Synthesis and structural insights into the binding mode of the albomycin δ1 core and its analogues in complex with their target aminoacyl-tRNA synthetase

Article abstract of DOI:10.1016/j.bmc.2020.115645

Despite of proven efficacy and well tolerability, albomycin is not used clinically due to scarcity of material. Several attempts have been made to increase the production of albomycin by chemical or biochemical methods. In the current study, we have synthesized the active moiety of albomycin δ1 and investigated its binding mode to its molecular target seryl-trna synthetase (SerRS). In addition, isoleucyl and aspartyl congeners were prepared to investigate whether the albomycin scaffold can be extrapolated to target other aminoacyl-tRNA synthetases (aaRSs) from both class I and class II aaRSs, respectively. The synthesized analogues were evaluated for their ability to inhibit the corresponding aaRSs by an in vitro aminoacylation experiment using purified enzymes. It was observed that the diastereomer having the 5′S, 6′R-configuration (nucleoside numbering) as observed in the crystal structure, exhibits excellent inhibitory activity in contrast to poor activity of its companion 5′R,6′S-diasteromer obtained as byproduct during synthesis. Moreover, the albomycin core scaffold seems well tolerated for class II aaRSs inhibition compared with class I aaRSs. To understand this bias, we studied X-ray crystal structures of SerRS in complex with the albomycin δ1 core structure 14a, and AspRS in complex with compound 16a. Structural analysis clearly showed that diastereomer selectivity is attributed to the steric restraints of the active site of SerRS and AspRS.

Full text of DOI:10.1016/j.bmc.2020.115645

Journal Pre-proofs
Synthesis and structural insights into the binding mode of the albomycin δ1
core and its analogues in complex with their target aminoacyl-tRNA synthe‐
tase
Bharat Gadakh, Gaston Vondenhoff, Luping Pang, Manesh Nautiyal, Steff De
Graef, Sergei V. Strelkov, Stephen D. Weeks, Arthur Van Aerschot
PII:
DOI:
Reference:
S0968-0896(20)30475-2
https://doi.org/10.1016/j.bmc.2020.115645
BMC 115645
To appear in:
Bioorganic & Medicinal Chemistry
Received Date:
Revised Date:
Accepted Date:
2 June 2020
2 July 2020
6 July 2020
Please cite this article as: Gadakh, B., Vondenhoff, G., Pang, L., Nautiyal, M., De Graef, S., Strelkov, S.V.,
Weeks, S.D., Van Aerschot, A., Synthesis and structural insights into the binding mode of the albomycin δ1 core
and its analogues in complex with their target aminoacyl-tRNA synthetase, Bioorganic & Medicinal Chemistry
(2020), doi: https://doi.org/10.1016/j.bmc.2020.115645
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 structural insights into the binding mode of the
albomycin δ1 core and its analogues in complex with their target
aminoacyl-tRNA synthetase
1
1,3
1,2
1
2,4
Bharat Gadakh , Gaston Vondenhoff , Luping Pang , Manesh Nautiyal , Steff De Graef ,
2
2,4
1,
Sergei V. Strelkov , Stephen D. Weeks and Arthur Van Aerschot *
1
Medicinal Chemistry, Rega Institute for Medical Research, KU Leuven, Herestraat 49 box
1
041, B-3000 Leuven, Belgium
2
Biocrystallography, Department of Pharmaceutical and Pharmacological Sciences, KU Leuven,
Herestraat 49 box 822, B-3000 Leuven, Belgium
3
Present address: Roche Diagnostics GmbH, DXRESA7G6164, Staffelseestrasse 2-8, 81477
Munich, Germany
4
Present address: OrthogonX, Gaston Geenslaan 1, 3001 Leuven, Belgium
E-mail: sweeks@orthogontherapeutics.com
*
To whom correspondence should be addressed. Tel: +32 16 37 26 24; Fax: +32 16 33 73 40;
Email: arthur.vanaerschot@kuleuven.be
1

Abstract
Despite of proven efficacy and well tolerability, albomycin is not used clinically due to
scarcity of material. Several attempts have been made to increase the production of albomycin by
chemical or biochemical methods. In the current study, we have synthesized the active moiety of
albomycin δ1 and investigated its binding mode to its molecular target seryl-trna synthetase
(SerRS). In addition, isoleucyl and aspartyl congeners were prepared to investigate whether the
albomycin scaffold can be extrapolated to target other aminoacyl-tRNA synthetases (aaRSs) from
both class I and class II aaRSs, respectively. The synthesized analogues were evaluated for their
ability to inhibit the corresponding aaRSs by an in vitro aminoacylation experiment using
purified enzymes. It was observed that the diastereomer having the 5’S, 6’R-configuration
(nucleoside numbering) as observed in the crystal structure, exhibits excellent inhibitory activity
in contrast to poor activity of its companion 5’R,6’S-diasteromer obtained as byproduct during
synthesis. Moreover, the albomycin core scaffold seems well tolerated for class II aaRSs
inhibition compared with class I aaRSs. To understand this bias, we studied X-ray crystal
structures of SerRS in complex with the albomycin δ1 core structure 14a, and AspRS in complex
with compound 16a. Structural analysis clearly showed that diastereomer selectivity is attributed
to the steric restraints of the active site of SerRS and AspRS.
Graphical abstract
2

Keywords
Albomycin, albomycin analogues, aminoacyl-tRNA synthetase inhibitors, serine-tRNA
synthetase, X-ray crystallography, structural analysis, natural product synthesis.
Highlights





Total synthesis of albomycin core scaffold targeting serRS was accomplished
Additional aspartyl- and isoleucyl-congeners were elaborated
Aspartyl analogue strongly inhibiting aspRS in contrast to isoleucyl analogue
3D-Structural analysis of the albomycin core scaffold-serRS interactions
Co-crystal structure of the aspartyl analogue and its aspRS target
3

1
. Introduction
Albomycin is a natural antibiotic produced by Streptomyces strain ATCC 700974 as a
mixture of closely related compounds (Fig 1, 1a-c) [1, 2]. It is also known as grisein [3] and was
discovered by Gause and Brazhnikovain 1951 from an Actinomyces subtropicus [4] culture,
while its correct structure was uncovered by Benz et al. in 1982 [5-7]. Albomycin is a Trojan-
horse antibiotic belonging to the sideromycins. Structurally, it consists of the active moiety
linked to a siderophore, an iron chelator essential for cellular uptake. The active moiety (shown in
red) comprises of a thioxylonucleoside linked to L-serine via an amide linkage mimicking the
seryl-adenylate intermediate and thus acts as a seryl-tRNA synthetase (SerRS) inhibitor. The
siderophore part consists of an L-ornithine tripeptide-based trihydroxamate (shown in blue) and
is attached to the active moiety via a cleavable amide bond [8]. Albomycins are classified as δ1,
δ2 and Ɛ based on their substituent at the C4-position of the pyrimidine base. The siderophore
part is responsible for uptake of the active part due to its ability to complex iron and being
recognized by the bacterial iron uptake receptor [8, 9].
Albomycin is highly effective against many gram-positive and gram-negative bacteria
with MIC for the isolated compound of 10 ng/mL against Streptococcus pneumoniae and 5
ng/mL against E. coli [8]. Shortly after its discovery, albomycin has been successfully used in
children to treat infections caused by pathogenic cocci. It proved to be safe and effective to treat
septic complications of dysentery, measles and meningitis caused by penicillin-resistant
pneumococci [4]. More recently, Pramanik et al. re-examined the activity spectrum and in vivo
efficacy of albomycin in a mouse model. Herein, it was found that albomycin is effective to treat
bacterial infections caused by gram-positive and gram-negative pathogens. In mice, the
maximum dose tested was 50 million units/Kg by either intravenous or subcutaneous
administration but without any sign of toxicity [4, 10] (with one unit of albomycin inhibiting the
growth of staphylococci in 1 ml. of nutritive agar, being approximately equal to 1/50 unit of
penicillin [4]). Despite of proven efficacy and safety of albomycin, its use as a potential antibiotic
was abandoned due to scarcity of the material. Therefore, ever since its discovery, efforts have
been made to improve the yield of albomycin production and to develop it into a commercial
drug [11, 12]. Indeed, Benz et al. reported the synthesis of albomycin δ1 analogue replacing
thioxylofuranose with xylofuranose. However, the newly synthesized analogue was reported to
be devoid of antibacterial activity [5]. In 1985, Fiedler et al. optimized the fermentation
4

conditions to increase the yield of albomycin from 1 mg/L to 25 mg/L [13]. However, this
amount is not sufficient for industrial application. Moreover, downstream processing proved
cumbersome to obtain pure albomycin. In 1991, Holzapfel et al. claimed to have synthesized the
thionucleoside moiety of albomycin δ1. However, synthetic details and analytical data were not
disclosed [14].
After identification of the albomycin biosynthesis gene cluster, it was envisioned that
implementation of bioengineering strategies could lead to production of intermediates which can
be used to produce albomycin in sufficient amounts [12]. However, no follow-up has been
disclosed. In parallel with efforts to improve the albomycin production along with better
purification strategies, we focused on a total synthetic strategy for albomycin production. Our
intent therefore was to follow a convergent approach where the active moiety and siderophore
part would be synthesized separately and will be coupled at a later stage. We further
hypothesized that we potentially can target various aminoacyl-tRNA synthetases (aaRSs) by
substituting L-serine with variant amino acids. Recently, He and coworkers focused on an
analogous strategy and successfully demonstrated total synthesis of albomycin and its base
modified congeners [15]. They confirmed the antibacterial potential and proved albomycin δ2
exhibited potent antimicrobial activities against clinical S. pneumoniae and S. aureus isolates
including MRSA. These results prompted us to share the details of our chemical efforts in
elaborating the albomycin δ1 core as well as its functionalization with aspartic acid and
isoleucine, thus successfully targeting other members of the aaRS family. We further disclose
biochemical and structural details of our enzymatic interaction studies, allowing further
improvement of the core structure binding mode via modeling.
2
. Chemistry
2
.1. Retrosynthetic analysis of albomycin
As shown in Figure 1, retrosynthetic analysis indicates the active moiety (A) and the
siderophore part (B) can be synthesized separately and can be coupled by a conventional peptide
coupling strategy. The siderophore part can be synthesized from L-ornithine tripeptide (F)
following indirect oxidation method reported by Miller [16]. The tripeptide (F) itself, is
synthesized from L-ornithine (G) following conventional peptide chemistry. In parallel, the
5

active moiety can be synthesized from 4’-thio-xylonucleoside (C, alcohol) by chain elongation
reaction of the nucleoside aldehyde and appropriately protected glycine (D) [17], while further
elongation is accomplished via reaction with appropriately protected active ester of L-serine (E)
or another amino acid. The albomycin core structure (X) depicts the scaffold before attaching
serine for targeting SerRS, and can be used to couple other amino acids broadening the potential
scope for inhibiting protein synthesis with a series of congeners.
Fig. 1. Structures of the natural albomycin congeners and retrosynthetic analysis of albomycin
δ1. The various starting building blocks are indicated with in the middle the key core structure X
6

enabling synthesis of albomycin analogues via coupling with amino acids at its 6’-amino
terminus.
2
.2. Synthesis of the albomycin δ1 core and its analogues targeting aaRSs
We previously already followed the outlined strategy for the siderophore part and used
this particular uptake module in our efforts for improvement of uptake of some aminoacyl
sulfamoyl adenosine derivatives (aaSAs), which displayed strong enzyme inhibitory activity in
cell extracts [18]. Here, in an effort to understand the binding mode of albomycin in the active
site of SerRS, we focused on synthesis of the active core of albomycin δ1, which is outlined in
o
Scheme 1. L-arabinose upon treatment with in situ generated HCl in methanol at 55 C yielded
methyl arabinoside (3) according to literature procedure [19, 20]. Next, the 3- and 4-hydroxyl
groups were selectively activated via their dibutyltin acetal followed by treatment with tosyl
chloride in presence of DMAP as a base to yield a mixture of regioisomers (4 and 5) in 1:2 ratio
[
21, 22]. The required 4-O-tosyl derivative (5) was taken further for synthesis and converted into
using BzCl in pyridine in presence of catalytic DMAP [23, 24]. The L-arabinoside was further
6
converted to its 4-thio-epimer, thio-D-xylopyranoside, by nucleophilic displacement of 4-O-tosyl
o
using potassium thioacetate at 80 C in DMF as a solvent with little modification to literature
procedure [25, 26]. Acidolysis of methyl glycoside followed by ring contraction (ring opening
and closing) and acetylation was achieved by treatment of 7 with conc. H SO in acetic acid and
2
4
acetic anhydride as solvent to yield 4-thio-D-xylofuranose 8 as a mixture of α/β anomers. [25, 27]
The thio-xylofuranose derivative 8 upon reaction with uracil under silyl-Hilbert-Johnson
glycosylation conditions followed by methylation using MeI yielded compound 10 [28]. Further,
selective removal of 5’-O-acetyl was accomplished by in situ generated HCl using acetyl chloride
in a mixture of methanol and DCM to yield appropriately protected thio-xylonucleoside 11.
7

Scheme 1. Total synthesis of albomycin δ1 core structure (14) and its isoleucine (15) and
aspartate (16) congeners as a mixture of the 5’S,6’R and 5’R,6’S isomers. Reagents and
o
o
conditions: (a) AcCl, MeOH, 0 C to 55 C, 4 h; 74%; (b) i. Bu SnO, dry MeOH, reflux
2
overnight; ii. DMAP, TsCl, 1,4-dioxane, rt, 24 h; 62%; (c) DMAP, BzCl, pyridine, 24 h, 92%; (d)
o
o
o
KSAc, DMF, 80 C, 6 h, 51%; (e) Conc. H SO , acetic acid, acetic anhydride, -10 C to 4 C,
2
4
o
overnight, 75%; (f) BSA, uracil, SnCl , ACN, 40 C, overnight, 72%; (g) K CO , MeI, acetone:
DMF, 70 C, overnight, 71%; (h) AcCl, DCM: MeOH, rt, overnight; 89%; (i) Dess-Martin, DCM,
C, 3 h; (j) i. nBuLi, DIPA, THF, -78 C, 2 h, ii. Boc GlyOMe, THF, -78 C, 1 h, 43%; (k) i.
NaOH, 1,4-dioxane and water, 0 C to rt, 2 h; ii. TFA; (l) Boc-aa-OSu, NaHCO , THF: water, rt,
4
2
3
o
o
o
o
0
2
o
3
overnight; (m) TFA/water (5:2)
The chain elongation reaction is the most challenging step for albomycin total synthesis in
view of its low yield, while obtaining the required stereoselectivity provides the second hurdle.
Numerous attempts were undertaken to optimize the two-step elongation reaction (oxidation and
C-C coupling). Several oxidizing agents such as IBX, Dess-Martin, PCC or Swern oxidation
were investigated to yield the intermediate aldehyde in good yield and purity. The Dess-Martin
periodinane afforded us the best results. Here, Boc GlyOMe was used as C-nucleophile in
2
presence of LDA to yield elongated product 12. The second step was optimized in changing the
equivalents of Boc GlyOMe, LDA, temperature, time and quenching reagent. Finally, the
2
8

elongated product 12 was obtained as a mixture of two diastereomers (~1:1 ratio) in up to 43%
yield (2-step reaction). Having compound 12 (as a mixture of two diastereomers) in hand,
separation of diastereomers was attempted using various HPLC conditions but without success.
We therefore decided to separate the diastereomers at a later stage of synthesis.
Acetylation of 5’-OH was attempted with the aim to be able to remove all esters at a later
stage in a single step. However, protection of 5’-hydroxyl moiety even at increased temperature
proved unsuccessful due to steric hindrance. Leaving the alcohol unprotected, the Boc groups of
1
2 were removed by 30% TFA in DCM at room temperature. However, following amino acid
coupling, we observed that the benzoyl group had migrated from 3'- to 5'-hydroxyl leading to
elimination reaction with the double bond residing between C5'-C6'. We therefore decided to
remove the benzoyl protecting groups before coupling. Unfortunately, deprotection of Boc
followed by removal of the benzoyl esters leads to undesired migration to the 6-amino position to
yield benzamide. Following optimization, one-pot deprotection of benzoyl groups and methyl
ester, followed by removal of Boc yielded most optimal results. However, purification of the
obtained nucleoside derivative 13 proved cumbersome. Notably, He and co-workers published
total synthesis of albomycins, which was enabled using the less known 2,3-butanedione for
protection of sugar hydroxyl moieties and proved more straightforward. We therefore believe the
choice for the right protecting groups is the key for successful albomycin total synthesis [15].
Having nucleoside 13 in hand, we evaluated several reaction conditions to obtain selective
aminoacylation of the 7-amine position. For example, using a transient protection strategy
protecting carboxyl and hydroxyl moieties using TMSCl [29], we found the N-hydroxy
succinimide ester was too weak to effect aminoacylation, whereas aminoacyl chloride is too
reactive and leads to partial racemization during the coupling reaction. Finally, using the N-
hydroxysuccinimide ester in the presence of sodium bicarbonate as the base in a mixture of THF
and water, provided us the desired N-aminoacylated product. Here, NaHCO is forming a salt
3
with the carboxylic acid and thus acts as a transient protection, while being too weak to effect
aminoacylation of the hydroxyl moiety. Using this methodology, we coupled three amino acids,
either serine, isoleucine or aspartic acid, to thionucleoside 13. The coupled products upon
t
acidolysis of Boc and side chain protecting groups (TBDMS in case of serine and Bu in case of
aspartic acid) yielded the final products (14-16) as a mixture of two diastereomers. Reverse phase
HPLC using Adamas C18-X-bond columns with ACN and water containing 50 mM TEAB as
9

eluens, provided separation of the serine and aspartic acid isomers. Isoleucine and aspartic acid
were coupled to the core albomycin scaffold as representative examples for obtaining potential
inhibitors for class I and other class II aaRSs, respectively.
3
. Enzyme inhibitory activity of the synthesized compounds
Without any reference compound in hand, the identity of the purified compounds and
presence of the correct stereoisomer could not be determined based on elution profile, but was
assigned upon evaluation of their enzyme inhibitory activity. Hereto, the isolated derivatives
were tested by an in vitro aminoacylation reaction using purified aaRSs and measuring the
transfer of radiolabelled amino acid to tRNA [30]. We later confirmed the correct
stereochemistry via X-ray analysis.
Aminoacylation reactions were carried out with a total tRNA pool and the appropriate
purified enzymes from E. coli. This obviously is an indirect method of assessing the right
stereoisomer, but we expected to see considerable differences in biological activity, with at least
strong inhibitory activity for the serine isomer corresponding to the albomycin warhead structure.
Indeed, one isomer of the serine congener ser-Xa (14a, 5’S,6’R), showed around 50% enzyme
inhibition at 0.1 M concentration, and 80% at 1 M (Figure 2) while the other isomer ser-Xb
(14b, 5’R,6’S) only displayed about 25% inhibition at 10 M, a more than 100-fold decrease in
activity. We therefore assume 14a to be endowed with the correct stereochemistry displaying
inhibition potential at the 100 nM level. The exact configuration of 14a was proved by
crystallographic data (vide infra). The very strong antibacterial effect of albomycin that was
reported in the past also finds an explanation in the preferential uptake of the compound as of its
conjugation with the siderophore part. The latter was clearly proven again by Lin et al [15].
1
0

Fig. 2. In vitro inhibitory activity of the obtained isomers on their respective aaRS. Results are
expressed as percentage inhibition versus the non-inhibited enzyme. Assay conditions were as
described in Nautiyal et al [31].
Our plan was to use the same scaffold for assembling congeners targeting different aaRSs.
Thus, we used the scarce material available for synthesizing asp-X (16) targeting the class II
AspRS, as well as compound ile-X (15) targeting the class I IleRS. While 16 proved separable in
two isomers 16a and 16b in analogy with 14, compound 15 could not be resolved under our
analytical conditions and therefore was evaluated as a mixture (approximately 50:50). However,
no enzyme inhibitory activity was noted up to 10 M for the ile-X (15) isomeric mixture,
meaning the “correct” isomer likewise is devoid of activity up to around 5 M. The very different
3
D architecture of the active site for class I and II enzymes has been described many times, and
was highlighted again recently in our work on 3-deaza-adenosine inhibitors [30], with complete
loss or poor activity upon targeting class II enzymes, while retaining very potent activity versus
class I enzymes. Here apparently, the albomycin δ1 scaffold is not well suited for the class I
architecture. Analogously,
sulfamoyladenosine by itself already was reported to be detrimental for the strong inhibitory
activity of the parent ligand [30].
a
3-methyluracil substitution of adenine in isoleucyl-
1
1

In contrast, 16a displayed over 60% inhibitory activity for AspRS at 0.1 M concentration, and
was still weakly inhibitory even at 10 nM. The second isomer 16b, as seen with the serine
congeners, was completely inactive even at 1 M concentration. The remarkable affinity for
AspRS as shown by 16a, albeit carrying a 3-methyluracil base, corresponds with our recent
findings on the inhibitory activity of the aspartyl conjugated sulfamoylated 3-methyluridine with
an IC of 31 nM for AspRS [32]. In view of the low amounts of available material, we have not
5
0
app
been able to generate dose-response data in order to derive some Ki values, but instead
preferred to use the material for soaking into corresponding aaRS crystals to characterize the
binding mode of compounds and protein-ligand interactions. Based on the obtained results, we
believe the albomycin core scaffold indeed can be used to target class II aaRS, and enzyme
inhibitory evaluations allowed us to predict the “correct” stereoisomer.
As described in the introduction, antibacterial activity of these polar warheads can only be
attained following attachment of the siderophore part as described many times before. The strong
antibacterial activity of the full albomycin structure was recently highlighted again by Lin et al
[
15].
4
. X-ray crystallographic studies
To further prove the configuration for the obtained steroisomers of the albomycin derivatives,
compounds 14a,b and 16a,b were soaked into Trypanosoma brucei gambiense (Tbg) SerRS and
Thermus thermophilus (Tt) AspRS, respectively. The obtained structural data showed
unambiguous density for compounds 14a and 16a and we noted the desired 5’S, 6’R chirality.
However, for compounds 14b and 16b we only observed electron density for the base moiety,
highlighting their reduced affinity for the target enzyme. The results demonstrate that a 5’S, 6’R
configuration for the albomycin core is required for targeting both SerRS and AspRS, which is as
found in the genuine albomycin structure. In addition, it proves compounds 14a and 16a have the
correct configuration as was already deduced from their more potent inhibitory activity.
4
.1. Structural analysis of protein-ligand interactions for the serine congener 14a
The above-obtained data allowed for structural analysis of 14a and 16a regarding
interactions with the target aaRS protein in comparison with the interactions made by the genuine
aminocyl-adenylate intermediates, Ser-AMP and Asp-AMP, respectively. Such analysis is highly
1
2

interesting in view of future inhibitor design, as we are confronted here with a natural inhibitor
equipped not only with the unusual base 3-methyluracil substituting for adenine, but having as
well a thiolated xylofuranosyl sugar ring instead of the ribose, and a one atom shorter
carboxylated linker substituting for the phosphate moiety. The overall interactions for the serine
containing compounds can be viewed in 2D interaction scheme (Figure 3A) and the Ligplots
(Figure SI-1) for both structures 14a and Ser-AMP. Superposition of both ligands is visualized in
Figure 3B.
Fig. 3. A: 2D interaction of Tbg-serRS with 14a. Hydrogen bonds are shown as purple lines.
Cation-π and π-π interactions are shown as red and green lines, respectively. Figure generated
using Maestro Schrodinger. B: Superposition of Tbg-serRS bound to 14a and ser-AMP. The
enzymes are shown as cartoon representations and colored orange and green, respectively.
Residues within 5Å of the compound are shown as stick model. Compound 14a is colored
magenta and ser-AMP is colored yellow.
4
.2. Comparison of the heterocyclic base interactions as seen for 14a and Ser-AMP
3
With the serine-containing product 14a, the N -methylated uracil base is sandwiched
between the Phe314 and Arg431 sidechain for the Tbg-SerRS complex (Figure 3B and for a
close-up Figure 4C), replicating a π-π interaction with the aryl ring on one face of the heterocycle
and a cation-π interaction on the other face. This is analogous to the positioning of the adenine
ring in the complex structure of Ser-AMP with SerRS (Figure 4C) [32]. The uracil ring is
positioned at the site where the imidazole ring of the adenine base resides, but is tilted around 20
9
1
degrees in comparison with the adenine ring plane. The adenine-N and uracil-N are almost
1
3

coinciding. Apart from this π-π interaction, there are fewer contributions from the modified base
with only the C2 oxygen making a water-bridged interaction with the backbone nitrogen of
3
Arg431, replicating the equivalent N -interaction made with adenine. The same water molecule
likewise bridges with the backbone nitrogen and hydroxyl substituent of Thr430. For the Ser-
AMP compound (Fig 4C) the adenine contributes more to the affinity compared to the
3
6
1
methyluracil ring via hydrogen bonding of N , N and N , and is likewise sandwiched between
Arg431 and Phe314.
4
.3. Comparison of the sugar rings
As seen from the ligplot (Figure SI-1) and the structural close-up (Figure 4B), each
hydroxyl moiety of the 5-membered thioxylofuranosyl ring in 14a interacts with the target
enzyme, while except for 3’-OH little interactions are noted for the ribose ring of Ser-AMP. The
positions for the sugar carbon atoms deviate substantially for both compounds, with a slight
displacement of the xylofuranose ring of 14a towards the serine site compared to the ribose ring
in the Ser-AMP structure. The C1’-position is already shifted about 1 Å and the further larger
thiofuranosyl ring results in C4’ of 14a to be positioned almost on the 5’-carbon of Ser-AMP.
The 2’ and 3’ secondary hydroxyl moieties however are largely positioned the same albeit with
slightly different orientation. For the thiolated sugar ring this results in a strong H-bond network
where the 2’-hydroxyl moiety is binding to the Arg431 side chain and Leu385 carbonyl and
where the 3’-OH is H-bonding to the Glu384 side chain. The unsubstituted 5’-OH as seen only in
1
4a is offset by 2.4 Å versus the 5’-oxygen in Ser-AMP, and binds to the same 2 oxygen of
Glu384 as well as to some water molecules.
The strong structuring of the sugar ring in 14a when in complex with serRS is remarkable
and is highlighted as well with the intramolecular H-bonding of 5’-OH with C6-H and C3’-H
(Figure 5A). The thioxylofuranosyl ring in 14a has the 3’-exo conformation (Figure 5B) with a
pseudorotation phase angle (P-value) of 186.79, while the seryl-adenylate has the expected 3’-
endo conformation of ribonucleosides (Figure 5C) with a P-value of 20.81. The thiolated ring is
larger resulting in a C4’-C1’ distance of 2.7 Å, versus 2.3 Å for the ribose ring.
1
4

Fig. 4. Interactions of Tbg-serRS with ser-AMP and 14a. The enzymes are shown as cartoon
representations and colored orange and green, respectively. Residues within 5Å of the compound
are shown as stick model. The serine congener 14a is colored magenta and ser-AMP is colored
yellow. Hydrogen bonds formed between serRS and 14a and ser-AMP are colored magenta and
yellow, respectively.
Fig. 5. A. Constrained conformation of 14a via intramolecular H-bonding of the 5’-hydroxyl
moiety to C6-H and C3’-H; B. Puckering of seryl-adenylate when in complex with serRS; C.
Puckering of the thio-xylofuranosyl ring as seen for 14a (both generated in pymol).
4
.4. Evaluation of the linker parts
The linker in the albomycin core part connecting the sugar ring to the amino acid carbonyl
group is one atom shorter (C5’-C6’-N) compared to the natural intermediate (C5’-O-P-O). The
former however carries the 5’-OH and a 6’-carboxylate moiety as additional substituents that are
anchoring its position. The hydroxyl moiety is bound to Glu384 as mentioned above, while the
6
’-carboxylate contributes for recognition via an intricate H-bonding network with water
molecules directed towards the Arg295 sidechain as well as electrostatic interaction. In analogy,
the acyl-phosphate linkage of the natural enzyme intermediate binds even stronger via an intricate
network of multiple H-bonds to Arg295 and Ser387 as well as water mediated interactions with
Glu384. In addition, Arg295 likewise is recognizing the serine carbonyl moiety for both ligands.
It underlines the importance of Arg295, recognizing both, the phosphate and serine carbonyl in
1
5

the active intermediate Ser-AMP, and binding to the reciprocal 6’-carboxylate and serine
carbonyl moiety for the albomycin core structure. As seen in the ligand overlay (SI-2B) and the
snapshot for the ligand protein interactions (SI-2C, Figure 4A), the 6’-carboxylate penetrates
somewhat deeper into the cavity resulting in a 2.82 Å H-bond length to Arg295, versus 3.26 Å
for the reciprocal phosphate, the latter however binding to both terminal nitrogen atoms of
Arg295. The linker fragment for 14a has the (5’S, 6’R)-5’-hydroxy-6’-carboxyl stereochemistry
(nucleoside numbering as seen in Figure 1, while its diastereomer 14b equipped with a (5’R,6’S)-
stereochemistry is unable to correctly position the 6’-carboxylate and 5’-OH (based on soaking
not providing a clear electron density map). Overall, the shorter linker of the albomycin core
structure is compensated by the small shift of the base and sugar ring towards the serine-
recognizing pocket. The firm anchoring of the thiosugar and the linker part allow for perfect
superposition of the serine part of both molecules (see below).
4
.5. Recognition of the serine part
In both structures the terminal serine moiety is well recognized, with the serine -
hydroxyl group being anchored with the Glu318 sidechain and Thr425 hydroxyl group, while the
-amine is positioned well via hydrogen bonding with the Thr264 and Glu318 sidechains.

Recognition of the carbonyl moiety was already disclosed above. Overall, the position of the
serine residues in 14a and in Ser-AMP corresponds entirely upon superposition of both complex
structures (Figure 4A). In a supplementary file (SI-Movie) the superimposed ligands 14a and
seryl-adenylate are rotated over 360° allowing inspection and comparison of the subtle changes
in orientation of the substituents, and highlighting the deeper positioning towards Arg295 of the
6
’-carboxylate versus the phosphate moiety in the natural intermediate.
4
.6. Analogy with the aspartyl congener 16a
Coupling of aspartate to the albomycin core structure provided the new congener 16a (and
its inactive diastereoisomer 16b), that strongly inhibited AspRS. A complex structure was
obtained upon soaking into Tt-AspRS crystals. Belonging to the same class II family of aaRSs,
the interactions of the base part are very similar with those for 14a, with a water-mediated bridge
from the C2-oxygen to both the backbone NH of Arg531 and sidechain and backbone NH of
Asp530 substituting for the threonine interactions as seen for the above serine congener (Figure
1
6

6
A and 7C). In this structure, the base is sandwiched via π-π and cation-π interactions mediated
by Arg231 and Phe235. Like with 14a, the uracil base is located at the position of the adenine
3
imidazole ring, with the N -methyl moiety coinciding with the C2 position of the adenine ring.
The sugar ring for this congener is anchored via Arg531 hydrogen bonding to the 2’-hydroxyl
and Glu476 sidechain (1) stabilization of the 3’- and 5’-hydroxyl group (Figure 7B). The 6’-
carboxylate likewise is stabilized as in the serine congener via Arg223 bonding and with weak
His442 interaction via its N3-H tautomer.
Fig. 6. A: 2D interaction of Tt-aspRS with BG760A. Hydrogen bonds are shown as purple lines.
Cation-π and π-π interactions are shown as red and green lines, respectively. Figure generated
using Maestro Schrodinger. B: Superposition of Tt-aspRS bound to 16a and asp-AMP. The
enzymes are shown as cartoon representations and colored orange and green, respectively.
Residues within 5Å of the compound are shown as stick model. Congener 16a is colored magenta
and ser-AMP is colored yellow. Data for Asp-AMP are taken from Poterszman et al. [33].
1
7

Fig. 7. Interactions of Tt-aspRS with asp-AMP and 16a. The enzymes are shown as cartoon
representations and colored orange and green, respectively. Residues within 5Å of the compound
are shown as stick model. Compound 16a is colored magenta and asp-AMP is colored yellow.
Hydrogen bonds are colored formed between aspRS and BG760A and asp-AMP are colored
magenta and yellow, respectively.
In analogy however with the Asp-AMP structure [33], the terminal aspartate of 16a is strongly
structured and held in place via its carbonyl backbone bonding with Arg223 and Gln237
sidechain amide, while the alpha-amine binds with Gln237 and Gln201 (Figure 7A). The terminal
carboxylate is strongly positioned via electrostatic interaction with Lys204 and twice with
Arg483. In comparison with the 14a complex structure, the sugar and linker parts of the
albomycin core for 16a are likewise recognized but a bit less bound apparently. Especially the
aspartate recognition boosts the affinity and results in overall slightly higher affinities for 16a
over 14a for their respective aaRS target.
5
. Conclusion
Total synthesis of the core scaffold for the natural product albomycin was accomplished
providing a mixture of two diastereomers, and the scaffold was functionalized with the amino
acids serine, aspartic acid and isoleucine, followed by isomer separation. The isoleucine
containing compound aimed at targeting a class I aaRS, but lacked inhibitory activity, while the
serine and aspartic acid congeners having the required 5’S,6’R-stereochemistry, in contrast
displayed strong inhibitory activity. Soaking 14a and 16a in crystals of the corresponding aaRS
provided 3D structural information highlighting the tight interaction of the scaffold with class II
aaRS enzymes. Overall, we conclude the albomycine core structure is well suited for targeting
class II aaRSs and could be used for development of a new family of antibiotics.
1
8

1
9

6
. Experimental section
6
.1. General materials and methods
Reagents and solvents were purchased from commercial suppliers (Acros, Sigma-Aldrich,
Bachem, Novabiochem) and used as provided, unless indicated otherwise. DMF and THF were
of analytical grade and were stored over 4Å molecular sieves. For reactions involving Fmoc-
protected amino acids and peptides, DMF for peptide synthesis (low amine content) was used.
All other solvents used for reactions were analytical grade and used as provided. Reactions were
carried out in oven-dried glassware under a nitrogen atmosphere with stirring at room
1
13
temperature, unless indicated otherwise. H and C NMR spectra of the compounds dissolved in
CDCl , CD OD, DMSO-d or D O were recorded on a Bruker UltraShield Avance 300 MHz or
3
3
6
2
5
00 MHz spectrometer. The chemical shifts are expressed as δ values in parts per million (ppm),
1
13
1
using the residual solvent peaks (DMSO: H, 2.50 ppm; C, 39.52 ppm; HOD: H, 4.79 ppm) as
a reference. Coupling constants are given in Hertz (Hz). The peak patterns are indicated by the
following abbreviations: bs = broad singlet, d = doublet, m = multiplet, q = quadruplet, s = singlet
and t = triplet. High resolution mass spectra were recorded on a quadrupole time-of-flight mass
spectrometer (Q-Tof-2, Micromass, Manchester, UK) equipped with a standard ESI interface;
-1
samples were infused in 2-propanol/H O (1:1) at 3 µL.min . For TLC, pre-coated aluminium
2
sheets were used (Merck, Silica gel 60 F ). The spots were visualized by UV light at 254 nm.
2
54
Column chromatography was performed on ICN silica gel 60A 60–200. Preparative HPLC of
final compound was done using an Adamas C18-X-bond 5 µm prep. (250 × 10 mm) column
connected to a Shimadzu LC-20AT pump using a Shimadzu SPD-20A UV-detector. Recordings
were performed at 254 nm and 214 nm. IUPAC nomenclature is used for the title compounds, but
NMR assignments are based on elongated nucleoside structures for convenience and consistency
with the basic starting blocks.
6
.2. Methyl-β-L-arabinopyranoside (3) [19, 20]
o
Acetyl chloride (42.5 mL, 588 mmol) was added dropwise to methanol (450 mL) at 0 C.
Next, L-arabinose (100 g, 666 mmol) was added in three portions. The suspension was
o
o
maintained at 0 C for 1 h and then heated to 55 C for 4 h. After cooling it to room temperature,
the reaction mixture was poured into a mixture of hexane: Et O (1:1 v/v, 1L). The precipitate was
2
2
0

filtered off and washed with a mixture of hexane: Et O (1:1 v/v). The precipitate was collected,
2
washed with Et O (3 times) to remove traces of HCl, and dried under vacuum to yield 81 g (493
2
1
mmol, 74%) of methyl arabinopyranoside 3 as a white solid. H NMR (300 MHz, CD OD) δ 3.38
3
(
s, OCH ), 3.55 (dd, 1H, J = 12.3 and 2.3 Hz), 3.73-3.81 (m, 3H), 3.82-3.86 (m, 1H), 4.66 (d, 1H,
3
1
3
J = 2.9 Hz, H-1); C NMR (75 MHz, CD OD) δ 55.8 (OCH ), 63.9 (C-5), 70.3 (C-3), 70.7 (C-2),
3
3
+
7
0.8 (C-4), 102.0 (C-1); HRMS for C H O Na [M+Na] calcd.: 187.0577; found: 187.0586
6 12 5
6
.3. Tosylation of L-β-arabinopyranoside (4 and 5) [21, 22]
A mixture of 3 (1 g, 6.09 mmol) and dibutyltin oxide (1.52 g, 6.09 mmol, 1.0 eq.) in dry
methanol (25 mL) was heated to reflux for overnight. Next, the solvent was evaporated to dryness
and the residue was co-evaporated with dioxane (3 times) to remove traces of methanol. The oily
residue was suspended in dry dioxane (25 mL) along with DMAP (149 mg, 1.22 mmol) and
o
cooled to 0 C, after which a solution of tosyl chloride (1.23 g, 6.7 mmol, 1.1 equivalent) in dry
dioxane (5 mL) was added dropwise. Next, the ice bath was removed and the reaction mixture
was stirred at room temperature for 30 h. The solvent was evaporated to dryness and the products
were purified by column chromatography to yield 300 mg (0.94 mmol, 16%) 3-O-tosyl (4) and
8
89 mg (2.79 mmol, 46%) 4-O-tosyl product (5). Both products were obtained as white solids.
1
Methyl-3-O-tosyl-β-L-arabinopyranoside (4): H NMR (600 MHz, CDCl ) δ 1.93 (d, 1H, J =
3
1
0.0 Hz, 2-OH), 2.45 (s, 3H, Tosyl-CH ), 2.51 (bs, 1H, 4-OH), 3.41 (s, 3H, OCH ), 3.72 (dd, 1H,
3 3
J = 2.3 Hz and 12.6 Hz, H-5a), 3.78 (d, 1H, J = 12.6 Hz, H-5b), 4.00 (td, 1H, J = 3.8 Hz and 9.9
Hz, H-2), 4.15-4.18 (m, 1H, H-4), 4.67 (dd, 1H, J = 3.2 Hz and 9.8 Hz, H-3), 4.79 (d, 1H, J = 3.8
1
3
Hz, H-1), 7.34 (m, 2H, meta-H, Ts), 7.84-7.87 (m, 2H, ortho-H, Ts); C NMR (150 MHz,
CDCl ) δ 21.7 (p-Ts-CH ), 55.6 (OCH ), 61.9 (C-5), 66.8 (C-2), 68.6 (C-4), 77.2 (C-3) 100.0 (C-
3
3
3
1
), 127.9 (meta-C, Ts), 129.9 (ortho-C, Ts), 133,4 (para-C, Ts)145.2 (ipso-C, Ts); HRMS for
+
C H O SN [M+NH ] calcd.: 336.1116; found: 336.1109.
13
22
7
4
1
Methyl-4-O-tosyl-β-L-arabinopyranoside (5): H NMR (600 MHz, CDCl ) δ 2.28-2.33 (m, 1H,
3
2
-OH), 2.45 (s, 3H, Tosyl-CH ), 2.54-2.58 (m, 1H, 3-OH), 3.41 (s, 3H, OCH ), 3.74-3.79 (m, 3H,
3 3
H-3 and H-5), 3.85-3.89 (m, 1H, H-2), 4.78 (d, 1H, J =3.6 Hz, H-1), 4.82-4.84 (m, 1H, H-4), 7.35
1
3
(
d, 2H, J = 7.8 Hz, meta-H, Ts), 7.84 (d, 2H, J = 8.4 Hz, ortho-H, Ts); C NMR (150 MHz,
CDCl ) δ 21.7 (Ts-CH ), 55.8 (C-5), 60.7 (OCH ), 68.6 (C-3), 69.5 (C-2), 78.8 (C-4), 78.8 (C-4),
3
3
3
2
1

9
9.5 (C-1), 127.9 (meta-C, Ts), 129.9 (ortho-C, Ts), 133.5 (para-C, Ts), 145.1 (ipso-C, Ts);
+
HRMS for C H O SN [M+NH ] calcd.: 336.1116; found: 336.1101.
13
22
7
4
6
.4. Methyl-2,3-di-O-benzoyl-4-O-tosyl-β-L-arabinopyranoside (6) [24]
To an ice cooled solution of 5 (37.4 g, 117.48 mmol) and DMAP (2.87 g, 23.5 mmol) in
dry pyridine (600 mL) was added dropwise a mixture of benzoyl chloride (54.6 mL, 469.9 mmol)
in dry pyridine (100 mL) over a period of 1.5 h. After 30 min, the ice bath was removed and the
reaction mixture was stirred at room temperature for 24 h. Next, the excess of benzoyl chloride
was quenched by adding saturated NaHCO (1 L) dropwise. The solvent was evaporated to
3
dryness and the residue was partitioned between saturated NaHCO (1 L) and ethyl acetate (1 L).
3
The layers were separated and the aqueous layer was further washed with ethyl acetate (2 x 500
mL). The organic layer was collected, dried over Na SO , filtered and evaporated to dryness. The
2
4
crude yellow residue was subjected to column chromatography to yield 56.8 g (107.87 mmol,
1
9
2%) of di-O-benzoylated product 6 as a white solid. H NMR (600 MHz, CDCl ) δ 2.17 (s, 3H,
3
Ts-CH ), 3.42 (s, 3H, OCH ), 3.96-4.05 (m, 2H, H-5), 5.08 (bs, 1H, H-4), 5.14 (d, 1H, J = 3.0 Hz,
3
3
H-1), 5.59 (t, 2H, J = 2.4 Hz, H-2 and H-3), 6.94-6.97 (m, 2H, ortho-H, Ts), 7.33-7.36 (m, 4H,
Bz), 7.48-7.53 (m, 2H, Bz), 7.65-7.68 (m, 2H, meta-H, Ts), 7.75-7.78 (m, 2H, Bz), 7.93-7.95 (m,
1
3
2
H, Bz); C NMR (150 MHz, CDCl ) δ 55.8 (OCH ), 60.6 (C-5), 67.8 (C-3), 68.4 (C-2), 76.8
3 3
(C-4), 97.7 (C-1), 127.7, 128.2, 128.4, 128.8, 129.2, 129.7, 129.78, 129.80, 133.1, 133.28,
+
1
5
33.29, 144.7 (aromatic), 165.5, 165.7 (C=O, Bz); HRMS for C H O SN [M+NH ] calcd.:
2
7
30
9
4
44.1641; found: 544.1631.
.5. Methyl-2,3-di-O-benzoyl-4-deoxy-4-thioacetyl-α-D-xylopyranoside (7) [25, 26]
6
A solution of 6 (55.5 g, 105 mmol) and potassium thioacetate (30.15 g, 264 mmol) in dry
o
DMF (600 mL) was stirred at 80 C for 6 h. Although, TLC indicated incomplete reaction, the
reaction was stopped due to formation of undesired products. Next, the solvent was evaporated to
dryness. The dark brown residue was partitioned between water (0.5 L) and DCM (0.5 L). The
aqueous layer was further washed with DCM (3 x 300 mL). The combined DCM layer was
washed with saturated NaCl solution. The organic layer was collected, dried over anhydrous
Na SO , filtered and evaporated to dryness. The brown residue was subjected to silica gel
2
4
chromatography to yield 23.1 g (53.66 mmol, 51%) of 4-thio-xylopyranoside 7 as white solid. In
2
2

addition, 26.2 g (49.76 mmol, 47%) of starting 4-O-tosyl derivative 6 was recovered as pale
1
yellow solid. H NMR (600 MHz, CDCl ) δ 2.26 (s, 3H, SCOCH ), 3.41 (s, 3H, OCH ), 3.76-
3
3
3
3
.82 (m, 2H, H-5), 4.03-4.09 (m, 1H, H-4), 5.15 (d, 1H, J = 3.5 Hz, H-1), 5.23 (dd, 1H, J = 3.5
Hz and 9.8 Hz, H-2), 5.82 (dd, 1H, J = 9.8 Hz and 11.3 Hz, H-3), 7.34-7.39 (m, 4H, Bz), 7.47-
1
3
7
.51 (m, 2H, Bz), 7.94-7.97 (m, 4H, Bz); C NMR (150 MHz, CDCl ) δ 30.7 (SCOCH ), 43.2
3 3
(
C-4), 55.5 (OCH ), 60.5 (C-5), 68.4 (C-3), 73.3 (C-2), 97.6 (C-1), 128.3, 128.4, 129.1, 129.2,
3
1
29.8, 133.2, 133.3 (aromatic), 165.75, 165,80 (C=O, Bz), 193.6 (SCOCH ); HRMS for
3
+
C H O SNa [M+Na] calcd.: 453.0979; found: 453.0975.
22
22
7
6
.6. 1,5-di-O-acetyl-2,3-di-O-benzoyl-4-deoxy-4-thio-α/β-D-xylofuranose (8) [25, 27]
A mixture of AcOH:Ac O:conc. H SO (10:10:1 v/v/v, 462 mL) was added to thioacetate
2
2
4
o
derivative 7 (35.1 g, 81.54 mmol) while maintaining below -10 C. The reaction mixture was
o
o
stirred at -10 C for 10 min. and was allowed to stand at +4 C for overnight. Next, sodium
o
acetate (30 g) was added at 0 C. The reaction mixture was then neutralized very slowly by
addition of saturated NaHCO to reach pH 8 (over 3-4 h). It was extracted with ethyl acetate (3 x
3
1
L). The combined organic layer was further washed with saturated NaCl. The organic layer was
collected, dried over Na SO , filtered and evaporated to dryness. The dark yellow viscous oil was
2
4
purified by silica gel chromatography to obtain 28.1 g (61.35 mmol, 75%) of xylofuranose
+
derivative as a mixture of α/β anomers of 8. HRMS for C H O SNa [M+Na] calcd.: 481.0928;
2
3
22
8
found: 481.0927.
1
6
.7. N -(5-O-acetyl-2,3-di-O-benzoyl-4-deoxy-4-thio-β-D-xylofuranos-1-yl)-uracil (9)
To a solution of uracil (8.13 g, 72.5 mmol, 1.2 eq.) in dry ACN (100 mL) was dropwise
added BSA (59.1 mL, 241.7 mmol, 4 eq.). The reaction mixture being a suspension was heated to
o
8
0 C for 45 min. during which it became clear. It was then cooled to room temperature and a
solution of thiofuranose 8 in ACN (27.7 g, 60.42 mmol) was added dropwise. Next a solution of
SnCl (1 M, 181.1 mL, 181.1 mmol, 3 eq.) was added dropwise. The reaction mixture was stirred
4
o
at 40 C for overnight after which it was cooled to room temperature and quenched by addition of
saturated NaHCO (pH~8). The gelatinous precipitate was filtered off through Celite® 545. The
3
filtrate was transferred to a separating funnel and washed with DCM (3 x 0.5 L). The organic
layer was further washed with brine. The DCM layer was collected, dried over anhydrous
Na SO , filtered and evaporated to dryness. The product was purified by column chromatography
2
4
2
3

1
to afford 22.07 g (43.23 mmol, 72%) of thionucleoside 9 as pale yellow solid. H NMR (500
MHz, CDCl ) δ 2.01 (s, 3H, COCH ), 4.09-4.17 (m, 1H, H-4’), 4.36 (dd, 1H, J = 5.8 Hz and
3
3
1
3
7
1.91 Hz, H-5’a), 4.56 (dd, 1H, J = 5.1 Hz and 11.88 Hz, H-5’b), 5.83-5.90 (m, 2H, H-5 and H-
’), 6.04 (t, 1H, J = 6.7 Hz, H-4’), 6.45 (d, 1H, J = 6.6 Hz, H-1’), 7.38-7.45 (m, 4H, Bz, 7.52-
1
3
.60 (m, 2H, Bz), 7.91-8.00 (m, 4H, Bz), 8.04 (d, 1H, J = 8.2 Hz, H-6), 9.37 (s, 1H, NH); C
NMR (125 MHz, CDCl ) δ 20.7 (COCH ), 43.3 (C-4’), 59.4 (C-1’), 63.4 (C-5’), 74.4 (C-3’), 77.4
3
3
(C-2’), 103.7 (C-5), 128.3, 128.4, 128.5, 128.6, 128.7, 129.7, 129.8, 129.86, 129.89, 130.0, 133.8,
1
1
34.0 (aromatic), 140.1 (C-6), 150.6 (C-2), 162.5 (C-4), 165.2 (C=O, Bz), 165.4 (C=O, Bz),
+
70.0 (C=O, Ac); HRMS for C H N O S [M+H] calcd.: 511.1169;. found: 511.1169.
2
5
23
2
8
3
1
6
.8. N -methyl-[N -(5-O-acetyl-2,3-di-O-benzoyl-4-deoxy-4-thio-β-D-xylofuranos-1-yl)]-uracil
(10)
Compound 9 (21.9 g, 42.90 mmol) and K CO (17.79 g, 128.7 mmol, 3 eq.) were
2
3
suspended in a mixture of dry DMF: acetone (1:1 v/v, 400 mL) to which methyl iodide (2.67 mL,
o
4
2.90 mmol, 1 eq.) was added dropwise. The suspension was heated to reflux (70 C) for
overnight. Next, the solvent was evaporated under reduced pressure. The residue was partitioned
between water (500 mL) and EtOAc (500 mL). The layers were separated and the aqueous layer
was further washed with ethyl acetate (3x 500 mL). The combined organic layer was further
washed with brine. The organic layer was collected, dried over Na SO , filtered and evaporated
2
4
to dryness. The product was purified by column chromatography to afford 15.96 g (30.42 mmol,
1
7
1%) of methylated derivative 10 as pale yellow semisolid. Protocol adapted from ref [28]. H
NMR (500 MHz, CDCl ) δ 2.02 (s, 3H, COCH ), 3.27 (s, 3H, NCH ), 4.16 (q, 1H, J = 6.0 Hz, H-
3
3
3
4
5
7
’), 4.38 (dd, 1H, J = 5.9 Hz and 11.8 Hz, H-5’a), 4.55 (dd, 1H, J = 5.2 Hz and 11.8Hz, H-5’b),
.83-5.92 (m, 2H, H-5 and H-3’), 6.01-6.07 (m, 1H, H-2’), 6.50 (d, 1H, J = 6.2 Hz, H-1’), 7.40-
13
.46 (m, 4H, Bz), 7.55-7.61 (m, 2H, Bz), 7.92-8.04 (m, 5H, Bz and H-6); C NMR (125 MHz,
CDCl ) δ 20.7 (COCH ), 28.1 (NCH ), 43.8 (C-4’), 60.7 (C-1’), 63.4 (C-5’), 74.6 (C-3’), 78.2
3
3
3
(C-2’), 103.0 (C-5), 128.3, 128.4, 128.6, 128.7, 129.8, 130.0, 133.9, 134.0 (aromatic), 137.9 (C-
6
), 151.6 (C-2), 162.2 (C-4), 165.2, 165.3 (C=O, Bz), 170.0 (C=O, Ac); HRMS for C H N O S
2
6
25
2
8
+
[
M+H] calcd.: 525.1326;.found: 525.1341.
3
1
6
.9. N -methyl-[N -(2,3-di-O-benzoyl-4-deoxy-4-thio-β-D-xylofuranos-1-yl)]-uracil (11)
2
4

The methylated nucleoside derivative 10 (15.86 g, 30.24 mmol) was dissolved in a
o
mixture of DCM:MeOH (1:1 v/v, 400 mL) and cooled to 0 C. Hereto, acetyl chloride (21.6 mL,
3
02 mmol, 10 eq.) was added dropwise. After 15 min, the ice bath was removed and the reaction
mixture was stirred at room temperature for overnight. The reaction mixture was cooled again to
o
0
C and quenched by addition of saturated NaHCO (pH~8). The volatiles were removed by
3
evaporation and the aqueous layer was extracted with DCM (3 x 500 mL). The combined organic
layer was washed with saturated NaCl. The organic layer was collected, dried over anhydrous
Na SO , filtered and evaporated to dryness. The product was isolated by column chromatography
2
4
1
to yield 13.0 g (26.94 mmol, 89%) of deacetylated product 11 as a white solid. H NMR (300
MHz, CDCl ) δ 2.83 (s, 1H, 5’-OH), 3.27 (s, 3H, NCH ), 3.81-3.89 (m, 1H, H-5’a), 3.90-3.97 (m,
3
3
1
8
H, J = 6.6 Hz, H-4’), 4.02-4.17 (m, 1H, H-5’b), 5.85 (t, 1H, J = 6.9 Hz, H-3’), 5.93 (d, 1H, J =
.1 Hz, H-5), 6.27 (t, 1H, J = 6.6 Hz, H-2’), 6.49 (d, 1H, J = 6.3 Hz, H-1’), 7.30-7.47 (m, 4H,
Bz), 7.48-7.61 (m, 2H, Bz), 7.91 (d, 2H, J = 7.5 Hz, Bz), 8.02 (d, 2H, J = 7.2 Hz, Bz), 8.47 (d,
1
3
1
H, J = 8.1 Hz, H-6); C NMR (75 MHz, CDCl ) δ 28.0 (NCH ), 46.4 (C-4’), 59.9 (C-4’), 59.9
3 3
(C-1’), 61.4 (C-5’), 75.4 (C-3’), 78.4 (C-2’), 102.8 (C-5), 128.45, 128.53, 128.6, 129.9, 133.6,
1
33.9 (aromatic), 139.3 (C-6), 151.8 (C-2), 163.0 (C-4), 165.4, 165.7 (C=O, Bz); HRMS for
+
C H N O S [M+H] calcd: 483.1220 found 483.1223.
24
23
2
7
6
.10. (2R,3S)-2-bis-(tert-butyloxycarbonyl)-amino-3-((2R,3R,4R,5R)-3,4-di-benzoyloxy-5-(3-
methyl-2,4-dioxo-3,4-dihydropyrimidin-1(2H)-yl)tetrahydrothiophen-2-yl)-3-hydroxypropanoic
acid methyl ester and its (2S,3R) diastereoisomer (12)
To an ice cooled solution of alcohol 11 (2 g, 4.14 mmol, 1.0 eq.) in dry DCM (50 mL)
was added dropwise a solution of Dess-Martin reagent (15% w/v, 17.6 mL, 6.21 mmol, 1.5
o
equiv.) and the reaction mixture was stirred at 0 C for 3 h. The reaction was quenched by
simultaneous addition of saturated NaHCO and 10% sodium thiosulphate. The biphasic mixture
3
was vigorously stirred for additional 7 min. The layers were separated and the aqueous layer was
extracted with DCM (2 x 300 mL). The combined organic layer was washed with saturated NaCl
and collected, dried over Na SO , filtered and evaporated to dryness. The crude aldehyde was
2
4
used without further purification.
3
1
1
N -methyl-[N -(2,3-di-O-benzoyl-4-deoxy-4-thio-5-aldehyde-β-D-xylofuranos-1-yl)]-uracil:
H
NMR (300 MHz, CDCl ) δ3.28 (s, 3H, NCH ), 4.61 (dd, 1H, J = 1.9 Hz and 6.8 Hz, H-4’), 5.95-
3
3
2
5

6
4
.07 (m, 2H, H-5, H-3’), 6.15-6.21 (m, 1H, H-2’), 6.56 (d, 1H, J = 6.4 Hz, H-1’), 7.38-7.47 (m,
1
3
H, Bz), 7.54-7.62 (m, 2H, Bz), 7.91-8.01 (m, 5H, Bz and H-6), 9.92 (s, 1H, CHO); C NMR
(
75 MHz, CDCl ) δ 28.0 (NCH ), 51.1 (C-4’), 60.6 (C-1’), 74.6 (C-3’), 77.3 (C-2’ merged in
3 3
solvent peak), 103.6 (C-5), 128.6, 128.7, 129.8, 129.9, 134.1 (Bz), 138.1 (C-6), 151.6 (C-2),
1
62.2 (C-4), 165.22, 165.25 (C=O, Bz), 194.6 (CHO).
In a separate vessel, nBuLi (2.5 M, 3.4 mL, 8.28 mmol) in dry THF (50 mL) was cooled
o
to 0 C (ice-bath). To this solution, DIPEA (1.2 mL, 8.28 mmol) was added slowly. The reaction
o
mixture was stirred at 0 C for 30 min and at room temperature for 15 min., and cooled again to -
o
7
8 C. Then a solution of Boc GlyOMe (2.4 g, 8.28 mmol) in dry THF (20 mL) was added
2
o
dropwise over a period of 10 min. The reaction mixture was stirred at -78 C for 2 h. Next, the
solution of crude aldehyde (4.14 mmol) in dry THF (50 mL) was added dropwise over a period of
o
1
5 min. and the reaction mixture was stirred at -78 C for an additional 1 h. After completion, it
was quenched by addition of glacial AcOH (100 mL) and was allowed to warm to room
temperature while stirring for overnight. The reaction mixture was partitioned between EtOAc
and water. The layers were separated and the aqueous layer was further extracted with EtOAc (2
x 100 mL). The combined EtOAc layer was washed with saturated NaCl, collected, dried over
Na SO , filtered and evaporated to dryness. The yellow oil was subjected to silica gel
2
4
chromatography using EtOAc: hexane with 0.2% TEA to remove excess Boc GlyOMe and other
2
impurities. The product was further purified by 16% EtOAc:0.2% TEA in toluene to yield 1.371
1
g (1.78 mmol, 43%) of the glycine adduct 12 as a mixture of two diastereomers. H NMR (300
MHz, CDCl ) (as a mixture of 2 diastereomers) δ 1.40-1.61 (m, 18H, Boc-CH ), 3.24-3.27 (m,
3
3
3
5
6
H, NCH ), 3.65-3.77 (m, 3H, COOCH ), 4.00-4.20 (m, 1H, H-6’), 4.65-4.75 (m, H-5’), 5.05-
3 3
.50 (m, 2H, H-4’ and 5’-OH), 5.87 (d, J = 8.2 Hz) and 5.94 (d, 1H, J = 8.2 Hz), (1H, H-5), 5.97-
.20 (m, 2H, H-2’and H-3’), 6.50 (d, J = 5.4 Hz) and 6.57 (d, J = 7.6 Hz) (1H, H-1’), 7.34-7.46
(m, 4H, Bz), 7.49-7.61 (m, 2H, Bz), 7.89-8.03 (m, 4H, Bz), 8.40 (d, J = 8.2 Hz) and 8.46 (d, J =
1
3
8
6
1
1
7
.2 Hz) (1H, H-6); C NMR (75 MHz, CDCl ) δ 27.5, 28.0, 44.7, 47.3, 52.9, 53.2, 55.5, 57.4,
3
0.0, 72.4, 72.8, 73.7, 74.3, 78.5, 80.7, 84.1, 84.2, 102.7, 103.1, 128.1, 128.2, 128.4, 128.46,
28.49, 128,57, 128.62, 129.88, 129.9, 130.0, 133.5, 133.6, 133.9, 151.6, 151.8, 151.9, 155.1,
+
62.5, 162.6, 164.9, 165.2, 165.3, 169.6, 170.1; HRMS for C H N O SNa[M+Na] calcd:
3
7
43
3
13
92.2409 found 792.2430.
2
6

6
.11. (2R,3S)-2-amino-3-((2R,3R,4R,5R)-3,4-dihydroxy-5-(3-methyl-2,4-dioxo-3,4-
dihydropyrimidin-1(2H)-yl)tetrahydrothiophen-2-yl)-3-hydroxypropanoic acid and its (2S,3R)
diastereoisomer (13)
To an ice-cooled solution of the elongated product 12 (1.22 g, 1.6 mmol) in 1,4-dioxane
(2o mL) was slowly added 1M NaOH (9.6 mL, 9.6 mmol, 6 eq.). After 10 min, ice bath was
removed and the reaction mixture was stirred at room temperature. After completion of the
o
reaction, it was quenched at 0 C by addition of TFA (10 mL, pH = 2). Next, the reaction mixture
was stirred at room temperature for 1h. After completion of the reaction, the solvent was
evaporated to dryness and the crude product was purified by column chromatography using
1
methanol: DCM with 1% acetic acid to yield 436 mg (79%) of the deprotected product 13. H
NMR (600 MHz, D O) (as a mixture of 2 diastereomers) δ 3.28, 3.34 (2s, N-CH ), 3.80-4.10 (m,
2
3
2
8
5
H, H-4’ and H-6’), 4.37-4.53 (m, 3H, H-2’, H-3’ and H-5’), 5.91-5.99 (m, 2H, H-1’ and H-5),
1
3
.43 (2d, 1H, J = 6.6 Hz and 6.0 Hz, H-6); C NMR (125 MHz, D O) δ 29.3, 29.4 (NCH ), 52.6,
2
3
3.7 (C-4’), 59.1, 59.5 (C-6’), 65.4, 67.7 (C-1’), 69.0, 69.4 (C-5’), 76.0, 76,5 (C-3’), 81.1, 82.2
(C-2’), 102.0, 102, 8 (C-5), 143.1, 143.5 (C-6), 154.2, 154.3 (C-4), 166.8, 166.9 (C-2), 172.5,
-
1
73.5 (COOH); HRMS for C H N O S [M-H] calcd: 346.0714 found 346.0696.
1
2
16
3
7
6
.12. Synthesis of the seryl-analogue 14: (2R,3S)-2-((S)-2-amino-3-hydroxypropanamido)-3-
((2R,3R,4R,5R)-3,4-dihydroxy-5-(3-methyl-2,4-dioxo-3,4-dihydropyrimidin-1(2H)-
yl)tetrahydrothiophen-2-yl)-3-hydroxypropanoic acid (14a)
HBTU (246 mg, 0.65 mmol) was added to a solution of Boc-Ser(OTBDMS)-OH (208
mg, 0.65 mmol) in THF (2 mL). The reaction mixture was stirred at room temperature for 1 h.
The HOBt activated ester was used as such in the next step. In a separate vessel, unprotected
nucleoside 13 (92 mg, 0.26 mmol) and NaHCO were dissolved in a mixture of THF and water
3
(1:1v/v, 2 mL) at room temperature. The active ester from above reaction was added slowly to
the reaction mixture and stirring was continued for overnight. Next, the solvent was evaporated to
dryness and the crude product was purified by silica gel column chromatography to yield 200 mg
+
semi-pure product. HRMS for C H N O SNa [M-TBDMS+Na] calcd: 557.1524 found
2
0
30
4
11
5
57.1520.
To a suspension of semi-pure coupled product (200 mg) in water (1 mL) was added TFA
1 mL). The reaction mixture was stirred at room temperature for 1 h. After completion of
(
2
7

reaction, the solvent was evaporated to dryness. Traces of TFA were removed by co-evaporation
with toluene. The product was further purified by RP-HPLC using C-18 column to yield 20.7 mg
-
(
18% over three steps, 0.05 mmol) of product. HRMS for C H N O S [M-H] calcd: 433.1034
15
21
4
9
found 433.1032. The obtained diastereomers were separated by using Adamas C18-X-bond 5 µm
250 × 10 mm) column using 50 mM TBAF in acetonitrile and water to yield 1.8 mg of the
(
desired isomer (14a) and 5.5 mg of undesired isomer (14b) (with opposite configuration at C-5’
and C-6’).
1
Analysis of 14a (the more polar diastereomer): H NMR (600 MHz, D O) δ 1.27 (t, 4H, J = 7.2
2
Hz, TEA-CH ), 3.19 (q, 2.5H, J = 7.2 Hz, TEA-CH ), 3.28 (s, 3H, N-CH ), 3.74-3.76 (m, 1H, H-
3
2
3
4
2
5
4
’), 3.85-3.90 (m, 2H, Ser-β-CH ), 3.92-3.94 (m, 1H, Ser-α-CH), 4.35-4.41 (m, 2H, H-3’ and H-
2
’), 4.44-4.50 (m, 2H, H-5’ and H-6’), 5.92 (d, 1H, J = 5.4 Hz, H-1’), 5.96 (d, 1H, J = 8.4 Hz, H-
1
3
), 8.47 (d, 1H, J = 7.2 Hz, H-6); C NMR (150 MHz, D O) δ 9.6 (TEA-CH ), 29.4 (N-CH ),
2
3
3
8.0 (TEA-CH ), 52.2 (C-4’), 56.5 (Ser-α-CH), 58.8 (C-6’), 60.3 (Ser-β-CH ), 65.4 (C-1’), 70.5
2
2
(C-5’), 76.4 (C-3’), 81.2 (C-2’), 102.7 (C-5), 143.2 (C-6), 154.3 (C-4), 165.6 (C-2), 166.9
-
(
CONH), 176.9 (COOH); HRMS for C H N O S [M-H] calcd: 433.1034 found 433.1027.
15
21
4
9
1
Analysis of 14b (less polar diastereomer): H NMR (600 MHz, D O) δ 1.26 (t, 3H, J = 12.6 Hz,
2
TEA-CH ), 3.20 (q, 1H, J = 12.6 Hz, TEA-CH ), 3.28 (s, 3H, N-CH ), 3.78-3.80 (m, 1H, H-4’),
3
2
3
4
.03 (d, 2H, J = 4.2 Hz, Ser-β-CH ), 4.21 (t, 1H, J = 4.2 Hz, Ser-α-CH), 4.34 (t, 1H, J = 4.8 Hz,
2
H-3’), 4.40 (t, 1H, J = 3.0 Hz, H-2’), 4.45 (d, 1H, J = 3.0 Hz, H-6’), 4.65-4.70 (m, 1H, H-5’),
13
5
.88 (d, 1H, J = 3.0 Hz, H-1’), 5.92 (d, 1H, J = 8.4 Hz, H-5), 8.44 (d, 1H, J = 7.8 Hz, H-6); C
NMR (150 MHz, D O) δ 8.2 (TEA-CH ), 27.9 (N-CH ), 46.6 (TEA-CH ), 54.5 (C-4’), 54.7 (Ser-
2
3
3
2
α-CH), 58.3 (C-6’), 60.2 (Ser-β-CH ), 67.8 (C-1’), 70.5 (C-5’), 75.5 (C-3’), 81.4 (C-2’), 100.2
2
(C-5), 142.4 (C-6), 152.8 (C-4), 165.5 (C-2), 167.9 (CONH), 175.5 (COOH); HRMS for
-
C H N O S [M-H] calcd: 433.1034 found 433.1030.
15
21
4
9
6
.13. Synthesis of the aspartyl-analogue 15: (2R,3S)-2-((S)-2-amino-succinamido)-3-
((2R,3R,4R,5R)-3,4-dihydroxy-5-(3-methyl-2,4-dioxo-3,4-dihydropyrimidin-1(2H)-
yl)tetrahydrothiophen-2-yl)-3-hydroxypropanoic acid (15a)
To a solution of unprotected thionucleoside (13) (90 mg, 0.26 mmol) in a mixture of THF
(
1 mL) and water (1 mL) was added NaHCO (87 mg, 1.04 mmol, 4 eq.). After 5 min, a solution
3
of Boc-Asp(OtBu)-OSu (251 mg, 0.65 mmol, 2.5 eq.) in THF (1 mL) was added. The reaction
2
8

mixture was stirred at room temperature for overnight. After completion of the reaction, the
solvent was evaporated to dryness. The crude product was purified by silica gel column using
methanol and DCM with 1% acetic acid to yield 50 mg (31%, 0.081 mmol) of the product as a
-
mixture of two diastereomers. HRMS for C H N O S [M-H] calcd: 617.2134 found 617.2150.
2
5
37
4
12
To a suspension of semi-pure coupled product (50 mg, 0.081 mmol) in water (1 mL) was
added TFA (1 mL) and the reaction mixture was stirred at room temperature for 1 h. After
completion of reaction, the solvent was evaporated to dryness. Traces of TFA were removed by
co-evaporation with toluene (2 x 2 mL). The product was purified by RP-HPLC using C-18
column to yield 15.7 mg (42%, 0.03 mmol) of the product as a white solid. Two diastereomers
were separated by using Adamas C18-X-bond 5 µm prep. (250 × 10 mm) column using 50 mM
TBAF in acetonitrile and water to yield 1.3 mg of the desired isomer (15a) and 2.3 mg of
undesired isomer (15b) (with opposite configuration at C-5’ and C-6’).
1
Analysis of 15a (more polar diastereomer): H NMR (600 MHz, D O) δ 1.26 (t, 6H, J = 7.2 Hz,
2
TEA-CH ), 2.74 (dd, 1H, J = 7.2 Hz and 17.4 Hz, Asp-β-CH , Ha), 2.81 (dd, 1H, J = 5.8 Hz and
3
2
1
3
2
1
7.4 Hz, Asp-β-CH , Hb), 3.19 (q, 4H, J = 7.2 Hz, TEA-CH ), 3.28 (s, N-CH ), 3.68 (dd, 1H, J =
2 2 3
.6 Hz and 6.6 Hz, H-4’), 4.27 (dd, 1H, J = 5.4 Hz and 7.2 Hz, Asp-α-CH), 4.25-4.39 (m, 2H, H-
’ and H-3’), 4.43 (d, 1H, J = 7.2 Hz, H-6’), 4.46-4.49 (m, 1H, H-5’), 5.94 (d, 1H, J = 6.0 Hz, H-
13
’), 5.97 (d, 1H, J = 8.4 Hz, H-5), 8.48 (d, 1H, J = 7.8 Hz, H-6); C NMR (150 MHz, D O) δ 9.6
2
(
TEA-CH ), 29.3 (NCH ), 38.4 (Asp-β-CH ), 48.0 (TEA-CH ), 51.8 (C-4’), 52.2 (Asp-α-CH),
3 3 2 2
5
8.5 (C-6’), 64.9 (C-1’), 70.3 (C-5’), 76.2 (C-3’), 81.0 (C-2’), 102.8 (C-5), 143.2 (C-6), 154.3 (C-
-
4
), 166.9 (C-2), 170.3 (CONH), 177.0, 177.2 (2 x COOH); HRMS for C H N O S [M-H]
1
6
22
4
10
calcd: 461.0984 found 461.0991.
1
Analysis of 15b (less polar diastereomer): H NMR (600 MHz, D O) δ 1.26 (t, 6H, J = 7.2 Hz,
2
TEA-CH ), 2.68 (dd, 1H, J = 9.6 Hz and 17.4 Hz, Asp-β-CH , Ha), 2.93 (dd, 1H, J = 4.8 Hz and
3
2
1
3
4
7.4 Hz, Asp-β-CH , Hb), 3.19 (q, 4H, J = 7.2 Hz, TEA-CH ), 3.28 (s, N-CH ), 3.78 (dd, 1H, J =
2 2 3
.6 Hz and 4.8 Hz, H-4’), 4.30-4.35 (m, 2H, Asp-α-CH and H-3’), 4.41 (t, 1H, J = 7.2 Hz, H-2’),
.45 (d, 1H, J = 3.6 Hz, H-6’), 4.63 (dd, 1H, J = 3.6 Hz and 3.6 Hz, H-5’), 5.88 (d, 1H, J = 3.6
1
3
Hz, H-1’), 5.93 (d, 1H, J = 7.8 Hz, H-5), 8.44 (d, 1H, J = 7.8 Hz, H-6); C NMR (150 MHz,
D O) δ 9.6 (TEA-CH ), 29.3 (NCH ), 38.2 (Asp-β-CH ), 48.0 (TEA-CH ), 52.34 (C-4’), 55.6
2
3
3
2
2
(Asp-α-CH), 59.8 (C-6’), 68.9 (C-1’), 71.8 (C-5’), 76.8 (C-3’), 82.7 (C-2’), 101.7 (C-5), 143.7
2
9