• Tidak ada hasil yang ditemukan

Ideally an immuno-diagnostic target molecule should be present throughout the malaria red blood cell cycle and expressed at relatively high levels to allow for detection of the target during early stages of infection while parasitemia is still low. The target molecule should also be parasite specific and not expressed in the host, or if protein in nature, have parasite specific peptides within its sequence, which could be exploited for diagnosis. The bioinformatics approach used in this study to identify such potential targets was outlined in this chapter.

The aim to identify highly abundant proteins was addressed using transcriptome and proteomic data (LeRoch et al., 2003; Foth et al., 2011; www.PlasmoDB.org). LeRoch et al.

(2003), Bozdech et al. (2003) and others hypothesised that the malaria parasite may have a

“just in time” strategy for gene expression. Simply put the genes of the proteins required during a specific growth stage are transcribed and translated as the parasite requires them.

Foth et al. (2011) studied the proteome of the parasite and found other possibilities. The “just in time” hypothesis was supported for some genes, but others exhibited delayed translation while they still mirrored their respective mRNA fluctuations. Finally a third group of genes clearly experienced alternate control of its proteome most likely involving post-translational modifications, meaning that their protein levels did not necessarily correlate with mRNA levels. The list of 35 potential proteins compiled here included proteins described by all three criteria.

Post translational modifications have been identified in LDH, GAPDH and PMT. Post translational modifications include acetylation, nitrosylation, phosphorylation, methylation and ubiquitination amongst others (Alam et al., 2014; Chung et al., 2009). In humans, post translational modifications result in approximately 1.8 million protein variants from only

~30000 open reading frames (Alam et al., 2014). Acetylation of lysine residues was long known to be essential in histone function (Chung et al., 2009), but has only been recognised as a regulatory modification in other malaria proteins in the last three to five years with Miao et al. publishing the malaria “acetylome” in 2013, which interestingly included LDH, GAPDH and PMT. Acetylation is thought to play a role in the regulation of glycolysis (Guarente, 2011), where alternate modifications such as phosphorylation seem to “switch on”

non-glycolytic moonlighting functions (Sirover 1999, 2005, 2012). Importantly post translational modifications such as acetylation and phosphorylation may affect antibody

84 affinities due the change in charge of the amino acid side chains and these possibilities are discussed in the general discussion (chapter 7).

Buehner et al. first described the common N-terminal nucleotide (NAD+(H)) binding motifs between mammalian LDH and GAPDH (Rossman fold) in 1973. This nucleotide binding ability has since implicated both mammalian enzymes in transcriptional regulation (Kim and Dang., 2005). Although this has not been empirically shown for Plasmodium LDH or GAPDH, both orthologs also share a common Rossman fold (Adams et al., 1973; Akinyi et al., 2008; Daubenberger et al., 2000; Granchi et al., 2010), suggesting they may also share this moonlighting function. Several additional moonlighting functions found with mammalian GAPDH, as a result of post translational modifications, have also been described (Alam et al., 2014). These may explain the apparent increased abundance of GAPDH in relation to LDH in Plasmodium. Some of these functions include: DNA repair, RNA binding, telomere binding, cell cycle regulation, histone expression, membrane fusion, phosphorylation, phosphatidyl serine binding, nitric oxide interaction, cytoskeletal binding and apoptosis (Demarse et al., 2009; Kaneda et al., 1997; Kim and Dang, 2005; Sirover M.A., 1999, 2012;

Tristan et al., 2011). Mammalian GAPDH has been shown to interact with Band 3 on the erythrocyte surface in an NADH and ATP dependent manner (Heard et al., 1998).

Structurally P. falciparum GAPDH shares 63.5% amino acid sequence identity with its human counterpart (Daubenberger et al., 2003), and forms a tetramer in solution (337 aa;

36651 Da; pI 7.59) (Berwal et al., 2008; Daubenberger et al., 2000; PlasmoDB; Satchell et al., 2005). P. falciparum GAPDH shares several post translational modification sites with its human counterpart, in addition to a few unique sites (Alam et al., 2014). In P. falciparum studies thus far, GAPDH has been detected in parasite membrane fractions in a GTPase (Rab2) dependent manner and the authors suggested roles in apicoplast formation and vesicle transport (Daubenberger et al., 2003), although sequence analysis suggests GAPDH to be cytosolic and not plastid targeted (Akinyi et al., 2008; Alam et al., 2014). These findings suggest there may be multiple additional functions for both “housekeeping” genes which may explain their different abundances within the parasite in spite of their linked role in cycling cellular NAD+(H).

The choice of peptides in this study was based on the approach taken by Hurdayal et al.

(2010). The current diagnostic target LDH was the subject of that study and was therefore included as a model for this study. The peptide selection strategy from the Hurdayal et al.

(2010) study was to choose specific and common peptide epitopes that were unique to the

85 Plasmodium proteins using multiple sequence alignment and epitope prediction programs.

Polyclonal antibodies raised against these peptides then allowed for the specific detection of Plasmodium LDH even to a species level. The GAPDH and PMT peptides were identified using a similar strategy and would serve as the target antigens for raising polyclonal chicken antibodies as well as chicken scFv antibodies as described in the later chapters. The chosen peptides were between ten and 16 amino acids in length, which is sufficient to allow specific detection of the parent protein (Hurdayal et al., 2010; Tomar et al., 2006). Multiple sequence alignments were used to identify the common and species specific targets. Predict7TM analysis was then used to analyse the potential peptides. Since each amino acid side chain has unique properties, this program uses algorithms to compare primary amino acid sequences and is able to predict secondary structure (Chou and Fasman, 1979); surface probability (Emini et al., 1985; Kyte and Doolittle, 1982); hydrophilicity (Hopp and Woods, 1981) and antigenicity (Jameson and Wolf, 1988). Since the recognition of proteins in an RDT or ELISA format uses antibodies, the peptides had to be located on the surface of their respective proteins, to facilitate antibody recognition (Saravanan et al., 2009) making the hydrophilicity plots very important. To complement these plots, target peptides were also located and shown to be on the surface of their respective parent protein crystal structures.

The greater the hydrophilicity also meant that the peptides were likely to be easily solubilised and stable in solution which is essential for use in an ELISA format. Due to the small size of peptides, they are not able to stimulate an immune response by themselves and require coupling to larger carrier proteins such as rabbit albumin used in this study (Hurdayal et al., 2010; Tomar et al., 2006). For this purpose and to allow coupling to an affinity resin, terminal cysteine residues were added to the selected peptides during synthesis. The Predict7TM plots were important in deciding whether the cysteines were added N- or C- terminally as it was preferred to expose the side with the greatest hydrophilicity, surface probability and flexibility.

Both sets of peptides selected from LDH were within regions that had insertions either in the Plasmodium proteins or the mammalian counterparts. The common peptide had a five amino acid insertion and formed part of the cofactor binding loop (Alam et al., 2014; Gomez et al., 1997). The specific peptides lacked a two amino acid insertion present in the mammalian proteins. The GAPDH peptides were selected from the regions of greatest variation, which is within the Rossman fold with 68% variance amongst the Plasmodium (Akinyi et al., 2008;

Fast et al., 2001). One of the major differences between human and Plasmodium GAPDH is

86 the presence of a two amino acid (K194; G195) insert within a structural region called the S- loop (residue 188 to 203) in the Plasmodium protein (Daubenberger et al., 2000; Satchell et al., 2005). The S-loop separates the Rossman folds of the adjacent subunits in the tetrameric form of the enzyme (Akinyi et al., 2008). Another difference is a substitution of two amino acids (L187; V188 for K187; T188) in the same region of the Plasmodium protein (Daubenberger et al., 2000). These changes are thought to be responsible for the ferriprotoporphyrin susceptibility of the Plasmodium protein in comparison to its human counterpart (Akinyi et al., 2008; Satchell et al., 2005).

PfPMT is not expressed in humans and its closest homolog is a histamine methyltransferase which is also a small-molecule S-adenocyl-L-methionine dependent methyl transferase with 7-16% sequence identity and approximately 31% sequence similarity around the substrate binding site (Horton et al., 2001). Amongst the different Plasmodium species PMT has only been confirmed in P. falciparum, P. vivax and P. knowlesi, with over 62% sequence identity between these species. Following genome sequencing PMT may also be expressed in P.

reichenowi and P. gallinaceum, but appears to be absent from all rodent malaria species (Dechamps et al., 2010). PMT homologs were identified in Burkholderia pseudomallei, B.

oklahomensis, Xenopus laevis, Xenopus tropicalis, Caenorhabditis briggsae, Danio rerio, Branchiostoma floridae, Caenorhabditis elegans and Anopheles gambiae, but critically, no human homologs exist (Pessi et al., 2004; Bobenchik et al., 2013). In theory, antibodies raised against such a target should have no cross-reactivity with the human proteome. The closest homolog to PfPMT in humans is a histamine methyltransferase which is also a small- molecule SAM-dependent methyl transferase with 7-16% sequence identity and approximately 31% sequence similarity around the substrate binding site (Horton et al., 2001).

Selection of P. falciparum specific peptide sequences was less challenging than selecting the P. vivax and P. knowlesi peptides. This was because the P. falciparum proteins had lower identity with their Plasmodium counterparts. As a result, the chosen P. vivax and P. knowlesi LDH species specific peptides aligned with 91% identity, the GAPDH sequences had around 80% identity and the PMT sequences 60%. It would be important to test the specificity of the antibodies produced against these targets using P. vivax and P. knowlesi proteins and parasite lysates. In this study, the P. falciparum and P. vivax LDH and PMT, P. falciparum GAPDH, as well as the P. yoelii LDH and GAPDH proteins were expressed and purified. The coding DNA for each recombinant protein was verified by sequencing. According to Hill et al.,

87 (2000); Kristensen et al., (1992) and Lamperti et al., (1992) Genbank entries may carry between 3.1 to 3.6% error rates. The sequences in this study were all within 4% identity of their respective database entries, where only the P. yoelii GAPDH sequence scored 91% and the P. vivax PMT sequenced scored 78 %. The P. vivax PMT sequence was codon optimised for expression in E. coli however, which explained its lower identity (Garg et al., 2015).

When translated, it scored 95% identity with the primary protein amino acid sequence and the P. yoelii GAPDH amino acid sequence scored 96%. Overall the sequencing results were satisfactory and validated the presence of the peptide targets on the recombinant proteins expressed and used in this study as described in the next chapter.

88