Eukaryotic porins are a group of membrane proteins whose best known role is to form an aqueous pore channel in the mitochondrial outer membrane. As opposed to the bacterial porins (a large family of protein whose 3D structure has been determined by X-ray diffraction), the structure of eukaryotic porins (also termed VDACs, voltage-dependent anion-selective channels) is still a matter of debate. We analysed the secondary structure of VDAC from the yeast Saccharomyces cerevisiae, the fungus Neurospora crassa and the mouse with different types of neural network-based predictors. The predictors were able to discriminate membrane L L-strands, globular K K-helices and membrane K K-helices and localised, in all three VDAC sequences, 16 L L-strands along the chain. For all three sequences the N-terminal region showed a high propensity to form a globular K K-helix. The 16 L L-strand VDAC motif was thus aligned to a bacterial porin-derived template containing a similar 16 L L-strand motif. The alignment of the VDAC sequence with the bacterial porin sequence was used to compute a set of 3D coordinates, which constitutes the first 3D prediction of a eukaryotic porin. All the predicted structures assume a L L-barrel structure composed of 16 L L-strands with the N-terminus outside the membrane. Loops are shorter in this side of the membrane than in the other, where two long loops are protruding. The shape of the pore varies between almost circular for Neurospora and mouse and slightly oval for yeast. Average values between 3 and 2.5 nm at the C-carbon backbone are found for the diameter of the channels. In this model VDAC shows large portions of the structure exposed on both sides of the membrane. The architecture we determine allows speculation about the mechanism of possible interactions between VDAC and other proteins on both sides of the mitochondrial outer membrane. The computed 3D model is consistent with most of the experimental results so far reported. ß 2002 Published by Elsevier Science B.V. on behalf of the Federation of European Biochemical Societies.

A method based on neural networks is trained and tested on a nonredundant set of ␤-barrel membrane proteins known at atomic resolution with a jackknife procedure. The method predicts the topography of transmembrane ␤ strands with residue accuracy as high as 78% when evolutionary information is used as input to the network. Of the transmembrane ␤-strands included in the training set, 93% are correctly assigned. The predictor includes an algorithm of model optimization, based on dynamic programming, that correctly models eight out of the 11 proteins present in the training/testing set. In addition, protein topology is assigned on the basis of the location of the longest loops in the models. We propose this as a general method to fill the gap of the prediction of ␤-barrel membrane proteins.

The VP2 genes of Italian canine parvovirus (CPV) type 2 strains isolated from dogs and wolves were sequenced and a three-dimensional model of the VP2 capsid protein was constructed. Two mutations were detected in the VP2 sequences of the Italian strains : one at residue 297 and one at residue 265. Variant 297 is the predominant CPV isolate in Europe, whereas variant 265 has never been detected before. The mutation at residue 265 causes a disruption in a G strand of the β-barrel in the VP2 protein. Data on strains isolated from wolves demonstrated that the same strain of CPV can circulate among domestic and wild canids ; therefore, this result leads us to exclude the possibility that a separate parvovirus pool exists in wild populations.Canine parvovirus type 2 (CPV-2) is an important pathogen in domestic dogs and several wild carnivore species. It was first identified in USA in 1978 (Appel et al., 1979) and was found later to have spread worldwide in domestic and wild canid populations.After its initial appearance, it was shown that antigenic drift continuously changes the antigenicity of CPV : the original CPV-2 strain has been completely replaced by the newer antigenic types CPV-2a and CPV-2b , which have also extended their host range to include cats (Mochizuki et al., 1996). The new types of CPV differ from the original type 2 strain in that there are some nucleotide changes (positions 3045, 3685, 3699, 4062 and 4449) in the gene encoding the VP2 coat protein Truyen et al., 1995). Sequences important for the determination of antigenic type and for the control of host range are located in the VP2 capsid protein (Parrish, 1991 ;Chang et al., 1992).

The most stringent test for predictive methods of protein secondary structure is whether identical short sequences that are known to be present with different conformations in different proteins known at atomic resolution can be correctly discriminated. In this study, we show that the prediction efficiency of this type of segments in unrelated proteins reaches an average accuracy per residue ranging from about 72 to 75% (depending on the alignment method used to generate the input sequence profile) only when methods of the third generation are used. A comparison of different methods based on segment statistics (2nd generation methods) and/or including also evolutionary information (3rd generation methods) indicate that the discrimination of the different conformations of identical segments is dependent on the method used for the prediction. Accuracy is similar when methods similarly performing on the secondary structure prediction are tested. When evolutionary information is taken into account as compared to single sequence input, the number of correctly discriminated pairs is increased twofold. The results also highlight the predictive capability of neural networks for identical segments whose conformation differs in different proteins. Proteins 2000;41:535–544. © 2000 Wiley‐Liss, Inc.

No abstract