Next-generation sequencing has enhanced the phage display process, allowing for the quantification of millions of sequences resulting from the biopanning process. In response, many valuable analysis programs focused on specificity and finding targeted motifs or consensus sequences were developed. For targeted drug delivery and molecular imaging, it is also necessary to find peptides that are selective—targeting only the cell type or tissue of interest. We present a new analysis strategy and accompanying software, PHage Analysis for Selective Targeted PEPtides (PHASTpep), which identifies highly specific and selective peptides. Using this process, we discovered and validated, both in vitro and in vivo in mice, two sequences (HTTIPKV and APPIMSV) targeted to pancreatic cancer-associated fibroblasts that escaped identification using previously existing software. Our selectivity analysis makes it possible to discover peptides that target a specific cell type and avoid other cell types, enhancing clinical translatability by circumventing complications with systemic use.

Introduction

Phage display enables the simultaneous, high throughput screening of billions of different peptides against diverse target molecules including proteins, miRNAs, polysaccharides, cells, and tissues. Phage display’s utility and versatility have made it an important technique in the identification of molecularly targeted affinity agents for imaging, targeted drug therapy, and biosensor applications. In addition to the biomedical field, phage display derived peptides that bind certain types of inorganic materials have been identified, such as gold-binding sequences, and have established insight into ligand recognition mechanisms and quantitative affinity analysis. Further, phage display has been used to develop self-assembling batteries.

Phage display offers a number of important advantages in identifying targeted peptides such as rapid and economical biological expansion, vast peptide diversity, a rapid screening process, and the availability of many types of phage clones and libraries (for review see). Another important advantage is that bacteriophage, unlike higher organisms, have only one copy of each gene, so it is easy to identify the displayed peptide of a clone by sequencing the appropriate portion of the phage genome. In a process called biopanning, a phage library is exposed to a target, non-bound phage are washed away, and bound phage are eluted off then amplified in their bacterial host. Theoretically, the amplified library has an increased proportion of phage clones that bind the target and can be further enriched by being subjected to additional rounds of panning. However, traditional screening protocols are hampered by false positive rates caused by non-specific phage binding and unequal rates of amplification as well as by loss of potential candidates early in the process due to low starting phage concentrations. Following the iterative selection process, it is necessary to differentiate between phage clones that bind the target specifically and those that have been co-extracted in the enriched library due to non-specific interactions. Initially, the standard technique was to pick and amplify individual clones from the final enriched phage pool and test them side-by-side in an ELISA to distinguish the specific binders. The identity of the displayed peptides with specific binding could be individually determined by Sanger sequencing.

Software programs have been developed to assist in the analysis of peptides identified through phage display. One of the earliest programs, REceptor LIgand Contacts (RELIC) is still popular in the phage display community. Although RELIC was limited by biases associated with the original phage display technique, it made the great contribution of enabling users to align sequences and find motifs from phage display experimental results. Since RELIC, databases such as PEPBANK and MimoDB, have been generated to browse for similar peptide sequences among those compiled from previously-conducted biopanning experiments. Other algorithms and programs were used to analyse results via methods that aligned sequences, performed epitope-mapping, and identified motifs, including MIMOP, PEPTIDE, SiteLight, Multiple Em for Motif Elicitation (MEME) [16], DNAStar, Short Linear Motif Finder (SLiMFinder), and Multiple Alignment using Fast Fourier Transform (MAFFT) [19]. Importantly, all of these methods were designed to interpret the small-scale results of traditional phage display. Thus they were still susceptible to many of the aforementioned weaknesses of the technique and a much deeper characterization of the post-biopanning enriched phage library would later prove to further enhance analytic capabilities.

The necessary deeper characterization became attainable with the advent of next generation sequencing (NGS) techniques, whereby it became possible to increase the yield of sequences from hundreds to millions and overcome some of the drawbacks of phage display. NGS has similarly been employed in other in vitro selection fields, including yeast display, mRNA display, antibody display, protein domains, and aptamer selection. In the peptide phage display field, NGS has highlighted the differences in amplification rates among phage clones that culminate in domination of the final enriched phage libraries by so-called parasitic sequences, leading to development of techniques that use modified amplification techniques or only one round of panning in order to circumvent this amplification bias. The effect of sequencing errors in the context of phage display has also been explored. Most importantly, the use of NGS with phage has been successful in moving forward the field of phage display. Yet, NGS also introduced the challenge of determining which sequences represent the most ideal ligands since the results still conflate enrichment due to ligand specificity with the enrichment due to non-specific binding and, when more than one round of panning is necessary, variation in amplification rate.

Since the onset of deep sequencing techniques, exquisite work has been carried out to build software for processing and translation of sequences and finding consensus sequences or motifs, including MATLAB-based translation software, Multiple Specificity Identifier (MUSI) , and target-binding motif analysis. Additionally, powerful analytic methods and software have been developed for processing data, clustering sequences and comparing selective versus non-selective libraries, affinity ranking, and statistical-based comparisons. While these tools provide valuable insight into the specificity and affinity of peptides, they do not address selectivity. Selectivity is an especially important parameter for use of phage display in finding clinically relevant targeted peptides. For example, phage display has been used to identify targeted peptides that can distinguish between diseased and healthy cells, which in turn can be used for molecular imaging or to alter the toxicity profile of drugs via targeted delivery. Due to the physiologic presence of numerous cell types and receptors, it is necessary to ensure that a given peptide is not only specific for the target of interest but also that it will not bind to other locations (i.e., the molecular target is only present on the cell of interest).