Differences between alcohol dehydrogenases. Structural properties and evolutionary aspects.

Comparisons of the primary structures of yeast and horse liver alcohol dehydrogenases reveal that the enzymes are homologous but distantly related. The overall positional identity is 25% between common regions, and several deletions/insertions occur in either enzyme, the longest apparently corresponding to 21 residues, showing that the different subunit sizes are largely explained by internal differences. Variabilities in the structural similarities can be coupled with functional requirements but not directly with whole domains in the previously known tertiary structure of the horse protein. The two most similar regions of the enzymes affect active-site segments and the two most dissimilar regions seem to affect a loop structure without known function, and a segment participating in subunit interactions. The dissimilarities may probably be correlated with changes in zinc-binding properties and quaternary structures. The extra region corresponding to the large internal chain-length difference shows an apparent coincidence in sequence to a following segment of the horse enzyme, and additional elements of internal coincidences, or superficial similarities with other dehydrogenases, are noticed. These characteristics are not fully distinguishable from chance distributions but in view of the extensive species variations in alcohol dehydrogenases some evolutionary considerations may not be excluded, in which case a model relating all regions of these and associated enzymes to a common ancestor is shown to be compatible with all known observations.     

High resolution X-ray structure of yeast hexokinase,an allosteric protein exhibiting a non-symmetric arrangement of subunits

The structure determination of yeast hexokinase has been extended to 3.5 Å resolution for the dimer and to 2.7 Å resolution for the monomer using multiple isomorphous replacement. The electron density maps of both the monomer and dimer crystal forms have been substantially improved by an averaging procedure. From these maps the course of the polypeptide backbone and some aspects of the dimer interaction have been established.The hexokinase subunit arrangement is contrary to a major tenet of the Monod et al. (1965) theory of allosteric proteins which postulated that only symmetric or isologous interactions of subunits would occur in oligomeric proteins. One subunit of the dimer is related to the other by a 156 ° rotation about and a 13.8 Å translation along a molecular screw axis. In the hexokinase dimer the set of residues in one subunit that is interacting with the other subunit is different from the set of residues in the second subunit that is interacting with the first subunit. This heterologous or non-symmetric interaction of subunits is associated with some small differences in the structure of the two subunits, particularly at the subunit interface, and accounts for some of this enzyme's non-symmetric interactions with substrates and activators. Indeed, the non-symmetric subunit association may play an important role in the control of this enzyme's activity.The overall structure of hexokinase is considerably different than the known structures of the other enzymes in the glycolytic pathway. Although there is a striking similarity between the domain of hexokinase that binds AMP and the domain of lactate dehydrogenase that binds NAD, the former structure contains both antiparallel and parallel β-pleated strands, while the latter contains only parallel β-structure. In an attempt to assess the significance of this structural similarity, the structure of the nucleotide binding domains of hexokinase and lactate dehydrogenase are compared to a portion of carboxypeptidase A. The observed similarities among these structures suggests that a central β-pleated sheet flanked by α-helices is a common supersecondary structure that probably arose by convergent as well as divergent evolution. Thus, there appears to be no compelling evidence at this time to support the hypothesis that a part of hexokinase has evolved from the same gene as the dinucleotide binding domain of lactate dehydrogenase.     

The role of structural intersubunit microheterogeneity in the regulation of the activity in hysteresis of ribulose 1, 5-bisphosphate carboxylase/oxygenase

Many enzymes are composed of subunits with the identical primary structure. It has been believed that the protein structure of these subunits is the same. Ribulose 1,5-bisphosphate carboxylase/oxygenase (RuBisCO) comprises eight large subunits with the identical amino acid sequence and eight small subunits. Rotation of the side chains of the lysine residues, Lys-21 and Lys-305, in each of the eight large subunits in spinach RuBisCO in two ways produces microheterogeneity among the subunits. These structures are stabilized through hydrogen bonds by water molecules incorporated into the large subunits. This may cause different effects upon catalysis and a hysteretic, time-dependent decrease in activity in spinach RuBisCO. Changing the amino acid residues corresponding to Lys-21 and Lys-305 in non-hysteretic Chromatium vinosum RuBisCO to lysine induces hysteresis and increases the catalytic activity from 8.8 to 15.8 per site per second. This rate is approximately five times higher than that of the higher-plant enzyme.     

Crystal structure of the human protein kinase CK2 regulatory subunit reveals its zinc finger-mediated dimerization.

Protein kinase CK2 is a tetramer composed of two alpha catalytic subunits and two beta regulatory subunits. The structure of a C-terminal truncated form of the human beta subunit has been determined by X-ray crystallography to 1.7 A resolution. One dimer is observed in the asymmetric unit of the crystal. The most striking feature of the structure is the presence of a zinc finger mediating the dimerization. The monomer structure consists of two domains, one entirely alpha-helical and one including the zinc finger. The dimer has a crescent shape holding a highly acidic region at both ends. We propose that this acidic region is involved in the interactions with the polyamines and/or catalytic subunits. Interestingly, conserved amino acid residues among beta subunit sequences are clustered along one linear ridge that wraps around the entire dimer. This feature suggests that protein partners may interact with the dimer through a stretch of residues in an extended conformation.     

Identification of functional subclasses in the DJ-1 superfamily proteins

Genomics has posed the challenge of determination of protein function from sequence and/or 3-D structure. Functional assignment from sequence relationships can be misleading, and structural similarity does not necessarily imply functional similarity. Proteins in the DJ-1 family, many of which are of unknown function, are examples of proteins with both sequence and fold similarity that span multiple functional classes. THEMATICS (theoretical microscopic titration curves), an electrostatics-based computational approach to functional site prediction, is used to sort proteins in the DJ-1 family into different functional classes. Active site residues are predicted for the eight distinct DJ-1 proteins with available 3-D structures. Placement of the predicted residues onto a structural alignment for six of these proteins reveals three distinct types of active sites. Each type overlaps only partially with the others, with only one residue in common across all six sets of predicted residues. Human DJ-1 and YajL from Escherichia coli have very similar predicted active sites and belong to the same probable functional group. Protease I, a known cysteine protease from Pyrococcus horikoshii, and PfpI/YhbO from E. coli, a hypothetical protein of unknown function, belong to a separate class. THEMATICS predicts a set of residues that is typical of a cysteine protease for Protease I; the prediction for PfpI/YhbO bears some similarity. YDR533Cp from Saccharomyces cerevisiae, of unknown function, and the known chaperone Hsp31 from E. coli constitute a third group with nearly identical predicted active sites. While the first four proteins have predicted active sites at dimer interfaces, YDR533Cp and Hsp31 both have predicted sites contained within each subunit. Although YDR533Cp and Hsp31 form different dimers with different orientations between the subunits, the predicted active sites are superimposable within the monomer structures. Thus, the three predicted functional classes form four different types of quaternary structures. The computational prediction of the functional sites for protein structures of unknown function provides valuable clues for functional classification.     

Solution structure of the disulfide-linked dimer of human intestinal trefoil factor (TFF3): the intermolecular orientation and interactions are markedly different from those of other dimeric trefoil proteins

The trefoil protein TFF3 forms a homodimer (via a disulfide linkage) that is thought to have increased biological activity over the monomer. The solution structure of the TFF3 dimer has been determined by NMR and compared with the structure of the TFF3 monomer and with other trefoil dimer structures (TFF1 and TFF2). The most significant structural differences between the trefoil domain in the monomer and dimer TFF3 are in the orientations of the N-terminal 3(10)-helix (residues 10-12) and in the presence in the dimer of an additional 3(10)-helix (residues 53-55) outside of the core region. The TFF3 dimer forms a more compact structure as compared with the TFF1 dimer where the two trefoil domains are connected by a flexible region with the monomer units being at variable distances from each other and in many different orientations. Although TFF2 is also a compact structure, the dispositions of its monomer units are very different from those of TFF3. The structural differences between the dimers result in the two putative receptor/ligand binding sites that remain solvent exposed in the dimeric structures having very different dispositions in the different dimers. Such differences have significant implications for the mechanism of action and functional specificity for the TFF class of proteins.     

Structural segments and residue propensities in protein-RNA interfaces: comparison with protein-protein and protein-DNA complexes

The interface of a protein molecule that is involved in binding another protein, DNA or RNA has been characterized in terms of the number of unique secondary structural segments (SSSs), made up of stretches of helix, strand and non-regular (NR) regions. On average 10-11 segments define the protein interface in protein-protein (PP) and protein-DNA (PD) complexes, while the number is higher (14) for protein-RNA (PR) complexes. While the length of helical segments in PP interaction increases with the interface area, this is not the case in PD and PR complexes. The propensities of residues to occur in the three types of secondary structural elements (SSEs) in the interface relative to the corresponding elements in the protein tertiary structures have been calculated. Arg, Lys, Asn, Tyr, His and Gln are preferred residues in PR complexes; in addition, Ser and Thr are also favoured in PD interfaces.     

Biochemical analysis of the N-terminal domain of human RAD54B

The human RAD54B protein is a paralog of the RAD54 protein, which plays important roles in homologous recombination. RAD54B contains an N-terminal region outside the SWI2/SNF2 domain that shares less conservation with the corresponding region in RAD54. The biochemical roles of this region of RAD54B are not known, although the corresponding region in RAD54 is known to physically interact with RAD51. In the present study, we have biochemically characterized an N-terminal fragment of RAD54B, consisting of amino acid residues 26–225 (RAD54B_26–225). This fragment formed a stable dimer in solution and bound to branched DNA structures. RAD54B_26–225 also interacted with DMC1 in both the presence and absence of DNA. Ten DMC1 segments spanning the entire region of the DMC1 sequence were prepared, and two segments, containing amino acid residues 153–214 and 296–340, were found to directly bind to the N-terminal domain of RAD54B. A structural alignment of DMC1 with the Methanococcus voltae RadA protein, a homolog of DMC1 in the helical filament form, indicated that these RAD54B-binding sites are located near the ATP-binding site at the monomer–monomer interface in the DMC1 helical filament. Thus, RAD54B binding may affect the quaternary structure of DMC1. These observations suggest that the N-terminal domain of RAD54B plays multiple roles of in homologous recombination.     

Role of the hinge peptide and the intersubunit interface in the swapping of N-termini in dimeric bovine seminal RNase.

Bovine seminal ribonuclease (BS-RNase) is the only known dimeric enzyme characterized by an equilibrium between two different 3D structures: MxM, with exchange (or swapping) of the N-terminal 1-20 residues, and M=M, without exchange. As a consequence, the hinge region 16-22 has a different tertiary structure in the two forms. In the native protein, the equilibrium ratio between MxM and M=M is about 7 : 3. Kinetic analysis of the swapping process for a recombinant sample shows that it folds mainly in the M=M form, then undergoes interconversion into the MxM form, reaching the same 7 : 3 equilibrium ratio. To investigate the role of the regions that are most affected structurally by the swapping, we expressed variant proteins by replacing two crucial residues with the corresponding ones from RNase A: Pro19, within the hinge peptide, and Leu28, located at the interface between subunits. We compared the structural properties of the monomeric forms of P19A-BS-RNase, L28Q-BS-RNase and P19A/L28Q-BS-RNase variants with those of the parent protein, and investigated the exchange kinetics of the corresponding dimers. The P19A mutation slightly increases the thermal stability of the monomer, but it does not alter the swapping tendency of the dimer. In contrast, the L28Q mutation significantly affects both the dimerization and swapping processes but not the thermal stability of the monomer. Overall, these results suggest that the structural determinants that control the exchange of N-terminal arms in BS-RNase may not be located within the hinge peptide, and point to a crucial role of the interface residues.     

DNA-induced conformational changes in type II restriction endonucleases: the structure of unliganded HincII

The 2.1A crystal structure of the unliganded type II restriction endonuclease, HincII, is described. Although the asymmetric unit contains only a single monomer, crystal lattice contacts bring two monomers together to form a dimer very similar to that found in the DNA bound form. Comparison with the published DNA bound structure reveals that the DNA binding pocket is expanded in the unliganded structure. Comparison of the unliganded and DNA liganded structures reveals a simple rotation of subunits by 11 degrees each, or 22 degrees total, to a more closed structure around the bound DNA. Comparison of this conformational change to that observed in the other type II restriction endonucleases where DNA bound and unliganded forms have both been characterized, shows considerable variation in the types of conformational changes that can occur. The conformational changes in three can be described by a simple rotation of subunits, while in two others both rotation and translation of subunits relative to one another occurs. In addition, the endonucleases having subunits that undergo the greatest amount of rotation upon DNA binding are found to be those that distort the bound DNA the least, suggesting that DNA bending may be less facile in dimers possessing greater flexibility.     

Molecular Dynamics on Complexes of ras-p21 and Its Inhibitor Protein, rap-1A, Bound to the ras-Binding Domain of the raf-p74 Protein: Identification of Effector Domains in the raf Protein

We have computed the average structures for the ras-p21 protein and its strongly homologous inhibitor protein, rap-1A, bound to the ras-binding domain (RBD) of the raf protein, using molecular dynamics. Our purpose is to determine the differences in structure between these complexes that would result in no mitogenic activity of rap-1A-RBD but full activity of p21-RBD. We find that despite the similarities of the starting structures for both complexes, the average structures differ considerably, indicating that these two proteins do not interact in the same way with this vital target protein. p21 does not undergo major changes in conformation when bound to the RBD, while rap-1 A undergoes significant changes in structure on binding to the RBD, especially in the critical region around residue 61. The p21 and rap-1A make substantially different contacts with the RBD. For example, the loop region from residues 55–71 of rap-la makes extensive hydrogen-bond contacts with the RBD, while the same residues of p21 do not. Comparison of the structures of the RBD in both complexes reveals that it undergoes considerable changes in structure when its structure bound to p21 is compared with that bound to rap-1A. These changes in structure are due to displacements of regular structure (e.g., α-helices and β-sheets) rather than to changes in the specific conformations of the segments themselves. Three regions of the RBD have been found to differ significantly from one another in the two complexes: the binding interface between the two proteins at residues 60 and 70, the region around residues 105–106, and 118–120. These regions may constitute effector domains of the RBD whose conformations determine whether or not mitogenic signal transduction will occur.     

Structural requirements for catalysis and dimerization of human methionine adenosyltransferase I/III

We have used site-directed mutagenesis to probe the structural requirements for catalysis and dimerization of human hepatic methionine adenosyltransferase (hMAT). We built a homology model of the dimeric hMAT III inferred by the crystal structure of the highly homologous Escherichia coli MAT dimer. The active sites of both enzymes comprise the same amino acids and are located in the inter-subunit interface. All of the amino acids predicted to be in the hMAT III active site were mutated, as well as residues in a conserved ATP binding region. All of the mutations except one severely affected catalytic activity. On the other hand, dimerization was affected only by single mutations of three different residues, all on one monomer. The homology model suggested that the side chains of these residues stabilized the monomer and participated in a bridge between subunits consisting of a network of metal and phosphate ions. In agreement with this observation, we demonstrated that dimerization cannot occur in the absence of phosphate.     

Structural similarity between legumin and vicilin storage proteins from legumes.

The primary structures for several members of both the vicilin and legumin families of storage proteins were examined using a computer routine based on amino acid physical characteristics. The comparison algorithm revealed that sequences from the two families could be aligned and share a number of predicted secondary structural features. The COOH-terminal half of the subunits in both families displayed a highly conserved core region that was largely hydrophobic and in which a high proportion of the residues were predicted to be in beta-sheet conformations. The central region of the molecules which contained mixed areas of predicted helical and sheet conformations showed more variability in residue selection than the COOH-terminal regions. The NH2-terminal segments of subunits from the two different families could not be aligned though they characteristically had a high proportion of residues predicted to be in helical conformations. The feature which most clearly distinguished subunits between the two families was an inserted span in the legumin group with a high proportion of acidic amino acids located between the central and COOH-terminal domains. Residues in this insertion were predicted to exist mainly in helical conformation. Since considerable size variation occurs in this area amongst the legumin subunits, alterations in this region may have a minimal detrimental effect on the structure of the proteins.     

Secondary structure of bovine beta-lactoglobulin B

Secondary-structure-prediction algorithms have been used to find the segments of beta-lactoglobulin sequence most likely to fit the circular dichroism assignment of 15% alpha-helix, 50% beta-sheet, and 15-20% reverse turn. A number of segments may have an alpha-helical conformation but the most prominent region of alpha-helix is from residue 129 to 143. A further probable alpha-helix segment is residues 65-76. The number of residues predicted to occur in segments of beta-sheet structure is less than expected. However, the most likely segments are for residues 1-6, 11-16, 39-45, 80-85, 92-96, 101-107, 117-123, and 145-151. Predicted reverse-turn tetrapeptides are residues 7-10, 49-52, 61-64, 88-91, and 112-115. These predicted secondary structures are consistent with the low-resolution structure of the molecule determined by X-ray diffraction studies.     

Organization of African green monkey DNA at junctions between alpha-satellite and other DNA sequences

Segments of African green monkey DNA containing sequences of the highly reiterated cryptic satellite DNA called α-satellite were selected from a library in λ bacteriophage. This λ library was constructed to enrich for monkey segments that contain (1) irregular regions of α-satellite and (2) α-satellite linked to other monkey sequences. At least 11 of 15 cloned monkey segments between 13 × 10³ and 16 × 10³ base-pairs in length, selected by hybridization to α-satellite, also include other monkey sequences.In general, α-satellite sequences close to the junctions with non-α-satellite DNA contain an abundance of divergent forms compared to the average frequency of such forms within total α-satellite. Many of the cloned segments are missing some of the HinIII sites that occur once in most monomer units of α-satellite, and likewise several of the cloned segments contain restriction sites that rarely occur in α-satellite as a whole. In some segments HinIII sites occur that are spaced at distances other than the basic multiple of 172 base-pairs. At least one of the cloned segments, however, is composed mainly of typical 172 base-pair long α-satellite monomer units.Several of these cloned DNAs have been mapped by restriction endonuclease digestion and Southern blot analysis and the arrangements of α-satellite and non-α-satellite sequences have been determined. In addition to segments that contain a boundary where satellite meets other types of sequence, some contain two such boundaries and thus satellite flanks a non-α-satellite segment. Further, two different types of non-α-satellite sequence appear to be common to more than one phage, perhaps indicating some recurring organization at boundaries.     

A comparison of structural relationships among α-crystallin, human Hsp27, γ-crystallins and βB2-crystallin

The 3D structures of α-crystallin, a major eye lens protein, and related small heat shock proteins are unresolved. It has been assumed that α-crystallin is primarily a β-sheet globular protein similar to γ-crystallin (Siezen and Argos, Biochim. Biophys. Acta, 1983, 748, 56–67) containing sequence repeats in its two domains (Wistow, FEBS Lett. 1985, 181, 1–6). Positional flexibility of amino acid residues and far UV-circular dichroism spectroscopy were used to investigate structural relationships among these proteins. The utility of flexibility plots for predicting protein structure is demonstrated by the excellent correlation of these plots with the known 3D X-ray structures of β/γ-crystallins. Similar analyses of α-crystallin subunits, αA and αB, and human heat shock protein 27 show that the C-terminal domains and connecting segments of these proteins are very similar while the N-terminal domains have significant structural differences. Unlike β/γ-crystallins, both Hsp27 and α-crystallin subunits are asymmetrical with highly flexible C-terminal domains. Flexibility is considered essential for protein functional activity. Therefore, the C-terminal region may play an active role in α-crystallin and small heat shock protein function. Differences in flexibility profiles and estimated secondary structure distribution in α-crystallin by three recent/updated algorithms from far UV-CD spectra support our predicted 3D structure and the concept that α-crystallin and members of β/γ-superfamily are structurally dissimilar.     

The topology of the Lcb1p subunit of yeast serine palmitoyltransferase

The structural organization and topology of the Lcb1p subunit of yeast and mammalian serine palmitoyltransferases (SPT) were investigated. In the yeast protein, three membrane-spanning domains were identified by insertion of glycosylation and factor Xa cleavage sites at various positions. The first domain of the yeast protein, located between residues 50 and 84, was not required for the stability, membrane association, interaction with Lcb2p, or enzymatic activity. Deletion of the comparable domain of the mammalian protein SPTLC1 also had little effect on its function, demonstrating that this region is not required for membrane localization or heterodimerization with SPTLC2. The second and third membrane-spanning domains of yeast Lcb1p, located between residues 342 and 371 and residues 425 and 457, respectively, create a luminal loop of approximately 60 residues. In contrast to the first membrane-spanning domain, the second and third membrane-spanning domains were both required for Lcb1p stability. In addition, mutations in the luminal loop destabilized the SPT heterodimer indicating that this region of the protein is important for SPT structure and function. Mutations in the extreme carboxyl-terminal region of Lcb1p also disrupted heterodimer formation. Taken together, these data suggest that in contrast to other members of the alpha-oxoamine synthases that are soluble homodimers, the Lcb1p and Lcb2p subunits of the SPT heterodimer may interact in the cytosol, as well as within the membrane and/or the lumen of the endoplasmic reticulum.     

Cryo-EM structure of a monomeric yeast S. cerevisiae complex IV isolated with maltosides: Implications in supercomplex formation

In mitochondria, complex IV (CIV) can be found as a monomer, a dimer or in association with other respiratory complexes. The atomic structure of the yeast S. cerevisiae CIV in a supercomplex (SC) with complex III (CIII) pointed to a region of significant conformational changes compared to the homologous mammalian CIV structures. These changes involved the matrix side domain of Cox5A at the CIII-CIV interface, and it was suggested that it could be required for SC formation. To investigate this, we solved the structure of the isolated monomeric CIV from S. cerevisiae stabilised in amphipol A8–35 at 3.9 Å using cryo-electron microscopy. Only a minor change in flexibility was seen in this Cox5A region, ruling out large CIV conformational shift for interaction with CIII and confirming the different fold of the yeast Cox5A subunit compared to mammalian homologues. Other differences in structure were the absence of two canonical subunits, Cox12 and Cox13, as well as Cox26, which is unique to the yeast CIV. Their absence is most likely due to the protein purification protocol used to isolate CIV from the III-IV SC.     

Biochemical and functional studies on the regulation of the Saccharomyces cerevisiae AMPK homolog SNF1

AMP-activated protein kinase (AMPK) is a master metabolic regulator for controlling cellular energy homeostasis. Its homolog in yeast, SNF1, is activated in response to glucose depletion and other stresses. The catalytic (α) subunit of AMPK/SNF1, Snf1 in yeast, contains a protein Ser/Thr kinase domain (KD), an auto-inhibitory domain (AID), and a region that mediates interactions with the two regulatory (β and γ) subunits. Previous studies suggested that Snf1 contains an additional segment, a regulatory sequence (RS, corresponding to residues 392-518), which may also have an important role in regulating the activity of the enzyme. The crystal structure of the heterotrimer core of Saccharomyces cerevisiae SNF1 showed interactions between a part of the RS (residues 460-498) and the γ subunit Snf4. Here we report biochemical and functional studies on the regulation of SNF1 by the RS. GST pulldown experiments demonstrate strong and direct interactions between residues 450-500 of the RS and the heterotrimer core, and single-site mutations in the RS-Snf4 interface can greatly reduce these interactions in vitro. On the other hand, functional studies appear to show only small effects of the RS-Snf4 interactions on the activity of SNF1 in vivo. This suggests that residues 450-500 may be constitutively associated with Snf4, and the remaining segments of the RS, as well as the AID, may be involved in regulating SNF1 activity.     

Identification and functional analysis of the nuclear localization signals of ribosomal protein L25 from Saccharomyces cerevisiae

The regions of the large subunit ribosomal protein L25 from Saccharomyces cerevisiae responsible for nuclear localization of the protein were identified by constructing fusion genes encoding various segments of L25 linked to the amino terminus of beta-galactosidase. Indirect immunofluorescence of yeast cells expressing the fusions demonstrated that amino acid residues 1 to 17 as well as 18 to 41 of L25 promote import of the reporter protein into the nucleus. Both nuclear localization signal (NLS) sequences appear to consist of two distinct functional parts: one showed relatively weak nuclear targeting activity, whereas the other considerably enhances this activity but does not promote nuclear import by itself. Microinjection of in vitro prepared intact and N-terminally truncated L25 into Xenopus laevis oocytes demonstrated that the region containing the two NLS sequences is indeed required for efficient nuclear localization of the ribosomal protein. This conclusion was confirmed by complementation experiments using a yeast strain that conditionally expresses wild-type L25. The latter experiments also indicated that amino acid residues 1 to 41 of L25 are required for full functional activity of yeast 60 S ribosomal subunits. Yeast cells expressing forms of L25 that lack this region are viable, but show impaired growth and a highly abnormal cell morphology.