Two-step chromatographic purification of recombinant Plasmodium falciparum circumsporozoite protein from Escherichia coli

Abstract

The Plasmodium falciparum circumsporozoite (PfCS) protein (aa 19–405) has been cloned and expressed in E. coli. The protein was purified in a two-step process that was rapid and reproducible. E. coli cells were grown to a high density before induction for 1 h. Cells were disrupted by high pressure microfluidization and the total bacterial protein solubilized in 6 M Gu-HCl. The protein was refolded while bound to Ni–NTA agarose by exchange of 6 M Gu-HCl for 8 M urea and then slow removal of the urea. The eluted protein was further purified on Q Sepharose Fast Flow using conditions developed to remove E. coli proteins and reduce endotoxin (to 10 EU/50 μg). Yield was 20 mg of PfCS protein from 10 g of wet cell paste. The final protein product bound to HepG2 liver cells in culture and inhibited the invasion of those cells by sporozoites in an ISI assay greater than 80% over control cultures when used at 10 μg/ml.

Introduction

The circumsporozoite (CS) protein, the major surface protein on the infective sporozoite stage of Plasmodium falciparum, is a major vaccine candidate against malaria. It has been shown by Nussenzweig et al. to play a critical role in the invasion of sporozoites into hepatocytes [1]. The gene encodes a protein of 405 amino acids that contains an amino-terminal signal peptide and a carboxy-terminal anchor domain typical of a membrane protein. The PfCS protein contains a large central repeat domain composed of tetrapeptides of sequence NANP or NVDP. The P. falciparum gene was first cloned by Dame et al. [2] but its recombinant expression and purification was difficult and disappointing. Even at the time of its initial cloning using the lambda-gt11 system, its stable expression proved precarious. Four out of five genes analyzed from that expression screening were in reverse order to the gene’s promoter and the fifth gene had an early nucleotide deletion and was out of reading frame. It was presumed, at the time, that the product of the gene was so toxic to the E. coli that only those clones with extremely low expression levels were able to grow and, thus, were selected by the anti-repeat domain specific monoclonal antibodies. Attempts to subclone the entire gene and express it in bacteria to achieve a full-length recombinant PfCS protein to be used for a malaria vaccine failed [3]. Consequently, the first recombinant malaria vaccine, FSV1, contained only 32 tetrapeptide repeat units expressed as a fusion protein with 32 amino acids derived from an open reading frame in the Tet^r gene.

While the development of a malaria vaccine based on recombinant expression of various portions of the PfCS protein has proceeded [3], [4], [5], [6], the analysis of the immune response to the vaccine has relied primarily on the recognition of synthetic peptides. These small sequences may not be able to reproduce the tertiary structure that is an inherent part of the full-length PfCS protein. Because the native PfCS molecule contains five cysteines and 54 prolines, amino acids that can greatly influence protein structure, the tertiary structure of the recombinant PfCS is likely to influence its presentation to the immune system. There is strong evidence to suggest disulfide bridge formation between some cysteines. Reduction and alkylation of the protein abolish binding to sulfatides [7] and mutagenesis of the cysteines to alanines greatly affects the binding of the PfCS protein to liver cells [8].

Eukaryotic proteins are frequently not folded correctly when synthesized in prokaryotes, resulting in an insoluble, non-native form of recombinant protein which is usually stored in inclusion bodies [9]. Published protocols for purification of inclusion bodies often include solubilization by detergents or by strong chaotropic salts. Some refolding strategies have included dialysis and stepwise or continuous removal of the denaturant while the target protein is bound to the resin. Although the refolding of globular proteins from inclusion bodies is used more often today, only a few membrane proteins have been refolded. Rogl et al. [10] reported refolding of two different membrane proteins using chaotropic salts and detergents. Frankel et al. [11] diluted protein with renaturation buffer containing oxidized glutathione and 0.5 M l-arginine HCl and refolding was allowed to proceed at 4°C for 44–60 h followed by 24 h of dialysis. Gupta et al. [12] tried to purify a protein immobilized on Ni–NTA resin under denaturating conditions using gradual removal of the denaturant followed by several dialysis steps. Another group, Holzinger et al. [13] accomplished solubilization of the target protein bound to a Ni–NTA column in a single step by omitting the denaturant from the last wash and elution buffers. Takacs et al. [14] successfully expressed and purified a fragment of the PfCS protein both as a fusion with a “merozoite protein” and as a 6×-His-tagged molecule. However, that procedure did not work to purify the full length the PfCS molecule produced here. Therefore, we report the development of a process that yields a significant amount of soluble, stable full-length PfCS protein. The purification protocol presented in this study allows for a gradual renaturation of the PfCS protein in an effort to refold it into a more native structure while at the same time achieving a purity acceptable for a human-use vaccine. The function of the CSP molecule on the sporozoite is to enable it to bind to receptors on the liver cells, therefore, in order to test the biological function of the recombinant CSP we allowed the protein to bind to heptocytes in culture and tested its ability to interfere with the invasion of sporozoites in an inhibition of invasion assay (ISI).