Computational tools help improve protein stability but with a solubility tradeoff

CYP2R1 and CYP27A1 genes: An in silico approach to identify the deleterious mutations, impact on structure and their differential expression in disease conditions

Mutations in CYP2R1 and CYP27A1 involved in the conversion of Cholecalciferol into Calcidiol were associated with the impaired 25-hydroxylase activity therefore affecting the Vitamin D metabolism. Hence, this study attempted to understand the influence of genetic variations at the sequence and structural level via computational approach. The non-synonymous mutations retrieved from dbSNP database were assessed for their pathogenicity, stability as well as conservancy using various computational tools. The above analysis predicted 11/260 and 35/489 non-synonymous mutations to be deleterious in CYP2R1 and CYP27A1 genes respectively. Native and mutant forms of the corresponding proteins were modeled. Further, interacting native and mutant proteins with cholecalciferol showed difference in hydrogen bonds, hydrophobic bonds and their binding affinities suggesting the possible influence of these mutations in their function. Also, expression of these genes in various disease conditions was investigated using GEO datasets which predicted that there is a differential expression in cancer and arthritis.

Complementary computational and experimental evaluation of missense variants in the ROMK potassium channel

The renal outer medullary potassium (ROMK) channel is essential for potassium transport in the kidney, and its dysfunction is associated with a salt-wasting disorder known as Bartter syndrome. Despite its physiological significance, we lack a mechanistic understanding of the molecular defects in ROMK underlying most Bartter syndrome-associated mutations. To this end, we employed a ROMK-dependent yeast growth assay and tested single amino acid variants selected by a series of computational tools representative of different approaches to predict each variants’ pathogenicity. In one approach, we used in silico saturation mutagenesis, i.e. the scanning of all possible single amino acid substitutions at all sequence positions to estimate their impact on function, and then employed a new machine learning classifier known as Rhapsody. We also used two additional tools, EVmutation and Polyphen-2, which permitted us to make consensus predictions on the pathogenicity of single amino acid variants in ROMK. Experimental tests performed for selected mutants in different classes validated the vast majority of our predictions and provided insights into variants implicated in ROMK dysfunction. On a broader scope, our analysis suggests that consolidation of data from complementary computational approaches provides an improved and facile method to predict the severity of an amino acid substitution and may help accelerate the identification of disease-causing mutations in any protein.

As the number of sequenced human genomes rises, a major challenge is to identify which single amino acid variations in a protein affect function and predispose individuals to disease. While predictive algorithms are available for this purpose, a comparative analysis of recently developed algorithms has not been adequately performed, nor is it clear whether combining algorithms would improve predictive power. To this end, we compared the efficacy of three publicly available algorithms and applied the results to Bartter syndrome, a human disease for which numerous poorly-characterized single amino acid variants have been identified and for which there is no cure. In silico saturation mutagenesis, i.e., the computational prediction of pathogenesis for every possible amino acid substitution, allowed us to experimentally test predictions by measuring the activity of an ion channel linked to Bartter syndrome. Based on data from blinded experiments, we discovered that Rhapsody and EVmutation successfully predicted deleterious mutations. Moreover, Rhapsody—which takes into account evolutionary as well as structural and dynamic considerations—predicted that >90% of known Bartter syndrome mutations are deleterious. Overall, our data will aid investigators who wish to test single amino acid variants in any protein and aid biomedical researchers who wish to develop hypotheses on the potential severity of genetic variants uncovered from genome databases.

Proteome Instability Is a Therapeutic Vulnerability in Mismatch Repair-Deficient Cancer

Deficient DNA mismatch repair (dMMR) induces a hypermutator phenotype that can lead to tumorigenesis; however, the functional impact of the high mutation burden resulting from this phenotype remains poorly explored. Here, we demonstrate that dMMR-induced destabilizing mutations lead to proteome instability in dMMR tumors, resulting in an abundance of misfolded protein aggregates. To compensate, dMMR cells utilize a Nedd8-mediated degradation pathway to facilitate clearance of misfolded proteins. Blockade of this Nedd8 clearance pathway with MLN4924 causes accumulation of misfolded protein aggregates, ultimately inducing immunogenic cell death in dMMR cancer cells. To leverage this immunogenic cell death, we combined MLN4924 treatment with PD1 inhibition and found the combination was synergistic, significantly improving efficacy over either treatment alone.

Model for Counterion Binding and Charge Reversal on Protein Surfaces

The structural stability and solubility of proteins in liquid therapeutic formulations is important, especially since new generations of therapeutics are designed for efficacy before consideration of stability. We introduce an electrostatic binding model to measure the net charge of proteins with bound ions in solution. The electrostatic potential on a protein surface is used to separately group together acidic and basic amino acids into patches, which are then iteratively bound with oppositely charged counterions. This model is aimed toward formulation chemists for initial screening of a range of conditions prior to lab-work. Computed results compare well with experimental zeta potential measurements from the literature covering a range of solution conditions. Importantly, the binding model reproduces the charge reversal phenomenon that is observed with polyvalent ion binding to proteins and its dependence on ion charge and concentration. Intriguingly, protein sequence can be used to give similarly good agreement with experiment as protein structure, interpreted as resulting from the close proximity of charged side chains on a protein surface. Further, application of the model to human proteins suggests that polyanion binding and overcharging, including charge reversal for cationic proteins, is a general feature. These results add to evidence that addition of polyanions to protein formulations could be a general mechanism for modulating solution stability.

Inherent Biophysical Properties Modulate the Toxicity of Soluble Amyloidogenic Light Chains

In light chain amyloidosis (AL), fibrillar deposition of monoclonal immunoglobulin light chains (LCs) in vital organs, such as heart, is associated with their severe dysfunction. In addition to the cellular damage caused by fibril deposition, direct toxicity of soluble prefibrillar amyloidogenic proteins has been reported, in particular, for cardiotoxicity. However, the molecular bases of proteotoxicity by soluble LCs have not been clarified. Here, to address this issue, we rationally engineered the amino acid sequence of the highly cardiotoxic LC H6 by introducing three residue mutations, designed to reduce the dynamics of its native state. The resulting mutant (mH6) is less toxic than its parent H6 to human cardiac fibroblasts and C. elegans. The high sequence and structural similarity, together with the different toxicity, make H6 and its non-toxic designed variant mH6 a test case to shed light on the molecular properties underlying soluble toxicity. Our comparative structural and biochemical study of H6 and mH6 shows closely matching crystal structures, whereas spectroscopic data and limited proteolysis indicate that H6 displays poorly cooperative fold, higher flexibility, and kinetic instability, and a higher dynamic state in its native fold. Taken together, the results of this study show a strong correlation between the overall conformational properties of the native fold and the proteotoxicity of cardiotropic LCs.

Optimized production of a biologically active Clostridium perfringens glycosyl hydrolase phage endolysin PlyCP41 in plants using virus-based systemic expression

BACKGROUND

Clostridium perfringens, a gram-positive, anaerobic, rod-shaped bacterium, is the third leading cause of human foodborne bacterial disease and a cause of necrotic enteritis in poultry. It is controlled using antibiotics, widespread use of which may lead to development of drug-resistant bacteria. Bacteriophage-encoded endolysins that degrade peptidoglycans in the bacterial cell wall are potential replacements for antibiotics. Phage endolysins have been identified that exhibit antibacterial activities against several Clostridium strains.

RESULTS

An Escherichia coli codon-optimized gene encoding the glycosyl hydrolase endolysin (PlyCP41) containing a polyhistidine tag was expressed in E. coli. In addition, The E. coli optimized endolysin gene was engineered for expression in plants (PlyCP41p) and a plant codon-optimized gene (PlyCP41pc), both containing a polyhistidine tag, were expressed in Nicotiana benthamiana plants using a potato virus X (PVX)-based transient expression vector. PlyCP41p accumulated to ~ 1% total soluble protein (100μg/gm f. wt. leaf tissue) without any obvious toxic effects on plant cells, and both the purified protein and plant sap containing the protein lysed C. perfringens strain Cp39 in a plate lysis assay. Optimal systemic expression of PlyCP41p was achieved at 2 weeks-post-infection. PlyCP41pc did not accumulate to higher levels than PlyCP41p in infected tissue.

CONCLUSION

We demonstrated that functionally active bacteriophage PlyCP41 endolysin can be produced in systemically infected plant tissue with potential for use of crude plant sap as an effective antimicrobial agent against C. perfringens.

Convolution Neural Network-Based Prediction of Protein Thermostability

Most natural proteins exhibit poor thermostability, which limits their industrial application. Computer-aided rational design is an efficient purpose-oriented method that can improve protein thermostability. Numerous machine-learning-based methods have been designed to predict the changes in protein thermostability induced by mutations. However, all of these methods have certain limitations due to existing mutation coding methods that overlook protein sequence features. Here we propose a method to predict protein thermostability using convolutional neural networks based on an in-depth study of thermostability-related protein properties. This method comprises a three-dimensional coding algorithm, including protein mutation information and a strategy to extract neighboring features at protein mutation sites based on multiscale convolution. The accuracies on the S1615 and S388 data sets, which are widely used for protein thermostability predictions, reached 86.4 and 87%, respectively. The Matthews correlation coefficient was nearly double those produced using other methods. Furthermore, a model was constructed to predict the thermostability of Rhizomucor miehei lipase mutants based on the S3661 data set, a single amino acid mutation data set screened from the ProTherm protein thermodynamics database. Compared with the RIF strategy, which consists of three algorithms, i.e., Rosetta ddg monomer, I Mutant 3.0, and FoldX, the accuracy of the proposed method was higher (75.0 vs 66.7%), and the negative sample resolution was simultaneously enhanced. These results indicate that our prediction method more effectively assessed the protein thermostability and distinguished its features, making it a powerful tool to devise mutations that enhance the thermostability of proteins, particularly enzymes.

Direct coupling analysis improves the identification of beneficial amino acid mutations for the functional thermostabilization of a delicate decarboxylase

The optimization of enzyme properties for specific reaction conditions enables their tailored use in biotechnology. Predictions using established computer-based methods, however, remain challenging, especially regarding physical parameters such as thermostability without concurrent loss of activity. Employing established computational methods such as energy calculations using FoldX can lead to the identification of beneficial single amino acid substitutions for the thermostabilization of enzymes. However, these methods require a three-dimensional (3D)-structure of the enzyme. In contrast, coevolutionary analysis is a computational method, which is solely based on sequence data. To enable a comparison, we employed coevolutionary analysis together with structure-based approaches to identify mutations, which stabilize an enzyme while retaining its activity. As an example, we used the delicate dimeric, thiamine pyrophosphate dependent enzyme ketoisovalerate decarboxylase (Kivd) and experimentally determined its stability represented by a T50 value indicating the temperature where 50% of enzymatic activity remained after incubation for 10 min. Coevolutionary analysis suggested 12 beneficial mutations, which were not identified by previously established methods, out of which four mutations led to a functional Kivd with an increased T50 value of up to 3.9°C.

Protein stability engineering insights revealed by domain-wide comprehensive mutagenesis

The accurate prediction of protein stability upon sequence mutation is an important but unsolved challenge in protein engineering. Large mutational datasets are required to train computational predictors, but traditional methods for collecting stability data are either low-throughput or measure protein stability indirectly. Here, we develop an automated method to generate thermodynamic stability data for nearly every single mutant in a small 56-residue protein. Analysis reveals that most single mutants have a neutral effect on stability, mutational sensitivity is largely governed by residue burial, and unexpectedly, hydrophobics are the best tolerated amino acid type. Correlating the output of various stability-prediction algorithms against our data shows that nearly all perform better on boundary and surface positions than for those in the core and are better at predicting large-to-small mutations than small-to-large ones. We show that the most stable variants in the single-mutant landscape are better identified using combinations of 2 prediction algorithms and including more algorithms can provide diminishing returns. In most cases, poor in silico predictions were tied to compositional differences between the data being analyzed and the datasets used to train the algorithm. Finally, we find that strategies to extract stabilities from high-throughput fitness data such as deep mutational scanning are promising and that data produced by these methods may be applicable toward training future stability-prediction tools.

Science and art of protein formulation development

Protein pharmaceuticals have become a significant class of marketed drug products and are expected to grow steadily over the next decade. Development of a commercial protein product is, however, a rather complex process. A critical step in this process is formulation development, enabling the final product configuration. A number of challenges still exist in the formulation development process. This review is intended to discuss these challenges, to illustrate the basic formulation development processes, and to compare the options and strategies in practical formulation development.

DDGun: an untrained method for the prediction of protein stability changes upon single and multiple point variations

Background

Predicting the effect of single point variations on protein stability constitutes a crucial step toward understanding the relationship between protein structure and function. To this end, several methods have been developed to predict changes in the Gibbs free energy of unfolding (∆∆G) between wild type and variant proteins, using sequence and structure information. Most of the available methods however do not exhibit the anti-symmetric prediction property, which guarantees that the predicted ∆∆G value for a variation is the exact opposite of that predicted for the reverse variation, i.e., ∆∆G(A → B) = −∆∆G(B → A), where A and B are amino acids.

Results

Here we introduce simple anti-symmetric features, based on evolutionary information, which are combined to define an untrained method, DDGun (DDG untrained). DDGun is a simple approach based on evolutionary information that predicts the ∆∆G for single and multiple variations from sequence and structure information (DDGun3D). Our method achieves remarkable performance without any training on the experimental datasets, reaching Pearson correlation coefficients between predicted and measured ∆∆G values of ~ 0.5 and ~ 0.4 for single and multiple site variations, respectively. Surprisingly, DDGun performances are comparable with those of state of the art methods. DDGun also naturally predicts multiple site variations, thereby defining a benchmark method for both single site and multiple site predictors. DDGun is anti-symmetric by construction predicting the value of the ∆∆G of a reciprocal variation as almost equal (depending on the sequence profile) to -∆∆G of the direct variation. This is a valuable property that is missing in the majority of the methods.

Conclusions

Evolutionary information alone combined in an untrained method can achieve remarkably high performances in the prediction of ∆∆G upon protein mutation. Non-trained approaches like DDGun represent a valid benchmark both for scoring the predictive power of the individual features and for assessing the learning capability of supervised methods.

Electronic supplementary material

The online version of this article (10.1186/s12859-019-2923-1) contains supplementary material, which is available to authorized users.

Focused dengue vaccine development: outwitting nature's design

The four DENV serotypes are mosquito-borne pathogens that belong to the Flavivirus genus. These viruses present a major global health burden, being endemic in over 120 countries, causing ∼390 million reported infections yearly, with clinical symptoms ranging from mild fever to severe and potentially fatal hemorrhagic syndromes. Development of a safe and efficacious DENV vaccine is challenging because of the need to induce immunity against all four serotypes simultaneously, as immunity against one serotype can potentially enhance disease caused by a heterotypic secondary infection. So far, live-virus particle-based vaccine approaches struggle with inducing protective tetravalent immunity, while recombinant subunit approaches that use the envelope protein (E) as the major antigen, are gaining promise in preclinical studies. However, E-based subunits require further development and characterization to be used as effective vaccine antigens against DENV. In this review, we will address the shortcomings of recombinant E-based antigens and will discuss potential solutions to enhance E-based subunit antigen immunogenicity and vaccine efficacy.

Mechanisms Influencing the Pharmacokinetics and Disposition of Monoclonal Antibodies and Peptides

Monoclonal antibodies (mAbs) and peptides are an important class of therapeutic modalities that have brought improved health outcomes in areas with limited therapeutic optionality. Presently, there more than 90 mAb and peptide therapeutics on the United States market, with over 600 more in various clinical stages of development in a broad array of therapeutic areas, including diabetes, autoimmune disorders, oncology, neuroscience, and cardiovascular and infectious diseases. Notwithstanding this potential, there is high clinical rate of attrition, with approximately 10% reaching patients. A major contributor to the failure of the molecules is often times an incomplete or poor understanding of the pharmacokinetics (PK) and disposition profiles leading to limited or diminished efficacy. Increased and thorough characterization efforts directed at disseminating mechanisms influencing the PK and disposition of mAbs and peptides can aid in improving the design for their intended pharmacological activity, and thereby their clinical success. The PK and disposition factors for mAbs and peptides are broadly influenced by target-mediated drug disposition and nontarget-related clearance mechanisms related to the interplay between the relationship of the structure and physiochemical properties of mAbs and peptides with physiologic processes. This review focuses on nontarget-related factors influencing the disposition and PK of mAbs and peptides. Contemporary considerations around the increasing in silico approaches to identify nontarget-related molecule limitations and enhancing the druggability of mAbs and peptides, including parenteral and nonparenteral delivery strategies that are geared toward improving patient experience and compliance, are also discussed.

Emerging technologies in protein interface engineering for biomedical applications

Protein interactions communicate critical information from the environment into cells to orchestrate functional responses relevant to health and disease. Whereas the natural repertoire of protein interfaces is finite, biomolecular engineering tools provide access to an unlimited scope of potential interactions that can be custom-designed for affinity, specificity, mechanism, or other properties of interest. This review highlights recent developments in protein interface engineering that offer insight into human physiology to inform the design of new pharmaceuticals, with a particular focus on immunotherapeutics. We cover three innovative and translationally promising approaches: (1) reprogramming receptor oligomerization to manipulate signaling pathways; (2) computational protein interface design strategies; and (3) engineering bioorthogonal protein interaction networks.

Insight into the Thermophilic Mechanism of a Glycoside Hydrolase Family 5 β-Mannanase

To study the molecular basis for thermophilic β-mannanase of glycoside hydrolase family 5, two β-mannanases, TlMan5A and PMan5A, from Talaromyces leycettanus JCM12802 and Penicillium sp. WN1 were used as models. The four residues, His112 and Phe113, located near the antiparallel β-sheet at the barrel bottom and Leu375 and Ala408 from loop 7 and loop 8 of PMan5A, were inferred to be key thermostability contributors through module substitution, truncation, and site-directed mutagenesis. The effects of these four residues on the thermal properties followed the order H112Y > A408P > L375H > F113Y and were strongly synergetic. These results were interpreted structurally using molecular dynamics (MD) simulations, which showed that improved hydrophobic interactions in the inner wall of the β-barrel and the rigidity of loop 8 were caused by the outside domain of the barrel bottom and proline, respectively. The TIM barrel bottom and four specific residues responsible for the thermostability of GH5 β-mannanases were elucidated.

KStable: A Computational Method for Predicting Protein Thermal Stability Changes by K-Star with Regular-mRMR Feature Selection

Thermostability is a protein property that impacts many types of studies, including protein activity enhancement, protein structure determination, and drug development. However, most computational tools designed to predict protein thermostability require tertiary structure data as input. The few tools that are dependent only on the primary structure of a protein to predict its thermostability have one or more of the following problems: a slow execution speed, an inability to make large-scale mutation predictions, and the absence of temperature and pH as input parameters. Therefore, we developed a computational tool, named KStable, that is sequence-based, computationally rapid, and includes temperature and pH values to predict changes in the thermostability of a protein upon the introduction of a mutation at a single site. KStable was trained using basis features and minimal redundancy–maximal relevance (mRMR) features, and 58 classifiers were subsequently tested. To find the representative features, a regular-mRMR method was developed. When KStable was evaluated with an independent test set, it achieved an accuracy of 0.708.

Third generation antibody discovery methods: in silico rational design

Owing to their outstanding performances in molecular recognition, antibodies are extensively used in research and applications in molecular biology, biotechnology and medicine. Recent advances in experimental and computational methods are making it possible to complement well-established in vivo (first generation) and in vitro (second generation) methods of antibody discovery with novel in silico (third generation) approaches. Here we describe the principles of computational antibody design and review the state of the art in this field. We then present Modular, a method that implements the rational design of antibodies in a modular manner, and describe the opportunities offered by this approach.

High proportion of transient neonatal zinc deficiency causing alleles in the general population

Loss of function (LoF) mutations in the zinc transporter SLC30A2/ZnT2 result in impaired zinc secretion into breast milk consequently causing transient neonatal zinc deficiency (TNZD) in exclusively breastfed infants. However, the frequency of TNZD causing alleles in the general population is yet unknown. Herein, we investigated 115 missense SLC30A2/ZnT2 mutations from the ExAC database, equally distributed in the entire coding region, harboured in 668 alleles in 60 706 healthy individuals of diverse ethnicity. To estimate the frequency of LoF SLC30A2/ZnT2 mutations in the general population, we used bioinformatics tools to predict the potential impact of these mutations on ZnT2 functionality, and corroborated these predictions by a zinc transport assay in human MCF-7 cells. We found 14 missense mutations that were markedly deleterious to zinc transport. Together with two conspicuous LoF mutations in the ExAC database, 26 SLC30A2/ZnT2 alleles harboured deleterious mutations, suggesting that at least 1 in 2334 newborn infants are at risk to develop TNZD. This high frequency of TNZD mutations combined with the World Health Organization-promoted increase in the rate of exclusive breastfeeding highlights the importance of genetic screening for inactivating SLC30A2/ZnT2 mutations in the general population for the early diagnosis and prevention of TNZD.

The state-of-the-art strategies of protein engineering for enzyme stabilization

Enzymes generated by natural recruitment and protein engineering have greatly contribute in various sets of applications. However, their insufficient stability is a bottleneck that limit the rapid development of biocatalysis. Novel approaches based on precise and global structural dissection, advanced gene manipulation, and combination with the multidisciplinary techniques open a new horizon to generate stable enzymes efficiently. Here, we comprehensively introduced emerging advances of protein engineering strategies for enzyme stabilization. Then, we highlighted practical cases to show importance of enzyme stabilization in pharmaceutical and industrial applications. Combining computational enzyme design with molecular evolution will hold considerable promise in this field.

Role of simple descriptors and applicability domain in predicting change in protein thermostability

The melting temperature (Tm) of a protein is the temperature at which half of the protein population is in a folded state. Therefore, Tm is a measure of the thermostability of a protein. Increasing the Tm of a protein is a critical goal in biotechnology and biomedicine. However, predicting the change in melting temperature (dTm) due to mutations at a single residue is difficult because it depends on an intricate balance of forces. Existing methods for predicting dTm have had similar levels of success using generally complex models. We find that training a machine learning model with a simple set of easy to calculate physicochemical descriptors describing the local environment of the mutation performed as well as more complicated machine learning models and is 2-6 orders of magnitude faster. Importantly, unlike in most previous publications, we perform a blind prospective test on our simple model by designing 96 variants of a protein not in the training set. Results from retrospective and prospective predictions reveal the limited applicability domain of each model. This study highlights the current deficiencies in the available dTm dataset and is a call to the community to systematically design a larger and more diverse experimental dataset of mutants to prospectively predict dTm with greater certainty.

Recent advances in automated protein design and its future challenges

INTRODUCTION

Protein function is determined by protein structure which is in turn determined by the corresponding protein sequence. If the rules that cause a protein to adopt a particular structure are understood, it should be possible to refine or even redefine the function of a protein by working backwards from the desired structure to the sequence. Automated protein design attempts to calculate the effects of mutations computationally with the goal of more radical or complex transformations than are accessible by experimental techniques. Areas covered: The authors give a brief overview of the recent methodological advances in computer-aided protein design, showing how methodological choices affect final design and how automated protein design can be used to address problems considered beyond traditional protein engineering, including the creation of novel protein scaffolds for drug development. Also, the authors address specifically the future challenges in the development of automated protein design. Expert opinion: Automated protein design holds potential as a protein engineering technique, particularly in cases where screening by combinatorial mutagenesis is problematic. Considering solubility and immunogenicity issues, automated protein design is initially more likely to make an impact as a research tool for exploring basic biology in drug discovery than in the design of protein biologics.

FoldX as Protein Engineering Tool: Better Than Random Based Approaches?

Improving protein stability is an important goal for basic research as well as for clinical and industrial applications but no commonly accepted and widely used strategy for efficient engineering is known. Beside random approaches like error prone PCR or physical techniques to stabilize proteins, e.g. by immobilization, in silico approaches are gaining more attention to apply target-oriented mutagenesis. In this review different algorithms for the prediction of beneficial mutation sites to enhance protein stability are summarized and the advantages and disadvantages of FoldX are highlighted. The question whether the prediction of mutation sites by the algorithm FoldX is more accurate than random based approaches is addressed.

Improvement in the Thermostability of a β-Amino Acid Converting ω-Transaminase by Using FoldX

ω-Transaminases (ω-TAs) are important biocatalysts for the synthesis of active, chiral pharmaceutical ingredients containing amino groups, such as β-amino acids, which are important in peptidomimetics and as building blocks for drugs. However, the application of ω-TAs is limited by the availability and stability of enzymes with high conversion rates. One strategy for the synthesis and optical resolution of β-phenylalanine and other important aromatic β-amino acids is biotransformation by utilizing an ω-transaminase from Variovorax paradoxus. We designed variants of this ω-TA to gain higher process stability on the basis of predictions calculated by using the FoldX software. We herein report the first thermostabilization of a nonthermostable S-selective ω-TA by FoldX-guided site-directed mutagenesis. The melting point (T_m ) of our best-performing mutant was increased to 59.3 °C, an increase of 4.0 °C relative to the T_m value of the wild-type enzyme, whereas the mutant fully retained its specific activity.