Smart, hollow, nanometer- and micrometer-size cylinders of biopolymers have attracted considerable scientific interest because of their potential applications^1–10 to molecular recognition devices,^{5a,7,9a,9c,9e,10c} enzymatic channel reactors,^{5a,5b,6,8,9b,9d,9f} drug delivery vehicles,^7,9a and others. Nanotubes present several important benefits beyond those of nanoparticles. (i) Multifunctionalities can be introduced independently to the interior and exterior surfaces. (ii) Numerous guest molecules can be loaded into the one-dimensional (1-D) pore space from open-end termini without structural deformation. (iii) Nanotubes show long-term blood retention in the circulatory system in vivo.¹¹ An efficient procedure to prepare such hollow biocylinders is template synthesis using layer-by-layer (LbL) assembly in track-etched polycarbonate (PC) membrane.^1,5–10 We demonstrated previously that human serum albumin (HSA) nanotubes (400 nm outer diameter, 200 nm inner diameter) with an avidin surface interior can capture biotin-labeled nanoparticles (100 nm diameter).^9a Currently, a challenging target for a nanotubular trap would be viruses, particularly influenza viruses. Among the influenza viruses of the three types (A, B, and C), type A is a virulent pathogen inducing severe diseases.¹² In 2009, swine-origin A (H1N1) virus spread worldwide causing a human pandemic.¹³ The infection of influenza A virus is initiated by the specific binding of viral hemagglutinin to a cellular surface receptor.¹⁴ The receptor contains sialyloligosaccharide with N-acetyl neuraminic acid (Neu5Ac) terminals. Fetuin from the fetal calf serum is a heavily glycosylated protein with tribranched oligosaccharides containing Neu5Ac residues.¹⁵ Influenza viruses are well known to interact with this naturally occurring glycoprotein.^16,17 If one were able to synthesize a unique nanotube for influenza virus trapping, then it would have a strong impact not only on biosupramolecular chemistry, but also on public health. This paper is the first to describe the synthesis and structure of HSA nanotubes that have a fetuin layer as an internal wall. Results emphasize that the influenza A virus is encapsulated in their cylindrical pores (Figure 1A).

Figure 1. (A) Schematic illustration of influenza virus A/PR8 (H1N1) and fetuin nanotube. (B, C) SEM images of fetuin nanotubes prepared using a porous PC template (800 nm pore diameter): (B) dried sample and (C) lyophilized sample after swelling in water.

Nanotubes were prepared using electrostatic LbL assembly.⁹ Positively charged poly-l-arginine (PLA, Mw: >70 kDa, pI = 12.5) and negatively charged HSA (Mw: 66.5 kDa, pI = 5.0) were alternately deposited (5.5-cycles) onto the channel surface of a track-etched PC membrane (800 nm pore size), followed by deposition of fetuin from fetal calf serum (Mw: 48.4 kDa). The isoelectric point of fetuin (pI = 3.3) is sufficiently low¹⁵ for adsorption on the positively charged PLA layer via electrostatic attraction. The dissolution of the PC template in N,N-dimethylformamide and freeze-drying of the liberated core yielded (PLA/HSA)₅PLA/fetuin nanotubes (fetuin nanotubes, Figure 1A) as a white powder. SEM measurements revealed the formation of uniform hollow cylinders with a 811 ± 13 nm outer diameter and 97 ± 3 nm wall thickness (Figure 1B). The maximum tube length (ca. 15 µm) coincided well with the PC membrane pore depth. In an aqueous solution, the fetuin nanotubes swelled markedly. Their wall thickness doubled (202 ± 5 nm) compared to that of the dried state (Figure 1C). The outer diameter was unchanged, thereby a reduction in the inner pore diameter (617 nm → 407 nm) was observed in water.

To prove the presence of fetuin at the internal wall, we introduce fluorescein-labeled fetuin (F-fetuin) as the final layer, yielding (PLA/HSA)₅PLA/F-fetuin nanotubes (F-fetuin nanotubes). Confocal laser scanning microscope (CLSM) images of F-fetuin nanotubes showed strong fluorescence, indicating the coating of the tube’s inner wall by F-fetuin (Figure 2A). Then, the amount of fetuin in the nanotube was estimated. The powder of fetuin nanotubes obtained from one PC membrane (25 mm diameter), in which ca. 1.6 × 10⁸ effective channels were observed, was suspended in acidic water (pH 3.5, 1.5 mL) to dissolve the multilayered wall. From absorbance at 280 nm, the fetuin concentration was estimated as approximately 0.2 µM, under the assumption that the fetuin nanotubes are composed of the 12-layer structure described above. Results show that one fetuin nanotube (length, 15 µm) contains ca. 1.7 × 10⁶ molecules of fetuin at the internal wall.

Figure 2. (A) CLSM image of F-fetuin nanotube (Ex. 488 nm). (B) CLSM image of F-WGA bound fetuin nanotube (Ex. 488 nm).

Wheat germ agglutinin (WGA, Mw: 38 kDa), one of the common lectins, binds to N-acetyl-d-glucosamine and sialic acid.¹⁸ Therefore, the fluorescein-labeled WGA (F-WGA) binding to the fetuin wall could be visualized in the tube’s interior by fluorescence microscopy. The F-WGA was injected into the fetuin nanotubes solution and observed by CLSM. As expected, the nanotubes fluoresced strongly, and the cylindrical wall was clearly visible (Figure 2B).

For trap experiments, we exploited influenza A virus [Puerto Rico/8/34 (PR8, H1N1)]. Moulès and co-workers described details of the morphology of PR8 virion in vitrified ice using cryogenic transmission electron microscopy (cryoTEM).¹⁹ The PR8 viruses comprised a mixture of spherical particles (94 ± 15 nm diameter) (51%) and elongated particles (49%). The inner diameter of a fetuin nanotube is sufficient to accommodate PR8 virions of both shapes. First, a fetuin nanotube dispersion (1 × 10⁸ tubes mL⁻¹, 0.25 mL) was added to the PR8 solutions with different concentrations of 2.5 × 10⁷, 5.0 × 10⁶, and 5.0 × 10⁵ pfu mL⁻¹ (0.25 mL). To prevent electrostatic adsorption of the virus particles onto the tube’s exterior surface, free HSA was added to the solution ([HSA] = 0.2 mM). After incubation for 3 h at 25 °C, the mixture was centrifuged to precipitate the tubes. The virus trapping capability of the fetuin nanotube was examined by direct enzyme-linked immunosorbent assay (ELISA) of the remaining PR8 in the supernatant by using the Influenza A Nucleoprotein Antigen Capture ELISA kit (Virusys Corp.).²⁰ Prior to the assay, we confirmed a linear relationship between the concentration of the PR8 virus and optical density in the range of 10⁴–10⁷ pfu mL⁻¹. The amount of PR8 in the sample solution mixed with fetuin nanotubes was significantly less than that of an identically treated virus solution without the tubes (control value). The remaining PR8 virus percentage against the control value was decreased by increasing the nanotube concentration (Figure 3A). The trapping ratio [100 − remaining PR8 (%)] reached 100% when [fetuin nanotube]/[PR8] was 100 (tubes/pfu). Results show that the supernatant of the PR8 solution treated with the fetuin nanotubes became completely virus free. We inferred that the influenza A virus PR8 diffused into the hollow space of fetuin nanotubes and that it bound to the inner surface wall.

Figure 3. (A) Remaining PR8 virus percentage after incubation with fetuin nanotubes and PLG nanotubes at a different ratio of [nanotube]/[PR8] (in PBS, pH 7.4, 0.2 mM HSA). Each bar represents the mean value measured in duplicate. ND: Not detected. (B) TEM image of PR8 viruses incorporated fetuin nanotube.

To clarify that the PR8 virus is truly drawn to the 1-D hollow nanospace composed of the fetuin protein, identical experiments were conducted using similar nanotubes having poly-l-glutamic acid (PLG) at the interior surface: (PLA/HSA)₅PLA/PLG nanotubes (PLG nanotubes). SEM measurements revealed that PLG nanotubes possess the same structure (810 nm outer diameter and ca. 15 µm length) as that of fetuin nanotubes. After incubation with the PLG nanotubes ([PLG nanotube]/[PR8] = 100 (tubes/pfu)) for 3 h at 25 °C, the mixture was centrifuged to spin down the tubes. The ELISA of the supernatants demonstrated that most of the PR8 had remained in the sample solution (Figure 3A). In light of these findings, we inferred that PR8 particles are trapped in the fetuin nanotube based on the specific binding of hemagglutinin on the virus surface to sialyloligosaccharide chains containing Neu5Ac residues of the fetuin wall.

TEM observation of the fetuin nanotubes incubated with PR8 demonstrated that the virus particles were accommodated into the cylinder (Figure 3B). The hollow and soft protein nanotubes were collapsed on the carbon grid by water evaporation, producing a flattened sheet with a uniform width of ca. 1.2 µm. The PR8 virion was attached to the inner surface wall of fetuin. The low contrast of the image was attributed to the preparation of the specimen without staining. Generally, a negative staining agent such as uranyl acetate (UO²⁺) is used to increase TEM image contrast. However, we reported previously that the addition of UO²⁺ to the aqueous (HSA/PLA)₃ nanotube solution induced positive staining of the tubular walls.^9a This reverse effect is caused by the strong binding of UO²⁺ to HSA.²¹ Unfortunately, our attempt to stain the fetuin nanotube using UO²⁺ to emphasize the presence of viruses failed.

In conclusion, homogeneous nanotubes composed of blood serum proteins, HSA and fetuin, completely entrapped influenza A virus PR8 (H1N1) particles. The removal efficiency by a single treatment with fetuin nanotubes reached −5.0 log order. This surprising result is expected to provoke interest in a new field of nanometer-scale-virus detection and removal devices. This system includes no antigen–antibody reaction. For that reason, fetuin nanotubes are applicable to various influenza viruses. The fetuin nanotubes are expected to be of great medical importance for advanced biomaterials and for public health.

T. Komatsu