Discovery of active proteins directly from combinatorial randomized protein libraries without display, purification or sequencing: identification of novel zinc finger proteins

We have successfully linked protein library screening directly with the identification of active proteins, without the need for individual purification, display technologies or physical linkage between the protein and its encoding sequence. By using ‘MAX’ randomization we have rapidly constructed 60 overlapping gene libraries that encode zinc finger proteins, randomized variously at the three principal DNA-contacting residues. Expression and screening of the libraries against five possible target DNA sequences generated data points covering a potential 40000 individual interactions. Comparative analysis of the resulting data enabled direct identification of active proteins. Accuracy of this library analysis methodology was confirmed by both in vitro and in vivo analyses of identified proteins to yield novel zinc finger proteins that bind to their target sequences with high affinity, as indicated by low nanomolar apparent dissociation constants.     

Chemically synthesized peptide libraries as a new source of BBB shuttles. Use of mass spectrometry for peptide identification

The blood–brain barrier (BBB) is a biological barrier that protects the brain from neurotoxic agents and regulates the influx and efflux of molecules required for its correct function. This stringent regulation hampers the passage of brain parenchyma‐targeting drugs across the BBB. BBB shuttles have been proposed as a way to overcome this hurdle because these peptides can not only cross the BBB but also carry molecules which would otherwise be unable to cross the barrier unaided. Here we developed a new high‐throughput screening methodology to identify new peptide BBB shuttles in a broadly unexplored chemical space. By introducing d‐ amino acids, this approach screens only protease‐resistant peptides. This methodology combines combinatorial chemistry for peptide library synthesis, in vitro models mimicking the BBB for library evaluation and state‐of‐the‐art mass spectrometry techniques to identify those peptides able to cross the in vitro assays. BBB shuttle synthesis was performed by the mix‐and‐split technique to generate a library based on the following: Ac‐d‐ Arg‐XXXXX‐NH₂, where X were: d‐ Ala (a), d‐ Arg (r), d‐ Ile (i), d‐ Glu (e), d‐ Ser (s), d‐ Trp (w) or d‐ Pro (p). The assays used comprised the in vitro cell‐based BBB assay (mimicking both active and passive transport) and the PAMPA (mimicking only passive diffusion). The identification of candidates was determined using a two‐step mass spectrometry approach combining LTQ‐Orbitrap and Q‐trap mass spectrometers. Identified sequences were postulated to cross the BBB models. We hypothesized that some sequences cross the BBB through passive diffusion mechanisms and others through other mechanisms, including paracellular flux and active transport. These results provide a new set of BBB shuttle peptide families. Furthermore, the methodology described is proposed as a consistent approach to search for protease‐resistant therapeutic peptides. Copyright © 2016 European Peptide Society and John Wiley & Sons, Ltd.     

Soluble Combinatorial Libraries: Extending the Range and Repertoire of Chemical Diversity

The generation and use of libraries made up of millions of chemical entities and the ability to identify the active compound in such libraries are at the forefront of a revolution in drug discovery and basic research. Such libraries, when made up of peptide sequences, offer a fundamental, practical advance in the study of interactions between peptides and their biochemical or pharmacological targets. The utility of soluble peptide libraries ranging from three to eight amino acids in length, and made up of mixtures from 361 tripeptides to 200 billion decapeptides, is described. These are readily usable in virtually all in vitro (and even in vivo) assay systems. The examples presented illustrate the utility of soluble peptide libraries for the study of antibody/antigen interactions, the identification of highly active opioid peptides in receptor binding studies using crude rat brain homogenates, and in vivo studies, in which the peptide mixtures making up the library are administered intravenously to determine peptide sequences that affect heart rate and blood pressure. A new class of library is also described, termed a modified peptide library, which is used to determine potent anti-Staphylococcus aureus compounds.     

Selection of biologically active peptides by phage display of random peptide libraries

Random peptide libraries displayed on phage are used as a source of peptides for epitope mapping, for the identification of critical amino acids responsible for protein—protein interactions and as leads for the discovery of new therapeutics. Efficient and simple procedures have been devised to select peptides binding to purified proteins, to monoclonal and polyclonal antibodies and to cell surfaces in vivo and in vitro.     

The Use of Phage-Displayed Peptide Libraries to Develop Tumor-Targeting Drugs

Monoclonal antibodies have been successfully utilized as cancer-targeting therapeutics and diagnostics, but the efficacies of these treatments are limited in part by the size of the molecules and non-specific uptake by the reticuloendothelial system. Peptides are much smaller molecules that can specifically target cancer cells and as such may alleviate complications with antibody therapy. Although many endogenous and exogenous peptides have been developed into clinical therapeutics, only a subset of these consists of cancer-targeting peptides. Combinatorial biological libraries such as bacteriophage-displayed peptide libraries are a resource of potential ligands for various cancer-related molecular targets. Target-binding peptides can be affinity selected from complex mixtures of billions of displayed peptides on phage and further enriched through the biopanning process. Various cancer-specific ligands have been isolated by in vitro, in vivo, and ex vivo screening methods. As several peptides derived from phage-displayed peptide library screenings have been developed into therapeutics in current clinical trials, which validates peptide-targeting potential, the use of phage display to identify cancer-targeting therapeutics should be further exploited.

Protein aggregation as an antibiotic design strategy

Taking advantage of the xenobiotic nature of bacterial infections, we tested whether the cytotoxicity of protein aggregation can be targeted to bacterial pathogens without affecting their mammalian hosts. In particular, we examined if peptides encoding aggregation‐prone sequence segments of bacterial proteins can display antimicrobial activity by initiating toxic protein aggregation in bacteria, but not in mammalian cells. Unbiased in vitro screening of aggregating peptide sequences from bacterial genomes lead to the identification of several peptides that are strongly bactericidal against methicillin‐resistant Staphylococcus aureus. Upon parenteral administration in vivo, the peptides cured mice from bacterial sepsis without apparent toxic side effects as judged from histological and hematological evaluation. We found that the peptides enter and accumulate in the bacterial cytosol where they cause aggregation of bacterial polypeptides. Although the precise chain of events that leads to cell death remains to be elucidated, the ability to tap into aggregation‐prone sequences of bacterial proteomes to elicit antimicrobial activity represents a rich and unexplored chemical space to be mined in search of novel therapeutic strategies to fight infectious diseases.     

The fluorescent protein iLOV as a reporter for screening of high-yield production of antimicrobial peptides in Pichia pastoris

The methylotrophic yeast Pichia pastoris is commonly used for the production of recombinant proteins at scale. The identification of an optimally overexpressing strain following transformation can be time and reagent consuming. Fluorescent reporters like GFP have been used to assist identification of superior producers, but their relatively big size, maturation requirements and narrow temperature range restrict their applications. Here, we introduce the use of iLOV, a flavin-based fluorescent protein, as a fluorescent marker to identify P. pastoris high-yielding strains easily and rapidly. The use of this fluorescent protein as a fusion partner is exemplified by the production of the antimicrobial peptide NI01, a difficult target to overexpress in its native form. iLOV fluorescence correlated well with protein expression level and copy number of the chromosomally integrated gene. An easy and simple medium-throughput plate-based screen directly following transformation is demonstrated for low complexity screening, while a high-throughput method using fluorescence-activated cell sorting (FACS) allowed for comprehensive library screening. Both codon optimization of the iLOV_NI01 fusion cassettes and different integration strategies into the P. pastoris genome were tested to produce and isolate a high-yielding strain. Checking the genetic stability, process reproducibility and following the purification of the active native peptide are eased by visualization of and efficient cleavage from the iLOV reporter. We show that this system can be used for expression and screening of several different antimicrobial peptides recombinantly produced in P. pastoris.     

A novel method for construction of gene fragment library to searching epitopes

Identification of the epitope sequence or the functional domain of proteins is a laborious process but a necessary one for biochemical and immunological research. To achieve intensive and effective screening of these functional peptides in various molecules, we established a novel screening method using a phage library system that displays various lengths and parts of peptides derived from target protein. Applying this library for epitope mapping, epitope peptide was more efficiently identified from gene fragment library than conventional random peptide library. Our system may be a most powerful method for identifying functional peptides.     

Antibody-free peptide substrate screening of serine/threonine kinase (protein kinase A) with a biotinylated detection probe

Being different from anti-phosphotyrosine antibodies, anti-phosphoserine- or anti-phosphothreonine-specific antibodies with high affinity for the detection of serine/threonine kinase substrates are not readily available. Therefore, chemical modification methods were developed for the detection of phosphoserine or threonine in the screening of protein kinase substrates based on β-elimination and Michael addition. We have developed a biotin-based detection probe for identification of the phosphorylated serine or threonine residue. A biotin derivative induced a color reaction using alkaline phosphate-conjugated streptavidin that amplified the signal. It was effective for the detection and separation of the target peptide on the resin. The detection probe was successfully used in identifying PKA substrates from peptide libraries on resin beads. The peptide library was prepared as a ladder-type, such that the active peptides on the colored resin beads were readily sequenced with the truncated peptide fragments by MALDI-TOF/MS analysis after releasing the peptides from the resin bead through photolysis.     

Parallel in Vivo and in Vitro Selection Using Phage Display Identifies Protease-dependent Tumor-targeting Peptides

We recently developed activatable cell-penetrating peptides (ACPPs) that target contrast agents to in vivo sites of matrix metalloproteinase activity, such as tumors. Here we use parallel in vivo and in vitro selection with phage display to identify novel tumor-homing ACPPs with no bias for primary sequence or target protease. Specifically, phage displaying a library of ACPPs were either injected into tumor-bearing mice, followed by isolation of cleaved phage from dissected tumor, or isolated based on selective cleavage by extracts of tumor versus normal tissue. Selected sequences were synthesized as fluorescently labeled peptides, and tumor-specific cleavage was confirmed by digestion with tissue extracts. The most efficiently cleaved peptide contained the substrate sequence RLQLKL and labeled tumors and metastases from several cancer models with up to 5-fold contrast. This uniquely identified ACPP was not cleaved by matrix metalloproteinases or various coagulation factors but was efficiently cleaved by plasmin and elastases, both of which have been shown to be aberrantly overexpressed in tumors. The identification of an ACPP that targets tumor expressed proteases without rational design highlights the value of unbiased selection schemes for the development of potential therapeutic agents.     

Extract of Caragana sinica as a potential therapeutic option for increasing alpha-secretase gene expression

BackgroundAlzheimer's disease represents one of the main neurological disorders in the aging population. Treatment options so far are only of symptomatic nature and efforts in developing disease modifying drugs by targeting amyloid beta peptide-generating enzymes remain fruitless in the majority of human studies. During the last years, an alternative approach emerged to target the physiological alpha-secretase ADAM10, which is not only able to prevent formation of toxic amyloid beta peptides but also provides a neuroprotective fragment of the amyloid precursor protein – sAPPalpha.PurposeTo identify novel alpha-secretase enhancers from a library of 313 extracts of medicinal plants indigenous to Korea, a screening approach was used and hits were further evaluated for their therapeutic value.MethodsThe extract library was screened for selective enhancers of ADAM10 gene expression using a luciferase-based promoter reporter gene assay in the human neuroblastoma cell line SH-SY5Y. Candidate extracts were then tested in wild type mice for acute behavioral effects using an open field paradigm. Brain and liver tissue from treated mice was biochemically analyzed for ADAM10 gene expression in vivo. An in vitro blood–brain barrier model and an in vitro ATPase assay were used to unravel transport properties of bioactive compounds from extract candidates. Finally, fractionation of the most promising extract was performed to identify biologically active components.ResultsThe extract of Caragana sinica (Buc'hoz) Rehder was identified as the best candidate from our screening approach. We were able to demonstrate that the extract is acutely applicable in mice without obvious side effects and induces ADAM10 gene expression in peripheral tissue. A hindered passage across the blood–brain barrier was detected explaining lack of cerebral induction of ADAM10 gene expression in treated mice. By fractionating C. sinica extract we identified alpha-viniferin as one of the biologically active components.ConclusionThe extract of C. sinica and alpha-viniferin as one of its bioactive constituents might serve as novel therapeutic options for treating Alzheimer's disease by increasing ADAM10 gene expression. The identification of alpha-viniferin represents a promising starting point to achieve blood–brain barrier penetrance in the future.     

A novel peptide specifically targeting ovarian cancer identified by in vivo phage display

Discovery of peptide ligands that can target human ovarian cancer and deliver chemotherapeutics offers new opportunity for cancer therapy. The advent of phage‐displayed peptide library facilitated the screening of such peptides. In vivo screening that set in a microanatomic and functional context was applied in our study, and a novel peptide WSGPGVWGASVK targeting ovarian cancer was isolated. The phage clone PC3‐1 displaying peptide WSGPGVWGASVK can gain effective access to accumulate in the tumor sites after intravenous injection while reducing its accumulation in normal organs. Positive immunostaining of PC3‐1 was located in both sites of tumor cells and tumor blood vessels, which resulted in a diffuse binding pattern through the tumor. In vitro study results confirmed the capability of peptide WSGPGVWGASVK binding to and being internalized by both tumor cells and angiogenic endothelial cells. Flow cytometry analysis revealed that the peptide bound to SKOV3 cells with Kd value of 5.43 ± 0.4 μM. Taken together, it suggested that peptide WSGPGVWGASVK is a lead candidate for delivering therapeutics to penetrate into tumors. Copyright © 2013 European Peptide Society and John Wiley & Sons, Ltd.     

A designed glycopeptide array for characterization of sugar-binding proteins toward a glycopeptide chip technology

For the realization of a practical high-throughput protein detection and analysis system, a novel peptide array has been constructed using a designed glycopeptide model library with an α-helical secondary structure. This study will contribute the increment of the diversity of such an array system and the application to focused proteomics and ligand screening by effective detection of sugar-binding proteins. Fluorescent glycopeptides with an α-helix, a β-strand, or a loop structure were designed initially to select a suitable scaffold for the detection of a model protein. After selection of the α-helical structure as the best scaffold, a small model library with various saccharides was constructed to have charge and hydrophobicity variations in the peptide sequences. When various sugar-binding proteins were added to the peptide library array, the fluorescent peptides showed different responses in fluorescence intensities depending on their sequences as well as saccharides. The patterns of these responses could be regarded as “protein fingerprints” (PFPs), which are able to establish the identities of the target proteins. The resulting PFPs reflected the recognition properties of the proteins. Furthermore, statistical data analysis from obtained PFPs was performed using a cluster analysis. The PFPs of sugar-binding proteins were clustered successfully depending on their families and binding properties. These studies demonstrate that arrays with glycopeptide libraries based on designed structures can be promising tools to detect and analyze the target proteins. Designed peptides with functional groups such as sugars will play roles as the capturing agents of high-throughput protein nano/micro arrays for focused proteomics and ligand screening studies.     

Identification of novel leads applying in silico studies for Mycobacterium multidrug resistant (MMR) protein

Multidrug efflux mechanism is the main cause of intrinsic drug resistance in bacteria. Mycobacterium multidrug resistant (MMR) protein belongs to small multidrug resistant family proteins (SMR), causing multidrug resistance to proton (H⁺)-linked lipophilic cationic drug efflux across the cell membrane. In the present work, MMR is treated as a novel target to identify new molecular entities as inhibitors for drug resistance in Mycobacterium tuberculosis. In silico techniques are applied to evaluate the 3D structure of MMR protein. The putative amino acid residues present in the active site of MMR protein are predicted. Protein–ligand interactions are studied by docking cationic ligands transported by MMR protein. Virtual screening is carried out with an in-house library of small molecules against the grid created at the predicted active site residues in the MMR protein. Absorption distribution metabolism and elimination (ADME) properties of the molecules with best docking scores are predicted. The studies with cationic ligands and those of virtual screening are analysed for identification of new lead molecules as inhibitors for drug resistance caused by the MMR protein.     

In vivo phage display: identification of organ-specific peptides using deep sequencing and differential profiling across tissues

In vivo phage display is widely used for identification of organ- or disease-specific homing peptides. However, the current in vivo phage biopanning approaches fail to assess biodistribution of specific peptide phages across tissues during the screen, thus necessitating laborious and time-consuming post-screening validation studies on individual peptide phages. Here, we adopted bioinformatics tools used for RNA sequencing for analysis of high-throughput sequencing (HTS) data to estimate the representation of individual peptides during biopanning in vivo. The data from in vivo phage screen were analyzed using differential binding—relative representation of each peptide in the target organ versus in a panel of control organs. Application of this approach in a model study using low-diversity peptide T7 phage library with spiked-in brain homing phage demonstrated brain-specific differential binding of brain homing phage and resulted in identification of novel lung- and brain-specific homing peptides. Our study provides a broadly applicable approach to streamline in vivo peptide phage biopanning and to increase its reproducibility and success rate.     

Improved protein sequence coverage by on resin deglycosylation and cysteine modification for biomarker discovery

Membrane proteins and secreted factors (soluble proteins or extracellular matrix components) are the targets of most monoclonal antibodies, which are currently in clinical development. These proteins are frequently post‐translationally modified, e.g. by the formation of disulfide bonds or by glycosylation, which complicates their identification using proteomics technologies. Here, we describe a novel methodology for the on resin deglycosylation and cysteine modification of proteins after in vitro, in vivo or ex vivo biotinylation. Biotinylated proteins are captured on streptavidin resin and all subsequent modifications, as well as the proteolytic digestion, which yields peptides for MS analysis, are performed on resin. Using biotinylated bovine fetuin‐A as a test protein, an improvement in sequence coverage from 7.9 to 58.7% could be shown, including the identification of all three glycosylation sites. Furthermore, a complex mixture derived from the ex vivo biotinylation of vascular structures in human kidney with cancer obtained by perfusion after surgical resection revealed almost a doubling of sequence coverage for all checked proteins when analyzed by LC‐MALDI TOF/TOF.     

Cloning, expression, purification, distribution and kinetics characterization of the bacterial β-galactosidase fused to the cytoplasmic transduction peptide in vitro and in vivo

Cytoplasmic transduction peptide (CTP) offers exciting therapeutic opportunities for the treatment of many diseases caused by cytoplasmic functional molecules. It can transduce large, biologically active proteins into the cytoplasmic compartment of several mammalian cells. However, other intriguing features of CTP, including its activity in vitro, and distribution and tissue infiltration abilities in vivo, remain to be explored. The present study was initiated to (1) further confirm the cytoplasmic localization preference and the enzymatic activity of the transduced CTP-β-gal in vitro and (2) examine the kinetics and tissue distribution of the CTP-β-gal fusion protein in mice. A CTP-β-gal fusion protein was expressed in Escherichia coli and either transduced into BaF3-BCR/ABL cells or administered intravenously into female Balb/C mice at a dose of 100 μg per mouse. Its localization in BaF3-BCR/ABL cells was evaluated by immunocytochemistry and in situ X-gal staining, and its distribution in various tissues was analyzed both by in situ X-gal staining and quantitative enzymatic activity assay. β-Galactosidase enzyme activity was observed in BaF3-BCR/ABL cells and in all tissues tested, with peak activity occurring at 15 min in most tissues and at 24 h in brain. These data will not only allow rational selection of delivery schedules for therapeutic CTP, but will also aid the use of CTP fusion protein transduction in the development of protein therapeutics targeting the cytoplasmic compartment both in vitro and in vivo.     

Characterization of New Milk-derived Inhibitors of Angiotensin Converting Enzyme In Vitro and In Vivo

Inhibition of angiotensin converting enzyme (ACE) has been observed with a variety of different peptides, and peptide fragments with inhibitory capabilities have been identified within many different proteins, including milk proteins. The purpose of this study therefore was to identify new short peptides with inhibitory properties from the primary structure of milk proteins and to characterize them in vitro and in vivo, since no milk derived ACE inhibitors have previously been evaluated for their ability to inhibit ACE in vivo. In vitro, 8 of 9 dipeptides were found to be competitive inhibitors of ACE. The IC₅₀ was significantly lower when an angiotensin I-like substrate was used, than when a bradykinin-like substrate was used. Using three different in vivo models for ACE inhibition, a very moderate effect was observed for three of the new peptides, but only for up to 6 or 12 minutes. Nothing was observed with two reference compounds that are reported to be hypotensive ACE-inhibitors derived from milk proteins. This raises the question whether the mechanism of hypotensive action is straightforward inhibition of ACE in vivo.     

A Technology for Developing Synbodies with Antibacterial Activity

The rise in antibiotic resistance has led to an increased research focus on discovery of new antibacterial candidates. While broad-spectrum antibiotics are widely pursued, there is evidence that resistance arises in part from the wide spread use of these antibiotics. Our group has developed a system to produce protein affinity agents, called synbodies, which have high affinity and specificity for their target. In this report, we describe the adaptation of this system to produce new antibacterial candidates towards a target bacterium. The system functions by screening target bacteria against an array of 10,000 random sequence peptides and, using a combination of membrane labeling and intracellular dyes, we identified peptides with target specific binding or killing functions. Binding and lytic peptides were identified in this manner and in vitro tests confirmed the activity of the lead peptides. A peptide with antibacterial activity was linked to a peptide specifically binding Staphylococcus aureus to create a synbody with increased antibacterial activity. Subsequent tests showed that this peptide could block S. aureus induced killing of HEK293 cells in a co-culture experiment. These results demonstrate the feasibility of using the synbody system to discover new antibacterial candidate agents.     

The immunoregulatory effects of edeine analogues in mice

The edeines analogs were tested in several in vitro and in vivo assays using the mouse model, with edeine B (peptide W1) and cyclosporine A as reference compounds. The peptides displayed moderate, stimulatory effects on concanavalin A-induced (ConA-induced) splenocyte proliferation, whereas their effects on pokeweed mitogen-induced (PWM-induced) splenocyte proliferation were inhibitory. The peptides inhibited lipopolysacharide-induced (LPS-induced) tumor necrosis factor alpha production but had little effect on interleukin 6 production. In the model of the humoral immune response in vitro to sheep red blood cells, peptide 1 was distinctly stimulatory in the investigated concentrations (1-100 μg/ml), whereas peptides 3 and 4 only stimulated the number of antibody-forming cells at the highest concentration (100 μg/ml). In the model of the delayed type hypersensitivity in vivo to ovalbumin, the peptides were moderately suppressive (3 being the most active). The reference peptide W1 stimulated ConA-induced cell proliferation at 1–10 μg/ml but was inhibitory at 100 μg/ml. It also inhibited PWM-induced cell proliferation in a dose-dependent manner. This peptide had no effect on the humoral immune response in vitro or on cytokine production, but inhibited DTH reaction in vivo. The relationship between structure and activity, and a possible mode of action of the peptides, is discussed in this paper.