Stereoselective ribosylation of amino acids

Article abstract of DOI:10.1021/ol400929c

The glycosylation properties of ribofuranosyl N-phenyltrifluoroacetimidates toward carboxamide side chains of asparagine and glutamine were investigated. Conditions were found that promote nearly exclusive formation of the α-anomerically configured N-glycosides. The strategy allows for the synthesis of Fmoc-amino acids suitably modified for the preparation of ADP-ribosylated peptides. Furthermore, ribosylation of serine with these donors proved to be completely α-selective, and for the first time, α-ribosylated glutamic and aspartic acid, the naturally occurring sites for poly-ADP-ribosylation, were synthesized.

Full text of DOI:10.1021/ol400929c

ORGANIC
LETTERS
XXXX
Vol. XX, No. XX
000–000
Stereoselective Ribosylation
of Amino Acids
Hans A. V. Kistemaker, Gerbrand J. van der Heden van Noort,
Herman S. Overkleeft, Gijsbert A. van der Marel,* and Dmitri V. Filippov*
Leiden Institute of Chemistry, Leiden University, P.O. Box 9502,
2300 RA Leiden, The Netherlands
marel_g@chem.leidenuniv.nl; filippov@chem.leidenuniv.nl
Received April 4, 2013
ABSTRACT
The glycosylation properties of ribofuranosyl N-phenyltrifluoroacetimidates toward carboxamide side chains of asparagine and glutamine were
investigated. Conditions were found that promote nearly exclusive formation of the R-anomerically configured N-glycosides. The strategy allows
for the synthesis of Fmoc-amino acids suitably modified for the preparation of ADP-ribosylated peptides. Furthermore, ribosylation of serine with
these donors proved to be completely R-selective, and for the first time, R-ribosylated glutamic and aspartic acid, the naturally occurring sites for
poly-ADP-ribosylation, were synthesized.
Glycosylation of proteins is a post-translational mod-
ification that plays a significant role in many biological
processes.1,2 Adenosine diphosphate ribosylation (ADPr)
(Figure 1) is a peculiar type of protein glycosylation that
occurs in both mono- and polymeric form and is consid-
ered to play an important role in a wide range of biological
processes including cell proliferation, immune response,
DNA repair, transcription regulation, and apoptosis.3ꢀ6
The process of ADP-ribosylation requires the enzymatic
transfer of a single ADP-ribose moiety from β-NADþ
to the nucleophilic side chain of an amino acid forming
an R-glycosidic linkage (dotted box, Figure 1). Enzymes
(called PARPs) that are able to transfer additional ADP-
ribosyl units to the 20-OH of the adenosine moiety effect
the formation of poly-ADP-ribose (Figure 1). The various
identified sites amenable to ADP-ribosylation are aspar-
agine, glutamic acid, aspartic acid, arginine, serine, and
cysteine.3,7 The construction of well-defined ADP-ribosy-
lated peptides and analogues thereof would be of signifi-
cant help in gaining a better understanding of the role and
function of ADP-ribosylation.8,9
(1) Walsh, C. T.; Garneau-Tsodikova, S.; Gatto, G. J. Angew. Chem.,
Int. Ed. 2005, 44, 7342.
(2) Khidekel, N.; Hsieh-Wilson, L. C. Org. Biomol. Chem. 2004, 2, 1.
(3) Hassa, P. O.; Haenni, S. S.; Elser, M.; Hottiger, M. O. Microbiol.
Mol. Biol. R. 2006, 70, 789.
Figure 1. Poly-ADP-ribose polymer linked to a peptide.
(4) Gibson, B. A.; Kraus, W. L. Nat. Rev. Mol. Cell Biol. 2012,
13, 411.
(5) Schreiber, V.; Dantzer, F.; Ame, J. C.; de Murcia, G. Nat. Rev.
Mol. Cell Biol. 2006, 7, 517.
(6) Adams-Phillips, L.; Briggs, A. G.; Bent, A. F. Plant. Physiol.
One of the crucial steps in the synthesis of ADP-ribosy-
lated peptides is construction of the R-glycosidic linkage
2010, 152, 267.
(8) Lin, H. Org. Biomol. Chem. 2007, 5, 2541.
(9) Hengel, S. M.; Goodlett, D. R. Int J. Mass. Spectrom. 2012,
312, 114.
(7) Cervanteslaurean, D.; Loflin, P. T.; Minter, D. E.; Jacobson,
E. L.; Jacobson, M. K. J. Biol. Chem. 1995, 270, 7929.
r
10.1021/ol400929c
XXXX American Chemical Society

(entry 1, Table 1). Application of identical conditions for
glutamine resulted in a decrease in selectivity (entry 3,
Table 1). Lowering the reaction temperature improved the
R-selectivity but also reduced yield (entry 2, Table 1).
Changing the solvent to nitromethane14 (CH3NO2) did
improve the yield but the R-selectivity remained poor
(entry 4, Table 1).
Table 1. Ribosylation of Asparagine and Glutaminea
We anticipated that the selectivity could be improved
by introducing a more bulky substituent at the
50-position. For this reason, two new donors with
tert-butyldiphenylsilyl (TBDPS), 11, or triisopropylsi-
lyl (TIPS) at the 50-position, 12, were synthesized
(Scheme 1).
TMSOTf
(equiv)
temp
yield
(%)
ratio
no. acceptor solvent
(°C)
(R/β)16
1
2
3
4
A
B
B
B
CH2Cl2
CH2Cl2
CH2Cl2
CH3NO2
0.2
0.3
0.2
0.3
ꢀ20 to rt
ꢀ78 to rt
ꢀ20 to rt
0
58
38
56
67
96/4
87/13
69/31
77/23
Scheme 1. Synthesis of 50-Silylated Imidate Donors
a Reaction conditions: 1 equiv of acceptor (50 mM), 1.25 equivof
donor, and 1.5 h reaction time.
between the ribofuranosyl moiety and the amino acid side
chain. We have reported the first fully synthetic approach
to ADP-ribosylated peptides via the incorporation of an
R-ribosyl containing glutamine or asparagine building
block by solid-phase peptide synthesis (SPPS) followed
byinstallment ofthe adenosine diphosphate linkage. These
modified amino acids were prepared by us10 and others11
via reduction of ribofuranosyl azide and in situ condensa-
tion of the formed amine to the side chain of suitably
protected glutamic acid or aspartic acid derivatives. The
condensation products in this key amide bond formation
were obtained as an anomeric mixture (75/25: R/β).10
We reasoned that the preparation of side-chain ribosy-
lated amino acids might proceed with better R-selectivity
when using acid-catalyzed glycosylation. Trifluoroaceti-
midate chemistry12 was deemed to be the most viable
for introducing the crucial R-linkage between the ribosyl
moiety and the amino acids.12ꢀ14 Our results in the use of
ribofuranosyl trifluoroacetimidates in the R-selective con-
struction of a series of ribosylated amino acids correspond-
ing to the sites of ADP-ribosylation are reported here.
We decided to use donor 1,13 since it is known from
literature13,15 that nonparticipating benzyl protection on
the ribofuranosyl donor predominantly renders the R-pro-
duct in glycosylation reactions. Donor 1 was reacted with
either N-Cbz- asparagine-OBn (acceptor A, Table 1) or
N-Cbz-glutamine-OBn (acceptor B, Table 1) under various
conditions. Ribosylation of asparagine using donor 1
and TMSOTf as activator gave an excellent R-selectivity
To investigate the glycosylation properties of donor 11
and 12, N-Cbz-glutamine-OBn was reacted with these
under various conditions. The most favorable conditions
for the ribosylation of glutamine turned out to be activa-
tion at ꢀ20 °C with 0.2 equiv of TMSOTf (entries 1 and 2,
Table 2).17 Using either of the two 50-silylated donors, 11
and 12, greatly improved the R-selectivity compared to the
use of 50-benzylated donor 1 (entry 2, Table 1). Ribosyla-
tion of asparagine (A) with TIPS-donor 12 still gave a
higher R-selectivity than that of glutamine (B) although the
difference in selectivity was greatly reduced compared to
the tri-O-benzyl donor (entry 3, Table 2). For these
Table 2. Optimized Ribosylation of Glutamine and Asparaginea
(10) van Noort, G. J. V.; van der Horst, M. G.; Overkleeft, H. S.;
van der Marel, G. A.; Filippov, D. V. J. Am. Chem. Soc. 2010, 132, 5236.
ꢀ
(11) Bonache, M. A.; Nuti, F.; Le Chevalier Isaad, A.; Real-Fernandez,
no. acceptor donor (R1=) product temp (°C) yield (%) ratio (R/β)
F.; Chelli, M.; Rovero, P.; Papini, A. M. Tetrahedron Lett. 2009, 50,
4151.
(12) Yu, B.; Tao, H. C. Tetrahedron Lett. 2001, 42, 2405.
(13) van Noort, G. J. V.; Overkleeft, H. S.; van der Marel, G. A.;
Filippov, D. V. Org. Lett. 2011, 13, 2920.
(14) Tanaka, H.; Iwata, Y.; Takahashi, D.; Adachi, M.; Takahashi,
T. J. Am. Chem. Soc. 2005, 127, 1630.
(15) Larsen, C. H.; Ridgway, B. H.; Shaw, J. T.; Smith, D. M.;
Woerpel, K. A. J. Am. Chem. Soc. 2005, 127, 10879.
1
2
3
B
B
A
11 (TBDPS)
12 (TIPS)
12 (TIPS)
13
14
15
ꢀ20 to rt
ꢀ20 to rt
ꢀ20 to rt
57
66
68
95/5
95/5
98/2
a Reaction conditions: 1 equiv of acceptor (50 mM), 1.25ꢀ1.5 equiv
of donor, 0.2 equiv of TMSOTf, DCM, and 1.5 h reaction time.
B
Org. Lett., Vol. XX, No. XX, XXXX

Table 3. Ribosylation of Fmoc-asparagine and Fmoc-glutamine Benzyl Estersa
no.
donor
acceptor
product
activator
solvent
temp (°C)
yield (%)
ratio (R/β)
1
2
3
4
5
6
7
8
11
11
11
11
12
16
16
16
A1
A1
B1
B2
B1
B
17
17
18
19
20
21
22
22
TMSOTf
dioxane/DCM
dioxane/DCM
dioxane/DCM
dioxane/DCM
dioxane/DCM
DCM
0
44
63
49
46
59
52
60
69
97/3
TMSOTf
rt
93/7
TMSOTf
0
72/28
73/27
96/4
TMSOTf
0
HClO4ꢀSiO2
TMSOTf
rt
ꢀ20 to rt
ꢀ20 to rt
rt
85/15
78/22
93/7
B1
B1
TMSOTf
dioxane/DCM
dioxane/DCM
HClO4ꢀSiO2
a Reaction conditions: 1 equiv of acceptor (50 mM), 1.5 equiv of donor, 0.2 equiv pf activator, and 1.5 h reaction time. b For the synthesis of donor 16,
see the Supporting Information.
ribosylated amino acids to be useful in SPPS, protective
group exchange was required and was exemplified for
compounds 14 and 15 (see the Supporting Information).
It occurred to us, at this stage, that direct ribosylation
of Fmoc-amino acids would give a more straightforward
entry to the target building blocks. The feasibility of this
approach was first tested in the ribosylation of N-Fmoc-
asparagine-OBn with donor 11. Because of the poor
solubility of the Fmoc-amino acid, the solvent system
had to be changed to methylene chloride/dioxane
(1:1, v/v). The ribosylation went almost completely R-selective
but in a moderate yield when performed at 0 °C (entry 1,
Table 3), whereas at room temperature the stereoselectivity
only slightly decreased while the yield significantly increased
(entry 2, Table 3). The yield and selectivity are comparable
with respect to the N-Cbz-protected asparagine derivative,
and this approach circumvents tedious protecting group
manipulations. However, these promising results were not
reflected in the ribosylation of N-Fmoc-glutamine-OBn,
which showed a modest excess of the R-anomer and an
average yield at 0 °C (entry 3, Table 3). Replacing the benzyl
ester with 4-methoxybenzyl (PMB), in an attempt to im-
prove the solubility of the amino acid, did not improve the
outcome of the reaction in any way (entry 4, Table 3).
Therefore, we investigated the use of perchloric acid on silica
(HClO4ꢀSiO2) as activator, for it has been used with glyco-
sylations in the presence of dioxane at room tempera-
ture.18,19 The HClO4ꢀSiO2 reagent was prepared according
to literature procedure20 and added to a mixture of
N-Fmoc-glutamine-OBn and donor 12 at room temperature.
The stereoselectivity increased tremendously along with a
good yield (entry 5, Table 3).
The hydrogenolysisofbenzyl ethersattheribosylmoiety
was found to be incompatible with the integrity of the
Fmoc protection. Therefore, we introduced the acid sensi-
tivePMB at the 20- and 30-position (donor16, Table 3). The
donor proved to be a little less R-selective in the ribosyla-
tion reaction of N-Cbz-glutamine-OBn (entry 6, Table 3)
compared to the benzylated donor 12 (95/5 vs 85/15 R/β).
Ribosylation of N-Fmoc-glutamine-OBn with donor 16 in
the presence of TMSOTf showed a comparable yield
and R-selectivity (entry 7, Table 3) to benzylated donor
11 (entry 3, Table 3), which could be again increased
significantly by activation with HClO4ꢀSiO2 (entry 8,
Table 3).
This optimized strategy allows for the synthesis of suitable
protected glutamine building blocks for SPPS in a straight-
forward approach and at a preparative scale (0.5 mmol). This
is demonstrated for compound 22, which was transformed in
three consecutive steps into compound 23 (Figure 2).21
The results of the glycosylation experiments encouraged
us to test the three donors (1, 11, and 12) on other amino
acid side chains. The anomeric ratio for the ribosylation
of serine (Ser, Table 4) with tribenzyl donor 1 (entry 1,
Table 4) was comparable to glutamine (entries 1 and 2,
Table 1) and to the reported data.22,23 Ribosylation with
the 50-silylated donors, 11 and 12, on the other hand gave
exclusively R-ribosylated serine (entry 2 and 3, Table 4). To
our knowledge, this is the first example of a completely
R-selective ribosylation for an amino acid.
Org. Lett., Vol. XX, No. XX, XXXX
C

Table 4. Ribosylation of Serine, Glutamic Acid, and Aspartic
Acida
Figure 2. Synthesis of Fmoc Building Block 23 Suitable for SPPS.
Next, glutamic acid and aspartic acid were subjected to
ribosylation with donor 12.24 The stereoselectivity for both
amino acids was mostly lost and in the best case a small
excess of R-ribosylated glutamic acid (77/23 R/β) was
obtained (entry 4, Table 4). Lowering the temperature
did not improve the stereochemical outcome of both
reactions and the best result for aspartic acid was a 50:50
mixture of R/β-anomers (entry 6, Table 4). Moreover,
using HClO4ꢀSiO2 as activator did not improve the out-
come of the ribosylations (entries 5 and 7, Table 4). Never-
theless, both anomers could easily be separated by silica gel
chromatography at this stage to afford novel R-ribosylated
glutamic acid and aspartic acid derivatives.
donor
(R1 =)
acceptor
(R2)
temp
yield
(%)
ratio
no.
product
(°C)
(R/β)
1
2
3
4
5
6
7
Bn
Ser
Ser
Ser
Glu
Glu
Asp
Asp
24
25
26
27
27
28
28
ꢀ50
ꢀ50
ꢀ50
ꢀ50
ꢀ50
ꢀ70
ꢀ78
78
60
56
84
73
63
74
75/25
100/0
100/0
77/23
60/40b
50/50
43/57b
TBDPS
TIPS
TIPS
TIPS
TIPS
TIPS
a Reaction conditions: 1 equiv of acceptor (50 mM), 1.25 equiv of
This first successful attempt to obtain R-ribosylated
glutamic acid ever reported might lead to ADP- ribosylated
peptides with glutamic acid,for example those derived from
histone H2B.3 So far, ribosylated glutamine is used as an
isostere of glutamic acid in the synthesis of such peptides.10
donor, 0.2 equiv of TMSOTf. b 0.05 equiv of HClO4ꢀSiO2, DCM for 1.5 h.
In conclusion, we have presented a study on the pre-
paration of various ribosylated amino acids in a highly
R-selective fashion by tuning the protective groups at the
20-, 30-, and 50-position of the ribofuranosyl imidate
donor. Furthermore, the number of synthetic steps to-
ward SPPS building blocks was minimized by using Fmoc
protected amino acids, simplifying the protective group
manipulation strategy. This new protocol allows for the
synthesis of ribosylated Fmoc building blocks in nine
steps starting from D-ribose at a preparative scale. In the
specific case of glutamine, the glycosylation reaction
using HClO4ꢀSiO2 as activator proved to offer great
improvement compared to TMSOTf, with respect to
R-selectivity. Finally, we showed a completely R-selective
ribosylation and prepared for the first time ribosylated
aspartic and glutamic acid using the trfluoroacetimidate
donors. Our method can conceivably be extended to the
ribosylation of other amino acids.
(16) The anomeric ratio of the glycosylation products was deter-
mined by H NMR, and the individual anomers could be assigned by
1
HECADE-NMR spectroscopy: Napolitano, J. G.; Gavın, J. A.; Garcıa,
ꢀ
ꢀ
C.; Norte, M.; Fernandez, J. J.; Hernandez Daranas, A. Chem.;Eur. J.
2011, 17, 6338.
(17) Although lower temperatures (∼ꢀ50 °C) resulted in highly
R-selective ribosylations (>98% R), yields were fairly low (<20%).
Silylation of glutamine, before ribosylation, with N,O-bis(trimethylsilyl)-
trifluoroacetamide (BSTFA) was employed in order to increase the
solubility, but this did not improve the solubility or yield. Replacing the
benzyl ethers with 2-naphthylmethyl or cyclic carbonate at the 20- and
30-positions improved neither the yield nor selectivity.
(18) Mukhopadhyay, B.; Maurer, S. V.; Rudolph, N.; van Well,
R. M.; Russell, D. A.; Field, R. A. J. Org. Chem. 2005, 70, 9059.
(19) Ludek, O. R.; Gu, W. L.; Gildersleeve, J. C. Carbohydr. Res.
2010, 345, 2074.
(20) Immobilized perchloric acid on silica was prepared by adding
HClO4 (2 mmol, as a 70% aqueous solution) to a slurry of silica gel
(5 g, 200 mesh) in Et2O (15 mL) and was stirred for 1 h at room
temperature. The solvent was removed under reduced pressure, and the
resulting powder was dried at 110þ°C for 2 h under reduced pressure to
obtain HClO4ꢀSiO2 (0.4 mmol H /g) and was used directly: Agarwal,
A.; Rani, S.; Vankar, Y. D. J. Org. Chem. 2004, 69, 6137.
(21) The use of HFIP as the solvent and a catalytic amount of HCl in
the acidolytic cleavage of the PMB ethers proved to be essential for
preserving the configuration of the anomeric center. For application of
HFIP as a solvent in acidolysis of protective groups, see: Palladino, P.;
Stetsenko, D. A. Org. Lett. 2012, 14, 6346.
Acknowledgment. The Netherlands Organization for
Scientific Research (NWO) is acknowledged for financial
support.
Supporting Information Available. Spectroscopic data
and experimental procedures. This material is available
free of charge via the Internet at http://pubs.acs.org.
(22) Suda, S.; Mukaiyama, T. Bull. Chem. Soc. Jpn. 1993, 66, 1211.
(23) Gravier-Pelletier, C.; Ginisty, M.; Le Merrer, Y. Tetrahedron:
Asymmetry 2004, 15, 189.
(24) The use of donor 16 in the ribosylation of glutamic- and aspartic
acid showed spontaneous and nonstereoselective product formation
without the addition of activator.
The authors declare no competing financial interest.
D
Org. Lett., Vol. XX, No. XX, XXXX