Synthesis and Pharmacokinetic Evaluation of Siderophore Biosynthesis Inhibitors for Mycobacterium tuberculosis

Abstract

MbtA catalyzes the first committed biosynthetic step of the mycobactins, which are important virulence factors associated with iron acquisition in Mycobacterium tuberculosis. MbtA is a validated therapeutic target for antitubercular drug development. 5′-O-[N-(salicyl)sulfamoyl]adenosine (1) is a bisubstrate inhibitor of MbtA and exhibits exceptionally potent biochemical and antitubercular activity. However, 1 suffers from sub-optimal drug disposition properties resulting in a short half-life (t_1/2), low exposure (AUC), and low bioavailability (F). Four strategies were pursued to address these liabilities including the synthesis of prodrugs, increasing the pK_a of the acyl-sulfonyl moiety, modulation of the lipophilicity, and strategic introduction of fluorine into 1. Complete pharmacokinetic (PK) analysis of all compounds was performed. The most successful modifications involved fluorination of the nucleoside that provided substantial improvements in t_1/2 and AUC. Increasing the pK_a of the acyl-sulfonyl linker yielded incremental enhancements while modulation of the lipophilicity and prodrug approaches led to substantially poorer PK parameters.

Introduction

Mycobacterium tuberculosis (Mtb), often considered the most successful human pathogen, infects more than two billion people and remains a leading source of infectious disease mortality and morbidity.¹ Tuberculosis (TB), the disease caused by Mtb or one of seven other closely related Mycobacterium species,² has received renewed attention since the World Health Organization (WHO) declared TB a global health emergency due the alarming rise of drug resistance, pandemic of HIV-TB co-infection, and lack of new therapeutics.¹ Multidrug resistant TB (MDR-TB) is defined as resistance to isoniazid and rifampin, which are the two most effective TB drugs. Approximately, 9% of MDR-TB cases are additionally resistant to one of the injectable antibiotics used for TB (aminoglycosides or capreomycins) and any fluoroquinoline; these TB strains are classified as extensively drug resistant (XDR-TB).¹ Mtb is extraordinarily difficult to treat, requiring prolonged therapy with multiple antibiotics for 6 months for simple drug susceptible infections³ and up to 20 months for drug-resistant infections, with poor treatment outcomes in the latter cases.⁴ Eradication of the global TB pandemic, a long-term goal of the WHO, will require improved diagnostics, vaccines, and new drugs. Excellent progress has been made over the last decade and, for the first time in over forty years, two new TB drugs (Bedaquiline and Delamanid) were approved for MDR-TB by US and European regulatory agencies; although initial enthusiasm has been marred by safety concerns.^{5, 6} While a specific target candidate profile has not yet been established in the TB field, ideal criteria for new antitubercular agents as advocated by the TB Alliance (http://www.tballiance.org) include: the potential to shorten and simplify the treatment regimen, effectiveness against multidrug resistant strains, and compatibility with HIV-TB co-infection.

In conceiving of new therapeutic strategies for TB we have focused on mycobacterial iron acquisition, an essential Mtb process with a long, storied history. In the mid-19th century, decades before Robert Koch identified Mtb as the etiological agent of TB, the famous French physician Armand Trousseau noticed some anemic patients rapidly succumbed to TB reactivation when given iron supplements.⁷ This observation is now mechanistically understood since Mtb, as well as most microorganisms, requires iron as a redox cofactor for numerous essential cellular processes. However, normal physiological conditions restrict the free ferric concentration to an estimated 10⁻²⁴ M, far below the concentration the bacteria needs for proliferation.^{8, 9} The process of withholding an essential nutrient is referred to as nutritional immunity, and represents an extremely effective strategy to prevent bacterial growth. To combat iron deficiency, bacteria produce small molecule iron chelators known as siderophores, often at great metabolic expense, that are secreted into the surrounding environment and abstract iron from host proteins.^{10, 11} The siderophore-ferric complexes are then reimported into the bacteria cytosol, usually by specific siderophore transporters, where iron is released through hydrolytic destruction of the siderophore or through reduction of Fe³⁺ to Fe²⁺. The tripodal mycobactins produced by Mycobacterium species were the first class of siderophores characterized in 1912 and found to be an essential substance for mycobacterial growth.^12–14 Siderophores of diverse chemical structures are now known to be produced by virtually all microorganisms, and even some plants.^{15, 16}

Following the publication of the Mtb genome in 1998, the Walsh group identified a locus of 10 genes (mbtA-J) encoding for a mixed nonribosomal peptide synthetase-polyketide synthase (NRPS-PKS) pathway responsible for biosynthesis of the mycobactin peptidic core. This was subsequently verified through comprehensive mutational analysis and functional characterization.^17–19 Gokhale and co-workers later identified a second locus of four genes (mbtK-N) required for synthesis of the lipid side-chain found in the mycobactins.²⁰ Mtb mutants deficient in mycobactin production or transport have been shown to have severe growth defects under iron limiting conditions, and are incapable of establishing an infection in vivo.^21–23 Collectively, these results provide compelling evidence for targeting iron acquisition in Mtb with small molecules. Indeed, two diametrically opposed approaches have been pursued through synthesis of mycobactin analogues and inhibitors of mycobactin biosynthesis.^{24, 25} In spectacular application of the first approach, Miller and co-workers synthesized a mycobactin-artemisinin conjugate that exploits the mycobactin uptake system for delivery of artemisinin, which generates toxic reactive oxygen species upon intracellular reductive release of Fe²⁺ from mycobactin.²⁶ Inhibition of mycobactin biosynthesis has also been fruitful, and inhibitors of enzymes within the pathway have been described for MbtA,^{25, 27, 28} MbtI,²⁹ MbtM,³⁰ and PptT.³¹

MbtA, responsible for incorporation of salicylic acid into the mycobactin core structure, has been the most extensively studied for inhibitor development. ^{28, 32–36} The bisubstrate inhibitor 5′-O-[N-(salicyl)sulfamoyl]adenosine (1, Figure 1C) was independently described by three groups and found to exhibit exceptionally potent enzyme inhibition of MbtA, with an apparent K_i of ∼6 nM under physiologically relevant super-saturating concentrations of its substrates ATP and salicylic acid.^{25, 27, 28} This compound possesses potent antitubercular activity, with a minimum inhibitory concentration of 0.39 µM under iron-limiting conditions in vitro, and was also effective in an acute murine TB infection model when dosed intraperitoneally, thereby chemically validating MbtA as an antitubercular target.³⁷ Unfortunately, 1 possesses sub-optimal pharmacokinetic (PK) parameters including poor oral bioavailability, low oral exposure, and rapid clearance.³⁷ We report herein four approaches to address the poor PK parameters of 1 including: the synthesis of prodrugs (2–3), analogs that increase the sulfamate pK_a (4–5), modulation of the lipophiliciy of the highly polar nucleoside (6–7), and strategic introduction of fluorine into the nucleoside (8–10) as shown in Figure 1C. Pharmacokinetic analysis of all compounds was performed in conjunction with selective in vitro studies to understand the PK behavior.

Figure 1.

Mycobactins, biosynthesis, and inhibitors. A. Structures of the lipid soluble mycobactins and water soluble mycobactins (refered to as carboxymycobactins) produced by Mtb. The mycobactins and carboxymycobactins are synthesized as a suite of compounds that differ in the chain length of the lipid tail on the central lysine residue. The most abundant chain length is depicted. B. Biosynthesis of all mycobactins is initiated with the ATP-dependent activation of salicylic acid by MbtA to form an intermediate acyl-adenylate. In a second half-reaction, MbtA ligates salicylic acid onto the N-terminal aryl acid carrier protein domain of MbtB. MbtB, in combination with MbtE, MbtC, MbtD, and MbtF (not shown), sequentially builds the mycobactin peptidic core scaffold. C. 5′-O-[N-(salicyl)sulfamoyl]adenosine (1) was the first described inhibitor of MbtA and is considered a bisubstrate inhibitor that mimics the intermediate acyl-adenylate. This modular inhibitor scaffold contains aryl (blue), linker (black), glycosyl (pink), and nucleobase (green) domains. Prior SAR studies have examined the importance of each atom of this scaffold and provided a comprehensive understanding of the structure-activity relationships (SAR) that govern enzyme inhibition and whole-cell activity. This information has been used to design prodrugs 2–3 and analogs 4–10 in order to explore the SAR that governs pharmacokinetic behavior.

Results and Discussion

Chemistry

The simple di-acetate prodrug 2, wherein the polar 2′ and 3′ alcohols are masked, as well as tri-acetate 3, in which the phenol is additionally acylated, were initially explored to improve the bioavailability and exposure of 1. Acetates are quickly hydrolyzed by serum esterases in vivo.³⁸ The di-acetate 2 was obtained by acylation of sulfamate 11³⁹ with N-hydroxysuccinimidyl 2-benzyloxybenzoate²⁸ to provide 12 (Scheme 1A). Trifluoracetic acid (TFA) mediated deprotection of the acetonide followed by acetylation of the resultant free 2′ and 3′ alcohols afforded 13. Catalytic hydrogenolysis and reverse-phase high performance liquid chromatography (HPLC) purification furnished di-acetate product 2. The tri-acetate derivative 3 was prepared by direct acetylation of 1 employing acetic anhydride and pyridine in dimethylformamide (DMF) (Scheme 1B). These conditions were notable since competitive acetylation at the N⁶ position of adenosine was not observed. Compound 3 was first purified on normal phase silica and isolated as an approximate 5:1 mixture of tri-and di-acetates 3 and 2. Further purification by reverse-phase HPLC was required to obtain pure 3.

Scheme 1. a. Synthesis of prodrugs 2 and 3.

^aReaction conditions: (a) N-hydroxysuccinimidyl 2-benzyloxybenzoate, Cs₂CO₃, DMF, 0 °C to rt, 21 h, 62%; (b) 80% aqueous TFA, 0 °C to rt, 4 h; (c) Ac₂O, pyridine, DMF, rt, 16 h, 63% (2 steps); (d) H₂, Pd/C, MeOH, rt, 2 h, 42%; (e) Ac₂O, pyridine, DMF, rt, 20 h, 46%.

In an effort to further increase the pK_a of the acylsulfonyl linker’s acidic nitrogen, the next analogues in this set replaced the sulfamate linker with a sulfamide and added a para-amino substitution on the salicyl moiety to provide 4 and 5, respectively (Scheme 2). We developed a new synthesis of 4 (i.e. 14→15→16→4) resulting in a 10-fold improved overall yield relative to our initially disclosed route.²⁸ Our revised synthesis features a Mitsunobu reaction of tert-butoxycarbonyl (Boc) sulfamide (29)⁴⁰ with N⁶,N⁶-bis(Boc)-isopropylideneadenosine (14)³⁹ to afford 15 in an impressive 96% yield for direct installation of the sulfamide group at C-5′ of the nucleoside. Here the N⁶-bis-Boc protecting group in 14 is essential to subdue cyclonucleoside formation that occurs from cyclization of N-3 onto C-5' during the Mitsunobu reaction by attenuating the nucleophilicity at N-3.^{41, 42} Coupling 15 to the N-hydroxysuccinimide (NHS) ester of 2-O-methoxymethylsalicyate (31)³² employing Cs₂CO₃ in DMF provided 16 in 60% yield. The adjacent Boc group on the sulfamide moiety is responsible for the reduced yields of this acylation reaction relative to that observed with simple sulfamate analogs, which typically proceed in 80–90% yields. Global deprotection with 80% aqueous TFA furnished sulfamide 4 that was purified by reverse-phase HPLC and isolated in 65% yield as the triethylammonium salt. Compound 5 was prepared in an analogous fashion (15→17→18→5) except sulfamide 15 was acylated with benzyloxycarbonyl (Cbz) protected p-aminosalicylate (32)⁴³ employing in situ activation with carbonyl diimidazole (CDI)^{25, 33} to afford 17. Sequential hydrogenolysis of the Cbz group with Pd/C in methanol to 18 followed by acidic deprotection furnished 5.

Scheme 2. a. Synthesis of 4–8.

^aReaction conditions: (a) 29, PPh₃, DIAD, THF, 0 °C to rt, 16 h, 15=96%, 24=91%, 27=89%; (b) i)30, CH₂Cl₂, rt, 14 h; ii) then H₂, Pd/C, MeOH–THF, rt, 22 h, 21=62%; (c) 31, Cs₂CO₃, DMF, 0 °C to rt, 16 h, 16=60%, 22=43%; (d) i)32, CDI, THF, 60 °C, 2 h; ii) then 15, Cs₂CO₃, THF, rt, 16 h, 17=55%; (e) H₂, Pd/C, MeOH, rt, 2 h; (f) 80% aqueous TFA, 0 °C to rt, 4 h, 4=65%, 5=34% (2 steps from 17), 6=61%, 7=18% (2 steps from 24), 8=38 % (2 steps from 27); (g) DPPA, NaN₃, 15-crown-5, THF, 0 °C to rt, 4.5 h; (h) H₂, Pd/C, MeOH, rt, 20=44%, (2 steps from 19). Nonstandard abbreviations: IPK = isopropylidene ketal.

N⁶-Cyclopropyl 6 and N⁶-cyclopropyl-2-phenyl 7 containing the optimal sulfamide linkages were prepared in an effort to increase the lipophilicity (Scheme 2). These individual modifications on a sulfamate template were previously shown to improve potency and selectivity.³⁴ For the synthesis of 6 we explored a complimentary synthetic route (19→20→21→22→6) employing sulfamoylation with the novel dimethylaminopyridinium-activated reagent 30 as reported by Tan and co-workers.⁴⁴ N⁶-Cyclopropyl-2′,3′-O-isopropylideneadenosine 19³⁴ was elaborated to 5′-amino derivative 20 through sequential azidation and hydrogenation. Sulfamoylation of aminonucleoside 20 was achieved with dimethylaminopyridinium-activated 30⁴⁴ followed by deprotection of the Cbz group to give the free sulfamide 21 that was converted to 6 through sequential salicylation and global deprotection using the standard conditions developed for 4. N⁶-cyclopropyl-2-phenyl 7 was prepared from 23 in analogy to 4 (23→24→25→7).

To complete our narrative, we explored fluorinated analogs 8–10, since fluorine is known to affect absorption-distribution-metabolism-excretion (ADME) properties.⁴⁵ The enhanced PK parameters of 8 prompted the synthesis of 9–10 (vide infra). Fluorinated analog 8 was prepared from 26 in analogy to 4 (26→27→28→8) as presented in Scheme 2. The synthesis of 2′-fluoro-2-phenyl analog 9 is illustrated in Scheme 3 starting from commercially available 2′-deoxy-2′-fluoroguanosine 33. Acetylation of the 3' and 5' alcohols provided 34, which was transformed to 35 by sequential chlorination employing POCl₃ in dimethylaniline (DMA) promoted by Et₄NCl followed by Sandmeyer-like reaction with isoamyl nitrite and diiodomethane.^{46, 47} Ammonolysis yielded adenosine derivative 36 that was immediately tert-butyldimethylsilyl (TBS) protected to provide 37. The requisite C-2 phenyl group was incorporated by Buchwald-Hartwig coupling with phenylboronic acid to give 38.⁴⁸ Regioselective monodesilylation of the 5′ TBS group with AcOH–THF–H₂O (9:1:1) afforded the key nucleoside intermediate 39. For installation of the sulfamide, we turned to new chemistry developed in our lab, which involves treatment of a 5′-aminonucleoside with sulfamide (NH₂SO₂NH₂), inspired by a side-reaction reported by Takai and co-workers.⁴⁹ In the present synthesis this was achieved by conversion of the 5′-alcohol in 39 to azide 40. The azide was reduced by catalytic hydrogenation to provide a crude amine that was not isolated, but treated directly with sulfamide in dioxane at 90 °C to cleanly afford 41 in 87% yield. This new sulfamide synthesis is high yielding, experimentally simpler than our Mitsunobu protocol since it avoids contamination with triphenylphosphine oxide, and obviates N⁶-protection on the nucleoside. The completion of the synthesis was accomplished using our standard salicylation method with 31 followed by TBS deprotection to furnish 9.

Scheme 3. a. Synthesis of 9.

^aReaction conditions: (a) Ac₂O, pyridine, DMF, rt, 12 h, 79%; (b) POCl₃, DMA, Et₄NCl, reflux, 20 min, 86%; (c) CH₂I₂, I₂, isoamylnitrite, 80 °C, 1 h, 95%; (d) NH₃, MeOH, 70 °C, 24 h, 56%; (e) TBSCl, imidazole, DMAP, DMF, rt, 24 h, 97%; (f) Pd(OAc)₂, CyJohnPhos, K₃PO₄, Ph(BOH)₂, 100 °C, 12 h, 67%; (g) AcOH–THF–H₂O (9:1:1), THF, 23 °C, 24 h, 70%; (h) DPPA, DBU, NaN₃, 15-crown-5, 24 h, 85%; (i) Pd/C, MeOH, rt, 1 h; (j) NH₂SO₂NH₂, dioxane, 90 °C, 3 h, 87% (2 steps); (k) 31, Cs₂CO₃, DMF, 0 °C to rt, 16 h; (l) 80% aq TFA, 0 °C, 3 h, 45% (2 steps from 41). Nonstandard abbreviations: CyJohnPhos = (2-biphenyl)dicyclohexylphosphine.

Based on the improved PK properties of 2′-fluoro analog 8 (vide infra), we also pursued the synthesis of 2′,3′-dideoxy-2′-fluoro derivative 10. Synthesis commenced from 2′-deoxy-2′-fluoroadenosine 42 that was regioselectively TBS protected at the C-5' alcohol to furnish 43.⁵⁰ Barton-McCombie deoxygenation was accomplished in two steps through conversion to the corresponding thionocarbonate 44 followed by treatment with Bu₃SnH in refluxing toluene to yield 45.⁵⁰ Deprotection of the TBS group with TFA provided alcohol 46, which was transformed to azide 47 using methodology reported by Liu and Austin.⁵¹ The sulfamide moiety at C-5' in 48 was then installed using our new protocol described in the previous section through azide reduction of 47 and in situ sulfamoylation of the crude aminonucleoside. The standard salicylation and TFA mediated deprotection sequence yielded the final analog 10.

Enzyme Inhibition and Antitubercular Activity

The new analogs 5–10, excluding the prodrugs 2 and 3, were evaluated for enzyme inhibition against recombinant MbtA under initial velocity conditions as previously described (see Experimental Section).³² The apparent inhibition constants (appK_i) were determined by fitting the concentration–response plots to the Morrison equation (eq 1, see Experimental Section) since all compounds exhibited tight-binding behavior (i.e. appK_i ≤10 × [MbtA]). The appK_i values for all analogs were potent, ranging from 0.78–8.9 nM (Table 1). These results are largely in agreement with prior findings and indicate that our targeted modifications to 1 did not adversely impact potency.

Table 1.

Enzyme inhibition and antimycobacterial activity of 1 analogs.

Compound	appKi, (nM)a	MIC, µM (µg/mL)b	MIC/appKic	CC50 (µM)d
1	6.6 ± 1.5	0.39 (0.20)	59	>200
4	3.7 ± 0.6	0.19 (0.11)	51	>200
5	8.9 ± 0.7	3.125 (1.81)	351	>200
6	1.5 ± 0.2	0.098 (0.06)	65	–
7	0.76 ± 0.40	6.25 (4.27)	8224	–
8	1.7 ± 0.1	0.19 (0.11)	112	>200
9	1.6 ± 0.1	1.56 (1.0)	975	>200
10	3.0 ± 0.2	0.78 (0.43)	260	–

^aAssay performed with 7 nM MbtA, 10 mM ATP, 250 µM salicylic acid, 1 mM PPi.

^bGrown in glycerol-alanine salts (GAS) medium without ferric ammonium citrate at pH 6.6.

^cboth in units of µM.

^dCell cytotoxicity of HepG2 cells resulting in loss of 50% viability.

Next, we evaluated the new analogs against whole-cell M. tuberculosis H37Rv under iron-deficient conditions as previously described.²⁸ The minimum inhibitor concentrations (MIC) that resulted in complete inhibition of observable growth are shown in Table 1. Notably, all new analogs 5–10 retained fairly potent antitubercular activity with MIC values ranging from 0.098– 6.25 µM. Good correlation between appK_i and MIC values was observed for 1, 4, 6 and 8 with the MIC/appK_i ratio (both units in µM) varying only slightly with the median ratio of 72 ± 16 for these four compounds. However, the relative whole-cell activity for 3′-deoxy-2′-fluoro 10, p-aminosalicyl 5, and 2′-fluoro-2-phenyl 9 was eroded, with the MIC increasing to 0.78–3.125 µM despite potent enzyme inhibition. This is reflected in the MIC/appK_i ratios of 260, 351 and 975, respectively. The most deleterious modification was found in the di-functionalized N⁶-cyclopropyl-2-phenyl derivative 7, whose MIC of 6.25 µM is nearly 10,000-fold greater than the appK_i.

Pharmacokinetic Analysis

All compounds were evaluated in a crossover experiment employing doubly cannulated female Sprague-Dawley rats to determine pharmacokinetic parameters. While mice are normally used for these studies, we elected to use rats based on the ability to serially sample a single animal and perform oral and intravenous crossover studies, which greatly minimized inter-animal variability. Each compound was orally dosed (p.o.) at 25 mg/kg in three rats, and sampling was conducted at the indicated time points (see Experimental Section). After a three-day washout period, each rat was then dosed intravenously (i.v.) at 2.5 mg/kg as a single bolus with more intensive sampling at early time points (see Experimental Section). Pharmacokinetic parameters shown in Table 2 were calculated from concentration-time profiles by non-compartmental analysis. From these studies, the area under the concentration-time curve (AUC) for each dose and absolute oral bioavailability (F) were determined. Additional relevant parameters that will be discussed are the maximum serum concentration (C_max) for the p.o. dose and the terminal elimination half-life (t_1/2) and clearance (CL) for the i.v. dose. We began our studies by administering a known antitubercular agent, isoniazid (INH). The obtained PK parameters were consistent with several published studies of INH oral dosing and detection in rat plasma.^{52, 53} INH was rapidly eliminated with a t_1/2 of 19 min and achieved a C_max of 13.5 µg/mL, which translated to a 65% oral bioavailability. Our minimal target PK candidate profile would provide an oral bioavailability of 20%, C_max at least 50-fold greater than the MIC, and a t_1/2 greater than 1 hour.

Table 2.

Pharmacokinetic parameters of 1 analogs.

	i.v. PK parametersa			p.o. PK parametersb
Compound	AUC0-inf (min·µg·mL−1)	CL (mL·min–1·kg−1)	t1/2 (min)	Cmax (µg·mL−1)	AUC0-inf (µg·min·mL−1)	F (%)
1	519 ± 115	4.9 ± 0.9	11 ± 2	0.4 ± 0.2	66 ± 41	1.3 ± 0.7
2	nd	nd	nd	<LD	<LD	na
3	nd	nd	nd	0.03 ± 0.01	2.9 ± 0.8	na
4	304 ± 60	8.4 ± 1.9	10.6 ± 0.7	0.70 ± 0.07	99.8 ± 13.6	3.5 ± 0.2
5	63 ± 47	46.3 ± 4.5	10.6 ± 2.1	0.09 ± 0.01	16.8 ± 4.2	3.7 ± 0.3
6	209 ± 76	9.6 ± 0.1	15.1 ± 1.3	0.14 ± 0.01	18.1 ± 2.1	1.0 ± 0.4
7	184 ± 30	13.6 ± 2	15.5 ± 10.6	0.10 ± 0.01	4.0 ± 0.9	0.3 ± 0.1
8	1043 ± 326	2.5 ± 0.9	32.5 ± 7.4	0.46 ± 0.30	188 ± 110	1.8 ± 0.9
9	1981 ± 200	1.3 ± 0.1	121 ± 18	0.50 ± 0.16	140 ± 46	0.80 ± 0.15
10	959 ± 135	2.6 ± 0.5	62 ± 6.4	0.56 ± 0.12	198 ± 49	2.0 ± 0.08

^ai.v. dose (D_i.v.) = 2.5 mg/kg.

^bp.o. dose (D_p.o.) = 25 mg/kg. Values are reported as the mean ± SD. AUC_0-inf: area under the plasma concentration-time curve from time 0 to infinity, CL: clearance, t_1/2: terminal elimination half-life, C_max: maximum plasma concentration, F: relative oral bioavailability calculated as follows F = 100 × [(AUC_p.o × D_i.v.)/(AUC_i.v. × D_p.o.)], nd: not determined, na: not applicable, LD: below the limit of detection.

PK results of 1 and prodrugs

Following an oral dose, 1 reached a maximum serum concentration (C_max) of 0.4 µg/mL at 60 minutes with an AUC value of 66 min·µg/mL (Table 2). To determine the absolute bioavailability (F), 1 was then dosed i.v. to provide an AUC of 519 min·µg/mL that yielded an oral bioavailability of 1.3%. The terminal elimination half-life and clearance were 11 min and 4.9 mL/min·kg, respectively, indicating rapid elimination. While these results were discouraging, we note the maximum serum concentration (C_max) achieved is nearly two times the MIC value. The PK parameters obtained herein are nearly identical to those obtained in BALB/c mice (C_max = 1.2 µg/mL and AUC = 58.6 min·µg/mL) validating the Sprague-Dawley rat PK model as a practical surrogate for the murine PK model.

Microsomal stability studies with mouse, rat, and human microsomes demonstrated 1 is extremely stable to oxidative metabolism (t_1/2 > 30 min) and is not glucuronidated (0% glucuronide detected at 60 min). The intra-molecular hydrogen-bonding of the phenol with the sulfamate nitrogen could account for the lack of observed glucuronidation. Moreover, 1 is highly water soluble (>50 mg/mL) and stabile under both acidic (pH 1) and alkaline (pH 12) conditions showing less than 5% degradation after 24 h at 37 °C. Collectively, these results suggest renal clearance of this highly polar nucleoside as the likely elimination mechanism. We also assessed the membrane permeability using the Caco-2 monolayer model to provide a benchmark for other analogs. These studies indicate moderate permeability with a flux (P_{app(A to B)}) of 1.2 × 10⁻⁶ cm/s, and a favorable efflux ratio of 0.78 (Table 3).

Table 3.

Caco-2 permeability studies of select compounds.

Compound	Papp, A to B 10−6 cm/s	Papp, B to A 10−6 cm/s	efflux ratio
1	1.23	0.96	0.78
2	0.62	0.97	1.55
4	1.23	0.86	0.69
7	0.52	1.8	3.45
8	4.20	1.34	0.32

The acetate prodrugs 2 and 3 were dosed orally, but neither of the parent compounds was detected in the serum samples and only a negligible amount of 1 could be detected with 3. Serum stability studies of tri-acetate 3 (see Figure S8) indicated the phenol acetate was rapidly cleaved to liberate di-acetate 2, which was slowly hydrolyzed to 1. Caco-2 studies with 2 showed a reduced transport rate (P_{app(A to B)}) of 6.2 × 10⁻⁷ cm/s, which is approximately 50% of 1, as well as an efflux ratio of 1.55 that is twice that of 1 (Table 3).

Modulation of the Sulfamate pK_a

Our next attempt to enhance oral exposure involved modulation of the pK_a of the charged linker in 1, whose estimated pK_a is ∼3 based on the experimentally determined value for a related analogue.²⁸ This acidic functional group is expected to be ionized under physiological conditions, which may impede membrane transport, thus analogs with pK_a values closer to 7 would be desirable. Replacement of the sulfamate linkage with the sulfamide linkage in 4 is predicted to raise the pK_a to ∼5 while installation of the electron-donating para-amino group in 5 is expected to further raise the pK_a to ∼6.²⁸ Sulfamide 4 achieved a C_max of 0.7 µg/mL or 1.4 µM, which is 7-fold higher than the MIC. The oral exposure (AUC_p.o.) was increased 50% to 100 min·µg/mL that led to a nearly 3-fold improved bioavailability of 3.5%. Caco-2 studies of 4 did not indicate enhanced permeability since the P_{app(A to B)} relative to 1 was unchanged, while the efflux ratio was nearly identical (Table 3). Based on this modest improvement all subsequent analogs were prepared with the sulfamide linkage.

The para-amino analog 5 was then evaluated. Unfortunately, this compound had significantly lower oral exposure with an AUC_p.o. of only 17 min·µg/mL, and reached a C_max of 0.09 µg/mL, which is far below the MIC value. The i.v. exposure was also diminished, and can be attributed to the substantially faster clearance of 46.4 mL/min·kg, which is nearly 10-fold greater than 1, and the highest value of any analog in this series. The enhanced clearance of 5 is likely caused by the increased polarity of this molecule due to introduction of the para-amino group.

Impact of Lipophilicity

To evaluate the impact of lipophilicity on absorption, we evaluated N⁶-cyclopropyl 6 and N⁶-cyclopropyl-2-phenyl 7. These modifications increased the clogP values from −0.89 for compound 1 to 0.5 and 2.9, respectively. The C_max for both analogs ranged from 0.1 to 0.14 µg/mL while the AUC_p.o. values ranged from 4 to 19 min·µg/mL, indicating low oral exposure. The half-life of both compounds marginally improved to 15 min. The oral bioavailability of 6 was unchanged relative to 1 while the oral bioavailability of 7 was a dismal 0.3%. The increased lipophilicity of 7 manifested as poor aqueous solubility and several excipients (carboxymethylcellulose, 25% hydroxypropyl-β-cyclodextrin, and 50% PEG400–5% tween-80) were evaluated to potentially improve oral absorption without success. These results suggest that enhancing the lipophilicity through addition of nonpolar groups is detrimental to improving PK. In retrospect, the high molecular weight (MW) of 1 leaves little room to maneuver, and both 6 and 7 exceed Lipinski's threshold MW value of 500. The permeability of 7 as measured using Caco-2 cell studies corroborate the low bioavailability and indicate low permeability with a flux of 5.2 × 10⁻⁷ cm/s, about two-fold less than either 1 or 4, and an efflux ratio of 3.45.

Fluorinated Analogs

Fluorine is commonly used to improve drug properties and can increase oral exposure through reduction of cytochrome P450-mediated metabolic clearance. Fluorination at the C-2' position of nucleosides is also known to enhance anomeric stability, which can lead to an improved PK profile if degradation by cleavage of the anomeric linkage occurs at acidic pH.⁵⁴ Introduction of a fluorine at the C-2' position in 8 led to a nearly 3-fold increase in oral exposure relative to 1, with an AUC_p.o. of 188 min·µg/mL that is concordant with a 3-fold increase in half-life to 32.5 min (relative to 11 min for 1). Since the AUC_i.v. is also dramatically increased to 1043 min·µg/mL due to the increased half-life, the oral bioavailability remains low at 2%. Caco-2 studies of 8 demonstrated enhanced permeability with a flux of 4.2 × 10⁻⁶ cm/s, or nearly 4-fold greater than 1 and 4, and a favorable efflux ratio of 0.32 (Table 3). Unfortunately, this did not translate to improved bioavailability, as noted above, and suggests a limitation of the Caco-2 in vitro model for predicting oral bioavailability for this compound class. The 2′-fluoro-2-phenyl 9 showed a dramatic improvement in half-life to 121 min, which is 11-fold greater than 1, and resulted in an astonishing i.v. exposure of 1981 min·µg/mL. Unfortunately, the oral exposure (AUC_p.o.) was diminished relative to 8, but still 2-fold greater than 1. The oral bioavailability therefore remained low at 0.8%. 2′,3′-Dideoxy-2′-fluoro 10 was the last compound evaluated with the anticipation that removal of the polar 3′-hydroxyl group may enhance membrane permeability. Compound 10 indeed realized the highest oral exposure with an AUC_p.o. of 198 min·µg/mL and a C_max of 0.56 µg/mL. Additionally, the half-life was improved to 62 min, but the oral bioavailability remained unchanged relative to 8. Overall, these results demonstrate that replacement of the 2′-hydroxyl group with a fluorine atom results in a dramatically enhanced half-life, which translates to improved oral exposure.

Conclusion

Four complimentary strategies were examined to improve the drug disposition and pharmacokinetic properties of a new class of antibiotics that inhibit siderophore biosynthesis in M. tuberculosis. Our first approach involved preparation of simple di- and tri-acetate prodrugs (2 and 3) to mask the polar hydroxyl groups of the prototypical inhibitor 1. Unfortunately, neither prodrug was orally bioavailable. Plasma stability studies suggested rapid cleavage of 3 to 2 that underwent slow cleavage to afford 1. Alternate pro-moieties with faster cleavage kinetics could be explored to mask the 2′ and 3′ hydroxyls. The complete lack of oral exposure with 2 and 3 suggested this would not be a viable strategy and further studies along this line were not pursued.

We hypothesized that the ionized acyl-sulfonyl linker of the adenylation inhibitors described herein may impede membrane permeability. The negative charge is required for potent binding due to critical electrostatic interactions in the active site of MbtA and any attempts to remove this charge have resulted in complete loss of activity.²⁸ Thus, we designed analogs to increase the pK_a of the acyl-sulfonyl nitrogen linker. The sulfamide analog 4 possessed 3-fold improved bioavailability relative to 1, as well as 50% increased oral exposure and a more than 3-fold higher C_max/MIC ratio. Incorporation of a para-amino group on the salicyl moiety in 5 led to a significant loss in oral exposure. This is attributed to dramatically enhanced clearance caused by the additional polar amino group. Given the modest improvements realized with this approach, coupled with the strict requirements for an acyl-sulfonyl linker moiety from our prior extensive SAR studies, further modifications to the salicyl moiety were not pursued. Moreover, we later recognized these salicyl-nucleoside analogs can engage in intramolecular hydrogen bonding between the sulfonyl nitrogen and the ortho-phenol of the salicyl moiety. This could partially shield the polarity through delocalization of the charge and mitigate our initial concerns about the ionized acyl-sulfonyl linker.⁵⁵

The third approach to enhance ADME and PK properties involved modulation of the lipophilicity. While it is recognized that many antibacterials have unique physicochemical properties relative to other drug classes, we nevertheless sought to increase the lipophilicity of the highly polar nucleoside derivative 1.⁵⁶ This was achieved through addition of non-polar groups onto the nucleoside at sites previously shown to enhance biochemical potency and antibacterial activity.³⁴ The calculated logP of N⁶-cyclopropyl 6 and N⁶-cyclopropyl-2-phenyl 7 was increased to +0.5 and +2.9, respectively. However, both compounds realized dramatically lower oral exposures and failed to even reach a C_max exceeding their MIC values. A more fruitful approach to increasing lipophilicity may be through strategic deletion of polar functional groups and atoms that are dispensable for activity, as this would additionally result in a reduction of the molecular weight.

Incorporation of a fluorine atom at C-2′ of the nucleoside in analogs 8–10 led to the most pronounced improvements in oral exposure, primarily through an increase in the terminal elimination half-life (t_1/2) following i.v. administration from 11 minutes for 1 to 32–121 minutes for 8–10. The apparent terminal elimination half-lives following oral dosing for all of the compounds described herein are much longer, presumably due to slow absorption, and do not reflect the rate of elimination. To calibrate these values, the intrinsic half-life of the first-line antitubercular agent isoniazid in a rat is only 18 minutes. Fluorine is commonly introduced to enhance t_1/2 through blocking P450-mediated metabolism.⁴⁵ In the case of nucleosides, a C-2′ fluoro group improves the stability of the anomeric linkage, which can be important for some nucleosides that degrade by anomeric cleavage under the acidic conditions of gastric fluids.⁵⁴ We have shown 1 and its derivatives to be extremely stable under strongly acidic conditions, while microsomal stability studies revealed little metabolism of 1. Thus, we hypothesize that the C-2′ fluoro group likely decreases renal clearance.

In order to provide an in vitro marker for oral bioavailability, we also measured select analogs using the standard Caco-2 model system for intestinal epithelial permeability. These experiments showed the Caco-2 model may be qualitatively useful, but is not quantitatively predictive of oral bioavailability. We identified a minimum threshold P_{app(A to B)} of 1 × 10⁻⁶ cm/s that can now be used to eliminate candidates prior to PK evaluation.

Through logical modification of 1 we explored multiple strategies to enhance pharmacokinetic parameters, and successfully increased the intrinsic half-life and oral exposure. While were unable to reach our minimal target PK candidate profile, we believe these studies have provided a foundation to more effectively address the PK limitations of 1. Our findings may also be applicable to other adenylation inhibitors, which have been described for dozens of other functionally related adenylating enzymes involved in diverse disease processes.^{57, 58}

EXPERIMENTALS

General materials and methods

Chemicals and solvents were purchased from Acros Organics, Alfa Aesar, Sigma-Aldrich, and TCI America and were used as received. The nucleosides 2′-deoxy-2′-fluoroguanosine 33 and 2′-deoxy-2′-fluoroadenosine 42 were obtained from Metkinen Chemistry (Kuopio, Finland). Compounds 1,²⁸ 11,³⁹ 14,³⁹ 19,³⁴ 23,³⁵ 26,⁵⁹ 27,⁵⁹ 29,⁴⁰ 30,⁴⁴ 31,³² 32,⁴³ and N-hydroxysuccinimidyl 2-benzyloxybenzoate²⁸ were prepared as described. All final compounds (1–10) were purified by preparative reverse-phase HPLC using the indicated method and purity (≥95%) of the lyophilized powders was confirmed by analytical reverse-phase HPLC. An anhydrous solvent dispensing system using two packed columns of neutral alumina was used for drying THF and CH₂Cl₂, while two packed columns of molecular sieves were used to dry DMF, and the solvents were dispensed under argon gas (Ar). Anhydrous grade MeOH, MeCN, pyridine, and DMA were purchased from Aldrich. EtOAc and hexanes were purchased from Fisher Scientific. All reactions were performed under an inert atmosphere of dry argon gas (Ar) in oven-dried (180 °C) glassware. TLC analyses were performed on TLC silica gel 60F254 plates from EMD Chemical Inc. and were visualized with UV light. Purification by flash chromatography was performed using a medium-pressure flash chromatography system equipped with flash column silica cartridges with the indicated solvent system. Reversed-phase HPLC (RP-HPLC) purification was performed on Phenomenex Gemini 10 µm C18 columns (250 × 21.2 or 10 or 4.6 mm) with detection at 254 nm employing the indicated solvent system. ¹H and ¹³C spectra were recorded on a 400 or 600 MHz Varian, or 700 MHz Bruker NMR spectrometer. Proton chemical shifts are reported in ppm from an internal standard of residual chloroform (7.26), methanol (3.31), dimethyl sulfoxide (2.50), or mono-deuterated water (HDO, 4.79); carbon chemical shifts are reported in ppm from an internal standard of residual chloroform (77.0), methanol (49.1), or dimethyl sulfoxide (39.5). Proton chemical data are reported as follows: chemical shift, multiplicity (s = singlet, d = doublet, dt = doublet of triplets, t = triplet, q = quartet, pentet = pent, m = multiplet, ap = apparent, br = broad, ovlp = overlapping), coupling constant(s), integration. High-resolution mass spectra were obtained on an LTQ Orbitrap Velos (Thermo Scientific, Waltham, MA).

General Procedure for Mitsunobu Reaction

To a solution of nucleoside (1.0 equiv), 29 (3.0 equiv) and PPh₃ (1.05 equiv) in THF (0.05 M for limiting reagent) at 0 °C was added neat DIAD (1.05 equiv) dropwise over 15 min. The solution was stirred at 0 °C for 45 min, then gradually warmed to 23 °C over 3 h and stirred for 12 h at 23 °C. The reaction was concentrated under reduced pressure to a dark yellow oil and immediately purified by flash chromatography (linear gradient 0–100% EtOAc–hexanes) to afford the title compound.

General Procedure for Salicylation

To a solution the sulfamide nucleoside (1.0 equiv) and Cs₂CO₃ (2.5 equiv) in DMF (0.5 M for limiting reagent) at 0 °C was added 31 (1.5 equiv). The reaction mixture was then stirred for 24 h at 23 °C during which time all starting material was consumed as monitored by electrospray mass spectrometry in the negative mode. DMF was removed by rotary evaporation under high vacuum (P = 0.01 torr) to afford crude salicylated nucleoside. The crude product was used in the next global deprotection reaction directly without further purification.

General Procedure for TFA Global Deprotection

To a solution of the crude salicylated nucleoside in H₂O (10 mL/mmol) at 0 °C was added TFA (40 mL/mmol) dropwise (final concentration 0.02 M). The reaction mixture was stirred for 8 h at 23 °C during which time all starting material was consumed as monitored by electrospray mass spectrometry in the negative mode. The solvent was removed under vacuum. The crude material was re-dissolved in 1:1 MeCN–50 mM TEAB (10–20 mg/mL) and filtered to remove insoluble solids. The resulting solution was purified by reverse phase HPLC employing the indicated method and column using 50 mM aqueous triethylammonium bicarbonate (TEAB) at pH 7.5 (solvent A) and acetonitrile (solvent B). The appropriate fractions were pooled and lyophilized to afford the final compound as the triethylammonium salt as a white foam.

5′ -O-[N-(2-Benzyloxybenzoyl)sulfamoyl]-N⁶,N⁶-bis(tert-butyloxycarbonyl)-2′, 3′ -O-isopropylideneadenosine triethylammonium salt (12)

To a stirring solution of 11 (0.54 g, 0.92 mmol, 1.0 equiv) in DMF (9.2 mL) cooled to 0 °C was added N-hydroxysuccinimidyl 2-benzyloxybenzoate (0.45 g, 1.38 mmol, 1.5 equiv) followed by Cs₂CO₃ (0.60 g, 1.84 mmol, 2.0 equiv) and the mixture was stirred at 23 °C for 21 h. The crude reaction mixture was concentrated under reduced pressure to a yellow oil that was taken up in EtOAc (200 mL) and filtered. The precipitate was washed with additional EtOAc (200 mL), and the combined filtrate was concentrated under reduced pressure and purified by flash chromatography (linear gradient 0–5% MeOH/EtOAc with 1% Et₃N) to afford the title compound (0.51 g, 62%) as a white foamy oil: R_f = 0.37 (1:9 MeOH–EtOAc with 1% Et₃N); ¹H NMR (600 MHz, CD₃OD) δ 1.13 (t, J = 7.2 Hz, 9H), 1.35 (s, 3H), 1.40 (s, 18H), 1.61 (s, 3H), 2.83 (q, J = 7.2 Hz, 6H), 4.19–4.20 (m, 2H), 4.41 (dd, J = 6.0, 3.6 Hz, 1H), 5.12–5.13 (m, 3H), 5.38 (dd, J = 6.0, 3.6 Hz, 1H), 6.37 (d, J = 3.0 Hz, 1H), 6.95 (t, J = 7.8 Hz, 1H), 7.05 (d, J = 8.4 Hz, 1H), 7.25 (t, J = 7.8 Hz, 1H), 7.31–7.34 (m, 3H), 7.46 (d, J = 7.8 Hz, 1H), 7.52 (d, J = 7.2 Hz, 2H), 8.84 (s, 1H), 8.89 (s, 1H); ¹³C NMR (150 MHz, CD₃OD) δ 10.2, 25.7, 27.8, 28.2, 47.4, 69.8, 71.6, 83.1, 85.4, 86.0, 86.2, 92.5, 115.3, 121.3, 121.9, 128.9, 129.2, 129.3, 129.6, 129.9, 130.3, 131.7, 131.9, 138.7, 146.8, 151.2, 151.7, 153.3, 154.6, 157.2, 176.9; HRMS (ESI–) calcd for C₃₇H₄₃N₆O₁₂S [M – H]⁻ 795.2665, found 795.2659 (0.8 ppm error).

2′, 3′ -O-di-Acetyl-5′ -O-[N-(2-hydroxybenzoyl)sulfamoyl]adenosine (2)

To a flask containing 12 (0.51 g, 0.56 mmol, 1.0 equiv) at 0 °C was added 80% aqueous TFA (6 mL) that had also been cooled to 0 °C and the mixture stirred at 0 °C for 1 h then at 23 °C for 3 h. The reaction mixture was concentrated under reduced pressure to a yellow oil and dried under high vacuum to remove all traces of TFA, then dissolved in DMF (6 mL) to which was added pyridine (0.17 mL, 2.1 mmol, 3.8 equiv) followed by acetic anhydride (0.33 mL, 3.5 mmol, 6.4 equiv) and the mixture stirred at 23 °C for 16 h. The reaction was concentrated under reduced pressure to remove all traces of DMF and give 13 as a crude yellow oil that was used directly in the next step without purification.

To a round-bottomed flask containing the crude product 13 prepared above (0.25 g, 0.33 mmol, 1.0 equiv) purged with argon was added anhydrous MeOH (3.4 mL) followed by Pd/C (0.030 g, 10% by weight) and the flask evacuated by house vacuum and back-filled with an H₂ balloon 3 times before being allowed to stir vigorously under an H₂ balloon at 23 °C for 2 h. The crude reaction mixture was filtered over a pad of Celite and concentrated under reduced pressure to a dark oil. Purification by semi-preparative reverse-phase HPLC on a Phenomenex Gemini 10 µm C18 (250 × 10.0 mm) column at a flow rate of 5 mL/min with isocratic elution of 35% MeCN–50 mM aqueous triethylammonium bicarbonate for 10 min, followed by 70% MeCN for 5 min. The retention time of the product was 5.4 minutes (k′ = 1.2) and the appropriate fractions were pooled and lyophilized to afford the title compound (90.0 mg, 42%) as a white solid: R_f = 0.18 (1:9 MeOH–CH₂Cl₂ with 1% Et₃N); ¹H NMR (600 MHz, CD₃OD) δ 1.23 (t, J = 7.2 Hz, 9H), 2.00 (s, 3H), 2.10 (s, 3H), 3.03 (q, J = 7.2 Hz, 6H), 4.44 (dd, J = 12.0, 3.0 Hz, 1H), 4.47 (dd, J = 11.4, 3.6 Hz, 1H), 4.53 (q, J = 3.6 Hz, 1H), 5.70 (dd, J = 5.4, 3.6 Hz, 1H), 5.89 (t, J = 6.0 Hz, 1H), 6.27 (d, J = 6.6 Hz, 1H), 6.75–6.79 (m, 2H), 7.28 (t, J = 7.2 Hz, 1H), 7.93 (d, J = 7.8 Hz, 1H), 8.18 (s, 1H), 8.56 (s, 1H); ¹H NMR (600 MHz, DMSO-d₆) δ 1.17 (t, J = 7.2 Hz, 9H), 1.99 (s, 3H), 2.10 (s, 3H), 3.09 (q, J = 7.2 Hz, 6H), 4.25 (dd, J = 11.4, 4.8 Hz, 1H), 4.33 (dd, J = 11.4, 4.2 Hz, 1H), 4.44 (q, J = 4.2 Hz, 1H), 5.56 (dd, J = 5.4, 4.2 Hz, 1H), 5.90 (t, J = 6.0 Hz, 1H), 6.20 (d, J = 6.0 Hz, 1H), 6.72–6.75 (m, 2H), 7.26 (t, J = 7.2 Hz, 1H), 7.36 (br s, 2H), 7.81 (d, J = 7.8 Hz, 1H), 8.14 (s, 1H), 8.46 (s, 1H), 13.5 (s, 1H); ¹³C NMR (150 MHz, CD₃OD) δ 9.9, 20.3, 20.6, 47.8, 69.1, 73.0, 75.3, 82.7, 87.1, 118.0, 119.4, 120.2, 120.7, 131.6, 134.5, 141.3, 150.8, 154.2, 157.5, 162.3, 171.0, 171.5, 175.3; HRMS (ESI–) calcd for C₂₁H₂₁N₆O₁₀S [M – H]⁻ 549.1045, found 549.1048 (0.5 ppm error).

5′-O-[N-(2-Acetoxybenzoyl)sulfamoyl]-2′, 3′-O-di-acetyladenosine (3)

To a stirring solution of 1 (0.10 g, 0.18 mmol, 1.0 equiv) in DMF was added pyridine (0.05 mL, 0.68 mmol, 3.8 equiv) followed by acetic anhydride (0.11 mL, 1.13 mmol, 6.4 equiv) and the mixture stirred at 23 °C for 21 h. The crude reaction mixture was concentrated under reduced pressure to a yellow oil and purified by flash chromatography (linear gradient 0–5% MeOH/DCM with 1% Et₃N) to afford an approximately 5:1 mixture of 3 and 2. The desired compound was purified by semi-preparative reverse-phase HPLC on a Phenomenex Gemini 10 µm C18 (250 × 10.0 mm) column at a flow rate of 5 mL/min with a linear gradient 20–70% MeCN–50 mM aqueous triethylammonium bicarbonate for 10 min, followed by 70% MeCN for 5 min. The retention time of the product was 6.6 minutes (k′ = 1.6) and the appropriate fractions were pooled and lyophilized to afford the title compound (57.8 mg, 46%) as a white solid: R_f = 0.25 (1:9 MeOH– CH₂Cl₂ with 1% Et₃N); ¹H NMR (600 MHz, CD₃OD) δ 1.27 (q, J = 7.2 Hz, 9H), 2.00 (s, 3H), 2.13 (s, 3H), 2.26 (s, 3H), 3.17 (t, J = 6.0 Hz, 6H), 4.40 (dd, J = 14.4, 3.6 Hz, 1H), 4.44 (dd, J = 9.6, 3.0 Hz, 1H), 4.54 (d, J = 3.0 Hz, 1H), 5.70–5.71 (m, 1H), 5.89 (t, J = 6.0 Hz, 1H), 6.29 (d, J = 6.0 Hz, 1H), 7.02 (d, J = 7.8 Hz, 1H), 7.06 (s, 1H), 7.23 (t, J = 7.8 Hz, 1H), 7.41 (t, J = 7.2 Hz, 1H), 7.86 (d, J = 7.8 Hz, 1H), 8.19 (s, 1H), 8.59 (s, 1H); ¹³C NMR (150 MHz, CD₃OD) δ 9.4, 20.3, 20.6, 21.4, 48.0, 69.1, 73.2, 75.3, 82.9, 87.0, 120.2, 124.1, 126.7, 131.9, 132.2, 133.1, 141.3, 150.8, 150.9, 154.2, 157.2, 171.0, 171.5, 172.0, 173.8; HRMS (ESI–) calcd for C₂₃H₂₃N₆O₁₁S [M – H]⁻ 591.1151, found 591.1170 (3.2 ppm error).

N⁶,N⁶-bis(tert-Butoxycarbonyl)-5′-N-[N-(tert-butoxycarbonyl)-N-sulfamoyl]amino-5′ -deoxy-2′, 3′-O-isopropylideneadenosine (15)

The title compound was prepared from 14 (0.60 g, 1.2 mmol) using the general procedure for the Mitsunobu reaction and isolated as a foamy oil (0.789 g, 96%): R_f = 0.25 (1:1 EtOAc–hexanes); ¹H NMR (600 MHz, CDCl₃) δ 1.40 (s, 21H), 1.46 (s, 9H), 1.49 (s, 3H), 1.61 (s, 3H), 3.94 (dd, J = 15.0, 7.8 Hz, 1H), 3.99 (dd, J = 15.6, 5.4 Hz, 1H), 4.51 (p, J = 4.8 Hz, 1H), 4.90 (p, J = 6.0 Hz, 1H), 5.24 (dd, J = 6.0, 3.6 Hz, 1H), 5.57 (dd, J = 6.0, 1.8 Hz, 1H), 6.38 (d, 1H, 1.8 Hz), 8.64 (s, 1H), 8.88 (s, 1H); ¹³C NMR (150 MHz, CDCl₃) δ 27.6, 28.2, 28.4, 28.5, 49.9, 83.6, 85.1, 85.4, 85.6, 87.5, 91.6, 115.6, 130.7, 147.6, 151.1, 151.5, 153.1, 153.2, 154.2; HRMS (ESI–) calcd for C₂₈H₄₂N₇O₁₁S [M – H]⁻ 684.2668, found 684.2685 (2.5 ppm error).

N⁶,N⁶-bis(tert-Butoxycarbonyl)-5′-N-(N-tert-butoxycarbonyl-N-{[(2-O-methoxymethoxy)benzoyl]sulfamoyl})amino-5′-deoxy-2′, 3′ -O-isopropylideneadenosine triethylammonium salt (16)

The title compound was prepared from 15 (0.52 g, 0.75 mmol, 1.0 equiv) using the general procedure for salicylation and isolated as an off-white amorphous solid (0.38 g, 53%): R_f = 0.37 (1:9 MeOH–CH₂Cl₂ with 1% Et₃N); ¹H NMR (600 MHz, CDCl₃) δ 1.26 (t, J = 7.2 Hz, 9H), 1.37 (s, 3H), 1.39 (s, 18H), 1.45 (s, 9H), 1.62 (s, 3H), 3.11 (q, J = 7.2 Hz, 6H), 3.45 (s, 2H), 4.06 (dd, J = 15.0, 6.0 Hz, 1H), 4.64 (dd, J = 18.0, 7.2, Hz, 1H), 4.60–4.63 (m, 1H), 5.14 (s, 2H), 5.28 (dd, J = 6.0, 3.0 Hz, 1H), 5.51 (dd, J = 6.6, 3.0 Hz, 1H), 6.32 (d, 2.4 Hz, 1H), 6.96 (t, J = 7.2 Hz, 1H), 7.11 (d, J = 8.4 Hz, 1H), 7.27 (t, J = 7.8 Hz, 1H), 7.51 (d, J = 7.8 Hz, 1H), 8.74 (s, 1H), 8.89 (s, 1H); ¹³C NMR (150 MHz, CDCl₃) δ 7.9, 27.7, 28.0, 46.1, 52.8, 56.3, 63.3, 81.4, 82.3, 83.7, 84.5, 85.0, 90.3, 96.5, 114.3, 118.0, 121.9, 129.0, 129.8, 131.3, 144.0, 150.0, 150.3, 151.6, 152.2, 152.5, 153.0, 154.5; HRMS (ESI–) calcd for C₃₇H₅₀N₇O₁₄S [M – H]⁻ 848.3142, found 848.3146 (0.5 ppm error).

5′-N-[N-(2-Hydroxybenzoyl)sulfamoyl]amino-5′ -deoxyadenosine triethylammonium salt (4)

The title compound was prepared (150 mg, 65%) from 16 (380 mg, 0.39 mmol) using the general procedure for TFA deprotection and purified by preparative-reverse phase HPLC as described. The analytical data (¹H NMR, ¹³C NMR, HRMS) agreed with previously reported values for 4 obtained by an alternate synthetic route.²⁸

5′-N-[N-(4-Benzyloxycarbonylamino-2-hydroxybenzoyl)-N-(tert-butoxycarbonyl)sulfamoyl]amino-N⁶,N⁶-bis(tert-butoxycarbonyl)-5′-deoxy-2′,3′-O-isopropylideneadenosine triethylammonium salt (17)

To a stirring solution of 32 (0.75 g, 2.6 mmol, 3.0 equiv) in DMF (4.4 mL) was added CDI (0.51 g, 3.2 mmol, 3.6 equiv) and the mixture heated at 60 °C for 2 h. The reaction mixture was cooled to 23 °C, and 15 (0.60 g, 0.87 mmol, 1.0 equiv) was added as a solution in DMF (4.4 mL) followed by solid Cs₂CO₃ (0.43 g, 1.31 mmol, 1.5 equiv), and stirred for 15 h. The crude reaction mixture was concentrated under reduced pressure to a dark oil and purified by flash chromatography (linear gradient 0–5% MeOH–CH₂Cl₂ with 1% Et₃N) to afford the title compound (0.46 g, 50%) as a waxy light yellow solid: R_f = 0.61 (1:9 MeOH–CH₂Cl₂ with 1% Et₃N); ¹H NMR (600 MHz, CD₃OD) δ 1.09 (t, J = 7.2 Hz, 9H), 1.36 (s, 9H), 1.38 (s, 18H), 1.40 (s, 3H), 1.61 (s, 3H), 2.67 (q, J = 7.2 Hz, 6H), 4.05 (dd, J = 15.0, 6.6 Hz, 1H), 4.18 (dd, J = 15.0, 7.2 Hz, 1H), 4.63 (td, J = 7.2, 3.0 Hz, 1H), 5.17 (s, 2H), 5.27 (dd, J = 5.4, 3.0 Hz, 1H), 5.57 (dd, J = 6.0, 3.0 Hz, 1H), 6.33 (d, J = 2.4 Hz, 1H), 6.82 (d, J = 9.0 Hz, 1H), 7.06 (s, 1H), 7.30 (d, J = 7.2 Hz, 1H), 7.36 (t, J = 7.8 Hz, 1H), 7.41 (d, J = 7.8 Hz, 1H), 7.75 (d, J = 8.4 Hz, 1H), 8.74 (s, 1H), 8.89 (s, 1H); ¹³C NMR (150 MHz, CD₃OD) δ 10.9, 25.7, 27.7, 28.2, 28.5, 47.3, 53.8, 67.7, 83.8, 84.0, 85.4, 87.3, 92.1, 106.8, 109.9, 115.5, 116.0, 129.1, 129.3, 129.7, 130.7, 131.9, 134.6, 138.1, 145.0, 147.3, 151.2, 151.6, 153.3, 154.4, 154.5, 155.4, 162.4, 173.6; HRMS (ESI–) calcd for C₄₃H₅₃N₈O₁₅S [M – H]⁻ 953.3357, found 953.3328 (3.0 ppm error).

5′-N-[N-(4-Amino-2-hydroxybenzoyl)sulfamoyl]amino-5′-deoxyadenosine (5)

To a Parr flask purged with Ar was added 17 (0.50 g, 0.52 mmol, 1.0 equiv) dissolved in MeOH (5.2 mL) followed by Pd/C (0.05 g, 10% by weight). The flask was evacuated and back-filled with H₂ five times before filling the flask to 40 psi H₂ and shaking on a Parr shaker at 23 °C. After 1 h, the reaction was filtered over a pad of Celite and concentrated under reduced pressure to afford 18 as a colorless oil that was used directly in the next step without purification.

To a flask containing the crude 18 prepared above at 0 °C was added 80% aqueous TFA (5.2 mL) that had also been cooled to 0 °C. The reaction was stirred for 1 h at 0 °C, 12 h at 4 °C, and 5 h at 23 °C. The reaction was concentrated under reduced pressure to remove all traces of TFA. Purification by preparative reverse-phase HPLC on a Phenomenex Gemini 10 µm C18 (250 × 21.2 mm) column was achieved at a flow rate of 20 mL/min with a gradient of 5–25% MeCN–50 mM aqueous triethylammonium bicarbonate over 10 min, followed by 25% MeCN for 5 min. The retention time of the product was 9.0 minutes (k′ = 2.0) and the appropriate fractions were pooled and lyophilized to afford the title compound (90 mg, 34%) as a white solid: ¹H NMR (600 MHz, CD₃OD) δ 1.20 (t, J = 7.2 Hz, 9H), 2.95 (q, J = 7.2 Hz, 6H), 3.29–3.31 (m, 2H), 4.23 (dd, J = 6.6, 4.2 Hz, 1H), 4.35 (dd, J = 5.4, 3.0 Hz, 1H), 4.85–4.86 (m, 1H), 5.93 (d, J = 6.6 Hz, 1H), 6.05 (s, 1H), 6.12 (d, J = 9.0 Hz, 1H), 7.62 (d, J = 9.0 Hz, 1H), 8.30 (s, 1H), 8.31 (s, 1H); ¹³C NMR as determined by HMBC (150 MHz, CD₃OD) δ 10.1, 46.6, 47.7, 73.2, 75.0, 86.1, 90.6, 101.9, 107.4, 110.9, 120.9, 132.4, 142.0, 150.7, 154.3, 154.6, 157.5, 163.7, 175.0; HRMS (ESI-) calcd for C₁₇H₁₉N₈O₇S [M – H]⁻ 479.1103, found 479.1095 (1.7 ppm error).

5′-Amino-N⁶-cyclopropyl-5′-deoxy-2′,3′-O-isopropylideneadenosine (20)

DIAD (3.4 mL, 17.3 mmol, 1.5 equiv) was dissolved in THF (300 mL) that had been cooled to 0 °C. Ph₃P (4.5 g, 17.3 mmol, 1.5 equiv) was added as a solution in THF (60 mL), and the solution stirred at 0 °C for 15 min. Compound 19 (4.0 g, 11.5 mmol, 1.0 equiv) was added as a solution in THF (60 mL), and the solution stirred at 0 °C for 15 min. DPPA (4.96 mL, 23.0 mmol, 2.0 equiv) was added directly and the solution stirred at 0 °C for 15 min then at 23 °C for 4.5 h. The reaction was quenched with H₂O (200 mL). The mixture was extracted with CH₂Cl₂ (3 × 200 mL), and the combined organic layers were dried over MgSO₄, filtered, and concentrated under reduced pressure to a clear oil that was dried under high vacuum for at least 4 h before being used directly in the next step.

To a Parr flask purged with Ar was added Pd/C (0.185 g, 10% by weight, added in 2 portions), followed by half the crude reaction mixture from above dissolved in anhydrous MeOH (50 mL). The flask was evacuated and back-filled with H₂ five times before filling the flask to 30 psi H₂ and shaking on a Parr shaker at 23 °C. After 1.75 h, the reaction was filtered over a pad of Celite, concentrated under reduced pressure to a dark oil, and purified by flash chromatography (linear gradient 0–10% MeOH–CH₂Cl₂). The same procedure was repeated with the second half of the crude reaction mixture to give the title compound (0.76 g, 44% over 2 steps) as a foamy oil: R_f = 0.05 (1:9 MeOH–CH₂Cl₂); ¹H NMR (600 MHz, CD₃OD) δ 0.61–0.64 (m, 2H), 0.85–0.88 (m, 2H), 1.35 (s, 3H), 1.57 (s, 3H), 2.92 (d, J = 5.4 Hz, 2H), 3.35 (s, 1H), 4.21–4.24 (m, 1H), 5.01 (dd, J = 6.6, 3.6 Hz, 1H), 5.46 (dd, J = 6.6, 3.0 Hz, 1H), 6.13 (d, J = 3.0 Hz, 1H), 8.21 (s, 1H), 8.28 (s, 1H); ¹³C NMR (150 MHz, CD₃OD) δ 7.7, 25.7, 27.7, 44.6, 55.0, 83.3, 85.0, 88.3, 91.70, 91.71, 115.7, 121.4, 141.7, 154.0, 157.2; HRMS (ESI+) calcd for C₁₆H₂₃N₆O₃ [M + H]⁺ 347.1826, found 347.1821 (1.4 ppm error).

N⁶-Cyclopropyl-5′-deoxy-2′,3′-O-isopropylidene-5′-N-(sulfamoyl)aminoadenosine (21)

To a stirring solution of 20 (0.50 g, 1.4 mmol, 1.0 equiv) in CH₂Cl₂ (20 mL) was added 30 (0.63 g, 1.9 mmol, 1.3 equiv) and the mixture stirred at 23 °C for 14 h. The crude reaction was concentrated under reduced pressure to a foamy oil that was used directly in the next step.

To a Parr flask purged with Ar was added Pd/C (0.075 g, 10% by weight), followed by the crude reaction mixture from above dissolved in 2:1 MeOH–THF (15 mL). The flask was evacuated and back-filled with H₂ five times before filling the flask to 30 psi H₂ and shaking on a Parr shaker at 23 °C for 22 h. The reaction was filtered over a pad of Celite, concentrated under reduced pressure to a clear oil, and purified by flash chromatography (linear gradient 0–10% MeOH–CH₂Cl₂) to afford the title compound (0.35 g, 62% over 2 steps) as a clear foamy oil: R_f = 0.13 (1:9 MeOH–CH₂Cl₂); ¹H NMR (600 MHz, CD₃OD) δ 0.63 (dd, J = 8.4, 4.2 Hz, 2H), 0.87 (dd, J = 7.2, 4.8 Hz, 2H), 1.34 (s, 3H), 1.59 (s, 3H), 3.35–3.36 (m, 1H), 3.37–3.41 (m, 2H), 4.46 (d, J = 2.4 Hz, 1H), 5.09 (dd, J = 6.0, 2.4 Hz, 1H), 5.31 (dd, J = 6.0, 4.2 Hz, 1H), 6.02 (d, J = 4.2 Hz, 1H), 8.15 (s, 1H), 8.34 (s, 1H); ¹³C NMR (150 MHz, CD₃OD) δ 7.7, 25.7, 27.8, 46.1, 50.0, 83.2, 84.5, 85.3, 93.1, 115.8, 121.5, 141.8, 154.1, 157.3; HRMS (ESI–) calcd for C₁₆H₂₂N₇O₅S [M – H]⁻ 424.1409, found 424.1415 (1.4 ppm error).

N⁶-Cyclopropyl-5′-deoxy-5′-N-[N-(2-hydroxybenzoyl)sulfamoyl]aminoadenosine triethylammonium salt (6)

The title compound was prepared from 21 (0.20 g, 0.47 mmol) using the general procedures for salicylation and TFA deprotection. Purification by preparative reverse-phase HPLC on a Phenomenex Gemini 10 µm C18 (250 × 21.2 mm) column at a flow rate of 20 mL/min isocratic at 22% MeCN in 50 mM aqueous triethylammonium bicarbonate for 10 min, followed by washing the column with 90% MeCN for 5 min prior to re-equilibration. The retention time of the product was 7.2 minutes (k′ = 1.4) and the appropriate fractions were pooled and lyophilized to afford the title compound (70.2 mg, 61%) as a white solid: ¹H NMR (600 MHz, CD₃OD) δ 0.63–0.66 (m, 2H), 0.87–0.90 (m, 2H), 1.12 (t, J = 7.8 Hz, 9H), 2.65 (s, 1H), 2.77 (q, J = 7.2 Hz, 6H), 3.33 (d, J = 4.2 Hz, 2H), 4.25 (q, J = 4.2 Hz, 1H), 4.37 (q, J = 3.0 Hz, 1H), 4.87 (solvent ovlp, 1H), 5.93 (d, J = 6.6 Hz, 1H), 6.75–6.79 (m, 2H), 7.26 (t, J = 7.8 Hz, 1H), 7.88 (d, J = 7.8 Hz, 1H), 8.24 (s, 1H), 8.40 (s, 1H); ¹³C NMR (176 MHz, CD₃OD) δ 7.7, 10.5, 40.6, 46.7, 47.4, 73.2, 74.9, 86.0, 90.8, 118.0, 119.2, 121.2, 121.5, 131.0, 134.0, 141.8, 154.2, 157.3, 162.1, 174.3; HRMS (ESI–) calcd for C₂₀H₂₂N₇O₇S [M – H]⁻ 504.1307, found 504.1313 (1.1 ppm error).

5′-N-[N-(tert-Butoxycarbonyl)-N-sulfamoyl]amino-N⁶-cyclopropyl-5′-deoxy-2′, 3′-O-isopropylidene-2-phenyladenosine (24)

The title compound was prepared from 23 (0.48 g, 1.1 mmol) using the general procedure for the Mitsunobu reaction and isolated as a foamy oil (0.868 g, 93%): R_f = 0.19 (1:1 EtOAc–hexane); ¹H NMR (400 MHz, CDCl₃) δ 0.74 (br s, 2H), 0.94 (d, J = 5.9 Hz, 2H), 1.28–1.25 (m, 1H), 1.41 (s, 3H), 1.48 (s, 9H), 1.62 (s, 3H), 3.22 (s, 1H), 4.02– 4.10 (m, 2H), 4.33 (q, J = 5.7 Hz, 1H), 5.26 (t, J = 5.4 Hz, 1H), 5.45 (s, 2H), 5.52 (d, J = 6.5 Hz, 1H), 6.12 (s, 1H), 7.44–7.52 (m, 3H), 7.89 (s, 1H), 8.42–8.46 (m, 2H); ¹³C NMR (100 MHz, CD₃OD) δ 6.2, 24.2, 26.1, 26.7, 27.1, 48.8, 82.5, 83.5, 84.1, 85.9, 90.0, 114.1, 118.3, 127.90 (2C), 127.93, 129.7, 140.8, 151.8, 155.0, 158.8 (missing 1 aryl C); HRMS (ESI+) calcd for C₂₇H₃₅N₇O₇SNa [M + Na]⁺ 624.2211, found 624.2255 (7.1 ppm error).

N⁶-Cyclopropyl-5′-deoxy-5′-N-[N-(2-hydroxybenzoyl)sulfamoyl]amino-2-phenyladenosine triethylammonium salt (7)

The title compound was prepared from 24 (0.45 g, 0.75 mmol) using the general procedure for salicylation and TFA deprotection. The crude product was purified by silica gel flash chromatography (10% MeOH–CH₂Cl₂ with 1% Et₃N) followed by preparative reverse-phase HPLC on a Phenomenex Gemini 10 µm C18 (250 × 21.2 mm) column at a flow rate of 20 mL/min with a gradient of 20–60% MeCN in 50 mM aqueous triethylammonium bicarbonate pH 7.5 over 20 min. The retention time of the product was 13.40 minutes (k′ = 3.9) and the appropriate fractions were pooled and lyophilized to afford the title compound (80 mg, 18% over 2 steps) as a white solid: ¹H NMR (600 MHz, CDCl₃) δ 0.64–0.57 (m, 2H), 0.77–0.86 (m, 2H), 1.11 (t, J = 7.3 Hz, 9H), 2.95 (q, J = 7.1 Hz, 6H), 3.01 (br s, 1H), 3.33–3.44 (m, 2H), 4.32–4.38 (m, 2H), 4.56–4.60 (m, 1H), 4.72 (t, J = 5.1 Hz, 1H), 6.00 (d, J = 5.4 Hz, 1H), 6.71 (t, J = 7.5 Hz, 1H), 6.81 (d, J = 8.2 Hz, 1H), 7.22 (t, J = 7.2 Hz, 1H), 7.35–7.40 (m, 3H), 7.85 (d, J = 7.3 Hz, 1H), 8.01 (s, 1H), 8.34 (d, J = 7.10 Hz, 2H); ¹³C NMR (150 MHz, CDCl₃) δ 7.1, 8.4, 23.8, 45.2, 46.3, 71.8, 74.9, 83.9, 88.9, 117.1, 118.1, 118.4, 119.2, 128.20, 128.22, 129.7, 129.8, 133.2, 138.3, 138.9, 149.6, 155.2, 159.0, 160.8, 173.3; HRMS (ESI–) calcd for C₂₆H₂₆N₇O₇S [M – Et₃NH]⁻ 580.1620, found 580.1589 (5.3 ppm error).

5′-N-[N-(2-Hydroxybenzoyl)sulfamoyl]amino-2′, 5′-dideoxy-2′-fluoroadenosine triethylammonium salt (8)

The title compound was prepared from 27 (0.62 g, 0.81 mmol) using the general procedure for salicylation and TFA deprotection. The crude product was purified by silica gel flash chromatography (10% MeOH–CH₂Cl₂ with 1% Et₃N) followed by preparative reverse-phase HPLC on a Phenomenex Gemini 10 µm C18 (250 × 21.2 mm) column at a flow rate of 20 mL/min with a gradient of 10–40% MeCN in 50 mM aqueous triethylammonium bicarbonate pH 7.5 over 20 min. The retention time of the product was 9.90 minutes (k′ = 2.3) and the appropriate fractions were pooled and lyophilized to afford the title compound (145 mg, 38% yield over 2 steps) as a white powder: ¹H NMR (600 MHz, CD₃OD) δ 1.29 (t, J = 7.3 Hz, 9H), 3.14 (q, J = 7.3 Hz, 6H), 3.37–3.39 (m, 1H), 3.44 (dd, J = 13.6, 3.2 Hz, 1H), 4.25–4.29 (m, 1H), 4.66 (dt, J = 14.6, 5.3, Hz, 1H), 5.52 (dt, J = 52.6, 4.0 Hz, 1H), 6.21 (dd, J = 16.1, 3.5, 1H), 6.80 (t, J = 7.5 Hz, 1H), 6.82 (d, J = 8.2 Hz, 1H), 7.25–7.33 (m, 1H), 7.91 (dd, J = 7.9, 1.7, Hz, 1H), 8.29 (s, 1H), 8.24 (s, 1H); ¹³C NMR (150 MHz, CD₃OD) δ 8.3, 44.3, 46.7, 70.1 (d, ²J_C-F = 15.9 Hz), 82.6, 87.1 (d, ²J_C-F = 33.0 Hz), 92.7 (d, ¹J_C-F = 190 Hz), 116.6, 118.1, 119.4, 129.7, 132.9, 139.9, 148.8, 152.9, 155.8, 160.2, 173.0 (missing 1 aryl C); HRMS (ESI–) calcd for C₁₇H₁₇FN₇O₆S [M – Et₃NH]⁻ 466.0951, found 466.0988 (7.9 ppm error).

3′, 5′-O-di-Acetyl-2′-deoxy-2′-fluoroguanosine (34)

To a stirring solution of 33 (4.00 g, 14 mmol, 1.0 equiv) in DMF (100 mL) at 23 °C were added sequentially pyridine (6.65 g, 84 mmol, 6.0 equiv) and acetic anhydride (8.6 g, 84 mmol, 6.0 equiv). The reaction mixture was stirred for 12 h at 23 °C and concentrated under reduced pressure to afford the title compound (4.1 g, 79%) as a white powder: ¹H NMR (400 MHz, DMSO-d₆) δ 2.03 (s, 3H), 2.14 (s, 3H), 4.22 (dd, J = 12.8, 6.4 Hz, 1H), 4.32–4.39 (m, 2H), 5.50–5.58 (m, 1H), 5.72 (ddd, J = 51.6, 5.3, 3.3 Hz, 1H), 6.11 (dd, J = 18.5, 3.2 Hz, 1H), 6.53 (s, 2H), 7.89 (s, 1H), 10.74 (s, 1H); ¹³C NMR (100 MHz, DMSO–d₆) δ 20.3, 20.5, 62.9, 70.26 (d, ²J_C-F = 14.5 Hz), 78.4, 85.5 (d, ²J_C-F = 33.4 Hz), 90.8 (d, ¹J_C-F = 191 Hz), 116.8, 135.6, 150.7, 153.8, 156.6, 169.4, 170.1; HRMS (ESI+) calcd for C₁₄H₁₆FN₅O₆Na [M + Na]⁺ 392.0977, found 392.0933 (4.9 ppm error).

3′, 5′-O-di-Acetyl-6-chloro-2′-deoxy-2′-fluoro-2-iodoadenosine (35)

To a stirring solution of 34 (1.8 g, 4.87 mmol. 1.0 equiv) and Et₄NCl (dried in a vacuum-oven over P₂O₅) (1.62 g, 9.74 mmol, 2.0 equiv) in acetonitrile (20 mL) at 23 °C were sequentially added freshly distilled POCl₃ (4.48 g, 29.2 mmol, 6.0 equiv) and freshly distilled dimethylaniline (0.59 g, 4.87 mmol, 1.0 equiv). The reaction mixture was heated for 20 min at 105 °C, cooled and poured into a beaker containing ice and stirred for 30 min. The organic layer was separated and the aqueous layer was extracted with EtOAc (2 × 50 mL) and the combined organic extracts were washed with H₂O, saturated aqueous NaHCO₃, saturated aqueous NaCl, dried (MgSO₄) and concentrated under reduced pressure. The crude product was recrystallized from isopropyl alcohol (100 mL) to afford 3′,5′-O-di-acetyl-6-chloro-2′-deoxy-2′-fluoroguanosine (1.77 g, 86%) as a white power that was used directly in the next step.

To a stirring solution of 3′,5′-O-di-acetyl-6-chloro-2′-deoxy-2′-fluoroguanosine prepared above (1.77 g, 4.56 mmol, 1.0 equiv) in THF (30 mL) at 23°C was sequentially added CH₂I₂ (12.2 g, 46 mmol, 10 equiv), I₂ (1.15 g, 4.6 mmol, 1.0 equiv), CuI (0.96 g, 5.0 mmol, 1.1 equiv) and isoamyl nitrite (1.6 g, 1.7 mmol, 3.0 equiv). The reaction mixture was then stirred at 80 °C for 1 h, cooled and concentrated under reduced pressure. Purification by flash chromatography (1:1 EtOAc–hexane) afforded the title compound (2.16 g, 95%) as a white powder: R_f = 0.45 (1:1 EtOAc–hexane); ¹H NMR (400 MHz, CD₃OD) δ 2.03 (s, 3H), 2.18 (s, 3H), 4.36 (dd, J = 12.4, 4.7 Hz, 1H), 4.45 (dd, J = 12.4, 3.1 Hz, 1H), 4.48–4.54 (m, 1H), 5.67 (ddd, J = 17.9, 7.8, 4.9 Hz, 1H), 5.81 (ddd, J = 51.4, 4.9, 2.0 Hz, 1H), 6.40 (dd, J = 18.7, 2.0 Hz, 1H), 8.58 (s, 1H); ¹³C NMR (100 MHz, CD₃OD) δ 20.4, 20.8, 63.6, 71.6 (d, ²J_C-F = 15.1 Hz), 80.7, 89.5 (d, ²J_C-F = 35.2 Hz), 92.6 (d, ¹J_C-F = 192 Hz), 117.8, 133.3, 146.9, 151.1, 153.1, 171.4, 172.2; HRMS (ESI+) calcd for C₁₄H₁₃ClFIN₄O₅Na [M + Na]⁺ 520.9495, found 520.9489 (1.2 ppm error).

2′-Deoxy-2′-fluoro-2-iodoadenosine (36)

To a solution of 35 (2.0 g, 4.02 mmol, 1.0 equiv) in MeOH (5 mL) in a high-pressure tube was added a 7 N ammonia solution in methanol (20 mL, 140 mmol, 35 equiv). The vessel was sealed and heated at 70 °C for 16 h. After cooling to rt, the reaction mixture was purged with N₂ for 1 h to remove ammonia and then concentrated under reduced pressure to afford the title compound (1.20 g, 56%) as a white powder contaminated with 2.3 equiv of acetamide (the calculated yield takes into account the acetamide): R_f = 0.18 (EtOAc); ¹H NMR (400 MHz, CD₃OD) δ 3.79 (dd, J = 12.5, 3.7 Hz, 1H), 3.94 (dd, J = 12.6, 2.4 Hz, 1H), 4.11–4.16 (m, 1H), 4.62 (ddd, J = 16.5, 6.4, 4.5 Hz, 1H), 5.39 (ddd, J = 52.7, 4.5, 2.9 Hz, 1H), 6.23 (dd, J = 16.3, 2.9 Hz, 1H), 8.26 (s, 1H); ¹³C NMR (100 MHz, CD₃OD) δ 62.7, 70.4 (d, ²J_C-F = 16.1 Hz), 85.9, 88.5 (d, ²J_C-F = 33.3 Hz), 94.9 (d, ¹J_C-F = 188 Hz), 120.5, 120.8, 141.2, 150.5, 157.2; HRMS (ESI+) calcd for C₁₀H₁₂FIN₅O₃ [M + H]⁺ 395.9963, found 395.9964 (0.3 ppm error).

3′, 5′-O-bis(tert-Butyldimethylsilyl)-2′-deoxy-2′-fluoro-2-iodoadenosine (37)

To a stirring solution of 36 (1.90 g, 4.8 mmol, 1.0 equiv) in DMF (20 mL) at 0 °C was sequentially added TBSCl (2.17 g, 14.4 mmol, 3.0 equiv), imidazole (1.64 g, 24 mmol, 5 equiv) and DMAP (59 mg, 0.48 mmol, 0.1 equiv). The reaction mixture was stirred at 23 °C for 24 h and concentrated under reduced pressure. Purification by flash chromatography (1:1 EtOAc–hexane) afforded the title compound (2.97 g, 97%) as a yellow oil: R_f = 0.5 (1:1 EtOAc–hexane); ¹H NMR (400 MHz, CD₃OD) δ –0.09 (s, 3H), 0.01 (s, 3H), 0.24 (s, 6H), 0.79 (s, 9H), 0.96 (s, 9H), 3.79 (dd, J = 11.7, 2.9 Hz, 1H), 3.94–4.05 (m, 2H), 4.97 (ddd, J = 20.8, 7.6, 4.3 Hz, 1H), 5.59 (ddd, J = 53.2, 4.4, 1.5 Hz, 1H), 6.15 (dd, J = 19.8, 1.5 Hz, 1H), 8.13 (s, 1H); ¹³C NMR (100 MHz, CD₃OD) δ –5.2, –5.1, –4.4, –4.0, 19.1, 19.3, 26.38, 26.47, 62.8, 70.7 (d, ²J_C-F = 15.9 Hz), 84.9, 89.2 (d, ²J_C-F = 34.6 Hz), 94.1 (d, ¹J_C-F = 188 Hz), 120.5, 121.0, 141.9, 150.5, 157.2; HRMS (ESI+) calcd for C₂₂H₃₉FIN₅O₃Si₂Na [M + Na]⁺ 646.1512, found 646.1540 (4.3 ppm error).

3′, 5′-O-tert-Butyldimethylsilyl-2′-deoxy-2′-fluoro-2-phenyladenosine (38)

To a stirring solution of 37 (1.1 g, 1.76 mmol, 1.0 equiv) in dioxane (10 mL) at 23 °C was sequentially added phenylboronic acid (0.65 g, 5.3 mmol, 3.0 equiv), K₃PO₄ (1.12 g, 5.3 mmol, 3.0 equiv), Pd(OAc)₂ (0.119 g, 0.53 mmol, 0.3 equiv) and CyJohnPhos (0.185 g, 0.53 mmol, 0.3 equiv). The reaction mixture was purged with argon for 15 min, sealed and heated at 100 °C for 12 h. The reaction was cooled to 23 °C, filtered through a pad of Celite and concentrated under reduced pressure. Purification by flash chromatography (1:1 EtOAc–hexane) afforded the title compound (0.68 g, 67%) as a yellow solid: R_f = 0.5 (1:1 EtOAc–hexane): ¹H NMR (400 MHz, CD₃OD) δ –0.09 (s, 3H), –0.02 (s, 3H), 0.17 (s, 6H), 0.77 (s, 9H), 0.95 (s, 9H), 3.83 (dd, J = 11.6, 2.9 Hz, 1H), 3.99–4.08 (m, 2H), 4.96 (ddd, J = 19.2, 7.2, 4.5 Hz, 1H), 5.71 (ddd, J = 53.4, 4.6, 1.9 Hz, 1H), 6.27 (dd, J = 19.4, 1.9 Hz, 1H), 7.39–7.44 (m, 3H), 8.24 (s, 1H), 8.30–8.36 (m, 2H); ¹³C NMR (100 MHz, CD₃OD) δ –5.24, –5.22, –4.5, –4.2, 19.0, 19.3, 26.29, 26.43, 62.4, 70.7 (d, ²J_C-F = 15.9 Hz), 85.1, 89.4 (d, ²J_C-F = 34.3 Hz), 93.9 (d, ¹J_C-F = 189 Hz), 119.4, 129.2, 129.5, 131.0, 139.8, 142.0, 151.5, 157.3, 161.3; HRMS (ESI+) calcd for C₂₈H₄₄FN₅O₃Si₂Na [M + Na]⁺ 596.2859, found 596.2853 (1.0 ppm error).

3′-O-tert-Butyldimethylsilyl-2′-deoxy-2′-fluoro-2-phenyladenosine (39)

To a stirring solution AcOH–THF–H₂O (9:1:1, 10 mL) at 0 °C was added 38 (0.674 g, 1.17 mmol, 1.0 equiv). The reaction mixture was warmed to 23 °C and stirred for 24 h and concentrated. Purification by flash chromatography (3:1 EtOAc–hexane) afforded the title compound (0.378 g, 70%) as a white solid. R_f = 0.28 (3:1 EtOAc–hexane): ¹H NMR (400 MHz, CD₃OD) δ 0.17 (s, 3H), 0.20 (s, 3H), 0.97 (s, 9H), 3.75 (dd, J = 12.4, 4.0, 1H), 3.94 (dd, J = 12.4, 2.7 Hz, 1H), 4.10–4.14 (m, 1H), 4.91–4.99 (m, 1H), 5.61 (ddd, J = 52.2, 4.8, 2.9, Hz, 1H), 6.32 (dd, J = 18.1, 2.9 Hz, 1H), 7.42–7.46 (m, 3H), 8.31–8.37 (m, 3H); ¹³C NMR (100 MHz, CD₃OD) δ –4.7, –4.5, 19.1, 26.3, 62.0, δ 71.8 (d, ²J_C-F = 15.5 Hz,), 86.2, 89.1 (d, ²J_C-F = 33.8 Hz), 94.1 (d, ¹J_C-F = 191 Hz), 119.6, 129.3, 129.5, 131.0, 139.9, 142.1, 151.6, 157.4, 161.4; HRMS (ESI+) calcd for C₂₂H₃₀FN₅O₃SiNa [M + Na]⁺ 482.1994, found 482.2123 (26 ppm error).

5′-Azido-3′-O-tert-butyldimethylsilyl-2′-deoxy-2′-fluoro-2-phenyladenosine (40)

To a stirring solution of 39 (0.268 g, 0.58 mmol, 1.0 equiv) in dioxane (5.0 mL) at 23 °C were added sequentially DPPA (0.376 mL, 1.74 mmol, 3.0 equiv) and DBU (0.175 mL, 1.16 mmol, 2.0 equiv). The reaction mixture was stirred for 12 h at 23 °C then NaN₃ (0.38 g, 5.8 mmol, 10 equiv) and 15-crown-5 (13 mg, 0.060 mmol, 0.1 equiv) were added and the reaction was heated at 90 °C for another 12 h. The reaction was then cooled and concentrated under reduced pressure. Purification by flash chromatography (3:1 EtOAc–hexane) afforded the title compound (0.240 g, 85%) as a white powder: R_f = 0.56 (3:1 EtOAc–hexane); ¹H NMR (400 MHz, CD₃OD) δ 0.16 (s, 3H), 0.22 (s, 3H), 0.95 (s, 9H), 3.41 (dd, J = 13.7, 3.9 Hz, 1H), 3.80 (dd, J = 13.8, 3.0, Hz, 1H), 4.15–4.21 (m, 1H), 5.33 (ddd, J = 19.1, 7.9, 5.0 Hz, 1H), 5.64 (ddd, J = 53.6, 5.0. 1.7 Hz, 1H), 6.27 (dd, J = 21.9, 1.7 Hz, 1H), 7.42–7.47 (m, 3H), 8.22 (s, 1H), 8.38–8.43 (m, 2H); ¹³C NMR (100 MHz, CD₃OD) δ –4.6, –4.4, 19.0, 26.2, 51.4, 71.9 (d, ²J_C-F = 16.1 Hz), 83.6, 89.8 (d, ²J_C-F = 35.7 Hz), 94.2 (d, ¹J_C-F = 189 Hz), 119.7, 129.3, 129.4, 131.0, 139.9, 142.5, 151.5, 157.4, 161.4; HRMS (ESI+) calcd for C₂₂H₂₉FN₈O₂SiNa [M + Na]⁺ 507.2059, found 507.2025 (6.7 ppm error).

2′, 5′-Dideoxy-2′-fluoro-5′-[N-(sulfamoyl)amino]-2-phenyladenosine (41)

To a round-bottomed flask containing 40 (0.230 g, 0.47 mmol, 1.0 equiv) purged with argon was added anhydrous MeOH (5.0 mL) followed by Pd/C (0.100 g, 10% by weight) and the flask evacuated by house vacuum and back-filled with a H₂ balloon 3 times before being allowed to stir vigorously under a H₂ balloon at 23 °C for 1 h. The crude reaction mixture was filtered over a pad of Celite and concentrated under reduced pressure to afford crude the amine, which was then dissolved in dioxane (6 mL). Sulfamide (0.136 g, 1.5 mmol, 3.0 equiv) was added and the reaction mixture was then heated at 90 °C for 3 h. The reaction was cooled, concentrated under reduced pressure and purified by flash chromatography (3:1 EtOAc–hexane) to afford the title compound (0.190 g, 87% over 2 steps) as a white powder: R_f = 0.3 (3:1 EtOAc–hexane); ¹H NMR (400 MHz, CD₃OD) δ 0.17 (s, 3H), 0.23 (s, 3H), 0.97 (s, 9H), 3.33–3.39 (m, 1H), 3.53 (dd, J = 13.6, 3.8 Hz, 1H), 4.18–4.24 (m, 1H), 5.04 (ddd, J = 15.9, 6.8, 4.9 Hz, 1H), 5.66 (ddd, J = 53.2, 5.0, 2.7 Hz, 1H), 6.28 (dd, J = 19.6, 2.7 Hz, 1H), 7.41–7.46 (m, 3H), 8.24 (s, 1H), 8.32– 8.38 (m, 2H); ¹³C NMR (100 MHz, CD₃OD) δ –4.5, –4.4, 19.1, 26.3, 45.5, 72.7 (d, ²J_C-F = 15.8 Hz), 83.8, 89.2 (d, ²J_C-F = 34.5 Hz), 93.8 (d, ¹J_C-F = 191 Hz), 119.6, 129.4, 129.5, 131.0, 139.8, 142.4, 151.6, 157.4, 161.5; HRMS (ESI+) calcd for C₂₂H₃₂FN₇O₄SSiNa [M + Na]⁺ 560.1882, found 560.1890 (1.4 ppm error).

2′, 5′-Dideoxy-2′-fluoro-5′-N-[N-(2-hydroxybenzoyl)sulfamoyl]amino-2-phenyladenosine triethylammonium salt (9)

The title compound was prepared from 41 (0.176 g, 0.33 mmol, 1.0 equiv) using the general procedure for salicylation and TFA deprotection. The crude product was purified by silica gel flash chromatography (10% MeOH–CH₂Cl₂ with 1% Et₃N) followed by preparative reverse-phase HPLC on a Phenomenex Gemini 10 µm C18 (250 × 21.2 mm) column at a flow rate of 25 mL/min with 30% MeCN in 50 mM aqueous triethylammonium bicarbonate pH 7.5 over 20 min. The retention time of the product was 12.0 minutes (k′ = 3.1) and the appropriate fractions were pooled and lyophilized to afford the title compound (95 mg, 45% over 2 steps) as a white powder: ¹H NMR (400 MHz, CD₃OD) δ 1.21 (t, J = 7.3 Hz, 14H, excess Et₃N), 2.98 (q, J = 7.3 Hz, 10H, excess Et₃N), 3.35–3.30 (ovlp m, 1H), 3.47 (dd, J = 13.5, 3.5 Hz, 1H), 4.19–4.25 (m, 1H), 4.80–4.86 (ovlp m, 1H), 5.57 (ddd, J = 53.2, 4.9, 2.2 Hz, 1H), 6.28 (dd, J = 19.1, 2.1 Hz, 1H), 6.71–6.78 (m, 2H), 7.22–7.27 (m, 1H), 7.37–7.45 (m, 3H), 7.83 (dd, J = 7.8, 1.7 Hz, 1H), 8.27 (s, 1H), 8.31–8.37 (m, 2H); ¹³C NMR (100 MHz, CD₃OD) δ 9.9, 46.0, 47.7, 71.8 (d, ²J_C-F = 16.2 Hz), 82.9, 88.6 (d, ²J_C-F = 34.3 Hz), 95.0 (d, ¹J_C-F = 187 Hz), 117.9, 119.2, 121.0, 129.32, 129.37, 129.39, 130.94, 130.97, 134.0, 139.7, 141.7, 151.6, 157.3, 161.3, 161.9, 174.6; HRMS (ESI+) calcd for C₂₃H₂₂FN₇O₆SNa [M – Et₃N + Na]⁺ 566.1229, found 566.1250 (3.7 ppm error).

5′-O-tert-Butyldimethylsilyl-2′-deoxy-2′-fluoroadenosine (43).⁵⁰

To a stirring solution of 42 (0.750 g, 2.8 mmol, 1.0 equiv) in DMF (25 mL) at 0 °C was added sequentially TBSCl (0.462 g, 4.5 mmol, 1.6 equiv) and imidazole (0.322 g, 4.7 mmol, 1.7 equiv). The reaction mixture was stirred for 7 h at 23 °C, then quenched with MeOH (20 mL) and concentrated. Purification by flash chromatography (4:1 EtOAc–hexane) afforded the title compound (0.79 g, 90%) as a white solid. The analytical data (¹H NMR, ¹⁹F NMR, MS) agreed with previously reported values.^{50 13}C NMR data and R_f have not been disclosed and are provided herein: R_f = 0.22 (EtOAc); ¹³C NMR (100 MHz, CD₃OD) δ –5.25, –5.22, 19.4, 26.5, 62.5, 69.4 (d, ²J_C-F = 16.6 Hz), 84.5, 88.2 (d, ²J_C-F = 33.6 Hz), 95.3 (d, ¹J_C-F = 187 Hz), 120.4, 140.8, 150.0, 154.1, 157.3.

5′-O-tert-Butyldimethylsilyl-2′-deoxy-2′-fluoro-3′-O-(phenoxythiocarbonyl)adenosine (44).⁵⁰

To a stirring solution of 43 (0.800 g, 2.1 mmol, 1.0 equiv) in MeCN (25 mL) was added sequentially O-phenyl chlorothionoformate (0.540 g, 3.1 mmol, 1.5 equiv) and DMAP (0.765 g, 6.3 mmol, 3.0 equiv). The reaction mixture was stirred for 1 h at 23 °C then quenched with H₂O (10 mL). The aqueous layer was extracted with EtOAc (2 × 20 mL) and the combined organic layers were washed with saturated aqueous NaCl, dried (MgSO₄), filtered and concentrated under reduced pressure. Purification by flash chromatography (4:1 EtOAc–hexane) afforded the title compound (0.79 g, 82%) as an orange solid. The analytical data (¹H NMR, ¹⁹F NMR, MS) agreed with previously reported values.^{50 13}C NMR data and R_f have not been disclosed and are provided herein: R_f = 0.38 (3:1 EtOAc–hexane); ¹³C NMR (100 MHz, CDCl₃) δ –5.5, –5.4, 18.4, 25.9, 61.8, 77.6 (d, ²J_C-F = 14.0 Hz), 81.6, 86.3 (d, ²J_C-F = 32.1 Hz), 90.7 (d, ¹J_C-F = 196.8 Hz), 119.9, 121.6, 126.9, 129.7, 138.5, 149.6, 153.38, 153.42, 155.6, 193.9.

2′, 3′-Dideoxy-2′-fluoroadenosine (46).⁵⁰

To a stirring solution of 44 (0.410 g, 0.80 mmol) in toluene (10 mL) was added sequentially AIBN (0.026 g, 0.16 mmol, 0.2 equiv) and Bu₃SnH (0.425 mL, 1.6 mmol, 2.0 equiv). The reaction mixture was degassed with Ar for 30 min then heated at 85 °C for 3 h. Upon completion, the reaction mixture was cooled, concentrated and filtered through a plug of silica gel, washing with EtOAc, to obtain 45 (0.235 g, 79% crude yield) as a yellow solid that was used directly in the next step.

To a stirring solution of 45 (0.210 g, 0.57 mmol, 1.0 equiv) in MeOH (5 mL) at 0 °C was added TFA (5 mL). The reaction mixture was stirred for 12 h at 0 °C and concentrated under reduced pressure. Purification by flash chromatography (1:19 MeOH–EtOAc) afforded the title compound (0.132 g, 92%) as a white powder: R_f = 0.18 (1:19 MeOH–EtOAc); ¹H NMR (400 MHz, DMSO-d₆) δ 2.23 (ddd, J = 20.6, 14.5, 5.5 Hz, 1H), 2.35–2.55 (m, 1H), 3.55 (ddd, J = 12.1, 5.5, 3.9 Hz, 1H), 3.74 (ddd, J = 12.1, 5.3, 3.0 Hz, 1H), 4.28–4.50 (m, 1H), 5.14 (t, J = 5.5 Hz, 1H), 5.59 (dd, J = 51.9, 4.3, Hz, 1H), 6.23 (d, J = 18.3 Hz, 1H), 7.30 (s, 2H), 8.14 (s, 1H), 8.34 (s, 1H); ¹³C NMR (100 MHz, DMSO-d₆) δ 32.3 (d, ²J_C-F = 20.3 Hz), 62.0, 81.8, 88.9 (d, ²J_C-F = 35.8 Hz), 97.3 (d, ¹J_C-F = 177.4 Hz), 119.4, 139.2, 149.1, 153.1, 156.5; HRMS (ESI+) calcd for C₁₀H₁₂FN₅O₂Na [M + Na]⁺ 276.0867, found 276.0860 (2.5 ppm error).

5′-Azido-2′, 3′, 5′-trideoxy-2′-fluoroadenosine (47)

To a stirring solution of 46 (0.20 g, 0.87 mmol, 1.0 equiv) in dioxane (5.0 mL) at 23 °C was added sequentially DPPA (0.56 mL, 2.6 mmol, 3.0 equiv) and DBU (0.39 mL, 2.6 mmol, 3.0 equiv). The reaction mixture was stirred for 12 h at 23 °C then NaN₃ (0.55 g, 8.6 mmol, 10 equiv) and 15-crown-5 (0.035 mL, 0.17 mmol, 0.2 equiv) were added and the reaction was heated at 90 °C for another 12 h. The reaction was then cooled and concentrated under reduced pressure. Purification by flash chromatography (EtOAc) afforded the title compound (0.145 g, 66%) as a white powder: R_f = 0.33 (EtOAc); ¹H NMR (400 MHz, CD₃OD) δ 2.38 (ddd, J = 20.4, 14.5, 5.4 Hz, 1H), 2.67 (dddd, J = 41.1, 15.0, 10.6, 4.7 Hz, 1H), 3.53 (dd, J = 13.4, 4.8, Hz, 1H), 3.72 (dd, J = 13.4, 3.3 Hz, 1H), 4.56–4.66 (m, 1H), 5.71 (dd, J = 51.6, 4.7, Hz, 1H), 6.31 (d, J = 19.3 Hz, 1H), 8.21 (s, 1H), 8.26 (s, 1H); ¹³C NMR (100 MHz, CD₃OD) 35.1 (d, ²J_C-F = 21.0 Hz), 54.1, 81.2, 91.2 (d, ²J_C-F = 37.3 Hz, H), 97.9 (d, ¹J_C-F = 180 Hz), 120.5, 140.9, 150.3, 154.1, 157.5; HRMS (ESI+) calcd for C₁₀H₁₁FN₈ONa [M + Na]⁺ 301.0932, found 301.0951 (6.3 ppm error).

2′, 3′, 5′-Trideoxy-2′-fluoro-5′-[N-(sulfamoyl)amino]adenosine (48)

To a round-bottomed flask containing 47 (0.145 g, 0.5 mmol, 1.0 equiv) purged with argon was added MeOH (4.0 mL) followed by Pd/C (0.055 g, 10% by weight) and the flask evacuated by house vacuum and backfilled with a H₂ balloon 3 times before being allowed to stir vigorously under a H₂ balloon at 23 °C for 1 h. The crude reaction mixture was filtered over a pad of Celite and concentrated under reduced pressure to afford crude amine, which was then dissolved in dioxane (4 mL). Sulfamide (0.145 g, 1.5 mmol, 3.0 equiv) was added and the reaction mixture was then heated at 90 °C for 3 h. The reaction was cooled, concentrated under reduced pressure and purified by flash chromatography (EtOAc) to afford the title compound (0.114 g, 60% over 2 steps) as a white powder: R_f = 0.21 (EtOAc); ¹H NMR (400 MHz, CD₃OD) δ 2.38 (dddd, J = 22.0, 14.5, 5.9, 1.8 Hz, 1H), 2.65 (dddd, J = 37.0, 14.9, 9.9, 5.3 Hz, 1H), 3.32 (ovlp dd, J = 13.8, 3.7 Hz, 1H), 3.45 (dd, J = 13.8, 3.7 Hz, 1H), 4.58–4.66 (m, 1H), 5.59 (dd, J = 52.3, 3.8, Hz, 1H), 6.23 (d, J = 18.8 Hz, 1H), 8.21 (s, 1H), 8.23 (s, 1H); ¹³C NMR (100 MHz, CD₃OD) δ 33.6 (d, ²J_C-F = 20.8 Hz), 45.2, 79.3, 90.2 (d, ²J_C-F = 36.7 Hz, H), 96.2 (d, ¹J_C-F = 181 Hz), 119.2, 139.8, 148.5, 152.6, 155.9; HRMS (ESI+) calcd for C₁₀H₁₄FN₇O₃SNa [M + Na]⁺ 354.0755, found 354.0772 (4.8 ppm error).

2′, 3′, 5′-Trideoxy-5′-N-[N-(2-hydroxybenzoyl)sulfamoyl]amino-2′-fluoroadenosine triethylammonium salt (10)

The title compound was prepared from 48 (0.075 g, 0.23 mmol, 1.0 equiv) using the general procedure for salicylation and TFA deprotection. The crude product was purified by silica gel flash chromatography (1:9 MeOH–CH₂Cl₂ with 1% Et₃N) followed by preparative reverse-phase HPLC on a Phenomenex Gemini 10 µm C18 (250 × 21.2 mm) column at a flow rate of 25 mL/min with 25% MeCN in 50 mM aqueous triethylammonium bicarbonate pH 7.5 over 20 min. The retention time of the product was 13.3 minutes (k′ = 3.5) and the appropriate fractions were pooled and lyophilized to afford the title compound (0.044 g, 34% yield over 2 steps) as a white powder: ¹H NMR (400 MHz, CD₃OD) δ 1.22 (t, J = 7.3 Hz, 9H), 2.30–2.47 (m, 1H), 2.59 (dddd, J = 39.0, 14.9, 10.2, 5.0 Hz, 1H), 2.99 (q, J = 7.2 Hz, 6H), 3.26 (dd, J = 13.6, 5.5 Hz, 1H), 3.39 (dd, J = 13.5, 3.8 Hz, 1H), 4.57–4.65 (m, 1H), 5.60 (dd, J = 52.1, 4.8 Hz, 1H), 6.24 (d, J = 18.3 Hz, 1H), 6.74–6.80 (m, 2H), 7.24–7.29 (m, 1H), 7.87 (dd, J = 7.8, 1.8 Hz, 1H), 8.21 (s, 1H), 8.31 (s, 1H); ¹³C NMR (100 MHz, CD₃OD) δ 9.4, 35.3 (d, ²J_C-F = 20.6 Hz) 47.4, 48.0, 81.1, 91.3 (d, ²J_C-F = 36.7 Hz), 97.9 (d, ¹J_C-F = 180 Hz), 117.9, 119.2, 121.0, 131.0, 134.1, 141.0, 150.2, 154.1, 157.4, 162.0, 174.4 (missing 1 aryl C); HRMS (ESI–) calcd for C₁₇H₁₇FN₇O₅S [M – Et₃NH]⁻ 450.1001, found 450.1007 (1.3 ppm error).

MbtA Enzyme Assay

MbtA was expressed in E. coli and purified as described.³² MbtA concentration was determined by active site titration with 1 as described.³² The inhibition assays were performed as reported in duplicate under initial velocity conditions.³² In brief, the reaction was initiated by adding 10 µL [³²P]PP_i with 7 nM MbtA in 90 µL reaction buffer (278 µM salicylic acid, 11.1 mM ATP, 1.11 mM PPi, 83.3 mM Tris-HCl, pH 7.5, 11.1 mM MgCl₂, 2.22 mM DTT) at 37 °C in the presence of eight different concentrations of the inhibitor (1.5-fold dilution from 100 nM down to 5.85 nM). The reaction was terminated at 20 min by the addition of 200 µL of quenching buffer (350 mM HClO₄, 100 mM PPi, 1.8% w/v activated charcoal). The charcoal was pelleted by centrifugation and washed once with 500 µL H₂O and analyzed by liquid scintillation counting as described. Fractional initial velocities were fit by nonlinear regression analysis to the Morrison equation⁶⁰ (eq 1) using GraphPad Prism 5.0:

$$\frac{v_i}{v_0}=1-\frac{([E]+[I]+K_i^{\mathit{\text{app}}})-\sqrt{{([E]+[I]+K_i^{\mathit{\text{app}}})}^2-4[E][I]}}{2[E]}$$ (1)

where I represents the concentration of inhibitor, $K_i^{\mathit{\text{app}}}$ is the apparent inhibition constant, E is the active enzyme concentration (determined by active-site titration), v_i is the initial rate at each [I], and v₀ is the initial rate of the DMSO control.

M. tuberculosis H37Rv MIC Assay

All compounds minimum inhibitory concentrations (MICs) were experimentally determined as previously described.²⁸ MICs were determined in triplicate in iron-deficient GAST according to the broth microdilution method using compounds from DMSO stock solutions or with control wells treated with an equivalent amount of DMSO. Isoniazid was used as a positive control while DMSO was employed as a negative control. All measurements reported herein used an initial cell density of 10⁴–10⁵ cells/assay and growth monitored at 10–14 days, with the untreated and DMSO-treated control cultures reaching an OD₆₂₀ 0.2–0.3. Plates were incubated at 37 °C (100 µL/well) and growth was recorded by measurement of optical density at 620 nm.

Procedures for the measurement of the oral bioavailability in Sprague-Dawley rats

Sample Collection

Sprague-Dawley rats (female, 250–274 g, Harlan Laboratories, Inc. Indianapolis, IN) were obtained with surgically implanted in-dwelling dual jugular vein/femoral vein catheters for compound infusion and blood withdrawal, respectively. Catheters were maintained by following the Harlan Surgically Modified Animal Models: Catheter Maintenance Recommendations provided by Harlan Laboratories. All animal care, housing, and laboratory procedures were approved by the University of Minnesota Institutional Animal Care and Use Committee (IACUC) prior to these investigations (IACUC #1208A18543). The animal care and housing was maintained by University of Minnesota Research Animal Resources (RAR) and veterinarian staff. Rats were allowed to acclimate for 7 d prior to participation in a study to ensure proper patency of the catheter lines as well as to observe that there were no adverse effects on the animals due to surgery or shipping. Oral doses of investigational compounds (25 mg/kg) were suspended in either 0.5% sodium carboxymethylcelluose (Aldrich, St. Louis, MO; for compounds 1, 3–6, 8–10) or dissolved in 25% wt/vol hydroxypropyl-β-cyclodextrin (Cargill, Inc. Cedar Rapids, IA; for compounds 2, 3) or in 50% PEG and 5% tween 80 (for compound 7) in sterile water as solubility allowed. Animals (3 per compound) were initially dosed via oral gavage with a ball-tipped feeding needle inserted through the mouth and esophagus with a 10 mg/mL solution of compound. The rats were restrained by holding them with 2 fingers over their shoulders with the thumb wrapped around underneath their shoulders to tilt their head up, and solution dispensed over 15–20 seconds. Rats were returned to their cages after drug administration with access to food and water ad libitum, and blood was removed in 0.25 mL aliquots from the left femoral vein catheter via tuberculin syringe with a sterile 22 gauge blunted needle (SAI Infusions, Inc. Lake Villa, IL) at the following time points: 0, 15, 30, 45, 60, 120, 240, and 480 min. Blood was immediately injected into a 2 mL vacutainer (Becton Dickinson, Franklin Lakes, NJ) containing 3.6 mg K₂EDTA. Subsequently, blood was centrifuged at 2,000 × g to separate the plasma from the hematocrit. Plasma was separated into two 50 µL aliquots that were immediately chilled on ice and transferred to −80 °C within 2 h.

Rats were given a 3-day period to wash any remaining compound from their system. Intravenous doses of investigational compounds (2.5 mg/kg) were prepared in either 0.05% sodium carboxymethylcellulose (Aldrich, St. Louis, MO; for compounds 1, 3–6, 8–10) or 25% wt/vol hydroxypropyl-β-cyclodextrin (Cargill, Inc. Cedar Rapids, IA; for compoounds 2, 3, 7) in sterile water as solubility allowed. The same triplicate set was then dosed via infusion in the right jugular vein catheter over 5 sec via a tuberculin syringe with a sterile 22 gauge blunted needle. Rats were returned to their cages after drug administration with access to food and water ad libitum, and 0.25 mL aliquots of blood were removed from the left femoral vein catheter via tuberculin syringe with a sterile 22 gauge blunted needle at the following time points: 0, 5, 10, 15, 30, 45, 60, 120, 240, and 480 min. The blood was processed in an identical manner as described for the oral dosing above.

Plasma Sample Preparation

For compounds 1–6, sample plasma (20 µL) was protein-precipitated with cold acetonitrile (MeCN, 200 µL) containing N⁶-cyclopropyl-2-phenyl-5′-O-[N-(salicyl)sulfamoyl]adenosine (CCA-086)³⁵ as the internal standard (IS) at a final concentration of 100 nM in MeCN. For compounds 7–10, sample plasma (20 µL) was protein precipitated with ice-cold 10% aqueous trichloroacetic acid solution (200 µL) containing the indicated internal standard at 100 nM (7: IS=1; 8: IS=CCA-086; 9: IS=8; 10: IS=8). For INH analysis, nicotinamide (500 nM) was used as the internal standard. The high concentrations for the IV samples of compounds 8–10 required dilution to fit within the standard curve. This was done using rat plasma in K₂EDTA (BioChemed, Winchester, VA) from pooled female Sprague-Dawley rats. Samples from 5, 10, 15, and 30 min time points were diluted 10-fold, while the 60, and 120 min time point samples were diluted 5-fold. In the final analysis the concentrations were adjusted accordingly. After adding either the MeCN or 10% TCA solution containing internal standard, samples were vortexed for 5 sec, centrifuged at 14,000 × g for 1 min at 4 °C, and the supernatant was analyzed via LC–MS/MS. Standards were prepared in pooled female Sprague-Dawley rat plasma in K₂EDTA (BioChemed, Winchester, VA) and processed in a similar fashion as animal plasma samples.

Instrumentation

LC-MS/MS analysis was performed using an Agilent (Santa Clara, CA) 1260 Infinity gradient solvent delivery system, an Agilent 1260 Infinity HiP ALS autosampler, an Agilent 1260 column oven; or by a Shimadzu UFLC XR, SIL-20AC autosampler, and prominence CTO-20A column oven with an AB SCIEX QTRAP 5500 (Framingham, MA) mass spectrometer fitted with an electrospray ionization source. The instruments were run using Analyst Software (version 1.5.2, AB SCIEX).

Chromatography

Reverse-phase HPLC analysis was performed by gradient HPLC on a Kinetix C18 column (50 mm × 2.1 mm, 2.6 µm particle size; Phenomenex, Torrance, CA) for 1– 10 and a Synergy Hydro RP (250 mm × 2.0 mm, 4 µm particle size; Phenomenex, Torrance, CA) for isoniazid. Mobile phase A was 0.1% aqueous formic acid while mobile phase B was 0.1% formic acid in acetonitrile. Initial conditions were 5% B from 0 to 0.5 min, after which the %B was increased to 95% from 0.5 to 1 min. The column was washed in 95% B for 3 min, returned to 5% over 0.2 min, and allowed to re-equilibrate for 3.8 min in 5% B to provide a total run time of 8 min. The flow rate was 0.5 mL/min and the column oven was maintained at 40 °C. The injection volume was 10 µL.

Mass spectrometry

All analytes were analyzed in positive ionization mode by Multiple Reaction Monitoring (MRM) with the following mass spectrometry settings: entrance potential = 10 V; curtain gas = 40 psi; collision gas = medium; ionspray voltage = 5500 V; temperature = 600 °C; ion source gas 1 = 45 psi; ion source gas 2 = 40 psi. Nitrogen was used for the nebuliser and collision gas. To determine the optimum MRM settings (Table 4), each analyte was infused at a concentration of 10 µM (in 1:1 water:acetonitrile containing 0.1% formic acid) onto the MS by a syringe pump at a flow of 10 µL/min.

Table 4.

Mass spectrometry ionization parameters.

Compound	Parent Ion (m/z)	Fragment Ion (m/z)	Declustering Potential (V)	Collision Energy (V)	Collision Cell Exit Potential (V)
INH	138.0	177.2	40	19	24
1	467.1	347.2	130	25	16
2	551.0	352.2	140	23	19
3	593.0	352.1	80	26	20
4	466.1	267.1	60	23	14
5	506.1	369.1	79	30	14
6	481.0	136.1	140	51	18
7	582.3	383.0	140	30	13
8	468.0	269.2	140	24	13
9	544.2	345.2	52	24	20
10	452.0	315.0	140	23	18

Data Analysis

Analyte and internal standard peak areas were calculated with MultiQuant Software (version 2.0.2, AB SCIEX, Framingham, MA). Analyte peak areas were normalized to internal standard peak areas, and the concentrations of the compounds as a triethylamine salt were calculated with the appropriate standard curve. Standard curves displayed good linearity with R² = 0.99. Pharmacokinetic parameters were calculated from concentration-time profiles by non-compartmental analysis using Phoenix WinNonLin (version 6.3, Pharsight). Linear trapezoidal linear interpolation was used to fit the data points and all the pharmacokinetic parameters were determined using the software. Bioavailability was calculated individually for each rat from the oral and intravenous data.

Caco-2 Studies

Studies to detemine permeability in Caco-2 monolayers followed the protocol described by Hubatsch and co-workers.⁶¹

Cultivation of Caco-2 cells on Transwell permeable supports

Caco-2 cells (ATCC® HTB-37™) of passage number (50–70) were used for studies of all compounds. Caco-2 cells from the stock were thawed and cultivated to the required passage number and cells were seeded on the filter support. Cells were grown using Dulbecco’s modified Eagle’s medium (DMEM) media containing 17% fetal calf serum and 1% PEST (penicillin 10,000 U mL⁻¹, streptomycin 10,000 µg mL⁻¹). Cells of required passage number were then transferred to the 12-well, 12 mm insert, 0.4 µm pore size Transwell membrane plate (Corning No. 3401) and were grown in the plates for 21 days (37 °C, CO₂ incubator, 5% CO₂, water saturated atmosphere). The media was changed every two days to support the required amount of growth.

Experimental method

The media in the plates was aspirated and the apical side washed twice with cell assay buffer (CAB) pH 6.0 (NaCl 3.56 g, NaHCO₃ 1.05 g, glucose 0.9 g, MES 0.975 g, KCl 0.112 g, MgSO₄ 0.147 g, CaCl₂ 0.10 g, K₂HPO₄ 0.035 g in 500 mL of water). The basolateral side was washed with CAB pH 7.4 (Same composition as CAB 6.0, except HEPES 1.19 g for MES). The apical side was filled with 0.5 mL of CAB pH 6.0 and the basolateral side was filled with 1.5 mL of CAB pH 7.4. Using an EVOM epithelial voltohmmeters, TEER values across the differentiated Caco-2 cells were measured in triplicate. Only wells having a TEER value of >300 were used for the studies. The plate was preincubated in an incubator shaker for 60 min at 65 rpm along with the CAB. Donor solution (2.5 mL) for the apical side was made with the compound of interest (12.5 µL, 20 mM DMSO stock), luciferrin yellow (12.5 µL, 20 mM in water) and CAB pH 6 (2.475 mL). Donor solution (6 mL) for the basolateral side was made with the compound of interest (30 µL, 20 mM in DMSO), luciferin yellow (30 µL, 20 mM in water) and CAB pH 7.4 (5.970 mL). Luciferrin yellow was added as an internal standard in the donor solutions. The CAB solutions were aspirated and the experiment was initiated by filling the apical side with 0.5 mL of donor solution and the basolateral side with 1.5 mL of donor solution. 100 µL of samples were taken out at 0, 30, 60 and 120 minutes in triplicates from both the apical and basolateral sides and the volume was made up with the appropriate CAB.

Instrumentation and Data analysis

The samples were analyzed using Shimadzu LC-10AD HPLC using a Phenomenex Gemini 5 µm C18 (250 × 4.6 mm) column at a flow rate of 1 mL/ min. Mobile phase A was 20 mM TEAB buffer (pH 7.5) and mobile phase B was acetonitrile. P_app A-B and P_app B-A, and efflux ratio were calculated based on the cumulative sampling. The P_app values were calculated using equation (2):

$$P_{\text{app}}=(\frac{\mathit{\text{dQ}}}{\mathit{\text{dt}}})(1/(\mathit{\text{ACo}})$$ (2)

where $(\frac{\mathit{\text{dQ}}}{\mathit{\text{dt}}})$ is the steady state flux, A is the surface area of the filter (cm²) and C₀ is the initial concentration in the donor chamber.

Microsomal Stability Studies

Rat, mouse, and pooled human liver microsomes (BD Biosciences, San Jose, CA) (0.5 mg/mL) were incubated in 100 mM potassium phosphate buffer, pH 7.4 with 10 mM MgCl₂ and 1.0 µM 1 in a volume of 450 µL. This solution was preincubated for 5 min at 4 °C followed by 2 min at 37 °C. Water (50 µL, control reaction) or 10 mM NADPH (50 µL, microsomal reaction) were added and the two solutions were incubated for 60 min at 37 °C. After 60 min, 100 µL from each reaction was added to 200 µL acetonitrile; the resulting suspension was vortexed, centrifuged (14,000g, 5 min) to remove precipitated protein, and the supernatant was analyzed via LC-MS/MS as described above.

Glucuronidation Studies

1 was incubated with human and rat liver microsomes in the presence of uridine 5’-diphosphoglucuronic acid (UDPGA) under standard glucuronidation condition. Mycophenolic Acid (MPA), which is known to be rapidly metabolized by glucuronidation, was used as a positive control. Under these conditions, MPA was 85% glucuronidated by human liver microsomes at 60 min, but no glucuronide species of 1 was detected. In rat liver microsomes, glucuronidation was slower even for MPA and only 30% was glucuronidated at 60 min. Once again, glucuronidation of 1 was not detected. This confirms 1 is not a substrate for UDP-gluguronosyl transferase (UDPGT) in both human and rat liver microsomes.

Reactions were set up in duplicate in 200 µl volumes (final concentration in parenthesis). A mixture of 5 µL 20 mg protein/mL human liver microsomes (Sigma Aldrich) (0.5 mg/mL), 80 µL 0.25 M Trizma buffer pH 7.4 (0.1 M), 73 µL H₂O, 10 µL 100 mM MgCl₂ (5 mM), and 10 µL 0.25 mg/ml alamethicin (50 µg/mg) was prepared in each vial and incubated on ice for 30 min. 10 Ll 100 mM Saccharolactone (5 mM), and 2 µL of 60 mM MPA (600 µM) or 2 µL of 10 mM 1 (100 µM) were added to corresponding vial. The vials were pre-incubated in a 37 °C water bath for 3 min and 10 µL 60 mM UDPGA (3 mM) was added. For thermal breakdown control (lacking UDPGA), 10 µL H₂O was added. The vials were incubated for 60 min in a 37 °C water bath. The reactions were quenched at 0, 5, 15, 30, and 60 minutes by adding 200 µL crash solution (for MPA incubations: ice cold methanol with 10 µM IS; for 1 incubations: 10% TCA with 10 µM IS) to each vial. Compound 9 was used as the internal standard (IS), which has significantly different retention time compared to 1. The vials were mixed by vortex, kept on ice for 5 min and centrifuged at 13,000 × g for 5 min. Supernatants were analyzed on LC-MS by monitoring the disappearance of 1 or MPA and the corresponding appearance of the glucuronide products at MW + 176. LC-MS was performed on an Agilent 1100 series LC-MSD trap instrument. Reverse phase LC was performed on a Zorbax SB-C18 (0.5 × 150 mm, 5 micron) column with a flow rate of 15 µL/min. The eluents and elution were as follows. Mobile phase A: 0.1% aqueous formic acid, mobile phase B: 0.1% formic acid in acetonitrile. Initial conditions were 5% B from 0 to 0.5 min, after which the %B was increased to 30% from 0.5 to 10 min, after which 30% to 95% B from 10 to 13 min. The column was washed with 95% B for 2 min, returned to 5% over 1 min, and allowed to re-equilibrate for 4 min in 5% B to provide a total run time of 20 min. The flow rate was 0.5 mL/min and the column oven was maintained at 40 °C. The injection volume was 5 µL. The MS was performed on positive ion mode using ESI.

Cell Cytotoxicity Assay

Human liver cells (HepG2, ATCC) cells were maintained in Eagle’s minimum essential medium (EMEM) supplemented with 10% fetal bovine serum (FBS), 100 U/mL penicillin, and 100 µg/mL streptomycin. All reactions were performed in triplicate. Compounds (1, 2, 4, 5, 8, and 9) were prepared as 20 mM stock solutions in DMSO. For the highest test concentration, 2 µL of the stock solution, 98 µL of EMEM and 100 µL of HepG2 cells in EMEM ((2.5–5.0) × 10⁴ cells per well) were plated in 96-well plates, yielding a final volume of 200 µL and compound concentration of 200 µM. A serial dilution to provide final compound concentration of 100, 50, 20, 10, 5, 2, 1 µM were also prepared. Control wells contained either 1% DMSO (negative control) or 1% triton X-100 (positive control). The plate was incubated for 48 h at 37 °C in a 5% CO₂/95% air humidified atmosphere. The EMEM media was aspirated and 100 µL serum-free, phenol red free DMEM was added to each well. Measurement of cell viability was carried out using a modified method of Mosmann based on 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT). The MTT solution was prepared fresh at 5 mg/mL in DI water. MTT solution (10 µL) was added to each well, and the plate was incubated as described above for 3 h. The MTT solution was removed, and the formazan crystals were solubilized with 200 µL of 2 N KOH. The plate was read on a M5e spectrophotometer (Molecular Devices) at 570 nm for formazan and 650 nm for background subtraction. Cell viability was estimated as the percentage absorbance of sample relative to the DMSO control. Percent cell viability was plotted against test concentration to determine CC₅₀.

Figure 2.

Mean plasma concentration vs. time curves of single dosage of compound 10 in 3 rats by both p.o. and i.v. administration. Error bars represent the standard deviation of the mean (n=3).

Scheme 4. a. Synthesis of 10.

^aReaction conditions: (a) TBSCl, imidazole, DMF, 23 °C, 7 h, 90%; (b) O-phenyl chlorothionoformate, DMAP, MeCN, 23 °C, 1 h, 82%; (c) Bu₃SnH, AIBN, toluene, 85 °C, 3 h,79%; (d) TFA, MeOH, 0 °C, 12 h, 92%; (e) i) DPPA, DBU, dioxane, 23 °C, 12 h; ii) NaN₃, 15-crown-5, dioxane, 90 °C, 12 h, 66%; (f) Pd/C, MeOH, 23 °C, 1 h; (g) NH₂SO₂NH₂, dioxane, 90 °C, 3 h, 60% (2 steps); (h) 31, Cs₂CO₃, DMF, 0 °C to rt, 16 h; (i) 80% aq TFA, 0 °C, 1 h, 34% (2 steps from 48).

ABBREVIATIONS USED

ADME

absorption distribution metabolism excretion

appK_i

apparent inhibition constant

AUC

area under the curve

Boc

tert-butoxycarbonyl

C_max

maximum serum concentration following an oral dose

Cbz

benzyloxycarbonyl

CDI

carbonyl diimidazole

CL

clearance

DMA

dimethylacetamide

DMF

dimethylformamide

F

bioavailability

HPLC

high performance liquid chromatography

INH

isoniazid

i.v.

intravenous

MDR-TB

multidrug-resistant tuberculosis

XDR-TB

extensively drug-resistant tuberculosis

MIC

minimum inhibitory concentration

Mtb

Mycobacterium tuberculosis

MW

molecular weight

NHS

N-hydroxysuccinimide

NRPS-PKS

nonribosomal peptide synthetase–polyketide synthase

PK

pharmacokinetic

p.o.

oral

Sal-AMS

5′-O-[N-(salicyl)sulfamoyl]adenosine

SAR

structure–activity relationship

t_1/2

terminal elimination half-life following an oral dose

TB

tuberculosis

TBS

tert-butyldimethylsilyl

TFA

trifluoroacetic acid