Koki Makabe

Design of protein function leaps by directed domain interface evolution

Most natural proteins performing sophisticated tasks contain multiple domains where an active site is located at the domain interface. Comparative structural analyses suggest that major leaps in protein function occur through gene recombination events that connect two or more protein domains to generate a new active site, frequently occurring at the newly created domain interface. However, such functional leaps by combination of unrelated domains have not been directly demonstrated. Here we show that highly specific and complex protein functions can be generated by joining a low-affinity peptide-binding domain with a functionally inert second domain and subsequently optimizing the domain interface. These directed evolution processes dramatically enhanced both affinity and specificity to a level unattainable with a single domain, corresponding to >500-fold and >2,000-fold increases of affinity and specificity, respectively. An x-ray crystal structure revealed that the resulting “affinity clamp” had clamshell architecture as designed, with large additional binding surface contributed by the second domain. The affinity clamps having a single-nanomolar dissociation constant outperformed a monoclonal antibody in immunochemical applications. This work establishes evolutionary paths from isolated domains with primitive function to multidomain proteins with sophisticated function and introduces a new protein-engineering concept that allows for the generation of highly functional affinity reagents to a predefined target. The prevalence and variety of natural interaction domains suggest that numerous new functions can be designed by using directed domain interface evolution.

Antitumor activity of interleukin-21 prepared by novel refolding procedure from inclusion bodies expressed in Escherichia coli

Interleukin‐21 (IL‐21) has recently been identified as a novel 4‐helix‐bundle type I cytokine possessing a cytokine receptor γ chain essential for the immune response. We report the preparation and functional characterization of Escherichia coli‐expressed recombinant human IL‐21 (rIL‐21). The rIL‐21, expressed as insoluble inclusion bodies in E. coli, was solubilized and then refolded by using a modified dialysis method. The introduction of redox reagents during refolding led to a dramatic increase in the refolding efficiency. Circular dichroism spectrum analysis showed that the refolded rIL‐21 had an α‐helix as a secondary structure, which is a characteristic of type I cytokines. Flow cytometry confirmed previous reports that rIL‐21 binds to CD3‐activated T cells (T‐LAK) and to cell lines Raji, HL60, and Jurkat. rIL‐21 stimulated the proliferation of T‐LAK but not peripheral blood mononuclear cells, and this effect seems to be identical to that of co‐stimulation with anti‐CD3 antibody. Growth inhibition assay indicated that enhancement of the cytotoxicity of T‐LAK to the human bile duct carcinoma TFK‐1 depended on the concentration of rIL‐21. Thus, refolded rIL‐21 had activity identical to that of authentic IL‐21 and enhanced the anti‐tumor activity of T‐LAK. These conclusions suggest the potential use of the refolded cytokine in adoptive immunotherapy using T‐LAK cells and in the discovery of other functions of the cytokine.

Thermodynamic Consequences of Mutations in Vernier Zone Residues of a Humanized Anti-human Epidermal Growth Factor Receptor Murine Antibody, 528

To investigate the role of Vernier zone residues, which are comprised in the framework regions and underlie the complementarity-determining regions (CDRs) of antibodies, in the specific, high affinity interactions of antibodies with their targets, we focused on the variable domain fragment of murine anti-human epidermal growth factor receptor antibody 528 (m528Fv). Grafting of the CDRs of m528Fv onto a selected framework region of human antibodies, referred to as humanization, reduced the antibodys affinity for its target by a factor of 1/40. The reduction in affinity was due to a substantial reduction in the negative enthalpy change associated with binding. Crystal structures of the ligand-free antibody fragments showed no noteworthy conformational changes due to humanization, and the loop structures of the CDRs of the humanized antibodies were identical to those of the parent antibodies. Several mutants of the CDR-grafted (humanized) variable domain fragment (h528Fv), in which some of the Vernier zone residues in the heavy chain were replaced with the parental murine residues, were constructed and prepared using a bacterial expression system. Thermodynamic analyses of the interactions between the mutants and the soluble extracellular domain of epidermal growth factor receptor showed that several single mutations and a double mutation increased the negative enthalpy and heat capacity changes. Combination of these mutations, however, led to somewhat reduced negative enthalpy and heat capacity changes. The affinity of each mutant for the target was within the range for the wild-type h528Fv, and this similarity was due to enthalpy-entropy compensation. These results suggest that Vernier zone residues make enthalpic contributions to antigen binding and that the regulation of conformational entropy changes upon humanization of murine antibodies must be carefully considered and optimized.

Humanization of the bispecific epidermal growth factor receptor x CD3 diabody and its efficacy as a potential clinical reagent.

Purpose: Bispecific antibodies (BsAb) have been exploited as both cancer immunodiagnostics and cancer therapeutics and show promise in clinical trials of cancer imaging and therapy. For development of BsAbs as clinical reagents, we have focused on construction of small recombinant BsAbs, called bispecific diabodies. Here, we constructed and characterized a humanized bispecific diabody. Experimental Design: We have reported significant antitumor activity of an anti-epidermal growth factor receptor (EGFR) × anti-CD3 bispecific diabody (Ex3) in in vitro cytotoxicity assays and in vivo. We humanized the Ex3 diabody (hEx3) by grafting the complementarity-determining region and compared its biological properties with those of Ex3. We also tested its physiologic stability and ability to alter survival in xenografted mice. Results: The final yield of hEx3 was 10 times that of Ex3, and refolded hEx3 and Ex3 showed identical binding profiles in EGFR-positive cell lines and EGFR-transfected Chinese hamster ovary cells. hEx3 showed dose-dependent cytotoxicity to EGFR-positive cell lines, which could be specifically inhibited by parental monoclonal antibody IgGs against EGFR or CD3 antigens. The heterodimeric structure was retained in PBS for 6 months, and growth inhibition was maintained after incubation under physiologic conditions. Coadministration of hEx3 with T-LAK cells and interleukin-2 prolonged the survival of nude mice with human colon carcinoma. Conclusions: The humanized diabody hEx3 is an attractive molecule for cancer therapy and may provide important insights into the development of EGFR-based cancer-targeting reagents.

Atomic structures of peptide self-assembly mimics

Although the β-rich self-assemblies are a major structural class for polypeptides and the focus of intense research, little is known about their atomic structures and dynamics due to their insoluble and noncrystalline nature. We developed a protein engineering strategy that captures a self-assembly segment in a water-soluble molecule. A predefined number of self-assembling peptide units are linked, and the β-sheet ends are capped to prevent aggregation, which yields a mono-dispersed soluble protein. We tested this strategy by using Borrelia outer surface protein (OspA) whose single-layer β-sheet located between two globular domains consists of two β-hairpin units and thus can be considered as a prototype of self-assembly. We constructed self-assembly mimics of different sizes and determined their atomic structures using x-ray crystallography and NMR spectroscopy. Highly regular β-sheet geometries were maintained in these structures, and peptide units had a nearly identical conformation, supporting the concept that a peptide in the regular β-geometry is primed for self-assembly. However, we found small but significant differences in the relative orientation between adjacent peptide units in terms of β-sheet twist and bend, suggesting their inherent flexibility. Modeling shows how this conformational diversity, when propagated over a large number of peptide units, can lead to a substantial degree of nanoscale polymorphism of self-assemblies.

Molecular mechanisms of the cytotoxicity of human α-lactalbumin made lethal to tumor cells (HAMLET) and other protein-oleic acid complexes

Background: The α-lactalbumin-oleic acid complex has a unique apoptotic activity for selectively killing tumor cells. Results: We identified the oleic acid-binding site in the human- and goat-α-lactalbumin complexes by two-dimensional NMR. Conclusion: Oleic acid is bound to a loosely organized hydrophobic core of the proteins in the molten globule state. Significance: The results are crucial for understanding the molecular mechanisms of the cytotoxicity of the protein-oleic acid complexes. Although HAMLET (human α-lactalbumin made lethal to tumor cells), a complex formed by human α-lactalbumin and oleic acid, has a unique apoptotic activity for the selective killing of tumor cells, the molecular mechanisms of expression of the HAMLET activity are not well understood. Therefore, we studied the molecular properties of HAMLET and its goat counterpart, GAMLET (goat α-lactalbumin made lethal to tumor cells), by pulse field gradient NMR and 920-MHz two-dimensional NMR techniques. We also examined the expression of HAMLET-like activities of complexes between oleic acid and other proteins that form a stable molten globule state. We observed that both HAMLET and GAMLET at pH 7.5 were heterogeneous, composed of the native protein, the monomeric molten globule-like state, and the oligomeric species. At pH 2.0 and 50 °C, HAMLET and GAMLET appeared in the monomeric state, and we identified the oleic acid-binding site in the complexes by two-dimensional NMR. Rather surprisingly, the binding site thus identified was markedly different between HAMLET and GAMLET. Furthermore, canine milk lysozyme, apo-myoglobin, and β2-microglobulin all formed the HAMLET-like complex with the anti-tumor activity, when the protein was treated with oleic acid under conditions in which their molten globule states were stable. From these results, we conclude that the protein portion of HAMLET, GAMLET, and the other HAMLET-like protein-oleic acid complexes is not the origin of their cytotoxicity to tumor cells and that the protein portion of these complexes plays a role in the delivery of cytotoxic oleic acid molecules into tumor cells across the cell membrane.

Flexible recognition of the tRNA G18 methylation target site by TrmH methyltransferase through first binding and induced fit processes

Transfer RNA (Gm18) methyltransferase (TrmH) catalyzes methyl transfer from S-adenosyl-l-methionine to a conserved G18 in tRNA. We investigated the recognition mechanism of Thermus thermophilus TrmH for its guanosine target. Thirteen yeast tRNAPhe mutant transcripts were prepared in which the modification site and/or other nucleotides in the D-loop were substituted by dG, inosine, or other nucleotides. We then conducted methyl transfer kinetic studies, gel shift assays, and inhibition experiments using these tRNA variants. Sites of methylation were confirmed with RNA sequencing or primer extension. Although the G18G19 sequence is not essential for methylation by TrmH, disruption of G18G19 severely reduces the efficiency of methyl transfer. There is strict recognition of guanosine by TrmH, in that methylation occurs at the adjacent G19 when the G18 is replaced by dG or adenosine. The fact that TrmH methylates guanosine in D-loops from 4 to 12 nucleotides in length suggests that selection of the position of guanosine within the D-loop is relatively flexible. Our studies also demonstrate that the oxygen 6 atom of the guanine base is a positive determinant for TrmH recognition. The recognition process of TrmH for substrate is inducible and product-inhibited, in that tRNAs containing Gm18 are excluded by TrmH. In contrast, substitution of G18 with dG18 results in the formation of a more stable TrmH-tRNA complex. To address the mechanism, we performed the stopped-flow pre-steady state kinetic analysis. The result clearly showed that the binding of TrmH to tRNA is composed of at least three steps, the first bi-molecular binding and the subsequent two uni-molecular induced-fit processes.

Atomic-resolution crystal structure of Borrelia burgdorferi outer surface protein A via surface engineering.

Outer surface protein A (OspA) from Borrelia burgdorferi has an unusual dumbbell‐shaped structure in which two globular domains are connected with a “single‐layer” β‐sheet (SLB). The protein is highly soluble, and it has been recalcitrant to crystallization. Only OspA complexes with Fab fragments have been successfully crystallized. OspA contains a large number of Lys and Glu residues, and these “high entropy” residues may disfavor crystal packing because some of them would need to be immobilized in forming a crystal lattice. We rationally designed a total of 13 surface mutations in which Lys and Glu residues were replaced with Ala or Ser. We successfully crystallized the mutant OspA without a bound Fab fragment and extended structure analysis to a 1.15 Å resolution. The new high‐resolution structure revealed a unique backbone hydration pattern of the SLB segment in which water molecules fill the “weak spots” on both faces of the antiparallel β‐sheet. These well‐defined water molecules provide additional structural links between adjacent β‐strands, and thus they may be important for maintaining the rigidity of the SLB that inherently lacks tight packing afforded by a hydrophobic core. The structure also revealed new information on the side‐chain dynamics and on a solvent‐accessible cavity in the core of the C‐terminal globular domain. This work demonstrates the utility of extensive surface mutation in crystallizing recalcitrant proteins and dramatically improving the resolution of crystal structures, and provides new insights into the stabilization mechanism of OspA.

The Catalytic Domain of Topological Knot tRNA Methyltransferase (TrmH) Discriminates between Substrate tRNA and Nonsubstrate tRNA via an Induced-fit Process

Background: Topologically knotted tRNA methyltransferases specifically recognize substrate tRNA. Results: Site-directed mutagenesis studies, chimeric protein analysis, and pre-steady state kinetics clarify the tRNA recognition sites of TrmH. Conclusion: The N- and C-terminal regions function in the initial binding process, and substrate tRNA is discriminated by the catalytic domain in an induced-fit process. Significance: Study of how proteins recognize RNA is crucial for understanding RNA maturation processes. A conserved guanosine at position 18 (G18) in the D-loop of tRNAs is often modified to 2′-O-methylguanosine (Gm). Formation of Gm18 in eubacterial tRNA is catalyzed by tRNA (Gm18) methyltransferase (TrmH). TrmH enzymes can be divided into two types based on their substrate tRNA specificity. Type I TrmH, including Thermus thermophilus TrmH, can modify all tRNA species, whereas type II TrmH, for example Escherichia coli TrmH, modifies only a subset of tRNA species. Our previous crystal study showed that T. thermophilus TrmH is a class IV S-adenosyl-l-methionine-dependent methyltransferase, which maintains a topological knot structure in the catalytic domain. Because TrmH enzymes have short stretches at the N and C termini instead of a clear RNA binding domain, these stretches are believed to be involved in tRNA recognition. In this study, we demonstrate by site-directed mutagenesis that both N- and C-terminal regions function in tRNA binding. However, in vitro and in vivo chimera protein studies, in which four chimeric proteins of type I and II TrmHs were used, demonstrated that the catalytic domain discriminates substrate tRNAs from nonsubstrate tRNAs. Thus, the N- and C-terminal regions do not function in the substrate tRNA discrimination process. Pre-steady state analysis of complex formation between mutant TrmH proteins and tRNA by stopped-flow fluorescence measurement revealed that the C-terminal region works in the initial binding process, in which nonsubstrate tRNA is not excluded, and that structural movement of the motif 2 region of the catalytic domain in an induced-fit process is involved in substrate tRNA discrimination.

Different Folding Pathways Taken by Highly Homologous Proteins, Goat α-Lactalbumin and Canine Milk Lysozyme

Is the folding pathway conserved in homologous proteins? To address this question, we compared the folding pathways of goat alpha-lactalbumin and canine milk lysozyme using equilibrium and kinetic circular dichroism spectroscopy. Both Ca(2+)-binding proteins have 41% sequence identity and essentially identical backbone structures. The Phi-value analysis, based on the effect of Ca(2+) on the folding kinetics, showed that the Ca(2+)-binding site was well organized in the transition state in alpha-lactalbumin, although it was not yet organized in lysozyme. Equilibrium unfolding and hydrogen-exchange 2D NMR analysis of the molten globule intermediate also showed that different regions were stabilized in the two proteins. In alpha-lactalbumin, the Ca(2+)-binding site and the C-helix were weakly organized, whereas the A- and B-helices, both distant from the Ca(2+)-binding site, were well organized in lysozyme. The results thus provide an example of highly homologous proteins taking different folding pathways. To understand the molecular origin of this difference, we investigated the native three-dimensional structures of the proteins in terms of non-local contact clusters, a parameter based on the residue-residue contact map and known to be well correlated with the folding rate of non-two-state proteins. There were remarkable differences between the proteins in the distribution of the non-local contact clusters, and these differences provided a reasonable explanation of the observed difference in the folding initiation sites. In conclusion, the protein folding pathway is determined not only by the backbone topology but also by the specific side-chain interactions of contacting residues.