Influenza is a highly infectious respiratory disease caused by the influenza A or B virus. Thus, its epidemiology is usually characterized by the sudden outbreak and rapid spreading of the virus. In general, more than 500 million people are infected by influenza virus annually, causing up to 250-500 thousands of deaths worldwide. In the past few years, the concirculation of influenza A H1N1 and H3N2 viruses with influenza B viruses in the human population and the outbreak of avian influenza (H5N1) in Southeast Asia, both contribute more opportunity for genetic reassortment of the viruses. A new type of influenza virus, which is a genetic reassortment from human, swine and avian influenza viruses, arose suddenly in Brazil in April 2009 and has spread rapidly worldwide. Many countries are taking various stringent measures to reduce the spread of the disease. Two classes of anti-influenza viral drugs targeting either the M2 protein (amantadine and rimantadine) or neuraminidase (NA; oseltamivir and zanamivir) are currently licensed, apart from Chinese herbs which action mechanisms have not been elucidated. These drugs are important in controlling potential pandemic influenza. The first two drugs are amine derivatives of adamantine and are M2 inhibitors, thus they are only effective against influenza A viruses. Amantadine and rimantadine have long been available for both prophylaxis and therapy of influenza A viral infections. Zanamivir (Relenza) and oseltamivir (Tamiflu) are potent inhibitors of influenza NA, and are both prophylactically and therapeutically effective against both influenza A and B viral infection in humans. Recent antiviral surveillance studies reported a global significant increase in amantadine drug-resistance among influenza viruses. This makes oseltamivir being generally regarded as the last resort in controlling the potential influenza pandemic, even in the recent H1N1 influenza pandemic despite the virus susceptibility to oseltamivir. However, the supply of oseltamivir is limited due to the shortage of raw material for manufacturing the drug. The news about influenza pandemic, drug resistance and shortage of antiviral drugs urge us to initiate this study to look for new lead compounds for developing new anti-influenza viral drug. The viral NA is essential for both the release of virus from infected cell and its transport through mucus in the respiratory track. Although the amino acid sequence homology between NAs of influenza A and B viruses is only about 25%, the active site of the enzyme is highly conserved and therefore presents a potential target for the design of new drugs. In this study, we focused on the discovery of candidates as NA inhibitors using several methods including high throughput screening (HTS), bioassay-guided isolation, virtual screening and structure-based drug design. Since each of the above methods has its own advantages, we have developed an integrative approach by combining these different methods to streamline the process for the discovery of new candidates with higher activities and lower toxicities. Briefly, several candidates such as flavonoids, stilbenoid, caffeic acid derivative and p-amino salicylic acid (PAS) with novel scaffolds were first identified using HTS of a plant extract library and vitural screening of a compound database. Among them, PAS was selected for further investigation by drug design. It was then chemically modified and tested for its activity. Eventually a derivative was found to exhibit dramatic increase in potency. A brief summary of the key finding of each method/approach was shown as follows: 1. HTS and structure-activity relationship analysis HTS of natural products for NA inhibitors After a fluorescence-based NA activity assay for HTS was established and validated, 1655 extracts, which satisfied with the polarity requirement of NA active sites, were screened for NA inhibitory activity. Finally, 15 active extracts were identified including Elsholtzia rugulosa Hemsl. and Caesalpinia sappan L., which were selected for further bioassay-guided isolation of active constituents. Bioassay-guided isolation The isolation and identification of the active constituents from E. rugulosa and C. sappan was guided by NA activity. Finally, 5 active constituents were obtained from E. rugulosa: apigenin, luteolin, apiin, galuteolin and luteolin-3′-glucuronyl acid methyl ester. On the other hand, 6 active constituents were obtained from C. sappan which include brazilein, brazilin, protosappanin A, 3-deoxysappanchalcone, sappanchalcone and rhamnetin. The results of in vitro antiviral assay confirmed that they all possessed anti-influenza viral activity. Structure-activity relationship (SAR) analysis The NA inhibitory and anti-influenza viral activities in vitro of 8 flavonoids from E. rugulosa and C. sappan suggested that a number of flavonoids possess anti-influenza viral activities. In order to reveal their structure-activity relationship (SAR), 25 flavonoids with 5 different basic scaffolds were studied. Density functional theory (DFT) was used to calculate the chemical parameters of these compounds. The results indicated that the order of potency for NA inhibition was as follows: aurones > flavon(ol)es > isoflavones > flavanon(ol)es and flavan(ol)es. The SAR analysis of flavonoids on influenza viral NAs revealed that for good inhibitory effect, the 4í-OH, 7-OH, C4=O and C2=C3 functionalities were essential and the presence of a glycosylation group greatly reduced NA inhibition. The in vitro anti-viral activities of 8 flavonoids were evaluated using cytopathic effect (CPE) reduction method, the assay results were confirmed with the SAR above. The total energy and molecular volume were also predictive of their activities. 2. Virtual screening The second approach used to further identify active NA inhibitors is computational based drug screening, namely virtual screening. A pharmacophore was created based on 4 active compounds, including one obtained from our SAR analysis of flavonoids, two well known antiviral drugs (zanamivir, and oseltamivir), and the natural substrate sialyc acid. It was then used to search the database containing more than 40,000 compounds for hits which partially satisfied with the pharmacophore. 2. Virtual screening The second approach used to further identify active NA inhibitors is computational based drug screening, namely virtual screening. A pharmacophore was created based on 4 active compounds, including one obtained from our SAR analysis of flavonoids, two well known antiviral drugs (zanamivir, and oseltamivir), and the natural substrate sialyc acid. It was then used to search the database containing more than 40,000 compounds for hits which partially satisfied with the pharmacophore. Eventually, these hits were selected to perform molecular docking and post-docking analysis, Furthermore, 240 potential compounds were chosen to undergo NA activity assay, and 5 of them were found to have NA inhibitory activity. Although their activities were weaker than that of zanamivir, their structural scaffolds are novel, and one of them displayed 70 times higher activity than the most active compounds above. Thus, these findings provide important information for the design of new drug. 3. Drug design Based on the active compounds obtained above and the features of NA active site, several series of new compounds were designed. Virtual screening was carried out, and the computational results suggested that the p-amino salicylic acid (PAS) derivatives have the highest potential among all. Therefore, these PAS derivatives were chemically synthesized, and then being evaluated for NA activity. A series of PAS derivatives containing a lipophilic side chain at C-2 and an amino or guanidine at C-5 were synthesized and evaluated for their abilities to inhibit NA isolated from influenza A virus (H3N2). All these compounds showed potent inhibitory activity against influenza NA and their SAR analysis showed that the order of increasing activity at C-5 was as follows: N=C(NH2)2 > ñNH2 > ñNO2 > ñH. While the side chain at C-2 did not affect the activity, possibly due to the R2 group stretching out of the active pocket. The most active compound of the series is compound 8, which activity was about 90 times higher than that of PAS. In summary, using multiple new technologies and new methods, 11 constituents with influenza NA inhibitory effects, for the first time, were systematically isolated from Elsholtzia rugulosa and C. sappan, and their structure-activity relationship was elucidated, which will provide important information for new drug design and for the exploitation and utilization of flavonoids as NA inhibitors for influenza prevention and treatment. Five active compounds with novel scaffolds, including 2 compounds from chemical synthesis and 3 natural products, including p-amino salicylic acid (PAS), a caffeic acid derivative and a stilbenoid, were firstly discovered and provided important information for new drug design. Based on the studies above, more than 20 new compounds were subjected to design, synthesis, activity evaluation and structure-activity relationship analysis, several compounds with highly increased activities and one predictable SAR model were obtained. These studies will provide important theory and information for the development of anti-viral drugs for influenza prevention treatment. Keywords: Influenza; Neuraminidase (NA); High throughput screening (HTS); Bioassay-guided isolation; Structure-activity relationship (SAR); Virtual screening; Drug design; Elsholtzia rugulosa Hemsl.; Caesipinia sappan L.; p-amino salicylic acid (PAS)