Synthesis of fluorophosphate nucleotide analogues and their characterization as tools for 19F NMR studies

Article abstract of DOI:10.1021/acs.joc.5b00337

To broaden the scope of existing methods based on ¹⁹F nucleotide labeling, we developed a new method for the synthesis of fluorophosphate (oligo)nucleotide analogues containing an O to F substitution at the terminal position of the (oligo)phosphate moiety and evaluated them as tools for ¹⁹F NMR studies. Using three efficient and comprehensive synthetic approaches based on phosphorimidazolide chemistry and tetra-n-butylammonium fluoride, fluoromonophosphate, or fluorophosphate imidazolide as fluorine sources, we prepared over 30 fluorophosphate-containing nucleotides, varying in nucleobase type (A, G, C, U, m⁷G), phosphate chain length (from mono to tetra), and presence of additional phosphate modifications (thio, borano, imido, methylene). Using fluorophosphate imidazolide as fluorophosphorylating reagent for 5'-phosphorylated oligos we also synthesized oligonucleotide 5'-(2-fluorodiphosphates), which are potentially useful as ¹⁹F NMR hybridization probes. The compounds were characterized by ¹⁹F NMR and evaluated as ¹⁹F NMR molecular probes. We found that fluorophosphate nucleotide analogues can be used to monitor activity of enzymes with various specificities and metal ion requirements, including human DcpS enzyme, a therapeutic target for spinal muscular atrophy. The compounds can also serve as reporter ligands for protein binding studies, as exemplified by studying interaction of fluorophosphate mRNA cap analogues with eukaryotic translation initiation factor (eIF4E).

Full text of DOI:10.1021/acs.joc.5b00337

Subscriber access provided by Umea University Library
Article
Synthesis of Fluorophosphate Nucleotide Analogs and
Their Characterization as Tools for 19F NMR Studies
Marek R. Baranowski, Anna Nowicka, Anna M. Rydzik, Marcin Warminski, Renata Kasprzyk,
Blazej A Wojtczak, Jacek Wojcik, Timothy D.W. Claridge, Joanna Kowalska, and Jacek Jemielity
J. Org. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.joc.5b00337 • Publication Date (Web): 27 Mar 2015
Downloaded from http://pubs.acs.org on April 6, 2015
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
The Journal of Organic Chemistry is published by the American Chemical Society.
1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 55
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
Synthesis of Fluorophosphate Nucleotide Analogs and Their
1
9
Characterization as Tools for F NMR Studies
1
1
2
1
Marek R. Baranowski, Anna Nowicka, Anna M. Rydzik, Marcin Warminski, Renata
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
3
4
2
Kasprzyk, Blazej A. Wojtczak, Jacek Wojcik, Timothy D. W. Claridge, Joanna
1,
3,
Kowalska, * and Jacek Jemielity *
1
Division of Biophysics, Institute of Experimental Physics, Faculty of Physics, University of
Warsaw, Zwirki i Wigury 93, 02ꢀ089 Warsaw, Poland
2
Chemistry Research Laboratory, University of Oxford, 12 Mansfield Road, OX1 3TA
Oxford, United Kingdom
3
Centre of New Technologies, University of Warsaw, Banacha 2c, 02ꢀ097 Warsaw, Poland
4
Laboratory of Biological NMR, Institute of Biochemistry and Biophysics, Polish Academy
of Sciences, Pawinskiego 5a, 02ꢀ106 Warsaw, Poland
1
9
ABSTRACT To broaden the scope of existing methods based on F nucleotide labeling, we
developed a new method for the synthesis of fluorophosphate (oligo)nucleotide analogs
containing an O to F substitution at the terminal position of the (oligo)phosphate moiety and
1
9
evaluated them as tools for F NMR studies. Using three efficient and comprehensive
synthetic approaches based on phosphorimidazolide chemistry and tetraꢀnꢀbutylammonium
1
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 2 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
fluoride, fluoromonophosphate or fluorophosphate imidazolide as fluorine sources, we
prepared over 30 fluorophosphateꢀcontaining nucleotides, varying in nucleobase type (A, G,
7
C, U, m G), phosphate chain length (from monoꢀ to tetraꢀ) and presence of additional
phosphate modifications (thioꢀ, boranoꢀ, imidoꢀ, methyleneꢀ). Using fluorophosphate
imidazolide as fluorophosphorylating reagent for 5′ꢀphosphorylated oligos we also
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
9
synthesized oligonucleotide 5′ꢀ(2ꢀfluorodiphosphates), which are potentially useful as
F
1
9
NMR hybridization probes. The compounds were characterized by F NMR, and evaluated
19
as F NMR molecular probes. We found that fluorophosphate nucleotide analogs can be used
to monitor activity of enzymes with various specificities and metalꢀion requirements,
including human DcpS enzyme – a therapeutic target for spinal muscular atrophy. The
compounds can also serve as reporter ligands for protein binding studies, as exemplified by
studying interaction of fluorophosphate mRNA cap analogs with eukaryotic translation
initiation factor (eIF4E).
INTRODUCTION
1
1
9
F nuclei have several beneficial properties for NMR applications, including a nuclear spin of
19
/2
, high magnetogyric coefficient, and 100% abundance, resulting in a F NMR sensitivity
1
19
equal to 83% of H NMR. However, unlike protons, F is almost never present in natural
biomolecules, has a higher chemical shift dispersion, and is more sensitive to changes in the
1
local environment. As such, fluorineꢀlabelled compounds and their transformations can be
1
9
2
19
selectively observed by F NMR even in complex mixtures. Consequently, Fꢀlabeled
nucleotides and nucleic acids have broad utility in biophysical and biological experiments that
19
utilize F NMR, including binding experiments, enzymatic activity assays, and structural and
3
functional studies. For example, 2ꢀfluoro ATP has been employed as a reporter substrate for
1
9
3b
F
NMRꢀbased screening of enzyme therapeutic targets.
A
macromolecular
trifluorothymidine prodrug has been recently designed as a potential anticancer theranostic
2
ACS Paragon Plus Environment

Page 3 of 55
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
1
9
4
19
agent with utility in F magnetic resonance imaging. For those various applications Fꢀ
labels are usually introduced within either the nucleobase or the ribose moiety. However, such
modifications sometimes disturb the physicochemical and biological properties of
(
oligo)nucleotides or require laborious synthetic procedures. Therefore, there is a need for
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
9
Fꢀlabeled (oligo)nucleotides with increased functionalities and/or synthetic availabilities.
We envisaged that the introduction of fluorine atom at the terminal position of an
(oligo)phosphate to form a fluorophosphate moiety by an efficient and straightforward
19
synthetic method would provide a viable alternative to existing Fꢀlabeling methods, which
usually require multistep syntheses of modified nucleosides and subsequent incorporation into
(oligo)nucleotides by chemical or chemoꢀenzymatic approaches.
The O to F substitution at the terminal phosphate produces a chemically stable analog that is
isostructural and isoelectronic to unmodified nucleotides, but has a smaller negative net
charge at physiological pH, a lower affinity to divalent metal cations, and properties similar to
5
those of corresponding phosphate methyl esters under hydrolytic conditions. As a result,
fluorophosphate analogs may be considered close mimics to either terminallyꢀprotonated
nucleoside 5′ꢀ(oligo)phosphates or nucleoside 5′ꢀphosphosulfates (Fig. 1) and may replace
6
nucleotides as substrates in some enzymatic reactions (Table S1). Several enzymes that
19
recognize fluorophosphate analogs are therapeutic targets and thus F NMRꢀbased assays that
facilitate the discovery of inhibitors for such targets are of general interest. To our knowledge,
1
9
fluorophosphate nucleotide analogs have not been used in F NMR assays to date, and one of
the reasons for that may be their limited synthetic availability.
3
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 4 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
n
n
n=0-3
protonated
(
oligo)phosphate
fluoro(oligo)phosphate
phosphosulfate
Figure 1. The general structures of fluorophosphate nucleotide analogs and structurally
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
related compounds found in nature. R is 5′ꢀnucleoside moiety.
Nucleoside 5′ꢀfluoromonophosphate derivatives (phosphorofluoridates
ꢀ
NMPF,
phosphorofluoridothioates and phosphorofluoridodithioates) are accessible by several, usually
efficient methods based either on phosphorus (V) or phosphorus (III) chemistry and
ꢀ
7
substitution by F or oxidative fluorination, respectively. Conversely, only few methods for
the synthesis of nucleoside fluoro diꢀ and triphosphates (NDPF and NTPF) have been
reported. The first synthesis of ATPF has been carried out by coupling of ADP and
fluorophosphate tributylammonium salts using Michelson method (diphenylchlorophosphate
5a
in pyridine) and isolated in about 35% yield after quite complex chemoenzymatic workꢀup.
GTPF has been synthesized directly from GTP by treatment with dinitrofluorobenzene, but
5d
isolated only in 5–10% yield, whereas ADPF has been synthesized in 25% yield by reacting
6a,8
adenosine 5′ꢀphosphomorpholidate and fluoromonophosphate in pyridine. Notably, all the
reported procedures gave moderate to low yields and have been demonstrated only on a single
4
target molecule. To our knowledge, nucleoside 5′ꢀfluorotetraphosphates (NP F) or nucleoside
oligophosphate analogs combining the fluorophosphate with other phosphateꢀmodifications
have not been reported so far.
In this work, our goal was to provide a comprehensive chemical method for the synthesis of
various fluorophosphateꢀcontaining nucleotide analogs – from monoꢀ to tetraphosphates – and
1
9
to evaluate them as tools for F NMR experiments in biological or medicinal context. We
assumed that our synthetic method should not only provide access to nucleotides differing in
nucleobase type and oligophosphate chain length, but also enable combining the presence of
4
ACS Paragon Plus Environment

Page 5 of 55
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
fluorophosphate group with other phosphate modifications. Due to resistance to enzymatic
degradation and differences in electronic structure, nucleotides containing phosphate
modifications such as imidodiphosphate, methylenebisphosphonate, phosphorothioate or
boranophosphate are valuable tools for structural and mechanistic studies on nucleotideꢀ
binding proteins and enzymes and often exhibit beneficial biological properties or inhibitory
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
9
effects. Combination of those modifications with fluorophosphate could be advantageous in
19
some F NMRꢀbased binding experiments and also provide new types of nucleotide mimics
for mechanistic studies.
To achieve our goal, we first developed three efficient synthetic methods for nucleotide
analogs containing the fluorophosphate moiety. The methods were based on
phosphorimidazolide chemistry and three different, easily available fluorine sources: tetraꢀnꢀ
butylammonium fluoride (TBAF), triethylammonium fluoromonophosphate (FP) or a new
reagent, fluoromonophosphate imidazolide lithium salt (FPIm). To demonstrate the utility of
our methods we synthesized a set of over 30 fluorophosphate nucleotide analogs of various
structural complexity, improving the efficiencies of the previously reported compounds and
preparing numerous previously unknown ones. We also, for the first time, obtained
oligonucleotides carrying a (2ꢀfluorodiphosphate) moiety at the 5′ꢀend. The synthesized
19
nucleotides were characterized by F NMR and their potential utility as reporter substrates in
1
9
F NMRꢀmonitored enzymatic experiments, reporter ligands for binding studies and
hybridization probes was also demonstrated.
RESULTS AND DISCUSSION
As a starting point for the syntheses we chose to explore the reactivity of imidazole activated
nucleotides (NMPIm and NDPIm) in aprotic polar solvents in the presence of divalent metal
chlorides (MCl
2
; usually MgCl
2
or ZnCl
2
) as Lewis acid reaction mediators. Such approach
5
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 6 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
has been previously used with success for pyrophosphate bond formation reactions leading to
various monoꢀ and dinucleoside oligophosphates and their phosphateꢀmodified
9c,10
derivatives.
After initial optimization studies we identified three complementary synthetic
approaches that could lead to fluorophosphate nucleotide analogs (Fig. S1). The first approach
employed a nucleophilic substitution of imidazoleꢀactivated nucleotides with a fluoride anion
(from TBAF) in the presence of excess divalent metal chloride (MgCl or ZnCl ) yielding a
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
2
fluorophosphate derivative of a nucleotide with the same phosphate chain length (Table 1).
The second approach, was similar, but utilized triethylammonium fluoromonophosphate as a
nucleophile yielding a fluorophosphateꢀcontaining product with the phosphate chain
elongated by one subunit (Table 2). Finally, in the third approach a reversed activation
scheme was applied, i.e. an imidazole activated fluorophosphate was coupled with nonꢀ
activated nucleotide to afford, similarly as in the second approach, a product with the
phosphate chain elongated by one subunit (Table 3).
The synthesis employing fluoride as a nucleophile was tested in the presence of 4–10ꢀfold
excess of TBAF, in either DMSO or DMF, and in the presence of excess ZnCl
2
, or of an
excess MgCl or in the absence of divalent metal chloride. The reaction was efficient only in
2
DMSO and with ZnCl
analysis of the UVꢀabsorbing peaks (Fig. 2), adenosine phosphorimidazolide (2a; t
treated with 6 equiv. of TBAF in DMSO at r.t. and in the presence of 8 equiv. of ZnCl
nearly quantitatively converted into AMPF (3a; t = 10.8 min, m/z = 348) within 2 h (the
reaction could be accelerated by addition of higher excess of TBAF). In the absence of
mediator or in the presence of MgCl no reaction was observed. Adenosine, guanosine,
2
as a mediator. As monitored by RP HPLC combined with MS(ꢀ)ES
= 12 min)
was
R
2
R
2
uridine, cytidine and 7ꢀmethylguanosine 5′ꢀfluoromonophosphates were synthesized under
these conditions and isolated after DEAE Sephadex ionꢀexchange purification in very good
6
ACS Paragon Plus Environment

Page 7 of 55
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
1
31
yields (58–88%) and high purity by HPLC, H and P NMR (Table 1, entries 1–5 and
Supporting Information).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2h
3h
3i
Table 1. Synthesis of fluorophosphate nucleotide
analogs from nucleotide imidazolide derivatives
and TBAF.
9
51 (64)
39 (57)
TBAF
ZnCl2
2
i
j
n
m
DMSO
1
0
1
2
2
3
Nucleotides unmodified in the phosphate chain
2
3j
Yield
N H2
N
a
b
Entry
1
Starting material
Product
% (%)
–
N
N
BH3
1
1
F
P
O
N
49 (86)
25 (65)
NH2
N
O
O
–
N
86 (91)
O
P
O H O H
F
O
N
O
N
2j
3k
O
OH O H
2a
3a
2
3
88 (95)
83 (99)
2k
3
l
a
b
Isolated yield; conversion calculated based on the reaction RPHPLC profile
2b
3b
NH2
–
N
O
P
F
O
N
O
O
O
O H OH
2c
3c
O
–
N H
O
P
4
5
78 (95)
58 (90)
67 (83)
F
O
N
O
O
O
O H O H
2
d
e
3d
2
3e
3f
6
7
2
f
42 (69)
Yield
2
g
3g
Nucleotides modified in the phosphate chain
a
b
Entry
8
Starting material Product
% (%)
82 (99)
7
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 8 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Imidazolide derivatives of nucleoside diphosphates, ADP (2f) and GDP (2g) were under
similar conditions converted into corresponding 5′ꢀfluorodiphosphates, ADPF (3f) and GDPF
(3g), respectively, within 2–4 h (Table 1, entries 6, 7). However, after longer reaction times
(>24 h) or in the presence of 10ꢀfold or higher TBAF excess, a fluorideꢀmediate cleavage of
the pyrophosphate was observed as a side reaction. For example, the initially formed ADPF
(3f) was partially cleaved to a mixture of AMPF (m/z 348) by HPLC and AMP (m/z 346) as
11
the only UVꢀabsorbing byꢀproducts.
NH2
N
NH2
N
Table 2. Synthesis of fluorophosphate nucleotide
analogs from nucleotide imidazolide derivatives
and FP (4).
–
N
N
–
–
N
N
O
P
O
P
O
P
17
18
65 (81)
N
O
N
F
O
O
N
N
O
O
P
O
O
O
P
O
O
O
O
O
O
–
–
O
O
MCl2
2l
3o
n
DMF
n
4
M=Zn or Mg
n.d.(65)
n.d.(66)
2
3
Nucleotides modified in the phosphate chain
Yaield
% (%)
2f
3p
b
Entry
Starting material
Product
19
1
1
3
4
73 (90)
3
q
2
g
a
b
Isolated yield; conversion calculated based on reaction RP HPLC profile
2
a
b
3f
n.d.(68)
54 (82)
2
3g
NH2
N
–
–
O
P
O
15
P
N
O
F
O
O
O
O
O
O H OH
2c
3m
1
6
76 (80)
2
d
3n
8
ACS Paragon Plus Environment

Page 9 of 55
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
The reaction with TBAF was also performed on activated phosphateꢀmodified NDP
derivatives, including imidazolides of nucleoside α,βꢀmethylenebisphosphonates (2h, 2i), αꢀ
boranodiphosphate (2j) and α,βꢀimidodiphosphate (2k), (Table 1, entries 8–12). For
compound 2h (t
fluoromethylenebisphosphonate (3h, t
chemical stability PꢀCH ꢀP bond no fluorideꢀmediated cleavage was observed even in the
presence of high TBAF concentrations. Adenosine αꢀboranodiphosphate imidazolide (2j),
which was a mixture of two (S and R ) diastereomers, was also efficiently converted into two
diastereomers of ADPαBH F (3j).However, if reaction with TBAF was prolonged or higher
excess of TBAF was used, 3j was efficiently cleaved by fluoride producing two diastereomers
of AMBH F (3k) as major products (Table 1, entries 10 and 11; Fig. S2). Imidodiphosphate
R
6.5 min, Fig. 2B) the conversion into corresponding
R
5.0 min) was nearly quantitative, and due to the high
2
P
P
3
3
derivative 2k was also converted into corresponding fluoroimidodiphosphate (3l), however,
the reaction was notably slower, taking over 24 h and reaching only 65% HPLC conversion
(Fig. 2C). The major UVꢀabsorbing byꢀproduct (t
61). The fluorideꢀbased approach appeared to be less practical for NTPF (and NP
the synthesis of the required starting materials (corresponding NTP or NP imidazolides) is
R
= 4.8 min) was identified as GPNH
2
(m/z =
3
4
F) since
4
often problematic as they are more polar and less stable than imidazolides of monoꢀ and
diphosphates.
9
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 10 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Therefore, we tested the second approach, in which triethylammonium fluoromonophosphate
(TEA FP, 4) was used as a nucleophile instead of fluoride. Reaction of NMPIm and NDPIm
with 4 resulted in the formation of a fluorophosphateꢀcontaining product with oligophosphate
chain one subunit longer compared to the starting imidazolide derivative (NDPF and NTPF,
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
respectively). The reactions were performed in DMF in the presence of excess MgCl
ZnCl
2
or
2
.
Nucleotides modified in the phosphate chain
Table 3. Synthesis of fluorophosphate nucleotide
analogs from nucleotides and FPIm (5).
Yield
a
b
Entry
Substrate
Product
% (%)
MCl2
n
n
DMF
2
2
3
8
9
0
72 (82)
63 (72)
74 (79)
M=Zn or Mg
5
1
3
Nucleotides unmodified in the phosphate chain
1i
3v
Yaield
b
Entry
20
Starting material
Product
% (%)
n.d.(87)
69 (75)
1
j
3w
1
a
b
3f
2
1
2
1k
1l
3x
3y
1
3g
N H2
N
31
32
63 (75)
65 (74)
–
O
P
–
N
O
O
O
2
O
O
64 (74)
60 (87)
O H
1
c
3r
23
1
m
1n
1o
1p
3z
1d
3p
3q
3s
33
68 (90)
24
56 (78)
28 (64)
3aa
3ab
1
e
34
35
70 (80)
60 (74)
25
1
f
–
O
+
–
N
N
N
O
–
P
3ac
O
O
N
NH2
26
O
O
57 (75)
58 (70)
a
b
Isolated yield; conversion calculated based on reaction RP HPLC profile
O H O H
1
g
3t
–
O
N
+
–
–
N
N
N
O
P
O
P
27
–
O
O
O
NH2
O
O
O
OH OH
1h
3u
10
ACS Paragon Plus Environment

Page 11 of 55
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
A pilot coupling reaction of 2a (AMPIm; Fig. 2D) with 1.5 equiv. of 4 in the presence of 8ꢀ
fold excess of MgCl
isolated yield of 73%), indicating this approach may represent a more straightforward
alternative for NDPF. The reaction was similarly efficient in the presence of ZnCl (Figure
2
resulted in 90% HPLC conversion into ADPF (3f) within 40 min (and
2
S2), indicating that for compound 4 as a nucleophile both metal chlorides are efficient
reaction mediators. The utility of this approach was further confirmed by the synthesis of 3g
(GDPF), 3n (UDPF), 3m (CDPF) and 3o (PADPF) in good yields (68–82) by HPLC, 40–76%
isolated, Table 2, entries 14–17; Fig. 2E). The approach was also tested for the synthesis of
NTPFs from corresponding NDPIms. In the coupling reactions of ADPIm (2f) and GDPIm
(2g) with 4, the maximum HPLC conversions reached 65 and 66 %, respectively. Hence, the
synthesis was feasible, however, the yield was slightly lower as a result of imidazolide
hydrolysis and selfꢀcoupling side reactions, which were more abundant in the case of NDPIms
because of their lower stability compared to NMPIms. Hence, we developed the third
approach avoiding the problems with synthetically challenging or chemically labile
imidazolides. This includes not only some of NDP and NTP (e.g. 7ꢀmethylguanosine
derivatives), but also nucleotides containing the phosphorothioate group, which decompose
11
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 12 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
with
the
loss
of
sulfur
under
standard
activation
protocols
(i.e.
dithiodipyridine/triphenylphosphine system used in this work or activation with CDI). The
approach was based on the use of nonꢀactivated nucleotides as nucleophiles and, thus,
required the use of an electrophilic fluorophosphorylating agent. As such, we synthesized
fluorophosphate phosphorimidazolide lithium salt (FPIm, 5), in a single step from
triethylammonium fluorophosphate (4), without chromatographic purification and in high
yield (75–80%). The compound was stable when stored at 4 °C in the absence of moisture for
at least one month. The compound acted as a powerful fluorophosphorylating reagent for
various (oligo)phosphates. Several nucleoside fluorodiphosphates (3f, 3g, 3r, 3t),
fluorotriphosphates (3p, 3q, 3u) and even tetraphosphates (3s) were synthesized from
corresponding nonꢀactivated nucleotides (1) using reagent 5. The analogs additionally
modified with methylenebisphosphonate (3v, 3w) phosphorothiate (3x, 3ac), boranophosphate
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
(3y, 3z, 3aa) and imidodiphoshate (3ab) groups were also synthesized. The reagent gave
usually very good HPLC conversion within 0.5 to 2 h with nearly no other UVꢀabsorbing byꢀ
products (Fig. 2F,G). However, prolonged reaction time resulted in formation of more polar
byꢀproduct of m/z value 82 units higher than the expected product, suggesting introduction of
an additional fluorophosphate group. MS/MS analysis indicated that the second
fluorophosphate group is present at the 2′ or 3′ꢀhydroxyl groups (not shown). Finally, we also
found that 5 efficiently reacts with 5′ꢀphosphorylated oligonucleotides in DMSO in the
presence of MgCl
2
excess to produce oligonucleotide 5′ꢀfluorodiphosphates in 80–90%
conversions. For example, 10 nt 5′ꢀphosphorylated oligo ON (t = 15.7 min) treated with
1
R
excess FPIm (5) and MgCl in DMSO was converted within 24 h and in 83% yield into
2
1
fluorophosphorylated FꢀON (6a) (Table 4, Figure 2H), which then was purified by RP HPLC.
Table 4. Synthesis of oligonucleotide 5′ꢀ
fluorodiphosphates from oligonucleotide 5′ꢀ
phosphates and FPIm
5
'
5'
MgCl2
DMSO
(1)
(5)
(6)
Yield
12
ACS Paragon Plus Environment

Page 13 of 55
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
a
ba
Entry
Substrate
Product
5'
% (%)
3
5
5'
20.4 (83)
21.2 (86)
ON
1q
1
F-ON
1
6a
3
6
5'
5'
ON
2
F-ON
2
1r
6b
a
b
Isolated yield; conversion calculated based on reaction RP HPLC profile
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Altogether, using the three approaches, we synthesized a set of over 30 fluorophosphate
nucleotide analogs (mononucleotides 3 and oligonucleotides 6; Figure 1, Table S2), varying
in oligophosphate chain length (from monoꢀ to tetraphosphate), nucleobase type (A, G, C, U,
7
m G), and phosphate modifications which may be useful for conferring enzymatic resistance
or tuning their affinity towards proteins. Although neither of the three synthetic approaches
was fully universal, they were complementary, enabling us to cover a broad spectrum of
nucleotides varying in size, polarity (and thus solubility in organic solvents), chemical
reactivity and stability. The compound structures and homogeneities were confirmed by
1
31
HRMS, H NMR, P NMR and RP HPLC. In contrast to O,Oꢀdialkyl phosphorofluoridates,
5b
which are rapidly hydrolyzed to phosphate diesters under aqueous conditions, the
fluorophosphate analogs of mononucleotides showed high stability, both in aqueous solutions
12
and upon storage as solid samples. NMPF (except 3k), NDPF and NTPF were stable in
water and in aqueous buffers of pH 5–8 at RT for several weeks or even months. Moreover,
during storage NDPF and NTPF appeared to be less susceptible to pyrophosphate bond
hydrolysis with the loss of terminal phosphate, which is commonly observed for unmodified
NDP and NTP.
13
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 14 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 2. Representative RP HPLC profiles from the syntheses of fluorophosphate nucleotide
analogs. Synthesis of A) AMPF (3a) from AMPIm (2a) and TBAF; B) ApCH
2
pF (3h) from
ApCH pIm (2h) and TBAF C) GpNHpF (3l) from GpNHpIm (2k) and TBAF D) ADPF (3f)
2
from AMPIm (2a) and FP (4) E) CDPF from AMPIm (2a) and FP (4) F) GTPF from GDP
(1e) and FPIm (5) G) Gp pF (two Pꢀdiastereomers) from GMPS (1k) and FPIm (5) H) 5′ꢀ
S
1
fluorodiphosphorylated oligo FꢀON (6a) from 5′ꢀphosphorylated oligo (1q) and FPIm (5).
Asterisks indicate always t
R
of the starting material.
14
ACS Paragon Plus Environment

Page 15 of 55
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
31
19
Figure 3. P and F NMR spectra of compound 3g (GDPF).
19
Figure 4. Duplex formation of 5′ꢀfluorodiphosphate oligos can be monitored by F NMR. a)
Hybridization of 240 µM 5′ꢀfluorodiphosphorylated 10 nt oligo (FꢀON , 6a) with increasing
concentrations of unmodified complementary ON (6b; 0 240 µM); b) Hybridization of 120
µM FꢀON with complementary 120 µM FꢀON and subsequent replacement of FꢀON
1
2
ꢀ
1
2
2
by
15
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 16 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
increasing concentrations of ON (0ꢀ540 µM). In each case, only the downfield component
2
of the doublet is shown for clarity.
1
9
The next step in our study was characterization of the compounds by F NMR (Table 5,
Figure S3). The δ values were strongly dependent on the length of the phosphate chain and
phosphateꢀmodifications, whereas structural alterations in the ribose or nucleobase moiety
influenced the δ values to a smaller extent (Figures S3 and S4). As for nucleobases, the
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
F
F
greatest differences in the δF values were observed between purine and pyrimidine
nucleosides, whereas in other cases the differences were very small (<0.01 ppm) or
undetectable (the differences decreased with increasing oligophosphate chain length). Due to
19
scalar coupling with the adjacent phosphorus nucleus, the F NMR signal of each nucleotide
1
appeared as a doublet ranging from ꢀ53 to ꢀ79 ppm with JPꢀF ranging from 914 to 1144 Hz
1
(
Table 5). The J values for various nucleoside fluoro(oligo)phosphates were virtually the
PꢀF
1
same ( JPꢀF ~935 Hz, 3aꢀ3g, 3mꢀ3u). However, if one of the oxygen atoms in the
ꢀ
ꢀ
fluorophosphate moiety was replaced by another species (S , BH
3 2
, NH, or CH ; 3hꢀ3l, 3vꢀ3z,
3
acꢀ3ad) the value changed up to ±130 Hz. The δ values for fluorophosphate nucleotides
F
were almost independent of the buffer, pH, or presence of magnesium ions and slightly
dependent from temperature as exemplified on various nucleotide mixtures (Fig. S5), making
19
1
them good candidates for F NMR reporter molecules or binding probes. The δ
F
and JPꢀF
values of oligo 5′ꢀ(2ꢀfluorodiphosphates) were similar to those of the corresponding
mononucleotides; however, the δ values varied depending on whether the oligonucleotide
F
was in the single or double stranded form. Notably, the 0.02 ppm difference in δ was
F
sufficient to monitor the hybridization of a 5′ꢀfluorodiphosphorylated 10 nt oligonucleotide
1 2
(FꢀON , 6a) with a complementary unlabeled oligo (ON , 1q) (Figure 4). In the case of
duplex formation between two complementary 5′ꢀfluorodiphosphorylated oligonucleotides,
16
ACS Paragon Plus Environment

Page 17 of 55
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
1
9
namely FꢀON (6a) and FꢀON (6b), two separate F resonances were observed for each
1
2
strand. The displacement of FꢀON by ON could also be monitored.
2
2
19
a
Table 5. Representative F NMR data for fluorophosphate nucleotides.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
a
1
a
Entry Number Abbreviation
δ
F
(ppm)
ꢀ78.98
ꢀ52.74/ꢀ52.95 1140
JPꢀF (Hz)
1
2
3
4
5
6
7
8
9
3b
3k
3f
GMPF
AMPBH
ADPF
933
b
b
b
3
F
ꢀ72.50
934
934
914
973
935
935
936
6a
3l
Fppꢀd(GTCAATGTCA) ꢀ72.59
GpNHpF
ꢀ61.54
ꢀ53.96
ꢀ72.26
ꢀ72.23
ꢀ72.00
3h
3p
3ac
3s
ApCH pF
2
ATPF
b
b
b
Ap
Ap
S
ppF
4
F
a 19
F NMR spectra recorded at 376 MHz in D O at 25 °C; chemical shifts reported relative to
2
b
aqueous NaF, as the secondary external standard (ꢀ121.5 ppm). Mixture of R
P
and S
P
diastereomers.
The next goal in our study was to preliminarily verify whether the fluorophosphate nucleotide
19
analogs can be of utility in F NMR monitored enzymaticꢀactivity assays and binding
experiments. For this purpose, we took advantage of commercially available broad specificity
enzymes – snake venom phosphodiesterase (SVPDE) or ribonuclease T2 (RNAse T2), as well
as of highly specific enzyme that constitutes a real therapeutic target – mRNA decapping
7
scavenger (DcpS), which is an m Gꢀnucleotide specific pyrophosphatase that has been
13
identified as a target in spinal muscular atrophy (SMA). Moreover, using eukaryotic
7
translation initiation factor 4E (eIF4E), a nonꢀenzymatic protein recognizing m Gꢀcapped
mRNAs during initiation of translation, we demonstrated possible use of fluorophosphate
nucleotide analogs as reporter molecules in binding assays. Finally, we also quantitatively
assessed how the terminal O to F substitution influences the proteinꢀnucleotide interactions.
17
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 18 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
First, we subjected four fluorophosphate nucleotides, (AMPF, ADPF, ATPF and AP F;
4
1
9
compounds 3a, 3f, 3p, and 3s, respectively), to HPLC and F NMRꢀmonitored degradation
by snake venom phosphodiesterase (SVPDE), which cleaves NTPs, NDPs, and oligos to
release nucleoside 5′ꢀmonophosphates. All tested nucleotides were viable substrates for
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
F
SVPDE. Furthermore, there was a sufficient difference in the chemical shift (ꢀδ > 0.1 ppm)
of the fluorophosphorylated substrate (ADPF, ATPF and AP F) and product (i.e.
4
fluoromonophosphate, fluorodiphosphate and fluorotriphosphate, respectively) to adequately
monitor the reaction progress (Figure 5A, Figure S6). However, the signal of fluoride released
from AMPF was not detected, probably as a result of significant line broadening caused by
2
+
2+
the presence of Mg ions, which are required by SVPDE. The presence of Mg caused also
slight line broadening and δ changes of the fluorophosphate signals, especially in the
F
products, which disappeared upon quenching the reaction with EDTA. Next, we subjected a
mixture of all four adenine nucleotides to incomplete hydrolysis by SVPDE, yielding a
sample containing four fluorinated substrates and four products. Each fluorophosphateꢀ
F
containing species was characterized by a distinctive δ value, meaning that each substrateꢀ
1
9
product combination could be monitored by F NMR (Figure S7, Table S2) and indicating
the potential applicability of the analogs in different adenine nucleotideꢀprocessing enzymes.
18
ACS Paragon Plus Environment

Page 19 of 55
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
9
Figure 5. Fluorophosphate nucleotide analogs can be used as reporter substrates in F NMR
monitored activityꢀbased enzymatic assays. The time course of several reactions was
19
determined by F NMR including a) cleavage of 1 mM ATPF (3p) by SVPDE, b) cleavage of
7
7
0
.5 mM m GDPF (3t) by DcpS, c) cleavage of 0.5 mM m GMPF (3e) by DcpS, and d)
cleavage of 2 mM cPAPPF (3o) by RNAse T2.
7
7
7ꢀMethylguanosine (m G) derivatives 3e, 3t, 3u may be considered analogs of the m G cap
structure present on the 5′ end of mRNA. Therefore, we tested their viability as substrates for
human mRNA decapping scavenger enzyme (DcpS). DcpS is a member of the HIT family of
pyrophosphatases and hydrolyzes mRNA cap derivatives formed after mRNA degradation by
14
exosomes. It has also been identified as a molecular target in spinal muscular atrophy
19
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 20 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
3,15
7
(SMA).
The canonical substrates of DcpS are diꢀ or oligonucleotides with an m GpppN
n
7
structure, which are cleaved between the γ and β phosphates to produce m GMP and
7
downstream (oligo)nucleotides. Among unmodified mononucleotides, m GDP is not a
7
substrate but is a strong inhibitor of DcpS, whereas m GTP is cleaved by DcpS, but less
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
6
7
7
7
efficiently than dinucleotides. Therefore, m GMPF (3e), m GDPF (3t), and m GTPF (3u)
are analogs of the reaction product, inhibitor, and substrate, respectively. Surprisingly, as
1
9
confirmed by HPLC and F NMR, all of these compounds were cleaved by DcpS, releasing
7
m GMP and fluoride, fluoromonophosphate, or fluorodiphosphate, respectively. In each case,
F
the δ values of the reaction substrates and products differed sufficiently to monitor the
reaction progress (Figure 5B,C; Figure S8). As such, these novel DcpS substrates offer the
possibility to develop novel DcpS assays that are an alternative to previously described
13
radioactive probes.
F
To test if δ changes of the fluorophosphate moiety could also be used to monitor enzymatic
processes in which more distant parts of the molecule are modified, we subjected 2′,3′ꢀcycloꢀ
phosphoadenosine 5′ꢀ(2ꢀfluorodiphosphate) (3o, cPADPF) to cleavage by RNAse T2,
producing 3′ꢀphosphoadenosine 5′ꢀ(2ꢀfluorodiphosphate) (PADPF). PADPF is an
6
a
isoelectronic and isostructural analog of phosphoadenosine 5′ꢀphosphosulfate (PAPS),
17
which serves as a source of ′active sulfate′ in cells, as ATP serves as ′active phosphate′. The
1
9
enzymatic reaction was easily monitored by F NMR due to the appearance of a 0.08 ppm
downfield shifted signal upon cleavage of the 2′ꢀphosphoester bond of 3o by RNAse T2
(
Figure 5D).
Apart from activityꢀbased assays, monitoring ligand binding is also of interest, especially in
inhibitor discovery. This is particularly important for proteins lacking enzymatic activity or
when targeting a particular binding site. As such, we investigated the use of fluorophosphateꢀ
19
containing nucleotides as reporter molecules in F NMRꢀbased binding assays. As a model
20
ACS Paragon Plus Environment

Page 21 of 55
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
system, we used eukaryotic translation initiation factor 4E (eIF4E), a protein responsible for
the recognition of the mRNA 5′ cap during the recruitment of ribosomes onto mRNA. eIF4E
18
7
is also a marker protein and therapeutic target in cancer. It has been shown that m Gꢀ
containing monoꢀ and dinucleotides can bind efficiently to eIF4E and that modifications
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
9
c
within the phosphate moiety can affect binding affinity. Therefore, we investigated how the
fluorophosphate moiety influences recognition by eIF4E. First, we determined the affinities of
fluorophosphate cap analogs 3e, 3t, 3u for eIF4E using fluorescence quenching titration
19
(
Figure S10). For comparison, we also determined the K
D
value for the complex between
7
DcpS and an unhydrolyzable analog of 3t, m GpNHpF (3ad), which was synthesized by
3
treating GpNHpF (3l) with CH I in DMSO. Replacing the terminal oxygen atom with fluorine
resulted in a 10 to 7ꢀfold decrease in affinity towards both proteins as compared to the
unmodified parent compound (Table 6), corresponding to a ꢁG° loss of 1.2–1.3 kcal/mol. The
7
7
affinity of m GDPF (3t) for eIF4E was 8ꢀfold lower than the affinity of m GDP, and only 2–
7
7
2
.5ꢀfold higher than that of m GMP and 7ꢀmethylguanosine phosphosulfate (m GPS), which
7
have the same net charge as m GDPF at pH 7.4. Similarly, the affinity of 30 for hDcpS was
7
8
.5ꢀfold lower than that of m GpNHp.
Table 6. Binding affinities of fluorophosphate cap analogs and their unmodified or
phosphosulfate counterparts for capꢀbinding proteins, eIF4E and DcpS.
D
K for eIF4E / nM
Ligand
ꢀ
20
ꢀ 21
X=F
(3e) (6.67± 0.98) 10
X=O
667 ± 44
X=OSO
3
7
.
3
m GMPX
575 ± 145
21
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 22 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
7
m GDPX
(3t)
267.4 ± 5.0
52.6 ± 3.3
34.6 ± 1.8
6.9 ± 0.4
n.d.
n.d.
7
m GTPX
(3u)
K
D
for DcpS / nM
9.6 ± 0.4
7
m GpNHpX
(3ad)
81.8 ± 4.3
n.d.
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
o
22
This effect can be attributed to the loss of a single Hꢀbond (∆∆G = ꢀ1.3 ± 0.6 kcal/mol ) and,
thus, well corresponds with the literature suggestions that fluorophosphate closely mimics
7
either a protonated phosphate or phosphosulfate. Next, we used m GTPF (3u) as a reporter
19
molecule for F NMRꢀmonitored binding experiments with eIF4E. Typically, a significant
19
broadening of the F signals of the reporter molecule is observed upon protein binding due to
increased transverse relaxation rates. Additionally, the CarrꢀPurcellꢀMeiboomꢀGill (CPMG)
sequence is used to attenuate broad (protein bound) resonances of reporter molecules.
Subsequently, the proteinꢀbound reporter ligand (in our case 3u) is displaced with a nonꢀ
fluorinated ligand of interest, leading to the recovery of the signal belonging to the unbound
7
7
reporter. In our case, displacement was carried out using m GpppG or m GppCH ppG, two
2
23
mRNA 5′ cap analogs with differing protein affinities. The apparent dissociation constants
app, defined as the concentration of free ligand when 50% of the protein is bound to the
(
K
19
ligand in the presence of reporter ligand 3u) obtained by F NMRꢀmonitored titration of
7
7
eIF4E in the presence of m GTPF and increasing concentrations of either m GpppG or
7
m GppCH ppG (Figure S11) are shown in Table 7. The values were in qualitative agreement
2
7
with the fluorescence titration data, as m GppCH ppG proved to be about a 4ꢀfold tighter
2
7
binder than m GpppG.
19
Table 7. Binding affinities of nonꢀfluorinated nucleotides for eIF4E determined by F NMR
7
19
ligand displacement with m GTPF (3u) as a Fꢀlabeled reporter ligand.
K
D
/ µM
7
7
m GpppG
m GppCH ppG
2
22
ACS Paragon Plus Environment

Page 23 of 55
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
Apparent 7 ± 1
30 ± 4
1
9
F NMR
Calculated 0.012 ± 0.002
0.054 ± 0.008
23
Fluorescence
0.0206 ± 0.0009 0.0800 ± 0.0019
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Since the affinity of 3u for eIF4E was known from the fluorescence titration, the actual K
D
values (defined as the concentration of free ligand when 50% of the protein is bound to the
7
7
2
ligand in the absence of reporter) for m GpppG and m GppCH ppG ligands could also be
calculated and showed good agreement with the literature data (Table 7).
The presented examples support our initial assumption that fluorophosphate nucleotide
19
analogs are potentially useful as tools in various biologicallyꢀoriented F NMR studies. The
sensitivity of the fluorophosphate moiety to the alterations of local environment is sufficient
to observe various enzymatic and nonꢀenzymatic transformations due to chemical shift or
relaxation rate changes. We expect fluorophosphate nucleotide analogs to be particularly
useful in studying enzymes processing and proteins binding the oligophosphate chain as well
as for receptors that are particularly sensitive to the changes within the nucleobase or ribose
moieties. Although the introduction of only a single F atom implicates a rather moderate
19
sensitivity, we were able to collect the F NMR spectra at concentrations as low as 80 µM
with standard NMR probe and only few µM using cryoprobe (the eIF4E binding experiment).
Even though such labeling method would not always be competitive to more sensitive labels
3 3 3
such as CF or C(CF ) , at least in some cases, the lower sensitivity can be compensated for
by higher synthetic availability.
CONCLUSION
In summary, a novel, more general and high yielding synthetic methodologies for the
preparation of fluorophosphateꢀcontaining nucleotides were developed. The first and second
23
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 24 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
approach were based on divalentꢀmetal cationꢀmediated coupling of imidazole activated
nucleotides with either tetraꢀnꢀbutylammonium fluoride or triethylammonium
fluorophosphate, respectively. The first approach gave access to nucleoside 5′ꢀfluoromonoꢀ,
5′ꢀfluorodiphosphates and some of their derivatives containing additional phosphateꢀchain
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
modifications (O to CH , O to NH and O to BH substitutions). The second approach was
2
3
exemplified by the synthesis phosphate unmodified nucleoside 5′ꢀfluorodiphosphates and 5′ꢀ
fluorotriphosphates. The third approach encompassed reactions of nonꢀactivated nucleotides
with a novel electrophilic reagent, fluorophosphate imidazolide, and was particularly useful
for the syntheses starting from nucleotides that were difficult to activate with imidazole, either
because of high polarity or chemical lability (e.g., triphosphates, 5′ꢀphosphorylated oligos,
and phosphateꢀmodified). The three approaches gave altogether access to a collection of
fluorophosphate nucleotide analogs including nucleotides varying in oligophosphate chain
7
length (from monoꢀ to tetraphosphate), nucleobase type (A, G, C, U, m G), and presence of
additional phosphate modifications. The majority of compounds were reported for the first
time (including 5′ꢀfluorodiphosphorylated oligonucleotides), and some may be considered
representatives of novel classes of nucleotide mimics, such as nucleoside 5′ꢀ
fluoroimidodiphosphates (3l, 3ad), 5′ꢀfluoromethylenebisphosphonates (3h, 3i) or 5′ꢀ
fluoroboranophosphates (3j) and thus, are interesting objects for further studies. Notably, the
1
9
fluorophosphate nucleotide analogs served as versatile reporter molecules in F NMR
enzymatic activity and binding assays, revealing their potential utility in inhibitor discovery
19
and ligandꢀbinding studies. The introduction of the F label into the oligophosphate is
advantageous because virtually the same synthetic protocols can be used for structurally
varying nucleotides, whereas for ribose or base fluorinated nucleotides, a change of structure
implicates often development of completely new synthetic route. The possibility of using 5′ꢀ
fluorophosphorylated oligonucleotides for hybridization/dehybridization events is also
24
ACS Paragon Plus Environment

Page 25 of 55
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
potentially interesting, however, more detailed studies into scope and limitations of such
application are required. We hope that the improved, efficient synthetic methods and
19
application highlights presented here may encourage broader use of such compounds in F
NMR studies on nucleotideꢀprocessing enzymes and proteins, which may contribute to the
development of efficient screening platforms. Further studies on the scope and limitations of
applications of nucleotides and oligonucleotides labelled with fluorophosphate moieties are in
progress.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
EXPERIMENTAL SECTION
I.
Chemical synthesis
1
1
. General information
.1. Starting materials and chemical reagents
Solvents, chemical reagents and starting materials, including nucleotides: 5′ꢀAMP (free acid
form), 3′ꢀAMP (free acid form), 5′ꢀGMP (disodium salt), 5′ꢀCMP (disodium salt), 5′ꢀdCMP
(disodium salt), 5′ꢀUMP (disodium salt) and ATP (trisodium salt), were from commercial
sources. Other nucleotide starting materials (1 and 2) were synthesized as described
10c
previously:
guanosine
5′ꢀmonoboranophosphate
(GMPBH
3
,
adenosine
5′ꢀ
10c
10c
monoboranophosphate, AMPBH3, adenosine 5′ꢀ(1ꢀthiodiphosphate), ADPαS, adenosine
1
0c
10c
5′ꢀ(1ꢀboranodiphosphate), ADPαBH
3
,
3
guanosine 5′ꢀ(1ꢀboranodiphosphate), GDPαBH ,
24
guanosine 5′ꢀimidodihosphate, GpNHp,
adenosine 5′ꢀ(methylenebisphosphonate),
2
5
25
ApCH
2
p,
guanosine 5′ꢀ(methylenebisphosphonate), GpCH
2
p
and 2′,3′ꢀcyclicꢀ
and pꢀON
21
phosphoadenosine imidazolide, cPAPIm. 5′ꢀPhosphorylated nucleotides pꢀON
1
2
(5′ꢀPꢀd(GTCAATGTCA) and 5′ꢀPꢀd(TGACATTGAC), respectively, in a desalted form were
1
from commercial source. Nonꢀphosphorylated nucleotide ON (d(TGACATTGAC)) was
synthesized using standard solid phase protocol and purified by RP HPLC.
25
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 26 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1.2. Nucleotide purification by ionꢀexchange chromatography using DEAE Sephadex Aꢀ
25 and conversion into sodium salts
The synthesized nucleotides (3) were purified by ionꢀexchange chromatography on DEAE
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
ꢀ
Sephadex Aꢀ25 (HCO
3
form) column. A column was loaded with reaction mixture and
solution) to
washed thoroughly with water (until the eluate did not precipitate with AgNO
3
elute solvents and reagents that do not bind to the column. Then, nucleotides were eluted
using 0–0.7 M, 0–1.0 M, 0–1.2 M or 0–1.4 M gradient of TEAB in deionized water for
nucleoside monoꢀ, diꢀ and triꢀ and tetraphosphates, respectively. Collected fractions were
analyzed spectrophotometrically at 260 nm and fractions containing the desired nucleotide
were analyzed by RP HPLC and poured together. The yields were calculated on the basis of
.
optical density miliunit measurements (mOD = absorbance of the solution volume in mL) of
isolated products and corresponding staring materials (nucleotides or nucleotide Pꢀ
imidazolide derivatives). Optical unit measurements were performed in 0.1M phosphate
buffer pH 7.0 at 260 nm for all nucleotides except 7ꢀmethylguanosine nucleotides, and in
7
0
.1M phosphate buffer pH 6.0 at 260 nm for m G nucleotides. After evaporation under
reduced pressure with repeated additions of 96% and then 99.8 % ethanol (to decompose
TEAB and remove residual water, respectively), nucleotides were isolated as
triethylammonium (TEA) salts. These were converted either into sodium salts by precipitation
4
with acetone solution of NaClO or into ammonium salts (especially if additional purification
or diastereomer separation was required) by purification on semiꢀpreparative RP HPLC (as
described in paragraph 1.3). In the latter case, the products were isolated after repeated freezeꢀ
drying of the collected HPLC fractions.
1
.3. Analytical and preparative reversedꢀphase (RP) HPLC
26
ACS Paragon Plus Environment

Page 27 of 55
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
Analytical HPLC was performed using Supelcosil LCꢀ18ꢀT HPLC column (4.6 x 250 mm,
flow rate 1.3 mL/min) with a linear gradient 0–25 % of methanol in 0.05 M ammonium
acetate buffer (pH 5.9) in 15 min. Semiꢀpreparative HPLC was performed on Discovery RP
Amide Cꢀ16 HPLC column (25 cm x 21.2 mm, 5ꢂm, flow rate 5.0 mL/min) with linear
gradients of acetonitrile in 0.05 M ammonium acetate buffer (pH 5.9). In both cases UVꢀ
detection at 254 nm was used.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1.4. Spectroscopic analysis of the synthesized compounds
The structure and homogeneity of each final product was confirmed by reꢀchromatography on
RP HPLC, high resolution mass spectrometry using negative electrospray ionization (HRMS(ꢀ
1
31
19
)
ES), and H NMR, P NMR and F NMR spectroscopy. Intermediate compounds were
characterized by RP HPLC and low resolution MS(ꢀ)ES. High resolution mass spectra were
recorded on LTQ Orbitrap. NMR spectra were recorded on a 400 MHz or 500 MHz
spectrometer equipped with a high stability temperature unit using 5mm 4NUC probe, at
1
19
31
3
99.94/500.61 MHz ( H NMR), 376.28/471.00 MHz ( F NMR) and 161.90/202.65 MHz ( P
o
1
NMR) and at 25 C if not stated otherwise. The H NMR chemical shifts were reported to
sodium 3ꢀ(trimethylsilyl)ꢀ2,2′,3,3′ꢀtetradeuteropropionate (TSP) (δ = 0 ppm) as an internal
3
1
standard. The P NMR chemical shifts were referenced to 20 % phosphorus acid in D O (δ 0
2
P
19
ppm) as an external standard. The F NMR chemical shifts were reported to external 10 mM
NaF in D O (δ ꢀ121.5 ppm). Typical parameters for proton spectra: pulse width 6.4 ꢂs,
2
F
acquisition time 2.5 s, equilibration delay 0.5 s, sweep width 4k, 20k points, resolution 0.25
Hz, number of transients ꢀ128. Typical parameters for phosphorus spectra: pulse width 15 ꢂs,
acquisition time 2.5 s, equilibration delay 0.5 s, sweep width 32k, 160k points, resolution 0.25
Hz, number of transients ꢀ512. Typical parameters for fluorine spectra: pulse width 27 ꢂs,
acquisition time 0.6 s, equilibration delay 4.0 s, sweep width 64k, 76k points, resolution 1.0
1
9
Hz, number of transients 4–128. F NMR binding assays were recorded on a 600 MHz
27
ACS Paragon Plus Environment

The Journal of Organic Chemistry
Page 28 of 55
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
1
spectrometer with BBꢀF/ H Prodigy N cryoprobe operating at 298 K using 5 mm diameter
2
NMR tubes.
2
. Preparation of fluorophosphate triethylammonium salt (4):
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
The commercially available fluorophosphate sodium salt was converted before the synthesis
into DMFꢀsoluble triethylammonium salt. The sodium salt (5 g) dissolved in water (ca. 75
mL) was passed through Dowex 50 W x 8 cationite in the triethylammonium form. The
collected eluate was evaporated under reduced pressure to dryness and then reꢀevaporated few
times with ethanol, the residue was dried in vacuum over P
2 5
O to yield the fluorophosphate
triethylammonium salt as a white solid (4.4–4.7 g, 0.022–0.024 mol).
3
. Synthesis of fluorophosphate imidazolide (FPIm, 5)
Fluorophosphate triethylammonium salt (4) (1.00 g, 4.95 mmol), imidazole (3.37 g, 49.5
mmol), 2,2′ꢀdithiodipyridine (3.27 g, 14.8 mmol) were mixed in 5 mL of DMF. Triethylamine
(
9.9 mmol) and triphenylphosphine (3.89 mg, 14.8 mmol) were added and the mixture was
stirred for 6–8 h. The product was precipitated from reaction mixture as a lithium salt by
addition of solution of anhydrous LiClO (2.64 g, 24.75 mmol) in dry acetonitrile (200 mL).
After cooling at 4 ºC for 30 min, the precipitate was filtered off, washed repeatedly with cold,
4
dry acetonitrile and dried in vacuum over P
4
O
10 to afford 750–800 mg of fluorophosphate
1
lithium salt (90–100%). MS(ꢀ)ES m/z 148.5 calc. for C
MHz, D
2.61 (d, J=951 Hz ,1 P); F NMR (376 MHz, D
3
H
3
N
2
O
2
PF: 148.9; H NMR (400
3
1
2
O) δ = 7.99 (s, 1 H), 7.36 (s, 1 H), 7.15 (s, 1 H); P NMR (162 MHz, D
2
O) δ = ꢀ
19
1
2
O) δ = ꢀ63.44 (d, J=951 Hz, 1 F)
4
. Synthesis of nucleotide imidazolide derivatives (2a–2l)
28
ACS Paragon Plus Environment

Page 29 of 55
The Journal of Organic Chemistry
1
2
3
4
5
6
7
8
9
All imidazolides except 2l were prepared using triphenyl phospihine/dithiodipyride
21
condensation system according to wellꢀestablished procedures. The commercially available
nucleotides were converted before the synthesis into triethylammonium salts. GMP disodium
salt, CMP disodium salt and UMP disodium salt were converted by passing their aqueous
solutions (ca. 1 g/15 mL) through Dowex 50 W x 8 cationite in the triethylammonium form.
In each case, the collected eluate was evaporated under reduced pressure with repeated
additions of ethanol, the residue dried in vacuum over P O to yield the nucleotide
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
2
5
triethylammonium salt as a white solid. AMP (free acid) was suspended in water/ethanol
mixture and neutralized to pH 7 by a stepwise addition of triethylamine under vigorous
stirring. The resulting solution was evaporated and dried analogously as described for GMP,
CMP and UMP. The nucleotides obtained by chemical synthesis, were isolated as
triethylammonium salts after DEAE Sephadex purification and, hence, no conversion was
required in their case. Then, an appropriate nucleotide (1 mmol, TEA salt), imidazole (8
mmol), 2,2′ꢀdithiodipyridine (3 mmol) were mixed in DMF (~2.5 mL/100 mg of nucleotide).
Triethylamine (2 mmol) and triphenylphosphine (3 mmol) were added and the mixture was
stirred for 6–24 h. The product was precipitated from reaction mixture as a sodium salt by
addition of a solution of anhydrous NaClO (4 mmol) in dry acetone (~8 volumes of DMF
4
volume). After cooling at 4 ºC, the precipitate was filtered off, washed repeatedly with cold,
4
dry acetone and dried in vacuum over P O10. Yields 90–100 %. Because of the chemically
labile nature of the compounds, the purity of the products was checked only by RP HPLC and
by measuring optical density miliunit per mg of dry compound. Compound 35 was prepared
by a different route, a oneꢀpot twoꢀstep procedure starting from adenosine as described
21
earlier. Compound 2a sodium salt (AMPIm): analytical RP HPLC t
R
12.0 min, MS(ꢀ)ES
396.0; 29.0 mOD/mg; 2b sodium salt (GMPIm): analytical RP HPLC t
R
9.2 min, MS(ꢀ)ES
4
12.1; 24.0 mOD/mg; 2d sodium salt (UMPIm): analytical RP HPLC t 8.7 min, MS(ꢀ)ES
R
29
ACS Paragon Plus Environment