Acylated sulfonamide adenosines as potent inhibitors of the adenylate-forming enzyme superfamily

Article abstract of DOI:10.1016/j.ejmech.2019.04.045

The superfamily of adenylate-forming enzymes all share a common chemistry. They activate a carboxylate group, on a specific substrate, by catalyzing the formation of a high energy mixed phosphoanhydride-linked nucleoside intermediate. Members of this diverse enzymatic family play key roles in a variety of metabolic pathways and therefore many have been regarded as drug targets. A generic approach to inhibit such enzymes is the use of non-hydrolysable sulfur-based bioisosteres of the adenylate intermediate. Here we compare the activity of compounds containing a sulfamoyl and sulfonamide linker respectively. An improved synthetic strategy was developed to generate inhibitors containing the latter that target isoleucyl- (IleRS)and seryl-tRNA synthetase (SerRS), two structurally distinct representatives of Class I and II aminoacyl-tRNA synthetases (aaRSs). These enzymes attach their respective amino acid to its cognate tRNA and are indispensable for protein translation. Evaluation of the ability of the two similar isosteres to inhibit serRS revealed a remarkable difference, with an almost complete loss of activity for seryl-sulfonamide 15 (SerSoHA)compared to its sulfamoyl analogue (SerSA), while inhibition of IleRS was unaffected. To explain these observations, we have determined a 2.1 ? crystal structure of Klebsiella pneumoniae SerRS in complex with SerSA. Using this structure as a template, modelling of 15 in the active site predicts an unfavourable eclipsed conformation. We extended the same modelling strategy to representative members of the whole adenylate-forming enzyme superfamily, and were able to disclose a new classification system for adenylating enzymes, based on their protein fold. The results suggest that, other than for the structural and functional orthologues of the Class II aaRSs, the O to C substitution within the sulfur-sugar link should generally preserve the inhibitory potency.

Full text of DOI:10.1016/j.ejmech.2019.04.045

Accepted Manuscript
Acylated sulfonamide adenosines as potent inhibitors of the adenylate-forming
enzyme superfamily
Dries De Ruysscher, Luping Pang, Steff De Graef, Manesh Nautiyal, Wim M. De
Borggraeve, Jef Rozenski, Sergei V. Strelkov, Stephen D. Weeks, Arthur Van
Aerschot
PII:
S0223-5234(19)30363-0
DOI:
https://doi.org/10.1016/j.ejmech.2019.04.045
EJMECH 11280
Reference:
To appear in: European Journal of Medicinal Chemistry
Received Date: 13 December 2018
Revised Date: 11 March 2019
Accepted Date: 16 April 2019
Please cite this article as: D. De Ruysscher, L. Pang, S. De Graef, M. Nautiyal, W.M. De Borggraeve,
J. Rozenski, S.V. Strelkov, S.D. Weeks, A. Van Aerschot, Acylated sulfonamide adenosines as potent
inhibitors of the adenylate-forming enzyme superfamily, European Journal of Medicinal Chemistry
(2019), doi: https://doi.org/10.1016/j.ejmech.2019.04.045.
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to
our customers we are providing this early version of the manuscript. The manuscript will undergo
copyediting, typesetting, and review of the resulting proof before it is published in its final form. 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.

ACCEPTED MANUSCRIPT
Acylated sulfonamide adenosines as potent inhibitors of the adenylate-
forming enzyme superfamily
Dries De Ruysscher1, Luping Pang1,2, Steff De Graef2, Manesh Nautiyal1, Wim M. De
Borggraeve3, Jef Rozenski1, Sergei V. Strelkov2, Stephen D. Weeks2,# and Arthur Van
Aerschot1,#,*
1 Medicinal chemistry, KU Leuven, Herestraat 49 BOX 1030, 3000 Leuven, Belgium
2 Biocrystallography, KU Leuven, Herestraat 49 BOX 822, 3000 Leuven, Belgium
3 Molecular design and synthesis, KU Leuven, Celestijnenlaan 200F BOX 2404, 3001
Leuven, Belgium
* To whom correspondence should be addressed. Tel: +32 16 37 26 24; Fax: +32 16 33 73 40;
Email: arthur.vanaerschot@kuleuven.be
# Both authors equally supervised and gave guidance to this work
Abstract
The superfamily of adenylate-forming enzymes all share a common chemistry. They activate a
carboxylate group, on a specific substrate, by catalyzing the formation of a high energy mixed
phosphoanhydride-linked nucleoside intermediate. Members of this diverse enzymatic family
play key roles in a variety of metabolic pathways and therefore many have been regarded as
drug targets. A generic approach to inhibit such enzymes is the use of non-hydrolysable
sulfur-based bioisosteres of the adenylate intermediate. Here we compare the activity of
compounds containing a sulfamoyl and sulfonamide linker respectively. An improved
synthetic strategy was developed to generate inhibitors containing the latter that target
isoleucyl- (IleRS) and seryl-tRNA synthetase (SerRS), two structurally distinct representatives
of class I and II aminoacyl-tRNA synthetases (aaRSs). These enzymes attach their respective
amino acid to its cognate tRNA and are indispensable for protein translation. Evaluation of
the ability of the two similar isosteres to inhibit serRS revealed a remarkable difference, with
an almost complete loss of activity for seryl-sulfonamide 15 (SerSoHA) compared to its
sulfamoyl analogue (SerSA), while inhibition of IleRS was unaffected. To explain these
observations, we have determined a 2.1 Å crystal structure of Klebsiella pneumoniae SerRS in
complex with SerSA. Using this structure as a template, modelling of 15 in the active site
predicts an unfavourable eclipsed conformation. We extended the same modelling strategy to
representative members of the whole adenylate-forming enzyme superfamily, and were able to
disclose a new classification system for adenylating enzymes, based on their protein fold. The
results suggest that, other than for the structural and functional orthologues of the class II
aaRSs, the O to C substitution within the sulfur-sugar link should generally preserve the
inhibitory potency.
Keywords
Adenylating enzymes, aminoacyl-tRNA synthetases, hydrolytically stable inhibitors, eclipsed
interactions, homo-adenosine derivatives
1. Introduction

ACCEPTED MANUSCRIPT
Adenylate-forming enzymes all catalyze the same reaction: the condensation between ATP
and a weakly nucleophilic carboxylate group of a substrate molecule, with the concomitant
release of a pyrophosphate byproduct. The resulting unstable acyl-AMP intermediate (Figure
1A) can be converted into a carboxylic acid derivative as it readily reacts with common
nucleophiles such as alcohols, thiols or amines, giving rise to esters, thioesters and amides
respectively. The large superfamily of enzymes, functionally linked by this chemistry, are
involved in a variety of essential cellular processes, that include but are not limited to
nucleotide and enzyme co-factor biosynthesis, lipid catabolism, and the modification and
aminoacylation of tRNAs necessary for protein translation [1]. The critical roles of these
enzymes have stimulated considerable interest in targeting them for the development of new
antimicrobials [2,3].
Figure 1 (190 mm width; for convenience, all figures and schemes are provided separately as
well): Chemical structures of the acyl-adenylate intermediate and different non-hydrolysable
alternatives. [A] The natural acyl-adenylate; a variety of R-groups are possible as shown in
Figure 2. [B] The high affinity acyl-SA, mimicking the reaction intermediate, is prone to
cyclative degradation into N-cycloadenosine (blue arrows). [C] Alternative sulfur-based
intermediate analogues. Left: 5’-deoxyadenosine (SoA); Middle: 3-deazadenosine (S3DA)
and the reported sulfonamide (SoHA) that target aminoacyl-tRNA synthetases, where the R-
group corresponds to the side chain of a proteinogenic amino acid (aa); Right: Stable
sulfamide based analogue targeting biotin ligase.
Despite the common chemistry these enzymes share, they can be divided into five structurally
distinct classes (Figure 2). While a previously reported classification system employed a
mixture of structural and functional similarity to make such divisions [1], we have updated
this to account for newly available X-ray crystallographic structures and separated the clusters
purely based on their protein fold. Class I members inherit their grouping as a result of older
nomenclature in the literature for the main representatives of this group, the HIGH-motif
containing Class I aminoacyl-tRNA synthetases (aaRS) [4,5]. Furthermore, this class also
includes enzymes belonging to the N-type ATP pyrophosphatase family, such as GMP and
NAD synthetase [6]. Both enzyme families are part of a larger functionally diverse group
collectively known as the HUP family (CATH 3.40.50.620 [7]). Similar to the first class, the
previously annotated Class II aminoacyl-tRNA synthetases designate the numbering of the
second structural cluster (CATH 3.30.930.10 [4,5], Figure 2). This family also includes biotin
and lipoate ligases that attach their respective substrate to protein targets [8]. Members of the
Class III cluster (originally termed Class I in the older classification scheme [1]) are part of
the functionally diverse ANL family (CATH 3.40.50.980 [9]). This family includes acetyl-

ACCEPTED MANUSCRIPT
CoA synthetase, fatty acid ligase and the non-ribosomal peptide synthetases (NRPS). Class IV
adenylate-forming enzymes encompass the NRPS-independent peptide (NIS) synthetases
involved in siderophore biosynthesis [10]. Finally, the Class V group contains enzymes that
transfer an activated carbamoyl moiety, which is necessary in some antibiotic biosynthetic
pathways as well as for the formation of the universal N6-threonylcarbamoyl adenosine tRNA
modification [11].
Figure 2 (140 mm width): Proposed structure-based classification system of the adenylate-
forming enzyme superfamily. The tertiary structure of an example member for each class is
shown as a cartoon representation, with β-strands colored magenta and α-helices in cyan.
The molecular structure of the corresponding acyl group is shown for each. Class I, tyrosyl-
tRNA synthetase (PDB ID 1N3L); Class II, alanyl-tRNA synthetase (PDB ID 4XEM); Class
III, firefly luciferase (PDB ID 2D1S); Class IV, achromobactin synthetase protein D (PDB ID
2X3J); Class V, L-threonylcarbamoyladenylate synthase (SUA5, PDB ID 4E1B). For clarity
only a monomeric subunit is shown for the representatives that form higher-order oligomers.
In general, mimicking the acyl-AMP intermediates by hydrolytically stable competitive
inhibitors has proven a potent way of targeting the adenylate-forming enzymes from the
different structural classes [12,13,14]. Focusing on aaRSs, enzymes which catalyse the
activation and esterification of amino acids to their cognate tRNA, there have been many
successful approaches to enhance the stability of the aminoacyl-adenylate intermediate by
substitution of the mixed phosphoanhydride linker with a sulfur based isoster [15,16,17,18].
Amongst these the aminoacyl-sulfamate isosteres (Figure 1B, where R is a decarboxylated
amino acid) have been established as the most potent aaRS inhibitors during the last two
decades, but further elaboration of this scaffold to improve uptake has been limited as the
sulfamate linker is prone to N-cycloadenosine formation (Figure 1B) [12]. Numerous attempts
have been made to improve the stability of these isosteres. Our group deleted the 5’-oxygen
atom resulting in aaSoA analogues (Figure 1C, aaSoA) [19]. This leads to a less electrophilic

ACCEPTED MANUSCRIPT
5’-carbon atom because the resulting sulfonamide moiety is a poorer leaving group compared
to the sulfamoyl group. Unfortunately, the majority of the resulting compounds no longer
efficiently inhibited their target aaRS. More recently, we applied another strategy involving
the synthesis of 3-deazaadenosine analogues (Figure 1C, aaS3DA). These compounds lack the
N3 atom of the adenine base, which is responsible for the nucleophilic attack on the 5’-carbon
atom. An evaluation of six different analogues highlighted a specific class dependent
inhibitory activity [20].
Alternative stabilization strategies have been applied to sulfamate based acyl-AMP
bioisosteres targeting other adenylate-forming enzymes. In an attempt to limit
disproportionation of biotinoyl-sulfamoyladenosine, a potent inhibitor of the Class II member
biotin ligase, modifications of the adenine base [12] or addition of a methyl group to the C5’-
position of the ribose [13] resulted in stabilization of the compounds. This was rationalized as
likely the result of steric hindrance, preventing the syn-conformation required for the adenine
N3 attack on the C5’ atom. In both cases the modifications had a minimal effect on binding
affinity, but the base modifications dramatically reduced the activity on whole cells [12]. In
addition to these scaffold elaborations, attempts have been made to substitute the 5’-oxygen
with alternative atoms [12,21]. Substitution with an amine, yielding biotinoyl-AMS resulted
in a more stable compound that is a potent and selective antitubercular [12]. Although
chemically more difficult, replacement of the same atom with a methylene group also resulted
in a potent inhibitor that was resistant to chemical and biochemical induced degradation
[12,13].
The reported stability and effectiveness of the adenosine sulfonamide scaffold raises an open
question as to whether it can be applied to other adenylate-forming enzymes. We developed
an improved synthetic strategy to generate two aminoacyl analogues (Figure 1C, aaSoHA)
targeting the Class I isoleucyl-tRNA synthetase (IleRS) and the Class II seryl-tRNA
synthetase (SerRS). Following synthesis of IleSoHA 14 and SerSoHA 15 (Scheme 1), the
activity against the corresponding enzymes was evaluated. Combined with X-ray
crystallographic studies and modelling we extended this characterization to representatives of
each class of the adenylate-forming enzyme superfamily.

ACCEPTED MANUSCRIPT
2. Results
2.1. Synthesis of aminoacylated sulfonamide homo-adenosines (aaSoHAs)
Reagents and conditions: [A] (a) DMP, PTSA·H2O, acetone, rt, overnight; (b) TBDMS-Cl,
imidazole, anh. DMF, rt, 3 d; (c) Boc2O, DMAP, TEA, anh. DMF, 0 °C - rt, 3 d; (d)
TBAF·3H2O, anh. THF, rt, overnight; (e) Dess-Martin periodinane, anh. DCM, rt, 1.5 h; (f)
(1) 9, LiHMDS, anh. DMF/THF, -78 °C, 2.5 h; (2) 6, anh. THF, -78 °C - rt, 2 d; (g) H2, Pd/C,
anh. MeOH, rt, 3 d; (h) TFA, anh. DCM, 0 °C, 2 h; (i) Boc-aa-(Bn)-OSu, Cs2CO3, anh. DMF,
0 °C – rt, 2 d - 4 d; (j) H2, Pd/C, anh. MeOH, rt, 2 d; (k) TFA/H2O, 0 °C - rt, 3 h; [B] (a)
Boc2O, DMAP, TEA, anh. DCM, 0°C - rt, overnight; (b) (1) (iPr)2NH, n-BuLi, anh. THF, -78
°C - 0 °C, 1 min.; (2) 8, anh. THF, -78 °C, 10 min.; (3) Ph2P(O)Cl, -78 °C, 1.5 h.
Scheme 1 (190 mm width): [A] Successful reaction sequence leading to the desired
sulfonamide nucleosides IleSoHA 14 and SerSoHA 15; [B] Synthesis of Wittig-Horner
reagent 9.
The synthetic scheme for the desired compounds 14 and 15 (Scheme 1A) started with
selective introduction of the isopropylidene moiety. TBDMS protection of the remaining
primary alcohol and subsequent di-Boc protection of the heterocyclic amine gave compound
4. Selective deprotection of the silyl moiety resulted in protected nucleoside 5. In analogy
with the work of Shi et al., 5 was used in a Wittig-Horner reaction which, in our hands,
afforded only 10 with a yield of 23% [12]. In spite of several attempts, we could not
reproduce the previously published results whereby only 11 was obtained with a better yield
than what we achieved for 10. The employed Wittig-Horner reagent 9 was synthesized,

ACCEPTED MANUSCRIPT
utilizing methanesulfonamide 7 as the starting material in which the amino group was first
Boc protected (Scheme 1B). The di-anion of 8 was then prepared with freshly made LDA,
whereafter it was treated with diphenylphosphinic chloride in a substitution reaction (Scheme
1B). An important issue in this Wittig-Horner reaction was the instability of aldehyde 6
(Scheme 1A). This compound was synthesized in a high (isolated) yield (88%) utilizing Dess-
Martin periodinane, but could not be stored overnight. Degradation of 6 was confirmed by
NMR analysis and as a consequence, the crude aldehyde had to be used immediately upon
production in order to obtain 10. In addition, we observed that the selective (E)-alkene
forming reaction limited the amount of starting material 5 which can be used, because
attempts to scale-up (0.5 g) in our hands no longer afforded the desired product 10.
Quantitative reduction of the double bond using hydrogen gas resulted in compound 12,
which was confirmed by mass analysis. Hence, 12 was immediately deprotected under
anhydrous acidic conditions. Unfortunately, the desired compound 13 in our hands
decomposed during work-up, in contrast to previously reported literature [12].
Scheme 2 (90 mm width): Two unsuccessful C-C coupling reactions (similar as in Scheme 1A,
steps e and f) using either synthesized Wadsworth-Horner-Emmons reagent 16 or Wittig-
Horner compound 18 (similar as in Scheme 1B).
To avoid the use of a Boc protecting group on the sulfonamide moiety, we chose to prepare
Wadsworth-Horner-Emmons and Wittig-Horner compounds 16 and 18 respectively (Scheme
2) in a similar reaction sequence as described for Boc protected Wittig-Horner product 9
(Scheme 1B). Both benzhydryl (16) and Cbz (18) protecting groups can be removed upon
hydrogenolysis, thus avoiding the use of acid. Compounds 16 and 18 were both utilized for
the C-C coupling reaction with crude aldehyde 6, but both reactions failed to give the desired
products 17 or 19 respectively (Scheme 2). The reaction of 6 with 16 resulted in the formation
of two unknown (and incorrect) compounds, while reaction of 6 with 18 unexpectedly caused
a Boc transfer from the adenine base to the sugar part of 5, resulting in protected 5’-O-Boc-
adenosine.
Reagents and conditions: (a) (1) DMAP, SOCl2, anh. MeCN, 0 °C - rt, 4 h; (2) anh. MeCN, rt,
overnight; (b) Boc2O, DMAP, TEA, anh. MeCN, 0 °C - rt, 3 d; (c) (1) (iPr)2NH, n-BuLi, anh.

ACCEPTED MANUSCRIPT
THF, -78 °C - 0 °C, 1 min.; (2) N-benzhydrylmethanesulfonamide, anh. DMF, -78 °C, 1 h; (3)
21, anh. DMF, -78 °C, 1 h; (4) anh. DMF/THF, -78 °C - rt, 3 d.
Scheme 3 (190 mm width): Unsuccessful nucleophilic substitution approach which would
have resulted in the desired sulfonamide 22.
A different approach for these modified Wittig reactions had to be found for obtaining the C-
C linked molecules. Therefore, a nucleophilic substitution was envisioned in order to extend
the 5’-carbon side chain (Scheme 3). Hereto, the remaining alcohol group of 2 was substituted
by a chlorine atom, using thionyl chloride as the halogen donor. Chlorine would act as a
stable, but still good leaving group (installing better leaving groups immediately results in the
formation of N-cycloadenosine (Figure 1A), as seen with the sulfamate analogues of 14 and
15). The adenine base was then di-Boc protected to obtain 21 devoid of acidic protons. This
compound was subsequently reacted with the di-anion of N-benzhydrylmethanesulfonamide,
which should lead to sulfonamide 22. Unfortunately, no reaction occurred, even at room
temperature, while at elevated temperatures, decomposition of the nucleoside was observed.
In view of decomposition of nucleosides at elevated temperatures in acidic medium, prior to
work-up, the mixture, as obtained in step h of Scheme 1, was neutralized immediately
following reaction with an excess of bicarbonate. Following conventional purification, 13 was
isolated as its TFA salt in moderate yield. Amino acids isoleucine and serine were then
coupled to the sulfonamide moiety using Cs2CO3. Following removal of all remaining
protecting groups, crude products 14 and 15 were purified by preparative reversed-phase
HPLC, resulting in the desired analogues IleSoHA 14 and SerSoHA 15.
2.2. Inhibitory activity
The ability of 14 and 15 to inhibit the respective aaRS was evaluated using an in vitro
aminoacylation assay, monitoring their effect on the transfer of the appropriate C14
radiolabelled amino acid to tRNA (Figure 3). In both cases the purified recombinant E. coli
protein was utilized with a total tRNA pool from the same organism. Compound 14 showed
potent inhibitory potential for the target E. coli IleRS. The inhibition profile, including the
steep increase in aaRS activity between 50 and 10 nM of the compound, is in line with
previous studies of the sulfamoyl analogue [20]. This observation is also reflected in an even
slightly improved apparent Ki value of 1.2 nM for 14 (Supplemental Figure S1) compared to
its sulfamoyl analogue (Ki,app = 1.9 nM [20]). Surprisingly, 15 demonstrated little inhibitory
activity. At a final concentration of 2 µM, it only reduced the ability of E. coli SerRS to
aminoacylate tRNA by 30 percent. This is more than a 200-fold reduction in inhibitory
capacity when compared to 14, and a marked reduction to the previously reported activity of
the analogous sulfamoyl, which showed an apparent Ki of 0.18 nM [20].

ACCEPTED MANUSCRIPT
Figure 3 (90 mm width): Aminoacylation activity of E. coli ileRS and serRS in the presence of
the respective sulfonamide linked intermediate analogue. Values were normalized based on
the measured activity of each enzyme in the absence of inhibitor. Experiments were performed
in triplicate, the mean and standard error are presented.
2.3. Structural studies
To understand why SerSoHA 15 showed a dramatically reduced activity relative to 14, we
solved the X-ray crystallographic structure of Klebsiella pneumoniae SerRS (KpSerRS) in
complex with the canonical SerSA analogue (Figure 4A and Supplemental Table 1).
Examination of datasets of similar crystals soaked with high concentrations of 15 failed to
reveal any suitable electron density corresponding to the compound (data not shown).
KpSerRS shares 93.7% sequence identity (97.4% similarity) with the E. coli homologue,
therefore interpretation of the ligand bound structure can be compared to the in vitro studies
with the latter enzyme. The asymmetric unit of the crystal contains one protomer of the
dimeric enzyme (Figure 4A), the canonical homodimer formed by a crystallographic 2-fold
rotation axis. The KpSerRS structure shows high similarity to the homologous enzyme from
other bacterial species [22,23], with an N-terminal antiparallel coiled-coil arm, involved in
tRNA binding, and a mixed α/β catalytic domain (Figure 4A and Supplemental Figure S2).
The 2.1 Å crystal structure, currently the highest resolution for a bacterial SerRS orthologue
in the PDB, unambiguously reveals the conformation of the intermediate analogue as
evidenced by the polder.omit map (Figure 4A, insert). The compound makes numerous
interactions with the protein, showing high complementarity with the large active site. Careful
examination of the 5’-oxygen atom suggests it only makes weak Van der Waals interactions
with the sidechain NH1 and NH2 atoms of Arg268. This residue, which is part of the catalytic
core, is found in all Class II aaRSs where it anchors the N-terminal side of a conserved
sequence motif [5,20]. In the ligand bound structure Arg268 makes H-bond and salt bridge
interactions with the sulfamoyl and carbonyl group of the intermediate analogue (Figure 4B).
The latter is mediated by a strong ionic interaction as the tautomer of conjugated peptide
bond, with an estimated pKa of 3.1 using Schrodinger Epik software [24], will be
deprotonated at physiological pH. This calculated value is comparable to what was
determined in previous literature for equivalent molecules [25]. In the case of 15, calculations
suggest a slight increase in the pKa of this group to 4.1, therefore it is also predicted to be
deprotonated under the native conditions that were evaluated and therefore is likely not the
cause of the observed loss of activity.
Using the empirically determined structure, SerSoHA 15 was modelled in the enzyme active
site maintaining the same conformation of SerSA. The substitution of the O5’ atom with a

ACCEPTED MANUSCRIPT
methylene group is readily accommodated, with no inter- or intramolecular clashes.
Maintaining a torsional angle of 127.8° between the C4’ atom and the sulfone group, the
minimized model shows that the additional protons introduced by the methylene substitution
are found in an energetically unfavorable (nearly) eclipsed position (Figure 4C). To start with,
a torsional angle of 13° is observed between the C5’ and C6’ protons (Figure 4C, insert).
Secondly, the other C6’ proton is offset from the C4’-C5’ bond by a 16° torsional angle.
Finally, a similar torsional angle of 19° is seen between the second C5’ proton and the C6’-S
bond, a near overlap of atoms that is also observed in the experimental SerSA bound structure
between the C5’ proton and the O5’-S bond. Hence this structural pose of SerSoHA 15 is
energetically unfavorable.

ACCEPTED MANUSCRIPT
Figure 4 (90 mm width): (A) The crystal structure of K. pneumoniae SerRS bound to the
intermediate analogue SerSA (PDB ID 6H9X). The protein is depicted as a cartoon
representation and the SerSA shown as spheres. The insert shows the phenix.polder omit map
for the bound SerSA contoured at 6σ. (B) Complementarity of SerSA with the enzyme active
site. The protein surface has been calculated using a 1.4 Å probe. The side chains of
interacting residues are shown as stick representations, the carbon atoms are coloured based
on the secondary structure in panel A. R268 interacts with the carbonyl and sulfamate oxygen
atoms of SerSA (yellow dashed lines). (C) Model of SerSoHA 15 based on the pose of SerSA,
showing the 3’-endo ribose conformation and the eclipsing protons and bonds. The insert
shows the torsional angle between the eclipsed C5’ and C6’ protons when looking down the
C6’-C5’ bond. The predicted protonation state at physiological pH is shown.
2.4. Comparative modelling studies
As IleSoHA 14 demonstrated high affinity for its target aaRS (Figure 3), we wanted to
examine whether the conformation of the atoms around the C5’-C6’ bond of this compound
was similar to that observed for SerSoHA 15 bound to SerRS. As there are no high resolution
(sub 2.5 Å) crystal structures of IleRS in complex with an aminoacyl-AMP intermediate, or
analogues thereof, we used the 1.85 Å resolution structure of the related leucyl-tRNA
synthetase (LeuRS) in complex with LeuSA for comparative analysis (Figure 5A; PDB ID
2V0C). Using a re-refined structure of this protein-ligand complex LeuSoHA was modelled in
the active site, by substituting the O5’ atom of LeuSA with a methylene group whilst
maintaining the same bound conformation (Figure 5B). In the original LeuSA structure one of
the oxygen atoms of the sulfamoyl group makes a H-bond with a backbone atom which is
maintained in the LeuSoHA model. Further examination of the model shows no clashing
atoms between the compound and the protein suggesting this pose is a viable conformation.
Focusing on the intramolecular contacts of LeuSoHA a staggered relationship for the C5’ and
C6’ protons (torsional angle of 42°) was observed (Figure 5B, insert). Similar values were
also observed between the other C5’ proton and the C6’-S bond (torsional angle of 51°). The
second C6’ proton and the C4’-C5’ bond exhibited a torsional angle of 35°.
In order to investigate if we can extrapolate the above analyses of LeuSoHA and SerSoHA 15
to all other Class I and Class II aaRSs respectively, we examined the available intermediate-
bound or non-hydrolysable analogue-bound crystal structures of the different aaRS
orthologues (Figure 5C and 5D). The ligands in the Class I representatives were superposed
using the C1’, C4’ and O4’ atoms of the ribose, which all demonstrated the same 2’-endo
pucker (Figure 5C). It is clear from this overlap that the position of the sulfamate group and
the adenine ring is not completely fixed between enzymes. Rotating around the base N9 atom,
the angle between the most displaced adenine rings is 16°, resulting in a maximum distance of
1.98 Å between their distal ends. Similarly, with a mean value of 161.6° but a standard
deviation of 13.7°, the torsion angle between the C4’-C5’ and the C6’-S bonds also
demonstrate some variability. In all cases though the staggered conformation of the atoms
circling the C5’-C6’ bond is maintained, suggesting that the O5’ to C substitution will not
result in a detrimental effect on binding of the inhibitor to any Class I aaRS. The same ring
based superposition strategy was also applied to all available Class II bound ligands, each
showing the same 3’-endo ribose pucker (Figure 5D). The overall conformation of the Class II
inhibitors is more strictly fixed, with an all non-H atom RMSD of 0.417 Å. In particular the
torsion angle between the C4’-C5’ and the C6’-S bonds, with a mean value of 127.4°, has a
standard deviation of 7.5°- just over half that was observed for the Class I intermediate

ACCEPTED MANUSCRIPT
analogues. This strongly suggests that the energetically unfavorable conformation observed
for the SerSoHA model would be replicated with an O5’ to C substitution in each of the
examined structures, likely negatively affecting the binding affinity of all Class II aaRS
targeting sulfonamide-linked intermediate analogues.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT
Figure 5 (90 mm width): [A] Structure of re-refined T. thermophilus LeuRS, shown as a
cartoon representation, bound to leucyl-sulfamoyl-adenosine (LeuSA - green spheres); [B]
Stick representation of the bound pose of LeuSA. The insert in the lower panel shows the
torsional angle between the staggered C5’ and C6’ protons when looking down the C6’-C5’
bond. The predicted protonation state at physiological pH is shown; [C] Superposition of
class I aaRS intermediate analogues, making use of the available ligand bound structures for
IleRS (PDB ID 1JZQ – red), TyrRS (PDB ID 6I5Y– yellow), LeuRS (PDB ID 2V0C – green),
MetRS (PDB ID 1PFY– blue), GlnRS (PDB ID 1QTQ – white), GluRS (PDB ID 3AFH– grey)
and TrpRS (PDB ID 1I6K – black); [D] Overlap of the bound intermediate analogues of class
II aaRS representatives LysRS (PDB ID 3E9H – red ), SerRS (PDB ID 6H9X – green), ProRS
(PDB ID 2J3L – magenta), ThrRS (PDB ID 1NYQ– cyan), GlyRS (PDB ID 5F5W– blue),
AsnRS (PDB ID 1X55– white), AspRS (PDB ID 1G51– grey) and AlaRS (PDB ID 3HXZ–
orange). The stated colors correspond to the carbon atoms.
As the two aaRS classes are members of a larger superfamily of adenylate-forming enzymes
(Figure 2), we extended our structural analyses to other representatives, to evaluate whether
the acyl-sulfonamide linker strategy can be employed to this structurally and functionally
diverse group of enzymes (Figure 6). A search for suitable ligand bound crystal structures,
specifically of enzymes that have been previously considered as druggable targets, was
performed. In all cases the O5’ of the acyl-SA intermediate analogue was replaced by a
methylene group in the minimised structure. For the Class I NAD+ synthetase, a member of
the N-type ATP pyrophosphatase family, the conformation of the acyl-adenylate analogue is
distinct to that observed for the Class I aaRS (1KQP, Figure 6A). These variances include the
pucker of the ribose and the position of the adenine relative to the sugar, and reflect
differences in the binding mode between these two class members. Despite this, the
sulfonamide based analogue of the NAD+-adenylate intermediate displays a staggered
conformation with a 164° torsion angle between the C4’-C5’ and C6’-S bonds, close to the
mean value seen for the Class I tRNA synthetases. This positioning of the sulfonamide group
within the active site, relative to the ribose, results in a 50° offset between the C6’and C5’ H-
atoms (Figure 6A).
A biotinoyl-SoHA intermediate analogue, targeting the alternative Class II member biotin-
protein ligase BirA (Figure 6B), has previously been described as a potent inhibitor of this
enzyme with demonstrable antitubercular activity [12,13]. This contrasts with our results for
SerRS and what is predicted for the other Class II aaRS. However, while BirA shares the core
six-stranded antiparallel β-sheet flanked by structurally conserved α-helices seen in SerRS
and other class II aaRS, the orientation of the loop region involved in nucleoside recogition is
significantly different (Supplemental Figure S2). This difference leads to an alternative
conformation of the intermediate whereby the S-C6’-C5’-C4’ torsion angle of biotinoyl-
SoHA is 201.8°, resulting in no eclipsing interactions around the C6’-C5’ bond (Figure 6B).
Comparatively, the conformation of the biostere linker in biotinoyl-SoHA resembles more
that of the intermediate analogues bound to the structurally distinct members of the Class III
adenylate-forming enzymes (Figure 6C). Both share the same 3’-endo conformation of the
sugar ring, and a similar C5’-C6’ bond position relative to the ribose. In the case of the
inhibitor conformation bound to the Mycobacterium tuberculosis long-chain-fatty-acid-AMP
ligase the sulfone group is positioned in an anti configuration, relative to the C4’ atom, with a
S-C6’-C5’-C4’ torsion angle of 179.2°. This pose results in a staggered relation of the C5’
and C6’ protons similar to that seen for biotinoyl-SoHA (Figure 6B and 6C).

ACCEPTED MANUSCRIPT
The Class IV NIS members of the adenylate-forming family, are all involved in siderophore
biosynthesis [10]. These chelating molecules sequester iron, in nutrient-poor environments
and are important virulence factors in pathogenic bacteria [10]. Despite the recognition that
this family are suitable targets for antimicrobial development, no structures are available of a
representative enzyme bound to the reaction intermediate, or an analogue thereof. Therefore
N-citryl-ethylenediamine-SoHA was modeled and minimised, by linking the natural
substrates observed in the crystal structure of achromobactin synthetase protein D (PDB ID
2X3J, Figure 6D). This intermediate seems to be an outlier compared to all other adenylate
classes in that the adenine base adopts a syn conformation. The sugar ring has a 2’-endo
conformation, observed in the intermediate of the Class I aaRSs (Figure 5B,C), but because of
the position of the adenine base, the C6’-C5’ bond is rotated away from the nucleoside
scaffold. Despite this difference, the S-C6’-C5’-C4’ torsion angle is 172.0°, similar to that
observed for LeuSoHA, and the angular offset of the C5’ and C6’ protons is 55° reflecting a
staggered conformation (Figure 6D). Finally, the threonylcarbamoyl-adenylate intermediate in
the Class V representative L-threonylcarbamoyl-adenylate synthase (4E1B, Figure 6E) has a
similar nuceloside binding pose as the Class III inhibitor. Uniquely though, amongst the
adenylate-forming superfamily, the sulfone group is positioned at an acute angle to the C4’
atom with a torsion angle of 76.9°. However, the C5’ and C6’ protons demonstrate a
staggered offset with a torsion angle of 39° for the closest proton pair (Figure 6E).
Combined, this family-wide analysis indicates that the O5’ to C substitution in the
sulfonamide bioisosteres or the acyl-adenylate would likely be tolerated by all classes of the
adenylate-forming enzyme superfamily except for the Class II aaRSs.

ACCEPTED MANUSCRIPT

ACCEPTED MANUSCRIPT
Figure 6 (90 mm width): Conformation analysis of sulfonamide-linked biosters for the
different intermediates of class representative adenylate-forming enzymes. (A) Left: C-5’-
(nicotinamide-adenine-dinucleotidyl)-adenosyl-methansulfonamide (R1 = -ribose-ADP),
Right: Class I crystal structure of Bacillus subtilis NH3-dependent NAD+ synthetase (PDB ID
1KQP); (B) Left: C-5’-(biotinoyl)-adenosyl-methansulfonamide, Right: Class II crystal
structure of Mycobacterium tuberculosis biotin-protein ligase (PDB ID 3RUX); (C) Left: C-
5'-(11-phenoxyundecanoyl)-adenosyl-methansulfonamide (R2 = -(CH2)5-O-Ph), Right: Class
III crystal structure of Mycobacterium tuberculosis long-chain-fatty-acid-AMP ligase (PDB
ID 5HM3); (D) Left: C-5’-(citryl)-adenosyl-methansulfonamide, Right: Class IV crystal
structure of Pectobacterium chrysanthemi achromobactin synthetase protein D (PDB ID
2X3J); (E) Left: C-5’-(threonylcarbamoyl)-adenosyl-methansulfonamide, Right: Class V
crystal structure of Sulfolobus tokodaii L-threonylcarbamoyl adenine synthase (PDB ID
4E1B).
3. Discussion and conclusion
Non-hydrolysable acyl-adenylate analogues of the enzymatic intermediate of a variety of
adenylate-forming enzymes have been consistently shown to have inhibitory activity [2,3].
Functioning as bi-substrate competitive inhibitors, the potency of these compounds emanates
from the entropic favorability of binding over the unlinked substrates. Sulfur-based
bioisosteres of the labile phosphoanhydrides have been typically employed, of which the
sulfamate containing compounds have proven to be particularly effective [26,27]. In the case
of targeting aminoacyl-tRNA synthetases subnanomolar binding has been observed in vitro,
but unfortunately this has not translated to in vivo activity due to the high polarity of the
compounds. Attempts to modify this scaffold, for prodrug strategies [28,29], has been limited
due to the chemical instability of the compound [12]. In this work, we synthesized
aminoacylated sulfonamide adenosines IleSoHA 14 and SerSoHA 15 (Scheme 1), aimed to
inhibit their corresponding Class I and II aminoacyl-tRNA synthetase. The rationale for this
oxygen-carbon substitution was to make hydrolytically stable aaSA analogues that are
resistant to degradation at the 5’-position due to the less electrophilic 5’-carbon atom (Figure
1A).
Inspired by the work of Shi et al. [12], we first attempted a Boc-Wittig-Horner procedure in
order to obtain the C-C linked sulfonamide nucleosides. Unfortunately, compound 13 in our
hands decomposed during work-up, in contrast to the previous report. Therefore, we explored
two related reaction routes, using compounds 16 and 18, but both failed to provide the desired
products (Scheme 2). A nucleophilic substitution approach was subsequently tried in vain in
order to extend the amino acid side chain (Scheme 3). With the intention of solving the
degradation problem, we slightly modified the Boc-Wittig-Horner reaction procedure
avoiding acidic circumstances. This alteration in the reported reaction procedure enabled us to
obtain the desired products 14 and 15 (Scheme 1) in a reproducible way.
An evaluation of the in vitro inhibitory activity of 14 and 15 surprisingly demonstrated a class
correlated difference, with at least a 10000-fold decrease in the IC50 for the SerSoHA 15 over
that reported for the analogous sulfamoyl [20] (Figure 3). Detailed examination of the
modeled SerSoHA bound conformation, employing a 2.1 Å resolution crystal structure of K.
pneumoniae SerRS soaked with SerSA suggests that an accumulation of torsional penalties,
resulting from the energetically unfavorable eclipsed positions of the C5’ and C6’ protons
with regard to each other, and with the C6’-S and C4’-C5’ bonds respectively, are in part
likely to cause the observed loss of activity (Figure 4C). By itself this appears surprising, as

ACCEPTED MANUSCRIPT
the relative energy between the staggered and eclipsed conformations of butane is only about
4 kcal/mol [30]. Examination of the other properties of the molecule suggest though that it is
the major determinant.
An estimate of the pKa of the sulfonamide nitrogen predicts it to be deprotonated under
physiological conditions in analogy with the sulfamoyl moiety. This same bioisosteric linker
has been successfully employed in compounds targeting M. tuberculosis biotin ligase, an
alternative Class II adenylate-forming enzyme (Figure 6B and Supplemental Figure S2). In
this particular enzyme, the phosphate makes numerous charge-based interactions with basic
residues in the protein active site including a conserved lysine that is essential for activity
[21,31]. As the deprotonated sulfonamide effectively replicates these ionic interactions in this
birA active site, by extrapolation this should also apply in the case of the aminoacyl-SoHA
analogues.
The switch from the S-O (1.577 Ǻ) to a S-C (1.758 Ǻ) bond results in a small shift in bond
length (∆ = 0.181 Å) which may have an effect on the overall conformation of the resultant
compound [32]. This value though is within one standard deviation of the observed S-C6’-
C5’-C4’ torsion angles from available Class II ligand bound crystal structures suggesting it
can be accommodated. Ultimately, improved binding of the two bisubstrate-based competitive
inhibitors over the amino acid and ATP substrates, which cumulatively make a similar
number of interactions with the aaRS, is dictated by an entropic benefit [33]. It therefore
appears that the torsion angle penalty caused by the O5’-C substitution in the bound state of
15 affects this profoundly.
A comprehensive examination of all Class II aaRSs shows that the conformation modelled for
SerSoHA 15 bound to SerRS is not just restricted to this orthologue. Superposition of this
scaffold on the intermediate observed in the available structures shows a highly restricted bent
conformation, where the O5’-S bond (or O5’-P in the natural adenylate) is found always
eclipsing the C6’ protons (Figure 5D). This torsion penalty, even for the regular intermediate,
appears to be overcome by extensive interactions that are made between the aminoacyl-
adenylate and the active sites of the different Class II aaRSs [34,35]. In contrast, examination
of the intermediate conformation in all available Class I aaRSs shows no eclipsed
conformation for the side-groups of the C5’-C6’ (or C5’-O), as would be predicted based on
the observed inhibitory activity results of IleSoHA 14 (Figure 3). Together this suggests that
the use of the sulfonamide linker in aminoacyl-adenylate intermediate analogues, is suitable
for Class I aaRS but should be avoided for Class II enzymes performing the equivalent
function.
To evaluate the further use of the sulfonamide-linked bioisosteres as pharmacologically
relevant molecules we examined the binding conformation of the acyl-adenylate analogue to
additional representatives of all five structurally distinct classes of the adenylate-forming
enzyme superfamily (Figures 2 and 6). In all cases we observed no steric clashes or torsion
angle penalties with the introduction of this linker suggesting it could be a viable alternative
for this diverse family. Interestingly, as stated above, a sulfonamide based intermediate
analogue has been reported to be effective against biotin ligase (BirA), with demonstrable
antitubercular activity [12]. Like SerRS, this particular enzyme is a member of the Class II
adenylate-forming enzymes which all share a core conserved β-sheet structure (Figures 2, 4,
6). That said, the binding mode in BirA, particular of the nucleoside portion, is significantly
different to that seen within the Class II aaRSs (Supplemental Figure S2). This ultimately
results in a distinct conformation of the intermediate analogue that is more similar to that

ACCEPTED MANUSCRIPT
observed in the functionally diverse Class III family enzymes, where there are no observed
issues with torsion angle overlaps.
As a final conclusion, we have observed that a theoretically innocuous atom substitution, from
a sulfamate to a sulfonamide, in a known potent class II aaRS inhibitor, has a dramatic
negative effect on the inhibitory activity of the resultant compound. Our comprehensive
analysis points to a torsion angle penalty as being the likely cause of this loss. This problem is
manifested due to conserved conformation of the aminoacyl-adenylate intermediate in all
class II aaRS, which is ultimately dictated by the enzymatic spatial arrangement of the α-
phosphate of ATP and the amino acid carboxylate group during catalysis [34,35]. A
comprehensive examination of the acyl-adenylate bound to representative members of all
adenylate-forming enzyme classes, shows considerable variation in the conformation of this
reaction intermediate between the different groups. Our analysis shows that, with the
exception of the class II aaRS, the sulfonamide-based phosphate bioisostere is likely a viable
non-hydrolysable linker for developing family-wide bisubstrate inhibitors that work by
mimicking the acyl-adenylate. The incorporation of the sulfonamide has been shown to
improve chemical stability over other sulfur-based linkers [12], and was more recently shown
to limit enzymatic-based disproportionation of the compound [13]. The latter is an important
consideration in developing inhibitors that can avoid natural resistance mechanisms [36]. The
optimized chemical synthesis strategy outlined here should permit elaboration of such
compounds, increasing opportunities to develop much needed antimicrobials.
4. Experimental section
4.1. Reagents and analytical procedures
Reagents and solvents were purchased from commercial suppliers (Acros, Sigma-Aldrich)
and used as provided, unless indicated otherwise. DMF, THF, DCM and MeOH were of
analytical grade and were stored over 4 Å molecular sieves. 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 temperature, unless
indicated otherwise. 14C-radiolabeled amino acids and scintillation liquid were purchased
from Perkin Elmer.
13
1H and C NMR spectra of the compounds dissolved in CDCl3, CD3OD, DMSO-d6 or D2O
were recorded on a Bruker UltraShield Avance 300 MHz, 400 MHz or when needed on a 600
MHz spectrometer. The chemical shifts are expressed as δ values in parts per million (ppm),
1
13
1
using the residual solvent peaks (CDCl3: H, 7.26 ppm; C, 77.16 ppm; DMSO: H, 2.50
13
1
1
13
ppm; C, 39.52 ppm; D2O: H, 4.79 ppm; CD3OD: H, 3.31 ppm; C, 49.00 ppm) as a
reference. Coupling constants are reported in Hertz (Hz). The peak patterns are indicated by
the following abbreviations: br s = broad singlet, d = doublet, m = multiplet, q = quadruplet, s
= singlet and t = triplet. High resolution mass spectra were recorded on an ion-mobility mass
spectrometer (SYNAPT G2-Si HDMS, Waters, Milford, US) equipped with a standard ESI
interface; samples were infused in 2- propanol/H2O (1:1) at 3mL min-1. For TLC, precoated
aluminum sheets were used (Merck, Silica gel 60 F254). The spots were visualized by UV
light at 254 nm. Chromatography was performed on ICN silica gel 60A 60-200. Final
products (acylated sulfonamide nucleosides) were purified using a semi-prep PLRP-S 100 Å
column (7.5 x 300 mm) connected to a Merck-Hitachi L6200A Intelligent pump. Eluent
compositions are expressed as v/v. The characterization of the final compounds by NMR and
mass are provided in the supplementary file. Compounds 2, 3, 8 and 9 were prepared

ACCEPTED MANUSCRIPT
according to literature procedures (references 37, 38, 39 and 40, respectively) and proved
identical.
4.2. Chemical synthesis of intermediates and final compounds
4.2.1. N6,N6-bis-Boc-5’-O-tertbutyldimethylsilyl-2’-3’-O–isopropylideneadenosine (4)
Compound 3 (3.46 g, 8.22 mmol, 1.0 eq.), 4-dimethylaminopyridine (0.201 g, 1.64 mmol, 0.2
eq.) and triethylamine (2.41 mL, 17.25 mmol, 2.1 eq.) were combined and put under an
atmosphere of nitrogen, whereafter dissolving the reactants in a 2.4:1 mixture of anhydrous
acetonitrile:DMF (48:20 mL, 0.12 M). Di-t-butyl dicarbonate (6.28 g, 28.8 mmol, 3.5 eq.)
was dissolved in DMF (6.0 mL) and added dropwise to the solution. The mixture was stirred
for 3 days at room temperature under nitrogen. The solvents were subsequently removed
under reduced pressure. The residue was purified by silica gel chromatography (heptane:ethyl
acetate 50:50) to give compound 4 (4.611 g, 90%) as a light-yellow foam with Rf = 0.68
1
(heptane:ethyl acetate 50:50). H NMR (300 MHz, CDCl3): δ = 8.86 (s, 1H), 8.32 (s, 1H),
6.23 (d, J = 2.6 Hz, 1H), 5.19 (dd, J = 6.1, 2.6 Hz, 1H), 4.93 (dd, J = 6.1, 2.4 Hz, 1H), 4.43
(dd, J = 6.2, 3.6 Hz, 1H), 3.82 (ddd, J = 33.3, 11.3, 3.8 Hz, 2H), 1.63 (s, 3H), 1.41 (s, 18H),
1.39 (s, 3H), 0.83 (s, 9H), 0.00 (s, 6H) ppm.
4.2.2. N6,N6-bis-Boc-2’-3’-O-isopropylideneadenosine (5).
Compound 4 (4.61 g, 7.42 mmol, 1.0 eq.) was put under an atmosphere of nitrogen and was
dissolved in anhydrous tetrahydrofuran (THF, 42.4 mL, 0.18 M). Tetrabutylammonium
fluoride trihydrate (4.68 g, 14.83 mmol, 2.0 eq.) was dissolved in anhydrous THF (14.83 mL)
and added dropwise to the solution. The mixture was stirred overnight at room temperature
under nitrogen. The solvent was removed under reduced pressure. The residue was purified by
silica gel chromatography (heptane:ethyl acetate 20:80) to give compound 5 (3.61 g, 96%) as
a white solid with Rf = 0.61 (heptane:ethyl acetate 20:80). 1H NMR (300 MHz, DMSO): δ =
8.88 (s, 1H), 8.82 (s, 1H), 6.29 (d, J = 2.6 Hz, 1H), 5.45 (dd, J = 6.1, 2.6 Hz, 1H), 5.13 (t,
1H), 5.00 (dd, J = 6.2, 2.4 Hz, 1H), 4.29 (td, J = 2.5 Hz, 1H), 3.55 – 3.50 (m, 2H), 1.56 (s,
3H), 1.39 (s, 18H), 1.34 (s, 3H) ppm.
4.2.3. N6,N6-bis-Boc-5'-deoxy-2',3'-O-isopropylidene-5'-oxoadenosine (6).
Alcohol 5 (0.1 g, 0.197 mmol, 1.0 eq.) was dissolved in anhydrous dichloromethane (7.88
mL, 0.025 M) and Dess-Martin periodinane (0.125 g, 0.296 mmol, 1.5 eq.) was added in 1
portion. The mixture was put under an atmosphere of nitrogen and stirred for 1.5 h at room
temperature. The reaction solution was cooled to 0 °C and quenched through adding 1.0 M
aqueous sodium thiosulfate (4.0 mL) and aqueous saturated NaHCO3 (5.0 mL). The mixture
was then stirred for 1.5 h at room temperature under nitrogen after which the aqueous layer
was extracted with dichloromethane (3x). The combined organic layers were washed with
water and brine and dried over anhydrous MgSO4. Following filtration, the solvent was
removed under reduced pressure. The residue was purified by silica gel chromatography
(heptane:ethyl acetate 20:80) to give compound 6 (0.088 g, 88%) with Rf = 0.42
1
(heptane:ethyl acetate 20:80). H NMR (300 MHz, DMSO): δ = 9.29 (s, 1H), 8.78 (s, 1H),
8.77 (s, 1H), 6.59 (s, 1H), 5.55 (dd, J = 6.1, 1.7 Hz, 1H), 5.51 (d, J = 6.1 Hz, 1H), 4.85 (d, J =
1.7 Hz, 1H), 1.55 (s, 3H), 1.39 (s, 18H), 1.38 (s, 3H) ppm.
4.2.4. 6-Amino-N6,N6-bis-Boc-9-[5,6-dideoxy-6-(ethoxysulfonyl)-2,3-O-isopropylidene-β-D-
ribo-hex-5-enofuranosyl]-9H-purine (11) and 4.2.5. 6-Amino-N6-Boc-9-[5,6-dideoxy-6-
(ethoxysulfonyl)-2,3-O-isopropylidene-β-D-ribo-hex-5-enofuranosyl]-9H-purine (10).

ACCEPTED MANUSCRIPT
Using alcohol 5 (0.250 g, 0.493 mmol, 1.0 eq.), crude aldehyde 6 was obtained as a yellow oil
with Rf = 0.42 (heptane:ethyl acetate 20:80) as decribed above.
Wittig-Horner reactant 9 (0.214 g, 0.542 mmol, 1.1 eq.) was put under nitrogen and dissolved
in a 1:1 mixture of anhydrous THF:DMF (5.0 mL, 0.1 M). The solution was cooled to -78 °C
and lithium bis(trimethylsilyl)amide (1.08 mL, 2.2 eq., 1.0 M in THF) was added dropwise.
The mixture was stirred for 2.5 h at -78 °C. Crude aldehyde 6 was dissolved in anhydrous
THF (4.0 mL) and was added to the reaction mixture. The solution was allowed to warm to
room temperature and was stirred for 2 days at room temperature. The solvents were removed
under reduced pressure and the residue was resuspended into water (20 mL). The aqueous
layer was acidified with aqueous HCl (1.0 M) to pH = 4 and was extracted with ethyl acetate
(3x). The combined organic layers were washed with water and brine and dried over
anhydrous MgSO4. After filtration, the solvent was removed under reduced pressure. The
residue was purified by silica gel chromatography (ethyl acetate) to give compounds 11
(0.080 g, 24%) with Rf = 0.89 (ethyl acetate) as a yellow solid and 10 (0.086 g, 30%) with Rf
1
= 0.64 (ethyl acetate) as a yellow oil. H NMR (400 MHz, CDCl3) of 11: δ = 8.84 (s, 1H),
8.11 (s, 1H), 6.82 (dd, J = 15.2, 3.6 Hz, 1H), 6.39 (dd, J = 15.4, 1.8 Hz, 1H), 6.21 (s, 1H),
5.77 (d, J = 5.9 Hz, 1H), 5.35 (dd, J = 5.9, 2.1 Hz, 1H), 4.94 (d, J = 2.2 Hz, 1H), 1.62 (s, 3H),
1.51 (s, 18H), 1.50 (s, 9H), 1.44 (s, 3H) ppm. 13C NMR (101 MHz, CDCl3) of 11: δ = 152.49,
150.96, 150.78, 149.50, 144.21, 141.76, 129.14, 128.57, 114.58, 90.95, 85.33, 84.56, 84.40,
83.97, 83.37, 27.86, 27.82, 26.78, 25.18 ppm. MS (ESI): calcd. for C29H41N6O11S1 [M - H]-
681.3; found 681.2. 1H NMR (400 MHz, CDCl3) of 10: δ = 8.72 (s, 1H), 8.63 (s, 1H), 8.20 (s,
1H), 7.04 (dd, J = 15.1, 4.2 Hz, 1H), 6.42 (d, J = 13.6 Hz, 1H), 6.26 (s, 1H), 5.53 (d, J = 6.2
Hz, 1H), 5.25 (dd, J = 5.9, 3.5 Hz, 1H), 4.93 (s, 1H), 1.63 (s, 3H), 1.56 (s, 9H), 1.43 (s, 9H),
1.41 (s, 3H) ppm. 13C NMR (101 MHz, CDCl3) of 10: δ = 153.05, 150.32, 150.27, 149.96,
149.92, 142.61, 142.24, 128.51, 121.90, 114.96, 90.67, 85.27, 84.02, 83.84, 83.78, 82.18,
28.10, 27.85, 26.95, 25.21 ppm. MS (ESI): calcd. for C24H33N6O9S1 [M - H]- 581.2; found
581.1.
4.2.6. C-(2’,3’-O-isopropylidene-5’-adenosyl)methansulfonamide (13).
Palladium on carbon (10 wt.% loading, 0.015 g, 0.143 mmol) was put under nitrogen. Wittig-
Horner compound 10 (0.086 g, 0.147 mmol, 1.0 eq.) was dissolved in anhydrous methanol
(15 mL, 0.01 M) and added to the catalyst. The atmosphere was exchanged for hydrogen gas
and the reaction mixture was vigorously stirred for 3 days at room temperature. After
completion of the reaction (ESMS), the mixture was filtered over celite. The solvent of the
filtrate was removed under reduced pressure, affording crude reduced compound 12 with Rf =
0.10 (heptane:ethyl acetate 50:50) as a colourless oil. MS (ESI): calcd. for C24H35N6O9S1 [M -
H]- 583.2; found 583.2.
Crude compound 12 was put under nitrogen and was dissolved in anhydrous dichloromethane
(5.66 mL, 0.026 M) and the mixture was cooled to 0 °C. Trifluoroacetic acid (TFA, 0.82 mL,
11.03 mmol, 75 eq.) was added dropwise after which the mixture was stirred for 2 h at 0 °C.
The mixture was then poured in cooled (0 °C) aqueous saturated NaHCO3, whereafter the
aqueous phase was washed with dichloromethane (3x). The aqueous layer was subsequently
extracted with ethyl acetate (3x). The combined organic layers were washed with water and
brine and dried over anhydrous MgSO4. After filtration, the solvent was removed under
reduced pressure. The residue was purified by silica gel chromatography (ethyl
acetate:methanol 100:0 → 90:10) to give the TFA salt of compound 13 (0.062 g, 69%) with
Rf = 0.20 (ethyl acetate) as a colourless oil. 1H NMR (400 MHz, CD3OD): δ = 8.28 (s, 1H),
8.25 (s, 1H), 6.18 (s, 1H), 5.52 (s, 1H), 5.00 (s, 1H), 4.31 (s, 1H), 3.13 (s, 2H), 2.25 (s, 2H),

ACCEPTED MANUSCRIPT
1.61 (s, 3H), 1.40 (s, 3H) ppm. 13C NMR (101 MHz, CD3OD): δ = 155.96, 152.61, 148.89,
140.48, 118.11, 115.21, 89.74, 84.67, 83.80, 83.78, 50.90, 27.66, 26.05, 24.18 ppm.
4.2.7. N-L-isoleucyl-C-5’-adenosyl-methansulfonamide (14).
Cesium carbonate (0.171 g, 0.52 mmol, 6.37 eq.) was put under nitrogen and cooled to 0 °C.
Compound 13 (0.050 g, 0.082 mmol, 1.0 eq.) and t-butyloxycarbonyl-L-isoleucine
hydroxysuccinimide ester (0.086 g, 0.26 mmol, 3.18 eq.) were dissolved in anhydrous DMF
(4.37 mL, 0.019 M) and were added to the base. The mixture was gradually warmed to room
temperature and was stirred for 4 days at room temperature. The solvent was then removed
under reduced pressure and the residue was resuspended in dichloromethane:methanol 90:10.
The suspension was filtered over celite, whereafter the solvents were removed under reduced
pressure, giving the crude isoleucine coupled compound with Rf
(dichloromethane:methanol 90:10) as a yellow solid.
=
0.64
The crude isoleucine coupled compound was put under nitrogen and was cooled to 0 °C. An
8:2 (v/v) mixture of trifluoroacetic acid:water (0.21 mL, 2.81 mmol, 34 eq.:0.05 mL, 2.90
mmol, 35 eq.) was added dropwise. The mixture was stirred for 3 h at room temperature after
which water was added and the mixture was lyophilized, affording a yellow oil.
The crude product was purified by preparative reversed-phase high-performance liquid
chromatography with a PLRP-S 100 Å column connected to a Merck-Hitachi L6200A
intelligent pump at a flow rate of 8.0 mL/min for 40 min. Hereto, following eluent was used:
milli-Q purified water: 5 min. acetonitrile:water 5:95 (v/v); followed by a linear gradient of
acetonitrile in 30 min. from 5% to 95% and re-equilibrating in 5 min. from 95% to 5%. The
retention time of the product was 18.0 min. The appropriate fractions were pooled and
lyophilized to afford the final compound 14 (IleSoHA, 0.003 g, 5%) as a white powder. 1H
NMR (600 MHz, D2O): δ = 8.35 (s, 1H), 8.32 (s, 1H), 6.00 (d, J = 5.1 Hz, 1H), 4.78 (t, J =
5.2 Hz, 1H), 4.26 (t, J = 4.9 Hz, 1H), 4.16 (dd, J = 11.2, 6.4 Hz, 1H), 3.61 (d, J = 4.2 Hz, 1H),
3.31 – 3.25 (m, 2H), 2.14 (dd, J = 15.8, 6.9 Hz, 2H), 1.94 – 1.87 (m, 1H), 1.39 – 1.31 (m,
1H), 1.15 – 1.08 (m, 1H), 0.89 (d, J = 7.0 Hz, 3H), 0.80 (t, J = 7.4 Hz, 3H) ppm. 13C NMR
(151 MHz, D2O): δ = 174.57, 150.68, 148.32, 145.52, 142.66, 119.03, 88.55, 82.81, 73.45,
72.87, 59.85, 48.19, 36.29, 26.92, 23.96, 14.43, 10.86 ppm. MS (ESI): calcd. for
C17H28N7O6S1 [M + H]+ 458.2; found 458.1.
4.2.8. N-L-seryl-C-5’-adenosyl-methansulfonamide (15).
Cesium carbonate (0.116 g, 0.36 mmol, 4.0 eq.) was put under an atmosphere of nitrogen and
was cooled to 0 °C. Compound 13 (0.043 g, 0.089 mmol, 1.0 eq.) and t-butyloxycarbonyl-O-
benzyl-L-serine hydroxysuccinimide ester (0.070 g, 0.18 mmol, 2.0 eq.) were dissolved in
anhydrous DMF (2.93 mL, 0.03 M) and were added to the base. The mixture was gradually
warmed to room temperature and was stirred for 2 days at room temperature. The solvent was
then removed under reduced pressure and the residue was resuspended in
dichloromethane:methanol 90:10. The suspension was filtered over celite, whereafter the
solvents were removed under reduced pressure. The residue was purified by silica gel
chromatography (ethyl acetate:methanol 95:5 → 90:10) to give serine coupled compound
(0.021 g, 31%) with Rf = 0.43 (ethyl acetate:methanol 90:10) as a white solid. HRMS (ESI):
calcd. for C34H47N7O11S1Na1 [M + Na]+ 784.29465; found 784.3362.
Palladium on carbon (10 wt.% loading, 0.003 g, 0.028 mmol) was placed under a nitrogen
blanket, while the serine coupled compound (0.021 g, 0.028 mmol, 1.0 eq.) was dissolved in
anhydrous methanol (2.69 mL, 0.01 M) and added to the catalyst. The atmosphere was
exchanged for hydrogen gas and the reaction mixture was vigorously stirred for 2 days at

ACCEPTED MANUSCRIPT
room temperature. The mixture was then filtered over celite. The filtrate was evaporated
under reduced pressure, giving the crude debenzylated compound.
The crude debenzylated compound was put under nitrogen and a 8:2 (v/v) mixture of
trifluoroacetic acid:water (0.04 mL, 0.59 mmol, 21 eq.:0.01 mL, 0.61 mmol, 22 eq.) was
added. The mixture was then stirred for 3 h at room temperature. Excess water was then
added and the reaction mixture was lyophilized, giving a transparent oil.
The crude product was purified by preparative RP-HPLC as described for 14. The retention
time of the product was 17.6 min. The appropriate fractions were pooled and lyophilized to
afford the desired compound 15 (SerSoHA, 0.002 g, 17%) as a white powder. 1H NMR (600
MHz, D2O): δ = 8.30 (s, 1H), 8.16 (s, 1H), 8.10 (s, 1H), 5.91 (d, J = 5.3 Hz, 1H), 4.16 (t, J =
4.9 Hz, 1H), 4.10 (dt, J = 9.0, 4.6 Hz, 1H), 3.84 – 3.77 (m, 2H), 3.68 (t, J = 4.5 Hz, 1H), 3.23
(pd, J = 14.0, 5.6 Hz, 2H), 2.14 – 1.98 (m, 2H) ppm. 13C NMR (151 MHz, D2O): δ = 170.86,
160.16, 155.46, 152.69, 148.72, 139.81, 118.72, 87.31, 82.36, 73.17, 72.74, 60.68, 56.88,
47.81, 27.08 ppm. HRMS (ESI): calcd. for C14H20N7O7S1 [M - H]- 430.1150; found 430.1144.
4.3. Cloning, expression and purification of E. coli SerRS and IleRS and K. pneumoniae
SerRS
The E. coli SerRS and IleRS were recombinantly expressed and purified using the previously
described procedures [20]. In the case of the Klebsiella pneumoniae SerRS the encoding
sequence was amplified by PCR from a genomic DNA preparation from the strain K.
pneumoniae BAA 1705. The reaction was performed using Phusion polymerase (New England
Biolabs) and the primers 5’-gcgaacagattggtggtATGCTCGATCCCAATCTGCTG-3’ and 5’-
ttgttagcagaagcttaTTAGCCGATATATTCGAGGCCG-3’, where the capitalized letters
correspond to the annealing sequence. The purified PCR product was cloned into the
linearized pETRUK plasmid, an in-house derivative of the SUMO fusion vector pETHSUL
[41] containing a modified sequence for a pI enhanced SUMO tag. The integrity of the
pETRUK-KpSerRS construct was confirmed by sequencing. This construct was transformed
into Rosetta 2 (DE3) pLysS strain (EMD biosciences). A single clone was grown in 2 L of
ZYP-5052 auto-induction medium [42] supplemented with chloramphenicol (10 µg/mL),
ampicillin (100 µg/mL) and 1ml of antifoam SE-15 (Sigma Aldrich). The culture was grown
at 24 °C overnight until it reached an OD600nm of 4. At this point, the temperature was
decreased to 18 °C. After a further 24 h incubation, cells were harvested by centrifugation at
6000 g at 4 °C. The cell pellet was resuspended in approximately 4 times w/v of cation
exchange buffer A (CEXA; 25 mM Hepes pH 8, 200 mM NaCl, 5 mM β-mercaptoethanol)
and then stored at -20 °C. The frozen cell resuspension was quickly thawed and diluted a
further two-fold in CEXA buffer. 100 U cryonase (Takara) and 10 mM (final concentration)
MgCl2 were added to the slurry and the cells were lysed by sonication on ice. The resulting
lysate was clarified by centrifugation at 18,000 g at 4 °C for 30 min and the supernatant was
loaded onto a HiTrap SP HP column (GE Healthcare Life Sciences). The tagged KpSerRS
was eluted from the column using a linear gradient of cation exchange buffer B (CEXB; 25
mM Hepes pH 8, 1000 mM NaCl, 5 mM β-ME). Protein fractions were combined and the
SUMO tag removed by addition of recombinant SUMO hydrolase in a 1:250 (hydrolase :
fusion protein) mass ratio. The cleavage reaction was immediately placed in Spectra/Por
dialysis tubing (standard regenerated cellulose tubing with size 12-14 kDa) and the sample
was dialyzed against 1 L of anion exchange buffer A (AEXA; 20 mM Tris pH 7, 10% w/v
Glycerol and 5 mM DTT) overnight at 4 °C. The dialysis bag was placed in a fresh 1 L of
AEXA buffer for an additional 2 h at 4 °C to remove the remaining salt, then the retentate was
applied to the same HiTrap SP HP column to subtract SUMO tag and initial contaminating
proteins. Further purification was carried out by anion exchange chromatography using a

ACCEPTED MANUSCRIPT
HiTrap Q HP column (GE Healthcare Life Sciences). Fractions corresponding to the
KpSerRS were pooled and the protein precipitated by addition of 2.8 M ammonium sulfate
and 1 mM EDTA. This sample was stored at 4 °C until required.
4.4. Total tRNA purification and aminoacylation assay
The total tRNA pool was extracted and purified from BL21 (DE3) using the previously
described procedure with a minor modification [20]. Following resuspension of the final
isopropanol precipitated tRNA in assay buffer (20 mM Tris pH 7.5, 10 mM MgCl2, 100 mM
KCl, 2.5 mM β-ME) the sample was incubated at 85 °C for 5 min and then allowed to cool
down to room temperature on the bench before performing the aminoacylation assay.
The inhibitory activity of SerSoHA 15 and IleSoHA 14 (and the SA analogs) were assessed
using the previously described radiolabel transfer assay [20]. Briefly, purified 10ꢀnM E. coli
IleRS or 2ꢀnM E. coli SerRS in assay buffer as mentioned before was pre-incubated with the
corresponding compound, at different concentrations, at 37ꢀ°C in the presence of 50ꢀꢁM of the
appropriate 14C-labeled amino acid, 2ꢀmg/mL tRNA and 0.5ꢀmg/mL inorganic
pyrophosphatase. After 10 min, the reaction was initiated by adding pre-warmed ATP at a
final concentration of 500 ꢁM. Upon reaching 6 min, 4 ꢁL quenching buffer (0.2ꢀM sodium
acetate pH 4, 0.1% N-lauroylsarcosine and 5ꢀmM unlabeled amino acid) was added to the
mixture. 20ꢀꢁL was spotted on 3ꢀMM Whatmann paper, precipitated using cold 10% TCA,
washed twice with 10% TCA and once with acetone and dried. Addition of 12 mL
scintillation liquid was followed by measurement of the radioactivity using scintillation
counter.
4.5. Crystallization and structure determination
Prior to crystallization, the ammonium sulphate precipitated KpSerRS was pelleted and
resuspended in size exclusion chromatography (SEC) buffer (10 mM Tris pH 7, 100 mM
NaCl, 5 mM DTT). The sample was loaded onto a Superdex 200 10/300 GL column (GE
Healthcare Life Sciences) equilibrated in the same buffer. Fractions containing the eluted
protein were concentrated using a microspin centrifugal filter device with PES 10 KDa (VWR
international) to 10 mg/ml. Initial crystallization screens was performed using mosquito liquid
transfer robot and sitting drop vapor diffusion method. Single crystals were produced by
hanging drop vapor diffusion mixing 1 ꢁL KpSerRS (10 mg/mL) with 1 ꢁL reservoir solution
(100 mM MES/imidazole buffer system 1 pH 6.5, 50 mM CaCl2, 7.5% w/v PEG 8000 and 20%
v/v ethylene glycol) at 20 °C. These crystals were soaked with 1 mM 5'-O-(N-(L-seryl)-
sulfamoyl)adenosine [20] for 1 h in similar solution contain 10% w/v PEG 8000. The soaked
crystals were transferred to cryoloops and flash frozen in liquid nitrogen. Diffraction data was
collected to on ESRF beamline MASSIF-3 (Grenoble, France). The data were processed using
autoPROC [43]. The initial structure, containing a single protomer in the asymmetric unit,
was solved by molecular replacement with Phaser using Aquifex aeolicus SerRS structure
2DQ3 as a model [44]. Further iterative improvement of the model was performed by manual
building in coot followed by refinement using Buster 2.10.3 [45].
4.6 Computational analysis
As the bound LeuSA in the available structure of Thermus thermophilus LeuRS (PDB ID
2V0C) is split into two chemical moieties (the amino acid and 5’-sulfamoyl modified
nucleoside) the structure was re-refined with a new geometrical restraints file encapsulating

ACCEPTED MANUSCRIPT
the full chemical entity. A suitable crystallographic information file for LeuSA, corresponding
to the ligand ID LSS in the PDB, was generated using the grade webserver
(http://grade.globalphasing.org). Using the deposited model and structure factor data, the full
LeuSA was placed into the calculated electron density map. This model was then refined
using Buster [45] applying a single TLS group to the whole protein chain. The resulting R and
Rfree values of 0.1730 and 0.1927, respectively, were 0.7 % better than those reported for the
original model [46].
Modelling of the O-C of substitution of the various acyl-adenylate analogues was carried out
using the Schrodinger Maestro software. The software was first used place or replace H-atoms
in the structures of K. pneumoniae SerRS, the re-refined T. thermophilus LeuRS-LeuSA
complex, Bacillus subtilis NH3-dependent NAD+ synthetase (PDB ID 1KQP [47];
Mycobacterium tuberculosis biotin-protein ligase (PDB ID 3RUX, [21]), M. tuberculosis
long-chain-fatty-acid-AMP ligase (PDB ID 5HM3 [48]); Pectobacterium chrysanthemi
achromobactin synthetase protein D (PDB ID 2X3J [49]) and the Sulfolobus tokodaii L-
threonylcarbamoyl adenine synthase (PDB ID 4E1B, [11]). The H-bonding was then
optimized before performing a restrained minimization of the protein and ligand employing
the OPLS3e force field [50]. The resultant structure was then used to perform the appropriate
atom substitutions. In the case of the P. chrysanthemi achromobactin synthetase protein D, the
original structure contained ATP and (3’ S)-N-citryl-ethylenediamine [49], therefore the
ethylenediamine atoms were removed and the citrate molecule linked to the α-phosphate,
keeping the resultant pyrophosphate. This structure then went under a second round or
restrained energy minimization prior to atom substitution.
Funding
This work was supported by the Research Fund Flanders [Fonds voor Wetenschappelijk
Onderzoek, 1109117N to D.D.R., G077814N to S.S. and A.V., G0A4616N to S.W. and A.V.,
1S53516N to S. D. G.]; and the KU Leuven Research Fund [3M14022 to S.W. and A.V.]; and
the Chinese Scholarship Council [to L.P.].
Conflicts of interest
No conflicts of interest are declared.
Acknowledgement
The authors thank the Belgian FWO for providing a PhD Fellowship of the Research
Foundation - Flanders to Dries De Ruysscher and a SB Fellowship to Steff De Graef, the
China Scholarship Council for providing a scholarship to Luping Pang and KU Leuven for
financial support. Mass spectrometry was made possible by the support of the Hercules
Foundation of the Flemish Government [20100225E7]. We are also indebted to Luc
Baudemprez for help with NMR spectra measurement.
Supplementary data
Supplementary Data related to this article can be found online.
References

ACCEPTED MANUSCRIPT
[1] S. Schmelz, J.H. Naismith, Adenylate-forming enzymes, Curr. Opin. Struct. Biol. 19
(2009) 666-671.
[2] B.P. Duckworth, K.M. Nelson, C.C. Aldrich, Adenylating Enzymes in Mycobacterium
tuberculosis as Drug Targets, Curr. Top. Med. Chem. 12 (2012) 766-796.
[3] G.H.M. Vondenhoff, A. Van Aerschot, Aminoacyl-tRNA synthetase inhibitors as
potential antibiotics, Eur. J. Med. Chem. 46 (2011) 5227-5236.
[4] J.J. Perona, A. Hadd, Structural Diversity and Protein Engineering of the Aminoacyl-
tRNA Synthetases, Biochemistry 51 (2012) 8705-8729.
[5] G. Eriani, M. Delarue, O. Poch, J. Gangloff, D. Moras, Partition of transfer-RNA
synthetases into 2 classes based on mutually exclusive sets of sequence motifs, Nature
347 (1990) 203-206.
[6] L. Aravind, V. Anantharaman, E.V. Koonin, Monophyly of class I aminoacyl tRNA
synthetase, USPA, ETFP, photolyase, and PP-ATPase nucleotide-binding domains:
Implications for protein evolution in the RNA world, Proteins: Struct., Funct., Genet. 48
(2002) 1-14.
[7] B.H. Dessailly, O.C. Redfern, A.L. Cuff, C.A. Orengo, Detailed Analysis of Function
Divergence in a Large and Diverse Domain Superfamily: Toward a Refined Protocol of
Function Classification, Structure 18 (2010) 1522-1535.
[8] P.A. Reche, Lipoylating and biotinylating enzymes contain a homologous catalytic
module, Protein Sci. 9 (2000) 1922-1929.
[9] A.M. Gulick, Conformational Dynamics in the Acyl-CoA Synthetases, Adenylation
Domains of Non-ribosomal Peptide Synthetases, and Firefly Luciferase, ACS Chem. Biol.
4 (2009) 811-827.
[10] C.S. Carroll, M.M. Moore, Ironing out siderophore biosynthesis: a review of non-
ribosomal peptide synthetase (NRPS)-independent siderophore synthetases, Crit. Rev.
Biochem. Mol. Biol. 53 (2018) 356-381.
[11] C. Parthier, S. Gorlich, F. Jaenecke, C. Breithaupt, U. Brauer, U. Fandrich, D. Clausnitzer,
U.F. Wehmeier, C. Bottcher, D. Scheel, M.T. Stubbs, The O-Carbamoyltransferase TobZ
Catalyzes an Ancient Enzymatic Reaction, Angew. Chem. Int. Ed. 51 (2012) 4046-4052.
[12] C. Shi, D. Tiwari, D.J. Wilson, C.L. Seiler, D. Schnappinger, C.C. Aldrich, Bisubstrate
Inhibitors of Biotin Protein Ligase in Mycobacterium tuberculosis Resistant to
Cyclonucleoside Formation, ACS Med. Chem. Lett. 4 (2013) 1213-1217.
[13] M.R. Bockman, C.A. Engelhart, S. Dawadi, P. Larson, D. Tiwari, D.M. Ferguson, D.
Schnappinger, C.C. Aldrich, Avoiding Antibiotic Inactivation in Mycobacterium
tuberculosis by Rv3406 through Strategic Nucleoside Modification, ACS Infect. Dis. 4
(2018) 1102-1113.
[14] C.H. Qiao, A. Gupte, H.I. Boshoff, D.J. Wilson, E.M. Bennett, R.V. Somu, C.E. Barry,
C.C. Aldrich, 5 '-O- (N-Acyl)sulfamoyl adenosines as antitubercular agents that inhibit
MbtA: An adenylation enzyme required for siderophore biosynthesis of the mycobactins,
J. Med. Chem. 50 (2007) 6080-6094.
[15] J.G. Hurdle, A.J. O'Neill, I. Chopra, Prospects for aminoacyl-tRNA synthetase inhibitors
as new antimicrobial agents, Antimicrob. Agents Chemother. 49 (2005) 4821-4833.
[16] J.S. Tao, P. Schimmel, Inhibitors of aminoacyl-tRNA synthetases as novel anti-infectives,
Expert Opin. Investig. Drugs. 9 (2000) 1767-1775.
[17] J. Lee, S.U. Kang, S.Y. Kim, S.E. Kim, M.K. Kang, Y.J. Jo, S. Kim, Ester and
hydroxamate analogues of methionyl and isoleucyl adenylates as inhibitors of methionyl-
tRNA and isoleucyl-tRNA synthetases, Bioorg. Med. Chem. Lett. 11 (2001) 961-964.
[18] J. Lee, S.E. Kim, J.Y. Lee, S.Y. Kim, S.U. Kang, S.H. Seo, M.W. Chun, T. Kang, S.Y.
Choi, H.O. Kim, N-alkoxysulfamide, N-hydroxysulfamide, and sulfamate analogues of
methionyl and isoleucyl adenylates as inhibitors of methionyl-tRNA and isoleucyl-tRNA

ACCEPTED MANUSCRIPT
synthetases, Bioorg. Med. Chem. Lett. 13 (2003) 1087-1092.
[19] B. Gadakh, S. Smaers, J. Rozenski, M. Froeyen, A. Van Aerschot, 5 '-(N-aminoacyl)-
sulfonamido-5 '-deoxyadenosine: Attempts for a stable alternative for aminoacyl-
sulfamoyl adenosines as aaRS inhibitors, Eur. J. Med. Chem. 93 (2015) 227-236.
[20] B. Zhang, S. De Graef, M. Nautiyal, L.P. Pang, B. Gadakh, M. Froeyen, L. Van Mellaert,
S.V. Strelkov, S.D. Weeks, A. Van Aerschot, Family-wide analysis of aminoacyl-
sulfamoyl-3-deazaadenosine analogues as inhibitors of aminoacyl-tRNA synthetases, Eur.
J. Med. Chem. 148 (2018) 384-396.
[21] B.P. Duckworth, T.W. Geders, D. Tiwari, H.I. Boshoff, P.A. Sibbald, C.E. Barry, D.
Schnappinger, B.C. Finzel, C.C. Aldrich, Bisubstrate Adenylation Inhibitors of Biotin
Protein Ligase from Mycobacterium tuberculosis, Chem. Biol. 18 (2011) 1432-1441.
[22] S. Cusack, C. Berthetcolominas, M. Hartlein, N. Nassar, R. Leberman, A second class of
synthetase structure revealed by X-ray-analysis of escherichia coli seryl-tRNA-synthetase
at 2.5 Å, Nature 347 (1990) 249-255.
[23] M. Fujinaga, C. Berthetcolominas, A.D. Yaremchuk, M.A. Tukalo, S. Cusack, Refined
crystal structure of seryl-tRNA synthetase from thermus thermophilus at 2.5 Å resolution,
J. Mol. Biol. 234 (1993) 222-233.
[24] J.C. Shelley, A. Cholleti, L.L. Frye, J.R. Greenwood, M.R. Timlin, M. Uchimaya, Epik: a
software program for pK (a) prediction and protonation state generation for drug-like
molecules, J. Comput. Aided Mol. Des. 21 (2007) 681-691.
[25] R.V. Somu, H. Boshoff, C.H. Qiao, E.M. Bennett, C.E. Barry, C.C. Aldrich, Rationally
designed nucleoside antibiotics that inhibit siderophore biosynthesis of Mycobacterium
tuberculosis, J. Med. Chem. 49 (2006) 31-34.
[26] A.K. Forrest, R.L. Jarvest, L.M. Mensah, P.J. O'Hanlon, A.J. Pope, R.J. Sheppard,
Aminoalkyl adenylate and aminoacyl sulfamate intermediate analogues differing greatly
in affinity for their cognate Staphylococcus aureus aminoacyl tRNA synthetases, Bioorg.
Med. Chem. Lett. 10 (2000) 1871-1874.
[27] M.J.B. Brown, L.M. Mensah, M.L. Doyle, N.J.P. Broom, N. Osbourne, A.K. Forrest,
C.M. Richardson, P.J. O'Hanlon, A.J. Pope, Rational design of femtomolar inhibitors of
isoleucyl tRNA synthetase from a binding model for pseudomonic acid-A, Biochemistry
39 (2000) 6003-6011.
[28] P. Van de Vijver, G.H.M. Vondenhoff, T.S. Kazakov, E. Semenova, K. Kuznedelov, A.
Metlitskaya, A. Van Aerschot, K. Severinov, Synthetic Microcin C Analogs Targeting
Different Aminoacyl-tRNA Synthetases, J. Bacteriol. 191 (2009) 6273-6280.
[29] G.H.M. Vondenhoff, B. Blanchaert, S. Geboers, T. Kazakov, K.A. Datsenko, B.L.
Wanner, J. Rozenski, K. Severinov, A. Van Aerschot, Characterization of Peptide Chain
Length and Constituency Requirements for YejABEF-Mediated Uptake of Microcin C
Analogues, J. Bacteriol. 193 (2011) 3618-3623.
[30] P.Y. Bruice, Organic chemistry, sixth ed., Pearson, Santa Barbara, 2011.
[31] Q.J. Ma, Y. Akhter, M. Wilmanns, M.T. Ehebauer, Active site conformational changes
upon reaction intermediate biotinyl-5'-AMP binding in biotin protein ligase from
Mycobacterium tuberculosis, Protein Sci. 23 (2014) 932-939.
[32] F.H. Allen, O. Kennard, D.G. Watson, L. Brammer, A.G. Orpen, R. Taylor, Tables of
bond lenghts determined by X-ray and neutron-diffraction. 1. Bond lenghts in organic-
compounds, J. Chem. Soc., Perkin Trans. 2 (1987) S1-S19.
[33] W.P. Jencks, On the attribution and additivity of binding energies, P. Natl. Acad. Sci.
USA 78 (1981) 4046-4050.
[34] E.A. First, Catalysis of the tRNA Aminoacylation Reaction, in: M. Ibba, C. Francklyn, S.
Cusack (Eds.), The Aminoacyl-tRNA Synthetases, Landes Bioscience, Georgetown,
2005, pp. 328–352.

ACCEPTED MANUSCRIPT
[35] J.J. Perona, I. Gruic-Sovulj, Synthetic and Editing Mechanisms of Aminoacyl-tRNA
Synthetases, Aminoacyl-Trna Synthetases in Biology and Medicine, 344 (2014) 1-41.
[36] V. Agarwal, G. Vondenhoff, B. Gadakh, K. Severinov, A. Van Aerschot, S.K. Nair,
Exploring the substrate promiscuity of an antibiotic inactivating enzyme, Medchemcomm
5 (2014) 1567-1570.
[37] K. Devkota, B. Lohse, Q. Liu, M.W. Wang, D. Staerk, J. Berthelsen, R.P. Clausen,
Analogues of the Natural Product Sinefungin as Inhibitors of EHMT1 and EHMT2, ACS
Med. Chem. Lett. 5 (2014) 293-297.
[38] C. Sheu, P. Kang, S. Khan, C.S. Foote, Low-temperature photosensitized oxidation of a
guanosine derivative and formation of an imidazole ring-opened product, J. Am. Chem.
Soc. 124 (2002) 3905-3913.
[39] J. Vannada, E.M. Bennett, D.J. Wilson, H.I. Boshoff, C.E. Barry, C.C. Aldrich, Design,
synthesis, and biological evaluation of beta-ketosulfonamide adenylation inhibitors as
potential antitubercular agents, Org. Lett. 8 (2006) 4707-4710.
[40] D.C. Reuter, J.E. McIntosh, A.C. Guinn, A.M. Madera, Synthesis of vinyl sulfonamides
using the Horner reaction, Synthesis-Stuttgart (2003) 2321-2324.
[41] S.D. Weeks, M. Drinker, P.J. Loll, Ligation independent cloning vectors for expression of
SUMO fusions, Protein Expr. Purif. 53 (2007) 40-50.
[42] F.W. Studier, Protein production by auto-induction in high-density shaking cultures,
Protein Expr. Purif. 41 (2005) 207-234.
[43] C. Vonrhein, C. Flensburg, P. Keller, A. Sharff, O. Smart, W. Paciorek, T. Womack, G.
Bricogne, Data processing and analysis with the autoPROC toolbox, Acta Crystallogr. D.
Biol. Crystallogr. 67 (2011) 293-302.
[44] A.J. McCoy, R.W. Grosse-Kunstleve, P.D. Adams, M.D. Winn, L.C. Storoni, R.J. Read,
Phaser crystallographic software, J. Appl. Crystallogr. 40 (2007) 658-674.
[45] G. Bricogne, E. Blanc, M. Brandl, C. Flensburg, P. Keller, W. Paciorek, P. Roversi, A.
Sharff, O.S. Smart, C. Vonrhein, T.O. Womack, (2017), BUSTER version 2.10.3.
Cambridge, United Kingdom: Global Phasing Ltd.
[46] F.L. Rock, W.M. Mao, A. Yaremchuk, M. Tukalo, T. Crepin, H.C. Zhou, Y.K. Zhang, V.
Hernandez, T. Akama, S.J. Baker, J.J. Plattner, L. Shapiro, S.A. Martinis, S.J. Benkovic,
S. Cusack, M.R.K. Alley, An antifungal agent inhibits an aminoacyl-tRNA synthetase by
trapping tRNA in the editing site, Science 316 (2007) 1759-1761.
[47] J. Symersky, Y. Devedjiev, K. Moore, C. Brouillette, L. DeLucas, NH3-dependent
NAD(+) synthetase from Bacillus subtilis at 1 angstrom resolution, Acta Crystallogr. D
Biol. Crystallogr. 58 (2002) 1138-1146.
[48] M.L. Kuhn, E. Alexander, G. Minasov, H.J. Page, Z. Warwrzak, L. Shuvalova, K.J.
Flores, D.J. Wilson, C. Shi, C.C. Aldrich, W.F. Anderson, Structure of the Essential Mtb
FadD32 Enzyme: A Promising Drug Target for Treating Tuberculosis, ACS Infect. Dis. 2
(2016) 579-591.
[49] S. Schmelz, C.H. Botting, L.J. Song, N.F. Kadi, G.L. Challis, J.H. Naismith, Structural
Basis for Acyl Acceptor Specificity in the Achromobactin Biosynthetic Enzyme AcsD, J.
Mol. Biol. 412 (2011) 495-504.
[50] E. Harder, W. Damm, J. Maple, C.J. Wu, M. Reboul, J.Y. Xiang, L.L. Wang, D. Lupyan,
M.K. Dahlgren, J.L. Knight, J.W. Kaus, D.S. Cerutti, G. Krilov, W.L. Jorgensen, R. Abel,
R.A. Friesner, OPLS3: A Force Field Providing Broad Coverage of Drug-like Small
Molecules and Proteins, J. Chem. Theory Comput. 12 (2016) 281-296.

ACCEPTED MANUSCRIPT
• Improved synthesis strategy of N-acylated sulfonamide adenosines developed
• Two aminoacylated compounds evaluated as aminoacyl-tRNA synthetase (aaRS) inhibitors
• Inhibitory activity depends on the aaRS structural class
• Structural studies suggest an unfavorable pose of the compound in class II aaRS
• Modelling suggests a general targeting potential towards adenylate-forming enzymes