High-Level Production of Human Papillomavirus (HPV) Type 16 L1 in Escherichia coli

Abstract

Human papillomavirus (HPV), a non-enveloped, double-stranded DNA tumor virus, is a primary etiological agent of cervical cancer development. As a potential tool for prophylactic vaccination, the development of virus-like particles (VLPs) containing the HPV16 L1 capsid protein is highly desired. In this study, we developed a high-level expression system of the HPV16 L1 in Escherichia coli for the purpose of VLP development. The native gene of HPV16 L1 has many rare codons that cause the early termination of translation and result in the production of truncated forms. First, we optimized the codon of the HPV16 L1 gene to the preferable codons of E. coli, and we succeeded in producing the full-size HPV16 L1 protein without early termination. Next, to find the best host for the production of HPV16 L1, we examined a total of eight E. coli strains, and E. coli BL21(DE3) with the highest yield among the strains was selected. With the selected host-vector system, we did a fed-batch cultivation in a lab-scale bioreactor. Two different feeding solutions (complex and defined feeding solutions) were examined and, when the complex feeding solution was used, a 6-fold higher production yield (4.6 g/l) was obtained compared with that with the defined feeding solution.

Introduction

Human papillomavirus (HPV) is a non-enveloped, doublestranded DNA tumor virus that is in the subfamily of papovaviruses and is a primary etiological agent of cervical cancer development [28]. More than 120 types of HPV have been reported, and among them, types 16 and 18 are the genotypes most frequently associated with cervical cancer development, which is the second most common cancer in women worldwide [6,31]. Currently, the prophylactic vaccination, which can promote a reduction in disease burden, is the most effective prevention method for HPV infection, and for this reason, the development of virus-like particles (VLPs) against HPV is considered as an attractive tool owing to their easy production and high safety [24].

HPV has two major structural proteins, L1 and L2. Among them, the L1 protein is the major structural protein (360 copies) and assembled into 72 pentameric capsomeres that are arranged in an icosahedral array, whereas the L2 protein, the minor capsid protein (72 copies), is not necessary to form viral particles [22]. Recombinant L1 protein can be self-assembled to form VLPs when expressed in heterologous hosts [4,7,32]. The synthesized HPV VLPs have a size of about 60 nm in diameter, which is almost the same as that of native virions (50-60 nm), and they have already been proven to be immunologically comparable to natural virions. Various heterologous hosts, including bacteria, yeast, insects, plants, etc., have been used for the production of HPV VLPs [1-3,5,20]. Currently, yeast and insect cells have been preferentially used for VLP production: two VLP-based prophylactic vaccines, including Gardasil (Merck Sharpe and Dohme) and Cervarix (GlaxoSmithKline), which have already been released to market, are also produced in insect and yeast cells, respectively [12,27]. Although the use of eukaryotic cells can provide highly effective vaccines, the production of VLPs using eukaryotic cells has several drawbacks – mainly high production and purification costs [17,23]. In contrast, bacterial hosts (especially Escherichia coli) have many advantages in the production of heterogeneous proteins: fast growth, easy gene manipulation, low production cost, easy scale-up, etc. [25,30]. Generally, HPV L1 proteins are not self-assembled into VLPs in bacterial hosts, but the produced proteins can be assembled into VLPs by in vitro assembly after purification of the L1 proteins. To achieve a higher yield of VLP, we first need to develop an efficient host-vector system for high-level gene expression.

In this study, we optimized the codon usage of the HPV16 L1 gene for high-level production in E. coli. With the codon-optimized gene, a gene expression system was constructed, and gene expression was examined in different E. coli host strains to determine the best host for the production of HPV16 L1. Finally, with the selected hostvector system, we also performed a fed-batch cultivation for the large-scale production of the HPV L1 protein. Two different feeding solutions were tested and the culture conditions were optimized.

 

Materials and Methods

Bacterial Strains and Plasmids

All strains and plasmids used in this study are listed in Table 1. E. coli MG1655 was used for gene cloning and protein production. The other E. coli strains, including DH5α, JM109, Jude-1, XL1-Blue, BL21(DE3), Origami(DE3), and Rosetta(DE3), were used for the production of the HPV L1 proteins. A polymerase chain reaction (PCR) was performed using a C1000 Thermal Cycler (Bio-Rad, R ichmond, CA, U SA) with the P rimeStar H S polymerase (Takara Bio, Shiga, Japan). Primers used for the gene amplification are listed in Table 2. The wild-type HPV type 16 L1 gene was kindly provided by Dr. CH Kim (Korea Research Institute of Biology and Biotechnology, Daejeon, Korea), and the codon-optimized HPV L1 gene was synthesized by GenScript Co. (Piscataway, NJ, USA). Both wild-type and codon-optimized HPV L1 genes were amplified by PCR. The PCR products were digested with BamHI and NotI restriction enzymes and then were ligated into the pGEX-4T-1 vector, yielding pGST-non-HPV and pGST-opt-HPV, respectively. The nucleotide sequence used for the comparison among the various types of HPV in this study was deposited in GenBank under the accession no. NC_001526. All DNA manipulations, including restriction digestion, ligation, and agarose gel electrophoresis, were done according to standard procedures [21].

Table 1.aThe Coli Genetic Stock Center, Yale University, USA. bStratagene Cloning Systems, La Jolla, CA, USA. cNovagen, Inc., Madison, WI, USA. dGE Healthcare, Pittsburgh, PA, USA.

Table 2.aRestriction enzyme sites are shown in bold.

Cultivations in Test Tubes

E. coli BL21(DE3) cells harboring plasmids were inoculated into liquid Luria-Bertani (LB) medium (10 g/l tryptone, 5 g/l yeast extract, and 10 g/l NaCl) containing 2% (w/v) glucose and 100 μg/ml ampicillin. After overnight cultivation at 37℃ and 200 rpm, 100 μl of inoculum was transferred into 5 ml of fresh LB medium in a test tube and incubated under the same conditions. When the optical density reached an OD600 of 0.5–0.7, isopropyl-β-ᴅ-thiogalactopyranoside (IPTG) was added to a final concentration of 1 mM to induce gene expression. After incubation for an additional 6 h, the cells were harvested by centrifugation at 6,000 rpm for 10 min at 4℃, and cell pellets were stored at −20℃ until further analysis.

Fed-Batch Cultivation

Fed-batch cultivations of E. coli BL21(DE3) harboring pGST-opt-HPV were done in a 5 L bioreactor (BioCNS, Daejeon, Korea) with R/2 medium [13]. The R/2 medium (pH 6.8) contained 6.75 g/l KH2PO4, 2 g/l (NH4)2HPO4, 0.85 g/l citric acid, 0.7 g/l MgSO4·7H2O, 5 ml/l Trace metal solution (TMS; 10 g/l FeSO4·7H2O, 2.2 g/l ZnSO4·7H2O, 2 g/l CaCl2·2H2O, 1 g/l CuSO4·5H2O, 0.58 g/l MnSO4·5H2O, 0.1 g/l (NH4)6Mo7O24·4H2O, and 0.02 g/l Na2B4O7·10H2O adjusted with HCl), and 20 g/l glucose as the sole carbon source. The bioreactor was equipped with a built-in digital controller for pH, temperature, agitation, and dissolved oxygen (DO) and with peristaltic pumps for ammonia and the feeding solutions. A seed culture was prepared in four 250 ml baffled flasks containing 50 ml of R/2 medium in each flask. The working culture volume in the bioreactor was 2 L of R/2 medium containing 10% (v/v) inoculum and 100 mg/l ampicillin. The pH was controlled at 6.8 by adding 50% (v/v) ammonia solution when the pH was lower than 6.77 and by adding nutrient solution when the pH was higher than 6.86 through the controller. For the supplementation of nutrients during the cultivation, two feeding solutions were tested: i) complex feeding solution (500 g/l glucose, 75 g/l yeast extract, and 20 g/l MgSO4·7H2O) and (ii) defined feeding solution (700 g/l glucose and 20 g/l MgSO4·7H2O). The temperature was maintained at 37℃, and the DO concentration was kept at 40% by automatically increasing the agitation rate up to 1,000 rpm and by mixing pure oxygen into the fed-batch cultures during the cultivation period. Sterilized antifoam agent (Sigma-Aldrich) was manually added. Culture samples were collected at an OD600 of 4 periodically during the cultivation, followed by the cell pelleting, and the supernatants were stored at −80℃ for further analysis.

SDS-PAGE and Western Blot Analysis

The harvested cells at an OD600 of 4 were disrupted with 600 μl of BugBuster Master Mix (Novagen, San Diego, CA, USA) and 80 μl of each cell lysate was saved as a total fraction. The remainder of all the samples was centrifuged to separate the soluble and insoluble fractions (10,000 rpm, 10 min at 4℃). To analyze the proteins in the culture medium, 300 μl of each culture supernatant was mixed with the same volume of cold acetone. After storing at −20℃ for 10 h, proteins were precipitated by centrifugation (13,000 rpm, 30 min at 4℃), and the pellets were air-dried and then solubilized in 8 M urea solution. Each of the total, soluble, insoluble, and supernatant fractions was loaded on a 12% SDS-PAGE gel, and after gel electrophoresis, all protein bands were transferred to a polyvinylidene fluoride (Roche, Basel, Switzerland) membrane with Bio-Rad Trans-blot SD (Bio-Rad) at 70 mA per gel for 90 min. The membrane was blocked with Tris-buffered saline containing Tween-20 (TBS-T; 10 mM Tris, 150 mM NaCl, and 0.05% Tween-20, pH 8.0) with 5% (w/v) skim milk at room temperature for 1 h, followed by incubation with a specific antibody dissolved in blocking solution (TBS-T) for 1 h. For the detection of GST tag, anti-GST antibody conjugated with horseradish peroxidase (HRP; Abcam, Cambridge, MA, USA) was used. For the detection of HPV L1, anti-HPV L1 antibody (Santa Cruz Biotechnology, Inc., Dallas, TX, USA) and peroxidase-conjugated anti-mouse antibody (Jackson ImmunoResearch Labs, West Grove, PA, USA) were used as the primary and secondary antibodies, respectively. After incubation, the membrane was washed with TBS-T for 10 min at least four times and treated with ECL western blotting detection reagent (Bionote, Hwaseong, Korea) for detection of the proteins on X-ray film.

 

Results and Discussion

Synthesis and Expression of the Codon-Optimized HPV L1 Gene

It is well known that codon usage in a gene greatly influences the translation rate and overall yield of protein production [9,26]. The presence of rare codons (particularly tandem repeats of rare codons) causes slow translation or early termination of translation and results in a low yield of protein production. Therefore, optimization of codon usage for a target gene is highly recommended to improve the production yield of target proteins [16]. In this regard, we first examined the codon usage of the HPV L1 gene (1,593 bp) using E. coli codon usage analyzer 2.1 (http://www.faculty.ucr.edu/~mmaduro/codonusage/usage.htm). From this analysis, it was found that 27 codons among the 531 codons were rare codons with frequencies below 5% (Fig. 1). In particular, several tandem/triple rare codons, which are highly critical in translation as described above, were also found (Fig. 1). To investigate possible early termination of the HPV L1 gene in E. coli, an expression system with the HPV L1 gene (pGST-non-HPV) was constructed in pGEX-4T-1 in which the HPV L1 gene was linked to the 3'-end of the GST gene, and its gene expression was controlled by an IPTG-inducible tac promoter. The constructed pGST-non-HPV was transformed into E. coli BL21(DE3), and the production of GST-fused HPV L1 was examined in shake-flask cultivations. From the western blot analysis after SDS-PAGE, it was clearly observed that, instead of the entire HPV L1 protein fused to GST (approx. 85 kDa), a smaller protein (approx. 50 kDa) was detected as the sole protein band in the western blot (Fig. 2). This result indicates that the translation of HPV L1 might be terminated early during the translation process. To prevent early termination and produce the full-size HPV L1 protein in E. coli, all the codons of the HPV gene were optimized to E. coli-preferable codons. The non-codon-optimized and codon-optimized sequences of DNA and corresponding amino acids are shown in Fig. 3, and the changed codons are shown in red. With the codon-optimized sequences, a gene expression system was constructed. The resulting plasmid, pGST-opt-HPV, was transformed into E. coli BL21(DE3), and the expression level was compared with that of the non-optimized gene. We observed that the full-size HPV L1 protein (85 kDa) was successfully produced in E. coli harboring the pGST-opt-HPV, and the expression level was also higher than that of the non-optimized gene (Fig. 2).

Fig. 1.Analysis of codon usage in the original gene encoding HPV16 L1. X- and Y-axes indicate the amino acid sequences and codon frequency (%) in E. coli codon usage. Red-color bars indicate the rare codons with a frequency below 5% in E. coli. Red-color arrows indicate the tandem repeats of rare codons.

Fig. 2.Analysis of the production of GST-fused HPV16 L1 proteins in shake flask cultivations using (A) SDS-PAGE (Coomassie-blue staining) and (B) western blot with anti-GST-HRP. Lane M, molecular weight size markers (kDa); lane 1, BL21(DE3) (no plasmid); lane 2, BL21(DE3) harboring pGST-non-HPV; lane 3, BL21(DE3) harboring pGST-opt-HPV. Closed and open arrowheads indicate full GST-tagged HPV L1 and its truncated form, respectively.

Fig. 3.Comparison of gene sequences between non-optimized HPV type 16 L1 gene and the codon-optimized one and their corresponding amino acid sequences. Changed codons are indicated in red.

Comparison of HPV L1 Expression Levels in Various E. coli Strains

For higher-level production of recombinant proteins in E. coli, it is important to select the best strain for the production of target protein because using the same expression system and cultivation under the same conditions for different strains may result in very different yields of protein production [8,14,19]. To determine which E. coli strain has the highest production of HPV L1, a total of eight different strains, BL21(DE3), Jude-1, XL1-Blue, MG1655, Origami(DE3), DH5α, JM109, and Rosetta(DE3), which all are widely used for recombinant protein production, were examined. All strains were transformed with pGST-opt-HPV in which the codon-optimized HPV L1 gene was linked to the C-terminus of the GST-tag under the tac promoter. After cultivation in shake flasks, cells were harvested at an OD600 of 2 and disrupted for comparison of the strains with SDS-PAGE and western blot analysis. Most of the examined strains except for the Rosetta(DE3) strain had relatively good production yields (Fig. 4). In several strains, including Jude-1, XL1-Blue, and MG1655, truncated bands were also found believed to be GST itself (~27 kDa). Among the eight strains, the E. coli BL21(DE3) strain had the highest production level of GST-fused HPV L1, and no degradation products were observed. Based on this result, the E. coli BL21(DE3) strain was selected for the production of the GST-fused HPV L1 in the fed-batch cultivation.

Fig. 4.Comparison of HPV L1 production in eight different E. coli strains by (A) SDS-PAGE (Coomassie brilliant blue staining) and (B) western blot with anti-GST-HRP. Lane M, molecular weight size markers (kDa); lane 1, BL21(DE3); lane 2, Jude-1; lane 3, XL1-Blue; lane 4, MG1655; lane 5, Origami(DE3); lane 6, DH5α; lane 7, JM109; lane 8, Rosetta(DE3). Closed arrowheads indicate GST-tagged HPV L1.

Fed-Batch Cultivation for the Production of HPV L1

For large-scale production of GST-tagged HPV L1 proteins, fed-batch cultivations were done with E. coli BL21(DE3) harboring pGST-opt-HPV in a 5-L-scale bioreactor system. Cells were cultivated in R/2 medium. In fed-batch cultivations, the nutrient feeding strategy is important in the final production yield of target proteins [14,15,18]. Two different feeding solutions, complex and defined feeding solutions, were examined, and the production yields of HPV L1 proteins were compared. For both fed-batch cultivations, cells were cultivated at 37℃, and IPTG was added to a final concentration of 1 mM when the cell density reached an OD600 of 80. When the complex feeding solution containing yeast extract with glucose was supplied, cells continued to grow up to an OD600 of 159.6 at 12 h (3 h after induction) and then the cell density decreased (Fig. 5A). The cell specific growth rate (μ) was 0.406 h-1. Immediately after IPTG induction at an OD600 of 80, HPV L1 proteins began to be produced and the protein content continued to increase (Fig. 5B). A maximum production yield as high as 4.6 g/l was achieved at 13 h (4 h after induction). Next, when the defined feeding solution was used, the cells also grew well with a high growth rate (μ = 0.417 h-1) (Fig. 6A). Compared with the fed-batch cultivation with the complex feeding, a higher maximum cell density (OD600 of 172) was obtained at 11 h (3 h after induction). Immediately after IPTG induction at an OD600 of 80, HPV L1 proteins began to be produced and the protein content continued to increase (Fig. 6B). A maximum production yield as high as 0.77 g/l was achieved at 11 h (3 h after induction) with a production yield almost 6-fold lower than that by fermentation with the complex feeding solution.

Fig. 5.Fed-batch cultivation with complex feeding condition. (A) Time profiles of cell growth and production yield of HPV L1. Closed circles indicate cell growth (OD600), and open circles indicate production yield of HPV L1 (g/l). Dashed line indicates the induction point. (B) SDS-PAGE analysis of periodically harvested samples after induction with Coomassie staining. Lanes 1 to 6, 0 h, 1 h, 2 h, 3 h, 4 h, and 5 h, respectively. Closed arrowhead indicates GST-fused HPV L1.

Fig. 6.Fed-batch cultivation with defined feeding condition. (A) Time profiles of cell growth and production yield of HPV L1. Closed circles indicate cell growth (OD600), and open circles indicate production yield of HPV L1 (g/l). Dashed line indicates the induction point. (B) SDS-PAGE analysis of periodically harvested samples after induction with Coomassie staining. Lanes 1 to 6, 0 h, 1 h, 2 h, 3 h, 4 h, and 5 h, respectively. Closed arrowhead indicates GST-fused HPV L1.

In conclusion, for the production of HPV L1 proteins in E. coli, we developed an E. coli host-vector system enabling high-level expression of the entire gene without early termination. Fed-batch cultivations were done with the selected host-vector system (E. coli BL21(DE3) harboring pGST-opt-HPV), and, by supplying a complex feeding solution, the production of GST-fused HPV L1 as high as 4.6 g/l was achieved. To the best of our knowledge, this is the highest production yield for the production of HPV-related proteins. The purification of GST-fused proteins produced in E. coli is well established, and after purification, the HPV-L1 proteins can be used for in vitro assembly of VLPs. Although E. coli cannot produce the entire VLP and additional steps are required for VLP assembly, other properties such as low-cost, and high-level production make this host-vector system attractive for commercial applications. We believe our host-vector system could be a good tool in the production of HPV vaccines.