A Parallel Synthesis Approach to the Identification of Novel

Feb 15, 2016 - British Columbia Centre for Excellence in HIV/AIDS, 608-1081 Burrard Street, ... people die from AIDS per year.2,3 In practice, the adv...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/jmc

A Parallel Synthesis Approach to the Identification of Novel Diheteroarylamide-Based Compounds Blocking HIV Replication: Potential Inhibitors of HIV‑1 Pre-mRNA Alternative Splicing Peter K. Cheung,‡ David Horhant,† Laura E. Bandy,† Maryam Zamiri,† Safwat M. Rabea,† Stoyan K. Karagiosov,† Mitra Matloobi,† Steven McArthur,§ P. Richard Harrigan,‡,∥ Benoit Chabot,⊥ and David S. Grierson*,† †

Faculty of Pharmaceutical Sciences, §Department of Microbiology and Immunology, and ∥Department of Medicine, The University of British Columbia, Vancouver, British Columbia V6T 1Z3, Canada ‡ British Columbia Centre for Excellence in HIV/AIDS, 608-1081 Burrard Street, Vancouver, British Columbia V6Z 1Y6, Canada ⊥ Département de microbiologie et d’infectiologie, Faculté de médecine et des sciences de la santé, Université de Sherbrooke, 3201, rue Jean-Mignault, Sherbrooke, Québec J1E 4K8 Canada S Supporting Information *

ABSTRACT: A 256-compound library was evaluated in an anti-HIV screen to identify structural “mimics” of the fused tetracyclic indole compound 1 (IDC16) that conserve its anti-HIV activity without associated cytotoxicity. Four diheteroarylamide-type compounds, containing a common 5-nitroisobenzothiazole motif, were identified as active. In subsequent screens, the most potent compound 9 (1C8) was active against wild-type HIV-1IIIB (subtype B, X4-tropic) and HIV-1 97USSN54 (subtype A, R5-tropic) with EC50’s of 0.6 and 0.9 μM, respectively. Compound 9 also inhibited HIV strains resistant to drugs targeting HIV reverse transcriptase, protease, integrase, and coreceptor CCR5 with EC50’s ranging from 0.9 to 1.5 μM. The CC50 value obtained in a cytotoxicity assay for compound 9 was >100 μM, corresponding to a therapeutic index (CC50/ EC50) of approximately 100. Further comparison studies revealed that, whereas the anti-HIV activity for compound 9 and the parent molecule 1 are similar, the cytotoxic effect for compound 9 was, as planned, markedly suppressed.



INTRODUCTION

that act through unexplored mechanisms of action, which bypass resistance, display minimal toxicity, and address the problem of activating the latent viral pool.9 We are exploring the potential of new anti-HIV agents that block HIV replication through perturbation of HIV pre-mRNA alternative splicing. Through random screening of compound libraries several molecules10−16 have been identified that either directly or indirectly interfere with the events whereby HIV pre-mRNA is alternatively spliced.17−20 Pertinent to the present work, screening of the Institut Curie compound collection21 identified the fused tetracyclic indole compound 1 (IDC16) (Figure 1) as a molecule that blocks replication of HIV through inhibition of the exonic splicing enhancer activity of SRSF1.10 This SR protein plays a crucial role in human RNA splicing, and it has been implicated in alternative splicing, mRNA stability, and translation.22 In HIV-1 infected cells it specifically regulates alternative splicing events that are vital to the production of

Anti-retroviral therapy (ART) has changed the prognosis of HIV infection to the point where, in developed countries, it is considered a chronic infection. In light of this success in reducing morbidity,1 significant steps are now being taken to bring ART to the developing world where the majority of HIV infected people reside. The necessity of these initiatives is underscored by the fact that, in spite of progress, an estimated 2 million new infections occur and approximately 1.2 million people die from AIDS per year.2,3 In practice, the advantage of ART therapy is that the use of a combination of drugs reduces the chance of viral escape, which results from acquired drug resistance to each individual drug. However, there are inherent long-term limitations to ART, which include the necessity for strict adherence to treatment schedules,4 drug toxicity,5 and the emergence of drug-resistant HIV strains,6 which are no longer sensitive to the drug cocktails employed. These issues take on added significance as the scale and diversity of the populations receiving treatment increases.7,8 There is need for the continued development of new drugs, and in particular those © XXXX American Chemical Society

Received: September 2, 2015

A

DOI: 10.1021/acs.jmedchem.5b01357 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Figure 1. Strategy for the parallel synthesis of compound 1 mimics, showing the different linker motifs 2−7 that were investigated, the general formula 8 for the series of 119 diheteroarylamides synthesized, and the structures of the compounds 9−12 identified as active in the preliminary antiHIV screen.

supplementary binding interactions not exploited by 1 could be located, which improve potency and/or target selectivity. As described in the present report, the screening of a small molecule library of structural mimics of 1 led to the identification of four active diheteroarylamide (DHA)-type compounds (Figure 1). The most potent hit in this new class of anti-HIV agents was carried forward to determine its anti-HIV profile against HIV mutant strains resistant to the current drugs used in ART.

critical HIV proteins including Tat, Rev, and Nef. The RNAbinding proteins Tat and Rev are key regulators for the expression of HIV-1 viral genes, for the synthesis of full-length genomic RNA and, ultimately, for the production of progeny virions.17,23 Compound 1 and other pyridocarbazole alkaloid analogues have been developed primarily as cytotoxic agents, acting, among other mechanisms, as DNA intercalators and/or topoisomerase II inhibitors.24−28 Consequently, although 1 blocks HIV replication at concentrations lower than the threshold of its cytotoxic effect (see ref 10 and results in Figure 5), it is not a suitable molecule to carry down the drug development process as an anti-HIV drug destined for longterm use. Be this as it may, the significant anti-HIV activity displayed by this molecule strongly suggests that it interacts with its target SR protein in a way that it disrupts its contact with other spliceosome components. It is possible, therefore, that the information contained in the structure of 1 (geometry, polarity, presence of aromatic rings), which is at the root of its anti-HIV activity, could be translated into a different and more flexible type of molecule that maintains this mechanism of action, without displaying affinity for DNA. Further, it is anticipated that such molecules would bind more strongly in the putative site than the rigid and flat compound 1. Our approach to identify molecules with these features has been to use parallel synthesis to construct a library of ringopened, and more conformationally mobile “mimics” of 1 and to evaluate the library components for anti-HIV activity. In the absence of high-resolution structural data for the SR proteins and only a limited vision of their interactions at the molecular level, the library design strategy adopted was to construct molecules that retain a structural resemblance to the outline of 1, when projected onto a flat (two-dimensional) surface. Of importance was to also include molecules in the library that incorporate functionality and/or structural motifs that project beyond the outline of 1. It was expected that, in this way,



RESULTS Chemistry and Preliminary Screen. The tetracyclic indole compound 1 contains three nitrogen atoms in an otherwise highly lipophilic polyaromatic structure. Taking as a supposition that these polar heteroatoms play an important role in interactions with the SR protein, the principle objective in the synthesis of mimics of this molecule was to generate structures that position nitrogen and/or oxygen functionality into the same space occupied by its indole and A/D-ring nitrogen atoms. Further, as the B- and C-rings in 1 may be contributing little to binding, the idea was to replace one or both of these rings by an appropriate spacer element that connects the different motifs used as alternatives to the A- and D-rings (Figure 1). Spacer elements explored, which replace rings B and C, included a 1,2,4-oxadiazole motif 2, an extended 1,2,4-oxadiazole-3-carboxamide system 3, and a simple amide function 4. Fused bicyclic A/B-type ring systems, where the Bring is part of the spacer element and the C-ring is eliminated included imidazopyridine 5, pyrazolopyridine 6, and pyridoimidazolone 7 systems. Using a parallel synthesis approach, a diverse library of 256 compound 1 mimics was prepared by combining A-and A/B-ring components with different D-ring motifs (see Table 1 in the Supporting Information for a complete listing of all compounds in the library). Focusing in on the subset of IDC16 mimics containing an amide linker, a total of 119 compounds, represented by general formula 8, were prepared by reaction of the acid chloride

B

DOI: 10.1021/acs.jmedchem.5b01357 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Figure 2. Structures of the different A- and D-ring subunits employed in the synthesis of the 119 dihereroarylamides 8 present in the compound library.

(3C2), 11 (2D3) and 12 (1E5) (microplate number and compound location) (Figure 1), were identified as active, inhibiting HIV-1 replication in the low micromolar range (see Figure 1a−d in the Supporting Information for the screening results). The immediately apparent feature of these molecules was that, whereas the A-ring motif could be varied among different pyridine/pyridinone structures, they all had in common the presence of a 5-nitro substituted benzo[d]isothiazole motif as the D-ring subunit. The four step preparation of compound 9 involved (Scheme 1a): (i) the conversion of 4-chloropyridine 13 to 4chloronicotinic acid 14 through reaction with LDA followed by CO2 gas;30 (ii) refluxing this intermediate in H2O at reflux;31 (iii) reaction of the derived 4-pyridinone 1532 with excess methyl iodide to give 1633 (using KHCO3 rather than NaOH as the base; 90% yield); and (iv) the reaction of the acid chloride derivative of N-methyl-4-pyridinone-3-carboxylic acid 16 with 5-nitrobenzo[d]isothiazol-3-amine. Compound 10 was prepared (Scheme 1a) by reaction of 4chloronicotinic acid 14 with the anion of 2-methoxyethanol to give 17, followed by subsequent conversion of 17 to the acid chloride and its reaction with 5-nitrobenzo[d]isothiazol-3-

derivative of the selected A-ring carboxylic acid components with a diverse set of functionalized aromatic and heteroaromatic amines (Figure 2). Looking at the set of D-ring components explored, one see that, in some cases, substituents were situated at different positions of the aromatic and heteroaromatic ring that are similar in size to the chlorine atom in 1. In other cases, more extended heteroatom-containing side chains were explored, and in still other cases the D-ring alternatives corresponded to fused and nonfused bicylic motifs. Preliminary evaluation of these and the other library molecules was carried out in a cell-based anti-HIV screen, in which reporter CEM-GXR cells were infected by coculture with 1% HIV-1NL4‑3 infected (GFP positive) cells at the start of the assay (day 0). CEM-GXR cells are an immortalized CD4+ Tlymphocyte line that expresses GFP due to the activation of the Tat-dependent promoter upon HIV-1 infection.29 A different library molecule in DMSO was added to the culture in each microplate well immediately after the inoculation by infected cells. The final library compound concentration was fixed at 2 μM, and the final DMSO concentration was 95% purity, as determined by 1H NMR. In the 1H NMR spectra of the four active compounds, the three peaks for the benzisothiazole motif (singlet between δ 8.85 and 9.40 for H-4′, and doublets for the adjacent H-6′/7′ protons between δ 8.13 and 8.17 and δ 7.75 and 7.77) were readily distinguished from the absorptions for the pyridine/pyridinone protons of the A-ring components. Anti-HIV Activity of 9 against Wild-Type and DrugResistant HIV Strains. Assessment of the four initial hits by dose−response relations indicated that compounds 10 and 12 did not respond to the increasing doses. This correlates with their poor DMSO-H2O solubility. In contrast, compounds 9 and 11 displayed good response. As the antiviral potency of 9 was approximately 3-fold higher than that for 11, compound 9 D

DOI: 10.1021/acs.jmedchem.5b01357 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Figure 3. Anti-HIV-1 activity of 9 against wild-type subtypes A and B and drug-resistant viruses. (A) HIV-1IIIB (subtype B), (B) 97USSN54 (subtype A), (C) (N)NRTI-resistant isolate, (D) PI-resistant isolate, (E) INI-resistant isolate, (F) Maraviroc-resistant R5 strain. The anti-HIV activity for emtricitabine (FTC) as a positive control is also shown in (A). In the presence of increasing 9 concentrations, HIV-1 infectivity on CEM-GXR cells was assessed by measuring GFP positive cells 3 days after infection. The inhibition curves were then fitted by nonlinear regression with GraphPad Prism software, allowing EC50 calculation. Results represent the average of three independent determinations.

therapeutic index signifies desirable antiviral activity with minimal cell toxicity. The potential adverse effect of 9 on CEM-GXR cells viability was determined by the Guava ViaCount assay. This assay differentially stains viable and nonviable cells based on their permeability to the DNA-binding dyes and uses the forward scatter (FSC) properties to distinguish free nuclei and cellular debris from intact cells to quantify cell count. CEM-GXR cells were incubated with 9 at concentrations ranging between 0.5 and 100 μM for 24 h. At the highest concentration tested (100 μM of 9), a decrease in cell viability from 98.0 to 68.8% (±6.8) (Figure 4) was observed. The cytotoxic concentration resulting in the 50% death of the host cells (CC50) was, thus, >100 μM. As the EC50s in all the tested viruses ranged between 0.6 and 1.5 μM, the calculated therapeutic indexes are >66−100 (Table 2). Comparative Study of 1 and 9 on Cell Viability and Anti-HIV Activity. To study further cell viability versus anti-

Table 1. Mutations in Drug-Resistant HIV-1 Strains Associated with Resistance to ARV Drugsa drugresistant HIV-1 strain

viral protein

primary mutations

E00443

reverse transcriptase

2948 11845

protease integrase

K103N, D67N, K70R, Y115F, Q151M, M184V, K219Q G48V, L90M G140S, Q148H

MVC-res

envelop-V3

A19T, L20F, T22A, E25D, I26V

resistance to ARV drugs EFV, NVP, 3TC, ABC, AZT, FTC ATV, LPV RAL, EVG, DTG MVC

a

Abbreviations: efavirenz (EFV), nevirapine (NVP), lamivudine (3TC), abacavir (ABC), zidovudine (AZT), emtricitabine (FTC), atazanavir (ATV), lopinavir (LPV), raltegravir (RAL), elvitegravir (EVG), dolutegravir (DTG), and maraviroc (MVC). E

DOI: 10.1021/acs.jmedchem.5b01357 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

four initial hit molecules. Although, at this stage, the screening results provide only a primitive guide to future SAR studies on this subunit, it is interesting to note the absence of activity for the 4-pyridinone-based compounds 22 (1B8) and 23 (1D8) and the 2-pyridinone molecules 24 (4B2) and 25 (4F3), which contain a single five-membered heterocycle as the D-ring subunit (Figure 6). It is further noteworthy that the isomeric benzothiazole compounds 26 (4C2) and 27 (4A4), where the benzo-ring has an “extended” spatial orientation versus the “bent orientation found in 9 and 12, were also inactive. However, replacement of the benzisothiazole subunit by a related isoquinoline motif, as in the 12-type compound 28 (4C3), also resulted in a loss of activity. Important to future SAR studies on the active mimics of 1 was the observation that considerable latitude exists to replace the A-ring of 1 by alternate substructures. Indeed, the position of the pyridine nitrogen atom relative to the carboxamide function can be changed, a 2-pyridinone system is tolerated, and the indole nitrogen in 1 can be replaced by the 4pyridinone carbonyl oxygen atom, as in 9, or a 2methoxyethanol side chain, as in 10 and 12. Interestingly, compound 29 (2C3), lacking this oxygen substituted side chain, is inactive. Importantly, the activity versus viability data in Figure 5 revealed a distinct advantage of 9 over 1 as an anti-HIV agent. Both compounds exhibited comparable EC50 values and had little effect on cell viability at 1 μM concentration. However, at higher concentrations, toxicity toward 1 was strongly manifested, whereas, the mimic 9 remained minimally toxic to the CEM-GXR cells. This result was highly encouraging, as it confirmed that, unlike 1, compound 9 displays no significant affinity for DNA. However, when taking into consideration long-term anti-HIV therapy, an important issue in moving forward with development of this molecule is the presence of the 5-nitro substituent on the isobenzoisothiazole motif. A priority will thus be given to establishing whether the nitro is crucial to activity, and if so, whether it can be replaced by a suitable bioisostere motif, which is less prone to bioactivation.41 Finally, having discovered a new prototype anti-HIV agent, which was inspired by the structure of 1, it remains to determine whether, like the parent compound, 9 blocks HIV replication through perturbation of HIV-1 pre-mRNA alternative splicing. In this context, it is noteworthy that the antiHIV agent 30 (ABX-464), whose development was also inspired by the structure of 1, interfered with HIV alternative splicing in a manner that does not implicate SRSF1.16 Considering that 1 and 9 are, in several ways, structurally dissimilar, one cannot, therefore, assume that they share the same mechanism of action. This line of reasoning was reinforced by the fact that compound 31 (1H7) (Figure 6), which mostly closely follow the outline of 1, was inactive. This suggested that the amide linker in the DHA compounds may not correctly emulate the B- and C-rings of 1 (i.e., capture the same interactions these rings are implicated in upon binding to the SR protein) and/or that the extra phenyl ring in the bennzoisothiazole system captured additional interactions, which compensate for this possible deficiency. To establish the mode of action of 9, very detailed experiments have been carried out, which both eliminated SRSF1 as the target of 9 and confirmed that it deregulates HIV-1 pre-mRNA alternative splicing in a complex process involving other SR proteins. The results of these ongoing studies42 serve to validate the approach taken to design 9 as HIV alternative splicing inhibitor. They

Figure 4. Evaluation of 9 on CEM-GXR cell viability in the Guava ViaCount assay. Cells were analyzed after 24 h incubation with compound 9 in concentrations ranging between 0.5 and 100 μM. Results are expressed as the percentage (%) of viable cells ± SEM of three independent experiments.

Table 2. Antiviral Potencies of Compound 9 in Cells Infected with Wild-Type or Drug-Resistant Virus Strainsa WT and drug resistant virus

resistance to drug category

CC50 (μM)b

EC50 (μM)c

therapeutic indexd

HIV-1IIIB 97USSN54 E00443 2948 11845 MVC-res

none none RTI PI INI CRI

>100 >100 >100 >100 >100 >100

0.6 0.9 1.3 1.4 1.5 0.9

>166 >111 >77 >71 >66 >111

a

Abbreviations: reverse transcriptase inhibitor (RTI), protease inhibitor (PI), integrase inhibitor (INI), and coreceptor inhibitor (CRI). bCytotoxic concentration resulting in the death of 50% of the host cells (CC50) measured by ViaCount assay (Millipore). cEffective concentration resulting in 50% reduction of viral infected cells measured by GFP expression signal. dTherapeutic index calculated as the ratio of CC50 to EC50.

HIV activity, a comparison study was made between 1 and the mimic 9 (Figure 5A,B). Serial dilutions of both molecules in final concentrations between 62.5 nM and 16 μM were added in the culture containing either CEM-GXR cells alone for the viability study or in one percent of HIV-1IIIB infected CEMGXR cells for antiviral activity assessment. At 1 μM concentration, both molecules showed comparable antiviral activities with 18.4% infected cells detected in 9 and 17.7% detected in 1. The half-maximal effective concentration (EC50) values for both molecules were calculated using the Igor Pro software. Compounds 1 and 9 exhibited equivalent activity against HIV-1IIIB with EC50 values of 0.959 ± 0.026 and 0.902 ± 0.113 μM, respectively. Similar effects on cell viability were observed for both molecules at 1 μM, as 89.6% viable cells were detected in 9 and 85.5% were found in 1. At higher concentrations, however, the cell viability dropped considerably for 1 (33.2% at 4 μM, 19.9% at 8 μM, and 11.9% at 16 μM) reflecting its cytotoxicity, while only minimal effect on viability was observed for 9 (81.5% at 4 μM, 80.9% at 8 μM, and 78.7% at 16 μM).



DISCUSSION AND CONCLUSIONS Taken together, the anti-HIV assay results strongly suggest that compound 9 does not inhibit HIV replication through a mechanism involving inhibition of the four current HIV protein drug targets. Looking at the structure of this molecule, the importance of the 5-nitro isobenzoisothiazole system to activity was evident from its appearance as the D-ring component in all F

DOI: 10.1021/acs.jmedchem.5b01357 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

Figure 5. Evaluation of anti-HIV-1 activity (column) and GXR-CEM cell viability (line) for 1 (A) and 9 (B) at concentrations between 62.5 nM and 16 μM (0 provides the value when no compound was added). HIV-1 infection in CEM-GXR cells was assessed by measuring GFP positive cells 3 days after infection. CEM-GXR cell viability was measured after 24 h incubation within the same range of concentrations. Results were expressed from three independent experiments for both the anti-HIV-1 activity and cell viability assays. compound 11 contained trace amounts (approximately 5%) of an unidentified impurity (see 1H NMR spectrum in the Supporting Information). For the biological assays: nelfinavir, efavirenz and 3′azido-3′-deoxythymidine (AZT) were obtained from Sigma-Aldrich (St. Louis, MO). Maraviroc, raltegravir, elvitegravir, dolutegraver, and emtricitabine were obtained from Cedarlane Laboratories (Ontario, Canada). Indicator cell line CEM-GXR was obtained from Dr. Mark Brockman (Simon Fraser University). The (N)NRTI-resistant virus (E00443) was obtained from Dr. Zabrina Brumme (Simon Fraser University). The HIV-1 CXCR4-tropic laboratory-adapted strain NL43, and IIIB, HIV-1 CCR5-tropic BaL, integrase inhibitor-resistant virus (11845), protease inhibitor-resistant virus (2948), and HIV-1 subtype A (4114) were obtained through the NIH AIDS Research and Reference Reagent Program, Division of AIDS, NIAID, NIH). 1-Methyl-N-(5-nitrobenzo[d]isothiazol-3-yl)-4-oxo-1,4-dihydropyridine-3-carboxamide (9). Thionyl chloride (0.23 mL, 3.13 mmol, 1.2 equiv) was added dropwise to an ice-bath cooled suspension of 1methyl-4-oxo-1,4-dihydropyridine-3-carboxylic acid 16 (400 mg, 2.6 mmol) in dry CH2Cl2 (6 mL) under a flow of nitrogen. After the addition, the reaction mixture was warmed to room temperature and then stirred at room temperature for 2 h. The resulting solution was then concentrated under reduced pressure, and the crude acyl chloride was dried in vacuo for 1 h. The solid product was dissolved in DMF (3 mL) containing Et3N (4.5 mL) and cooled to 0 °C. Then a solution of 5-nitrobenzo[d]isothiazol-3-amine (612.5 mg, 3.13 mmol) in DMF (8

further provide the tools necessary for continued research to identify the optimal 9 analog that blocks HIV replication without detrimental effect on normal cellular splicing/ alternative splicing processes.



EXPERIMENTAL SECTION

Experimental General. All chemicals were purchased from SigmaAldrich or Oakwood Chemicals, and they were used without purification unless mentioned. All solvents were dried and kept under N2. 1H and 13C NMR spectra were recorded at 400 and 100 MHz, respectively, on a Bruker AC 400 Ultrashield 10 spectrophotometer. Chemical shifts are expressed in ppm, (δ scale). When peak multiplicities are reported, the following abbreviations are used: s (singlet), d (doublet), m (multiplet), dd (doublet of doublet). Coupling constants are reported in Hz. Low- and high-resolution mass spectra were recorded on a AB Sciex UHPLC/MS/MS System and a Thermo Scientific Q Exactive Orbitrap High-Resolution Mass Spectrometer, respectively. Flash column chromatography was performed using silica gel (Silicycle, Siliaflash F60, 40−63 μm, 230− 400 mesh) or on a Biotage Isolera purification system, (PartnerTech Atvidaberg AB) using prepacked silica gel columns (Biotage, part no. FSKO-1107-0010, FSKO-1107-0025, or FSKO-1107-0050). Intermediates 15−21 and the final compounds 9, 10, and 12 were confirmed to be >95% pure by 1H NMR spectroscopy. Final G

DOI: 10.1021/acs.jmedchem.5b01357 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

2-Methoxy-N-(5-nitrobenzo[d]isothiazol-3-yl)isonicotinamide (11). Thionyl chloride (10 mL) was added to acid 18 (500 mg, 3.27 mmol) (commercially available), and the mixture was heated at 90 °C for 1h. The excess of thionyl chloride was removed under reduced pressure, and the crude acyl chloride was suspended in THF (30 mL). Potassium carbonate (902 mg, 6.54 mmol, 2 equiv) and 5nitrobenzo[d]isothiazol-3-amine (637 mg, 3.27 mmol) were then added, and the resulting mixture was stirred at room temperature for 10 h. The mixture was then filtered, and the solid material was washed with THF. The filtrate was concentrated in vacuo and flash column chromatographed (silica-gel: 0−2% MeOH in CH2Cl2), providing 11 as a yellow solid (20 mg, 2%). 1H NMR (DMSO-d6) δ 13.18 (br s, 1H), 9.70 (s, 1H), 8.47 (d, J = 5.2 Hz, 1H), 8.16 (dd, J = 1.2, 9.6 Hz, 1H), 7.77 (d, J = 9.6 Hz, 1H), 7.63 (d, J = 5.2 Hz, 1H), 7.55 (s, 1H), 3.97 (s, 3H); 13C NMR (DMSO-d6) δ 164.04, 148.12, 141.74, 120.70, 120.03, 115.48, 109.86, 53.81 (6 peaks for quaternary carbons are missing); HRMS (HESI) m/z: calcd for C14H9N4O4S [M − H]− 329.03500, found 329.03531. 4-(2-Methoxyethoxy)-1-methyl-N-(5-nitrobenzoisothiazol-3-yl)6-oxo-1,6-dihydropyridine-3-carboxamide (12). 4-(2-methoxyethoxy)-1-methyl-6-oxo-1,6-dihydropyridine-3-carboxylic acid 21 (100 mg, 0.40 mmol) was dissolved in an excess of thionyl chloride (1 mL), and the reaction mixture was heated at 85 °C for 1 h. The residual thionyl chloride was then removed in vacuo. 5-nitrobenzo[d]isothiazol-3-amine (66 mg, 0.34 mmol) in THF (3 mL) containing triethylamine (0.077 mL, 0.548 mmol) was then added, and the reaction mixture was stirred under nitrogen at room temperature overnight. The precipitate formed was removed by vacuum filtration, and the concentrated filtrate was washed with acetone. The acetone washed concentrate was silica-gel flash column chromatographed (EtOAc/MeOH; 9:1), providing compound 12 as a yellow solid (80 mg, 58%). 1H NMR (DMSO-d6) δ 12.34 (s, 1H), 9.36 (s, 1H), 8.54 (s, 1H), 8.16 (d, 1H, J = 10 Hz), 7.75 (d, 1H, J = 10 Hz), 6.05 (s, 1H), 4.28 (s, 2H), 3.81 (s, 2H), 3.50 (s, 3H), 3.22 (s, 3H); 13C NMR (DMSO-d6) δ 163.43, 162.93, 162.71, 161.83, 157.79, 145.62, 141.62, 122.27, 122.03, 119.95, 119.38, 104.40, 96.15, 69.58, 68.18, 58.06, 36.50; HRMS (HESI) m/z: calcd for C17H15N4O6S [M − H]− 403.07178; found 403.06946. 4-Oxo-1,4-dihydropyridine-3-carboxylic acid (15). Adapting the procedure by Ross,31 4-chloronicotinic acid 14 (1.29 g, 0.77 mmol) was dissolved in 20 mL of water and refluxed for 2 h. After the solution was cooled in an ice bath, a sufficient quantity of conc. HCl was added (if necessary) to ensure that the pH ≤ 4. The precipitate that formed was collected by suction filtration and washed with water. The aqueous filtrate was then concentrated and suspended in approximately 10 mL of water. After cooling at 0 °C, the precipitated material was again collected by suction filtration, washed with Et2O, and dried in vacuo to afford 15 as a colorless solid (800 mg, 75%). 1H NMR (DMSO-d6) δ 12.94 (bs, 1H), 8.60 (s 1H), 8.07 (m, 1H), 6.73 (d, J = 7.1 Hz, 1H); 13 C NMR (DMSO-d6) δ 179.43, 166.73, 143.38, 141.34, 118.13, 115.77; HRMS (HESI) m/z: calcd for C6H4NO3 [M − H]− 138.01967, found 138.01811. 1-Methyl-4-oxo-1,4-dihydropyridine-3-carboxylic acid (16). Adapting the procedure in ref 32, 4-oxo-1,4-dihydropyridine-3carboxylic acid 15 (792 mg, 5.7 mmol) was dissolved in DMF/water (5/1) (12 mL), and KHCO3 (590 mg, 5.9 mmol) was added. The mixture was heated at 40 °C for 1 h. After cooling to room temperature, a white solid was collected by suction filtration and washed with DMF (2 mL). To the filtrate was added MeI (1.42 mL, 0.023 mol, 4 equiv), and the mixture was heated at 100 °C for 1 h. It was then concentrated in vacuo, and the residue was washed with acetone, filtered, and rinsed with a 1/1 mixture of acetone/MeOH. The solid material was triturated with hot MeOH, and the MeOH washes were concentrated and dried in vacuo to afford 16 as a colorless solid (0.745 g, 85%). 1H NMR (DMSO-d6) δ 8.70 (d, J = 2.3 Hz, 1H), 8.10 (dd, J = 7.8, 2.3 Hz, 1H), 6.76 (d, J = 7.8 Hz, 1H), 3.89 (s, 3H); 13C NMR (DMSO-d6) δ 178.29, 165.54, 147. 31, 145.43, 118.56, 115.60, 34.88; HRMS (HESI) m/z: calcd for C7H8NO3 [M + H]+ 154.04987, found 154.04977.

Figure 6. Library components providing cursory SAR on compound 9 and the structure of 30 (ABX-464). mL) was added dropwise, and the reaction mixture was warmed to room temperature and stirred for 24 h. The DMF was removed under reduced pressure, and the solid mixture was washed with acetone (30 mL), filtered and concentrated. The solid obtained was purified twice by flash column chromatography (DCM/MeOH 95:5) to afford 9 as a red-brown solid (80 mg, 9%). 1H NMR (DMSO-d6) δ 15.58 (s, 1H), 8.85 (s, 1H), 8.75 (d, J = 2.0 Hz, 1H), 8.16 (dd, J = 9.6, 2.3 Hz, 1H), 8.04 (d, J = 7.10 Hz, 1H), 7.77 (d, J = 9.6 Hz, 1H), 6.75 (d, J = 7.10 Hz), 1H), 3.92 (s, 3H); 13C NMR (DMSO-d6) δ 176.41, 157.91, 147.09, 143.80, 141.60, 122.45, 122.15, 120.08, 119.69, 118.22 (3 peaks for quaternary carbons are missing); HRMS (HESI) m/z: calcd for C14H9N4O4S [M + H]− 329.03500, found 329.03531 4-(2-Methoxyethoxy)-N-(5-nitrobenzo[d]isothiazol-3-yl)nicotinamide (10). Thionyl chloride (1.8 mL) was added to acid 17 (50 mg, 0.25 mmol), and the mixture was heated at 90 °C for 1 h. The excess of thionyl chloride was removed under reduced pressure, and the crude acyl chloride was suspended in THF (3 mL). Potassium carbonate (70 mg, 0.51 mmol) and 5-nitrobenzo[d]isothiazol-3-amine (50 mg, 0.26 mmol) were then added, and the resulting mixture was stirred at room temperature for 10 h. The solvent was removed in vacuo, and the solid obtained was silica-gel flash column chromatographed (EtOAc to EtOAc/MeOH 9:1). Compound 10 was isolated pure as a yellow solid (5 mg, 5%). 1H NMR (DMSO-d6) δ 13.05 (br s, 1H), 9.45 (d, J = 2.3 Hz, 1H), 8.79 (br s, 1H), 8.66 (d, J = 5.9 Hz, 1H), 8.16 (dd, J = 9.7, 2.3 Hz, 1H), 7.77 (d, J = 9.7 Hz, 1H), 7.37 (d, J = 5.9 Hz 1H,), 4.37 (m, 2H), 3.71 (m, 2H), 3.20 (s, 3H); 13C NMR (DMSO-d6) δ 163.56, 162.82, 162.41, 157.79, 154,30, 150.66, 141,77, 122.27, 122.03, 120.02, 119.64, 118.36, 108.69, 69.79, 68.24, 58.17; HRMS (HESI) m/z: calcd for C16H13N4O5S [M − H]− 373.06121, found 373.06107. H

DOI: 10.1021/acs.jmedchem.5b01357 J. Med. Chem. XXXX, XXX, XXX−XXX

Journal of Medicinal Chemistry

Article

4-(2-Methoxyethoxy)nicotinic acid (17). A suspension of sodium (1.15 g, 50 mmol) in 2-methoxyethan-1-ol (14 mL) was heated at 90°−100 °C under nitrogen until the sodium was consumed.33 4Chloronicotinic acid 14 (1.07 g, 6.79 mmol) was then added in one portion, and the reaction mixture was heated at 90 °C for 4 h. After cooling to room temperature, the reaction mixture was poured into ice water (75 mL) and extracted with CH2Cl2 (3 × 30 mL). The aqueous phase was concentrated under reduced pressure to about 1/5 of the initial volume and acidified with 6 N HCl at 0 °C until pH 3.5−4.0 was attained. The precipitate formed was filtered, washed with ice water and ether, and dried in vacuo. Subsequent crystallization from ethanol/ether gave 17 (545 mg, 41%) as a white solid. 1H NMR (DMSO-d6) δ 12.96 (br s, 1H), 8.67 (s, 1H), 8.53 (d, J = 5.7 Hz, 1H), 7.18 (d, J = 5.9 Hz, 1H), 4.26 (m, 2H), 3.68 (m, 2H), 3.31 (s, 3H); 13 C NMR (MeOH-d4) δ 171.96, 163.91, 146.39, 145.92, 121.17, 113.20, 72.58, 71,21, 59.53; HRMS (HESI) m/z: calcd for C9H12NO4 [M + H]+ 198.07608, found 198.07617. Ethyl 4-(2-Methoxyethoxy)-1-methyl-6-oxo-1,6-dihydropyridine3-carboxylate (20). To a solution of 1934 (4 g, 20.30 mmol) in DMF (120 mL) were added Cs2CO3 (8.4 g, 25.78 mmol) and 2-bromoethyl methyl ether (2.4 mL, 25.54 mmol). The yellow colored mixture was stirred for 4 h at 100 °C and then concentrated under reduced pressure. The resulting orange solid was taken-up in water and CH2Cl2 and extracted with CH2Cl2 (4 × 50 mL). The combined organic layers were dried over Na2SO4 and concentrated. The crude product mixture was silica-gel flash chromatographed (0−2% of MeOH in EtOAc), providing compound 20 as a colorless solid (4.45 g, 86%). 1H NMR (CDCl3) δ 8.07 (s, 1H), 5.83 (s, 1H), 4.22 (dd, 2H, J = 14.3 Hz, J = 7.1 Hz), 4.04 (t, 2H, J = 4.5 Hz), 3.75 (t, 2H, J = 4.2 Hz), 3.49 (s, 3H), 3.41 (s, 3H), 1.30 (t, 3H, J = 7.1 Hz); 13C NMR (CDCl3) δ 165.74, 163.92, 163.36, 145.14, 104.62, 97.09, 70.23, 68.35, 60.88, 59.37, 37.29, 14.24; HRMS (HESI) m/z: [M + H]+: calcd for C12H18NO5 256.11795; found 256.11765. 4-(2-Methoxyethoxy)-1-methyl-6-oxo-1,6-dihydropyridine-3-carboxylic acid (21). A mixture of 20 (7.8 g, 30.6 mmol) and LiOH (2.2 g, 52.8 mmol) in THF/H2O (2−1) (470 mL) was stirred for 4 h at room temperature. The mixture was then acidified to pH 1 using concentrated HCl. The product was extracted with CH2Cl2, and the aqueous layer was washed with CH2Cl2 (4 × 25 mL). The combined organic layers were dried over Na2SO4, filtrated, and concentrated to afford 21 as a colorless solid (6.18 g, 89%). 1H NMR (CDCl3) δ 8.34 (s, 1H), 5.99 (s, 1H), 4.26 (t, 2H, J = 9 Hz), 3.80 (t, 2H, J = 9 Hz), 3.57 (s, 3H), 3.45 (s, 3H); 13C NMR (CDCl3) δ 164.18, 163.85, 163.85, 147.03, 102.77, 97.58, 69.54, 68.96, 59.38, 37.73; HRMS (HESI) m/z: [M − H]+: calcd for C10H14NO5 228.08665 ; found 228.08621. Preliminary Library Screen for Anti-HIV-1 Activity. The antiHIV-1 activity of the library molecules was measured using a human T-cell reporter assay based on CEM-GXR cells.29 These cells naturally express the HIV-1 receptor CD4 and coreceptor CXCR4 and have been engineered to expresses the other HIV-1 coreceptor, CCR5. Upon HIV-1 infection, CEM-GXR cells express GFP due to the activation of an exogenous Tat-driven LTR-GFP expression cassette, and the level of infection was monitored using flow cytometric analysis (guavaSoft 2.2 software, Guava HT8, Millipore). Cells were grown in RPMI-1640 media (Invitrogen) supplemented with 10% fetal calf serum (Invitrogen) and 100 U per ml penicillin-streptomycin. During propagation, CEM-GXR cells were selected with 0.2 mg/mL G418 (Invitrogen) and 0.1 mg/mL hygromycin B (Calbiochem). Antiviral activity was evaluated in the assay that measures inhibition of HIV-1 spread in a coculture of CEM-GXR cells containing 1% of HIV-1NL4−3 infected (GFP positive) cells. Infection was performed in T25 flasks or in 96-well plates containing 1 × 106 or 8 × 104 CEM-GXR cells, respectively. The library molecules at 2 μM concentration and commercially available anti-HIV-1 reagents zidovudine 2 μM and nelfinavir 100 nM were added to the culture immediately after the inoculation by infected cells. All compounds were prepared in DMSO at the final concentration in culture that were