11.2 Permissive Sites and Their dentification in a Protein 219virtually any user‐defined sequence can be added into the transposon design
such that it remains in the gene after excision of the selection marker.
By these random approaches, permissive sites that accept short three‐residue
linker insertions have been explored for the enzymes β‐lactamase and β‐
galactosidase, revealing a variety of phenotypes depending on the nature of the
inserted residues [23, 25]: insertions with similar physicochemical character as
the neighboring aa’s (regarding hydrophobicity, acidity, and charge) had less
effect on enzyme functionality than physicochemical distant residues.
In another study, linker insertion mutagenesis of TEM1 β‐lactamase revealed
that two residue insertions into predicted β‐sheets abolished enzymatic activity,
while insertions into predicted reverse turns only affected the degree of activity
but did not completely cause loss of function [26]. Further, in some cases, inser-
tions of four residues abolished enzymatic activity, while insertion of two resi-
dues into the same site did not cause complete loss of lactamase activity.
Permissive sites accepting longer insertions – like the seven‐residue‐long
tobacco etch virus (TEV) protease cleavage site – and were in addition accessible
for efficient cleavage by the corresponding protease were explored for a variety
of integral membrane proteins (like the pullulanase secretin PulD from Klebsiella
oxytoca [27], the protein transporter FhaC of Bordetella pertussis [28], and
Escherichia coli lac permease [20]) in order to study structure–function relation-
ships. Further, random insertions of a 31‐residue mostly hydrophilic peptide – so‐
called i31 libraries – were studied for the membrane‐inserted maltose transporters
MalG and MalF [29]. In both cases – for TEV cleavage site insertions as well as
for i31 insertions, permissive sites allowing functional insertions were found to
be mostly located in periplasmic turns or surface loops but not in parts spanning
the membrane or in regions necessary for multimerization with interaction part-
ners. Sequence insertions into nonpermissive sites affected folding, membrane
insertion, multimerization, and overall functionality.
Only in a few cases permissive sites were successfully explored for cytosolic
proteins. The same i31 libraries as mentioned previously were used for the map-
ping of functional domains and further for the identification of permissive sites
in the cytosolic adenosine triphosphate (ATP)‐binding component of the malt-
ose ATP‐binding cassette (ABC) transporter of E. coli MalK, the regulator of the
lac operon LacI, and the F‐plasmid‐derived relaxase TraI [29–32].
Further, a random transposon‐based approach was used to successfully iden-
tify permissive sites in the essential E. coli chaperonin GroEL by delivering a TEV
cleavage site through transposon mutagenesis [9] as well as in the essential
Saccharomyces cerevisiae glycosylphosphatidylinositol (GPI)‐anchored mem-
brane protein Dcw1 [10].
Identification of permissive sites within the mentioned proteins, all of which
are spatially rather complex assemblies, indicates that other less challenging pro-
teins might be able to accept even larger insertions at certain positions. However,
as it was already shown for short insertions [23], the permissiveness of a certain
site depended on the size and the character of the inserted sequence and the
functionality of a certain insertion needed to be evaluated for each case.
Still, the current literature suggests the widespread existence of permissive
sites for peptides of a length between a few and a few dozen residues.