Extensive antibody search with whole spectrum black-box optimization.

In designing functional biological sequences with machine learning, the activity predictor tends to be inaccurate due to shortage of data. Top ranked sequences are thus unlikely to contain effective ones. This paper proposes to take prediction stability into account to provide domain experts with a reasonable list of sequences to choose from. In our approach, multiple prediction models are trained by subsampling the training set and the multi-objective optimization problem, where one objective is the average activity and the other is the standard deviation, is solved. The Pareto front represents a list of sequences with the whole spectrum of activity and stability. Using this method, we designed VHH (Variable domain of Heavy chain of Heavy chain) antibodies based on the dataset obtained from deep mutational screening. To solve multi-objective optimization, we employed our sequence design software MOQA that uses quantum annealing. By applying several selection criteria to 19,778 designed sequences, five sequences were selected for wet-lab validation. One sequence, 16 mutations away from the closest training sequence, was successfully expressed and found to possess desired binding specificity. Our whole spectrum approach provides a balanced way of dealing with the prediction uncertainty, and can possibly be applied to extensive search of functional sequences.

Synthesis of epitope-targeting nanobody based on native protein-protein interactions for FtsZ filamentation suppressor.

Phage display and biopanning are powerful tools for generating binding molecules for a specific target. However, the selection process based only on binding affinity provides no assurance for the antibody's affinity to the target epitope. In this study, we propose a molecular-evolution approach guided by native protein-protein interactions to generate epitope-targeting antibodies. The binding-site sequence in a native protein was grafted into a complementarity-determining region (CDR) in the nanobody, and a nonrelated CDR loop (in the grafted nanobody) was randomized to create a phage display library. In this construction of nanobodies by integrating graft and evolution technology (CAnIGET method), suitable grafting of the functional sequence added functionality to the nanobody, and the molecular-evolution approach enhanced the binding function to inhibit the native protein-protein interactions. To apply for biological tool with growth screening, model nanobodies with an affinity for filamenting temperature-sensitive mutant Z (FtsZ) from Staphylococcus aureus were constructed and completely inhibited the polymerization of FtsZ as a function. Consequently, the expression of these nanobodies drastically decreased the cell division rate. We demonstrate the potential of the CAnIGET method with the use of native protein-protein interactions for steady epitope-specific evolutionary engineering.

Recent research advances on non-linear phenomena in various biosystems.

All biological phenomena can be classified as open, dissipative and non-linear. Moreover, the most typical phenomena are associated with non-linearity, dissipation and openness in biological systems. In this review article, four research topics on non-linear biosystems are described to show the examples from various biological systems. First, membrane dynamics of a lipid bilayer for the cell membrane is described. Since the cell membrane separates the inside of the cell from the outside, self-organizing systems that form spatial patterns on membranes often depend on non-linear dynamics. Second, various data banks based on recent genomics analysis supply the data including vast functional proteins from many organisms and their variable species. Since the proteins existing in nature are only a very small part of the space represented by amino acid sequence, success of mutagenesis-based molecular evolution approach crucially depends on preparing a library with high enrichment of functional proteins. Third, photosynthetic organisms depend on ambient light, the regular and irregular changes of which have a significant impact on photosynthetic processes. The light-driven process proceeds through many redox couples in the cyanobacteria constituting chain of redox reactions. The fourth topic focuses on a vertebrate model, the zebrafish, which can help to understand, predict and control the chaos of complex biological systems. In particular, during early developmental stages, developmental differentiation occurs dynamically from a fertilized egg to divided and mature cells. These exciting fields of complexity, chaos, and non-linear science have experienced impressive growth in recent decades. Finally, future directions for non-linear biosystems are presented.

Selection of target-binding proteins from the information of weakly enriched phage display libraries by deep sequencing and machine learning.

Despite the advances in surface-display systems for directed evolution, variants with high affinity are not always enriched due to undesirable biases that increase target-unrelated variants during biopanning. Here, our goal was to design a library containing improved variants from the information of the "weakly enriched" library where functional variants were weakly enriched. Deep sequencing for the previous biopanning result, where no functional antibody mimetics were experimentally identified, revealed that weak enrichment was partly due to undesirable biases during phage infection and amplification steps. The clustering analysis of the deep sequencing data from appropriate steps revealed no distinct sequence patterns, but a Bayesian machine learning model trained with the selected deep sequencing data supplied nine clusters with distinct sequence patterns. Phage libraries were designed on the basis of the sequence patterns identified, and four improved variants with target-specific affinity (EC50 = 80-277 nM) were identified by biopanning. The selection and use of deep sequencing data without undesirable bias enabled us to extract the information on prospective variants. In summary, the use of appropriate deep sequencing data and machine learning with the sequence data has the possibility of finding sequence space where functional variants are enriched.

Comparison of the stability of Mucor-derived flavin adenine dinucleotide-dependent glucose dehydrogenase and glucose oxidase.

Long-term stability at near-body temperature is important for continuous glucose monitoring (CGM) sensors. However, the stability of enzymes used in CGM sensors has often been evaluated by measuring their melting temperature (Tm) values and by short heat treatment but not at around 37 °C. Glucose oxidase (GOD) is used in current CGM sensors. In this study, we evaluated the stability of modified Mucor-derived flavin adenine dinucleotide-dependent glucose dehydrogenase (designated Mr144-297) with improved thermal stability at medium to high temperatures and compared it with that of GOD. The Tm value of Mr144-297 was 61.6 ± 0.3 °C and was similar to that of GOD (61.4 ± 1.2 °C). However, Mr144-297 was clearly more stable than GOD at 40 °C and 55 °C. At 37 °C, the stability of a carbon electrode with immobilized Mr144-297 was higher than that of an electrode with GOD. Our data indicate that Mr144-297 is a more suitable enzyme for CGM sensors than is GOD.

A streamlined strain engineering workflow with genome-wide screening detects enhanced protein secretion in Komagataella phaffii.

Expression of secreted recombinant proteins burdens the protein secretion machinery, limiting production. Here, we describe an approach to improving protein production by the non-conventional yeast Komagataella phaffii comprised of genome-wide screening for effective gene disruptions, combining them in a single strain, and recovering growth reduction by adaptive evolution. For the screen, we designed a multiwell-formatted, streamlined workflow to high-throughput assay of secretion of a single-chain small antibody, which is cumbersome to detect but serves as a good model of proteins that are difficult to secrete. Using the consolidated screening system, we evaluated >19,000 mutant strains from a mutant library prepared by a modified random gene-disruption method, and identified six factors for which disruption led to increased antibody production. We then combined the disruptions, up to quadruple gene knockouts, which appeared to contribute independently, in a single strain and observed an additive effect. Target protein and promoter were basically interchangeable for the effects of knockout genes screened. We finally used adaptive evolution to recover reduced cell growth by multiple gene knockouts and examine the possibility for further enhancing protein secretion. Our successful, three-part approach holds promise as a method for improving protein production by non-conventional microorganisms.

G6P-capturing molecules in the periplasm of Escherichia coli accelerate the shikimate pathway.

Escherichia coli, the most studied prokaryote, is an excellent host for producing valuable chemicals from renewable resources as it is easy to manipulate genetically. Since the periplasmic environment can be easily controlled externally, elucidating how the localization of specific proteins or small molecules in the periplasm affects metabolism may lead to bioproduction development using E. coli. We investigated metabolic changes and its mechanisms occurring when specific proteins are localized to the E. coli periplasm. We found that the periplasmic localization of ß-glucosidase promoted the shikimate pathway involved in the synthesis of aromatic chemicals. The periplasmic localization of other proteins with an affinity for glucose-6-phosphate (G6P), such as inactivated mutants of Pgi, Zwf, and PhoA, similarly accelerated the shikimate pathway. Our results indicate that G6P is transported from the cytoplasm to the periplasm by the glucose transporter protein EIICBGlc, and then captured by ß-glucosidase.

Enzymatic ligation of an antibody and arginine 9 peptide for efficient and cell-specific siRNA delivery.

A fusion protein comprising an antibody and a cationic peptide, such as arginine-9 (R9), is a candidate molecule for efficient and cell-specific delivery of siRNA into cells in order to reduce the side effects of nucleic acid drugs. However, their expression in bacterial hosts, required for their development, often fails, impeding research progress. In this study, we separately prepared anti-EGFR nanobodies with the K-tag sequence MRHKGS at the C-terminus and R9 with the Q-tag sequence LLQG at the N-terminus, and enzymatically ligated them in vitro by microbial transglutaminase to generate Nanobody-R9, which is not expressed as a fused protein in E. coli. Nanobody-R9 was synthesized at a maximum binding efficiency of 85.1%, without changing the binding affinity of the nanobody for the antigen. Nanobody-R9 successfully delivered siRNA into the cells, and the cellular influx of siRNA increased with increase in the ratio of Nanobody-R9 to siRNA. We further demonstrated that the Nanobody-R9-siRNA complex, at a 30:1 ratio, induced an approximately 58.6% reduction in the amount of target protein due to RNAi in mRNA compared to lipofectamine.

Chemically Crosslinked Bispecific Antibodies for Cancer Therapy: Breaking from the Structural Restrictions of the Genetic Fusion Approach.

Antibodies are composed of structurally and functionally independent domains that can be used as building blocks to construct different types of chimeric protein-format molecules. However, the generally used genetic fusion and chemical approaches restrict the types of structures that can be formed and do not give an ideal degree of homogeneity. In this study, we combined mutation techniques with chemical conjugation to construct a variety of homogeneous bivalent and bispecific antibodies. First, building modules without lysine residues-which can be chemical conjugation sites-were generated by means of genetic mutation. Specific mutated residues in the lysine-free modules were then re-mutated to lysine residues. Chemical conjugation at the recovered lysine sites enabled the construction of homogeneous bivalent and bispecific antibodies from block modules that could not have been so arranged by genetic fusion approaches. Molecular evolution and bioinformatics techniques assisted in finding viable alternatives to the lysine residues that did not deactivate the block modules. Multiple candidates for re-mutation positions offer a wide variety of possible steric arrangements of block modules, and appropriate linkages between block modules can generate highly bioactive bispecific antibodies. Here, we propose the effectiveness of the lysine-free block module design for site-specific chemical conjugation to form a variety of types of homogeneous chimeric protein-format molecule with a finely tuned structure and function.

Identification of Indium Tin Oxide Nanoparticle-Binding Peptides via Phage Display and Biopanning Under Various Buffer Conditions.

BACKGROUND: By recent advances in phage-display approaches, many oligopeptides exhibiting binding affinities for metal oxides have been identified. Indium tin oxide is one of the most widely used conductive oxides, because it has a large band gap of 3.7-4.0 eV. In recent years, there have been reports about several ITO-based biosensors. Development of an ITO binding interface for the clustering of sensor proteins without complex bioconjugates is required. OBJECTIVE: In this article, we aimed to identify peptides that bind to indium tin oxide nanoparticles via different binding mechanisms. METHODS: Indium tin oxide nanoparticles binding peptide ware selected using phage display and biopanning against indium tin oxide, under five different buffer conditions and these peptides characterized about binding affinity and specificity. RESULTS: Three types of indium tin oxide nanoparticles-binding peptides were selected from 10 types of peptide candidates identified in phage display and biopanning. These included ITOBP8, which had an acidic isoelectric point, and was identified when a buffer containing guanidine was used, and ITOBP6 and ITOBP7, which contained a His-His-Lys sequence at their N-termini, and were identified when a highly concentrated phosphate elution buffer with a low ionic strength was used. Among these peptides, ITOBP6 exhibited the strongest indium tin oxide nanoparticlesbinding affinity (dissociation constant, 585 nmol/L; amount of protein bound at saturation, 17.5 nmol/m 2 - particles). CONCLUSION: These results indicate that peptides with specific binding properties can be obtained through careful selection of the buffer conditions in which the biopanning procedure is performed.

Complementary Design for Pairing between Two Types of Nanoparticles Mediated by a Bispecific Antibody: Bottom-Up Formation of Porous Materials from Nanoparticles.

Recent advances in biotechnology have enabled the generation of antibodies with high affinity for the surfaces of specific inorganic materials. Herein, we report the synthesis of functional materials from multiple nanomaterials by using a small bispecific antibody recombinantly constructed from gold-binding and ZnO-binding antibody fragments. The bispecific antibody-mediated spontaneous linkage of gold and ZnO nanoparticles forms a binary gold-ZnO nanoparticle composite membrane. The relatively low melting point of the gold nanoparticles and the solubility of ZnO in dilute acidic solution then allowed for the bottom-up synthesis of a nanoporous gold membrane by means of a low-energy, low-environmental-load protocol. The nanoporous gold membrane showed high catalytic activity for the reduction of p-nitrophenol to p-aminophenol by sodium borohydride. Here, we show the potential utility of nanoparticle pairing mediated by bispecific antibodies for the bottom-up construction of nanostructured materials from multiple nanomaterials.

Nanoparticle Assisted Remodeling of Proteotoxic SOD1 Mutants Alters the Biointerface of the Functional Interaction of Microtubules and Kinesin Motors.

Transport deficits with motor neuron degeneration have been implicated in amyotrophic lateral sclerosis (ALS). We report a biomimetic system composed of microtubules/kinesin motor that mimics the dysregulated motor dynamics of ALS. Pathogenic ALS mutants A4V SOD1 completely shut off motility. Treatment with 5 nm citrate coated gold nanoparticles recovers the impaired motor stepping by remodeling the A4V SOD1 rather than stabilizing microtubules or protein folding. Instead, gold nanoparticles alter the protein by a mechanism that reforms protein elements of A4V SOD1, in turn fixing the aberrant interaction of kinesin with microtubules. Reinstating kinesin motility holds potential for managing debilitating ALS.

Machine-Learning-Guided Mutagenesis for Directed Evolution of Fluorescent Proteins.

Molecular evolution based on mutagenesis is widely used in protein engineering. However, optimal proteins are often difficult to obtain due to a large sequence space. Here, we propose a novel approach that combines molecular evolution with machine learning. In this approach, we conduct two rounds of mutagenesis where an initial library of protein variants is used to train a machine-learning model to guide mutagenesis for the second-round library. This enables us to prepare a small library suited for screening experiments with high enrichment of functional proteins. We demonstrated a proof-of-concept of our approach by altering the reference green fluorescent protein (GFP) so that its fluorescence is changed into yellow. We successfully obtained a number of proteins showing yellow fluorescence, 12 of which had longer wavelengths than the reference yellow fluorescent protein (YFP). These results show the potential of our approach as a powerful method for directed evolution of fluorescent proteins.

Large-scale chirality in an active layer of microtubules and kinesin motor proteins.

During the early developmental process of organisms, the formation of left-right laterality requires a subtle mechanism, as it is associated with other principal body axes. Any inherent chiral feature in an egg cell can in principal trigger this spontaneous breaking of chiral symmetry. Individual microtubules, major cytoskeletal filaments, are known as chiral objects. However, to date there lacks convincing evidence of a hierarchical connection of the molecular nature of microtubules to large-scale chirality, particularly at the length scale of an entire cell. Here we assemble an in vitro active layer, consisting of microtubules and kinesin motor proteins, on a glass surface. Upon inclusion of methyl cellulose, the layered system exhibits a long-range active nematic phase, characterized by the global alignment of gliding MTs. This nematic order spans over the entire system size in the millimeter range and, remarkably, allows hidden collective chirality to emerge as counterclockwise global rotation of the active MT layer. The analysis based on our theoretical model suggests that the emerging global nematic order results from the local alignment of MTs, stabilized by methyl cellulose. It also suggests that the global rotation arises from the MTs' intrinsic curvature, leading to preferential handedness. Given its flexibility, this layered in vitro cytoskeletal system enables the study of membranous protein behavior responsible for important cellular developmental processes.

High-throughput cytotoxicity and antigen-binding assay for screening small bispecific antibodies without purification.

The cytotoxicity of T cell-recruiting antibodies with their potential to damage late-stage tumor masses is critically dependent on their structural and functional properties. Recently, we reported a semi-high-throughput process for screening highly cytotoxic small bispecific antibodies (i.e., diabodies). In the present study, we improved the high-throughput performance of this screening process by removing the protein purification stage and adding a stage for determining the concentrations of the diabodies in culture supernatant. The diabodies were constructed by using an Escherichia coli expression system, and each diabody contained tandemly arranged peptide tags at the C-terminus, which allowed the concentration of diabodies in the culture supernatant to be quantified by using a tag-sandwich enzyme-linked immunosorbent assay. When estimated diabody concentrations were used to determine the cytotoxicity of unpurified antibodies, results comparable to those of purified antibodies were obtained. In a surface plasmon resonance spectroscopy-based target-binding assay, contaminants in the culture supernatant prevented us from conducting a quantitative binding analysis; however, this approach did allow relative binding affinity to be determined, and the relative binding affinities of the unpurified diabodies were comparable to those of the purified antibodies. Thus, we present here an improved high-throughput process for the simultaneous screening and determination of the binding parameters of highly cytotoxic bispecific antibodies.

Impact in stability during sequential CDR grafting to construct camelid VHH antibodies against zinc oxide and gold.

Biomolecules which recognize inorganic materials and metal surfaces gain much attention for creating new type of nanomaterials and sensors. 4F2, a camelid VHH antibody, recognizes ZnO surface and has been applied for sensor applications. 4F2 was constructed sequential complementarity determining region (CDR) replacement on the parental VHH antibody, termed the Construction of Antibody by Integrating Grafting and Evolution Technology; CAnIGET procedure. Here, we evaluate the influence of CDR replacements during 4F2 generation using calorimetric technique. We found that the initial peptide grafting at CDR1 results in the stability reduction and subsequent CDR3 randomize and selection restore the stability during the construction of 4F2. Further examination using anti-gold VHH, AuE32, revealed that the final CDR3 randomize and selection step has little effect in stability while the initial CDR1 grafting reduces the stability as same as the case for 4F2. Our results showing here provide the detailed view of the stability alteration during the CAnIGET procedure.

Use of a Phage-Display Method to Identify Peptides that Bind to a Tin Oxide Nanosheets.

BACKGROUND: Nanosheets of SnO2 which an n-type semiconductor with a rutile-type crystalline structure are predominantly used as gas sensors. SnO2 nanosheets have a tetragonal crystal structure where growth along the c-axis is suppressed to form a sheet. The major exposed facets of SnO2 nanosheets have {110}, {101} and {211} crystal planes along the a-axis, with the reduced {110} surface having a particularly high surface energy. Identifying peptides that bind to specific crystal planes by using peptide phage-display approach will increase the potential applications of metal oxide nanomaterials by fusing proteins with desirable active sites to peptides that adsorb at high density on the major exposed crystal plane of nanosheets. It may be possible to construct highly sensitive biosensors. OBJECTIVES: The main objective of the present study is to identify peptides that adsorb preferentially to a SnO2 nanosheet by using peptide-phage display approach. METHODS: Four milligrams of SnO2 nanosheet were mixed with 1011 plaque-forming units of Ph.D.-12 Phage Display Peptide Library. Phage-bound nanosheet particles were washed 10 times with 1 mL of phosphatebuffered saline containing 0.5% Tween 20. Phages bound to the nanosheet were eluted with three different buffers: (1) high-salt buffer containing 2 M NaCl (pH 7.5); (2) acidic buffer containing 200 mM Gly-HCl (pH 2.2); and (3) high-phosphate-ion buffer containing 500 mM NaH2PO4 (pH 7.5). The eluted phages were subjected to four or five rounds of biopanning. At each round, individual plaques were picked from the plates, and the amino acid sequences of the peptides were identified by DNA sequencing. The identified SnO2-binding peptides labeled with fluorescein isothiocyanate were synthesized. Adsorption isotherms were constructed at peptide concentrations ranging from 0.25 to 2.0 µM with 4mg of nanomaterials. RESULTS: We were determined the sequences of 11 clones with the high-salt buffer, 7 with the high-phosphateion buffers, and 6 with the acidic buffer and three peptides (SnO2BPn1, 2, and 3), two peptides (SnO2BPa1 and SnO2BPa2), and one peptide (SnO2BPp1) concentrated under each condition were selected respectively. All six selected peptides contained at least one histidine residue. In addition, the His-Asn-Leu (HNL) sequence was found in two of the peptides (SnO2BPa1 and SnO2BPa2). We constructed adsorption isotherms for the six selected peptides using 4mg of nanosheets. All six peptides were well adsorbed on the SnO2 nanosheet. The adsorption isotherms for SnO2 material with different structure revealed that SnO2BPn1, -2, and -3, and SnO2BPp1, preferentially bound to the spherical SnO2 nanoparticles. SnO2BPa2 preferentially bound to the SnO2 nanosheet, and SnO2BPa1 bound equally to both materials. This result suggested that SnO2BPa2 bound to a specific crystal plane of the nanosheet. The major exposed facet of the SnO2 crystal was the {110} plane, suggesting that SnO2BPa2 likely adsorbed on the {110} plane. SnO2BPn1, SnO2BPn2, SnO2BPn3, SnO2BPa1, and SnO2BPp1 also bound to the other metal oxides, in particular to ZrO2. At pH 7.5, peptides with a negative charge at pH 7.5 (pI 8.5-12) can bind to ZrO2 and SnO2, if the binding is mediated by electrostatic interactions. Thus, it is likely that these five peptides bind to metal oxides via electrostatic interactions. In contrast, SnO2BPa2 had a structurally specific affinity, binding more with the SnO2 nanosheet than with the spherical SnO2 nanoparticles or other metal oxides. CONCLUSION: We identified six peptides that adsorbed on a SnO2 nanosheet. Five of the selected peptides bound preferentially to spherical SnO2 nanoparticles rather than to the SnO2 nanosheet. Whereas, SnO2BPa2 exhibited specifically binding to the SnO2 nanosheet. Our results suggest that crystal plane recognition and material recognition by these peptides are mediated via different, independent mechanisms.

A semi high-throughput method for screening small bispecific antibodies with high cytotoxicity.

Small bispecific antibodies that induce T-cell-mediated cytotoxicity have the potential to damage late-stage tumor masses to a clinically relevant degree, but their cytotoxicity is critically dependent on their structural and functional properties. Here, we constructed an optimized procedure for identifying highly cytotoxic antibodies from a variety of the T-cell-recruiting antibodies engineered from a series of antibodies against cancer antigens of epidermal growth factor receptor family and T-cell receptors. By developing and applying a set of rapid operations for expression vector construction and protein preparation, we screened the cytotoxicity of 104 small antibodies with diabody format and identified some with 103-times higher cytotoxicity than that of previously reported active diabody. The results demonstrate that cytotoxicity is enhanced by synergistic effects between the target, epitope, binding affinity, and the order of heavy-chain and light-chain variable domains. We demonstrate the importance of screening to determine the critical rules for highly cytotoxic antibodies.

Generation of camelid VHH bispecific constructs via in-cell intein-mediated protein trans-splicing.

Production of various combinations of bispecific variable domain of heavy chain of heavy chain-only antibody (VHH) constructs to evaluate their therapeutic potential usually requires several gene-engineering steps. Here, we present an alternative method of creating bispecific VHH constructs in vivo through protein trans-splicing (PTS) reaction; this method may reduce the number of gene manipulation steps required. As a proof-of-concept, we constructed a bispecific antibody (bsAb) containing an anti-epidermal growth factor receptor VHH and anti-green fluorescent protein VHH, and we evaluated and confirmed its bispecificity. We also tested antibody labeling by fluorescent protein tagging using the PTS reaction. Compared with the conventional gene construction method, bsAb construction via PTS is a promising alternative approach for generating multiple bsAb combinations.

Salt-Switchable Artificial Cellulase Regulated by a DNA Aptamer.

A novel artificial cellulase was developed by conjugating a DNA aptamer to an endoglucanase catalytic domain, thereby substituting the natural carbohydrate-binding module. Circular dichroism spectroscopy and adsorption isotherm showed the binding motif of cellulose-binding DNA aptamer (CelApt) was G-quadruplex and stem-loop structures stabilized in the presence of salts, and CelApt binding preferred the amorphous region of the solid cellulose. By introducing the revealed salt-switchable cellulose-binding nature of CelApt into a catalytic domain of a cellulase, we created CelApt-catalytic domain conjugate possessing both controllable adsorption on the solid substrates and equal enzymatic activity to the wild-type cellulase. Thus potential use of a responsive DNA aptamer for biocatalysis at a solid surface was demonstrated.