A good introduction into membrane proteins and their structural organization can be found in the course "Principles of Protein Structure Using the Internet" at the Birkbeck University of London.
Membrane proteins constitute a significant portion of all proteins. The first complete genome sequences allowed to estimate the percentage of cytoplasmic membrane proteins; for Escherichia coli a value of 20 % was predicted (Boyd et al., 1998). Erik Wallin and Gunnar von Heijne (1998) estimated that 20 to 30 % of all Open Reading Frames (ORFs) code for integral alpha-helical membrane proteins (cytoplasmic membrane proteins). Similar values were suggested by David T. Jones (1998) upon analysis of five genomes using MEMSAT. Fifteen single cell organisms were analyzed using the program SOSUI and led to the prediction of 15-20 % polytopic membrane proteins (Mitaku et al., 1999). Die Verfügbarkeit ganzer Genomsequenzen erlaubte auch die Beantwortung der Frage, ob komplexe Organismen einen höheren Anteil an Membranproteinen aufweisen als niedere Organismen. Diese Frage konnte verneint werden (Stevens and Arkin, 2000).
Membrane proteins fulfill a multitude of functions, such as signal transduction or transport of substrates. Milton Saier suggested a procedure to classify transport systems (Saier, 1998). Von Ian Paulsen und Mitarbeitern stammt ein genomischer Vergleich von Membrantransportsystemen (Paulsen et al., 2000).
The following websites predict membrane-spannening protein segments, using different algorithms. The lab of Gunnar von Heijne developed several algoriths to predict the topology of cytoplasmic membrane proteins and tested the corresponding rules experimentally. Several comparisons of different algorithms are found below in the list of references.
Several studies have shown that the combination of different algorithms can improve the prediction reliability:
The polypeptide chain of cytoplasmic membrane proteins traverses the lipid bilayer in the form of α helices. By contrast, the polypeptide chain of bacterial outer membrane proteins traverses the lipid bilayer in the form of β strands (Koebnik et al., 2000). Estimations using the program Hunter suggest that outer membrane proteins constitute 1.5 % to 2.4 % of the proteome of Gram-negative bacteria (Casadio et al., 2003).
Membrane-spanning β strands are less easy to predict and existing algorithms are less precise for this class of membrane proteins. Still, some of them are listet below:
Horst Vogel and Fritz Jähnig (1986)
Tilman Schirmer and Sandra W. Cowan (1993)
Andrew F. Neuwald, Jun S. Liu, and Charles E. Lawrence (1995)
M. Michael Gromiha, R. Majumdar, and P.K. Ponnuswamy (1997)
To some extent, identification of classic N-terminal signal sequences for secretion across the cytoplasmic membrane is a variation of predicting membrane-spanning protein segments. These sequence segments are as well hydrophobic but are enzymatically cleaved by the signal peptidase upon transport and are thus not present in the mature protein. The specific recognition of the cleavage site has been described by the "(-3, -1) rule" (von Heijne, 1983; Perlman and H.O. Halvorson, 1983), i.e. positions -3 und -1 upstream of the cleavage site are strongly enriched in certain amino acids (-3: Ala, Gly, Ser, Leu, Val, Ile; -1: Ala, Gly, Ser).General Secretion Pathway:
Latest update: October 14, 2009
Back to main page