Synthesis, structure and bioactivity of primary sulfamate-containing natural products

Article abstract of DOI:10.1016/j.bmcl.2018.04.038

Here we report the synthesis of natural products (NPs) 5′-O-sulfamoyl adenosine 1 and 5′-O-sulfamoyl-2-chloroadenosine 2. As primary sulfamates these compounds represent an uncommon class of NPs, furthermore there are few NPs known that contain a N–S bond. Compounds 1 and 2 were evaluated for inhibition of carbonic anhydrases (CA), a metalloenzyme family where the primary sulfamate is known to coordinate to the active site zinc and form key hydrogen bonds with adjacent CA active site residues. Both NPs were good to moderate CA inhibitors, with compound 2 a 20–50-fold stronger CA inhibitor (K_i values 65–234 nM) than compound 1. The protein X-ray crystal structures of 1 and 2 in complex with CA II show that it is not the halogen-hydrophobic interactions that give compound 2 a greater binding energy but a slight movement in orientation of the ribose ring that allows better hydrogen bonds to CA residues. Compounds 1 and 2 were further investigated for antimicrobial activity against a panel of microbes relevant to human health, including Gram-negative bacteria (4 strains), Gram-positive bacteria (1 strain) and yeast (2 strains). Antimicrobial activity and selectivity was observed. The minimum inhibitory concentration (MIC) of NP 1 was 10 μM against Gram-positive Staphylococcus aureus and NP 2 was 5 μM against Gram-negative Escherichia coli. This is the first time that NP primary sulfamates have been assessed for inhibition and binding to CAs, with systematic antimicrobial activity studies also reported.

Full text of DOI:10.1016/j.bmcl.2018.04.038

Accepted Manuscript
Synthesis, structure and bioactivity of primary sulfamate-containing natural
products
Prashant Mujumdar, Silvia Bua, Caludiu T. Supuran, Thomas S. Peat, Sally-
Ann Poulsen
PII:
S0960-894X(18)30350-0
https://doi.org/10.1016/j.bmcl.2018.04.038
BMCL 25785
DOI:
Reference:
To appear in:
Bioorganic & Medicinal Chemistry Letters
Received Date:
Revised Date:
Accepted Date:
14 March 2018
11 April 2018
14 April 2018
Please cite this article as: Mujumdar, P., Bua, S., Supuran, C.T., Peat, T.S., Poulsen, S-A., Synthesis, structure and
bioactivity of primary sulfamate-containing natural products, Bioorganic & Medicinal Chemistry Letters (2018),
doi: https://doi.org/10.1016/j.bmcl.2018.04.038
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.

Synthesis, structure and bioactivity of primary sulfamate-containing natural products
a
b
b
c
Prashant Mujumdar , Silvia Bua , Caludiu T. Supuran , Thomas S. Peat , and Sally-Ann
a,*
Poulsen
a
Griffith Institute for Drug Discovery, Griffith University, Don Young Road, Nathan,
Brisbane, Queensland 4111, Australia
b
NEUROFARBA Dept., Sezione di Scienze Farmaceutiche, Università degli Studi di Firenze,
Via Ugo Schiff 6, 50019 Sesto Fiorentino, Florence, Italy
c
CSIRO, 343 Royal Parade, Parkville, Melbourne, Victoria 3052, Australia
*
Corresponding author. Tel.: +61 7 3735 7825; e-mail: s.poulsen@griffith.edu.au

Abstract
Here we report the synthesis of natural products (NPs) 5ꞌ-O-sulfamoyl adenosine 1 and 5ꞌ-O-
sulfamoyl-2-chloroadenosine 2. As primary sulfamates these compounds represent an
uncommon class of NPs, furthermore there are few NPs known that contain a N-S bond.
Compounds 1 and 2 were evaluated for inhibition of carbonic anhydrases (CA), a
metalloenzyme family where the primary sulfamate is known to coordinate to the active site
zinc and form key hydrogen bonds with adjacent CA active site residues. Both NPs were
good to moderate CA inhibitors, with compound 2 a 20 to 50-fold stronger CA inhibitor (Ki
values 65 ‒ 234 nM) than compound 1. The protein X-ray crystal structures of 1 and 2 in
complex with CA II show that it is not the halogen-hydrophobic interactions that give
compound 2 a greater binding energy but a slight movement in orientation of the ribose ring
that allows better hydrogen bonds to CA residues. Compounds 1 and 2 were further
investigated for antimicrobial activity against a panel of microbes relevant to human health,
including Gram-negative bacteria (4 strains), Gram-positive bacteria (1 strain) and yeast (2
strains). Antimicrobial activity and selectivity was observed. The minimum inhibitory
concentration (MIC) of NP 1 was 10 µM against Gram-positive Staphylococcus aureus and
NP 2 was 5 µM against Gram-negative Escherichia coli. This is the first time that NP
primary sulfamates have been assessed for inhibition and binding to CAs, with systematic
antimicrobial activity studies also reported.
Keywords: carbonic anhydrase, inhibitor, natural product, sulfamate, nucleoside,
antimicrobial

Carbonic anhydrases (CA, EC 4.2.1.1) are zinc metalloenzymes that catalyse the reversible
-
+
hydration of carbon dioxide to bicarbonate and a proton: CO + H O  HCO + H . This
2
2
3
equilibrium is critical for human health, and blocking the endogenous chemistry catalysed by
1
CAs is a current target for therapeutic intervention. Most reported small molecule CA
inhibitors incorporate a primary sulfonamide (-SO NH ) or primary sulfamate (-O-SO NH )
2
2
2
2
as a zinc binding functional group that blocks the enzyme activity by coordination to the
2
catalytic zinc. Natural products (NPs) offer extraordinary chemical diversity, and in recent
times have been explored as a source of novel CA inhibitors. Given the prominence of
primary sulfonamides and primary sulfamates as CA inhibitors we conducted a literature
search of the Dictionary of Natural Products (DNP) database against these functional
3
groups. We identified just seven compounds, consistent with the known low representation
4
of N‒S bonds in nature. We recently reviewed the discovery, total synthesis, and known
5
bioactivity of the seven NPs and have reported extensive CA inhibition studies of a
6
representative NP primary sulfonamide, Psammaplin C. Here we sought to further build our
knowledge of the properties of NP primary sulfamates, 5ꞌ-O-sulfamoyl adenosine 1 and 5ꞌ-O-
5
sulfamoyl-2-chloroadenosine 2 (Figure 1). These 5'-sulfamate substituted adenosine
nucleosides have not before been investigated for interaction with CA enzymes so the aim of
this study was to synthesise 1 and 2 then evaluate the enzyme inhibition properties against a
comprehensive panel of human CA isozymes. We were also keen to determine the binding
pose of the NPs in the CA active site using protein X-ray crystallography. Lastly, as there is
sporadic literature available that indicates some antimicrobial potential of 5'-sulfamate
substituted adenosine nucleosides, we were interested to assess the antibacterial and
antifungal activity of these compounds.
Figure 1: Natural product primary sulfamates 5ꞌ-O-sulfamoyl adenosine 1 and 5ꞌ-O-
sulfamoyl-2-chloroadenosine 2.

Compounds 1 and 2 are 5ꞌ-O-sulfamoyl adenosine compounds that differ at the C-2
substituent. Compound 1 is unsubstituted at C-2, while compound 2 has a chlorine at C-2
(Figure 1). To synthesise 1 and 2 the required 2ꞌ,3ꞌ-O-isopropylidene protected adenosine
precursors 5 and 6 were prepared from commercially available adenosine 3 and 2-
chloroadenosine 4 using standard conditions of acetone and tosylic acid (Scheme 1).
Compounds 3 and 4 were treated with chlorosulfonyl amine 7 (prepared from chlorosulfonyl
7
8
isocyanate and formic acid) in the presence of NaH , DMAP or DMA as base to give a
mixture of desired product (sulfamates 8 and 9) and uncharacterised products. We propose
that unreacted formic acid carried forward from the preparation of 7 and/or the strong base
7
,8
promoted cyclonucleoside formation as a side reaction. The successful sulfamoylation of
2
ꞌ,3ꞌ-O-isopropylidene compounds 5 and 6 has been described using 7 prepared from
9
chlorosulfonyl isocyanate and water in acetonitrile as solvent. Compound 7 is formed
instantaneously under these conditions, the reaction mixture is concentrated by azeotroping
with toluene under reduced pressure and the crude 7 used immediately. Sulfamoylation with
crude 7 in the presence of DBU as base gave the corresponding 5ꞌ-O-sulfamoyl adenosine
compounds 8 and 9 in moderate yields without cyclonucleoside byproducts. The 2ꞌ,3ꞌ-O-
isopropylidene protecting group of compounds 8 and 9 was removed using TFA and water
solution (4:2) to afford the target NPs 1 and 2 (Scheme 1).

Scheme 1: Synthesis of NP sulfamates 1 and 2. Reagents and conditions: (i) p-TsOH, acetone,
rt, 4 h; (ii) ClSO NH 7, DBU, DCM, rt, 15 h; (iii) TFA:H O (4:2), 0 °C‒rt, 4 h.
2
2
2
Four different hCA isozymes were selected to investigate the inhibition characteristics of the
NP sulfamates 1 and 2, CA I, II, IV and CA IX, Table 1. CA I is an abundant enzyme and
typically considered an off-target for CA inhibitor applications in clinical use. CA II is also
abundant and is a well-established biological target for anti-glaucoma and anti-edema
1
0
drugs. Membrane-anchored isozyme CA IV has recently been implicated in wound
1
1
healing while CA IX is implicated in the invasion and metastasis of tumours through a pH
1
2
modulating role. The CA isozymes have a similar funnel shaped active site architecture
2
such that isozyme selective small molecule inhibitors are challenging to develop. The NPs
were moderate to good CA inhibitors, with the chlorine substituent at C-2 of 2 leading to a
consistent increase (20 to 50-fold) in CA inhibition across the CA enzyme panel (K values
i
6
5.2 - 234 nM) compared to the unsubstituted analogue 1 (K values 3100 - 6350 nM). This
i
data indicates that the chlorine of 2 may occupy a site in the enzyme cavity that is not
occupied by the corresponding aryl hydrogen of 1.
Table 1: CA inhibition for NPs 1 and 2 at hCA I, II, IV and IX.
a,b
K (nM)
i
Compound
hCA I
6350
167
hCA II
3410
65.2
hCA IV
4450
hCA IX
3100
1
2
234
143
a
b
Errors in the range of ± 5% of the reported value, from three determinations. Obtained
using a stopped flow assay that monitors the physiological reaction (CA catalysed hydration
of CO2).
Having established that NP sulfamates were indeed inhibitors of CA enzymes we were keen
to discern the detailed structural parameters underlying interactions of the sulfamate with the
CA active site. Here we have utilised protein X-ray crystallography to determine the crystal
structures of sulfamates 1 and 2 in complex with CA II. The crystal structures were
determined to 1.28 Å (CA II:1) and 1.50 Å (CA II:2) resolution, respectively (Figure 2 and 3).

There was clear electron density for the sulfamate moiety of compounds 1 and 2 directly
adjacent to the zinc atom in the CA II active site. The two protein-compound structures can
be superposed with an rmsd of 0.11 Å, with the residues within the active site being
essentially identical in position and orientation (His94, His96, His119, Leu198 to Pro202),
Figure 2. The density seen for compound 1 is clearer than the density for compound 2 as
compound 2 may have a second (minor) pose in the active site (and has been refined with an
occupancy of 90%, unlike compound 1 which was refined at full occupancy), Figure 3. The
direct sulfamate-zinc interaction is strong evidence in support of our original hypothesis that
NP sulfamates should inhibit CA activity. In addition to the sulfamate-zinc interaction several
other interactions of NPs 1 and 2 with the CA II residues are observed as the ligand extends
beyond the catalytic pocket. The Gln92 side chain is slightly rotated between the two
structures (Figure 2) and makes a distinctly tighter interaction with compound 2 (the side
chain nitrogen has bond distances of 2.5 and 2.9 Å to the ribose hydroxyl residues whereas
the distances are 3.2 and 3.7 Å to the same atom in the complex with compound 1). This side
chain may also have a minor alternate conformation that puts it in the same orientation as
seen with the compound 1 structure. It seems that the chlorine in the purine ring interacts
sterically with two hydrophobic residues near the top of the funnel-like active site, residues
Phe131 (3.8 Å) and Val135 (3.4 Å), and this interaction may push the ribose ring slightly
over to give the tighter hydrogen bonding interactions to Gln92. The purine ring of
compound 1 sits slightly deeper in the hydrophobic pocket and makes interactions with
Phe131 (3.5-3.7 Å) but is 4 Å or more away from Val135. We suggest that it is not the
halogen-hydrophobic interactions that give compound 2 a greater binding energy but the
slight movement in orientation which allows much better hydrogen bonds between the ribose
ring hydroxyls and the Gln92 side chain.

Figure 2: X-ray crystal structures. A: overview of the CAII structures with 1 and 2
superposed. B: a zoomed in view of panel A with carbons in 1 coloured cyan and carbons in
2
coloured grey. Important residues mentioned in the text that are involved in compound
binding are highlighted as sticks and labelled. C: 5ꞌ-O-sulfamoyl adenosine 1 in complex
with hCA II with distances from the Gln92 side chain to the ribose sugar hydroxyl atoms
shown. D: 5ꞌ-O-sulfamoyl-2-chloroadenosine 2 in complex with hCA II with distances from
the Gln92 side chain to the ribose sugar hydroxyl atoms shown. The small pink sphere
represents the zinc atom in the active site of the hCAII protein in all panels.

Figure 3: Difference electron density shown for the compounds 1 and 2 in the active site of
CAII. The difference density is shown at a 3 level and coloured as green wireframe,
important residues for binding of the compounds are labelled (except Val135 which is
partially blocked from view in this orientation).
Although NPs 1 and 2 have closely related chemical structures, they possess differing but
4
,13,14
15-17
18
broad spectrum antimicrobial
, antitubercular
, and herbicidal
bioactivity.
Furthermore, these compounds are inhibitors of serotonin-induced platelet aggregation and
1
9,20
21
have potential to aid development of antithrombotic therapeutics.
Recently , the gene
cluster that encodes the biosynthesis of 2 was discovered leading to a proposed biosynthesis
for S-N bond formation to incorporate a sulfamate group onto the ribose moiety of adenosine.
As drug resistant bacteria and fungi are prevalent concerns for human health we screened
NPs 1 and 2 against a panel of five bacteria strains (Gram-positive and Gram-negative) and
two fungi strains in collaboration with the Community for Open Antimicrobial Drug
Discovery, CO-ADD. Compounds with activity >50% growth inhibition at 20 µM were
classified as hits, Table 2. NP 1 was a hit against Gram-positive Staphylococcus aureus and
displayed low activity against Klebsiella pneumoniae (~50% growth inhibition at 20 µM).
The chlorinated NP 2 was a hit only against Gram-negative Escherichia coli. Hits were
followed up to determine a minimum inhibitory concentration (MIC) against the microbe,
with 1 displaying a value of 10 µM and 2 a value of 5 µM against S. aureus and E. coli,

respectively, Table 2. A counter-screen to assess mammalian cell cytotoxicity results showed
that 1 and 2 had CC (concentration with 50% cytotoxicity) values 0.8 µM and <0.15 µM,
5
0
respectively, against eukaryotic human embryonic kidney cells (HEK-293), Table 3.
Although there is good selectivity with respect to antibacterial activity the safety window for
mammalian cell cytotoxicity was considered insufficient for follow up hit validation studies
of the NPs.
a
Table 2: Primary screening of antimicrobial activity of NPs 1 and 2 at 20 µM.
Name
Strain
Organsim Type 1 2
Staphylococcus aureus
Escherichia coli
ATCC 43300
ATCC 25922
ATCC 700603
ATCC 19606
ATCC 27853
ATCC 90028
Bacteria
Bacteria
Bacteria
Bacteria
Bacteria
Fungi
G+ve
G-ve
G-ve
G-ve
G-ve
Yeast
Yeast
Klebsiella pneumoniae
Acinetobacter baumannii
Pseudomonas aeruginosa
Candida albicans
Cryptococcus neoformans var. grubii
H99; ATCC 208821 Fungi
Compounds with activity >50% growth inhibition at 20 µM were classified as hits (n=2),
green shading. Non-hits, red shading. G+ve: Gram-positive, G-ve: Gram-negative.
Table 3: Secondary screening of antimicrobial hit activity and cytotoxicity of NPs 1 and 2.
1
2
MIC (µM)
S. aureus ATCC 43300
E. coli ATCC 25922
10
>20
5
>20
CC (µM)
5
0
Human embryonic kidney cells HEK-293 ATCC
0
.8
<0.15
CRL-1573
a
Bacterial strains and cell line sources are detailed in Supplementary Information.

This is the first time NP sulfamates have been evaluated for in vitro inhibition of CA
enzymes for which sulfamates are well known as inhibitors. Herein we have synthesised 5ꞌ-
O-sulfamoyl adenosine 1 and its chlorinated derivative 5ꞌ-O-sulfamoyl-2-chloroadenosine 2,
as representative compounds of the few NP sulfamates that are known. Both 1 and 2 inhibited
CAs at moderate to good levels, with the chlorinated derivative 2 having 20 to 50-fold greater
CA inhibition compared to 1. Using protein X-ray crystallography we confirmed that the
sulfamate group of 1 and 2 coordinate to the CA active site zinc in a manner consistent with
CA:sulfamate structures previously reported. A different orientation of the ribose ring of NP
2
allows shorter hydrogen bond distances to CA residues compared to NP 1. The NPs
displayed good selectivity across a panel of microbes, with 1 a hit (MIC 10 µM) only against
Gram-positive S. aureus and NP 2 a hit (MIC 5 µM) only against Gram-negative E. coli.
Both compounds displayed mammalian cell cytotoxicity, precluding further development of
this compound class for human drug development, although compounds may be useful for
other antimicrobial applications. The NPs have been submitted to the Compounds Australia
2
2
Compound Collection (Academic) (www.compoundsaustralia.com).
Acknowledgements
This research was financed by the Australian Research Council (Grant Number
DP140101495 to S.-A.P. and T.P, Fellowship FT10100185 to S.-A.P.). We are grateful to
Griffith University for PhD Scholarship to P.M. We thank Prof. Ian Jenkins for helpful
discussions. We thank Dr Panagiotis K. Chrysanthopoulos for high resolution mass
spectrometry measurements. We thank CO-ADD (The Community for Antimicrobial Drug
Discovery), funded by the Wellcome Trust (UK) and The University of Queensland
(Australia) for performing antimicrobial screening of compounds. We thank Compounds
Australia staff at Griffith Institute for Drug Discovery for facilitating collaboration with CO-
ADD. We also thank the beamline staff at the Australian Synchrotron for help with data
collection and OpenEye Scientific Software for a license to use AFITT.
Supplementary Data

1
13
Supplementary data associated with this article contains experimental procedures, H, C
NMR and HRMS spectra of compounds, X-ray crystallography statistics and CA inhibition
and activity assay information.
All crystallography data were deposited into the PDB with accession codes 6C7W (1) and
6
C7X (2).
References
1
2
.
.
Neri, D.; Supuran, C. T. Nat. Rev. Drug Discov. 2011, 10, 767-777.
Alterio, V.; Di Fiore, A.; D'Ambrosio, K.; Supuran, C. T.; De Simone, G. Chem. Rev.
2
012, 112, 4421-4468.
3
4
.
.
Dictionary of Natural Products (DVD) version 21.1, 2012.
Waldman, A. J.; Ng, T. L.; Wang, P.; Balskus, E. P. Chem. Rev. 2017, 117, 5784-
5
863.
5
6
.
.
Mujumdar, P.; Poulsen, S.-A. J. Nat. Prod. 2015, 78, 1470-1477.
Mujumdar, P.; Teruya, K.; Tonissen, K.; Vullo, D.; Supuran, C.; Peat, T.; Poulsen, S.-
A. J. Med. Chem. 2016, 59, 5462-5470.
7
.
Lukkarila, J. L.; da Silva, S. R.; Ali, M.; Shahani, V. M.; Xu, G. W.; Berman, J.;
Roughton, A.; Dhe-Paganon, S.; Schimmer, A. D.; Gunning, P. T. ACS Med. Chem.
Lett. 2011, 2, 577-582.
8
9
.
.
Redwan, I. N.; Ljungdahl, T.; Grøtli, M. Tetrahedron 2012, 68, 1507-1514.
Tuck, K. L.; Saldanha, S. A.; Birch, L. M.; Smith, A. G.; Abell, C. Org. Biomol.
Chem. 2006, 4, 3598-3610.
1
1
0.
1.
Supuran, C. T. Nat. Rev. Drug Discov. 2008, 7, 168-181.
Barker, H.; Aaltonen, M.; Pan, P.; Vähätupa, M.; Kaipiainen, P.; May, U.; Prince, S.;
Uusitalo-Järvinen, H.; Waheed, A.; Pastoreková, S.; Sly, W. S.; Parkkila, S.; Järvinen,
T. A. H. Exp. Mol. Med. 2017, 49, e334.
1
2.
Monti, S. M.; Supuran, C. T.; De Simone, G. Expert Opin. Ther. Pat. 2013, 23, 737-
7
49.
1
1
1
3.
4.
5.
Bloch, A.; Coutsogeorgopoulos, C. Biochemistry 1971, 10, 4394-4398.
Osada, H. I., K. Antimicrob. Agents Chemother. 1985, 27, 230-233.
Grimes, K. D.; Aldrich, C. C. Anal. Biochem. 2011, 417, 264-273.

1
1
1
1
6.
Vannada, J.; Bennett, E. M.; Wilson, D. J.; Boshoff, H. I.; Barry, C. E.; Aldrich, C. C.
Org. Lett 2006, 8, 4707-4710.
7. Ferreras, J. A.; Ryu, J.-S. D. L., F.; Tan, D. S.; Quadri, L. E. N. Nat. Chem. Biol. 2005,
, 29-32.
Scacchi, A. B., R.; Cassani, G.; Nielsen, E. . Pest. Biochem. Physiol. 1994, 50, 149-
58.
1
8.
9.
1
Gough, G. R.; Nobbs, D. M.; Middleton, J. C.; Penglis-Caredes, F.; Maguire, M. H. J.
Med. Chem. 1978, 21, 520-525.
2
2
0.
1.
Takahashi, E.; Beppu, T. J. Antibiot. 1982, 35, 939-947.
Zhao, C.; Qi, J.; Tao, W.; He, L.; Xu, W.; Chan, J.; Deng, Z. PLOS ONE 2014, 9,
e114722.
2
2.
Simpson, M.; Poulsen, S.-A. ACS Chem. Biol. 2014, 9, 28-33.

Graphical abstract






Natural products have extraordinary chemical diversity yet few contain a N-S bond
Natural products with a N-S bond include 5'-sulfamate adenosine nucleosides
Carbonic anhydrase inhibition and protein X-ray structures were determined
The compounds display antimicrobial activity
Compounds have been submitted to Compounds Australia
(www.compoundsaustralia.com)