Strategies for Selection from Protein Libraries Composed of de Novo Designed Secondary Structure Modules

As more and more protein structures are determined, it has become clear that there is only a limited number of protein folds in nature. To explore whether the protein folds found in nature are the only solutions to the protein folding problem, or that a lack of evolutionary pressure causes the paucity of different protein folds found, we set out to construct protein libraries without any restriction on topology. We generated different libraries (all alpha-helix, all beta-strand and alpha-helix plus beta-strand) with an average length of 100 amino acid residues, composed of designed secondary structure modules (alpha-helix, beta-strand and beta-turn) in various proportions, based primarily on the patterning of polar and non-polar residues. From the analysis of proteins chosen randomly from the libraries, we found that a substantial portion of pure alpha-helical proteins show properties similar to native proteins. Using these libraries as a starting point, we aim to establish a selection system which allows us to enrich proteins with favorable folding properties (non-aggregating, compactly folded) from the libraries. We have developed such a method based on ribosome display. This selection is based on two concepts: (1) misfolded proteins are more sensitive to proteolysis, (2) misfolded and/or aggregated proteins are more hydrophobic. We show that by applying each of these selection criteria proteins that are compactly folded and soluble can be enriched over insoluble and random coil proteins.     

High-affinity binders selected from designed ankyrin repeat protein libraries

We report here the evolution of ankyrin repeat (AR) proteins in vitro for specific, high-affinity target binding. Using a consensus design strategy, we generated combinatorial libraries of AR proteins of varying repeat numbers with diversified binding surfaces. Libraries of two and three repeats, flanked by 'capping repeats,' were used in ribosome-display selections against maltose binding protein (MBP) and two eukaryotic kinases. We rapidly enriched target-specific binders with affinities in the low nanomolar range and determined the crystal structure of one of the selected AR proteins in complex with MBP at 2.3 A resolution. The interaction relies on the randomized positions of the designed AR protein and is comparable to natural, heterodimeric protein-protein interactions. Thus, our AR protein libraries are valuable sources for binding molecules and, because of the very favorable biophysical properties of the designed AR proteins, an attractive alternative to antibody libraries.     

Design of compound libraries based on natural product scaffolds and protein structure similarity clustering (PSSC)

Recent advances in structural biology, bioinformatics and combinatorial chemistry have significantly impacted the discovery of small molecules that modulate protein functions. Natural products which have evolved to bind to proteins may serve as biologically validated starting points for the design of focused libraries that might provide protein ligands with enhanced quality and probability. The combined application of natural product derived scaffolds with a new approach that clusters proteins according to structural similarity of their ligand sensing cores provides a new principle for the design and synthesis of such libraries. This article discusses recent advances in the synthesis of natural product inspired compound collections and the application of protein structure similarity clustering for the development of such libraries.     

Selection of proteins with desired properties from natural proteome libraries using mRNA display

mRNA display is a powerful yet challenging in vitro selection technique that can be used to identify proteins with desired properties from both natural proteome and combinatorial polypeptide libraries. The physical conjugation between a protein and its own RNA presents unique challenges in manipulating the displayed proteins at a low nanomolar scale in an RNase-free environment. The following protocol outlines the generation of cDNA libraries derived from natural organisms as well as the steps required for generation of mRNA-protein fusion molecules, in vitro functional selection and regeneration of the selected cDNA library. The selection procedures for the identification of protease substrates and Ca(2+)-dependent calmodulin-binding proteins from natural cDNA libraries are presented as examples. The method can be generally applied to the identification of protein sequences with desired properties from various natural proteome libraries. One round of mRNA display-based selection can be accomplished in ~7 d.     

Phosphorylation state-dependent interactions of hepadnavirus core protein with host factors

Dynamic phosphorylation and dephosphorylation of the hepadnavirus core protein C-terminal domain (CTD) are required for multiple steps of the viral life cycle. It remains unknown how the CTD phosphorylation state may modulate core protein functions but phosphorylation state-dependent viral or host interactions may play a role. In an attempt to identify host factors that may interact differentially with the core protein depending on its CTD phosphorylation state, pulldown assays were performed using the CTD of the duck hepatitis B virus (DHBV) and human hepatitis B virus (HBV) core protein, either with wild type (WT) sequences or with alanine or aspartic acid substitutions at the phosphorylation sites. Two host proteins, B23 and I2PP2A, were found to interact preferentially with the alanine-substituted CTD. Furthermore, the WT CTD became competent to interact with the host proteins upon dephosphorylation. Intriguingly, the binding site on the DHBV CTD for both B23 and I2PP2A was mapped to a region upstream of the phosphorylation sites even though B23 or I2PP2A binding to this site was clearly modulated by the phosphorylation state of the downstream and non-overlapping sequences. Together, these results demonstrate a novel mode of phosphorylation-regulated protein-protein interaction and provide new insights into virus-host interactions.     

Generation and evaluation of mammalian secreted and membrane protein expression libraries for high-throughput target discovery

Expressed protein libraries are becoming a critical tool for new target discovery in the pharmaceutical industry. In order to get the most meaningful and comprehensive results from protein library screens, it is essential to have library proteins in their native conformation with proper post-translation modifications. This goal is achieved by expressing untagged human proteins in a human cell background. We optimized the transfection and cell culture conditions to maximize protein expression in a 96-well format so that the expression levels were comparable with the levels observed in shake flasks. For detection purposes, we engineered a 'tag after stop codon' system. Depending on the expression conditions, it was possible to express either native or tagged proteins from the same expression vector set. We created a human secretion protein library of 1432 candidates and a small plasma membrane protein set of about 500 candidates. Utilizing the optimized expression conditions, we expressed and analyzed both libraries by SDS-PAGE gel electrophoresis and Western blotting. Two thirds of secreted proteins could be detected by Western-blot analyses; almost half of them were visible on Coomassie stained gels. In this paper, we describe protein expression libraries that can be easily produced in mammalian expression systems in a 96-well format, with one protein expressed per well. The libraries and methods described allow for the development of robust, high-throughput functional screens designed to assay for protein specific functions associated with a relevant disease-specific activity.     

Peptide arrays with designed secondary structures for protein characterization using fluorescent fingerprint patterns

To realize a practical high-throughput protein-detection system, novel peptide arrays have been constructed using designed peptide libraries with loop, alpha-helix, or beta-strand structures. Here, we describe the overview of the reported designed peptide arrays with loop and alpha-helix structures and the new results of those with beta-strand structures. Initially, several model peptides known to interact with model structured proteins were selected to establish the present strategy for high-throughput detection of proteins. The fluorescent probes and suitable scaffolds of peptides were examined for the effective detection of proteins. The detection methods were established in solution and in an immobilized manner using the model systems. In the case of alpha-helix peptide, the response of a peptide with fluorescent resonance energy transfer between two probes at both termini was several times higher than that of a peptide with a single probe. In the cases of peptides with other structures, however, proteins were effectively detectable even by the fluorescent change of one probe. Furthermore, structurally focused libraries consisting of a total of ca. 250 different peptides based on the model peptides with secondary and/or tertiary structures were constructed with systematic replacement of residues. Using these libraries, various proteins were characterized effectively to give their own fluorescent "protein fingerprint" patterns. The resulting protein fingerprints correlated with the recognition properties of the proteins. These studies demonstrate that arrays with peptide libraries based on designed structures can be promising tools for detecting the target proteins. Designed synthetic peptides play roles as the capturing agents to be developed for practical protein chips.     

Computational protein design: structure, function and combinatorial diversity

Computational protein design has blossomed with the development of methods for addressing the complexities involved in specifying the structure, sequence and function of proteins. Recent applications include the design of novel functional membrane and soluble proteins, proteins incorporating non-biological components and protein combinatorial libraries.     

Modeling Sequence-Space Exploration and Emergence of Epistatic Signals in Protein Evolution

During their evolution, proteins explore sequence space via an interplay between random mutations and phenotypic selection. Here, we build upon recent progress in reconstructing data-driven fitness landscapes for families of homologous proteins, to propose stochastic models of experimental protein evolution. These models predict quantitatively important features of experimentally evolved sequence libraries, like fitness distributions and position-specific mutational spectra. They also allow us to efficiently simulate sequence libraries for a vast array of combinations of experimental parameters like sequence divergence, selection strength, and library size. We showcase the potential of the approach in reanalyzing two recent experiments to determine protein structure from signals of epistasis emerging in experimental sequence libraries. To be detectable, these signals require sufficiently large and sufficiently diverged libraries. Our modeling framework offers a quantitative explanation for different outcomes of recently published experiments. Furthermore, we can forecast the outcome of time- and resource-intensive evolution experiments, opening thereby a way to computationally optimize experimental protocols.     

Incremental truncation as a strategy in the engineering of novel biocatalysts

The application and success of combinatorial approaches to protein engineering problems have increased dramatically. However, current directed evolution strategies lack a combinatorial methodology for creating libraries of hybrid enzymes which lack high homology or for creating libraries of highly homologous genes with fusions at regions of non-identity. To create such hybrid enzyme libraries, we have developed a series of combinatorial approaches that utilize the incremental truncation of genes, gene fragments or gene libraries. For incremental truncation, Exonuclease III is used to create a library of all possible single base-pair deletions of a given piece of DNA. Incremental truncation libraries (ITLs) have applications in protein engineering as well as protein folding, enzyme evolution, and the chemical synthesis of proteins. In addition, we are developing a methodology of DNA shuffling which is independent of DNA sequence homology.     

Rat sulfite oxidase antibodies cross-react with two gene family-related proteins: albumin and vitamin D-binding protein.

Screening lambda cDNA libraries from rat liver with antibody to native rat liver sulfite oxidase (RLSO) showed cross-reaction with two proteins that belong to the same gene family: serum albumin and vitamin D-binding protein. Antibodies raised against native RLSO or sodium dodecyl sulfate-denatured protein cross-reacted with these proteins by Western blot analysis. The relative effectiveness of RLSO antibody binding was estimated to be 1/5 for rat serum albumin and 1/10 for rat vitamin D-binding protein. This result was not caused by contaminating proteins in the RLSO used for immunization as the RLSO preparation did not react with rat serum albumin antibody. RLSO antibodies, selected for their ability to bind rat serum albumin immobilized on nitrocellulose, recognized both rat serum albumin and RLSO. RLSO antibody, with albumin-reactive antibody removed, still recognized vitamin D-binding protein, suggesting that multiple determinants specific to each protein are involved in the cross-reaction. Comparison of RLSO antibody binding to the rat and human proteins indicated that the determinants were species-specific. cDNA clones identified by screening cDNA libraries with RLSO antibody demonstrated that these determinants reside in the C-terminal domain of these proteins. These results suggest that these proteins contain some common immunological features and may be evolutionarily related.     

A yeast one-hybrid system to screen for methylated DNA-binding proteins

We had previously exploited a method for targeted DNA methylation in budding yeast to succeed in one-hybrid detection of methylation-dependent DNA–protein interactions. Based on this finding, we developed a yeast one-hybrid system to screen cDNA libraries for clones encoding methylated DNA-binding proteins. Concurrent use of two independent bait sequences in the same cell, or dual-bait system, effectively reduced false positive clones, which were derived from methylation-insensitive sequence-specific DNA-binding proteins. We applied the dual-bait system to screen cDNA libraries and demonstrated efficient isolation of clones for methylated DNA-binding proteins. This system would serve as a unique research tool for epigenetics.     

Fluorescent microplate-based analysis of protein-DNA interactions. II: Immobilized DNA

A simple protein-DNA interaction analysis has been developed using both a high-affinity/high-specificity zinc finger protein and a low-specificity zinc finger protein with nonspecific DNA binding capability. The latter protein is designed to mimic background binding by proteins generated in randomized or shuffled gene libraries. In essence, DNA is immobilized onto the surface of microplate wells via streptavidin capture, and green fluorescent protein (GFP)-labeled protein is added in solution as part of a crude cell lysate or protein mixture. After incubation and washing, bound protein is detected in a standard microplate reader. The minimum sensitivity of the assay is approximately 0.4 nM protein. The assay format is ideally suited to investigate the interactions of DNA binding proteins from within crude cell extracts and/or mixtures of proteins that may be encountered in protein libraries generated by codon randomization or gene shuffling.     

Discovery of improved EGF agonists using a novel in vitro screening platform

Directed evolution is a powerful strategy for protein engineering; however, evolution of pharmaceutical proteins has been limited by the reliance of current screens on binding interactions. Here, we present a method that identifies protein mutants with improved overall cellular efficacy, an objective not feasible with previous approaches. Mutated protein libraries were produced in soluble, active form by means of cell-free protein synthesis. The efficacy of each individual protein was determined at a uniform dosage with a high-throughput protein product assay followed by a cell-based functional assay without requiring protein purification. We validated our platform by first screening mock libraries of epidermal growth factor (EGF) for stimulation of cell proliferation. We then demonstrated its effectiveness by identifying EGF mutants with significantly enhanced mitogenic activity at low concentrations compared to that of wild-type EGF. This is the first report of EGF mutants with improved biological efficacy despite much previous effort. Our platform can be extended to engineer a broad range of proteins, offering a general method to evolve proteins for improved biological efficacy.     

Protein structure prediction based on fragment assembly and parameter optimization

We propose a novel method for ab-initio prediction of protein tertiary structures based on the fragment assembly and global optimization. Fifteen residue long fragment libraries are constructed using the secondary structure prediction method PREDICT, and fragments in these libraries are assembled to generate full-length chains of a query protein. Tertiary structures of 50 to 100 conformations are obtained by minimizing an energy function for proteins, using the conformational space annealing method that enables one to sample diverse low-lying local minima of the energy. Then in order to enhance the performance of the prediction method, we optimize the linear parameters of the energy function, so that the native-like conformations become energetically more favorable than the non-native ones for proteins with known structures. We test the feasibility of the parameter optimization procedure by applying it to the training set consisting of three proteins: the 10-55 residue fragment of staphylococcal protein A (PDB ID 1bdd), a designed protein betanova, and 1fsd.     

Development of in vitro transposon assisted signal sequence trapping and its use in screening Bacillus halodurans C125 and Sulfolobus solfataricus P2 gene libraries

To identify genes encoding extracytosolic proteins, a minitransposon, TnSig, containing a signal-less beta-lactamase ('bla) as reporter gene, was constructed and used for in vitro transposition of genomic libraries made in Escherichia coli. The 'bla gene was cloned into a bacteriophage Mu minitransposon enabling translational fusions between 'bla and target genes. Fusion of TnSig in the correct reading frame to a protein carrying transmembrane domains or signal peptides resulted in ampicillin resistance of the corresponding clone. Prokaryotic gene libraries from the alkaliphilic bacterium Bacillus halodurans C125 and the hyperthermophilic archaeon Sulfolobus solfataricus P2 were tagged with TnSig. The genomic sequences, which are publicly available (EMBL and EMBL ), were used for rapid open reading frame (ORF) identification and prediction of protein localisation in the cell. Genes for secreted proteins, transmembrane proteins and lipoproteins were successfully identified by this method. In contrast to previous transposon based identification strategies, the method described here is fast and versatile and essentially enables any selectable marker compatible library to be tagged. It is suited for identifying genes encoding extracytosolic proteins in gene libraries of a wide range of prokaryotic organisms.     

Identification of peptides mimicking the ligands of proteins phosphorylated by protein kinase CK2

Synthetic peptides containing a phosphorylation site for protein kinase CK2 were used to investigate their binding properties to other peptides/proteins. The aim of this work was to find an efficient procedure to search for these peptide/protein ligands. The goal was successfully achieved through screening of random peptide libraries displayed on phage. Peptides corresponding to the amino terminal region of topoisomerase I were synthesized in both phosphorylated and unphosphorylated form and used to screen the libraries. Four of the selected sequences were also tested for their reactivity with synthetic peptides corresponding to the carboxy terminal region of the largest subunit of RNA polymerase II. The positive reaction detected supports the hypothesis that the isolated sequences may represent mimics of ligands of proteins phosphorylated by protein kinase CK2.     

Hidden Markov models-based system (HMMSPECTR) for detecting structural homologies on the basis of sequential information

HMMSPECTR is a tool for finding putative structural homologs for proteins with known primary sequences. HMMSPECTR contains four major components: a data warehouse with the hidden Markov models (HMM) and alignment libraries; a search program which compares the initial protein sequences with the libraries of HMMs; a secondary structure prediction and comparison program; and a dominant protein selection program that prepares the set of 10-15 "best" proteins from the chosen HMMs. The data warehouse contains four libraries of HMMs. The first two libraries were constructed using different HHM preparation options of the HAMMER program. The third library contains parts ("partial HMM") of initial alignments. The fourth library contains trained HMMs. We tested our program against all of the protein targets proposed in the CASP4 competition. The data warehouse included libraries of structural alignments and HMMs constructed on the basis of proteins publicly available in the Protein Data Bank before the CASP4 meeting. The newest fully automated versions of HMMSPECTR 1.02 and 1.02ss produced better results than the best result reported at CASP4 either by r.m.s.d. or by length (or both) in 64% (HMMSPECTR 1.02) and 79% (HMMSPECTR 1.02ss) of the cases. The improvement is most notable for the targets with complexity 4 (difficult fold recognition cases).     

Identification of the key residues responsible for the assembly of selenodeiodinases

Type I deiodinase is the best characterized member of a small family of selenoenzymes catalyzing the bioactivation and disposal of thyroid hormone. This enzyme is an integral membrane protein composed of two 27-kDa subunits that assemble into a functional enzyme after translation using a highly conserved sequence of 16 amino acids in the C-terminal half of the polypeptide, (148)DFLXXYIXEAHXXDGW(163). In this study, we used alanine scanning mutagenesis to identify the key residues in this domain required for holoenzyme assembly. Overexpression of sequential alanine-substituted mutants of a dimerization domain-green fluorescent protein fusion showed that sequence (152)IYI(154) was required for type I enzyme assembly and that a catalytically active monomer was generated by a single I152A substitution. Overexpression of the sequential alanine-substituted dimerization domain mutants in type II selenodeiodinase-expressing cells showed that five residues ((153)FLIVY(157)) at the beginning and three residues ((164)SDG(166)) at the end of this region were required for the assembly of the type II enzyme. In vitro binding analysis revealed a free energy of association of -60 +/- 5 kJ/mol for the noncovalent interaction between dimerization domain monomers. These data identify and characterize the essential residues in the dimerization domain that are responsible for the post-translational assembly of selenodeiodinases.