Prediction of transmembrane segment

Persson, B., and Argos, P. (1994). Prediction of transmembrane segments in proteins utilising multiple sequence alignments./. Mol. Biol. 237, 182-192. 

False-positive predictions of transmembrane segments are not found in porins (no transmembrane helical segments predicted) and are rare in soluble proteins (11.5%),... 

Communication between a cell and other cells in its surroundings is based almost exclusively on proteins that are embedded in the cell s membrane. Many proteins pass through the cellular membrane and can therefore interact with molecules on the intracellular side of the membrane as well as with molecules on the extracellular side. These transmembrane proteins and their molecular mechanisms are of particular interest in biomedicine (Ofiran and Rost, 2005). It is particularly difficult to decipher the structure of transmembrane segments (helices) of proteins experimentally, which makes in silico prediction particularly valuable. 

Wo, Z. G., and Oswald, R. E. (1995a). A topological analysis of goldfish kainate receptors predicts three transmembrane segments./. Biol. Chem. 270, 2000-2009. 

Fig. 4.1. Topology, conservation, andRHPmotif oftheStel4poI5. cerevisiae. (A) Hydropathy plots predict six transmembrane segments (TMs). In this model, the N-and C-termini are disposed toward the cytosol. TM 5 and 6 are proposed to form a helix-turn-helix hehcal hairpin within the membrane [25]. Fifteen unique Icmt protein sequences were ahgned using ClustalW 2.0.1.1 [27]. The blue residues denote amino acid identity and the magenta residues denote amino acid similarity. The C-terminal portion of the enzyme (136-239) contains the majority of the identical amino acids. (B) Sequence of the RHP motif, a C-terminal consensus sequence common to Icmt enzymes, a number of bacterial open reading frames, and two phosphatidyl-ethanolamine methyltransferases. Numbers denote the amino acid position in Stel4p.

The conformation of TMS seems to be a-helical for most membrane proteins [12,13,25-27], but there are some proteins, such as porins, that have TMS in the P-sheet conformation [28]. There may also exist proteins with both helical and P-strand transmembrane segments or with transmembrane helical segments combined with still unknown topology of membrane buried P-strands [29-31]. This work is focused on the prediction of transmembrane helical segments (TMH), but our algorithms do allow the prediction of transmembrane or surface attached P-strands (TMBS) as well... 

By using our standard training procedure the tests were performed on membrane proteins of known or partially known structure with transmembrane (i-strands and on soluble proteins of known structure. For seven tested porins and two defensins (PORINS, Methods) we tested 12 best scales used in Table 7. Only the scale 4 [59] predicted one transmembrane segment in the a-helix conformation (residues 119 to 133 in the porin sequence from Rhodobacter capsulatus). For two different sets of soluble proteins SOLUl and SOLU2 (Methods) prediction results are collected in Table 8 as percentage of proteins falsely predicted to be membrane proteins. The best scales for TMH prediction in membrane proteins still falsely predicted 11-12% of soluble proteins as being membrane proteins with at least one transmembrane helix. 

The observation that conformational preferences are specified by the contexts - local segment primary structure, amino acid attributes, the three-dimensional environment in protein and environmental media, has been discussed before [102-105] Algorithms that do take into account context-dependence of preferences [106] generally perform better for secondary structure prediction In this report simple mathematical representation of context dependence is obtained through preference flmctions that are analytical fiinctions of the surrounding sequence hydrophobicity or of any other amino acid attribute. Furthermore, preference functions are used to predict secondary structure motifs. It has turned out that for integral membrane proteins preference functions are excellent predictors of transmembrane segments in helical conformation In fact preference functions are much better predictors than the hydrophobicity scale chosen to extract these functions. 

The main goal of this work was accurate prediction of transmembrane helical structures, but we do realize that membrane proteins may exist that have both a-helices and P-strands as transmembrane structure. Preference function method is capable of predicting separately a-helical and P-strand conformation of segments that have potential to become... 

Accurate prediction of transmembrane helical segments is superimposed on the prediction of all other secondary structure elements of interest. 

In predicting the transmembrane segments, we distinguish the numbers of residues that are associated with positive correct prediction p, negative correct prediction n, underprediction u, and overprediction o. Out of these four data it is possible to define and compute many performance parameters. We consider as the best parameter the performance parameter A2 defined by Ponnuswamy and Gromiha ... 

---