidant
yielda
1
2
3
4
5
6
7
Br2
I2
67%
66%
66%
71%
18%
34%
37%
PyrHBr3
NaBrO3/NaBr
NaOCl
Reducing the amount of base to 4.0 equiv (sufficient to
neutralize the HCl generated from the PCl3) resulted in an
improved yield (65%). At this stage, the conditions reliably
generated a 60À70% HPLC assay yield of the desired
compound on a multigram scale.
Oxone
DAIB
a HPLC assay yield using an external standard.
Although bromine is inexpensive and performed well in
the oxidation, a more user-friendly oxidant was desired for
handling reasons. Therefore, a screen of oxidants was
performed (see Table 1). Many reagents are known to
oxidize phosphites to phosphates; however many, such as
t-BuOOH, employ anhydrous conditions, which are not
suitable for our system.
Our attention turned next to the preparation of other
phosphorylated serine derivatives (Table 3). Fmoc-O-methyl-
phospho-L-serine 2 (entry 1) was prepared in 69% HPLC
assay yield and could be isolated by preparative reversed-
phase HPLC chromatography in 39% yield (not optimized).
A previous synthesis of this molecule involved the use of an
additional phosphate protecting group.10
In our hands, iodine (entry 2) resulted in a similar yield,
while aqueous NaOCl, Oxone (potassium monoperoxy-
sulfate), and diacetoxyiodobenzene (DAIB) all provided
<50% HPLC assay yield of 1. Bromine-related oxidants
gave the highest yield, with pyridinium tribromide provid-
ing an equivalent yield to bromine. The most suitable
conditions proved to be an aqueous mixture of sodium
bromide and sodium bromate, known to form molecular
bromine at low pH. To our knowledge this is the first use of
these reagents for the oxidation of phosphorus in either
phosphoamino acids or nucleotides.
The use of the (()-1-(2-nitrophenyl)ethyl (NPE) pro-
tecting group (entry 2) was also investigated. Imperiali and
co-workers have shown that phosphoserine derivatives
employing this protecting group can be incorporated into
phosphopeptides.11 The NPE group can be removed in the
presence of far UV light (>350 nm), unmasking the free
phosphate group. A related reagent had previously been
prepared as a doubly protected phosphate in several steps.
In our hands, the one-pot procedure gave a 51% isolated
yield of 3 as a mixture of diastereomers after preparative
chromatography.
Other bases were also screened (see Table 2). The study was
limited to organic bases due to the insolubility of inorganic
bases in the reaction medium (THF). In addition, secondary
amines were ruled out due to their ability to remove the Fmoc
protecting group. The highest yield was obtained with 2,6-
lutidine (entry 3), although other pyridine-related bases also
Table 2. Effect of Base on the Yield of 1
€
performed well (entries 1,4,5). Hunig’s base (entry 2) gave a
slightly lower yield, while DBU resulted in an insoluble
precipitate that inhibited stirring of the mixture.
The unique properties of 1 made the isolation difficult
initially. Most literature preparations involve direct pre-
cipitation of1; howeverthiswasnot feasibleinthe presence
of other related impurities. The use of 2-MeTHF proved
critical to extract 1 from the biphasic reaction mixture
present at the end of the reaction. Fortuitously, upon
concentration of the solvent, 1 crystallized as a 1:1 complex
entry
base
yielda
1
2
3
4
5
2,4,6-collidine
i-Pr2EtN
63%
59%
71%
68%
67%
2,6-lutidine
2-picoline
pyridine
a HPLC assay yield using an external standard.
1
with 2-MeTHF, as determined by H NMR, in >95%
purity in 55À60% yield. The pure, desolvated compound
could be obtained by stirring 1 in the presence of solvents
such as EtOAc or CH2Cl2.9 Gratifyingly, analysis of the
isolated product by chiral HPLC indicated that no race-
mization occurred.
The use of the CBz protecting group (entry 3) resulted in
a 63% HPLC assay yield of 4. The use of Boc-L-serine
resulted in a 45% HPLC assay yield; however, chromato-
graphy resulted in decomposition.
1208
Org. Lett., Vol. 14, No. 5, 2012

Table 3. Preparation of Phosphoserine Derivatives
entry
R1
R2
product
yielda
1
2
3
4
Me
Fmoc
Fmoc
Boc
2
3
4
5
69% (39%)
79% (51%b)
45% (nd)
1-(2-NO2-Ph)Et
Bn
Bn
Cbz
63% (16%)
Figure 3. Additional phosphoamino acids prepared with HPLC
assay yields (nonoptimized isolated yields are in parentheses).
a HPLC assay yield using an external standard; isolated yields (not
optimized) are in parentheses. b 93% pure by HPLC.
prepared in a multistep procedure involving protection
of the carboxylate group as well as phosphorylation with a
phosphoramidite reagent.12
In summary, a highly efficient, one-pot synthesis of
protected phosphoamino acids was developed using inex-
pensive, readily available reagents. This method allows the
rapid preparation of these useful building for use in the
synthesis of phosphopeptides.
Finally, the phosphorylation of other Fmoc-hydroxy-
amino acids was performed. Fmoc-O-benzyl-phospho-L-
threonine (6) was prepared in 72% HPLC assay yield and
crystallized in 51% yield (Figure 3). This important build-
ing block had previously been prepared via a multistep
synthesis, using phosphoramidite chemistry.4b The iso-
meric Fmoc-L-allo-threonine (7) was also prepared in
88% HPLC assay yield, 55% isolated yield. In addition,
the racemic phenyl substituted derivative 9 was prepared
in 96% assay yield, 48% isolated yield.
The L-homoserine derivative 8 was also prepared in 70%
HPLC assay yield. Interestingly, Fmoc-L-tyrosine (10),
which is not capable of forming a cyclic phosphite, also
resulted in a high HPLC assay yield (86%) and good
isolated yield (69%). This intermediate was previously
Acknowledgment. Mark Olsen, Brendan Mahoney,
Chris Neville, and Emily Kordwitz (Cephalon Analytical
Development) are acknowledged for performing HPLC/
HRMS analyses, preparative HPLC separations, NMR
analyses, and optical rotations, respectively. Vikram Purohit
(Cephalon Chemical Research and Development) is thanked
ꢀ
ꢀ
for helpful discussions. Sebastien Rose and Stephanie
Graf (Cephalon France) are acknowledged for performing
the hazard analysis and for further development of this
chemistry.
(9) The details of the preparation and isolation of 1 on a kilogram
scale will be detailed. Mowrey, D. R. Petrillo, D. E.; Allwein, S. P.; Graf,
S.; Bakale, R. P. Manuscript in preparation.
(10) Bhandari, R.; Saiardi, A.; Ahmadibeni, Y.; Snowman, A. M.;
Resnick, A. C.; Kristiansen, T. Z.; Molina, H.; Pandey, A.; Werner,
J. K., Jr.; Juluri, K. R.; Xu, Y.; Prestwich, G. D.; Parang, K.; Snyder,
S. H. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 15305.
(11) (a) Rothman, D. M.; Vazquez, M. E.; Vogel, E. M.; Imperiali, B.
Org. Lett. 2002, 4, 2865. (b) Rothman, D. M.; Vazquez, M. E.; Vogel,
E. M.; Imperiali, B. J. Org. Chem. 2003, 68, 6795. (c) Goguen, B. N.;
Aemissegger, A.; Imperiali, B. J. Am. Chem. Soc. 2011, 133, 11038.
(12) (a) Perich, J. W.; Reynolds, E. C. Synlett 1991, 577. (b) Handa,
B. K.; Hobbs, C. J. J. Pept. Sci. 1998, 4, 138.
Supporting Information Available. Experimental pro-
cedures, spectral characterization, and copies of 1H, 13C,
and 31P NMR spectra for all compounds. This material is
available free of charge via the Internet at http://pubs.acs.
org.
The authors declare no competing financial interest.
Org. Lett., Vol. 14, No. 5, 2012
1209