Multi-compartment heart segmentation in CT angiography using a spatially varying gaussian classifier.

OBJECTIVE: A fully automated and efficient method for segmenting ten major structures within the heart in Cardiac CT Angiography data for the purposes of display or cardiac functional analysis. MATERIALS AND METHODS: A spatially varying Gaussian classifier is a flexible model for segmentation, combining the advantages of atlas-based frameworks, with supervised intensity models. It is composed of an independent Gaussian classifier at each voxel and uses non-rigid registration for the initial spatial alignment. We show how this large model can be trained efficiently and present a novel smoothing technique based on normalised convolution to mitigate inherent overfitting issues. The 30 datasets used in this study are selected from a variety of different scanners in order to test the robustness and stability of the algorithm. The datasets were manually segmented by a trained clinician. RESULTS: The method was evaluated in a leave-one-out fashion, and the results were compared to other state of the art methods in the field, with a mean surface-to-surface distance of between 0.61 and 2.12 mm for different compartments. CONCLUSION: The accuracy of this method is comparable to other state of the art methods in the field. Its benefits lie in its conceptual simplicity and its general applicability. Only one non-rigid registration is required, giving it a speed advantage over multi-atlas approaches. Further accuracy may be achievable through the incorporation of an explicit shape model.

Involvement of the C-terminal end of the prostrate-specific antigen in a conformational epitope: characterization by proteolytic degradation of monoclonal antibody-bound antigen and mass spectrometry.

Prostate-specific antigen (PSA), a 237-amino acid glycoprotein, encoded by the hKLK3 gene, is widely used as a serum marker for the diagnosis and management of prostate cancer. We report here the localization of a conformational epitope recognized by the anti-total PSA monoclonal antibody (mAb) 11E5C6, by proteolytic degradation of mAb-bound antigen followed by mass spectrometric analyses of the peptides generated. These two technologies, combined with molecular display, allowed the identification of amino acid residues contained within three different peptides distant on the PSA sequence, but close in the PSA three-dimensional structure, that may be part of the mAb 11E5C6 epitope. The last four C-terminal amino acid residues are included in this epitope, as well as certain other C-terminal residues between Y225 and T232. The involvement of the PSA C-terminal end in the mAb 11E5C6 epitope was confirmed by western blotting experiments with the recombinant protein proPSA-RP1, resulting from the cloning of an alternative transcript of the hKLK3 gene, in which the PSA C-terminal end was deleted and replaced by another sequence. Although the anti-total PSA mAb 5D5A5 used as a control bound proPSA-RP1, mAb 11E5C6 did not. The requirement of the C-terminal end for the recognition by mAb 11E5C6 may be useful for the discrimination of PSA-related forms.

Deacylation kinetics analysis of Streptococcus pneumoniae penicillin-binding protein 2x mutants resistant to beta-lactam antibiotics using electrospray ionization- mass spectrometry.

Penicillin-binding proteins (PBPs) catalyze the transpeptidase reaction involved in peptidoglycan synthesis and are covalently inhibited by the beta-lactam antibiotics. In a previous work we have focused on acylation efficiency measurements of various Streptococcus pneumoniae PBP2x* mutants to study the molecular determinants of resistance to beta-lactams. In the present paper we have developed a method to improve an accurate determination of the deacylation rate constant using electrospray ionization-mass spectrometry. This method is adaptable to the analysis of deacylation of any beta-lactam. Compared to the fluorographic technique, the ESI-MS method is insensitive to variations in the concentration of functional proteins and is therefore more reliable. We have established that the resistance of PBPs to beta-lactams is mostly due to a decrease of the acylation efficiency with only marginal effects on the deacylation rates.

Structural and kinetic studies of the pyruvate-ferredoxin oxidoreductase/ferredoxin complex from Desulfovibrio africanus.

The pyruvate-ferredoxin oxidoreductase (PFOR)/ferredoxin (Fd) system of Desulfovibrio africanus has been investigated with the aim of understanding more fully protein-protein interaction and the kinetic characteristics of electron transfer between the two redox partners. D. africanus contains three Fds (Fd I, Fd II and Fd III) able to function as electron acceptors for PFOR. The complete amino acid sequence of Fd II was determined by automatic Edman degradation. It revealed a striking similarity to that of Fd I. The protein consists of 64 residues and its amino acid sequence is in agreement with a molecular mass of 6822.5 Da as measured by electrospray MS. Fd II contains five cysteine residues of which the first four (Cys11, Cys14, Cys17 and Cys54) are likely ligands for the single [4Fe-4S] cluster. A covalently cross-linked complex between PFOR and Fd I or Fd II was obtained by using a water soluble carbodiimide. This complex exhibited a stoichiometry of one ferredoxin for one PFOR subunit and is dependent on the ionic strength. The second-order rate constants for electron transfer between PFOR and Fds determined electrochemically using cyclic voltammetry are 7 x 107 M-1.s-1 for Fd I and 2 x 107 M-1.s-1 for Fd II and Fd III. The Km values of PFOR for Fd I and Fd II measured both by the electrochemical and the spectrophotometric method have been found to be 3 microM and 5 microM, respectively. The three-dimensional modelling of Fd II and surface analysis of Fd I, Fd II and PFOR suggest that a protein-protein complex is likely to be formed between aspartic acid/glutamic acid invariant residues of Fds and lysine residues surrounding the distal [4Fe-4S] cluster of PFOR. All of these studies are indicative of the involvement of electrostatic interactions between the two redox partners.

A [2Fe-2S] protein from the hyperthermophilic bacterium Aquifex aeolicus.

Overexpression in Escherichia coli of the fdx4 gene from Aquifex aeolicus has allowed isolation and characterization of the first hyperthermophilic [2Fe-2S](Scys)(4) protein, a homodimer of M = 2 x 12.4 kDa with one [2Fe-2S] cluster per subunit. This protein is undamaged by heating to 100 degrees C for at least three hours. The primary structure, in particular the characteristic distribution of the four cysteine ligands of the metal site, and the spectroscopic properties of the A. aeolicus protein relate it to well characterized [2Fe-2S] proteins from Clostridium pasteurianum and Azotobacter vinelandii. These proteins are also homologous to subunits or domains of hydrogenases and NADH-ubiquinone oxidoreductase (Complex I) of respiratory chains. The A. aeolicus [2Fe-2S] protein is thus representative of a presumably novel protein fold involved in a variety of functions in very diverse cellular backgrounds.

Refined X-ray structures of the oxidized, at 1.3 A, and reduced, at 1.17 A, [2Fe-2S] ferredoxin from the cyanobacterium Anabaena PCC7119 show redox-linked conformational changes.

The chemical sequence of the [2Fe-2S] ferredoxin from the cyanobacterium AnabaenaPCC7119 (Fd7119) and its high-resolution X-ray structures in the oxidized and reduced states have been determined. The Fd7119 sequence is identical to that of the ferredoxin from the PCC7120 strain (Fd7120). X-ray diffraction data were collected at 100 K with an oxidized trigonal Fd7119 crystal, at 1.3 A resolution, and with an orthorhombic crystal, previously reduced with dithionite and flash frozen under anaerobic conditions, at 1.17 A resolution. The two molecular models were determined by molecular replacement with the [2Fe-2S] ferredoxin from the strain PCC7120 (Rypniewski, W. R., Breiter, D. R., Benning, M. M., Wesenberg, G., Oh, B.-H., Markley, J. L., Rayment, I., and Holden, H. M. (1991) Biochemistry 30, 4126-4131.) The final R-factors are 0. 140 (for the reduced crystal) and 0.138 (for the oxidized crystal). The [2Fe-2S] cluster appears as a significantly distorted lozenge in the reduced and oxidized redox states. The major conformational difference between the two redox forms concerns the peptide bond linking Cys46 and Ser47 which points its carbonyl oxygen away from the [2Fe-2S] cluster ("CO out") in the reduced molecule and toward it ("CO in") in the oxidized one. The "CO out" conformation could be the signature of the reduction of the iron atom Fe1, which is close to the molecular surface. Superposition of the three crystallographically independent molecules shows that the putative recognition site with the physiological partner (FNR) involves charged, hydrophobic residues and invariant water molecules.

Identification of a structural determinant for resistance to beta-lactam antibiotics in Gram-positive bacteria.

Streptococcus pneumoniae is the main causal agent of pathologies that are increasingly resistant to antibiotic treatment. Clinical resistance of S. pneumoniae to beta-lactam antibiotics is linked to multiple mutations of high molecular mass penicillin-binding proteins (H-PBPs), essential enzymes involved in the final steps of bacterial cell wall synthesis. H-PBPs from resistant bacteria have a reduced affinity for beta-lactam and a decreased hydrolytic activity on substrate analogues. In S. pneumoniae, the gene coding for one of these H-PBPs, PBP2x, is located in the cell division cluster (DCW). We present here structural evidence linking multiple beta-lactam resistance to amino acid substitutions in PBP2x within a buried cavity near the catalytic site that contains a structural water molecule. Site-directed mutation of amino acids in contact with this water molecule in the "sensitive" form of PBP2x produces mutants similar, in terms of beta-lactam affinity and substrate hydrolysis, to altered PBP2x produced in resistant clinical isolates. A reverse mutation in a PBP2x variant from a clinically important resistant clone increases the acylation efficiency for beta-lactams and substrate analogues. Furthermore, amino acid residues in contact with the structural water molecule are conserved in the equivalent H-PBPs of pathogenic Gram-positive cocci. We suggest that, probably via a local structural modification, the partial or complete loss of this water molecule reduces the acylation efficiency of PBP2x substrates to a point at which cell wall synthesis still occurs, but the sensitivity to therapeutic concentrations of beta-lactam antibiotics is lost.

Large-scale preparation, purification, and crystallization of UDP-N-acetylmuramoyl-L-alanine: D-glutamate ligase from Escherichia coli.

The UDP-N-acetylmuramoyl-L-alanine:D-glutamate ligase from Escherichia coli, an enzyme involved in the biosynthesis of the bacterial peptidoglycan monomer unit, was overproduced and purified to homogeneity on a large scale, yielding 4 mg of protein per liter of bacterial culture. Crystals of the complex with the substrate UDP-MurNAc-L-Ala were grown by the hanging drop method using ammonium sulfate as the precipitant. They are tetragonal with cell dimensions a = b = 65.5 A and c = 134.59 A, space group P4(1) or P4(3), and contain one monomer of 46,842 Da in the asymmetric unit. In order to use the multiple-wavelength anomalous diffraction method for phasing, a selenomethionine derivative of the protein has also been overproduced, purified, and crystallized.

Dipeptidyl aminotransferase activity and in vitro O-glycosylation of MUC5AC mucin motif peptides by human gastric microsomal preparations.

The in vitro O-glycosylation reaction of the MUC5AC mucin motif peptide, TTSAPTTS (in one-letter code), was achieved with human gastric microsomal homogenates. The analyses using capillary electrophoresis online coupled with electrospray mass spectrometry and further Edman degradation of the purified products (obtained by capillary electrophoresis at preparative scale) allowed us to distinguish two components at close masses: the addition of a mass of 202 corresponded to an N-terminal elongation of the peptide TTSAPTTS with the dipeptide (TT) and the addition of a mass of 203 corresponded to an N-acetylgalactosamine O-linkage. Using different peptidase inhibitors, a dipeptidyl peptidase/transferase activity was further characterized. A thiol dependence and an inhibition by H-Gly-PheCHN2 (specific to cathepsin C activity) were found. Moreover, besides TTSAPTTS, other MUC5AC motif peptides (GTTPSPVP, TSAPTTS) were also dipeptide donors (GT and TS, respectively) and our results suggested the involvement of a single dipeptidyl peptidase/transferase activity. Finally, this latter activity modified the in vitro GalNAc incorporation rates when using our selected MUC5AC motif peptides. Our study therefore shows that caution must be taken to prevent peptidic substrate elongation while performing in vitro O-glycosylation with microsomal preparations as the enzyme source. In fact, the results of the N-acetylgalactosamine incorporation rates and thus the microsomal N-acetylgalactosamine transferase affinity can be misinterpreted if dipeptidyl peptidase/transferase activity is not inhibited by the thiol inhibitor E-64 or the cathepsin C inhibitor H-Gly-PheCHN2.

Polypeptide:N-acetylgalactosaminyltransferase activities towards the mucin MUC5AC peptide motif using microsomal preparations of normal and tumoral digestive mucosa.

The selected-acceptor substrate peptide (TTSAPTTS), deduced from the human mucin gene MUC5AC (expressed essentially in the human gastric and tracheobronchial mucosa), was used to assay polypeptide:N-acetylgalactosaminyltransferases (GalNAc transferases) of different microsomal preparations, obtained from gastric and colonic mucosa in normal and tumoral situations. The O-glycosylated products, analyzed by capillary electrophoresis and electrospray mass spectrometry, showed a variable number of GalNAc O-linked to the different hydroxy amino acids of TTSAPTTS, depending on the tissue studied. Our observations were consistent with the existence of more than one form of GalNAc transferases which were expressed differentially in the gastrointestinal tract (stomach and/or colon). The levels of enzyme activities showed a tissue-specific pattern as they were high in normal colonic tissue and low in colon cancer. On the other hand, in the tumoral gastric tissue (displaying intestinal metaplasia) a high level of GalNAc transferase activities was obtained, similar to that found in the normal colon. Moreover, slight discrepancies (activities and number of O-linked GalNAc) were only detected between normal gastric and tumoral colonic preparations. Thus, the data indicated that the dedifferentiation of the gastric cancer tissue may induce GalNAc transferase activities similar to those in the normal colonic, tissue and that colonic and gastric tissues may contain families of glycosyltransferases involved specifically in reaction towards particular peptide or protein substrates. In addition, the analysis by capillary electrophoresis and electrospray mass spectrometry revealed, in tumoral gastric as well as in normal colonic tissues, a high dipeptidylaminotransferase activity inducing an elongation of TTSAPTTS by dithreonine. This activity was low in normal gastric and tumoral colonic tissues.

The protein sequence of an archaeal catalase-peroxidase.

The gene encoding a catalase-peroxidase of archaeal origin, the halophilic catalase-peroxidase from Haloarcula marismortui, was sequenced. The primary structure proposed was confirmed by Edman degradation and mass spectrometry analyses of proteolytic fragments of the purified protein. The open reading frame in the gene corresponds to 731 amino acids and the calculated mass of the mature protein (deleted of the N-terminal methionine) is 81,253.65 Da, in reasonable agreement with the value of 81,292 +/- 9 Da previously measured by mass spectrometry. Southern and Northern blot analyses showed that the protein is encoded by a single gene as a monocistronic transcript. The protein sequence shows a high level of identity with bacterial catalase-peroxidases, with strongly conserved regions around the heme binding histidines. Similarly to other soluble halophilic proteins, it shows the excess of acidic residues that has been associated with solvation in halophilic adaptation.

Further characterization of the two tetraheme cytochromes c3 from Desulfovibiro africanus: nucleotide sequences, EPR spectroscopy and biological activity.

The genes encoding the basic and acidic tetraheme cytochromes c3 from Desulfovibrio africanus have been sequenced. The corresponding amino acid sequences of the basic and acidic cytochromes c3 indicate that the mature proteins consist of a single polypeptide chain of 117 and 103 residues, respectively. Their molecular masses, 15102 and 13742 Da, respectively, determined by mass spectrometry, are in perfect agreement with those calculated from their amino acid sequences. Both D. africanus cytochromes c3 are synthesized as precursor proteins with signal peptides of 23 and 24 residues for the basic and acidic cytochromes, respectively. These cytochromes c3 exhibit the main structural features of the cytochrome c3 family and contain the 16 strictly conserved cysteine + histidine residues directly involved in the heme binding sites. The D. africanus acidic cytochrome c3 differs from all the other homologous cytochromes by its low content of basic residues and its distribution of charged residues in the amino acid sequence. The presence of four hemes per molecule was confirmed by EPR spectroscopy in both cytochromes c3. The g-value analysis suggests that in both cytochromes, the angle between imidazole planes of the axial histidine ligands is close to 90 degrees for one heme and much lower for the three others. Moreover, an unusually high exchange interaction (approximately 10[-2] cm[-1]) was evidenced between the highest potential heme (-90 mV) and one of the low potential hemes in the basic cytochrome c3. The reactivity of D. africanus cytochromes c3 with heterologous [NiFe] and [Fe] hydrogenases was investigated. Only the basic one interacts with the two types of hydrogenase to achieve efficient electron transfer, whereas the acidic cytochrome c3 exchanges electrons specifically with the basic cytochrome c3. The difference in the specificity of the two D. africanus cytochromes c3 has been correlated with their highly different content of basic and acidic residues.

Assembly of a [2Fe-2S]2+ cluster in a molecular variant of Clostridium pasteurianum rubredoxin.

The rubredoxin from Clostridium pasteurianum contains a single iron atom bound to the polypeptide chain by cysteines 6, 9, 39, and 42. The C42A variant of this protein has been prepared by site-directed mutagenesis and heterologous expression of the gene in Escherichia coli. The mutated protein was found to contain an unexpected chromophore that has been characterized by a variety of techniques. UV-visible absorption and resonance Raman spectra were strongly reminiscent of those of [2Fe-2S] proteins. Mössbauer spectra of the oxidized chromophore isolated in oxygen-free conditions indicated low-temperature diamagnetism resulting from antiferromagnetically coupled high-spin ferric ions. Analysis of X-ray absorption fine structure spectra yielded an Fe-Fe distance of 2.68 A. Colorimetric assays of iron and inorganic sulfide showed that the two elements are present in a 1:1 ratio. Electrospray-ionization mass spectra displayed a major component at M = 6190 Da, i.e. the molecular mass of the C42A apoprotein plus two atomic masses of iron and two atomic masses of sulfur. Taken together, these data show that a mere point mutation allows the stabilization of a binuclear [2Fe-2S] cluster in a protein that normally accommodates a mononuclear Fe(Scys)4 site. Assembly of a [2Fe-2S] cluster may occur because rubredoxin assumes a similar fold around its metal center as the [2Fe-2S] Rieske protein. Alternatively, a more extensive structural rearrangement of the polypeptide chain of the C42A rubredoxin variant may be considered as well.

Biochemical and spectroscopic characterization of two new cytochromes isolated from Desulfuromonas acetoxidans.

The multimeric cytochromes described to date in sulfate- and sulfur-reducing bacteria are associated with diverse respiratory modes involving the use of elemental sulfur or oxidized sulfur compounds as terminal acceptors. They exhibit no structural similarity with the other cytochrome c classes and are characterized by a bis-histidinyl axial iron coordination and low redox potentials. We have purified two new cytochromes c with markedly different molecular masses (10 000 and 50 000) from the bacterium Desulfuromonas acetoxidans, which uses anaerobic sulfur respiration as its sole energy source. The characterization by electrochemistry and optical and EPR spectroscopies revealed the cytochrome c (Mr = 10 000) to be the first monohemic cytochrome c exhibiting a bis-histidinyl axial coordination and a low redox potential (-220 mV). The cytochrome c (Mr = 50 000) contains four hemes of low potential (-200, -210, -370, and -380 mV) with the same axial coordination. The N-terminal amino acid sequences were compared with that of the trihemic cytochrome c7, previously described in D. acetoxidans and which is related to tetrahemic cytochrome c3 from sulfate reducing bacteria. Some homology was found between cytochrome c (Mr = 10 000) and cytochrome c7. Both D. acetoxidans cytochromes c are located in the periplasmic space and their biochemical and spectroscopic properties indicate that they belong to the class III cytochromes.

Probing the influence of mutations on the stability of a ferredoxin by mass spectrometry.

Hydrogen/deuterium exchange, which depends on solvent accessibility, can be probed by mass spectrometry (MS) to get information on protein conformation or protein-ligand interaction. In this work, the conformational properties of the cyanobacterium Anabaena wild-type ferredoxin as well as of two single-site mutants (Phe 65 Ala and Arg 42 Ala) were studied. After incubation of the wild type and mutant proteins in deuterated water and quenching of the exchange at low pH, the proteins were rapidly digested at high enzyme-to-substrate ratio using immobilized pepsin, and the resulting peptides were characterized using ESI-MS. We have identified specific regions for which the H-bonding or solvent accessibility properties were perturbed by the mutations. These results show that this approach can provide local information on the influence of mutations, even for a highly structured protein like ferredoxin, and sometimes in regions distant from the mutation point.

Purification, crystallization and preliminary X-ray analysis of Bacteroides fragilis Zn2+ beta-lactamase.

The Zn(2+) beta-lactamase from Bacteroides fragilis (E.C. 3.5.2.6) was overexpressed in Escherichia coli using an isopropylthiogalactoside-inducible T7 RNA polymerase expression system. Crystallization trials by the hanging-drop vapour-diffusion method have yielded two different crystal forms from two slightly different conditions. Crystals of form I belong to the monoclinic space group C2 with unit-cell dimensions a = 56.03, b = 43.98, c = 105.32 A, beta = 112 degrees and diffracted only up to 4.0 A. Crystals of form II are orthorhombic, space group P2(1)2(1)2(1) with unit-cell dimensions a = 48.10, b = 98.05, c = 111.76 A, diffract to at least 2.0 A and are suitable for high-resolution structural analysis.

Chemical synthesis and characterization of the epidermal growth factor-like module of human complement protease C1r.

C1r is one of the two serine proteases of C1, the first component of complement, in which it is associated in a calcium-dependent manner to the homologous serine protease C1s. This interaction is mediated by the N-terminal region of C1r, which comprises a single epidermal growth factor (EGF)-like module containing the consensus sequence required for calcium binding, surrounded by two CUB modules. With a view to determine the structure of the EGF-like module of C1r and evaluate its contribution to calcium binding, this module [C1r(123-175)] was synthesized by automated solid-phase methodology using the Boc strategy. A first synthesis using the Boc-His(Z) derivative gave very low yield, due to partial deprotection of His residues leading to chain termination by acetylation, and to insertion of glycine residues. This could be circumvented by using the Boc-His(DNP) derivative and by condensation of appropriate glycine-containing segments. The synthetic peptide was efficiently folded under redox conditions to the species with three correct disulfide bridges, as determined by mass spectrometry and N-terminal sequence analyses of thermolytic fragments. The homogeneity of the synthetic peptide was assessed by reversed-phase HPLC and electrospray mass spectrometry. One-dimensional 1H NMR spectroscopic analysis provided evidence that the EGF-like module had a well defined structure, and was able to bind calcium with an apparent Kd of 10 mM. This value, comparable to that found for the isolated EGF-like modules of coagulation factors IX and X, is much higher than that measured for native C1r. As already proposed for factors IX and X, it is suggested that neighbouring module(s), most probably the N-terminal CUB module, contribute(s) to the calcium binding site.

Composite versus multichannel binary phase-only filtering.

Multichannel filtering and its inherent capacity for the implementation of data-fusion algorithms for high-level image processing, as well as composite filtering and its capacity for distortion-invariant pattern-recognition tasks, are discussed and compared. Both approaches are assessed by use of binary phase-only filters to simplify implementation issues. We discuss similarities and differences of these two solutions and demonstrate that they can be merged efficiently, giving rise to a new category of filters that we call composite-multichannel filters. We illustrate this comparison and the new filter design for the case of rotation-invariant fingerprint recognition. In particular, we show that the gain in terms of encoding capacity in the case of the composite-multichannel approach can be used efficiently to introduce multichannel-filter reconfigurability.

Can immunoglobulin C(H)1 constant region domain modulate antigen binding affinity of antibodies?

Although the switch process is frequently associated with affinity maturation, the constant region is not assumed to play a role in Ag-Ab binding. In the present work, we demonstrate that two clonally related human monoclonal Igs sharing identical V(H) and V(L) sequences, but expressing different isotypes (IgA1kappa(PER) and IgG1kappa(PER)), bind tubulin with significantly different affinities. This difference was mainly accounted for by a disparity in the association rate constants. These results suggest that affinity maturation of this clone could be achieved through class switching in the absence of further somatic mutations. Since the differences observed were found at the Fab level, they also suggest a role for the C(H)1 domain in structuring the Ag-binding site into a more kinetically competent form.