A Simple Add-and-Display Method for Immobilisation of Cancer Drug on His-tagged Virus-like Nanoparticles for Controlled Drug Delivery

pH-responsive virus-like nanoparticles (VLNPs) hold promising potential as drug delivery systems for cancer therapy. In the present study, hepatitis B virus (HBV) VLNPs harbouring His-tags were used to display doxorubicin (DOX) via nitrilotriacetic acid (NTA) conjugation. The His-tags served as pH-responsive nanojoints which released DOX from VLNPs in a controlled manner. The His-tagged VLNPs conjugated non-covalently with NTA-DOX, and cross-linked with folic acid (FA) were able to specifically target and deliver the DOX into ovarian cancer cells via folate receptor (FR)-mediated endocytosis. The cytotoxicity and cellular uptake results revealed that the His-tagged VLNPs significantly increased the accumulation of DOX in the ovarian cancer cells and enhanced the uptake of DOX, which improved anti-tumour effects. This study demonstrated that NTA-DOX can be easily displayed on His-tagged VLNPs by a simple Add-and-Display step with high coupling efficiency and the drug was only released at low pH in a controlled manner. This approach facilitates specific attachment of any drug molecule on His-tagged VLNPs at the very mild conditions without changing the biological structure and native conformation of the VLNPs.

FA was conjugated to the HisHBcAg nanoparticles by using EDC and Sulfo-NHS. Figure 2a shows the absorbance (A) measurements taken from wavelength 240 to 600 nm. The FA-conjugated HisHBcAg nanoparticles (FA-HisHBcAg) exhibited a significantly higher A 360 value as compared with the non-conjugated HisHBcAg. About 468 FA molecules were conjugated to each HisHBcAg nanoparticle with a conjugation efficiency (CE FA ) of 4.1 ± 0.2%. FA:HisHBcAg stoichiometry was determined to be 2:1. Transmission electron microscopy (TEM) revealed that conjugation of FA did not affect the spherical structure of the HisHBcAg nanoparticles (Fig. 2b).
The internalisation efficacy of FA-HisHBcAg nanoparticles was studied using OVCAR-3 cells. The HBcAg was detected with the rabbit anti-HBcAg antibody followed by anti-rabbit IgG conjugated to Alexa Fluor ® 488 ( Fig. 3a), and the His-tag was probed with the anti-His monoclonal antibody followed by anti-mouse IgG conjugated to Alexa Fluor ® 488 (Fig. 3b). Immuno-fluorescence microscopy clearly showed that the FA-HisHBcAg nanoparticles translocated into OVCAR-3 cells by emitting bright green fluorescence. The green fluorescence was Figure 1. Schematic representation of the Add-and-Display method for immobilisation of doxorubicin non-covalently on His-tagged VLNPs. The HisHBcAg VLNP is made up of many copies of HBcAg dimers (blue). The His-tag fused at the N-terminal end of the HBcAg monomer forms trimeric spikes (magenta) and are exposed on the surface of the HisHBcAg VLNP. NTA-DOX is synthesised from nitrilotriacetic acid and doxorubicin hydrochloride. In the presence of Zn 2+ , the NTA-DOX interacts with histidine residues and is displayed on the surface of the VLNP.
observed in the cytoplasm of OVCAR-3 cells but it was not detected in the nuclei of the cells labelled with Hoechst ( Fig. 3a and b). This demonstrates that internalisation of the FA-HisHBcAg nanoparticles into the OVCAR-3 cells was mediated by FA. The cells incubated with the HisHBcAg nanoparticles which served as a negative control did not emit fluorescence, indicating the HisHBcAg nanoparticles without FA conjugation could not internalise OVCAR-3 cells (Fig. 3a).

Synthesis and characterisation of His-tagged VLNPs conjugated covalently with FA and non-covalently with NTA-DOX (FA-HisHBcAg-NTA-DOX). NTA-DOX containing a His-tag-targeting
NTA segment and a DOX segment (Fig. 4a), was synthesised from an NTA-derivative and doxorubicin hydrochloride in a two-step reaction. First, the amine group of alanine (Ala) was fully protected with ethylbromoacetate using K 2 CO 3 under reflux condition to obtain diester-Ala (compound 1, Supplementary Figure S1). The structure of compound 1 was confirmed using 1 H and 13 C nuclear magnetic resonance (NMR) spectroscopy (Supplementary Figure S2). Liquid chromatography-mass spectrometry (LC-MS) analysis of compound 1 revealed a major signal with m/z value of 262.13488 Da (M + H) + , which corresponded well with the calculated molecular mass of 261.12 Da (Supplementary Figure S3). Then, compound 1 was mixed with DOX in the presence of 4-Dimethylaminopyridine (DMAP) and N,N'-Diisopropylcarbodiimide (DIC) to synthesise NTA-DOX (Supplementary Figure S4). The structure of NTA-DOX was confirmed using 1 H and 13 C NMR (Supplementary Figure S5) and the molecular mass of the NTA-DOX was determined using LC-MS (Supplementary Figure S6). The most abundant ion detected was 731.85509 Da (M + 2H) +2 , which is in good agreement with the calculated molecular mass of 730.22 Da (Supplementary Figure S6). The NTA-DOX was incubated with the HisHBcAg and FA-HisHBcAg nanoparticles separately, in the presence of Zn 2+ and these nanoparticles were analysed with sucrose density gradient ultracentrifugation (Fig. 4b). In the absence of NTA-DOX, the HisHBcAg nanoparticles migrated into the sucrose gradient and accumulated in fractions 9-19. The migration profile is in good agreement with that demonstrated by Yap et al. 10 . In the presence of NTA-DOX and Zn 2+ , the HisHBcAg-NTA-DOX and FA-HisHBcAg-NTA-DOX nanoparticles migrated faster in the gradient and accumulated in fractions 11-22 and 13-25, respectively, indicating the presence of denser nanoparticles. Incubation of His-tagged VLNPs with NTA-DOX in the absence of Zn 2+ [HisHBcAg + (NTA-DOX)] did not show any difference in the migration profile as compared to that of HisHBcAg (Fig. 4b), demonstrating that in the absence of Zn 2+ , NTA-DOX cannot be conjugated to the HisHBcAg nanoparticles. In this experiment, His-tagged VLNP incubated with Zn 2+ (HisHBcAg + Zn 2+ ) was used as a control.
The fractions containing the HisHBcAg-NTA-DOX and FA-HisHBcAg-NTA-DOX nanoparticles were pooled separately, concentrated and viewed under a TEM. Figure 4c shows that the HisHBcAg-NTA-DOX and FA-HisHBcAg-NTA-DOX formed VLNPs morphologically similar with the HisHBcAg nanoparticles. Approximately 622 DOX molecules were conjugated non-covalently to each HisHBcAg nanoparticle with a conjugation efficiency (CE DOX ) of 7.17 ± 0.40%.
A 495 measurement was used to determine the DOX conjugated to the HisHBcAg nanoparticles and their migration profiles in sucrose gradient are shown in Fig. 5a. The result showed that the free DOX, NTA-DOX, and NTA-DOX from HisHBcAg + (NTA-DOX) samples were retained at the top fractions of the gradients (Fig. 5a). The HisHBcAg nanoparticles were not detected at A 495 . The His-tagged VLNP incubated with Zn 2+ (HisHBcAg + Zn 2+ ) showed a negligible absorbance at 495 nm. In contrast, the HisHBcAg-NTA-DOX and FA-HisHBcAg-NTA-DOX were detected in fractions 11-21 and 14-23, respectively (Fig. 5b). The positions of these peaks corresponded well with those of HisHBcAg nanoparticles measured with the Bradford assay (Fig. 4b).
Native agarose gel electrophoresis of intact VLNPs further verified that the particles were successfully conjugated with DOX ( Fig. 5c and d). In the electrophoresis, the HisHBcAg nanoparticles migrated towards the anode, while the DOX migrated towards the cathode. The gel visualised under UV light showed that the HisHBcAg-NTA-DOX and FA-HisHBcAg-NTA-DOX nanoparticles migrated to the anode and the protein bands fluoresced, demonstrating that DOX was conjugated to these nanoparticles ( Fig. 5c; lanes 2 and 3). The migration of the NTA-DOX incubated with HisHBcAg in the absence of Zn 2+ [HisHBcAg + (NTA-DOX)] was similar with the free DOX, indicating that NTA-DOX did not bind on the surface of the HisHBcAg nanoparticles in the absence of Zn 2+ . In the gel stained with Coomassie Brilliant Blue (CBB), the HisHBcAg and DOX-conjugated HisHBcAg nanoparticles (HisHBcAg-NTA-DOX) were detected (Fig. 5d). The migration of the non-conjugated HisHBcAg nanoparticles was faster than the conjugated ones (Fig. 5d), demonstrating the conjugation of HisHBcAg nanoparticles with DOX and further with FA significantly reduced the mobility of the VLNPs towards the anode.
In vitro release of doxorubicin from His-tagged VLNPs conjugated covalently with FA and non-covalently with NTA-DOX. To study the release behaviour of the HisHBcAg-NTA-DOX and FA-HisHBcAg-NTA-DOX formulations, in vitro release experiments were carried out under simulated tumour tissue conditions (pH 5.4, 37 °C) and physiological conditions (pH 7.4, 37 °C). The HisHBcAg-NTA-DOX and FA-HisHBcAg-NTA-DOX formulations exhibited a significant release of DOX at pH 5.4, in which a rapid DOX release appeared instantly upon contacting the release medium, and about 80% of the drug was released at 16 h (Fig. 6). On the other hand, the HisHBcAg-NTA-DOX and FA-HisHBcAg-NTA-DOX formulations exhibited a slow release of DOX at pH 7.4, only about 10% was released at 16 h (Fig. 6). The free DOX diffused rapidly through the dialysis membrane, resulting in 80% cumulative release of DOX after 5 h (Fig. 6). The cumulative release rate of free DOX was not significantly different in both pH solutions. Cellular uptake and cytotoxicity of His-tagged VLNPs conjugated covalently with FA and non-covalently with NTA-DOX. The cellular uptake of free DOX, HisHBcAg-NTA-DOX and FA-HisHBcAg-NTA-DOX nanoparticles in ovarian cancer and normal fibroblast cells was quantified spectrophotometrically and the localisation of these formulations in both cells was studied using live cell imaging microscopy. Quantitative data showed a 3-fold higher uptake of DOX from FA-HisHBcAg-NTA-DOX by OVCAR-3 cells compared with the addition of free DOX (Supplementary Figure S7a). However, the uptake of DOX from HisHBcAg-NTA-DOX and FA-HisHBcAg-NTA-DOX nanoparticles by 3T3 normal cells was lower compared with DOX alone (Supplementary Figure S7b). The cellular uptake of DOX, HisHBcAg-NTA-DOX and FA-HisHBcAg-NTA-DOX nanoparticles was further confirmed by live cell imaging microscopy. OVCAR-3 cells incubated with free DOX exhibited intense red fluorescence and the fluorescence intensity increased significantly when the cells were incubated with the FA-HisHBcAg-NTA-DOX nanoparticles (Fig. 7a). The results of live cell imaging microscopy provide evidence for the FR-mediated uptake of FA-HisHBcAg-NTA-DOX nanoparticles by OVCAR-3 cells. Incubation of 3T3 normal cells with free DOX exhibited a stronger red fluorescent intensity as compared to the same cells incubated with the FA-HisHBcAg-NTA-DOX nanoparticles (Fig. 7b).
MTT assay was performed to evaluate the cytotoxicity of various DOX formulations against the OVCAR-3 and 3T3 cells after 3 h of treatment. As compared with free NTA-DOX, the FA-HisHBcAg-NTA-DOX nanoparticles showed a higher cytotoxicity against OVCAR-3 cells (Fig. 8a). The IC 50DOX of free NTA-DOX and FA-HisHBcAg-NTA-DOX formulation was 1.00 ± 0.08 μM and 0.31 ± 0.03 μM, respectively (Fig. 8a). The IC 50DOX of free NTA-DOX and FA-HisHBcAg-NTA-DOX formulation in 3T3 cells was 2.92 ± 0.28 μM and 6.30 ± 0.30 μM, respectively (Fig. 8b). The HisHBcAg nanoparticles did not show any cytotoxic effect on both the cancer and normal cells. The amine group of alanine (Ala) was fully protected with ethylbromoacetate using K 2 CO 3 under reflux condition to obtain diester-Ala (compound 1). Then, the free carboxylic acid of diester-Ala was activated by N,N′-diisopropylcarbodiimide (DIC) and 4-Dimethylaminopyridine (DMAP). The activated diester-Ala was reacted with doxorubicin (DOX) to produce diester-Ala-DOX, which was then converted to dicarboxylic-Ala-DOX (NTA-DOX; compound 2) by the hydrolysis method using NaOMe in methanol.

Discussion
In conventional drug delivery systems, drugs are either packaged inside or conjugated covalently on the surface of carriers. The packaging and conjugation steps are tedious, laborious and time-consuming. To simplify these steps, a novel Add-and-Display method for immobilisation of cancer drug non-covalently on His-tagged VLNPs was established in this study. The His-tag was employed as a nanojoint to display DOX because many viral proteins including VLNPs produced via recombinant DNA technology are fused with this tag to facilitate protein detection and purification. Most importantly, the His-tag interacts tightly with a transition metal ion and this interaction is reversible by reducing the pH of a surrounding environment. As the contents of intracellular lysosomes and endosomes (pH 5-5.5) are slightly acidic compared to extracellular fluid with a physiological pH 7.4 8,9,26 , the protonation of His residues at a lower pH weakens the interaction between the His-tag and metal ions, causing the release of the attached drug, DOX.
DOX was chosen to establish the Add-and-Display method because it is a widely used anti-tumour drug for the treatments of a broad range of human cancers, including breast, ovarian, colon and prostate cancers 15 . However, this drug has several drawbacks such as poor selectivity and high cardiotoxicities 27 . Hence, the major challenge in DOX chemotherapy is to develop a safe and highly effective delivery system, which can deliver the drug into tumour cells with lower systemic toxicity and higher anti-tumour efficiency. In this study NTA-DOX was synthesised from NTA derivative and doxorubicin hydrochloride. The NTA allows the drug to interact with the His-tag via a transition metal, Zn 2+ . By using this approach, approximately 622 DOX molecules were conjugated non-covalently to each HisHBcAg nanoparticle. The HisHBcAg nanoparticles with a triangulation number T = 4, is made up of 240 subunits of HBcAg. Cryoelectron microscopy and image reconstruction revealed that the His-tag fused at the N-terminal end of HBcAg monomer formed trimeric spikes and exposed on the surface of the HisHBcAg nanoparticles 11 . Twenty and sixty of these trimeric spikes are located at the icosahedral threefold and fivefold symmetry axes, respectively 11 . Our finding showed that each His-tag associated with 2 to 3 DOX molecules.
The His-tag interacts strongly with Zn 2+ and Ni 2+ . However, Zn 2+ was chosen in the present study because several studies conducted in animals and humans demonstrated that zinc could prevent and treat a number of cancers including colorectal, pancreatic, esophageal, head and neck cancers, as well as basal cell carcinoma [28][29][30] . Zinc is a key constituent or cofactor of over 300 mammalian proteins, which indicates its importance in host defence against the initiation and progression of cancer, therefore it was used in the management and chemoprevention of cancer 28,29 . Conversely, nickel is hazardous to human body and may cause lung, nasal, sinus, throat and stomach cancers 31 .
pH-responsive nanocarriers have had a significant impact on targeted drug delivery by taking advantage of different pH gradients in extracellular environments and intracellular compartments 6 . Here, we demonstrated that the HisHBcAg nanoparticles conjugated non-covalently with NTA-DOX retained the drug at physiological conditions (pH 7.4, 37 °C), and released the drug in an acidic condition (pH 5.4). The cumulative release of DOX in the tumour tissue conditions was significantly higher (P < 0.01) than that of physiological conditions (pH 7.4, 37 °C). The release of DOX can be explained by protonation of His residues at pH lower than 6.0. The imidazole ring of His has a pKa of ~6 32 , thus at the physiological pH 7.4, the His-tag is mostly deprotonated and uncharged 14 . When the His-HBcAg nanoparticles arrive at acidic environments, such as endosomes and lysosomes, the His residues are protonated, which resulted in the release of DOX.
In order to confer cell-target specificity, FA was conjugated to the Lys residues of HisHBcAg nanoparticles to mediate internalisation of the nanoparticles into tumour cells. Accumulation of DOX in OVCAR-3 cells increased considerably when they were treated with the FA-HisHBcAg-NTA-DOX nanoparticles as compared to free DOX. The result indicated that the FA-HisHBcAg-NTA-DOX nanoparticles bound efficiently to FR, and delivered the drug into FR-positive tumour cells. However, the 3T3 normal cells treated with FA-HisHBcAg-NTA-DOX nanoparticles showed a lower DOX uptake as compared with the free DOX. The higher uptake of FA-HisHBcAg-NTA-DOX by the OVCAR-3 cancer cells compared with the 3T3 normal cells could be due to the fact that the FR is highly expressed in the tumour cells 33,34 . Furthermore, cancer cells overexpress the α-FR which has a high affinity to the free R-carboxylic acid of the conjugated FA molecules 9,35,36 . Whereas, normal cells mostly express the β-FR which has a higher binding affinity for 5-methyltetrahydrofolate, the reduced form of FA 9,35,37,38 . As such, FA conjugated drug-delivery-carrier can internalise the cancer cells at a higher rate compared with normal cells 9 . This is in good accord with our finding in which FA-HisHBcAg-NTA-DOX nanoparticles were taken up at a higher rate by OVCAR-3 cells compared with 3T3 cells.
The cytotoxicity of FA-HisHBcAg-NTA-DOX nanoparticles on OVCAR-3 cells was evaluated by MTT assay. The IC 50DOX of FA-HisHBcAg-NTA-DOX nanoparticles (0.31 ± 0.03 μM) was about 3-fold lower compared to that of IC 50DOX of free NTA-DOX (1.00 ± 0.08 μM), indicating that the former is more cytotoxic to OVCAR-3 cells compared to the latter. In contrast, the FA-HisHBcAg-NTA-DOX nanoparticles were less toxic to 3T3 cells than free NTA-DOX as the IC 50DOX of this formulation (6.30 ± 0.30 μM) increased by approximately 2-fold compared with the IC 50DOX of free NTA-DOX (2.92 ± 0.28 μM), demonstrating a conferred protection of the normal cells against the cancer drug.
In summary, a novel Add-and-Display method to conjugate DOX non-covalently on His-tagged VLNPs was established. This method could be used to display other anticancer drugs on His-tagged VLNPs, avoiding laborious and time consuming drug packaging steps. Further, this approach has numerous advantages compared to covalent linkages. It provides facile attachment of drugs on the VLNPs with mild conditions and high coupling efficiencies, without changing the structural conformation of VLNPs.

Materials and Methods
Purification of HisHBcAg VLNPs produced in E. coli. The HisHBcAg was produced in E. coli strain BL21 (DE3) harbouring plasmid pHis-β-L-HBcAg, as described by Yap et al. 10 . A fast protein liquid chromatography (FPLC) system (Akta Purifier; GE Healthcare, Sweden) was used to purify the HisHBcAg nanoparticles. The purity of the HisHBcAg was found to be more than 95% pure as analysed with SDS-PAGE. The amount of protein was quantified using the Bradford assay 39 .
Conjugation of FA to HisHBcAg VLNPs. The carboxylate group of FA was activated by Sulfo-NHS and EDC as described by Biabanikhankahdani et al. 9 . The activated carboxyl group was then reacted with the amine group of Lys residues located on the outer surface of HisHBcAg VLNPs, in sodium phosphate buffer (10 mM NaH 2 PO 4 /Na 2 HPO 4 , pH 7.4). The mixture was stirred gently for 8-16 h at 4 °C and separated with sucrose density gradient ultracentrifugation (8-40%) as described by Tan et al. 40 . A 360 of the HisHBcAg nanoparticles (HisHBcAg) and FA-conjugated HisHBcAg nanoparticles (FA-HisHBcAg) was measured using the NanoDrop TM 1000 spectrophotometer (Thermo Scientific, Rockford, IL, USA). The conjugation efficiency of FA (CE FA ) and the number of FA molecules conjugated to each nanoparticle (N FA ) were calculated using an extinction coefficient of 5312 mol −1 cm −1 , as described by Biabanikhankahdani et al. 9 .
Synthesis of NTA-DOX. Figure 4a summarises the steps used in the synthesis of NTA-DOX. Alanine (Ala; Sigma-Alderich, Louis, MO, USA) was fist reacted with ethylbromoacetate to obtain the diester compound (1). The pH of the reaction mixture was adjusted to 10.5-11.00 using K 2 CO 3 under reflux condition. The reaction was monitored by thin-layer chromatography (TLC) until the reaction was completed and the amine groups of Ala were fully protected with ethylbromoacetate. Then, the free carboxylic acid of Ala diester was activated by N,N′-diisopropylcarbodiimide (DIC) and 4-Dimethylaminopyridine (DMAP). The activated diester-Ala was reacted with DOX (Merck Millipore, Billerica, MA, USA) under controlled pH. The pH of reaction must be in the range of 8-8.5. Diester-Ala-DOX was converted to the dicarboxylic-Ala-DOX compound by hydrolysis method using NaOMe in methanol. The obtained product was purified with silica gel chromatography using MeOH:CHCl 3 (2:8) as the mobile phase (supplementary information).
Immobilisation of NTA-DOX on HisHBcAg VLNPs. The HisHBcAg VLNPs (50 nM) were mixed with NTA-DOX (15 μM) and ZnCl 2 (20 μM). The mixture was incubated at room temperature for 4 h, and at 4 °C overnight. The mixture was then dialysed against dialysis buffer (50 mM Tris-HCl, 100 mM NaCl, pH 8.0; 1 L, two times) at 4 °C. The HisHBcAg and FA-HisHBcAg nanoparticles conjugated with NTA-DOX were purified as described by Tan et al. 40 . The samples (500 μL) were applied on sucrose density gradient (8-40%, w/v), centrifuged (210,000 × g, for 5 h at 4 °C), and fractionated (400 µL per fraction). The protein concentration in each fraction was analysed with the Bradford assay 39 while the amount of DOX was measured at A 495 . Sucrose fractions containing the HisHBcAg-NTA-DOX and FA-HisHBcAg-NTA-DOX nanoparticles were then pooled, dialysed in dialysis buffer and the morphology of the nanoparticles was observed under a TEM. In this experiment, the HisHBcAg nanoparticles (HisHBcAg) and the HisHBcAg nanoparticles added with NTA-DOX in the absence of Zn 2+ [HisHBcAg + (NTA-DOX)] were also purified on sucrose density gradient and used as controls. The conjugated DOX was quantified at A 495 using an extinction coefficient of 8030 cm −1 M −1 . The conjugation efficiency of DOX (CE DOX ) and the number of DOX conjugated to each nanoparticle (N DOX ) were calculated using equations 1 and 2, respectively.

Internalisation of HisHBcAg VLNPs into OVCAR-3 cells. The internalisation of FA-conjugated
HisHBcAg nanoparticles into OVCAR-3 cells was detected using the anti-His monoclonal antibody and the anti-HBcAg serum. Native agarose gel electrophoresis (NAGE). The migration profiles of the HisHBcAg nanoparticles conjugated with NTA-DOX and FA were examined with NAGE. The samples were electrophoresed on a native agarose gel as described by Biabanikhankahdani et al. 9 and Yoon et al. 41 . DOX bands were visualised by ultraviolet illumination (UV) using the GelDoc 2000 Imaging System (Bio-Rad, Philadelphia, PA, USA), while proteins were stained with CBB.
In vitro release of DOX from HisHBcAg VLNPs. Drug release experiment was performed by the dial- and constant stirring at 37 °C. To quantify the released DOX, 1 mL of the release buffer was collected at specific time point and A 495 was measured. For each collection, the collected buffer was replaced with the same volume of fresh medium.

Localisation of HisHBcAg VLNPs in normal and cancer cells. OVCAR-3 and 3T3 cells were
subcultured in six well plates at 2.0 × 10 5 and 1.0 × 10 5 cell/mL, respectively, and incubated for 24 h. After that, each medium was replaced with fresh medium containing free DOX, HisHBcAg-NTA-DOX or FA-HisHBcAg-NTA-DOX (1 mL; with a constant concentration of 5 μg/mL DOX). The cells were incubated for 1 h, washed and fixed with 3.7% (w/v) paraformaldehyde in PBS (pH 7.4) for 10 min at 25 °C. Cell nuclei were stained with Hoechst 33342. The stained substrates were imaged by using the Olympus Live Cell Imaging (EX 480 nm and Em 535 nm ; Center Valley, PA, USA). In this experiment, the HisHBcAg nanoparticles and untreated cells served as negative controls. Statistical analysis. Statistical analysis was performed using the SPSS programme. Values of p < 0.01 are considered to be significant.