South (S)- and north (N)-methanocarba-7-deazaadenosine analogues as inhibitors of human adenosine kinase

Article

pubs.acs.org/jmc

South (S)- and North (N)-Methanocarba-7-Deazaadenosine

Analogues as Inhibitors of Human Adenosine Kinase

Kiran S. Toti,^†Danielle Osborne,^‡Antonella Ciancetta,^†Detlev Boison,^‡and Kenneth A. Jacobson*^,†

^†Molecular Recognition Section, Laboratory of Bioorganic Chemistry, National Institute of Diabetes and Digestive and Kidney

Diseases, National Institutes of Health, Bldg. 8A, Rm. B1A-19, Bethesda, Maryland 20892-0810, United States

^‡Robert Stone Dow Neurobiology Laboratories, Legacy Research Institute, 1225 NE Second Avenue, Portland, Oregon 97232,

United States

S

* Supporting Information

ABSTRACT: Adenosine kinase (AdK) inhibitors raise endogenous adenosine levels, particularly in disease states, and have

potential for treatment of seizures, neurodegeneration, and inﬂammation. On the basis of the South (S) ribose conformation and

molecular dynamics (MD) analysis of nucleoside inhibitors bound in AdK X-ray crystallographic structures, (S)- and North (N)-

methanocarba (bicyclo[3.1.0]hexane) derivatives of known inhibitors were prepared and compared as human (h) AdK inhibitors.

5′-Hydroxy (34, MRS4202 (S); 55, MRS4380 (N)) and 5′-deoxy 38a (MRS4203 (S)) analogues, containing 7- and N⁶-NH

phenyl groups in 7-deazaadenine, robustly inhibited AdK activity (IC₅₀∼ 100 nM), while the 5′-hydroxy derivative 30 lacking the

phenyl substituents was weak. Docking in the hAdK X-ray structure and MD simulation suggested a mode of binding similar to

5′-deoxy-5-iodotubercidin and other known inhibitors. Thus, a structure-based design approach for further potency enhancement

is possible. The potent AdK inhibitors in this study are ready to be further tested in animal models of epilepsy.

modulating the concentration of adenosine in the brain, and

therefore the basal levels of activation of the ARs, is by

administering brain-penetrant human (h) AdK inhibitors. The

raised level of adenosine in the brain counteracts seizures by

activating the neuroprotective A₁AR and attenuates epilepsy

progression by decreasing S-adenosyl methionine dependent

DNA methylation as an epigenetic mechanism of action.^8,10

Increased DNA methylation is a pathological hallmark of

chronic epilepsy and associated with disease progression and

maintenance of the epileptic state.^5,11Eﬃcient transmethylation

reactions, for example, in the liver, require the removal of

adenosine by AdK,¹¹and this eﬀect has been demonstrated to

occur in the brain as well.^5,8Unlike adenosine itself, synthetic

A₁AR agonists, which are also proposed for seizure treat-

ment,^10,13would not inhibit DNA methylation. Overexpression

of AdK in the brain is both a result of astroglial activation and a

contributing factor to epileptic seizures. Thus, an inhibitor of

INTRODUCTION

■

Endogenous adenosine acts on its G protein-coupled receptors

(adenosine receptors, ARs)¹in the central nervous system to

suppress seizures²and pain³and to blunt the eﬀects of

ischemia.⁴In addition, adenosine has AR-independent

epigenetic eﬀects based on interactions with the trans-

methylation pathway.⁵Extracellular adenosine may originate

from intracellular pools or from the action of ectonucleotidases

on extracellular ATP.⁶There is a dynamic equilibrium between

extracellular adenosine levels and its intracellular content that is

mediated by either equilibrative (ENTs) or concentrative

(CNTs) transporters of nucleosides.⁷Within the brain, the

concentration of adenosine is largely under the control of

metabolic clearance through astrocytic adenosine kinase (AdK),

which converts adenosine to 5′-AMP.⁸AdK exists in two

isoforms derived from alternative splicing and alternative

promoter use. The short isoform AdK-S resides in the

cytoplasm, whereas the long isoform Adk-L is located in the

cell nucleus.⁹By inhibiting AdK, the adenosine concentration

can be exogenously raised. Thus, an indirect means of

Received: May 5, 2016

A

DOI: 10.1021/acs.jmedchem.6b00689

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Chart 1. Known Nucleoside and Nonnucleoside Inhibitors of hAdK That Have Been Examined in Models of Pain and/or

a

Seizures

a

Published potencies (IC₅₀values) for inhibition of the hAdK catalysis of the conversion of adenosine to AMP are given in nM.^7,29

AdK might have an advantage in seizure control because it

would combine a pharmacological mechanism (increased A₁AR

activation) with epigenetic mechanisms (decreased DNA

methylation) and might preferentially act on pathologically

increased AdK as opposed to normal baseline levels of the

enzyme. The advantages of “adenosine augmentation therapies”

for epilepsy and its associated comorbidities have already been

discussed.¹⁴⁻¹⁶

the AdK inhibitor 3 was found to promote rodent and porcine

islet β-cell replication, which suggests the possible application

of such inhibitors to the treatment of diabetes.³⁰However,

other, undesired eﬀects of the inhibitor 1a have been noted; it

seems to indirectly inhibit acetyl-CoA carboxylase to promote

oxidation of hepatic fatty acids and reduce de novo synthesis of

lipids and cholesterol, which raises the AMP/ATP ratio.³¹

Thus, there might be a need to increase selectivity for AdK

within this nucleoside series.

Several classes of AdK inhibitors have already been

introduced and explored for the treatment of seizures and

pain (Chart 1, 1−6).¹⁷⁻²¹One class consists of nucleosides

derived from the known AdK inhibitor 7-iodo-7-deazaadeno-

sine 1a, otherwise known as 5-iodotubercidin (5-IT). N⁷-

[(1′R,2′S,3′R,4′S)-2′,3′-Dihydroxy-4′-aminocyclopentyl]-4-

amino-5-iodopyrrolopyrimidine (2, A-134974) acts in the

spinal cord to reduce carrageenan-induced inﬂammatory

hyperalgesia.²²Both 1a and its potent analogue, e.g., the AdK

i n h i b i t o r 4 - ( N - p h e n y l a m i n o ) - 5 - p h e n y l - 7 - ( 5 ′ -

deoxyribofuranosyl)pyrrolo[2,3-d]pyrimidine 4, inhibited max-

imal electroshock (MES) seizures in rats.²Compound 4 also

reduced the volumes of infarction in a model of focal cerebral

ischemia in rats.²³Another class of nonnucleoside, heterocyclic

inhibitors includes the widely used AdK inhibitor 4-amino-5-(3-

bromophenyl)-7-(6-morpholino-pyridin-3-yl)pyrido[2,3-d]-

pyrimidine (3, ABT-702). Potent and selective inhibitors of the

AdK of Mycobacterium tuberculosis that do not aﬀect human

AdK were found to have antimicrobial activity.^24,25AdK

inhibitors were being considered for clinical trials for pain and

seizure treatment in the early 2000s, but this eﬀort was

discontinued, with one of the inhibitors causing brain

hemorrhage in dogs.²⁶Thus, AdK inhibition holds interest

for the control of infectious, as well as neurological diseases.

AdK inhibitors also induce anti-inﬂammatory eﬀects that are

adenosine-dependent.²⁷A potent AdK inhibitor, 4-amino-3-

iodo-1-β-^D-ribofuranosylpyrazolo[3,4-d]pyrimidine 1b, was also

shown to indirectly reduce the expression of inducible nitric

oxide synthase and tumor necrosis factor (TNF) in glial cells

via activation by adenosine of G_s-coupled ARs.^28,29Recently,

A common approach in medicinal chemistry to enhance the

activity or selectivity of ﬂexible biologically active, small

molecules is to introduce a conformational constraint to

achieve a desired conformation for interacting with a target

biopolymer, i.e., here an enzyme. This lowers the energy barrier

of the binding process and can eliminate undesired interactions

with other molecular targets that prefer a diﬀerent con-

formation of the ligand. One means of sterically constraining

the ribose ring of nucleoside derivatives, as already applied to

antiviral agents and to receptor ligands, is to incorporate a

bicyclic ribose substitute in a conformation that is preferred

when the molecule is bound to the protein target.^32,33The

methanocarba ([3.1.0]bicyclohexane) ring system is applied to

hold the ribose-like ring in either a North (N) or a South (S)

conformation. The X-ray structure of human AdK shows a

bound nucleoside inhibitor 1c containing a ribose in the (S)

conformation, which is similar to the ribose conformation

preferred by other nucleoside kinases.³⁴⁻³⁷This prompted us

to explore the eﬀects of sterically constraining nucleoside

inhibitors of human AdK using methanocarba rings.

RESULTS

■

Chemical Synthesis. The intermediate 24 containing the

(S)-methanocarba ring with 1′-amino functionalization was

required as an intermediate for the target compounds (Scheme

1). We adapted our previously reported synthesis of

enantiomerically pure (S)-methanocarba nucleosides via

bicyclic intermediate 24^38,39to a larger scale preparation of

this intermediate (Supporting Information, Scheme S1).

B

DOI: 10.1021/acs.jmedchem.6b00689

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

a

Scheme 1. Synthesis of Nucleobase Modiﬁed (S)-Methanocarba Analogues of 5-Iodotubercidin

a

Reagents and conditions: (a) EtOH, 2-(4,6-dichloropyrimidin-5-yl)acetaldehyde (25), NEt₃, reﬂux, 18 h, 76%; (b) anhydrous DMF, NIS, 60 °C, 6

h, 92%; (c) 70% TFA (aq), rt, 1.5 h, from 27, 83%; (d) EtOH, 7N NH₃-MeOH, sealed tube, 110 °C, 18 h, 89%; (e) anhydrous THF, aniline, 1M

tBuOK in THF, −20 °C, 1.5 h, 70−80%; (f) 6:1 DME−EtOH, phenylboronic acid, Pd(PPh₃)₄, satd Na₂CO₃(aq), 90 °C, 7 h, 75−86%; (g)

anhydrous THF, 1M TBAF in THF, rt, 18 h, 85−96%; (h) (i) anhydrous CH₃CN, DMAP, O-(p-tolyl) chlorothionoformate (36), rt, 18 h, (ii)

anhydrous toluene, Bu₃SnH, AIBN, reﬂux, 2 h, 52−90%.

Using the (S)-methanocarba intermediate 24 as a precursor,

analogues of known AdK nucleoside inhibitors were prepared.

The 7-deazaadenine core was constructed by reacting 24 with

symmetrical dichloropyrimidine bearing an acetaldehyde

moiety (25),⁴⁰which on iodination using NIS followed by

the removal of protecting groups in aqueous triﬂuoroacetic acid

resulted in 28. Similarly, the aminolysis of 27 and deprotection

rendered 30, the (S)-methanocarba analogue of 1a, in

moderate yield.

Substitution of chlorine on the 6-position of 7-deazapurine

using aniline and sodium acetate achieved a complete reaction,

but the use of a stronger base, i.e., potassium tert-butoxide at

lower temperature produced N⁶-phenyl 7-iodo derivatives 31a

and 31b in increased yields. A Suzuki reaction involving

arylboronic acids and 31a,b, followed by the removal of a silyl

protecting group, gave 35a−c, which were subjected to

Barton−McCombie deoxygenation⁴¹to yield 5′-deoxy com-

pounds 37a−c. Removal of the protecting groups from 31, 33,

and 37a−c gave 32, 34, and 38a−c, respectively, in low to

moderate yields.^17,42Compounds 38b and 38c diﬀer from 38a

only in the presence of a ﬂuorine atom in the p-position of

either the 4-phenyl-amino or 5-phenyl ring, respectively.

In an attempt to synthesize 5′-azido compounds (Scheme 2),

a Mitsunobu reaction involving dipenylphosphoryl azide

(DPPA) and 29 yielded disubstituted 40 exclusively. Surpris-

ingly, this compound was suﬃciently stable to be isolated using

silica gel column chromatography. 40 on heating with sodium

azide in DMF formed undesired product 41, and attempts to

convert this phosphine to 42 or 43 using aqueous TFA were

unsuccessful. Alternatively, mesylation of 29 followed by

substitution with sodium azide at elevated temperature aﬀorded

the desired compound 42 in 96% yield, which under Staudinger

C

DOI: 10.1021/acs.jmedchem.6b00689

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

a

Scheme 2. Synthesis of 4′-Position Modiﬁed (S)-Methanocarba Analogues of 5-Iodotubercidin

a

Reagents and conditions: (a) anhydrous THF, 1M TBAF in THF, rt, 18 h, quantitative yield; (b) anhydrous THF, PPh₃, DEAD, DPPA, 0 °C → rt,

18 h; (c) anhydrous DMF, NaN₃, 65 °C, 24 h; (d) (i) anhydrous pyridine, MsCl, 0 °C → rt, 4 h, (ii) anhydrous DMF, NaN₃, 65 °C, 4 h, 95%; (e)

70% TFA (aq), rt, 1.5 h, 35−42%; (f) (i) anhydrous THF, PPh₃, rt, 18 h, (ii) 25% NH₄OH (aq), 65 °C, 4 h, 80%; (g) 4:1 CH₃CN−H₂O, BAIB,

TEMPO, rt, 18 h, 49%; (h) DPPA, TEA, anhydrous t-BuOH, rt, 30 min, 90 °C, 2 h; (i) (i) DPPA, TEA, anhydrous THF, rt, 18 h, (ii) toluene,

benzyl alcohol, 110 °C, 5 h, 17%; (j) anhydrous CH₂Cl₂, 33% HBr-CH₃COOH, 0 °C → rt, 2.5 h, 17%.

reaction conditions gave 44 in good yield.⁴³The deprotection

of the acetonide group from 42 and 44 using aqueous TFA led

to 5′-azido and 5′-amino congeners 43 and 45, respectively.

To arrive at 49 from 39, a Curtius rearrangement strategy

was adapted, and to realize this, the 4′-hydroxylmethyl group in

39 was converted to the corresponding carboxylic acid 46

employing a TEMPO-BAIB oxidation⁴⁴in aqueous acetonitrile.

Initial eﬀorts to prepare the Boc-protected analogue of

urethane 48 from 46 using DPPA and tert-butanol failed, but

instead, after deprotection of the isopropylidine group, resulted

in dimer 47. However, a similar reaction with benzyl alcohol

formed Cbz-protected compound 48,³⁹which on deprotection

using 33% HBr in acetic acid provided the desired compound

49, as a minor product, and 50 as the major product.

Unfortunately, an attempt to convert 50 to the desired

compound 49 using NIS in DMF at elevated temperature did

not materialize.

The synthesis of several corresponding (N)-methanocarba

analogues was performed as shown in Scheme 3. A protected

bicyclic intermediate 51, prepared previously for studies of

adenosine receptors,^38,39,45was subjected to a Mitsunobu

reaction with 6-chloro-7-iodo-7-deazapurine. The next step

provided two divergent pathways leading to target inhibitors:

nucleophilic substitution of the 4-chloro group with either

ammonia or aniline (followed by a Suzuki coupling at the 5-

position). Deprotection of the hydroxyl groups yielded

compounds 55 and 57, which are the (N)-methanocarba

equivalent of compounds 34 and 30, respectively.

Biochemical Evaluation. The nucleoside derivatives were

tested for inhibitory potential at hAdK using a commercial,

nonisotopic assay for the measurement of hAdK activity based

on continuous monitoring at 340 nm (appearance of reduced

nicotinamide-adenine dinucleotide from a coupled enzymatic

reaction). The hAdK-catalyzed phosphorylation of inosine to

D

DOI: 10.1021/acs.jmedchem.6b00689

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

a

Scheme 3. Synthesis of (N)-Methanocarba Analogue 57 of 5-Iodotubercidin and Its N⁶,7-Diphenyl Congener 55

a

Reagents and conditions: (a) 6-chloro-7-iodo-7-deazapurine, anhydrous THF, DEAD, PPh₃, 0 °C → rt, 18 h, 88%; (b) anhydrous THF, aniline, 1M

tBuOK in THF, −20 °C, 1.5 h, 63%; (c) 6:1 DME−EtOH, phenylboronic acid, Pd(PPh₃)₄, satd Na₂CO₃(aq), 90 °C, 7 h, 82%; (d) 70% TFA (aq),

rt, 1.5 h, 77−79%; (e) EtOH, 7N NH₃-MeOH, sealed tube, 110 °C, 18 h, 82%.

Figure 1. Concentration−response study for all inhibitors compared to 1a and 2 at 50 μM, 20 μM, 1 μM, 500 nM, 100 nM, 20 nM, and 10 nM.

Percent residual activity of 38a was commensurate with 2 at most concentrations. Slope activity was determined between 0 and 40 min of assay

activity time and determined as a percentage relative to DMSO.

inosine 5′-monophosphate is coupled to the inosine 5′-

monophosphate dehydrogenase-dependent oxidation of the

product. Thus, the assay was performed in the presence of

dithiothreitol, oxidized nicotinamide-adenine dinucleotide

(NAD, 2.5 mM) and ATP (2.75 mM) as cofactors and inosine

(2.5 mM) as substrate. hAdK activity was continuously

measured via absorption at 340 nm for 4 h, where

measurements were made at 5 min intervals; however, for

graphing and analysis purposes, we focused on the ﬁrst 40 min,

which was mostly linear, for determining the slope of initial

AdK activity (Figure 1). Percent inhibition of hAdK activity was

calculated by determining the change in the slope of the hAdK

reaction in the absence or presence of inhibitor during the

linear phase of the enzymatic reaction (ﬁrst 40 min) (Table 1).

We included two known AdK inhibitors, compound 1a and 2,

as positive controls. Each inhibitor was analyzed at the

E

DOI: 10.1021/acs.jmedchem.6b00689

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Because the 5′-OH derivative 30 was a poor inhibitor of

hAdK, we hypothesized that it might serve as a substrate for

hAdK, unlike the corresponding riboside 1a, which is not

primarily a substrate.^29,43The K_mof adenosine as substrate of

hAdK is 700 nM,²⁹and the nucleoside analogue concentrations

that we used in inhibition experiments far exceeded that

concentration. By comparison in another kinase system, (S)-

methanocarbathymidine, but not (N)-methanocarbathymidine,

is a good substrate for herpes simplex virus type 1 thymidine

kinase.³⁵Moreover, (S)-methanocarbathymidine is not a

substrate for human cytosolic thymidine kinase isoenzyme

1.³⁵Thus, it was important to probe the substrate qualities of

compound 30, (S)-methanocarbaadenosine derivative, at

hAdK. Therefore, we performed three diﬀerent assays under

similar conditions: ﬁrst, with substrate in the presence of 1 μM

30; second, without substrate in the presence of 1 μM 30; and

third as a control, without 30. After 4 h incubation, the

resulting mixtures were studied using LC-MS in negative and

positive mode (Supporting Information), which could identify

and qualitatively measure the levels of nucleosides and their

phosphorylated products. In the ﬁrst and last cases, we detected

the ﬁnal enzymatic product, i.e., xanthosine-5′-phosphate (MW

364 + 1). However, in the ﬁrst two cases, we found no evidence

from the LC-MS analysis that 30 serves as a substrate of hAdK.

There was a prominent signal (MW 402 + 1) from unchanged

compound 30, and no detectable or prominent signal of the

corresponding 5′-phosphate (MW 482 1).

Table 1. Percent Inhibition of hAdK Activity for All 14

Inhibitors, Relative to Uninhibited Activity, At

Concentrations of 50 μM, 20 μM, 500 nM, and 100 nM, and

the Derived IC₅₀Values

Molecular Modeling. The selection of novel AdK

inhibitors was informed using molecular modeling studies. To

date, three X-ray structures of the hAdK in complex with both

nucleoside and nonnucleoside inhibitors have been solved,

which revealed at least two diﬀerent enzyme conformations. A

closed form was present in the complexes of hAdK with

nucleoside 1c³⁶(PDB ID 2I6A) and 7-ethynyl-7-deazaadeno-

sine³⁴(PDB ID 4O1L), but the enzyme adopted an open

conformation in the complex with a bulkier alkynylpyrimidine

inhibitor³⁶(PDB ID 2I6B). Given the structural similarity of

our compounds to 1c, we used the corresponding cocrystallized

structure as macromolecular starting point of our analysis.

MD Simulation of the X-ray Structure of the 1c-AdK

Complex. In a ﬁrst instance, we subjected the experimentally

determined complex (PDB ID 2I6A) to 30 ns of all atom

molecular dynamics (MD) simulation, in order to identify the

residues mostly involved in the interaction and to explore in

detail the conformational space of the inhibitor. In the starting

structure,³⁶1c bound with the glycosidic bond (χ) in the anti

conformation (χ = −134.7°) and the sugar moiety in the C1′-

exo conformation (P = 125.3°). The analysis of the trajectory

(Supporting Information, Video S1, left panel) revealed that the

anti conformation was retained throughout the simulation,

while the pseudosugar ring explored diﬀerent conformational

states (Supporting Information, Figure S2). The anti

conformation of the glycosidic bond seemed to be compatible

with the charge distribution of the residues surrounding the

enzyme active site. Indeed, the inhibitor established persistent

H-bond interactions with negatively charged residues through

the C2′ and C3′ hydroxyl groups and a stable π−π stacking

interaction with Phe170 through the purine core. Moreover,

the inhibitor was anchored in the active site of the enzyme

through a network of H-bond interactions consisting of the N3

atom of the purine core and the C2′ hydroxyl groups

associating with the backbone of Ser65 and Gly64, respectively.

Concerning the ribose ring conformations of 1c, the starting

a

Based on slope of the linear portion (0−40 min interval) of the

activity plot of hAdK, compared to control with DMSO vehicle. ND,

not determined.

following concentrations (μM): of 50, 20, 1, 0.50, 0.10, 0.020,

and 0.010. Percent residual hAdK activity showed a signiﬁcant

interaction (F₇₈₄₉₀= 71, p < 0.0001) and main eﬀects for both

inhibitor (F₁₃₄₉₀= 626, p < 0.0001) and concentration (F₆₄₉₀

=

3366, p < 0.0001). Post hoc analysis used Bonferroni multiple

comparison tests to determine drug diﬀerences relative to 2

(most potent known inhibitor used here) at each of the

concentrations investigated. At the highest concentration, 28,

30, 45, 49, and 57 signiﬁcantly varied from 2 with poor

inhibition of hAdK. Compound 43 was slightly more potent,

varying from 2 at 20 μM. 32 and 55 were moderately potent,

beginning to vary from 2 at 1 μM and 500 nM, respectively.

Only at a concentration of 100 nM did the inhibition by 34,

38b, and 38c diﬀer from 2, while 38a never signiﬁcantly varied

from 2, indicating that 38a as low as 20 nM was equipotent to 2

in inhibition of hAdK. At lower concentrations, 34 and 38a

maintained more potent hAdK inhibition; therefore, it was

necessary to determine full concentration−response curves to

distinguish between the most potent compounds. We

performed a detailed concentration−response study of most

of the analogues for precise determination of IC₅₀, using the

earlier data with the addition of three concentrations (10, 0.20,

and 0.050 μM). The IC₅₀values for potent (S)-methanocarba

inhibitors 34 and 38a were 114 and 88 nM, respectively.

F

DOI: 10.1021/acs.jmedchem.6b00689

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Figure 2. Docking poses of the selected methanocarba derivatives in complex with the closed form of hAdK (PDB ID 2I6A) obtained after IFD: 30

(A) and 57 (B), both derivatives of prototypical AdK inhibitor 1a; 34 (C) and 57 (D), both containing two phenyl rings. Ligand and side chains of

residues important for ligand recognition are represented as line and sticks, respectively. H-Bonds are pictured as yellow dashed lines, while π−π

stacking interactions are indicated by blue dashed lines, with the centroids of the aromatic rings displayed as cyan spheres. Nonpolar hydrogen atoms

are omitted.

C1′-exo conformation (Supporting Information, Figure S2(i)),

featuring a bidentate H-bond interaction between the C2′ and

C3′ hydroxyl groups and the side chain of Asp18 (Supporting

Information, Video S1), was the most favorable in terms of

ligand−protein interaction energy (IE) during the simulation

(right panel in Video S1 and upper panel in Supporting

Information, Figure S2). However, after approximately 12 ns of

MD simulations, the H-bond network was lost and the sugar

ring adopted a C2′-endo (S) conformation (P = 156.7°,

Supporting Information, Figure S2 (ii)) with the C2′ and C3′

hydroxyl groups interacting with the Asp18 side chain and

water molecules, respectively (Supporting Information, Video

S1). At the end of the simulation (Supporting Information,

Video S1), the ring adopted a less favorable, in terms of IE,

C3′-endo (N) conformation (P = 36.4°, Supporting Informa-

tion, Figure S2 (iii)) that was accompanied by a rotameric

switch of the Asp18 side chain to establish a bidentate H-bond

interaction with Arg132 while still interacting with the C3′

hydroxyl group (data not shown). To assess whether this

conformation was persistent, we restarted the simulation for

another 30 ns in which the inhibitor remained in the (N)

conformation (Supporting Information, Figure S2). These

results prompted us to consider nucleoside inhibitors con-

strained by the methanocarba ring system in the (N) as well as

the (S) conformation. We therefore analyzed from a molecular

point of view two isomeric pairs by inspecting in detail the

comparison between 30 and 57 and between 34 and 55.

Docking of Selected Methanocarba-Nucleoside Deriva-

tives. The compounds were docked into the closed form of

hAdK by using the induced ﬁt docking (IFD) procedure (see

Methods section), because a preliminary attempt to dock bulky

diphenyl derivatives 34 and 55 into the rigid enzyme in the

closed form failed. In all four docking poses (Figure 2), the

purine core established a π−π stacking interaction with Phe170

and H-bond interactions with the side chain of Asn14 and the

backbone of Ser65, through the N1 and N3 atoms, respectively.

Furthermore, the phenyl rings of 34 and 55 (Figure 2C,D)

interacted with residues located in the small lid domain (Leu16,

Leu40, Leu134, Ala136, Leu138, and Val174) by means of

extended hydrophobic contacts. In the derivatives in the (S)

conformation (30 and 34, parts A and C of Figure 2,

respectively), the methanocarba ring was involved in a tight

network of H-bond interactions. The backbone NH of Gly64

and side chain of Asn68, respectively, served as H-bond donors

to the C2′ and C3′ hydroxyl groups. This coordination

contributed to placing the same hydroxyl groups in a favorable

orientation to act as H-bond donors in a bidentate interaction

with the side chain of Asp18. Moreover, several additional H-

bond interactions involving the C5′ hydroxyl group further

stabilized this proposed binding mode. However, in the

docking poses of the (N) derivatives (57 and 55, parts B and

D of Figure 2, respectively), we detected a slightly diﬀerent

interaction pattern to compensate for the loss of the bidentate

interaction with Asp18. In particular, the hydroxyl groups at

C2′ engaged in H-bond interactions with Gly64 (backbone)

and Asn68, while the hydroxyl group at C5′ established a H-

bond interaction with the side chain of Gln38 and with Asn296

in the docking pose of 57 and 55, respectively. As shown below,

this diﬀerent orientation of the (N) isomer was constrained by

the closed form of the rigid enzyme and the bidentate

interaction was partially restored by relaxing the protein in a

dynamic environment.

MD Simulations of Methanocarba-Nucleoside hAdK

Complexes. The docking poses described above were subjected

to 30 ns of all-atom MD simulation. The visualization of the

trajectory of the comparison between the 30-hAdK and 57-

hAdK complexes (Supporting Information, Video S2), revealed

that in the 30-hAdK complex (Supporting Information, Video

S2, left panel), the H-bond interaction network anchoring the

C2′ and C3′ hydroxyl groups in a favorable orientation to

G

DOI: 10.1021/acs.jmedchem.6b00689

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

establish a bidentate interaction with Asp18 side chain was lost

after approximately 14 ns. This conformational change followed

the opening of the small lid domain, probably triggered by the

steric hindrance imposed by the fused three-membered ring.

However, in the trajectory of the 57-hAdK complex

(Supporting Information, Video S2, right panel), the H-bond

network was maintained throughout the simulation. A

comparison of trajectories of the 34-hAdK and 55-hAdK

complexes (Supporting Information, Video S3) highlighted the

almost immediate opening of the small lid domain for both

complexes. This opening established π−π stacking interactions

between Phe170 (purine core) and Phe201 (phenyl-amino

group) and van der Waals interactions with the side chains of

Gln38 (phenyl group, data not shown), Leu40, and Leu138

(phenyl-amino group), which persisted throughout the

simulation. In the 34-hAdK complex (Supporting Information,

Video S3, left panel), the C2′ and C3′ hydroxyl groups and the

Asp18 side chain persisted in a reciprocal favorable orientation

for almost the entire duration, through the interplay of Asp300,

Arg332, and water molecules. The analysis of the trajectory of

the 55-hAdK complex (Supporting Information, Video S3, right

panel), indeed, featured less enduring H-bond networks and

more predominant van der Waals interactions that contributed

to a net IE proﬁle that was slightly more favorable (Supporting

Information, Figure S3).

domain during MD simulation of nucleoside inhibitors agreed

with modeling studies reported by other authors.^46,47

To further support our hypothesis, we subjected 1a to the

same computational protocol (IFD followed by 30 ns of MD

simulation run in triplicate, see Supporting Information, Table

S1). The obtained docking pose (Supporting Information,

Figure S5A) superimposed well with the experimental binding

mode for 1c (RMSD = 0.37 Å), and both the purine core and

C2′ and C3′ hydroxyl groups featured the same interaction

pattern observed for the methanocarba derivatives. Additional

H-bond interactions were detected for the side chains of Gln38

and Asp300 with the ribose ether oxygen and the C5′ hydroxyl

group, respectively. The analysis of the trajectory (Supporting

Information, Video S4) was consistent with the conformational

analysis of the 1c-hAK complex: in particular, 1a persisted for

most of the simulation time in the C1′-exo and C2′-endo (S)

conformations (65% and 25% of the trajectory, respectively)

and maintained a stable network of interactions, such as the

bidentate H-bond with the side chain of Asp18, H-bonds to the

backbone of Gly64, Ser65, and the side chain of Asn14, and the

π−π stacking interaction with Phe170. Concerning the enzyme

conformation, superimposition of MD average protein structure

with the enzyme in the closed conformation (magenta and gray

ribbons in Supporting Information, Figure S5B, respectively)

revealed that the enzyme persisted in the closed conformation.

Similarly to what was observed for the 57-hAdK complex, no

dynamic domain was found for the 1a-hAK complex.

hAdK Conformation during MD Simulations. As men-

tioned above, initial attempts to dock the 34 and 55 in the

closed conformation of the enzyme failed. We therefore

resorted to an IFD protocol to obtain initial docking poses

that we subsequently subjected to MD simulation (30 ns). The

analysis of the trajectories revealed that the opening of the

small lid domain occurred after a few ns of simulation for the

hAK complexes with 34 and 55 (Supporting Information,

Video S3, left and right panels) and after approximately 14 ns

of simulation for the 30-hAK complex (Supporting Informa-

tion, Video S2, left panel). Superimposition of MD average

protein structures with respect to the enzyme starting structure

(PDB ID 2I6A, gray ribbons) is depicted in Supporting

Information, Figure S4. First, we compared the X-ray structures

of the enzyme in the closed conformation (PDB ID 2I6A, gray

ribbons in Supporting Information, Figure S4) and the open

conformation (PDB ID 2I6B, green ribbons in Supporting

Information, Figure S4A). Comparing each MD average

structure with the starting conformation highlighted that the

enzyme remained in the closed conformation for the hAdK

complexes with 1c (Supporting Information, Figure S4B, pink

ribbons) and 57 (Supporting Information, Figure S4D, green

ribbons). However, in the complexes with 30 (Supporting

Information, Figure S4C, yellow ribbons), 34 (Supporting

Information, Figure S4E, cyan ribbons), and 55 (Supporting

Information, Figure S4F, orange ribbons), the enzyme

approached the open conformation. In particular, in the MD

average structures, the small lid domain was rotated outward by

about 24° with respect to the large domain for both the 34-

hAdK and 55-hAdK complexes (comparison with the enzyme

starting structure obtained after IFD, C-terminal domain

excluded from the analysis), while the rotation angle was

approximately 9° for the 1c-hAdK and 30-hAdK complexes.

Notably, no dynamic domain was found for the 57-hAdK

complex; thus, the enzyme persisted in its closed conformation

throughout the simulation. The preference for an open enzyme

conformation as well as the rotation angle of the small lid

DISCUSSION

■

The application of AdK inhibitors to CNS disorders such as

epilepsy and chronic pain experienced a hiatus during the past

decade, due to possible side eﬀects of the known inhibitors.

Critical gaps to be addressed are isoform selectivity

(cytoplasmic versus nuclear AdK), cell-type selectivity (neuron

versus astrocyte), and trans-membrane transport. Thus, there is

a need for a new class of inhibitors with enhanced selectivities,

such as nucleosides that have a major distinction from the

known riboside inhibitors such as 1a and 2.

We have utilized an alternative ribose-like ring system, e.g.,

two isomeric methanocarba pseudoribose substitutions, in

nucleoside derivatives to demonstrate that inhibition of AdK

is compatible with these major structural changes. By virtue of

the bridged bicycloaliphatic ring, a rigid (S) 3′-exo or (N) 2′-

exo envelope conformation is permanently locked in these

nucleoside analogues. The most potent inhibitors, (S)-

methanocarba derivatives 34, 38a, 38b, and 38c and (N)-

methanocarba derivative 55 were approximately 2-fold less

potent than the reference compound 2. The apparent

discrepancy for the IC₅₀of 2 in our study of 48 nM, compared

to its previously reported sub-nM IC₅₀,²²is because of the use

of a diﬀerent assay system. The previous assay used was based

on a radioactive assay using adenosine as a substrate,¹⁷and ours

is a commercial spectrophotometric assay based on inosine as

substrate.

The presence of a p-F substitution in either the 5-phenyl 38b

or 4-phenylamino 38c ring in the (S)-methanocarba series

maintained the inhibitory potency of analogue 38a. We

synthesized and compared potencies of two pairs of

corresponding (S) and (N)-methanocarba 4′-CH₂OH deriva-

tives. We conclude that both conformations maintain potency

at hAdK and can provide a path to novel nucleoside inhibitors.

5-Iodo analogues 32 (S) and 57 (N) displayed comparable

potency, and the pair of more potent 4-phenylamino-5-phenyl

H

DOI: 10.1021/acs.jmedchem.6b00689

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

analogues 34 (S) and 55 (N) also displayed similar IC₅₀values.

This conclusion was consistent with X-ray structures of known

inhibitors bound to hAdK^36,37and with molecular docking and

MD analyses of several of the present derivatives.

in a bidentate interaction with the hydroxyl group in C2′ and

C3′ position of adenosine.⁸⁰

Thus, we have discovered novel high potency inhibitors of

hAdK, which can now be evaluated in vivo, for example, in

eﬀects on DNA methylation, in comparison to riboside

inhibitors such as 1a.⁵The brain penetration by these analogues

remains undetermined, which could be a complicating factor

because nucleosides often have low entry into the CNS.⁴⁹

Nevertheless, the slightly greater hydrophobicity of the (S)-

methanocarba derivatives (cLogP of 49 is −0.11; tPSA is 120

Å) compared to ribosides (cLogP of the tetrahydrofuryl

equivalent of 49 is −1.15; tPSA is 130 Å) might be beneﬁcial

for crossing the blood−brain barrier. Furthermore, there

remains the possibility that the speciﬁcity for AdK is enhanced

in these (S)-methanocarba derivatives, which is crucial for

avoiding possible side eﬀects already noted for known AdK

inhibitors. Although we have shown the feasibility of using this

ribose conformational constraint in AdK inhibitors, there

remains room for structural optimization to improve the

inhibitory potency, considering that some of the known

inhibitors achieve sub-nM aﬃnity.

A similar attempt to constrain analogues of 2′-deoxynucleo-

side inhibitors of adenosine deaminase (ADA) using the

[3.1.0]bicyclohexane ring system failed.⁴⁸Although the enzyme

preferred the (N) methanocarba nucleoside over the (S), the

relative rate of deamination was ∼100-fold lower than

adenosine, which suggested a possible role of the 4′-oxygen

atom of native ribose in an anomeric eﬀect to assist hydrolysis

by ADA. In (S)-methanocarba nucleosides, the syn-conforma-

tion of the pseudoglycosidic bond is thought to be more stable

than the anti-conformation, but in the X-ray structure of hAdK

complexes, the nucleoside anti-conformation is present. The

analysis of MD simulation starting from the X-ray complex

suggests that the anti conformation best suits the electrostatic

potential distribution of the enzyme active site, featuring a

highly hydrophobic cavity that hosts the purine core next to a

region full of charged residues that anchors the ribose ring.

As far as the conformational preference of the enzyme is

concerned, the active site of AdK appears to be highly ﬂexible,

as both (S)-methanocarba (C3′-exo, P = 198°) compound 34

and its (N)-methanocarba (C2′-exo, P = 342°) analogue 55 are

equipotent inhibitors. It has to be noted that the synthesized

locked conformers deviate by 18° from the (N) and (S) ideal

(N) and (S) conformations of ribose.^32,35Interestingly, the

near equipotency of (N) and (S) methanocarba isomers did not

apply to conformationally locked analogues of 1a lacking the

phenyl moieties. (S)-Conformers 30 and 49 were weak

inhibitors of hAdK, while the inhibitory potencies of 1a and

its (N)-methanocarba counterpart 57 were comparable. A

possible explanation is provided by the modeling results: the

bulky N⁶,C7-diphenyl substituents do not ﬁt into the hAdK

closed form and, as emerged from MD simulations, induced the

opening of the small lid domain. The enzyme open

conformation is expected to exhibit a higher plasticity with

respect to closed conformation to which smaller 1a mimics are

predicted to bind. As a consequence, this enzyme conformation

could accommodate the pseudosugar ring locked in either the

(S) or (N) conformation. Thus, hAdK might indeed prefer the

(N)-conformer as inhibitor in the closed/catalytic-phase of the

enzyme.

When comparing pairs of methanocarba isomers, a

distinction needs to be made between those having or lacking

bulky phenyl groups, which largely determine the enzyme

conformational preference. Indeed, when phenyl groups are

attached to the purine core, the opening of the small lid domain

occurs to host them. Those groups establish additional van der

Waals interactions and hydrophobic contacts with the lid,

which compensate for any geometrical factors in the ribose

region. However, for the small nucleoside inhibitors, the

conformation of the pseudoribose ring predominates. In

compound 30, the interactions with the backbone NH groups

of Gly64 and Ser65 appear to be missing because of steric

hindrance of the cyclopropyl ring of the (S) isomer. Those

interactions impose a speciﬁc conformation of adenosine

analogues and might help maintain the enzyme in the closed

conformation by allowing the C2′ and C3′ hydroxyl groups of

the inhibitors to H-bond with Asp18 of the lid. A mutagenesis

study of AdK from Leishmania donovani reported the

indispensable role of Asp16 (the equivalent of Asp18 in the

hAdK) for the catalytic activity and suggested its involvement

The interaction of these (S)-methanocarba analogues with

nucleoside transporters that are relevant to adenosine

derivatives, such as ENT1 and CNT2,⁵⁰remains to be

characterized. The ability to serve as substrate or inhibitor of

nucleoside transporters could aﬀect the biodistribution or

availability of the compounds in vivo.^49,51An analogue of the

potent ENT1 inhibitor S-(4-nitrobenzyl)-thioinosine contain-

ing the opposite ring twist conformation ((N) methanocarba)

was shown to inhibit ENT1, and other ﬁxed conformations of

2′-deoxynucleosides were evaluated.^52,53(S)-Methanocarba 2′-

deoxyadenosine inhibited both ENTs and CNTs, although less

potently than 2′-deoxyadenosine.

The bioavailability and the in vivo activity of these inhibitors

remain to be determined. Other nucleoside derivatives were

protective in seizure models when administered peripherally.¹⁷

For example, the Br analogue of carbocyclic nucleoside 2 was

found to be orally active in vivo in models of pain and

inﬂammation.²¹5′-Deoxynucleoside analogues of 1a were

noted to be more potent in vivo than 5′-amino analogues,

possibly because of enhanced passage across the blood−brain

barrier by virtue of being less polar.¹⁷Among the new (S)-

methanocarba AdK inhibitors, 5′-deoxy analogue 38a and

especially the ﬂuoro analogues 38b and 38c appear to be the

least polar based on their tendency to dissolve in organic

solvents.

It is possible that transient treatment with AdK inhibitors

would have long-lasting therapeutic beneﬁts for treatment of

CNS disorders, not only by raising the basal level of AR

activation but also through epigenetic reprogramming. A high

level of adenosine in the brain drives the enzymatic equilibrium

in the presence of S-adenosylhomocysteine (SAH) hydrolase in

the direction of increased SAH formation. SAH in turn inhibits

DNA methyltransferases through product inhibition.^39,54

Because the epigenetic eﬀects related to changes in the DNA

methylation status would persist, it might be possible to reduce

the duration of drug administration, thus avoiding toxicities

associated with prolonged, chronic dosing. A transient dosing

regimen might also avoid possible side eﬀects such as liver

toxicity.¹²Compounds 38a and 55 were submitted to the

Psychoactive Drug Screening Program (PDSP) for screening at

45 receptors, channels, and transporters (Supporting Informa-

tion).⁵⁵38a was found to inhibit radioligand binding at the

I

DOI: 10.1021/acs.jmedchem.6b00689

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

screened compounds is >95% as determined by HPLC with detection

at 254 nm. tPSA and cLogP were calculated using ChemDraw

Professional version 15.0 (PerkinElmer, Boston, MA). All chemical

reagents were obtained from Sigma-Aldrich (St. Louis, MO), except 2-

(4,6-dichloropyrimidine-5-yl)acetaldehyde 25 (Small Molecules, Inc.,

Hoboken, NJ).

human 5HT₇(serotonin) receptor with a K_ivalue of 0.71 μM

and did not substantially inhibit binding at any of the other oﬀ-

target sites examined (<50% inhibition at 10 μM). 55 was

found to inhibit radioligand binding at the human 5HT_2B

receptor with a K_ivalue of 51 nM and did not substantially

inhibit binding at any of the other oﬀ-target sites examined.

Also, 38a was found to be inactive (10 μM) as agonist or

antagonist at human P2Y₁, P2Y₂, P2Y₄, and P2Y₁₁Rs (calcium

transients) expressed in 1321N1 astrocytoma cells and

protease-activated receptor (PAR)1 expressed in mouse

KOLF cells. Nevertheless, the oﬀ-target interactions of the

present set of compounds would have to be examined more

extensively.

7-((3aS,3bS,4aS,5R,5aR)-5-(((tert-Butyldiphenylsilyl)oxy)methyl)-

2,2-dimethyltetrahydrocyclopropa[3,4]cyclopenta[1,2-d][1,3]-

dioxol-3b(3aH)-yl)-4-chloro-7H-pyrrolo[2,3-d]pyrimidine (26). The

amine 24 (200 mg, 0.46 mmol) was dissolved in ethanol (4 mL). To

this was added 4,6-dichloropyrimidin-5-acetaldehyde 25 (88 mg, 0.46

mmol) and triethylamine (TEA, 128 μL, 0.92 mmol). The reaction

mixture was heated to 90 °C under reﬂux condenser overnight,

volatiles evaporated under reduced pressure, and residue puriﬁed by

silica gel ﬂash column chromatography to aﬀord compound 26 as a

white foam (200 mg, 76%, R_f= 0.45, TLC eluent = 20% ethyl acetate

1

in hexanes). H NMR (400 MHz, CDCl₃) δ 8.40 (s, 1H), 7.79−7.65

CONCLUSION

■

(m, 4H), 7.49−7.34 (m, 6H), 7.28 (d, J = 3.6 Hz, 1H), 6.51 (d, J = 3.6

Hz, 1H), 4.98 (dd, J = 1.6, 7.1 Hz, 1H), 4.60 (dd, J = 0.8, 7.1 Hz, 1H),

4.12 (d, J = 7.9 Hz, 2H), 2.58 (t, J = 7.9 Hz, 1H), 2.11 (ddd, J = 1.4,

5.5, 9.9 Hz, 1H), 1.62 (t, J = 5.6 Hz, 1H), 1.58 (s, 3H), 1.44 (ddd, J =

1.8, 5.8, 9.8 Hz, 1H), 1.22 (s, 3H), 1.08 (s, 9H). ¹³C NMR (101 MHz,

CDCl₃) δ 151.72, 151.64, 150.44, 135.63, 135.61, 133.56, 133.50,

130.37, 129.81, 129.79, 127.75, 118.19, 111.51, 99.44, 84.57, 83.24,

65.03, 49.68, 47.51, 31.32, 26.90, 26.36, 24.06, 19.34, 16.63. HRMS m/

z [M + H]⁺for C₃₂H₃₆ClN₃O₃Si calculated 574.2293, found 574.2290.

7-((3aS,3bS,4aS,5R,5aR)-5-(((tert-Butyldiphenylsilyl)oxy)methyl)-

2,2-dimethyltetrahydrocyclopropa[3,4]cyclopenta[1,2-d][1,3]-

dioxol-3b(3aH)-yl)-4-chloro-5-iodo-7H-pyrrolo[2,3-d]pyrimidine

(27). To a dried mixture of compound 26 (200 mg, 0.35 mmol) and

N-iodosuccinimide (NIS, 135 mg, 0.42 mmol) was added anhydrous

DMF (5 mL) under nitrogen atmosphere. The solution was heated to

65 °C with stirring for 6 h. The volatiles were evaporated under

vacuum, and the residue was puriﬁed by silica gel ﬂash column

chromatography to aﬀord 27 as a white foam (225 mg, 92%, R_f= 0.50,

In conclusion, the development of novel AdK inhibitors, by

virtue of their ability to raise the level of endogenous adenosine,

particularly in disease states, remains of interest for the

potential treatment of seizures and neurodegenerative and

inﬂammatory conditions. We demonstrated that the class of

constrained bicyclic ribonucleoside analogues retain inhibitory

activity at this enzyme and are amenable to structural

modiﬁcation to enhance the potency. The SAR for ring-

constrained analogues deviates the SAR determined previously

for ribosides; analogues with aryl groups at the 7-deaza and N⁶

positions of adenine are greatly favored, with (S) ≈ (N). We

determined that the (S) conformation permits a range of

substitutions, but amino derivatives 45 and 49 were much less

potent than expected from the ribose equivalents. However, a

diﬀerence between the (N)- and (S)-methanocarba series is

that the methanocarba equivalent of reference riboside 1a is

more potent in the (N) than in the (S) series (55 and 30,

respectively). We have identiﬁed compounds 34, 38a, 38b, 38c,

and 55 as hAdK inhibitors with IC₅₀values of ∼100 nM. The

successful docking and MD simulation of selected members of

this series in the enzyme structure suggests that a structure-

based design approach for further enhancement is possible.

Although we have not yet explored all of the potential oﬀ-target

eﬀects and adenosine-receptor related side eﬀects of this

structural class, it is possible that the novel nonribose ring

system will provide a cleaner pharmacological proﬁle. The

potent AdK inhibitors in this study are now ready for further

tests in animal models of epilepsy and its development.

1

TLC eluent = 20% ethyl acetate in hexanes). H NMR (400 MHz,

CDCl₃) δ 8.36 (s, 1H), 7.80−7.68 (m, 4H), 7.53−7.34 (m, 7H), 4.94

(d, J = 7.0 Hz, 1H), 4.59 (d, J = 7.0 Hz, 1H), 4.08 (d, J = 7.9 Hz, 2H),

2.57 (t, J = 7.8 Hz, 1H), 2.10 (dd, J = 5.5, 9.9 Hz, 1H), 1.61 (t, J = 5.6

Hz, 1H), 1.57 (s, 3H), 1.42 (dd, J = 5.8, 10.0 Hz, 1H), 1.21 (s, 3H),

1.08 (s, 9H). ¹³C NMR (101 MHz, CDCl₃) δ 152.47, 151.39, 150.77,

135.62, 135.60, 135.58, 134.79, 133.47, 133.44, 129.84, 129.82, 129.64,

127.77, 127.77, 127.71, 117.52, 111.61, 84.40, 83.14, 64.92, 49.87,

47.43, 31.38, 26.89, 26.56, 26.34, 24.04, 19.33, 16.64. HRMS m/z [M

+ H]⁺for C₃₂H₃₅ClIN₃O₃Si calculated 700.1259, found 700.1253.

(1R,2S,3R,4R,5S)-1-(4-Chloro-5-iodo-7H-pyrrolo[2,3-d]pyrimidin-

7-yl)-4-(hydroxymethyl)bicyclo[3.1.0]hexane-2,3-diol (28). To com-

pound 27 (20 mg, 0.029 mmol) was added 70% aqueous

triﬂuoroacetic acid (TFA, 1.5 mL) and the mixture stirred at room

temperature for 1.5 h. The volatiles were evaporated under vacuum,

and the residual TFA was chased ﬁrst with water (2 × 2 mL) and then

neutralized with a mixture of water (2 mL) and a few drops of

triethylamine. The residue after evaporation was puriﬁed by silica gel

ﬂash column chromatography to aﬀord 28 as a white solid (10 mg,

83%, R_f= 0.50, TLC eluent = 10% methanol in dichloromethane). ¹H

NMR (400 MHz, CD₃OD) δ 8.63 (s, 1H), 7.86 (s, 1H), 4.69 (dd, J =

2.0, 6.5 Hz, 1H), 4.14−4.02 (m, 2H), 3.93 (dd, J = 4.9, 11.3 Hz, 1H),

2.34 (t, J = 4.6 Hz, 1H), 2.01−1.92 (m, 2H), 1.31 (ddd, J = 2.0, 4.3,

8.2 Hz, 1H). ¹³C NMR (101 MHz, CD₃OD) δ 153.63, 152.31, 151.19,

139.06, 119.11, 77.64, 73.84, 65.49, 51.54, 51.31, 50.55, 26.50, 15.48.

HRMS m/z [M + H]⁺for C₁₃H₁₃ClIN₃O₃calculated 421.9768, found

421.9771.

EXPERIMENTAL SECTION

■

Chemical Synthesis. Materials and Methods. All chemicals and

anhydrous solvents were obtained directly from commercial sources.

All reactions were carried out under nitrogen atmosphere using

anhydrous solvents unless speciﬁed otherwise. Room temperature or rt

refers to 25 5 °C. Silica-gel precoated with F254 on aluminum plates

were used for TLC. The spots were examined under ultraviolet light at

254 nm and further visualized by anisaldehyde or cerium ammonium

molybdate stain solution. Column chromatography was performed on

silica gel (40−63 μm, 60 Å). NMR spectra were recorded on a Bruker

400 MHz spectrometer. Chemical shifts are given in ppm (δ),

calibrated to the residual solvent signals or TMS. High resolution mass

(HRMS) measurements were performed on a proteomics optimized

Q-TOF-2 (Micromass-Waters). The RP-HPLC was performed using

Phenomenex Luna 5 μm C18(2)100A, AXIA, 21.2 mm × 250 mm

column. Purity was determined using Agilent ZORBAX SB-Aq, 5 μm,

4.6 mm × 150 mm column attached to Agilent 1100 HPLC system

and a linear gradient of acetonitrile/water as mobile phase with a ﬂow

rate of 1.0 mL/min (acetonitrile/5 mM tetrabutylammonium

dihydrogen phosphate in the case of 45 and 49). The purity of all

7-((3aS,3bR,4aS,5R,5aR)-5-(((tert-Butyldiphenylsilyl)oxy)methyl)-

2,2-dimethyltetrahydrocyclopropa[3,4]cyclopenta[1,2-d][1,3]-

dioxol-3b(3aH)-yl)-5-iodo-7H-pyrrolo[2,3-d]pyrimidin-4-amine (29).

Compound 27 (75 mg, 0.11 mmol) was placed in a glass pressure tube

with stir bar and treated with ethanol (5 mL) followed by 7 N

ammonia in methanol (5 mL). The tube was capped tightly and heated

to 120 °C for 18 h with stirring. The solvents were evaporated and the

residue puriﬁed by silica gel ﬂash column chromatography to aﬀord 29

as a white solid (65 mg, 89%, R_f= 0.45, TLC eluent = 5% methanol in

J

DOI: 10.1021/acs.jmedchem.6b00689

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

dichloromethane). ¹H NMR (400 MHz, CDCl₃) δ 7.82 (s, 1H), 7.75−

7.65 (m, 3H), 7.51−7.44 (m, 2H), 7.43−7.36 (m, 5H), 7.31 (s, 1H),

4.90 (dd, J = 1.6, 7.1 Hz, 1H), 4.58 (dd, J = 1.3, 7.2 Hz, 1H), 4.00 (d, J

= 7.6 Hz, 1H), 2.56 (t, J = 7.5 Hz, 1H), 2.12 (ddd, J = 1.4, 5.4, 9.9 Hz,

1H), 1.62 (t, J = 5.7 Hz, 1H), 1.57 (s, 3H), 1.40 (ddd, J = 1.8, 5.9, 10.0

Hz, 1H), 1.22 (s, 4H), 1.08 (s, 9H). ¹³C NMR (101 MHz, CDCl₃) δ

152.60, 148.78, 141.90, 135.61, 135.55, 133.35, 133.24, 132.89, 130.02,

129.94, 127.84, 127.80, 111.83, 103.11, 84.65, 83.15, 64.89, 50.27,

47.28, 31.53, 26.91, 26.86, 26.32, 24.04, 19.33, 16.72. HRMS m/z [M

+ H]⁺for C₃₂H₃₇IN₄O₃Si calculated 681.1758, found 681.1749.

(1R,2S,3R,4R,5S)-1-(4-Amino-5-iodo-7H-pyrrolo[2,3-d]pyrimidin-

7-yl)-4-(hydroxymethyl)bicyclo[3.1.0]hexane-2,3-diol (30). Follow-

ing the procedure described to synthesize 28, compound 29 (13 mg,

0.019 mmol) aﬀorded 30 as a white solid which was further puriﬁed by

RP-HPLC (3.6 mg, 47%, R_f= 0.25, TLC eluent = 10% methanol in

dichloromethane; RP-HPLC (C18) R_t= 18.9 min, water/acetonitrile

75/25 → 25/75 in 40 min at 5 mL/min). ¹H NMR (400 MHz,

CD₃OD) δ 8.08 (s, 1H), 7.35 (s, 1H), 4.57 (dd, J = 2.0, 6.6 Hz, 1H),

4.05−3.97 (m, 2H), 3.79 (dd, J = 3.6, 11.5 Hz, 1H), 2.25 (t, J = 2.9 Hz,

1H), 1.84−1.76 (m, 2H), 1.18−1.09 (m, 1H). ¹³C NMR (101 MHz,

CD₃CN) δ 158.58, 152.56, 150.84, 132.31, 105.11, 77.55, 74.12, 65.56,

51.15, 50.56, 49.18, 26.79, 14.94. HRMS m/z [M + H]⁺for

C₁₃H₁₅IN₄O₃calculated 403.0267, found 403.0269.

7-((3aS,3bR,4aS,5R,5aR)-5-(((tert-Butyldiphenylsilyl)oxy)methyl)-

2,2-dimethyltetrahydrocyclopropa[3,4]cyclopenta[1,2-d][1,3]-

dioxol-3b(3aH)-yl)-5-iodo-N-phenyl-7H-pyrrolo[2,3-d]pyrimidin-4-

amine (31a). To a solution of predried compound 27 (75 mg, 0.11

mmol) and aniline (12 μL, 0.13 mmol) in anhydrous THF (1 mL) at

−20 °C under nitrogen atmosphere was added potassium t-butoxide

solution (1 M in THF, 160 μL, 0.16 mmol) dropwise and stirred at

this temperature for 1 h. The reaction was quenched by adding satd

NH₄Cl (1 mL) and the products extracted into ethyl acetate (3 × 2

mL). The separated organic layers were combined, washed with brine,

dried over anhydrous Na₂SO₄, and evaporated, and the residue was

puriﬁed by silica gel ﬂash column chromatography to aﬀord 31a as a

white foam (65 mg, 80%, R_f= 0.30, TLC eluent = 10% ethyl acetate in

hexanes). ¹H NMR (400 MHz, CDCl₃) δ 8.27 (s, 1H), 8.23 (brs, 1H),

7.78−7.64 (m, 8H), 7.50−7.34 (m, 7H), 7.21 (s, 1H), 4.99 (dd, J =

1.6, 7.1 Hz, 1H), 4.59 (dd, J = 1.1, 7.0 Hz, 1H), 4.11 (d, J = 7.8 Hz,

2H), 2.55 (t, J = 7.8 Hz, 1H), 2.08 (ddd, J = 1.3, 5.4, 9.8 Hz, 1H),

1.63−1.54 (m, 4H), 1.46−1.38 (m, 1H), 1.22 (s, 3H), 1.09 (s, 9H).

HRMS m/z [M + H]⁺for C₃₈H₄₁IN₄O₃Si calculated 757.2071, found

757.2067.

7-((3aS,3bR,4aS,5R,5aR)-5-(((tert-Butyldiphenylsilyl)oxy)methyl)-

2,2-dimethyltetrahydrocyclopropa[3,4]cyclopenta[1,2-d][1,3]-

dioxol-3b(3aH)-yl)-N,5-diphenyl-7H-pyrrolo[2,3-d]pyrimidin-4-

amine (33a). A mixture of DME (1.7 mL) and ethanol (0.3 mL) was

heated to 90 °C for 10 min and cooled to room temperature. This

solution was added to a vacuum-dried mixture of 31a (75 mg, 0.13

mmol), phenylboronic acid (19 mg, 0.16 mmol), and Pd(PPh₃)₄(7.5

mg, 0.007 mmol). A sonicated aqueous saturated Na₂CO₃solution was

added to this mixture, and the ﬂask immersed in to a preheated oil

bath at 90 °C and stirred for 7 h. The solvents were evaporated under

vacuum, and the residue was partitioned between water and

dichloromethane (3 × 10 mL), the organic layer separated, dried

over anhydrous Na₂SO₄, and then evaporated. The residue was

puriﬁed by silica gel ﬂash column chromatography to aﬀord pure

product 33a as a white foam (60 mg, 86%, R_f= 0.25, TLC eluent =

10% ethyl acetate in hexanes). ¹H NMR (400 MHz, CDCl₃) δ 8.32 (s,

1H), 7.80−7.66 (m, 6H), 7.52−7.33 (m, 11H), 7.31−7.26 (m, 2H),

7.07 (s, 1H), 7.05−6.99 (m, 1H), 5.13 (dd, J = 1.5, 7.1 Hz, 1H), 4.64

(d, J = 7.0 Hz, 1H), 4.18 (d, J = 7.9 Hz, 2H), 2.58 (t, J = 7.9 Hz, 1H),

2.14 (ddd, J = 1.3, 5.5, 9.8 Hz, 1H), 1.64−1.56 (m, 4H), 1.49 (ddd, J =

1.7, 5.7, 9.8 Hz, 1H), 1.24 (s, 3H), 1.10 (s, 9H). HRMS m/z [M + H]⁺

for C₄₄H₄₆N₄O₃Si calculated 707.3417, found 707.3405.

7-((3aS,3bR,4aS,5R,5aR)-5-(((tert-Butyldiphenylsilyl)oxy)methyl)-

2,2-dimethyltetrahydrocyclopropa[3,4]cyclopenta[1,2-d][1,3]-

dioxol-3b(3aH)-yl)-N-(4-ﬂuorophenyl)-5-phenyl-7H-pyrrolo[2,3-d]-

pyrimidin-4-amine (33b). Following the procedure described for the

synthesis of 33a, compound 31b (50 mg, 0.0645 mmol) gave 33b as a

white foam (35 mg, 75%, R_f= 0.15, TLC eluent = 10% ethyl acetate in

1

hexanes). H NMR (400 MHz, CDCl₃) δ 8.25 (s, 1H), 7.73 (ddt, J =

1.7, 4.0, 8.1 Hz, 5H), 7.56−7.29 (m, 13H), 7.07 (s, 1H), 7.03−6.81

(m, 2H), 5.11 (dd, J = 1.6, 7.1 Hz, 1H), 4.69−4.57 (m, 1H), 4.16 (d, J

= 7.9 Hz, 2H), 2.58 (t, J = 7.8 Hz, 1H), 2.14 (ddd, J = 1.3, 5.4, 9.9 Hz,

1H), 1.61 (t, J = 6.0 Hz, 1H), 1.58 (s, 3H), 1.49 (ddd, J = 1.8, 5.8, 9.8

Hz, 1H), 1.24 (s, 3H), 1.09 (s, 9H). ¹⁹F NMR (376 MHz, CDCl₃) δ

−119.7. HRMS m/z [M + H]⁺for C₄₄H₄₅FN₄O₃Si calculated

725.3323, found 725.3322.

7-((3aS,3bR,4aS,5R,5aR)-5-(((tert-Butyldiphenylsilyl)oxy)methyl)-

2,2-dimethyltetrahydrocyclopropa[3,4]cyclopenta[1,2-d][1,3]-

dioxol-3b(3aH)-yl)-5-(4-ﬂuorophenyl)-N-phenyl-7H-pyrrolo[2,3-d]-

pyrimidin-4-amine (33c). Following the procedure described for the

synthesis of 33a, compound 31a (50 mg, 0.0645 mmol) gave 33c as a

white foam (40 mg, 75%, R_f= 0.15, TLC eluent = 10% ethyl acetate in

hexanes). ¹H NMR (400 MHz, CDCl₃) δ 8.31 (s, 1H), 7.78−7.70 (m,

4H), 7.42 (dddd, J = 2.0, 6.0, 8.3, 14.5 Hz, 10H), 7.28 (t, J = 7.8 Hz,

3H), 7.15 (d, J = 8.4 Hz, 2H), 7.07−7.00 (m, 2H), 5.12 (dd, J = 1.5,

7.1 Hz, 1H), 4.64 (dd, J = 1.3, 7.2 Hz, 1H), 4.19 (d, J = 7.9 Hz, 2H),

2.58 (t, J = 7.8 Hz, 1H), 2.14 (ddd, J = 1.3, 5.4, 9.8 Hz, 1H), 1.61 (t, J

= 3.8 Hz, 1H), 1.59 (s, 3H), 1.49 (ddd, J = 1.7, 5.7, 9.8 Hz, 1H), 1.24

(s, 3H), 1.10 (s, 9H). ¹⁹F NMR (376 MHz, CDCl₃) δ −114.31 (q, J =

3.7, 4.4 Hz). HRMS m/z [M + H]⁺for C₄₄H₄₅FN₄O₃Si calculated

725.3323, found 725.3317.

7-((3aS,3bR,4aS,5R,5aR)-5-(((tert-Butyldiphenylsilyl)oxy)methyl)-

2,2-dimethyltetrahydrocyclopropa[3,4]cyclopenta[1,2-d][1,3]-

dioxol-3b(3aH)-yl)-N-(4-ﬂuorophenyl)-5-iodo-7H-pyrrolo[2,3-d]-

pyrimidin-4-amine (31b). Following the procedure described for the

synthesis of 31a, compound 27 (60 mg, 0.086 mmol) gave 31b as a

white foam (47 mg, 70%, R_f= 0.2, TLC eluent = 10% ethyl acetate in

1

hexanes). H NMR (400 MHz, CDCl₃) δ 8.23 (s, 1H), 7.89 (s, 1H),

7.77−7.64 (m, 6H), 7.48−7.35 (m, 6H), 7.17 (s, 1H), 7.09 (dd, J =

8.3, 9.0 Hz, 2H), 5.01 (dd, J = 1.6, 7.1 Hz, 1H), 4.62−4.56 (m, 1H),

4.13 (d, J = 8.0 Hz, 2H), 2.54 (t, J = 8.0 Hz, 1H), 2.07 (ddd, J = 1.3,

5.4, 9.8 Hz, 1H), 1.57 (s, 3H), 1.55 (m, 1H), 1.41 (ddd, J = 1.7, 5.6,

9.8 Hz, 1H), 1.21 (s, 3H), 1.08 (s, 9H). ¹⁹F NMR (376 MHz, CDCl₃)

δ −119.03. HRMS m/z [M + H]⁺for C₃₈H₄₀FIN₄O₃Si calculated

775.1977, found 775.1971.

(1R,2S,3R,4R,5S)-4-(Hydroxymethyl)-1-(5-phenyl-4-(phenylami-

no)-7H-pyrrolo[2,3-d]pyrimidin-7-yl)bicyclo[3.1.0]hexane-2,3-diol

(34). Following the reaction procedure described for the synthesis of

28, compound 33a (17 mg, 0.024 mmol) gave 34 as a white solid

which was further puriﬁed by RP-HPLC (4.7 mg, 46%, R_f= 0.25, TLC

eluent = 5% methanol in dichloromethane; RP-HPLC (C18) R_t= 29.8

1

min, water/acetonitrile 50/50 → 10/90 in 40 min at 5 mL/min). H

(1R,2S,3R,4R,5S)-4-(Hydroxymethyl)-1-(5-iodo-4-(phenylamino)-

7H-pyrrolo[2,3-d]pyrimidin-7-yl)bicyclo[3.1.0]hexane-2,3-diol (32).

Following the reaction procedure described for the synthesis of 28,

compound 31a (20 mg, 0.035 mmol) gave 32 as a white solid which

was further puriﬁed by RP-HPLC (3.35 mg, 27%, R_f= 0.45, TLC

eluent = 10% methanol in dichloromethane; RP-HPLC (C18) R_t=

25.6 min, water/acetonitrile 50/50 → 10/90 in 40 min at 5 mL/min).

¹H NMR (400 MHz, CD₃OD) δ 8.27 (s, 1H), 7.74 (dd, J = 1.1, 8.7

Hz, 1H), 7.45 (s, 1H), 7.41−7.33 (m, 2H), 7.17−7.09 (m, 1H), 4.60

(dd, J = 2.0, 6.6 Hz, 1H), 4.08−3.99 (m, 2H), 3.81 (dd, J = 3.7, 11.5

Hz, 1H), 2.29−2.25 (m, 1H), 1.87−1.79 (m, 2H), 1.17 (ddp, J = 2.1,

3.6, 5.7 Hz, 1H). HRMS m/z [M + H]⁺for C₁₉H₁₉IN₄O₃calculated

479.0580, found 479.0589.

NMR (400 MHz, CD₃OD) δ 8.32 (s, 1H), 7.65−7.51 (m, 4H), 7.51−

7.42 (m, 3H), 7.33 (s, 1H), 7.30−7.22 (m, 2H), 7.03 (ddt, J = 1.1, 7.1,

7.7 Hz, 1H), 4.70 (dd, J = 2.0, 6.6 Hz, 1H), 4.14−4.01 (m, 2H), 3.84

(dd, J = 3.6, 11.5 Hz, 1H), 2.33−2.27 (m, 1H), 1.93−1.81 (m, 2H),

1.24 (ddd, J = 2.1, 5.1, 8.9 Hz, 1H). HRMS m/z [M + H]⁺for

C₂₅H₂₄N₄O₃calculated 429.1927, found 429.1935.

((3aS,3bR,4aS,5R,5aR)-2,2-Dimethyl-3b-(5-phenyl-4-(phenylami-

no)-7H-pyrrolo[2,3-d]pyrimidin-7-yl)hexahydrocyclopropa[3,4]-

cyclopenta[1,2-d][1,3]dioxol-5-yl)methanol (35a). Compound 33a

(60 mg, 0.089 mmol) was dissolved in anhydrous THF (2 mL),

treated with tetrabutylammonium ﬂuoride (1 M in THF, 93 μL, 0.093

mmol) at 0 °C, and stirred overnight at room temperature. The

solvent was evaporated and the residue puriﬁed by silica gel ﬂash

K

DOI: 10.1021/acs.jmedchem.6b00689

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

column chromatography to aﬀord 35a as a white foamy solid (38 mg,

96%, R_f= 0.15, TLC eluent = 30% ethyl acetate in hexanes). ¹H NMR

(400 MHz, CDCl₃) δ 8.44 (s, 1H), 7.59−7.47 (m, 6H), 7.47−7.40 (m,

1H), 7.36−7.26 (m, 2H), 7.13 (s, 1H), 7.11−7.00 (m, 2H), 5.07 (dd, J

= 1.8, 6.9 Hz, 1H), 4.90 (dd, J = 1.6, 6.9 Hz, 1H), 4.12 (d, J = 11.5 Hz,

1H), 3.89 (dt, J = 3.14, 12.1 Hz, 1H), 2.50 (dd, J = 1.2, 3.2 Hz, 1H),

2.10 (ddd, J = 1.5, 5.2, 9.7 Hz, 1H), 1.64−1.55 (m, 4H), 1.40−1.21

(m, 4H). HRMS m/z [M + H]⁺for C₂₈H₂₈N₄O₃calculated 469.2240,

found 469.2241.

((3aS,3bR,4aS,5R,5aR)-3b-(4-((4-Fluorophenyl)amino)-5-phenyl-

7 H - p y r r o l o [ 2 , 3 - d ] p y r i m i d i n - 7 - y l ) - 2 , 2 -

dimethylhexahydrocyclopropa[3,4]cyclopenta[1,2-d][1,3]dioxol-5-

yl)methanol (35b). Following the procedure described for the

synthesis of 35a, compound 33b (35 mg, 0.048 mmol) gave 35b as

a white foam (20 mg, 85%, R_f= 0.35, TLC eluent = 50% ethyl acetate

in hexanes). ¹H NMR (400 MHz, CDCl₃) δ 8.53 (s, 1H), 7.51 (d, J =

4.4 Hz, 4H), 7.44 (ddd, J = 2.5, 4.6, 9.0 Hz, 3H), 7.14 (s, 1H), 6.99 (t,

J = 8.6 Hz, 2H), 5.06 (dd, J = 1.7, 6.9 Hz, 1H), 4.89 (dd, J = 1.5, 6.9

Hz, 1H), 4.13 (dd, J = 1.4, 11.4 Hz, 1H), 3.93 (dd, J = 3.1, 11.4 Hz,

1H), 2.51 (dd, J = 1.3, 3.0 Hz, 1H), 2.10 (ddd, J = 1.5, 5.2, 9.7 Hz,

1H), 1.64−1.60 (m, 1H), 1.58 (s, 3H), 1.37−1.31 (m, 1H), 1.29 (s,

3H). ¹³C NMR (101 MHz, CDCl₃) δ 160.33, 157.92, 154.57, 151.42,

150.10, 134.83, 134.80, 134.31, 129.39, 129.27, 128.04, 124.57, 122.33,

122.25, 116.28, 115.85, 115.63, 111.32, 102.48, 85.80, 84.72, 65.72,

50.36, 46.90, 32.12, 26.59, 24.41, 16.44. ¹⁹F NMR (376 MHz, CDCl₃)

δ −118.65. HRMS m/z [M + H]⁺for C₂₈H₂₇FN₄O₃calculated

487.2145, found 487.2148.

((3aS,3bR,4aS,5R,5aR)-3b-(5-(4-Fluorophenyl)-4-(phenylamino)-

7 H - p y r r o l o [ 2 , 3 - d ] p y r i m i d i n - 7 - y l ) - 2 , 2 -

dimethylhexahydrocyclopropa[3,4]cyclopenta[1,2-d][1,3]dioxol-5-

yl)methanol (35c). Following the procedure described for the

synthesis of 35a, compound 33c (40 mg, 0.055 mmol) gave 35c as

a white foam (25 mg, 93%, R_f= 0.15, TLC eluent = 30% ethyl acetate

in hexanes). ¹H NMR (400 MHz, CDCl₃) δ 8.51 (s, 1H), 7.49 (ddd, J

= 2.1, 3.2, 8.7 Hz, 4H), 7.35−7.28 (m, 2H), 7.24−7.17 (m, 2H), 7.12

(s, 1H), 7.07 (tt, J = 1.1, 7.1 Hz, 1H), 5.05 (dd, J = 1.7, 6.9 Hz, 1H),

4.89 (dd, J = 1.4, 6.9 Hz, 1H), 4.12 (dd, J = 1.4, 11.4 Hz, 1H), 3.91

(dd, J = 3.0, 11.4 Hz, 1H), 2.10 (ddd, J = 1.5, 5.1, 9.8 Hz, 1H), 1.61 (t,

J = 5.3 Hz, 1H), 1.58 (s, 3H), 1.35−1.30 (m, 1H), 1.29 (s, 3H). ¹³C

NMR (101 MHz, CDCl₃) δ 163.88, 161.41, 154.55, 151.55, 150.10,

138.73, 131.07, 130.99, 130.31, 130.27, 129.18, 124.65, 123.71, 120.41,

116.45, 116.24, 115.12, 111.33, 102.65, 85.79, 84.72, 65.71, 50.35,

46.90, 32.11, 26.59, 24.41, 16.44. ¹⁹F NMR (376 MHz, CDCl₃) δ

−113.56. HRMS m/z [M + H]⁺for C₂₈H₂₇FN₄O₃calculated

487.2145, found 487.2146.

(3aH)-yl)-7H-pyrrolo[2,3-d]pyrimidin-4-amine (37b). Following the

procedure described for the synthesis of 37a, compound 35b (20 mg,

0.041 mmol) gave 37b as a white foam (5 mg, 57% based on 11 mg

starting material recovered, R_f= 0.75, TLC eluent = 30% ethyl acetate

1

in hexanes). H NMR (400 MHz, CDCl₃) δ 8.45 (s, 1H), 7.58−7.29

(m, 8H), 7.08 (s, 1H), 6.99−6.86 (m, 2H), 5.28 (dd, J = 1.6, 7.0 Hz,

1H), 4.54−4.46 (m, 1H), 2.41 (q, J = 7.5 Hz, 1H), 1.93 (ddd, J = 1.4,

5.4, 9.7 Hz, 1H), 1.59−1.52 (m, 7H), 1.41 (ddd, J = 1.9, 5.8, 9.8 Hz,

1H), 1.27−1.26 (m, 3H). ¹⁹F NMR (376 MHz, CDCl₃) δ −119.6.

HRMS m/z [M + H]⁺for C₂₈H₂₇FN₄O₂calculated 471.2196, found

471.2188.

5-(4-Fluorophenyl)-N-phenyl-7-((3aS,3bR,4aS,5S,5aR)-2,2,5-tri-

methyltetrahydro-cyclopropa[3,4]cyclopenta[1,2-d][1,3]dioxol-3b-

(3aH)-yl)-7H-pyrrolo[2,3-d]pyrimidin-4-amine (37c). Following the

procedure described for the synthesis of 37a, compound 35c (25 mg,

0.051 mmol) gave 37c as a white foam (7 mg, 90% based on 12 mg

starting material recovered, R_f= 0.75, TLC eluent = 30% ethyl acetate

1

in hexanes). H NMR (400 MHz, CDCl₃) δ 8.47 (s, 1H), 7.39 (brs,

4H), 7.31−7.20 (m, 3H), 7.11 (t, J = 7.8 Hz, 2H), 7.05 (s, 1H), 7.05−

7.00 (m, 1H), 5.28 (dd, J = 1.7, 7.0 Hz, 1H), 4.50 (dd, J = 1.4, 7.1 Hz,

1H), 2.41 (q, J = 7.4 Hz, 1H), 1.94 (ddd, J = 1.5, 5.4, 9.7 Hz, 1H),

1.59−1.52 (m, 7H), 1.43−1.37 (m, 1H), 1.27 (s, 3H). ¹⁹F NMR (376

MHz, CDCl₃) δ −114.30. HRMS m/z [M + H]⁺for C₂₈H₂₇FN₄O₂

calculated 471.2196, found 471.2204.

(1R,2S,3R,4S,5S)-4-Methyl-1-(5-phenyl-4-(phenylamino)-7H-

pyrrolo[2,3-d]pyrimidin-7-yl)bicyclo[3.1.0]hexane-2,3-diol (38a).

Following the reaction procedure described for the synthesis of 28,

compound 37 (5 mg, 0.011 mmol) gave 38a as a white solid which

was further puriﬁed by RP-HPLC (3 mg, 66%, R_f= 0.30, TLC eluent =

5% methanol in dichloromethane; RP-HPLC (C18) R_t= 43.8 min,

water/acetonitrile 50/50 → 10/90 in 40 min at 5 mL/min). ¹H NMR

(400 MHz, CD₃OD) δ 8.34 (s, 1H), 7.62−7.49 (m, 4H), 7.46 (dq, J =

1.0, 7.0 Hz, 3H), 7.32−7.21 (m, 3H), 7.07−6.97 (m, 1H), 3.80 (dt, J =

1.3, 6.3 Hz, 1H), 2.19 (q, J = 7.2 Hz, 1H), 1.92 (t, J = 5.2 Hz, 1H),

1.81 (ddd, J = 1.4, 4.8, 9.1 Hz, 1H), 1.43 (d, J = 7.4 Hz, 2H), 1.21

(ddd, J = 1.8, 5.4, 9.0 Hz, 1H). ¹³C NMR (101 MHz, CD₃OD) δ

155.44, 152.66, 152.18, 140.38, 136.00, 130.36, 130.28, 129.90, 128.75,

126.02, 124.25, 121.26, 116.76, 103.85, 77.89, 77.12, 43.13, 30.78,

30.35, 19.94, 16.73. HRMS m/z [M + H]⁺for C₂₅H₂₄N₄O₂calculated

413.1978, found 413.1977.

(1R,2S,3R,4S,5S)-1-(4-((4-Fluorophenyl)amino)-5-phenyl-7H-

pyrrolo[2,3-d]pyrimidin-7-yl)-4-methylbicyclo[3.1.0]hexane-2,3-diol

(38b). Following the procedure described for the synthesis of 28,

compound 37b (5 mg, 0.011 mmol) gave 38b as a white solid which

was puriﬁed further by RP-HPLC (2.4 mg, 53%, R_f= 0.3, TLC eluent

= 5% methanol in dichloromethane; RP-HPLC (C18) R_t= 42.7 min,

water/acetonitrile 50/50 → 10/90 in 40 min at 5 mL/min). ¹H NMR

(400 MHz, CD₃OD) δ 8.32 (s, 1H), 7.61−7.50 (m, 4H), 7.50−7.40

(m, 3H), 7.24 (s, 1H), 7.05−6.97 (m, 2H), 4.85−4.82 (m, 1H), 3.80

(dt, J = 1.3, 6.3 Hz, 1H), 2.23−2.15 (m, 1H), 1.95−1.89 (m, 1H), 1.81

(ddd, J = 1.4, 4.7, 9.3 Hz, 1H), 1.43 (d, J = 7.4 Hz, 3H), 1.21 (ddd, J =

1.9, 5.4, 9.1 Hz, 1H). ¹⁹F NMR (376 MHz, CD₃OD) δ −121.81 (tt, J

= 4.8, 8.4 Hz). HRMS m/z [M + H]⁺for C₂₅H₂₃FN₄O₂calculated

431.1883, found 431.1881.

(1R,2S,3R,4S,5S)-1-(5-(4-Fluorophenyl)-4-(phenylamino)-7H-

pyrrolo[2,3-d]pyrimidin-7-yl)-4-methylbicyclo[3.1.0]hexane-2,3-diol

(38c). Following the procedure described for the synthesis of 28,

compound 37c (7 mg, 0.015 mmol) gave 38c as a white solid which

was puriﬁed further by RP-HPLC (4 mg, 63%, R_f= 0.3, TLC eluent =

5% methanol in dichloromethane; RP-HPLC (C18) R_t= 41.9 min,

water/acetonitrile 50/50 → 10/90 in 40 min at 5 mL/min). 38c was

soluble in chloroform but not methanol. ¹H NMR (400 MHz, DMSO-

d₆) δ 8.37 (s, 1H), 7.63−7.48 (m, 5H), 7.37−7.23 (m, 5H), 6.98 (tt, J

= 1.2, 7.4 Hz, 1H), 4.78−4.73 (m, 1H), 3.63 (d, J = 6.1 Hz, 1H),

2.10−2.02 (m, 1H), 1.72 (t, J = 5.0 Hz, 1H), 1.67−1.62 (m, 1H), 1.37

(d, J = 7.4 Hz, 3H), 1.24−1.16 (m, 1H). ¹⁹F NMR (376 MHz, DMSO-

d₆) δ −115.78 (tt, J = 5.5, 9.1 Hz). HRMS m/z [M + H]⁺for

C₂₅H₂₃FN₄O₂calculated 431.1883, found 431.1887.

N , 5 - D i p h e n y l - 7 - ( ( 3 a S , 3 b R , 4 a S , 5 S , 5 a R ) - 2 , 2 , 5 -

trimethyltetrahydrocyclopropa[3,4]-cyclopenta[1,2-d][1,3]dioxol-

3b(3aH)-yl)-7H-pyrrolo[2,3-d]pyrimidin-4-amine (37a). To a sol-

ution of compound 35a (10 mg, 0.022 mmol) and DMAP (8 mg,

0.065 mmol) in anhydrous acetonitrile (0.5 mL) was added O-p-

tolylchlorothionoformate (36, 8 μL, 0.052 mmol) dropwise at room

temperature. The mixture was stirred for 4 h, and then the volatile

organics were evaporated under reduced pressure. The residue was

suspended in ethyl acetate (5 mL) and washed with water and brine.

The organic layer was dried over anhydrous sodium sulfate and the

solvent evaporated to dryness. The residue obtained was dried under

high vacuum for 2 h and then suspended in anhydrous toluene (1.5

mL). Tributyltin hydride (17 μL, 0.065 mmol) and azoisobutyronitrile

(AIBN, 10 mg, 0.065 mmol) was added, and the ﬂask was immersed in

to an oil bath at 120 °C for 2 h. Volatile materials were evaporated,

and the residue was puriﬁed by silica gel ﬂash column chromatography

to aﬀord compound 37a as a white foam (5 mg, 52%, R_f= 0.60, TLC

1

eluent = 30% ethyl acetate in hexanes). H NMR (400 MHz, CDCl₃)

δ 8.49 (s, 1H), 7.58−7.46 (m, 6H), 7.46−7.38 (m, 1H), 7.33−7.26 (m,

2H), 7.05 (s, 1H), 7.04−6.99 (m, 1H), 6.94 (s, 1H), 5.31 (dd, J = 1.7,

7.0 Hz, 1H), 4.50 (dd, J = 1.4, 7.0 Hz, 1H), 2.40 (q, J = 7.5 Hz, 1H),

1.92 (ddd, J = 1.4, 5.4, 9.7 Hz, 1H), 1.56 (d, J = 7.3 Hz, 3H), 1.55 (s,

3H), 1.41 (ddd, J = 1.8, 5.6, 9.7 Hz, 1H), 1.27 (s, 3H). HRMS m/z [M

+ H]⁺for C₂₈H₂₈N₄O₂calculated 453.2291, found 453.2285.

(3aS,3bR,4aS,5R,5aR)-3b-(4-Amino-5-iodo-7H-pyrrolo[2,3-d]-

pyrimidin-7-yl)-2,2-dimethylhexahydrocyclopropa[3,4]cyclopenta-

[1,2-d][1,3]dioxol-5-yl)methanol (39). Following the reaction proce-

N-(4-Fluorophenyl)-5-phenyl-7-((3aS,3bR,4aS,5S,5aR)-2,2,5-tri-

methyltetrahydro-cyclopropa[3,4]cyclopenta[1,2-d][1,3]dioxol-3b-

L

DOI: 10.1021/acs.jmedchem.6b00689

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

dure described for the synthesis of 35, compound 29 (65 mg, 0.095

mmol) gave 39 as a white solid (45 mg, quantitative yield, R_f= 0.20,

TLC eluent = 5% methanol in dichloromethane). ¹H NMR (400

MHz, CDCl₃) δ 8.23 (s, 1H), 7.19 (s, 1H), 6.81 (dd, J = 1.8, 12.5 Hz,

1H), 5.70 (s, 3H), 4.94 (dd, J = 1.8, 6.9 Hz, 1H), 4.83 (dd, J = 1.5, 6.9

Hz, 1H), 4.05 (dt, J = 1.5, 11.5 Hz, 1H), 3.91−3.78 (m, 1H), 2.51−

2.42 (m, 1H), 2.01 (ddd, J = 1.5, 5.2, 9.7 Hz, 1H), 1.58−1.53 (m, 4H),

1.27 (s, 3H), 1.23 (ddd, J = 1.9, 5.5, 9.7 Hz, 1H). HRMS m/z [M +

H]⁺for C₁₆H₁₉IN₄O₃calculated 443.0580, found 443.0585.

((3aS,3bS,4aS,5R,5aR)-3b-(5-Iodo-4-((triphenyl-l5-

phosphanylidene)amino)-7H-pyrrolo[2,3-d]pyrimidin-7-yl)-2,2-

dimethylhexahydrocyclopropa[3,4]cyclopenta[1,2-d][1,3]dioxol-5-

yl)methyl Diphenyl Phosphate (40). To an ice-cold solution of 39

(20 mg, 0.045 mmol) and triphenylphosphine (17 mg, 0.063 mmol) in

anhydrous THF (1 mL) was added diisopropyl azodicarboxylate

(DIAD, 13 μL, 0.063 mmol) dropwise. After 5 min, diphenylphos-

phorylazide (DPPA, 14 μL, 0.063 mmol) was added and stirred at 0

°C for 30 min and then at room temperature overnight. The reaction

mixture was heated to 60 °C for 4 h. Solvents were evaporated to

dryness, and residue was suspended in ethyl acetate, washed with water

and brine, and dried over anhydrous Na₂SO₄. Solvent was evaporated

and the residue puriﬁed by silica gel ﬂash column chromatography to

aﬀord the product 40 as a white solid (R_f= 0.80, TLC eluent = 5%

methanol in dichloromethane). ¹H NMR (400 MHz, CD₃OD) δ

7.98−7.88 (m, 4H), 7.86 (s, 1H), 7.63 (s, 1H), 7.55 (ddd, J = 1.6, 5.2,

7.3 Hz, 2H), 7.47 (ddd, J = 3.3, 6.8, 8.7 Hz, 5H), 7.36−7.08 (m, 12H),

7.07−6.98 (m, 2H), 4.89 (dd, J = 1.6, 7.2 Hz, 1H), 4.69−4.63 (m,

2H), 4.54−4.48 (m, 1H), 2.55 (t, J = 6.9 Hz, 1H), 1.83−1.75 (m, 1H),

1.57−1.40 (m, 5H), 1.14 (s, 3H). HRMS m/z [M + H]⁺for

C₄₆H₄₂IN₄O₆P₂calculated 935.1624, found 935.1628.

the reaction procedure described for the synthesis of 28, compound

42 (8 mg, 0.017 mmol) gave 43 as a white solid which was further

puriﬁed by RP-HPLC (3.1 mg, 42%, R_f= 0.25, TLC eluent = 5%

methanol in dichloromethane; RP-HPLC (C18) R_t= 28.5 min, water/

1

acetonitrile 75/25 → 25/75 in 40 min at 5 mL/min). H NMR (400

MHz, CD₃OD) δ 8.11 (s, 1H), 7.31 (s, 1H), 4.66 (dd, J = 1.9, 6.5 Hz,

1H), 3.94−3.84 (m, 2H), 3.73 (dd, J = 7.4, 12.2 Hz, 1H), 2.30−2.22

(m, 1H), 1.91−1.81 (m, 2H), 1.23−1.17 (m, 1H). HRMS m/z [M +

H]⁺for C₁₃H₁₄IN₇O₂calculated 428.0332, found 428.0336.

7 - ( ( 3 a S , 3 b S , 4 a S , 5 R , 5 a R ) - 5 - ( A m i n o m e t h y l ) - 2 , 2 -

dimethyltetrahydrocyclopropa[3,4]-cyclopenta[1,2-d][1,3]dioxol-

3b(3aH)-yl)-5-iodo-7H-pyrrolo[2,3-d]pyrimidin-4-amine (44). Com-

pound 42 (8 mg, 0.017 mmol) was dissolved in anhydrous THF (0.5

mL) and to this was added triphenylphosphine (9 mg, 0.034 mmol)

and then stirred at room temperature overnight. Ammonium

hydroxide (25%, 30 μL) was added and heated to 65 °C under reﬂux

condenser for 4 h. Solvent was evaporated, and the residue was

puriﬁed by silica gel ﬂash column chromatography to aﬀord 44 (6 mg,

80%, R_f= 0.10, TLC eluent = 10% methanol and 1% triethylamine in

dichloromethane). HRMS m/z [M + H]⁺for C₁₆H₂₀IN₅O₂calculated

442.0740, found 442.0736.

(1S,2S,3R,4R,5S)-1-(4-Amino-5-iodo-7H-pyrrolo[2,3-d]pyrimidin-

7-yl)-4-(aminomethyl)bicyclo[3.1.0]hexane-2,3-diol acetate (45).

Following the reaction procedure described for the synthesis of 28,

compound 44 (6 mg, 0.0136 mmol) after puriﬁcation by RP-HPLC

gave 45 as a white solid (1.91 mg, 35%. RP-HPLC (C18) R_t= 29.6

min, 10 mM triethylammonium acetate in water/acetonitrile 90/10 →

65/35 in 40 min at 5 mL/min). ¹H NMR (400 MHz, CD₃OD) δ 8.11

(s, 1H), 7.34 (s, 1H), 4.56 (d, J = 6.8 Hz, 1H), 4.00 (d, J = 6.6 Hz,

1H), 3.38 (t, J = 5.9 Hz, 2H), 2.43 (t, J = 5.0 Hz, 1H), 1.95−1.70 (m,

5H), 1.27 (t, J = 7.2 Hz, 1H). HRMS m/z [M + H]⁺for C₁₃H₁₆IN₅O₂

calculated 402.0427, found 402.0421.

N-(7-((3aS, 3bS, 4aS, 5R, 5aR)-5-(Azidomethyl)-2, 2-

dimethyltetrahydrocyclopropa[3,4]-cyclopenta[1,2-d][1,3]dioxol-

3b(3aH)-yl)-5-iodo-7H-pyrrolo[2,3-d]pyrimidin-4-yl)-1,1,1-triphenyl-

l5-phosphanimine (41). The intermediate compound 40 was

dissolved in anhydrous DMF and treated with excess of NaN₃. The

mixture was heated to 65 °C overnight. Solvent was evaporated, and

the residue waspuriﬁed by silica gel ﬂash column chromatography to

aﬀord the title compound 41 as a white solid (R_f= 0.30, TLC eluent =

(3aS,3bS,4aS,5S,5aR)-3b-(4-Amino-5-iodo-7H-pyrrolo[2,3-d]-

pyrimidin-7-yl)-2,2-dimethylhexahydrocyclopropa[3,4]cyclopenta-

[1,2-d][1,3]dioxole-5-carboxylic Acid (46). To a solution of

compound 39 (40 mg, 0.09 mmol) in acetonitrile−water (4:1, 2.0

mL) was added bis(acetoxy)iodobenzene (BAIB, 64 mg, 0.20 mmol)

and 2,2,6,6-tetramethylpiperidine-1-oxyl (TEMPO, 12 mg, 0.077

mmol) at room temperature. After stirring for 18 h at room

temperature, the volatiles were evaporated under high vacuum. The

residue was puriﬁed by silica gel column chromatography to aﬀord 46

as a light-yellow solid (20 mg, 49%, R_f= 0.20, TLC eluent = 5%

1

20% ethyl acetate in hexanes). H NMR (400 MHz, CDCl₃) δ 8.02−

7.91 (m, 7H), 7.55−7.49 (m, 3H), 7.48−7.40 (m, 6H), 7.05 (s, 1H),

5.07 (dd, J = 1.6, 7.1 Hz, 1H), 4.54 (dd, J = 1.4, 7.1 Hz, 1H), 3.85 (qd,

J = 8.1, 12.1 Hz, 2H), 2.41 (t, J = 8.1 Hz, 1H), 1.86 (ddd, J = 1.4, 5.4,

9.7 Hz, 1H), 1.55 (s, 3H), 1.49 (t, J = 5.6 Hz, 1H), 1.41 (ddd, J = 1.7,

5.8, 9.7 Hz, 1H), 1.22 (s, 3H). HRMS m/z [M + H]⁺for

C₃₄H₃₂IN₇O₂P calculated 728.1400, found 728.1406.

1

methanol in dichloromethane). H NMR (400 MHz, CDCl₃) δ 8.23

(s, 1H), 7.21 (s, 1H), 6.10 (s, 2H), 5.02 (dd, J = 1.7, 7.1 Hz, 1H), 4.94

(dt, J = 1.2, 6.8 Hz, 1H), 3.30 (s, 1H), 2.31 (ddd, J = 1.5, 5.3, 10.0 Hz,

1H), 1.67 (t, J = 5.7 Hz, 1H), 1.57 (s, 3H), 1.48 (ddd, J = 1.8, 6.0, 9.9

Hz, 1H), 1.25 (s, 4H). HRMS m/z [M + H]⁺for C₁₆H₁₇IN₄O₄

calculated 457.0373, found 457.0381.

7 - ( ( 3 a S , 3 b S , 4 a S , 5 R , 5 a R ) - 5 - ( A z i d o m e t h y l ) - 2 , 2 -

dimethyltetrahydrocyclopropa[3,4]-cyclopenta[1,2-d][1,3]dioxol-

3b(3aH)-yl)-5-iodo-7H-pyrrolo[2,3-d]pyrimidin-4-amine (42). To a

solution of compound 41 (15 mg, 0.034 mmol) in anhydrous pyridine

(0.5 mL) under a nitrogen atmosphere was added methanesulfonyl

chloride (MsCl, 6 μL, 0.068 mmol) at 0 °C. After stirring the reaction

mixture for 4 h, volatiles were evaporated under vacuum and the

residue partitioned between water and ethyl acetate (3 × 5 mL). The

organic layer was separated, combined, dried over anhydrous Na₂SO₄,

and the solvent evaporated. The residue was dissolved in anhydrous

DMF (0.5 mL), sodium azide (22 mg, 0.34 mmol) was added, and the

mixture was heated to 60 °C for 4 h. The solvent was evaporated

under high vacuum, and the residue was suspended in ethyl acetate,

washed with water and brine, and dried over anhydrous Na₂SO₄. The

residue after evaporation was puriﬁed by silica gel ﬂash column

chromatography to aﬀord title compound 42 as a white foamy solid

(15 mg, 95%, R_f= 0.45, TLC eluent = 5% methanol in

dichloromethane). ¹H NMR (400 MHz, CDCl₃) δ 8.24 (s, 1H),

7.12 (s, 1H), 5.63 (s, 2H), 5.08 (dd, J = 1.7, 7.1 Hz, 1H), 4.57 (dd, J =

1.4, 7.1 Hz, 1H), 3.92 (dd, J = 7.9, 12.1 Hz, 1H), 3.83 (dd, J = 8.3, 12.1

Hz, 1H), 2.46 (t, J = 8.1 Hz, 1H), 1.95 (ddd, J = 1.4, 5.4, 9.8 Hz, 1H),

1.59−1.52 (m, 4H), 1.41 (ddd, J = 1.7, 5.9, 9.8 Hz, 1H), 1.25 (s, 3H).

HRMS m/z [M + H]⁺for C₁₆H₁₈IN₇O₂calculated 468.0645, found

468.0647.

1,3-Bis((1S,2S,3R,4S,5S)-5-(4-Amino-5-iodo-7H-pyrrolo[2,3-d]-

1

pyrimidin-7-yl)-3,4-dihydroxybicyclo[3.1.0]hexan-2-yl)urea (47). H

NMR (400 MHz, CD₃OD) δ 8.21 (s, 1H), 7.35 (s, 1H), 4.74 (dd, J =

2.1, 6.3 Hz, 1H), 4.19 (s, 1H), 3.92−3.86 (m, 1H), 2.91−2.77 (m,

6H), 2.04 (t, J = 5.4 Hz, 1H), 1.88 (s, 4H), 1.28 (td, J = 3.6, 6.3, 7.2

Hz, 2H), 1.16 (t, J = 7.3 Hz, 9H). HRMS m/z [M + H]⁺for

C₂₅H₂₆I₂N₁₀O₅calculated 801.0255, found 801.0250.

Benzyl ((3aS,3bS,4aS,5S,5aR)-3b-(4-Amino-5-iodo-7H-pyrrolo-

[2,3-d]pyrimidin-7-yl)-2,2-dimethylhexahydrocyclopropa[3,4]-

cyclopenta[1,2-d][1,3]dioxol-5-yl)carbamate (48). To a solution of

compound 46 (20 mg, 0.0437 mmol) in anhydrous THF (0.5 mL)

was added triethylamine (12 μL, 0.0875 mmol) and diphenylphos-

phoryl azide (DPPA, 15 μL, 0.065 mmol) at room temperature. After

stirring overnight at room temperature, the solvent was evaporated

under reduced pressure and the residue was dissolved in anhydrous

toluene (1.0 mL). To this mixture was added benzyl alcohol (50 μL,

0.48 mmol) and heated to 120 °C for 5 h. The volatile compounds

were evaporated under reduced pressure and the residue puriﬁed by

silica gel column chromatography to aﬀord 48 as a light-yellow solid

(4.0 mg, 17%, R_f= 0.65, TLC eluent = 5% methanol in

dichloromethane). ¹H NMR (400 MHz, CDCl₃) δ 8.22 (s, 1H),

7.56 (d, J = 8.8 Hz, 2H), 7.49−7.29 (m, 4H), 7.15 (s, 1H), 5.78 (s,

2H), 5.25−5.12 (m, 2H), 4.96 (dd, J = 1.9, 6.7 Hz, 1H), 4.59 (dd, J =

(1S,2S,3R,4R,5S)-1-(4-Amino-5-iodo-7H-pyrrolo[2,3-d]pyrimidin-

7-yl)-4-(azidomethyl)bicyclo[3.1.0]hexane-2,3-diol (43). Following

M

DOI: 10.1021/acs.jmedchem.6b00689

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

1.9, 6.6 Hz, 1H), 4.36 (d, J = 8.7 Hz, 1H), 1.97 (ddd, J = 1.9, 5.8, 10.3

Hz, 1H), 1.66 (t, J = 6.0 Hz, 1H), 1.55 (s, 3H), 1.41 (ddd, J = 1.9, 6.3,

10.3 Hz, 1H), 1.22 (s, 3H). HRMS m/z [M + H]⁺for C₂₃H₂₄IN₅O₄

calculated 562.0951, found 562.0955.

(400 MHz, CDCl₃) δ 8.53 (s, 1H), 7.66−7.56 (m, 4H), 7.50 (d, J =

8.0 Hz, 3H), 7.45−7.34 (m, 7H), 7.34−7.24 (m, 6H), 7.07−7.00 (m,

1H), 5.45 (s, 1H), 5.36 (dd, J = 1.4, 7.2 Hz, 1H), 4.62 (dd, J = 1.5, 7.3

Hz, 1H), 4.29 (d, J = 11.0 Hz, 1H), 3.41 (d, J = 11.0 Hz, 1H), 1.76

(ddd, J = 1.5, 4.4, 9.3 Hz, 1H), 1.55 (s, 3H), 1.27 (s, 3H), 1.18 (dd, J =

4.4, 5.6 Hz, 1H), 0.94 (s, 9H), 0.92−0.83 (m, 1H). HRMS m/z [M +

H]⁺for C₄₄H₄₆N₄O₃Si calculated 707.3417, found 707.3414.

(1R,2R,3S,4R,5S)-1-(Hydroxymethyl)-4-(5-phenyl-4-(phenylami-

no)-7H-pyrrolo[2,3-d]pyrimidin-7-yl)bicyclo[3.1.0]hexane-2,3-diol

(55). Following the procedure described for the synthesis of 28,

compound 54 (15 mg, 0.021 mmol) aﬀorded 55 as a white solid which

was further puriﬁed by RP-HPLC (7 mg, 77%, R_f= 0.15, TLC eluent =

5% methanol in dichloromethane; RP-HPLC (C18) R_t= 24.9 min,

water/acetonitrile 50/50 → 10/90 in 40 min at 5 mL/min). ¹H NMR

(400 MHz, 9:1 CD₃OD−CDCl₃) δ 8.34 (s, 1H), 7.75 (d, J = 1.8 Hz,

1H), 7.63−7.40 (m, 7H), 7.28 (t, J = 7.7 Hz, 2H), 7.03 (td, J = 1.1, 7.4

Hz, 1H), 5.11 (s, 1H), 4.80 (dd, J = 1.5, 6.8 Hz, 1H), 4.28 (d, J = 11.7

Hz, 1H), 3.87 (dd, J = 1.3, 6.8 Hz, 1H), 3.28 (d, J = 6.9 Hz, 1H), 1.63

(dd, J = 3.8, 8.9 Hz, 1H), 1.54 (t, J = 4.6 Hz, 1H), 0.77−0.69 (m, 1H).

¹³C NMR (101 MHz, 9:1 CD₃OD−CDCl₃) δ 155.20, 151.61, 150.10,

(1S,2S,3R,4S,5S)-4-Amino-1-(4-amino-5-iodo-7H-pyrrolo[2,3-d]-

pyrimidin-7-yl)bicyclo[3.1.0]hexane-2,3-diol (49). To a solution of

compound 48 (3.0 mg, 5.34 μmol) in anhydrous dichloromethane (0.1

mL) was added 33% HBr in acetic acid (5 μL, 27.5 μmol) at 0 °C and

stirred at room temperature for 0.5 h. The volatiles were evaporated

under vacuum, and the residual acid was chased by evaporating

repeatedly with dichloromethane (HRMS indicated 2′,3′- isopropyli-

dene deprotected compound). The above procedure was repeated and

stirred for 1.5 h [Note: a straight 2 h reaction degraded the starting

material]. The volatiles were evaporated under vacuum, and the

residual acid was removed by evaporating repeatedly with dichloro-

methane. The product mixture was dissolved in methanol and

subjected to RP-HPLC (C18) puriﬁcation to aﬀord 49 and 50 as a

white solid. Compound 49 acetate salt (0.35 mg, 17%; RP-HPLC

(C18) R_t= 29.4 min, 10 mM triethylammonium acetate in water/

1

acetonitrile 90/10 → 65/35 in 40 min at 5 mL/min). H NMR (400

MHz, CD₃OD) δ 8.13 (s, 1H), 7.33 (s, 1H), 4.81 (dd, J = 2.2, 6.4 Hz,

1H), 3.90 (dd, J = 1.9, 6.4 Hz, 1H), 1.95 (t, J = 5.5 Hz, 1H), 1.88−1.81

(m, 1H), 1.35−1.25 (m, 2H). HRMS m/z [M + H]⁺for C₁₂H₁₄IN₅O₂

calculated 388.0270, found 388.0268. (1R,2S,3R,4S,5S)-4-amino-1-(4-

amino-7H-pyrrolo[2,3-d]pyrimidin-7-yl)bicyclo[3.1.0]hexane-2,3-diol

(50): RP-HPLC (C18) R_t= 15.8 min, 10 mM triethylammonium

acetate in water/acetonitrile 90/10 → 65/35 in 40 min at 5 mL/min).

¹H NMR (400 MHz, CD₃OD) δ 8.10 (s, 1H), 7.13 (d, J = 3.6 Hz,

1H), 6.51 (d, J = 3.6 Hz, 1H), 4.86−4.83 (m, 1H), 3.96−3.88 (m,

1H), 1.95 (dd, J = 5.0, 5.8 Hz, 1H), 1.83 (ddd, J = 1.9, 5.0, 9.5 Hz,

1H), 1.30 (ddd, J = 2.0, 5.7, 9.3 Hz, 1H). HRMS m/z [M + H]⁺for

C₁₂H₁₅N₅O₂calculated 262.1299, found 262.1303.

7-((3aR,3bR,4aS,5R,5aS)-3b-(((tert-Butyldiphenylsilyl)oxy)-

methyl)-2,2-dimethylhexahydrocyclopropa[3,4]cyclopenta[1,2-d]-

[1,3]dioxol-5-yl)-4-chloro-5-iodo-7H-pyrrolo[2,3-d]pyrimidine (52).

Compound 51 (46 mg, 0.105 mmol) was dried by evaporating with

anhydrous toluene (2 × 1 mL). To the dried compound was added

anhydrous THF (1 mL), triphenylphosphine (83 mg, 0.315 mmol),

and 6-chloro-7-iodo-7-deazapurine (88 mg, 0.315 mmol). The mixture

was cooled to 0 °C and diisopropylazodicarboxylate (62 μL, 0.315

mmol) added dropwise. After 30 min, the reaction mixture was

brought to room temperature and stirred overnight (18 h). The

volatiles were evaporated under reduced pressure and the residue

puriﬁed by silica gel chromatography to aﬀord 52 as light-yellow foam

(65 mg, 88%, R_f= 0.55, TLC eluent = 20% ethyl acetate in hexanes).

¹H NMR (400 MHz, CDCl₃) δ 8.61 (s, 1H), 7.74 (s, 1H), 7.71−7.63

(m, 4H), 7.47−7.34 (m, 6H), 5.35 (s, 1H), 5.31 (dd, J = 1.4, 7.3 Hz,

1H), 4.51−4.45 (m, 1H), 4.29 (dd, J = 0.7, 11.0 Hz, 1H), 3.42 (d, J =

11.0 Hz, 1H), 1.60 (ddd, J = 1.5, 4.4, 9.3 Hz, 1H), 1.53 (s, 3H), 1.23

(s, 3H), 1.12−1.15 (m, 1H), 1.16 (s, 9H), 0.92−0.84 (m, 1H). HRMS

m/z [M + H]⁺for C₃₂H₃₅ClIN₃O₃Si calculated 700.1259, found

700.1266.

139.46, 135.42, 129.94, 129.93, 129.63, 128.42, 124.27, 122.91, 121.24,

116.99, 103.22, 77.84, 72.04, 64.34, 63.18, 37.74, 24.46, 11.96. HRMS

m/z [M + H]⁺for C₂₅H₂₄N₄O₃calculated 429.1927, found 429.1924.

7-((3aR,3bR,4aS,5R,5aS)-3b-(((tert-Butyldiphenylsilyl)oxy)-

methyl)-2,2-dimethylhexahydrocyclopropa[3,4]cyclopenta[1,2-d]-

[1,3]dioxol-5-yl)-5-iodo-7H-pyrrolo[2,3-d]pyrimidin-4-amine (56).

Following the procedure described for the synthesis of 29, compound

52 (20 mg, 0.028 mmol) gave 56 as a white foam (16 mg, 82%, R_f=

1

0.45, TLC eluent = 50% ethyl acetate in hexanes). H NMR (400

MHz, CDCl₃) δ 8.25 (s, 1H), 7.76−7.60 (m, 4H), 7.49 (s, 1H), 7.47−

7.33 (m, 6H), 6.20 (s, 2H), 5.37−5.26 (m, 2H), 4.53−4.40 (m, 1H),

4.30 (d, J = 11.0 Hz, 1H), 3.36 (d, J = 11.0 Hz, 1H), 1.59 (ddd, J = 1.5,

4.4, 9.3 Hz, 1H), 1.51 (s, 3H), 1.25−1.21 (m, 3H), 1.16 (s, 9H), 1.13

(dd, J = 4.5, 5.7 Hz, 1H), 0.82 (ddd, J = 1.5, 5.6, 9.3 Hz, 1H). HRMS

m/z [M + H]⁺for C₃₂H₃₇IN₄O₃Si calculated 681.1758, found

681.1769.

(1R,2R,3S,4R,5S)-4-(4-Amino-5-iodo-7H-pyrrolo[2,3-d]pyrimidin-

7-yl)-1-(hydroxymethyl)bicyclo[3.1.0]hexane-2,3-diol (57). Follow-

ing the procedure described for the synthesis of 28, compound 56 (16

mg, 0.019 mmol) aﬀorded 57 as a white solid (7.5 mg, 79%, R_f= 0.25,

1

TLC eluent = 10% methanol in dichloromethane). H NMR (400

MHz, CD₃OD) δ 8.11 (s, 1H), 7.79 (s, 1H), 4.99 (s, 1H), 4.72 (dd, J

= 1.6, 6.7 Hz, 1H), 4.26 (dd, J = 1.0, 11.7 Hz, 1H), 3.74 (dt, J = 1.2,

6.7 Hz, 1H), 3.33 (d, J = 15.5 Hz, 1H), 3.27 (d, J = 11.7 Hz, 1H),

1.57−1.43 (m, 2H), 0.70 (ddd, J = 1.8, 4.5, 8.0 Hz, 1H). ¹³C NMR

(101 MHz, CD₃OD) δ 158.81, 152.52, 150.32, 129.15, 105.15, 78.26,

72.29, 64.43, 63.50, 50.37, 37.97, 24.66, 12.12. HRMS m/z [M + H]⁺

for C₁₃H₁₅IN₄O₃calculated 403.0267, found 403.0271.

Quantiﬁcation of hAdK Inhibition. The nucleoside derivatives

were tested for inhibitory activity at hAdK using a commercial kit for

the measurement of hAdK activity (ADK Phosphorylation Assay Kit,

ref. no. K0507-02, NovoCib, Lyon, France). For use in the assay,

compounds 28, 30, 32, 43, 45, and 49 were dissolved in 60% DMSO

in puriﬁed water, 34 was prepared in 80% DMSO solution, and 38a

was prepared in 88% DMSO solution, while 38b, 38c, 55, and 57 were

prepared in 100% DMSO. Reactions were performed at 37 °C in a

total volume of 200 μL containing 100 mM Tris/HCl, 250 mM KCl,

10 mM MgCl₂, 2.5 mM NAD, 2.75 mM ATP, 20 mU/mL IMPDH,

2.2 mU hAdK. All assays were run in triplicate and then replicated on a

diﬀerent plate. hAdK activity was assayed by measuring speciﬁc

absorption at 340 nm at 5 min time intervals using a Spectromax plate

reader (Molecular Devices, CA, USA). All reactions were started by

the addition of 2.5 mM inosine (t₀) and maintained for a total of 4 h.

Baseline absorption prior to the addition of inosine was subtracted

from all measurements. Percent inhibition of hAdK activity was

calculated based on the slope during the linear phase of the

progression of the enzymatic reaction (quantiﬁed between 0 and 40

min after inosine addition) and normalized to the slope of the

enzymatic reaction in the presence of the vehicle DMSO (1% v/v ﬁnal

concentration of reaction mix).

7-((3aR,3bR,4aS,5R,5aS)-3b-(((tert-Butyldiphenylsilyl)oxy)-

methyl)-2,2-dimethylhexahydrocyclopropa[3,4]cyclopenta[1,2-d]-

[1,3]dioxol-5-yl)-5-iodo-N-phenyl-7H-pyrrolo[2,3-d]pyrimidin-4-

amine (53). Following the procedure described for the synthesis of

31a, compound 52 (125 mg, 0.178 mmol) gave 53 as a white foam (85

1

mg, 63%, R_f= 0.3, TLC eluent = 10% ethyl acetate in hexanes). H

NMR (400 MHz, CDCl₃) δ 8.47 (s, 1H), 8.19 (s, 1H), 7.79−7.64 (m,

6H), 7.53 (s, 1H), 7.48−7.34 (m, 8H), 7.20−7.13 (m, 1H), 5.36 (s,

1H), 5.33 (dd, J = 1.4, 7.2 Hz, 1H), 4.49 (dd, J = 1.5, 7.2 Hz, 1H), 4.32

(d, J = 11.2 Hz, 1H), 3.36 (d, J = 11.1 Hz, 1H), 1.62 (ddd, J = 1.5, 4.4,

9.3 Hz, 1H), 1.52 (s, 3H), 1.24 (s, 3H), 1.17 (s, 8H), 1.16−1.13 (m,

1H), 0.83 (ddd, J = 1.5, 5.6, 9.3 Hz, 1H). HRMS m/z [M + H]⁺for

C₃₈H₄₁IN₄O₃Si calculated 757.2071, found 757.2072.

7-((3aR,3bR,4aS,5R,5aS)-3b-(((tert-Butyldiphenylsilyl)oxy)-

methyl)-2,2-dimethylhexahydrocyclopropa[3,4]cyclopenta[1,2-d]-

[1,3]dioxol-5-yl)-N,5-diphenyl-7H-pyrrolo[2,3-d]pyrimidin-4-amine

(54). Following the procedure described for the synthesis of 33a,

compound 53 (85 mg, 0.112 mmol) gave 54 as a white foam (65 mg,

82%, R_f= 0.25, TLC eluent = 10% ethyl acetate in hexanes). ¹H NMR

N

DOI: 10.1021/acs.jmedchem.6b00689

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Statistical Analysis. Two-way ANOVA was used to determine

signiﬁcances between compound and concentration with percent

residual hAdK activity based on slope as the primary dependent

measure. Slopes were calculated for the ﬁrst 40 min of activity for each

compound and concentration combination. Post hoc analyses were

used to determine speciﬁc between group diﬀerences relative to 2

using Bonferroni multiple corrections test.

implemented in VMD⁷²exploiting NAMD 2.10.⁷³From the IE_tot

values so obtained, we derived a modiﬁed version of the dynamic

scoring function⁷⁴(herein deﬁned as wDSF_tot) by computing the

cumulative sum of ligand−protein IE_totand dividing the value for the

ligand RMSD with respect to the starting conformation as follows:

n

IE_tot

RMSD

wDSF

=

∑

tot

Molecular Modeling. Computational Facilities. Ligand geometry

optimization and molecular docking simulations were carried out using

a 8 Intel Xeon X5482 CPU workstation. Membrane MD simulations

were run on four NVIDIA Tesla K20X by exploiting the computa-

tional resources of the NIH HPC Biowulf cluster (http://biowulf.nih.

gov).

i=1

wDSF_totslope values were derived as previously reported,⁷⁴and for

each ligand−protein system the replica that returned the highest

absolute slope value (Supporting Information, Table S1) was selected.

IE vs simulation time graphs were generated with Gnuplot.⁷⁵

Sugar pseudorotation phase angle⁷⁶(P) during MD simulations was

computed with a in house tcl script with the endocyclic torsions υ₀−υ₄

angles deﬁned as reported in Supporting Information, Figure S6, by

using the following equation:

Ligand Preparation. The compounds were built using Maestro⁵⁶

and subjected to gas phase geometry optimization with the Jaguar 8.9

quantum chemistry package⁵⁷using density functional theory (DFT)

with the B3LYP hybrid functional and the 6-31G** (34 and 55) or

LACVP** (1c, 1a, 30, and 57) basis set. Frequency calculations were

performed to ensure the structures represented minima on the

potential energy surfaces.

(υ + υ₁) − (υ₃+ υ₀)

2υ₂(sin 36° + sin 72°)

4

P = tan⁻¹

Protein Preparation. The structure of hAdK in complex with 1c

was retrieved from the RCSB PDB database⁵⁸(http://www.rcsb.org,

PDB ID 2I6A) and prepared as follows: after manual removal of

cocrystallized ligand and water molecules, ionization states of protein

side chains and hydrogen positions were assigned with the Protein

Preparazion Wizard tool.⁵⁹

Induced Fit Docking. The ligands were docked by means of the

IFD procedure based on Glide search algorithm⁶⁰using the Standard

Protocol and OPLS3 force ﬁeld.⁶¹The centroid of the cocrystallized

5I5 ligand residue was selected as center of the Glide grid (inner box

side = 10 Å; outer box side = auto). Ligands were initially docked

rigidly into the receptor by applying a scaling factor of 0.5 to both

ligand and protein van der Waals radii. Up to 20 poses per ligand were

collected and the side chains of residues within 5 Å of the ligand were

reﬁned with Prime.⁶²Ligands were redocked into the newly generated

receptor conformations with Glide⁶⁰by generating up to 10 poses

using the SP scoring function and reverting the vdW radii scaling

factors to their default values.

Molecular Dynamics. MD simulations were carried out with the

ACEMD program⁶³using periodic boundaries conditions and the

AMBER14SB⁶⁴/General Amber Force Field (GAFF)⁶⁵force ﬁelds for

the protein and ligand atoms, respectively. To derive force ﬁeld

parameters, ligands were subjected to energy minimization with

Gaussian 09⁶⁶at the HF/6-31G* level of theory. After geometry

optimization, ligand parameters and RESP partial charges were derived

with antechamber and parmchk tools as implemented in amber-

tools2014.⁶⁴Protein−ligand complexes were solvated by a cubic water

box with cell borders placed at least 12 Å away from any protein atom

using TIP3P as water model.⁶⁷The systems were neutralized with

Na⁺/Cl⁻counterions to a ﬁnal salt concentration of 0.150 M. After

2000 steps of energy minimization (conjugate-gradient method), the

systems were equilibrated with 50000 steps (100 ps) of NVE followed

by 1 ns of NPT simulations by applying harmonic positional

constraints on protein and ligand atoms that were gradually reduced

(scaling factor = 0.1). During the equilibration, the temperature was

maintained at 310 K using a Langevin thermostat⁶⁸with a low

damping constant of 1 ps⁻¹, and the pressure was maintained at 1 atm

using a Berendensen barostat.⁶⁹Bond lengths involving hydrogen

atoms were constrained using the M-SHAKE⁷⁰algorithm with an

integration time step of 2 fs. The equilibrated systems were then

subjected to 30 ns of unrestrained MD simulations in a NVT

ensemble. Long-range Coulomb interactions were handled using the

particle mesh Ewald summation method (PME)⁷¹with mesh spacing

set to 1.0 Å. A nonbonded cutoﬀ distance of 9 Å with a switching

distance of 7.5 Å was used. Each simulation was run in triplicate.

Trajectory Analysis. Selection of representative trajectories was

based upon the total ligand−protein interaction energy (IE_tot). In

particular, for each ligand−protein system, the IE_totexpressed as the

sum of van der Waals (IE_vdW) and electrostatic (IE_ele) contribution was

computed at frames extracted every 50 ps with the mdenergy function

Ligand root-mean-square deviation and ﬂuctuation (RMSD and

RMSF, respectively) and protein α carbon atoms RMSD with respect

to the starting docking pose were computed with the RMSD trajectory

tool (RSMDTT) implemented in VMD. MD average protein

structures were derived with the g_covar function implemented in

GROMACS,^77,78and rigid body rotation angles were computed

through the DYNDOM web server.⁷⁹

ASSOCIATED CONTENT

* Supporting Information

■

S

The Supporting Information is available free of charge on the

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

chem.6b00689.

1

Procedures for synthesis of intermediate 24, H NMR

and mass spectra of ﬁnal compounds, supplementary

modeling ﬁgures (conformational analysis and IE proﬁle

of 60 ns of MD simulation of hAdK in complex with 1c,

IE proﬁles of 30 ns of MD simulation of hAdK in

complex with 34 and 55, superimposition between

average MD and X-ray protein structures, docking pose

of 1a and superimposition between average MD and X-

ray protein structures, deﬁnition of endocyclic torsion

angles) and table (MD replicas RMSD and wDSF_tot

values), and enzymatic assay data (PDF)

3D coordinates of the hAdK in complex with 30 (PDB)

3D coordinates of the hAdK in complex with 57 (PDB)

3D coordinates of the hAdK in complex with 34 (PDB)

3D coordinates of the hAdK in complex with 55 (PDB)

Video S1 of 30 ns of MD simulation of the hAdK in

complex with 1c (AVI)

Video S2 of the comparison of 30 ns of MD simulation

of the hAdK in complex with 30 and 57 (AVI)

Video S3 of the comparison of 30 ns of MD simulation

of the hAdK in complex with 34 and 55 (AVI)

Video S4 of 30 ns of MD simulation of the hAdK in

complex with 1a (AVI)

Molecular formula strings (CSV)

AUTHOR INFORMATION

Corresponding Author

*Phone: 301-496-9024. Fax: 301-480-8422. E-mail:

kajacobsathelix.nih.gov.

■

Notes

The authors declare no competing ﬁnancial interest.

O

DOI: 10.1021/acs.jmedchem.6b00689

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

adenosine kinase and showing novel genetic and biochemical

characteristics. BMC Biochem. 2011, 12, 22.

(10) Fedele, D. E.; Li, T.; Lan, J. Q.; Fredholm, B. B.; Boison, D.

Adenosine A₁receptors are crucial in keeping an epileptic focus

localized. Exp. Neurol. 2006, 200, 184−190.

ACKNOWLEDGMENTS

■

We thank Dr. John Lloyd and Dr. Noel Whittaker (NIDDK)

for mass spectral determinations. This research was supported

by the Intramural Research Program of the NIH, National

Institute of Diabetes and Digestive and Kidney Diseases (ZIA

DK031117-28), National Institute of Neurological Disorders

and Stroke (R21 NS088024), and National Institute of Mental

Health (R01 MH083973). We thank Prof. Artem Melman

(Clarkson College) for helpful discussions. We thank Dr. Bryan

L. Roth (University of North Carolina at Chapel Hill) and

National Institute of Mental Health’s Psychoactive Drug

Screening Program (contract no. HHSN-271-2008-00025-C)

for screening data.

(11) Miller-Delaney, S. F.; Bryan, K.; Das, S.; McKiernan, R. C.; Bray,

I. M.; Reynolds, J. P.; Gwinn, R.; Stallings, R. L.; Henshall, D. C.

Differential DNA methylation profiles of coding and non-coding genes

define hippocampal sclerosis in human temporal lobe epilepsy. Brain

2015, 138, 616−631.

(12) Boison, D.; Scheurer, L.; Zumsteg, V.; Rulicke, T.; Litynski, P.;

̈

Fowler, B.; Brandner, S.; Mohler, H. Neonatal hepatic steatosis by

disruption of the adenosine kinase gene. Proc. Natl. Acad. Sci. U. S. A.

2002, 99, 6985−6990.

(13) Tosh, D. K.; Paoletta, S.; Deflorian, F.; Phan, K.; Moss, S. M.;

Gao, Z. G.; Jiang, X.; Jacobson, K. A. Structural sweet spot for A₁

adenosine receptor activation by truncated (N)-methanocarba nucleo-

sides: Receptor docking and potent anticonvulsant activity. J. Med.

Chem. 2012, 55, 8075−8090.

(14) Boison, D. Adenosine augmentation therapy for epilepsy. In

Jasper’s Basic Mechanisms of the Epilepsies; Noebels, J. L., Avoli, M,

Rogawski, M. A., Olsen, R. W., Delgado-Escueta, A. V., Eds.; Oxford

University Press: Oxford, UK, 2012; pp 1150−1160, .

(15) Boison, D.; Huber, A.; Padrun, V.; Deglon, N.; Aebischer, P.;

Mohler, H. Seizure suppression by adenosine-releasing cells is

independent of seizure frequency. Epilepsia 2002, 43, 788−796.

(16) Boison, D.; Singer, P.; Shen, H. Y.; Feldon, J.; Yee, B. K.

Adenosine hypothesis of schizophrenia - opportunities for pharmaco-

therapy. Neuropharmacology 2012, 62, 1527−1543.

(17) Ugarkar, B. G.; Castellino, A. J.; Dare, J. M.; Kopcho, J. J.;

Wiesner, J. B.; Schanzer, J. M.; Erion, M. D. Adenosine kinase

inhibitors. 2. Synthesis, enzyme inhibition, and antiseizure activity of

diaryltubercidin analogues. J. Med. Chem. 2000, 43, 2894−2905.

(18) Lee, C.-H.; Jiang, M.; Cowart, M.; Gfesser, G.; Perner, R.; Kim,

K. H.; Gu, Y. G.; Williams, M.; Jarvis, M. F.; Kowaluk, E. A.; Stewart,

A. O.; Bhagwat, S. S. Discovery of 4-amino-5-(3-bromophenyl)-7-(6-

morpholino-pyridin-3-yl)pyrido[2,3-d]pyrimidine, an orally active,

non-nucleoside adenosine kinase inhibitor. J. Med. Chem. 2001, 44,

2133−2138.

ABBREVIATIONS USED

■

5-IT, 5-iodotubercidin; ADA, adenosine deaminase; AdK,

adenosine kinase; AIBN, 2,2′-azobis(2-methylpropionitrile);

AMP, adenosine 5′-monophosphate; ATP, adenosine 5′-

triphosphate; BAIB, (diacetoxyiodo)benzene; CHO, Chinese

hamster ovary; DCM, dichloromethane; DEAD, diethyl

azodicarboxylate; DIAD, diisopropyl azodicarboxylate; DIPEA,

diisopropylethylamine; DMAP, N,N-dimethylaminepyridine;

DME, 1,2-dimethoxyethane; DMF, N,N-dimethylformamide;

DPPA, diphenylphosphoryl azide; ENT1, equilibrative nucleo-

side transporter1; hAdK, human adenosine kinase; HEPES, 4-

(2-hydroxyethyl)-1-piperazineethanesulfonic acid; HRMS, high

resolution mass spectroscopy; IE, interaction energy; IFD,

induced ﬁt docking; LC-MS, liquid chromatography mass

spectroscopy; NAD, nicotinamide adenine dinucleotide; NIS,

N-iodosuccinimide; NMR, nuclear magnetic resonance; RP-

HPLC, reversed phase high pressure liquid chromatography;

TBAF, tetrabutylammonium ﬂuoride; TEA, triethylamine;

TEMPO, (2,2,6,6-tetramethylpiperidin-1-yl)oxyl; TFA, tri-

ﬂuoroacetic acid; TLC, thin layer chromatography; tPSA,

total polar surface area

(19) (a) McGaraughty, S.; Cowart, M.; Jarvis, M. F. Recent

developments in the discovery of novel adenosine kinase inhibitors:

Mechanism of action and therapeutic potential. CNS Drug Rev. 2001,

7, 415−432. (b) Butini, S; Gemma, S.; Brindisi, M.; Borrelli, G.;

Lossani, A.; Ponte, A. M.; Torti, A.; Maga, G.; Marinelli, L.; La Pietra,

V.; Fiorini, I.; Lamponi, S.; Campiani, G.; Zisterer, D. M.; Nathwani, S.

M.; Sartini, S.; La Motta, C.; Da Settimo, F.; Novellino, E.; Focher, F.

Non-nucleoside inhibitors of human adenosine kinase: synthesis,

molecular modeling, and biological studies. J. Med. Chem. 2011, 54,

1401−1420.

(20) Bookser, B. C.; Matelich, M. C.; Ollis, K.; Ugarkar, B. G.

Adenosine kinase inhibitors. 4. 6,8-Disubstituted purine nucleoside

derivatives. Synthesis, conformation, and enzyme inhibition. J. Med.

Chem. 2005, 48, 3389−3399.

(21) Jarvis, M. F.; Yu, H.; McGaraughty, S.; Wismer, C. T.; Mikusa,

J.; Zhu, C.; Chu, K.; Kohlhaas, K.; Cowart, M.; Lee, C.-H.; Stewart, A.

O.; Cox, B. F.; Polakowski, J.; Kowaluk, E. A. Analgesic and anti-

inflammatory effects of A-286501, a novel orally active adenosine

kinase inhibitor. Pain 2002, 96, 107−118.

(22) McGaraughty, S.; Chu, K. L.; Wismer, C. T.; Mikusa, J.; Zhu, C.

Z.; Cowart, M.; Kowaluk, E.; Jarvis, M. F. Effects of A-134974, a novel

adenosine kinase inhibitor, on carrageenan-induced inflammatory

hyperalgesia and locomotor activity in rats: evaluation of the sites of

action. J. Pharmacol. Exp. Ther. 2001, 296, 501−509.

REFERENCES

■

(1) Chen, J. F.; Eltzschig, H. K.; Fredholm, B. B. Adenosine receptors

as drug targets - what are the challenges? Nat. Rev. Drug Discovery

2013, 12, 265−286.

(2) Wiesner, J. B.; Ugarkar, B. G.; Castellino, A. J.; Barankiewicz, J.;

Dumas, D. P.; Gruber, H. E.; Foster, A. C.; Erion, M. D. Adenosine

kinase inhibitors as a novel approach to anticonvulsant therapy. J.

Pharmacol. Exp. Ther. 1999, 289, 1669−1677.

(3) Zylka, M. J. Pain-relieving prospects for adenosine receptors and

ectonucleotidases. Trends Mol. Med. 2011, 17, 188−196.

(4) Choi, I. Y.; Lee, J. C.; Ju, C.; Hwang, S.; Cho, G. S.; Lee, H. W.;

Choi, W. J.; Jeong, L. S.; Kim, W. K. A₃adenosine receptor agonist

reduces brain ischemic injury and inhibits inflammatory cell migration

in rats. Am. J. Pathol. 2011, 179, 2042−2052.

(5) Williams-Karnesky, R. L.; Sandau, U. S.; Lusardi, T. A.; Lytle, N.

K.; Farrell, J. M.; Pritchard, E. M.; Kaplan, D. L.; Boison, D. Epigenetic

changes induced by adenosine augmentation therapy prevent epilepto-

genesis. J. Clin. Invest. 2013, 123, 3552−3563.

(6) Zimmermann, H.; Zebisch, M.; Strater, N. Cellular function and

molecular structure of ectonucleotidases. Purinergic Signalling 2012, 8,

̈

437−502.

(7) Parkinson, F. E.; Damaraju, V. L.; Graham, K.; Yao, S. Y. M.;

Baldwin, S. A.; Cass, C. E.; Young, J. D. Molecular biology of

nucleoside transporters and their distributions and functions in the

brain. Curr. Top. Med. Chem. 2011, 11, 948−972.

(23) Tatlisumak, T.; Takano, K.; Carano, R. A. D.; Miller, L. P.;

Foster, A. C.; Fisher, M.; Hsu, C. Y.; Lin, W. Delayed treatment with

an adenosine kinase inhibitor, GP683, attenuates infarct size in stroke.

Stroke 1998, 29, 1952−1958.

(8) Boison, D. Adenosine kinase: exploitation for therapeutic gain.

Pharmacol. Rev. 2013, 65, 906−943.

(9) Cui, X. A.; Agarwal, T.; Singh, B.; Gupta, R. S. Molecular

characterization of Chinese hamster cells mutants affected in

(24) Malnuit, V.; Slavet

̌ ́

̌ ́ ̌ ́

ínska, L. P.; Naus, P.; Dzubak, P.; Hajduch,

M.; Stolaríkova, J.; Snas l, J.; Pichova,

̌

́

e

̌

́

I.; Hocek, M. 2-Substituted 6-

P

DOI: 10.1021/acs.jmedchem.6b00689

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

(het)aryl-7-deazapurine ribonucleosides: Synthesis, inhibition of

adenosine kinases, and antimycobacterial activity. ChemMedChem

2015, 10, 1079−1093.

(25) 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: II. Modifications to the

ribofuranosyl moiety. Biochem. Pharmacol. 2008, 75, 1588−1600.

(26) McGaraughty, S.; Cowart, M.; Jarvis, M. F.; Berman, R. F.

Anticonvulsant and antinociceptive actions of novel adenosine kinase

inhibitors. Curr. Top. Med. Chem. 2005, 5, 43−58.

(27) Cronstein, B. N.; Naime, D.; Firestein, G. The antiinflammatory

effects of an adenosine kinase inhibitor are mediated by adenosine.

Arthritis Rheum. 1995, 38, 1040−1045.

(28) Lee, J. K.; Won, J. S.; Singh, A. K.; Singh, I. Adenosine kinase

inhibitor attenuates the expression of inducible nitric oxide synthase in

glial cells. Neuropharmacology 2005, 48, 151−160.

(29) Cottam, H. B.; Wasson, D. B.; Shih, H. C.; Raychaudhuri, A.; Di

Pasquale, G.; Carson, D. A. New adenosine kinase inhibitors with oral

antiinflammatory activity: synthesis and biological evaluation. J. Med.

Chem. 1993, 36, 3424−3430.

(30) Annes, J. P.; Ryu, J. H.; Lam, K.; Carolan, P. J.; Utz, K.;

Hollister-Lock, J.; Arvanites, A. C.; Rubin, L. L.; Weir, G.; Melton, D.

A. Adenosine kinase inhibition selectively promotes rodent and

porcine islet β-cell replication. Proc. Natl. Acad. Sci. U. S. A. 2012, 109,

3915−3920.

(31) García-Villafranca, J.; Castro, J. Effects of 5-iodotubercidin on

hepatic fatty acid metabolism mediated by the inhibition of acetyl-CoA

carboxylase. Biochem. Pharmacol. 2002, 63, 1997−2000.

bicyclo[3.1.0]hexane-carboxylate from L-ribose: A versatile chiral

synthon for preparation of adenosine and P2 receptor ligands.

Nucleosides, Nucleotides Nucleic Acids 2008, 27, 279−291.

(39) Melman, A.; Zhong, M.; Marquez, V. E.; Jacobson, K. A.

Synthesis of enantiomerically pure (S)-methanocarbaribo uracil

nucleoside derivatives for use as antiviral agents and P2Y receptor

ligands. J. Org. Chem. 2008, 73, 8085−8088.

(40) Smith, D. E.; Marquez, I.; Lokensgard, M. E.; Rheingold, A. L.;

Hecht, D. A.; Gustafson, J. L. Exploiting atropisomerism to increase

the target selectivity of kinase inhibitors. Angew. Chem., Int. Ed. 2015,

54, 11754−11759.

(41) Boyer, S. H.; Ugarkar, B. G.; Solbach, J.; Kopcho, J.; Matelich,

M. C.; Ollis, K.; Gomez-Galeno, J. E.; Mendonca, R.; Tsuchiya, M.;

Nagahisa, A.; Nakane, M.; Wiesner, J. B.; Erion, M. D. Adenosine

kinase inhibitors. 5. Synthesis, enzyme inhibition, and analgesic activity

of diaryl-erythro-furanosyltubercidin analogues. J. Med. Chem. 2005,

48, 6430−6441.

(42) Toti, K. S.; Derudas, M.; Pertusati, F.; Sinnaeve, D.; Van den

Broeck, F.; Margamuljana, L.; Martins, J. C.; Herdewijn, P.; Balzarini,

J.; McGuigan, C.; Van Calenbergh, S. Synthesis of an apionucleoside

family and discovery of a prodrug with anti-HIV activity. J. Org. Chem.

2014, 79, 5097−5112.

(43) Ugarkar, B. G.; DaRe, J. M.; Kopcho, J. J.; Browne, C. E., III;

Schanzer, J. M.; Wiesner, J. B.; Erion, M. D. Adenosine kinase

inhibitors. 1. Synthesis, enzyme inhibition, and antiseizure activity of 5-

iodotubercidin analogues. J. Med. Chem. 2000, 43, 2883−2893.

(44) Epp, J. B.; Widlanski, T. S. Facile preparation of nucleoside-5′-

carboxylic acids. J. Org. Chem. 1999, 64, 293−295.

(32) Marquez, V. E.; Siddiqui, M. A.; Ezzitouni, A.; Russ, P.; Wang,

J.; Wagner, R. W.; Matteucci, M. D. Nucleosides with a twist. Can

fixed forms of sugar ring pucker influence biological activity in

nucleosides and oligonucleotides? J. Med. Chem. 1996, 39, 3739−3747.

(33) Maruoka, H.; Barrett, M. O.; Ko, H.; Tosh, D. K.; Melman, A.;

Burianek, L. E.; Balasubramanian, R.; Berk, B.; Costanzi, S.; Harden, T.

K.; Jacobson, K. A. Pyrimidine ribonucleotides with enhanced

selectivity as P2Y₆receptor agonists: Novel 4-alkyloxyimino, (S)-

methanocarba, and 5′-triphosphate γ-ester modifications. J. Med. Chem.

2010, 53, 4488−4501.

(45) Tosh, D. K.; Padia, J.; Salvemini, D.; Jacobson, K. A. Efficient,

large-scale synthesis and preclinical studies of MRS5698, a highly

selective A₃adenosine receptor agonist that protects against chronic

neuropathic pain. Purinergic Signalling 2015, 11, 371−387.

(46) Bhutoria, S.; Ghoshal, N. Deciphering ligand dependent degree

of binding site closure and its implication in inhibitor design: A

modeling study on human adenosine kinase. J. Mol. Graphics Modell.

2010, 28, 577−591.

(47) Dong, L.; Shi, J.; Wang, J.; Liu, Y. Theoretical studies on the

conformational change of adenosine kinase induced by inhibitors. Int.

J. Quantum Chem. 2011, 111, 3980−3990.

(34) Snas

Bourderioux, A.; Duse

H.; Pohl, R.; Konecn

́

e

̌

l, J.; Naus,

̌

P.; Dostal

k, M.; Dvora

y, P.; Dzubak

́

, J.; Hnízda, A.; Fanfrlík, J.; Brynda, J.;

ova, H.; Stolaríkova, J.; Zabranska,

, P.; Votruba, I.; Hajduch, M.;

̌

k

́

̌

́

(48) Hernandez, S.; Ford, H., Jr.; Marquez, V. E. Is the anomeric

effect an important factor in the rate of adenosine deaminase catalyzed

hydrolysis of purine nucleosides? A direct comparison of nucleoside

analogues constructed on ribose and carbocyclic templates with

equivalent heterocyclic bases selected to promote hydration. Bioorg.

Med. Chem. 2002, 10, 2723−2730.

̌

́

̌

́

́ ̌ ́ ́

Rezacova, P.; Veverka, V.; Hocek, M.; Pichova, I. 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.

(35) (a) Schelling, P.; Claus, M. T.; Johner, R.; Marquez, V. E.;

Schulz, G. E.; Scapozza, L. Biochemical and structural characterization

of (South)-methanocarbathymidine that specifically inhibits growth of

herpes simplex virus type 1 thymidine kinase-transduced osteosarcoma

cells. J. Biol. Chem. 2004, 279, 32832−32838. (b) Marquez, V. E.;

Choi, Y.; Comin, M. J.; Russ, P.; George, C.; Huleihel, M.; Ben-Kasus,

T.; Agbaria, R. Understanding how the herpes thymidine kinase

orchestrates optimal sugar and nucleobase conformations to

accommodate its substrate at the active site: A chemical approach. J.

Am. Chem. Soc. 2005, 127, 15145−15150. (c) Sun, G.; Voigt, J. H.;

Filippov, I. V.; Marquez, Y. E.; Nicklaus, M. C. PROSIT: Pseudo-

Rotational Online Service and Interactive Tool, applied to a

conformational survey of nucleosides andnucleotides. J. Chem. Inf.

Comp. Sci. 2004, 44, 1752−1762.

(36) Muchmore, S. W.; Smith, R. A.; Stewart, A. O.; Cowart, M. D.;

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.

(37) Mathews, I. I.; Erion, M. D.; Ealick, S. E. Structure of human

adenosine kinase at 1.5 A resolution. Biochemistry 1998, 37, 15607−

15620.

(49) Schaddelee, M. P.; Read, K. D.; Cleypool, C. G.; IJzerman, A. P.;

Danhof, M.; de Boer, A. G. Brain penetration of synthetic adenosine

A₁receptor agonists in situ: role of the rENT1 nucleoside transporter

and binding to blood constituents. Eur. J. Pharm. Sci. 2005, 24, 59−66.

(50) Pardridge, W. M. Blood−brain barrier endogenous transporters

as therapeutic targets: a new model for small molecule CNS drug

discovery. Expert Opin. Ther. Targets 2015, 19, 1059−1072.

̀

(51) Sinclair, C. J. D.; Powell, A. E.; Xiong, W.; LaRiviere, C. G.;

Baldwin, S. A.; Cass, C. E.; Young, J. D.; Parkinson, F. E. Nucleoside

transporter subtype expression: effects on potency of adenosine kinase

inhibitors. Br. J. Pharmacol. 2001, 134, 1037−1044.

(52) Lee, K.; Cass, C.; Jacobson, K. A. Synthesis using ring closure

metathesis and effect on nucleoside transport of a (N)-methanocarba

S-(4-nitrobenzyl)thioinosine derivative. Org. Lett. 2001, 3, 597−599.

(53) Damaraju, V. L.; Mowles, D.; Smith, K. M.; Yao, S. Y.; Young, J.

D.; Marquez, V. E.; Cass, C. E. Influence of sugar ring conformation

on the transportability of nucleosides by human nucleoside trans-

porters. ChemBioChem 2011, 12, 2774−2778.

(54) Mato, J. M.; Martínez-Chantar, M. L.; Lu, S. C. Methionine

metabolism and liver disease. Annu. Rev. Nutr. 2008, 28, 273−293.

(55) Besnard, J.; Ruda, G. F.; Setola, V.; Abecassis, K.; Rodriguiz, R.

M.; Huang, X. P.; Norval, S.; Sassano, M. F.; Shin, A. I.; Webster, L. A.;

Simeons, F. R.; Stojanovski, L.; Prat, A.; Seidah, N. G.; Constam, D.

B.; Bickerton, G. R.; Read, K. D.; Wetsel, W. C.; Gilbert, I. H.; Roth, B.

(38) Joshi, B. V.; Melman, A.; Mackman, R. L.; Jacobson, K. A.

Synthesis of ethyl (1S,2R,3S,4S,5S)-2,3-O-(isopropylidene)-4-hydroxy-

Q

DOI: 10.1021/acs.jmedchem.6b00689

J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

L.; Hopkins, A. L. Automated design of ligands to polypharmacological

ligand recognition: The human A_2Aadenosine receptor as a key study.

J. Chem. Inf. Model. 2014, 54, 169−183.

profiles. Nature 2012, 492, 215−220.

(75) Williams, T.; Kelley, C. Gnuplot 4.6: An Interactive Plotting

Program, version 4.6.6, 2014; http://gnuplot.info (accessed May

2016).

(56) Maestro, version 10.3; Schrodinger, LLC: New York, 2015.

̈

(57) Jaguar, version 8.9; Schrodinger, LLC: New York, 2015.

̈

(58) Bernstein, F. C.; Koetzle, T. F.; Williams, G. J.; Meyer, E. E., Jr.;

Brice, M. D.; Rodgers, J. R.; Kennard, O.; Shimanouchi, T.; Tasumi,

M. The protein data bank: A computer-based archival file for

macromolecular structures. J. Mol. Biol. 1977, 112, 535−542.

(59) Sastry, G. M.; Adzhigirey, M.; Day, T.; Annabhimoju, R.;

Sherman, W. Protein and ligand preparation: parameters, protocols,

and influence on virtual screening enrichments. J. Comput.-Aided Mol.

Des. 2013, 27, 221−234.

(76) Altona, C.; Sundaralingam, M. Conformational analysis of the

sugar ring in nucleosides and nucleotides. A new description using the

concept of pseudorotation. J. Am. Chem. Soc. 1972, 94, 8205−8212.

(77) van der Spoel, D.; Lindahl, E.; Hess, B.; GROMACS

Development Team. GROMACS User Manual, version 4.6.5, 2013;

www.gromacs.org.

(78) Hess, B.; Kutzner, C.; van der Spoel, D.; Lindahl, E. GROMACS

4: Algorithms for highly efficient, load-balanced, and scalable

molecular simulation. J. Chem. Theory Comput. 2008, 4, 435−447.

(79) Girdlestone, C.; Hayward, S. The DynDom3D webserver for the

analysis of domain movements in multimeric proteins. J. Comput. Biol.

2015, 23, 21−26.

(80) Datta, R.; Das, I.; Sen, B.; Chakraborty, A.; Adak, S.; Mandal, C.;

Datta, A. K. Homology-model-guided site-specific mutagenesis reveals

the mechanisms of substrate binding and product-regulation of

adenosine kinase from Leishmania donovani. Biochem. J. 2006, 394,

35−42.

(60) Glide, version 6.8; Schrodinger, LLC: New York, 2015.

̈

(61) Prime, version 4.1; Schrodinger, LLC: New York, 2015.

̈

(62) Harder, E.; Damm, W.; Maple, J.; Wu, C.; Reboul, M.; Xiang, J.

Y.; Wang, L.; Lupyan, D.; Dahlgren, M. K.; Knight, J. L.; Kaus, J. W.;

Cerutti, D. S.; Krilov, G.; Jorgensen, W. L.; Abel, R.; Friesner, R. A.

OPLS3: A force field providing broad coverage of drug-like small

molecules and proteins. J. Chem. Theory Comput. 2016, 12, 281−296.

(63) Harvey, M.; Giupponi, G.; De Fabritiis, G. ACEMD:

Accelerated molecular dynamics simulations in the microseconds

timescale. J. Chem. Theory Comput. 2009, 5, 1632−1639.

(64) Case, D.; Babin, V.; Berryman, J.; Betz, R.; Cai, Q.; Cerutti, D.;

Cheatham Iii, T.; Darden, T.; Duke, R.; Gohlke, H. Amber14, 2014;

http://ambermd.org/ (accessed May 2016).

(65) Wang, J.; Wolf, R. M.; Caldwell, J. W.; Kollman, P. A.; Case, D.

A. Development and testing of a general amber force field. J. Comput.

Chem. 2004, 25, 1157−1174.

(66) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;

Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci,

B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H.

P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.;

Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,

T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A.;

Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin,

K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.;

Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega,

N.; Millam, J. M.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.;

Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.;

Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.;

Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.;

Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas; Foresman, J.

B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09, revision B.01;

Gaussian, Inc.: Wallingford, CT, 2009.

(67) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R.

W.; Klein, M. L. Comparison of simple potential functions for

simulating liquid water. J. Chem. Phys. 1983, 79, 926−935.

(68) Loncharich, R. J.; Brooks, B. R.; Pastor, R. W. Langevin

dynamics of peptides: the frictional dependence of isomerization rates

of N-acetylalanyl-N′-methylamide. Biopolymers 1992, 32, 523−535.

(69) Berendsen, H. J. C.; Postma, J. P. M.; van Gunsteren, W. F.;

DiNola, A.; Haak, J. R. Molecular dynamics with coupling to an

external bath. J. Chem. Phys. 1984, 81, 3684−3690.

(70) Krautler, V.; Van Gunsteren, W. F.; Hunenberger, P. H. A Fast

̈

SHAKE algorithm to solve distance constraint equations for small

molecules in molecular dynamics simulations. J. Comput. Chem. 2001,

22, 501−508.

(71) Essmann, U.; Perera, L.; Berkowitz, M. L.; Darden, T.; Lee, H.;

Pedersen, L. G. A. Smooth particle mesh Ewald method. J. Chem. Phys.

1995, 103, 8577−8593.

(72) Humphrey, W.; Dalke, A.; Schulten, K. VMD - Visual molecular

dynamics. J. Mol. Graphics 1996, 14, 33−38.

(73) Phillips, J. C.; Braun, R.; Wang, W.; Gumbart, J.; Tajkhorshid,

E.; Villa, E.; Chipot, C.; Skeel, R. D.; Kale, L.; Schulten, K. Scalable

molecular dynamics with NAMD. J. Comput. Chem. 2005, 26, 1781−

1802.

(74) Sabbadin, D.; Ciancetta, A.; Moro, S. Bridging molecular

docking to membrane molecular dynamics to investigate GPCR−

R

DOI: 10.1021/acs.jmedchem.6b00689

J. Med. Chem. XXXX, XXX, XXX−XXX

Article Doi

South (S)- and north (N)-methanocarba-7-deazaadenosine analogues as inhibitors of human adenosine kinase

DOI: 10.1021/acs.jmedchem.6b00689

Source and publish data:

Authors:

Article abstract of DOI:10.1021/acs.jmedchem.6b00689

Full text of DOI:10.1021/acs.jmedchem.6b00689

Products guided by the article

R&D Labs maybe for 1069040-47-6

Relevant to this article

Hot Product