Biomedical Engineering Reference
In-Depth Information
4.2 OVERVIEW OF GENERAL PROPERTIES
OF LINKERS DERIVED FROM NATURALLY
OCCURRING MULTIDOMAIN PROTEINS
TABLE 4.1 Linker Propensities
George and Heringa [17]
Argos [18]
Pro
1.299
1.35
Arg
1.143
0.84
To design a suitable linker for fusion proteins, a direct and
reliable way is to find an example in nature. Many multi-
domain proteins are composed of several functional domains
that are joined by linker peptides. Evolution has selected the
optimal linker peptides to achieve the desired biological
functions of the protein domains. These linkers allow for the
correct folding of protein modules, and also keep the
necessary distance between domains either to avoid strong
steric hindrance or to permit interaction between the protein
domains [16].
With the aim to elucidate the general properties of linkers
in natural proteins, two studies were performed by George
and Heringa [17] and Argos [18]. These groups built data-
bases that contain natural multidomain proteins from which
oligopeptide linkers were extracted. By examining these
databases, the classification and properties of the linkers
were elucidated. Their results suggested that length, com-
position, hydrophobicity, as well as conformation of the
linkers are all important in exerting proper functions such as
separating the protein domains and providing structure
stability.
Phe
1.119
0.69
Leu
1.085
N/A
Glu
1.051
0.87
Gln
1.047
1.13
Met
1.032
0.75
Thr
1.017
1.55
His
1.014
0.55
Tyr
1.000
0.75
Ala
0.964
1.05
Val
0.955
1.00
Ser
0.947
1.46
Asn
0.944
1.09
Lys
0.944
1.16
Ile
0.922
0.81
Asp
0.916
1.25
Trp
0.895
0.23
Gly
0.835
1.25
Cys
0.778
0.35
Calculated from the ratio of single amino acid occurrence in the linker set
compared to its occurrence in the full protein set [17,18]. Values greater than
1 (italics) indicate larger than average occurrences in linker sequences
compared to the full protein sequences.
Source: Reproduced from Reference [17] by permission of Oxford Univer-
sity Press and reprinted from Reference [18], Copyright (1990), with
permission from Academic Press, Elsevier.
4.2.1 Length of Linker Sequence
The average length of oligopeptide linkers in natural multi-
domain proteins calculated by the Argos study was 6.5
residues [18], while the length calculated by George and
Heringa was slightly longer (10.0 residues) [17]. In the latter
study, the linkers were further split into small, medium, and
large linkers, which have average length of 4.5
also investigated. George and Heringa [17] suggested pro-
line (Pro), arginine (Arg), phenylalanine (Phe), threonine
(Thr), glutamic acid (Glu), and glutamine (Gln) were pre-
ferred in the linkers, in the order of descending occurrence.
On the other hand, the study by Argos [18] considered Thr,
serine (Ser), glycine (Gly), alanine (Ala), as well as Pro,
aspartic acid (Asp), lysine (Lys), Gln, and asparagine (Asn)
to be desirable linker constituents (Table 4.1). Both studies
showed that Pro, Thr, and Gln were the desirable amino
acids. Pro was favored probably because it had no amide
hydrogen to form a hydrogen bond with other amino acids,
and therefore, it structurally prevented the interaction
between the linkers and the protein domains. By grouping
the linkers into helical and nonhelical linkers, Pro showed a
higher preference to constitute nonhelical linkers, with a
propensity value of 1.81, compared to 0.8 for helical linkers.
Having Pro residues in the nonhelical linkers could increase
the stiffness, and therefore, allowed for an effective separa-
tion of the protein domains. Small or polar amino acids, such
as Thr (both studies) and Ser and Gly (Argos's study), were
found to be desirable for the linker sequence, since they
provided good flexibility, and also maintained stability of the
structure in the solvent through the formation of hydrogen
bonds with water or the peptide chain. However, there was a
0.7,
9.1
7.6 residues, respectively. The normal-
ized solvent accessibility was then calculated to give insight
into the structural environment of each of the groups (i.e.,
whether the residues in the linkers were buried within the
core of the domains giving a value below 20%, or exposed on
the surface giving a value above 20%). The average normal-
ized solvent accessibility was calculated to be 14.4% for
small linkers, 26.3% for medium sized linkers, and 42.5%
for large linkers [17]. This result suggested that longer
linkers were more exposed to the solvent compared to the
shorter ones. Therefore, when designing a peptide linker for
joining functional domains, the length of the linker may
affect whether it is buried between the functional domains,
or exposed to the solvent.
2.4, and 21.0
4.2.2 Preference of Amino Acid Residues
in Linker Sequence
The preference of amino acid residues used in the linker
sequences in the natural multidomain protein databases was
Search WWH ::




Custom Search