Atomic model of human cystic fibrosis transmembrane conductance regulator: membrane-spanning domains and coupling interfaces.

We describe herein an atomic model of the outward-facing three-dimensional structure of the membrane-spanning domains (MSDs) and nucleotide-binding domains (NBDs) of human cystic fibrosis transmembrane conductance regulator (CFTR), based on the experimental structure of the bacterial transporter Sav1866. This model, which is in agreement with previous experimental data, highlights the role of some residues located in the transmembrane passages and directly involved in substrate translocation and of some residues within the intracellular loops (ICL1-ICL4) making MSD/NBD contacts. In particular, our model reveals that D173 ICL1 and N965 ICL3 likely interact with the bound nucleotide and that an intricate H-bond network (involving especially the ICL4 R1070 and the main chain of NBD1 F508) may stabilize the interface between MSD2 and the NBD1F508 region. These observations allow new insights into the ATP-binding sites asymmetry and into the molecular consequences of the F508 deletion, which is the most common cystic fibrosis mutation.

Hydrophobic cluster analysis and modeling of the human Rh protein three-dimensional structures.

Rh (Rhesus) is a major blood group system in man, which is clinically significant in transfusion medicine. Rh antigens are carried by an oligomer of two major erythroid specific polypeptides, the Rh (D and CcEe) proteins and the RhAG glycoprotein, that shared a common predicted structure with 12 transmembrane a-helices (M0 to M11). Non erythroid homologues of these proteins have been identified (RhBG and RhCG), notably in diverse organs specialized in ammonia production and excretion, such as kidney, liver and intestine. Phylogenetic studies and experimental evidence have shown that these proteins belong to the Amt/Mep/Rh protein superfamily of ammonium/methylammonium permease, but another view suggests that Rh proteins might function as CO2 gas channels. Until recently no information on the structure of these proteins were available. However, in the last two years, new insight has been gained into the structural features of Rh proteins (through the determination of the crystal structures of bacterial AmtB and archeaebacterial Amt-1. Here, models of the subunit and oligomeric architecture of human Rh proteins are proposed, based on a refined alignment with and crystal structure of the bacterial ammonia transporter AmtB, a member of the Amt/Mep/Rh superfamily. This alignment was performed considering invariant structural features, which were revealed through Hydrophobic Cluster Analysis, and led to propose alternative predictions for the less conserved regions, particularly in the N-terminal sequences. The Rh models, on which an additional Rh-specific, N-terminal helix M0 was tentatively positioned, were further assessed through the consideration of biochemical and immunochemical data, as well as of stereochemical and topological constraints. These models highlighted some Rh specific features that have not yet been reported. Among these, are the prediction of some critical residues, which may play a role in the channel function, but also in the stability of the subunit structure and oligomeric assembly. These results provide a basis to further understand the structure/function relationships of Rh proteins, and the alterations occurring in variant phenotypes.

Nucleotide binding domains of human CFTR: a structural classification of critical residues and disease-causing mutations.

Defective function of the cystic fibrosis (CF) transmembrane conductance regulator (CFTR) causes CF, the most frequent lethal inherited disease among the Caucasian population. The structure of this chloride ion channel includes two nucleotide-binding domains (NBDs), whose ATPase activity controls channel gating. Recently, the experimental structures of mouse and human CFTR NBD1 and our model of the human CFTR NBD1/NBD2 heterodimer have provided new insights into specific structural features of the CFTR NBD dimer. In the present work, we provide a structural classification of CF-causing mutations which may complement the existing functional classification. Our analysis also identified amino acid residues which may play a critical role in interdomain interaction and are located at the NBD1-NBD2 interface or on the surface of the dimer. In particular, a cluster of aromatic amino acids, which includes F508 and straddles the two NBDs, might be directly involved in the interaction of the NBD1/NBD2 heterodimer with the channel-forming membrane-spanning domains.

Nucleotide-binding domains of human cystic fibrosis transmembrane conductance regulator: detailed sequence analysis and three-dimensional modeling of the heterodimer.

The cystic fibrosis transmembrane conductance regulator (CFTR) protein is encoded by the gene that is defective in cystic fibrosis, the most common lethal inherited disease among the Caucasian population. CFTR belongs to the ABC transporter superfamily, whose members form macromolecular architectures composed of two membrane-spanning domains and two nucleotide-binding domains (NBDs). The experimental structures of NBDs from several ABC transporters have recently been solved, opening new avenues for understanding the structure/function relationships and the consequences of some disease-causing mutations of CFTR. Based on a detailed sequence/structure analysis, we propose here a three-dimensional model of the human CFTR NBD heterodimer. This model, which is in agreement with recent experimental data, highlights the specific features of the CFTR asymmetric active sites located at the interface between the two NBDs. Moreover, additional CFTR-specific features can be identified at the subunit interface, which may play critical roles in active site interdependence and are uncommon in other NBD dimers.

Structural features of prions explored by sequence analysis I. Sequence data.

Animal prion proteins (PrPs) form at the sequence level a very homogenous and 'closed' family. Therefore, few of their structural and functional features can be gleaned from sequence comparison as is now possible on a wide scale for other protein families. To detect putatively related proteins (at the structural and/or functional level), we used a battery of sequence analysis tools. This analysis resulted in (i) the identification of a putative 'prion-like' domain within the envelope of foamy retroviruses, (ii) the detection of putative similarities between prions and an interferon-inducible membrane protein, and (iii) the proposal that of the TATA-box-binding protein is a structural scaffold, which might allow understanding of a key event leading to the structural conversion from PrP(C) (normal cellular prion structure) towards PrP(Sc) (pathogenic structure).

Structural features of prions explored by sequence analysis. II. A PrP(Sc) model.

Prion diseases are neurodegenerative disorders associated with a conformational conversion of the prion PrP protein, in which the beta strand content increases and that of the a helix decreases. However, the structure of the pathogenous form PrP(Sc), occurring after conformational conversion of the normal cellular form PrP(C), is not yet known. From sequence analysis, we have previously proposed that helix H2 of the prion PrP(C) structure might be a key region for this structural conversion. More recently, we identified the TATA box-binding protein fold as a putative scaffold that may locally satisfy the predicted secondary-structure organisation of PrP(Sc). In the present analysis, we detail the schematic construction of PrP(Sc) monomeric and dimeric models, based on this hypothesis. These models are globally compatible with available data and therefore may provide further insights into the structurally and functionally elusive PrP protein. Some comments are also devoted to a comparison of the yeast Ure2p prion and animal prions.

Glycosidase active site mutations in human alpha-L-iduronidase.

Mucopolysaccharidosis type I (MPS I; McKusick 25280) results from a deficiency in alpha-L-iduronidase activity. Using a bioinformatics approach, we have previously predicted the putative acid/base catalyst and nucleophile residues in the active site of this human lysosomal glycosidase to be Glu182 and Glu299, respectively. To obtain experimental evidence supporting these predictions, wild-type alpha-L-iduronidase and site-directed mutants E182A and E299A were individually expressed in Chinese hamster ovary-K1 cell lines. We have compared the synthesis, processing, and catalytic properties of the two mutant proteins with wild-type human alpha-L-iduronidase. Both E182A and E299A transfected cells produced catalytically inactive human alpha-L-iduronidase protein at levels comparable to the wild-type control. The E182A protein was synthesized, processed, targeted to the lysosome, and secreted in a similar fashion to wild-type alpha-L-iduronidase. The E299A mutant protein was also synthesized and secreted similarly to the wild-type enzyme, but there were alterations in its rate of traffic and proteolytic processing. These data indicate that the enzymatic inactivity of the E182A and E299A mutants is not due to problems of synthesis/folding, but to the removal of key catalytic residues. In addition, we have identified a MPS I patient with an E182K mutant allele. The E182K mutant protein was expressed in CHO-K1 cells and also found to be enzymatically inactive. Together, these results support the predicted role of E182 and E299 in the catalytic mechanism of alpha-L-iduronidase and we propose that the mutation of either of these residues would contribute to a very severe clinical phenotype in a MPS I patient.

uDENN, DENN, and dDENN: indissociable domains in Rab and MAP kinases signaling pathways.

DENN domains are found in a variety of signaling proteins but their exact function remains undefined. Some of the DENN-containing proteins, such as rat Rab3GEP (Rab3 GDP/GTP exchange protein) or mouse Rab6IP1 (Rab6 interacting protein 1) interact with GTPases of the Rab family. Others, such as human MADD (MAP (Mitogen-activated protein) kinase activating protein containing death domain) and human ST5 (Suppressor of tumoreginicity 5) gene products are involved in regulation of MAPKs (Mitogen-activated protein kinases) signaling pathways. Using a combination of profile-based and bidimensional analyses, we show here that DENN domains are much larger than described to date in domain databases, always encircled on both sides by more divergent domains, that we called uDENN and dDENN. These however share conserved amino acids which could play a key role in the DENN functions.

Distribution of tightened end fragments of globular proteins statistically matches that of topohydrophobic positions: towards an efficient punctuation of protein folding?

Using a set of 372 proteins representative of a variety of 56 distinct globular folds, a statistical correlation was observed between two recently revealed features of protein structures: tightened end fragments or 'closed loops', i.e. sequence fragments that are able in three-dimensional (3D) space to nearly close their ends (a current parameter of polymer physics), and 'topohydrophobic positions', i.e. positions always occupied in 3D space by strong hydrophobic amino acids for all members of a fold family. Indeed, in sequence space, the distribution of preferred lengths for tightened end fragments and that for topohydrophobic separation match. In addition to this statistically significant similarity, the extremities of these 'closed loops' may be preferentially occupied by topohydrophobic positions, as observed on a random sample of various folds. This observation may be of special interest for sequence comparison of distantly related proteins. It is also important for the ab initio prediction of protein folds, considering the remarkable topological properties of topohydrophobic positions and their paramount importance within folding nuclei. Consequently, topohydrophobic positions locking the 'closed loops' belong to the deep cores of protein domains and might have a key role in the folding process.

RUN domains: a new family of domains involved in Ras-like GTPase signaling.

RUN domains are present in several proteins that are linked particularly to the functions of GTPases in the Rap and Rab families. They could hence play an important role in multiple Ras-like GTPase signaling pathways.

Molecular and structural analysis of two novel mutations in a patient with mut(-) methylmalonyl-CoA deficiency.

Inherited defects in the gene encoding the methylmalonyl-CoA mutase (MCM) result in the mut forms of methylmalonic aciduria (MMA). Twelve mutations have been identified associated with the mut(-) phenotype. We report two novel mutations (K621N and D156N) in a compound heterozygote mut(-) patient. These two mutations and three previously published ones (H627N, A191E, Y231N) were mapped onto a three-dimensional homology model of the human MCM constructed from the crystal structure of the Propionibacterium shermanii enzyme.

Molecular characterization of the major membrane skeletal protein in the ciliate Tetrahymena pyriformis suggests n-plication of an early evolutionary intermediate filament protein subdomain.

Epiplasmin C is the major protein component of the membrane skeleton in the ciliate Tetrahymena pyriformis. Cloning and analysis of the gene encoding epiplasmin C showed this protein to be a previously unrecognized protein. In particular, epiplasmin C was shown to lack the canonical features of already known epiplasmic proteins in ciliates and flagellates. By means of hydrophobic cluster analysis (HCA), it has been shown that epiplasmin C is constituted of a repeat of 25 domains of 40 residues each. These domains are related and can be grouped in two families called types I and types II. Connections between types I and types II present rules that can be evidenced in the sequence itself, thus enforcing the validity of the splitting of the domains. Using these repeated domains as queries, significant structural similarities were demonstrated with an extra six heptads shared by nuclear lamins and invertebrate cytoplasmic intermediate filament proteins and deleted in the cytoplasmic intermediate filament protein lineage at the protostome-deuterostome branching in the eukaryotic phylogenetic tree.

Human glucocerebrosidase: heterologous expression of active site mutants in murine null cells.

Using bioinformatics methods, we have previously identified Glu235 and Glu340 as the putative acid/base catalyst and nucleophile, respectively, in the active site of human glucocerebrosidase. Thus, we undertook site-directed mutagenesis studies to obtain experimental evidence supporting these predictions. Recombinant retroviruses were used to express wild-type and E235A and E340A mutant proteins in glucocerebrosidase-deficient murine cells. In contrast to wild-type enzyme, the mutants were found to be catalytically inactive. We also report the results of various studies (Western blotting, glycosylation analysis, subcellular fractionation, and confocal microscopy) indicating that the wild-type and mutant enzymes are identically processed and sorted to the lysosomes. Thus, enzymatic inactivity of the mutant proteins is not the result of incorrect folding/processing. These findings indicate that Glu235 plays a key role in the catalytic machinery of human glucocerebrosidase and may indeed be the acid/base catalyst. As concerns Glu340, the results both support our computer-based predictions and confirm, at the biological level, previous identification of Glu340 as the nucleophile by use of active site labeling techniques. Finally, our findings may help to better understand the molecular basis of Gaucher disease, the human lysosomal disease resulting from deficiency in glucocerebrosidase.

Voronoï tessellation reveals the condensed matter character of folded proteins.

The packing geometry of amino acids in folded proteins is analyzed via a modified Voronoï tessellation method which distinguishes bulk and surface. From a statistical analysis of the Voronoï cells over 40 representative proteins, it appears that the packings are in average similar to random packings of hard spheres encountered in condensed matter physics, with a quite strong fivefold local symmetry. Moreover, the statistics permits one to establish a classification of amino acids in terms of increasing propensity to be buried in agreement with what is known from chemical considerations.

HYR, an extracellular module involved in cellular adhesion and related to the immunoglobulin-like fold.

Domains belonging to the immunoglobulin-like fold are responsible for a wide variety of molecular recognition processes. Here we describe a new family of domains, the HYR family, which is predicted to belong to this fold, and which appears to be involved in cellular adhesion. HYR domains were identified in several eukaryotic proteins, often associated with Complement Control Protein (CCP) modules or arranged in multiple copies. Our analysis provides a sequence and structural basis for understanding the role of these domains in interaction mechanisms and leads to further characterization of heretofore undescribed repeated domains with similar folds found in several bacterial proteins involved in enzymatic activities (some chitinases) or in cell surface adhesion (streptococcal C-alpha antigen).

Beta-sheet modeling by helical surfaces.

We present a topological description of a beta-sheet in terms of a piece of helical surface. It requires only two easy-to-handle parameters: the twist, i.e. the turn of the helical surface per residue, and the coiling, which is a curvature along the strands or in the direction perpendicular to the strands of the sheet. This method applies fairly well to three- and four-strand sheets, forming a too limited structure to be able to build a barrel. From an analysis of beta-sheets derived from a structural database, we show that this picture can even be reduced to the use of one main value, the twist angle. The dependence of beta-sheet twisting on the number of strands in a sheet, and also on the length and direction of strands, has been demonstrated. The applications of such a description may include the rapid modeling of 3D structures.

Structural features of normal and mutant human lysosomal glycoside hydrolases deduced from bioinformatics analysis.

Lysosomal storage diseases are due to inherited deficiencies in various enzymes involved in basic metabolic processes. As with other genetic diseases, accurate structure data for these enzymatic proteins should help in better understanding the molecular effects of mutations identified in patients with the corresponding lysosomal diseases; however, no such three-dimensional (3D) structure data are available for many lysosomal enzymes. Thus, we herein intend to illustrate for an audience of molecular geneticists how structure information can nonetheless be obtained via a bioinformatics approach in the case of five human lysosomal glycoside hydrolases. Indeed, using the two-dimensional hydrophobic cluster analysis method to decipher the sequence information available in data banks for the large group of glycoside hydrolases (clan GH-A) to which these human lysosomal enzymes belong, we could deduce structure predictions for their catalytic domains and propose explanations for the molecular effects of mutations described in patients. In addition, in the case of human beta-glucuronidase for which experimental 3D data have been reported, we also show here that bioinformatics methods relying on the available 3D structure information can be used to obtain further insights into the effects of various mutations described in patients with Sly disease. In a broader perspective, our work stresses that, in the context of a rapid increase in protein sequence information through genome sequencing, bioinformatics approaches might be highly useful for generating structure-function predictions based on sequence-structure interrelationships.

Sequence and structural features of the T-fold, an original tunnelling building unit.

A similar fold has been found in four archetype enzymes that perform different functions. This new fold has been named the T-fold because it is found in multimeric proteins crossed by a tunnel. The T-fold consists of an antiparallel beta-sheet of four sequential strands, and two antiparallel helices between the second and third strand, layered on the concave side of the beta-sheet. The presently known T-fold proteins share a high structural similarity (a mean of 1.4 A root mean square (r.m.s.) deviation on the common core) while they only exhibit a low level of sequence identity (a mean of 10.5% on the aligned regions). They bind to substrates belonging to the purine or pterin families, and share a fold-related binding site with a glutamate or glutamine residue anchoring the substrate and a lot of conserved interactions. They also share a similar oligomerization mode: several T-folds join together to form a beta(2n)alpha(n) barrel, then two barrels join together in a head-to-head fashion to made up the native enzymes. The T-fold has the characteristics of a globular domain, with a hydrophobic core and a clearly defined topohydrophobic network. It defines a new class of common folds or recurrent domains found in distantly related proteins. However, it is likely not stable in monomeric form and until now is only observed in association with other T-folds through multimerization. Proteins 2000;39:142-154.

The uteroglobin fold.

Uteroglobin (UTG) forms a fascinating homodimeric structure that binds small- to medium-sized ligands through an internal hydrophobic cavity, located at the interface between the two monomers. Previous studies have shown that UTG fold is not limited to the UTG/CC10 family, whose sequence/structure relationships are highlighted here, but can be extended to the cap domain of Xanthobacter autotrophicus haloalkane dehalogenase. We show here that UTG fold is adopted by several other cap domains within the alpha/beta hydrolase family, making it a well-suited "geode" structure allowing it to sequester various hydrophobic molecules. Additionally, some data about a new crystal form of oxidized rabbit UTG are presented, completing previous structural studies, as well as results from molecular dynamics, suggesting an alternative way for the ligand to reach the internal cavity.

Structure modelling and site-directed mutagenesis of the rat aromatic L-amino acid pyridoxal 5'-phosphate-dependent decarboxylase: a functional study.

The pyridoxal-5'-phosphate-dependent enzymes (B6 enzymes) are grouped into three main families named alpha, beta, and gamma. Proteins in the alpha and gamma families share the same fold and might be distantly related, while those in the beta family exhibit specific structural features. The rat aromatic L-amino acid decarboxylase (AADC; EC(4.1.1.28)) catalyzes the synthesis of two important neurotransmitters: dopamine and serotonin. It binds the cofactor pyridoxal-5'-phosphate and belongs to the alpha family. Despite the low level of sequence identity (approximately 10%) shared by the rat AADC and the sequences of the enzymes belonging to the B6 enzymes family, including the known three-dimensional structures, a multiple sequence alignment was deduced. A model was built using segments belonging to seven of the eleven known structures. By homology, and based on knowledge of the biochemistry of the aspartate aminotransferase, structurally and functionally important residues were identified in the rat AADC. Site-directed mutagenesis of the conserved residues D271, T246, and C311 was carried out in order to confirm our predictions and highlight their functional role. Mutation of D271A and D271N resulted in complete loss of enzyme activity, while the D271E mutant exhibited 2% of the wild-type activity. Substitution of T246A resulted in 5% of the wild-type activity while the C311A mutant conserved 42% of the wild-type activity. A functional model of the AADC is discussed in view of the structural model and the complementary mutagenesis and labelling studies.