Search

In analogy to the cystine knots present in natural collagens, a simplified disulfide cross-link was used to analyse the conformational effects of a C-terminal artificial cystine knot on the folding of collagenous peptides consisting of solely (Pro-Hyp-Gly) repeating units. Assembly of the alpha chains into a heterotrimer by previously applied regioselective disulfide-bridging strategies failed because of the high tendency of (Pro-Hyp-Gly)(5) peptides to self-associate and form homotrimers. Only when side-chain-protected peptides were used, for example in the Hyp(tBu) form, and a new protection scheme was adopted, selective interchain-disulfide cross-linking into the heterotrimer in organic solvents was successful. This unexpected strong effect of the conformational properties on the efficiency of well-established reactions was further supported by replacing the Hyp residues with (4S)-fluoroproline, which is known to destabilise triple-helical structures. With the related [Pro-(4S)-FPro-Gly](5) peptides, assembly of the heterotrimer in aqueous solution proceeded in a satisfactory manner. Both the intermediates and the final fluorinated heterotrimer are fully unfolded in aqueous solution even at 4 degrees C. Conversely, the disulfide-crossbridged (Pro-Hyp-Gly)(5) heterotrimer forms a very stable triple helix. The observation that thermal unfolding leads to scrambling of the disulfide bridges was unexpected. Although NMR experiments support an extension of the triple helix into the cystine knot, thermolysis is not associated with the unfolding process. In fact, the unstructured fluorinated trimer undergoes an equally facile thermodegradation associated with the intrinsic tendency of unsymmetrical disulfides to disproportionate into symmetrical disulfides under favourable conditions. The experimental results obtained with the model peptides fully support the role of triple-helix nucleation and stabilisation by the artificial cystine knot as previously suggested for the natural cystine knots in collagens.

The aim of this paper is to present a unifying view of four SVD-neural-computation techniques found in the scientific literature and to present some theoretical results on their behavior. The considered SVD neural algorithms are shown to arise as Riemannian-gradient flows on double Stiefel manifold and their geometric and dynamical properties are investigated with the help of differential geometry.

We present an overview of independent component analysis, an emerging signal processing technique based on neural networks, with the aim to provide an up-to-date survey of the theoretical streams in this discipline and of the current applications in the engineering area. We also focus on a particular application, dealing with a remote sensing technique based on synthetic aperture radar imagery processing: we briefly review the features and main applications of synthetic aperture radar and show how blind signal processing by neural networks may be advantageously employed to enhance the quality of remote sensing data.

Collagen type IV is a specialized form of collagen that is found only in basement membranes. It is involved in integrin-mediated cell-adhesion processes, and the responsible binding sites for the alpha1beta1 integrin cell receptor have been identified as Asp461 of the two alpha1 chains and Arg461 of the alpha2 chain. In the most plausible stagger of native collagen type IV the alpha2 chain is the tailing one. This has recently been confirmed by the differentiated binding affinities of synthetic heterotrimeric collagen peptides in which the chains were staggered in this native register as well as in the less plausible alpha1alpha2alpha1' register with an artificial cystine knot. In the present work, two heterotrimeric collagen peptides with chain registers identical to the previous ones were synthesized for fluorescence resonance energy transfer and emission anisotropy measurements, exploiting the native Phe464 in the alpha2 chain as donor and an Ile467Tyr mutation in the alpha1' chain as acceptor fluorophore. This fluorophore pair allowed extraction of more detailed information on the conformational properties of the cell-adhesion epitope incorporated into the central part of the trimeric collagen model peptides. A comparison of the experimentally derived values of the interfluorophore distance and of the orientation factor kappa(2) with the values extracted from the molecular model of the trimer in the native stagger confirmed a triple-helical structure of the adhesion-site portion at low temperature. The thermal unfolding of this central domain was specifically monitored by emission anisotropy, allowing unambiguous assignment of the three structural domains of the trimeric collagen molecules detected by microcalorimetry, with the integrin binding site as the portion of weakest triple-helical stability flanked by two more stable triple-helical regions. The results are consistent with the picture of a conformational microheterogeneity as the responsible property for selective recognition of collagens by interacting proteins.

This article investigated the behavior of a single-input, single-unit neuron model of the Bell-Sejnowski class, which learn through the maximum-entropy principle, in order to understand its probability density function and matching ability.

In the present paper we investigate iterative minor subspace analysis computation by describing a neural approach based on weight flow on Stiefel manifold and by discussing four neural algorithms and a purely algebraic algorithm known from the scientific literature. A comparison of numerical experimental results and computational complexity estimates confirms the effectiveness and efficiency of the proposed approach.

Collagens contain sequence- and conformation-dependent epitopes responsible for their digestion by collagenases at specific loci. A synthetic heterotrimer construct containing the collagenase cleavage site of collagen type I was found to mimic perfectly native collagen in terms of selectivity and mode of enzymatic degradation. The NMR conformational analysis of this molecule clearly revealed the presence of two structural domains, i.e. a triple helix spanning the Gly-Pro-Hyp repeats and a less ordered portion corresponding to the collagenase cleavage site where the three chains are aligned in extended conformation with loose interchain contacts. These structural properties allow for additional insights into the very particular mechanism of collagen digestion by collagenases., ((c) 2002 Elsevier Science Ltd.)

The active-site hexapeptides of glutaredoxin (Grx), thioredoxin (Trx), protein disulfide isomerase (PDI), and thioredoxin-reductase (Trr) containing the common motif Cys-Xaa-Yaa-Cys were conformationally restricted by backbone cyclization, and their redox potentials were found to increase in the rank order of Trr < Grx < Trx < PDI peptide, with E'(0) values ranging between -204 mV and -130 mV. In each peptide the thiol pK(a) of one Cys residue was found to be lower than the other (e.g., 7.3 against 9.6 in the PDI peptide). Both the yield and rate of refolding of reduced RNase A in the presence of the bis(cysteinyl)peptides increased with the oxidizing character of the cyclic compounds. These results show that small peptides can function as adjuvants for the in vitro oxidative folding of proteins.

The present paper aims at introducing the concepts and mathematical details of unsupervised neural learning with orthonormality constrains. The neural structures considered are single non-linear layers and the learnable parameters are organized in matrices, as usual, which gives the parameters spaces the geometrical structure of the Euclidean manifold. The constraint of orthonormality for the connection-matrices further restricts the parameters spaces to differential manifolds such as the orthogonal group, the compact Stiefel manifold and its extensions. For these reasons, the instruments for characterizing and studying the behavior of learning equations for these particular networks are provided by the differential geometry of Lie groups. In particular, two sub-classes of the general Lie-group learning theories are studied in detail, dealing with first-order (gradient-based) and second-order (non-gradient-based) learning. Although the considered class of learning theories is very general, in the present paper special attention is paid to unsupervised learning paradigms.

The aim of this paper is to present an efficient implementation of unsupervised adaptive-activation function neurons dedicated to one-dimensional probability density estimation, with application to independent component analysis. The proposed implementation is a computationally light improvement to adaptive pseudo-polynomial neurons, recently presented in Fiori, S. (2000a). Blind signal processing by the adaptive activation function neurons. Neural Networks, 13(6), 597-611, and is based upon the concept of 'look-up table' (LUT) neurons.

In a recent work, we introduced the concept of pseudo-polynomial adaptive activation function neuron (FAN) and presented an unsupervised information-theoretic learning theory for such structure. The learning model is based on entropy optimization and provides a way of learning probability distributions from incomplete data. The aim of the present paper is to illustrate some theoretical features of the FAN neuron, to extend its learning theory to asymmetrical density function approximation, and to provide an analytical and numerical comparison with other known density function estimation methods, with special emphasis to the universal approximation ability. The paper also provides a survey of PDF learning from incomplete data, as well as results of several experiments performed on real-world problems and signals.

The aim of this paper is to study an Information Theory based learning theory for neural units endowed with adaptive activation functions. The learning theory has the target to force the neuron to approximate the input-output transference that makes it flat (uniform) the probability density function of its output or, equivalently, that maximizes the entropy of the neuron response. Then, a network of adaptive activation function neurons is studied, and the effectiveness of the new structure is tested on Independent Component Analysis (ICA) problems. The new ICA neural algorithm is compared with the closely related 'Mixture of Densities' (MOD) technique by Xu et al.. Both simulation results and structural comparison show the new method is effective and more efficient in computational complexity.

By replacing two cysteine residues in apamin with selenocysteine, the three possible isomers related to the side-chain connectivities of a bis-cystinyl-peptide were synthesized in regioselective manner exploiting the low redox potential of the diselenide bond. Nuclear magnetic resonance conformational analysis of monoselenocystine analogue apamin with the natural diselenide/disulfide network confirmed the highly isomorphous character of the sulfur replacement with selenium despite its slightly larger atomic radius and increased bond lengths. The comparative conformational analysis of the apamin analogues containing the non-natural side-chain links with wild type apamin clearly revealed retention of the main structural fold and thus the high propensity of these small molecules to adopt the secondary structure elements present in natural apamin. These findings offered interesting hints for a better understanding of the oxidative refolding pathway of the bis-cystinyl peptide that leads exclusively to the correct natural isomer., (Copyright 2000 John Wiley & Sons, Inc.)

The newly discovered endomorphin-1 (Tyr-Pro-Trp-Phe-NH2) and endomorphin-2 (Tyr-Pro-Phe-Phe-NH2) are potent opioid peptides with the highest affinity and selectivity for the mu receptor among all known endogenous ligands. To investigate a possible correlation between these biological properties and the conformational preferences of the small peptides, a comparative structural analysis was performed of endomorphin-1 in aqueous buffer and in membrane-mimicking SDS and AOT normal and reverse micelles by the use of CD, FT-IR, fluorescence and(1)H-NMR spectroscopy. It is well established for opioid peptides that, independently of the receptor selectivity, the Tyr1 residue plays the role of the primary pharmacophore and that the orientation of the second aromatic pharmacophore relative to the tyrosine side-chain dictates the mu or delta-receptor selectivity. By varying the environment of endomorphin-1 from water to the amphipathic SDS micelles and even more efficiently to the AOT reverse micelles, the display of the aromatic side-chains changes from an interaction of the Tyr1 and Phe4 residues to a switch of the Trp3 indole group into close contact with the phenolic moiety to prevent this type of interaction and to force an orientation of the Phe4 side-chain into the opposite direction. This conformational switch is accompanied by a stabilization of the cis -Pro2 isomer and the resulting spatial array of the pharmacophoric groups correlate well with the structural model of mu receptor-bound opioid peptides. The results indicate that AOT reverse micelles with a woof 10, where almost exclusively ordered water is secluded in the cavity, constitute with their electrostatic and hydrophobic potential an excellent mimetic of amphipathic surfaces as present on lipid bilayers and on ligand-recognition and ligand-binding sites of proteins., (Copyright 1999 Academic Press.)

The sequence of apamin, an 18 residue bee venom toxin, encloses all the information required for the correct disulfide-coupled folding into the cystine-stabilized alpha-helical motif. Three apamin analogs, each containing a pair of selenocysteine residues replacing the related cysteines, were synthesized to mimic the three possible apamin isomers with two crossed, parallel, or consecutive disulfides, respectively. Refolding experiments clearly revealed that the redox potential of selenocysteine prevails over the sequence encoded structural information for proper folding of apamin. Thus, selenocysteine can be used as a new device to generate productive and nonproductive folding intermediates of peptides and proteins. In fact, disulfides are selectively reduced in presence of the diselenide and the conformational features derived from these intermediates as well as from the three-dimensional (3D) structures of the selenocysteine-containing analogs with their nonnatural networks of diselenide/disulfide bridges allowed to gain further insight into the subtle driving forces for the correct folding of apamin that mainly derive from local conformational preferences.

The aim of this paper is to present a study of polynomial functional-link neural units that learn through an information-theoretic-based criterion. First the structure of the neuron is presented and the unsupervised learning theory is explained and discussed, with particular attention being paid to its probability density function and cumulative distribution function approximation capability. Then a neural network formed by such neurons (the polynomial functional-link artificial neural network, or PFANN) is shown to be able to separate out linearly mixed eterokurtic source signals, i.e. signals endowed with either positive or negative kurtoses. In order to compare the performance of the proposed blind separation technique with those exhibited by existing methods, the mixture of densities (MOD) approach of Xu et al, which is closely related to PFANN, is briefly recalled; then comparative numerical simulations performed on both synthetic and real-world signals and a complexity evaluation are illustrated. These results show that the PFANN approach gives similar performance with a noticeable reduction in computational effort.

Air re-oxidation of fully reduced human endothelin-1 under optimized conditions yields the natural isomer with parallel disulfide bridges and the non-natural isomer with crossed disulfide bridges at a ratio of 3:1. In view of the recently determined highly reducing redox potential of selenocysteine (-381 mV) in peptides, the half-cystine residues Cys3 and Cys11 of the natural isomer of endothelin-1 were replaced by selenocysteine. Taking advantage of the high stability of the diselenide group toward reducing agents for disulfides a regioselective disulfide bridging of the second cysteine pair allowed for straightforward preparation of the [Sec3,Sec11, Nle7]-endothelin-1. NMR structural analysis showed conformational preferences of this endothelin analog that were identical to those of the natural hormone. Similarly, the bioactivity data confirmed that replacement of cysteine residues with selenocysteine was without detectable effect on receptor recognition and signal transduction. Both findings strongly support that the exchange of sulfur against selenium produces a fully isomorphous molecule as recently observed for similar exchanges at the level of methionine residues in proteins. Moreover, oxidative refolding of the fully reduced [Sec3,Sec11,Nle7]-endothelin-1 fulfilled the expectation that the redox potential of the selenocysteines would dictate quantitative formation of the natural isomer. These results suggest that the selenocysteine approach, besides offering an interesting chemical tool for induction of correct oxidative folding of multiple cysteine-containing peptides, should even allow for the preparation of non-natural isomers and thus for studying conformational preferences of folding intermediates in peptides and proteins., (Copyright 1998 Academic Press)

The conformation of segments corresponding to the four alpha-helical stretches found in human granulocyte-macrophage colony-stimulating factor was studied in water solution in the presence of different amounts of 2,2,2-trifluoroethanol (TFE). The CD spectra reveal the onset of secondary structure upon addition of TFE. The final amount of helical conformation varies among the four peptides. In all cases, the conformational transition is complete before 50% TFE (v/v). 1H-NMR studies were conducted at this solvent composition, leading to the assignment of all the resonances and to the definition of the secondary structure for all four fragments.

This article presents an easy technique to enable the dentist or dental laboratory technician to transfer the path of insertion from the diagnostic cast to multiple working casts.