Skip to main content
NCBI home page
Protein Science : A Publication of the Protein Society logoLink to Protein Science : A Publication of the Protein Society
. 2016 Jun 13;25(8):1507–1516. doi: 10.1002/pro.2953

Design strategies to address the effect of hydrophobic epitope on stability and in vitro assembly of modular virus‐like particle

Alemu Tekewe 1, Natalie K Connors 1, Anton P J Middelberg 1, Linda H L Lua 2,✉
PMCID: PMC4972206  PMID: 27222486

Abstract

Virus‐like particles (VLPs) and capsomere subunits have shown promising potential as safe and effective vaccine candidates. They can serve as platforms for the display of foreign epitopes on their surfaces in a modular architecture. Depending on the physicochemical properties of the antigenic modules, modularization may affect the expression, solubility and stability of capsomeres, and VLP assembly. In this study, three module designs of a rotavirus hydrophobic peptide (RV10) were synthesized using synthetic biology. Among the three synthetic modules, modularization of the murine polyomavirus VP1 with a single copy of RV10 flanked by long linkers and charged residues resulted in the expression of stable modular capsomeres. Further employing the approach of module titration of RV10 modules on each capsomere via Escherichia coli co‐expression of unmodified VP1 and modular VP1‐RV10 successfully translated purified modular capomeres into modular VLPs when assembled in vitro. Our results demonstrate that tailoring the physicochemical properties of modules to enhance modular capsomeres stability is achievable through synthetic biology designs. Combined with module titration strategy to avoid steric hindrance to intercapsomere interactions, this allows bioprocessing of bacterially produced in vitro assembled modular VLPs.

Keywords: rotavirus, synthetic biology, linkers, module titration, co‐expression, Escherichia coli

Introduction

Virus‐like particles (VLPs) are highly organized multimeric protein complexes that self‐assemble from viral structural protein(s) without the viral genome. VLPs are undergoing research as tools for vaccination, gene therapy, drug delivery, diagnostics, nanomaterials, and immune therapy.1, 2 Particularly, their application in the field of vaccinology increased interest following the successful development and approval of three VLP‐based vaccines for human use against hepatitis B virus infection,3, 4 human papillomavirus‐induced cervical cancer,5 and hepatitis E virus infection.6 Numerous other VLP‐based vaccines against many infectious diseases have shown promising results under preclinical and clinical studies.7

VLP subunits, termed capsomeres, have also gained attention for their potential as alternative second‐generation cheap vaccines to VLPs.8, 9, 10, 11, 12, 13 A human papillomavirus 16/18 L1 capsomere vaccine is currently undergoing a phase II clinical trial.14 Under preclinical studies, human papillomavirus L1/L2 capsomeres11, 15, 16, 17 and modular murine polyomavirus VP1 capsomeres8, 10, 18 have shown promising immunogenicity and protective efficacy results in different animal models.

Both VLPs and capsomeres can serve as platforms for surface display of foreign epitopes or antigens in a modular architecture.8, 10 The murine polyomavirus major capsid protein VP1 has been used for presentation of a single or multiple modules of M2e‐peptide epitope from influenza A virus8, 10 and J8i‐peptide epitope from group A streptococcus M1 protein,8, 19 respectively. These modular architectures were produced using a low‐cost E. coli expression system and purified in high yields using both chromatographic8, 10, 19, 20, 21 and non‐chromatographic methods.22 Purified modular VP1 capsomeres can form VLPs via in vitro assembly in a cell‐free bioreactor.8, 23 This approach reduces VLP manufacturing and processing costs often linked to eukaryotic expression systems,24 thus enables low‐resource countries to use the final product at affordable costs.13 However, the production of a modular capsomere from E. coli and the in vitro modular VLP assembly can highly be dependent on the physicochemical properties of the inserted modules. As previously observed, hydrophobic modules at the surface‐exposed loop affected the capsomere stability during downstream processing.25, 26

Although nature employs hydrophobicity for its role in protein–protein interactions27 and in stabilization of proteins,28, 29, 30 inserting hydrophobic modules into the surface‐exposed loop of viral capsid protein can significantly change protein surface hydrophobicity. The surface hydrophobicity is often associated with changes in protein conformations31 and related to the ease with which a protein unfolds.32 Engineering surface‐exposed hydrophobic modules is highly desirable to promote correct protein folding during expression and to maintain the solubility and stability of modular capsomeres, thus allowing the formation of modular VLPs via in vitro assembly. Insertion of negatively charged residues at both ends of hydrophobic protein domains has shown to enhance the stability of proteins.26, 33 Polar and/or charged amino acids, particularly, glutamic acid, aspartic acid and serine, have contributed most favorably to protein solubility at high net charge.34

Linkers are the other important elements that have gained considerable success in construct design for the production of proteins.35 Linkers can improve protein stability,36, 37 folding, expression, and purification yield,35, 38 and for targeting fusion proteins to specific sites in vivo to increase their desired biological activity.35 Linkers are used to separate different moieties of fusion proteins spatially in order to alleviate structural perturbation to moieties.36, 39 More recently, Lua et al.39 designed longer linkers to ensure structural separation and independence between a rotavirus (RV) 18 kDa VP8* antigenic module and the base VLP. Alleviating structural perturbations via incorporation of longer linkers and eliminating steric barrier by VP8* module titration using a baculovirus‐insect cell co‐expression strategy have facilitated VLP assembly in vivo.

In this study, we described three specific module designs containing hydrophobic RV10 peptide from a RV VP8 subunit protein,40 to address the effect of hydrophobic modules on the stability and in vitro assembly of VLPs. In addition, we demonstrated that module titration approach using E. coli co‐expression strategy is feasible, and necessary for the production of in vitro assembled modular VLP displaying hydrophobic RV10. A synthetic biology‐based module design and E. coli co‐expression strategy developed in this study can further enable a rapid and a low‐cost processing of modular capsomeres and VLPs presenting modules with adverse physicochemical properties for low‐cost vaccine delivery against target diseases at a global scale.

Results and Discussion

Module designs

Murine polyomavirus VP1 was engineered to present heterologous antigenic modules on its surface loops.8 The surface loop S4, position 293 of VP1, contains an AfeI site to allow molecular insertion of foreign sequences [Fig. 1(A)]. This is a VLP forming platform. For the capsomere platform [Fig. 1(C)], the first 28 and the last 63 amino acid residues of VP1 were excluded. ΔVP1 contains engineered PmlI‐, NaeI‐, AfeI‐, and SnaBI‐restriction enzyme sites at positions 28 (N‐terminus), 86 (S1 loop), 293 (S4 loop), and 380 (C‐terminus) of VP1 positions, respectively, for insertion of foreign modules. Previous studies have demonstrated good tolerance of heterologous peptide module insertions on both VLP and capsomere platforms.8, 10, 19 The physicochemical properties of modules for insertion, such as hydrophobicity and charge, can affect the expression, solubility, and stability of modular proteins.25, 26

Figure 1.

Figure 1

Open in a new tab

Construct designs for modular capsomeres. (A) VLP platform with engineered insertion site at VP1 surface‐exposed S4 loop,8 with VP1 protein expressed as GST fusion protein. (B) Modular constructs VLP‐(RV10)3, VLP‐(RV10)3ESE, and VLP‐RV10 with inserted modules (RV10)3, (RV10)3ESE, and G4S‐Q25‐E4‐RV10‐E4‐P6‐G4S at S4 loop of VP1, respectively. (C) Capsomere platform with engineered N‐terminus, S1 loop, S4 loop and C‐terminus insertion sites on truncated VP1,10 expressed as GST fusion protein. (D) Modular constructs Cap(RV10)3 and Cap(RV10)3ESE with inserted modules (RV10)3 and (RV10)3ESE, respectively, at N‐terminus, S4 loop and C‐terminus insertion sites of truncated VP1. (E) Dual expression construct, pET‐VP1‐RV10, carrying both wild‐type VP1 and modular VP1‐G4S‐Q25‐E4‐RV10‐E4‐P6‐G4S sequence.

The RV10 peptide comprises amino acid residues 1‐10 of the RV VP8 subunit protein, and is one of the linear B‐cell epitopes involved in human RV neutralization.40, 41 RV10, a hydrophobic peptide, is a highly conserved epitope among human and animal RV strains.40 In this work, three different designs of module RV10 [(RV10)3, (RV10)3ESE, and G4S‐Q25‐E4‐RV10‐E4‐P6‐G4S] were synthesized at the DNA level using synthetic biology strategies, with the aim to generate modular constructs for production of stable modular capsomeres and VLPs.

(RV10)3 is a module containing three tandem copies of RV10. Insertion of (RV10)3 at the DNA level into the VLP and capsomere platforms produced constructs VLP‐(RV10)3 and Cap(RV10)3, as illustrated in Fig. 1. Module (RV10)3ESE contains three tandem copies of RV10 with additional E, ESE, and ES residues. Constructs VLP‐(RV10)3ESE [Fig. 1(B)] and Cap(RV10)3ESE [Fig. 1(D)] were designed to investigate if the addition of polar and/or charged amino acids to the element (RV10)3 can reduce module hydrophobicity, thus enhance the expression of soluble modular proteins. Previous studies have demonstrated that substitution or incorporation of polar and/or charged amino acids into the sequence of surface exposed residues contribute most favorably to protein solubility34, 42, 43 and suppress protein aggregation.26, 33

Module G4S‐Q25‐E4‐RV10‐E4‐P6‐G4S was synthesized by incorporating a single element of RV10 and ionic linkers to shorten the hydrophobic stretch in comparison to (RV10)3 and (RV10)3ESE. Insertion of G4S‐Q25‐E4‐RV10‐E4‐P6‐G4S into the VLP platform at the DNA level produced construct VLP‐RV10 [Fig. 1(B)]. Longer flanking linkers, G4S‐Q25 and G4S‐P6, were used for spatial separation of the module from the carrier to alleviate potential structural perturbation, as observed in another study displaying VP8* protein domain on VLP.39

Module density

Figure 1(E) illustrates the construct pET‐VP1‐RV10 designed for dual expression of unmodified VP1 and modular VP1 inserted with G4S‐Q25‐E4‐RV10‐P6‐G4S module, in E. coli co‐expression. High surface density of G4S‐Q25‐E4‐RV10‐E4‐P6‐G4S modules on each capsomere (five copies per capsomere) may cause a steric barrier to intercapsomere interactions, resulting in poor VLP assembly. This construct was designed to allow titration of module density on the surface of capsomere by co‐expressing both VP1 and modular VP1 in a single bacterial cell. The resulting capsomeres may contain a mixture of 5VP1, 4VP1:1VP1‐RV10, 3VP1:2VP1‐RV10, 2VP1:3VP1‐RV10, 1VP1:4VP1‐RV10, and 5VP1‐RV10. We hypothesized stable capsomeres with reduced module density will lead to in vitro assembled VLPs. This co‐expression strategy for the reduction of module density on the surface of carrier protein is adopted from the success of a baculovirus‐insect cell co‐expression system in reducing the surface density of VP8* module to avoid steric barrier to VLP formation.39

Effect of module hydrophobicity and charge

The expression and solubility of glutathione‐S‐transferase (GST) fusion modular proteins in comparison with GST fusion VP1 and ΔVP1 were analyzed using SDS‐PAGE (Fig. 2). No basal expression of the target GST fusion proteins was observed from the cultures harvested before induction with isopropyl‐β‐D‐thiogalactoside (IPTG). The expression and solubility of GST fusion VP1 and ΔVP1 were as previously reported.8, 10 VLP‐(RV10)3 was poorly expressed and no expression was detected for Cap(RV10)3. Modularization of the hydrophobic (RV10)3 modules likely increased the hydrophobicity of the modular proteins, thus resulting in poor or no target protein expression. Low level expression in E. coli was previously reported for a protein with its highly hydrophobic characteristics.44 Highly hydrophobic proteins may have adverse effect on host cells, thus prematurely terminated polypeptides, trapped folding intermediates and partially folded proteins may be consistently targeted for degradation to avoid their accumulation in cells.44, 45

Figure 2.

Figure 2

Open in a new tab

Expression and protein solubility of GST fusion target proteins. The target proteins were detected and visualized by SDS‐PAGE from un‐induced cultures (U), total cell lysate (T), and soluble fraction (S) of induced cultures. Novex sharp pre‐stained protein marker (M) was used as a ladder. Arrows indicate the GST‐tagged target proteins.

The addition of charge residues to RV10 element resulted in soluble expression of modular construct VLP‐(RV10)3ESE (Fig. 2, Lane 10). Total expression level of modular VLP‐(RV10)3ESE and Cap(RV10)3ESE improved significantly in comparison to VLP‐(RV10)3 and Cap(RV10)3 (Fig. 2). Addition of ESE, E, and ES residues to (RV10)3 has enhanced the total protein expression, consistent with previous studies.26, 33, 34, 42, 43

GST‐tagged VLP‐(RV10)3ESE was first purified with GST affinity chromatography, followed by TEVp‐mediated release of the GST tag and purification of capsomeres by SEC. Both SEC chromatograms [Fig. 3(A)] and SDS‐PAGE analysis [Fig. 3(B)] show that mostly soluble aggregates were obtained post removal of GST tag from construct VLP‐(RV10)3ESE. Unmodified VP1 capsomeres were obtained after TEVp‐mediated release of the GST tag as expected and consistent with previous reports.20, 21, 55 This suggests that VLP‐(RV10)3ESE modular capsomeres were not stable and prone to aggregation in L‐buffer. Modular VLP‐(RV10)3ESE capsomeres purified in the presence of TX‐100 as stabilizing additive in L‐buffer did not form modular VLPs via in vitro assembly [Fig. 4(B)]. When analyzed on AF4 post VLP assembly, only unassembled capsomeres and soluble aggregates were detected. The specific causes preventing modular VLP assembly for construct VLP‐(RV10)3ESE are unknown. One possibility is the hydrophobic–hydrophobic interactions between surface exposed RV10 elements, driving the formation of aggregates. Poor capsomere and VLP stability can also arise due to module insertion, causing a reduction of stabilization energy.46 Thus, this necessitates specific design of RV10 modules that will maintain capsomeres stability in compatible VLP assembly buffer.

Figure 3.

Figure 3

Open in a new tab

Downstream processing of capsomeres. (A) Size exclusion chromatograms of modular capsomeres following TEVp‐mediated release of GST tag. P1, P2, and P3 represent the aggregate, capsomere, and GST peaks, respectively. (B) SDS‐PAGE analysis on downstream processing of capsomeres. Marker protein (M), GST‐tagged soluble protein after GST affinity purification (S), TEVp‐mediated tag cleavage (D), P1 aggregate peak fraction (A), and P2 capsomere peak fraction (C).

Figure 4.

Figure 4

Open in a new tab

Characterization of capsomeres and in vitro assembled products. (A) SEC‐HPLC/LS analysis of VP1 and VLP‐RV10 capsomeres following TEVp‐mediated release of GST tag. P1, P2, and P3 represent peaks for VP1 capsomeres, VLP‐RV10 capsomeres, and GST dimers, respectively. (B) AF4 fractograms of in vitro assembled products of VP1, VLP‐(RV10)3ESE, and VLP‐RV10. P1, P2, and P3 represent peaks for non‐assembled proteins, assembled VLPs and aggregates, respectively.

Flanking linker and charged element for surface display of hydrophobic peptide

Construct VLP‐RV10 contains module G4S‐Q25‐E4‐RV10‐P6‐G4S and was expressed as soluble GST fusion target proteins [Fig. 2(B), Lane 13). Using a single copy of RV10 element with additional surface charged E4 residues likely decreased the hydrophobicity of modular proteins and resulted in good expression level of soluble proteins. Previously, insertion of E4 residues as ionic flanking elements has enhanced expression of soluble modular VP1 containing a hydrophobic H190 module from influenza virus.47 Studies have demonstrated positive correlation of decreasing hydrophobicity with high protein expression and solubility levels.48, 49

In addition to high level of soluble proteins with construct VLP‐RV10, stable capsomeres were obtained post release of GST tag (Fig. 3). The SEC chromatogram suggests a more compact capsomere structure for VLP‐RV10 as the capsomeres eluted later than unmodified VP1 capsomeres. The MW of VP1 and modular VLP‐RV10 capsomeres analyzed by static light scattering were 219.2 and 270.2 kDa, respectively, confirming their pentameric structure forms [Fig. 4(A)]. Despite obtaining stable VLP‐RV10 capsomeres, no VLP was detected after in vitro VLP assembly [Fig. 4(B)]. Assembled unmodified VP1 VLPs were detected as expected on the AF4.

Decreasing module hydrophobicity by reducing elements of RV10 in a module can minimize hydrophobic interaction forces that drive protein aggregation.50 The E4 ionic linkers increased the surface charge of the module; consequently the electrostatic repulsive forces between protein molecules can counteract the hydrophobic interaction forces, enhancing the stability of modular capsomeres. Introduction of longer flanking linkers G4S‐Q25 and P6‐G4S might have alleviated structural perturbation of capsomeres. We speculate that the lack of VLP assembly is caused by either strong electrostatic–repulsive interactions or steric hindrance between modular capsomeres due to high module density on capsomere surface.

Module titration for in vitro assembly of VLP

As shown in Fig. 5(A), soluble VP1 and VP1‐RV10 proteins were expressed in E. coli without GST tag. When both proteins were co‐expressed in a dual expression system, VP1‐RV10 was expressed at higher level than VP1. Co‐expression of two or more genes from a single expression vector might lead to imbalance ratio of expressed proteins due to differences in the rates of translation,51 transcription, translocation, and the stability of RNA and protein products.52, 53

Figure 5.

Figure 5

Open in a new tab

Co‐expression strategy for bacterially produced stable modular capsomeres and in vitro assembled modular VLPs presenting module RV10. (A) Detection and visualization of non‐tagged target proteins of pET‐VP1 and pET‐VP1‐RV10 from cell lysates. (B) Size exclusion chromatograms of pET‐VP1 and pET‐VP1‐RV10 capsomeres following purification by selective salting‐out precipitation. P1, P2, and P3 represent the aggregate, capsomere, and co‐precipitated E. coli protein peaks, respectively. (C) SDS‐PAGE analysis of the target proteins. Marker protein (M), purified protein by selective salting‐out precipitation (P), P1 aggregate peak fraction (A), and P2 capsomere peak fraction (C). (D) AF4 fractograms and TEM micrographs of assembled products (i) VP1VLP and (ii) RV10VLP. P1, P2, and P3 represent peaks for non‐assembled proteins, assembled VLPs, and aggregates, respectively.

Capsomeres of pET‐VP1 (CapVP1) and pET‐VP1‐RV10 (CapRV10) were purified from clarified cell lysates by selective salting‐out precipitation followed by a further polishing SEC step. Figures 5(B,C) show that stable modular capsomeres were obtained after purification. Co‐precipitated E. coli proteins, in the aggregated fractions, were separated from the capsomeres during the SEC polishing step [Fig. 5(C)]. After in vitro assembly, modular VLPs (RV10VLP) comprising VP1 and VP1‐RV10 capsomeres were obtained, as analyzed on AF4 and TEM [Fig. 5(D)]. A small peak of unassembled proteins was detected on AF4 with a VLP peak detected at 18–21 min elution time. The average root‐mean‐square (r.m.s) radius for the RV10VLP was 21.4 nm, similar to the size distribution of unmodified VP1VLP (r.m.s radius 22.5 nm). Under TEM, RV10VLP was similar in morphology to the unmodified VP1VLP [Fig. 5(D)]. This co‐expression strategy of both VP1 and VP1‐RV10 proteins to reduce the RV10 module density per capsomere was a successful approach to obtain stable RV10VLPs. Reducing the density of modules on the surface of capsomeres potentially minimize hydrophobic interactions, electrostatic repulsive interactions and overcome steric hindrance of modules to the formation of modular VLPs. Here, we have demonstrated that co‐expression strategy to reduce module density is feasible for bacterially produced modular capsomeres displaying hydrophobic peptide RV10.

Conclusion

Using synthetic biology, hydrophobic RV10 peptide epitope of RV VP8 spike protein was modularized on murine polyomavirus VP1 capsid protein. The hydrophobicity of element RV10 deters the tandem copies display strategy to increase the ratio of antigenic module to base protein VP1. Flanking hydrophobic element with charged glutamic acid (E4) can reduce hydrophobicity, to prevent hydrophobic interaction driven protein aggregation. However, the addition of long flanking linkers (G4S‐Q25 and P6‐G4S) potentially displays the module away from base protein VP1, thus avoid structural perturbation by the module. The module design G4S‐Q25‐E4‐RV10‐E4‐P6‐G4S yielded high expression of modular proteins and stable modular capsomeres after purification. Attempts to assemble these stable modular capsomeres via in vitro VLP assembly were not successful. The results suggest that titrating module density down by applying co‐expression strategy of both unmodified VP1 and modular VP1 in a single bacterial cell allows an optimal ratio of VP1:VP1‐RV10 capsomere formation, which then lead to VLP assembly. Module titration can minimize hydrophobic interactions, electrostatic repulsive interactions or prevent steric barrier to modular VLP assembly. The strategies employed in this study provide a powerful approach for modularization of VLP and capsomere with various hydrophobic antigenic modules, such as T‐cell epitopes, that are potentially challenging due to their physiochemical properties. The design strategies can also be used to modulate the hydrophobic domains of other proteins for enhanced expression of soluble and stable proteins using a low‐cost prokaryotic expression system.

Materials and Methods

Plasmid construction

pGEX‐4T‐1 plasmid (GE Healthcare Biosciences, Chalfont St. Giles, UK) with inserted murine polyomavirus VP1 sequence (M34958) was generously provided by Professor Robert Garcea (University of Colorado, CO, USA). This construct was designated as GST‐VP1 and used for expression of GST‐tagged wild‐type murine polyomavirus VP1 protein. Plasmid GST‐VP1‐S4 [Fig. 1(A)] was generated previously by inserting AfeI restriction enzyme site at position 293 of VP1.8 Construct GST‐ΔVP1 was generated previously by excluding the first 28 amino acids from the N‐terminus and the last 63 amino acids from the C‐terminus of VP1 sequence, and inserting PmlI, NaeI, AfeI, and SnaBI restriction enzyme sites at positions 28 (N‐terminus), 85 (loop S1), 293 (loop S4), and 380 (C‐terminus) of VP1 positions, respectively.10

DNA sequences encoding three tandem copies of residues 1‐10 (MASLIYRQLL, RV10) from the human RV spike protein VP8 subunit domain [MASLIYRQLLMASLIYRQLLMA SLIYRQLL, (RV10)3] was prepared by annealing of complementary oligos (5′atggcaagcctaatatacagacaactactaatggcaagcctaatatacagacaactactaatggcaagcctaatatacagacaactacta3′). The prepared gene insert was ligated into AfeI‐linearized GST‐VP1‐S4 vector [Fig 1(A)], and into PmlI‐, AfeI‐, or SnaBI‐linearized GST‐ΔVP1 vector [Fig. 1(C)] sequentially to generate modular constructs VLP‐(RV10)3 [Fig. 1(B)] and Cap(RV10)3 [Fig. 1(D)], respectively. The gene insert encoding three tandem copies of RV10 with ESE residues inserted between the RV10 peptides, and flanked by E and ES [EMASLIYRQLLESEMASLIYRQLLESEMASLIYRQL LES, (RV10)3ESE] was prepared by annealing complementary oligos (5′gagatggcgagcctcatctatcgccaactcctcgaaagcg aaatggcctctctgatctaccgccagctgctggagtctgaaatggcgtccctgatttaccgtcaactgctcgaatcc3′). The prepared gene insert was cloned into AfeI‐linearized GST‐VP1‐S4 vector, and into PmlI‐, AfeI‐, or SnaBI‐linearized GST‐ΔVP1 vector sequentially to generate modular constructs VLP‐(RV10)3ESE [Fig. 1(B)] and Cap(RV10)3ESE [Fig. 1(D)], respectively.

Similarly, gene insert encoding E4‐RV10‐E4, a module comprising RV10 with tetra glutamic acid residues (EEEE or E4), was prepared by assembling a set of oligos (5′tataatgggctggagag ttacaa3′, 5′gatgaggctcgccatctcttcctcttcgcttcttgtaactctccagcccattata3′, 5′agagatggcgagcctcatctatcgccaactcctcgaagaagaggaagcttatgat3′, 5′gccctctccagtgatggacatcataagcttcctcttcttcga3′) generated from DNAWorks (http://helixweb.nih.gov/dnaworks/). Linker sequences, G4S‐Q25 (residues GGGGS‐QGVSDLVGLPNQICLQKTTSTILKP) and P6‐G4S (residues PAQCSE‐GGGGS), were inserted into the N‐and C‐terminus of E4‐RV10‐E4 sequence, respectively, at the DNA level using PCR‐based gene assembly, annealing, and amplification. The prepared gene insert encoding module, G4S‐Q25‐E4‐RV10‐E4‐P6‐G4S, was cloned into AfeI‐linearized GST‐VP1‐S4 vector by homologous recombination to generate construct VLP‐RV10 [Fig. 1(B)].pETDuet‐1 vector with multiple cloning sites (MCS1 and MCS2) was purchased from Novagen (Madison, WI, USA). VP1 gene was cloned into MCS1 of pETDuet‐1 vector between EcoRI and SalI restriction sites. This vector was designated as pET‐VP1 and used for expression of VP1 without a tag. Another construct, designated as pET‐VP1‐RV10 [Fig. 1(E)], was generated to carry both VP1 and VP1‐RV10 gene inserts for co‐expression of the proteins. The NaeI restriction enzyme site was inserted into construct VLP‐RV10 at position 86 of VP1 as described previously.39 The VP1‐RV10 gene insert with RV10 flanking linkers G4S‐Q25‐E4 and E4‐P6‐G4S was amplified from construct VLP‐RV10 and cloned into MSC2 of pET‐VP1 construct between NedI and PacI restriction sites, to generate construct pET‐VP1‐RV10. All cloned constructs were verified by DNA sequencing at the Australian Genome Research Facility (Brisbane, Australia).

Protein expression

Constructs GST‐VP1, GST‐ΔVP1, VLP‐(RV10)3, VLP‐(RV10)3ESE, VLP‐RV10, Cap(RV10)3, and Cap(RV10)3ESE were transformed into E. coli Rosetta(DE3) pLysS chemically competent cells (Novagen, San Diego, CA, USA), separately. Bacterial expression of all GST‐tagged proteins were as previously described,20, 21 except cultures were induced with 0.1 mM IPTG (Astra Scientific Pty. Ltd., Gymea, NSW, Australia) at 20°C for 16 h. Constructs pET‐VP1 and pET‐VP1‐RV10 were transformed into chemically competent E. coli Rosetta 2 (DE3) cells (Novagen, San Diego, CA, USA). The transformed cells were grown separately at 37°C to an OD600 of 0.5 using Luria Bertani (LB) broth containing 50 µg mL−1 ampicillin (aMResco, Solon, OH, USA) and 34 µg mL−1 chloramphenicol (Astral Scientific Pty. Ltd., Gymea, NSW, Australia). Expression of non‐tagged VP1 and co‐expression of VP1 and modular VP1‐RV10 were induced with 0.5 mM IPTG at 26°C for 16 h. The expression level and solubility of the target proteins were detected using SDS‐PAGE that was performed using 10% gel as reported previously.54

Purification of capsomeres from GST‐tagged constructs

Purification of GST‐tagged proteins was performed as described previously for purification of GST‐tagged wild‐type VP1.20, 21, 55 Release of GST from GST‐tagged wild‐type VP1 (GST‐VP1) using thrombin‐mediated cleavage and VP1 capsomere purification was carried out as previously reported.10, 55 Tobacco etch virus protease (TEVp) was produced from a recombinant E. coli and purified by immobilized metal ion affinity chromatography as reported previously.56 The pure enzyme was used for release of GST from all other GST‐tagged proteins. TEVp‐mediated release of the tag protein was performed at 25:1 ratio (w/w) for 2 h at room temperature. The cleavage products were centrifuged (22,000g, 5 min, 4°C) and the capsomeres were recovered from 1.0 mL of supernatants for each protein with a Superdex 200 10/300 GL column (GE Healthcare Biosciences) operated with an AKTAexplorer 10 (GE Healthcare Biosciences) liquid chromatography system. The column was pre‐equilibrated with L‐buffer [40 mM Tris (pH 8.5), 500 mM NaCl, 1 mM ethylenediaminetetra‐acetic acid (EDTA) disodium, 5% (v/v) glycerol, and 5 mM dithiothreitol (DTT)] at a flow rate of 0.5 mL min−1. The column was also pre‐equilibrated with L‐buffer containing 0.5% (v/v) triton x‐100 (TX‐100) at a flow rate of 0.5 mL min−1 for purification of VLP‐(RV10)3ESE capsomeres after TEVp‐mediated release of the GST tag in the presence of 0.05% (v/v) TX‐100. The Unicorn software (Version 7.0) (GE Healthcare) was used for monitoring sample runs and for analysis and evaluation of the chromatogram data. Target proteins were detected with SDS‐PAGE in elution fractions corresponding to the aggregate and capsomere peak fractions of SEC chromatograms.

Purification of capsomeres from non‐tagged constructs

Soluble cell lysates were prepared as described previously,20, 21, 55 except cell pellets were resuspended in L‐buffer containing 500 mM NaCl. Capsomeres from constructs pET‐VP1 and pET‐VP1‐RV10 were purified from soluble cell lysate by selective salting‐out precipitation for 2 h at 4°C using 1 M Na2SO4. The protein pellets were collected by centrifugation (22,000g, 5 min at 4°C) and resuspended in L‐buffer. The soluble portion of the resuspension was separated from the insoluble fraction by centrifugation (22,000g, 5 min at 4°C). The capsomeres were further purified by SEC through a Superdex 200 10/300 GL column that was pre‐equilibrated with L‐buffer at a flow rate of 0.5 mL min−1.

Static light scattering

Molecular weight (MW) of modular capsomere VLP‐RV10 was estimated on multiangled detectors DAWN EOS (Wyatt Technology Corporation) light scattering instrument connected to a high‐performance liquid chromatography system (Agilent Technology System). The molecular weight of VP1 capsomeres was estimated under the same conditions as a control. Protein solutions of VP1 and VLP‐RV10 (V = 100 µL, C = 10.0–15.0 mg mL−1) after release of GST by enzyme‐mediated cleavage were loaded on a Superdex 200 10/300 GL column and eluted with L‐buffer at a flow rate of 0.5 mL min−1. The 90° light scattering and UV absorbance at 280 nm of the eluting materials were recorded on a computer and analyzed with the Astra software supplied by Wyatt Technology Corporation. The 90° light scattering detector was calibrated using bovine serum albumin (100 µL at 2.0 mg mL−1, MW: 66 kDa] as a standard.

VLP assembly and characterization

In vitro assembly of stable modular capsomeres into modular VLPs was as previously described,8 except dialysis was performed for 24 h at room temperature and for 16 h at 4°C against assembly buffer GL1 and PBS, respectively. Assembled products were analyzed using Asymmetric Flow Field‐Flow Fractionation (AF4) coupled with multiangle light scattering as described previously.57 Visualization of VLPs with transmission electron microscope (TEM) from in vitro assembly was performed as previously reported.57

Acknowledgments

The authors acknowledge the technical assistance provided by Yuanyuan Fan for obtaining the TEM images. University of Queensland (UQ) filed patents on the use of murine polyomavirus as a vaccine platform. L.H.L.L. and A.P.J.M. contributed to those patents and, through their employment with UQ, hold an indirect interest in this intellectual property. Other authors declare that there are no conflicts of interest.

References

  • 1. Lua LHL, Connors NK, Sainsbury F, Chuan YP, Wibowo N, Middelberg APJ (2014) Bioengineering virus like‐particles as vaccines. Biotechnol Bioeng 111:425–440. [DOI] [PubMed] [Google Scholar]
  • 2. Zeltins A (2012) Constructions and characterizations of virus‐like particles: a review. Mol Biotechnol 53:92–107. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 3. Adkins J, AJ W (1998) Recombinant hepatitis B vaccine: a review of its immunogenicity and protective efficacy against hepatitis B. Biodrugs 10:137–158. [DOI] [PubMed] [Google Scholar]
  • 4. Keating G, Noble S (2003) Recombinant hepatitis B vaccine (Engerix‐B): a review of its immunogenicity and protective efficacy against hepatitis B. Drugs 63:1021–1051. [DOI] [PubMed] [Google Scholar]
  • 5. Siddiqui M, Perry C (2006) Human papillomavirus quadrivalent (types 6, 11, 16, 18) recombinant vaccine (Gardasil). Drugs 66:1263–1271. [DOI] [PubMed] [Google Scholar]
  • 6. Wu T, Li SW, Zhang J, Ng MH, Xia NS, Zhao Q (2012) Hepatitis E vaccine development: a 14 year odyssey. Hum Vaccines Immunother 8:823–827. [DOI] [PubMed] [Google Scholar]
  • 7. Roldao A, Mellado M, Castilho L, Carrondo M, Alves P (2010) Virus‐like particles in vaccine development. Expert Rev Vaccines 9:1149–1176. [DOI] [PubMed] [Google Scholar]
  • 8. Middelberg APJ, Rivera‐Hernandez T, Wibowo N, Lua LHL, Fan Y, Magor G, Chang C, Chuan YP, Good MF, Batzloff MR (2011) A microbial platform for rapid and low‐cost virus‐like particle and capsomere vaccines. Vaccine 29:7154–7162. [DOI] [PubMed] [Google Scholar]
  • 9. Schädlich L, Senger T, Kirschning CJ, Müller M, Gissmann L (2009) Refining HPV 16 L1 purification from Escherichia coli: reducing endotoxin contaminations and their impact on immunogenicity. Vaccine 27:1511–1522. [DOI] [PubMed] [Google Scholar]
  • 10. Wibowo N, Chuan YP, Lua LHL, Middelberg APJ (2012) Modular engineering of a microbially‐produced viral capsomere vaccine for influenza. Chem Eng Sci 103:12–20. [Google Scholar]
  • 11. Dell K, Koesters R, Linnebacher M, Klein C, Gissmann L (2006) Intranasal immunization with human papillomavirus type 16 capsomeres in the presence of non‐toxic cholera toxin‐based adjuvants elicits increased vaginal immunoglobulin levels. Vaccine 24:2238–2247. [DOI] [PubMed] [Google Scholar]
  • 12. Senger T, Schädlich L, Textor S, Klein C, Michael KM, Buck CB, Gissmanna L (2010) Virus‐like particles and capsomeres are potent vaccines against cutaneous alpha HPVs. Vaccine 28:1583–1593. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 13. Chuan YP, Wibowo N, Lua LHL, Middelberg APJ (2014) The economics of virus‐like particles and capsomere vaccines. Biochem Eng J 90:255–263. [Google Scholar]
  • 14. Kumar SVAS, Biswas BM, Jose CT (2015) HPV vaccine: current status and future directions. Med J Armed Forces India 71:171–177. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 15. Yuan H, Estes PA, Chen Y, Newsome J, Olcese VA, Garcea RL, Schlegel R (2001) Immunization with a pentameric L1 fusion protein protects against papillomavirus infection. J Virol 75:7848–7853. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 16. Wu WH, Gersch E, Kwak K, Jagu S, Karanam B, Huh WK, Garcea RL, Roden RBS (2011) Capsomer vaccines protect mice from vaginal challenge with human papillomavirus. PloS One 6:e27141. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 17. Jagua S, Kwaka K, Garceab RL, Rodena RBS (2010) Vaccination with multimeric L2 fusion protein and L1 VLP or capsomeres to broaden protection against HPV infection. Vaccine 28:4478–4486. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 18. Wibowo N, Hughes FK, Fairmaid EJ, Lua LHL, Brown LE, Middelberg APJ (2014) Protective efficacy of a bacterially produced modular capsomere presenting M2e from influenza: extending the potential of broadly cross‐protecting epitopes. Vaccine 32:3651–3655. [DOI] [PubMed] [Google Scholar]
  • 19. Rivera‐Hernandez T, Hartas J, Wu Y, Chuan YP, Lua LHL, Good M, Batzloff MR, Middelberg APJ (2013) Self‐adjuvanting modular virus‐like particles for mucosal vaccination against group A streptococcus (GAS). Vaccine 31:1950–1955. [DOI] [PubMed] [Google Scholar]
  • 20. Liew MWO, Rajendran A, Middelberg APJ (2010) Microbial production of virus‐like particle vaccine protein at gram‐per‐litre levels. J Biotechnol 150:224–231. [DOI] [PubMed] [Google Scholar]
  • 21. Chuan YP, Lua LHL, Middelberg APJ (2008) High‐level expression of soluble viral structural protein in Escherichia coli . J Biotechnol 134:64–71. [DOI] [PubMed] [Google Scholar]
  • 22. Wibowo N, Wu Y, Fan Y, Meers J, Lua LHL, Middelberg APJ (2015) Non‐chromatographic preparation of a bacterially produced single‐shot modular virus‐like particle capsomere vaccinefor avian influenza. Vaccine 33:5960–5965. [DOI] [PubMed] [Google Scholar]
  • 23. Liew MWO, Chuan YP, Middelberg APJ (2012) High‐yield and scalable cell‐free assembly of virus‐like particles by dilution. Biochem Eng J 67:88–96. [Google Scholar]
  • 24. Palkova Z, Adamec T, Liebl D, Stokrova J, Forstova J (2000) Production of polyomavirus structural protein VP1 in yeast cells and its interaction with cell structures. FEBS Lett 478:281–289. [DOI] [PubMed] [Google Scholar]
  • 25. Tekewe A, Connors NK, Sainsbury F, Wibowo N, Lua LHL, Middelberg APJ (2015) A rapid and simple screening method to identify conditions for enhanced stability of modular vaccine candidates. Biochem Eng J 100:50–58. [Google Scholar]
  • 26. Abidin RS, Lua LHL, Middelberg APJ, Sainsbury F (2015) Insert engineering and solubility screening improves recovery of virus‐like particle subunits displaying hydrophobic epitopes. Protein Sci 24:1820–1828. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 27. Chanphai P, Bekale L, Tajmir‐Riahi HA (2015) Effect of hydrophobicity on protein‐protein interactions. Eur Polym J 67:224–231. [Google Scholar]
  • 28. Fagain C (1995) Understanding and increasing protein stability. Biochim Biophys Acta 1252:1–14. [DOI] [PubMed] [Google Scholar]
  • 29. Pace CN (1992) Contribution of the hydrophobic effect to globular protein stability. J Mol Biol 226:29–35. [DOI] [PubMed] [Google Scholar]
  • 30. Pace CN, Fu H, Fryar KL, Landua J, Trevino SR, Shirley BA, Hendricks MM, Iimura S, Gajiwala K, Scholtz JM, Grimsley GR (2011) Contribution of hydrophobic interactions to protein stability. J Mol Biol 408:514–528. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 31. Chao CC, Yan‐Shan M, Stadtman ER (1997) Modification of protein surface hydrophobicity and methionine oxidation by oxidative systems. Proc Nat Acad Sci USA 94:2969–2974. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32. Mann DF, Shah K, Stein D, Snead GA (1984) Protein hydrophobicity and stability support the thermodynamic theory of protein degradation. Biochim Biophys Acta 788:17–22. [DOI] [PubMed] [Google Scholar]
  • 33. Perchiacca JM, Ladiwala AR, Bhattacharya M, Tessier PM (2012) Aggregation‐resistant domain antibodies engineered with charged mutations near the edges of the complementarity‐determining regions. Protein Eng Des Sel 25:591–601. [DOI] [PubMed] [Google Scholar]
  • 34. Trevino SR, Scholtz JM, Pace CN (2007) Amino acid contribution to protein stability: Asp, Glu and Ser contribute more favorably than the other hydrophilic amino acids in RNase Sa. J Mol Biol 366:449–460. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 35. Chen X, Zaro JL, Shen W‐C (2013) Fusion protein linkers: Property, design and functionality. Adv Drug Deliv Rev 65:1357–1369. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 36. Zhao HL, Yao XQ, Xue C, Wang Y, Xiong XH, Liu ZM (2008) Increasing the homogeneity, stability and activity of human serum albumin and interferon‐α2b fusion protein by linker engineering. Protein Expr Purif 61:73–77. [DOI] [PubMed] [Google Scholar]
  • 37. Hao H‐J, Jiang Y‐Q, Zheng Y‐L, Ma R, Yu D‐W (2005) Improved stability and yield of Fv targeted superantigen by introducing both linker and disulfide bond into the targeting moiety. Biochimie 87:661–667. [DOI] [PubMed] [Google Scholar]
  • 38. Klement M, Liu C, Loo BLW, Choo AB‐H, Siak‐Wei Ow D, Lee D‐Y (2015) Effect of linker flexibility and length on the functionality of a cytotoxic engineered antibody fragment. J Biotechnol 199:90–97. [DOI] [PubMed] [Google Scholar]
  • 39. Lua LHL, Fan Y, Chang C, Connors NK, Middelberg APJ (2015) Synthetic biology design to display an 18 kDa rotavirus large antigen on a modular virus‐like particle. Vaccine 33:5937–5944. [DOI] [PubMed] [Google Scholar]
  • 40. Kovacs‐Nolan J, Yoo DW, Mine Y (2003) Fine mapping of sequential neutralization epitopes on the subunit protein VP8 of human rotavirus. Biochem J 376:269–275. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 41. Kovacs‐Nolan J, Mine Y (2006) Tandem copies of a human rotavirus VP8 epitope can induce specific neutralizing antibodies in BALB/c mice. Biochim Biophys Acta 1760:1884–1893. [DOI] [PubMed] [Google Scholar]
  • 42. Vernet E, Kotzsch A, Voldborg B, Sundsstrom M (2011) Screening of genetic parameters for soluble protein expression in Escherichia coli . Protein Expr Purif 77:104–111. [DOI] [PubMed] [Google Scholar]
  • 43. Pokkuluri PR, Raffen R, Dieckman L, Boogaard C, Stevens FJ, Schiffer M (2002) Increasing protein stability by polar surface residues: domain‐wide consequences of interactions within a loop. Biophys J 82:391–398. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 44. Ciccaglione A, Marcantonio C, Costantino A, Equestre M, Geraci A, Rapicetta M (2000) Expression and membrane association of hepatitis C virus envelope 1 protein. Virus Genes 21:223–226. [DOI] [PubMed] [Google Scholar]
  • 45. Niiranen L, Espelid S, Karlsen CR, Mustonen M, Paulsen SM, Heikinheimo P, Willassen NP (2007) Comparative expression study to increase the solubility of cold adapted Vibrio proteins in Escherichia coli . Protein Expr Purif 52:210–218. [DOI] [PubMed] [Google Scholar]
  • 46. Zhang L, Tang R, Bai S, Connors NK, Lua LHL, Chuan YP, Middelberg APJ, Sun Y (2014) Energetic changes caused by antigenic module insertion in a virus‐like particle revealed by experiment and molecular dynamic simulations. PloS One 9:e107313. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 47. Anggraeni MR (2014) Engineering of virus‐like particles for alternative vaccine candidate targeting hyper‐variable peptide antigen‐element. PhD Thesis, The University of Queensland. DOI: 10.14264/uql.2015.135.
  • 48. Idicula‐Thomas S, Balaji PV (2004) Understanding the relationship between the primary structure of proteins and its propensity to be soluble on overexpression in Escherichia coli . Protein Sci 14:582–592. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 49. Price WN, Handelman SK, Everett JK, Tong SN, Bracic A, Luff JD, Naumov V, Acton T, Manor P, Xiao R, Rost B, Montelione GT, Hunt JF (2011) Large‐scale experimental studies show unexpected amino acid effects on protein expression and solubility in vivo in Escherichia coli . Microb Inform Exp 1: DOI: 10.1186/2042-5783-1-6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 50. Valerio M, Colosimo A, Conti F, Giuliani A, Grottesi A, Manetti C, Zbilut JP (2005) Early events in protein aggregation: molecular flexibility and hydrophobicity/charge interaction in amyloid peptides as studied by molecular dynamics simulations. Proteins: Struct Funct Bioinf 58:110–118. [DOI] [PubMed] [Google Scholar]
  • 51. Tsao K, Waugh D (1997) Balancing the production of two recombinant proteins in Escherichia coli by manipulating plasmid copy number: high‐level expression of heterodimeric Ras farnesyltransferase. Protein Expr Purif 11:233–240. [DOI] [PubMed] [Google Scholar]
  • 52. Kerrigan JJ, Xie Q, Ames RS, Lu Q (2011) Production of protein complexes via co‐expression. Protein Expr Purif 75:1–14. [DOI] [PubMed] [Google Scholar]
  • 53. Karamitros CS, Konrad M (2014) Bacterial co‐expression of the α and β protomers of human L‐asparaginase‐3: achieving essential N‐terminal exposure of a catalytically critical threonine located in the β‐subunit. Protein Expr Purif 93:1–10. [DOI] [PubMed] [Google Scholar]
  • 54. Laemmli U (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680–685. [DOI] [PubMed] [Google Scholar]
  • 55. Lipin DL, Lua LHL, Middelberg APJ (2008) Quaternary size distribution of soluble aggregates of glutathione‐S‐transferase purified viral protein as determined by asymmetrical flow field flow fractionation and dynamic light scattering. J Chromatogr A 1190:204–214. [DOI] [PubMed] [Google Scholar]
  • 56. Cabrita LD, Gilis D, Robertson AL, Dehouck Y, Rooman M, Bottomley SP (2007) Enhancing the stability and solubility of TEV protease using in silico design. Protein Sci 16:2360–2367. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 57. Chuan YP, Fan Y, Lua LHL, Middelberg APJ (2008) Quantitative analysis of virus‐like particle size and distribution by field‐flow fractionation. Biotechnol Bioeng 99:1425–1433. [DOI] [PubMed] [Google Scholar]

Articles from Protein Science : A Publication of the Protein Society are provided here courtesy of The Protein Society

RESOURCES