Targeting Mycobacterium tuberculosis Biotin Protein Ligase (MtBPL) with Nucleoside-Based Bisubstrate Adenylation Inhibitors

Article abstract of DOI:10.1021/acs.jmedchem.5b00719

Mycobacterium tuberculosis (Mtb), responsible for both latent and symptomatic tuberculosis (TB), remains the second leading cause of mortality among infectious diseases worldwide. Mycobacterial biotin protein ligase (MtBPL) is an essential enzyme in Mtb and regulates lipid metabolism through the post-translational biotinylation of acyl coenzyme A carboxylases. We report the synthesis and evaluation of a systematic series of potent nucleoside-based inhibitors of MtBPL that contain modifications to the ribofuranosyl ring of the nucleoside. All compounds were characterized by isothermal titration calorimetry (ITC) and shown to bind potently with K_Ds ≤ 2 nM. Additionally, we obtained high-resolution cocrystal structures for a majority of the compounds. Despite fairly uniform biochemical potency, the whole-cell Mtb activity varied greatly with minimum inhibitory concentrations (MIC) ranging from 0.78 to >100 μM. Cellular accumulation studies showed a nearly 10-fold enhancement in accumulation of a C-2′-α analogue over the corresponding C-2′-β analogue, consistent with their differential whole-cell activity.

Full text of DOI:10.1021/acs.jmedchem.5b00719

Subscriber access provided by Nanyang Technological Univ
Article
Targeting Mycobacterium tuberculosis Biotin Protein Ligase
(MtBPL) with Nucleoside-Based Bisubstrate Adenylation Inhibitors
Matthew R. Bockman, Alvin S. Kalinda, Riccardo Petrelli, Teresa De la Mora-
Rey, Divya Tiwari, Feng Liu, Surendra Dawadi, Madhumitha Nandakumar,
Kyu Y. Rhee, Dirk Schnappinger, Barry C. Finzel, and Courtney C. Aldrich
J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.5b00719 • Publication Date (Web): 24 Aug 2015
Downloaded from http://pubs.acs.org on August 26, 2015
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of Medicinal Chemistry is published by the American Chemical Society. 1155
Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 73
Journal of Medicinal Chemistry
1
2
3
4
5
Targeting Mycobacterium tuberculosis Biotin Protein Ligase (MtBPL) with
6
NucleosideꢀBased Bisubstrate Adenylation Inhibitors
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Matthew R. Bockman,a‡ Alvin S. Kalinda,a,b‡ Riccardo Petrelli,b ,d Teresa De la MoraꢀRey,a Divya Tiwari,c Feng
Liu,a Surrendra Dawadi,a Madhumitha Nandakumar,c Kyu Y. Rhee,c Dirk Schnappinger,c Barry C. Finzel,a
Courtney C. Aldrich a,b*
aDepartment of Medicinal Chemistry, University of Minnesota, Minneapolis, MN 55455, USA
bCenter for Drug Design, Academic Health Center, University of Minnesota, MN 55455 USA
cDepartment of Microbiology and Immunology, Weill Cornell Medical College, New York, NY 10021, USA
dPresent Address: School of Pharmacy, Medicinal Chemistry Unit, University of Camerino, Via S. Agostino 1,
62032 Camerino, Italy
‡These authors have contributed equally
Abstract
Mycobacterium tuberculosis (Mtb) responsible for both latent and symptomatic tuberculosis (TB)
remains the second leading cause of mortality among infectious diseases worldwide. Mycobacterial
biotin protein ligase (MtBPL) is an essential enzyme in Mtb and regulates lipid metabolism through the
postꢀtranslational biotinylation of acyl coenzyme A carboxylases. We report the synthesis and
evaluation of a systematic series of potent nucleosideꢀbased inhibitors of MtBPL that contain
modifications to the ribofuranosyl ring of the nucleoside. All compounds were characterized by
isothermal titration calorimetry (ITC) and shown to bind potently with KD's below 2 nM. Additionally,
we obtained highꢀresolution coꢀcrystal structures for a majority of the compounds. Despite fairly
uniform biochemical potency, the wholeꢀcell Mtb activity varied greatly with minimum inhibitory
concentrations (MIC) ranging from 0.78 to >100 µM. Cellular accumulation studies showed a nearly
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 2 of 73
1
2
3
4
10ꢀfold enhanced accumulation of a Cꢀ2′ꢀα analog over the corresponding Cꢀ2′ꢀβ analog, consistent
5
with their differential wholeꢀcell activity.
6
7
8
9
Keywords: Biotin protein ligase, Mycobacterium tuberculosis, tuberculosis, bisubstrate adenylation
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
inhibitors.
Corresponding Author Footnote: To whom correspondence should be addressed. Phone 612ꢀ625ꢀ
7956. Fax 612ꢀ626ꢀ3114. Email: aldri015@umn.edu
ACS Paragon Plus Environment

Page 3 of 73
Journal of Medicinal Chemistry
1
2
3
Introduction
4
5
6
7
8
9
Tuberculosis (TB) is a leading cause of bacterial infectious disease mortality and morbidity. The
bacillus Mycobacterium tuberculosis (Mtb) is responsible for both latent and symptomatic TB and
belongs to the greater Mycobacterium tuberculosis complex, which includes M. bovis, M. africanum, M.
canettii, and M. microti.1 The current treatment for drug sensitive TB requires combination therapy with
the first line drugs isoniazid, rifampin, pyrazinamide and ethambutol for 6–9 months.2 Successful
treatment of TB necessitates (1) stringent supervision by wellꢀtrained care givers, (2) an ample and
consistent supply of high quality medicines and (3) drug intake fidelity by patients to minimize the risks
of recrudescence and drug resistance. These are already tall obstacles to surmount in remote and under
developed regions where TB is endemic. Consequently, the emergence of multidrugꢀ and extensively
drugꢀresistant TB (MDRꢀTB and XDRꢀTB) strains which are minimally resistant to the two most
effective drugs, isoniazid and rifampin, has alarmed clinicians and threatens the positive gains made in
combatting TB.3 Hence, it is imperative to develop new classes of antiꢀtubercular agents for this
worldwide clinical need that are effective against drugꢀresistant TB and ideally possess a novel mode of
action.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
From the onset Mtb has both intrigued and frustrated researchers. Its cellular envelope contains a
panoply of structurally diverse lipids that provide a protective barrier and modulate hostꢀpathogen
interactions.4,5 The fortified waxy cell envelope in turn makes TB a particularly challenging and
intransigent infection to treat. The mycolic acids (1) are the most abundant lipids in Mtb comprising
approximately 30% of its dry cellular weight and also the largest lipids known in nature containing up to
90 carbon atoms (Figure 1).6,7 These essential lipids are covalently attached to the arabinogalactan core
of the cell wall and anchor the outer membrane. Other lipids and lipidated metabolites such as the
phenolic glycolipids (PGLs, 2), phthiocerol dimycocerosate A (PDIMꢀA, 3), sulfolipids (SLs, 4) and
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 4 of 73
1
2
mycobactins (MBTs, 5) are nonꢀcovalently associated with the cell envelope and critical for virulence.8ꢀ
3
4
5
6
10
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 1. Unique lipids and lipidated metabolites found in cell envelope of Mycobacterium tuberculosis.
All of the molecules shown exist as a suite of related isomers that vary in the lipid chain length. If
reported, the major isomer is shown otherwise a representative molecule is depicted. The mycolic acids
are represented by the most abundant αꢀmycolic acid (αꢀMA, 1), the phenolic glycolipids are
represented by 2, the phthiocerol dimycocerosates are exemplified by PDIMꢀA (3), the sulfolipids are
represented by SLꢀ1 (4) and the mycobactins by 5.
All mycobacterial lipids are derived from simple malonyl coenzyme A (CoA) building blocks that
are synthesized by acyl CoA carboxylases (ACCs). Distinct from most bacteria, Mtb possesses multiple
ACCs for preparation of malonyl CoA, methylmalonyl CoA, and longꢀchained malonyl CoAs needed
for synthesis of simple fatty acids, methyl brached lipids, and the mycolic acids, respectively.11ꢀ13 The
multimeric ACCs require postꢀtranslational biotinylation in order to become functionally active, which
is catalyzed by biotin protein ligase (BPL) encoded by birA (Rv3273c) in Mtb.14 The biotin cofactor
mediates carboxy group transfer onto the acyl CoA substrates (Figure 2B). Consequently, MtBPL
universally regulates lipid metabolism in Mtb and therefore represents an attractive biochemical target
for therapeutic development.15
ACS Paragon Plus Environment

Page 5 of 73
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 2. (A) MtBPL catalyzed biotinylation of ACC, a biotin dependent enzyme proceeds in two steps
by sequential adenylation of biotin (6) to form BioꢀAMP (7) followed by acylation of the biotin
carboxylase carrier protein domain of ACCs to furnish holoꢀACC (8). (B) Biotinylated ACC proteins
mediate carboxyl group transfer onto acylꢀCoAs forming malonylꢀCoAs for lipid biosynthesis. The Nꢀ1
atom of biotin is carboxylated in an ATPꢀdependent manner by the biotinꢀcarboxylase domain of ACCs
to furnish 9. In the second halfꢀreaction, ACCs transfer the carboxy group onto an acylꢀCoA, mediated
by the biotin carboxylase domain of ACCs.
MtBPL catalyzes a twoꢀstep reaction initiated when the substrate biotin (6) and ATP bind to form an
acylꢀadenylate intermediate (7, BioꢀAMP) and pyrophosphate (Figure 2A).16 In the succeeding step,
MtBPL binds the ACC and catalyzes the transfer of biotin from 7 onto the terminal amine of a conserved
lysine residue to afford an amide bound biotinylatedꢀACC complex (8). We have reported the design of
a bisubstrate inhibitor of MtBPL termed BioꢀAMS (14) as well as several related analogs including 15,
16, and 17.15, 17 These analogs mimic 7 through substitution of the labile acylꢀphosphate linkage in 7
with stabile acylꢀsulfamide moieties (Figure 3). Compound 14 is exceptionally potent, binding MtBPL
with a KD of approximately 0.5 nM or more than 1,700 times more tightly than biotin, and derives its
potency through interactions with the biotin and adenosine substrate binding pockets.17 The sulfamide is
a near perfect phosphate isostere and is also negatively charged at pH 7 (the estimated pKa of the acylꢀ
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 6 of 73
1
2
3
4
5
6
7
8
9
sulfamide NH group that is Nꢀacylated is ~3), which ensures preservation of the critical electrostatic
interactions with R69, R72, and K138 of MtBPL. Several other related bisubstrate biotinꢀadenosine
inhibitors of BPLs from other organisms have also been described including: alkylphosphate 10,18 βꢀ
ketophosphonate 11,19 1,2,3ꢀtriazole 12,20 and acylsulfamate 13. More recently, Tieu, Abell and
coworkers identified a novel series of Staphylococcus aureus BPL inhibitors containing a heterocyclic
benzoxazoline moiety expemplifid by 18 wherein the adenosine nucleobase has been replaced
altogether.21,22 Furthermore, biotin analogs (19 and 20) with selective S. aureus BPL inhibitory activity
have also been reported.23
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
In addition to its potent biochemical activity, 14 was shown to possess promising antimycobacterial
activity against ten MDRꢀTB and XDRꢀTB strains with minimum inhibitory concentrations (MICs)
ranging from 0.16 to 0.625 ꢁM.15 Proteomic analysis demonstrated 14 inhibited biotinylation in wholeꢀ
cell Mtb while overexpression of MtBPL conferred resistance in Mtb and M. smegmatis. Collectively,
these results provide strong evidence for the designed mechanism of action.15 The structure of 14 and 7
bound to MtBPL reveals a large number of electrostatic interactions with the biotin, adenine, and
acylsulfamate linkage as well as limited space in the active site to further extend the ligand at these
positions.15,24 By contrast, the ribose moiety of the nucleoside has relatively few interactions with the
protein and an adjacent accessible crevice to grow the ligand. As a result, we believe modification of
the ribose offers the greatest opportunity to further increase potency and improve drug disposition
properties. Based on these considerations, we describe herein the design, synthesis and evaluation of a
systematic series of rationally designed bisubstrate inhibitors of MtBPL that probe the consequences of
modifications to the ribofuranosyl ring of 14.
ACS Paragon Plus Environment

Page 7 of 73
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Figure 3. (A) Representative examples of biotinꢀAMP nucleosideꢀbased bisubstrate inhibitors of BPL
including: alkylphosphate 10, βꢀketophosphonate 11, 1,2,3ꢀtriazole 12, acylsulfamate 13, acylsulfamide
14, acylsulfonamide 15, 3ꢀdeaza acylsulfamate 16, and 2ꢀtertꢀbutylethynylꢀacylsulfamate 17. (B)
Benzoxazoline moiety 18, a selective inhibitor of S. aureus BPL. (C) Biotinꢀbased inhibitors 19 and 20
of S. aureus BPL.
RESULTS AND DISCUSSION
Chemistry. The 4′ꢀribofuranosyl oxygen of the lead compound 14 does not appear to have any
interactions with MtBPL.15 To investigate the impact of deleting the 4′ꢀribofuranosyl oxygen we
proposed the synthesis of a trio of analogs that include carbocyclic BioꢀAMS 28, carbocyclic 2′,3′ꢀ
dideoxyꢀBioꢀAMS 36 and carbocyclic 2′,3′ꢀdideoxyꢀ2′,3′ꢀdehydroꢀBioꢀAMS 34. Analogs 28, 34 and 36
are all derived from the common intermediate 23. Scheme 1 illustrates the synthesis of 28. The
aristeromycin derivative 24 was synthesized from Vince lactam (21),25,26 which was first elaborated to
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 8 of 73
1
2
3
4
5
6
7
8
9
the cyclopentenylamine intermediate 22 in three steps following methodology developed by Jung and
Rhee.27,28 The installation of the adenine moiety was accomplished via a palladium catalyzed allylic
amination to afford adenine intermediate 23. The use of [PdCl(C3H5)]2/triisopropylphosphite as the
catalyst and Cs2CO3 as the base was found to be superior to previous reported conditions for related
compounds.28 Pivalate hydrolysis of 23 followed by dihydroxylation and acetonide protection afforded
a mixture of diasteromeric carbocyclic nucleosides 24α and 24β in an approximately 1:1 ratio. The
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1
relative stereochemistry was confirmed by NOESY studies, comparison of H NMR with literature
reported values, and coꢀinjection with an authentic standard of 24α obtained from the natural product
aristeromycin.29ꢀ31 The aristeromycin derivative 24α was then converted into the bisꢀBocꢀprotected
intermediate 25 in two steps, followed by desilylation to afford 26. A Mitsunobu reaction with tertꢀ
butyl sulfamoylcarbamate was employed for direct installation of the sulfamide moiety at Cꢀ5′. It was
discovered that excess triphenylphosphine and DIAD resulted in a significant generation of an
iminophosphorane byproduct of 27 (not shown). This was avoided using three equivalents of tertꢀbutyl
sulfamoylcarbamate (3.0 equiv) and slow addition via syringe pump of a slight excess of DIAD (1.05
equiv). Biotinylation employing
Dꢀ(+)ꢀbiotin Nꢀhydroxysuccinimide ester (BioꢀNHS) mediated by
Cs2CO3 followed by global deprotection with 50% aqueous TFA, reverseꢀphase HPLC purification and
lyophilization afforded the carbocylic BioꢀAMS 28 in 56% over two steps as the triethylammonium salt.
ACS Paragon Plus Environment

Page 9 of 73
Journal of Medicinal Chemistry
1
2
3
Scheme 1. Synthesis of CarbocyclicꢀBioꢀAMS (28).
4
5
6
7
8
[Pd(C3H5)Cl]2, P(OiPr)3
adenine, Cs2CO3
O
O
Piv
S
ref. 27
N
Ph
NH
PivO
DMF, 25 °C
92%
O
21
22
9
NH2
NH2
N
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
1. K2CO3, MeOH, 25 °C, 89%
2. OsO4, NMO, THF-H2O, 0 °C
N
N
N
N
N
N
HO
3. p-TSA•H2O, (CH3O)C(CH3)2
acetone, 25 °C, 53% 2-steps
α:β = 1:1
N
PivO
O
O
23
24α
O
O
N(Boc)2
S
H2N
NHBoc
1. TBSCl, ImH
DMF, 25 °C
N
N
DIAD, PPh3,
N
THF, 0 to 25 °C
83%
2. (Boc)2O, DMAP
NEt3, DMF, 25 °C
92% 2-steps
N
RO
TBAF, THF
25 °C, 93 %
25, R = TBS
26, R = H
O
O
N(Boc)2
1. Biotin-NHS, Cs2CO3
0 to 25 oC, 24 h, DMF
N
N
N
O
O
S
N
2. 50% aq. TFA, 0 to 25 o
C
H2N
N
Boc
52% 2-steps
O
O
O
NH2
HN
NH
N
N
O
N
O
O
27
S
N
N
N
S
H
Et3NH
28
HO
OH
The retention of the C2′ꢀC3′ olefinic bond in analog 34 and saturation of the olefin in cyclopentyl
analog 36, were designed to explore the impact of deletion of the corresponding hydroxyl groups and
examine the subtle conformational implications on binding that result from saturation of the olefin.
Continuing from intermediate 23, hydrolysis of the Cꢀ5′ pivaloyl ester proceeded with facility to afford
29 (Scheme 2). Reprotection of the Cꢀ5′ alcohol as the silyl ether 30 followed by bisꢀBoc protection of
the N6 amine provided 31. Desilylation of the primary alcohol yielded 32 that was utilized in the
succeeding Mitsunobu reaction to give sulfamide intermediate 33. Biotinylation provided the
penultimate crude material, which was followed by global deprotection, reverseꢀphase HPLC
purification, and lyophilization to achieve analog 34 in 52% over two steps. Analog 36 was synthesized
over three synthetic steps from previously synthesized unsaturated intermediate 33. Hydrogenation
using 10 wt% Pd/C in EtOAc supplied the saturated carbocyclic Bocꢀsulfamide intermediate 35.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 10 of 73
1
2
3
4
Biotinylation followed by direct global deprotection of the crude material with 80% aqueous TFA,
5
6
reverseꢀphase HPLC purification and lyophilization furnished final analog 36 in 58% over two steps.
7
8
9
Scheme 2. Synthesis of Cyclopentenyl BioꢀAMS (34) and Cyclopentyl BioꢀAMS (36).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
NH2
NH2
N
N
N
N
N
Et3N, (Boc)2O
N TBSCl, ImH
0 to 25 o
10 h, DMF
83%
C
DMAP, DMF
0 to 25 oC, 16 h
91%
N
N
RO
TBSO
30
23, R = Piv
1M NaOH, MeOH
80 oC, 4 h, 95%
29, R = H
N(Boc)2
N(Boc)2
N
O
O
S
N
N
N
N
TBAF
H2N
NHBoc
0 oC, 0.5 h
94%
N
DIAD, PPh3, THF
0 to 25 oC, 18 h
70%
N
N
TBSO
HO
31
32
N(Boc)2
1. Biotin-NHS, Cs2CO3
DMF, 0 to 25 oC, 10 h
N
N
O
O
S
N
2. 80% aq. TFA,
0 to 25 oC, 4-6 h
57-58% 2-steps
H2N
N
N
Boc
33
35
H2/Pd/C, EtOAc
25 oC, 16 h
O
S
NH2
HN
NH
N
N
O
O
O
S
N
N
N
N
H
X
X
Et3NH
34
X = CH
36 X = CH2
To directly target and manipulate the interactions that exist between the 2′ꢀhydroxyl moiety of our
lead compound 14 and an Asp167 residue in the binding pocket and provide a functional handle to grow
ligands into the adjacent crevice, we envisioned the introduction of alternate functional groups at the Cꢀ
2′ carbon. Analogs 46, 47, 57, 63 and 69 are representative archetypal 2′ꢀdeoxyꢀmolecules that contain
αꢀazido, αꢀamino, βꢀazido, αꢀfluoro and βꢀfluoro substituents, respectively. The synthesis of 2′ꢀαꢀazido
(46) and 2′ꢀαꢀamino (47) analogs is shown in Scheme 3. Vidarabine (37) was protected simultaneously
at the Cꢀ3′ and Cꢀ5′ alcohols using the Markiewicz protecting group yielding cyclic siloxane 38.32
Conversion of the Cꢀ2′ alcohol into the corresponding triflate 39 occurred smoothly. SN2 displacement
ACS Paragon Plus Environment

Page 11 of 73
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
of the triflate by NaN3 in DMF at 65 °C proceeded with inversion of configuration to provide 40.33 bisꢀ
Boc protection of the N6 amine furnished 41.34 The selective monoꢀdeprotection of either the Cꢀ3′ or Cꢀ
5′ alcohol nucleosides protected by the Markiewicz group is extensively described in the literature.35
Utilizing a TFA, H2O and THF (1:1:4) solution it is possible to exclusively deprotect the Cꢀ5′ alcohol
while leaving the Cꢀ3′ alcohol protected.36 We were successful in the aforementioned deprotection only
to discover that any subsequent chemistry on the Cꢀ5′ alcohol was hampered by the steric bulk of the
Markiewicz group remnants still attached to Cꢀ3′. It was decided that the complete removal of the
Markiewicz protecting group with TBAF and 1% AcOH followed by reprotection was the best avenue
to pursue. Thus, removal of the Markiewicz group afforded 42 followed by TBS reprotection of the Cꢀ
3′ and the Cꢀ5′ alcohols using standard protocols achieved 43. Regioselective removal of the Cꢀ5′ TBS
group employing a 3:1:1 AcOH–THF–H2O mixture provided the free Cꢀ5′ alcohol 44.37 Despite the
formation of inseparable side products, formation of sulfamide intermediate 45 via the Mitsunobu
reaction proceeded smoothly. The standard biotinylation reaction described previously yielded the
biotinylated product that was immediately followed by global deprotection in 80% aqueous TFA. The
crude product was purified via reverse phase HPLC in a yield of 62% over two steps to furnish 46.
Further elaboration of the 2′ꢀαꢀazide to the 2′ꢀαꢀamine, an isostere of the C2′ꢀhydroxyl on 14, was
affected by the Staudinger reduction to provide 47.38,39
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 12 of 73
1
2
3
4
5
6
7
8
9
Scheme 3. Synthesis of 2′ꢀαꢀAzidoꢀ2′ꢀdeoxyꢀBioꢀAMS (46), 2′ꢀαꢀAminoꢀ2′ꢀdeoxyꢀBioꢀAMS (47) and
2′ꢀβꢀazidoꢀ2′ꢀdeoxyꢀBioꢀAMS (57).
NH2
N
NH2
N
N
N
N
N
Tf2O
TPDSCl2, pyr
0 to 25 o
N
HO
N
O
O
O
DMAP, CH2Cl2
C
R1
Si
R1
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
0 o
C
O
HO
R2
O
R2
Si
37 R1 = OH, R2 = H
48 R1 = H, R2 = OH
38 R1 = OH, R2 = H, 86%
49 R1 = H, R2 = OH, 89%
NH2
N
NH2
N
N
NaN3, DMF
65 o
Et3N, (Boc)2O
N
C
DMAP, DMF
N
N
N
O
O
N
O
O
0 to 25 o
C
Si
Si
R1
R1
O
O
R2
O
R2
O
Si
Si
39 R1 = OTf, R2 = H, 62%
50 R1 = H, R2 = OTf, 58%
40 R1 = H, R2 = N3, 88%
51 R1 = N3, R2 = H, 98%
N(Boc)2
N(Boc)2
N
N
TBAF, 1% AcOH
N
N
TBSCl, Et3N
DMAP, DMF
N
N
THF, 0 to 25 o
C
N
O
N
HO
0 to 25 o
C
O
O
Si
R1
R1
O
O
R2
HO
R2
Si
41 R1 = H, R2 = N3, 75%
52 R1 = N3, R2 = H, 76%
42 R1 = H, R2 = N3, 79%
53 R1 = N3, R2 = H, 72%
N(Boc)2
N
N(Boc)2
3: 1: 1
AcOH–H2O–THF
N
H2NSO2NHBoc
N
N
0 to 25 o
C
DIAD, PPh3, THF
N
N
N
HO
N
TBSO
O
O
or
R1
R1
H2NSO2Cl, DMA
TBSO
R2
TBSO
R2
43 R1 = H, R2 = N3, 72%
54 R1 = N3, R2 = H, 76%
44 R1 = H, R2 = N3, 63%
55 R1 = N3, R2 = H, 65%
N(Boc)2
1. Biotin-NHS, Cs2CO3
0 to 25 oC, 24 h, DMF
N
N
N
O
O
S
2. 80% aq. TFA,
N
H2N
X
O
0 to 25 o
C
R1
O
R2
NH2
N
TBSO
45 R1 = H, R2 = N3, X = NBoc
HN
NH
N
N
O
O
O
56 R1 = N3, R2 = H, X = O
S
N
N
X
O
S
R1
Et3NH
HO
R2
PPh3, H2O, THF
0 to 24 oC, 6 h, 75%
46 R1 = H, R2 = N3, X = NH, 28% 3-steps
47 R1 = H, R2 = NH2, X = NH
57 R1 = N3, R2 = H, X = O, 56% 3-steps
The only difference between 2′ꢀαꢀazido (46) described above and 2′ꢀβꢀazido (57) is the βꢀfacing
appendage at the Cꢀ2′ carbon. This minor alteration precluded us from utilizing the more sterically
demanding Mitsunobu reaction to install the sulfamide arm. Therefore, we were forced to pursue
sulfamate analogs for our βꢀCꢀ2′ analogs. The sulfamates were installed using the Okada protocol.40
ACS Paragon Plus Environment

Page 13 of 73
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Sulfamoyl chloride is sterically smaller and vastly more reactive than tertꢀbutyl sulfamoylcarbamate.
Towards the synthesis of analog 57 shown in Scheme 3, adenosine (48) was analogously elaborated to
55, which was treated with sulfamoyl chloride in dimethylacetamide (DMA) to afford sulfamate 56.
Biotinylation of the sulfamate intermediate was followed by acidic global deprotection, HPLC
purification and lyophilization to yield 57 in 76% over two steps. Our previous experiences with a
biotinylated acylꢀsulfamate analog had been marred with rapid decomposition via an intraꢀmolecular
displacement reaction to form the biologically inactive 3′,5′ꢀcycloꢀ5′ꢀdeoxyadenosine and Nꢀ
(biotinyl)sulfamic acid.15 Remarkably in the cases of analogs 57 as well as 69 and 73 (vide infra), we
were pleasantly surprised to discover that they resisted any cyclonucleoside formation. Their stability
can be attributed to the βꢀfacing Cꢀ2′ moieties, which preclude the adenine base from adopting the synꢀ
conformation around the glycosidic linkage from which the intramolecular displacement reaction occurs.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Introduction of a fluorine atom at the Cꢀ2′ position of nucleosides is well known to increase
anomeric stability, modulate ring pucker and binding affinity as well as enhance pharmacokinetic
properties.41,42 Scheme 4 illustrates our synthesis of the 2′ꢀdeoxyꢀ2′ꢀαꢀfluoro analog 63 starting with
commercially available 2′ꢀdeoxyꢀ2′ꢀαꢀfluoroꢀadenosine 58. TBS protection of the Cꢀ3′ and Cꢀ5′
alcohols furnished 59. Again, bisꢀBoc protection of the N6 amine cleanly afforded 60 and monoꢀ
desilylation, using our previously described conditions, achieved the desired product 61. Formation of
the sulfamide intermediate 62 using our optimized Mitsunobu procedure gave slightly impure product,
which was used in the next reaction. Biotinylation, followed by global deprotection in 80% aqueous
TFA, furnished the final analog 63 in 56% over the final two steps. We also pursued the synthesis of the
2′ꢀdeoxyꢀ2′ꢀβꢀfluoro biotinylated analog 69 (Scheme 4). Compound 64 was analogously elaborated to
67. Sulfamate formation (68) was followed by biotinylation and then global deprotection with 80%
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 14 of 73
1
2
3
4
aqueous TFA. Reverse phase HPLC purification and lyophilization furnished the final analog 69 in a
5
6
threeꢀstep yield of 51%.
7
8
9
Scheme 4. Synthesis of 2′ꢀDeoxyꢀ2′ꢀαꢀfluoroꢀBioꢀAMS (63) and C2′ꢀdeoxyꢀ2′ꢀβꢀfluoroꢀBioꢀAMS (69).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
NH2
N
NH2
N
N
N
N
N
Et3N, (Boc)2O
DMAP, DMF
TBSCl, ImH,DMAP
DMF, 0 to 25 o
C
N
N
TBSO
O
0 to 25 o
C
HO
O
R1
R1
R2
TBSO
HO
R2
58 R1 = H, R2 = F
64 R1 = F, R2 = H
59 R1 = H, R2 = F, 85%
65 R1 = F, R2 = H, 92%
N(Boc)2
N(Boc)2
3: 1: 1
AcOH–H2O–THF
N
N
N
N
H2NSO2NHBoc
N
N
0 to 25 o
C
DIAD, PPh3, THF
N
N
TBSO
HO
O
O
or
R1
R1
H2NSO2Cl, DMA
R2
R2
TBSO
TBSO
60 R1 = H, R2 = F, 92%
66 R1 = F, R2 = H, 86%
61 R1 = H, R2 = F, 66%
67 R1 = F, R2 = H, 66%
N(Boc)2
1. Biotin-NHS, Cs2CO3
0 to 25 oC, 20 h, DMF then
N
N
N
O
O
2. 80% aq. TFA, 0 to 25 o
C
S
N
H2N
X
O
R1
R2
TBSO
O
NH2
N
62 R1 = H, R2 = F, X = NBoc,
68 R1 = F, R2 = H, X = O
HN
NH
N
N
O
O
O
S
N
N
X
O
S
R1
Et3NH
R2
HO
63 R1 = H, R2 = F, X = NH, 35% 3-steps
69 R1 = F, R2 = H, X = O, 56% 3-steps
Further modifications of the glycoside unit steered us towards the synthesis of analogs 73 in Scheme
5 and 82 in Scheme 6. Addition of the βꢀmethyl to Cꢀ2′ facilitated our investigation into whether the
application of a small lipophilic arm would have any effect on binding. Analog 73 was synthesized over
9ꢀsteps from the tribenzoyl protected ribose 70. Compound 71 was attained over six arduous chemical
steps.43 Sulfamate formation in DMA afforded 72 in good yield.28,30 Subsequent biotinylation was
followed by global deprotection to afford analog 73 in 45% yield over three steps.
ACS Paragon Plus Environment

Page 15 of 73
Journal of Medicinal Chemistry
1
2
3
Scheme 5. Synthesis of Cꢀ2′ꢀβꢀmethylꢀBioꢀAMS (73).
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
An even more extreme alteration can be seen in derivative 82 in which the ribosyl ring has been
altogether removed in favor of a morpholine ring. Analog 82 was designed primarily as a platform for
preparation of derivatives based upon modification of the morpholine nitrogen. The synthetic route to
morpholine analog 82 (Scheme 9) began with commercially available N6ꢀbenzoyl adenosine (74). After
selective TBDPS protection of the Cꢀ5′ hydroxyl to afford 75, reaction with sodium metaꢀperiodate in
MeOH resulted in oxidative cleavage of the Cꢀ2′–Cꢀ3′ bond to reveal the nonisolable dihemiacetal
intermediate 76, which underwent in situ reaction with ammonium biborate to afford dihemiaminal 77.
Reduction with sodium cyanoborohydride in the presence of glacial acetic acid provided crude
morpholine 78.44,45 NꢀTritylation furnished 79 in 43% yield over four synthetic steps.44,46 Desilylation
followed by sulfamoylation of 80 in DME–MeCN afforded sulfamate 81 that was biotinylated using our
standard protocol. The synthesis was completed by ammonolysis to remove the N6ꢀbenzoyl group and
by subsequent treatment with 80% aqueous acetic acid to deprotect the Nꢀtrityl group to give 82 in 36%
yield over three steps from 81.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 16 of 73
1
2
3
4
Scheme 6. Synthesis of MorpholinylꢀBioꢀAMS (82).
5
6
7
8
NHBz
N
NHBz
N
N
N
N
N
N
TBDPSCl, DMAP
N
HO
TBDPSO
O
O
py., 25 °C
98%
9
HO
OH
HO
OH
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
74
75
NHBz
N
N
(NH4)2B2O7 • 4H2O
MeOH, 25 °C
NaIO4
MeOH, 25 °C
N
N
TBDPSO
HO
O
O
OH
76
NHBz
N
NHBz
N
N
N
N
NaBH3CN, AcOH
N
N
N
TBDPSO
HO
TBDPSO
O
O
4 Å sieves,
MeOH, 25 °C
N
H
OH
N
H
77
78
NHBz
N
N
N
N
TBAF
TrtCl, Et3N
DMF, 25 o
43% 4-steps
TBDPSO
O
C
THF, 25 °C
>99%
N
Trt
79
NHBz
NHBz
N
N
O
O
N
N
N
N
O
O
S
S
H2N
Cl
N
N
HO
H2N
O
O
O
DMA:MeCN (1:1)
0 to 25 o
78%
C
N
N
Trt
Trt
80
81
O
NH2
1. Biotin-NHS, Cs2CO3
0 to 25 oC, 16 h, DMF,
then NH3, pyridine, rt, 20 h
HN
NH
N
N
N
O
O
O
S
N
N
O
O
S
2. 80% aq. TFA, 25 oC, 2 h
36% 2-steps
Et3NH
N
H
82
To examine the importance of the 2′ and 3′ꢀhydroxyls for biological activity, we attempted to
prepare 2′ꢀdeoxyꢀ and 3′ꢀdeoxyꢀBioꢀAMS analogs. We initially synthesized 2′ꢀdeoxyꢀBioꢀAMS (not
shown), but discovered the molecule was innately unstable and rapidly decomposed at room
temperature. We attribute this instability to a vastly less pronounced anomeric effect which is derived
from poor orbital overlap between Cꢀ1′ and Cꢀ2′ resulting in a high propensity to depurinate.47
However, 3′ꢀdeoxyꢀBioꢀAMS 87 was successfully synthesized over six steps from commercially
ACS Paragon Plus Environment

Page 17 of 73
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
available 3′ꢀdeoxyꢀadenosine (83, Scheme 7). TBS protection of the Cꢀ2′ and Cꢀ5′ alcohols was
followed by bisꢀBoc protection of N6 furnished 84 in 82% over two steps. Monoꢀdesilylation yielded
the desired product 85. Formation of the sulfamide intermediate using our optimized Mitsunobu
protocol afforded slightly impure product 86, which was used directly in next reaction. Biotinylation
was followed by global deprotection in 80% aqueous TFA to supply the final analog 87 in 62% over the
last three steps.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Scheme 7. Synthesis of 3′ꢀDeoxyꢀBioꢀAMS (87).
NH2
N
N(Boc)2
1. TBSCl, DMAP
ImH, DMF
N
N
N
N
N
N
0 oC to 25 oC, 10h
N
HO
TBSO
O
O
2. Boc2O, DMAP
Et3N, DMF
0 oC to 25 oC, 16 h
82% 2 steps
OH
OTBS
84
83
N(Boc)2
O
O
N
N
N
3: 1: 1
AcOH–H2O–THF
S
H2N
NHBoc
N
HO
O
0 to 25 oC, 7 h
63%
DIAD, PPh, THF
0 oC to 25 oC, 16h
>80%
OTBS
85
N(Boc)2
N
N
N
N
1. Biotin-NHS, Cs2CO3
0 to 25 oC, 24 h, DMF
O
O
S
H2N
N
O
2. 80% aq. TFA,
Boc
0 to 25 o
C
62% 3-steps
OTBS
O
NH2
N
86
HN
NH
N
N
O
O
O
S
N
N
N
O
S
H
HNEt3
OH
87
To complete our narrative we examined the consequence of altogether eliminating the ribose moiety
with acycloꢀBioꢀAMS analog 90 (Scheme 8). Acycloꢀsulfamide analog 90 was prepared from the
reported acycloadenosine (88),48 that was converted to sulfamide intermediate 89, which was
biotinylated and deprotected to provide 90 in 52% over three steps.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 18 of 73
1
2
3
4
Scheme 8. Synthesis of AcycloꢀBioꢀAMS (90).
5
6
7
8
O
O
NH2
N
NH2
N
S
N
N
N
N
H2N
NHBoc
O
O
S
DIAD, PPh3, THF
0 to 25 oC, 18 h
>85%
N
N
H2N
N
HO
O
O
Boc
9
88
89
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
O
NH2
N
1. Biotin-NHS, Cs2CO3
0 to 25 oC, 8 h, DMF
HN
NH
N
O
O
O
S
N
2. 80% aq. TFA,
0 to 25 oC, 4 h
52% 3-steps
N
N
N
O
S
H
Et3NH
90
Ligand characterization by coꢀcrystallization with MtBPL and isothermal titration calorimetry.
The previously described coꢀcrystal structures with 1415 and 724 revealed a binding site with two deep
and largely enclosed subsites for the biotinyl and adenosyl groups, and the sidechain of the Trp74
stacked between them. The ribose lies across the open mouth of this pocket. The 2′ꢀhydroxyl of 14 is
engaged in a hydrogen bond with the carboxylate of Asp167, while the 3′ꢀhydroxyl makes a water
mediated contact with Gly73 (Figure 4A). In order to characterize possible structural consequences of
glycosyl substitutions, coꢀcrystal structures with compounds 36, 46, 57, 63, 69, 73, 82, 87, and 90 have
been determined at high resolution (Diffraction data and summary refinement statistics are provided in
Table S2). The various complexes with bisubstrate inhibitors are remarkably similar. The presence of
an acylsulfamate with the 5′ oxygen in 57, 69, 73, and 82 instead of an acylsulfamide with the 5′
nitrogen in 14, 36, 46, 63, 87, and 90 seems structurally inconsequential; a hydrogen bond between the
deprotonated acylsulfamate/acylsulfamide nitrogen and the amide NH of Arg69 is conserved in all
structures.
A rotameric shift of Asp167 might have been anticipated upon loss of the 2′ꢀhydroxyl, but no change
is observed. A hydrogen bond from OD265 to the amide NH of Asp169 holds the carboxylate in position
in all structures (Figure 4A). No shift occurs even in the case of analogs missing the 2′ꢀhydroxyl (36 or
90). Altered stereochemistry at the 2′ꢀposition is also well tolerated. The 2′ꢀαꢀfluoro and 2′ꢀβꢀfluoro
ACS Paragon Plus Environment

Page 19 of 73
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
compounds (63 and 69, respectively) bind so similarly that all other atoms in the vicinity of the ligands
almost perfectly align (Figure 4B, C). Both azido analogs 46 and 57 also compare similarly. In cases
where a 2′ꢀsubstituent is aliphatic, as in the 2′ꢀβꢀmethyl of 73 or the morpholine ring of 82, the Asp167
side chain shifts just far enough (without changing rotamers) to extend the contact distance between
OD165 and the carbon to the sum of their van der Waals radii (3.4 Å). The similarity of the structures is
reflected in the flat SAR in binding affinity observed by ITC for these compounds.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Analogs 82 and 90 are interesting in that they represent the most diverse departures from the
ribofuranosyl ring, yet are well tolerated. Even the acyclic analog (90) follows a trajectory through Cꢀ1′,
O, and Cꢀ5′ adopted by the ribose in BioꢀAMS (14) (Figure 4D). The morpholine of 82 also
superimposes well upon this triad. The conformation is likely necessary to allow the adenosine and
biotin moieties to retain the same relative orientations upon binding.
Little density is observed for the azido functional group in complexes with 46 and 57. Sufficient
density confirms the stereochemistry of the 2′ substitution, but the remainder of the azido group appears
disordered in rotation about the Cꢀ2′ꢀN bond. In each complex, this group has been modeled in the
conformation that gives rise to at least (–) density in a FoꢀFc difference maps, but any assignment of a
preferred rotamer based on diffraction data is uncertain at best.
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 20 of 73
1
2
3
4
5
D169
D167
A
N158
6
7
8
K138
9
A75
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
W74
G73
R72
R69
N158
N158
B
C
D167
A75
A75
W74
W74
K138
K138
G73
G73
R72
R72
D
D169
D167
N158
K138
Figure 4. Comparison of coꢀcrystal structures. A) BioꢀAMS (14, white) binds with adenosyl and
biotinyl groups in deep pockets with Trp74 stacked between them. Protein carbon atoms of this
complex are green included in other panels for comparison. Hydrogen bonds are illustrated with brown
dashed lines. The ribose is positioned across the open mouth of a larger unoccupied space. B and C)
Binding site and ribose ring pucker are unchanged upon binding of the 2′ꢀfluoro analogs with standard
(63, salmon) or reversed (69, magenta) stereochemistry, despite the loss of the hydrogen bond to Asp167.
Azido analogs (46, 57) bind similarly (not shown). D) Morpholino (82, cyan) and acyclo (90, orange)
analogs adopt conformations that mimic the ribose.
ACS Paragon Plus Environment

Page 21 of 73
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Because of the exceedingly tightꢀbinding nature (KD values ≤ 2 nM) of our family of analogs to
MtBPL we were unable to determine their potency using a steadyꢀstate kinetic assay. Thus we enlisted a
technique that has gained popularity to determine ligandꢀprotein binding affinities. Originally described
in 1965 by Christensen and coworkers, isothermal titration calorimetry (ITC) was first used to
concurrently determine the Keq and ꢀH of various weak acid–base equilibria and metal ion complexation
reactions.49 Today, ITC has become a powerful and commonplace tool to directly characterize the
thermodynamics of ligandꢀprotein interactions. ITC enables determination of the dissociation constant
(KD), Gibbs free energy (ꢀG), enthalpy (ꢀH) and entropy (ꢀS) of ligandꢀprotein interactions allowing
researchers to tease out nuanced information of ligand binding.50 Using 14 as our benchmark we
evaluated a series of analogs for their activity against MtBPL using ITC (Table 1). Unfortunately, the
remarkably tightꢀbinding nature of our analogs exceeded the limit (KD ≤ 10ꢀ8 M) of accurate
determination for binding constants by ITC.51 Therefore, we employed a more reliable approach for
determining the dissociation constants less than 10ꢀ8 M by using displacement ITC, in which our high
affinity ligands are titrated into a solution of protein that is bound to a weaker ligand (in this case,
biotin). Enthalpies (ꢀH) were determined by direct titration for all ligands while dissociation constants
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
(KD) and Gibbs free energies (ꢀG) were measured by displacement ITC experiments (see methods,
Figures S1–S14, and Table S1).
All ligands bound tightly to MtBPL with KD's ranging from 27 pM to 2.14 nM. Binding was
ethalpically driven and the ꢀH's ranged from ꢀ18.8 to ꢀ29.3 kcal/mol (Table 1). Carbocycle 28 that
varies by replacement of the ribofuranosyl ring oxygen with a CH2 moiety was nearly equipotent to 14
with a KD of 0.75 nM. 28 had a 1.2 kcal/mol decrease in ꢀH that was compensated by an equal increase
in the entropic term, resulting in an identical ꢀG of ꢀ12.4 kcal/mol. Remarkably, cyclopentenyl 34,
cyclopentanyl 36 and acyclo 90 all are missing key hydrogen bond donors and acceptors (Cꢀ2′ and Cꢀ3′
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 22 of 73
1
2
3
4
5
6
7
8
9
hydroxyls), yet still bound tightly with KD's of 2.14, 0.54, and 0.315 nM, respectively. These analogs
confirmed our notion that the ribofuranosyl oxygen alone played a limited role in the binding pocket.
The results also suggest that deletion of the Cꢀ2′ and Cꢀ3′ hydroxyls in addition to the removal of the 4′ꢀ
ribofuranosyl oxygen is also well tolerated inferring that the numerous other electrostatic hydrogen
bonding interactions between the ligand and protein are sufficient to maintain affinity. Interestingly,
compound 90 has improved enthalpic contributions, which could be attributed to improved interactions
with MtBPL as a result of its increased flexibility. Cyclopentyl 34 is the least potent of all ligands
evaluated. The presumed constrained Cs envelope conformation of the cyclopentene ring likely does not
allow optimal interactions within the binding pocket. Compounds 46, 47, and 57 all replace the
hydrogen bond donating 2′ꢀhydroxyl with a 2′ꢀamino or 2′ꢀazido group. Both 2′ꢀα substituted
compounds 46 and 47 were nearly equipotent to 14, with 47 being slightly better than 46. Surprisingly,
the 2′ꢀβꢀazido analog (57) proved to be the tightest binder, with a KD of 27 pM, which is 32ꢀfold more
potent than 14. Additionally, 57 bound with the highest enthalpy of ꢀ29.3 kcal/mol, which cannot be
easily reconciled. We speculate the azide may directly interact with MtBPL, however, it is not clear
from the crystallographic data since there is little electron density for the azide. Both compounds 63 and
69 are epimeric with respect to their Cꢀ2′ꢀfluoro substituents and bind with KD's of 0.614 nM and 0.153
nM. Although 63 and 69 are similar in potency to 14, they exhibit unique thermodynamic binding
signatures driven by large enthalpic contributions of ꢀ24.3 kcal/mol and ꢀ28.4 kcal/mol. This could be
attributed to the improved dipoleꢀdipole interactions with MtBPL by the incorporation of the fluorine
atom into the structure. Compounds 73 and 87 both had slight ring modifications, incorporating a 2′ꢀβꢀ
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
methyl into the ring or removal of the 3′ꢀhydroxyl, respectively. Both compounds exhibited similar
binding profiles in terms of KD with 87 displaying slightly tighter binding affinities. The improved
enthalpic contribution of 73 can most likely be attributed to van der Waals interactions from the 2′ꢀβꢀ
ACS Paragon Plus Environment

Page 23 of 73
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
methyl group. Finally, the morpholine 82 analog with the most significant structural change to the ring
structure was well tolerated and bound with a KD of 0.324 nM as well as an identical thermodynamic
binding signature compared to 14. This result indicates MtBPL is quite tolerant to the perturbations of
the glycosyl ring.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
We also investigated how well our analogs stabilize MtBPL through a general binding experiment
using differential scanning fluorimetry (DSF).52 They all had significant stabilization properties as
evident by their change in melting temperature ranging from a change of 12.2 to 19.2 °C (Table 1).
There seems to be limited correlation between the change in melting temperature and binding affinity,
and these data could only be interpreted as a qualitative measure of ligand stabilization to MtBPL.
Antiꢀtubercular Activity and Cytotoxicity. We next determined the minimum inhibitory
concentration that inhibited greater than 90% of cell growth (MIC) of our analogs against Mtb H37Rv.
We previously published the MIC for 14 against H37Rv as 0.78 µM.15 The MIC values of our Bioꢀ
AMS derivatives are as diverse as the molecules themselves (Table 1). The parent 14 still reigns
supreme as the most potent inhibitor as determined by its wholeꢀcell activity. The carbocyclic
molecules 28, 34 and 36 are all deprived of the 4′ꢀribofuranosyl oxygen. Analog 34 incorporates an
olefin while 36 is a saturated cyclopentane. Interestingly, analog 28 does not show strong activity
against Mtb, even at 100 µM, but to our amazement both analogs 34 and 36 displayed MIC values that
were comparable to 14.
Analogs 46 and 57 along with compounds 63 and 69 are respective Cꢀ2′ epimers, but differ in their
substituents. Yet this minor modification in the stereochemistry had significant implications on the antiꢀ
tubercular activity of the molecules. The Cꢀ2′ꢀα substituted compounds 46, 47 and 63 are roughly as
active as 14 against Mtb H37Rv, most likely due to the linearity and directionality of the new
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 24 of 73
1
2
3
4
5
6
7
8
9
appendages which mimic that of the Cꢀ2′ hydroxyl in BioꢀAMS. This can be further attested by
comparing the subtle changes in analogs 46 and 47; 2′ꢀαꢀamino analog 47 is slightly more potent than
2′ꢀαꢀazido analog 46, likely due to amino groups being significantly superior bioisosteres of hydroxyls
than azides. Alternatively, compounds 57 with the 2′ꢀβꢀazido and 69 with the 2′ꢀβꢀfluoro handles are
void of any biological activity. The poor antiꢀtubercular activity of compounds 57 and 69 is mirrored by
analog 73 which also contains a βꢀfacing attachment (CH3) at the Cꢀ2′ position. These results suggest a
steric clash between the βꢀfacing Cꢀ2′ appendages and a putative nucleoside transporting protein
responsible for uptake or association.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Of the current set of BioꢀAMS analogs, the morpholine analog 82 and acyclo analog 90, which
exhibited the greatest structural changes, also exhibited potent antitubercular activity. The increased
flexibility of the acyclo moiety in 90 and 6ꢀmembered ring in 82,53 as opposed to a 5ꢀmembered
ribofuranosyl ring, could allow for perturbations without sacrificing wholeꢀcell activity or binding
affinity. We were also confident that deletion of the Cꢀ3′ hydroxyl would result in a marginal change to
the MIC when juxtaposed to BioꢀAMS. As expected, compound 87 showed an equipotent MIC value of
0.78 ꢁM to that of 14. Moreover, the synthesis of 87 from 3′ꢀdeoxyꢀadenosine was facile in comparison
by circumventing the obligatory Cꢀ3′ and Cꢀ5′ hydroxyl protection with the Markiewicz protecting
group which rendered the syntheses of 46, 47 and 57 cumbersome.
To assess the potential cytotoxicity of our compounds, BioꢀAMS and a subset of the most active
analogs (14, 36, 63, 82, and 90) were evaluated against HepG2 cells using an MTT assay. No inhibition
of growth was observed (IC50 > 100 ꢁM) for any of the tested compounds.
Cellular Uptake of Cꢀ2′ꢀfluoroꢀBioꢀAMS Analogs. The discordant wholeꢀcell MIC data compared
to binding data for Cꢀ2′ epimers 63 and 69 prompted investigation of their cellular accumulation.
ACS Paragon Plus Environment

Page 25 of 73
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
Utilizing the uptake methodology developed by Rhee and coworkers,54,55 the amounts of compounds 63
and 69 associated with the cells were quantified by LCꢀMS/MS. The amount of cellꢀassociated
compound 63 was determined to be approximately 7ꢀfold higher than compound 69 when Mtb H37Rv
was incubated with 25 ꢁM of each analog. This method, however, does not discern between the
amounts of compound that is intracellular versus cellꢀassociated. Nevertheless, there is still a substantial
difference between the amounts of compound 63 and 69 that are accumulated giving credence to the
hypothesis that the disparate MIC and binding data of the Cꢀ2′ fluoro analogs can be potentially
explained by differential cellular accumulation.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
Table 1. MIC90 and thermodynamic data of biotinylated nucleosides.
O
NH2
N
N
HN
NH
N
O
O O
S
N
N
X
S
H
glycosyl
Compound
KD, nMa
ꢁG, kcal/mola
ꢁH, kcal/mola
-TꢁS, kcal/mol
ꢁTm, oC
PDB
MIC90, ꢁMb
14, BioꢀAMS
28
0.78
0.865 ± 0.040
0.753 ± 0.002
ꢀ12.4 ± 0.5
ꢀ12.4 ± 0.0
ꢀ18.9 ± 0.1
ꢀ20.2 ± 0.2
6.6 ± 0.1
7.8 ± 0.2
14.5 ± 1.0
12.2 ± 1.0
3rux
n.d. c
>100
34
36
3.12
1.56
2.14 ± 0.08
ꢀ11.8 ± 0.2
ꢀ12.6 ± 0.2
ꢀ22.1 ± 0.2
ꢀ23.5 ± 0.4
10.3 ± 0.2
10.7 ± 0.4
12.5 ± 0.5
14.5 ± 1.0
n.d.c
4xtv
0.540 ± 0.007
46
3.12
0.933 ± 0.045
ꢀ12.3 ± 0.6
ꢀ23.2 ± 0.2
10.9 ± 0.3
13.0 ± 0.5
4xyw
47
57
63
1.25
>100
1.25
0.627 ± 0.016
0.027 ± 0.003
0.614 ± 0.018
ꢀ12.6 ± 1.0
ꢀ14.4 ± 0.2
ꢀ12.6 ± 0.3
ꢀ19.1 ± 0.1
ꢀ29.3 ± 0.1
ꢀ24.3 ± 0.5
6.5 ± 0.2
14.9 ± 0.1
11.8 ± 0.5
19.2 ± 0.6
17.0 ± 1.0
15.2 ± 0.8
n.d. c
4xtx
4xty
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 26 of 73
1
2
3
4
5
6
7
8
9
69
73
50.0
50.0
0.153 ± 0.000
0.769 ± 0.006
ꢀ13.4 ± 0.0
ꢀ12.4 ± 0.4
ꢀ28.4 ± 0.6
ꢀ20.3 ± 0.4
14.7 ± 0.5
8.1 ± 0.5
17.7 ± 0.8
14.2 ± 0.3
4xtz
4xu0
4xu1
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
82
3.12
0.324 ± 0.023
ꢀ12.9 ± 0.9
ꢀ18.8 ± 0.0
5.9 ± 0.1
13.8 ± 0.8
87
90
0.78
3.12
0.592 ± 0.009
0.315 ± 0.003
ꢀ12.6 ± 0.1
ꢀ13.0 ± 0.1
ꢀ21.4 ± 0.5
ꢀ25.5 ± 0.7
8.8 ± 0.6
15.0 ± 1.0
14.7 ± 0.3
4xu2
4xu3
12.6 ± 0.7
aCompetitive ITC experiments to determine KD and ꢀG were performed at least in triplicate with 2 different enzyme batches while direct
titration experiments were done in triplicate. All analogs showed n values of 1 ± 0.2. Minimum inhibitory concentrations (MIC) that
resulted in >90% growth inhibition of M. tuberculosis H37Rv were determined by a broth microdilution assay in GAS medium.
Experiments were performed twice independently in triplicate. The MIC is defined as the lowest concentration of inhibitors that prevented
growth, as determined by measuring the endpoint OD600 values. cNot determined
b
CONCLUSION
We have successfully designed and synthesized a focused series of MtBPL bisubstrate inhibitors by
strategically modifying the ribofuranosyl ring. All analogs bound tightly to MtBPL, with KD values
ranging from 27 pM to 2.14 nM. Aside from the weakest binding 34 and strongest binding 57, all
compounds exhibited a narrow range of potency from 0.153 to 0.933 nM. ITC also illustrated the
subtleties in the thermodynamic binding signatures of each analog that cannot be easily reconciled.
These data, along with Xꢀray coꢀcrystal structures, demonstrate modification or substitution of the
ribofuranosyl ring does not substantially impact the binding of the analogs to MtBPL. However, the
relatively flat biochemical potencies sharply contrast with wholeꢀcell antitubercular activity. Notably,
all of the analogs containing a Cꢀ2′ꢀβ substituent including 57, 69 and 73 had substantially reduced
wholeꢀcell activities with MIC ranging from 50 to >100 µM despite potent binding affinities. This can
only be attributed to differences in cellular accumulation through enhanced efflux or reduced uptake,
and when examining the differences in cellular accumulation between 63 and 69, there was substantial
ACS Paragon Plus Environment

Page 27 of 73
Journal of Medicinal Chemistry
1
2
3
4
5
6
7
8
9
preference for cellular association of compound 63 over 69. The other analogs in the series, except 28,
possessed wholeꢀcell activity commensurate with 14 with MIC values ranging from 0.78–3.12 µM.
Both 47 and 82 represent promising ligands for further optimization since they can be readily
functionalized through ‘click chemistry’ or substitution of the endocyclic nitrogen, respectively.
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment

Journal of Medicinal Chemistry
Page 28 of 73
1
2
3
4
EXPERIMENTALS
5
6
7
8
9
General materials and methods. Chemicals and solvents were purchased from Acros Organics,
Alfa Aesar, SigmaꢀAldrich, and TCI America and were used as received. The nucleosides 9ꢀβꢀ
arabinofuranosyladenine 37 and 3ꢀdeoxyadenosine 83 were obtained from Berry & Associates (Dexter,
D
ꢀ
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
MI) while 2′ꢀdeoxyꢀ2′ꢀfluoroadenosine 58 and 9ꢀ(2′ꢀdeoxyꢀ2′ꢀfluoroꢀβꢀ
were obtained from Metkinen Chemistry (Kuopio, Finland), and N6ꢀbenzoyladenosine 74 was obtained
from Carbosynth (Berkshire, UK). tertꢀButyl sulfamoylcarbamate,56 sulfamoyl chloride,57
ester,58
Dꢀarabinofuranosyl)adenine 64
D
ꢀ(+)ꢀbiotin
Nꢀhydroxysuccinimide
5′ꢀ[Nꢀ( ꢀbiotinoyl)sulfamoyl]aminoꢀ5′ꢀdeoxyadenosine
D
triethylammonium salt (BioꢀAMS) 14,15 {(1S,4R)ꢀ4ꢀ[Nꢀ(phenylsulfonyl)pivalamido]cycloꢀpentꢀ2ꢀ
enyl}methyl pivalate 22,28 2′,3′ꢀOꢀisoproylideneꢀβꢀ2'ꢀmethyladenosine 71,59 and 9ꢀ[(2ꢀ
hydroxyethoxy)methyl]adenine 88,48 were prepared as described. An anhydrous solvent dispensing
system using two packed columns of neutral alumina was used for drying THF and CH2Cl2, while two
packed columns of molecular sieves were used to dry DMF, and the solvents were dispensed under
argon gas (Ar). Anhydrous grade MeOH, MeCN, pyridine, and DMA were purchased from Aldrich.
EtOAc and hexanes were purchased from Fisher Scientific. All reactions were performed under an inert
atmosphere of dry argon gas (Ar) in ovenꢀdried (180 °C) glassware. TLC analyses were performed on
TLC silica gel plates 60F254 from EMD Chemical Inc. and were visualized with UV light. Optical
rotations values were obtained on a polarimeter using a 1 dm cell. Purification by flash chromatography
was performed using a mediumꢀpressure flash chromatography system equipped with flash column
silica cartridges with the indicated solvent system. Preparative reversedꢀphase HPLC purification was
performed on a Phenomenex Gemini 10 ꢁm C18 250 × 20 mm column operating at 30.0 mL/min with
detection at 254 nm with the indicated solvent system. Analytical reversedꢀphase HPLC purification
was performed on a Phenomenex Gemini 5 ꢁm C18 250 × 4.6 mm column operating at 1 mL/min with
ACS Paragon Plus Environment

Page 29 of 73
Journal of Medicinal Chemistry
1
2
3
detection at 254 nm employing a linear gradient from 5% to 50% MeCN in 50 mM aqueous
4
5
6
13
triethylammonium bicarbonate (TEAB) at pH 7.5 for 30 min (Method A). 1H and C spectra were
7
8
9
recorded on 400, 600 or 900 MHz NMR spectrometers. Proton chemical shifts are reported in ppm from
an internal standard of residual chloroform (7.26), methanol (3.31), dimethyl sulfoxide (2.50), or monoꢀ
deuterated water (HDO, 4.79); carbon chemical shifts are reported in ppm from an internal standard of
residual chloroform (77.0), methanol (49.1), or dimethyl sulfoxide (39.5). Proton chemical data are
reported as follows: chemical shift, multiplicity (s = singlet, d = doublet, dt = doublet of triplets, t =
triplet, q = quartet, pentet = pent, m = multiplet, ap = apparent, br = broad, ovlp = overlapping), coupling
constant(s), integration. Highꢀresolution mass spectra were obtained on an LTQ Orbitrap Velos
(Thermo Scientific, Waltham, MA). All compounds were determined to be >95 % by analytical reverseꢀ
phase HPLC (purities for each final compound are given in the experimental section below).
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
General Procedure for Mitsunobu Reaction. To a solution of nucleoside (1.0 equiv), tertꢀbutyl
sulfamoylcarbamate (3.0 equiv) and PPh3 (1.5 equiv) in THF (0.05 M for the limiting reagent) at 0 °C
was added DIAD (1.05 equiv) dropwise over 15 min. The solution was stirred at 0 °C for 15 min, then
gradually warmed to 23 °C over 3 h and stirred for 12 h at 23 °C. The reaction was concentrated in
vacuo to afford the crude sulfamide, which was purified by flash chromatography using the indicated
solvent system. The product was typically contaminated with triphenylphosphine oxide (TPPO). In this
case the calculated yield was obtained by a simple linear equation relating the total isolated weight to the
ratio of product:TPPO and formula weights for each compound.
General Procedure for Okada Sulfamoylation Protocol. To a solution of nucleoside (1.0 equiv)
in DMA (0.4 M) was added freshly prepared sulfamoyl chloride (4.0 equiv) at 0 °C. The reaction was
slowly warmed to 23 °C over 6 h and then stirred an additional 18 h at 23 °C. The reaction was diluted
with EtOAc (200 mL/mmol) and washed successively with H2O (5 × 100 mL/mmol), saturated aqueous
ACS Paragon Plus Environment