Structure-Guided Drug Design of 6-Substituted Adenosine Analogues as Potent Inhibitors of Mycobacterium tuberculosis Adenosine Kinase

This is an open access article published under a Creative Commons Attribution (CC-BY)

License, which permits unrestricted use, distribution and reproduction in any medium,

provided the author and source are cited.

Article

pubs.acs.org/jmc

Cite This: J. Med. Chem. XXXX, XXX, XXX −XXX

Structure-Guided Drug Design of 6‑Substituted Adenosine

Analogues as Potent Inhibitors of Mycobacterium tuberculosis

Adenosine Kinase

†

,¶

§,#

†

‡

†

Roberto A. Crespo, Qun Dang,_† Nian E. Zhou, Liam M. Guthrie, Thomas C. Snavely,

†

∥

⊥

Wen Dong, Kimberly A. Loesch, Takao Suzuki, Lanying You, Wei Wang, Theresa O’Malley,

Tanya Parish,^⊥David B. Olsen,* and James C. Sacchettini*

,§

,†

†

Department of Biochemistry and Biophysics, Texas A&M University, College Station, Texas 77843, United States

College of Medicine, Texas A&M University Health Science Center, Bryan, Texas 77807, United States

Merck Sharp Dohme Corporation, West Point Pennsylvania 19486, United States

‡

§

∥

WuXi AppTec, 288 Fute Zhong Road, Shanghai 200131, China

⊥

TB Discovery Research, Infectious Disease Research Institute, 1616 Eastlake Avenue E, Seattle, Washington 98102, United States

*

S Supporting Information

ABSTRACT: Mycobacterium tuberculosis adenosine kinase (MtbAdoK) is an essential enzyme of Mtb and forms part of the

purine salvage pathway within mycobacteria. Evidence suggests that the purine salvage pathway might play a crucial role in Mtb

survival and persistence during its latent phase of infection. In these studies, we adopted a structural approach to the discovery,

structure-guided design, and synthesis of a series of adenosine analogues that displayed inhibition constants ranging from 5 to

1

20 nM against the enzyme. Two of these compounds exhibited low micromolar activity against Mtb with half maximal eﬀective

inhibitory concentrations of 1.7 and 4.0 μM. Our selectivity and preliminary pharmacokinetic studies showed that the

compounds possess a higher degree of speciﬁcity against MtbAdoK when compared with the human counterpart and are well

tolerated in rodents, respectively. Finally, crystallographic studies showed the molecular basis of inhibition, potency, and

selectivity and revealed the presence of a potentially therapeutically relevant cavity unique to the MtbAdoK homodimer.

INTRODUCTION

corresponding purine nucleotides by the purine salvage

enzymes. Although the de novo and purine salvage pathways

have not been extensively studied in Mtb, it is known that Mtb

possesses all of the enzymes required for both pathways. It is

currently unknown if there is restrictive regulation of the two

pathways. However, by switching to the salvage pathway Mtb

■

Mycobacterium tuberculosis (Mtb), the bacterium that causes

pulmonary tuberculosis (TB), represents one of the major

leading causes of death worldwide by a single infectious agent.

One-third of the world’s population is thought to harbor latent

TB, and around 5−10% of these infected individuals are

expected to develop the active disease sometime during their

2

can bypass several chemically demanding steps. This has led

1

to the hypothesis that the salvage pathway might be the most

likely source of nucleotides within the hostile and nutrient-

deprived microenvironment encountered by Mtb during its

lifetime. The rapid emergence of multidrug resistant and

extensively drug-resistant (XDR) TB demands the develop-

ment of novel chemotherapeutic agents with novel molecular

targets.

2

,3

latent phase of infection.

The purine salvage pathway is a druggable pathway within

mycobacteria. In this pathway, preformed nucleobases from

the product of nucleic acid breakdown are converted to their

Received: January 4, 2019

©

XXXX American Chemical Society

A

DOI: 10.1021/acs.jmedchem.9b00020

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Figure 1. Positions of interest for structure-guided drug design and tool compounds utilized in crystallography studies. (a) Adenosine bound to the

active site of MtbAdoK (PDB ID 2PKM). (b) Tool compounds used to explore the chemical space surrounding the positions highlighted in (a).

M. tuberculosis adenosine kinase (MtbAdoK) performs a

critical step in the purine salvage pathway within mycobacteria.

The enzyme catalyzes the conversion of adenosine to

adenosine monophosphate in a Mg²and adenosine 5′-

kinases that diﬀers from its eukaryotic counterparts by its

4,12,13

unique quaternary structure and regulatory mechanisms.

Furthermore, our preliminary eﬀorts to ﬁnd novel drugs and

drug targets identiﬁed MtbAdoK as the target of the nucleoside

analogue iodotubercidin. Mtb H37Rv-resistant mutants raised

against iodotubercidin identiﬁed three nonsynonymous muta-

tions, that is, D251N, Q172P, and G262S; all of which

displayed minimum inhibitory concentrations (MICs) > 100

μM when compared with WT (18.0 ± 10.0 μM). The

essentiality, diﬀerences between hAdoK and MtbAdoK and on-

target activity of the nucleoside analogue described above

make MtbAdoK an attractive and novel drug target.

+

4

triphosphate (ATP)-dependent manner. The crystal structure

of MtbAdoK has been previously solved at high resolution with

the substrate (adenosine), substrate analogue 2-ﬂuoroadeno-

sine, ATP analogue AMP−PCP and without substrate (apo) at

5

resolutions of 1.90, 1.93, 1.90, and 1.50 Å, respectively.

The crystallographic data showed that the apo and the

AMP−PCP structures adopted the opened conformation of

the protein where the active site is solvent exposed. In contrast,

the substrate or substrate analogue complexes revealed that the

lid domain of MtbAdoK undergoes a 30° movement upon

To date, most reported inhibitors of MtbAdoK have been

designed as substrate surrogates to elicit the production of

14−18

6

toxic metabolites.

For example, Parker et al. showed that

substrate binding. This conformational change eﬀectively

the H37Ra Mtb strain was able to uptake the adenosine

analogue 2-methyl-adenosine and that MtbAdoK was respon-

sible for its activation into the toxic metabolite 2-methyl-

brings lid domain residues Asp12, Phe116, and Phe102 in close

contacts with adenosine, thereby completing the active site

5

,7−9

(Figure S1a ).

As previously described, stabilization of the

16

AMP. Despite having some advantages, AMP or ATP toxic

metabolites could inadvertently lead to undesirable and cross-

species oﬀ-target eﬀects exacerbating the identiﬁcation and

diminishing the value of a potential drug target.

ribose and adenine rings are primarily mediated by π-stacking

interactions with residues Phe102 and Phe116, respectively.

Hydrogen bonding networks with the adenine moiety were

described to occur with residues Gln172, Gln173, Ser8, and

Ser36′ while the ribose forms extensive hydrogen bonding

interactions with residues Gln172, Asp12, Gly48, Asn52, and

In this work, we adopted a structure-guided approach to the

design of very potent and safe adenosine analogues. To gain

insights into the chemical space surrounding the active site, we

solved the crystal structures of MtbAdoK complexed to several

adenosine analogues at resolutions between 1.70 and 2.23 Å.

This initial approach laid the foundation for the structure-

guided design and synthesis of several very potent 6-

substituted adenosine analogues as selective inhibitors of

MtbAdoK. Several of the synthesized compounds displayed

low micromolar anti-Mtb activity in a whole-cell assay, which

supports the concept that inhibition of MtbAdoK can be used

as an approach to treat TB. Finally, these compounds exhibited

a higher degree of speciﬁcity against MtbAdoK when

compared with the human counterpart, and a 6-substituted

adenosine analogue from our series was well tolerated in Swiss

Webster mice at high levels of exposure while demonstrating

adequate pharmacokinetic properties.

5

catalytic base Asp257 (Figure S1b).

MtbAdoK is signiﬁcantly diﬀerent when compared with

8

eukaryotic adenosine kinases. Although the human adenosine

kinase (hAdoK) and MtbAdoK are composed of a small-lid

like domain and large domain, hAdoK shares less than 20%

sequence identity with MtbAdoK. The hAdoK structure has

been previously solved at 1.4 Å with substrate bound. The

structure was initially solved with two molecules of adenosine

embedded within the enzyme, one in the active site and

8

another in cofactor (ATP) site. A critical diﬀerence between

MtbAdoK and its eukaryotic counterparts is that MtbAdoK is a

5

,8

functional homodimer. In addition, prior to its discovery and

characterization by Long et al. in 2003, the adoK gene was

thought to be unique to eukaryotic organisms and was

4

annotated as a general carbohydrate kinase (cbhK) in Mtb.

On the basis of its amino acid sequence, MtbAdoK is more

closely related to members of the phosphofructokinase B

family of sugars kinases, which includes ribokinase, a

RESULTS AND DISCUSSION

■

Structural analysis of the MtbAdoK−adenosine (1) complex

6

homodimer and the prototypical member of this family.

showed a high potential for the structure-guided drug design

5

MtbAdoK is essential for the survival of the bacilli when

utilizing cholesterol as a carbon source and in infected mouse

(Figure 1a). Of particular interest are positions N6 and N7 of

the adenine ring and position O5′ of the ribose moiety; all of

1

0,11

models.

Even though the precise role of the purine salvage

which have been the focal point of several chemistry-based

1

2−15

pathway under these conditions is currently unknown,

biochemical and structural characterization suggests that

MtbAdoK might represent a new class of bacterial adenosine

structural−activity relationship (SAR) studies.

studies have highlighted that substitutions at the aforemen-

tioned positions lead to inhibitory activity against MtbAdoK

These

B

DOI: 10.1021/acs.jmedchem.9b00020

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Figure 2. Crystal structure complexes of compounds 2−5. (a) 2 bound to the active site of MtbAdoK. (b) 3 bound to the active site of MtbAdoK.

c,d) N7-substituents of compounds 2−3 are buried in an active site pocket composed of residues from chain A and chain B. (e) Chimney-like

(

cavity observed above position N6 formed by residues from chain A and chain B. (f) 4 bound to the active site of MtbAdoK. (g) Methylmercapto

group of 4 is accommodated in a compound-induced pocket formed by the movement of Arg176 (red). (h) Chimney-like cavity observed in the

cocrystal structures of 2 and 3 is blocked by the conformational change of Arg176 (red). (i) 5 bound to the active site of MtbAdoK. In all cases,

chain A residues are colored by heteroatom and chain B residues are color magenta.

and hAdoK. The chemistry-based eﬀorts have also suggested

that there is a higher degree of selectivity to MtbAdoK when

substitutions are located at the N6-position when compared

with the N7-analogues. However, the structural and molecular

basis for the reported selectivity has remained unexplained.

The 5′-position oﬀers a more practical approach to inhibition

given that the critical hydroxyl group is required for catalysis.

To guide our drug discovery eﬀorts and to investigate the

molecular basis of selectivity and inhibition, we solved the

crystal structure complexes of several known AdoK inhibitors.

Iodotubercidin (2) and sangivamycin (3) were selected to

investigate the conformational ﬂexibility and chemical proper-

ties surrounding the binding pocket near the N7-position of

the adenosine scaﬀold. Additionally, 6-methylmercaptopurine

riboside (4) and 5′-aminoadenosine (5) cocrystal structures

were solved to explore the 6- and 5′-positions of the adenosine

scaﬀold, respectively (Figure 1b).

inhibitory activity against MtbAdoK, hAdoK, MAP kinases,

and Ser/Thr kinases while compound 3 has been previously

observed to be an inhibitor of protein kinase C, MtbAdoK, and

19−21

hAdoK.

Cocrystallization of MtbAdoK with compounds 2 and 3 was

produced via vapor diﬀusion, and the structures were solved

using the molecular replacement (MR) method utilizing the

5

previously reported adenosine-bound structure. The crystals

of both complexes were determined to be in the P4 crystal

1

space group with two molecules in the asymmetric unit (ASU).

The MtbAdoK-2 complex was reﬁned to 2.10 Å resolution and

displayed a closed conformation of the enzyme with a

backbone rmsd of 1.14 Å amongst all Cα atoms when

superimposed to the MtbAdoK−adenosine structure (Figure

2a, Tables S1 and S2 ). The MtbAdoK-3 complex reﬁned to

2.10 Å resolution displayed a closed conformation of the

enzyme with a backbone rmsd of 1.15 Å amongst all Cα atoms

when superimposed to the MtbAdoK−adenosine structure

(Figure 2b, Tables S3 and S4 ). As in the case of the hAdoK−

adenosine structure, both crystal structures showed positive

electron density for the inhibitor bound to the active site in a

similar position to adenosine and electron density in the

cofactor (ATP) binding site (Figure S2a,b ). However, only

compound 2 displayed full electron density in the ATP binding

Our structural studies commenced with the cocrystallization

of the N7-substituted analogues. Iodotubercidin possesses an

iodine group at position 7 and is a semisynthetic derivative of

the natural compound tubercidin, whereas sangivamycin is a

natural compound derived from Streptomyces rimosus and

12,19,20

contains an amide group at position 7.

Compound 2 has

been noted to be a general kinase inhibitor, displaying an

C

DOI: 10.1021/acs.jmedchem.9b00020

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

site while compound 3 displayed partial electron density for

the adenine ring only. Overall, the crystal structures of

compounds 2 and 3 complexed to MtbAdoK displayed many

similarities with the previously reported MtbAdoK−adenosine

complex. The inhibitors were bound in the same conformation

and locations as we observed in the adenosine-bound structure.

Residues Phe116 and Phe102 were involved in π-stacking

interactions with the adenine and ribose rings, respectively.

The ribose moiety was found to be stabilized by hydrogen

bonding with several residues. Asp12 oxygen OD1 H-bonds

with the O2′ and O3′ of ribose, OD2 of catalytic base Asp257

H-bonds with O5′, ND2 of Asn52 H-bonds with O3′, and

backbone N of Gly48 H-bonds O2′; all of which are likely to

contribute signiﬁcantly to the binding aﬃnity (Tables S1, S3,

and S5 ).

The most notable diﬀerences between the cocrystal

structures of adenosine, 2, and 3 were observed in the

interactions formed by residues Gln172, Gln173, and Ser36′.

In the adenosine-bound structure, Gln172 NE2 H-bonds with

O5′ (2.62 Å) and O4′ (3.26 Å) while backbone O of Gln172

H-bonds with N6 (3.26 Å). Although interactions with Gln172

were observed in the MtbAdoK-3 complex (NE2−O4′

distance of 3.23 Å and NE2−O5′ distance of 2.83 Å), no

interactions formed with Gln172 were observed in the

MtbAdoK-2 complex. A similar scenario was seen for

Gln173. In the MtbAdoK-1 structure, NE2 of Gln173 H-

bonds with N1 (2.99 Å) while OE1 of Gln173 H-bonds with

N6 (2.96 Å) of the adenine scaﬀold. On the other hand, as

observed in the MtbAdoK-3 and MtbAdoK-2 binary

complexes, Gln173 only formed H-bonds with N6 with

distances of 3.25 Å (OE1 −N6) and 3.29 Å (OE1−N6),

respectively (Tables S1, S3, and S5 ).

substituted adenosine analogues, where Snas

́

e

̌

l et al. reported

that several N7-anthracene or N7-phenanthrene derivatives

1

5

displayed high potency against MtbAdoK. The same group

synthesized several N7-chain-extended bulky groups that

displayed low potency against the enzyme more than likely

due to occlusion by chain B, validating our structural

15

assessment. The bulky iodine group of 2 is observed to be

completely buried in a pocket formed by residues Phe116,

Phe102, Gln172, Gln173, Ala114, Ser115, Leu38′ and Ser36′

leading to several polar and vdW interactions (Figure 2c). In

contrast, the amide group of 3 occupies a smaller portion of

the pocket (Figure 2d). We also observed that the amide group

of 3 participates in many of the conserved H-bond interactions

(Gln172 with O5′ and O4′ and Gln173 with N6) when

compared with the MtbAdoK-1 complex (Tables S3 and S5 ).

In contrast, the bulky iodine group of 2 sterically hinders many

of the conserved interactions described in the MtbAdoK-1

complex, including all of the H-bonds formed by Gln172 and

Ser36′ (Gln172 with O5′, O4′ and N6 and Ser36 ′ with N6 and

N7) and the H-bond between Gln173 and N1 (Tables S1 and

S5 ). Our enzymatic assay also showed that compound 2 is

more potent than compound 3 displaying 50% inhibitory

concentrations (IC ) of 1.2 and 16.5 μM, respectively, further

validating our structural assessment. Finally, the crystal

structure complexes of the N7 analogues revealed the presence

of a “chimney-like” cavity situated above position N6 of the

adenine scaﬀold and that is formed by residues from both

chains of the MtbAdoK homodimer, thereby representing a

unique structural feature of MtbAdoK (Figure 2e).

Next, we decided to explore the chemical space surrounding

the binding pocket near the N6-position by cocrystallizing

MtbAdoK with 6-methylmercaptopurine riboside. Compound

4 is a derivative of mercaptopurine, a drug used to treat acute

lymphatic leukemia and possess a methylmercapto group at the

N6-position of the adenosine scaﬀold. The binary complex of

MtbAdoK-4 was reﬁned to 2.0 Å resolution with two

5

0

The crystal structures of compounds 2 and 3 showed that

the iodine group of 2 and the amide group of 3 was oriented

toward Ser36′. One unique aspect of MtbAdoK, being a

homodimer, is that residues from one subunit presumably

complete the active of the other. It has been shown that Ser36

from chain B completes the active site of chain A and vice

versa by hydrogen bonding with positions N6 (OG−N6

distance of 3.43 Å) and N7 (OG−N7 distance of 2.65 Å) of

molecules in the ASU in the P4 crystal space group (Figure

1

2f, Tables S6 and S7 ). Like the cocrystals structures of 2 and 3,

the MtbAdoK-4 binary complex displayed a closed con-

formation of the enzyme with a backbone rmsd of 1.46 Å

amongst all Cα atoms when compared with the MtbAdoK-1

complex. Positive electron density was also observed for a

molecule of 4 bound to the ATP site. As in the case of the

MtbAdoK-3 structure, we were only able to model the adenine

moiety due to the absence of electron density of the ribose

(Figure S3).

The MtbAdoK-4 structure was very similar to the cocrystal

structures of compounds 2−3 and MtbAdoK−adenosine

complex. The aforementioned binding interactions formed by

residues Phe102, Phe116, Asp12, Asp257, Ser8, Asn52, Gly48,

and Val49 were maintained. The most notable diﬀerences were

found with residues Gln173 and Ser36′. Foremost, the critical

H-bonds formed by NE2 and OE1 of Gln173 with N1 and N6

of compound 4 were not observed. Also, the H-bonds formed

by OG of Ser36′ with N6 and N7 of compound 4 were not

observed in the cocrystal structure. The lack of interactions

with residues Ser36′ and Gln173 can be attributed to steric

shielding by the methylmercapto group of 4. Detailed

inspection of the binary complex revealed that the methyl-

mercapto group of 4 was oriented toward a pocket formed by

residues Gln173, Gln172, Met121, Phe116, Arg176, Leu38′,

and Ser36′ (Figure 2g). Arg176 was observed to be oriented

toward the solvent in the MtbAdoK−adenosine and crystals

5

the adenine ring. Ser36 forms part of the lid-domain of

MtbAdoK, and as reported before, the dimerization interface of

MtbAdoK occurs via extensive van der Waals (vdW) contacts

5

formed by the lid-domain of each chain. The observed

orientation of both N7-substituted adenosine analogues (2−3)

suggests that the N7-position is obstructed by chain B. That is,

chain B represents an imposing physical barrier at this position,

thus limiting drug discovery eﬀorts to the active site of the

enzyme (Figure 2c,d). The MtbAdoK-3 structure showed that

the −NH of the amide group of 3 formed a H-bond with OG

2

of Ser36′ (3.00 Å, Table S3 ). In contrast, H-bond interactions

mediated by Ser36′ with compound 2 were not observed in the

MtbAdoK-2 complex. Instead, the interaction is of vdW nature

with OG−I distance of 3.07 Å (Table S1 ). Finally, the N7-

substituents of 2 and 3 were observed to be buried in a pocket

formed by several conserved active site residues of MtbAdoK

including Phe102 (Cys123 in hAdoK) and Phe116 (Phe170 in

hAdoK).

Overall, structural analysis of the N7-analogues (2−3)

suggested that bulky substitutions at the N7-position are likely

to be favorably accommodated in the active site (Figure 2c,d).

Our structural ﬁndings go well in accordance with previously

reported chemistry-based SAR eﬀorts focused on N7-

D

DOI: 10.1021/acs.jmedchem.9b00020

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Figure 3. Crystal structure of the MtbAdoK-6 complex. (a) Chemical structure of compound 6. (b) 6 bound to the active site of MtbAdoK. (c)

Superimposition of the cocrystal structures of compound 6 (gray) and adenosine (blue). (d) Surface representation of residues forming the pocket

where the thiophene group gets accommodated. Chain B is colored magenta, Arg176 is colored red, and chain A is colored by heteroatom. (e)

Arg176 adopts a diﬀerent conformation in the cocrystal structure of 4 (blue) when compared with the cocrystal structure of 6 (gray).

structures of compounds 2 and 3 (Figure 2c,d). However, the

MtbAdoK-4 structure revealed that Arg176 had well-deﬁned

electron density for the side chain oriented toward the

methylmercapto group of 4. The reorientation of Arg176 led to

the formation of a new compound-induced pocket between

residues Phe116, Met121, Gln172, Gln173, Arg176, Leu38′,

and Ser36′, consequently leading to extensive vdW and polar

interactions between the residues mentioned above and

methylmercapto group (Figure 2g). Remarkably, the com-

pound-induced movement of Arg176 closed the “chimney-like”

cavity we saw in the cocrystal structures of compounds 2 and

with the adenine atoms mentioned above and that are formed

by residues Ser36′, Gln172, and Gln173. In contrast, the

methylmercapto group of 4 not only sterically blocked the

aforementioned interactions but also was accommodated in a

new compound-induced pocket that is comprised in part by

residues from both chains of the MtbAdoK homodimer

(Gln173, Gln172, Met121, Phe116, Arg176, Leu38′, and

Ser36′). It should be noted that compound 2 has been

reported to be a very potent inhibitor of hAdoK (K

≈ 30

ihAdoK

4

nM and K_i_M_t_b_A_d_o_K≈ 210 nM). The diﬀerent degrees of

potencies can be attributed to the way compound 2 interacts

with the enzymes. In our MtbAdoK-2 complex, the compound

is shown to be partially solvent exposed because of the

presence the “chimney-like” cavity. In contrast, the previously

solved hAdoK-2 complex showed that the compound is

completely buried within the hAdoK monomer because of a

unique latch-like region of hAdoK comprised by residues 23−

3, indicating that residues surrounding the N6-position are

ﬂexible (Figure 2c−e,g,h). The orientation and proximity (4.4

Å) of the cationic CZ of Arg176 with respect to the sulfur

atom of the methylmercapto group suggests the possibility of

cation−polar interactions. Finally, it should be noted that the

methylmercapto group seemed to occupy a small portion of

the pocket, indicating that bulky groups might be good

candidates to further improve contacts at the position 6.

Just like compound 2, compound 4 displayed good potency

22

57 of the enzyme’s lid-domain (Figure S4 ). Also, the lack of

selectivity of compound 2 can be attributed to the fact that the

compound is buried within several conserved residues located

in the active site. Overall, the N7 versus N6 substituents,

exempliﬁed by compounds 2−3 and 4, seemed to be

accommodated in mutually exclusive pockets. The N7 pocket

is formed by several conserved active site residues while the N6

pocket is compound-induced and formed by unique MtbAdoK

residues Gln172, Gln173, Ser36′, Leu38′, and ﬂexible Arg176.

Finally, to explore the potential for modifying the 5′-position

and to evaluate if the ATP channel can be utilized as a viable

route for drug design, we solved the crystal structure of 5′-

(

IC of 2.3 μM) in our enzymatic assay. The similar potencies

50

can be attributed to the comparable vdW and polar contacts

formed by the substituents within their respective pockets.

Although these two inhibitors have substitutions at diﬀerent

positions within the adenine ring, their mode of interaction

with the protein is similar. That is, both substitutions

prevented H-bonds with positions N1, N6, and N7 of the

adenine moiety. Our structure-guided SAR studies revealed

that the iodine group of 2 sterically hinders the interactions

E

DOI: 10.1021/acs.jmedchem.9b00020

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

aminoadenosine bound to MtbAdoK (Figure S5). As in the

case of the compounds described above, compound 5 has been

previously shown to possess inhibitory activity against

AdoKs. The MtbAdoK-5 cocrystal was obtained in the

same condition as the reported adenosine-bound structure and

pocket showed that Arg176 reorients once more to

accommodate the bulkier thiophene group, conﬁrming the

plasticity of this residue (Figure 3e). Furthermore, the

orientation of the aromatic thiophene ring with respect to

Phe116 suggests the possibility of parallel-displaced π-staking

interactions, whereas the orientation of cationic CZ of Arg176

with respect to the sulfur atom of the thiophene group (CZ−S

distance of 3.97 Å) suggests the possibility of T-shaped

stacking interactions. Overall, the extensive vdW interactions

within the pocket coupled to the stacking interactions are the

most likely molecular basis for the signiﬁcant increase in

potency of 6 when compared with compounds 2−4. Finally,

the orientation of Arg176, when compared with the MtbAdoK-

4 structure, suggested that the “chimney-like” cavity we

observed in the crystal structure complexes described above

might be accessible if larger substitutions are utilized to force

Arg176 to give access into the cavity.

1

3

was reﬁned to 1.95 Å resolution in the P3 21 crystal space

1

group with one molecule in the ASU. The binary complex

displayed a closed conformation of the enzyme with a

backbone rmsd of 0.28 Å amongst all Cα backbone atoms

when compared with the MtbAdoK−adenosine complex, and

the inhibitor was bound in a very similar fashion as adenosine

(Figure 2i, Tables S8 and S9 , and Figure S6 ). Unlike, the

cocrystal structures of compounds 2−4; the MtbAdoK-5

structure had no electron density for a molecule of 5 bound to

the ATP site. The interactions observed between 5 and

MtbAdoK were remarkably similar to those previously

described in the adenosine-bound complex (Tables S8 and

5

S5 ). Indeed, the main diﬀerence was the interaction formed

To further characterize compound 6, we determined the

antimycobacterial and cytotoxic proﬁle of the compound. The

antimycobacterial assay with compound 6 showed a half

maximal eﬀective concentration (EC ) of 25.6 ± 2.4 μM with

by the 5′-NH , displaying a H-bond with OD2 of Asp257

2

(

2.61 Å). Here, enzymatic inhibition (IC₅of 8.0 μM) is

0

attributed to the absence of the hydroxyl group that is required

for catalysis.

5

0

a 50% cytotoxic concentration (CC ) > 100 μ M against

5

0

On the basis of the preliminary structure-guided SAR eﬀorts

with compounds 2−5, we screened a focused library of ∼140

nucleoside analogues with substitutions at N6, N7, and/or 5′

positions. The very potent compound identiﬁed in this screen

human dermal ﬁbroblasts (HDF) (Figures S8 and S9 ).

Indicating that, in general, 6-substituted adenosine analogues

might show selectivity to MtbAdoK over the human

counterpart.

(

compound 6, IC₅₀of 196.5 ± 20.3 nM) had a bulky

Overall, the structure-based SAR eﬀorts showed that the N6,

N7, and 5′ positions are indeed conducive to inhibition of the

MtbAdoK enzyme through diﬀerent molecular and structural

means. The 5′-position provides a mechanistic approach for

inhibition by taking advantage of the critical hydroxyl group

required for catalysis and provides a platform to design

inhibitors that might take advantage of the ATP channel. On

the other hand, the crystal structures of the N7- and N6-

substituted adenosine analogues showed that the substitutions

sterically hindered conserved H-bonds formed by residues

Gln172, Gln173, and Ser36′ with key positions N1, N6, and

N7 of the adenine ring; all of which have been noted to be

critical for substrate binding and recognition for human, rabbit,

thiophene group at position 6, validating the pharmacological

relevance of N6-substituted adenosine analogues (Figure 3a).

To investigate the structural basis of the ∼10× increase in

potency when compared with compounds 2 and 4 and to

advance our structure-guided drug discovery eﬀorts, we

crystallized the MtbAdoK-6 binary complex. The cocrystal

structure of compound 6 was solved by MR and reﬁned to a

resolution of 2.10 Å in the P2 2 2 crystal space group with

1

1 1

two molecules in the ASU (Figure 3b, Tables S10 and S11). As

observed in the structures described above, the MtbAdoK-6

complex displayed a closed conformation with a backbone

rmsd of 2.18 Å amongst all Cα atoms when compared with the

MtbAdoK−adenosine complex. Unlike the structures of

compounds 2−5, the MtbAdoK-6 complex revealed that the

compound binds in a diﬀerent orientation with respect to

adenosine (Figure 3c) and that the bulky substituent prevented

full lid domain closure by leaving a gap of 2.94 Å when

compared with the fully closed MtbAdoK−adenosine complex.

As in the case of the crystal structure complexes of compounds

6

and Toxoplasma gondii adenosine kinase. The crystal

structures of the N7-substituted analogues (2−3) revealed

the presence of a “chimney-like” cavity that extends above the

N6-position, and that is formed by the MtbAdoK homodimer.

However, the MtbAdoK-4 complex showed that the cavity is

closed by a conformational movement of Arg176 to interact

with the N6-substituent of 4 thereby forming a new and

compound-induced pocket. Identiﬁcation of the very potent

compound 6 revealed that the bulky aromatic thiophene group

is not only completely buried within the newly identiﬁed

pocket but also suggested that bulkier substitutions could open

back up the “chimney-like” cavity.

2

−4, the MtbAdoK-6 structure showed electron density in the

active site and partial electron density in the ATP site (Figure

S7 ).

The cocrystal structure of compound 6 showed that the

interactions formed by residues Phe116, Phe102, Ser8, Asp12,

Asn52, Asp257, Val49, and Gly48 were preserved. The

signiﬁcant diﬀerences were found once more with residues

Gln172, Gln173, and Ser36′. In fact, none of the H-bond

interactions formed between Gln172 (with O4′, O5′ and N6),

Gln173 (with N1 and N6), and Ser36′ (with N6 and N7) were

observed (Table S10 ). Instead, these residues were involved in

weaker vdW interactions with the adenine ring and ribose

rings. The thiophene moiety formed unique contacts. This

group was observed to be completely buried within the newly

identiﬁed and compound-induced pocket that is formed by

residues Phe102, Phe116, Gln172, Gln173, Arg176, Ser36′,

Leu38′ and Phe37′ (Figure 3d). Closer inspection of the

Despite oﬀering inhibitory activity against MtbAdoK, the

N7-substituents were observed to be buried in a pocket that is

formed by several conserved active site residues and that the

substituent itself is physically blocked by the dimerization

event of MtbAdoK. Consequently, limiting medicinal chem-

istry eﬀorts to the conserved active site. In contrast, the N6-

position oﬀers a better route for structure-guided drug design

by providing the possibility of exploiting many contacts formed

by the unique oligomerization state of MtbAdoK. In addition,

it is possible that the presence of this cavity is the structural

basis behind the reported higher degree of speciﬁcity conferred

12

by N6-substituted analogues when compared with hAdoK.

F

DOI: 10.1021/acs.jmedchem.9b00020

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

a

Scheme 1. Synthesis of Adenosine Analogues

a

Reagents and conditions: (a) TBSCl, imidazole, DCM, 15 °C, 17 h, 99%; (b) 1-phenylpiperazine or 1-([1,1′-biphenyl]-4-yl)piperazine or 1-(4-

bromophenyl)piperazine, DIEA, EtOH, 70−80 °C, 17 h, 13−90%; (c) ArB(OH) , K PO , XPhos Pd G2, THF/H O, 70 °C, 17 h, 49−87%; (d)

2

3

4

2

TFA, THF/H O, 15−25 °C, 2−17 h, 13−67%; (e) 4,4,5,5-tetramethyl-2-(4-(phenylethynyl)phenyl)-1,3,2-dioxaborolane, K PO , XPhos Pd G2,

2

3

4

THF, 70 °C, 12 h, 95% or 4-ethynyl-1,1′-biphenyl, Cs CO , CuI, XPhos Pd G2, CH CN, 90 °C, 17 h, 27%; (f) 1-([1,1′-biphenyl]-4-yl)piperazine,

2

3

Cs CO , 1-([1,1′-biphenyl]-4-yl)piperazine, RuPhos Pd G2, tert-amyl-OH, 100 °C, 17 h, 50%; and (g) NH OH, MeOH, 15 °C, 24 h, 21%.

2

3

4

Taken together, we hypothesized that chain extension via

bulky substitutions at the N6-position might trigger a

conformational change of Arg176, thereby opening the

imidazopyridine A4, C−N coupling in the presence of second

generation RuPhos precatalyst proceeded smoothly and global

deprotection of acetyl groups by ammonium hydroxide gave

the triol 8 (Scheme 1 and chemistry section of the Supporting

“

chimney-like” cavity, thus providing a unique route to

26

increase potency and selectivity.

Information).

We next set out to systematically synthesize a series of 6-

substituted adenosine analogues. Synthesis of the novel

adenosine analogues (7−18) is shown in Scheme 1. Starting

from the known 6-chloro-9H-purine A0 or A1 and TBS

protected A2, 4-biphenyl and 4-Br-phenylpiperazinyl moieties

were introduced by substitution in EtOH to aﬀord B1 and B2

that was converted to variety of 4-arylphenyl(R)piperazine by

using diﬀerent arylboronic acids followed by removal of

acetonide with/without the TBS group which gave the triol

All synthesized adenosine analogues were characterized for

inhibitory activity against MtbAdoK and hAdoK. The

compounds were also tested against Mtb and HDF cells to

proﬁle their antimycobacterial and cytotoxic properties,

respectively (Table 1). We began our eﬀorts with compound

7. This compound showed excellent potency and selectivity

against MtbAdoK when compared with hAdoK. We hypothe-

sized that the piperazine group might lead to favorable contacts

with ﬂexible residue Arg176 while the benzene ring could

extend toward the unique cavity, thereby conferring selectivity.

To investigate the structural basis of inhibition and

selectivity of 7, we cocrystallized the compound with

MtbAdoK. The crystal structure of the MtbAdoK-7 binary

23,24

adenosines, 7, 9, 10, 12, 13, 14, 16, and 18, respectively.

The cyclopentane 15 was directly synthesized from

1R,2S,3R,5R)-3-(6-chloro-9H-purin-9-yl)-5-(hydroxymethyl)-

(

25

cyclopentane-1,2-diol A3 without alcohol protection. The

C−C bond at 6-position of purine was formed by Suzuki−

Miyaura or Sonogashira cross-coupling to lead to alkyne

analogues 11 and 17. To the tri-acetyl protected 7-Cl-

complex was solved by MR in the P4 crystal space group with

1

two molecules in the ASU and reﬁned to a resolution of 1.70 Å

(Figure 4a−c). Like the structures described above, the

G

DOI: 10.1021/acs.jmedchem.9b00020

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

a

Table 1. SAR Data for Synthesized Adenosine Analogues with Substitutions at Position 6

ID

R or R⁶

Y

Z

MtbAdoK K (nM)

hAdoK K (μM)

Mtb EC₅₀(μM)

HDF CC₅₀(μM)

i

7

8

9

R

O

N

120.2 ± 0.9

16.3 ± 0.1

18.5 ± 0.06

18.9 ± 0.5

19.9 ± 0.03

21.3 ± 0.2

23.0 ± 1.1

25.5 ± 2.8

27.5 ± 3.5

32.6 ± 0.3

48.0 ± 0.6

5.3 ± 0.07

5.1 ± 0.5

>50.0

≥15.1

0.41 ± 0.07

>50.0

≥65.0

>50.0

≥30.0

≥50.0

>50.0

≥25.0

>50.0

1.7 ± 0.02

4.0 ± 0.2

30.9 ± 4.3

>50.0

CH

N

10

11

12

13

14

15

16

17

18

6

R

CH₂

O

1.6 ± 0.1

>50.0

3.5 ± 0.4

>50.0

6

R

O

a

The error is reported as ± SD of 2 experiments.

Figure 4. Crystal structure of the MtbAdoK-7 complex. (a) 7 bound to the active site to MtbAdoK. Chain A is colored gray and chain B is colored

magenta. (b) The bulky substituent of 7 (gray) forces the compound to bind in a diﬀerent orientation with respect to adenosine (blue; PDB ID

2

PKM). (c) The bulky substituent forces Arg176 to open the cavity while the distal benzene ring is reoriented back into the active site groove that

is formed by residues of chain A (heteroatom) and chain B (magenta).

MtbAdoK-7 complex displayed a closed conformation of the

enzyme with a backbone rmsd of 1.52 Å amongst all Cα atoms

when compared with the MtbAdoK−adenosine structure.

Unlike the structures described above, the MtbAdoK-7

structure had no electron density for a molecule bound to

the ATP site (Figure S10). Overall, the interactions seen

between the adenine ring and ribose scaﬀold were very similar

to those observed in the MtbAdoK−adenosine complex

diﬀerences were found once more with residues Gln172,

Gln173, and Ser36′. Here, Gln173 and Ser36′ were not

observed to make any of the critical H-bonds with positions

N1, N6, and N7 while Gln172 retains the H-bonds with O4′ of

ribose (NE2−O4′ distance of 3.03 Å) and O5′ (NE2−O5′

distance of 3.29 Å) but lacks the H-bond with N6. As in the

case of the MtbAdoK-6 structure, superimposition of the

crystal structure complexes of MtbAdoK−adenosine and

MtbAdoK-7 showed that the bulky substitution of 7 causes

(Tables S12 and S13). Interestingly, the most signiﬁcant

H

DOI: 10.1021/acs.jmedchem.9b00020

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Figure 5. Summary of SAR results for compounds 8−17. Compound 7 bound to the active site of MtbAdoK. Based on the MtbAdoK-7 complex,

compounds 8−17 where synthesized. Chain A is colored white, and chain B is colored magenta.

Figure 6. Crystal structure of the MtbAdoK-17 complex. (a) 17 bound to the active site of MtbAdoK. (b) The large bulky substitution gets

accommodated in the “chimney-like” cavity. (c) Compound 17 (gray) binds in a diﬀerent orientation with respect to adenosine (blue). (d)

Superimposition of crystal structure complexes of 17 (gray) and 7 (blue). Chain B residues forming the distal part of the “chimney-like” cavity

colored magenta while chain A residues are colored by heteroatom.

the compound to bind in a diﬀerent orientation with respect to

adenosine with notable diﬀerences in critical lid domain

residues Asp12, Phe102, and Phe116 (Figure 4b). Here, the lid

domain gap with respect to the MtbAdoK-1 complex was an

average of 2.45 Å. We also observed that the phenylpiperazine

forces Arg176 to open the “chimney-like” cavity by reorienting

the residue back to the solvent. The crystal structure also

showed that the piperazine group redirects the distal benzene

group within the enzyme’s active site groove and that there was

suﬃcient space to incorporate additional groups at the distal

benzene ring to fully occupy the cavity (Figure 4c).

SAR with compound 5 suggested that bisubstrate-like

inhibitors might be able to take advantage of the ATP channel.

To investigate if the ATP channel is a viable route for drug

design, we synthesized several 5 ′-substituted adenosine

analogues (Figure S11 and Table S14). Despite having

suﬃcient space to accommodate large groups, bulky phenyl-

substituents at the 5′-position resulted in inhibitors with

negligible inhibitory activity. Indeed, only one of the

synthesized compounds displayed some measurable inhibitory

activity against the enzyme with ∼53% inhibition at 12.5 μM of

the compound. The lack of inhibitory activity of these

analogues may be due to suboptimal angles adopted by the

bulky substitution.

As noted above, another position of interest was the 5′-

position of the ribose scaﬀold. The preliminary structure-based

I

DOI: 10.1021/acs.jmedchem.9b00020

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

−

1

Figure 7. Preliminary pharmacokinetics of compound 18. (a) Oral PK study conducted at 10 mg·kg utilizing 9 Swiss Webster Female mice with

−

1

three animal bleeds per time point. Maximum plasma concentration of 0.13 ± 0.02 μg·mL observed at 4 h following gavage. The AUC₀₋₂₄_hof 1.8

μg·h·mL⁻was calculated using the linear trapezoid rule method. (b) Snapshot PK studies conducted from plasma samples taken from tolerability

1

−1

mice at 50, 100, and 200 mg·kg following ﬁrst initial dose. 10 mg·kg data included for comparison. (c) C_m_a_x, t_m_a_x, and AUC of compound 18 at

four dose levels.

The MtbAdoK-7 complex suggested that further extension

at the distal benzene group could be accommodated in the

cavity and that rigid substituents at the 6-position might orient

the substituent within the cavity. On the basis of these

observations, we synthesized compounds 8−17. As shown in

Table 1, all compounds displayed high potency against

crystal structure complexes of compounds 17 and adenosine

also showed that the bulky substituent of 17 causes the

compound to bind in a diﬀerent orientation with respect to

adenosine, leaving lid domain gap of 1.81 Å when compared

with the MtbAdoK−adenosine complex (Figure 6c).

The MtbAdoK-17 binary complex showed that the

compound binds to the active site of the enzyme forming

similar contacts with the protein as observed with the

adenosine-bound complex (Figure 6a, Tables S15 and S5).

Structural diﬀerences were found again with residues Gln172,

Gln173, and Ser36′. Only Gln173 was observed to form a weak

H-bond with endocyclic N1 of the adenine ring (NE2−N1

distance of 3.21 Å) while residues Ser36′ and Gln172 were not

observed to make any of the critical H-bond interactions

described above. The major diﬀerence between the MtbAdoK-

7 and MtbAdoK-17 structures is the orientation that the 6-

substituent adopts as it goes into the “chimney-like” cavity.

The cocrystal structure of the MtbAdoK-7 complex suggests

that the piperazine-containing adenosine analogues might

reorient to the active site groove that is composed of

hydrophobic residues from both chains including Ala175.A,

Leu38.B, Leu35.B, Phe37.B, and polar residues Ser36.B and

Arg176.A. In contrast, the rigidity imparted by the triple bond

of compound 17 limits the substitution to adopt a bond angle

of 180° with respect to the 6-position. Consequently, one side

of the distal benzene ring of 17 was seen to be oriented toward

backbone and main chain atoms of Phe37.B while the other

side was observed to be partially solvent exposed (Figures 6d

and S12 ).

MtbAdoK with K values ranging from ∼16 to 48 nM. Overall,

i

the SAR results are in accordance previous observations where

it was noted that 6-substituted adenosine analogues displayed a

higher degree of speciﬁcity to MtbAdoK when compared with

12,14

hAdoK (Figure 5).

The apparent lack of correlation

between enzymatic inhibition and intracellular activity might

be due to several reasons such as a lack of permeability,

ineﬃcient import, or rapid export. For example, it is known

that Mtb possesses over 30 ATP-binding cassettes (ABC)

transporters which are involved in the uptake and export of

27,28

many molecules including nucleosides.

In addition, Mtb

possesses several classes of drug eﬄux pumps including ABC

transporters, the small multidrug resistance superfamily, the

major facilitator superfamily, and the resistance nodulation cell

division superfamily; all of which have been attributed to eﬄux

29−32

of many classes of antimycobacterial drugs.

It is currently

unknown how Mtb is able to uptake, discriminate, import, and

export this series of compounds. Further experiments such as

resistant mutant isolation, knockdown or overexpression

strains, and systematic interrogation of potential transporters

and eﬄux pumps will be needed to investigate the intracellular

fate of the series.

Compounds 11 and 17 were designed to test diﬀerent

orientations of the 6-substituent within the cavity. These

compounds possess a rigid triple bond instead of the ﬂexible

piperazine group. Because compound 17 displayed antimyco-

bacterial activity, we decided to determine the cocrystal

structure of MtbAdoK with compound 17. The crystal

structure of the MtbAdoK-17 complex was solved by MR in

Finally, on the basis of the chemistry and structure-guided

SAR eﬀorts with compounds 7−17, we synthesized compound

18. This compound is an analogue of compounds 8 and 15.

The major diﬀerence is that 18 contains endocyclic atoms N1

and O4′ which we observed in the structures described above

that they were involved in key H-bond interactions with

residues Gln172 and Gln173, and the SAR was suggested to be

relevant in potency. Indeed, compound 18 was our most

the P4 crystal space group with two molecules in the ASU and

1

re ﬁ ned to a resolution of 2.23 Å (Figure 6a−c and Tables S15

and S16). The binary complex displayed a closed conformation

of the enzyme with a backbone rmsd of 1.40 Å amongst all Cα

atoms when compared with the MtbAdoK−adenosine

structure. Just like the MtbAdoK-7 structure, the MtbAdoK-

potent derivative, exhibiting a K of ∼5 nM against MtbAdoK,

i

no cytotoxicity, no activity against hAdoK, and an EC of

5

0

∼4.0 μM against Mtb. Overall, the observed diﬀerences in

binding modalities of reorientation within the enzyme versus

solvent exposed could explain the overall higher potency of the

piperazine-substituted analogues. Furthermore, the predom-

17 binary complex had no electron density for another

molecule of 17 bound to the ATP site. Superimposition of the

J

DOI: 10.1021/acs.jmedchem.9b00020

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

inantly hydrophobic character of the “chimney-like” cavity goes

in accordance with the SAR studies where hydrophobic

substitutions displayed a higher degree of potency when

compared with the polar substituted analogues (Table 1).

To further characterize our most potent derivative (18), we

performed steady-state kinetic experiments and in vivo

pharmacokinetics and tolerability studies. The kinetics experi-

ments showed that increasing the concentration of adenosine

in the presence of a ﬁxed inhibitor concentration led to an

piperazine group might lead to more favorable contacts with

the enzyme, thereby leading to a higher degree of potency. The

predominantly hydrophobic character of the cavity goes in

accordance with the SAR studies demonstrating that hydro-

phobic substitutions were preferred at the 6-position. More-

over, the lid domain gaps and diﬀerent orientation of

compounds 6, 7, and 17 with respect to the MtbAdoK−

adenosine structure suggest that bulky substitutions at position

6 might exert its inhibitory eﬀects by sterically preventing full

lid domain closure. Our preliminary in vivo pharmacokinetics

and tolerability studies showed that our most potent derivative

is very well tolerated and that MIC values are indeed

achievable. However, future studies regarding SAR derivatiza-

tion will be needed to optimize pharmacokinetic parameters.

Overall, we showed that the unique “chimney-like” cavity

formed by the unique oligomerization state of MtbAdoK could

be utilized as a canvas for future medicinal chemistry eﬀorts to

further improve the in vitro and in vivo potency and selectivity.

increase of the K , consistent with a competitive mode of

m

inhibition (Figure S13a,b and Table S17). The K value for

m

adenosine determined with our enzymatic assay was 1.71 ±

0

.02 μM, well in agreement with previously reported values of

4,33

.8−3.4 μM.

To determine in vivo pharmacokinetic

parameters, we administered compound 18 to female Swiss

−

1

Webster mice via oral gavage as a single dose at 10 mg·kg

and quantiﬁed plasma concentrations at 0, 0.25, 0.5, 1, 2, 4, 6,

, and 24 h by liquid chromatography−mass spectrometry

−1

8

34,35

(

LC/MS, Figure 7a).

Compound 18 dosed at 10 mg·kg

1

−

achieved a C_m_a_xof 0.13 ± 0.02 μg·mL₋at approximately 4 h

EXPERIMENTAL SECTION

■

1

with an AUC₀₋₂₄_hof 1.80 μg·h·mL . At this dose level,

General Synthetic Chemistry Methods. Solvents, reagents, and

intermediates that are commercially available were used as received.

Reagents and intermediates that were not commercially available were

prepared in the manner as described below. Compounds 1, 3, and 4

were purchased at Sigma, with speciﬁc purities of >99.0, >98, and

36

compound 18 was well tolerated. In vivo tolerability studies

were also performed by administering compound 18 to cohorts

of female Swiss Webster mice (n = 3) at 50, 100, and 200 mg·

−

1

kg via oral gavage once daily for 3 days. We evaluated mice

twice daily for signs of acute toxicity and observed no adverse

eﬀects or weight loss (Table S18 ). Blood collected at 0, 1, 4,

and 24 h after the ﬁrst administered dose was used to

determine PK for compound 18 at these three higher dose

levels (Figure 7b). Data points collected from the 10 mg·kg

study was included for comparison. Compound 18 achieved a

>

99%, respectively. Compound 2 was obtained from ChemScene with

speciﬁc purity of >95%, and compounds 5 and 6 are part of Merck’s

nucleoside collection library. H NMR spectra were measured on

1

either a Varian VNMR System 400 or Bruker AVANCE 400

spectrometer at 400 MHz and are reported as ppm downﬁeld from

36

−1

Me Si with number of protons, multiplicities, and coupling constants

4

in hertz indicated parenthetically. Where LC/MS data are presented,

analyses were performed using an Agilent 6110A MSD or an Applied

Shimadzu 2020MSD. The parent ion is given. Unless otherwise

stated, all compounds possess a purity ≥ 95% as determined by LC/

MS. Puriﬁcation was conducted with reversed-phase HPLC (Waters

Xbridge Prep OBD C18 (150 × 30 mm, 5 μm), YMC-Actus Pro C18

(150 × 30 mm, 5 μm), and Agela ASB (150 × 25 mm, 5 μm)

column) or silica gel column chromatography (Liang Chen Gui Yuan

maximum plasma concentration for all dose levels (50, 100,

−

1

and 200 mg·kg ) at approximately 4 h with C of 0.54 ±

max

−1

−

1

0

.22 μg·mL (1.1 ± 0.4 μM), 0.92 ± 0.09 μg·mL (1.9 ± 0.2

−

1

μM), and 1.58 ± 0.38 μg·mL (3.2 ± 0.8 μM), respectively

(

Figure 7b,c). Calculated exposure for these higher dose levels

−

1

improved overall exposure to 9.8, 15.3, and 23.2 μg·h·mL .

On the basis of our preliminary pharmacokinetic data,

compound 18 is very well tolerated during acute exposure.

Most notably, at the highest concentration tested, 18 falls

within the range of EC₅₀value we determined in our in vitro

assay (Table 1).

2

00−300 mesh) or PTLC (Liang Chen Gui Yuan, acryloid cement;

0

.5 × 200 × 200 mm). For the speci ﬁ c chemical synthesis of reaction

intermediates, see the Supporting Information .

(2R,3R,4S,5R)-2-(7-(4-([1-Phenyl]-4-yl)piperazin-1-yl)-3H-

imidazo[4,5-b]pyridin-3-yl)-5-(hydroxymethyl)tetrahydrofuran-3,4-

diol (7). To a suspension of 1-phenylpiperazine (210 mg, 1.3 mmol)

in anhydrous EtOH (6 mL) were added DIEA (0.3 mL, 1.7 mmol)

and 6-chloro-purine riboside A0 (287 mg, 1 mmol) at room

temperature, and it was stirred at 70 °C for 17 h after cooling to

room temperature. The resulting solid was ﬁltered and washed with

dichloromethane (DCM, 6 mL × 2) and dried under reduced

pressure to aﬀord (2R,3R,4S,5R)-2-(7-(4-([1-phenyl]-4-yl)piperazin-

CONCLUSIONS

■

These studies represent a systematic structure-guided approach

to the rational design of very potent MtbAdoK inhibitors with

low micromolar activity against Mtb. Furthermore, our studies

oﬀer a structural explanation behind the speciﬁcity of

12−15

1

-yl)-3H-imidazo[4,5-b]pyridin-3-yl)-5-(hydroxymethyl)-

previously reported N6-substituted adenosine analogues.

tetrahydrofuran-3,4-diol (7), (288 mg, 70% yield) as a white solid;

That is, substitutions at the 6-position of the adenine scaﬀold

might be accommodated in a compound-induced pocket that

is formed within the unique oligomerization state of MtbAdoK

when compared with hAdoK and other eukaryotic adenosine

kinases. Unlike the crystal structures of compounds 2−6, the

crystal structures of 7 and 17 did not display any electron

density for a compound bound to the ATP site suggesting that

there might be a size threshold conferring speciﬁcity to the

active site versus ATP site. Therefore, future strategies to

design bisubstrate-like inhibitors should focus on small

substituents at the N6-position with bulky groups at the 5′-

position.

LC/MS purity of 97%; Scheme A0 Supporting Information . MS (ESI)

+

1

m/z: 412.9 [M + H] . H NMR (DMSO-d , 400 MHz): δ 8.43 (s,

6

1

5

H), 8.27 (s, 1H), 7.21−7.26 (m, 2H), 6.99−7.01 (m, 2H), 6.81 (m,

H), 5.93 (d, J = 6.0 Hz, 1H), 5.42 (d, J = 6.0 Hz, 1H), 5.27 (m, 1H),

.14 (d, J = 4.8 Hz, 1H), 4.58 (m, 1H), 4.38 (br s, 4H), 4.16 (m, 1H),

3.97 (m, 1H), 3.65−3.70 (m, 1H), 3.55−3.60 (m, 1H), 3.20−3.35

(m, 4H).

(2R,3R,4S,5R)-2-(7-(4-([1,1′-Biphenyl]-4-yl)piperazin-1-yl)-3H-

imidazo[4,5-b]pyridin-3-yl)-5-(hydroxymethyl)tetrahydrofuran-3,4-

diol (8). To a solution of (2R,3R,4R,5R)-2-(7-(4-([1,1′-biphenyl]-4-

yl)piperazin-1-yl)-3H-imidazo[4,5-b]pyridin-3-yl)-5-(acetoxymethyl)-

tetrahydrofuran-3,4-diyl diacetate B4 (130 mg, 0.082 mmol (38.8%

purity)) in MeOH (3 mL) was added ammonium hydroxide (1.5 mL)

dropwise at 15 °C, and it was stirred at 15 °C for 24 h. The mixture

was concentrated and puriﬁed by reversed-phase HPLC [Waters

The crystal structures of the bulky 6-substituted adenosine

analogues suggested that the reorientation conferred by the

K

DOI: 10.1021/acs.jmedchem.9b00020

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Xbridge Prep OBD C18 (150 × 30 mm, 5 μm) column, elution of

acetonitrile/water (0.05% ammonia hydroxide)] to aﬀord

J = 3.2 Hz, 1H), 3.95 (d, J = 3.2 Hz, 1H), 3.65−3.64 (m, 1H), 3.55−

+

3.53 (m, 1H), 3.37 (s, 4H). MS (ESI) m/z: 557.2 [M + H] .

(

2R,3R,4S,5R)-2-(7-(4-([1,1′-biphenyl]-4-yl)piperazin-1-yl)-3H-

(2R,3R,4S,5R)-2-(6-(4-(4-(1-(Diﬂuoromethyl)-1H-pyrazol-4-yl)-

phenyl)piperazin-1-yl)-9H-purin-9-yl)-5-(hydroxymethyl)-

tetrahydrofuran-3,4-diol (13 ). Yellow solid. Following the similar

procedure shown in Scheme C with diﬀerent boronic acids

imidazo[4,5-b]pyridin-3-yl)-5-(hydroxymethyl)tetrahydrofuran-3,4-

diol 8 (8.81 mg, 21.3% yield) as a white solid; LC/MS purity of 97%;

1

Scheme B Supporting Information . H NMR (DMSO-d , 400 MHz):

6

(

RB(OH) ), compound 13 was synthesized; LC/MS purity of 97%.

δ 8.40 (s, 1H), 7.99 (d, J = 5.6 Hz, 1H), 7.62−7.56 (m, 4H), 7.43−

2

1

H NMR (DMSO-d , 400 MHz): δ 8.58 (s, 1H), 8.46 (s, 1H), 8.29

7

=

.39 (m, 2H), 7.29−7.27 (m, 1H), 7.10 (d, J = 8.8 Hz, 2H), 6.66 (d, J

6

(

s, 1H), 8.19 (s, 1H), 7.80 (t, J = 59.2 Hz, 1H), 7.57 (d, J = 8.4 Hz,

H), 7.05 (d, J = 8.8 Hz, 2H), 5.93 (d, J = 5.6 Hz, 1H), 4.59−4.57

6.0 Hz, 1H), 5.69−5.67 (m, 1H), 5.42 (d, J = 6.0 Hz, 1H), 5.17 (d,

2

J = 4.4 Hz, 1H), 4.66−4.65 (m, 1H), 4.15 (s, 1H), 4.01 (s, 4H), 3.98

(m, 1H), 4.39 (br s, 4H), 4.15 (t, J = 3.2 Hz, 1H), 3.96 (s, 1H), 3.67−

(

4

d, J = 2.4 Hz, 1H), 3.66−3.65 (m, 1H), 3.56−3.55 (m, 1H), 3.37 (s,

3

.66 (m, 1H), 3.57−3.54 (m, 1H), 3.32 (s, 4H). MS (ESI) m/z: 529.2

H). MS (ESI) m/z: 488.1.

2R,3S,4R,5R)-2-(Hydroxymethyl)-5-(6-(4-(4-(6-methoxypyridin-

-yl)phenyl)piperazin-1-yl)-9H-purin-9-yl)tetrahydrofuran-3,4-diol

+

[M + H] .

(

2R,3S,4R,5R)-2-(Hydroxymethyl)-5-(6-(4-(3′-(morpholinometh-

3

yl)-[1,1′-biphenyl]-4-yl)piperazin-1-yl)-9H-purin-9-yl)-

tetrahydrofuran-3,4-diol (14 ). Yellow solid. Following the similar

procedure shown in Scheme C with diﬀerent boronic acids

(

9). The mixture of 9-((3aR,4R,6R,6aR)-6-(((tert-butyldimethylsilyl)-

oxy)methyl)-2,2-dimethyltetrahydrofuro[3,4-d][1,3]dioxol-4-yl)-6-

(

4-(4-(6-methoxypyridin-3-yl)phenyl)piperazin-1-yl)-9H-purine B3-1

(

RB(OH) ), compound 14 was synthesized; LC/MS purity of 95%.

80 mg, 0.119 mmol) in THF (1.0 mL) and H O (0.2 mL) was added

2

1

H NMR (CD OD, 400 MHz): δ 8.34 (d, J = 12.4 Hz, 2H), 7.77−

TFA (0.5 mL, 6.49 mmol), and the mixture was stirred at 20 °C for

1

3

7

.75 (m, 2H), 7.64−7.61 (m, 2H), 7.54−7.52 (m, 1H), 7.43−7.42

7 h, concentrated, and puriﬁed by reversed-phase HPLC [Waters

(m, 1H), 7.16 (d, J = 8.8 Hz, 2H), 6.01 (d, J = 5.6 Hz, 1H), 4.67−

Xbridge Prep OBD C18 (150 × 30 mm, 5 μm) column, elution of

acetonitrile/water (0.05% ammonia hydroxide)] to give

4

3

.65 (m, 1H), 4.54 (br s, 4H), 4.42 (s, 2H), 4.33−4.32 (m, 1H),

.20−4.19 (m, 1H), 4.19−3.98 (m, 2H), 3.89−3.77 (m, 4H), 3.46−

(

2R,3S,4R,5R)-2-(hydroxymethyl)-5-(6-(4-(4-(6-methoxypyridin-3-

yl)phenyl)piperazin-1-yl)-9H-purin-9-yl)tetrahydrofuran-3,4-diol 9

26.4 mg, 42.7% yield) as a yellow solid; LC/MS purity of 99%;

+

.40 (m, 6H), 3.31−3.20 (2H). MS (ESI) m/z: 588.2 [M + H] .

(

1R,2S,3R,5R)-3-(6-(4-([1,1′-Biphenyl]-4-yl)piperazin-1-yl)-9H-

(

1

purin-9-yl)-5-(hydroxymethyl)cyclopentane-1,2-diol (15). The mix-

ture of 1-([1,1′-biphenyl]-4-yl)piperazine (56.3 mg, 0.236 mmol),

Scheme C Supporting Information . H NMR (DMSO-d , 400 MHz):

6

δ 8.45−8.42 (m, 2H), 8.28 (s, 1H), 7.96−7.93 (m, 1H), 7.54 (d, J =

(

1R,2S,3R,5R)-3-(6-chloro-9H-purin-9-yl)-5-(hydroxymethyl)-

8

.4 Hz, 2H), 7.09 (d, J = 8.4 Hz, 2H), 6.86 (d, J = 8.4 Hz, 1H), 5.93

cyclopentane-1,2-diol A3 (100 mg, 0.197 mmol), and DIEA (0.103

mL, 0.590 mmol) in EtOH (5 mL) was stirred at 70 °C for 17 h. The

mixture was cooled and concentrated, puriﬁed by reversed-phase

HPLC [Waters Xbridge Prep OBD C18 (150 × 30 mm, 5 μm)

column, elution of acetonitrile/water (0.05% ammonia hydroxide)] to

give (1R,2S,3R,5R)-3-(6-(4-([1,1′-biphenyl]-4-yl)piperazin-1-yl)-9H-

purin-9-yl)-5-(hydroxymethyl)cyclopentane-1,2-diol 15 (12.26 mg,

(d, J = 5.6 Hz, 1H), 5.48 (d, J = 5.6 Hz, 1H), 5.33 (t, J = 6.0 Hz, 1H),

5

.21 (d, J = 4.4 Hz, 1H), 4.59−4.58 (m, 1H), 4.40 (br s, 4H), 4.15 (d,

J = 3.2 Hz, 1H), 3.96 (s, 1H), 3.87 (s, 3H), 3.69−3.66 (m, 1H),

+

3.57−3.56 (m, 1H), 3.35 (s, 4H). MS (ESI) m/z: 520.2 [M + H] .

(

2R,3S,4R,5R)-2-(Hydroxymethyl)-5-(6-(4-(4′-(2-hydroxypropan-

2

-yl)-[1,1′-biphenyl]-4-yl)piperazin-1-yl)-9H-purin-9-yl)-

tetrahydrofuran-3,4-diol (10 ). Yellow solid. Following the similar

procedure shown in Scheme C with diﬀerent boronic acids

1

2.77% yield) as a white solid; LC/MS purity of 99%; Scheme F

1

Supporting Information. H NMR (DMSO-d , 400 MHz): δ 8.29 (s,

6

(

RB(OH) ), compound 10 was synthesized; LC/MS purity of 99%.

2

1

1H), 8.27 (s, 1H), 7.62−7.55 (m, 4H), 7.42−7.39 (m, 2H), 7.29−

H NMR (DMSO-d , 400 MHz): δ 8.46 (s, 1H), 8.29 (s, 1H), 7.55−

7

6

7

4

3

1

.27 (m, 1H), 7.09 (d, J = 8.8 Hz, 2H), 4.93 (d, J = 6.8 Hz, 1H),

.75−4.66 (m, 3H), 4.40−4.15 (m, 5H), 3.84−3.83 (m, 1H), 3.51−

.46 (m, 2H), 3.35 (br s, 4H), 2.25−2.22 (m, 1H), 2.06−2.01 (m,

.48 (m, 6H), 7.09 (d, J = 8.8 Hz, 2H), 5.94 (d, J = 5.6 Hz, 1H), 5.48

(

d, J = 5.6 Hz, 1H), 5.33 (t, J = 5.6 Hz, 1H), 5.21 (d, J = 5.2 Hz, 1H),

5

1

3

.01 (s, 1H), 4.60−4.58 (m, 1H), 4.40 (br s, 4H), 4.16−4.15 (m,

+

H), 1.75−1.70 (m, 1H). MS (ESI) m/z: 487.3 [M + H] .

H), 3.97−3.96 (m, 1H), 3.68−3.67 (m, 1H), 3.58−3.56 (m, 1H),

(2R,3S,4R,5R)-2-(Hydroxymethyl)-5-(6-(4-(4-(6-(triﬂuoromethyl)-

+

.31 (s, 4H), 1.44 (s, 6H). MS (ESI) m/z: 547.2 [M + H] .

pyridin-3-yl)phenyl)piperazin-1-yl)-9H-purin-9-yl)tetrahydrofuran-

,4-diol (16). Yellow solid. Following the similar procedure shown in

(

2R,3S,4R,5R)-2-(Hydroxymethyl)-5-(6-(4-(phenylethynyl)-

3

phenyl)-9H-purin-9-yl)tetrahydrofuran-3,4-diol (11). To a solution

of 9-((3aR,4R,6R,6aR)-6-(((tert-butyldimethylsilyl)oxy)methyl)-2,2-

dimethyltetrahydrofuro[3,4-d][1,3]dioxol-4-yl)-6-(4-(phenylethynyl)-

phenyl)-9H-purine C1-1 (103 mg, 0.18 mmol) in THF (1.5 mL) and

Scheme C with diﬀerent boronic acids (RB(OH) ), compound 16

2

1

was synthesized; LC/MS purity of 97%. H NMR (CD OD, 400

3

MHz): δ 8.94 (s, 1H), 8.38 (d, J = 16.4 Hz, 2H), 8.23 (d, J = 7.6 Hz,

1H), 7.83 (d, J = 8.0 Hz, 1H), 7.69 (d, J = 8.8 Hz, 2H), 7.20 (d, J =

H O (0.2 mL) was added TFA (2.5 mL), and it was stirred at 20 °C

2

8.8 Hz, 2H), 6.01 (d, J = 5.6 Hz, 1H), 4.66−4.65 (m, 1H), 4.56 (br s,

for 17 h. The mixture was diluted with dimethylformamide (DMF, 2

mL), and it was concentrated under reduced pressure. The residue

was puriﬁed by reversed-phase HPLC [TFA condition: Agela ASB

4

3

H] .

H), 4.33−4.32 (m, 1H), 4.20−4.19 (m, 1H), 3.90−3.89 (m, 1H),

.82−3.81 (m, 1H), 3.53−3.50 (m, 5H). MS (ESI) m/z: 558.2 [M +

+

(

150 × 25 mm, 5 μm) column, elution of acetonitrile/water (0.1%

(

2R,3R,4S,5R)-2-(6-([1,1′-Biphenyl]-4-ylethynyl)-9H-purin-9-yl)-5-

TFA)] and then lyophilized to give (2R,3S,4R,5R)-2-(hydroxymeth-

yl)-5-(6-(4-(phenylethynyl)phenyl)-9H-purin-9-yl)tetrahydrofuran-

(hydroxymethyl)tetrahydrofuran-3,4-diol (17). Following similar

procedure for 11 and Scheme D, compound 17 was synthesized.

The reaction of 6-([1,1′-biphenyl]-4-ylethynyl)-9-((3aR,4R,6R,6aR)-

6-(((tert-butyldimethylsilyl)oxy)methyl)-2,2-dimethyltetrahydrofuro-

[3,4-d][1,3]dioxol-4-yl)-9H-purine C1-2 (36 mg, 0.062 mmol) gave

(2R,3R,4S,5R)-2-(6-([1,1′-biphenyl]-4-ylethynyl)-9H-purin-9-yl)-5-

3

,4-diol 11 (17.48 mg, 23.3% yield) as a yellow solid; LC/MS purity

1

of 99%; Scheme D Supporting Information . H NMR (DMSO-d ,

6

4

7

6

4

00 MHz): δ 9.04 (s, 1H), 8.97 (s, 1H), 8.92 (d, J = 8.4 Hz, 2H),

.80 (d, J = 8.4 Hz, 2H), 7.62−7.61 (m, 2H), 7.47−7.46 (m, 3H),

.11 (d, J = 5.6 Hz, 1H), 4.67−4.65 (m, 1H), 4.23−4.21 (m, 1H),

(hydroxymethyl) tetrahydrofuran-3,4-diol 17 (9.86 mg, 37.2% yield)

as a yellow solid; LC/MS purity of 99%. H NMR (DMSO-d , 400

1

.01 (d, J = 3.6 Hz, 1H), 3.71−3.70 (m, 1H), 3.62−3.59 (m, 1H). MS

6

+

(ESI) m/z: 429.0 [M + H] .

MHz): δ 8.96 (d, J = 8.4 Hz, 2H), 7.86−7.75 (m, 6H), 7.53−7.50 (m,

2H), 7.42−7.41 (m, 1H), 6.07 (d, J = 5.2 Hz, 1H), 5.61 (d, J = 4.8

Hz, 1H), 5.30 (s, 1H), 5.13 (s, 1H), 4.63 (s, 1H), 4.21 (s, 1H), 4.00

(

2R,3S,4R,5R)-2-(Hydroxymethyl)-5-(6-(4-(4′-(triﬂuoromethyl)-

1,1′-biphenyl]-4-yl)piperazin-1-yl)-9H-purin-9-yl)tetrahydrofuran-

,4-diol (12). Yellow solid. Following the similar procedure shown in

[

3

(

d, J = 3.6 Hz, 1H), 3.73−3.70 (m, 1H), 3.61−3.58 (m, 1H). MS

+

Scheme C with diﬀerent boronic acids (RB(OH) ), compound 12

(ESI) m/z: 429.1 [M + H] .

2

1

was synthesized; LC/MS purity of 99%. H NMR (DMSO-d , 400

(2R,3R,4S,5R)-2-(6-(4-([1,1′-Biphenyl]-4-yl)piperazin-1-yl)-9H-

purin-9-yl)-5-(hydroxymethyl)tetrahydrofuran-3,4-diol (18). To a

solution of ((3aR,4R,6R,6aR)-6-(6-(4-([1,1′-biphenyl]-4-yl)piperazin-

1-yl)-9H-purin-9-yl)-2,2-dimethyltetrahydrofuro[3,4-d][1,3]dioxol-4-

6

MHz): δ 8.44 (s, 1H), 8.28 (s, 1H), 7.82 (d, J = 8.4 Hz, 2H), 7.72 (d,

J = 8.4 Hz, 2H), 7.64 (d, J = 8.4 Hz, 2H), 7.11 (d, J = 8.4 Hz, 2H),

5.92 (d, J = 6.0 Hz, 1H), 4.58−4.55 (m, 1H), 4.38 (br s, 4H), 4.15 (t,

L

DOI: 10.1021/acs.jmedchem.9b00020

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

yl)methanol B1 (3.1 g, 5.50 mmol) in THF (8 mL) and water (1.5

mL) was added TFA (30 mL) dropwise, and it was stirred at 15 °C

(s, 1H), 8.22 (s, 1H), 7.84 (d, J = 7.6 Hz, 2H), 7.55−7.45 (m, 5H),

7.13 (d, J = 8.4 Hz, 2H), 5.91 (d, J = 5.6 Hz, 1H), 4.76 (t, J = 5.6 Hz,

2H), 4.38 (br s, 4H), 4.19−4.10 (m, 2H), 3.65−3.45 (m, 5H). MS

for 2 h. After NH OH was added to adjust pH to 7−8, water (70 mL)

4

+

was added and the mixture was ﬁltered. The resulting solid was dried

under reduced pressure, and the residue was puriﬁed by silica gel

column chromatography (DCM: MeOH = 100:1 to 30:1) to give

(ESI) m/z: 584.3 [M + H] .

N-(4-((((2R,3S,4R,5R)-3,4-Dihydroxy-5-(6-(4-(4-(triﬂuoromethyl)-

phenyl)piperazin-1-yl)-9H-purin-9-yl)tetrahydrofuran-2-yl)methyl)-

amino)phenyl)methanesulfonamide (22). The mixture of N-(4-

((((3aR,4R,6R,6aR)-2,2-dimethyl-6-(6-(4-(4-(triﬂuoromethyl)-

phenyl)piperazin-1-yl)-9H-purin-9-yl)tetrahydrofuro[3,4-d][1,3]-

dioxol-4-yl)methyl)amino)phenyl)methanesulfonamide D7 (50 mg,

0.036 mmol) and TFA (0.5 mL, 6.49 mmol) in THF (1.0 mL) and

(

2R,3R,4S,5R)-2-(6-(4-([1,1′-biphenyl]-4-yl)piperazin-1-yl)-9H-

purin-9-yl)-5-(hydroxymethyl)tetrahydrofuran-3,4-diol (1.327 g,

8.2% yield) as a white solid; LC/MS purity of 98%; Scheme A

4

1

Supporting Information . H NMR (DMSO-d , 400 MHz): δ 8.46 (s,

6

1

H), 8.29 (s, 1H), 7.62−7.55 (m, 4H), 7.43−7.39 (m, 2H), 7.29−

7.27 (m, 1H), 7.09 (d, J = 8.4 Hz, 2H), 5.94 (d, J = 6.0 Hz, 1H), 5.48

H O (0.2 mL) was stirred at 20 °C for 17 h. The mixture was

2

(

d, J = 6.0 Hz, 1H), 5.35−5.32 (m, 1H), 5.20 (d, J = 4.8 Hz, 1H),

concentrated and puriﬁed by reversed-phase HPLC [TFA condition:

Agela ASB (150 × 25 mm, 5 μm) column, elution of acetonitrile/

water (0.1% TFA)] and then lyophilized to give N-(4-

((((2R,3S,4R,5R)-3,4-dihydroxy-5-(6-(4-(4-(triﬂuoromethyl)phenyl)-

piperazin-1-yl)-9H-purin-9-yl)tetrahydrofuran-2-yl)methyl)amino)-

phenyl)methanesulfonamide 21 (15.03 mg, 61.2% yield) as a yellow

4

3

.59 (d, J = 5.2 Hz, 1H), 4.40 (br s, 4H), 4.16 (d, J = 3.6 Hz, 1H),

.97 (d, J = 2.8 Hz, 1H), 3.70−3.67 (m, 1H), 3.58−3.56 (m, 1H),

+

.35 (s, 4H). MS (ESI) m/z: 489.2 [M + H] .

N-(((2R,3S,4R,5R)-3,4-Dihydroxy-5-(6-(4-(4-(triﬂuoromethyl)-

phenyl)piperazin-1-yl)-9H-purin-9-yl)tetrahydrofuran-2-yl)methyl)-

4

R,6aR)-2,2-dimethyl-6-(6-(4-(4-(triﬂuoromethyl)phenyl)piperazin-1-

yl)-9H-purin-9-yl)tetrahydrofuro[3,4-d][1,3]dioxol-4-yl)methyl)-4-

methylbenzenesulfonamide D3 (72 mg, 0.090 mmol) in THF (1 mL)

-methylbenzenesulfonamide (19). To a solution of N-(((3aR,4R,6-

1

solid; LC/MS purity of 96%; Scheme S4 Supporting Information . H

NMR (CD OD, 400 MHz): δ 8.30 (d, J = 9.2 Hz, 2H), 7.52 (d, J =

3

7

.2 Hz, 2H), 7.28−7.08 (m, 6H), 6.08 (br s, 1H), 4.58−4.26 (m, 6H),

3.89 (br s, 1H), 3.76−3.49 (m, 5H), 2.97 (s, 3H). MS (ESI) m/z:

and H O (0.1 mL) was added TFA (1 mL, 12.98 mmol). The mixture

+

2

649.2 [M + H] .

was stirred at 15 °C for 17 h and then DMF was added (2 mL), and

the mixture was puriﬁed by reversed-phase HPLC [TFA condition:

Agela ASB (150 × 25 mm, 5 μm) column, elution of acetonitrile/

water (0.1% TFA)] and then lyophilized to give N-(((2R,3S,4R,5R)-

3

9

mide (31.74 mg, 38.5% yield) as a white solid; LC/MS purity of 99%;

Scheme S1 Supporting Information . H NMR (DMSO-d , 400

MHz): δ 8.38 (s, 1H), 8.27−8.21 (m, 2H), 7.65 (d, J = 8.8 Hz, 2H),

(

2R,3R,4S,5R)-2-(6-(4-([1,1′-Biphenyl]-4-yl)piperazin-1-yl)-9H-

purin-9-yl)-5-((benzyloxy)methyl)tetrahydrofuran-3,4-diol (23).

The solution of 6-(4-([1,1′-biphenyl]-4-yl)piperazin-1-yl)-9-

(

( 3 a R , 4 R , 6 R , 6 a R ) - 6 - ( ( b e n z y l o x y ) m e t h y l ) - 2 , 2 -

,4-dihydroxy-5-(6-(4-(4-(triﬂuoromethyl)phenyl)piperazin-1-yl)-

H-purin-9-yl)tetrahydrofuran-2-yl)methyl)-4-methylbenzenesulfona-

dimethyltetrahydrofuro[3,4-d][1,3]dioxol-4-yl)-9H-purine B5 (700

mg, 1.131 mmol) and con. HCl (1 mL, 4.00 mmol) in MeOH (5

mL) was stirred at 25 °C for 1 h. Then, the mixture was neutralized

with ammonium hydroxide (5 mL), concentrated, puriﬁed by

reversed-phase HPLC [neutral condition: Waters Xtimate C18 (150

1

6

7

2

4

.52 (d, J = 8.4 Hz, 2H), 7.35 (d, J = 8.4 Hz, 2H), 7.11 (d, J = 8.4 Hz,

H), 5.85 (d, J = 5.6 Hz, 1H), 4.76 (t, J = 5.6 Hz, 2H), 4.36 (br s,

×

25 mm, 5 μm) column, elution of acetonitrile/water (10 mM

NH HCO )] to give(2R,3R,4S,5R)-2-(6-(4-([1,1′-biphenyl]-4-yl)-

piperazin-1-yl)-9H-purin-9-yl)-5-((benzyloxy)methyl)-

tetrahydrofuran-3,4-diol 23 (600 mg, 88% yield) as a white solid; LC/

4

3

H), 4.07−3.96 (m, 2H), 3.47 (br s, 4H), 3.07−2.97 (m, 1H), 2.34

+

(s, 3H). MS (ESI) m/z: 634.2 [M + H] .

(

2R,3S,4R,5R)-2-(((4-Nitrophenyl)amino)methyl)-5-(6-(4-(4-

1

MS purity of 96%; Scheme S5 Supporting Information . H NMR

(

triﬂuoromethyl)phenyl)piperazin-1-yl)-9H-purin-9-yl)-

(

DMSO-d , 400 MHz): δ 8.37 (s, 1H), 8.29 (s, 1H), 7.41−7.31 (m,

6

tetrahydrofuran-3,4-diol (20). The solution of N-(((3aR,4R,6-

R,6aR)-2,2-dimethyl-6-(6-(4-(4-(triﬂuoromethyl)phenyl)piperazin-1-

yl)-9H-purin-9-yl)tetrahydrofuro[3,4-d][1,3]dioxol-4-yl)methyl)-4-

nitroaniline D4 (50 mg, 0.078 mmol) and TFA (0.5 mL, 6.49 mmol)

8

5

4

H), 7.09 (d, J = 8.4 Hz, 2H), 5.97 (d, J = 6.0 Hz, 1H), 5.55 (d, J =

.6 Hz, 1H), 5.31 (d, J = 5.2 Hz, 1H), 4.58−4.54 (m, 3H), 4.40 (br s,

H), 4.22−4.07 (m, 2H), 3.74−3.67 (m, 1H), 3.31 (s, 4H). MS (ESI)

+

m/z: 579.2 [M + H] .

in THF (1.0 mL) and H O (0.2 mL) was stirred at 20 °C for 17 h,

and then, it was concentrated and puriﬁed by reversed-phase HPLC

2

(

2R,3S,4R,5R)-2-(((4-Fluorophenyl)amino)methyl)-5-(6-(4-(4-

triﬂuoromethyl)phenyl)piperazin-1-yl)-9H-purin-9-yl)-

tetrahydrofuran-3,4-diol (24). The mixture of N-(((3aR,4R,6R,6aR)-

,2-dimethyl-6-(6-(4-(4- (triﬂuoromethyl)phenyl)piperazin-1-yl)-9H-

(

[

TFA condition: Agela ASB (150 × 25 mm, 5 μm) column, elution of

acetonitrile/water (0.1% TFA)] and then lyophilized to give

2

(

2R,3S,4R,5R)-2-(((4-nitrophenyl)amino)methyl)-5-(6-(4-(4-

triﬂuoromethyl)phenyl)piperazin-1-yl)-9H-purin-9-yl)-

purin-9-yl)tetrahydrofuro[3,4-d][1,3]dioxol-4-yl)methyl)-4-ﬂuoroani-

line D9 (200 mg, 0.228 mmol) and con. HCl (0.5 mL, 2.0 mmol) in

MeOH (5 mL) was stirred at 25 °C for 1 h. Then, the mixture was

neutralized with ammonium hydroxide (5 mL), concentrated, puriﬁed

by reversed-phase HPLC (basic condition: Waters Xbridge Prep OBD

C18 (150 × 30 mm, 5 μm) column, elution of acetonitrile/water

tetrahydrofuran-3,4-diol 20 (24.24 mg, 51.5% yield) as a yellow solid;

1

LC/MS purity of 99%; Scheme S2 Supporting Information . H NMR

(

CD OD, 400 MHz): δ 8.29 (s, 1H), 8.10 (s, 1H), 7.83 (d, J = 9.2

3

Hz, 2H), 7.50 (d, J = 8.8 Hz, 2H), 7.09 (d, J = 8.8 Hz, 2H), 6.57 (d, J

9.2 Hz, 2H), 5.96 (d, J = 4.0 Hz, 1H), 4.80−4.74 (m, 1H), 4.51−

=

4

(

0.05% ammonia hydroxide)) to give (2R,3S,4R,5R)-2-(((4-

.32 (m, 5H), 4.28−4.21 (m, 1H), 3.71−3.62 (m, 1H), 3.56−3.40

+

ﬂuorophenyl)amino)methyl)-5-(6-(4-(4-(triﬂuoromethyl)phenyl)-

piperazin-1-yl)-9H-purin-9-yl)tetrahydrofuran-3,4-diol 23 (43.09 mg,

(m, 5H). MS (ESI) m/z: 601.2 [M + H] .

N-(((2R,3S,4R,5R)-3,4-Dihydroxy-5-(6-(4-(4-(triﬂuoromethyl)-

3

2.6% yield) as a white solid; LC/MS purity of 99%; Scheme S6

phenyl)piperazin-1-yl)-9H-purin-9-yl)tetrahydrofuran-2-yl)methyl)-

benzamide (21). To a solution of N-(((3aR,4R,6R,6aR)-2,2-

dimethyl-6-(6-(4-(4-(triﬂuoromethyl)phenyl)piperazin-1-yl)-9H-

purin-9-yl)tetrahydrofuro[3,4-d][1,3]dioxol-4-yl)methyl)benzamide

1

Supporting Information . H NMR (CD OD, 400 MHz): δ 8.31 (s,

3

1H), 8.14 (s, 1H), 7.50 (d, J = 8.8 Hz, 2H), 7.10 (d, J = 8.8 Hz, 2H),

6.89−6.79 (m, 2H), 6.73−6.62 (m, 2H), 5.97 (d, J = 5.2 Hz, 1H),

4.81 (t, J = 5.6 Hz, 1H), 4.49−4.26 (m, 6H), 3.47−3.41 (m, 6H). MS

D2 (68 mg, 0.090 mmol) in THF (1.0 mL) and H O (0.1 mL) was

2

+

added TFA (1.0 mL, 12.98 mmol). The mixture was stirred at 15 °C

for 17 h, and then, it was concentrated under reduced pressure,

puriﬁed by reversed-phase HPLC (TFA condition: Agela ASB [150 ×

(ESI) m/z: 574.1 [M + H] .

Mtb Resistant Mutant Isolation. Isolation of Mtb H37Rv-

resistant mutants to iodotubercidin on a solid medium was performed

37

2

5 mm, 5 μm) column, elution of acetonitrile/water (0.1% TFA)]

and then lyophilized to give N-(((2R,3S,4R,5R)-3,4-dihydroxy-5-(6-

4-(4-(triﬂuoromethyl)phenyl)piperazin-1-yl)-9H-purin-9-yl)-

as described by Ioerger et al. Three mutants were subjected to

whole genome sequencing. Single nucleotide polymorphisms were

identiﬁed in Mtb adoK and conﬁrmed by sequencing from genomic

(

38

tetrahydrofuran-2-yl)methyl)benzamide 21 (25.7 mg, 46.4% yield) as

DNA. Resistant mutants were conﬁrmed on the solid medium by

conﬁrming the growth on plates containing 5× MIC and in liquid

a white solid; LC/MS purity of 95%; Scheme S3 Supporting

Information . H NMR (DMSO-d , 400 MHz): δ 8.68 (s, 1H), 8.47

1

39

medium using the Microplate Alamar Blue Assay (MABA).

6

M

DOI: 10.1021/acs.jmedchem.9b00020

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Cloning, Expression, and Puriﬁcation of Recombinant

MtbAdoK and hAdoK. The WT MtbAdoK gene (Rv2202c) was

ampliﬁed by the polymerase chain reaction from total genomic DNA

of M. tuberculosis H37Rv. The following oligonucleotides were used in

the polymerase chain reaction 5′-GGAATTCCATATGGTGAC-

GATCGCGGTAACC-3′ and 5′-CTTAAGCTTCTAGGCCAG-

CAC-3′, respectively. The ampliﬁed DNA fragment was digested

with NdeI and HindIII restriction enzymes (New England Biolabs)

and subcloned into the corresponding restriction sites of the pET28b

vector containing an N-terminal Tobacco Etch Virus cleavable His-

i

y

j

a

z

j

z

specific activity = j

dz/c

j

z

^εNADH × b

(

3)

k

{

where a is the change in absorbance over time, ε

is the millimolar

NADH

extinction coeﬃcient of NADH, b is the pathlength, d is the dilution

factor of the enzyme in the assay, and c is the concentration of enzyme

stock used for the assay.

Finally, the inhibition constant (K ) was determined by using the

Cheng−Prusoﬀ relationship for competitive inhibition (eq 4).

i

41,42

5

tag. hAdoK gene was ampliﬁed from clone HsCD00042641

i

y

j

[S]

z

j

z

(

DNASU plasmid repository) using the following synthetic

K = IC /j

z

i

50 j

z

j

^Km + 1

oligonucleotides in the polymerase chain reaction: 5′-GGAATTCCA-

TATGATGACGTCAGTCAGAGAAAATATTC-3′ and 5′-

CCCAAGCTTCTAGTGGAAGTCTGGC-3′. The ampliﬁed frag-

ment was subsequently cloned into the same pET28b using the

same procedures described above. In all cases, gene ﬁdelity was

conﬁrmed by DNA sequencing, and sequenced plasmids were used to

transform Escherichia coli BL21 (DE3) cells for protein expression.

For protein expression, cell cultures were grown in an LB medium at

k

{

(4)

In all cases, the compounds were serially diluted in 100.0% DMSO

and added to the respective enzymatic reaction to a ﬁnal

concentration of 2.5% DMSO.

Crystallization, Data Collection, and Crystal Structure

Determination. Puriﬁed MtbAdoK was concentrated to 18 mg/

mL before crystallization trials. All cocrystals were obtained by the

vapor diﬀusion method. MtbAdoK-2, 3, 4, and 7 cocrystals were

achieved by mixing 2.0 μL of protein solution, preincubated with 4.0

mM of the compound for 1.0 h at 25.0 °C, with 1.0 μL of 100.0 mM

HEPES pH 7.5, 2.0 M ammonium sulfate and 2.0% PEG 400.

MtbAdoK-5 cocrystals were obtained by mixing 2.0 μL of protein

solution, preincubated with 4.0 mM of 5 for 1.0 h at 25.0 °C, with 1.0

μL of 100.0 mM HEPES pH 7.5 and 1.2 M sodium citrate tribasic

dihydrate. MtbAdoK-6 cocrystals were obtained by mixing 2.0 μL of

protein solution, preincubated with 4.0 mM of the compound for 1.0

h at 25.0 °C, with 1.0 μL of 5.0 M sodium formate. Finally,

MtbAdoK-17 cocrystals were obtained by mixing 2.0 μL of protein

solution, preincubated with 2.0 mM 17 for 1 h at 25 °C, with 1.0 μL

of 100 mM Bis−Tris pH 6.5, 2.0 M ammonium sulfate and 2.0% PEG

3

7.0 °C. Cells were induced with 0.5 mM isopropyl β-D-1-

thiogalactopyranoside (IPTG) when the cell density reached A₆₀₀

.6−1.0. Cell cultures were incubated for 18 h at 18.0 °C before

harvesting.

≈

0

Harvested cells were lysed using a French press, and the lysate was

centrifuged at 17 000 rpm for 1 h. Recombinant MtbAdoK and

hAdoK were puriﬁed by using a HisTrap HP Nickel column (GE

Healthcare). Puriﬁcation buﬀers A and B contained, 50.0 mM

HEPES, pH 7.5, 500.0 mM NaCl, 500.0 mM imidazole (buﬀer B

only), and 5.0% glycerol. For crystallization studies, MtbAdoK was

dialyzed in 20.0 mM HEPES, pH 7.5, 50.0 mM NaCl, 2.0 mM DTT,

and 5.0% glycerol. For enzymatic assays, the proteins were dialyzed in

5

0.0 mM HEPES, 50.0 mM NaCl, 100.0 mM KCL, 4.0 mM DTT,

4

00. In all cases, crystals were grown at 20.0 °C. Prior to data

and 20.0% glycerol. Finally, the proteins were aliquoted and stored in

collection, crystals were cryoprotected with Paratone (Hampton

research) and ﬂash frozen in liquid nitrogen. X-ray diﬀraction data

were collected at Argonne’s National Lab Advanced Photon Source

beamlines 19ID and 23ID. Diﬀraction data were indexed, scaled, and

−

80.0 °C for subsequent crystallization and enzymatic assays.

Enzymatic Assay, IC₅₀Determination, Steady-State Kinetics,

and K Determination. Compounds were tested against MtbAdoK

i

using the pyruvate kinase−lactate dehydrogenase coupled assay

system in a Cary100 UV−Vis spectrophotometer by monitoring the

43

integrated using HKL2000. Initial phases were obtained by MR in

MOLREP using the high-resolution structure of the MtbAdoK-1

complex with PDB accession code 2PKM. Reﬁnement was

performed in PHENIX followed by iterative runs of inspection and

manual modiﬁcation using coot.

were created in ELBOW from the PHENIX suite and ﬁtted into the

electron density using COOT. Images and ﬁgures were rendered

using Molsoft ICM, Chimera and PyMOL.

Antimycobacterial Assay. Antitubercular testing and EC₅₀

40

decrease in absorbance of NADH at 340 nm. The reaction was

started by the addition of 60 nM of enzyme into a ﬁnal volume of 200

μL master mix containing 50.0 mM HEPES pH 7.5, 50.0 mM KCl,

44

45,46

Ligand model and restraint ﬁles

6

.0 mM MgCl (4.0 mM MgCl for hAdoK), 3.0 mM ATP (2.0 mM

2

ATP for hAdoK), 200.0 μM NADH, 1.0 mM phosphoenolpyruvate,

.0 mM DTT, 12.0 U/mL pyruvate kinase, 12.0 U/mL lactate

1

47−49

dehydrogenase, and 15.0 μM adenosine. IC₅₀values for each

compound were determined by varying the concentration of the

inhibitor at ﬁxed concentrations of the enzyme and by ﬁtting the

dose−response data into the four-parameter logistic curve (eq 1)

model of GraphPad prism 7.02, as follows

determination was performed using the MABA assay in a 96-well

50,51

2

format as previously described.

was grown in 7H9 media supplemented with OADC (Middlebrook),

.5% dextrose, 0.085% NaCl, 0.05% Tyloxapol (Sigma), 0.25 μg/mL

Starter culture of Mtb mc 7000

0

(

logIC₅₀−I)H

malachite green (Sigma), and 25 μg/mL pantothenate. Once cells

reached an optical density of OD₆₀₀≈ 1.5, cells were diluted to an

OD₆₀₀of 0.01 in the same media composition without OADC.

Compounds were serially diluted in 100.0% DMSO and added to the

cells to a ﬁnal concentration of DMSO of 2.5%. Plates were incubated

for 6 days before staining with resazurin (Sigma). After staining, plates

were incubated for 2 additional days for developing. Finally,

developed plates were read using a POLARstar Omega spectropho-

tometer (BMG Labtech) and by monitoring the ﬂuorescence

emission of resazurin (excitation = 570.0 nm, emission = 585.0

nm). In all cases, rifampicin was used as a negative control. The anti-

Tb activity was determined as a percent control, and EC₅₀values for

each compound were determined by utilizing eq 1.

Y = Y_m_i_n+ (Y_m_a_x− Y_m_i_n)/(1 + 10

)

(1)

where I is the logarithm of inhibitor concentration, H is the Hill slope

and Y, Y_m_a_x, and Y_m_i_nare the speciﬁc activity, maximum speciﬁc

activity, and minimum speciﬁc activity, respectively. Kinetic assays for

MtbAdoK were performed essentially as described above with the

following exceptions: the master mix contained 60.0 nM MtbAdoK,

and the reaction was started by the addition of varying concentrations

of adenosine in the presence of constant concentrations of compound

0.0, 20.0, and 40.0 nM). Kinetic data were obtained by ﬁtting the

initial velocity data into GraphPad Prism 7.02 nonlinear regression

function of Michaelis−Menten model (eq 2), as follows

(

V_o= (Y_m_a_x)[S]/(K_m+ [S])

(2)

Human Dermal Fibroblast Cytotoxicity Assay. HDFs were

purchased from ATCC (Manassas, VA). HDF cells were cultured in

DMEM (Lonza) media supplemented with 10.0% fetal bovine serum

(Lonza) and penicillin/streptomycin (Lonza). For the cytotoxicity

assay, compound stocks were serially diluted in phosphate-buﬀered

saline plus 10.0% DMSO. On the day of assay, HDF cells were

trypsinized, counted, and resuspended at a concentration of 64 000

where V is the initial velocity, Y is the maximum speciﬁc activity, S

o

max

is the substrate concentration, and K_mis the Michaelis−Menten

constant.

Speciﬁc activity values for kinetics and IC₅₀measurements were

determined using (eq 3)

N

DOI: 10.1021/acs.jmedchem.9b00020

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

cells/mL in media. Cells were plated, overlaid with the compound

serial dilutions, and incubated at 37.0 °C. After 48.0 h, resazurin dye

was added, and the assay plates cultured for another 24.0 h. On the

next day, the absorbance of the resazurin was measured on a

microplate reader to assess cell death. Cytotoxicity was determined as

a percent control, and CC₅₀values for each compound were

determined by utilizing eq 1.

Pharmacokinetic and Tolerability Studies. ***All animal

experiments described in this manuscript followed protocols approved

by the Texas A&M University Institutional Animal Care and Use

Committee under Animal Use Protocol “Preliminary in vivo PK/Tox

on newly developed drug compounds” #2017-0368.***

6C67, MtbAdoK-3 complex-PDB ID 6C9N, MtbAdoK-4

complex-PDB ID 6C9P, MtbAdoK-5 complex-PDB ID

6C9Q, MtbAdoK-6 complex-PDB ID 6C9R, MtbAdoK-7

complex-PDB ID 6C9V, MtbAdoK-17 complex-PDB ID 6C9S.

AUTHOR INFORMATION

Corresponding Authors

*E-mail: david_olsen@merck.com. Phone: 215-652-5250

(D.B.O.).

■

*

E-mail: sacchett@tamu.edu. Phone: (979) 845-8548 (J.C.S.).

ORCID

Eight-week old female ND4 Swiss Webster mice (∼28 g each)

Envigo, Indianapolis, IN) were used for PK and combination

Tanya Parish: 0000-0001-7507-0423

James C. Sacchettini: 0000-0001-5767-2367

Present Addresses

(

tolerability-snapshot PK studies. Compound 18 was formulated in 5%

DMSO, 95% canola oil and administered as a single 100 μL oral

−

1

gavage at 10 mg·kg for PK and as a 200 μL oral gavage at 50, 100, or

#

Lilly China R&D Center, Eli Lilly and Company, Shanghai

1

2

00 mg·kg⁻QD × 3 days for tolerability-snapshot PK. Mice were

China.

anesthetized using isoﬂurane and 50 μL collected at 0, 0.25, 0.5, 1, 2,

and 4 h survival bleed times for traditional PK or 0, 1, 4, and 24 h for

tolerability-snapshot PK studies after the ﬁrst administered dose.

Around 200 μL was collected at 6, 8, and 24 h terminal bleed times

for the traditional PK. Survival bleeds were drawn from the facial vein

¶

GSK, Swedeland Road, King of Prussia, PA.

Funding

This work was supported by grants from the Bill and Melinda

Gates Foundation (OPP1024055), Welch Foundation (A-

and terminal bleeds via cardiac puncture following CO asphyx-

2

0

015), NIAID-NIH (TB Structural Genomics grant

35

iation. Three mice were used per time point.

P01A1095208) to J.C.S. and the Bill and Melinda Gates

Foundation (OPP1024038) to T.P.

Blood samples were centrifuged (3000g, 15 min) for plasma

separation. Aliquots of plasma (10 μL) were extracted with two 500

μL equivalents of methanol containing 0.1% formic acid to ensure

maximum recovery of compound 18 from plasma. The supernatant

was collected from precipitated protein pellet following centrifugation

Notes

The authors declare no competing ﬁnancial interest.

(

3000g, 5 min) and evaporated to dryness using a temperature-

ACKNOWLEDGMENTS

■

controlled Eppendorf Vacufuge at 30 °C for 4 h. The dry samples

The authors would like to acknowledge the Argonne National

Lab staﬀ at beamlines 19-ID and 23-ID, the TB Drug

Accelerator Program (TBDA), the Bill and Melinda Gates

Foundation and the Welch Foundation for their ﬁnancial and/

or intellectual support.

were reconstituted with 100 μL of methanol (1:10 original analyte

_d_i_l_u_t_i_o_n_f_a_c_t_o_r₎_s_p_i_k_e_d_w_i_t_h₂₅₀_n_g_·_m_L₋1 verapamil (Sigma-Aldrich)

as an internal standard and subjected to LC/MS analysis on a Bruker

Daltronics micrOTOF-Q II mass spectrometer coupled with an

Agilent 1200 Inﬁnity series HPLC with temperature controlled

autosampler (24 °C) and photodiode array detector. Standard

solutions of compound 18 ranging from 1000 to 7.8125 ng·mL

spiked into mouse control plasma was also analyzed to generate a

calibration curve corrected for matrix eﬀects. A linear regression

generated from calibration samples was used to quantify samples.

The LOD of the LC/MS method was approximately 2 ng·mL . A 4.6

−

1

ABBREVIATIONS

■

Mtb, Mycobacterium tuberculosis; MtbAdoK, Mycobacterium

tuberculosis adenosine kinase; hAdoK, human adenosine

kinase; vdW, van der Waals; HDF, human dermal ﬁbroblasts;

MR, molecular replacement; TBSCl, tert-butyldimethylsilyl

chloride; EtOH, ethanol; DIEA, N,N-diisopropylethylamine;

34

−1

×

100 mm Kinetex 2.6 μm EVO C18 100 Å column at a ﬂow rate of

.5 mL·min⁻was used in the analysis. The mobile phase consisted of

water containing 0.1% formic acid as solvent A and acetonitrile with

.1% formic acid as solvent B. The gradient conditions were

1

0

ArB(OH) , aryl boronic acids; XPhos Pd G2, chloro(2-

2

0

dicyclohexylphosphino-2′,4′,6′-triisopropyl-1,1′-biphenyl)[2-

maintained as follows: 90% A, 10% B to 100% B in 8 min; 100%

held for 4 min as a washing step; 100% B back to 90% A in 2 min;

9

Each analytical run was automatically calibrated using a secondary line

injection of 10 μL of sodium acetate external standard at 13 min. The

injection volume of the analyte was 10 μL, and MS was operated in

positive mode with electrospray ionization at source. Mass spectra

were monitored in a range of 50−1000m/z. Data were processed

using DataAnalysis 4.1 and QuantAnalysis (Bruker Daltronics).

(

2′-amino-1,1′-biphenyl)]palladium(II); MeOH, methanol;

RuPhos, 2-dicyclohexylphosphino-2′,6′-diisopropoxybiphenyl;

0% A, 10% B held for 3 min to reset and equilibrate the column.

AUC, area under the curve; PK, pharmacokinetic; C

,

max

^m^a^xⁱ^m^u^m^c^oⁿ^c^eⁿ^t^r^a^tⁱ^oⁿ^;^tmax, time at maximum concentration.

REFERENCES

■

(1) WHO. Global Tuberculosis Report, 2016.

(2) Ducati, R. G.; Breda, A.; Basso, L. A.; Santos, D. S. Purine

Salvage Pathway in Mycobacterium tuberculosis. Curr. Med. Chem 2011,

18, 1258−1275.

3) Villela, A. D.; Sanchez-Quitian, Z. A.; Ducati, R. G.; Santos, D.

S.; Basso, L. A. Pyrimidine Salvage Pathway in Mycobacterium

tuberculosis. Curr. Med. Chem 2011, 18, 1286−1298.

4) Long, M. C.; Escuyer, V.; Parker, W. B. Identification and

Characterization of a Unique Adenosine Kinase from Mycobacterium

tuberculosis. J. Bacteriol. 2003, 185, 6548−6555.

(5) Reddy, M. C. M.; Palaninathan, S. K.; Shetty, N. D.; Owen, J. L.;

Watson, M. D.; Sacchettini, J. C. High Resolution Crystal Structures

ofMycobacterium tuberculosisAdenosine Kinase. J. Biol. Chem. 2007,

ASSOCIATED CONTENT

■

(

*

S

Supporting Information

The Supporting Information is available free of charge on the

ACS Publications website at DOI: 10.1021/acs.jmed-

chem.9b00020 .

(

Schemes and reagents for chemical synthesis of

intermediates and supporting ﬁ gures and tables (PDF)

Molecular formula strings (CSV )

Accession Codes

Authors will release the atomic coordinates and experimental

data upon article publication: MtbAdoK-2 complex-PDB ID

2

82, 27334−27342.

(6) Park, J.; Gupta, R. S. Adenosine kinase and ribokinase - the RK

family of proteins. Cell. Mol. Life Sci. 2008, 65, 2875−2896.

O

DOI: 10.1021/acs.jmedchem.9b00020

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

(

7) Schumacher, M. A.; Scott, D. M.; Mathews, I. I.; Ealick, S. E.;

(23) Zemlicka, J.; Endo, T. O6-(4-Nitrophenyl)inosine and

-Guanosine as Chromogenic Substrates for Adenosine Deaminase.

Nucleosides Nucleotides 1996, 15, 619−629.

Roos, D. S.; Ullman, B.; Brennan, R. G. Crystal Structures of

Toxoplasma gondii Adenosine Kinase Reveal a Novel Catalytic

Mechanism and Prodrug Binding. J. Mol. Biol. 2000, 298, 875−893.

(24) Camp, D.; Li, Y.; McCluskey, A.; Moni, R. W.; Quinn, R. J.

Diimidazo[1,2-c:4′,5′-e]pyrimidines: N6-N1 conformationally re-

stricted adenosines. Bioorg. Med. Chem. Lett. 1998, 8, 695−698.

(25) Katagiri, N.; Matsuhashi, Y.; Kokufuda, H.; Takebayashi, M.;

Kaneko, C. A Highly Efficient Synthesis of the Antiviral Agent

(+)-cyclaradine Involving the Regioselective Cleavage of Epoxide by

Neighboring Participation. Tetrahedron Lett. 1997, 38, 1961−1964.

(

8) Mathews, I. I.; Erion, M. D.; Ealick, S. E. Structure of Human

Adenosine Kinase at 1.5 Å Resolution†,‡. Biochemistry 1998, 37,

5607−15620.

9) Romanello, L.; Bachega, J. F. R.; Cassago, A.; Brandao

1

(

̃

-Neto, J.;

DeMarco, R.; Garratt, R. C.; Pereira, H. D. M. Adenosine kinase

fromSchistosoma mansoni: structural basis for the differential

incorporation of nucleoside analogues. Acta Crystallogr., Sect. D:

Biol. Crystallogr. 2013, 69, 126−136.

(

26) Itoh, T.; Sugawara, T.; Mizuno, Y. A Novel Synthesis of 1-

Deazaadenosine. Heterocycles 1982, 17, 305−309.

27) Wilkens, S. Structure and Mechanism of ABC transporters.

F1000Prime Rep. 2015, 7, 14.

28) Braibant, M.; Gilot, P.; Content, J. The ATP Binding Cassette

ABC) Transport Systems of Mycobacterium tuberculosis. FEMS

Microbiol. Rev. 2000, 24, 449−467.

29) Sandhu, P.; Akhter, Y. The Drug Binding Sites and Transport

(

10) Griffin, J. E.; Gawronski, J. D.; DeJesus, M. A.; Ioerger, T. R.;

Akerley, B. J.; Sassetti, C. M. High-Resolution Phenotypic Profiling

Defines Genes Essential for Mycobacterial Growth and Cholesterol

Catabolism. PLoS Pathog. 2011, 7, No. e1002251.

(

11) Zhang, Y. J.; Reddy, M. C.; Ioerger, T. R.; Rothchild, A. C.;

(

Dartois, V.; Schuster, B. M.; Trauner, A.; Wallis, D.; Galaviz, S.;

Huttenhower, C.; Sacchettini, J. C.; Behar, S. M.; Rubin, E. J.

Tryptophan Biosynthesis Protects Mycobacteria from CD4 T-Cell-

Mediated Killing. Cell 2013, 155, 1296−1308.

Mechanism of the RND Pumps from Mycobacterium tuberculosis:

Insights from Molecular Dynamics Simulations. Arch. Biochem.

Biophys. 2016, 592, 38−49.

(30) De Rossi, E.; Aínsa, J. A.; Riccardi, G. Role of Mycobacterial

(

12) Long, M. C.; Parker, W. B. Structure-activity relationship for

Efflux Transporters in Drug Resistance: an Unresolved Question.

nucleoside analogs as inhibitors or substrates of adenosine kinase

from Mycobacterium tuberculosis. Biochem. Pharmacol. 2006, 71,

FEMS Microbiol. Rev. 2006, 30, 36−52.

(31) Balganesh, M.; Dinesh, N.; Sharma, S.; Kuruppath, S.; Nair, A.

1

(

671−1682.

V.; Sharma, U. Efflux Pumps of Mycobacterium tuberculosis Play a

Significant Role in Antituberculosis Activity of Potential Drug

Candidates. Antimicrob. Agents Chemother. 2012, 56, 2643−2651.

13) Long, M. C.; Shaddix, S. C.; Moukha-Chafiq, O.; Maddry, J. A.;

Nagy, L.; Parker, W. B. Structure-activity relationship for adenosine

kinase from Mycobacterium tuberculosis. Biochem. Pharmacol. 2008,

(32) Pule, C. M.; Sampson, S. L.; Warren, R. M.; Black, P. A.; van

7

5, 1588−1600.

14) Malnuit, V.; Slavet

M.; Stolaríkova, J.; Snas l, J.; Pichova,

Het)aryl-7-deazapurine Ribonucleosides: Synthesis, Inhibition of

Adenosine Kinases, and Antimycobacterial Activity. ChemMedChem

015, 10, 1079−1093.

15) Snas l, J.; Naus,

J.; Bourderioux, A.; Duse

Zabranska, H.; Pohl, R.; Konecn

Hajduch, M.; Rezac

Helden, P. D.; Victor, T. C.; Louw, G. E. Efflux Pump Inhibitors:

Targeting Mycobacterial Efflux Systems to Enhance TB Therapy. J.

Antimicrob. Chemother. 2016, 71, 17−26.

(

̌

ínska,

́

L. P.; Naus,

I.; Hocek, M. 2-Substituted 6-

P.; Dzubak, P.; Hajduch,

̌ ́

̌

́

e

̌

́

(

(33) Park, J.; Singh, B.; Gupta, R. S. Mycobacterial Adenosine

Kinase is not a Typical Adenosine Kinase. FEBS Lett. 2009, 583,

231−2236.

34) Ahmed, T. R.; Shashank, R. Application of LC-MS in

2

(

2

(

́

e

̌

P.; Dostal

k, M.; Dvora

y, P.; Dzu

P.; Veverka, V.; Hocek, M.; Pichova,

́

, J.; Hnízda, A.; Fanfrlík, J.; Brynda,

̌

k

́

ova,

́

H.; Stolarí

̌

kova,

́

J.;

Supporting PK/PD Studies During Drug Discovery and Develop-

́

̌

́

̌

bak

́

, P.; Votruba, I.;

ment. Pharm. Res. 2012, 5, 2514−2526.

̌

́

o

̌

va,

́

I.

(

35) Diehl, K.-H.; Hull, R.; Morton, D.; Pfister, R.;

Structural Basis for Inhibition of Mycobacterial and Human

Adenosine Kinase by 7-Substituted 7-(Het)aryl-7-deazaadenine

Ribonucleosides. J. Med. Chem. 2014, 57, 8268−8279.

Rabemampianina, Y.; Smith, D.; Vidal, J.-M.; Vorstenbosch, C. V.

D. A Good Practice Guide to the Administration of Substances and

Removal of Blood, Including Routes and Volumes. J. Appl. Toxicol.

(16) Parker, W. B.; Barrow, E. W.; Allan, P. W.; Shaddix, S. C.; Long,

2

(

001, 21, 15−23.

M. C.; Barrow, W. W.; Bansal, N.; Maddry, J. A. Metabolism of 2-

methyladenosine in Mycobacterium tuberculosis. Tuberculosis 2004, 84,

36) Liu, B.; Chang, J.; Gordon, W. P.; Isbell, J.; Zhou, Y.; Tuntland,

T. Snapshot PK: a Rapid Rodent in vivo Preclinical Screening

3

(

27−336.

Approach. Drug Discovery Today 2008, 13, 360−367.

17) Chen, C.-K.; Barrow, W. W.; Parker, W. B.; Barrow, E. W.;

(37) Ioerger, T. R.; O’Malley, T.; Liao, R.; Guinn, K. M.; Hickey, M.

Maddry, J. A.; Suling, W. J.; Allan, P. W.; Bansal, N. The Metabolism

of 2-methyladenosine in Mycobacterium smegmatis. Microbiology 2002,

J.; Mohaideen, N.; Murphy, K. C.; Boshoff, H. I. M.; Mizrahi, V.;

Rubin, E. J.; Sassetti, C. M.; Barry, C. E., III; Sherman, D. R.; Parish,

T.; Sacchettini, J. C. Identification of New Drug Targets and

Resistance Mechanisms in Mycobacterium tuberculosis. PLoS One

1

(

48, 289−295.

18) Barrow, E. W.; Westbrook, L.; Bansal, N.; Suling, W. J.;

Maddry, J. A.; Parker, W. B.; Barrow, W. W. Antimycobacterial

Activity of 2-methyl-adenosine. J. Antimicrob. Chemother. 2003, 52,

2

(

013, 8, No. e75245.

38) Ioerger, T. R.; Feng, Y.; Ganesula, K.; Chen, X.; Dobos, K. M.;

8

(

01−808.

Fortune, S.; Jacobs, W. R.; Mizrahi, V.; Parish, T.; Rubin, E.; Sassetti,

C.; Sacchettini, J. C. Variation among Genome Sequences of H37Rv

Strains of Mycobacterium tuberculosis from Multiple Laboratories. J.

Bacteriol. 2010, 192, 3645−3653.

19) Massillon, D.; Stalmans, W.; van de Werve, G.; Bollen, M.

Identification of the Glycogenic Compound 5-iodotubercidin as a

General Protein Kinase Inhibitor. Biochem. J. 1994, 299, 123−128.

(

Schanzer, J. M.; Wiesner, J. B.; Erion, M. D. Adenosine Kinase

Inhibitors. 1. Synthesis, Enzyme Inhibition, and Antiseizure Activity of

20) Ugarkar, B. G.; DaRe, J. M.; Kopcho, J. J.; Browne, C. E.;

(39) Cho, S.; Lee, H. S.; Franzblau, S. Microplate Alamar Blue Assay

(MABA) and Low Oxygen Recovery Assay (LORA) for Mycobacte-

rium tuberculosis. In Mycobacteria Protocols; Parish, T., Roberts, D. M.,

Eds.; Springer New York: New York, NY, 2015; pp 281−292.

(40) Blondin, C.; Serina, L.; Wiesmuller, L.; Gilles, A. M.; Barzu, O.

Improved Spectrophotometric Assay of Nucleoside Monophosphate

Kinase Activity Using the Pyruvate Kinase/Lactate Dehydrogenase

Coupling System. Anal. Biochem. 1994, 220, 219−221.

(41) Cer, R. Z.; Mudunuri, U.; Stephens, R.; Lebeda, F. J. IC50-to-

Ki: a web-based tool for converting IC50 to Ki values for inhibitors of

enzyme activity and ligand binding. Nucleic Acids Res. 2009, 37,

W441−W445.

5

(

-Iodotubercidin Analogues. J. Med. Chem. 2000, 43, 2883−2893.

21) Loomis, C. R.; Bell, R. M. Sangivamycin, a Nucleoside

Analogue, is a Potent Inhibitor of Protein Kinase C. J. Biol. Chem.

988, 263, 1682−1692.

22) Muchmore, S. W.; Smith, R. A.; Stewart, A. O.; Cowart, M. D.;

1

(

Gomtsyan, A.; Matulenko, M. A.; Yu, H.; Severin, J. M.; Bhagwat, S.

S.; Lee, C.-H.; Kowaluk, E. A.; Jarvis, M. F.; Jakob, C. L. Crystal

Structures of Human Adenosine Kinase Inhibitor Complexes Reveal

Two Distinct Binding Modes. J. Med. Chem. 2006, 49, 6726−6731.

P

DOI: 10.1021/acs.jmedchem.9b00020

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

(42) Yung-Chi, C.; Prusoff, W. H. Relationship between the

inhibition constant (KI) and the concentration of inhibitor which

causes 50 per cent inhibition (I50) of an enzymatic reaction. Biochem.

Pharmacol. 1973, 22, 3099−3108.

(43) Otwinowski, Z.; Minor, W. Processing of X-ray Diﬀraction Data

Collected in Oscillation Mode. In Methods in Enzymology; Carter, C.

W., Jr., Sweet, R. M., Eds.; Academic Press: New York, 1997; Vol.

2

(

76: Macromolecular Crystallography, Part A, pp 307−326.

44) Collaborative Computational Project, Number 4 The CCP4

Suite: Programs for Protein Crystallography. Acta Crystallogr., Sect. D:

Biol. Crystallogr. 1994, 50 (), 760. DOI: 10.1107/

s0907444994003112

(45) Adams, P. D.; Grosse-Kunstleve, R. W.; Hung, L.-W.; Ioerger,

T. R.; McCoy, A. J.; Moriarty, N. W.; Read, R. J.; Sacchettini, J. C.;

Sauter, N. K.; Terwilliger, T. C. PHENIX: Building New Software for

Automated Crystallographic Structure Determination. Acta Crystal-

logr., Sect. D: Biol. Crystallogr. 2002, 58, 1948−1954.

(

46) Emsley, P.; Lohkamp, B.; Scott, W. G.; Cowtan, K. Features

and development ofCoot. Acta Crystallogr., Sect. D: Biol. Crystallogr.

010, 66, 486−501.

47) Pettersen, E. F.; Goddard, T. D.; Huang, C. C.; Couch, G. S.;

2

(

Greenblatt, D. M.; Meng, E. C.; Ferrin, T. E. UCSF Chimera?A

visualization system for exploratory research and analysis. J. Comput.

Chem. 2004, 25, 1605−1612.

(

48) PyMOL The PyMOL Molecular Graphics System, version 1.8;

Schro

49) Abagyan, R.; Totrov, M.; Kuznetsov, D. ICM?A new method

̈

dinger, LLC.

(

for protein modeling and design: Applications to docking and

structure prediction from the distorted native conformation. J.

Comput. Chem. 1994, 15, 488−506.

(50) Franzblau, S. G.; Witzig, R. S.; McLaughlin, J. C.; Torres, P.;

Madico, G.; Hernandez, A.; Degnan, M. T.; Cook, M. B.; Quenzer, V.

K.; Ferguson, R. M.; Gilman, R. H. Rapid, Low-Technology MIC

Determination with Clinical Mycobacterium tuberculosis Isolates by

Using the Microplate Alamar Blue Assay. J. Clin. Microbiol 1998, 36,

3

(

62−366.

51) Krieger, I. V.; Freundlich, J. S.; Gawandi, V. B.; Roberts, J. P.;

Gawandi, V. B.; Sun, Q.; Owen, J. L.; Fraile, M. T.; Huss, S. I.;

Lavandera, J.-L.; Ioerger, T. R.; Sacchettini, J. C. Structure-Guided

Discovery of Phenyl-diketo Acids as Potent Inhibitors of M.

tuberculosis Malate Synthase. Chem. Biol. 2012, 19, 1556−1567.

Q

DOI: 10.1021/acs.jmedchem.9b00020

J. Med. Chem. XXXX, XXX, XXX−XXX

Article Doi

Structure-Guided Drug Design of 6-Substituted Adenosine Analogues as Potent Inhibitors of Mycobacterium tuberculosis Adenosine Kinase

DOI: 10.1021/acs.jmedchem.9b00020

Source and publish data:

Authors:

Article abstract of DOI:10.1021/acs.jmedchem.9b00020

Full text of DOI:10.1021/acs.jmedchem.9b00020

Products guided by the article

R&D Labs maybe for 1365531-08-3

Relevant to this article

Hot Product