In vivo fragment complementation of a (β/α)8barrel protein: generation of variability by recombination

Recurring sequence-structure motifs in (βα)8-barrel proteins and experimental optimization of a chimeric protein designed based on such motifs

An interesting way of generating novel artificial proteins is to combine sequence motifs from natural proteins, mimicking the evolutionary path suggested by natural proteins comprising recurring motifs. We analyzed the βα and αβ modules of TIM barrel proteins by structure alignment-based sequence clustering. A number of preferred motifs were identified. A chimeric TIM was designed by using recurring elements as mutually compatible interfaces. The foldability of the designed TIM protein was then significantly improved by six rounds of directed evolution. The melting temperature has been improved by more than 20°C. A variety of characteristics suggested that the resulting protein is well-folded. Our analysis provided a library of peptide motifs that is potentially useful for different protein engineering studies. The protein engineering strategy of using recurring motifs as interfaces to connect partial natural proteins may be applied to other protein folds.

Half-Barrels Derived from a (β/α)8 Barrel β-Glycosidase Undergo an Activation Process

The evolution of (β/α)₈ barrel proteins is currently thought to have involved the fusion of two (β/α)₄ half-barrels, thereby conferring stability on the protein structure. After the formation of a whole (β/α)₈ barrel, this structure could evolve and diverge to form fully active enzymes. Interestingly, we show here that isolated (β/α)₄ half-barrels derived from the N- and C-terminal domains of the β-glucosidase Sfβgly (Sfβgly-N: residues 1 to 265; Sfβgly-C: residues 266 to 509) undergo an activation process, which renders them catalytically active. The rate constants of the activation process were calculated to be 0.029 and 0.032 h^-1 for Sfβgly-N and Sfβgly-C, respectively. Moreover, the Sfβgly-N and Sfβgly-C activation processes were simultaneous with modifications in their initial structure, which reduced the exposure of their tryptophan residues. Importantly, this activation was also coincident with an increase in the sizes of Sfβgly-N and Sfβgly-C particles. These novel observations suggest that the change in catalytic activity associated with the transition from a half to whole (β/α)₈ barrel might also have driven such an evolutionary process.

Alternative splice variants in TIM barrel proteins from human genome correlate with the structural and evolutionary modularity of this versatile protein fold

After the surprisingly low number of genes identified in the human genome, alternative splicing emerged as a major mechanism to generate protein diversity in higher eukaryotes. However, it is still not known if its prevalence along the genome evolution has contributed to the overall functional protein diversity or if it simply reflects splicing noise. The (βα)₈ barrel or TIM barrel is one of the most frequent, versatile, and ancient fold encountered among enzymes. Here, we analyze the structural modifications present in TIM barrel proteins from the human genome product of alternative splicing events. We found that 87% of all splicing events involved deletions; most of these events resulted in protein fragments that corresponded to the (βα)₂, (βα)₄, (βα)₅, (βα)₆, and (βα)₇ subdomains of TIM barrels. Because approximately 7% of all the splicing events involved internal β-strand substitutions, we decided, based on the genomic data, to design β-strand and α-helix substitutions in a well-studied TIM barrel enzyme. The biochemical characterization of one of the chimeric variants suggests that some of the splice variants in the human genome with β-strand substitutions may be evolving novel functions via either the oligomeric state or substrate specificity. We provide results of how the splice variants represent subdomains that correlate with the independently folding and evolving structural units previously reported. This work is the first to observe a link between the structural features of the barrel and a recurrent genetic mechanism. Our results suggest that it is reasonable to expect that a sizeable fraction of splice variants found in the human genome represent structurally viable functional proteins. Our data provide additional support for the hypothesis of the origin of the TIM barrel fold through the assembly of smaller subdomains. We suggest a model of how nature explores new proteins through alternative splicing as a mechanism to diversify the proteins encoded in the human genome.

Folding mechanism of an extremely thermostable (βα)(8)-barrel enzyme: a high kinetic barrier protects the protein from denaturation

HisF, the cyclase subunit of imidazole glycerol phosphate synthase (ImGPS) from Thermotoga maritima, is an extremely thermostable (βα)(8)-barrel protein. We elucidated the unfolding and refolding mechanism of HisF. Its unfolding transition is reversible and adequately described by the two-state model, but 6 weeks is necessary to reach equilibrium (at 25 °C). During refolding, initially a burst-phase off-pathway intermediate is formed. The subsequent productive folding occurs in two kinetic phases with time constants of ~3 and ~20 s. They reflect a sequential process via an on-pathway intermediate, as revealed by stopped-flow double-mixing experiments. The final step leads to native HisF, which associates with the glutaminase subunit HisH to form the functional ImGPS complex. The conversion of the on-pathway intermediate to the native protein results in a 10(6)-fold increase of the time constant for unfolding from 89 ms to 35 h (at 4.0 M GdmCl) and thus establishes a high energy barrier to denaturation. We conclude that the extra stability of HisF is used for kinetic protection against unfolding. In its refolding mechanism, HisF resembles other (βα)(8)-barrel proteins.

Mechanisms of protein evolution and their application to protein engineering

Protein engineering holds great promise for the development of new biosensors, diagnostics, therapeutics, and agents for bioremediation. Despite some remarkable successes in experimental and computational protein design, engineered proteins rarely achieve the efficiency or specificity of natural enzymes. Current protein design methods utilize evolutionary concepts, including mutation, recombination, and selection, but the inability to fully recapitulate the success of natural evolution suggests that some evolutionary principles have not been fully exploited. One aspect of protein engineering that has received little attention is how to select the most promising proteins to serve as templates, or scaffolds, for engineering. Two evolutionary concepts that could provide a rational basis for template selection are the conservation of catalytic mechanisms and functional promiscuity. Knowledge of the catalytic motifs responsible for conserved aspects of catalysis in mechanistically diverse superfamilies could be used to identify promising templates for protein engineering. Second, protein evolution often proceeds through promiscuous intermediates, suggesting that templates which are naturally promiscuous for a target reaction could enhance protein engineering strategies. This review explores these ideas and alternative hypotheses concerning protein evolution and engineering. Future research will determine if application of these principles will lead to a protein engineering methodology governed by predictable rules for designing efficient, novel catalysts.

Protein design through systematic catalytic loop exchange in the (beta/alpha)8 fold

Protein engineering by directed evolution has proven effective in achieving various functional modifications, but the well-established protocols for the introduction of variability, typically limited to random point mutations, seriously restrict the scope of the approach. In an attempt to overcome this limitation, we sought to explore variant libraries with richer diversity at regions recognized as functionally important through an exchange of natural components, thus combining design with combinatorial diversity. With this approach, we expected to maintain interactions important for protein stability while directing the introduction of variability to areas important for catalysis. Our strategy consisted in loop exchange over a (beta/alpha)(8) fold. Phosphoribosylanthranilate isomerase was chosen as scaffold, and we investigated its tolerance to loop exchange by fusing variant libraries to the chloramphenicol acetyl transferase coding gene as an in vivo folding reporter. We replaced loops 2, 4, and 6 of phosphoribosylanthranilate isomerase with loops of varied types and sizes from enzymes sharing the same fold. To allow for a better structural fit, saturation mutagenesis was adopted at two amino acid positions preceding the exchanged loop. Our results showed that 30% to 90% of the generated mutants in the different libraries were folded. Some variants were selected for further characterization after removal of chloramphenicol acetyl transferase gene, and their stability was studied by circular dichroism and fluorescence spectroscopy. The sequences of 545 clones show that the introduction of variability at "hinges" connecting the loops with the scaffold exhibited a noticeable effect on the appearance of folded proteins. Also, we observed that each position accepted foreign loops of different sizes and sequences. We believe our work provides the basis of a general method of exchanging variably sized loops within the (beta/alpha)(8) fold, affording a novel starting point for the screening of novel activities as well as modest diversions from an original activity.

Identification of viral peptide fragments for vaccine development

We report a simple method for identifying foldable viral surface protein fragments in a random but systematic manner. The method involves digestion and reassembly of a target gene to generate a pool of smaller DNA fragments with random ends but controllable lengths, followed by screening for foldable fragments using green fluorescent protein (GFP) as a folding reporter. The surface glycoproteins of SARS-CoV and HIV-1 were used as model proteins. Two foldable fragments for SARS-CoV spike protein were identified, which coincide with various anti-SARS-CoV peptides. A similar treatment of the HIV-1 gp120 yielded a number of fragments that are associated with the critical CD4 binding site, or the partially buried CCR5 binding site of the protein. The random dissection approach described here should be applicable to other viral proteins for isolating soluble viral surface protein fragments, and may provide alternatives to the full-length proteins (subunits) or linear short peptides in search for antigen or vaccine candidates.

Folding and unfolding of gammaTIM monomers and dimers

Kinetic simulations of the folding and unfolding of triosephosphate isomerase (TIM) from yeast were conducted using a single monomer gammaTIM polypeptide chain that folds as a monomer and two gammaTIM chains that fold to the native dimer structure. The basic protein model used was a minimalist Gō model using the native structure to determine attractive energies in the protein chain. For each simulation type--monomer unfolding, monomer refolding, dimer unfolding, and dimer refolding--thirty simulations were conducted, successfully capturing each reaction in full. Analysis of the simulations demonstrates four main conclusions. First, all four simulation types have a similar "folding order", i.e., they have similar structures in intermediate stages of folding between the unfolded and folded state. Second, despite this similarity, different intermediate stages are more or less populated in the four different simulations, with 1), no intermediates populated in monomer unfolding; 2), two intermediates populated with beta(2)-beta(4) and beta(1)-beta(5) regions folded in monomer refolding; 3), two intermediates populated with beta(2)-beta(3) and beta(2)-beta(4) regions folded in dimer unfolding; and 4), two intermediates populated with beta(1)-beta(5) and beta(1)-beta(5) + beta(6) + beta(7) + beta(8) regions folded in dimer refolding. Third, simulations demonstrate that dimer binding and unbinding can occur early in the folding process before complete monomer-chain folding. Fourth, excellent agreement is found between the simulations and MPAX (misincorporation proton alkyl exchange) experiments. In total, this agreement demonstrates that the computational Gō model is accurate for gammaTIM and that the energy landscape of gammaTIM appears funneled to the native state.

Stabilisation of a (betaalpha)8-barrel protein designed from identical half barrels

It has been suggested that the common (betaalpha)(8)-barrel enzyme fold has evolved by the duplication and fusion of identical (betaalpha)(4)-half barrels, followed by the optimisation of their interface. In our attempts to reconstruct these events in vitro we have previously linked in tandem two copies of the C-terminal half barrel HisF-C of imidazole glycerol phosphate synthase from Thermotoga maritima and subsequently reconstituted in the fusion construct HisF-CC a salt bridge cluster present in wild-type HisF. The resulting recombinant protein HisF-C*C, which was produced in an insoluble form and unfolded with low cooperativity at moderate urea concentrations has now been stabilised and solubilised by a combination of random mutagenesis and selection in vivo. For this purpose, Escherichia coli cells were transformed with a plasmid-based gene library encoding HisF-C*C variants fused to chloramphenicol acetyltransferase (CAT). Stable and soluble variants were identified by the survival of host cells on solid medium containing high concentrations of the antibiotic. The selected HisF-C*C proteins, which were characterised in vitro in the absence of CAT, contained eight different amino acid substitutions. One of the exchanges (Y143C) stabilised HisF-C*C by the formation of an intermolecular disulfide bond. Three of the substitutions (G245R, V248M, L250Q) were located in the long loop connecting the two HisF-C copies, whose subsequent truncation from 13 to 5 residues yielded the stabilised variant HisF-C*C Delta. From the remaining substitutions, Y143H and V234M were most beneficial, and molecular dynamics simulations suggest that they strengthen the interactions between the half barrels by establishing a hydrogen-bonding network and an extensive hydrophobic cluster, respectively. By combining the loop deletion of HisF-C*C Delta with the Y143H and V234M substitutions, the variant HisF-C**C was generated. Recombinant HisF-C**C is produced in soluble form, forms a pure monomer with its tryptophan residues shielded from solvent and unfolds with similar cooperativity as HisF. Our results show that, starting from two identical and fused half barrels, few amino acid exchanges are sufficient to generate a highly stable and compact (betaalpha)(8)-barrel protein with wild-type like structural properties.

A remote but significant sequence homology between glycoside hydrolase clan GH-H and family GH31

Although both the alpha-amylase super-family, i.e. the glycoside hydrolase (GH) clan GH-H (the GH families 13, 70 and 77), and family GH31 share some characteristics, their different catalytic machinery prevents classification of GH31 in clan GH-H. A significant but remote evolutionary relatedness is, however, proposed for clan GH-H with GH31. A sequence alignment, based on the idea that residues equivalent in the primordial catalytic GH-H/GH31 (beta/alpha)(8)-barrel may not be found in the present-day GH-H and GH31 structures at strictly equivalent positions, shows remote sequence homologies covering beta3, beta4, beta7 and beta8 of the GH-H and GH31 (beta/alpha)(8)-barrels. Structure comparison of GH13 alpha-amylase and GH31 alpha-xylosidase guided alignment of GH-H and GH31 members for construction of evolutionary trees. The closest sequence relationship displayed by GH31 is to GH77 of clan GH-H.

Generation of variability by in vivo recombination of halves of a (beta/alpha)8 barrel protein

Similar to what has been achieved with nucleic acids, directed evolution of proteins would be greatly facilitated by the availability of large libraries and efficient selection methods. So far, host cell transformation efficiency has been a bottleneck, practically limiting libraries to sizes less than 10(9). One way to circumvent this problem has been implemented with antibody systems, where contribution to the binding site is provided by two different polypeptides (light and heavy chains). The central concept is the construction of binary systems in which the gene from the two chains are separated by a cre-lox recombinase recognition site, packaged in a phage, and subsequently introduced, by multiple infection, into a recombinase expressing cell [Sblattero D, Bradbury A. Nat Biotechnol 2000;18(1):75-80]. Here, we describe the development of a system which applies the same concept to a single-domain enzyme, the cytoplasmic (beta/alpha)8 barrel protein phosphoribosyl anthranilate isomerase (PRAI) from E. coli. For that purpose, we identified the site at which a loop containing the recognition sequence for cre-lox recombinase could be inserted yielding a functional enzyme. We evaluated the effect of this insertion on the capability of the engineered gene to complement a trp F-E. coli strain and the efficiency of the system to recover the original sequence from an abundance of non-functional mutant genes.

In vitro selection and characterization of a stable subdomain of phosphoribosylanthranilate isomerase

The (beta(alpha))8-barrel is the most common enzyme fold and it is capable of catalyzing an enormous diversity of reactions. It follows that this scaffold should be an ideal starting point for engineering novel enzymes by directed evolution. However, experiments to date have utilized in vivo screens or selections and the compatibility of (beta(alpha))8-barrels with in vitro selection methods remains largely untested. We have investigated plasmid display as a suitable in vitro format by engineering a variant of phosphoribosylanthranilate isomerase (PRAI) that carried the FLAG epitope in active-site-forming loop 6. Trial enrichments for binding of mAb M2 (a mAb to FLAG) demonstrated that FLAG-PRAI could be identified from a 10(6)-fold excess of a FLAG-negative competitor in three rounds of in vitro selection. These results suggest PRAI as a useful scaffold for epitope and peptide grafting experiments. Further, we constructed a FLAG-PRAI loop library of approximately 10(7) clones, in which the epitope residues most critical for binding mAb M2 were randomized. Four rounds of selection for antibody binding identified and enriched for a variant in which a single nucleotide insertion produced a truncated (beta(alpha))8-barrel consisting of (beta(alpha))1-5beta6. Biophysical characterization of this clone, trPRAI, demonstrated that it was selected because of a 21-fold increase in mAb M2 affinity compared with full-length FLAG-PRAI. Remarkably, this truncated barrel was found to be soluble, structured, thermostable and monomeric, implying that it represents a genuine subdomain of PRAI and providing further evidence that such subdomains have played an important role in the evolution of the (beta(alpha))8-barrel fold.

Equilibrium and kinetic folding pathways of a TIM barrel with a funneled energy landscape

The role of native contact topology in the folding of a TIM barrel model based on the alpha-subunit of tryptophan synthase (alphaTS) from Salmonella typhimurium (Protein Data Bank structure 1BKS) was studied using both equilibrium and kinetic simulations. Equilibrium simulations of alphaTS reveal the population of two intermediate ensembles, I1 and I2, during unfolding/refolding at the folding temperature, Tf = 335 K. Equilibrium intermediate I1 demonstrates discrete structure in regions alpha0-beta6 whereas intermediate I2 is a loose ensemble of states with N-terminal structure varying from at least beta1-beta3 (denoted I2A) to alpha0-beta4 at most (denoted I2B). The structures of I1 and I2 match well with the two intermediate states detected in equilibrium folding experiments of Escherichia coli alphaTS. Kinetic folding simulations of alphaTS reveal the sequential population of four intermediate ensembles, I120Q, I200Q, I300Q, and I360Q, during refolding. Kinetic intermediates I120Q, I200Q, and I300Q are highly similar to equilibrium alphaTS intermediates I2A, I2B, and I1, respectively, consistent with kinetic experiments on alphaTS from E. coli. A small population (approximately 10%) of kinetic trajectories are trapped in the I120Q intermediate ensemble and require a slow and complete unfolding step to properly refold. Both the on-pathway and off-pathway I120Q intermediates show structure in beta1-beta3, which is also strikingly consistent with kinetic folding experiments of alphaTS. In the off-pathway intermediate I(120Q), helix alpha2 is wrapped in a nonnative chiral arrangement around strand beta3, sterically preventing the subsequent folding step between beta3 and beta4. These results demonstrate the success of combining kinetic and equilibrium simulations of minimalist protein models to explore TIM barrel folding and the folding of other large proteins.