The plasma membrane of a cell is essential to its identity and survival, but at the same time presents a barrier to intracellular delivery of potentially diagnostic or therapeutic cargoes. Therefore, the development of approaches to deliver functional cargoes, be they peptides, proteins, nucleic acids, or nanoparticles across cell membranes, a process termed protein transduction, has wide-reaching research and clinical implications. The ability of Trans-Activator of Transcription (Tat) protein of the Human Immunodeficiency Virus (HIV) to transduce cultured cells and lead to viral gene expression ^[1][2][1,2] without requiring a receptor was the first example of a protein that naturally employs a portion of itself to achieve cell penetration and lead to intracellular delivery of the entire HIV viral particle. The chemical cross-linking of a full-length Tat protein to multiple different proteins such as horseradish peroxidase, ß-galactosidase, RNase A and domain III of Pseudomonas exotoxin A served to demonstrate the ability of Tat protein to ferry other large cargoes across the cell membrane [3]. Similarly, the homeobox Antennapedia (Antp) transcription factor of Drosophila melanogaster was demonstrated to enter nerve cells in a receptor independent manner where it could then regulate neural morphogenesis [4]. Mapping of the domains within Tat and Antp responsible for the observed transduction led to the identification of the first two cell penetrating peptides (CPPs): the 11 amino acid cationic domain of HIV-1 Tat protein (YGRKKRRQRRR) [5] and the 16 amino acid sequence from the third helix of the Antennapedia domain (RQIKIWFQNRRMKWKK) termed Antp or penetratin [6]. Subsequently, the ability of the small part of the full-length Tat protein to deliver cargoes, including other full length proteins and even large multimeric protein complexes across cell membranes in culture and in vivo following systemic delivery in mice [7] was documented, further highlighting the delivery potential of these unique peptides. Since then, the number of peptides, both cell-specific and non-specific, reported as having cell penetrating properties has increased exponentially [8]. This is particularly true for a wide spectrum of cationic peptides that primarily rely on their cationic charge to interact with proteoglycans on the cell surface (see below). There has been intense interest in identifying both new cell-specific CPPs, as well as strategies to make Tat and other non-specific CPPs act in a more cell-specific manner by taking advantage of tissue characteristics, mostly in the context of cancer.

Cell-specific CPPs were identified predominantly by using peptide phage display libraries to screen for peptides able to target specific cell types. The concept of phage display was first proposed by Smith in 1985 ^[39][29]. Following this initial report, combinatorial peptide libraries of various lengths using different types of phages (M13, T7) have been used successfully to identify peptides able to facilitate internalization of intact phage. Alternatively, plasmid, antibodies, microorganism surface or ribosome displays of peptide libraries have been employed as well. Phage display requires exposing the target cell or tissue of interest to a large, randomized phage library in which one of the envelope proteins used by the phage for internalization has been modified to display linear or cyclic peptides of various lengths and randomized amino acid sequences ^[40][30]. The internalized phage can then be isolated, expanded and used in subsequent rounds of screening. Usually 3–5 rounds of screening results in enrichment of a small number of peptides identifiable by DNA sequencing of the recovered phage. This approach requires enriching for a specific, small subset of phage from a very large library. As such, false positives are a concern and have to be discerned from phage that is indeed bound and internalized by the target cell type. One approach to circumvent this problem is to carry out the first cycle in cell culture using relevant cell types in order to reduce the likelihood of false positives. Subsequent cycles can then be carried out in vivo using the enriched pool from the in vitro cycle ^[40][23][41][30,31,32]. Such an approach has led to the identification of peptides targeting vascular endothelium ^[28][33], synovial tissue ^[26][27], dendritic cells ^[42][34], pancreatic islet cells ^[43][35] and cardiac myocytes ^[23][31], and has identified NRG (Asparginine-Arginine-Glycine) and RGD (Arginine-Glycine-Aspartic acid) motifs that target phage to tumor vasculature in nude tumor-bearing mice ^[44][36].

Despite intense study of CPPs, the specific pathway(s) involved in facilitating transduction remain elusive. CPPs are short in length, making the use of standard techniques for identifying binding targets on the cell surface more challenging. Additionally, their rapid cell entry, occurring within minutes at physiological conditions, makes analysis difficult. It is likely that a non-cell specific CPP such as Tat that crosses the blood brain barrier will not share a cell entry pathway with a cell-specific CPP. Even for a particular CPP the mechanism of transduction likely varies depending on the specific cargo fused to it, with changes in parameters such as local milieu and pH almost certainly playing a part. This is further complicated by recent data suggesting that the local concentration of a CPP may influence the internalization pathway used ^[45][46][47][37,38,39]. It should be noted that elucidation of the mechanism of transduction is not only of theoretical interest as the loading of cargoes must be achieved in a way as to not interfere with either the binding or cell internalization mechanism of CPPs. For example changing the hydrophobicity, but not the cationic charge, of a guanidine-rich homo-polymer significantly changed its transduction abilities and ability to internalize cargoes ^[48][40].

Although the exact transduction mechanism of CPPs remains elusive, extensive work from multiple investigators has shed considerable light. There is evidence both for mechanisms that are non-endocytic/energy independent or endocytic/energy-dependent ^[49][41]. The broad range of cells that are readily transducible by non-specific CPPs such as Tat suggests a role for ubiquitously shared cellular structures such as surface binding to plasma membrane phospholipids or, in particular proteoglycans through electrostatic interactions, as a first step towards cell entry ^[50][51][42,43]. Cell lines deficient in heparan sulfate have significantly reduced transduction by cationic peptides ^[10][52][53][13,44,45]. This reduction suggests that electrostatic interactions on the cell surface, separate from CPP-lipid interactions, contribute to protein transduction. It is likely that transduction is a two-step process, with the first step being electrostatic interaction of non-specific CPPs with anionic elements, such as glycosaminoglycans on the cell surface that draw the CPPs into close proximity to the plasma membrane. Subsequently, cationic CPPs bearing small cargoes likely enter cells via direct translocation, with uptake of larger cargoes mediated by micropinocytosis, a more energy-dependent and slower process ^[54][46]. Transduction has been shown to occur at 4 °C and after depletion of the adenosine triphosphate (ATP) pool ^[10][13], albeit at a reduced level, suggesting that it is not exclusively an energy-dependent process. Research also suggests that increasing hydrophobic characteristics of a CPP, as in the case of Tat, increases its efficiency as a transporter ^[55][47].

Although CPPs have been used as vectors for delivery of drugs ^{[56][57][58][59]}[48,49,50,51], other peptides of therapeutic potential ^{[60][61][62][63][64][65][66][67][68]}[52,53,54,55,56,57,58,59,60], proteins ^{[64][69][70][71][72][73]}[56,61,62,63,64,65], radioisotopes ^{[24][74][75][76]}[66,67,68,69], quantum dots ^[77][78][70,71] and photosensitizers ^[79][80][72,73], this entry will focus on use of CPPs as vectors for nucleic acid delivery, be they genes, oligonucleotides, peptide nucleic acid conjugates, small interfering RNA (siRNA) or the newest application with DNA origami. Although there is literature on all of these applications, most interest and work done to date has been with siRNA. The necessary factors to consider in these applications is the platform’s ability to successfully conjugate nucleic acids to CPPs, escape from enzymatic degradation in serum, escape the reticuloendothelial compartments, successfully cross cell membranes, escape from endocytic degradation and, in cases of gene delivery, achieve nuclear localization.