Green fluorescent protein as a scaffold for high efficiency production of functional bacteriotoxic proteins in Escherichia coli

Abstract

The availability of simple, robust, and cost-effective methods for the large-scale production of bacteriotoxic peptides such as antimicrobial peptides (AMPs) is essential for basic and pharmaceutical research. However, the production of bacteriotoxic proteins has been difficult due to a high degree of toxicity in bacteria and proteolytic degradation. In this study, we inserted AMPs into the Green fluorescent protein (GFP) in a loop region and expressed them as insoluble proteins in high yield, circumventing the inherent toxicity of AMP production in Escherichia coli. The AMPs inserted were released by cyanogen bromide and purified by chromatography. We showed that highly potent AMPs such as Protegrin-1, PMAP-36, Buforin-2, and Bactridin-1 are produced in high yields and produced AMPs showed similar activities compared to chemically synthesized AMPs. We increased the yield more than two-fold by inserting three copies of Protegrin-1 in the GFP scaffold. The immunogold electron micrographs showed that the expressed Protegrin-1 in the GFP scaffold forms large and small size aggregates in the core region of the inclusion body and become entirely nonfunctional, therefore not influencing the proliferation of E. coli. Our novel method will be applicable for diverse bacteriotoxic peptides which can be exploited in biomedical and pharmaceutical researches.

Figure 1. Comparison of protein expression levels and bacterial growth of E. coli r5M-172-PG1-173 and KSI-PG1 fusion systems.

(a) Comparison of protein production efficiency of r5M-172-PG1-173 in pET30b and KSI-PG1 in pET31b. Total cellular protein samples were analyzed by 12% SDS-PAGE at different expression times. Lane UI: uninduced protein sample, lanes 1–3 were induced protein samples collected at 3, 4, and 5 h of expression, respectively, and M is molecular weight marker. The boxes indicate the expected size of r5M-172-PG1-173 (31 kDa) and KSI-PG1 (18.4 kDa). Results showed >7 fold increase in 172PG-1173 production compared to KSI-PG1. (b) Growth of E.coli carrying pET31b with KSI and PG-1 fused to KSI (KSI-PG1) after induction and r5M-172-PG1-173 in pET30b with and without induction with IPTG. The OD₆₀₀ was measured for all of the constructs with or without induction every hour up to 5 h (I = induced and UI = uninduced).

Figure 2. The vector map of the r5M-172-AMP-173 expression construct and the production strategy of antimicrobial peptides using r5M-GFP system in E. coli.

(a) AMPs were inserted between 172 and 173 amino acids positions of r5M-GFP. Loop structure forming linker sequences, GGSGT and GSGG, are used for cloning and indicated in amino acids. The letter “M” flanking both 5′ and 3′ side of AMP indicates methionine. The restriction sites and 6x His-tag at both N- and C-terminals are indicated. (b) AMPs are cloned into a loop region of methionine-free GFP (r5M-172-AMP-173) and expressed as insoluble proteins in E. coli. The insoluble proteins are extracted and purified by Ni-NTA and cleaved at methionine sites using cyanogen bromide. Finally, AMPs are purified by high performance liquid chromatography.

Figure 3. Purification of PG-1, PMAP-36, and Buforin-2 and confirmation of PG-1 expression.

Insoluble protein purification by affinity chromatography, CNBr digestion of affinity-purified target proteins, and reverse-phase high performance liquid chromatography (RP-HPLC) of CNBr-digested r5M-172-PG1-173 (a) r5M-172-PMAP36-173 (b) and r5M-172-Bf2-173 (c) respectively. Lane 1: affinity purification of insoluble protein; lane 2: CNBr digestion of purified insoluble protein; lane 3: final RP-HPLC purification of PG-1 (a) PMAP-36 (b) and Buforin-2 (c) respectively, and lane M is molecular weight marker. The boxes indicate the expected size of the target protein after the various purification steps for PG-1 (2.4 kDa), PMAP-36 (4.2 kDa), and Buforin-2 (2.3 kDa). In lane 3, the purified peptides with higher molecular weights (indicated by an arrow) show the occurrence of AMP multimerization. The samples were loaded onto either 16% Tris-Tricine or 12% SDS-PAGE gels. (d) Confirmation of the presence of purified PG-1 using a rabbit anti-PG-1 antibody.

Figure 4. Immunogold transmission electron micrographs of expressed r5M-172-PG1-173 inclusion bodies in E. coli using PG-1 specific antibodies.

(A) Expression of r5M-172-PG1-173 as insoluble inclusion bodies was clearly visible inside the E. coli cytoplasm. (B–D) The expressed inclusion bodies of r5M-172-PG1-173 were immunogold labeled using rabbit-anti PG-1 antibody. The expressed PG-1 was shown as dark spots within the inclusion bodies and indicated by arrows.