Identifying interacting residues using Boolean Learning and Support Vector Machines: case study on mRFP and DsRed proteins

In a protein, interactions exist between amino acid residues that influence the protein's structural integrity or stability and thus affect its catalytic function. The loss of this interaction due to mutations in these amino acids usually leads to a non-functional protein. Probing the sequence space of a protein through mutations or recombinations, as performed in directed evolution to search for an improved variant, frequently results in such inactive sequences. In this work, we demonstrate the use of machine learning to identify such interacting residues and the use of template engineering strategies to increase the fraction of active variants in a library. We show that using the sequences from recombination of monomeric red fluorescent protein (mRFP) and Discosoma red fluorescent protein (DsRed), we were able to identify a pair of interacting residues using an algorithm based on Boolean Learning and Support Vector Machines. The interaction between the identified residues was verified through point mutations on the mRFP and DsRed genes. We also show that it is possible to use such results to alter the parental genes such that the probability of disrupting the important interactions is minimized. This will result in a larger fraction of active variants in the recombinant library and allow us to access more functional space. We demonstrate this effect by comparing the recombinant library of wild-type (WT) DsRed, mRFP and an altered sequence of DsRed with mRFP WT genes.     

Crystal structure of the Melampsora lini effector AvrP reveals insights into a possible nuclear function and recognition by the flax disease resistance protein P

The effector protein AvrP is secreted by the flax rust fungal pathogen (Melampsora lini) and recognized specifically by the flax (Linum usitatissimum) P disease resistance protein, leading to effector‐triggered immunity. To investigate the biological function of this effector and the mechanisms of specific recognition by the P resistance protein, we determined the crystal structure of AvrP. The structure reveals an elongated zinc‐finger‐like structure with a novel interleaved zinc‐binding topology. The residues responsible for zinc binding are conserved in AvrP effector variants and mutations of these motifs result in a loss of P‐mediated recognition. The first zinc‐coordinating region of the structure displays a positively charged surface and shows some limited similarities to nucleic acid‐binding and chromatin‐associated proteins. We show that the majority of the AvrP protein accumulates in the plant nucleus when transiently expressed in Nicotiana benthamiana cells, suggesting a nuclear pathogenic function. Polymorphic residues in AvrP and its allelic variants map to the protein surface and could be associated with differences in recognition specificity. Several point mutations of residues on the non‐conserved surface patch result in a loss of recognition by P, suggesting that these residues are required for recognition.     

PTPN11 (Shp2) mutations in LEOPARD syndrome have dominant negative, not activating, effects

Multiple lentigines/LEOPARD syndrome (LS) is a rare, autosomal dominant disorder characterized by Lentigines, Electrocardiogram abnormalities, Ocular hypertelorism, Pulmonic valvular stenosis, Abnormalities of genitalia, Retardation of growth, and Deafness. Like the more common Noonan syndrome (NS), LS is caused by germ line missense mutations in PTPN11, encoding the protein-tyrosine phosphatase Shp2. Enzymologic, structural, cell biological, and mouse genetic studies indicate that NS is caused by gain-of-function PTPN11 mutations. Because NS and LS share several features, LS has been viewed as an NS variant. We examined a panel of LS mutants, including the two most common alleles. Surprisingly, we found that in marked contrast to NS, LS mutants are catalytically defective and act as dominant negative mutations that interfere with growth factor/Erk-mitogen-activated protein kinase-mediated signaling. Molecular modeling and biochemical studies suggest that LS mutations contort the Shp2 catalytic domain and result in open, inactive forms of Shp2. Our results establish that the pathogenesis of LS and NS is distinct and suggest that these disorders should be distinguished by mutational analysis rather than clinical presentation.     

Sequence determinants of thermodynamic stability in a WW domain—An all‐β‐sheet protein

The stabilities of 66 sequence variants of the human Pin1 WW domain have been determined by equilibrium thermal denaturation experiments. All 34 residues composing the hPin1 WW three‐stranded β‐sheet structure could be replaced one at a time with at least one different natural or non‐natural amino acid residue without leading to an unfolded protein. Alanine substitutions at only four positions within the hPin1 WW domain lead to a partially or completely unfolded protein—in the absence of a physiological ligand. The side chains of these four residues form a conserved, partially solvent‐inaccessible, continuous hydrophobic minicore comprising the N‐ and C‐termini. Ala mutations at five other residues, three of which constitute the ligand binding patch on the concave side of the β‐sheet, significantly destabilize the hPin1 WW domain without leading to an unfolded protein. The remaining mutations affect protein stability only slightly, suggesting that only a small subset of side chain interactions within the hPin1 WW domain are mandatory for acquiring and maintaining a stable, cooperatively folded β‐sheet structure.     

Characterisation of the BRCT domains of the breast cancer susceptibility gene product BRCA1

The breast cancer susceptibility gene product BRCA1 is a tumour suppressor but the biochemical and biological functions that underlie its role in carcinogenesis remain to be determined. Here, we characterise the solution properties of the highly conserved C terminus of BRCA1, consisting of a tandem repeat of the BRCT domain (BRCT-tan), that plays a critical role in BRCA1-mediated tumour suppression. The overall free energy of unfolding of BRCT-tan is high (14.2 kcal mol(-1) at 20 degrees C in water) but unfolding occurs via an aggregation-prone, partly folded intermediate. A representative set of cancer-associated sequence variants was constructed and the effects on protein stability were measured. All of the mutations were highly destabilising and they would be expected to cause loss of function for this reason. Over half could not be purified in a soluble form, indicating that these residues are critical for maintaining structural integrity. The remaining mutants exhibited much greater aggregation propensities than the wild-type, which is most likely a consequence of their reduced thermodynamic stability relative to the partly folded intermediate. The mutations characterised here are located at different sites in the BRCT-tan structure that do not explain fully their effects on the protein's stability. Thus, the results indicate an important role for biophysical studies in assessing the significance of sequence variants and in determining how they cause disease.     

Reconstruction of mycobacterial dehalogenase Rv2579 by cumulative mutagenesis of haloalkane dehalogenase LinB

The homology model of protein Rv2579 from Mycobacterium tuberculosis H37Rv was compared with the crystal structure of haloalkane dehalogenase LinB from Sphingomonas paucimobilis UT26, and this analysis revealed that 6 of 19 amino acid residues which form an active site and entrance tunnel are different in LinB and Rv2579. To characterize the effect of replacement of these six amino acid residues, mutations were introduced cumulatively into the six amino acid residues of LinB. The sixfold mutant, which was supposed to have the active site of Rv2579, exhibited haloalkane dehalogenase activity with the haloalkanes tested, confirming that Rv2579 is a member of the haloalkane dehalogenase protein family.     

Folding and aggregation are selectively influenced by the conformational preferences of the alpha-helices of muscle acylphosphatase

The native state of human muscle acylphosphatase (AcP) presents two alpha-helices. In this study we have investigated folding and aggregation of a number of protein variants having mutations aimed at changing the propensity of these helical regions. Equilibrium and kinetic measurements of folding indicate that only helix-2, spanning residues 55-67, is largely stabilized in the transition state for folding therefore playing a relevant role in this process. On the contrary, the aggregation rate appears to vary only for the variants in which the propensity of the region corresponding to helix-1, spanning residues 22-32, is changed. Mutations that stabilize the first helix slow down the aggregation process while those that destabilize it increase the aggregation rate. AcP variants with the first helix destabilized aggregate with rates increased to different extents depending on whether the introduced mutations also alter the propensity to form beta-sheet structure. The fact that the first alpha-helix is important for aggregation and the second helix is important for folding indicates that these processes are highly specific. This partitioning does not reflect the difference in intrinsic alpha-helical propensities of the two helices, because helix-1 is the one presenting the highest propensity. Both processes of folding and aggregation do not therefore initiate from regions that have simply secondary structure propensities favorable for such processes. The identification of the regions involved in aggregation and the understanding of the factors that promote such a process are of fundamental importance to elucidate the principles by which proteins have evolved and for successful protein design.     

Toward Classification of BRCA1 Missense Variants Using a Biophysical Approach

Carriers of germ line mutations in breast cancer susceptibility gene BRCA1 have an increased risk of developing breast and ovarian cancers; missense mutations have, however, been difficult to assess for disease association. Here we have used a biophysical approach to classify these variants. We established an assay for measuring the thermodynamic stability of the BRCA1 BRCT domains and investigated the effects of 36 missense mutations. The mutations show a range of effects. Some do not change the stability, whereas others destabilize the protein by as much as 6 kcal mol⁻¹; one-third of the mutants could not be expressed in soluble form in Escherichia coli, and we conclude that these destabilize the protein by an even greater amount. We tested several computer algorithms for their ability to predict the mutant effects and found that by grouping them into two classes (destabilizing by less than or more than 2.2 kcal mol⁻¹), the algorithms could predict the stability changes. Importantly, with the exception of the few mutants located in the binding site, none showed a significant reduction in affinity for phosphorylated substrate. These results indicate that despite very large losses in stability, the integrity of the structure is not compromised by the mutations. Thus, the majority of mutations cause loss of function by reducing the proportion of BRCA1 molecules that are in the folded state and increasing the proportion of molecules that are unfolded. Consequently, small molecule stabilization of the structure could be a generally applicable preventative therapeutic strategy for rescuing many BRCA1 mutations.     

Analysis of allosteric effector binding sites of potato ADP-glucose pyrophosphorylase through reverse genetics

ADP-glucose pyrophosphorylase (AGPase) is a key regulatory enzyme of bacterial glycogen and plant starch synthesis as it controls carbon flux via its allosteric regulatory behavior. Unlike the bacterial enzyme that is composed of a single subunit type, the plant AGPase is a heterotetrameric enzyme (alpha2beta2) with distinct roles for each subunit type. The large subunit (LS) is involved mainly in allosteric regulation through its interaction with the catalytic small subunit (SS). The LS modulates the catalytic activity of the SS by increasing the allosteric regulatory response of the hetero-oligomeric enzyme. To identify regions of the LS involved in binding of effector molecules, a reverse genetics approach was employed. A potato (Solanum tuberosum L.) AGPase LS down-regulatory mutant (E38A) was subjected to random mutagenesis using error-prone polymerase chain reaction and screened for the capacity to form an enzyme capable of restoring glycogen production in glgC(-) Escherichia coli. Dominant mutations were identified by their capacity to restore glycogen production when the LS containing only the second site mutations was co-expressed with the wild-type SS. Sequence analysis showed that most of the mutations were decidedly nonrandom and were clustered at conserved N- and C-terminal regions. Kinetic analysis of the dominant mutant enzymes indicated that the K(m) values for cofactor and substrates were comparable with the wild-type AGPase, whereas the affinities for activator and inhibitor were altered appreciably. These AGPase variants displayed increased resistance to P(i) inhibition and/or greater sensitivity toward 3-phosphoglyceric acid activation. Further studies of Lys-197, Pro-261, and Lys-420, residues conserved in AGPase sequences, by site-directed mutagenesis suggested that the effectors 3-phosphoglyceric acid and P(i) interact at two closely located binding sites.     

Examining protein surface structure in highly conserved sequence variants with mass spectrometry

A simple mass spectrometry-based method capable of examining protein structure called SNAPP (selective noncovalent adduct protein probing) is used to evaluate the structural consequences of point mutations in naturally occurring sequence variants from different species. SNAPP monitors changes in the attachment of noncovalent adducts to proteins as a function of structural state. Mutations that lead to perturbations to the electrostatic surface structure of a protein affect noncovalent attachment and are easily observed with SNAPP. Mutations that do not alter the tertiary structure or electrostatic surface structure yield similar results by SNAPP. For example, bovine, porcine, and human insulin all have very similar backbone structures and no basic or acidic residue mutations, and the SNAPP distributions for all three proteins are very similar. In contrast, four variants of cytochrome c (cytc) have varying degrees of sequence homology, which are reflected in the observed SNAPP distributions. Bovine and pigeon cytc have several basic or acidic residue substitutions relative to horse cytc, but the SNAPP distributions for all three proteins are similar. This suggests that these mutations do not significantly influence the protein surface structure. On the other hand, yeast cytc has the least sequence homology and exhibits a unique, though related, SNAPP distribution. Even greater differences are observed for lysozyme. Hen and human lysozyme have identical tertiary structures but significant variations in the locations of numerous basic and acidic residues. The SNAPP distributions are quite distinct for the two forms of lysozyme, suggesting significant differences in the surface structures. In summary, SNAPP experiments are relatively easy to perform, require minimal sample consumption, and provide a facile route for comparison of protein surface structure between highly homologous proteins.     

Modelling the structure of the fusion protein from human respiratory syncytial virus

The fusion protein of respiratory syncytial virus (RSV-F) is responsible for fusion of virion with host cells and infection of neighbouring cells through the formation of syncytia. A three-dimensional model structure of RSV-F was derived by homology modelling from the structure of the equivalent protein in Newcastle disease virus (NDV). Despite very low sequence homology between the two structures, most features of the model appear to have high credibility, although a few small regions in RSV-F whose secondary structure is predicted to be different to that in NDV are likely to be poorly modelled. The organization of individual residues identified in escape mutants against monoclonal antibodies correlates well with known antigenic sites. The location of residues involved in point mutations in several drug-resistant variants is also examined.     

Determinants of the factor IX mutational spectrum in haemophilia B: an analysis of missense mutations using a multi-domain molecular model of the activated protein

A multi-domain molecular model of factor IXa was constructed by comparative methods. The quaternary structure of the protein was assembled by docking individual domains through consideration of their shape complementarity, polaric properties and the location of cross-reacting material positive/negative (CRM^+/–) variants on domain surfaces. Some 217 different missense mutations in the factor IX (F9) gene were then selected for study. Using maximum likelihood analysis, missense mutations affecting highly conserved amino acid residues of factor IX were shown to be 15–20 times more likely to result in haemophilia B than those affecting non-conserved residues. However, about one quarter of this increase in likelihood of clinical observation could be attributed to the magnitude of the amino acid exchange. Missense mutations in structurally conserved residues were found to be 2.1-fold more likely to come to clinical attention than those in structurally variable residues. Missense mutations in residues whose side chains were inwardly pointing were 3.6-fold more likely to be observed than those in surface residues. These observations imply a complex hierarchy of sequence/structure conservation in the protein. The severity of the clinical phenotype correlated with both the extent of the evolutionary sequence conservation of the residue at the site of mutation and the magnitude of the amino acid exchange. Further, the substitution of residues exhibiting minimal side chain solvent accessibility was associated disproportionately with severe haemophilia compared with that of surface residues. Clusters of CRM⁺ mutations were observed at factor IX-specific residues on the surface of the molecule. These clusters may reflect factor IX-specific docking interactions. The likelihood that a given factor IX mutation will come to clinical attention is therefore a complex function of the sequence characteristics of the F9 gene, the nature of the amino acid substitution, its precise location and immediate environment within the protein molecule, and its resulting effects on the structure and function of the protein.This paper is dedicated to the memory of Andrew Wacey     

News from the Protein Mutability Landscape

Some mutations of protein residues matter more than others, and these are often conserved evolutionarily. The explosion of deep sequencing and genotyping increasingly requires the distinction between effect and neutral variants. The simplest approach predicts all mutations of conserved residues to have an effect; however, this works poorly, at best. Many computational tools that are optimized to predict the impact of point mutations provide more detail. Here, we expand the perspective from the view of single variants to the level of sketching the entire mutability landscape. This landscape is defined by the impact of substituting every residue at each position in a protein by each of the 19 non-native amino acids. We review some of the powerful conclusions about protein function, stability and their robustness to mutation that can be drawn from such an analysis. Large-scale experimental and computational mutagenesis experiments are increasingly furthering our understanding of protein function and of the genotype–phenotype associations. We also discuss how these can be used to improve predictions of protein function and pathogenicity of missense variants.     

The study of the role of mutations M182T and Q39K in the TEM-72 β-lactamase structure by the molecular dynamics method

Synthesis of β-lactamases is one of the common mechanisms of bacterial resistance to β-lactam antibiotics such as penicillins and cephalosporins. The widespread use of antibiotics resulted in appearance of numerous extended-spectrum β-lactamase variants or inhibitor-resistant β-lactamases. In TEM type β-lactamases mutations of 92 residues have been described. Several mutations are functionally important and they determine the extended substrate specificity. However, roles of the most so-called associated mutations, located far from the active site, remain unknown. We have investigated the role of associated mutations in structure of β-lactamase TEM-72, which contains two key mutations (G238S, E240K) and two associated mutations (Q39K, M182T) by means of molecular dynamics simulation. Appearance of the key mutations (in 238 and 240 positions) caused destabilization of the protein globule, characterized by increased mobility of amino acid residues. Associated mutations (Q39K, M182T) exhibited opposite effect on the protein structure. The mutation M182T stabilized, while the mutation Q39K destabilized the protein. It appears that the latter mutation promoted optimization of the conformational mobility of β-lactamase and may influence the enzyme activity.     

Investigation of the Interaction between the Large and Small Subunits of Potato ADP-Glucose Pyrophosphorylase

ADP-glucose pyrophosphorylase (AGPase), a key allosteric enzyme involved in higher plant starch biosynthesis, is composed of pairs of large (LS) and small subunits (SS). Current evidence indicates that the two subunit types play distinct roles in enzyme function. Recently the heterotetrameric structure of potato AGPase has been modeled. In the current study, we have applied the molecular mechanics generalized born surface area (MM-GBSA) method and identified critical amino acids of the potato AGPase LS and SS subunits that interact with each other during the native heterotetrameric structure formation. We have further shown the role of the LS amino acids in subunit-subunit interaction by yeast two-hybrid, bacterial complementation assay and native gel. Comparison of the computational results with the experiments has indicated that the backbone energy contribution (rather than the side chain energies) of the interface residues is more important in identifying critical residues. We have found that lateral interaction of the LS-SS is much stronger than the longitudinal one, and it is mainly mediated by hydrophobic interactions. This study will not only enhance our understanding of the interaction between the SS and the LS of AGPase, but will also enable us to engineer proteins to obtain better assembled variants of AGPase which can be used for the improvement of plant yield.     

Reconstruction of Mycobacterial Dehalogenase Rv2579 by Cumulative Mutagenesis of Haloalkane Dehalogenase LinB

The homology model of protein Rv2579 from Mycobacterium tuberculosis H37Rv was compared with the crystal structure of haloalkane dehalogenase LinB from Sphingomonas paucimobilis UT26, and this analysis revealed that 6 of 19 amino acid residues which form an active site and entrance tunnel are different in LinB and Rv2579. To characterize the effect of replacement of these six amino acid residues, mutations were introduced cumulatively into the six amino acid residues of LinB. The sixfold mutant, which was supposed to have the active site of Rv2579, exhibited haloalkane dehalogenase activity with the haloalkanes tested, confirming that Rv2579 is a member of the haloalkane dehalogenase protein family.     

The zinc finger motif of Escherichia coli RecQ is implicated in both DNA binding and protein folding

The RecQ family of DNA helicases has been shown to be important for the maintenance of genomic integrity. Mutations in human RecQ genes lead to genomic instability and cancer. Several RecQ family of helicases contain a putative zinc finger motif of the C4 type at the C terminus that has been identified in the crystalline structure of RecQ helicase from Escherichia coli. To better understand the role of this motif in helicase from E. coli, we constructed a series of single mutations altering the conserved cysteines as well as other highly conserved residues. All of the resulting mutant proteins exhibited a high level of susceptibility to degradation, making functional analysis impossible. In contrast, a double mutant protein in which both cysteine residues Cys397 and Cys400 in the zinc finger motif were replaced by asparagine residues was purified to homogeneity. Slight local conformational changes were detected, but the rest of the mutant protein has a well defined tertiary structure. Furthermore, the mutant enzyme displayed ATP binding affinity similar to the wild-type enzyme but was severely impaired in DNA binding and in subsequent ATPase and helicase activities. These results revealed that the zinc finger binding motif is involved in maintaining the integrity of the whole protein as well as DNA binding. We also showed that the zinc atom is not essential to enzymatic activity.     

The fibrillogenic L178H variant of apolipoprotein A-I forms helical fibrils

A number of amyloidogenic variants of apoA-I have been discovered but most have not been analyzed. Previously, we showed that the G26R mutation of apoA-I leads to increased β-strand structure, increased N-terminal protease susceptibility, and increased fibril formation after several days of incubation. In vivo, this and other variants mutated in the N-terminal domain (residues 26 to ～90) lead to renal and hepatic accumulation. In contrast, several mutations identified within residues 170 to 178 lead to cardiac, laryngeal, and cutaneous protein deposition. Here, we describe the structural changes in the fibrillogenic variant L178H. Like G26R, the initial structure of the protein exhibits altered tertiary conformation relative to wild-type protein along with decreased stability and an altered lipid binding profile. However, in contrast to G26R, L178H undergoes an increase in helical structure upon incubation at 37°C with a half time (t(1/2)) of about 12 days. Upon prolonged incubation, the L178H mutant forms fibrils of a diameter of 10 nm that ranges in length from 30 to 120 nm. These results show that apoA-I, known for its dynamic properties, has the ability to form multiple fibrillar conformations, which may play a role in the tissue-specific deposition of the individual variants.     

Polygalacturonase-inhibiting protein interacts with pectin through a binding site formed by four clustered residues of arginine and lysine

Polygalacturonase-inhibiting protein (PGIP) is a cell wall protein that inhibits fungal polygalacturonases (PGs) and retards the invasion of plant tissues by phytopathogenic fungi. Here, we report the interaction of two PGIP isoforms from Phaseolus vulgaris (PvPGIP1 and PvPGIP2) with both polygalacturonic acid and cell wall fractions containing uronic acids. We identify in the three-dimensional structure of PvPGIP2 a motif of four clustered arginine and lysine residues (R183, R206, K230, and R252) responsible for this binding. The four residues were mutated and the protein variants were expressed in Pichia pastoris. The ability of both wild-type and mutated proteins to bind pectins was investigated by affinity chromatography. Single mutations impaired the binding and double mutations abolished the interaction, thus indicating that the four clustered residues form the pectin-binding site. Remarkably, the binding of PGIP to pectin is displaced in vitro by PGs, suggesting that PGIP interacts with pectin and PGs through overlapping although not identical regions. The specific interaction of PGIP with polygalacturonic acid may be strategic to protect pectins from the degrading activity of fungal PGs.     

Deep mutational scanning of an RRM domain of the Saccharomyces cerevisiae poly(A)-binding protein

The RNA recognition motif (RRM) is the most common RNA-binding domain in eukaryotes. Differences in RRM sequences dictate, in part, both RNA and protein-binding specificities and affinities. We used a deep mutational scanning approach to study the sequence-function relationship of the RRM2 domain of the Saccharomyces cerevisiae poly(A)-binding protein (Pab1). By scoring the activity of more than 100,000 unique Pab1 variants, including 1246 with single amino acid substitutions, we delineated the mutational constraints on each residue. Clustering of residues with similar mutational patterns reveals three major classes, composed principally of RNA-binding residues, of hydrophobic core residues, and of the remaining residues. The first class also includes a highly conserved residue not involved in RNA binding, G150, which can be mutated to destabilize Pab1. A comparison of the mutational sensitivity of yeast Pab1 residues to their evolutionary conservation reveals that most residues tolerate more substitutions than are present in the natural sequences, although other residues that tolerate fewer substitutions may point to specialized functions in yeast. An analysis of ∼40,000 double mutants indicates a preference for a short distance between two mutations that display an epistatic interaction. As examples of interactions, the mutations N139T, N139S, and I157L suppress other mutations that interfere with RNA binding and protein stability. Overall, this study demonstrates that living cells can be subjected to a single assay to analyze hundreds of thousands of protein variants in parallel.