Construction and use of Plasmodium falciparum phage display libraries to identify host parasite interactions

Background The development of Plasmodium falciparum within human erythrocytes induces a wide array of changes in the ultrastructure, function and antigenic properties of the host cell. Numerous proteins encoded by the parasite have been shown to interact with the erythrocyte membrane. The identification of new interactions between human erythrocyte and P. falciparum proteins has formed a key area of malaria research. To circumvent the difficulties provided by conventional protein techniques, a novel application of the phage display technology was utilised. Methods P. falciparum phage display libraries were created and biopanned against purified erythrocyte membrane proteins. The identification of interacting and in-frame amino acid sequences was achieved by sequencing parasite cDNA inserts and performing bioinformatic analyses in the PlasmoDB database. Results Following four rounds of biopanning, sequencing and bioinformatic investigations, seven P. falciparum proteins with significant binding specificity toward human erythrocyte spectrin and protein 4.1 were identified. The specificity of these P. falciparum proteins were demonstrated by the marked enrichment of the respective in-frame binding sequences from a fourth round phage display library. Conclusion The construction and biopanning of P. falciparum phage display expression libraries provide a novel approach for the identification of new interactions between the parasite and the erythrocyte membrane.


Background
Malaria remains one of the most serious and widespread causes of pathogen-specific mortality in humans. Between 300 and 500 million people are affected globally with the mortality rate estimated to be 2.5 million deaths each year http://www.malaria.org/. Control of malaria is becoming increasingly difficult. The spread of resistance of both the parasite and the mosquito vector to anti-malaria drugs and insecticides, ensures that malaria remains an expanding health risk. Therefore, new approaches are needed to combat the disease. These include the identification of novel therapeutic agents and the development of vaccines targeted to various stages of the parasite life cycle [1].
The virulence and widespread prevalence of Plasmodium falciparum has made this parasite the best-studied species of malaria protozoa that infect humans. The invasion of erythrocytes by P. falciparum alters the host cells morphologically and functionally. P. falciparum has been shown to extensively modify the erythrocyte membrane in order to facilitate merozoite entry, intra-erythrocytic parasite development and exit from the erythrocytes. Only a handful of parasite proteins have been shown to interact with erythrocyte cytoskeletal proteins, including key structural modulators such as spectrin and protein 4.1 [2]. Despite our increasing knowledge, many questions remain unanswered concerning protein-protein interactions between host and parasite.
The isolation of proteins from P. falciparum using conventional protein techniques has limited the discovery of new protein interactions. The inability to maintain a very high parasitaemia of P. falciparum in continuous culture has hindered means to purify parasite-derived proteins in large quantities. Furthermore, the presence of contaminating erythrocyte membrane proteins in P. falciparum protein extracts makes it extremely difficult to obtain preparations of sufficient purity. To circumvent these limitations, phage display technology was used for characterizing biomolecular interactions.
Phage display involves the construction and assembly of peptide libraries, which comprise millions of short, variable amino acid sequences that are displayed on the surface of bacteriophage virions [3]. To date, research into P. falciparum phage display has focused only on two applications: the expression of random peptide fragments derived from a selected malaria protein [4] and production of antibody libraries for biopanning against a malaria protein [5]. In this study, we demonstrate a novel approach using P. falciparum phage display libraries for mapping unidentified protein interactions between human erythrocyte membrane proteins and the malaria parasite.

Construction of P. falciparum phage display libraries
The P. falciparum FCR-3 strain was cultured in vitro according to the method of Trager and Jensen [6]. Parasites were extracted from infected erythrocytes and total RNA isolated using guanidinium isothiocyanate [7]. Messenger RNA was extracted using the Dynal ® (Oslo, Norway) mRNA direct kit. Two-base anchored (oligodT 12 VN) primers (2 µg) were used for reverse transcribing 4 µg of P. falciparum mRNA. cDNA was synthesized and end modified according to Novagen's (Wisconsin, USA) T7Select10-3 Orient Express ™ cDNA cloning (random primer) system. Excess linkers and cDNA smaller than 300 base pairs (bp) were removed by gel filtration. P. falciparum phage display libraries were created by cloning the modified cDNA into the T7Select10-3b EcoRI/HindIII vec-tor arms. Recombinant vectors were subsequently packaged in T7 bacteriophage packaging extracts and propagated in Escherichia coli (strain BLT5403) during which the fusion proteins were expressed and displayed on the phage surface.
Biopanning P. falciparum phage display cDNA libraries were screened by biopanning against immobilized human erythrocyte spectrin and protein 4.1. These proteins were isolated from erythrocyte membranes by using a glycine-NaOH (pH 9.8) buffer containing Tween 20 [8]. Spectrin and protein 4.1 were subsequently concentrated using Slide-Alyzer cassettes (Pierce, USA) and polyethylene glycol 20,000 (Sigma-Aldrich, USA) and biotinylated with D-Biotin-N-hydroxysuccinimide ester (Roche, Germany). Thirty to 120 µg biotinylated protein (0.6 µg/µl-2.0 µg/µl) were immobilized on streptavidin-coated magnetic beads (Roche, Germany) at room temperature for 2 hours. Unbound protein was removed by washing with Tris-buffered saline (TBS), pH 7.5. Libraries were reacted overnight at 4°C with streptavidin-coated beads to pre-select background binding. The remaining phage were then incubated overnight against spectrin and protein 4.1 coated streptavidin beads at 4°C. Unbound phage were removed by washing with TBS plus 0.1 % (v/v) Tween 20. Bound phage were eluted from beads for 10-20 minutes at room temperature by adding 200 µl 1% SDS, and then amplified overnight at 37°C in 50 ml early log-phase BLT5403 E. coli cultures. Following amplification, sodium chloride was added to lysed cultures to a final concentration of 0.5 M and cellular debris cleared by centrifugation at 8000 × g for 10 minutes. The supernatant was subjected to another round of biopanning. After fourth round biopanning, the final lysate was titred and DNA was extracted from individual plaques by incubation in 50 µl 10 mM EDTA (pH 8.0) at 65°C for 10 minutes. P. falciparum insert sizes were determined by PCR using 1 µl phage lysate, 12.5 µl PCR Master Mix (Roche, Germany) and 0.25 µl T7Select specific primers. The DNA sequence and correct reading frames of insert sequences were determined using the PCR product pre-sequencing and Sequenase Version 2.0 DNA sequencing kits (Amersham Biosciences, UK). Sequences were verified for correct reading frame and matched to annotated P. falciparum genes by performing bioinformatic analyses in the PlasmoDB database http://www.plasmodb.org/.

Results and Discussion
The average size of P. falciparum cDNA synthesized was greater than 722 bp (Fig. 1). The recommended molar ratio of insert to vector (1:1 to 3:1 insert: vector for an average cDNA size of 1.5 kbp) was, therefore, used for cloning P. falciparum cDNA into the T7Select vector. However, the majority of P. falciparum cDNA inserts after fourth round biopanning were 100-200 bp in length, and the largest insert was 600 bp (Fig. 2). The presence of small inserts was unexpected, since gel filtration of the modified cDNA should have excluded fragments smaller than 300 bp, and electrophoretic analysis (Fig. 1) did not reveal any smaller fragments. However, smaller cDNA fragments would have incorporated a very low amount of radiolabel, resulting in their apparent absence from the gel (Fig. 1). Inefficient size exclusion during column chromatography presumably resulted in a higher molar ratio of smaller to larger cDNA fragments, which would therefore have facilitated a favourable cloning efficiency for the smaller fragments. The efficiency of in vitro packaging and/ or ligation between P. falciparum cDNA inserts and T7Select10-3b vector was significantly lower (10 4 -10 6 plaque forming units/µg vector), when compared to a positive control (10 9 plaque forming units/µg vector). This may be accounted for by having selected an insert: vector ratio favourable for cloning fragments of ± 1.5 kbp.
Sequencing analyses of fourth round virions and subsequent blast searches in PlasmoDB, revealed in-frame insert sequences which were mapped to predicted P. falciparum genes. Seventeen parasite sequences were in-frame, representing approximately one third of the inserts sequenced. However, several of these inserts were found more than once and only seven represented unique open reading frames. This was attributed to the significant degree of enrichment of each of these clones from the initial library. This demonstrates that the in-frame P. falciparum binding sequences identified by biopanning are highly specific for human erythrocyte spectrin and protein 4.1. The seven annotated proteins are shown in Table 1. Ebl-1 like protein, a Duffy binding-like protein is Gel filtration of P. falciparum cDNA Figure 1 Gel filtration of P. falciparum cDNA. Lanes 1 and 2 represent end-modified cDNA, labeled with [α-32 P] dATP (Amersham, Biosciences, UK), after gel filtration. To determine the average size of P. falciparum cDNA, size markers were synthesized by PCR utilizing primers specific for human erythrocyte α-spectrin (exon 2) and band 3 (exon 18 and 19) sequences (lanes 3 and 4). P. falciparum cDNA appears as a smear of an average size greater than 722 bp in length. cDNA and PCR products were evaluated using an 8 % denaturing polyacrylamide gel and autoradiography.
PCR amplification of P. falciparum cDNA inserts cloned into the T7Select10-3b vector Figure 2 PCR amplification of P. falciparum cDNA inserts cloned into the T7Select10-3b vector. All fragments include 107 bp of vector sequence. Lanes 2-12 depict amplified DNA from selected phage recombinants. Lane 1 is a 100 bp DNA ladder (Promega, USA). PCR products were evaluated using 1% agarose gel electrophoresis and ethidium bromide staining.
synthesized in the merozoite micronemes and may be involved in erythrocyte entry [9]. The biochemical modification of spectrin and protein 4.1 by the putative kinase, aminopeptidase and phosphotransferase may result in the structural re-organisation of the erythrocyte skeleton, thus facilitating the survival of P. falciparum. P. falciparum phage display libraries were successfully used to identify malaria protein sequences involved in binding important structural components of the erythrocyte membrane, namely spectrin and protein 4.1. The expression and analyses of the identified P. falciparum proteins/binding domains will aid in unravelling the binding kinetics of these recombinant proteins with spectrin and protein 4.1.

Conclusions
P. falciparum phage display libraries provide a powerful tool and novel approach for identifying new protein interactions between the parasite and the human erythrocyte membrane. These libraries were used to identify seven parasite proteins that bind human erythrocyte spectrin and/or protein 4.1. The function of these P. falciparum proteins remains to be determined, however, they are potentially involved in the entry and/or exit of merozoites from the human erythrocyte. Furthermore, these proteins may also be involved in parasite growth and survival during intra-erythrocytic development.