Biomedical Engineering Reference
In-Depth Information
hydrophobicity of one side of an assumed membrane spanning
b
-strand is
computed using the equation
H
b
ðÞ¼
i
1
=
5
½
h
ð
i
4
Þþ
h
ð
i
2
Þþ
hi
ðÞ;
where h(i) is the hydrophobicity of the residue based on Kyte and Doolittle [
12
]. An
amphipathic
-strand is identified with a period of two residues, so that the criterion
for prediction is H
b
(i)
b
1.6 and H
b
(i
1)
<<
1.6. The membrane spanning
b
-strands have been assigned with the minimal length of ten residues.
Gromiha and Ponnuswamy [
68
] used a “hydrophobicity profile method” for
predicting the membrane spanning
-strand segments as described below (1)
construct hydrophobicity profiles of window lengths 12, 6, and 1; (2) the high
peaks in the profile correspond to transmembrane
b
-strands of around 12 residues,
(3) the low peaks have been divided into seven overlapping six-residue segments
and selected the segment showing the highest amphipathicity; and (4) the end
residue is fixed when it breaks the amphipathic character (alternating high and
low hydrophobicity) in the profile of window length one. This method has been
successfully applied to predict the transmembrane segments of porin, maltoporin,
ompA, and ompX, and we found a good agreement with experimental observations.
Gromiha et al. [
69
] developed a “rule-based approach” for predicting transmem-
brane
b
-strands using three features (1) preference of amino acid residues in
membrane spanning
b
) and (2) hydrophobic
character (H
p
) (3) and amphipathicity. A set of five primary rules have been
designed to assign the priority of each residue to be in transmembrane
b
-strands (conformational parameters,
b
-strand
and four secondary rules to pick up the membrane spanning segments. The primary
rules for assigning the priority of each residue, i, are:
b
b
(i)
>
1.0 (average confor-
mational parameter), 1/6
P
6
1
b
ð
i
Þ >
1.0, H
p
(i)
>
13.34 (average hydrophobicity),
i
¼
1/6
P
6
13.34 and 1/2
P
2
H
p
ð
i
Þ >
H
p
ð
i
Þ¼
13.34
0.5 (oscillating around the
i
¼
1
i
¼
1
average hydrophobicity). If these conditions are satisfied, the priority is one and
zero, otherwise. The secondary rules for picking up the membrane spanning
segments are: if any residue has the priority of 5, two consecutive residues have
the priority of 4 or three consecutive residues have the priority of
3, there is a
possibility of a transmembrane
-strand segment around the residue(s). Extend the
length in both directions so that there may not be two consecutive low priority
residues (
b
3) or a residue of zero priority. If the segment is longer than 20 residues,
cut into two smaller segments at the residue of highest hydrophobicity. The method
has been applied to three different porins and the transmembrane
<
-strand segments
are predicted with the accuracy of 82%. Recently, Freeman and Wimley [
70
]
proposed a statistical method based on the physicochemical properties of experi-
mentally characterized
b
-barrel membrane protein structures for predicting their
membrane spanning segments.
b
