Madin Darby Canine Kidney (MDCK) cells were obtained from the American Type Culture Collection (Manassas, VA) and grown at 37˚C with 5% CO₂ in Roswell Park Memorial Institute medium (RPMI; Invitrogen, No: 22400-105), supplemented with 10% foetal bovine serum (FBS; Invitrogen, No: 16140-071) and 1% Penicillin-Streptomycin (Invitrogen, No: 15140-122). Before the addition of compounds or the virus, or while determining the results, the monolayers of MDCK cells were washed twice with phosphate buffered-saline (PBS, pH 7.4 at room temperature). The following controls were included in all experiments: cell control (cells that were neither infected with the virus nor treated with the plant extracts), virus control (cells that were infected

^*Information derived from; Chai 2006 [18] , Salleh 2002 [21] , Yaacob 2009 [22] , Maji 2010 [20] , Focho 2010 [17] [19] .

only with the virus but not with the plant extracts), and the positive controls (virus-infected cells treated with NA inhibitor drugs, Zanamivir or Oseltamivir).

Six influenza viruses (Table 2) that were isolated, plaque purified and cultured in MDCK cells were provided by Dr. Aeron Hurt, WHO Collaborating Centre for Reference and Research on Influenza, Victorian Infectious Diseases Reference Laboratory (VIDRL), Melbourne, Australia. Virus stocks were developed in MDCK cells using RPMI medium supplemented with 4 µg/mL trypsin (Sigma, No: T1426) at 37˚C in 5% CO₂ for three days as described elsewhere [23] . Supernatants containing virus were collected after cytopathic effects (CPE) were observed and antiviral titres were determined using 50% Tissue Culture Infectious Dose (TCID₅₀), according to Reed and Muench’s endpoint method [24] , and a colorimetric endpoint was used to obtain quantitative results as described elsewhere [25] . All aliquots of virus stocks that were collected were stored at −80˚C until use.

The medicinal plant extracts were originally extracted with a mixture of dichloromethane and methanol in a 1:1 (v/v) ratio and subsequently concentrated using a rotary evaporator. The yield was dependent on the part of plant used; approximately 0.05 to 0.10 g of extract was obtained from 6 g of whole plant, whereas the yield was reduced to 0.01 to 0.05 g when extracts were obtained from stems only. Prior to use, the extracts were reconstituted in PBS with 10% dimethyl sulfoxide (DMSO, SIGMA No: D 5879) and filtered using a 0.45 µm filter (Sartorius Stedium Australia No: 16533K). Selection of plant extracts for this study was based on the pathways of viral inhibition involving inhibition of hemagglutination demonstrated in previous studies [17] .

The antiviral activities of plant extracts against influenza viruses were evaluated according to a method described elsewhere [17] . Briefly, 96-well plates seeded with 3 × 10⁴ cells/well were incubated for 24 h at 37˚C with 5% CO₂ until a confluent monolayer was attained. The cells were washed twice with PBS, and two-fold serial dilutions of plant extracts (3.13 - 100 µg/mL) in RPMI medium challenged with 100 TCID₅₀ of influenza virus were added to the wells in triplicate for each concentration. Wells containing 10% DMSO with media only and virus only, devoid of plant extract were also included in the study. All wells were supplemented with virus growth medium (100 µL of RPMI medium containing 2 µg/mL trypsin). After three days of incubation at 37˚C/5% CO₂, the results were quantified as previously explained. The antiviral activity curve was then constructed by plotting percentages of virus inhibition against concentrations of extracts. The inhibitory concentration (IC₅₀) of extract required to reduce virus-induced CPE by 50% was expressed relative to the virus control employing dose-response curves. A trend line that best suited the curve was selected using regression analysis of antiviral activity curves (in Microsoft excel) and the corresponding equation was used to calculate IC₅₀ values [26] .

In order to test the effects of extracts on the viral NA of Types A and B viruses, the NA-Fluor™ Influenza Neuraminidase Assay Kit (Life Technologies, No: 4457091) was employed as per the manufacturer’s instructions. The optimum virus dilution for the NA inhibition assay was selected by titration of the virus stock, employing the NA activity assay. Two-fold serial dilutions of plant extracts (0.3 - 25 µg/mL) were tested for NA inhibitory activity. NAIs, Zanamivir and Oseltamivir, were included as positive controls in the assay and tested at nanomolar concentrations (10⁻² to 10⁴ nM). A POLARstar Omega fluorescence polarization microplate reader (excitation 355 nm, emission 460 nm) was used to measure fluorescence. A sigmoidal curve generated by dose- response data and analysed using GraphPad Prism Software was used in the determination of IC₅₀ values.

The effect of extracts on virus adsorption was determined by an HI assay [27] . Briefly, two-fold serial dilutions of the extract (12.5 - 100 µg/mL) prepared in PBS and an equal volume (25 µL/well containing 4 HAU) of the virus stock were added to each well in a round-bottomed 96-well microtitre plate in triplicate, followed by the addition of 50 µL of 5% chicken red blood cells (CRBC). The following controls were included in every plate i) CRBC without virus; ii) CRBC with virus devoid of extract and iii) CRBC with extracts devoid of virus. The hemagglutination reactions were observed after 30 minutes incubation at room temperature.

In vitro micro inhibition and HI assays were employed to study the effect of Receptor Destroying Enzyme (RDE, Denka Seiken Co. Ltd., Tokyo, Japan) treatment upon the antiviral activity of the extracts. The extracts added to RDE solutions in the ratio 1:3 were incubated at 37˚C for 20 hours according to the manufacturer’s instructions. The objective of this experiment was to eliminate the activity of compounds that may possess sialic acid (SA)-like structures which imitate the receptors of RBC and compete for hemagglutinin [28] . The RDE and extract mixture was inactivated at 56˚C for 60 minutes before performing the in vitro micro inhibition and HI assays.

Crude extracts 8, 41, 42 and 43 were analysed using a gas chromatograph coupled to a mass spectrometer as detector (GCMS-QP2010 Ultra, Shimadzu), equipped with an Rxi^®-5SIL-MS column. Helium gas was used as a carrier gas with total flow, column flow and purge flow set to 7.5 mL/min, 1.50 mL/min and 3.0 mL/min respectively; 2 µL of the sample was injected. Injector temperature was set to 200˚C and split injection mode (1:2) was selected at pressure 88.9 kPa. The column oven temperature was programmed as follows; held at 50˚C for the first two minutes and ramping at a rate of 20˚C/min to a final temperature of 250˚C, then held for eight minutes. Total GC run time was 20 min. The chromatogram obtained after the run was used to generate a qualitative table of the 50 most abundant compounds using peak integrate option. A similarity search was then performed to determine the possible structural units. The compounds were identified by comparison with the NIST and Wiley 8.0 MS libraries. Only peaks with a match greater than 75% and area % greater than 0.5% were reported.

All treatments were performed in triplicate and each experiment was independently repeated at least twice. The data were expressed as mean ± standard error of the mean (SEM). The results of the antiviral activity assays were analysed with a one-way ANOVA test and a significance level (p value) of 0.05 or 0.01 was considered to compare the means.

The plant extracts 8, 41, 42 and 43 were active against both the wild-type seasonal Type A (H1N1) and mutant seasonal Type A (H1N1) (H275Y) viruses. As shown in Table 3, IC₅₀ values for extract 8 were <3.13 µg/mL

IC₅₀ represents the concentration of plant extract needed to reduce the viral inhibition by 50% relative to virus control wells without test compound, calculated from dose-response data of virus inhibition. Plant extracts (3.13 - 100 µg/mL) in RPMI medium were challenged with 100 TCID₅₀ of influenza viruses.

against both viruses. Extracts 41 and 42, obtained from different parts of the same plant, showed similar IC₅₀ values against the wild-type virus and a relatively higher IC₅₀ value was noted when the mutant virus was tested. Extract 43 was shown to have an IC₅₀ of 16.45 µg/mL against the wild-type virus, while 27.63 µg/mL were required to reduce the activity of the mutant virus to 50%. Among the NAI drugs that were included as controls, Zanamivir showed antiviral inhibition against both viruses while Oseltamivir was active only against the wild- type virus and showed an IC₅₀ value > 100 µg/mL against mutant seasonal A virus.

As shown in Figure 1, Extract 8 demonstrated similar activity against both the wild-type and mutant seasonal A (H1N1) viruses between 25 - 100 µg/mL, but at concentrations <12.5 µg/mL a lower percentage of viral inhibition was exhibited against the mutant strain; this result was statistically significant (p < 0.05). Extracts 41, 42 and 43 also showed a similar trend in viral inhibition against both wild-type and mutant seasonal Type A viruses, however inhibition against the mutant strain was relatively lower. DMSO, at a concentration of 10% in media did not affect the cell viability of cells with media only and virus only, devoid of plant extracts and cell viability obtained was similar to that of cell control and virus control respectively without 10% DMSO.

Extracts 8, 41, 42 and 43 demonstrated antiviral activities against both the wild-type pandemic A (H1N1) and the mutant pandemic A (H1N1) (H275Y) viruses. As evidenced in Figure 2, Extract 8 showed similar activities against the wild-type and mutant pandemic viruses with a demonstrated IC₅₀ value of <3.13 µg/mL. Although similar trends in virus inhibition were shown by extracts 41, 42 and 43, at concentrations below 6.25 µg/mL a lower percentage of virus inhibition was demonstrated against the mutant virus compared to the the wild-type. As shown in Table 3, IC₅₀ values were <3.13 µg/mL against the wild-type for all the extracts while against the mutant virus IC₅₀ values of 9.99 µg/mL and 12.2 µg/mL were obtained for extracts 41 and 43, respectively. Even though extracts 41 and 42 were obtained from different parts of the same plant, the IC₅₀ value against the mutant virus was <3.13 µg/mL for Extract 42, while it was 9.99 µg/mL for Extract 41; this result was statistically significant (p < 0.05).

Since the H275Y mutation confers Oseltamivir-resistance, the NAI inhibitor Oseltamivir was only active against the wild-type virus with an IC₅₀ value of <3.13 µg/mL while the IC₅₀ was >100 µg/mL against the Oseltamivir-resistant mutant virus. Zanamivir was shown to be active against both viruses with IC₅₀ values of <3.13 µg/mL.

Extracts 8, 41, 42 and 43 were also shown to be active against both the wild-type and the mutant Type B (D197E) viruses. As evidenced in Figure 3, Extract 8 showed similar trends in viral inhibition against both wild-type and mutant viruses; this behaviour was also observed with extracts 41 and 43. However, Extract 42 was active between 6.25 - 100 µg/mL against both viruses; while at 3.13 µg/mL the antiviral activity against the wild-type virus was only 17% which was significantly less than the activity against mutant virus. As shown in Table 2, Extract 8 demonstrated an IC₅₀ value of <3.13 µg/mL against both viruses while Extract 43 showed

IC₅₀ values of 13.55 µg/mL and 9.19 µg/mL against the wild-type and mutant viruses, respectively. Extract 41 demonstrated a similar IC₅₀, of 11.16 µg/mL and 11.29 µg/mL against the wild type and mutant viruses, respectively, while Extract 42 showed IC₅₀ values of 4.6 µg/mL and <3.13 µg/mL for the same viruses.

Since the Type B virus with D197E mutation confers resistance against all NA inhibitors, both Oseltamivir and Zanamivir were shown to demonstrate significantly lower antiviral activity against the mutant virus relative to the wild-type virus (Figure 3), but the IC₅₀ value of Zanamivir showed no significant difference for both viruses, even though the D197E mutation affects the NAI potential of the drug in the order of <10 fold. However, the IC₅₀ value of Oseltamivir, which is known to be affected by the D197E mutation in the order of 10 - 50 fold or higher, was shown to have a calculated IC₅₀ value greater than 22.67 µg/mL against the mutant virus while that against the wild-type virus was 6.23 µg/mL.

Influenza NA serves as a principal drug target for the treatment of influenza [29] . Extracts 8, 41, 42 and 43 were tested for NA inhibitory activity against the six influenza viruses included in the study. Increasing concentrations of extracts were associated with decreased relative fluorescence, consistent with the inhibition of NA activity (Table 4). Extract 8 reduced the NA activity of the six viruses at lower concentrations (3.14 - 7.99 µg/mL) compared to other extracts. The NAI potential of Extracts 41 and 42 collected from the same plant source were not similar. As reported in Table 4, Extract 41 showed lower IC₅₀ values than Extract 42 against wild-type and mutant seasonal A (H1N1), pandemic A (H1N1) (H275Y) and Type B viruses, while Extract 42 showed lower IC₅₀ values than Extract 41 against pandemic A (H1N1) and Type B (D197E) viruses; this result was statistically significant (p < 0.05). The IC₅₀ value for Extract 43 ranged between 4.90 - 10.02 µg/mL against all the viruses that were tested.

NAI activity of the plant extracts was measured at concentrations ranging from 0.3 to 25 µg/mL, while the controls, Zanamivir and Oseltamivir, were measured at 0.01 to 10,000 nM, as recommended by the manufacturer. The optimum virus dilution for the NAI assay was selected by titration of virus stock in an NA activity assay; 1:4 dilutions of the virus strains were selected in the NA activity assay to perform NAI assay.

Zanamivir and Oseltamivir were included as positive controls in the assay and tested at nanomolar concentrations, as recommended by the manufacturer; therefore the results should not be compared directly with the crude plant extracts that were tested at µg/mL concentrations. IC₅₀ values obtained in our studies for the NAIs were in agreement with those reported in the literature except for minor variations in the NAI potential of Oseltamivir against pandemic A (H1N1) and Type B (D197E) viruses and Zanamivir against seasonal A (H1N1) (Table 5).

The critical step in the initiation of influenza infection is the binding of viral HA to the sialic acid (SA) residues expressed by host cell glycoproteins and glycolipids [30] . The viral HA also binds to SA residues expressed on the surface of erythrocytes, resulting in hemagglutination. The ability of plant extracts to inhibit influenza virus-induced hemagglutination was studied in an HI assay.

As shown in Figure 4, Extract 8 did not demonstrate HI activity against any of the six viruses studied. At concentrations ranging between 50 - 100 µg/mL, extracts 41 and 42 (collected from different parts of the same plant) inhibited virus-induced hemagglutination of both wild-type and mutant strains of pandemic A (H1N1). Although extracts 41 and 42 demonstrated HI activity against Type B virus only at 100 µg/mL, both extracts demonstrated effective HI activity against Type B (D197E) virus at 50 - 100 µg/mL; however, no HI activity

^*The IC₅₀ range was based on results determined by seven different laboratories using a range of different fluorescence-based protocols [39] .

was evidenced against the seasonal A (H1N1) viruses. Extract 43 inhibited HI activity of both wild-type and mutant pandemic A (H1N1) viruses at 100 µg/mL while exhibiting HI activity only against Type B (D197E) at the same concentration. Nevertheless, Extract 43 did not exhibit HI activity against wild-type and mutant seasonal A (H1N1) viruses and wild-type B viruses. None of the plant extracts caused hemolysis of CRBC at 12.5 - 100 µg/mL as evidenced in previous studies that were performed on different batches of extracts 8, 41, 42 and 43 [17] [31] .

Extracts 8, 41 and 42 and 43 were treated with RDE in order to eliminate SA mimics that may demonstrate antiviral potential against the viruses similar to the experiments against Mem-Bel and PR8 viruses in our previous study [17] . An in vitro micro-inhibition assay using the six viruses was performed with the RDE-treated extracts. As shown in Figure 5, antiviral activities originally exhibited by all four extracts were either eliminated or reduced

following RDE treatment. The antiviral activity of Extract 8 was suppressed against all viruses following RDE treatment, with the exception of mutant pandemic A (H1N1) (H275Y) and Type B viruses where viral inhibitions of 51% and 57%, respectively, were exhibited. RDE-treated extracts 41 and 42 showed loss of antiviral activity against all the six viruses. On the contrary, the antiviral activity of RDE-treated Extract 43 was reduced by only 14% against pandemic A (H1N1) (H275Y) virus (down to 78% virus inhibition), 22% against Type B (45% virus inhibition) and 30% against Type B (D197E) virus (48% virus inhibition). However, the activity against Type B viruses was considered to be negligible since the 50% virus inhibition threshold was not reached. There was no evidence of antiviral activity against either the wild-type or the mutant seasonal A and pandemic A (H1N1) viruses; this result was statistically significant (p < 0.05).

Extracts 41 and 42, which demonstrated HI activity against pandemic H1N1 and Type B viruses, were treated with RDE in order to eliminate SA-like components that may compete with the CRBC receptors for viral HA. Extract 43, which showed HI activity against pandemic H1N1 and mutant Type B (D197E) viruses was also subjected to RDE treatment. An HI assay was then performed with the RDE-treated extracts against wild-type and mutant strains of pandemic A (H1N1) and Type B viruses. As shown in Figure 6, RDE treatment suppressed the HI activity originally exhibited by extracts 41, 42 and 43.

RDE-treated extracts were also studied in an NAI assay to determine whether NAI SA mimics were present in the extracts. As shown in Figure 7, the NAI activity originally present in the extracts was also affected by RDE treatment. The NA activity of the virus-only wells were set to 100% and the loss in the NA activity of the virus by treatment with plant extracts in the native form and those that were treated with RDE is shown in Figure 7. In its native form, Extract 8 was shown to significantly reduce the NA activity to less than 30% against all but Type B viruses where the extract reduced the NA activity to 43%, while RDE-treated Extract 8 showed decreased NAI potential. Similarly, RDE treatment decreased the NAI potential of extracts 41 and 42; NA activity of virus in wells treated with RDE treated extracts was close to 100%, similar to the wells containing only the virus. Extract 43 also showed a similar trend in the loss of NAI activity upon RDE treatment. Only with Type B virus, the difference in the NAI activity of extracts 8, 41 and 42 in native and RDE-treated forms was less than 25%. RDE-treated extracts were shown to contain relatively lowered NAI potential than native extracts; the differences between means for each concentration were statistically significant (p < 0.05).

The crude extracts 8, 41, 42 and 43 were analysed using GC-MS in order to determine their chemical profile. Several components were present in the medicinal extracts and the first 50 peaks present in the fractions were identified using NIST and Wiley 8.0 MS libraries based on the area (%), however, only compounds that showed greater than 75% matching and 0.5% area were reported. Since the crude extracts demonstrated the presence of several components with multiple peaks of minor area, 50 peaks were chosen for each sample in order to determine the chemical nature. The reason for choosing 50 peaks was to ensure that compounds which contributed only to a minor area were not neglected; hence compounds that contribute to the antiviral potential of the extract through synergy may be detected.

As shown in Figure 8, Extract 8 showed a major proportion of esters (20.63%), hydrocarbons (19.34%), cyclitols (16.06%) and alcohols (14.84%); Quinic acid constituted 9.25% of Extract 8. Although extracts 41 and 42 were obtained from different parts of the same plant, their chemical profile showed variations in the amount of chemical components contained. Extract 41 showed more amounts of alcohol (11.2%), aldehyde (4.8%) and ketone (19.37%) and Extract 42 showed relatively higher amount of esters (11.7%) and phenolics (8.2%) compared to Extract 41. Extract 43 showed major proportion of phenolic compounds (20.39%), esters (19.1%), alcohols (11.36%) and cyclitols (10.34%).