Computational Modeling of Structurally Conserved Cancer Mutations in the RET and MET Kinases: The Impact on Protein Structure, Dynamics, and Stability

Structural and biochemical characterization of protein kinases that confer oncogene addiction and harbor a large number of disease-associated mutations, including RET and MET kinases, have provided insights into molecular mechanisms associated with the protein kinase activation in human cancer. In this article, structural modeling, molecular dynamics, and free energy simulations of a structurally conserved mutational hotspot, shared by M918T in RET and M1250T in MET kinases, are undertaken to quantify the molecular mechanism of activation and the functional role of cancer mutations in altering protein kinase structure, dynamics, and stability. The mechanistic basis of the activating RET and MET cancer mutations may be driven by an appreciable free energy destabilization of the inactive kinase state in the mutational forms. According to our results, the locally enhanced mobility of the cancer mutants and a higher conformational entropy are counterbalanced by a larger enthalpy loss and result in the decreased thermodynamic stability. The computed protein stability differences between the wild-type and cancer kinase mutants are consistent with circular dichroism spectroscopy and differential scanning calorimetry experiments. These results support the molecular mechanism of activation, which causes a detrimental imbalance in the dynamic equilibrium shifted toward the active form of the enzyme. Furthermore, computer simulations of the inhibitor binding with the oncogenic and drug-resistant RET mutations have also provided a plausible molecular rationale for the observed differences in the inhibition profiles, which is consistent with the experimental data. Finally, structural mapping of RET and MET cancer mutations and the computed protein stability changes suggest a similar mechanism of activation, whereby the cancer mutations which display the higher oncogenic activity tend to have the greatest destabilization effect on the inactive kinase structure.     

A novel mechanism for preventing mutations caused by oxidation of guanine nucleotides

MutT-related proteins, including the Escherichia coli MutT and human MutT homologue 1 (MTH1) proteins, degrade 8-oxo- 7,8-dihydrodeoxyguanosine triphosphate (8-oxo-dGTP) to a monophosphate, thereby preventing mutations caused by the misincorporation of 8-oxoguanine into DNA. Here, we report that human cells have another mechanism for cleaning up the nucleotide pool to ensure accurate DNA replication. The human Nudix type 5 (NUDT5) protein hydrolyses 8-oxo-dGDP to monophosphate with a K_m of 0.77 µM, a value considerably lower than that for ADP sugars, which were originally identified as being substrates of NUDT5. NUDT5 hydrolyses 8-oxo-dGTP only at very low levels, but is able to substitute for MutT when it is defective. When NUDT5 is expressed in E. coli mutT⁻ cells, the increased frequency of spontaneous mutations is decreased to normal levels. Considering the enzymatic parameters of MTH1 and NUDT5 for oxidized guanine nucleotides, NUDT5 might have a much greater role than MTH1 in preventing the occurrence of mutations that are caused by the misincorporation of 8-oxoguanine in human cells.     

Consequences of single-site mutations in the intestinal fatty acid binding protein

The intestinal fatty acid binding protein (IFABP) is a small (15 kDa) protein consisting mostly of 10 antiparallel beta-strands (A-J) and a small helical region that serves as a portal for the ligand. Two beta-sheet structures (strands A-E and F-J) surround a cavity into which the ligand binds. In this work, we investigated how changes in the side chains of specific residues are propagated through the structure. To determine what these changes were and how they relate to changes in stability, (15)N chemical shift perturbations were measured and compared to those of the wild-type protein. Seven mutations, five of which change either valine or leucine to glycine, have been examined. All these mutants were less stable than wild-type IFABP, suggesting some structural changes. For five of the mutants, the data suggest that destabilization of a small region of the protein propagates throughout the structure, resulting in an overall decrease in stability. In two (Leu38Gly and Leu89Gly), the loss of cooperativity in the equilibrium denaturation curves suggests that the destabilization of one region may not be transmitted to other regions in a cooperative manner. It is shown that the effect of mutating hydrophobic residues is much greater than that observed upon mutation of a solvent-exposed polar residue.     

Structural effects of Parkinson's disease linked DJ-1 mutations

Mutations in the protein DJ-1 are associated with familial forms of Parkinson's disease, indicating that DJ-1 may be involved in pathways related to the etiology of this disorder. Here we have used solution state NMR and circular dichroism spectroscopies to evaluate the extent of structural perturbations associated with five different Parkinson's disease linked DJ-1mutations: L166P, E64D, M26I, A104T, and D149A. Comparison of the data with those obtained for the wild-type protein shows that the L166P mutation leads to severe and global destabilization and unfolding of the protein structure, while the structure of the E64D mutation, as expected, is nearly unperturbed. Interestingly, the remaining three mutants all show different degrees of structural perturbation, which are accompanied by a reduction in the thermodynamic stability of the protein. The observed structural and thermodynamic differences are likely to underlie any functional variations between these mutants and the wild type, which in turn are likely responsible for the pathogenicity of these mutations.     

Theoretical studies of the response of a protein structure to cavity-creating mutations

We have investigated the response of a protein structure to cavity-creating mutations by molecular dynamics (MD) simulations for the wild-type and the five mutants of phage T4 lysozyme. Essential dynamics (ED) analysis and the methods for calculating different components of local interaction energies are used to examine the structural and energetic characteristics associated with the mutations. In agreement with the x-ray results, it is found that the structural changes due to the replacements of a bulky side chain such as Leu or Phe with Ala within the hydrophobic core can be characterized as slight adjustments rather than substantial reorganization of the protein. The relative stability of different mutant structures can be related with the extent of structural readjustments in response to the mutation. The destabilization of the mutant Leu-->Ala proteins relative to the wild-type is closely related with the loss of van der Waals contacts due to the cavity-creating mutations.     

Structural and Functional Impact of Cancer-Related Missense Somatic Mutations

A number of large-scale cancer somatic genome sequencing projects are now identifying genetic alterations in cancers. Evaluation of the effects of these mutations is essential for understanding their contribution to tumorigenesis. We have used SNPs3D, a software suite originally developed for analyzing nonsynonymous germ-line variants, to identify single-base mutations with a high impact on protein structure and function. Two machine learning methods are used: one identifying mutations that destabilize protein three-dimensional structure and the other utilizing sequence conservation and detecting all types of effects on in vivo protein function. Incorporation of detailed structure information into the analysis allows detailed interpretation of the functional effects of mutations in specific cases.  Data from a set of breast and colorectal tumors were analyzed. In known cancer genes, mutations approaching 100% of mutations are found to impact protein function, supporting the view that these methods are appropriate for identifying driver mutations. Overall, 50-60% of all somatic missense mutations are predicted to have a high impact on structural stability or to more generally affect the function of the corresponding proteins. This value is similar to the fraction of all possible missense mutations that have a high impact and is much higher than the corresponding one for human population single-nucleotide polymorphisms, at about 30%. The majority of mutations in tumor suppressors destabilize protein structure, while mutations in oncogenes operate in more varied ways, including destabilization of less active conformational states. The set of high-impact mutations encompasses the possible drivers.     

Ensemble-based computational approach discriminates functional activity of p53 cancer and rescue mutants

The tumor suppressor protein p53 can lose its function upon single-point missense mutations in the core DNA-binding domain (“cancer mutants”). Activity can be restored by second-site suppressor mutations (“rescue mutants”). This paper relates the functional activity of p53 cancer and rescue mutants to their overall molecular dynamics (MD), without focusing on local structural details. A novel global measure of protein flexibility for the p53 core DNA-binding domain, the number of clusters at a certain RMSD cutoff, was computed by clustering over 0.7 µs of explicitly solvated all-atom MD simulations. For wild-type p53 and a sample of p53 cancer or rescue mutants, the number of clusters was a good predictor of in vivo p53 functional activity in cell-based assays. This number-of-clusters (NOC) metric was strongly correlated (r² = 0.77) with reported values of experimentally measured ΔΔG protein thermodynamic stability. Interpreting the number of clusters as a measure of protein flexibility: (i) p53 cancer mutants were more flexible than wild-type protein, (ii) second-site rescue mutations decreased the flexibility of cancer mutants, and (iii) negative controls of non-rescue second-site mutants did not. This new method reflects the overall stability of the p53 core domain and can discriminate which second-site mutations restore activity to p53 cancer mutants.     

Genetic and Physiological Characterization of met15 Mutants of SACCHAROMYCES CEREVISIAE: a Selective System for Forward and Reverse Mutations

One hundred and thirty-three spontaneous and induced mutants of the met15 locus in Saccharomyces cerevisiae were characterized with respect to temperature sensitivity, osmotic remediability, interallelic complementation, and suppressibility by amber and ochre suppressors. Forty mutants are osmotic remedial; 17 of these, and no others, are also temperature-sensitive. Seven of 133 mutations are suppressible by an amber suppressor and 11 are suppressible by an ochre suppressor. Seventy percent of the mutants exhibited interallelic complementation, suggesting that the functional gene product of the met15 gene is a multimeric protein. Relative map positions of 30 met15 were estimated from the frequencies of X-ray-induced mitotic reversion of various heteroallelic diploids. All complementing nonsense mutations are located near one end of the gene in contrast to other nonsense mutations which span most of the gene, thus relating the direction of translation of the mRNA with respect to the fine-structure map. Recombination studies indicated that two of 30 mutants contained deletions of the entire met15 locus.—It was established that a variety of mutational types, including missense, nonsense, and deletions, are recovered with this unique system in which both forward and reverse mutations can be selected on the basis of methyl mercury resistance and methionine requirement of the met15 mutants.     

Molecular dynamics as a tool to detect protein foldability. A mutant of domain B1 of protein G with non-native secondary structure propensities

The usefulness of molecular dynamics to assess the structural integrity of mutants containing several mutations has been investigated. Our goal was to determine whether molecular dynamics would be able to discriminate mutants of a protein having a close-to-wild-type fold, from those that are not folded under the same conditions. We used as a model the B1 domain of protein G in which we replaced the unique central alpha-helix by the sequence of the second beta-hairpin, which has a strong intrinsic propensity to form this secondary structure in solution. In the resulting protein, one-third of the secondary structure has been replaced by a non-native one. Models of the mutants were built based on the three-dimensional structure of the wild-type GB1 domain. During 2 ns of molecular dynamics simulations on these models, mutants containing up to 10 mutations in the helix retained the native fold, while another mutant with an additional mutation unfolded. This result is in agreement with our circular dichroism and NMR experiments, which indicated that the former mutants fold into a structure similar to the wild-type, as opposed to the latter mutant which is partly unfolded. Additionally, a mutant containing six mutations scattered through the surface of the domain, and which is unfolded, was also detected by the simulation. This study suggests that molecular dynamics calculations could be performed on molecular models of mutants of a protein to evaluate their foldability, prior to a mutagenesis experiment.     

Modification of [PSI +] prion properties by combining amino acid changes in N-terminal domain of Sup35 protein

The prion [PSI ⁺] is an amyloid isoform of the release factor eRF3 encoded by the SUP35 gene in Saccharomyces cerevisiae yeast. Naturally occurring amyloid complexes have been studied for a long time, yet their structural organization is still not well understood. The formation of amyloid forms of the wild-type Sup35 protein (Sup35p) is directed by its N-terminal portion, which forms a superpleated β-sheet structure. We previously constructed five mutants, each of which carried a replacement in two consecutive amino acids, one in each of the oligopeptide repeats (OR) and in the Sup35p N-terminal region. Mutations sup35-M1 (YQ46-47KK) and sup35-M2 (QQ61-62KK) lead to the compete loss of prion conformation. Three other mutants, i.e., sup35-M3 (QQ70-71KK), sup35-M4 (QQ80-81KK), and sup35-M5 (QQ89-90KK), formed functional prions. In the current study, we investigated the contribution of each mutant peptide to the stability of the prion and aggregation properties, and compared the effects of single mutants and combinations of different mutant alleles. Studies were carried out in yeast strains designed to carry single or a combination of different SUP35 alleles. Based on our analysis, we propose a model that clarifies the 3D organization of the β-sheet within the prion. We also provide evidence that sup35-M2 and sup35-M4 mutations change the 3D structure of prion complexes. We propose that the destabilization of prion complexes in these mutants is due to the decreased efficiency of the fragmentation of the prion aggregates by chaperone complexes.     

Structure and Functional Characterization of Human Histidine Triad Nucleotide-Binding Protein 1 Mutations Associated with Inherited Axonal Neuropathy with Neuromyotonia

Inherited peripheral neuropathies are a group of neurodegenerative disorders that clinically affect 1 in 2500 individuals. Recently, genetic mutations in human histidine nucleotide-binding protein 1 (hHint1) have been strongly and most frequently associated with patients suffering from axonal neuropathy with neuromyotonia. However, the correlation between the impact of these mutations on the hHint1 structure, enzymatic activity and in vivo function has remained ambiguous. Here, we provide detailed biochemical characterization of a set of these hHint1 mutations. Our findings indicate that half of the mutations (R37P, G93D and W123*) resulted in a destabilization of the dimeric state and a significant decrease in catalytic activity and HINT1 inhibitor binding affinity. The H112N mutant was found to be dimeric, but devoid of catalytic activity, due to the loss of the catalytically essential histidine; nevertheless, it exhibited high affinity to AMP and a HINT1 inhibitor. In contrast to the active-site mutants, the catalytic activity and dimeric structure of the surface mutants, C84R and G89V, were found to be similar to the wild-type enzyme. Taken together, our results suggest that the pathophysiology of inherited axonal neuropathy with neuromyotonia can be induced by conversion of HINT1 from a homodimer to monomer, by modification of select surface residues or by a significant reduction of the enzyme's catalytic efficiency.     

Phenotypic Interactions among Mutations in a Thermus thermophilus 16S rRNA Gene Detected with Genetic Selections and Experimental Evolution

During protein synthesis, the ribosome undergoes conformational transitions between functional states, requiring communication between distant structural elements of the ribosome. Despite advances in ribosome structural biology, identifying the protein and rRNA residues governing these transitions remains a significant challenge. Such residues can potentially be identified genetically, given the predicted deleterious effects of mutations stabilizing the ribosome in discrete conformations and the expected ameliorating effects of second-site compensatory mutations. In this study, we employed genetic selections and experimental evolution to identify interacting mutations in the ribosome of the thermophilic bacterium Thermus thermophilus. By direct genetic selections, we identified mutations in 16S rRNA conferring a streptomycin dependence phenotype and from these derived second-site suppressor mutations relieving dependence. Using experimental evolution of streptomycin-independent pseudorevertants, we identified additional compensating mutations. Similar mutations could be evolved from slow-growing streptomycin-resistant mutants. While some mutations arose close to the site of the original mutation in the three-dimensional structure of the 30S ribosomal subunit and probably act directly by compensating for local structural distortions, the locations of others are consistent with long-range communication between specific structural elements within the ribosome.     

NUDT15 polymorphism and NT5C2 and PRPS1 mutations influence thiopurine sensitivity in acute lymphoblastic leukaemia cells

In chemotherapy for childhood acute lymphoblastic leukaemia (ALL), maintenance therapy consisting of oral daily mercaptopurine and weekly methotrexate is important. NUDT15 variant genotype is reportedly highly associated with severe myelosuppression during maintenance therapy, particularly in Asian and Hispanic populations. It has also been demonstrated that acquired somatic mutations of the NT5C2 and PRPS1 genes, which are involved in thiopurine metabolism, are detectable in a portion of relapsed childhood ALL. To directly confirm the significance of the NUDT15 variant genotype and NT5C2 and PRPS1 mutations in thiopurine sensitivity of leukaemia cells in the intrinsic genes, we investigated 84 B-cell precursor-ALL (BCP-ALL) cell lines. Three and 14 cell lines had homozygous and heterozygous variant diplotypes of the NUDT15 gene, respectively, while 4 and 2 cell lines that were exclusively established from the samples at relapse had the NT5C2 and PRPS1 mutations, respectively. Both NUDT15 variant genotype and NT5C2 and PRPS1 mutations were significantly associated with DNA-incorporated thioguanine levels after exposure to thioguanine at therapeutic concentration. Considering the continuous exposure during the maintenance therapy, we evaluated in vitro mercaptopurine sensitivity after 7-day exposure. Mercaptopurine concentrations lethal to 50% of the leukaemia cells were comparable to therapeutic serum concentration of mercaptopurine. Both NUDT15 variant genotype and NT5C2 and PRPS1 mutations were significantly associated with mercaptopurine sensitivity in 83 BCP-ALL and 23 T-ALL cell lines. The present study provides direct evidence to support the general principle showing that both inherited genotype and somatically acquired mutation are crucially implicated in the drug sensitivity of leukaemia cells.     

Thermostable Bacillus subtilis lipases: in vitro evolution and structural insight

In vitro evolution methods are now being routinely used to identify protein variants with novel and enhanced properties that are difficult to achieve using rational design. However, one of the limitations is in screening for beneficial mutants through several generations due to the occurrence of neutral/negative mutations occurring in the background of positive ones. While evolving a lipase in vitro from mesophilic Bacillus subtilis to generate thermostable variants, we have designed protocols that combine stringent three-tier testing, sequencing and stability assessments on the protein at the end of each generation. This strategy resulted in a total of six stabilizing mutations in just two generations with three mutations per generation. Each of the six mutants when evaluated individually contributed additively to thermostability. A combination of all of them resulted in the best variant that shows a remarkable 15 °C shift in melting temperature and a millionfold decrease in the thermal inactivation rate with only a marginal increase of 3 kcal mol⁻¹ in free energy of stabilization. Notably, in addition to the dramatic shift in optimum temperature by 20 °C, the activity has increased two- to fivefold in the temperature range 25-65 °C. High-resolution crystal structures of three of the mutants, each with 5° increments in melting temperature, reveal the structural basis of these mutations in attaining higher thermostability. The structures highlight the importance of water-mediated ionic networks on the protein surface in imparting thermostability. Saturation mutagenesis at each of the six positions did not result in enhanced thermostability in almost all the cases, confirming the crucial role played by each mutation as revealed through the structural study. Overall, our study presents an efficient strategy that can be employed in directed evolution approaches employed for obtaining improved properties of proteins.     

The Role of Protein Denaturation Energetics and Molecular Chaperones in the Aggregation and Mistargeting of Mutants Causing Primary Hyperoxaluria Type I

Primary hyperoxaluria type I (PH1) is a conformational disease which result in the loss of alanine:glyoxylate aminotransferase (AGT) function. The study of AGT has important implications for protein folding and trafficking because PH1 mutants may cause protein aggregation and mitochondrial mistargeting. We herein describe a multidisciplinary study aimed to understand the molecular basis of protein aggregation and mistargeting in PH1 by studying twelve AGT variants. Expression studies in cell cultures reveal strong protein folding defects in PH1 causing mutants leading to enhanced aggregation, and in two cases, mitochondrial mistargeting. Immunoprecipitation studies in a cell-free system reveal that most mutants enhance the interactions with Hsc70 chaperones along their folding process, while in vitro binding experiments show no changes in the interaction of folded AGT dimers with the peroxisomal receptor Pex5p. Thermal denaturation studies by calorimetry support that PH1 causing mutants often kinetically destabilize the folded apo-protein through significant changes in the denaturation free energy barrier, whereas coenzyme binding overcomes this destabilization. Modeling of the mutations on a 1.9 Å crystal structure suggests that PH1 causing mutants perturb locally the native structure. Our work support that a misbalance between denaturation energetics and interactions with chaperones underlie aggregation and mistargeting in PH1, suggesting that native state stabilizers and protein homeostasis modulators are potential drugs to restore the complex and delicate balance of AGT protein homeostasis in PH1.     

NMR structures of two variants of bovine pancreatic trypsin inhibitor (BPTI) reveal unexpected influence of mutations on protein structure and stability

Here we determined NMR solution structures of two mutants of bovine pancreatic trypsin inhibitor (BPTI) to reveal structural reasons of their decreased thermodynamic stability. A point mutation, A16V, in the solvent-exposed loop destabilizes the protein by 20 degrees C, in contrast to marginal destabilization observed for G, S, R, L or W mutants. In the second mutant introduction of eight alanine residues at proteinase-contacting sites (residues 11, 13, 17, 18, 19, 34, 37 and 39) provides a protein that denatures at a temperature about 30 degrees C higher than expected from additive behavior of individual mutations. In order to efficiently determine structures of these variants, we applied a procedure that allows us to share data between regions unaffected by mutation(s). NOAH/DYANA and CNS programs were used for a rapid assignment of NOESY cross-peaks, structure calculations and refinement. The solution structure of the A16V mutant reveals no conformational change within the molecule, but shows close contacts between V16, I18 and G36/G37. Thus, the observed 4.3kcal/mol decrease of stability results from a strained local conformation of these residues caused by introduction of a beta-branched Val side-chain. Contrary to the A16V mutation, introduction of eight alanine residues produces significant conformational changes, manifested in over a 9A shift of the Y35 side-chain. This structural rearrangement provides about 6kcal/mol non-additive stabilization energy, compared to the mutant in which G37 and R39 are not mutated to alanine residues.     

Structure of the periplasmic stress response protein CpxP

CpxP is a novel bacterial periplasmic protein with no homologues of known function. In Gram-negative enteric bacteria, CpxP is thought to interact with the two-component sensor kinase, CpxA, to inhibit induction of the Cpx envelope stress response in the absence of protein misfolding. CpxP has also been shown to facilitate DegP-mediated proteolysis of misfolded proteins. Six mutations that negate the ability of CpxP to function as a signaling protein are localized in or near two conserved LTXXQ motifs that define a class of proteins with similarity to CpxP, Pfam PF07813. To gain insight into how these mutations might affect CpxP signaling and/or proteolytic adaptor functions, the crystal structure of CpxP from Escherichia coli was determined to 2.85-Å resolution. The structure revealed an antiparallel dimer of intertwined α-helices with a highly basic concave surface. Each protomer consists of a long, hooked and bent hairpin fold, with the conserved LTXXQ motifs forming two diverging turns at one end. Biochemical studies demonstrated that CpxP maintains a dimeric state but may undergo a slight structural adjustment in response to the inducing cue, alkaline pH. Three of the six previously characterized cpxP loss-of-function mutations, M59T, Q55P, and Q128H, likely result from a destabilization of the protein fold, whereas the R60Q, D61E, and D61V mutations may alter intermolecular interactions.