Preparation and Characterization of porous gold nanoparticle (PoGNP) and other metallic nanoparticles
In contrast to conventional gold nanoparticle synthesis, PoGNP was synthesized using the surfactant-free emulsion method. The Au3+ ion in HAuCl4 and anilinium ion (C6H5NH3+) formed the PANI-Au complex nanoparticle by reducing the Au3+ ion and oxidizing the anilinium ion to PANI through redox reaction 29. After fabrication of the PANI-Au complex nanoparticle, PANI was selectively etched by NMP from the PANI-Au nanocomplex, shaping a mesoporous gold structure. Non-porous sGNP was synthesized via the seed-mediated growth method to compare whether nanoparticle morphology could affect inactivation of the virus. AgNP was also prepared as an antiviral control group as AgNP has been widely studied for its antiviral properties.
The morphology of the synthesized nanoparticles was observed by TEM and their average size was determined by dynamic light scattering (DLS) analysis. TEM images of synthesized PoGNP exhibited 6 nm-sized nanospheres with rough surfaces and nanobundled structure (Fig. 1A), whereas sGNP and AgNP exhibited narrow monodispersed spherical structures with smooth surfaces (Fig. 1B, 1C). Moreover, PoGNP was stable in buffer solution (DPBS) for over 7 days despite being synthesized without surfactant. Mean of cumulative sizes of PoGNP, AgNP, and sGNP that was measured triple times were 156.65 4.47 nm, 20.75 2.04 nm and 129.96 26.83 nm, respectively (Fig. 1D).
Moreover, the biocompatibility of the synthesized nanoparticles was evaluated by MDCK cell viability assay. MDCK cells in 96-well plates were exposed to various nanoparticle concentrations and cell viability was measured by the WST-1 assay, with the optical density being measured at 450 nm. The measured cell viability after nanoparticle treatment was greater than 95%, except for 0.2 mg/mL PoGNP and 0.1-0.2 mg/mL AgNP, which were observed to have 80% cell viability. The results of the nanoparticle cell viability assay implied that the infectivity of nanoparticle-treated viruses could be evaluated by MDCK cytotoxicity assay because the nanoparticles exhibited good biocompatibility with the MDCK cells (Fig. S2).
Interaction between synthesized nanoparticle and influenza virus
Interaction between the prepared nanoparticles and IAV HA was evaluated, as we hypothesized inactivation of IAV to be induced by cleaving disulfide bonds required for viral membrane fusion with the host cell. TEM analysis was first conducted for a simple confirmation of the interaction between the virus and the nanoparticles. IAV suspension was incubated with 2 × 10-1 mg/mL of nanoparticles for 10 min, and then the mixtures were prepared for TEM analysis. However, interaction was difficult to observe in the TEM images with larger NPs such as PoGNP and sGNP. On the other hand, AgNP agglomerated on the viral surface as expected, since the small size of AgNP facilitated migration (Fig. 2). Another technique was considered to prove the affinity between HA and the larger sGNPs. As the isolation of IAV requires ultracentrifugation over 20,000 g31 or gradient centrifugation 32, we centrifuged the samples at 6000 g for 10 min to monitor whether H3N2 precipitated with the nanoparticles. We assumed that the nanoparticle-treated H3N2 virus would precipitate with the nanoparticles in contrast to the H3N2 virus sample alone. Definitively, the real-time cycle quantification (Cq) value of PoGNP-treated H3N2 virus in redispersed precipitate solution was much lower than that in the supernatant, indicating that H3N2 virus could interact with PoGNP (Fig. 3). In addition, PoGNP attracted more H3N2 virus than sGNP at lower concentration according to the real-time Cq values of precipitated samples, which indicated that PoGNP had much higher affinity for HA compared with sGNP. The difference in attraction resulted from their surface structure; the foam-shaped porous outer surface of PoGNP created more surface area for interaction with HA than the sGNP surface.
Antiviral effect of PoGNP compared with other metallic nanoparticles
To observe the antiviral effect of nanoparticles, h1n1 virus was exposed to each nanoparticle suspension for 10 min and 60 min prior to infection of the MDCK cells. The antiviral effect of the nanoparticles was determined by WST-1 cytotoxicity assay by observing the optical density of treated cells at 450 nm. Compared with the other nanoparticles, PoGNP showed much higher antiviral activity on H1N1 virus, whereas AgNP showed only minor antiviral activity over 0.1 mg/mL AgNP. 0.2 mg/mL PoGNP successfully inactivated H1N1 virus after exposure for 60 min. In contrast, sGNP had no antiviral effect regardless of its concentration or exposure time (Fig. 4). Comparing PoGNP with sGNP, the difference in nanoparticle antiviral activity is the result of differences in their specific surface areas despite similar diameters; each nanoframe of PoGNP behaved as a single reactant for disulfide bonds that could interact with HA. AgNP was able to agglomerate on viral HA and had extensive specific surface area for interaction compared with sGNP due to its small size; however, PoGNP showed higher inactivation of the virus at 0.2 mg/mL. PoGNP’s superior virus inactivation ability compared with AgNP and sGNP is due to both its stability under saline conditions and its high affinity for HA. The antiviral effectiveness of AgNP is restricted because AgNP aggregated in the culture media at higher concentration and it should only be treated at concentrations lower than 0.1 mg/mL due to toxicity concerns. As observed by lower attraction of sGNP to viral HA in the binding efficiency test with centrifugation, sGNP exhibited lower antiviral activity, as expected. Indeed, the inactivation of HINI virus by sGNP was much lower than that of PoGNP, although sGNP also exhibited affinity for HA at higher concentration. This result endorsed the fact that inactivation of HINI virus was caused by the cleavage of disulfide bonds ensuing from the influence of metallic nanoparticles because the relatively flat surface of sGNP made contact with the disulfide bridge difficult in contrast to PoGNP’s rough surface.
Antiviral effect of PoGNP towards various influenza A virus subtypes
To confirm the inactivation of PoGNP regardless of genetic mutation, 3 IAV strains (H1N1, H3N2, and H9N2) were treated with PoGNP (Fig. 5 and Fig. S3). According to the phylogenic tree of IAV, the selected virus strains had low sequence similarity, so they were considered representative of general viral treatment 33. For the 10 min-treated sample, only 0.2 mg/mL PoGNP showed over 50% cell viability. Under 0.1 mg/mL PoGNP concentration, the treated virus was only slightly inactivated compared to the non-treated virus. After the 60 min treatment, the antiviral effect of treatment with 0.2 mg/mL PoGNP on H1N1, H3N2, and H9N2 increased to 74%, 76%, and 56%, respectively. 54% of MDCK cells survived H1N1 infection when treated with 0.1 mg/mL PoGNP for 60 min. H3N2 and H9N2 were minorly affected under 0.1 mg/mL PoGNP and still exhibited higher cell viability than the control group. PoGNP displayed antiviral activity on various virus strains, which is important for on-the-spot preprocessing of IAV for further analysis. Here, we could conclude that 0.2 mg/mL PoGNP could attenuate the infectivity of multiple IAV strains.
Intracellular viral RNA quantification
We had inferred the inactivation mechanism of the virus as the cleavage of disulfide bonds, because the high affinity of the disulfide-gold interaction and abundant presence of disulfide bonds in HA would expedite inhibition by PoGNP. The viral inhibition mechanism of PoGNP was then identified by quantitative analysis of intracellular viral RNA from MDCK cells infected with nanoparticle-treated viruses (Fig. 6). MDCK cells were infected with PoGNP-treated H3N2 virus for 24 h and the cells were lysed with Qiagen RNeasy mini kit, followed by calculation of logEID50/mL values by standard EID curve (Fig. S1). PoGNP-treated H3N2 virus showed lower intracellular viral RNA levels indicating that the amount of IAV in the endosome of MDCK cells was clearly reduced in the nanoparticle-treated samples. Also, the amount of intracellular viral RNA was clearly influenced by the exposure time. The gene content level of viral RNA in the host cells was higher for viruses treated for 10 min compared with 60 min. Intracellular viral RNA quantification data corresponded well with the MDCK cell viability test results, revealing that the antiviral activity of PoGNP was affected by time and concentration, especially 0.2 mg/mL PoGNP was effective after exposure for 10 min. Furthermore, the result also confirmed that PoGNP blocked viral infection by inhibiting viral entry.
Demonstrating the importance of disulfide bond in viral infection
We determined that viral infection was suppressed by obstructing viral attachment, therefore we further experimented to prove that the disulfide bonds in HA have important role in viral infection. As described above, gold atoms tended to participate in gold-sulfur interactions by cleaving disulfide bonds in HA. TCEP was thus selected as a disulfide-reducing agent to cleave the disulfide bonds in HA. Results of the WST-1 assay revealed that TCEP-treated IAVs lost their infectivity similar to PoGNP-treated viruses. Samples treated with under 10-3 M TCEP exhibited nearly 60% cell viability, which was much higher than the 33% viability of H3N2 virus-infected MDCK cells (Fig. S4). On the other hand, 10-2 M TCEP-treated samples displayed lower cell viability compared to other TCEP-treated samples because of the cytotoxicity of TCEP itself 34. The result suggested that the cleavage of disulfide bond before membrane fusion reduce viral infectivity toward MDCK cells. Likewise, as the viral infectivity of PoGNP-reacted influenza virus was suppressed to the level similar to that of the TCEP interacted one, which derives that PoGNP also cleaves the disulfide bonds in HA as TCEP.