key: cord-0043783-wwphy054
authors: Pagadala, Nataraj Sekhar; Bhat, Rakesh; Kumar D, Jagadeesh; Landi, Abdolamir
title: Discovery of anti-influenza nucleoside triphosphates targeting the catalytic site of A/PR/8/34/H1N1 polymerase
date: 2020-05-24
journal: Med Chem Res
DOI: 10.1007/s00044-020-02561-0
sha: 079c584ccfa936df16a04090f490222f0cffe8ab
doc_id: 43783
cord_uid: wwphy054

In an effort to develop potent anti-influenza drugs that inhibit the activity of influenza virus RNA-dependent RNA polymerase (IAV RdRp), a database of nucleoside triphosphates with ~800 molecules were docked with the homology model of IAV RdRp from A/PR/8/34/H1N1 strain. Out of top 12 molecules that bind with higher affinities to the catalytic site of IAV RdRp above and below the PB1 priming loop, only seven molecules decreased the transcriptional activity of the viral RNA polymerase with an IC(50) in the range of 0.09–3.58 µM. Molecular docking combining with experimental study indicated that the molecules with linear chain are more effective in inhibiting IAV RdRp replication than the molecules with V-shaped and are cyclic in nature. A correlation between ΔG and LogIC(50) for these seven compounds resulted an R(2) value of 0.73. Overall, these newly developed seven nucleoside triphosphates lay a strong foundation for the future development of a new therapeutics that can satisfy the Lipinski’s rule of five exhibiting high specificity to the catalytic site of influenza-A viruses.

Influenza is a highly contagious airborne life-threatening viral infection causing recurrent outbreaks of humans responsible for respiratory diseases and death. It was estimated during 2017-2018 that there were 959,000 hospitalizations and 79,400 deaths (CDC 2018) alone in the United States related to influenza illness. Moreover, WHO estimates that 3-5 million people suffer from the disease every year with death rate ranging between 290,000 and 650,000 due to influenza epidemics (Influenza (Seasonal) 2018) worldwide (Sherman et al. 2019) . Of note, previous epidemiological studies on pandemic A/H1N1 during the year 2009 clearly showed that the severity of influenza was not so high in individuals effected with HIV compared with HIV-negative patients (Martinez et al. 2011; Perez et al. 2010) . This clearly shows that influenza remains a major target for vaccine and antiviral treatment and prophylaxis. One of the causes of the severity of the disease is due to influenza replication and transcription played by viral RNAdependent RNA polymerase (RdRp) which is composed of three subunits i.e., polymerase basic1 (PB1), polymerase basic 2 (PB2), and polymerase acidic (PA) (Fodor 2013; Resa-Infante et al. 2011) . Both PB1 and the N-terminus of PB2 form a large central cavity for catalytic mechanism of influenza-A virus (IAV) viral replication (Hengrung et al. 2015; Pflug et al. 2014; Reich et al. 2014) . The priming loop of (a β-hairpin structure) PB1 thumb domain protrudes into the central cavity and supports the sugar base of the first NTP during de nova initiation Appleby et al. 2015; Butcher et al. 2001; Tao et al. 2002) . Moreover, primer-independent replication on the viral RNA (vRNA) template was also initiated by the priming loop (Te Velthuis et al. 2016) . However, no drug exits presently in the market that blocks the entry of putative NTP and inhibits the initiation of vRNA priming loop occupying to the catalytic site of IAV RdRp. Apart from the catalytic site, IAV RdRp contains multiple sites for potential antivirus drug development due to its high conservation among different strains (Babar et al. 2014) .

Recently, four different active sites of IAV RdRp were predicted computationally (Pagadala 2019) . At present, three major classes of drugs are under development for influenza: RNA synthesis inhibitors, cap-snatching inhibitors, and inhibitors targeting protein-protein interactions between the three polymerase subunits (e.g., PA-PB1, PB1-PB2). Inhibition of IAV RNA synthesis during chain elongation is mainly inhibited by nucleoside mimicking's such as favipiravir (T-705), 2′-deoxy-2′-fluoroguanosine (Furuta et al. 2009; Tisdale et al. 1995) . On the other hand, the pyrazine carboxamide derivatives (T-1105 and T-1106) show different antiviral activity on the type of host cells used. Both ribavirin and its amidine prodrug viramidine are converted to ribavirin monophosphate and acts on cellular enzyme inosine 5′-monophosphate dehydrogenase leading to the inhibition of vRNA synthesis (Sidwell et al. 2005; Ghanem et al. 2007 ). In comparison with ribavirin, cytotoxicity of viramidine is lesser which lead to further evaluation as a possible anti-influenza viral therapy. Unlike ribavirin, the influence of T-705 on cellular DNA or RNA synthesis was negative, indicating its high specificity and lesser cytotoxicity. Recently, a set of compounds that were screened targeting IAV RNA synthesis revealed that 2′F-2′ dNTP nucleoside analogs inhibit cPol and vPol in a completely different way (Reich et al. 2017) . A combination therapy of amantadine and oseltamivir as well as favirpiravir and oseltamivir in H5N1 infected mice has shown increased in drug efficacy (Ilyushina et al. 2007 ). Although the combination of oseltamivir and ribavirin was shown effective, ribavirin has shown some undesirable toxic effects. Recently, combination therapy with three antiviral drugs (amantadine, ribavirin, and oseltamivir) has shown a strong effect on drug-resistant viruses including pH1N1 influenza in Korea (Kim et al. 2011) . In continuation, the TCDA therapy also showed similar pharmacokinetics compared with mono-drug therapy with increased safety in immune compromised patients (Kim et al. 2011; Seo et al. 2013) .

In continuation of the previous publication of known molecules of anti-influenza nucleoside analogs that were approved for the treatment of influenza viral infection and their mechanism of binding, in the present study, (Sherman et al. 2019 ) the compound database of~800 nucleoside triphosphates that are commercially available was developed for molecular docking against the built homology model in three stages of replication mechanism I. In the presence of influenza-A vRNA; II. In the presence of Influenza-A vRNA with extended primer; and III. In the presence of influenza-B vRNA (Pagadala 2019) . Based on the docking studies, the top ranking 11 nucleoside triphosphates that bind commonly to the catalytic site in all the three stages with higher affinities and are predicted to inhibit the IAV viral replication were chosen and their ligand protein interactions were studied further. In addition, the binding modes of 11 NTPs against IAV RdRp were predicted and their dihedral angles Φ T C (2′) -C (1′) -N (1) -N (5)/C (6) and Φ P O (5′) -C (1′) -N (1) -N (5)/C (6) that play an important role in ligand binding were analyzed. Furthermore, 25% (IC 25 ), 50% (IC 50 ), and 95% (EC 95 ) effective concentrations of these 11 NTPs were calculated for IAV RdRp inhibition using trichloroacetic acid (TCA) precipitation assay (Pagadala 2019) . Finally, LogIC 50 of the molecules that showed inhibition was correlated with ΔG and their regression analysis R 2 was calculated.

The database of~800 nucleoside triphosphates developed from different chemical companies were used for molecular docking using the same methodology that was published earlier with the homology model of A/PR/8/34/H1N1 polymerase in three different stages of replication mechanism; I: In the presence of influenza-A vRNA, II. In the presence of Influenza-A vRNA with extended primer, and III. In the presence of influenza-B vRNA using MOE (Chemical computing group, Montreal, Canada) (Pagadala 2019) . Initially, the ligand was optimized by energy minimization with TAFF forcefield until the energy gradient is below 0.001 Kcal/Mol (Clark et al. 1989) . Totally 1000 docking poses for each ligand were generated using implicit generalized born solvation model and ligand placement method "Alpha PMI" with London dG scoring function. The initial and final scoring of the top docking poses were refined using the same London dG scoring function calculated with TAFF forcefield (Mackerell et al. 2004) . At the end, global minimum of the ligand was predicted using the simulated annealing based on the Monte Carlo method (Metropolis et al. 1953) . Finally, the affinity of the docking poses as the sum of the electrostatic, and Van der Waals energies were ranked with ΔG (U total in kcal/Mol). The reaction model cutoff between 8 and 10 Å as a dielectric function with the pocket radius of 6 Å was included in our docking studies. In total, the 30 docking conformations were retained as a cutoff.

We then investigated the effect of top ranking 11 compounds with higher MW and TPSA and bind to the catalytic site of RdRp inhibiting the incorporation of GTP into TCA-precipitable material with the A/PR/8/34 (PR8) strain (Table 1 and Fig. 1 ). The purified influenza virus A/PR/8/ 34/H1N1 obtained from Charles River (MD) was disrupted with 2.5% Triton N-101 and diluted 1:2 with 0.25% Triton N-101. Disruption provided the source of influenza ribonucleoprotein (RNP) containing the IAV RdRp and template vRNA. Samples were stored on ice until use in the assay. Six serial half-log dilutions of the test drugs (twelve drugs) at concentrations of 0.001, 0.01, 0.1, 1, 10, and 100 μM and positive control 2′Deoxy-2′fluoroguanosine 5′triphosphate (high tests of 100 µM) were tested in triplicate. Each polymerase reaction contained the following: disrupted RNP, Tris-HCl, KCl, MgCl 2 , Dithiothreitol, 0.25% Triton N-101, [α-32 P] GTP, ATP, CTP, UTP, GTP, and Adenyl (3′-5′) Guanosine. For testing the inhibitor, the reactions contained the inhibitor and the same was done for reactions containing the positive control. The reaction was incubated at 30°C for 1 h, transferred onto glass-fiber filter plates and subsequent precipitation of nucleic acids with 10% TCA. The filters were then washed with 5% TCA followed by 95% ethanol and air dried. Once the filter has dried, incorporation of [α-32 P] GTP was measured using a scintillation counter (Micro beta). Radioactivity was measured in a Micro beta liquid scintillation counter. Negative control reactions were prepared by omitting RNP complexes, whereas, positive control reactions for polymerase inhibitors contained the specific inhibitor 2′-deoxy-2′-Fluoroguanosine 5′-triphosphate. Inhibition of viral polymerase activity was measured by IC 25 , IC 50 , and IC 95 in triplicate (Pagadala 2019) .

Binding modes of 11 NTPs against the homology model of human A/PR/8/34/H1N1 RdRp

In the docked complex, compound 1 binds to the catalytic site with both ribose sugar and cytidine nitrogen interacting with Ser 494 and Ala 648 of the priming loop through 2′ and backbone oxygens. The central 1H-indole ring and sulfonyl oxygen atoms contacts with Asn 42 and Leu 45 of PB2-N involved in interactions with the 3′activator strand of vRNA once it enters the polymerase active site. However, the terminal 1H-indole ring contacts with Pro 641 of PB1 priming loop through ethyl carbon. Also, the α and γ phosphates contact with side chains of Asn 536 and Arg 142 at the end of the second tunnel (Fig. 2a) . Due to its bigger size, compound 2, occupies more space towards the catalytic site showing interactions between cyanine and Met 646 of the priming loop. In addition, 7-Allyl also shows contact with Met 646 and 507 while the ribose sugar towards the exit channel makes contacts with Asn 504 , Glu 508 , and Asn 537 of PB1 in the central region of a typical right-handed RdRp fold. The central sulfonyl at the C-terminal region contacts with Asn 42 , Pro 43 , and Ala 44 of PB2-N through backbone atoms (Fig. 2b ). In comparison, the terminal 1H-indole ring of compound 3 shows H-Pi stack with Glu 232 and Arg 233 near the NTP tunnel placing the sulfonyl oxygens close to Mg ++ . In addition, the central 1H-indole ring and the attached sulfonyl oxygen also interacts with Met 646 and Ser 511 of PB1 central domain through H-Pi and hydrogen bond. This orientation is further stabilized by additional contacts with Met 646 and Ala 648 of the priming loop through the central 5-5-propargylamino carbon and oxygen atoms. Other than these interactions, the ribose sugar, α and γ phosphates also interacts with Thr 493 , Asn 537 , and Asn 276 of PB1 central domain (Fig. 2c) .

Furthermore, it was demonstrated that compound 4 occupies the binding pocket at the end of tunnel 2 with -phosphate and oxymethyl of ribose sugar makes contacts with PB2 linker Asn 138 and Arg 142 that leads to an Nterminal subdomain. The cytidine which is placed far away from the PB2 linker shows contacts with Val 502 of PB1 central domain. This allows 6-aminoally aliphatic chain to enter deeply into the pocket forming a U-shape, exposing desthiobiotin at the other end of the molecule to interact with Lys 670 of PB2 627 linker (Fig. 2d) . When desthiobiotin-6-Aminoallyl was replaced with N 4 -biotin-OBEA, compound 5 penetrates much deeper than compound 4 in the same binding site. The sulfur atom of biotin contacts with Val 139 , Leu 668 , and Lys 670 of PB2-N subdomain and PB2 627 linkers, while the oxo group interact with Phe 501 of PB1 central domain. In addition, the OBEA aliphatic chain also shows contacts with Val 502 , Ala 503 , and Asn 536 through backbone carbon, and oxygen atoms. Moreover, the cytidine and ribose sugar along with α, β, and γ phosphates shows contacts with Ser 494 , Met 507 , Ala 648 , Gly 275 , Asn 276 , and Asn 536 of PB1 central and priming loop forming more or less a circular fold. This shows that compound 5 binds with a greater number of contacts than compound 4 in the same binding site influencing both PB1 priming loop and PB2 linker domain involved in RNA catalysis (Fig. 2e) .

Along with the above, compound 6 also binds to the same binding site that was identified previously with compounds 4 and 5 with one end of adenosine moiety shows Pi stack with Gly 275 and the adenosine at the other end orients towards the PB2 627 linker domain without showing any contacts with IAV RdRp. The terminal oxygens of both the α and γ phosphates show interactions with Lys 279 and Ala 503 of PB1 central domain and Thr 238 of PB2-N-terminal domain keeping both the domains intact. On the other hand, the ribose sugars of both the sides also show contacts with Asn 276 , Asn 536 , and Phe 501 of PB1 central domain through carbon, and oxygen atoms (Fig. 2f) . However, the benzyl ring of compound 7 shows H-Pi stacking with Met 507 with sulfonyl oxygen interacting with Phe 495 of the PB1 central domain. Also, the ribose sugar interacts with Lys 279 and Val 502 of PB1 central domain through 2′ oxygen and carbon. In addition, the 5′-dimethoxytrityl group also contacts with Val 502 through phenyl carbon. In comparison, the adenosine ring forms H-Pi interactions with Asn 137 of PB2 near the entry channel. In addition, the adenosine and ribose rings also to show contacts with Asp 671 and Gly 669 of PB2 linker domain through backbone and methyl carbons. Moreover, the attached α and γ triphosphates also orient towards the linker domain showing contacts with Gly 500 of PB1 central domain and Lys 670 of PB2 627 linker through terminal oxygens (Fig. 2g) . However, both the anthraniloyl rings of compound 8 in 2′ and 3′ shows contacts with Val 502 Arg 142 both PB1 and PB2 domains through terminal chlorine. Other than these contacts, oxo group of 2′ ribose sugars also interact with Thr 238 of PB2 through a backbone carbon (Fig. 2h) .

When compound 6 is replaced with guanosyl moiety limiting the number of phosphates to four allowing the compound 9 to occupy both the tunnels with a priming loop in between. The 5′ adenosyl group is placed in tunnel 1 orienting towards the C-terminal domain of PB2 interacting with Asn 42 , Ala 44 , and Leu 45 while 5′ guanosyl placed in tunnel 2 shows contacts with Ser 494 near to the exit channel. In addition, the α and β phosphates also contact with Ala 648 of the priming loop through terminal oxygens (Fig. 2i ). The propionamide-N-allyl of compound 10 shows interactions with Arg 233 , Met 507 , and Met 646 of the PB1 priming loop through oxo, nitro, and carbon atoms, leaving the indole ring to move freely in the catalytic site of IAV RdRp. Both the ribose sugar and α, β, and γ phosphates in tunnel 2 interacts with Gly 275 and Phe 495 , Lys 279 , Ala 503 , Asn 537 , and Phe 501 of PB1 through 3′ and terminal oxygens ( (Fig. 2k) .

The bind modes of compounds 1-3 show that cyanine moiety is placed below the priming loop perpendicular to the tunnel 1 with the rest of the chain appears in parallel to the tunnel 2 almost showing a V-shape. This reveals that compounds 1 and 3 with cytidine moiety are more effective in binding to IAV RdRp than compound with uridine moiety. These compounds follow the order in binding to IAV RdRp, compound 1 > compound 3 > compound 2. In comparison, the binding orientation of compounds 4-6, 9, and compounds 10-11 shows linearity in the catalytic site interacting with PB1 priming loop. In case of compounds 4 and 5, both occupy tunnel 2 with biotin orienting towards the exit channel. However, compound 4 is close to the PB2 linker domain while the same moiety stays nearer to PB1 priming loop. Surprisingly, compound 6 with Penta phosphate completely occupies tunnel 2 with respect to compound 9 having tetra phosphate which occupies tunnel 2 partially. This may be due to the presence of guanosyl at one end in compound 10 instead of adenosyl at both the ends in compound 6. Furthermore, comparison of compounds 10 and 11 clearly indicates that 2′-deoxyuridine-5′triphosphate of both these compounds in tunnel 2 are placed quite opposite to each other with 5-[(3-Indolyl) propionamide-N-allyl] of compound 10 is placed close to PB1 3a) . Due to its bigger size, compound 2 occupies more space in the tunnel 2 just above the priming loop of PB1. The oxygen atoms of uridine and ribose sugar shows contact with Phe 501 and Asn 536 of PB1 and Lys 140 of PB2 at the end of the tunnel 2. In addition, the 7-Allyl group also shows contact with Asn 536 and Asn 276 of PB1 catalytic site above the priming loop. However, triphosphate orients towards the exit channel and makes contact with the positive changed lys 140 of PB2-N and Lys 670 of PB2 627 domains. Both the sulfonyl groups attached to 1H-indole ring contacts with Ser 494 of PB1 central, Ala 648 and His 649 of PB1 priming loop and Arg 216 of PB2-N laying almost in parallel to the priming loop (Fig. 3b) . The cyanine group of compound 3 also lies in almost parallel to the priming loop of PB1 with a sulfonyl group attached at the two ends of the indole interacts with Met 507 and Asn 536 of PB1 central and Ala 648 of the PB1 priming loop. The 5-5-propargylamino and 2′-deoxycytidine interacts with Asn 536 and Phe 496 while the terminal phosphates interaction with Lys 279 , Val 502 , and Val 503 of PB1 central domain at the end of the tunnel 2 (Fig.  3c) . The γ-phosphate of compound 4 penetrates deeper at the end of the tunnel 2 showing contacts with Asn 16 and Ser 594 of both PB1 central and PA-C through terminal oxygens. Both cytidine and desthiobiotin rings including the aminoallyl contacts with Gly 500 of PB1 central, Gln 138 and Arg 142 of PB2-N through oxo and nitro groups (Fig.  3d) . Replacing desthiobiotin-6-Aminoallyl with N4-Biotin-OBEA allows compound 5 to also bind at the end of the tunnel 2 forming a circular structure. The terminal biotin contacts with Lys 279 and Lys 140 of both PB1 Central and PB2-N, while the cytidine interacts with Asn 16 of PB1-N through oxo and nitrogen atoms. In addition, the β and γ triphosphate interacts with Ile 535 , Asp 538 of PB1 central, and Thr 238 of PB2-N through terminal oxygens (Fig. 3e) .

In comparison with compounds 4 and 5, compound 6 with two ends of adenosine interaction with Met 507 of PB1 central and Met 645 of PB2 627 domains. The attached furanosyls also shows contacts with Asn 537 and Gly 500 of PB1 domains through 2′ and 3′ oxygens. In addition, the α, β, and ε triphosphates in-between shows contact with Asn 276 , Asn 536 , and Pro 219 of PB1 central and PB2-N with terminal oxygens (Fig. 3f) . However, the sulfonyl group of compound 7 is placed just below to the tip of the priming loop interacting with Gly 650 of PB1 priming loop. The 5′-dimethoxytrityl group also makes contact with Arg 126 through H-Pi stacking along with Met 646 of PB1 and Ly 41 of PB2-N near to the entry channel of tunnel 1. The N2-isobutyrylguanosine protrudes towards NTP tunnel showing no contacts with the viral enzyme (Fig. 3g) . In comparison, the compound 8 with 5-chloroanthraniloyl groups attached to 2′ and 3′ positions of ribose sugar make H-Pi interactions with Lys 140 and Arg 142 of PB2-N at the end of the second tunnel. In addition, the adenosine and triphosphate groups attached to 1′ and 4′ positions also show contacts with Gln 138 and Thr 238 of PB2-N far away from the linker domain through H-Pi and weak hydrogen bonds (Fig. 3h) . However, the two ends of adenosine and guanosyl moieties of compound 9 show H-Pi stack with Met 507 of PB1 central, Thr 238 of PB2-N, and Lys 670 of PB2 627 domains, while the attached furanosyls shows contact with Val 502 , Asn 504 , Met 507 of PB1, and Gly 669 of PB2 627 domains, respectively. In addition, the α and γ phosphates also show contact with Lys 279 , Asn 504 , Asn 536 , and Asp 538 of PB1 central through terminal oxygens. In addition, adenine also interacts with Ala 648 of the priming loop through the amino group (Fig.  3i) . The indole and uridine groups of compound 10 forms H-Pi stacking with Ala 139 and Arg 126 of PB1 central. The propionamide-N-allyl shows contacts with the backbone atoms of Gly 125 and Gln 39 of both PB1 central and PB2-N through backbone nitrogen. In addition, both the sugar, and the base which occupies in tunnel 1 contacts with Lys 229 , Asp 230 , and Ala 231 of PB1 central and Lys 41 of PB2-N with both stronger and weaker hydrogen bonds. The attached 5′triphosphate also shows contacts with Ala 231 of PB1 central and with Lys 41 and Asn 42 of PB2-N domain (Fig. 2j) . Furthermore, compound 11 also binds similar to compound 10 in perpendicular to the catalytic site. Biotin-16-5-aminoallyl interacts with Leu 122 , Thr 128 and Ala 139 and Met 646 of PB1 central while 2′-deoxyuridine shows H-Pi stacking with Met 645 of the priming loop. This binding clearly shows that Biotin-16-5-aminoallyl orient away from the catalytic site with triphosphate projecting into the entry channel (Fig. 3k) .

The binding orientations of these compounds reveal that compound 1 occupy tunnel 1 while compounds 2 and 3 Fig. 3 Ligand protein interaction of top 11 compounds represented in codes 3-AA-CTP (a), 7-AA-CTP (b), cyanine-5-dCTP (c), Desthiobion-6-dCTP (d), N4-Biotin-dCTP (e) Penta phosphate (f), 5′-DMT-2′-Ts-N2-Biotin-Ibu-G (g), Bis-Cl-ANT-ATP (h), Tetra phosphate (i), 5′-Indoyl-AA-dUTP (j), and Biotin-16-dUTP (k) with the homology model of IAV RdRp from A/PR/8/34/H1N1 strain in the presence of Flu A Primer occupied tunnel 2 above the priming loop with cyanine moiety placing in perpendicular to the tunnels 1 and 2 forming a V-shape. This proves that compound with cytidine base is more effective against polymerase inhibition than compound with uridine base. The order of compounds inhibiting RdRp is as follows, compound 3 > compound 2 > compound 1. Surprisingly, compound 4 also forms a Vshape while compound 5 shows circular form binding close to the PB2 linker domain with triphosphates is on the opposite sides. In comparison, the binding orientation of compounds 6 and 9 shows linearity above the catalytic site in the tunnel 2 interacting with a priming loop of PB1. The other linear shape compounds 10 and 11 binds away from tunnel 1 with 2′-deoxyuridine-5′-triphosphate placing in tunnel 1 close to NTP tunnel. However, compound 8 mostly occupies tunnel 2 near to the PB2 linker domain. Compared with dicyclic group of compound 8, a tricyclic group of compound 7 binds in tunnel 1 below the PB1 priming loop with 5′-dimethoxytrityl orients towards the 3′end of the primer and 2′-tosyl-N2-isobutyryl-guanosine towards motif-B involved in stabilizing the base pair between the incoming NTP and the primer template.

These docking studies also show that the residues Arg 126 , Lys 279 , Gly 410 , Phe 495 , Phe 501 , Val 502 , Asn 532 , Asn 536 , Met 646 , and Ala 652 from PB1; Asn 41 , Asn 137 , Lys 140 , and Arg 142 from PB2 act as common pharmacophores of these compounds preventing the activity of A/PR/8/34/H1N1 influenza-A polymerase in the presence of capped primer.

When influenza-A vRNA was replaced with influenza-B vRNA, both the nitrogenous base and the ribose sugar of compound 1 interacts with Asn 504 and Asn 537 through C 60 and O 91 atoms. The 5′ triphosphate contacts with Asn 276 and Lys 279 at the exit channel of the RdRp through terminal oxygens. The 1H-indole interacts with positively charged Lys 576 and Arg 583 of PB1 central domain through sulfonyl oxygens 124,125&133 . In addition, the cytidine ring also forms, contacts with Arg 233 and Ala 648 at the NTP tunnel and PB1 priming loop though terminal oxygen and amino nitrogen (Fig. 4a) . Also, the uridine base O 100 in compound 2 shows contact with Ala 648 at the tip of the priming loop while the 3′carbon of ribose sugar along with triphosphate forms a Uturn in tunnel 2 interacting with Gly 275 and Asn 537 of PB1 central. The terminal 1H-indole ring points towards the 3′ activation strand of the vRNA interacting with Asn 42 and Pro 43 of PB2-N through sulfonyl oxygens. In addition, the 7-Aminoally also shows contacts with Pro 647 and Met 646 of the PB1 priming loop (Fig. 4b) . The compound 3 cyanine lies below the PB1 priming loop with central 1H-indole ring interacting with Met 645 and Met 646 of PB1 priming loop while the terminal IH-indole ring contacts with positively charged Ly 235 and Arg 239 of PB2 222-246 loop through sulfonyl oxygens. The 5-5-propargylamino lies perpendicular to the priming loop showing contacts with Met 507 , Glu 508 , and Met 645 of PB1 central domain. This orientation of 2′-deoxycytidine-5′-triphosphate towards the exit channel allows to interact with Gly 275 , Asn 276 , and Asn 536 of PB1 and Arg 142 of PB2 central (Fig. 4c) . The α and β phosphates of compound 4 contacts with Gly 275 and Asn 537 and the attached ribofuranosyl interacts with Val 502 and Asn 536 of PB1 central. Moreover, both the oxo and nitro group of cytidine base also shows contacts with Phe 501 and Asn 536 of PB1 central and Pro 219 of PB2-N2 at the end of the tunnel 2 through H-Pi interaction. In addition, the desthiobiotin also shows interaction with Lys 140 of the PB2-N through oxo group (Fig. 4d) . The compound 5 also binds at the end of the tunnel 2 with the terminal ends of biotin and cytidine contacts with Lys 279 of PB2-N and Ala 643 of the priming loop through oxo groups. In addition, cytidine and furan contacts with Phe 505 , Glu 505 , and Ala 648 through nitrogen and 2′ carbon while the α and γ triphosphate interacts with Gly 275 and Asn 537 through terminal oxygens (Fig. 4e) .

In comparison with compounds 4 and 5, compound 6 occupies both tunnel 1 and tunnel 2 almost in parallel to the catalytic site of RdRp. One end of adenosine moiety contacts along with the attached ribose sugar shows interactions with Val 644 , Met 645 , and Met 646 of PB1 priming loop while the other end of adenosine orients towards the exit channel contacting with Ala 503 of the PB1 central. The phosphate groups in between also interacting with Glu 508 of the PB1 along with Pro 647 and Ala 648 of the priming loop (Fig. 4f) . However, compound 7 is placed below the exit channel with 2′-tosyl showing contacts with Ala 503 and Asn 504 and the ribose sugar interacting with Lys 279 and Phe 495 with another two hydrogen bonds. The 5′-dimethoxytrityl group also makes contact with Ile 535 and Asn 536 below the exit channel Fig. 4g . Similarly, the 5-chloroanthraniloyl makes contact with Asn 536 of PB1 below the exit channel of RdRp. In addition, the adenosine and triphosphate groups also show contacts with Ser 601 of PA-C, Gly 500 of PB1, and Gly 669 of PB2 627 linker (Fig. 4h ). In comparison with compound 6, the compound 9 with adenosyl at one end orients towards the priming loop of PB1 while the other end with guanosyl orients very close to Leu 668 of PB2 linker domain. The 2′ribose sugar of guanosyl shows contacts with Arg 142 of PB2-N2 subdomain, while the 2′ribose sugar of adenosyl contacts with Gly 275 of PB1 at the exit channel. In addition, the phosphate groups also show contacts with Lys 279 , Val 502 , and Ala 503 of PB1 central fold through terminal oxygens (Fig. 4i) . The indole and propionamide-Nallyl of 5- propionamide-N-allyl] of compound 10 contacts with Glu 508 and Thr 493 of PB1 central while uridine base interacts with Ala 648 of the priming loop. In addition, the 2′-deoxyuridine which occupies in tunnel 1 shows contacts with Pro 510 of PB1 central and the attached 5′-triphosphate shows contacts with Ser 515 of PB1, Pro 43 , Ala 44 , and Asn 42 of PB2_C domain (Fig. 4j) . However, compound 11 interacts with Lys 279 , Phe 501 , Glu 508 , and Gly 125 of PB1 and Met 646 of the priming loop while 2′-deoxyuridine-5′-triphosphate shows contacts with Arg 233 , Gly 513 , and Ser 515 of PB1 and Asn 42 of PB2 at the Cterminal end of RdRP (Fig. 4k) .

The predicted V-shape orientations for compounds 1, 2, and 3 shows that cyanine moiety is placed almost in parallel to the PB1 priming loop while the rest of the molecule projects towards the exit channel in parallel to the catalytic site of RdRp polymerase. This reveals that compounds 1 and 3 with cytidine moiety binds with higher affinity compared with compound with uridine moiety. These compounds inhibiting RdRp are as follows, compound 1 > compound 3 > compound 2. The orientation of binding shows that desthiobiotin-6-aminoallyl appears in parallel to the catalytic site of RdRp in tunnel 2 with triphosphate orientation towards the priming loop while the orientation of compound 5 appears in circular with triphosphate pointing towards the exit channel. In comparison, the binding orientations of compounds 6 and 9 shows linearity above and below the catalytic site. In these orientations, it was carefully observed that the phosphate groups of compound 6 primarily interact with the priming loop which is not seen with compound 9. The other linear shape compounds 10 and 11 binds below the priming loop occupying both tunnels 1 and 2 with triphosphates towards the entry channel. However, compounds 7 and 8 mostly occupy tunnel 2 nearer to the PB2 linker domain. The predicted orientation shows that bulky groups of 5′-dimethoxytrityl and 5-chloroanthraniloyl orients towards the catalytic site while compound 7 (isobutyryl-guanosine) and compound 8 (adenosine-5′-O-triphosphate) orients towards the exit channel and the linker domain. These two binding conformations clearly indicate that compound 8 is far away from the catalytic site and PB2 linker domain compared with compound 7 which is close to the catalytic site interacting with PB1 priming loop.

The docking results of these compounds show that the residues Arg 125 , Glu 126 , Ala 651 , and Ser 652 from PA; Thr 39 , Ala 231 , Ala 503 , Pro 510 , and Met 646 from PB1 and Arg 17 , Glu 18 , Gln 39 , and Lys 41 from PB2 act as common pharmacophores of these compounds preventing the activity of A/PR/8/34/H1N1 influenza-A polymerase after the translocation of 3′end of the vRNA into the catalytic site.

The activity of the compounds against IAV replication using TCA precipitation assay Further experimental studies using TCA precipitation assay showed that compound 1 inhibited polymerase activity with an IC 25 an IC 50 of 0.03 and 1.69 μM with higher rates of inhibition between 0.1 and 1 µM compared with other concentrations used in the study (Fig. 5a) . Compounds 2 (Figs. 5b) and 3 (Fig. 5c) weakly inhibited polymerase activities, having an IC 25 and IC 50 of 0.8 and 3.58 and 0.6 and 2.13 μM, whereas compounds 4 (Fig. 5d) and 8 (Fig.  5h) had a significant effect on Flu A polymerase activity with an IC 25 and IC 50 of 0.06 and 0.21 and 0.27 and 0.68 μM, respectively. However, compounds 5 (Fig. 5e) and 6 ( Fig. 5f ) had a very large effect on polymerase inhibition with an IC 25 and IC 50 of 0.02 and 0.09 and 0.03 and 0.09 μM. Thus, compounds 5 and 6 displays broadspectrum antiviral activity against Flu A virus with IC 50 at the same concentration of 0.09 µM. 2′-deoxy-2′-fluoroguanosine 5′-triphosphate, a known inhibitor of Flu A polymerase exhibited a 50% (IC 50 ) effective concentration at 1.2 μM. For compounds 2 and 3, the higher rate of polymerase inhibition was seen between 0.1 and 1 µM while for the compounds 4, 5, and 7 the higher rate of polymerase inhibition was seen at the lesser concentration between 0.1 and 1 µM. However, for compound 6, the higher rate of polymerase inhibition was seen completely at a very low concentration between 0.01 and 0.1 µM compared with all other compounds used in the study. Moreover, all these eight NTPs show 95% (EC 95 ) of viral polymerase inhibition greater than 100 µM. However, no significant effect of compounds 8-11 (Fig. 5i-k) was observed against the activity of the influenza polymerase even at higher concentration of 100 μM.

A regression analysis of logIC 50 with the docked scores, ΔG for the compounds 1-7 including the positive control 2′deoxy-2′-fluoroguanosine 5′-triphosphate in the presence of influenza-A vRNA (Stage-I) ( Table 2) , influenza-A vRNA with extended primer (Stage-II) ( Table 3) and influenza-B vRNA (Stage-III) (Table 4 ), were carried out and their scatter plots was drawn. It was found that these compounds showed a positive correlation with an R 2 value of 0.21, and 0.20 in Stage-I and Stage-III and a negative correlation of 0.001 in Stage-II. However, the best correlation was obtained with an R 2 value of 0.73 and 0.57 (Fig. 6) when ΔG for the positive control 2′-deoxy-2′-fluoroguanosine 5′triphosphate were removed from Stage-I and Stage-III. However, no change in R 2 value was seen in Stage-II even after removing the ΔG for the positive control.

Here we report the identification of 11 NTPs that binds to the homology model of A/PR/8/34/H1N1 influenza virus RNA polymerase with higher affinities. The docking studies in the Stage-I and III confirm that all these linear and Vshaped compounds 1-3 except compound 11 with influenza-A vRNA commonly binds to the catalytic site around the priming loop of PB1 domain. However, compounds 10 and 11 occupies the catalytic site while the rest of the molecules bind to tunnel 2 in the presence of capped primer from ɸ6 polymerase. Both the cyclic compounds 7 and 8 along bind near to the exit channel below the linker domain in Stage-III, although compound 7 binds to catalytic site in Stage-I. Previous studies also shown that the known compound AZT-TP bind near to the exit channel below the linker domain (Pagadala 2019) . Further, the results obtained using regression analysis with R 2 = 0.73 shows that binding Ligand protein interactions show that the compounds 1-3 in Stage-I primarily interact with PB2-N, residue nearby motif-II and motif-IV, and the PB1 priming loop that plays an important role in stabilizing the ligand-target interaction. In particular, Met 646 , Pro 647 , and Ala 648 of the priming loop of PB1 are the key residues stabilizing the compounds 1 and 3 scaffolds within the catalytic site of the IAV RdRp. This binding orientation in turn shields Asp 445 and Asp 446 essential for polymerase activity of influenza virus within the catalytic site (Zamyatkin et al. 2008 ). These two residues coordinate the metal ions involved in the catalysis. When either of these Asp-Asp residues were (Biswas and Nayak 1994) . The Asp 446 may deprotonate the 3′hydroxy group of the priming NTP by acting as a general base and helps the α-phosphate of the NTP to be incorporated into the activator strand. The suitable geometry for the reaction to happen is provided by His 649 of the priming loop. Mutational analysis of His 649 A and Pro 647 A individually, impair terminal de novo initiation on a vRNA promoter with P 651 A having the strongest effect (Te Velthuis et al. 2016) . To coordinate with the catalytic magnesium ion, the conformational change of Asp 446 side chain with 3.5 Å is required to allow the polymerase to initiate polymerization. Apart from these interactions, compounds 1 and 3 have the ability to disturb the structural stability of priming loop by interacting with nearby residues Pro 647 and Ala 648 of PB1 central domain. However, compound 2 may disturb the structural stability of PB2-N required for subunit binding to PB1-C by interacting with Asn 42 and Ala 44 of PB2-N. Structural analysis has shown that N-terminus of PB1 established a crucial high-affinity interaction with Cterminal domain of the PA protein (Reuther et al. 2011 ). Based on the binding modes, it might be speculated that compounds 1-3 may induce some structural changes of motifs-II and IV involved in the formation of catalytic mechanism and putative NTP tunnel formation. Mutational analysis of Lys 481 of motif-IV to Gln and Arg caused a drastic reduction in catalytic activity of PB1 ranging from 1 to 7% compared with wild-type which shows the importance of this amino acid in the viral replicase transcriptase activity of PB1 (Biswas and Nayak 1994) . The conserved residues of motif D, Lys 480 and Lys 481 involved in NTP binding is further stabilized by contacts with PA helix α20 (656-663) and the PA peptide 671-684 ). Thus, our compounds 1-3 interferes in stacking interactions between His 649 , P 651 with the incoming NTP by binding to the catalytic site and inhibit the activity of the viral polymerase. Furthermore, binding mode analysis revealed that compounds 4, 5, and 6 may disturb the structural stability of motif-IV by interacting the nearby residues mainly with Phe 501 and Val 502 of PB1 central. More likely PB1 Phe 501 plays a main role in IAV RdRp inhibition of these compounds 4, 5, and 6 in tunnel 2 above the PB1 priming loop. However, this residue acts as an electron donor for compound 5 inhibition, while the same residue acts as an electron acceptor for IAV RdRp inhibition by compounds 4 and 6. Plausibly, the side chains of Arg 142 , Lys 279 , and Lys 670 create a favorable environment with a positive electrostatic potential for the binding of these compounds with high potency in TCA precipitation assay. In addition, compounds 4 and 5 might be associated with less conformational changes of PB2 linker domain by tight binding with Lys 670 of PB2 627 domain. This contact may show the impact on the viral genome transcription and replication by influenza polymerase. These results also reveal that compounds with 2′deoxycytidine-5′triphosphate and diadenosine Penta phosphate are more potent than other molecules used in the study with 0.09 and 0.21 µM. This may be due to inhibition of influenza virus RNA polymerase by incorporating in place of CTP and ATP. Previously, McGeoch and Kitron showed that IAV RdRp was stimulated by GpG and GpC (McGeoch and Kitron 1975) . Later, plotch and krug had reported that initiation of IAV RdRp cRNA can be stimulated by ApG and GpG dinucleotides which are incorporated at the 5′end of the cRNA (Plotch and Krug 1977) . This enhancement of IAV RdRp by GpC and ApG is abolished by compounds 4, 5, and 6 more likely by compounds 5 and 6 than compound 4 in the series. Even though, the binding site of compounds 7 and 8 is the same at the end of the tunnel 2, compound 7 may have a different mechanism of action by donating electrons to Gly 275 of linker that leads to the PB2-N2 subdomain and induce conformational changes of PA endonuclease.

By determining the IC 50 profiles for each of these seven compounds tested experimentally, it was proved that compounds 1-7 and AZT-TP that was published previously were competitively inhibiting the incorporation of GTP into viral RNA of influenza polymerase (Pagadala 2019) . This is in good correlation between our docking studies showing that these compounds bind to catalytic and linker domains and prevent the activity of influenza RdRp similar to positive control 2′-deoxy-2′-fluoroguanosine 5′-triphosphate. This also implies that compounds 1-3 may disturb the initiation of replication by inhibiting the incoming NTPs and compounds 4-7 may terminate viral elongation by adding the next correct nucleotide. The results also suggest that the compounds 1-7 after binding to the catalytic site and exit channel inhibit the nucleophilic attack of the 3′-OH on the α-phosphorous of the incoming NTP. For all the compounds 4, 5, 6, and 7 that demonstrated inhibitory activities, the IC 50 values were less than onefold better compared with IC 50 of 1.66 μM of the known compound 2′deoxy-2′-fluoroguanosine 5′-triphosphate. Even though compounds 8-11 did not inhibit viral replication in TCA precipitation assay, they all bound to the influenza catalytic site. Similar to compounds 1-7, these molecules might not exhibit anti-influenza activity, perhaps due to premature degradation or they may bind away from the catalytic site as explained in Stage-II and III. Structural analysis of these compounds indicates that both indole ring and sulfonyl groups of V-shaped compounds 1-3 might play a major role in binding to the catalytic site of the viral polymerase. Previously, it was also shown that the effect of viral inhibition was high with compound A3 that contains an indole ring when compared with its derivate A3 that lack the indole ring. On the other hand, the high IC 50 of linear chain compounds 4 and 5 shows that high flexibility of desthiobiotin and N4-biotin might play an important role in destabilizing PB2 linker of the viral polymerase. In addition, destiobiotin may form destiobiotin-GTP complex through conjugation while N4-biotin may form N4-biotin-ε-Lys 670 affecting the viral replication in TCA precipitation assay. However, the high IC 50 of linear chain compound 6 may be due the presence of di-adenosine that leads to an increase in prolongation of transcription elongation competing with the natural substrates. Considering the binding orientation, it clearly appears that a π-stacking between Phe 495 and 2′-tosyl-N2-isobutyryl-guanosine of compound 7 may play an additional role in the inhibition of viral polymerase activity with an IC 50 of 0.68 µM. These docking combing through experimental studies clearly indicate that linear chain molecules inhibit polymerase activity with lower IC 50 compared with V-shaped and cyclic compounds used in the study. Conceivably, the pharmacophores of these compounds could be useful to screen chemical libraries of nucleoside molecules to identify novel anti-influenza drugs that bind to the catalytic site and still satisfy Lipinski's rule of five. Thus, these findings clearly reveal that these compounds have the strong potential as therapeutic leads for further drug development to treat patients infected with influenza virus.

In this study, we have successfully combined docking and experimental study to identify eight NTPs that bind to the catalytic site of IAV RdRp from A/PR/8/34/H1N1 strain. Docking studies indicated that linear shape compounds 4, 5, and 6 may induce conformational changes of PB1 motif-IV and PB2 linker domain with an IC 50 of 0.21 and 0.09 µM. Whereas, V-shaped compounds 1, 2, and 3 destabilize the PB1 priming loop involved in initiation of viral replication showing an IC 50 of 1.69, 3.58, and 2.13 µM. However, cyclic compound 8 even though binds nearer to motif-IV of PB1 central domain showed a moderate inhibition with an IC 50 of 0.68 µM. This in turn indicates that the known compound 3′azido-3′-deoxy-thymidine-5′-triphosphate binds to the viral promoter while the positive control 2′deoxy-2′fluoroguanosine triphosphate binds near to the exit channel inhibiting the transcriptional activity of viral polymerase with the same IC 50 of 1.12 µM (Pagadala 2019) . A good correlation between the ΔG and IC 50 values with an R 2 = 0.73 indicate that these molecules bind to the catalytic site inhibiting the incorporation of natural NTPs for replication initiation and elongation of IAV RdRp. The compounds 5 and 6 exhibit significant antiviral effect inhibiting the viral replication with lower IC 50 of 0.09 µM compared with other compounds used in the study. Thus, this study clearly demonstrates the importance of catalytic site below and above the priming loop for the development of effective anti-influenza drugs. The evaluation of these compounds in further experimental studies will pay the way to develop small molecule therapeutic that can inhibit the IAV viral replication binding to the catalytic site and satisfy Lipinski's rule of five.

Conflict of interest The authors declare that they have no conflict of interest.

Publisher's note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Viral replication. Structural basis for RNA replication by the hepatitis C virus polymerase

Global geno-proteomic analysis reveals cross-continental sequence conservation and druggable sites among influenza virus polymerases

Mutational analysis of the conserved motifs of influenza A virus polymerase basic protein 1

A mechanism for initiating RNA-dependent RNA polymerization

Estimated Influenza Illnesses

Validation of the general-purpose tripos 5.2 force-field

The RNA polymerase of influenza a virus: mechanisms of viral transcription and replication

T-705 (favipiravir) and related compounds: Novel broad-spectrum inhibitors of RNA viral infections

Peptide-mediated interference with influenza A virus polymerase

Crystal structure of the RNA-dependent RNA polymerase from influenza C virus

Amantadine-oseltamivir combination therapy for H5N1 influenza virus infection in mice

Influenza (Seasonal)

Triple-combination antiviral drug for pandemic H1N1 influenza virus infection in critically ill patients on mechanical ventilation

Extending the treatment of backbone energetics in protein force fields: limitations of gas-phase quantum mechanics in reproducing protein conformational distributions in molecular dynamics simulations

McGeoch D, Kitron N (1975) Influenza virion RNA-dependent RNA polymerase: stimulation by guanosine and related compounds

Equation of state calculations by fast computing machines

AZT acts as an anti-influenza nucleotide triphosphate targeting the catalytic site of A/PR/8/34/H1N1 RNA dependent RNA polymerase

Pandemic influenza A (H1N1) in HIV-1-infected patients

Structure of influenza A polymerase bound to the viral RNA promoter

Influenza virion transcriptase: synthesis in vitro of large, polyadenylic acid-containing complementary RNA

An in vitro fluorescence based study of initiation of RNA synthesis by influenza B polymerase

Structural insight into cap-snatching and RNA synthesis by influenza polymerase

The influenza virus RNA synthesis machine: advances in its structure and function

Targeting of the influenza A virus polymerase PB1-PB2 interface indicates strain-specific assembly differences

Combination therapy with amantadine, oseltamivir and ribavirin for influenza A infection: safety and pharmacokinetics

The future of flu: a review of the human challenge model and systems biology for advancement of influenza vaccinology

vitro and in vivo influenza virus-inhibitory effects of viramidine

RNA synthesis in a cage-structural studies of reovirus polymerase lambda3

The role of the priming loop in influenza A virus RNA synthesis

Inhibition of influenza virus transcription by 2'-deoxy-2'-fluoroguanosine

Structural insights into mechanisms of catalysis and inhibition in Norwalk virus polymerase