2.3  RNA methyltransferase activity assays
Compounds 1–16 were tested for their ability to inhibit the methylation of the RNA cap structure. The inhibition induced by each compound (50 μM) was determined by a radioactive MTase assay (filter binding assay) which consists in measuring the [3H] radiolabeled methyl transferred from the methyl donor SAM onto RNA substrate (GpppAC4) synthetized by using T7 primase [33]. Compounds 1–16 designed as mimics of the transition state of RNA 2′-O-methylation were screened against several viral RNA 2′-O-MTases from SARS-CoV (nsp10/nsp16 complex), Zika, West-Nile, Dengue, Vaccinia (VP39) viruses. At the same time, the compounds were tested against human RNA N7-MTase (hRNMT) and selected viral N7-MTases such as SARS-CoV nsp14 and Vaccinia D1-D12 complex to evaluate their specificity (Table 1 ).
Table 1 Screening for inhibitory activity of sinefungin and compounds 1–16 at 50 μM on N7-MTases.
Compounds Percentage of inhibition at 50 μM (%)a
SARS-CoV nsp14 Vaccinia virusD1-D12 hRNMT
Sinefungin 98.3 ± 0.2 99.8 ± 0.1 99.8 ± 0.2
1 31.0 ± 6.8 20.3 ± 0.8 35.2 ± 4.9
2 72.0 ± 1.2 85.8 ± 2.5 77.4 ± 1.2
3 30.6 ± 9.3 32.1 ± 2.4 33.2 ± 4.3
4 13.1 ± 13.3 53.2 ± 2.6 12.2 ± 2.1
5 n.i n.i 27.5 ± 6.6
6 38.4 ± 11.7 11.6 ± 7.1 23.1 ± 9.7
7 n.i 69.2 ± 1.9 32.8 ± 16.1
8 43.0 ± 4.0 n.i n.i
9 88.6 ± 1.3 49.8 ± 3.2 66.0 ± 6.1
10 96.6 ± 0.9 4.6 ± 0.3 31.8 ± 3.3
11 47.6 ± 2.8 5.3 ± 4.3 44.2 ± 8.5
12 94.6 ± 1.1 10.1 ± 5.5 23.3 ± 3.6
13 97.2 ± 2.7 2.8 ± 0.8 33.9 ± 3.3
14 96.2 ± 1.5 19.7 ± 3.8 20.2 ± 9.4
15 94.0 ± 1.1 4.3 ± 3.9 n.i
16 75.9 ± 2.5 4.5 ± 15.1 14.7 ± 1.3
aValues are the mean of three independent experiments. The MTase activity was measured using a filter binding assay. Assays were carried out in reaction mixture [40 mM Tris-HCl (pH 8.0), 1 mM DTT, 1 mM MgCl2, 2 μM SAM and 0.1 μM 3H-SAM] in the presence of 0.7 μM GpppAC4 synthetic RNA and incubated at 30 °C. SARS-CoV nsp14 (50 nM), vaccinia virus capping enzyme (D1-D12) (41 U), human RNA N7 MTase (hRNMT) (50 nM). Compounds were previously dissolved in 100% DMSO. n.i: no inhibition detected at 50 μM.
Unexpectedly, all the bisubstrate compounds were barely active against the 2′-O MTases of flaviviruses or coronavirus SARS-CoV. In contrast, most of the compounds displayed inhibition of N7-MTases. Dinucleoside 2 bearing the amino acid chain of the SAM showed some significant inhibition of both viral N7-MTases with a better activity on Vaccinia D1-D12 than on SARS-CoV nsp14. However, compound 2 also displayed a potent inhibition of hRNMT in the same range as the viral MTases displaying a lack of specificity against human and viral enzymes. The amino acid group of 2 seems essential for inhibition since compound 1 with a non-substituted NH linker weakly inhibited the three MTases. The replacement of the amino acid group with an α-amino-ester at the extremity in compound 3 is detrimental for the inhibitory activity. Interestingly, the dinucleoside 4 bearing an α-amino-amide specifically inhibited the viral protein Vaccinia D1-D12 complex whereas did not show any inhibition of SARS-CoV nsp14 or hRNMT. Replacing the amino acid chain by a more hydrophobic butyl or phenylpropyl chain in dinucleosides 5 and 6, respectively, we aimed at favoring the Van der Waals interactions in hydrophobic pockets of the protein. Only compound 6 showed a moderate but specific inhibition of SARS-CoV nsp14. The removal of the NH2 of the amino acid chain of the broader spectrum inhibitor 2 with an ester-ended butyl chain in compound 7 or with an acid-ended butyl chain in 8 induced weaker but more specific inhibitions of Vaccinia D1-D12 MTase and SARS-CoV nsp14. In the synthetic pathway of dinucleoside 1, the intermediate 19 bearing a 4-Ns-amide group was prepared. In view of the valuable properties of such motif in some antiviral or anticancer drugs [19,20] it seemed interesting to us to obtain dinucleoside 9 by simple acidic deprotection of 19. Of special interest, compound 9 showed a good and specific inhibition on SARS-CoV nsp14 confirming that the nosyl group contributes to the inhibitory activity with specificity. Then, we modulated the initial nosyl moiety with the nitro group in “para” position by introducing diverse hydrophobic substituents (Cl, OMe, CF3) at different positions on the phenyl ring and/or by varying the position of the nitro group in compounds 10–13. The addition of such substituents aimed at increasing the interactions with proteins. Like 9, the four dinucleosides 10–13 maintained a high inhibitory activity on SARS-CoV nsp14. The role of the nitro group on the phenyl ring was demonstrated by removing it in dinucleoside 14 bearing solely one chlorine atom in “para” position, thus the inhibitory effect was slightly decreased. These data indicated the importance of the hydrophobic Cl in “para” position. Finally, the resulting decreased inhibition when the sulfone moiety of 14 was replaced by a methylene group in compounds 15 and 16 stressed the importance of the N-arylsulfonylbenzene motif in the dinucleoside structure to maintain an effective inhibition.