Prodrug design for the potent cardiovascular agent Nω- hydroxy-l-arginine (NOHA): Synthetic approaches and physicochemical characterization

Article abstract of DOI:10.1039/c0ob01117g

N^ω-Hydroxy-l-arginine (NOHA) - the physiological nitric oxide precursor - is the intermediate of NO synthase (NOS) catalysis. Besides the important fact of releasing NO mainly at the NOS-side of action, NOHA also represents a potent inhibitor of arginases, making it an ideal therapeutic tool to treat cardiovascular diseases that are associated with endothelial dysfunction. Here, we describe an approach to impart NOHA drug-like properties, particularly by wrapping up the chemically and metabolically instable N-hydroxyguanidine moiety with different prodrug groups. We present synthetic routes that deliver several more or less highly substituted NOHA derivatives in excellent yields. Versatile prodrug strategies were realized, including novel concepts of bioactivation. Prodrug candidates were primarily investigated regarding their hydrolytic and oxidative stabilities. Within the scope of this work, we essentially present the first prodrug approaches for an interesting pharmacophoric moiety, i.e., N-hydroxyguanidine. The Royal Society of Chemistry 2011.

Full text of DOI:10.1039/c0ob01117g

View Article Online / Journal Homepage / Table of Contents for this issue
Organic &
Biomolecular
Chemistry
Dynamic Article Links
Cite this: Org. Biomol. Chem., 2011, 9, 5249
www.rsc.org/obc
PAPER
Prodrug design for the potent cardiovascular agent Nx-hydroxy-L-arginine
(NOHA): Synthetic approaches and physicochemical characterization†
Dennis Schade,*‡ Ju¨rke Kotthaus, Nikola Klein, Joscha Kotthaus and Bernd Clement*
Received 6th December 2010, Accepted 20th April 2011
DOI: 10.1039/c0ob01117g
Nw-Hydroxy-L-arginine (NOHA)—the physiological nitric oxide precursor—is the intermediate of NO
synthase (NOS) catalysis. Besides the important fact of releasing NO mainly at the NOS-side of action,
NOHA also represents a potent inhibitor of arginases, making it an ideal therapeutic tool to treat
cardiovascular diseases that are associated with endothelial dysfunction. Here, we describe an approach
to impart NOHA drug-like properties, particularly by wrapping up the chemically and metabolically
instable N-hydroxyguanidine moiety with different prodrug groups. We present synthetic routes that
deliver several more or less highly substituted NOHA derivatives in excellent yields. Versatile prodrug
strategies were realized, including novel concepts of bioactivation. Prodrug candidates were primarily
investigated regarding their hydrolytic and oxidative stabilities. Within the scope of this work, we
essentially present the first prodrug approaches for an interesting pharmacophoric moiety, i.e.,
N-hydroxyguanidine.
NO levels, with the nitric oxide synthase (NOS) representing the
central enzyme of NO formation.
NOSs catalyze the 5-electron oxidation of L-arginine
Introduction
Nitric oxide (NO) is involved in a myriad of (patho)physiological
processes.1,2 An impaired NO bioavailability is well-recognized in
the pathology of endothelial dysfunction and atherosclerosis, and
therefore is relevant for many cardiovascular diseases.3 To date,
to L-citrulline and NO via intermediate formation of
Nw-hydroxy-L--arginine (NOHA).5 Although NOS is a desirable
target, the NOS active site has demonstrated very little promiscu-
multiple up- and downstream targets have been identified that are
ity, and to date few alternative substrates or modulators have been
identified.4 We consider the physiological N-hydroxyguanidine
available for pharmacologic intervention within the NO/cGMP
system. Among these, new developments suggest a paradigm
NOHA the ideal NOS substrate for therapeutic applications.
shift from purely symptomatic to more causative therapeutic ap-
Besides the important fact of releasing NO mainly at the NOS-site
proaches. Interference with the NO generating/regulating system
of action, NOHA also represents a potent inhibitor of arginases,
opens up opportunities to not only treat ischemic heart disease
with the latter being well recognized as contributing to the
development of endothelial dysfunction.6
but many more diseases that are characterized by endothelial
dysfunction.4 It is highly desirable to more selectively affect
pathologically altered processes of NO generation and availability,
thereby providing a safe and effective therapy. As a consequence,
a great deal of research has focused on modulation of the
predominant enzymes involved in the regulation of endogenous
Although NOHA can serve as a substrate for NOS and
thereby stimulate NO generation, direct therapeutic application is
prohibited by its little drug-like character. The major problem to be
solved is NOHA’s deleterious chemical and metabolic properties
(see Fig. 1).
N-Hydroxyguanidines in general are thermolabile as they are
already unstable at room temperature and should be stored at
-20 ◦C. Usually, they are most stable in the form of salts of
strong acids and are labile towards bases.7,8 N-Hydroxyguanidines
undergo two and three electron oxidative degradation reactions,
with different oxidants leading to distinct products, such as
cyanamides or urea derivatives.9–14 Furthermore, NOHA and other
N-hydroxyguanidines are known to be readily metabolized by the
N-reductive biotransformation pathway.15–19
Department of Pharmaceutical Chemistry, Pharmaceutical Institute,
Christian-Albrechts-University of Kiel, Gutenbergstraße 76-78, D-
24118 Kiel, Germany. E-mail: dschade@sanfordburnham.org, bclement@
pharmazie.uni-kiel.de; Fax: +49-880-1352; Tel: +49-880-1126
† Electronic supplementary information (ESI) available: Complete syn-
thetic protocols and spectral data of all synthesized compounds (including
intermediates); synthetic details for attempts towards O-acylated NOHA
prodrugs; determination of enantiomeric purity by HPLC; in vitro bioac-
tivation of 13; detailed hydrolytic stability graphs of selected prodrugs. See
DOI: 10.1039/c0ob01117g
Although the formed metabolite L-arginine represents the actual
substrate for NOSs and may be considered an active metabolite
from this N-reduction, it would be favorable to “smuggle” the
‡ Current address: Sanford-Burnham Medical Research Institute and
Human BioMolecular Research Institute, 5310 Eastgate Mall, San Diego,
CA 92121, USA. Fax: +1-858-459-9311; Tel: +1-858-459-9305
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 5249–5259 | 5249
©

View Article Online
employs a concept that has been first described by Linton
et al. for the synthesis of guanidines.24 Martin et al. could
successfully apply this strategy to the preparation of NOHA and
Nw-amino-L-arginines.25,26 Essentially, a highly substituted guani-
dine (or N-hydroxyguanidine) is built up by reacting a car-
bamoylated thiourea with the desired amine or hydroxylamine
in the presence of a desulfurizing agent, such as EDCI. This
strategy provided an attractive basis for the preparation of
the herein described NOHA derivatives as it avoids direct
O-substitution of the N-hydroxyguanidine group. Direct substi-
tution was expected to be hampered by the unstable character of
the N-hydroxyguanidine as rather harsh reaction conditions are
required.
Our general synthetic approach to a diverse set of protected
NOHA derivatives is outlined in Scheme 1. Thioureas 2 could
be conveniently obtained by reaction of the protected L-ornithine
1 with the respective isothiocyanate. That way, either a simple
protecting group (Cbz) could be inserted or a prodrug group (Eoc)
that was supposed to be retained for some of the final molecules.
Fig. 1 Chemical and metabolic degradation of N-hydroxyguanidines.
direct NO-precursor NOHA through the first-pass effect passage
without having significant amounts of L-arginine formed. The
background is that L-arginine underlies a plethora of metabolic
pathways,20 whereas NOHA physiologically only serves NO
production, as is so far known.
To circumvent the above stability issues associated with NOHA,
we decided to synthetically pursue chemically stabilized NOHA
analogues. O- and N-substitutions were designed which may
suppress radical formation, abstraction of protons, or nucle-
ophilic attack at the hydroxyguanidine-carbon. Concurrently, such
modifications were expected to increase membrane permeation
processes, either active or passive ones. Synthetic feasibility for
novel NOHA prodrug candidates and model compounds was
explored, along with their physicochemical stabilities.
Synthesis of O-alkylated prodrugs
One approach to increase the stability of N-hydroxyguanidines
was by suitable O-alkylations. A similar approach has already
been described for amidoximes, such as pafuramidine (DB289), an
O-methylated amidoxime with antiprotozoal activity.27 The pri-
mary aim of this prodrug concept was to increase oral bioavail-
ability and permeation of the blood-brain barrier rather than
improving its overall stability.
Our first attempts to prepare such O-methylated NOHA
prodrugs used the corresponding cyanamide and methoxylamine
as starting materials in accordance with a literature protocol.28 But
even after optimization of this reaction (80% yield), the general
applicability to other derivatives remained limited.
Results and discussion
Using the outlined synthetic scheme (see Scheme 1), we were
not only able to further increase the overall yields but also gained
access to quite diversely substituted derivatives (Fig. 2). Thus, we
could realize free amino acid prodrug candidates 8a and 8c as
well as the corresponding ethylesters 8b and 8d. For complete
deprotection, 6a was treated with TFA/thioanisole. For esterifi-
cation, dry HCl gas was care◦fully bubbled through an ethanolic
General synthetic concept
Within the course of our ongoing studies with various
Nd- and Nw-substituted L-arginines,21,22 we have already reported
on a synthetic strategy that allowed access to Nw-substituted
L-arginines.23 Among these, one O-methylated NOHA derivative
could be prepared in excellent yields. The underlying reaction
◦
solution of 8a (or 8c) at -10 C, which was then left at 4 C for
Fig. 2 Synthesis of O-alkylated prodrugs of NOHA. Reagents and conditions: (a) TFA/thioanisole, 30 min, r.t.; (b) HCl(g) in EtOH, 1 h at -10 ◦C, 36 h
at 4–8 ◦C; (c) TFA, 30 min at 0 ◦C, 3 h at r.t.
5250 | Org. Biomol. Chem., 2011, 9, 5249–5259
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
Scheme 1 General synthetic approach to highly substituted NOHA prodrugs. Reagents and conditions: (a) in CH2Cl2, r.t., 2 h; (b) DIPEA, EDCI, r.t.,
24–48 h, in CH2Cl2; (c) for details see Fig. 2 and 3. a Overall isolated yield, starting from 1.
1.5 days until the reaction was complete. The desired ethylesters
8b/d were obtained in high yield and purity after lyophilization.
TFA was simultaneously esterified by these procedures, and the
ester was subsequently removed in vacuo. In addition, with 8c and
8d, Eoc could be introduced in the molecules as a carbamate-based
prodrug moiety. For chemoselective deprotection of the a-amino
acid group in 7a, thioanisole was omitted and dry conditions
guaranteed to prevent loss of the Eoc moiety. The product 8c could
be conveniently purified by flash chromatography on RP-18 silica
gel. With 8c the N-hydroxyguanidine group was already highly
substituted, but a further esterification of the carboxy function
to 8d was expected to increase the lipophilicity of the molecule
and thereby its general tendency to pass biological membranes by
passive diffusion processes.
Therefore, we envisioned a similar O-alkyl prodrug ap-
proach for NOHA, utilizing a novel mechanism of bioactiva-
tion via action of the peptidylglycine a-amidating monooxy-
genase (PAM).30 PAM exhibits a broad substrate specificity
and recognizes many glycine-like structures, including N-,
O- and S-carboxymethylated compounds.31 We previously
demonstrated that O-carboxymethylated amidoximes and simple
N-hydroxyguanidines are accepted as substrates by PAM.30 For
the synthesis of the respective O-carboxymethylated NOHA
prodrugs 9a–c, we essentially applied the same chemistry as
described for the O-methyl prodrugs 8a–d, but using aminooxy-
acetic acid derivatives 4 as the hydroxylamine component (see
Scheme 1). As an exception, the synthesis of the fully deprotected
Nw-carboxymethoxy-L-arginine (9a) (as a reference compound
for stability and pharmacokinetic studies) had to be slightly
altered. The O-carboxymethylester was unexpectedly stable to-
wards TFA/thioanisole treatment, and complete deprotection
required 5 h incubation at 50–60 ◦C with 6 N aqueous HCl.
The optical purity of 9a was assessed by chromatography on
One possible disadvantage of simple O-methylation may be
seen in the involvement of cytochrome P450 enzymes in the
bioactivation cascade as has been shown for DB289.29 The con-
sequence might be an interaction potential with other substances
metabolized by the respective enzyme system.
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 5249–5259 | 5251
©

View Article Online
a Crownpak Cr(+) column and was preserved in this synthetic
sequence (>99%ee). Several other O-alkylated NOHA prodrugs
were also analyzed on chiral stationary phase and shown to be
pure by >99%ee (see the ESI†).
galactose conjugates 7c and 12 in very good yields (70% and 90%,
see Fig. 3).
Further deprotection of 12 was achieved with a strong anion
exchange resin to afford the deacetylated b-D-galactose conjugate
13 as the model compound (see B in Fig. 3). On the basis of NOHA
derivative 7c we achieved chemoselective deprotection of the
a-amino acid moiety without affecting the acid-labile glycosidic
group by careful bubbling of HCl gas through an absolute diethyl
ether solution, after which the amino acid 10 precipitated as a
hydrochloride salt (see A in Fig. 3). To avoid an autocatalyzed
hydrolysis of this hydrochloride upon contact with water, we
attempted isolation of the respective sodium salt by dissolving the
powder in 0.5 M sodium bicarbonate and subsequent purification
on reversed-phase (RP-18) silica gel. Using this method, the
desired O-galactosyl NOHA prodrug 10 could be prepared in
good yields in its peracetylated form. Compound 10 represented
the preferred form as a prodrug candidate due to its higher stability
towards presystemic inactivation by chemical and/or enzymatic
degradation.
Synthesis of O-acetalic prodrugs
Considering that O-alkylations should contribute most to the
overall stability of the N-hydroxyguanidine moiety, another modi-
fication of this kind was realized. Inclusion of the N-hydroxy group
into an acetal was expected to afford stable hydroxyguanidine
prodrugs that can be bioactivated by monooxygenase-independent
hydrolytic mechanisms.
Especially, the incorporation into a carbohydrate-based acetal
was deemed attractive as (1) they could be bioactivated by gly-
cosidases in addition to slow chemical hydrolysis, (2) the released
monosaccharides are rather uncritical from a toxicological point
of view, and (3) the conjugate might be taken up into cells by
facilitated transport.
This approach was particularly inspired by the work of Melisi
et al. who used galactosyl-conjugates of D- and L-arginine
(attached at the a-amino acid group) as a prodrug principle to
increase permeation in, and thereby NO release within, cancer
cells.32 Other similar examples in the field of NO-donors are
carbohydrate conjugates of the diazeniumdiolates (NONOates)
and S-nitrosothiols, that were successfully optimized in terms of
stability and pharmacokinetics.33,34
Because such strategies were not previously applied to
N-hydroxyguanidines, we also synthesized a simple model com-
pound for in vitro characterization purposes. The only described
preparation of an O-glycosylic hydroxyguanidine was reported by
our group for the N-hydroxylated metabolite of the antihypertonic
agent debrisoquine.16 However, the actual glycoside formation
was attempted by direct substitution of the N-hydroxy function
using the respective sugar-1-bromide, with resulting yields of
<1%. Following the synthetic strategy described here, we used the
1-aminooxyl derivative of peracetylated b-D-galactose (5, prepared
according to a modified literature procedure)35 in order to build
up the glycosylated N-hydroxyguanidine under mild conditions.
Importantly, the b-configuration is required for the targeted
galactosidases.36 Using this strategy, we could prepare the desired
Synthesis of O-acylated prodrugs
A final set of NOHA prodrugs was envisioned that was solely
comprised of carboxylic ester- and carbamate-based prodrug
groups. In terms of bioactivation, this group of compounds was
most appealing as esterases are ubiquitously present and reveal
high catalytic activities as well as broad substrate specificities.37
However, we faced substantial problems with the preparation of
these NOHA derivatives (see Scheme 2).
Initial attempts to apply the above described synthetic strat-
egy (see Scheme 1) failed since, in our hands, the required
O-acylated hydroxylamines could not be prepared according to
a literature procedure.38 We were unable to deprotect several
O-acylated ethyl imidates using HCl in different solvents, such
as nitromethane, diethyl ether, or 1,4-dioxane. They were not
commercially available, and thus, other approaches were pursued.
A second strategy was to acylate the protected NOHA deriva-
tive 14 (see A in Scheme 2). Compound 14 was prepared in
two steps according to a literature protocol.39 However, despite
efforts exploring several different solvent systems under various
conditions with acid chlorides and with anhydrides, none of the
Fig. 3 Synthesis of O-galactosylated N-hydroxyguanidines. Reagents and conditions: (a) DIPEA, EDCI, r.t., 24 h, in CH2Cl2; (b) basic anion exchanger,
MeOH, r.t., 24 h; (c) Et2O, -15 ◦C, argon, HCl(g) for 5 min, 4–8 ◦C overnight; (d) 0.5 M NaHCO3, RP-18 flash chromatography.
5252 | Org. Biomol. Chem., 2011, 9, 5249–5259
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
different conditions. Other attempts included the preparation of
O-THP-protected, Nw-Eoc-substituted NOHA (and simple model
compound analogs) in order to induce cyclization (elimination of
EtOH) after in situ chemoselective deprotection of the THP group
(see the ESI†).
The third strategy was to prepare an agent that should transfer
an O-acylamidoxime moiety (19) or a 1,2,4-oxadiazol-5-one group
(20) to the side-chain amine of protected L-ornithine 1 to afford
the respective O-acylhydroxyguanidine 16 or cyclic carbamate 15,
respectively (see B in Scheme 2). According to the various pio-
neering applications described by the Katritzky group, we deemed
benzotriazole to be a suitable leaving group for this purpose.
In fact, Katritzky et al. have described similar applications to
the synthesis of amidoximes and N-hydroxyguanidines.44,45 After
slight modifications of the literature protocols we could prepare
N¢-hydroxy-1H-1,2,3-benzotriazole-1-carboximidamide (18) and
from there, different O-acyl derivatives (19; acetyl, pivaloyl and
ethoxycarbonyl) in good yields. However, neither under conven-
tional nor microwave conditions did these reagents react with
1 to give the desired products 16. The only detected reaction
was elimination of the hydroxylamine with replacement by the
amine (i.e., the protected L-ornithine 1), as demonstrated by TLC
and LC/MS analysis of crude products. Moreover, reagent 20
could not be prepared using CDI, ethylchloroformate, phosgene
or phosgene-like reagents.
Scheme 2 Attempts to prepare O-acyl type prodrugs of NOHA. Reagents
and conditions: (a) carbonyldiimidazole (CDI) or ethylchloroformate, in
THF, DMF or EtOH; or via O-THP protected Eoc-substituted 14 (see
the ESI†); (b) acid chlorides or anhydrides, in dry DCM, Et2O, pyridine,
temperatures from -30 ◦C to r.t.; (c) NH2OH (free base), in dioxane,
60 ◦C, 2 h; (d) acid chloride, TEA, CH2Cl2; (e) conventional or microwave
conditions, from r.t. up to 100 ◦C, neat or in CH2Cl2, THF or EtOH;
(f) carbonyldiimidazole (CDI) or ethylchloroformate in THF, DMF or
EtOH, or phosgene in toluene. Experimental data for these reactions are
found in the ESI.†
In vitro characterization
Hydrolytic stability. As indicated in the introduction,
N-hydroxyguanidines are particularly susceptible to base-
catalyzed hydrolysis. Therefore, the pH stability of different
NOHA prodrugs was assessed at 37 ◦C for up to 24 h, and the
mixtures were analyzed by HPLC (see Fig. 4, A). A more detailed
analysis including the analysis of samples at several time points
over the 24 h incubation is included in the ESI.†
When comparing the most simple (i.e., least substituted)
prodrugs 8a and 9a with NOHA, it is apparent that O-alkylations
improve the stability at pH 9. The O-methylated Nw-carbamate
prodrug 8c also exhibits increased stability at this pH suggesting
that Nw-substitution does not have deleterious effects on stability
but rather improves it. Surprisingly, NOHA underwent only about
10% degradation after 24 h at 37 ◦C at physiological pH (i.e., 7.4).
NOHA prodrugs 8b and 9c revealed an expected hydrolytic
degradation due to hydrolysis of the a-amino acid ester. Interest-
ingly, this ester appears to be more sensitive towards hydrolysis
than the ethyloxycarbonylmethoxy group in 9b.
NOHA prodrug 10 was not included in this stability study, as the
presence of four acetyl groups was expected to overly complicate
interpretation of the data. Instead, the deacetylated b-D-galactose-
conjugate 13 was used, allowing for focused analysis of the stability
of the sugar-acetal function. Unfortunately, we were not yet
able to isolate the corresponding deacetylated NOHA prodrug
in sufficient purity for the herein presented stability studies.
Strikingly, 13 proved very stable at all three pHs. Considering
the actual lability of acetals towards acidic media, the stability of
13 at pH 2 was quite surprising and might be due to steric as well
as electronic effects.
compound 16 O-acyl derivatives could be prepared. We usually
obtained complex product mixtures instead. One persistent side re-
action was the incidental N-acylation of the N-hydroxyguanidine
moiety, even when using the less reactive pivalic acid chloride.
N-Acetylation occurred after formation of a mixed anhydride with
acetic acid, which was present in the starting material 14 after
chromatographic workup. Furthermore, no significant amounts
of the desired O-acyl derivatives could be obtained from the free
base form of 14, which was found to be unstable. This is consistent
with a study by Belzecki et al. concerning peracetylations of
similar N-hydroxyguanidines, which states that successful selective
O-acetylation was only achieved for very few highly substituted
N-hydroxyguanidines.40–42
The cyclic carbamate (1,2,4-oxadiazol-5-one) of NOHA (15)
represents an extension of the “O-acyl” prodrug principle for
NOHA and was based on a similar compound described by Rehse
et al.43 However, their NOHA-based oxadiazolone turned out to
be inactive as a NO donor, a characteristic likely attributable to
the chemically and metabolically highly stable Na-acetyl group. In
addition, the racemic nature of their compound eventually lowers
efficacy 2-fold. We therefore wanted to circumvent a synthetic
protocol that delivers racemic material or Na-substituted deriva-
tives. To date, our attempts to cyclize 14 to afford the respective
cyclic carbamate 15 failed, using CDI or ethylchloroformate under
Oxidative stability. We selected a set of distinct oxidants that
have been reported in the literature to be capable of oxidizing
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 5249–5259 | 5253
©

View Article Online
to inhibit peroxynitrite-induced oxidation.48 NOHA showed little
sensitivity to H2O2.
Simple O-alkylation resulted in a remarkable increase of
stability as can be seen for prodrugs 8a and 9a. An additional
Nw-carbamate group (see data for 8c) provides a similarly stable
prodrug. This is also observed for 9b but with an increased lability
-
towards ONO2 . Data for 8b and 9c should be carefully interpreted
considering their general stability profile without addition of
oxidants, i.e., at pH 7–9 for 24 h at 37 ◦C (see A in Fig. 4). Only
8b turned out exceptionally labile in incubations with cupric ions,
a phenomenon that does not parallel with other results of this
study (e.g., 8a as an O-methyl analog, or 9c as a Nw-carbamate
analog). The apparent stability of galactose-conjugate 13 towards
all examined oxidants is remarkable. This result again underscores
the utility of this prodrug concept for N-hydroxyguanidines.
It is apparent that O-substitution (alkyl-, acetal-type) efficiently
protects from oxidation. Several groups have suggested that initial
formation of an iminoxy radical (IIIa/b, and the less favored
C-centered radical IIIc) is key to oxidative degradation reactions
of N-hydroxyguanidines (see Fig. 5).9–11
Fig. 4 In vitro stability studies. (A) pH stability: 250 mM compounds
in 10 mM potassium phosphate buffer pH 2, pH 7.4 and pH 9 after
incubation at 37 ◦C for 24 h (mean value of at least three independent
experiments SD); (B) oxidative stability: 250 mM compounds in 10 mM
phosphate buffer pH 7.4 and 500 mM potassium hexacyanoferrate(III),
500 mM copper(II) acetate, 500 mM peroxynitrite or 29 mM hydrogen
peroxide at 37 ◦C for 24 h (mean value of at least three independent
experiments SD); control = concentration at t = 0 h.
Fig. 5 First step of N-hydroxyguanidine oxidation according to litera-
ture-proposed mechanisms.
N-hydroxyguanidines, with a special focus on chemicals that are
of significance in NO-related biochemistry. Because oxidation
of NOHA by these compounds has been proposed to lead to
a variety of products, we opted to monitor the stability of the
parent compound, rather than formation of products. Hydrogen
Our data therefore suggest formation of this iminoxy radical
is prohibited by O-substitution. Cleavage of the O–H bond in
NOHA or N-hydroxyguanidines is an expected requisite for
formation of the stabilized radical (III), and proton abstraction
is greatly limited by O-substitutions. Although IIIb does represent
a possible radical intermediate for an O-alkyl/acetalic NOHA
prodrug, it lacks the possibility of tautomeric stabilization.
Stabilization of this radical would require dimerization, or re-
action with other radicals, such as NO.12 Since cleavage of the
O–C bond is thermodynamically much less favorable, it effectively
protects from this oxidation. Moreover, prodrug candidates with
an electron-withdrawing substituent (such as Eoc) exhibit a lower
electron density rendering them less prone to initial 1e- oxidation
(preferably at the double-bonded nitrogen).
Another consequence of O-substitution is that the
N-hydroxyguanidine group has no chance any more to chelate
Cu(II) and Fe(III). Especially for cupric ions, this has been
discussed as the mechanistic basis for its oxidative potency since
the calculated redox potentials for N-hydroxyguanidines present
in solution make Cu(II)-mediated reduction thermodynamically
prohibitive.12
-
peroxide (H2O2) and peroxynitrite (ONO2 ) were used to mimic
oxidative stress conditions. Elevated levels of such reactive oxygen
species (ROS) are especially of interest in the targeted tissues/cells
(e.g., endothelium) under pathophysiological conditions, and are
known to affect the bioavailability of NO.46,47 Copper and iron
were also used as oxidants, as two- and three-valent metal
ions that are considered to contribute to the metabolic fate of
NOHA and of N-hydroxyguanidines in general. Normally, metal
metabolism is highly regulated. However, NO has been shown
to elicit metal ions from metalloproteins and to disrupt metal
metabolism (see Cho et al. and references therein).12 The Fukuto
lab pointed at the physiological relevance of interactions between
N-hydroxyguanidines and cupric ions, providing comprehensive
insights into the underlying reactions and mechanisms.12 As an
example of three-valent ions, iron(III) was used in the form of its
ferricyanide complex (K3[Fe(CN)6]). This oxidant has also been
reported in the literature to oxidize N-hydroxyguanidines.13
The oxidative stability of NOHA and its different prodrugs was
monitored over 24 h at 37 ◦C by HPLC (Fig. 4). Under the tested
conditions, NOHA exhibited large sensitivity towards copper(II)
acetate and ferricyanide, but was moderately insensitive towards
Conclusions
Our ultimate goal is to establish a prodrug concept for NOHA
that ensures its proper bioavailability, and thereby, sufficient NO
-
ONO2 . In fact, N-hydroxyguanidines have been demonstrated
5254 | Org. Biomol. Chem., 2011, 9, 5249–5259
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
levels in pathophysiologically altered situations that are associated
with endothelial dysfunction. We postulate that physiological
conditions could be ideally mimicked with a suitable prodrug
candidate that ensures a prolonged and sustained release of
NOHA.
spectrometer BioApex II equipped with an ESI-ion source (Ag-
ilent); substances were dissolved in 60% acetonitrile in water
and introduced by flow injection analysis. Elemental analyses
were performed on a CHNS analysator (HEKAtech GmbH).
Reactions were monitored by TLC on precoated silica gel plates
(SiO2 60, F254). All compounds could be visualized by either UV
detection, ninhydrine spray or Ce(IV)-sulfate staining and heating
at 120 ◦C. Purification of synthesized compounds was performed
by column chromatography using silica gel (particle size 40–
Here, we have demonstrated synthetic routes to diversely
substituted NOHA derivatives. The overall yields are very good
for O-alkylated compounds 8 and 9 in this 3–4 step sequence (82–
89%), and reasonable for galactose conjugate 10 (49%). The se-
lected prodrug groups addressed different pharmaceutical aspects,
particularly with regards to NOHA’s chemical and metabolic sta-
bility profiles. Our difficulties with the preparation of O-acylated
prodrug candidates indicates that this type of modification may
not lead to compounds with improved pharmaceutical properties.
The performed hydrolytic and oxidative stability stud-
ies clearly demonstrated that O- and N-substitution of the
N-hydroxyguanidine moiety benefits the drug-likeness of these
NOHA derivatives. In summary, especially compared to unsub-
stituted NOHA , its lability towards bases, two- and three-valent
ions as well as the ROS peroxynitrite was solved primarily by
O-substitution (O-alkyl, O-acetal). However, it needs to be clari-
fied whether NOHA’s stability at pH 2–7 and its moderate stability
towards hydrogen peroxide and peroxynitrite is sufficient for in vivo
(peroral) applications. It might still be challenged by distinct
oxidative situations, particularly under conditions of temporarily
high local concentrations of oxidatively relevant species.
Notably, we not only wanted to address stability issues but
also the general pharmacokinetic profile with respect to passive
diffusion processes and targeted active transport mechanisms
(such as amino acid and carbohydrate transporters) for a good oral
bioavailability. Therefore, novel prodrug concepts were pursued as
there are none reported for the class of N-hydroxyguanidines so
far.
R
63 mm). Flash chromatography on a RP-18 RediSepꢀ column
R
(43 g) was performed with a CombiFlashꢀRETRIEVE system.
Chromatography on chiral stationary phase was carried out using
a Crownpak Cr(+) column (see the ESI for details†). All starting
materials were commercially available and used without further
purification. Na-Boc-L-ornithine tert-butylester hydrochloride (1)
was purchased from Bachem. CbzNCS was prepared according
to literature procedures.23–25 All solvents were distilled and dried
according to standard procedures.
Na-(t-Butyloxycarbonyl)-Nx-benzyloxycarbonyl-L-thiocitrulline
t-Butylester (2a)23,25
. Has been prepared in analogy to the
thiourea 2b described below. Yield: 93%.
Na -(t-Butyloxycarbonyl)-Nx -ethoxycarbonyl-L -thiocitrulline
t-butylester
(2b). Na-(t-Butyloxycarbonyl)-L-ornithine-t-
butylester (2.02 g, 7.0 mmol) is dissolved in 250 mL of dry CH2Cl2.
The solution is chilled on ice and ethoxycarbonylisothiocyanate
(917 mg, 7.0 mmol), dissolved in 15 mL of dry CH2Cl2, is added
dropwise for about 30 min. The reaction mixture is stirred for 2 h
at room temperature, concentrated to one third of the original
volume and washed with 25 mL 1% HCl, water and brine. The
organic phase is dried with Na2SO4 and concentrated in vacuo.
Thiourea 2b is at this point already pure by >95% (TLC). The
product was dissolved in a small amount of eluent (Cy/EtOAc,
4 : 1), charcoal added and purified on a short silica gel column
(ca. 20 g SiO2). Yield: 2.76 g of a colorless oil (94%) that solidifies
upon standing in the refrigerator; mp. 81 ◦C; TLC: Rf 0.31
(Cy/EtOAc, 4 : 1); 1H NMR (CDCl3): d (ppm) = 1.31 (t, 3H, 3J =
7.1 Hz, CH2–CH3), 1.45, 1.47 (2 ¥ s, 9H, C(CH3)3), 1.61–1.90 (m,
Our future efforts are dedicated to investigate the different
bioactivation pathways and general metabolic profiles, respec-
tively. Using model compound 13 we were already able to show
very efficient bioactivation by b-galactosidases in vitro (see the
ESI†). However, without a comprehensive metabolic study it is
difficult to name the best prodrug candidate at this stage. Each
implicated concept has its assets and drawbacks depending on the
desired route of application and formulation (pharmaceutical), ve-
locity of drug release (retardation principles), clinical application
(disease), et cetera.
3
4H, b,g-CH2), 3.66 (pseudo q, 2H, N–CH2), 4.22 (q, 2H, J =
7.1 Hz, CH2–CH3), 5.08 (m, 1H, a-CH), 8.06, 9.70 (2 ¥ br s,
1H, NH); 13C NMR (CDCl3): d (ppm) = 14.9 (CH2–CH3), 24.9
(g-CH2), 28.7, 29.0 (2 ¥ C(CH3)3), 31.0 (b-CH2), 45.8 (N–CH2),
54.3 (a-CH), 63.4 (O–CH2), 82.8 (C(CH3)3), 153.4 (CO-Eoc),
156.2 (CO-Boc), 172.5 (COOtBu), 180.1 (C S); MS (ESI): m/z =
442 [M + Na]+, 420 [M + H]+, 308 [M - 2 ¥ C4H8 + H]+, 264 [M -
2 ¥ C4H8 - CO2 + H]+; Anal. calcd for C18H33N3O6S (419.54): C
51.53, H 7.93, N 10.02, S 7.64; Found: C 51.68, H 7.97, N 10.06,
S 7.62.
Experimental section
Syntheses
General remarks. Melting points are uncorrected. 1H
(300 MHz) and 13C (75 MHz) NMR spectra were obtained
General procedure for the synthesis of fully protected NOHA
derivatives 6 and 7. A literature protocol for the preparation of
carbamoylguanidines 6 and 7 was used for the herein presented
protected NOHA derivatives.23–25 1.5 Equivalents of DIPEA,
O-substituted hydroxylamine (3–5) and EDCI were reacted
with 0.5 mmol thiourea (2a,b) in 10 mL of dry CH2Cl2. For
hydroxylamines that were applied as hydrochlorides (= methoxy-
lamine hydrochloride, 3) 3 equivalents of DIPEA were used.
Unless noted otherwise, reactions were complete after stirring
overnight. The organic phase was diluted with 10 mL of CH2Cl2
on
a Bruker ARX 300 spectrometer at 300 K. Chemi-
cal shifts (d values) are reported in ppm relative to TMS,
3-(trimethylsilyl)-1-propanesulfonic acid-d6 sodium salt (TPS) as
an internal standard, or alternatively, relative to the residual
solvent signal. All coupling constants (J values) were obtained
by first order analysis of the multiplets. Low resolution mass
spectra were recorded on a Bruker–Esquire–LC with an elec-
trospray ionization (ESI). Recordings of high mass resolution
spectra were conducted on a Bruker 7.4 Tesla FTICR mass
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 5249–5259 | 5255
©

View Article Online
3
and washed with small amounts of 1% aqueous HCl, water and
brine. The resulting oils were purified by column chromatography
on silica gel.
5.06 (m, 2H, 6¢-CH2), 5.25 (dd, J = 10.4, 8.3 Hz, 1H, 2¢-CH),
5.39 (dd, 3J = 3.5, 1.0 Hz, 1H, 4¢-CH), 6.47 (br t, 1H, NH), 7.65
(br s, 1H, NH); 13C NMR (CDCl3): d (ppm) = 14.3 (CH2–CH3),
20.55, 20.6, 20.7, 20.8 (4 ¥ COCH3), 24.9 (g-CH2), 28.0, 28.3
(2 ¥ C(CH3)3), 30.3 (b-CH2), 40.5 (N–CH2), 53.7 (a-CH), 61.0
(6¢-CH2), 62.1 (CH2–CH3), 66.9, 68.0, 70.5, 71.0 (2¢,3¢,4¢,5¢-CH),
79.6, 81.9 (2 ¥ C(CH3)3), 102.5 (1¢-CH), 150.5 (C N), 153.1
(COEoc), 158.9 (COBoc), 170.01, 170.03, 170.2, 170.3 (4 ¥
COCH3), 171.7 (COOtBu); MS (ESI): m/z = 772 [M + Na]+, 750
[M + H]+, 707 [M - C2H2O + H]+.
Nx -Benzyloxycarbonyl-Na-t-butyloxycarbonyl-Nx¢ -methoxy-L-
arginine t-butylester (6a)23. Yield: 95%.
Nx-Benzyloxycarbonyl-Na-(t-butyloxycarbonyl)-Nx¢-(methoxy-
carbonyl)methoxy-L-arginine
t-butylester
(6b). Eluent:
Cy/EtOAc (3 : 2), Rf 0.48; yield: 254 mg (92%) of a colorless oil;
1H NMR (CDCl3): d (ppm) = 1.44, 1.46 (2 ¥ s, 9H, C(CH3)3),
1.50–1.87 (m, 4H, b,g-CH2), 3.07 (m, 2H, N–CH2), 3.73 (s, 3H,
O–CH3), 4.16 (m, 1H, a-CH), 4.41 (s, 2H, O–CH2), 5.08 (m, 1H,
Nx-Methoxy-L-arginine bis(trifluoroacetate) (8a)23. Described
in the literature. Yield: 99%.
3
NH), 5.15 (s, 2H, CH2-Cbz), 6.37 (br t, J = 5.3 Hz, 1H, NH),
Nx-Methoxy-L-arginine ethylester dihydrochloride (8b). For
esterification 238 mg of the free amino acid precursor 8a
(0.55 mmol) were dissolved in 5 mL of absolute ethanol under
an argon atmosphere. This solution was stirred for 30 min at
-10 ◦C before gently bubbling HCl gas through for 5–10 min. For
completion of the reaction, it is stirred for one hour at 0 ◦C and left
in the refrigerator for 36 h. The mixture is carefully concentrated
under reduced pressure and lyophilized to obtain a hygroscopic
amorphous solid, that liquifies upon contact with air. Yield:
168 mg (99%) of a colorless oil; Rf 0.18 (i-propanol/H2O/AcOH,
7.30–7.39 (m, 5H, ArH), 8.23 (br s, 1H, NH); 13C NMR (CDCl3):
d (ppm) = 25.5 (g-CH2), 28.7 (b-CH2), 29.0, 30.9 (2 ¥ C(CH3)3),
41.2 (N–CH2), 52.4 (O–CH3), 54.5 (a-CH), 68.3 (CH2-Cbz), 71.2
(O–CH2), 80.3, 82.5 (2 ¥ C(CH3)3), 129.0, 129.28, 129.34 (ArCH),
135.9 (ArC), 150.8 (C N), 153.7 (CO-Cbz), 156.0 (CO-Boc),
171.7, 172.4 (COOtBu, COOMe); MS (ESI): m/z = 575 [M +
Na]+, 553 [M + H]+, 497 [M - C4H8 + H]+.
Na -(t-Butyloxycarbonyl)-Nx -ethoxycarbonyl-Nx¢ -methoxy-L-
arginine t-butylester (7a). Eluent: CH2Cl2/MeOH (98 : 2), Rf
1
1
0.26; yield: 203 mg (94%) of a colorless oil; H NMR (CDCl3):
8 : 1 : 1);%ee = 99.6 0.1; H NMR (DMSO-d6): d (ppm) = 1.24
3
(t, 3J = 7.2 Hz, 3H, CH2–CH3), 1.47–1.90 (m, 4H, b,g-CH2), 3.22
(m, 2H, N–CH2), 3.64 (s, 3H, O–CH3), 3.99 (m, 1H, a-CH), 4.21
d (ppm) = 1.27 (t, J = 7.1 Hz, 3H, CH2–CH3), 1.43, 1.45 (2 ¥ s,
9H, C(CH3)3), 1.56–1.89 (m, 4H, b,g-CH2), 3.09 (m, 2H, N–CH2),
3
3
3.66 (s, 3H, O–CH3), 4.16 (br q, J = 7.1 Hz, 3H, CH2–CH3,
(q, J = 7.2 Hz, 2H, CH2–CH3), 8.04 (br s, 2H, NH2), 8.31 (br
a-CH), 5.10, 6.26 (2 ¥ br m, 1H, NH), 7.80 (br s, 1H, NH); 13C-
NMR (CDCl3): d (ppm) = 14.9 (CH2–CH3), 25.6 (g-CH2), 29.7
(b-CH2), 29.0, 30.9 (2 ¥ C(CH3)3), 41.2 (N–CH2), 54.5 (a-CH),
62.0 (CH2–CH3), 62.6 (O–CH3), 80.2, 82.5, (2 ¥ C(CH3)3), 149.0
(C N), 153.8 (CO-Eoc), 156.0 (CO-Boc), 172.4 (COOtBu); MS
(ESI): m/z = 455 [M + Na]+, 433 [M + H]+, 377 [M - C4H8 + H]+;
Anal. calcd for C19H36N4O7 (432.52): C 52.76, H 8.39, N 12.95;
Found: C 53.65, H 8.57, N 13.24.
t, 1H, NH), 8.72 (br s, 3H, NH3 ), 11.33 (br s, 1H, NH+); 13C
+
NMR (DMSO-d6): d (ppm) = 13.9 (CH2–CH3), 24.0 (g-CH2),
27.0 (b-CH2), 40.0 (N–CH2), 51.4 (a-CH), 61.7 (CH2–CH3), 64.4
(O–CH3), 157.2 (C N), 169.2 (CO); HRMS (m/z): Calculated
for C9H21N4O3 [M + H]+ = 233.16082, found: 233.16064; Anal.
calcd for C9H20N4O3·2.0 HCl·0.7 H2O (317.82): C 34.01, H 7.42,
N 17.63; Found: C 33.65, H 7.66, N 18.20.
Nx-Ethoxycarbonyl-Nx¢-methoxy-L-arginine
bis(trifluoroace-
Na -(t-Butyloxycarbonyl)-Nx -ethoxycarbonyl-Nx¢ -(ethoxycar-
bonyl)methoxy-L-arginine t-butylester (7b). Eluent: Cy/EtOAc
tate) (8c). 200 mg of the fully protected precursor 7a (0.46 mmol)
are dissolved in 5 mL of TFA at 0 ◦C, stirred at that temperature
for 30 min and then 3 h at room temperature. TFA is removed in
a vacuum (not exceeding 20 ◦C) and the residue is taken up in
a minimum amount of water. The crude product is purified by
flash chromatography on a RP-18 column (0.1% TFA in water).
Ninhydrin-positive fractions are combined and concentrated on a
rotary evaporator (not exceeding 30 ◦C) to a volume of ca. 10 mL,
and were finally lyophilized. Yield: 225 mg (97%) of a colorless
oil; Rf 0.44 (i-propanol/H2O/AcOH, 8 : 1 : 1); %ee = 99.5
1
(3 : 2), Rf 0.51; yield: 250 mg (99%) of a colorless oil; H NMR
(CDCl3): d (ppm) = 1.26 (t, 3J = 7.1 Hz, 3H, CH2–CH3), 1.27 (t, 3J =
7.1 Hz, 3H, CH2–CH3), 1.42, 1.44 (2 ¥ s, 9H, C(CH3)3), 1.51–1.85
(m, 4H, b,g-CH2), 3.06 (m, 2H, N–CH2), 4.16 (q, 3J = 7.11 Hz, 2H,
CH2–CH3), 4.19 (q, 3J = 7.17 Hz, 2H, CH2–CH3), 4.39 (s, 2H, O–
CH2), 5.08 (m, 1H, NH), 6.40 (br t, 1H, NH), 8.19 (br s, 1H, NH);
13C NMR (CDCl3): d (ppm) = 14.1, 14.2 (2 ¥ CH2–CH3), 24.8 (g-
CH2), 27.9, 28.3 (2 ¥ C(CH3)3), 30.2 (b-CH2), 40.5 (N–CH2), 53.7
(a-CH), 60.8, 61.9 (2 ¥ CH2–CH3), 70.7 (O–CH2), 79.5, 81.8 (2 ¥
C(CH3)3), 150.4 (C N), 153.2 (CO-Eoc), 155.3 (CO-Boc), 170.7,
171.7 (COOEt, COOtBu); MS (ESI): m/z = 527 [M + Na]+, 505
[M + H]+, 449 [M - C4H8]+; Anal. calcd for C22H40N4O9 (504.59):
C 52.37, H 7.99, N 11.10; Found: C 53.09, H 7.90, N 11.44.
1
3
0.1; H NMR (DMSO-d6): d (ppm) = 1.22 (t, J = 7.1 Hz, 3H,
CH2–CH3), 1.50–1.87 (m, 4H, b,g-CH2), 3.12 (br t, 2H, N–CH2),
3
3.62 (s, 3H, O–CH3), 3.90 (m, 1H, a-CH), 4.13 (q, J = 7.1 Hz,
+
2H, CH2–CH3), 7.35 (br s, 1H, NH), 8.28 (br s, 3H, NH3 ); 13C
NMR (DMSO-d6): d (ppm) = 14.1 (CH2–CH3), 24.6 (g-CH2),
27.3 (b-CH2), 40.6 (N–CH2), 51.7 (a-CH), 61.6 (CH2–CH3), 62.1
(O–CH3), 149.0 (C N), 153.5 (CO-Eoc), 170.9 (CO); HRMS
(m/z): Calculated for C10H21N4O5 [M + H]+ = 277.15065, found:
277.15049; Anal. calcd for C10H20N4O5·2.0 CF3COOH·0.4 H2O
(513.56): C 32.74, H 4.87, N 10.91; Found: C 32.44, H 4.69, N
10.53.
Na-(t-Butyloxycarbonyl)-Nx-ethoxycarbonyl-Nx¢-(2,3,4,6-tetra-
O-acetyl-b-D-galactopyranos-1-yl)oxy-L-arginine t-butylester (7c).
Eluent: CH2Cl2/MeOH (97 : 3), Rf 0.35; yield: 262 mg (70%) of a
1
3
colorless oil; H NMR (CDCl3): d (ppm) = 1.29 (t, J = 7.1 Hz,
3H, CH2–CH3), 1.43, 1.45 (2 ¥ s, 9H, C(CH3)3), 1.52–1.86 (m,
4H, b,g-CH2), 1.98, 2.02, 2.05, 2.13 (4 ¥ COCH3), 3.08 (m, 2H,
3
N–CH2), 3.96 (br t, J = 6.8 Hz, 1H, 5¢-CH), 4.11–4.21 (m, 5H,
Nx-Ethoxycarbonyl-Nx¢-methoxy-L-arginine ethylester dihy-
drochloride (8d). The esterification of the free amino acid 8c was
a-CH, CH2–CH3, 3¢-CH, NH), 4.85 (d, 3J = 8.3 Hz, 1H, 1¢-CH),
5256 | Org. Biomol. Chem., 2011, 9, 5249–5259
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
3
carried out as described for 8b. Yield: 150 mg (99%) of a colorless
oil, starting from 200 mg (0.397 mmol) of 8c; 1H NMR (DMSO-
d6): d (ppm) = 1.23, 1.25 (2 ¥ t, J = 7.0 Hz, 3H, CH2–CH3),
3H, CH2–CH3), 1.26 (t, J = 7.10 Hz, 3H, CH2–CH3), 1.51–1.87
(m, 4H, b,g-CH2), 3.17 (m, 2H, N–CH2), 3.95 (m, 1H, a-CH),
4.14 (q, 3J = 7.12 Hz, 2H, CH2–CH3), 4.15 (q, 3J = 7.10 Hz, 2H,
CH2–CH3), 4.17–4.25 (m, 2H, CH2–CH3), 4.51 (s, 2H, O–CH2),
3
1.55–1.88 (m, 4H, b,g-CH2), 3.32 (br t, 2H, N–CH2), 3.71 (s, 3H,
O–CH3), 3.95 (m, 1H, a-CH), 4.19 (m, 4H, 2 ¥ CH2–CH3), 8.80
+
7.80 (br s, 1H, NH), 8.55 (br s, 1H, NH), 8.72 (br s, 3H, NH3 );
(br s, 4H, NH3 , NH); 13C NMR (DMSO-d6): d (ppm) = 13.9, 14.0
13C NMR (DMSO-d6): d (ppm) = 13.9, 14.0, 14.1 (3 ¥ CH2–CH3),
24.2 (g-CH2), 27.2 (b-CH2), 40.8 (N–CH2), 51.5 (a-CH), 60.5,
61.7, 61.9 (3 ¥ CH2–CH3), 71.1 (O–CH2), 150.3 (C N), 153.0
(CO-Eoc), 169.1, 169.3 (2 ¥ CO); HRMS (m/z): Calculated for
C15H29N4O7 [M + H]+ = 377.20308, found: 377.20287.
+
(2 ¥ CH2–CH3), 24.0 (g-CH2), 26.9 (b-CH2), 41.1 (N–CH2), 51.4
(a-CH2), 61.7, 62.5 (2 ¥ CH2–CH3), 63.9 (O–CH3), 150.7 (C N),
152.6 (CO-Eoc), 169.2 (COOEt); HRMS (m/z): Calculated for
C12H25N4O5 [M + H]+ = 305.18195; found: 305.18176; Anal. calcd
for C12H24N4O5·2.0 HCl·0.6 H2O (388.08): C 37.14, H 7.06, N
14.44; Found: C 36.67, H 7.59, N 15.00.
Sodium
Nx-ethoxycarbonyl-Nx¢-(2,3,4,6-tetra-O-acetyl-b-D-
galactopyranos-1-yl)oxy-L-argininate (10). In a Schlenk flask
120 mg (0.16 mmol) of the fully protected, thoroughly dried
precursor 7c is dissolved in absolute Et2O under an argon
atmosphere. The solution is stirred for ca. 30 min at -15 ◦C,
and HCl gas is carefully bubbled through the solution. The
reaction mixture is left in the refrigerator overnight and then
concentrated in a vacuum. The white-yellow solid residue is taken
up with ca. 1–2 mL of 0.5 M aqueous NaHCO3 and purified
by flash chromatography on a RP-18 column (eluent = water,
with a stepwise MeOH gradient of 10%, 25%, 50%). The product
containing fractions were combined, concentrated on a rotary
evaporator (not exceeding 30 ◦C) and lyophilized. Yield: 74 mg
of a fine, white powder (75%). 1H NMR (D2O, TPS): 10 appears
as isomers in the ratio of ca. 6 : 4 (at 300 K, referred to the CH3
triplet of the ethoxycarbonyl group). The stated ppm values
Nx-Carboxymethoxy-L-arginine dihydrochloride (9a). The
fully protected precursor 6b (270 mg, 0.489 mmol) is stirred for 4 h
at 50–60 ◦C in 5 mL aqueous 6 N HCl. The mixture is concentrated
to dryness, taken up with ca. 1–2 mL water and subjected to
flash chromatography on a RP-18 column (eluent = 0.1% TFA in
water). Ninhydrin-positive fractions are combined, concentrated
to a few residual millilitres and lyophilized. Yield: 150 mg of a
fine, amorphous solid (96%); Rf 0.53 (i-propanol/H2O/AcOH,
1
6 : 3 : 1); %ee = 99.1 0.1; H NMR (D2O): d (ppm) = 1.76–2.18
(m, 4H, b,g-CH2), 3.41 (br t, 3J = 6.7 Hz, 2H, N–CH2), 4.17 (br t,
3J = 6.2 Hz, a-CH), 4.63 (s, 2H, O–CH2); 13C NMR (D2O, TPS): d
(ppm) = 26.3 (g-CH2), 29.6 (b-CH2), 43.2 (N–CH2), 55.3 (a-CH),
75.5 (O–CH2), 161.0 (C N), 174.6, 175.3 (2 ¥ CO); MS (ESI):
m/z = 249 [M + H]+; Anal. calcd for C8H16N4O5·2.0 HCl·0.5 H2O
(330.17): C 29.10, H 5.80, N 16.97; Found: C 29.00, H 5.99, N
17.16.
3
refer to the main isomer; d (ppm) = 1.28 (t, J = 7.1 Hz, 3H,
CH2–CH3), 1.52–1.97 (m, 4H, b,g-CH2), 2.02, 2.08, 2.12, 2.23
3
(4 ¥ s, 3H, COCH3), 3.16 (br t, J = 6.8 Hz, 2H, 6¢-CH2), 3.75
Nx -Ethoxycarbonyl-Nx¢ -(ethoxycarbonyl)methoxy-L -arginine
bis(trifluoroacetate) (9b). Deprotection of 7b is carried out in
analogy to the protocol for 8c, but starting from 0.5 mmol (252 mg)
starting material 7b. The purification was performed by flash
chromatography on a RP-18 column using a stepwise gradient
(0.1% TFA in water with MeOH from 5–30%). Ninhydrin-positive
fractions were combined, concentrated (not exceeding 30 ◦C)
and lyophilized. Yield: 277 mg of a colorless oil (96%); Rf 0.62
(t, 3J = 6.3 Hz, 1H, 5¢-CH), 4.15–4.30 (m, 5H, CH2–CH3, a-CH,
3
N–CH2), 5.07 (d, J = 8.0 Hz, 1H, 1¢-CH), 5.21–5.33 (m, 2H,
2¢,3¢-CH), 5.47 (m, 1H, 4¢-CH); 13C NMR (D2O, TPS): d (ppm) =
16.2 (CH2–CH3), 22.7, 22.8, 22.9, (4 ¥ COCH3), 26.6 (g-CH2),
30.6 (b-CH2), 42.7 (N–CH2), 57.2 (a-CH), 64.6 (6¢-CH2), 65.8
(CH2–CH3), 70.6, 70.7, 73.4, 74.1 (2¢,3¢,4¢,5¢-CH), 104.1 (1¢-CH),
155.6 (C N), 156.8 (CO-Eoc), 175.4, 175.8, 176.2, 177.1 (4 ¥
COCH3); MS (ESI): m/z = 615 [M + Na]+, 593 [M + H]+, 551
[M - C2H2O + H]+, 331 [C14H19O9]+; HRMS (m/z): Calculated
for C23H33N4O14Na [M + Na]+ = 615.21202, found: 615.21164;
anal. calcd for C23H35N4NaO14 (614.53): C 44.95, H 5.74, N 9.12;
Found: C 45.08, H 6.21, N 8.77.
1
(i-propanol/H2O/AcOH, 8 : 1 : 1); %ee = 99.0 0.1; H NMR
(DMSO-d6): With DMSO-d6 as the solvent 9b appears as isomers
in the ratio of ca. 8.6 : 1.4 (at 300 K, referred to the CH2 singlet of
the ethoxycarbonylmethoxy group). The stated ppm values refer
3
to the main isomer; d (ppm) = 1.19 (t, J = 7.10 Hz, 3H, CH2–
N-Benzyl-N¢-ethoxycarbonyl-N¢¢-[(2,3,4,6-tetra-O-acetyl-b-D-
galactopyranos-1-yl)oxy]guanidine (12). 119 mg of thiourea 1124
(see also ESI for analytical data†) (0.5 mmol), 218 mg of the
protected 1-aminooxygalactose 5 (0.6 mmol) and 104.5 mL DIPEA
(0.6 mmol) are dissolved in 5 mL of dry CH2Cl2. The mixture is
cooled to 0 ◦C, after which 115 mg of EDCI (0.6 mmol) is added.
This mixture is stirred for 48 h at room temperature. For work up,
10 mL of CH2Cl2 are added and the organic phase is washed with
1% HCl, water and brine (5 mL each). The organic phase is dried
with Na2SO4 and concentrated in a vacuum. The crude product
is purified by column chromatography on SiO2 using Cy/EtOAc
(1 : 1) as the eluent. Yield: 255 mg of a colorless oil (90%); TLC:
3
CH3), 1.21 (t, J = 7.08 Hz, 3H, CH2–CH3), 1.48–1.87 (m, 4H,
b,g-CH2), 3.02 (m, 2H, N–CH2), 3.89 (m, 1H, a-CH), 4.10 (q, 3J =
3
7.04 Hz, 2H, CH2–CH3), 4.12 (q, J = 7.13 Hz, 2H, CH2–CH3),
+
4.37 (s, 2H, O–CH2), 6.54 (br s, 1H, NH), 8.24 (br s, 3H, NH3 ),
10.64 (br s, COOH); 13C NMR (DMSO-d6): d (ppm) = 14.0, 14.2
(2 ¥ CH2–CH3), 27.4 (CH2), 40.3 (N–CH2), 51.8 (a-CH), 60.2,
61.3 (2 ¥ CH2–CH3), 70.3 (O–CH2), 169.9, 171.0 (3 ¥ CO); HRMS
(m/z): Calculated for C13H25N4O7 [M + H]+ = 349.17178, found:
349.17155; Anal. calcd for C13H24N4O7·3.2 CF3COOH·1.5 H2O
(740.26): C 31.48, H 4.11, N 7.57; Found: C 31.52, H 4.24, N 7.40.
Nx -Ethoxycarbonyl-Nx¢ -(ethoxycarbonyl)methoxy-L -arginine
ethylester dihydrochloride (9c). The esterification of the free
amino acid 9b was carried out as described for 8b. Yield: 222 mg
(99%) of a colorless oil, starting from 288 mg (0.5 mmol) of 9b; Rf
1
Rf 0.54 (Cy/EtOAc, 1 : 1); H NMR (CDCl3): d (ppm) = 1.30 (t,
3J = 7.1 Hz, 3H, CH2–CH3), 1.99, 2.02, 2.06, 2.15 (4 ¥ s, 3H,
3
COCH3), 3.98 (br t, J = 6.7 Hz, 1H, 5¢-CH), 4.10–4.30 (m, 6H,
1
0,57 (i-propanol/H2O/AcOH, 8 : 1 : 1); H NMR (DMSO-d6): d
6¢-CH2, N–CH2, CH2–CH3), 4.90 (d, 3J = 8.3 Hz, 1H, 1¢-CH), 5.08
(m, 1H, 3¢-CH), 5.29 (br t, 3J = 9.6 Hz, 1H, 2¢-CH), 5.41 (m, 1H,
(ppm) = 1.20 (t, 3J = 7.16 Hz, 3H, CH2–CH3), 1.23 (t, 3J = 7.13 Hz,
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 5249–5259 | 5257
©

View Article Online
4¢-CH), 6.79 (br t, 1H, NH), 7.22–7.35 (m, 5H, ArH), 7.72 (br s,
1H, NH); 13C NMR (CDCl3): d (ppm) = 14.3 (CH2–CH3), 20.6,
20.7, 20.9 (4 ¥ COCH3), 45.1 (N–CH2), 61.1 (6¢-CH2), 62.2 (CH2–
CH3), 67.0, 68.1, 70.7, 71.0 (2¢,3¢,4¢,5¢-CH), 102.7 (1¢-CH), 127.4,
127.8, 128.6 (ArCH), 138.3 (ArC), 150.4 (C N), 153.2 (CO-Eoc),
170.0, 170.2, 170.3 (4 ¥ COCH3); MS (ESI): m/z = 590 [M + Na]+,
568 [M + H]+, 526 [M - C2H2O + H]+, 331 [C14H19O9]+.
immediately at the desired time points. Subsequent samples were
analyzed by HPLC as described above.
Oxidative stability. Stability was determined at 37 ◦C in 10 mM
phosphate buffer pH 7.4 containing either 500 mM potassium
hexacyanoferrate(III), 500 mM copper(II)-acetate, 500 mM perox-
ynitrite or 29 mM hydrogen peroxide, respectively. Prodrugs were
applied in a concentration of 250 mM. After 24 h samples were
taken and immediately analyzed via HPLC as described above.
N -Benzyl-N¢-ethoxycarbonyl-N¢¢-[(b-D-galactopyranos-1-yl)-
oxy]guanidine (13). 150 mg of the protected precursor 12
(0.26 mmol) are dissolved in 5 mL MeOH and gently stirred
with ca. 30 mg of strong anion exchanger resin (Merck Ionenaus-
tauscher III, OH-) over night. The solution is filtered, washed with
small amounts of MeOH and concentrated on a rotary evaporator.
The product is further purified by flash chromatography on SiO2
using a step-wise gradient of EtOAc/MeOH (9 : 1 to 8 : 2). Yield:
84 mg of a fine white powder (80%); TLC: Rf 0.66 (EtOAc/MeOH,
8 : 2); 1H NMR (MeOH-d4): In MeOH-d4 as the NMR solvent the
substance occurs as isomers at a ratio of ca. 8 : 2 (300 K, referred
to the triplet of the ethyloxycarbonyl methyl-group). In DMSO-d6
this ratio is ca. 6 : 4. The stated ppm values refer to the main isomer;
d (ppm) = 1.38 (t, 3J = 7.1 Hz, 3H, CH2–CH3), 3.58–3.67 (m, 2H,
Acknowledgements
We thank Dr Ulrich Girreser for performing NMR and MS
experiments and gratefully acknowledge the excellent technical
assistance of Ms. Melissa Zietz, Ms. Johanna Blauth and Mr.
Sven Wichmann. We are also indebted to Dr. Erik Ralph for his
meticulous and valuable proofreading.
Notes and references
1 K. Bian, M. F. Doursout and F. Murad, J. Clin. Hypertens., 2008, 10,
304–310.
2 S. Moncada and E. A. Higgs, Br. J. Pharmacol., 2006, 147(Suppl 1),
S193–201.
3 C. Napoli, F. de Nigris, S. Williams-Ignarro, O. Pignalosa, V. Sica and
L. J. Ignarro, Nitric Oxide, 2006, 15, 265–279.
4 D. Schade, J. Kotthaus and B. Clement, Pharmacol. Ther., 2010, 126,
279–300.
5 D. J. Stuehr, N. S. Kwon, C. F. Nathan, O. W. Griffith, P. L. Feldman
and J. Wiseman, J Biol Chem, 1991, 266, 6259–6263.
6 W. Durante, F. K. Johnson and R. A. Johnson, Clin. Exp. Pharmacol.
Physiol., 2007, 34, 906–911.
7 J. B. Walker and M. S. Walker, J. Biol. Chem., 1959, 234, 1481–1484.
8 J. M. Fukuto, Methods Enzymol., 1996, 268, 365–375.
9 T. Cai, M. Xian and P. G. Wang, Bioorg. Med. Chem. Lett., 2002, 12,
1507–1510.
3
6¢-CH2), 3.71–3.96 (m, 4H, 2¢,3¢,4¢,5¢-CH), 4.28 (q, J = 7.1 Hz,
2H, CH2–CH3), 4.36 (br s, 2H, N–CH2), 4.61 (d, 3J = 8.1 Hz, 1H,
1¢-CH), 7.31–7.44 (m, 5H, ArH); 13C NMR (MeOH-d4): d (ppm) =
14.6 (CH2–CH3), 45.8 (N–CH2), 62.5 (6¢-CH2), 63.2 (CH2–CH3),
70.2, 71.2, 75.0, 76.8 (2¢,3¢,4¢,5¢-CH), 106.6 (1¢-CH), 128.3, 128.6,
129.6 (ArCH), 139.8 (ArC), 152.8 (C N), 155.1 (CO); HRMS
(m/z): Calculated for C17H26N3O8 [M + H]+ = 400.17144, found:
400.17147; Anal. calcd for C17H33N4O14Na [M + Na]+ = 422.15339,
found: 422.15327.
In vitro assays
10 A. Jousserandot, J. L. Boucher, Y. Henry, B. Niklaus, B. Clement and
D. Mansuy, Biochemistry, 1998, 37, 17179–17191.
11 H. G. Korth, R. Sustmann, C. Thater, A. R. Butler and K. U. Ingold,
J. Biol. Chem., 1994, 269, 17776–17779.
12 J. Y. Cho, A. Dutton, T. Miller, K. N. Houk and J. M. Fukuto, Arch.
Biochem. Biophys., 2003, 417, 65–76.
Materials and methods. Commercially available materials were
purchased from either Sigma Aldrich, Fluka, Roth or Merck.
HPLC-grade methanol was obtained from Mallinckrodt Baker.
D-Galactose concentration was analysed using the testkit “Lac-
tose/Galactose” of Boehringer Mannheim.
13 J. M. Fukuto, G. C. Wallace, R. Hszieh and G. Chaudhuri, Biochem.
Pharmacol., 1992, 43, 607–613.
Analytical reversed phase HPLC. Amino acids were deriva-
tized with o-Pa and separated at 37 ◦C on a Nova–Pak RP18 4 mm
(4 ¥ 150 mm, VDS Optilab) with a Phenomenex C18, 4 ¥
3 mm guard column. HPLC system: Waters autosampler 717plus,
Waters 600 Controller and Waters 470 scanning fluorescence
detector, set at lex: 338 nm, lem: 425 nm. For derivatization, the
autosampler was set to mix 14 mL of o-Pa reagent with 10 mL of
sample and allowed to react for 2 min at room temperature before
injection. o-Pa reagent was prepared by dissolving 50 mg o-Pa in
9 mL of 0.2 M potassium borate buffer pH 9.4 and 53 mL of 2-
mercaptoethanol. Eluent A: 10 mM potassium phosphate buffer
(pH 4.7)/MeOH (78%/22%); eluent B: methanol/aqua bidest.
(80%/20%). The following elution conditions were used: flow-rate
was kept at 1 mL min-1; 0–12 min isocratic with 90% eluent A and
10% eluent B; 12–25 min linear gradient to 100% eluent B, 25–
27 min, linear gradient to 90% eluent A, 27–35 min reequilibration
with 90% eluent A and 10% eluent B.
14 J. M. Fukuto, D. J. Stuehr, P. L. Feldman, M. P. Bova and P. Wong,
J. Med. Chem., 1993, 36, 2666–2670.
15 B. Clement, M. Demesmaeker and S. Linne, Chem. Res. Toxicol., 1996,
9, 682–688.
16 A. K. Froehlich, U. Girreser and B. Clement, Drug Metab. Dispos.,
2005, 33, 1532–1537.
17 B. Clement, M. H. Schultze-Mosgau and H. Wohlers, Biochem.
Pharmacol., 1993, 46, 2249–2267.
18 B. Clement, T. Kunze and S. Heberling, Biochem. Biophys. Res.
Commun., 2006, 349, 869–873.
19 J. Kotthaus, B. Wahl, A. Havemeyer, J. Kotthaus, D. Schade, D. Garbe-
Scho¨nberg, R. Mendel, F. Bittner and B. Clement, Biochem. J., 2011,
433, 383–391.
20 S. M. Morris Jr., Am. J. Clin. Nutr., 2006, 83, 508S–512S.
21 J. Kotthaus, D. Schade, K. Topker-Lehmann, E. Beitz and B. Clement,
Bioorg. Med. Chem., 2008, 16, 2305–2312.
22 J. Kotthaus, D. Schade, N. Muschick, E. Beitz and B. Clement, Bioorg.
Med. Chem., 2008, 16, 10205–10209.
23 D. Schade, J. Kotthaus and B. Clement, Synthesis, 2008, 15, 2391–2397.
24 B. R. Linton, A. J. Carr, B. P. Orner and A. D. Hamilton, J. Org. Chem.,
2000, 65, 1566–1568.
25 N. I. Martin, J. J. Woodward and M. A. Marletta, Org. Lett., 2006, 8,
pH stability. Compounds were dissolved in 10 mM potassium
phosphate buffer pH 2, pH 7.4 and pH 9 to a final concentration of
250 mM. Samples were incubated at 37 ◦C and reactions analyzed
4035–4038.
26 N. I. Martin, W. T. Beeson, J. J. Woodward and M. A. Marletta, J. Med.
Chem., 2008, 51, 924–931.
5258 | Org. Biomol. Chem., 2011, 9, 5249–5259
This journal is
The Royal Society of Chemistry 2011
©

View Article Online
27 T. Wenzler, D. W. Boykin, M. A. Ismail, J. E. Hall, R. R. Tidwell and
R. Brun, Antimicrob. Agents Chemother., 2009, 53, 4185–4192.
28 F. L. Wagenaar and J. F. Kerwin, J. Org. Chem., 1993, 58, 4331–
4338.
29 M. Z. Wang, J. Q. Wu, A. S. Bridges, D. C. Zeldin, S. Kornbluth, R. R.
Tidwell, J. E. Hall and M. F. Paine, Drug Metab. Dispos., 2007, 35,
2067–2075.
37 B. Testa and J. M. Mayer, in Hydrolysis in Drug and Prodrug
Metabolism. Chemistry, Biochemistry and Enzymology, Helvetica
Chimica Acta, Zu¨rich, 2003, pp. 365–387.
38 G. Zinner, Arch. Pharm., 1960, 293, 657–661.
39 B. Clement, E. Schnorwangen, T. Kampchen, P. Mordvintcev and A.
Mulsch, Arch. Pharm., 1994, 327, 793–798.
40 C. Belzecki, B. Hintze and S. Kwiatkowska, Bull. Acad. Pol. Sci. Chem.,
1970, 18, 569–572.
30 D. Schade, J. Kotthaus, H. Hungeling and B. Clement, ChemMedChem,
2009, 4, 1595–1599.
41 C. Belzecki and J. Trojnar, Bull. Acad. Pol. Sci. Chem., 1970, 18, 379–
31 A. G. Katopodis and S. W. May, Biochemistry, 1990, 29, 4541–
4548.
383.
42 C. Belzecki, K. Piotrowska and B. Hintze, Bull. Acad. Pol. Sci. Chem.,
1971, 19, 367–370.
43 K. Rehse and S. Bade, Arch. Pharm., 1996, 329, 535–540.
44 A. R. Katritzky, N. M. Khashab, N. Kirichenko and A. Singh, J. Org.
Chem., 2006, 71, 9051–9056.
45 A. R. Katritzky, N. M. Khashab, S. Bobrov and M. Yoshioka, J. Org.
Chem., 2006, 71, 6753–6758.
46 Y. Higashi, K. Noma, M. Yoshizumi and Y. Kihara, Circ. J., 2009, 73,
411–418.
32 D. Melisi, A. Secondo, P. Montoro, S. Piacente, M. G. Rimoli, M.
Minale, P. de Caprariis and L. Annunziato, J. Med. Chem., 2006, 49,
4826–4833.
33 P. G. Wang, M. Xian, X. Tang, X. Wu, Z. Wen, T. Cai and A. J. Janczuk,
Chem. Rev., 2002, 102, 1091–1134.
34 X. X. T. Wu, M. Xian and P. G. Wang, Tetrahedron Lett., 2001, 42,
3779–3782.
35 O. Renaudet and P. Dumy, Tetrahedron Lett., 2001, 42, 7575–
7578.
36 B. Testa and J. M. Mayer, in Hydrolysis in Drug and Prodrug
Metabolism. Chemistry, Biochemistry and Enzymology, Helvetica
Chimica Acta, Zu¨rich, 2003, pp. 664–672.
47 C. Szabo, H. Ischiropoulos and R. Radi, Nat. Rev. Drug Discovery,
2007, 6, 662–680.
48 G. J. Southan, A. L. Salzman and C. Szabo, Free Radical Biol. Med.,
1998, 25, 914–925.
This journal is
The Royal Society of Chemistry 2011
Org. Biomol. Chem., 2011, 9, 5249–5259 | 5259
©