Synthesis and biological evaluation of triazolyl-substituted benzyloxyacetohydroxamic acids as LpxC inhibitors

Article abstract of DOI:10.1016/j.bmc.2020.115529

The bacterial deacetylase LpxC is a promising target for the development of antibiotics selectively combating Gram-negative bacteria. To improve the biological activity of the reported benzyloxyacetohydroxamic acid 9 ((S)-N-hydroxy-2-{2-hydroxy-1-[4-(phenylethynyl)phenyl]ethoxy}acetamide), its hydroxy group was replaced by a triazole ring. Therefore, in divergent syntheses, triazole derivatives exhibiting rigid and flexible lipophilic side chains, different configurations at their stereocenter, and various substitution patterns at the triazole ring were synthesized, tested for antibacterial and LpxC inhibitory activity, and structure-activity relationships were deduced based on docking and binding energy calculations.

Full text of DOI:10.1016/j.bmc.2020.115529

Journal Pre-proofs
Synthesis and biological evaluation of triazolyl-substituted benzyloxyacetohy‐
droxamic acids as LpxC inhibitors
Katharina Hoff, Sebastian Mielniczuk, Oriana Agoglitta, Maria Teresa Iorio,
Manlio Caldara, Emre F. Bülbül, Jelena Melesina, Wolfgang Sippl, Ralph
Holl
PII:
DOI:
Reference:
S0968-0896(20)30355-2
https://doi.org/10.1016/j.bmc.2020.115529
BMC 115529
To appear in:
Bioorganic & Medicinal Chemistry
Received Date:
Revised Date:
Accepted Date:
20 March 2020
20 April 2020
21 April 2020
Please cite this article as: Hoff, K., Mielniczuk, S., Agoglitta, O., Teresa Iorio, M., Caldara, M., Bülbül, E.F.,
Melesina, J., Sippl, W., Holl, R., Synthesis and biological evaluation of triazolyl-substituted
benzyloxyacetohydroxamic acids as LpxC inhibitors, Bioorganic & Medicinal Chemistry (2020), doi: https://
doi.org/10.1016/j.bmc.2020.115529
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover
page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version
will undergo additional copyediting, typesetting and review before it is published in its final form, but we are
providing this version to give early visibility of the article. Please note that, during the production process, errors
may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
©
2020 Published by Elsevier Ltd.

Synthesis
and
biological
evaluation
of
triazolyl-substituted
benzyloxyacetohydroxamic acids as LpxC inhibitors
Katharina Hoff,a,b Sebastian Mielniczuk,a,b Oriana Agoglitta,a,b,c,d Maria Teresa Iorio,d
a
e
e
e
Manlio Caldara, Emre F. Bülbül, Jelena Melesina, Wolfgang Sippl, and Ralph
a,b,
Holl *
aInstitute of Organic Chemistry, University of Hamburg, Martin-Luther-King-Platz 6,
20146 Hamburg, Germany
bGerman Center for Infection Research (DZIF), partner site Hamburg-Lübeck-Borstel-
Riems
cNRW Graduate School of Chemistry, University of Münster, Germany
dInstitute of Pharmaceutical and Medicinal Chemistry, University of Münster,
Corrensstr. 48, 48149 Münster, Germany
eInstitute of Pharmacy, Martin Luther University of Halle-Wittenberg, Wolfgang-
Langenbeck Str. 4, 06120 Halle (Saale), Germany
th
Dedicated to Prof. Dr. Bernhard Wünsch on the occasion of his 60 birthday
*To whom correspondence should be addressed. Tel.: +49-40-42838-2825; Fax:
+
49-40-42838-4325; E-mail: ralph.holl@chemie.uni-hamburg.de

Graphical Abstract
Key words
LpxC inhibitors, structure-activity relationships, azide-alkyne cycloadditions, triazole
derivatives, hydroxamic acids, enantioselective synthesis
Abstract
The bacterial deacetylase LpxC is a promising target for the development of antibiotics
selectively combating Gram-negative bacteria. To improve the biological activity of the
reported benzyloxyacetohydroxamic acid
9
((S)-N-hydroxy-2-{2-hydroxy-1-[4-
(phenylethynyl)phenyl]ethoxy}acetamide), its hydroxy group was replaced by a triazole
ring. Therefore, in divergent syntheses, triazole derivatives exhibiting rigid and flexible
lipophilic side chains, different configurations at their stereocenter, and various
substitution patterns at the triazole ring were synthesized, tested for antibacterial and
LpxC inhibitory activity, and structure-activity relationships were deduced based on
docking and binding energy calculations.

1. Introduction
The development of bacterial resistance to commonly used antibiotics is a natural
process of adaptation.1-2 However, the inappropriate and unnecessary use of the
available antibiotics enhances the emergence of bacterial resistance, which develops
through a multitude of mechanisms.3-5 Various bacterial species have already
developed resistance against most of the available antibiotics.6-7 These multidrug-
resistant bacteria lead to severe difficulties in the treatment of even common infections,
cause an increase in the number of fatalities and are associated with higher costs for
health care systems.8-10 Thus, in order to combat the infections caused by multidrug-
resistant bacteria, there is an essential need for the development of new antibiotics
with novel mechanisms of action.11-12 However, despite the urgent need for new
antibiotics, there is a big lack of new antibacterial drugs in the development pipeline,
in particular of those needed to counteract the increasing number of multidrug-resistant
Gram-negative bacteria.13-17
The bacterial deacetylase LpxC represents a promising target for the development of
18
novel antibiotics selectively combating Gram-negative bacteria. In Escherichia coli,
2+
the Zn -dependent enzyme catalyzes the irreversible deacetylation of UDP-3-O-[(R)-
-hydroxymyristoyl]-N-acetylglucosamine (2), representing the first irreversible step of
3
lipid A biosynthesis (Figure 1), which is highly conserved in Gram-negative bacteria.19
Lipid A is the hydrophobic membrane anchor of lipopolysaccharides (LPS), which are
the main component of the outer leaflet of the outer membrane of Gram-negative
bacteria.20 Being necessary for the maintenance of an effective outer membrane
barrier, lipid A is essential for the growth and viability of Gram-negative bacteria.19, 21
Thus, bacteria with a defective lipid A biosynthesis grow slowly and are hypersensitive
towards several antibiotics, whereas the complete inhibition of lipid A biosynthesis is
lethal to Gram-negative bacteria.20, 22 In consequence, due to the essential role of LpxC
in the biosynthesis of lipid A, inhibitors of the deacetylase represent a new class of
potential antibiotics.18 Whereas LpxC orthologues from different Gram-negative
species share substantial sequence similarity, the deacetylases exhibit no sequence
homology with any mammalian protein.22 The enzyme displays a novel “β-α-α-β
sandwich” fold, which is formed by two domains showing similar topologies.23 The
2+
active site, in which the catalytic Zn -ion is located, can be found at the interface of
24
the two domains on one side of the sandwich. Additionally, the enzyme exhibits a

hydrophobic tunnel leading out of the active site, which binds the fatty acyl chain of the
enzyme’s natural substrate 2 during catalysis.25
Figure 1: Chemical structure of Kdo -lipid A and the first two steps of its biosynthesis
2
in E. coli.18, 23, 26
Several classes of LpxC inhibitors have already been reported in the literature.18, 26-35
2+
Most of them share a Zn -chelating hydroxamate moiety and a lipophilic side chain
binding to the hydrophobic tunnel of LpxC, like e.g. the potent threonine-based LpxC
inhibitors CHIR-090 (5) and LPC-009 (6), tyrosine analogue LPC-051 (7), as well as
the 3-amino-valine derivative ACHN-975 (8), which was the first LpxC inhibitor to enter
human clinical trials and was thereby found to exhibit dose‐limiting cardiovascular
toxicity (Figure 2).36-39
Recently, we have described the synthesis and the biological evaluation of
benzyloxyacetohydroxamic acid 9 (Figure 2).40 The compound showed promising
inhibitory activity against E. coli LpxC and was found to exhibit antibacterial properties
against several E. coli strains. Therefore, this compound was chosen as lead for further

optimization steps. Docking studies with the benzyloxyacetohydroxamic acid had
2+
revealed that the compound’s hydroxamate moiety chelates the catalytic Zn -ion of
LpxC and forms hydrogen bonds with conserved residues of the active site, whereas
40
the diphenylacetylene moiety is placed in the hydrophobic tunnel. The hydroxy group
as well as the ether oxygen of the linker region undergo hydrogen-bonding interactions
with polar residues of the active site of LpxC. However, the compound leaves the UDP-
binding site of the enzyme unoccupied. In order to further optimize hydroxamic acid 9,
its hydroxy group should be replaced by another functional group, which is also able
to form hydrogen bonds with the enzyme and additionally offers the possibility of
introducing further substituents addressing the so far unoccupied UDP-binding site. As
triazoles are able to undergo hydrogen bonds and can be easily accessed via azide-
alkyne cycloadditions, thereby enabling the introduction of various substituents, a
series of triazole derivatives (10) was envisaged to test the feasibility of the triazole
moiety as possible linker group. Besides 1-monosubstituted triazole derivatives, 1,4-
and 1,5-disubstituted regioisomers should be synthesized to find the optimal position
of the substituent at the triazole ring. Additionally, the influence of polar and non-polar
substituents at the triazole ring should be investigated. As hydroxamic acid 9 showed
considerably higher antibacterial and inhibitory activity than its (R)-configured
enantiomer ent-9, mainly (S)-configured triazole derivatives should be synthesized.
However, to unequivocally confirm that also in case of the triazole derivatives the (S)-
enantiomers represent the eutomers, for selected triazole derivatives both enantiomers
should be synthesized and tested for inhibitory activity against LpxC and antibacterial
properties.

Figure 2: Structures of reported and envisaged LpxC inhibitors.

2
. Results and Discussion
.1. Chemistry
2
The envisaged triazole derivatives were synthesized from 4-bromostyrene (13) and 4-
benzyloxystyrene (12). The latter compound could be easily accessed from 4-
(benzyloxy)benzaldehyde (11) via a Wittig reaction with methyltriphenylphosphonium
bromide and potassium tert-butoxide (Scheme 1). To obtain the (S)-configured vicinal
diols 14 and 15, styrene derivatives 12 and 13 were subjected to asymmetric Sharpless
dihydroxylation reactions employing AD-mix-α. Subsequently, the primary alcohol
groups were transformed into better leaving groups by performing Bu SnO-catalyzed
2
regioselective tosylations of diols 14 and 15. Then, the thereby obtained tosylates 16
and 17 were transformed into azides 18 and 19 via nucleophilic substitutions with
sodium azide. Alkylation of the secondary alcohols 18 and 19 with ethyl bromoacetate
gave the benzyloxyacetic acid derivatives 20 and 21, representing the central
intermediates of the envisaged syntheses. ent-21, the (R)-configured enantiomer of
21, was obtained from 4-bromostyrene (13) in principally the same way by using AD-
mix-β in the dihydroxylation step.
Scheme 1: Reagents and conditions: (a) methyltriphenylphosphonium bromide,
KOtBu, THF, -10 °C → rt, 12 94%; (b) AD-mix-α, tBuOH/H O (1:1), 0 °C, 14 80%, 15
2
88%; (c) p-TsCl, Bu SnO, NEt , THF, rt, 16 76%, 17 85%; (d) NaN , DMSO, 80 °C, 18
2 3 3
7
0%, 19 85%; (e) ethyl bromoacetate, base, NBu I, DMF, 20 68%, 21 64%.
4

In order to obtain the 4-phenyl-substituted 1,2,3-triazole derivatives 23 and 27, azide
0 was subjected to a Cu(I)-catalyzed azide-alkyne cycloaddition with phenylacetylene
Scheme 2). The resulting triazole derivative 22 could be transformed into hydroxamic
2
(
acid 23 by performing an aminolysis with hydroxylamine. Additionally, ester 22 was
used to access diphenylacetylene derivative 27. At first, its benzyl protective group was
hydrogenolytically cleaved to obtain phenol 24. Subsequently, phenol 24 was treated
with trifluoromethanesulfonic anhydride to give triflate 25, which was then subjected to
a Sonogashira coupling with phenylacetylene yielding diphenylacetylene derivative 26.
A final aminolysis of ester 26 with hydroxylamine gave hydroxamic acid 27.
Scheme 2: Reagents and conditions: (a) phenylacetylene, sodium ascorbate, CuSO4,
tBuOH/H O (1:1), rt, 92%; (b) NH OH·HCl, NaOMe, MeOH, rt, 26%; (c) H , Pd/C,
2
2
2
MeOH, rt, 78%; (d) Tf O, NEt , CH Cl , -20 °C, 92%; (e) phenylacetylene, Pd(PPh ) ,
2
3
2
2
3 4
CuI, NEt , Δ, 63%; (f) NH OH·HCl, NaOMe, MeOH, rt, 54%; (g) trimethylsilylacetylene,
3
2
sodium ascorbate, CuSO , tBuOH/H O (1:1), rt, 38%; (h) NH OH·HCl, NaOMe, MeOH,
4
2
2
rt, 64%.

Benzyl ether 29 was also accessed from azide 20 (Scheme 2). In order to obtain the
respective triazole derivative, a Cu(I)-catalyzed azide-alkyne cycloaddition with azide
20 and trimethylsilylacetylene was performed. However, under the conditions of the
cycloaddition, additionally the cleavage of the trimethylsilyl protective group occurred
yielding the 1-monosubstituted triazole derivative 28. Subsequently, ester 28 was
transformed into hydroxamic acid 29 by treating the compound with hydroxylamine.
In order to shorten the synthesis of the respective diphenylacetylene derivative 32, the
hydroxamic acid was synthesized from 4-bromophenyl derivative 21 (Scheme 3). After
reacting azide 21 with trimethylsilylacetylene in a Cu(I)-catalyzed azide-alkyne
cycloaddition, the resulting crude triazole derivative was directly treated with
tetrabutylammonium fluoride to completely cleave the trimethylsilyl protective group.
The thereby obtained 1-monosubstituted triazole derivative 30 was subjected to a
Sonogashira coupling with phenylacetylene. The reaction was performed in refluxing
triethylamine using tetrakis(triphenylphosphine)palladium(0) and copper(I) iodide as
catalysts. The resulting diphenylacetylene derivative 31 was finally transformed into
hydroxamate 32 by performing an aminolysis with hydroxylamine.
The 4-hydroxymethyl-substituted 1,2,3-triazole derivative 35 was obtained in
principally the same way. After a Cu(I)-catalyzed azide-alkyne cycloaddition of azide
21 and propargyl alcohol, the resulting triazole derivative 33 was coupled with
phenylacetylene. In the Sonogashira coupling of aryl bromide 33
bis(triphenylphosphine)palladium(II) dichloride and copper(I) iodide were used as
catalysts and the transformation was conducted in a refluxing mixture of
diisopropylamine and THF. Under these conditions, a higher yield could be obtained
compared to the respective C-C coupling of aryl bromide 30. The obtained
diphenylacetylene derivative 34 was subjected to a final aminolysis with
hydroxylamine, yielding hydroxamic acid 35.

Scheme 3: Reagents and conditions: (a) 1. trimethylsilylacetylene, sodium ascorbate,
CuSO , tBuOH/H O (1:1), rt, 2. NBu F, rt, 92%; (b) phenylacetylene, Pd(PPh ) , CuI,
4
2
4
3 4
NEt , Δ, 58%; (c) NH OH·HCl, NaOMe, MeOH, rt, 72%; (d) propargyl alcohol, sodium
3
2
ascorbate, CuSO , tBuOH/H O (1:1), rt, 93%; (e) phenylacetylene, PdCl (PPh ) , CuI,
4
2
2
3 2
diisopropylamine, THF, Δ, 93%; (f) NH OH·HCl, NaOMe, MeOH, rt, 52%; (g)
2
phenylacetylene, Cp*RuCl(PPh ) , THF, Δ, 61%; (h) 1. phenylacetylene, PdCl (PPh ) ,
3
2
2
3 2
CuI, diisopropylamine, THF, Δ, 2. NH OH·HCl, NaOMe, MeOH, rt, 25%; (i) propargyl
2
alcohol, Cp*RuCl(PPh ) , dioxane, 60 °C (microwave), 26%; (j) phenylacetylene,
3
2
PdCl (PPh ) , CuI, THF, DIPA, 80 °C, 71%; (k) NH OH aq., iPrOH/THF, rt, 73%.
2
3 2
2

The 1,5-disubstituted 1,2,3-triazole derivatives 37 and 40 were finally obtained via
ruthenium-catalyzed azide-alkyne cycloadditions. The cycloaddition of azide 21 and
phenylacetylene gave the 5-phenyl-substituted 1,2,3-triazole 36, which was subjected
to a Sonogashira coupling and a subsequent aminolysis with hydroxylamine to yield
hydroxamic acid 37. The 5-hydroxymethyl-substituted 1,2,3-triazole derivative 38 was
obtained via a ruthenium-catalyzed azide-alkyne cycloaddition of azide 21 and
propargyl alcohol. A subsequent Sonogashira coupling of aryl bromide 38 with
phenylacetylene yielded diphenylacetylene derivative 39. The performance of the final
aminolysis in a mixture of isopropanol and THF, employing a 50% aqueous solution of
hydroxylamine, gave the desired hydroxamic acid 40.
Scheme 4: Reagents and conditions: (a) O-THP-hydroxylamine, LiHMDS, THF, -
7
8 °C, 90%; (b) phenylacetylene, sodium ascorbate, CuSO , tBuOH/H O (1:1), rt, 82%;
4 2
(
c) phenylacetylene, Pd(PPh ) , CuI, NEt , Δ, 64%; (d) HCl, MeOH, rt, 38%; (e)
3 4 3
acetylene, sodium ascorbate, CuSO , tBuOH/H O (1:1), rt, 48%; (f) phenylacetylene,
4
2
Pd(PPh ) , CuI, NEt , Δ, 24%; (g) HCl, MeOH, rt, 62%; (h) NH OH·HCl, NaOMe,
3
4
3
2
MeOH, rt, 42%.

ent-27 and ent-32, the (R)-configured enantiomers of 27 and 32, were synthesized
from ent-21 via a different synthetic route to test the feasibility of establishing the
hydroxamate moiety at an earlier reaction step. At first, ester ent-21 was reacted with
O-THP-hydroxylamine yielding the tetrahydropyranyl-protected hydroxamic acid ent-
41. After the establishment of the 4-phenyl-substituted 1,2,3-triazole ring and the
diphenylacetylene side chain, the THP protective group of compound ent-43 was
cleaved under acidic conditions yielding hydroxamic acid ent-27. Similarly, in order to
obtain hydroxamic acid ent-32, azide ent-41 was reacted with gaseous acetylene in a
Cu(I)-catalyzed azide-alkyne cycloaddition. Subsequently, a Sonogashira coupling of
the resulting aryl bromide ent-44 with phenylacetylene was performed. Finally, the THP
protective group of the obtained diphenylacetylene derivative ent-45 should be
cleaved. However, the reaction of the 1-monosubstituted triazole derivative in HCl-
saturated methanol led to a solvolysis yielding methyl ester ent-46. From this
compound, the desired hydroxamic acid ent-32 could be accessed via an aminolysis
with hydroxylamine.

2.2. Biological evaluation
Table 1: Antibacterial and LpxC inhibitory activities of the investigated hydroxamic acids.
*
The compound could not be tested at a concentration of 200 µM due to solubility issues.
R'
O
HO
O
zone of inhibition [mm]
MIC [µg/mL]
IC [µM]
N
50
H
cmpd.
R
config.
R
R’
E. coli
E. coli
BL21(DE3)
E. coli
D22
E. coli
LpxCC63A
E. coli D22
BL21(DE3)
(
S)
<6
<6
21.0 ± 1.0
11.0 ± 1.4
>64
1
36.0 ± 1.7
>200
3
2
N
N
N
(R)
>64
>64
32
ent-32
N
N
(
S)
S)
<6
15.0 ± 2.0
>64
>200
2
9
7
O
N
(
<6
<6
12.3 ± 0.6
8.7 ± 1.2
>64
>64
0.25
32
10.4 ± 9.7
>200
2
N
N
N
(R)
ent-27
N
N
(S)
<6
10.0 ± 1.0
>64
32
>200
23
O
N

(S)
<6
10.7 ± 0.6
>64
8
>20*
37
N
N
N
N
N
(
(
(
S)
S)
S)
<6
20.7 ± 0.6
29.5 ± 3.9
>64
>64
4
1
8.5 ± 1.3
3
5
0
OH
OH
N
10.3 ± 1.2
23.4 ± 5.8
4
N
N
N
9.5 ± 0.4
<6
20.5 ± 0.2
13.0 ± 1.7
37.5 ± 1.5
64
1
86.2 ± 1.9
>200
9
OH
(R)
>64
0.25
32
ent-9
3
1.7 ± 1.7
0.032
0.076 ± 0.003
CHIR-090 (5)

The antibacterial properties of lead compound 9 and the synthesized triazole
derivatives were evaluated by performing broth microdilution susceptibility tests as well
as disk diffusion assays against E. coli BL21(DE3) and the defective E. coli D22 strain,
which due to an impaired lipid A biosynthesis is more sensitive towards various
41
antibiotics (Table 1). Additionally, a fluorescence-based in vitro enzyme assay was
carried out to determine the inhibitory activity of the hydroxamic acids toward LpxC.29
The biological evaluation of the triazole derivatives revealed that the (S)-configured
hydroxamic acids 32 and 27 show superior antibacterial as well as LpxC inhibitory
activities compared to their (R)-configured enantiomers ent-32 and ent-27. The same
trend can be also observed in case of lead compound 9 and its enantiomer ent-9.
Additionally, it was found that triazole derivatives 29 and 23, possessing a flexible 4-
(
benzyloxy)phenyl-based lipophilic side chain, show diminished antibacterial and LpxC
inhibitory activities compared to the respective diphenylacetylene derivatives 32 and
7, which exhibit a linear and rigid side chain. Both observations are in agreement with
2
our previous findings.40, 42
The reevaluation of the LpxC inhibitory activity of lead compound 9 revealed an IC -
50
value being considerably higher than the one observed in previous investigations.40
The comparison of the biological activities of the 1-monosubstituted triazole derivative
32 with the ones of lead compound 9 gave contradictory results. Whereas the
replacement of the hydroxy group of compound 9 by a triazole ring leads to an increase
in LpxC inhibitory activity, no change in the antibacterial activity against E. coli D22
and a decrease in the diameter of the zone of growth inhibition in the disk diffusion
assays against E. coli BL21(DE3) can be observed.
The introduction of a phenyl substituent in position 4 of the triazole ring, leading to
compound 27, seems to be beneficial as it causes an increase in inhibitory activity
toward LpxC and a reduction of the MIC against E. coli D22 in comparison with the 1-
monosubstituted triazole derivative 32. The decreased diameter of the zone of growth
inhibition against the latter strain might result from a decreased diffusion rate of the
substituted compound in the agar medium. In contrast, the 5-phenyl-substituted
derivative 37 was found to be unable to inhibit the enzymatic activity of LpxC at the
highest concentration tested and to exhibit a higher MIC-value against E. coli D22 than
the 1-monosubstituted triazole derivative 32 as well as the 1,4-disubstituted derivative
27.

With respect to the inhibitory activity toward LpxC, similar trends can be observed for
the hydroxymethyl-substituted triazole derivatives 35 and 40, with the 1,4-disubstituted
compound 35 being a more potent LpxC inhibitor than its 1,5-disubstituted regioisomer
40. Both hydroxymethyl-substituted compounds exceed the inhibitory activity of the
4,5-unsubstituted triazole derivative 32. When assaying the antibacterial activities of
regioisomers 35 and 40 contradictory results were observed. Whereas the 1,5-
disubstituted triazole derivative 40 was found to exhibit the same MIC-value against E.
coli D22 as compound 32, the more potent LpxC inhibitor 35 was shown to exhibit an
increased MIC-value. Both hydroxymethyl-substituted compounds caused larger halos
of growth inhibition in the disc diffusion assays than the respective phenyl-substituted
compounds, indicating that the presence of a polar hydroxy group leads to a higher
diffusion rate in the agar medium.
Most of the synthesized (S)-configured triazole derivatives exhibiting a
diphenylacetylene-based side chain are more potent LpxC inhibitors than lead
compound 9. Among those compounds, the 4-phenyl-substituted derivative 27 shows
a particularly lower MIC-value against E. coli D22, whereas the 5-hydroxymethyl-
substituted triazole derivative 40 caused considerably larger halos of growth inhibition
in the disc diffusion assays against E. coli D22 as well as E. coli BL21(DE3) than lead
compound 9. However, all triazole derivatives were significantly less potent with
respect to antibacterial and inhibitory activity than reference compound CHIR-090 (5).

2.3. Molecular docking studies
In order to rationalize the experimentally obtained IC -values for E. coli LpxC,
50
molecular docking of all synthesized compounds into the LpxC structure (PDB ID:
P3G) co-crystallized with the potent inhibitor LPC-009 (6, Figure 2) was performed.36
3
All ligands except the two inactive compounds 23 and 29 could be docked into the
binding pocket of LpxC. Only 23 and 29 possess a 4-(benzyloxy)phenyl group instead
of a diphenylacetylene moiety, which did not fit to the hydrophobic tunnel, although
different docking protocols were tried. The lipophilic distal parts of all other compounds
are placed in the hydrophobic tunnel formed by L18, F192, I198, F212, and V217. The
2+
hydroxamic acid group of the docked inhibitors shows a bidentate chelation of the Zn -
ion. The hydroxy and carbonyl group of the hydroxamate moiety in all compounds
interact with the side chains of E78 and T191, respectively. Different interactions,
however, are observed for the middle part of the inhibitors, which together with
calculated binding free energy values allow to discriminate active and inactive
compounds.
The (S)-configured hydroxamic acid 32 having a monosubstituted triazole ring and its
enantiomer ent-32 exhibit a similar binding mode in LpxC (Fig. S1, Supporting
information). In both cases, the triazole ring is directed to the UDP-binding site and is
placed in a small pocket formed by conserved hydrophobic patch residue F192 and
basic patch residue K239. The triazole ring of the active compound 32 is turned toward
F192 and is able to make an edge-to-face aromatic interaction similar to the potent E.
coli LpxC inhibitor LPC-051 (7, Figure 2) in its crystal structure of LpxC from
Pseudomonas aeruginosa (PDB ID: 4LCH).39 In contrast, the triazole ring of the
inactive compound ent-32 is oriented less favorably and the aromatic interaction with
F192 is not observed. The other difference observed is the solvent exposure of the
carbon-oxygen-carbon (C-O-C) linker. In the active compound 32 the oxygen atom is
directed toward T191 while the carbon atoms are buried in the hydrophobic part of the
pocket. In the inactive compound ent-32, the orientation of these atoms is reversed.
That is, the oxygen atom is directed toward the hydrophobic C63 residue, while the
carbon atoms are solvent-exposed.

Figure 3: Predicted binding mode of LpxC inhibitors 27 (colored cyan, A) and 35
colored magenta, B). The surface of the binding pocket is displayed and colored grey.
(
LpxC-inhibitor hydrogen bonds and Zinc-coordination are shown as orange colored
lines.
The active inhibitor 27 having a 4-(phenyl)triazole group and its (R)-isomer ent-27
demonstrate docking poses similar to their analogues 32 and ent-32. The C-O-C linker
and the triazole ring of 27 are placed in the same region as observed for 32. The
triazole ring is making aromatic interactions with F192, while the attached phenyl ring
is additionally involved in cation- interactions with K239. These interactions might
contribute to the improved activity of 27 which is also reflected by a more favourable
calculated binding energy (-159.1 kcal/mol) (Figure 3, Table 2). The triazole ring of the
inactive compound ent-27 is flipped due to the bulky phenyl substituent and the
aromatic interaction with F192 is lost (Fig. S1, Supporting information). The aromatic
interaction with F192 seems to play a crucial role for the activity of the studied
inhibitors. This also explains why the replacement of the hydroxy group of 9 by a
triazole ring in 32 increased the inhibitory activity. Due to the lack of an aromatic
substituent in 9, the interaction with F192 is missing, the UDP-binding site remains
unoccupied, and the calculated interaction energy value is increased (-143.6 kcal/mol).

The hydroxy group of 9 forms hydrogen bonds with the backbone carbonyl oxygen of
M61 and the backbone nitrogen of C63 as reported previously.33
In 37 the shifted position of the phenyl ring from position 4 to 5 of the triazole ring
prevents the ligand from making an interaction with F192, which results in decreased
activity. The triazole ring of 37 is turned toward C63 and is not engaged in any direct
interaction (Fig. S1, Supporting information). The binding energy calculation confirms
the less favourable binding mode of 37 (-138.1 kcal/mol).
The same trend is observed for the hydroxymethyl-substituted compounds when
shifting the hydroxymethyl group from position 4 in compound 35 to position 5 in 40
which slightly reduces the activity (Figure 3). The orientation of the triazole ring in the
most active compound from this series, 35, is similar to the other active triazole
containing derivatives 32 and 27. In addition, the hydroxy group attached to the triazole
ring in 35 is engaged in a hydrogen bond interaction with K239. An interaction with this
residue is also observed for the phosphate and glucosamine moieties of the
deacetylated substrate 3 (Figure 1) as well as for the potent reference inhibitor LPC-
0
09 (6, Figure 2).25, 36 In contrast, the triazole ring of 40 is flipped by 180 degrees (Fig.
S1, Supporting information). The triazole orientation observed for other active
compounds leading to an interaction with F192 is not possible due to the
hydroxymethyl substituent at position 5. Nevertheless, the triazole ring of 40 stays in
the UDP-binding site. Its hydroxy group interacts with the backbone carbonyl of M61
and backbone nitrogen of C63 like in case of 9. Additionally, the nitrogen of the triazole
ring is interacting with K239. These three hydrogen bond interactions are probably
compensating for the lost aromatic interaction with F192. Favourable interaction
energies have been calculated for both active compounds 35 (-150.6 kcal/mol) and 40
(-159.1 kcal/mol).
The calculated interaction energy values and docking scores are listed in Table 2.
Whereas the docking scores were not able to discriminate between active and inactive
inhibitors, the calculated binding energies worked well. More favourable binding free
energies (< -150 kcal/mol) were observed for the active inhibitors compared to the
inactive ones. The only exception is the moderately active non-triazole inhibitor 9 (-
143.6 kcal/mol), which, however gave better binding energy value than its inactive
enantiomer ent-9 (-136.7 kcal/mol).

Table 2: Calculated docking scores and protein-ligand interaction energies
(Amber12:EHT/MOE) of the studied compounds.
calculated ligand
interaction
binding free
experimental
Glide docking
score
compound
energy
pIC50
energy ΔH

G [kcal/mol]
[kcal/mol]
4
.4
-6.07
> -5.10
> -5.10
6.89
-6.7
-6.8
n.a.
-6.3
-6.4
n.a.
-5.4
-6.7
-6.7
-6.5
-6.4
-150.2
3
2
<
<
3.7
3.7
-141.7
n.a.
ent-32
2
9
7
5.0
-159.1
-144.1
n.a.
2
<
<
<
3.7
3.7
4.7
> -5.10
> -5.10
> -6.48
-7.03
ent-27
23
37
35
40
-138.1
-150.6
-159.1
-143.6
-136.7
5
.1
.6
.1
4
-6.34
4
-5.65
9
<
3.7
> -5.10
ent-9

3. Discussion and conclusions
In divergent syntheses, a series of triazole-based LpxC inhibitors was accessed.
Starting from the enantiomerically pure azides 20, 21, and ent-21, which represent the
central intermediates of the synthetic routes, via copper- or ruthenium-catalyzed azide-
alkyne cycloadditions, Sonogashira couplings with phenylacetylene and final
aminolyses with hydroxylamine, most of the synthesized hydroxamic acids were
obtained. When synthesizing the various hydroxamic acids, different reaction
conditions for analogous reaction steps were employed.
Thus, it was observed, that when performing the Sonogashira coupling of the aryl
bromides in the presence of bis(triphenylphosphine)palladium(II) chloride in a mixture
of THF and diisopropylamine (cf. syntheses of 34 and 39) higher yields were obtained
compared to the usage of tetrakis(triphenylphosphine)palladium(0) as catalyst and
triethylamine as solvent (cf. syntheses of ent-43, ent-45, and 31). The route via triflate
25 was found to be disadvantageous. Although the triflate should be more reactive
than the respective aryl bromides, the Sonogashira coupling of compound 25 gave the
desired diphenylacetylene derivative in a similar yield as the analogous coupling of aryl
bromide ent-42. Additionally, this synthetic route required further reaction steps,
including the debenzylation of ether 22, the triflation of phenol 24, and the synthesis of
styrene derivative 12, which in contrast to 4-bromostyrene (13) was not commercially
available.
Most of the performed aminolyses with hydroxylamine were performed employing
hydroxylamine hydrochloride and sodium methoxide in methanol and gave the final
hydroxamic acids in variable yields ranging from 26% to 72%. However, when the
reaction was performed in a mixture of isopropanol and THF using an aqueous solution
of hydroxylamine, hydroxamic acid 40 was obtained in a relatively high yield (73%). In
an alternative strategy, a THP-protected hydroxamate moiety was established in an
early reaction step and the protective group should be cleaved in the end. Whereas
the reaction of ester ent-21 with O-(tetrahydro-2H-pyran-2-yl)hydroxylamine gave
THP-protected hydroxamic acid ent-41 in high yield (90%), the subsequent acid-
catalyzed cleavage of the protective group in case of triazole derivative ent-43 gave
hydroxamic acid ent-27 in only 38% yield and completely failed in case of triazole
derivative ent-45 yielding ester ent-46 instead.

When synthesizing the disubstituted triazole derivatives, the performed Cu(I)-catalyzed
azide-alkyne cycloadditions gave the 1,4-disubstituted regioisomers in higher yields
than the ruthenium-catalyzed cycloadditions leading to the respective 1,5-disubstituted
triazole derivatives. When synthesizing the 1-monosubstituted triazole derivatives, it
was found that a two-step sequence including a Cu(I)-catalyzed azide-alkyne
cycloaddition with trimethylsilylacetylene and a subsequent desilylation (cf. syntheses
of 30) gave the desired triazole derivative in a higher yield than the direct coupling with
acetylene (cf. syntheses of ent-44).
The biological evaluation of the triazolyl-substituted benzyloxyacetohydroxamic acids
revealed that (S)-configuration at the stereocenter as well as a linear and rigid
diphenylacetylene-based side chain lead to superior antibacterial as well as LpxC
inhibitory activities.
The replacement of the hydroxy group of lead compound 9 by a triazole ring generally
led to an increase in inhibitory activity against LpxC. Particularly, the 1,4-disubstituted
triazole derivatives 27 and 35, bearing a phenyl ring and a hydroxymethyl group in
position 4, respectively, showed relatively high inhibitory activities, exhibiting lower IC50
values than their 1,5-disubstituted regioisomers as well as the 4,5-unsubstituted
triazole derivative 32.
With respect to antibacterial activities, the 4-phenyl substituted triazole derivative 27
exhibited a considerably low MIC value against the defective E. coli D22 strain,
whereas the 5-hydroxymethyl-substituted compound 40 was found to cause the largest
halos of growth inhibition in the disc diffusion assays against E. coli BL21(DE3) and
the D22 strain.
Molecular docking studies revealed that aromatic interactions between the triazole ring
of the synthesized compounds and F192, which is part of a highly conserved
hydrophobic patch within the active site of LpxC, play a crucial role for the activity of
the studied inhibitors. In the predicted binding modes, the triazole ring of the active
compounds 27, 32, and 35 is turned toward this important residue and is able to make
edge-to-face aromatic interactions with it. Additionally, interactions of the substituent
at the triazole ring with the basic patch residue K239 (cation- or hydrogen bond
interactions) seem to enhance activity.

Finally, in case of the studied benzyloxyacetohydroxamic acids, active inhibitors could
be discriminated from inactive ones based on their calculated protein-ligand interaction
energy values.
In conclusion, the triazole moiety was found to be a suitable functional group to replace
the hydroxy group of hydroxamic acid 9 and to link various substituents to its main
scaffold, which can address additional binding sites of LpxC.

4. Experimental Section
Chemistry, general
Unless otherwise mentioned, THF was dried with sodium/benzophenone and was
freshly distilled before use. Thin layer chromatography (TLC): Silica gel 60 F254 plates
(Merck). Reversed phase thin layer chromatography (RP-TLC): Silica gel 60 RP-18
F S plates (Merck). Flash chromatography (fc): Silica gel 60, 40 – 64 µm (Macherey-
254
Nagel); parentheses include: diameter of the column, fraction size, eluent, R value.
f
Automatic flash column chromatography: IsoleraTM One (Biotage ); brackets include:
eluent, cartridge-type; product-containing fractions were freeze-dried using a Christ
Alpha 2-4 LDplus freeze-dryer. Microwave assisted reactions were conducted in a
CEM discover system in closed-vessel mode. Melting point (m.p.): Melting point
apparatus SMP 3 (Stuart Scientific), uncorrected. Optical rotation α [deg] was
determined with a Polarimeter 341 (Perkin Elmer); path length 1 dm, wavelength 589
®
nm (sodium D line); the unit of the specific rotation [훼] [deg mL dm g-1] is omitted;
2
0
.
.
-1 .
퐷
.
-1
the concentration of the sample c [mg mL ] and the solvent used are given in
1
13
brackets. H NMR (400 MHz), C NMR (100 MHz): Mercury plus 400 spectrometer
(Varian); δ in ppm related to tetramethylsilane; coupling constants are given with 0.5
Hz resolution. IR: IR Prestige-21(Shimadzu). HRMS: MicrOTOF-QII (Bruker). HPLC
methods for the determination of product purity: Method 1: Merck Hitachi Equipment;
UV detector: L-7400; autosampler: L-7200; pump: L-7100; degasser: L-7614; column:
®
®
LiChrospher 60 RP-select B (5 μm); LiChroCART 250-4 mm cartridge; flow rate: 1.00
mL/min; injection volume: 5.0 µL; detection at λ = 210 nm for 30 min; solvents: A: water
with 0.05% (V/V) trifluoroacetic acid; B: acetonitrile with 0.05% (V/V) trifluoroacetic
acid: gradient elution: (A%): 0 – 4 min: 90% , 4 – 29 min: gradient from 90% to 0%, 29
–
31 min: 0%, 31 – 31.5 min: gradient from 0% to 90%, 31.5 – 40 min: 90%. Method

2: Merck Hitachi Equipment; UV detector: L-7400; pump: L-6200A; column:
®
phenomenex Gemini 5 µm C6-Phenyl 110 Å; LC Column 250 × 4.6 mm; flow rate:
1.00 mL/min; injection volume: 5.0 µL; detection at λ = 254 nm for 20 min; solvents: A:
acetonitrile : 10 mM ammonium formate = 10 : 90 with 0.1% formic acid; B: acetonitrile
:
10 mM ammonium formate = 90 : 10 with 0.1% formic acid; gradient elution: (A%): 0
5 min: 100% , 5 – 15 min: gradient from 100% to 0%, 15 – 20 min: 0%, 20 – 22 min:
–
gradient from 0% to 100%, 22 – 30 min: 100%.
Synthetic procedures
1-(Benzyloxy)-4-vinylbenzene (12)
Under N atmosphere, a solution of methyltriphenylphosphonium bromide (10 g, 28
2
mmol) in dry THF (100 mL) was cooled to -10 °C and potassium tert-butoxide (4.0 g,
35 mmol) was added. After 5 min, 4-(benzyloxy)benzaldehyde (5.0 g, 24 mmol) was
added and the reaction mixture was stirred at ambient temperature for 2 h. Then, water
was added and the mixture was extracted with ethyl acetate (3×). The combined
organic layers were dried over Na SO , filtered, and the solvent was evaporated in
2
4
vacuo. The residue was purified by flash column chromatography (Ø = 6 cm, h = 15
cm, V = 50 mL, cyclohexane/ethyl acetate = 95/5, R = 0.29) to give 12 as colorless
f
1
solid (4.6 g, 22 mmol, 94%). m.p. = 67 °C; H NMR (D COD): δ [ppm] = 5.07 (s, 2H,
3
PhCH OPh), 5.08 (dd, J = 10.9/1.1 Hz, 1H, CH=CH ), 5.61 (dd, J = 17.6/1.1 Hz, 1H,
2
2
CH=CH ), 6.66 (dd, J = 17.6/10.9 Hz, 1H, CH=CH ), 6.92 – 6.96 (m, 2H, Harom.), 7.27
2
2
13
–
7.39 (m, 5H, Harom.), 7.41 – 7.44 (m, 2H, Harom.); C NMR (D COD): δ [ppm] = 71.0
3
(
1C, PhCH O), 111.7 (1C, CH=CH ), 115.9 (2C, Carom.), 128.4 (2C, Carom.), 128.5 (2C,
2 2
Carom.), 128.9 (1C, Carom.), 129.5 (2C, Carom.), 132.1 (1C, Carom.), 137.6 (1C, CH=CH2),
-1
1
38.7 (1C, Carom.), 160.0 (1C, Carom.); IR (neat): 휈 [cm ] = 3036, 2859, 1605, 1508,

+
1
454, 1381, 1238, 1169, 1022, 995, 899, 833, 729, 694; HRMS (m/z): [M+H] calcd for
C H O: 211.1117, found: 211.1113; HPLC (method 1): t = 21.5 min, purity 97.1%.
15
15
R
(S)-1-[4-(Benzyloxy)phenyl]ethane-1,2-diol (14)
AD-mix-α (28 g) was added to a mixture of tert-butyl alcohol (100 mL) and water (100
mL). The mixture was cooled to 0 °C, 12 (4.2 g, 20 mmol) and TBME (23.7 mL) were
added. After stirring at 0 °C for 16 h, sodium sulfite (30 g, 240 mmol) was added and
the reaction mixture was stirred for 1 h. Ethyl acetate was added and after the
separation of the layers, the aqueous phase was further extracted with ethyl acetate
(
3×). The combined organic layers were dried over Na SO , filtered, and the solvent
2 4
was removed in vacuo. The residue was purified by flash column chromatography (Ø
=
6 cm, h = 15 cm, V= 50 mL, cyclohexane/ethyl acetate = 1/2, R = 0.48) to give 14
f
2
D
0
as colorless solid (3.9 g, 16 mmol, 80%). m.p. = 130 °C;



= +27.5 (7.8, methanol);
1
H NMR (D COD): δ [ppm] = 3.59 (d, J = 6.1 Hz, 2H, HOCHCH OH), 4.62 (t, J = 6.1
3
2
Hz, 1H, HOCHCH OH), 5.07 (s, 2H, PhCH OPh), 6.94 – 6.98 (m, 2H, Harom.), 7.26 –
2
2
13
7
.31 (m, 3H, Harom.), 7.33 – 7.38 (m, 2H, Harom.), 7.40 – 7.44 (m, 2H, Harom.); C NMR
(
D COD): δ [ppm] = 68.7 (1C, HOCHCH OH), 71.0 (1C, PhCH OPh), 75.5 (1C,
3
2
2
HOCHCH OH), 115.7 (2C, Carom.), 128.5 (2C, Carom.), 128.6 (2C, Carom.), 128.8 (1C,
2
Carom.), 129.5 (2C, Carom.), 135.6 (1C, Carom.), 138.8 (1C, Carom.), 159.7 (1C, Carom.); IR
-1
(
neat): 휈 [cm ] = 3302, 2905, 2866, 1609, 1508, 1454, 1377, 1238, 1165, 1072, 1042,
+
1
007, 818, 741, 694, 629; HRMS (m/z): [M+Na] calcd for C H NaO : 267.0992,
15
16
3
found: 267.0996; HPLC (method 1): t = 15.8 min, purity 99.8%.
R
(S)-2-[4-(Benzyloxy)phenyl]-2-hydroxyethyl 4-methylbenzenesulfonate (16)

Dibutyltin oxide (0.20 g, 0.8 mmol), p-toluenesulfonyl chloride (3.7 g, 19 mmol) and
triethylamine (6.5 mL, 47 mmol) were added to a solution of 14 (3.5 g, 14 mol) in THF
(75 mL). After stirring the mixture for 16 h at ambient temperature, water was added
and the mixture was extracted with ethyl acetate (3×). The combined organic layers
were dried over Na SO , filtered, and the solvent was evaporated in vacuo. The
2
4
residue was purified by flash column chromatography (Ø = 4 cm, h = 15 cm, V= 30 mL,
cyclohexane/ethyl acetate = 2/1, R = 0.49) to give 16 as colorless solid (4.3 g, 11
f
2
D
0
1
mmol, 76%). m.p. = 89 °C;



= +22.1 (3.4, methanol); H NMR (CDCl ): δ [ppm] =
3
2
1
5
–
.45 (s, 3H, PhCH ), 4.03 (dd, J = 10.4/8.5 Hz, 1H, HOCHCH OSO ), 4.11 (dd, J =
3
2
2
0.4/3.4 Hz, 1H, HOCHCH OSO ), 4.92 (dd, J = 8.5/3.4 Hz, 1H, HOCHCH OSO ),
2
2
2
2
.05 (s, 2H, PhCH OPh), 6.91 – 6.96 (m, 2H, Harom.), 7.21 – 7.24 (m, 2H, Harom.), 7.30
2
13
7.35 (m, 3H, Harom.), 7.36 – 7.44 (m, 4H, Harom.), 7.75 – 7.79 (m, 2H, Harom.); C NMR
(
CDCl ): δ [ppm] = 21.8 (1C, PhCH ), 70.2 (1C, PhCH OPh), 71.7 (1C,
3 3 2
HOCHCH OSO ), 74.5 (1C, HOCHCH OSO ), 115.2 (2C, Carom.), 127.6 (2C, Carom.),
2
2
2
2
127.7 (2C, Carom.), 128.1 (2C, Carom.), 128.2 (1C, Carom.), 128.8 (2C, Carom.), 130.1 (2C,
Carom.), 130.7 (1C, Carom.), 132.8 (1C, Carom.), 136.9 (1C, Carom.), 145.2 (1C, Carom.),
-
1
1
1
4
59.1 (1C, Carom.); IR (neat): 휈 [cm ] = 3480, 3429, 1612, 1512, 1377, 1323, 1246,
+
169, 1096, 1011, 914, 814, 745, 691; HRMS (m/z): [M+NH ] calcd for C H NO S:
4
22 26
5
16.1526, found: 416.1567; HPLC (method 1): t = 20.7 min, purity 96.0%.
R
(
S)-2-Azido-1-[4-(benzyloxy)phenyl]ethan-1-ol (18)
Sodium azide (1.4 g, 21 mmol) was added to a solution of 16 (4.2 g, 11 mmol) in DMSO
53 mL) and the reaction mixture was heated to 80 °C for 3 h. Then, the reaction
(
mixture was cooled to ambient temperature, water was added and the mixture was
extracted with ethyl acetate (3×). The combined organic layers were dried over

Na SO , filtered and the solvent was removed in vacuo. The residue was purified by
2
4
flash column chromatography (Ø = 4 cm, h = 15 cm, V = 30 mL, cyclohexane/ethyl
acetate = 8/2, R = 0.40) to give 18 as colorless solid (2.0 g, 7.4 mmol, 70%). m.p. =
f
2
D
0
1
65 °C;



= +37.8 (2.6, methanol); H NMR (CDCl ): δ [ppm] = 3.41 (dd, J = 12.6/4.0
3
Hz, 1H, HOCHCH N ), 3.48 (dd, J = 12.6/8.2 Hz, 1H, HOCHCH N ), 4.83 (dd, J =
2
3
2
3
8
7
5
.2/4.0 Hz, 1H, HOCHCH N ), 5.07 (s, 2H, PhCH OPh), 6.96 – 7.00 (m, 2H, Harom.),
2 3 2
13
.27 – 7.35 (m, 3H, Harom.), 7.36 – 7.45 (m, 4H, Harom.); C NMR (CDCl ): δ [ppm] =
3
8.2 (1C, HOCHCH N ), 70.2 (1C, PhCH OPh ), 73.2 (1C, HOCHCH N ), 115.2 (2C,
2
3
2
2
3
Carom.), 127.4 (2C, Carom.), 127.6 (2C, Carom.), 128.2 (1C, Carom.), 128.8 (2C, Carom.),
-
1
1
2
8
2
33.1 (1C, Carom.), 137.0 (1C, Carom.), 159.0 (1C, Carom.); IR (neat): 휈 [cm ] = 3414,
978, 2920, 2087, 1609, 1508, 1454, 1385, 1304, 1234, 1173, 1115, 1080, 1057, 1015,
+
80, 864, 826, 748, 698, 625; HRMS (m/z): [M+Na] calcd for C H N NaO :
15
15
3
2
92.1056, found: 292.1057; HPLC (method 1): t = 19.0 min, purity 99.0%.
R
Ethyl (S)-2-{2-azido-1-[4-(benzyloxy)phenyl]ethoxy}acetate (20)
Under N atmosphere, sodium hydride (55% suspension in paraffin oil, 0.10 g, 2.3
2
mmol), tetrabutylammonium iodide (0.13 g, 0.35 mmol) and ethyl bromoacetate (0.78
mL, 7.0 mmol) were added to a solution of 18 (0.94 g, 3.5 mmol) in DMF (30 mL). After
stirring at ambient temperature for 16 h, water was added and the mixture was
extracted with ethyl acetate (3×). The combined organic layers were dried over
Na SO , filtered and the solvent was removed in vacuo. The residue was purified by
2
4
flash column chromatography (Ø = 4 cm, h = 15 cm, V = 30 mL, cyclohexane/ethyl
acetate = 20/1 → 9/1) to give 20 as colorless oil (0.84 g, 2.4 mmol, 68%). TLC
2
D
0
1
(
cyclohexane/ethyl acetate = 8:2): R = 0.54;



= +91.9 (5.4, methanol); H NMR
f
(
CDCl ): δ [ppm] = 1.27 (t, J = 7.2 Hz, 3H, OCH CH ), 3.25 (dd, J = 12.9/4.1 Hz, 1H,
3
2
3

OCHCH N ), 3.59 (dd, J = 12.9/8.1 Hz, 1H, OCHCH N ), 3.91 (d, J = 16.2 Hz, 1H,
2
3
2
3
OCH CO Et), 4.07 (d, J = 16.2 Hz, 1H, OCH CO Et), 4.14 – 4.24 (m, 2H, OCH CH ),
2
2
2
2
2
3
4.62 (dd, J = 8.1/4.1 Hz, 1H, OCHCH N ), 5.07 (s, 2H, PhCH OPh), 6.96 – 7.00 (m,
2 3 2
13
2
H, Harom.), 7.23 – 7.28 (m, 2H, Harom.), 7.31 – 7.45 (m, 5H, Harom.); C NMR (CDCl ):
3
δ [ppm] = 14.3 (1C, OCH CH ), 56.5 (1C, OCHCH N ), 61.2 (1C, OCH CH ), 65.9
2
3
2
3
2
3
(
1C, OCH CO Et), 70.2 (1C, PhCH OPh), 81.3 (1C, OCHCH N ), 115.3 (2C, Carom.),
2 2 2 2 3
127.6 (2C, Carom.), 128.2 (1C, Carom.), 128.4 (2C, Carom.), 128.7 (2C, Carom.), 129.9 (1C,
-1
Carom.), 136.9 (1C, Carom.), 159.3 (1C, Carom.), 170.1 (1C, CO Et); IR (neat): 휈 [cm ] =
2
2982, 2099, 1751, 1732, 1609 1508, 1454, 1377, 1285, 1238, 1204, 1173, 1123, 1022,
+
9
18, 864, 829, 737, 698; HRMS (m/z): [M+NH ] calcd for C H N O : 373.1870,
4
19 25
4
4
found: 373.1907; HPLC (method 1): t = 21.1 min, purity 95.6%.
R
Ethyl
(S)-2-{1-[4-(benzyloxy)phenyl]-2-(4-phenyl-1H-1,2,3-triazol-1-
yl)ethoxy}acetate (22)
Phenylacetylene (0.11 mL, 1.0 mmol), sodium ascorbate (20 mg, 0.1 mmol) and
copper(II) sulfate (10 mg, 0.06 mmol) were added to a solution of 20 (200 mg, 0.56
mmol) in a tBuOH/H O mixture (1:1; 45 mL). After stirring at ambient temperature for
2
2
4 h, a saturated aqueous solution of NaHCO was added and the mixture was
3
extracted with ethyl acetate (3×). The combined organic layers were dried over
Na SO , filtered, and the solvent was evaporated in vacuo. The residue was purified
2
4
by flash column chromatography (Ø = 2 cm, h = 15 cm, V= 10 mL, cyclohexane/ethyl
acetate = 2:1, R = 0.30) to give 22 as colourless solid (240 mg, 0.52 mmol, 92%). m.p.
f
2
D
0
1
=
106 °C;



= +42.2 (2.1, methanol); H NMR (CDCl ): δ [ppm] = 1.18 (t, J = 7.2
3
Hz, 3H, OCH CH ), 3.83 (d, J = 16.2 Hz, 1H, OCH CO Et), 4.03 (d, J = 16.2 Hz, 1H,
2
3
2
2
OCH CO Et), 4.08 – 4.16 (m, 2H, OCH CH ), 4.60 (dd, J = 14.2/8.5 Hz, 1H,
2
2
2
3