A new antiviral scaffold for human norovirus identified with computer-aided approaches on the viral polymerase

Human norovirus is the leading cause of acute gastroenteritis worldwide, affecting every year 685 million people. In about one third of cases, this virus affects children under five years of age, causing each year up to 200,000 child deaths, mainly in the developing countries. Norovirus outbreaks are associated with very significant economic losses, with an estimated societal cost of 60 billion dollars per year. Despite the marked socio-economic consequences associated, no therapeutic options or vaccines are currently available to treat or prevent this infection. One promising target to identify new antiviral agents for norovirus is the viral polymerase, which has a pivotal role for the viral replication and lacks closely homologous structures in the host. Starting from the scaffold of a novel class of norovirus polymerase inhibitors recently discovered in our research group with a computer-aided method, different new chemical modifications were designed and carried out, with the aim to identify improved agents effective against norovirus replication in cell-based assays. While different new inhibitors of the viral polymerase were found, a further computer-aided ligand optimisation approach led to the identification of a new antiviral scaffold for norovirus, which inhibits human norovirus replication at low-micromolar concentrations.

resembles the one of other positive-strand RNA viruses 19 , and its activity of RNA synthesis can be initiated de novo RNA or via a VPg-primed mechanism 20 . Due to its essential role in the viral replication, and to the repeatedly proven success of targeting viral polymerases in antiviral drug discovery 21 , norovirus RdRp has been previously chosen in our research group as a promising target for the identification of new anti-norovirus agents, focussing in particular on the identification of novel non-nucleoside inhibitors (NNIs) of this enzyme. A limited number of inhibitors of this type has been reported so far for norovirus RdRp, but the majority of these compounds lack any activity against the viral replication in cellular systems, possibly due to poor cell permeability and drug-like properties 22 . As several crystal structures are available for human and murine norovirus polymerase, including ternary complexes with nucleotide analogues and with allosteric non-nucleoside inhibitors [23][24][25][26][27][28] , the study of these structures has been the starting point for a structure-based virtual screening study that led to the identification of our broad-spectrum Caliciviridae RdRp inhibitor 1 (Fig. 1) 29 . As is the case for other reported NNIs of norovirus RdRp, despite showing an enzyme inhibition in the low micromolar range, 1 was associated with a very mild effect against norovirus replication in cell-based systems, possibly due to its poor aqueous solubility. Moreover, this compound showed some cytotoxicity at relatively low concentrations, with a CC 50 of ~64 μM, possibly due, at least in part, to its low solubility and precipitation from the assay medium.
In the present study, the structure of 1 has been rationally modified in order to improve its drug-like properties and achieve an antiviral effect against norovirus replication in cell-systems. The novel structural modifications carried out have allowed a better understanding of the functional groups required for enzymatic and antiviral activity, and the successful identification of a new anti-norovirus scaffold with antiviral EC 50 values in the low micromolar range. This new scaffold represents a promising starting point for further optimisations and for the potential development of a viable treatment for norovirus infections.

Results and Discussion
Rational modifications on compound 1. 1 is characterised by a central 5-phenylfuran-2-ylmethylene-pyrazolidine-3,5-dione core (planar central linker in Fig. 1), substituted at position 1 of the pyrazolidine with a benzene ring (terminal hydrophobic ring 1), and at position 4 of the phenyl ring with a N-phenylsulfonamide (terminal hydrophobic ring 2). These structural features render 1 relatively hydrophobic (calculated logP (o/w) 4.3) and poorly soluble, limiting its potential as a drug.
As described by Hashimoto et al. 30 , decreasing the logP of a molecule by introducing a hydrophilic group is a classical and general strategy for improving its aqueous solubility. Following this rationale, two main structural modifications have been planned for 1 to achieve this result: 1) the introduction of a carboxylic acid group at different positions of terminal hydrophobic ring 1 (compounds 2 and 3 in Fig. 1), which should also provide the potential for extra interactions with the RdRp active site; 2) the replacement of the terminal hydrophobic ring 2 with different heteroaromatics (compounds 4-6 in Fig. 1). In particular, this second approach was designed as a consequence of our previously published work 29 , in which a thiazole derivative of 1 showed a better antiviral

Chemistry
Synthesis of compounds with modifications on the two terminal hydrophobic rings. A seven-step synthetic pathway previously developed and optimised by our research group has been used for the preparation of a small family of seven new derivatives bearing modifications on the terminal hydrophobic ring 1 or on ring 2 29 .
As shown in Fig. 3, differently substituted phenylhydrazines (10)(11)(12)37) were reacted with ethyl-3-chloro-3-oxopropanoate 13 in THF and triethylamine (NEt 3 ), giving the desired hydrazides 14-16, 38. Ester-susbtituted hydrazines 10-11 were prepared in a good yield by converting the differently substituted hydrazineylbenzoic acids 8-9 into the corresponding ethyl esters via Fisher esterification using hydrochloric acid (HCl) and ethanol (EtOH). An alternative approach was required for the preparation of the ortho ethyl ester 37. In fact, under Fisher reaction conditions, an intramolecular reaction between the carboxylic acid and the hydrazine group occurs, leading to the formation of 3-indanzolinone. The desired ethyl 2-hydrazineylbenzoate 37 was obtained by reacting the ethyl 2-aminobenzoate with sodium nitrite (NaNO 2 ) in HCl, and then reducing the intermediate diazonium salt using tin chloride (SnCl 2 ). Hydrazides 14-16 were converted into the corresponding 1-arylpyrazolidine-3,5-diones 17-19 through an ester displacement reaction in the presence of sodium hydroxide (NaOH) and EtOH. Unfortunately, the synthesis of ortho substituted compound 39 could not be achieved, potentially due to steric hindrance that impedes the cyclization reaction. Treatment of 4-bromobenzene-1-sulfonyl chloride (24) with the appropriate aniline (20)(21)(22)(23) in pyridine produced sulfonamides 25-28, which were then converted into the aldehyde intermediates 30-33 by Suzuky coupling with (5-formylfuran-2-yl)boronic acid 29. In particular, for compound 30, bearing the original phenyl group, our formerly reported conditions using potassium phosphate (K 3 PO 4 ) as base, Pd(dppf) as catalyst, water/DMF as solvent and heating under microwave irradiation for 75 min at 130 °C 25 , gave the desired product, whereas no product could be obtained for derivatives 31-33 in these reaction conditions. After exploring various alternative procedures, the best reaction conditions were found using sodium carbonate (Na 2 CO 3 ) as base, Pd(OAc) 2 and PPh 3 as catalyst, water/DME as solvent, and heating the reaction mixture under microwave irradiation for 10 min at 85 °C, which gave the desired products in moderate yields (34-51%). The two portions were then linked together by reacting aldehydes 30-33 with the appropriate arylpyrazolidine-3,5-dione 17-19 according to a Knoevenagel condensation in acetic acid at 130 °C for 3 h, giving the desired products 5, 34-35. The derivatives showing a pyrimidine (4) and a 1H-benzoimidazole ring (6) appeared to degrade under these reaction conditions, therefore 4 was obtained using EtOH under reflux overnight, while 6 was prepared using methanol under reflux for 2 hours. In the last step, hydrolysis of the ethyl ester group was performed for 34 and 35. The para carboxylic acid derivative 2 was obtained by treating the starting ethyl ester with lithium hydroxide (LiOH) overnight at room temperature, while for the meta derivative 3 sodium hydroxide (NaOH) and dioxane were used. In fact, using LiOH for the hydrolysis of 35 led to the formation of a hydroxyl amine on the -NH group of the pyrazolidine-3,5-dione ring, with this hydroxylamine side product in a 50:50 ratio with the desired compound 3. The use of NaOH and dioxane limited significantly the side product formation.

Biological Evaluation and Molecular Modelling studies
Human norovirus RdRp activity inhibition. The newly synthesised compounds, including the carboxylate ester derivatives 34-35 and 7 from the Core Hopping approach, were evaluated for their enzymatic inhibition of human norovirus Sydney 2012 (HuNoV (GII.4)) RdRp activity in vitro, using a quantitative fluorescent assay 29 . Compounds were tested at six different concentrations (concentration range examined: 0.1-100 µM) and compared to the relative activity of mock treated samples containing the vehicle only (0.5% DMSO [vol/vol]). PPNDS [Pyridoxal-5′-phosphate-6-(2′-naphthylazo-6′-nitro-4′,8′-disulfonate) tetrasodium salt], a previously reported potent inhibitor of NoV RdRp activity 27 , and 1 were used as positive controls. Dose-dependent inhibitory response curves were used to establish IC 50 values for the eight test compounds, as reported in Table 1 (see also Supplementary Fig. S1). While IC 50 values could be attained for six of the eight compounds tested, 7 only reached 20% inhibition at the maximum concentration tested, while 4 reached 15% inhibition at the maximum concentration tested (Supplementary Fig. S2).
Compounds bearing modifications at different positions of the terminal hydrophobic ring 1 (2-3, 34-35) seem to retain the RdRp inhibitory activity with an IC 50 in the range of 13-22 µM, slightly higher in comparison with 1 (5.6 µM), but in the same order of magnitude. These results further prove the RdRp inhibitory activity associated with the general scaffold of 1 and suggest that introduction of hydrophilic groups into terminal hydrophobic ring 1 could improve aqueous solubility without disrupting RdRp inhibitory activity. A similar effect is obtained when replacing the terminal hydrophobic ring 2 with an oxazole (5) or benzoimidazole (6), with an IC 50 value of 12.8 and 26.7 µM, respectively. On the other hand, the insertion of a pyrimidine ring at this position (4) causes loss of activity, with only partial RdRp inhibition seen at 100 µM.
As mentioned above, 7 did not reach 50% inhibition at the test concentrations, showing only 20% RdRp inhibition at 100 µM. This result suggests that compound 1 planarity and rigidity might have an important role for its biochemical activity, as the flexibility conferred by the N,2-diphenylacetamide central core almost completely abolishes RdRp inhibition. This effect could be linked with a better overall occupation of the RdRp active site by the active inhibitors, as indicated by molecular docking results shown below.
To further confirm RdRp inhibition and exclude any potential false positives from the first fluorescence RdRp assay, the eight molecules were then assessed using a gel-shift enzyme activity assay at a fixed concentration of  100 μM, following a procedure previously adopted by our research group 29 . PPNDS was used as positive control and exhibited complete inhibition of transcription at 100 µM (no extension of 32 nucleotide RNA template (PE44-NoV) to 44 nucleotides in the presence of an active RdRp). Seven of the eight compounds exhibited complete or almost complete inhibition of norovirus RdRp activity at 100 μM, comparable to the effect observed for 1, further confirming their activity as pure polymerase inhibitors (Fig. 5a,b). On the contrary, 7 did not demonstrate inhibition of primed elongation activity in this assay (second higher band on gel shift almost identical to the 0.5% DMSO control, Fig. 5c), confirming that this linker modification abolishes RdRp inhibitory activity.
Molecular docking studies. In our previously published study 29 , both competitive binding studies with PPNDS and molecular docking simulations indicated that 1 likely occupies the central core of site-B and the initial part of site-A of the human norovirus RdRp (Fig. 6). Molecular docking evaluations of the new derivatives on the binding site of 1 in the structure of human norovirus RdRp were performed using Glide SP 33 . Compounds 4-6 are predicted to occupy the main nucleic acid binding site of HuNoV RdRp in a similar orientation compared to 1, with the N-phenyl/heteroaromatic-benzenesulfonamide portion pointing out from site-B towards the RdRp portion in which site-B and site-A overlap (area defined by Arg392), and the differently substituted phenyl-pyrazolidine portion occupying the area where PPNDS places its naphthylazo part (Fig. 7a-c), within the core area of site-B. The added carboxylic group appears to confer an extra interaction with Lys166, further stabilising the ligand-protein complex (Fig. 8a, compound 2). This interaction appears to be maintained also in the case of the ester analogues 34-35 (Fig. 8b, compound 34). An interesting observation can be made examining the predicted binding mode of 4. The presence of the pyrimidine ring seems to increase the attraction towards Arg182, pushing the compound away from the predicted binding site, potentially explaining its reduced enzymatic activity (Fig. 7a). This effect, which is also present at minor extent for 5, could be a consequence of a different electron/charge delocalisation on the terminal part of the molecule due to the different heteroaromatic rings. Interestingly, 7, due to the presence of the more flexible N,2-diphenylacetamide central core, assumes a series of conformations that do not allow an optimal occupation of the binding area, supporting the observed reduced inhibitory effect of RdRp activity ( Fig. 9).
Cell-based antiviral effect evaluation. The newly prepared compounds 2-7 and 34-35 were evaluated for their antiviral effect against the genogroup V mouse norovirus (MNV) using the mouse macrophage cell line RAW264.7. This assay evaluates the ability of the compounds to protect infected cells from the virus-induced cytopathic effect, thus being suitable to identify inhibition of virus replication at every step of the virus life cycle, and representing a robust cellular assay available for the rapid evaluation of potential norovirus inhibitors. Compounds identified using MNV have shown effects against HuNoV in vitro and in vivo, as in the case of nucleoside RdRp inhibitor 2′-C-methylcytidine (2CMC) [34][35][36] . Unfortunately, despite not presenting the aqueous solubility issues we had previously encountered for 1 29 , no significant antiviral effect could be observed in this assay for the new compounds up to the maximum concentration tested. The new structural modifications, despite retaining inhibition of the RdRp in an isolated enzyme activity assay and showing no cytotoxic effects at the test concentrations in the MNV antiviral assay (Supplementary Table S1), are correlated with a lack of antiviral activity. As these new structures present the same limitation of most non-nucleoside HuNoV polymerase inhibitors reported so far 22 , we were prompted to apply a different computer-aided strategy to further modify the scaffold of 1, with the aim to achieve an antiviral effect in cellular systems. prompting us to attempt further modifications on their structure.
Computer-aided flexible alignment approach: new structural modifications on the structure of 1.
Since both attempts at modifying the structure of 1 did not give the desired result of identifying antiviral hits against norovirus replication in cell-based systems, an alternative strategy to identify potential linker replacements for 1 was followed. A conformational search, using MOE conformational search tool 37 , was performed on 1 to identify its lowest energy conformation. This conformation was then kept rigid and used to run a flexible alignment analysis with the Flexible Alignment tool in MOE 37 , in order to search for potential three-dimensional and functional similarities with an in-house small-molecule database of previously reported non-nucleoside inhibitors of viral polymerases, including hepatitis C, Zika and Dengue virus RdRps [38][39][40] . In particular, the chosen database contains small-molecule compounds reported to have inhibitory activity against viral polymerases, and also to show antiviral effects in cell-based assays. From this study, www.nature.com/scientificreports www.nature.com/scientificreports/ -piperazinyl]phenyl}amino)carbonothioyl]-1-benzothiophene-2-carboxamide] (48), a recently identified inhibitor of Zika virus RdRp 39 , resulted as the best match for its structural overlapping with 1, as highlighted in Fig. 10.
Evaluation of inhibition of norovirus RdRp activity using an in vitro fluorescent, de novo RdRp activity assay at a fixed concentration of 20 μM revealed that 48, which we promptly synthesised in our laboratory, has a mild inhibition of human norovirus RdRp activity (20% inhibition at 20 μM), as reported in the biological evaluation studies section below. This preliminary encouraging result, together with the reported antiviral activity demonstrated for 48 in Vero cells against ZIKV replication 39 , prompted us to explore this scaffold in the search of new norovirus RdRp inhibitors, and most importantly new agents to inhibit HuNoV replication. The structure of 48 was therefore combined with some structural features from 1 and from some of our previously published norovirus RdRp inhibitors 29 , leading to the design of a series of new compounds with a non-planar central linker. These The added carboxylic acid group appears to confer an extra interaction with Lys166, further stabilising the compound 2-protein complex. This interaction appears to be maintained also in the case of the ester analogue 34. Moreover, the carboxylic acid group is placed at an optimal distance to interact with Arg419. The binding site area is represented as molecular surface. Compound 1 binding is reported as comparison (carbon atoms in lilac). www.nature.com/scientificreports www.nature.com/scientificreports/ new molecules could present improved solubility and better cell-based antiviral profiles in comparison with 1 and its analogues, as shown in Fig. 11. In addition, in order to verify the activity associated with the scaffold of 48, a few small structural modifications on its scaffold were also planned and carried out.

Chemistry
Synthesis of the newly designed "hybrid" molecules. The 12 newly designed analogues, along with the Flexible Alignment hit 48, were synthesised according to a four-step synthetic pathway, as shown in Fig. 12.
Briefly, commercial 1-(4-nitrophenyl)piperazine 61 was treated with the appropriately substituted aryl-carboxylic acids or aryl sulfonyl chlorides, to give intermediate nitro-compounds 62a-67a. In particular, amide-compounds 62a-65a were obtained through a TBTU-assisted coupling reaction in the presence of  www.nature.com/scientificreports www.nature.com/scientificreports/ DiPEA, in DMF at r.t., while sulfonamide intermediates 66a-67a were obtained in DCM at 0 °C to r.t., in the presence of NEt 3 . Nitro-intermediates 62a-67a were then converted into the corresponding substituted aromatic amines 62b-67b following a catalytic hydrogenation in EtOH, in the presence of wet activated palladium on carbon. Isothiocyanates intermediates 68b-72b were obtained by treating the corresponding aryl-carbonyl chlorides 68a-72a with ammonium isothiocyanate, in acetone under reflux conditions. These intermediates were not isolated, but treated in situ with the differently substituted aromatic amines 62b-67b, to give the desired final products 48-60. In most cases, formation of unwanted amide-byproducts was observed for this reaction, thus explaining the moderate yields obtained in the final step for most of these products. By-products 73-76 were formed in a particularly high amount in the course of the reaction, therefore they were isolated and fully characterised.

Biological Evaluation and Molecular Modelling Studies
Evaluation of the inhibitory activity of HuNoV RdRp for analogues 48-60. The newly prepared compounds 48-60 were initially evaluated against norovirus Sydney 2012 RdRp activity in vitro using a quantitative fluorescent assay. Compounds were tested at a fixed concentration of 20 µM and compared to the relative activity of mock treated samples. PPNDS and 1 were used as positive controls. Results of this assay are reported in Fig. 13. Removal of the chlorine atom from position 2 of the benzothiophene ring (49) or its replacement with a bigger trifluoromethyl group (50) appears not to affect much the RdRp inhibitory activity, with an observed inhibition which is similar to the effect of 48 (15-20% inhibition at 20 µM). A similar effect is obtained for 51 and 52 (~15% inhibition at 20 µM), in which the benzothiophene ring has been replaced with a phenyl ring and a 4-methylphenyl ring respectively. These two rings are structural features present in 1 and in one of its derivatives we have previously reported 25 . Replacement of the thiophene by a furan (53) or by an unsubstituted phenyl ring (54) increased the degree of HuNoV RdRp inhibition up to 60% at 20 µM, whereas introduction of a thiazole ring (56) does not affect enzymatic inhibition activity. From these observations, the thiophene ring appears not to influence inhibitory activity. Derivatives in which the central core of 48 has been used to link the two terminal hydrophobic rings of 1 or of its derivatives we have previously reported 25 (55, 57, 58) appear to retain the RdRp inhibitory activity found for 48 (~20% at 20 µM). A similar effect is obtained when a sulfonyl group, as in our hit 1, is used to link ring 2, regardless if it is a thiazole (59) or an unsubstituted phenyl ring (60), to the rest of the molecule. Interestingly, 73, one of the amide side products formed during the last synthetic step for this series of compounds, reduced the polymerase activity of ~40% at 20 µM. 53, the most active compound found in this series, also exhibits a significant inhibition of HuNoV RdRp activity at 100 µM in the gel-shift assay, confirming the polymerase inhibitory function of the new scaffold. Overall, even if showing a reduced biochemical activity in comparison with 1, some of the new derivatives can be considered as modest HuNoV RdRp inhibitors at 20 µM. This reduced activity in comparison with 1 could to some extent have been expected, since, as seen for 7, the reduction in central core planarity/rigidity in the new derivatives would likely reduce their interactions and  Cell-based antiviral effect evaluation. Since the main scope of this work was to find new norovirus inhibitors which show antiviral effects against both the HuNoV RdRp and in norovirus-infected cells, priority was given to the evaluation of most of the new molecules for their antiviral activity against MNV in infected RAW cells. The nucleoside 2′-C-methylcytidine was used as positive control. The test compounds were evaluated at eight different concentrations (range 0.6-100 µM) and their ability to reduce the virus-induced cytopathic effect was assessed.
Five out of 10 compounds tested were found to have an interesting antiviral EC 50 in the 20-100 µM range in this assay, as reported in Table 2. These new derivatives showed a slightly better EC 50 if compared with what we had previously found for 1 29 , but more interestingly they were not cytotoxic at the test concentrations, whereas 1 exhibited a mean half maximal cytotoxic concentration (CC 50 ) of 62.8 μM in a similar MNV cell-based assay 29 .
As a further step, the five hit compounds which showed activity in the MNV assay were tested against HuNoV using a replicon system, i.e. a human gastric tumor-1 (HGT-1) cell line which stably expresses the genome of a HuNoV GI.1 virus 41 . Given that in this system the gene encoding for the major capsid protein is replaced by a neomycin resistance gene, no new virus particles will be produced but the non-structural proteins are expressed and the replication of the genomic RNA can be studied 42 .
While 48 and 56 do not display any significant antiviral activity in this second assay, 52, 59 and 60 inhibit the viral replication of HuNoV GI with EC 50 values in the low micromolar range, as shown in Table 3. The lack of activity encountered for 48 and 56 in this assay could be a consequence of several factors, such as differences on the envelope and capsid proteins between MNV and HuNoV, which could affect the ability of the two compounds to reach the viral RdRp, or different readouts and sensitivity of the two assays in detecting the inhibitory activity of the compounds.
Even if showing an interesting EC 50 of 8 µM, 60 was found to form crystals starting from a 25 µM concentration, thus indicating that its solubility should still be improved. 52 and 59 display instead the most interesting antiviral activity profile found so far, with EC 50 values against HuNoV GI replication of 6 µM and 10 µM, respectively, providing a new antiviral scaffold for norovirus which shows promise for additional investigations. Overall, due to their low-micromolar inhibitory activity against norovirus replication in vitro, 52 and 59 represent one of the few successful examples of non-nucleoside HuNoV polymerase inhibitors which also show antiviral activity against the viral replication in a cell-based system. These compounds represent a promising starting point for further structural optimisation and SAR evaluations. In particular, current research efforts are now ongoing to further improve the drug-like properties of these compounds, and to increase their potency in both biochemical and cellular assays.

Conclusions and Future Work
Starting from a rational approach to modify the structure of one HuNoV RdRp inhibitor, 1, which was recently identified in our research group, we have explored different modifications on its scaffold, in order to increase its drug-like potential as an antiviral agent. Despite the fact that most of the newly modified compounds retain inhibitory activity on the viral polymerase, none of them resulted as an antiviral hit when tested in a MNV CPE reduction assay. In an attempt to further modify the scaffold of 1 and achieve an antiviral effect in cellular systems, we have performed a computer-aided approach in order to identify potential similarities between 1 and previously reported non-nucleoside inhibitors of viral polymerases, which also display good antiviral activities against the replication of different viruses in cell-based assays. Following this rationale, we have recognised significant structural similarities between our biochemical hit 1 and a recently reported inhibitor of Zika virus polymerase, 48. Rationally combining together different structural features of 1 and 48, we have designed and synthesised a new family of 12 modified analogues. Among them, different displayed at least partial inhibition of HuNoV RdRp activity in a quantitative fluorescent assay. Most interestingly, five of these compounds were found to reduce MNV-induced cytopathic effect with EC 50 values in the micromolar range, without being associated www.nature.com/scientificreports www.nature.com/scientificreports/ with significant cytotoxicity. These five hits were then evaluated in a human norovirus replicon assay harbouring a genogroup GI, and two of them, 52 and 59, were found to inhibit HuNoV replication with EC 50 values in the low micromolar range, revealing a new antiviral scaffold for human norovirus. These compounds represent one of the few examples of non-nucleoside polymerase inhibitors also showing significant antiviral effect against human norovirus replication in a cell-based system. Based on these findings, additional investigations on these new structures are currently ongoing to further explore their potential as antiviral agents. In particular, further structural optimisation is the main focus of present research efforts, which aim to identify suitable preclinical antiviral candidates for the treatment of norovirus infections.

Experimental
Chemistry. All solvents and reagents were used as obtained from commercial sources unless otherwise indicated. All solvents used for chromatography were HPLC grade from Fisher Scientific (UK). All reactions were performed under a nitrogen atmosphere 29 . 1 H and 13 C-NMR spectra were recorded with a Bruker Avance III HD spectrometer operating at 500 MHz for 1 H and 125 MHz for 13 C, with Me 4 Si as internal standard. Deuterated chloroform was used as the solvent for NMR experiments, unless otherwise stated. 1 H chemical shifts values (δ) are referenced to the residual non-deuterated components of the NMR solvents (δ = 7.26 ppm for CHCl 3 , etc.) 29 . The 13 C chemical shifts (δ) are referenced to CDCl 3 (central peak, δ = 77.0 ppm). TLC was performed on silica gel 60 F254 plastic sheets. Flash column chromatography was performed using silica an Isolera Biotage system. UPLC-MS analysis was conducted on a Waters UPLC system with both Diode Array detection and Electrospray (+′ve and −′ve ion) MS detection. The stationary phase was a Waters Acquity UPLC BEH C18 1.  Table 3. EC 50 values found for the six tested compounds the HuNoV GI replicon assay. A The mean values ± standard deviations of triplicate datasets are shown from at least three independent experiments.
The crude product was purified by automated flash column chromatography (Biotage Isolera One, SNAP KP Sil 10 g) eluting with n-hexane:EtOAc 100:0 v/v increasing to 0:100 v/v in 12 CV. Obtained in 63% yield as an off-white solid. 1
The Flexible Alignment was performed using MOE2018. 10. The MOE flexible alignment tool generates different possible conformations for each molecule present in the input database (in-house small-molecule database of reported non-nucleoside inhibitors of viral polymerases, previously prepared in MOE) that could overlap the conformation of the assigned template, which is kept rigid (compound 1). The quality of the alignment is evaluated by a score which is a sum of the internal strain of the obtained conformation (the smaller, the better) and the overlap of molecular features (aromatic regions, donors/acceptors). MOE, for each alignment performed, evaluates the average internal energy of the ligands U, the similarity score F (the lower value is, the better the two structures overlap) and the value S (sum of U and F values obtained for each alignment). A good alignment should present a dU value (the average strain energy of the molecules in the alignment in kcal/mol) lower than 1 kcal/mol meaning that the obtained conformations are not energetically disadvantaged. The obtained data from the Flexible Alignment with dU of 0.0 (no energy penalty) were kept and ranked according to the lowest S value. TPB (48) resulted the best match for its structural overlapping.
The conformation of 1 required for the Flexible Alignment was obtained using MOE conformational search tool, using the default parameters, selecting Stochastic as search method and setting RMS Gradient to 0.001 and the RMSD Limit to 0.15. 159 conformations were generated and among them the best in terms of dE (the strain of the conformation relative to the lowest energy conformation with the same stereochemistry configuration. Value of 0 indicate the best result) and E (value of the potential energy of the conformation) was chosen as template for the Flexible Alignment.
The Glide-Based Core Hopping tool of Maestro 32 was used to perform a scaffold replacement. The structure of compound 1 docked into the RdRp binding site was used as template ligand keeping as fixed the two terminal hydrophobic rings and let the program replacing the original scaffold with new optimal linkers. The new ligands are then automatically docked on the receptor and ranked and scored according to the docking results. The default parameters were used, using as docking grid the one previously generated for the SP docking studies (see above).