Sterically-Controlled Self-Immolation in Phosphoramidate Linkers Triggered by Light

Article abstract of DOI:10.1002/ejoc.201901882

Phosphoramidate prodrugs stand out for their applications as drug delivery systems. However, information on their activation remains limited to prodrug half-live and metabolite structure, which hinders their development. Considering the lack of data on ph

Full text of DOI:10.1002/ejoc.201901882

A Journal of
Accepted Article
Title: Sterically-Controlled Self-Immolation in Phosphoramidate Linkers
Triggered by Light
Authors: Eliška Procházková, Juraj Filo, Marek Cigáň, and Ondrej
Baszczynski
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: Eur. J. Org. Chem. 10.1002/ejoc.201901882
Link to VoR: http://dx.doi.org/10.1002/ejoc.201901882
Supported by

10.1002/ejoc.201901882
European Journal of Organic Chemistry
Sterically-Controlled Self-Immolation in Phosphoramidate Linkers Triggered by Light
Eliška Procházková,a Juraj Filo,b Marek Cigáň b and Ondřej Baszczyňski c,*
a Institute of Organic Chemistry and Biochemistry, The Czech Academy of Sciences, Prague 116 10,
Czech Republic
b Institute of Chemistry, Comenius University, Bratislava 842 15, Slovakia
c Faculty of Science, Charles University, Prague 128 43, Czech Republic
baszczyo@natur.cuni.cz
Abstract
Phosphoramidate prodrugs stand out for their applications as drug delivery systems.
However, information on their activation remains limited to prodrug half-live and metabolite
structure, which hinders their development. Considering the lack of data on phosphoramidate self-
immolation, we studied light-triggered self-immolation of phosphoramidate-based linkers bearing
various amino acid arms by NMR spectroscopy with in situ irradiation, complemented with
mathematical fitting, and by optical methods. By changing the alanine arm, we were able to
systematically fine-tune, slow down or speed up the cyclization rate. The cyclization rate of branched
amino acid arms increased with the steric demands of their α-substituents according to the Thorpe-
Ingold effect. The cyclization of β-alanine and proline was very slow or did not proceed at all, but
lactate may be a good alternative for the alanine arm because lactate exhibits a significantly slower
cyclization rate than alanine (ca. 60 times). Overall, our study demonstrates that phosphoramidate
self-immolation strongly depends on the structure of the amino acid arm and on the conditions used
(solvent, concentration and temperature), thereby providing a solid foundation for the rational
design of new self-immolative phosphoramidate systems with desired properties.
1
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201901882
European Journal of Organic Chemistry
Introduction
Phosphoramidates and other phosphate-based constructs have become useful tools in
chemical biology[1] as probes for biochemical processes, such as membrane transport,[2] protein
phosphorylation[3] and conjugation.[1a] In addition, these compounds can also be used as alternatives
to phosphorothioate antisense oligonucleotides[4] or as candidates for the development of smart
materials[5], including biodegradable polymers,[6] or sensors for medical applications[7] such as
diagnosis[1b, 3, 8] and drug delivery.[9]
In particular, phosphoramidate prodrugs[10] (ProTides),[11] which were introduced by
McGuigan[12] in the late 1990s, have been successfully used in drug delivery systems as effective
lipophilic masking units for polar phosphate/phosphonate groups, thus increasing the membrane
transport of various nucleotide drugs, see detailed reviews.[13] These compounds are enzymatically
activated inside target cells via metabolic pathways, in a cascade of decomposition reactions,[14]
which could be considered self-immolative (SI) processes.[15]
Phosphoramidate self-immolation was first addressed by Imbach’s group[16] when performing
a comparative study of 3’-azidothymidine prodrugs, which showed that a minor modification on a
phosphoramidate linker could hamper the SI process, thereby completely inactivating compounds.
More recently, thanks to the commercial success of phosphoramidate prodrugs, several studies on
their activation have been performed, primarily using one of the following methods: 1) prodrug
incubation in various cell extracts and subsequent metabolite analysis by HPLC/MS[12c] or 2) prodrug
activation by carboxyesterase in an NMR tube with simultaneous reaction monitoring by 31P NMR
spectroscopy.[10, 17] Although these experiments provide general information on prodrug half-lives
and metabolite structures, they fail to describe phosphoramidate self-immolation. Therefore, the
lack of specific knowledge on phosphoramidate cyclization has hindered the development of ProTide
technology.[18]
Based on the above, we designed and prepared a series of phosphoramidate-based linkers
differing in the structure of the amino acid/lactate arm triggered by light (Figure 1). Photoactivation
2
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201901882
European Journal of Organic Chemistry
was selected for triggering the SI process to overcome chemical or enzymatic activation, which could
either hamper the self-immolation of our substrates or could be substrate-specific, respectively. In
addition, information on activation reaction rate is usually masked by a complex reaction cascade
and by the presence of unstable intermediates or an enzyme(s). However, photoactivation allows us
to directly observe the SI process by NMR spectroscopy with in situ irradiation in real time. In our
recent study,[19] we experimentally described mechanism of the SI process in phosphoramidates
using alanine-based linker 1 as a model compound (Figure 1). In the present study, we used
compound 1 as a standard for comparative purposes. To describe phosphoramidate self-immolation,
we assumed that the entire reaction pathway consists of two sequential reactions: 1) photochemical
activation (cleavage of the dimethoxynitrobenzyl (DMNB) group)[20] and 2) cyclization (SI process)
followed by cargo (phenol) release (Figure 1). This reaction cascade is strongly sensitive to the
structure of the phosphoramidate linker and to the conditions used – such as solvent, concentration
and temperature.
Figure 1. Mechanism of the phosphoramidate self-immolative process and structure of the phosphate-based self-
immolative probes 1–7 studied in this work.
3
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201901882
European Journal of Organic Chemistry
Results and Discussion
Synthesis
Convergent synthesis was used to prepare target phosphoramidate linkers. Alanine
derivative 1 was prepared as described previously (45%).[19] Phosphorylation agent 9 was always
freshly prepared from ethyl dichlorophosphate 8 at a synthesis scale required for phosphorylation of
one equivalent of substrate (Scheme 1). General synthetic strategy started from NH-Boc-protected L-
amino acids 10–12, 16, 18 and from D, L-lactate 20, which were protected at the carboxylate group
as the corresponding 3,4-dimethoxy-5-nitrobenzyl (DMNB) esters 13–15, 17, 19 and 21. Deprotection
of the NH-Boc groups followed by phosphorylation gave the final compounds 2–7.
In the first set of NH-Boc-protected amino acids, glycine 10, valine 11 and phenylalanine 12,
the corresponding DMNB-protected intermediates 13 (84%), 14 (80%) and 15 (85%) were prepared in
the reactions of 10, 11, 12 with DMNB bromide in DMF, respectively (Scheme 1). The NH-Boc group
of 13, 14 and 15 was deprotected by TFA in DCM, completed in 2 h, which was directly followed by
phosphorylation using the phosphorochloridate 9. However, isolation and purification of the final
compounds 2, 3, and 4 was challenging. Compounds 2 (19%), 3 (16%) and 4 (6%) were sequentially
purified by multiple flash or preparative TLC chromatography - using a gradient or isocratic mode -
either on normal or reverse phase silica gel. Installation of the DMNB group in NH-Boc proline 16 and
β-alanine 18 gave the corresponding esters 17 (90%) and 19 (83%). The NH-Boc group of 17 and 19
was deprotected by TFA in DCM and phosphorylated by phosphorochloridate 9. Compounds 5 (35%)
and 6 (38%) were purified by single reversed-phase chromatography or by normal-phase preparative
TLC. Lactate 20 was neutralized to pH 7 using an aqueous solution of cesium carbonate (20% w/v) to
form the corresponding lactate cesium salt,[21] which was dried and alkylated by DMNB bromide,
yielding 21 (crude). Crude 21 was subsequently phosphorylated by phosphorochloridate 9 to yield
the target compound 7 (4%). All compounds were spectroscopically characterized (NMR and ESI-MS
spectra in the section Compound characterization in the SI).
4
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201901882
European Journal of Organic Chemistry
Scheme 1. Synthesis of target phosphoramidate linkers. Reaction conditions: a) ethyl dichlorophosphate 8, phenol, DCM,
Et3N, –78 °C to 25 °C, 16 h; b) 3,4-dimethoxy-5-nitrobenzyl bromide, Cs2CO3, DMF, 0 °C to 25 °C, 18 h; c) 1. TFA, DCM, 0 °C
to 25 °C, 2 h; 2. DCM, Et3N, ethyl phenyl chlorophosphate 9, –78 °C to 25 °C, 16 h; d) 1. neutralization to pH 7, Cs2CO3,
MeOH, H2O; 2. 3,4-dimethoxy-5-nitrobenzyl bromide, DMF, 18 h, 25 °C; e) DCM, Et3N, ethyl phenyl chlorophosphate 9, –78
°C to 25 °C, 16 h.
Optimization of reaction conditions
According to our previous research,[19] self-immolation is highly sensitive to the experimental
conditions. In searching for suitable conditions, we used alanine-based linker 1, which has optimal SI
parameters (e.g. reaction rate), as a model compound. Moreover, alanine-based arm 1 is used in
medical applications – ProTide prodrugs – and is therefore a good representative of current clinical
practice.[22]
The concentration of substrate is an important factor in photochemistry. Accordingly, we had to
find a compromise between a concentration sufficiently low for photochemistry and a concentration
5
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201901882
European Journal of Organic Chemistry
high enough for recording NMR signals in a reasonable time frame (reaction monitoring in a minute
scale). As shown in Figure S1, 0.5 mM is optimal for photo-cleavage when using an in situ irradiation
setup and recording 31P NMR spectra every 2.5–5 min. Two 31P NMR signals correspond to two
diastereoisomers induced by a chiral center on phosphorus atom and L-amino acid. A higher
concentration (2 mM) slowed photo-activation in this setup, most likely due to saturation of light
absorption.
The solvent also plays a key role in the entire reaction course, as demonstrated in our previous
study.[19] Using DMSO in a mixture with buffers (0.1 M, pH 7.4, 20% DMSO), such as HEPES,
phosphate buffer (PBS) or cacodylate (CACO), allows us to obtain the final product in a higher yield
than when using the unbuffered DMSO/D2O system (Figure S2 in the SI). In addition, PBS and CACO
systems provided significantly higher concentrations of the final product P than the HEPES system.
Cacodylate buffer, a sodium salt of dimethylarsinic acid which works in a pH range of 5.0–7.4,[23] has
been reported as a non-phosphate alternative to PBS suitable for biochemical studies.[24] Indeed, the
CACO/DMSO mixture performed similarly to PBS and afforded the final product P (Figures 1 and S2 in
the SI) almost quantitatively, but the broad phosphate signal in 31P NMR spectra could hamper self-
immolation monitoring of target phosphoramidates in a PBS system. Moreover, the increase in
DMSO content (50%) supporting photo-activation (Figures S3 and S4 in the SI) led to precipitation of
the PBS, in contrast to the CACO system. Therefore, CACO was the best candidate, and a mixture of
50% CACO/DMSO provided optimal conditions for a detailed investigation of both reactions: 1)
photochemical activation and 2) cyclization leading to phenol release.
Temperature: To investigate the temperature effect on phosphoramidate self-immolation for
possible medicinal applications, we examined the SI process on model compound 1 at both room (25
°C) and physiological (37 °C) temperatures, which gave the final product in quantitative yields (Figure
S5 in the SI), thus assessing the efficiency in activation of the phosphoramidates.
6
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201901882
European Journal of Organic Chemistry
1) Photochemical activation
Photochemical activation was studied by HPLC analysis combined with optical spectroscopy
(Figures S6–S12 in the SI) to determine quantum yields and to compare the photochemical behavior
of specific phosphoramidates. In contrast to 31P NMR reaction monitoring, wherein photochemical
activation was detected as an increase in the concentration of intermediate I, optical spectroscopy
monitored photochemical cleavage as a decrease in concentration of the initial phosphoramidates A.
The studied compounds 1–7 exhibited the typical characteristic of the dimethoxynitrobenzyl (DMNB)
photoremovable protecting group:[25] relatively low photolysis quantum yield () in the 0.5–1 %
range (Table 1). As expected, based on its similarity among the substrates, the chemical structure of
the amide linkage has almost no effect on the photochemical properties of the DMNB moiety; thus
no considerable differences in the photolysis quantum yield of derivatives 1–4 and 6 (primary amine)
were observed. The proline derivative 5 (secondary amine) stands out among amino acid linkers for
its slightly higher quantum yield. Interestingly, single-atom substitution of the nitrogen atom in
phosphoramidate by oxygen yields a phosphoric acid ester, which also doubles the  value and,
therefore, the overall photolytic efficiency of compound 7 (Table 1). Accordingly, the photochemical
behavior of the derivatives 5 and 7 indicates the increased stabilization of the o-nitrobenzyl-type
26]
radical formed during the photo-fragmentation process.[26] As shown by Solomek et al.,[ radical-
stabilizing groups decrease barrier of the photo-induced hydrogen atom transfer, and thus reduce
the strength of the C-H bond that is cleaved, and simultaneously hinder the back thermal hydrogen
transfer.
7
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201901882
European Journal of Organic Chemistry
Table 1. Photochemical parameters of the studied compounds 1–7: Molar absorption coefficient, photolysis
quantum yield, photolytic efficiency, and reaction rate of photochemical activation.[a]
A
A
kA
Compound
kp

[mol of photons/ (mol
[s-1]
[M-1 cm-1]
[M-1 cm-1]
of cpd  s-1)]
5.8  10-3
5.8  10-3
5.9  10-3
6.2  10-3
4.5  10-3
4.4  10-3
4.5  10-3
3.0  10-5
2.8  10-5
2.9  10-5
3.2  10-5
5.0  10-5
2.0  10-5
4.7  10-5
1
2
3
4
5
6
7
5250
4930
4990
4400
5590
6120
5300
0.0054 ± 0.0003
0.0048 ± 0.0002
0.0049 ± 0.0002
0.0052 ± 0.0003
0.0107 ± 0.0004
0.0045 ± 0.0001
0.0100 ± 0.0002
28.5 ± 1.4
23.7 ± 1.0
24.2 ± 0.8
22.9 ± 1.3
61.9 ± 2.3
27.5 ± 0.9
55.5 ± 1.1
Molar absorption coefficient A at absorption maximum, photolysis quantum yield  determined at irrad = 340 nm,
photolytic efficiency A, specific rate of light absorption kA (kA = IA/c; IA – absorbed light intensity, c – concentration of
compound), first-order rate constant for direct photolysis kp (kp =  kA).
2) Cyclization step and cargo (phenol) release
The self-immolation of 1–7 differs dramatically, depending on the structure of the amino
acid/lactate arm, as shown by 31P NMR spectroscopy with in situ UV irradiation (365 nm, Figure 2).
The starting compound A (1–5) with two 31P NMR signals (δ ~ 4 ppm) is activated by UV light; the
photo-removable DMNB group is cleaved off, and the intermediate I is detected as two new
downfield-shifted 31P signals. The intermediate I is then converted to the final product P
characterized by the chemical shift – usually higher than 6 ppm (except compound 5, as discussed
below). The chemical structure of the intermediates I and of the final products P was determined in
situ by careful analysis of 13C spectra, based on C-P spin-spin interactions visible in the 13C spectra as
a signal splitting, as described in our previous work in detail.[19]
By changing the alanine arm (Figure 2A), we can systematically tune, slow down (derivative 2,
Figure 2B) or speed up (derivatives 3 and 4, Figure 2C and 2D), the reaction cyclization rate according
8
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201901882
European Journal of Organic Chemistry
to the Thorpe-Ingold effect, which explains the key role of α-substitution on the cyclization
process.[27]
Glycine derivative 2, which is less sterically demanding than valine derivative 3, thus provides a
significantly lower concentration of the final product P (Figure 2B and 2C). Interestingly, in the case
of valine derivative 3, we were unable to detect the intermediate I, albeit preparing the final product
P in a quantitative yield. This means that the cyclization step is extremely fast and, therefore,
photoactivation is the rate-limiting step of this process. Hence, the rate constant of the cyclization
step could be calculated using a steady-state approximation assuming that 푑[퐼] = 0. Although phenyl
푑푡
alanine derivative 4 is more sterically hindered than valine analog 3, traces of intermediate I were
detected (Figure 2D). In this case, cyclization is the rate-limiting step, thus clearly showing that the
Thorpe-Ingold theory can partly explain this trend, but steric hindrance is not the only factor that
must be considered when analyzing this process. As mentioned above, the solvent plays a key role in
the entire reaction. Most likely, the intermediate derived from phenylalanine analog 4 could be
stabilized by solvation and, therefore, detected in 31P NMR spectra. π-π stacking between
phenylalanine and phenol benzene rings can also significantly stabilize intermediate I. Surprisingly,
no cyclization was observed in proline derivative 5, wherein the cyclic side chain is even more
sterically constrained than in 4, and no final product was detected even after several months. This
might be explained by the high rigidity of the proline arm in which cyclization is not energetically
preferred and, therefore, the intermediate is unable to release phenol.
9
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201901882
European Journal of Organic Chemistry
Figure 2. 31P NMR spectra of compounds 1–5 (A–E, respectively) measured during in situ irradiation by UV light (365 nm). In
all reactions, a concentration of 0.5 mM of the compound in 400 μL of a solvent mixture of 50% CACO (0.1M, pH
7.4)/[D6]DMSO was used at room temperature (25 °C).
β-alanine derivative 6 can be considered an elongated analog of glycine derivative 2 (β –amino
acid) and, surprisingly, needed more than 9 weeks to provide at least traces of the final product, in
contrast to 2, which yielded the final product within an hour (Figures 3 and S13 in the SI). This
considerable difference in cyclization rate shows that the cyclization rate of the studied compounds
closely depends on the arm length. Moreover, the sharp contrast between glycine 2 and β-alanine 6
derivatives also supports the theory of primordial protein synthesis, wherein only α-amino acids
were preferred in the formation of the primordial protein upon cyclization.[28]
Figure 3. 31P NMR spectra of 0.5 mM compounds 2 (left) and 6 (right) in 50% CACO/[D6]DMSO after 60 min of continuous
irradiation by UV light (365 nm) and after switching the light source off.
10
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201901882
European Journal of Organic Chemistry
For further insight into phosphoramidate cyclization, an alternative phosphotriester derivative 7,
wherein the amino acid-based arm was substituted by lactate, was prepared. The structure of lactate
derivative 7 differs from that of alanine derivative 1 in a single atom, the one directly attached to
phosphorus (NH- in 1, instead of O- in 7). Interestingly, this single-atom alteration has a significant
impact on the self-immolative process. Compound 1 provided the final product within 15 min,
whereas lactate derivative 7 provided the final product overnight (Figure 4). This finding can be
useful when designing linkers for delayed cargo release. The structure of the final product was
determined based on commonly used NMR methods.
All final products were stable at room temperature and showed no decomposition for several
months.
Figure 4. 31P NMR spectra at room temperature (25 °C). Self-immolation of compounds 1 (left) and 7 (right) in 50%
CACO/[D6]DMSO during irradiation by UV light (365 nm). While compound 1 provided the final product P within 15 min,
lactate derivative afforded product P overnight.
11
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201901882
European Journal of Organic Chemistry
Mathematical fitting of kinetic curves extracted from 31P NMR spectra
Mathematical fitting of the reaction course monitored by 31P NMR data enabled us to estimate
the reaction rates of both photolysis and cyclization reactions (Table 2). The rate constants were one
order of magnitude lower in HPLC/UV-Vis than in NMR spectroscopy, most likely due to differences
in light intensity between the methods (ca. 1 mW and 12 mW, respectively). For further details,
please refer to the SI and Figures S14–S20. It should also be noted that the rate constant of direct
photolysis kp and the rate constant of intermediate ka are very close to those of all studied photo-
triggered phosphoramidates (Table 2), but the  of 5 and 7 was twice higher than that of other
derivatives determined by UV-Vis/HPLC in DMSO. However, the determined kp values can be
influenced by the significant light absorption of the photoproduct at the irradiation wavelength,
although role of deuterated solvent or irradiation wavelength dependence of  (340 vs. 365 nm,
respectively) cannot be excluded. A more detailed assessment of deuterated solvent effects and of
the variation of  as a function of wavelength was outside the scope of this work.
Table 2. Kinetic parameters of photolysis for the studied compounds 1–7 determined from data collected by
31P NMR spectroscopy at 0.5 mM initial concentration of the starting compounds.
Cpd
1
kp
ka
kb
4.45  10-4
7.20  10-4
-
5.44  10-4
1.05  10-4
2
-
3.63  10-4
3
-
-
6.93  10-4
1.00  10-3
4
-
4.31  10-4
5.11  10-4
*1.39  10-4
5
-
-
-
-
-
-
6
7
Rate constant for direct photolysis kp, rate constant for intermediate ka, rate constant for product kb. *Measured at a
higher concentration of the starting compound (ca. 1.2 mM).
12
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201901882
European Journal of Organic Chemistry
Conclusions
We designed and synthetized several phosphoramidate-based linkers to study their self-
immolation. The linkers were altered by a set of various amino acids or lactate, and sterically
demanding valine- and phenylalanine-based linkers showed the highest cyclization rate (based on
phenol release), in line with the Thorpe-Ingold effect. Furthermore, we observed a significant solvent
effect. Photoactivation was faster in an organic solvent, whereas cyclization was faster in aqueous
solution. Therefore, we performed all experiments in a mixture of organic and aqueous solvent in a
1:1 ratio. Moreover, the presence of buffer markedly supports the entire reaction. Thus, as a non-
phosphate alternative to PBS, cacodylate buffer was selected as the best buffer for our purposes.
Photochemical analysis proved that all examined compounds were activated similarly, with quantum
yields ranging from 0.5 to 1% and with reaction rates in the same order of magnitude.
We believe that our results lay a solid foundation for the rational design of new self-
immolative systems, thereby applying the phosphoramidate concept to the development of various
drug delivery systems, smart materials or advanced tools for chemical biology.
Experimental Section
General. All reagents were purchased from Sigma–Aldrich (Prague, Czech Republic). Dry
solvents were purchased from Across-Organics (Czech Republic) and protected amino acids were
bought from TCI (Europe). 1-(Bromomethyl)-4,5-dimethoxy-2-nitrobenzene (53413-67-5) was
purchased from Fluorochem Ltd. (United Kingdom). All reactions were performed under an inert
atmosphere of argon. Thin layer chromatography (TLC) and preparative TLC was performed on TLC
aluminium sheets (silica gel 60 F254; Merck). Flash-column chromatography was performed on
CombiFlash® (Teledyne Isco) chromatography system using silica gel 230–400 mesh, 60 Å (Merck) or
reversed-phase C-18 RediSep Rf Gold® pre-packed columns. HPLC analyses were performed on
Agilent 1100 HPLC System (with G1311A quaternary pump, G1314A UV VWD detector) using
LiChrospher100, RP-18 column (Endcapped 5m, 4 mm  250 mm) in isocratic mode 55% MeCN/
13
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201901882
European Journal of Organic Chemistry
water (v/v, modified with 0.1% phosphoric acid, 1.0 mL/min,  = 350 nm). Final products were
recovered by solvent evaporation and freeze-drying.
Phosphorylation agent - Ethyl phenyl phosphorochloridate (9).[19] Phenol (94 µL, 0.95 mmol,
0.95 eq.) and TEA (138 µL, 0.95 mmol, 0.95 eq.) were dissolved in DCM (3 mL). The solution was
cooled down to –78 °C and ethyl phosphorodichloridate 8 (119 µL, 1 mmol, 1 eq.) was added
dropwise. The mixture was stirred at –78 °C for 15 min and then allowed to warm up to 25 °C and
stirred for 2 h. The crude reaction mixture was stored under argon and directly used in the next
reaction steps in the amount required for phosphorylation of protected amino acids, according to
procedure B. [M+H]+ 221.09.
4,5-Dimethoxy-2-nitrobenzyl (ethoxy(phenoxy)phosphoryl)-L-alaninate (1).[19] Compound 1
was prepared according to the literature[19] from Boc-L-Ala-OH. Yield (175 mg, 45%). NMR and HRMS
(ESI+) data confirmed the structure. 1H NMR (500 MHz, [D6]DMSO, 25 °C, 2 diastereoisomers) δ 7.71
and 7.70 (s, 3’’), 7.35–7.30 (m, 3’), 7.21 and 7.20 (s , 6’’), 7.18–7.11 (m, 3H, 2’, 4’), 6.06 and 6.01 (dd,
JNH-P = 13.1, JNH-2 = 10.1, 1), 5.45–5.36 (m, 2H, 4), 4.07–3.92 (m, 3H, OCH2CH3, 2), 3.89 and 3.88 and
3.87 and 3.87 (s, 2x 3H, 4’’-OCH3, 5’’-OCH3), 1.29 and 1.26 (d, 3H, JCH3-2 = 7.1, 2-CH3), 1.21 and 1.18 (t,
3H, JCH3-CH2 = 7.1, OCH2CH3); 31P NMR (202 MHz, 50% CACO/[D6]DMSO, 25 °C) δ 3.73 and 3.46 (2 x s, 2
diastereoisomers); HPLC (Rt) 6.59 min.
Procedure A: General procedure for the preparation of DMNB-protected amino acids. NH-
Boc protected amino acid (1 mmol, 1 eq.) was dissolved in DMF (4 mL), cesium carbonate (195 mg,
0.6 mmol, 0.6 eq.) was added and the suspension was stirred at 25 °C for 15 min. The suspension was
cooled down to 0 °C and 1-(bromomethyl)-4,5-dimethoxy-2-nitrobenzene 11 (290 mg, 1.05 mmol,
1.05 eq.) was added in one portion. Then, the mixture was stirred at 25 °C for 16 h and then
evaporated in vacuo. Chromatographic methods are described above.
4,5-Dimethoxy-2-nitrobenzyl (tert-butoxycarbonyl)glycinate (13). Boc-L-Gly-OH 10 (175 mg,
1.0 mmol) was treated using procedure A. Crude product was purified by Isco-Flash reverse-phase
14
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201901882
European Journal of Organic Chemistry
chromatography using a gradient (H2O-MeOH, 100:0 → 0:100, v/v) which afforded product 13 as an
off-white solid (311 mg, 84%). 1H NMR (400 MHz, [D6]DMSO, 25 °C) δ 7.71 (s, 1H, 3’’), 7.35 (t, 1H, JNH-2
= 6.1, NH), 7.19 (s, 1H, 6’’), 5.44 (s, 2H, 4), 3.94 and 3.87 (2 x s, 2 x 3H, 4’’-O-CH3, 5’’-O-CH3), 3.80 (d,
1H, J2-NH =6.1, 2), 1.37 (s, 9H, C(CH3)3); 13C {1H}NMR (100 MHz, [D6]DMSO, 25 °C) δ 170.4 (3), 156.1
(NH-CO), 153.6 and 148.1 (4’’, 5’’), 139.6 (2’’), 126.5 (1’’), 111.1 (6’’), 108.4 (3’’), 78.5 (C(CH3)3), 63.0
(4), 56.5 and 56.3 (4’’-O-CH3, 5’’-O-CH3), 42.2 (2), 28.3 (C(CH3)3); HRMS (ESI+) calcd for
C16H22O8N2Na 393.1268 [M+Na]+, found 393.1266.
4,5-Dimethoxy-2-nitrobenzyl (tert-butoxycarbonyl)-L-valinate (14). Boc-L-Val-OH 11 (217 mg,
1.0 mmol) was treated using the procedure A. The crude product was purified by Isco-Flash reverse-
phase chromatography using a gradient (H2O-MeOH, 100:0 → 0:100, v/v) which afforded product 14
1
as an off-white solid (330 mg, 80%). H NMR (400 MHz, [D6]DMSO, 25 °C) δ 7.71 (s, 1H, 3’’), 7.31 (d,
1H, JNH-2 = 8.0, NH), 7.24 (s, 1H, 6’’), 5.47 and 5.42 (2 x d, 1H, Jgem = 14.4, Jgem = 14.4, 4), 3.94 and 3.87
(2 x s, 2 x 3H, 4’’-O-CH3, 5’’-O-CH3), 3.9 (2, from HSQC), 2.04 (m, 1H, CHiPr), 1.37 (s, 9H, C(CH3)3), 0.89
and 0.88 (2 x d, 2 x 3H, JCH3-CH = 6.8, JCH3-CH = 6.8, CH(CH3)2); 13C {1H}NMR (100 MHz, [D6]DMSO, 25 °C) δ
171.9 (3), 156.0 (NH-CO), 153.5 and 148.1 (4’’, 5’’), 139.7 (2’’), 126.5 (1’’), 111.4 (6’’), 108.4 (3’’), 78.5
C(CH3)3), 63.0 (4), 59.8 (2), 56.5 and 56.3 (4’’-O-CH3, 5’’-O-CH3), 29.7 (CHiPr), 28.3 (C(CH3)3), 19.1 and
18.6 (CH(CH3)3); HRMS (ESI+) calcd for C19H28O8N2Na 435.1738 [M+Na]+, found 435.1737.
4,5-Dimethoxy-2-nitrobenzyl (tert-butoxycarbonyl)-L-phenylalaninate (15). Boc-L-Phe-OH 12
(265 mg, 1 mmol) was treated using procedure A. The crude product was purified by Isco-Flash
reverse-phase chromatography using a gradient (H2O-MeOH, 100:0 → 0:100, v/v) which afforded
1
product 15 as an off-white solid (390 mg, 85%). H NMR (400 MHz, [D6]DMSO, 25 °C) δ 7.71 (s, 1H,
3’’), 7.46 (d, 1H, JNH-2 = 7.9, NH), 7.30–7.18 (m, 6H, 6’’, 2’’’, 3’’’,4’’’), 5.46 (d, 1H, Jgem = 14.6, 4a), 5.37
(d, 1H, Jgem = 14.6, 4b), 4.30 (m, 1H, 2), 3.94 and 3.87 (2 x s, 2 x 3H, 4’’-O-CH3, 5’’-O-CH3), 3.05 (dd, 1H,
Jgem = 13.18, JCHa-2 = 5.3, 2-CHa), 2.92 (dd, 1H, Jgem = 13.8, JCHa-2 = 10.1, 2-CHb), 1.30 (s, 9H, C(CH3)3); 13C
{1H}NMR (100 MHz, [D6]DMSO, 25 °C) δ 172.1 (3), 155.7 (NH-CO), 153.6 and 148.1 (4’’, 5’’), 139.6
15
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201901882
European Journal of Organic Chemistry
(2’’), 137.7 (1’’’), 129.3 (2’’’), 128.4 (3’’’), 126.7 (4’’’), 126.6 (1’’), 111.0 (6’’), 108.4 (3’’), 78.6 (C(CH3)3),
63.2 (4), 56.5 and 56.3 (4’’-O-CH3, 5’’-O-CH3), 55.6 (2), 36.4 (2-CH2), 28.3 (C(CH3)3); HRMS (ESI+) calcd
for C23H28O8N2Na 483.1740 [M+Na]+, found 483.1730.
1-(tert-Butyl) 2-(4,5-dimethoxy-2-nitrobenzyl) (S)-pyrrolidine-1,2-dicarboxylate (17). Boc-L-
Pro-OH 16 (215 mg, 1 mmol) was treated using the procedure A. The crude product was purified by
Isco-Flash chromatography using isocratic mode (cyclohexane-EtOAc, 20:80, v/v) which afforded
product 17 as an off-white solid (367 mg, 90%). 1H NMR (400 MHz, [D6]DMSO, 25 °C) δ 7.71 and 7.70
(2 x s, 2 x 1H, 3’’), 7.21 (s, 2H, 6’’), 5.47–5.38 (m, 4H, 7), 4.31 (dd, 1H, J2-3b = 8.1, J2-3a = 3.3, 2), 4.24
(dd, 1H, J2-3b = 8.6, J2-3a = 4.1, 2), 3.95 and 3.90 and 3.87 (3 x s, 2 x 3H, 1 x 6H, 4’’-O-CH3, 5’’-O-CH3),
3.41–3.29 (m, 4H, 5), 2.29–2.16 (m, 2H, 3a), 1.94–1.78 (m, 6H, 4), 1.37 and 1.27 (2 x s, 2 x 9H,
C(CH3)3); 13C {1H}NMR (100 MHz, [D6]DMSO, 25 °C) δ 172.4 and 172.3 (6), 153.8–153.0 (NH-CO, 5’’ or
4’’), 148.3 and 148.1 (4’’ or 5’’), 140.3 and 140.1 (2’’), 126.5 and 126.6 (1’’), 112.0 and 111.2 (6’’),
108.5 and 108.4 (3’’), 79.2 and 79.1 (C(CH3)3), 63.4 and 63.3 (7), 58.9 and 58.7 (2), 56.5 and 56.3 and
56.3 (4’’-O-CH3, 5’’-O-CH3), 46.6 and 46.3 (5), 30.5 and 29.6 (3), 28.2 and 28.0 (C(CH3)3), 24.2 and 23.3
(4); HRMS (ESI+) calcd for C19H26O8N2Na 433.1581 [M+Na]+, found 433.1576.
4,5-Dimethoxy-2-nitrobenzyl 3-((tert-butoxycarbonyl)amino)propanoate (19). Boc-β-Ala-OH
18 (189 mg, 1 mmol) was treated using procedure A. Crude product was purified by Isco-Flash
reverse-phase chromatography using a gradient (H2O-MeOH, 100:0 → 0:100, v/v) which afforded
1
product 19 as an off-white solid (320 mg, 83%). H NMR (400 MHz, [D6]DMSO, 25 °C) δ 7.70 (s, 1H,
3’’), 7.19 (s, 1H, 6’’), 6.89 (t, 1H, JNH-2 = 5.3, NH), 5.38 (s, 2H, 5), 3.92 and 3.87 (2 x s, 2 x 3H, 4’’-O-CH3,
5’’-O-CH3), 3.20 (m, 2H, 2), 2.54 (t, 2H, J3-2 = 7.0, 3), 1.35 (s, 9H, C(CH3)3); 13C {1H}NMR (100 MHz,
[D6]DMSO, 25 °C) δ 171.1 (5), 155.7 (NH-CO), 153.4 and 148.1 (4’’, 5’’), 139.9 (2’’), 126.4 (1’’), 111.6
(6’’), 108.4 (3’’), 77.9 (C(CH3)3), 62.8 (5), 56.4 and 56.3 (4’’-O-CH3, 5’’-O-CH3), 36.2 (2), 34.4 (3), 28.3
(C(CH3)3); HRMS (ESI+) calcd for C17H24O8N2Na 407.1425 [M+Na]+, found 407.1423.
16
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201901882
European Journal of Organic Chemistry
Procedure B: General procedure for the phosphorylation of DMNB-protected amino acids.
Boc-group deprotection: DMNB-protected NH-Boc amino acid (1 eq.) was dissolved in DCM (3 mL)
and cooled to 0 °C. Trifluoroacetic acid (3 mL) was added dropwise and the mixture was stirred at 0
°C for 30 min. The reaction was allowed to warm up to 25 °C and stirred for additional 2 h. The
solvent was evaporated and crude residue was co-evaporated with DCM (3x), toluene (1x) and dried
on Hi-vac pump (0.1 mBar) for 2 h. Phosphorylation: Dry residue was dissolved in DCM (3 mL) and
TEA (280 µL, 2 mmol, 2 eq.) was added. Solution was cooled down to –78 °C and ethyl phenyl
phosphorochloridate 9 (1 eq.) was added dropwise. Reaction mixture was stirred at –78 °C for 30
min, and then it was allowed to warm up to 25 °C and stirred for 4 h. The crude reaction mixture was
purified using various methods; silica gel chromatography and/or reverse-phase silica gel
chromatography and/or preparative TLC.
4,5-Dimethoxy-2-nitrobenzyl (ethoxy(phenoxy)phosphoryl)glycinate (2). Compound 13
(250 mg, 0.68 mmol) was treated using procedure B. The solvent was evaporated, and the crude
material was purified by Isco-Flash reverse-phase silica gel chromatography (H2O/MeOH, 100:0 →
0:100, v/v), followed by isocratic silica gel preparative TLC (cyclohexane/EtOAc, 40:60, v/v) to give a
1
white gummy solid 2 (59 mg, 19%). H NMR (400 MHz, [D6]DMSO, 25 °C) δ 7.71 (s, 1H, 3’’), 7.33 (m,
2H, 3’), 7.20 (s, 1H, 6’’), 7.17 (m, 2H, 2’), 7.14 (m, 1H, 4’), 5.93 (dt, 1H, JNH-P = 13.3, JNH-2 = 7.3, NH),
5.43 (s, 2H, 4), 4.06 (m, 2H, O-CH2-CH3), 3.90 and 3.87 (2 x s, 2 x 3H, 4’’-O-CH3, 5’’-O-CH3), 3.78 (dd,
2H, J2-P = 14.3, J2-NH = 7.3, 2), 1.21 (dt, 3H, JCH3-CH2 = 7.1, JCH3-P = 0.6, O-CH2-CH3); 31P NMR (202 MHz,
50% CACO/[D6]DMSO, 25 °C) δ 4.82; 13C {1H}NMR (100 MHz, [D6]DMSO, 25 °C) δ 170.8 (J3-P = 3.9, 3),
153.5 (4’’ or 5’’), 151.0 (J1’-P = 6.2, 1’), 148.1 (5’’ or 4’’), 139.7 (1’’ or 2’’), 129.7 (3’), 126.3 (2’’ or 1’’),
124.6 (4’), 120.3 (J2’-P = 4.6, 2’), 111.4 (6’’), 108.4 (3’’), 63.1 (4), 62.5 (JCH2-P = 5.3, O-CH2-CH3), 56.5 and
56.3 (4’’-O-CH3 and 5’’-O-CH3), 42.8 (2), 16.1 (d, JCH3-P = 6.9, O-CH2-CH3); HRMS (ESI+) calcd for
C19H24O9N2P 455.1214 [M+H]+, found 455.1209; HPLC (Rt) 5.63 min.
17
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201901882
European Journal of Organic Chemistry
4,5-Dimethoxy-2-nitrobenzyl (ethoxy(phenoxy)phosphoryl)-L-valinate (3). Compound 14
(250 mg, 0.61 mmol) was treated using procedure B. The solvent was evaporated, and the crude
material was purified by Isco-Flash reverse-phase silica gel chromatography (H2O/MeOH, 100:0 →
0:100, v/v), followed by Isco-Flash gradient silica gel chromatography (cyclohexane/EtOAc, 100:0 →
1
40:60, v/v) to give a white amorphous solid 3 (49 mg, 16%). H NMR (400 MHz, [D6]DMSO, 25 °C) δ
7.70 and 7.69 (2 x s, 2 x 1H, 3’’), 7.34–7.27 (m, 4H, 3’), 7.22 and 7.21 (2 x s, 2 x 1H, 6’’), 7.16–7.09 (m,
6H, 2’, 4’), 5.94 and 5.88 (2 x dd, 2 x 1H, JNH-P = 12.1 and 11.6, JNH-2 = 10.9 and 10.9, NH), 5.43–5.33 (m,
4H, 4), 4.06–3.97 (m, 4H, O-CH2-CH3), 3.90 and 3.88 and 3.87 and 3.87 (4 x s, 4 x 3H, 4’’-O-CH3, 5’’-O-
CH3), 3.65–3.57 (m, 2H, 2), 2.00-1.90 (m, 2H, CH(CH3)2), 1.20 and 1.18 (2 x td, 2 x 3H, JCH3-CH2 = 7.1 and
7.1, JCH3-P = 0.7 and 0.6, O-CH2-CH3), 0.86 and 0.85 and 0.84 and 0.79 (4 x d, 4 x 3H, JCH3-CH = 7.0 and
7.0 and 6.8 and 6.8, CH(CH3)2); 31P NMR (202 MHz, 50% CACO/[D6]DMSO, 25 °C) δ 4.42 and 4.04 (2 x
s, 2 diastereoisomers); 13C {1H}NMR (100 MHz, [D6]DMSO, 25 °C) δ 172.4 and 172.2 (2 x d, J3-P = 2.8
and 2.6, 3), 153.4 (5’’ or 4’’), 151.1 (d, J1’-P = 6.6, 1’), 148.2 (4’’ or 5’’), 139.9 and 139.9 (2’’), 129.6 (3’),
126.1 and 126.0 (1’’), 124.5 and 124.4 (4’), 120.3 and 120.1 (2 x d, J2’-P = 5.2 and 4.8, 2’), 111.8 and
111.7 (6’’), 108.4 (3’’), 63.1 and 63.0 (4), 62.5 and 62.4 (2 x d, JCH2-P = 5.6 and 5.5, O-CH2-CH3), 60.6 (d,
J2-P = 3.3, 2), 56.5 and 56.5 and 56.3 and 56.3 (5’’-O-CH3, 4’’-O-CH3), 31.3 and 31.3 (2 x d, JCH-P = 7.5
and 7.4, 2-CH), 19.1 and 19.1 and 18.3 and 18.1 (2-CH-(CH3)2), 16.1 and 16.1 (2 x d, JCH3-P = 7.1 and
7.1, O-CH2-CH3); HRMS (ESI+) calcd for C22H30O9N2P 497.1684 [M+H]+, found 497.1679; HPLC (Rt)
10.28 min.
4,5-dimethoxy-2-nitrobenzyl (ethoxy(phenoxy)phosphoryl)-L-phenylalaninate (4). Compound
15 (250 mg, 0.54 mmol) was treated using procedure B. The solvent was evaporated, and the crude
material was purified by Isco-Flash silica gel chromatography using a gradient cyclohexane/EtOAc
(100:0 → 0:50, v/v), followed by isocratic preparative TLC (cyclohexane/EtOAc, 40:60, v/v) to give off-
white amorphous solid 4 (17 mg, 6%). 1H NMR (400 MHz, [D6]DMSO, 25 °C) δ 7.70 and 7.60 (2 x s, 2 x
1H, 3’’), 7.28–7.20 (m, 7H, 3’,2’’’, 3’’’, 4’’’), 7.18 and 7.16 (2 x s, 2 x 1H, 6’’), 7.11–7.05 (m, 1H, 4’),
6.99-6.92 (m, 2H, 2’), 6.19–6.08 (m, 1H, NH), 5.42-5.32 (m, 2H, 4), 4.13–3.95 (m, 1H, 2), 3.89–3.85 (m,
18
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201901882
European Journal of Organic Chemistry
6H, 4’’-O-CH3 and 5’’-O-CH3), 3.84–3.55 (m, 2H, O-CH2-CH3), 3.09–2.99 and 2.87–2.79 (2 x m, 2H, 2-
CH2), 1.07–1.00 (m, 3H, O-CH2-CH3); 31P NMR (202 MHz, 50% CACO/[D6]DMSO, 25 °C) δ 3.49 and 3.02
(2 x s, 2 diastereoisomers); 13C {1H}NMR (100 MHz, [D6]DMSO, 25 °C) δ 172.2 (3), 153.5 (4’’ or 5’’),
150.8 (1’), 148.2 (5’’ or 4’’), 139.6 (1’’ or 2’’), 137.3 (1’’’), 129.6 and 129.5 (2’’’, 3’), 128.4 (3’’’), 126.7
(4’’’), 126.2 and 126.1 (1’’ or 2’’), 124.4 (4’), 120.2 and 120.0 (2 x d, J2’-P = 4.6, J2’-P = 4.4, 2’), 111.4 and
111.3 (6’’’), 108.3 (3’’), 63.1 (4), 62.3–62.1 (m, O-CH2-CH3), 56.7 and 56.6 (2), 56.5 and 56.3 (4’’-O-
CH3, 5’’-O-CH3); HRMS (ESI+) calcd for C26H30O9N2P 545.1683 [M+H]+, found 545.1681; HPLC (Rt) 13.43
min.
4,5-Dimethoxy-2-nitrobenzyl (ethoxy(phenoxy)phosphoryl)-L-prolinate (5). Compound 17 (300
mg, 0.73 mmol) was treated using procedure B. The solvent was evaporated, and the crude material
was purified by Isco-Flash reverse phase chromatography using a H2O/MeOH gradient (100:0 →
1
0:100, v/v) to give white amorphous solid 5 (128 mg, 35%). H NMR (500 MHz, [D6]DMSO, 25 °C) δ
7.70 and 7.69 (2 x s, 2 x 1H, 3’’), 7.38–7.35 and 7.31–7.27 (2 x m, 2 x 2H, 3’), 7.22–7.09 (m, 6H, 6’’, 2’,
4’), 5.42 and 5.37 (2 x s, 2 x 2H, 7), 4.33 and 4.25 (2 x m, 1H, 2), 4.12–3.97 (m, 4H, O-CH2-CH3), 3.91
and 3.87 and 3.86 and 3.85 (4 x s, 4 x 3H, 4’’-O-CH3, 5’’-O-CH3), 3.27–3.18 (m, 4H, 5), 2.21–1.72 (m,
8H, 3, 4), 1.20 and 1.19 (2 x t, 2 x 3H, JCH3-CH2 = 7.0, JCH3-CH2 = 6.9, O-CH2-CH3); 31P NMR (202 MHz, 50%
CACO/[D6]DMSO, 25 °C) δ 2.08 and 1.98 (2 x s, 2 diastereoisomers); 13C {1H}NMR (126 MHz,
[D6]DMSO, 25 °C) δ 173.2 and 172.8 (6), 153.5 (5’’), 150.8 and 150.7 (2 x d, J1’-P = 6.5, J1’-P = 5.8, 148.2
and 148.1 (4’’), 139.8 and 139.7 (2’’), 130.0 and 129.8 (3’), 126.3 and 126.2 (1’’), 124.9 and 124.8 (4’),
120.3 and 120.1 (2 x d, 2 x J2’-P = 4.6, 2’), 111.6 and 111.4 (6’’), 108.4 and 108.4 (3’’), 63.5 and 63.4 (7),
62.7–62.6 (m, O-CH2-CH3), 60.6 and 60.1 (2 x d, J2-P = 6.8, J2-P = 6.3, 2), 56.6–56.3 (m, 4’’-O-CH3, 5’’-O-
CH3), 47.5 and 46.8 (2 x d, J5-P = 4.6, J5-P = 4.3, 5), 31.1 and 31.0 (2 x d, J3-P = 8.5, J3-P = 8.3, 5), 25.2 and
25.1 (2 x d, J4-P = 8.7, J4-P = 8.6, 4), 16.2 and 16.1 (O-CH2-CH3); HRMS (ESI+) calcd for C22H27O9N2NaP
517.1346 [M+Na]+, found 517.1339; HPLC (Rt) 10.33 and 12.08 min.
19
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201901882
European Journal of Organic Chemistry
4,5-Dimethoxy-2-nitrobenzyl 3-((ethoxy(phenoxy)phosphoryl)amino)propanoate (6). Compound 19
(300 mg, 0.78 mmol) was treated using procedure B. The solvent was evaporated, and the crude
material was purified by preparative TLC (4 x 100mg crude, cyclohexane/EtOAc, 80:20, v/v) to give an
off-white solid 6 (140 mg, 38%). 1H NMR (400 MHz, [D6]DMSO, 25 °C) δ 7.69 (s, 1H, 3’’), 7.34 (m, 2H,
3’), 7.18–7.12 (m, 4H, 6’’, 2’, 4’), 5.45 (dt, 1H, JNH-P = 12.3, JNH-2 = 6.7, NH), 5.36 (s, 2H, 5), 4.01 (m, 2H,
O-CH2-CH3), 3.90 and 3.87 (2 x s, 2 x 3H, 4’’-O-CH3, 5’’-O-CH3), 3.13 (m, 2H, 2), 2.54 (t, 2H, J3-2 = 6.9, 3),
1.23 (dt, 3H, JCH3-CH2 = 7.1, JCH3-P = 0.7, O-CH2-CH3); 31P NMR (202 MHz, 50% CACO/[D6]DMSO, 25 °C) δ
5.56; 13C {1H}NMR (100 MHz, [D6]DMSO, 25 °C) δ 170.9 (4), 153.4 (5’’ or 4’’), 151.1 (d, J1’-P = 6.5, 1’),
148.1 (4’’ or 5’’), 139.9 (1’’ or 2’’), 129.8 (3’), 126.3 (2’’ or 1’’), 124.5 (4’), 120.3 (d, J2’-P = 4.9, 2’), 111.5
(6’’), 108.4 (3’’), 62.8 (5), 62.3 (d, JCH2-P = 5.5, O-CH2-CH3), 56.4 and 56.3 (4’’-O-CH3 and 5’’-O-CH3),
37.2 (2), 36.1 (d, J3-P = 5.9, 3), 16.2 (d, JCH3-P = 6.8, O-CH2-CH3); HRMS (ESI+) calcd for
C20H26O9N2P 469.1370 [M+H]+, found 469.1366; HPLC (Rt) 5.99 min.
4,5-Dimethoxy-2-nitrobenzyl 2-((ethoxy(phenoxy)phosphoryl)oxy)propanoate (7). DMNB
group installation: D, L -Lactic acid 20 (75 µL, 1 mmol) was dissolved in aqueous methanol (5 mL,
MeOH/H2O, 1:9, v/v), and the solution was neutralized to pH 7 by adding a few drops of aqueous
solution of Cs2CO3 (20%, w/v). The solution was evaporated, and the solid residue was co-evaporated
with DMF (2x) and toluene (1x). Then, the solids were suspended in DMF (5 mL), cooled to 0 °C, and
1-(bromomethyl)-4,5-dimethoxy-2-nitrobenzene was added (290 mg, 1.05 mmol). The mixture was
stirred for 16 h at 25 °C. Crude 21 was evaporated, passed through a reverse-phase column, and
directly used for the next phosphorylation step. Phosphorylation: Crude product from the previous
reaction 21 (134 mg) was dissolved in DCM (3 mL) and TEA (140 µL, 1 mmol, 2 eq.) was added.
Solution was cooled down to –78 °C and ethyl phenyl phosphorochloridate 9 (1 eq.) was added
dropwise. Reaction mixture was stirred at –78 °C for 30 min, and then it was allowed to warm up to
25 °C and stirred for 4 h. The solvent was evaporated, and the raw material was purified by Isco-Flash
silica gel chromatography (cyclohexane-EtOAc, 60:40, v/v) to give an off-white solid 7 (20 mg, 4%). 1H
NMR (500 MHz, [D6]DMSO, 25 °C) δ 7.71 and 7.70 (2 x s, 2 x 1H, 3’’), 7.42–7.33 (m, 4H, 3’), 7.25–7.17
20
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201901882
European Journal of Organic Chemistry
(m, 8H, 6’’, 2’, 4’), 5.54–5.36 (m, 4H, 4), 5.17–5.08 (m, 2H, 1), 4.20–4.12 (m, 4H, O-CH2-CH3), 3.90–
3.85 (m, 12H, 4’’-O-CH3, 5’’-O-CH3), 1.29–1.19 (m, 12H, O-CH2-CH3, 2); 31P NMR (202 MHz, 50%
CACO/[D6]DMSO, 25 °C) δ –7.64 and –7.70 (2 x s, 2 diastereoisomers); 13C {1H}NMR (126 MHz,
[D6]DMSO, 25 °C) δ 153.5 and 153.4 (4’’ or 5’’), 148.4 and 148.4 (5’’ or 4’’), 140.0 and 139.9 (2’’),
130.3 and 130.1 (3’), 126.4 and 125.7 (1’’), 125.7 and 125.5 (4’), 120.3 and 120.2 (2 x d, J2’-P = 5.0, J2’-P
= 4.5, 2’), 111.8–111.6 (m, 6’’), 108.5 (3’’), 72.5–72.4 (m, 1), 65.2–65.0 (m, O-CH2-CH3), 64.1 and 62.9
(4), 56.6 and 56.4 (4’’-O-CH3, 5’’-O-CH3), 20.6 (2), 16.1–16.0 (m, O-CH2-CH3). C3 and C1’ were not
detected; HRMS (ESI+) calcd for C20H24O10NNaP 492.1030 [M+Na]+, found 492.1027; HPLC (Rt) 10.74
min.
Characterization of the final products P (after irradiation experiments)
1
(Ethoxy(hydroxy)phosphoryl)glycine (2-P). H NMR (600 MHz, 50% CACO/[D6]DMSO, 25 °C) δ
3.62 (m, 2H, O-CH2-CH3), 3.13 (d, 2H, J2-NH = 6.9, 2), 1.04 (t, 3H, JCH3-CH2 = 7.1, JCH3-P = 0.3, O-CH2-CH3);
31P NMR (202 MHz, 50% CACO/[D6]DMSO, 25 °C) δ 6.99; 13C {1H} NMR (151 MHz, 50%
CACO/[D6]DMSO, 25 °C) δ 61.5 (d, JCH2-P = 4.5, O-CH2-CH3), 46.8 (2), 17.4 (d, JCH3-P = 7.5, O-CH2-CH3);
HRMS (ESI-) calcd for C4H9O5NP 182.0224 [M-H]-, found 182.0224.
(Ethoxy(hydroxy)phosphoryl)-L-valine (3-P). 1H NMR (500 MHz, 50% CACO/[D6]DMSO, 25 °C) δ
4.47 (s, 1H, NH), 3.59 (m, 2H, O-CH2-CH3), 3.07 (m, 2H, 2), 1.67 (m, 4), 1.01 (t, 3H, JCH3-CH2 = 7.2, O-CH2-
CH3), 0.75 and 0.70 (2 x d, 2 x 3H, J5-4 = 7.0, J5-4 = 7.0, 5); 31P NMR (202 MHz, 50% CACO/[D6]DMSO, 25
°C) δ 7.02; 13C {1H}NMR (126 MHz, 50% CACO/[D6]DMSO, 25 °C) δ 63.6 (2), 61.2 (d, JCH2-P = 5.0, O-CH2-
CH3), 33.6 (d, J4-P = 5.8, 4), 20.7 (5), 17.6 (d, JCH3-P = 7.7, O-CH2-CH3); HRMS (ESI-) calcd for C7H15O5NP
224.0693 [M-H]-, found 224.0691.
(Ethoxy(hydroxy)phosphoryl)-L-phenylalanine (4-P). 1H NMR (500 MHz, 50% CACO/[D6]DMSO,
25 °C) δ 7.18–7.09 (m, 5H, 2’’, 3’’, 4’’), 3.57–3.54 (m, 1H, 2), 3.49–3.36 (m, 2H, O-CH2-CH3), 2.77–2.73
(m, 2H, 2-CH2-1’’), 0.94 (t, 3H, JCH3-CH2 = 7.0, O-CH2-CH3); 31P NMR (202 MHz, 50% CACO/[D6]DMSO, 25
°C) δ 6.04; 13C {1H}NMR (126 MHz, 50% CACO/[D6]DMSO, 25 °C) δ 180.9 (d, J3-P = 4.4, 3), 140.3 (1’’),
21
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201901882
European Journal of Organic Chemistry
131.3 (2’’), 129.5 (3’’), 127.5 (4’’), 61.3 (d, JCH2-P = 5.5, O-CH2-CH3), 59.4 (2), 42.3 (2-CH2-1’’), 17.6 (d,
JCH3-P = 8.2, O-CH2-CH3); HRMS (ESI-) calcd for C11H15O5NP 272.0693 [M-H]-, found 272.0695.
(Ethoxy(hydroxy)phosphoryl)-L-proline (5-P). 1H NMR (500 MHz, 50% CACO/[D6]DMSO, 25 °C)
δ 7.33–7.28 (m, 4H, 3’), 7.15–7.06 (m, 6H, 2’, 4’), 4.09–3.95 (m, 4H, O-CH2-CH3), 3.25–3.07 (m, 4H, 5),
1.92–1.69 (m, 6H, 3, 4a), 1.56–1.51 (m, 2H, 4b), 1.17 and 1.12 (2 x t, 2 x 3H, JCH3-CH2 = 7.0, JCH3-CH2 = 7.0,
O-CH2-CH3); 31P NMR (202 MHz, 50% CACO/[D6]DMSO, 25 °C) δ 4.25 and 3.55 (2 x s, 2
diastereoisomers); 13C {1H}NMR (126 MHz, 50% CACO/[D6]DMSO, 25 °C) δ 180.9 and 180.8 (2 x d, J6-P
= 1.8, 6), 151.9 and 151.9 (2 x d, J1’-P = 3.4, J1’-P = 2.5, 1’), 131.5 (3’), 126.7 and 126.6 (4’), 121.8 and
121.5 (2 x d, J2’-P = 4.6, J2’-P = 4.5, 2’), 65.2 and 64.9 (2 x d, JCH2-P = 5.4, JCH2-P = 5.4, O-CH2-CH3), 64.4 and
63.6 (2 x d, J2-P = 6.3, J2-P = 4.9, 2), 49.6 and 48.3 ( 2 x d, J5-P = 5.3, J5-P = 5.2, 5), 33.3 and 33.0 (2 x d, J3-P
= 9.2, J3-P = 8.5, 3), 26.5 and 26.4 (2 x d, J4-P = 8.2, J4-P = 8.2, 4), 17.3 and 17.2 (2 x d, JCH3-P = 7.1, JCH3-P
6.4, O-CH2-CH3); HRMS (ESI-) calcd for C11H15O5NP 222.0537 [M-H]-, found 222.0538.
=
(Ethyl(2-carboxyethyl)phosphoramidate (6-P). Compound not characterized (trace amount
only after 9 weeks).31P NMR (202 MHz, 50% CACO/[D6]DMSO, 25 °C) δ 7.74.
2-((Ethoxy(hydroxy)phosphoryl)oxy)propanoic acid (7-P). 1H NMR (500 MHz, 50%
CACO/[D6]DMSO, 25 °C) δ 4.2 (s, 1H, 1), 3.66 (m, 2H, O-CH2-CH3), 1.20 (d, 3H, J2-1 = 6.8, 2), 1.04 (t, 3H,
JCH3-CH2 = 7.1, O-CH2-CH3); 31P NMR (202 MHz, 50% CACO/[D6]DMSO, 25 °C) δ –1.21; 13C NMR
(126 MHz, [D6]DMSO, 25 °C) δ 179.6 (d, J3-P = 6.7, 3), 73.4 (d, J1-P = 5.7, 1), 62.2 (d, JCH2-P = 5.4, O-CH2-
CH3), 21.6 (d, J2-P = 2.7, 2), 17.5 (d, JCH3-P = 7.6, O-CH2-CH3); HRMS (ESI-) calcd for C5H10O6P 197.0221,
found 197.0225.
Methods
NMR spectroscopy. NMR spectra were recorded on a Bruker Avance III spectrometer with a
broad-band cryo probe with ATM module (5 mm CPBBO BB-1H/19F/15N/D Z-GRD) operating at
1
500 MHz for H and 125.7 MHz for 13C, on a 600 MHz Bruker Avance III spectrometer with inverse
broad-band cryo probe with ATM module (5 mm BBI 1H/D-BB Z-GRD) or on a Bruker Avance IIITM HD
22
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201901882
European Journal of Organic Chemistry
400 MHz spectrometer equipped with Prodigy cryo probe with ATM module (5 mm CPBBO BB-
1H/19F/D Z-GRD). For NMR signal assignment, standard Bruker pulse sequences for 1D (1H, 13C-APT,
31P) as well as 2D (COSY, ROESY, HSQC, HMBC) NMR experiments at corrected temperature 25 °C
were employed. In situ structure determination of the intermediates was performed as recently
reported.[19] All NMR data were interpreted using Topspin 3.5. For reference, the following solvent
signals were used: [D6]DMSO: 2.50 (1H) and 39.7 (13C) ppm. The 31P NMR spectra were referenced to
H3PO4 with 0 ppm.[29]
For NMR experiments with in situ irradiation, a light emitting diode (LED; Thorlabs, Germany)
was used at 365 nm. The light was guided into the spectrometer, directly into the NMR tube via a
multimode silica optical fiber with 1 mm diameter, 0.39 NA, high amount of OH (Thorlabs, Germany).
Mass spectrometry. Mass spectra were measured on a LTQ Orbitrap XL (Thermo Fisher
Scientific) using electrospray ionization (ESI).
Photochemical study. Electronic absorption spectra were recorded on an Agilent 8453 diode
array spectrophotometer (Hewlett Packard, USA) in cacodylate buffer (0.1 M, pH = 7.4). All
quantitative photochemical measurements were performed at 25°C in the dark, with UV-glass
window 340 nm LED diodes (LED341W; Thorlabs; Figure S21) as light sources with optical power of
P=0.33 mW. Photochemical measurements were performed using the apparatus described
elsewhere (Figure 7 in the reference[30]) without ultrasonic horn H and lens L1 and using an Ocean
Optics HR4000CG-UV-NIR spectrophotometer for initial and transmitted light intensity
determination. The light sources were three 340 nm LED diodes Thorlabs with overall incident
photon flux I0=1.320.02
10-6 mol s-1 dm-3. The actual concentration of the corresponding
phosphoramidate in freshly prepared air-saturated solutions during irradiation in a 1 cm quartz
fluorescence cuvette was determined by HPLC. HPLC analyses were performed using Agilent 1100
series HPLC system under the following conditions: LiChrospher 100, RP-18 e (5m), 4 mm  250
mm, 55% MeCN in water containing 0.1% phosphoric acid, 1.0 mL/min, 350 nm. Determination of
23
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201901882
European Journal of Organic Chemistry
the photocleavage quantum yield is described in more detail in the SI, section Determination of
photochemical quantum yields.
Acknowledgments
This work was funded by Experientia Start-Up grant (SG-2018-1, O.B.) and by the University Scientific
Park of Comenius University in Bratislava (ITMS 26240220086, M.C.). We would like to thank Carlos
V. Melo for proofreading the manuscript.
References
[1]
a) M. A. Kasper, M. Glanz, A. Oder, P. Schmieder, J. P. von Kries, C. P. R. Hackenberger, Chem.
Sci. 2019, 10 (25), 6322–6329; b) R. L. Hanna, S. M. Gryaznov, J. A. Doudna, Chem. Biol. 2000,
7 (11), 845–854.
[2]
[3]
[4]
J. A. Wiemer, F. D. Wiemer, Top. Curr. Chem. 2015, 360, 115–160.
A. Hauser, M. Penkert, C. P. R. Hackenberger, Acc. Chem. Res. 2017, 50 (8), 1883–1893.
a) S. K. Miroshnichenko, O. A. Patutina, E. A. Burakova, B. P. Chelobanov, A. A. Fokina, V. V.
Vlassov, S. Altman, M. A. Zenkova, D. A. Stetsenko, Proc. Natl. Acad. Sci. 2019, 116 (4), 1229–
1234; b) M. Faria, D. G. Spiller, C. Dubertret, J. S. Nelson, M. R. H. White, D. Scherman, C.
Hélène, C. Giovannangeli, Nat. Biotechnol. 2001, 19 (1), 40–44.
[5]
[6]
W. Wang, C. Alexander, Angew. Chemie - Int. Ed. 2008, 47 (41), 7804–7806.
a) A. Cankaya, M. Steinmann, Y. Bülbül, I. Lieberwirth, F. R. Wurm, Polym. Chem. 2016, 7 (31),
5004–5010; b) Z. Todorova, N. Koseva, I. Ugrinova, K. Troev, J. Polym. Sci. Part A Polym. Chem.
2017, 55 (10), 1730–1741.
[7]
[8]
[9]
V. L. Robinson, A. M. Stock, Structure 1999, 7 (3), 47–53.
H. Jung, S. H. Shin, J. M. Kee, ChemBioChem 2019, 20 (5), 623–633.
a) R. Walther, J. Rautio, A. N. Zelikin, Adv. Drug Deliv. Rev. 2017, 118, 65–77. b) M. Slusarczyk,
M. H. Lopez, J. Balzarini, M. Mason, W. G. Jiang, S. Blagden, E. Thompson, E. Ghazaly, C.
McGuigan, J. Med. Chem. 2014, 57 (4), 1531–1542.
24
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201901882
European Journal of Organic Chemistry
[10] a) M. Derudas, D. Carta, A. Brancale, C. Vanpouille, A. Lisco, L. Margolis, J. Balzarini, C.
McGuigan, J. Med. Chem. 2009, 52 (17), 5520–5530; b) C. McGuigan, P. Murziani, M.
Slusarczyk, B. Gonczy, J. V. Voorde, S. Liekens, J. Balzarini, J. Med. Chem. 2011, 54 (20), 7247–
7258.
[11] C. McGuigan, S. A. Harris, S. M. Daluge, K. S. Gudmundsson, W. E. McLean, T. C. Burnette, H.
Marr, R. Hazen, L. D. Condreay, L. Johnson, E. De Clerck, J. Balzarini, J. Med. Chem. 2005, 48
(10), 3504–3515.
[12] a) C. McGuigan, A. Hassan-Abdallah, S. Srinivasan, Y. Wang, A. Siddiqui, S. M. Daluge, K. S.
Gudmundsson, H. Zhou, E. W. McLean, J. P. Peckham, T. C. Burnette, H. Marr, R. Hazen, L. D.
Condreay, L. Johnson, J. Balzarini, J. Med. Chem. 2006, 49 (24), 7215–7226; b) C. McGuigan, D.
Cahard, H. M. Sheeka, E. De Clercq, J. Balzarini, J. Med. Chem. 1996, 39 (8), 1748–1753; c) D.
Saboulard, L. Naesens, D. Cahard, A. Salgado, R. Pathirana, S. Velazquez, C. McGuigan, E. De
Clercq, J. Balzarini, Mol. Pharmacol. 1999, 56 (4), 693–704.
[13] a) Y. Mehellou, J. Balzarini, C. McGuigan, ChemMedChem 2009, 4 (11), 1779–1791; b) M.
Slusarczyk, M. Serpi, F. Pertusati, Antiviral Chem. Chemother. 2018, 26, 1-31.
[14] G. Birkus, R. Wang, X. Liu, N. Kutty, H. MacArthur, T. Cihlar, C. Gibbs, S. Swaminathan, W. Lee,
M. McDermott, Antimicrob. Agents Chemother. 2007, 51 (2), 543–550.
[15] A. Alouane, R. Labruère, T. Le Saux, F. Schmidt, L. Jullien, Angew. Chemie - Int. Ed. 2015, 54
(26), 7492–7509.
[16] A. Aubertin, J. Imbach, Bioorg. Med. Chem. Lett. 1998, 8, 1045–1050.
[17] H. Hřebabecký, E. Procházková, M. Šála, P. Plačková, E. Tloušťová, O. Barauskas, Y.-J. Lee, Y.
Tian, R. Mackman, R. Nencka, Org. Biomol. Chem. 2015, 13 (35), 9300–9313.
[18] a) Y. Mehellou, H. S. Rattan, J. Balzarini, J. Med. Chem. 2018, 61 (6), 2221-2226; b) A. S.
Alanazi, E. James, Y. Mehellou, ACS Med. Chem. Lett. 2019, 10 (1), 2–5.
[19] E. Procházková, R. Navrátil, Z. Janeba, J. Roithová, O. Baszczyňski, Org. Biomol. Chem. 2019, 17
(2), 315–320.
25
This article is protected by copyright. All rights reserved.

10.1002/ejoc.201901882
European Journal of Organic Chemistry
[20] I. Aujard, C. Benbrahim, M. Gouget, O. Ruel, J. B. Baudin, P. Neveu, L. Jullien, Chem. - A Eur. J.
2006, 12 (26), 6865–6879.
[21] R. W. Clark, T. M. Deaton, Y. Zhang, M. I. Moore, S. L. Wiskur, Org. Lett. 2013, 15 (24), 6132–
6135.
[22] a) A. S. Ray, M. W. Fordyce, M. J. M. Hitchcock, Antiviral Res. 2016, 125, 63–70; b) T. K.
Warren, R. Jordan, M. K. Lo, A. S. Ray, R. L. Mackman, V. Soloveva, D. Siegel, M. Perron, R.
Bannister, H. C. Hui, et al. Nature 2016, 531 (7594), 381–385.
[23] R. C. Neuman, W. Kauzmann, A. Zipp, J. Phys. Chem. 1973, 77 (22), 2687–2691.
[24] V. S. Stoll, J. S. Blanchard, Methods Enzymol. 1990, 182, 24–38.
[25] P. Klán, T. Šolomek, C. G. Bochet, A. Blanc, R. Givens, M. Rubina, V. Popik, A. Kostikov, J. Wirz,
Chem. Rev. 2013, 113 (1), 119–191.
[26] T. Šolomek, S. Mercier, T. Bally, C. G. Bochet, Photochem. Photobiol. Sci. 2012, 11 (3), 548.
[27] M. E. Jung, G. Piizzi, Chem. Rev. 2005, 105 (5), 1735–1766.
[28] a) J. Ying, R. Lin, P. Xu, Y. Wu, Y. Liu, Y. Zhao, Sci. Rep. 2018, 8 (1), 936; b) A. Ricci, A. Brancale,
J. Comput. Chem. 2012, 33 (10), 1029–1037; c) M. Karki, C. Gibard, S. Bhowmik, R.
Krishnamurthy, Life 2017, 7 (3), 32.
[29] R. F. Fulmer, A. J. M. Miller, N. H. Sherden, H. E. Gottlieb, A. Nudelman, B. M. Stoltz, J. E.
Bercaw, K. I. Goldberg, Organometallics 2010, 29 (9), 2176–2179.
[30] A. Gáplovský, Š. Toma, J. Donovalová, J. Photochem. Photobiol. A Chem. 2007, 191 (2–3), 162–
166.
26
This article is protected by copyright. All rights reserved.