Neighborhood properties are important determinants of temperature sensitive mutations

Temperature-sensitive (TS) mutants are powerful tools to study gene function in vivo. These mutants exhibit wild-type activity at permissive temperatures and reduced activity at restrictive temperatures. Although random mutagenesis can be used to generate TS mutants, the procedure is laborious and unfeasible in multicellular organisms. Further, the underlying molecular mechanisms of the TS phenotype are poorly understood. To elucidate TS mechanisms, we used a machine learning method-logistic regression-to investigate a large number of sequence and structure features. We developed and tested 133 features, describing properties of either the mutation site or the mutation site neighborhood. We defined three types of neighborhood using sequence distance, Euclidean distance, and topological distance. We discovered that neighborhood features outperformed mutation site features in predicting TS mutations. The most predictive features suggest that TS mutations tend to occur at buried and rigid residues, and are located at conserved protein domains. The environment of a buried residue often determines the overall structural stability of a protein, thus may lead to reversible activity change upon temperature switch. We developed TS prediction models based on logistic regression and the Lasso regularized procedure. Through a ten-fold cross-validation, we obtained the area under the curve of 0.91 for the model using both sequence and structure features. Testing on independent datasets suggested that the model predicted TS mutations with a 50% precision. In summary, our study elucidated the molecular basis of TS mutants and suggested the importance of neighborhood properties in determining TS mutations. We further developed models to predict TS mutations derived from single amino acid substitutions. In this way, TS mutants can be efficiently obtained through experimentally introducing the predicted mutations.     

Analysis and prediction of the location of catalytic residues in enzymes

The catalytic residues of an enzyme are defined as the amino acids directly involved in chemical catalysis. They mainly act as a general acid--base, electrophilic or nucleophilic catalyst or they polarize and stabilize the transition state. An analysis of the structural features of 36 catalytic residues in 17 enzymes of known structure and with defined mechanism is reported. Residues that bind metal ions (Zn2+ and Cu2+) are considered separately. The features examined are: residue type, location in secondary structure, separation between the residues, accessibility to solvent, intra-protein electrostatic interactions, mobility as evaluated from crystallographic temperature factors, polarity of the environment and the sequence conservation between homologous enzymes of residues that were sequentially or spatially close to the catalytic residue. In general the environment of catalytic residues is similar to that of polar side chains that have low accessibility to solvent. Two algorithms have been developed to identify probable catalytic residues. Scanning an alignment of homologous enzyme sequences for peaks of sequence conservation identifies 13 out of the 16 catalytic residues with 50 residues overpredicted. When the conservation of the spatially close residues is used instead, a different set of 13 residues are identified with 47 residues overpredicted. A combination of the two algorithms identifies 11 residues with 36 residues overpredicted.     

Charge Composition Features of Model Single-span Membrane Proteins That Determine Selection of YidC and SecYEG Translocase Pathways in Escherichia coli

We have investigated the features of single-span model membrane proteins based upon leader peptidase that determines whether the proteins insert by a YidC/Sec-independent, YidC-only, or YidC/Sec mechanism. We find that a protein with a highly hydrophobic transmembrane segment that inserts into the membrane by a YidC/Sec-independent mechanism becomes YidC-dependent if negatively charged residues are inserted into the translocated periplasmic domain or if the hydrophobicity of the transmembrane segment is reduced by substituting polar residues for nonpolar ones. This suggests that charged residues in the translocated domain and the hydrophobicity within the transmembrane segment are important determinants of the insertion pathway. Strikingly, the addition of a positively charged residue to either the translocated region or the transmembrane region can switch the insertion requirements such that insertion requires both YidC and SecYEG. To test conclusions from the model protein studies, we confirmed that a positively charged residue is a SecYEG determinant for the endogenous proteins ATP synthase subunits a and b and the TatC subunit of the Tat translocase. These findings provide deeper insights into how pathways are selected for the insertion of proteins into the Escherichia coli inner membrane.     

Molecular packing in the feather keratin filament

Avian feathers have a filament-matrix texture and X-ray diffraction studies show that the filament has a helical structure with four repeating units per turn. Each repeating unit consists of a pair of twisted beta-sheets related by a perpendicular diad, and the twist in the sheets is of opposite hand to that of the helix. Each sheet is believed to comprise a 32-residue segment of the feather keratin molecule, which contains around 100 residues, the remainder constituting the matrix. In the present contribution, the sequence of emu feather is mapped to the low-resolution model derived earlier from X-ray studies. This shows that the inner surface of the "beta-sandwich" is densely populated by hydrophobic residues and that the charged residues and cysteine residues lie on the outer surface. In addition, the inner residues in the repeating unit mesh neatly together in layers oriented perpendicular to the filament axis. Amino acid sequences from a range of avian and reptilian keratins were collected and a 32-residue segment corresponding to the filament framework could be identified in every case, supporting the notion that there is a common plan for the filament framework in all of these materials. The hairpin turns in the beta-sheet were also identified and shown to be unusually rich in proline residues and also of variable composition. Two variants of the mapping were found which have complimentary conformations of the hairpin turns and these are illustrated and discussed. Since feather keratin yields a fiber rather than a crystalline X-ray pattern refinement of the model is restricted to trial-and-error methods and the assumptions made in its derivation are critically examined and some possible modifications discussed.     

Evolutionary conservation of RecA genes in relation to protein structure and function.

Functional and structural regions inferred from the Escherichia coli R ecA protein crystal structure and mutation studies are evaluated in terms of evolutionary conservation across 63 RecA eubacterial sequences. Two paramount segments invariant in specific amino acids correspond to the ATP-binding A site and the functionally unassigned segment from residues 145 to 149 immediately carboxyl to the ATP hydrolysis B site. Not only are residues 145 to 149 conserved individually, but also all three-dimensional structural neighbors of these residues are invariant, strongly attesting to the functional or structural importance of this segment. The conservation of charged residues at the monomer-monomer interface, emphasizing basic residues on one surface and acidic residues on the other, suggests that RecA monomer polymerization is substantially mediated by electrostatic interactions. Different patterns of conservation also allow determination of regions proposed to interact with DNA, of LexA binding sites, and of filament-filament contact regions. Amino acid conservation is also compared with activities and properties of certain RecA protein mutants. Arginine 243 and its strongly cationic structural environment are proposed as the major site of competition for DNA and LexA binding to RecA. The conserved acidic and glycine residues of the disordered loop L1 and its proximity to the RecA acidic monomer interface suggest its involvement in monomer-monomer interactions rather than DNA binding. The conservation of various RecA positions and regions suggests a model for RecA-double-stranded DNA interaction and other functional and structural assignments.     

Direct evidence that the carboxyl-terminal sequence of a bacterial chemoreceptor is an unstructured linker and enzyme tether

Sensory adaptation in bacterial chemotaxis involves reversible methylation of specific glutamyl residues on chemoreceptors. The reactions are catalyzed by a dedicated methyltransferase and dedicated methylesterase. In Escherichia coli and related organisms, control of these enzymes includes an evolutionarily recent addition of interaction with a pentapeptide activator located at the carboxyl terminus of the receptor polypeptide chain. Effective enzyme activation requires not only the pentapeptide but also a segment of the receptor polypeptide chain between that sequence and the coiled-coil body of the chemoreceptor. This segment has features consistent with a role as a flexible and presumably unstructured linker and enzyme tether, but there has been no direct information about its structure. We used site-directed spin labeling and electron paramagnetic resonance spectroscopy to characterize structural features of the carboxyl-terminal 40 residues of E. coli chemoreceptor Tar. Beginning ～ 35 residues from the carboxyl terminus and continuing to the end of the protein, spectra of spin-labeled Tar embedded in native membranes or in reconstituted proteoliposomes, exhibited mobilities characteristic of unstructured, disordered segments. Binding of methyltransferase substantially reduced mobility for positions in or near the pentapeptide but mobility for the linker sequence remained high, being only modestly reduced in a gradient of decreasing effects for 10-15 residues, a pattern consistent with the linker providing a flexible arm that would allow enzyme diffusion within defined limits. Thus, our data identify that the carboxyl-terminal linker between the receptor body and the pentapeptide is an unstructured, disordered segment that can serve as a flexible arm and enzyme tether.     

Comparison of the computed structures for the phosphate-binding loop of the p21 protein containing the oncogenic site Gly 12 with the X-ray crystallographic structures for this region in the p21 protein and EFtu. A model for the structure of the p21 protein in its oncogenic form

The GTP-binding p21 protein encoded by the ras-oncogene can be activated to cause malignant transformation of cells by substitution of a single amino acid at critical positions along the polypeptide chain. Substitution of any non-cyclic L-amino acid for Gly 12 in the normal protein results in a transforming protein. This substitution occurs in a hydrophobic sequence (residues 6-15) which is known to be involved in binding the phosphate moities of GTP (and GDP). We find, using conformational energy calculations, that the 6-15 segment of the normal protein (with Gly 12) adopts structures that contain a bend at residues 11 and 12 with the Gly in the D* conformation, not allowed energetically for L-amino acids. Substitution of non-cyclic L-amino acids for Gly 12 results in shifting this bend to residues 12 and 13. We show that many computed structures for the Gly 12-containing phosphate binding loop, segment 9-15, are superimposable on the corresponding segment of the recently determined X-ray crystallographic structure for residues 1-171 of the p21 protein. All such structures contain bends at residues 11 and 12 and most of these contain Gly 12 in the C* or D* conformational state. Other computed conformations for the 9-15 segment were superimposable on the structure of the corresponding 18-23 segment of EFtu, the bacterial chain elongation factor having structural similarities to the p21 protein in the phosphate-binding regions. This segment contains a Val residue where a Gly occurs in the p21 protein. As previously predicted, all of these superimposable conformations contain a bend at positions 12 and 13, not 11 and 12. If these structures that are superimposable on EFtu are introduced into the p21 protein structure, bad contacts occur between the sidechain of the residue (here Val) at position 12 and another phosphate binding loop region around position 61. These bad contacts between the two segments can be removed by changing the conformation of the 61 region in the p21 protein to the corresponding position of the homologous region in EFtu. In this new conformation, a large site becomes available for the binding of phosphate residues. In addition, such phenomena as autophosphorylation of the p21 protein by GTP can be explained with this new model structure for the activated protein which cannot be explained by the structure for the non-activated protein.     

Linked and threaded loops in proteins

Tertiary structure of globular proteins has traditionally been analyzed in terms of the organization of secondary structure elements. This paper presents a method for systematically identifying different topological features of the convolutions of the backbone. We define a loop as a segment of chain whose end residues are in contact. We find some loops which are threaded by another segment of chain passing through the loop or actually linked with another loop. Fifty-six loop threadings were found among the 20 proteins studied, all of them occurring in a subset of seven proteins. In our sample, threadings and linkings were generally found if and only if the protein has more than 200 residues. To account for the existence of these topological features, despite their apparent entropic unfavorability, we have proposed a number of kinetic mechanisms by which they may form without a thread actually passing through a loop. We have found that almost all loop threadings possess structural features that would make one of these mechanisms plausible.     

Properties and the primary structure of a new halorhodopsin from halobacterial strain mex.

A new halorhodopsin-like pigment from the new halobacterial strain mex (Otomo, J., Tomoika, H. and Sasabe, H. (1992) J. Gen. Microbiol. 138, 1027-1037) was partially purified, and its amino acid sequence from helices A to G was determined using PCR technique. Two arginine residues in the A-B interhelix loop segment, a series of six amino acid residues (EMPAGH) in the B-C interhelix segment and most of the residues near the Schiff base of the retinal were found to be conserved in three halorhodopsins (halobium, pharaonis and mex). This result strongly suggests that these residues are essential for anion pumping function in halorhodopsin. The light-induced ion-pump measurements have shown that the selectivity of anion transport between chloride and nitrate in mex halorhodopsin is lower than that of halobium halorhodopsin, but higher than that of pharaonis halorhodopsin. The number of amino acid residues in the B-C interhelix loop segments is different in each halorhodopsin, and it correlates with their anion (chloride and nitrate) selectivity. These results suggest that the length of the B-C segment affects the selectivity of anion transport in halorhodopsin.     

Co-evolving residues in membrane proteins

MOTIVATION: The analysis of co-evolving residues has been exhaustively evaluated for the prediction of intramolecular amino acid contacts in soluble proteins. Although a variety of different methods for the detection of these co-evolving residues have been developed, the fraction of correctly predicted contacts remained insufficient for their reliable application in the construction of structural models. Membrane proteins, which constitute between one-fourth and one-third of all proteins in an organism, were only considered in few individual case studies. RESULTS: We present the first general study of correlated mutations in alpha-helical membrane proteins. Using seven different prediction algorithms, we extracted co-evolving residues for 14 membrane proteins having a solved 3D structure. On average, distances between correlated pairs of residues lying on different transmembrane segments were found to be significantly smaller compared to a random prediction. Covariation of residues was frequently found in direct sequence neighborhood to helix-helix contacts. Based on the results obtained from individual prediction methods, we constructed a consensus prediction for every protein in the dataset that combines obtained correlations from different prediction algorithms and simultaneously removes likely false positives. Using this consensus prediction, 53% of all predicted residue pairs were found within one helix turn of an observed helix-helix contact. Based on the combination of co-evolving residues detected with the four best prediction algorithms, interacting helices could be predicted with a specificity of 83% and sensitivity of 42%. AVAILABILITY: http://webclu.bio.wzw.tum.de/helixcorr/     

Dynamic properties of pH-dependent structural organization of the amyloidogenic β-protein (1-40)

The structural organization of the amyloidogenic β-proteins containing 40 amino acid residues (Aβ40) was studied by the high temperature molecular dynamics simulations in the acidic (pH~3) and basic (pH~8) pH regions. The obtained data suggest that the central Ala21-Gly29 segment of Aβ40, can adopt folded and partially unfolded structures. At the basic pH, this segment forms folded structures, stabilized by electrostatic interactions and hydrogen bonds. At the acidic pH, it forms partially unfolded structures. Two other segments flanking to the central segment exhibit the propensity to adopt unstable inter-converting α-helical, 310-helical and turn-like structures. One of these segments is comprised of the Ala30-Val36 residues at both of the considered pHs. The second segment is comprised of the Glu11-Phe20 at the basic pH and of the Glu11-Val24 residues at the acidic pHs. The revealed pH-dependent structuration of the Aβ40 allowed us to suggest a possible scenario for initial Aβ aggregation. According to this scenario, the occurrence of the partially unfolded states of the Ala21-Gly29 segment plays main role in the Aβ oligomerization process.     

Hydrogen-bonded structure of the complex N-linked fetuin glycopeptide

The conformation of the N-linked complex glycopeptide of fetuin was examined with hydrogen-exchange techniques. The glycopeptide molecule contains eight acetamido hydrogens stemming from five N-acetylglucosamine residues and three N-acetylneuraminic acid residues and also one from the remaining sugar-peptide linkage. The hydrogen-exchange rates of these secondary amides were compared with small molecule model compounds having identical primary structures at their exchangeable hydrogen sites. Differences between the model rates and glycopeptide rates therefore cannot be accounted for by primary structure effects but reflect conformational features of the glycopeptide. Two glycopeptide hydrogens exhibit significantly hindered exchange; the rest exchange at the model rates. Removal of the three N-acetylneuraminic acid residues from terminal positions on the three branches of the glycopeptide removes the slowed hydrogens. The remaining ones continue to exchange at the model rate. These results indicate that two of the eight sugar acetamido hydrogens are involved in intramolecular hydrogen bonds. A likely structure includes two hydrogen bonds between the three N-acetylneuraminic acid residues. These two hydrogens, slowed to a moderate degree, reflect a preferred conformation stabilized by about 1 kcal/mol in free energy. The solution conformation of the glycopeptide suggested by these results is one that is partially ordered and can be easily modulated, owing to the relatively small amount of energy stabilizing the preferred conformation.     

High-resolution solid state 13C NMR of bacteriorhodopsin: characterization of [4-13C]Asp resonances.

Solid state 13C nuclear magnetic resonance measurements of bacteriorhodopsin labeled with [4-13C]Asp show that resonances of single amino acids can be resolved. In order to assign and characterize the resonances of specific Asp residues, three different approaches were used. (1) Determination of the chemical shift anisotropy from side-band intensities provides information about the protonation state of Asp residues. (2) Relaxation studies and T1 filtering allow one to discriminate between resonances with different mobility. (3) A comparison of the spectra of light- and dark-adapted bacteriorhodopsin provides evidence for resonances from aspartic acid residues in close neighborhood of the chromophore. In agreement with other investigations, four resonances are assigned to internal residues. Two of them are protonated in the ground state up to pH 10 (Asp96 and Asp115). All other detected resonances, including Asp85 and Asp212, are due to deprotonated aspartic acid. Two lines due to the two internal deprotonated groups change upon dark and light adaptation, whereas the protonated Asp residues are unaffected.     

A gene coding for a novel protein specific to the olfactory rosettes of Atlantic salmon

We have isolated a 1.6 kb clone from a cDNA library made from the olfactory rosettes of the Atlantic salmon (Salmo salar). The clone contains a 1200 bp, open reading frame (named OSC) which codes for a protein with 400 amino-acid residues (Oscp). The mRNA corresponding to OSC is strongly expressed in the olfactory rosettes and weakly expressed in gills but is expressed in only these two tissues. This suggests that Oscp may have a specific and important role in olfaction. The sequence of Oscp suggests that it is not globular. Predictions show only a small fraction of alpha-helix. Oscp is hydrophilic but with the number of positively charged residues equal to the number of negatively charged residues. No closely similar protein can be found on the basis of homology searches or hydrophobicity comparisons. However, a 44 residue segment (G300 through K343) is significantly homologous to a segment of alpha-lactalbumin (G51 through K94). The similarities include the 19 residues of the "alpha- lactalbumin-lysozyme C signature," the ten residues of the Ca(2+) binding elbow and the four cysteine residues which provide two key disulfide links in alpha-lactalbumin and lysozyme C. Two more Cys residues are also very similarly placed. We conclude that the gene OSC codes for a unique protein which most likely contains a specific site for binding Ca(2+) and plays a unique role in the signal pathway of olfaction in salmon.     

Designation of sequences involved in the “coiled-coil” interdomainal connections in fibrinogen: Construction of an atomic scale model

Primary structure studies on human fibrinogen (α₂β₂γ₂) have revealed certain unusual features which are compatible with the existence of a three-stranded set of supercoiled α-helices thought to be characteristic of the keratin family of fibrous proteins. In particular, each of the three non-identical chains has two characteristic braces of cysteines separated by 111 residues. The three chains are apparently bound together at these two junctures in unique six-cysteine rings. The amino acid sequences between these unusual cysteine pairs (themselves separated in all six cases by three residues) are helix-permissive over significant portions of their lengths. Moreover, the non-polar residues tend to vary rhythmically. To test the proposition that these sequences do indeed correspond to the “coiled-coils” long ago predicted on the basis of wide-angle X-ray diffraction studies, we constructed a detailed, atomic scale model of a part of this region. To this end, we fashioned three α-helical segments, each 29 residues long and corresponding to the designated sequences of the α, β and γ-chains, respectively. In each case we incorporated a pitch of approximately 200 Å. We were then able to fit the three helices together in the two possible combinations which yield a pseudo-3-fold axis. In either case all polar residues extend away from the parallel three-stranded rope, and almost all the non-polar side-chains are directed toward the interior. We also constructed a separate model showing how the six cysteines at each end of the proposed rod-like segment are best arranged. The co-ordinates from both models were collected and utilized in a computer-graphic Molecular Modeling System which can display features of the models selectively. Various projections were plotted automatically, some of which are reproduced here.     

1H NMR studies of the glucocorticoid receptor DNA-binding domain: sequential assignments and identification of secondary structure elements

Two protein fragments containing the DNA-binding domain (DBD) of the glucocorticoid receptor (GR) have been studied by two-dimensional 1H NMR spectroscopy. The two peptides (93 and 115 residues, respectively) contain a common segment corresponding to residues C440-I519 of the rat GR or residues C421-I500 of the human GR and include two Zn-binding "finger" domains. The structures of this segment are almost identical in the two protein fragments, as judged from chemical shifts and sequential NOE connectivities. More than 90% of all observable 1H resonances within a 71-residue segment encompassing C440-R510 (rat GR) could be sequentially assigned by standard techniques, and stereospecific assignments could be made for the methyl groups in four valine residues within this segment. Sequential NOE connectivities indicate several elements of secondary structure including two alpha-helical segments consisting of residues S459-E469 and P493-G504, a type I reverse turn between residues R479 and C482, a type II reverse turn between residues L475 and G478, and several regions of extended peptide conformation. No evidence for alpha-helical conformation was found within the two putative zinc-finger domains, indicating that the structures of these domains differ from that of TFIIIA-type zinc fingers. The observation of some very slowly exchanging amide protons in the N-terminal (CI) domain of the DBD in combination with slow rotation of the Y452 aromatic ring indicates that this domain has a restricted conformational flexibility compared to the C-terminal (CII) domain. We also observe several long-range NOE connectivities within C440-R510, suggesting that the sequential assignments presented here will provide a basis for a complete structure determination of this segment of the GR.     

Evolutionary constraints associated with functional specificity of the CMGC protein kinases MAPK, CDK, GSK, SRPK, DYRK, and CK2alpha

Amino acid residues associated with functional specificity of cyclin-dependent kinases (CDKs), mitogen-activated protein kinases (MAPKs), glycogen synthase kinases (GSKs), and CDK-like kinases (CLKs), which are collectively termed the CMGC group, were identified by categorizing and quantifying the selective constraints acting upon these proteins during evolution. Many constraints specific to CMGC kinases correspond to residues between the N-terminal end of the activation segment and a CMGC-conserved insert segment associated with coprotein binding. The strongest such constraint is imposed on a "CMGC-arginine" near the substrate phosphorylation site with a side chain that plays a role both in substrate recognition and in kinase activation. Two nearby buried waters, which are also present in non-CMGC kinases, typically position the main chain of this arginine relative to the catalytic loop. These and other CMGC-specific features suggest a structural linkage between coprotein binding, substrate recognition, and kinase activation. Constraints specific to individual subfamilies point to mechanisms for CMGC kinase specialization. Within casein kinase 2alpha (CK2alpha), for example, the binding of one of the buried waters appears prohibited by the side chain of a leucine that is highly conserved within CK2alpha and that, along with substitution of lysine for the CMGC-arginine, may contribute to the broad substrate specificity of CK2alpha by relaxing characteristically conserved, precise interactions near the active site. This leucine is replaced by a conserved isoleucine or valine in other CMGC kinases, thereby illustrating the potential functional significance of subtle amino acid substitutions. Analysis of other CMGC kinases similarly suggests candidate family-specific residues for experimental follow-up.     

When fold is not important: a common structural framework for adenine and AMP binding in 12 unrelated protein families

ATP is a ligand common to many proteins, yet it is unclear whether common recognition patterns do exist among the many different folds that bind ATP. Previously, it was shown that cAMP-dependent protein kinase, D-Ala:D-Ala ligase and the alpha-subunit of the alpha 2 beta 2 ribonucleotide reductase do share extensive common structural elements for ATP recognition although their folds are different. Here, we have made a survey of structures that bind ATP and compared them with the key features seen in these three proteins. Our survey shows that 12 different fold types share a specific recognition pattern for the adenine moiety, and 8 of these folds have a common structural framework for recognition of the AMP moiety of the ligand. The common framework consists of a tripeptide segment plus three additional residues, which provides similar polar and hydrophobic interactions between the protein and mononucleotide. Consensus interactions are represented by four key hydrogen bonds present in each fold type. Two of these four hydrogen bonds, together with three aliphatic residues, form a specific recognition pattern for the adenine moiety in all 12 folds. These similarities point to a structural-functional requirement shared by these different mononucleotide-binding proteins that represent at this time 28% of the adenine mononucleotide complexes found in the Brookhaven Protein Data Bank.