HepG2 cell was cultured with Dulbecco’s Modified Eagle Medium, 10% FBS, and penicillin/streptomycin until the confluent reached 90%. Then, 10⁶ cells were harvested and isolated whole RNA using High Pure RNA Tissue Kit (Roche) and cDNAs were synthesized by reverse transcription reaction using RevertAid First Strand cDNA Synthesis Kit (Thermo) under the manufacturer’s instructions.

GPC3-Csub coding sequence was amplified by PCR reaction carried out by Phusion High-Fidelity DNA Polymerase (Thermo) with the cDNAs of HepG2 cell line as templates and the primer pair Forward 5’ (CGCGGATCCAGATCTGC-TTATTATCCTGAAGAT) and Reverse 5’ (CCGCTCGAGGTGCACCAGGA). We designed the reverse primer that omitted the stop codon so our expressed protein will have 6xHis-tag fragment that can facilitate the purification steps. The PCR products were analyzed by gel electrophoresis and the gene of interest was isolated by gel extraction using GeneJET Gel Extraction Kit (Thermo). GPC3-Csub fragment was inserted into plasmid PET28a+ by restriction enzymes XhoI and BamHI and ligated by T4 ligase. The constructed plasmid has been sequenced and analyzed with the gene database by using the BLAST tool of NCBI (http://www.ncbi.nlm.nih.gov).

His-tag GPC3-Csub was expressed in Escherichia coli BL21(DE3) strain. Approximately 100 µg of pET28a(+)-GPC3-Csub plasmid was transformed to competent E. coli BL21 (DE3) by heat shock at 42˚C, 2 minutes, followed by incubation at 37˚C for one hour. The colony that appeared on LB/kanamycin plate was picked up and inoculated in 10 mL of LB medium with kanamycin (±3% Ethanol, Merck) as starter culture. Different culture conditions were investigated to maximize protein production (Table 1). In general, the starter culture was added to fresh LB medium (ratio 1:50) with or without an additional 3% Ethanol, inoculated at 37˚C, 180 rpm shaking. Until OD_600nm reached the target value, 1 mM of IPTG (Bioline) was added to the medium to induce protein production. The induced medium then was cultured at different time and temperature conditions, 180 rpm shaking as shown in Table 1.

The bacteria biomass was harvested by centrifugation and resuspended in lysis buffer II (50 mM Tris pH 8.0, 10% glycerol (Merck), 0.1% Triton X-100 (Biobasic), 5 mM MgCl2 (Merck), 1 mM NaCl (Merck) with the ratio 2 mL lysis buffer II/10 mL culture, then added 1 mM PMSF solution (PMSF, isopropyl alcohol) and lysozyme (1 mg lysozyme/1 mL lysis buffer II). The mixture was then incubated on ice for 1 hour before freezing by liquid nitrogen and thawing in 37˚C. After that, adding DNase (3 U/1 mL lysis buffer II) (Roche) to the solution and incubate at 37˚C for 30 minutes to digest genomic DNA. The mixture was frozen and thawed three more times to facilitate the cell wall decomposition, and then sonicated on ice at ultrasonic level 6.0, 10 × 30 seconds, followed by centrifugation at 12,000 g, 20 minutes. The collected pellet was dissolved in a buffer containing Urea, SDS 10%, pH pH12.

Ni/NTA affinity purification was performed on an AKTA FPLC system using 1 mL HisPurTMNi-NTA Chromatography Cartridge column (Thermo Scientific). After equilibrating with binding buffer (300 mM NaCl, 20 mM NaH₂PO₄, 10 mM Imidazole, 8M Urea), protein sample was applied to Ni-NTA column. The

^(*): culture until this OD to do expansion culture; ^(**): culture until this OD to do induction with IPTG 1 mM; ^(o/n): culture overnight.

column was then washed with binding buffer and refolded gradually with refolding buffer (300 mM NaCl, 20 mM NaH₂PO₄, 10 mM Imidazole). The protein was then eluted with elution buffer (300 mM NaCl, 20 mM NaH₂PO₄, 500 mM imidazole). The total protein after elution was combined and exchanged with refolding buffer (to reduce Imidazole concentration) using ultrafiltration membrane with a cut-off 3000 MWCO (Da).

Protein was subjected to SDS-PAGE. The gel then was incubated with staining solution (0.6 mM Coomassie brilliant blue (CBB G-250, Merck) in 50% methanol (Merck), 10% acetic acid (Merck)) for 1 hour. After that, the gel was de-stained with a solution containing 25% methanol and 75% acetic acid.

Protein was subjected to SDS-PAGE. The gel then was fixed with a fixer solution (50% Methanol (Merck), 12% acetic acid (Merck), 0.05% formalin (Merck)) for 2 hours, washed three times with 35% Ethanol (Merck) for 30 minutes. The gel was sensitized with 0.02% Na₂S₂O₃ (Biobasic) and washed with distilled water three times for 5 minutes. Then, the gel was stained with silver nitrate solution (0.2% AgNO₃ (Merck), 0.076% Formalin), washed two times with distilled water for 1 minute before developing in developing buffer (6% Na₂CO₃ (Merck), 0.05% Formalin, 0.0004% Na₂S₂O₃) until protein bands were visible and stop the reaction by 50% Methanol, 12% acetic acid for 5 minutes. ImageJ software was used to estimate the protein concentration before and after purification.

Protein was subjected to SDS-PAGE and then electro-transferred to Nitrocellulose membrane (Thermo). Membrane was blocked with skim milk for 30 minutes; wash 3 times by TBST before adding primary antibody His-tag (Invitrogen, 1:5000). Bound antibodies were detected with horseradish peroxidase-conjugated anti-mouse IgG (Invitrogen, 1:5000), and ImageQuant^TM LAS 500 System (GE Healthcare) was used for the detection of antibody reaction.

Figure 1A shows the gel electrophoresis results of the PCR products that amplified the GPC3-Csub fragment from cDNA of HepG2 cells. We observed a clear band between 600 bp and 700 bp, which is expected to be the size 669 bp of the GPC3-Csub fragment. We then isolated this PCR product fragment and cloned it into the PET28a(+) vector by restriction enzyme digestion with BamHI and XhoI and ligation with T4 ligase. The PCR screening results of colonies show four clones that had positive results with the GPC3-Csub gene among six tested colonies (Figure 1B). We then picked randomly the clone at line 4 for growing to isolate the constructed plasmid. The size of the insert was confirmed by restriction enzymes BamHI and XhoI (Figure 1C), and then the plasmid was sequenced and aligned with the NCBI Genebank database by BLAST tool. Figure 2 shows that our insert gene is 100% identical to the human GPC3-C sequence (GenBank Accession No: AK222766). Thus, we have successfully constructed the expression plasmid PET28a(+)-GPC3-Csub.

After successfully constructing recombinant plasmid to produce the C-subunit of GPC3, this protein was induced by IPTG and the protein expression was confirmed by Western Blot (Figure 2B, Figure 2D). To maximize protein production, different culture conditions were applied. There are several factors that need to be considered to achieve the highest protein products such as cell density (OD_600nm of medium in the starter culture and expansion culture), temperature, culture time, shaking rate… In this experiment setting, we first investigated how cell density, temperature, and time affect the protein expression. The results showed that GPC3-Csub protein was in inclusion bodies and when induction by IPTG at 37˚C, 4 hours, this protein expression was much higher than at 26˚C, 24 hours (Figure 2A, Figure 2B). Especially, at 37˚C, 4 hours culture condition, even incubation overnight when starter OD_600nm reached to death phase, protein expression level was not affected as compared to the differences of expansion OD_600nm. When expansion OD_600nm reached 0.8 (final-log phase to death phase), the targeted protein band was significantly lower as compared to OD_600nm 0.4 (mid-log phase). In general, the GPC3-Csub protein can be achieved the most when cell density reaches OD_600nm 0.4 before induction with 1 mM IPTG for 37˚C, 4 hours.

Recently, with the development of biofuel products, there are some researchers have investigated and concluded that the using of ethanol can increase the bioproducts from E. coli [5] . The presence of ethanol can increase DNA synthesis and plasmid number, which facilitates the increase of protein expression with established investigation in laboratory by Chhetri and colleagues [6] . Therefore, we tested the effect of ethanol on protein expression with 3% ethanol added to LB culture medium from the beginning of the culture and induction process.

With the addition of ethanol, after induction with 1 mM IPTG, culturing medium at 26˚C, 24 hours can increase GPC3-Csub protein expression as compared to 37˚C, 4 hours culture condition (Figure 2C, Figure 2D). We still also observed that the high cell density after expansion culture (OD_600nm 1.0) significantly reduced GPC3-Csub protein expression. Especially, when considering other protein bands in the inclusion proteins, the expression of GPC3-Csub was strongest in the total protein expression as compared to other culture conditions. Therefore, we chose expansion OD_600nm 0.4, 26˚C, 24 hours as our final culture conditions for further experiments.

After confirming culture conditions which led to achieve the most amount of GPC3-Csub inclusion body (IB), the solubilization of this IB was examined in different conditions. Firstly, as a well-known reagent for solubilization of inclusion protein, we used 8 M urea to dissolve protein after sonication. Urea 8 M is known as a traditional chaotropic agent to solubilize inclusion proteins by disrupting the hydrogen bonds in the networking of water molecules, reducing hydrophobic effect which maintains macromolecular structure as protein [7] . However, in our study, the inclusion bodies may be formed in more complex structures, which made the protein could not dissolve completely by only using 8 M urea. Therefore, GPC3-Csub together with other proteins were still in the pellet (Figure 3). Therefore, we tried using freeze-thawing method in combination with the addition of lysozyme to disrupt cell membrane proteins to expose internal proteins to lysis buffer. After the first free-thaw cycle, DNase was treated to degrade internal genome which also indicated that the cell membrane was broken down enough for releasing the composition inside their cytosol as compared to the old lysis buffer with sonication method (Figure 3). To check whether the sonication method after free-thawing process can help in separating GPC3-Csub to lysis buffer, we did sonication with the same level and time as previous method, and the protein after this treatment was even partially degraded without improving the solubilization of the IB.

To examine the role of denaturant in precipitate dissolution, urea was added with concentrations of 4 M, 6 M, and 8 M, respectively to the refolding buffer. The results in Figure 4A show that, under the condition of refolding buffer without the addition of urea, protein is almost present in pellet with a very small part in supernatant. As the concentration of urea gradually increased, the amount of protein dissolved into the soluble phase also increased. It can be seen through the density of the 30 kDa target protein band and the protein size between 40 - 50 kDa. This may be because the high concentration of urea makes it easier to denature proteins. However, the precipitated phase was still largely protein present, which was visible in the 30 kDa target protein, so it was necessary to examine whether increasing the pH value would lead to better conversion of the protein to the soluble phase. Besides, the results of Figure 4B showed that the protein was dissolved in urea-supplemented refolding buffer in pH 12 more than pH 7.4, but still had much amount of protein in pellet. It indicated that the high pH may help in increasing the negative charge of the inclusive proteins so that the urea 8 M can break the hydrogen bonds among these partially folded aggregates [8] ; We observed that the concentration of urea supplemented was proportional to the amount of protein dissolved into the soluble phase (can be seen in the protein line between 40 and 50 kDa). Moreover, we demonstrated that the addition of SDS detergent at a concentration of 10% to the refolding urea buffer 8 M, pH 12 helped to dissolve the protein better than no addition. This can be seen clearly that the amount of protein GPC-Csub in the soluble phase when treated with SDS is much higher than that without SDS and similarly, the protein in the pellet phase is less when SDS is added.

After identifying optimum conditions for expressing and solubilizing, protein GPC3-Csub was subsequently purified using affinity chromatography method. The pH of medium then was immediately neutralized as well as SDS concentration was reduced to 2% by diluting 5 times the supernatant with binding buffer (8 M urea in PBS with pH 7.4) to facilitate the binding of targeted protein on Ni-NTA column. GPC3-Csub protein was purified using 1 mL HisPurTM-Ni-NTA Chromatography Cartridge column managed by FPLC system (Figure 5A). The protein after purification was checked by Silver Staining (Figure 5B, Figure 5C) with expected GPC3-Csub protein was eluted starting from approximately 250 mM Imidazole until 500 mM Imidazole. The GPC3-Csub protein after buffer exchange was then confirmed by Western Blotting (Figure 5D).