Protein Spread - Introduction

Chapter 5

Protein Transduction Domain mediated protein spread

Aim: To enhance the efficiency of gene delivery through the intercellular spread of expressed protein.

5.1 Protein Spread

The term ‘protein spread’ used here refers to the secretion of a transgenically encoded protein from transfected cells and transduction into adjacent untransfected cells. This will enhance the biological response to a transfection strategy. This system is based upon the secretion of protein from transfected cells and subsequent transduction of adjacent cells by virtue of a protein transduction domain (PTD) peptide sequence.

5.2 Protein Secretion

Endoplasmic reticulum signal sequences (also known as secretory signals) guide nascent polypeptide chains to the endoplasmic reticulum (ER) through association with a signal recognition particle (SRP). The SRP transiently associates with the signal sequence peptide and ribosome, directing this complex to the SRP receptor located on the membrane of the rough ER. Here the SRP dissociates from the ribosome and nascent polypeptide chain complex. The complex is transferred to the translocon, which allows entry of the growing polypeptide chain into the lumen of the ER where it is folded. Signal sequences are located at the N-terminus of the growing polypeptide chain and generally contain a few cationic amino acids followed by a hydrophobic region. The signal sequence is cleaved once inside the lumen of the ER. Correctly folded proteins are transported to the Golgi complex where they become glycosylated. Hence, secreting PTD-cargo proteins in a mammalian system allows for correct folding and glycosylation. Proteins destined for continuous secretion (as apposed to regulated secretion) are targeted to transport vesicles that continuously fuse with the plasma membrane providing an exit route from the cell.

5.3 Mechanism of PTD cell entry

The cell membrane is relatively permeable to small-uncharged molecules, however provides a challenging barrier to charged macromolecules such as protein and DNA. The transport of macromolecules across the phospholipid bilayer is imperative for survival and the cell employs a number of methods to internalise extracellular macromolecules (Mukherjee, et al., 1997). Broadly, endocytosis can be divided into two: phagocytosis and pinocytosis. The former is a mechanism by which macrophages, monocytes and neutrophils ingest large microorganisms, insoluble particles, damaged or dead host cells, cell debris and activated clotting factors. Pinocytosis is a collective name describing the different modes of endocytosis involved in the internalisation of proteins, peptides and DNA.

Pinocytosis is divided into four types, each mode differing in mechanism, cell type and cargo. The best studied of these is that initiated through the formation of clathrin coated pits, during which receptor and bound ligand are constitutively internalised into coated pits. These coated pits are essentially invaginations in the phospholipid membrane, coated on the cytoplasmic side with the protein clathrin.

There are a number of non-clathrin dependant modes of endocytosis, which include macropinocytosis, caveolae-mediated endocytosis and constitutive non-clathrin uptake. Macropinosomes form from actin-driven protrusions from the phospholipid bilayer that capture extracellular milieu by collapsing onto the membrane. Macropinocytosis is a non-selective process of endocytosis, internalising fluid phase solutes (Swanson & Watts, 1995). Caveolae-mediated endocytosis involves the formation of flask shaped invaginations in the membrane. The protein caveolin binds cholesterol and organises the shape and structure of the caveloae. Caveolae-mediated endocytosis takes place in defined microdomains of the membrane (abundant on endothelial cells) that are particularly rich in cholesterol and sphingolipid (Anderson, 1998) and internalisation is thought to be regulated by dynamin (Nichols & Lippincott-Schwartz, 2001). Caveolae and clathrin coated regions of the cell membrane contain specific cell surface receptors. Hence, depending on which microdomain a particular receptor is in will determine the subsequent path of endocytosis. A fourth mode of pinocytosis is clathrin and caveloin independent endocytosis. Markers for these non-clathrin modes of endocytosis are found within cholesterol rich microdoamins on the cell membrane known as lipid rafts which can diffuse freely on the membrane (Brown & London, 1998, Simons & Ikonen, 1997). Presumably these lipid rafts can be captured by and internalised within any endocytic vesicle (Conner & Schmid, 2003). However, relatively little is known about this mode of endocytosis.

A study investigating cholesterol depleted membranes and visualisation of lipoplex (liposome and DNA) transfection identified clathrin dependant endocytosis is the main mechanism for cellular uptake of lipoplex formulations following electrostatic adsorption onto the negatively charged plasma membrane (Zuhorn, et al., 2002). However, direct fusion or a mode of fluid phase endocytosis may also contribute to internalisation. Internalised components are first targeted to the sorting endosome (early endosome), which contain Rab5 GTPase and early endosome antigen 1 (EEA1) markers. Much of the early work investigating endosome trafficking was done with the protein transferrin which is recycled back to the surface and low-density lipoprotein (LDL) which is delivered to the late endosomes. Early endosomes move as a whole unit and are subsequently delivered to late endosomes; this is termed the maturation model (Dunn & Maxfield, 1992). Late endosomes fuse with acidic lysosomes, which contain proteases and nucleases collectively referred to as acid hydrolases.

5.4 Lysomotrophic reagents

Chloroquine is a weak base and has been used as an endosomal buffering agent. Chloroquine aids early release from the endosome and acts to increase the intra-vesicle pH, which indirectly inhibits degradation of plasmid DNA. Viruses have evolved efficient mechanisms to facilitate early release from the endosome or completely bypass the endosome. Influenza virus haemagglutinin HA2 peptide inserts itself in the endosome membrane leading to lysis (Luo & Saltzman, 2000). Wadia and colleagues therefore designed a TAT-HA2 fusogenic peptide that acted as a good lysomotrophic agent without any noticeable cytotoxicity (Wadia, et al., 2004).

5.5 Protein Transduction Domains

Most protein transduction domains (PTD’s) consist of short basic peptide sequences (<20 amino-acids) that can cross the cell membrane of eukaryotic cells. PTD’s have been identified in a variety of proteins including the HIV-1 transactivator (TAT) protein (Frankel AD, 1988, Green M, 1988), a 16 amino-acid sequence corresponding to the third helix of the Antennapedia homeodomain (Antp) (Joliot, et al., 1991) and a short peptide (12 amino-acids) derived from the hydrophobic region of Kaposi fibroblast growth factor termed Membrane Translocating Sequence (MTS) (Lin, et al., 1995). PTD’s offer a versatile approach for drug delivery. In fact, over 95% of new therapeutic compounds are hindered by their low level of cellular transduction.

5.5.1 HIV-1 transactivator protein (TAT)

The first description of a protein with transduction activity was for the full-length transactivator (TAT) protein (86 amino acids) from the retrovirus HIV-1. It was shown that full length TAT protein is rapidly taken up from the cell medium and can stimulate HIV-LTR driven RNA syntheis in a variety of different cell types (Frankel AD, 1988, Green M, 1988). TAT mediated transactivation was enhanced in the presence of 100µM chloroquine (Frankel AD, 1988, Green M, 1988). The mechanism of TAT uptake was studied using radiolabelled and fluorescently purified TAT peptides. Both heparin and dextran sulphate were shown to inhibit cellular uptake suggesting that the basic region (TAT38-58) was necessary. This finding was reinforced with the demonstration that a monoclonal antibody to the same region also inhibited uptake (Mann & Frankel, 1991). Based on the demonstration of TAT association with the cell membrane and inhibition of uptake in some cells at 4˚C this suggested a mechanism of adsorptive endocytosis (Mann & Frankel, 1991). Fawell and colleagues utilised TAT peptides to deliver a variety of heterologous proteins to cells. TAT37-72 was chemically conjugated to β-galactosidase and delivery to cells in vitro and in vivo was shown (Fawell, et al., 1994). The highest levels of in vivo delivery following intravenous injection was seen in the endothelial cells surrounding blood vessels with some penetration into heart, liver and spleen (Fawell, et al., 1994). Synthesis of a range of peptides covering TAT37-60 identified the requirement of the basic region spanning amino acids 48-60 for transduction (Vives, et al., 1997) (this TAT sequence is referred to as TAT8 in this study). The authors monitored uptake of fluorescently labelled TAT peptides into the cell within 5 minutes. Schwarze and co-workers demonstrated delivery of a TAT-β-galactosidase fusion protein to all organs in the mouse including the brain (Schwarze, et al., 1999). The TAT peptide was flanked by glycine residues (GYGRKKRRQRRG) and fused in frame to the 120 kDa β-galactosidase protein in the pTATHA expression vector (Nagahara, et al., 1998). The authors attribute the improved results compared to previous in vivo attempts (Fawell, et al., 1994) to the in-frame fusion and protein purification strategy (Nagahara, et al., 1998). It should be noted that the pTATHA contains an in-frame positively charged histidine tag (His6) that is located at the 5’ of the TAT PTD sequence.

TAT mediated internalisation has been described to be arginine residue dependant with interaction of the guanidino head group (Mitchell, et al., 2000) and charged components of the cell membrane, a prerequisite to internalisation (Mitchell, et al., 2000). Charged surface heparan sulphate (HS) proteoglycans present on the membrane have been shown to be essential for TAT mediated uptake (Rusnati, et al., 1997, Rusnati, et al., 1998, Tyagi, et al., 2001). Recently TAT-mediated uptake into the cell has been proposed to be via fluid phase macropinocytosis (Kaplan, et al., 2005, Wadia, et al., 2004). In this study the authors used chloroquine to enhance delivery. TAT has been used to deliver liposomes (Hyndman, et al., 2004), DNA (Rudolph, et al., 2003), nanoparticles (Kleemann, et al., 2005), peptides and proteins (Lindsay, 2002, Wadia & Dowdy, 2005) into cells.

5.5.2 Antennapedia

A 60-aa peptide (pAntp) corresponding to the Drosophila antennapedia homeoprotein was shown to transduce the membrane of differentiated neurons and accumulate in the nucleus (Joliot, et al., 1991). Mechanistic studies have highlighted two amino-acids essential for pAntp internalisation (Roux, et al., 1993). The authors also present a mutant antennapedia peptide (pAntp48S) where two hydrophobic residues were removed and a conversion of Gln50 to Ser50. It was first thought that the hydrophobic residues were essential in formation of a structural motif necessary for internalisation (Roux, et al., 1993). Derossi and co-workers synthesised several peptides in an attempt to home in on the protein transduction domain. A 16 amino acid segment (also known as penetratin) corresponding to the third helix of the antennapedia homeoprotein was identified as the minimal essential peptide required for transduction (Derossi, et al., 1994). This peptide was shown to form an alpha helical structure in a hydrophobic environment. Alpha helical peptides have been shown to translocate polar cargos across the cell membranes in a non-endocytic process (Vlassov, et al., 1994). The possibility that this could account for pAntp uptake was discounted when proline amino acids introduced to disrupt the alpha helical structure had no effect on internalisation, neither did reversing the helix or composing the peptide of D-enantiomers (Derossi, et al., 1996). The authors concluded that penetratin internalisation did not take place via a receptor-mediated pathway. A two-step non-receptor dependant internalisation model has been proposed (Dom, et al., 2003). The first step requires association with and internalisation into the hydrophobic environment of the phospholipid bilayer. This is followed by a membrane destabilising step, providing transport into the cytoplasm (Dom, et al., 2003) in which the tryptophan amino-acid at position 6 is essential. Penetratin and TAT are similar in their requirement of basic amino-acids for uptake into the cell. However, TAT has been used more widely for the transduction of large cargo proteins both in vitro and in vivo. Penetratin has been used mainly to deliver peptides (Lindsay, 2002) but less so for larger protein cargos.

5.5.3 MTS

Membrane translocating sequence (MTS) and cell penetrating peptides (CPP) have been used as general terms for protein transduction domains. Here the term MTS is used as a name for a particular PTD sequence derived from the hydrophobic region of the Kaposi fibroblast growth factor signal sequence. The authors reasoned that if a signal sequence can lead to export of cargo out of the cell then it might also be able to transport cargo in the opposite direction (Lin, et al., 1995). Indeed, they demonstrate internalisation of the MTS peptide and an attached NLS from the intracellular signalling molecule NF-kB. Subsequently MTS has been used to deliver Cre recombinase protein (@40 kDa) into ROSA26R mice where CRE-mediated recombination drives expression of the lacZ gene and production of β-galactosidase protein (Jo, et al., 2001). Accomodation of β-galactosidase was found in all the organs including the brain following intraperitoneal injection of bacterially purified His6-NLS-CRE-MTS fusion protein.

5.5.4 Poly-arginine peptides

It is clear that both TAT and penetratin rely on their cationic amino-acid composition for effective translocation of the cell membrane. However, in recent years it has become apparent that arginine might be more effective than lysine in these sequences. Most of the studies performed with cationic peptides have concentrated on condensing the cargo and not actually transporting it. Mitchell and co-workers compared cationic peptides of arginine, lysine and histidine for uptake into cells. Using confocal microscopy and flow-cytometry they demonstrated increased uptake of arginine peptides into Jurkat T-cells (Mitchell, et al., 2000). The optimal length of the arginine peptide was between 6 and 15 amino acids. Substituting arginine for its isomer citrulline confirmed the requirement of the guanidino head group for uptake. Citrulline has an oxygen in the same position that arginine has a nitrogen and citulline peptides were not internalised (Mitchell, et al., 2000). The guanidine head group can potentially form stable hydrogen bonds with phosphate or sulphate groups on the phospholipid membrane whereas lysine and histidine cannot. A peptide of nine arginine residues (R9) was shown to be 100-fold faster in uptake compared to TAT47-57 (Wender, et al., 2000). Oligoarginine was chemically cross-linked to the small anti-inflammatory drug Cyclosporin A (CsA) demonstrating delivery to both mouse and human skin following topical delivery (Rothbard, et al., 2000).

A recent publication investigated the effect of the density of peptides in the transfection mixture. The authors showed that a high-density of ARG8 oligonucleotides in a lipoplex, directed internalisation via macropinocytosis whilst a low density was internalised through endocytosis (Khalil, et al., 2006). The authors attribute the lower level of gene expression in the latter case with increased degradation of transfection mix in the lysosome.