The IS. coli export pathway has been analysed using a combination of
genetic, biochemical and physiological techniques. Table 1.1. lists the known
components of the j!. coll export apparatus and their putative modes of
action.
A genetic approach was used to isolate mutants defective in protein
export. The approaches taken have recently been reviewed (Bieker and
Silhavy, 1990) and are described below. Hybrid proteins were constructed
between the 3' region of the lacZ gene of E. coii and the 5' end (containing
the signal-sequence encoding DNA) of genes encoding for the normally
exported proteins maiE and lamB. The hybrid proteins (MalE-LacZ and
LamB-LacZ) were targeted to the IM of _E. coll and their production was
Induced by maltose. The fusion proteins were recognised by the export
apparatus but could not complete the process because the J3-galactosidase
molecule adopted a conformation that could not be exported. This led to an
Inactive p-galactosldase (because it was unable to form functional
tetramers) and also resulted in the jamming o f the export apparatus. High
levels of production o f either of these hybrid proteins (controlled by maltose
levels) were lethal as they completely jammed the export apparatus. When
these fusions were introduced into a lac
A
strain of E. coli, the resulting phenotype was Lac- and Mals. Maltose resistant (Malr) mutants were thensearched for In order to identify intragenic targeting signals for the fusion
proteins. Such mutants would not target the hybrid protein to the IM and
therefore would not jam the export machinery. Indeed, the vast majority of
the mutations (conferring Malr) were located in the signal-sequence of the hybrid proteins.
Three approaches were then taken -to identify components of the
export apparatus. The first approach was to identify extragenlc suppressors of
Table 1.1. Components of the ji. coli protein export apparatus
Protein Location Size
(kD)
Function
SecB Cytoplasm
12
Form complexes with pre-Trigger factor 60 proteins to maintain
GroEL 910 translocation competence
[Chaperones]
SecA Membrane
(peripheral)
102
ATPase, directs [pre- protein/chaperone] complex to IMSecY Membrane 49 Interact with SecA,
SecE (integral) 14 translocators?
Lep Membrane 36 Process pre-proteins
LspA (integral) 18 Process pre-lipoproteins
SecD Membrane 65 Unknown, late step in
SecF (integral) 35 translocation?
Legend
The information for this table was obtained from the following sources; Bieker and Silhavy, (1990); Bassford et al., (1991), Lecker et al., (1989) and Crooke et al., (1988).
signal-sequence mutations. Strains carrying mutations in the signal-sequences
of either MalE or LamB were used. Restoration of the Mal+ phenotype was
sometimes produced by mutations in export proteins which interacted with
the (mutated) signal-sequence of MalG or LamB. These genes were termed
prl as they were involved in protein localisation. Three genes were identified
which were called prlA, prlD and prlG. Surprisingly, the £rl mutations were
not lethal even though they were in essential genes.
The second approach taken led to the identification of the sec genes,
some of which were allelic with £rl genes. In this approach general export
defects were screened for in the fusion protein (MalE-LacZ) in a
lacA background (as described earlier). Secretion mutants of the above
strain were identified on the basis of increased ^-galactosidase activity at
30°C. The production of an active p-galactosidase was in some cases
expected to result from mutations in the export machinery. Such mutations
would cause an inability of the export apparatus to recognise the
signal-sequence of the MalE-LacZ protein which would then reside (and form
active molecules) within the cytoplasm. Such mutants were then re-screened
for a conditionally lethal phenotype. This approach was used to identify secA
and secB. A similar approach, using PhoA-LacZ and LamB-LacZ fusions, was
used to identify secD and secE.
The third approach came from findings that secA gene expression was
de-repressed under conditions that inhibited protein export. A SecA-LacZ
fusion was constructed (in a merodiplold containing secA-*-) and conditional
lethal mutants were Isolated with raised levels o f p-galactosidase activity.
This approach led to the discovery of secE. (cold-sensitive). Mutations were
mutations in secB were isolated because secB mutations do not cause over-
expression of secA.
Biochemical studies have focussed on isolating the components of the
export apparatus. The reconstitution of an hi vitro synthesis/transport system
then followed which enabled the study o f the individual components of the
export apparatus. The in vitro system has relied heavily on the use of
inverted plasma membrane vesicles. Under the correct conditions and with
the necessary components it is possible to direct various, normally exported,
proteins into such vesicles (Swidersky et al., 1990). Once proteins are
internalised into vesicles they are immune from proteinase attack. This
feature can serve as an assay to monitor the progress of protein
translocation. This type of experiment has been used to demonstrate the
necessity of SecA (Swidersky et al., 1990, Cunningham et al., 1989), SecB
(Watanabe and Blobel, 1989) and SecY/E (Brundage £t al., 1990) in protein
export. A summary of the components o f the jl. coli export machinery and
their proposed functions is shown in Table 1.1. and described below.
The translocation of proteins across the IM of IS. coll involves several
cytoplasmic and membrane protein factors (see above). These include six sec
gene products (SecA [PrlD], SecB, SecD, SecE [PrlG], SecF, SecY [PrlA ]) and
signal-peptidase(s). Other proteins, as well as the Sec proteins, have been
implicated in the export of some classes o f protein. Trigger factor has been
demonstrated to be needed for the in vitro 'export' o f pre-OmpA into
E. coli IM vesicles (Crooke and Wickner, 1987). The E. coll heat shock
proteins, GroEL and GroES, have also been implicated in protein export in
E. coli (Rusukawa et al., 1989).
Biochemical approaches have been used to analyse the functions of
SecA and SecB and have been reviewed (Bieker and Silhavy, 1990). An
elaborate genetic approach, termed suppressor-directed inactivation (SDI),
was used to investigate SecY and SecE. The LamB-LacZ hybrid protein with a
defective signal was used in a merodiploid £ . coli strain carrying a
suppressing SecE (recognising the mutated LamB-LacZ signal-sequence)
protein and SecE+. The cell functions normally because the SecE+ protein
does not recognise (and is not blocked by) the mutant hybrid protein. The
hybrid protein is, however, trapped by the mutated (suppressing) SecE. The
suppressor directed inactivation step can be studied to identify the stage of
the block. The signal-sequence of The LamB-LacZ protein was not processed
when blocked at the stage of action o f SecE. However, when blocked at
SecY, the protein had been processed. This indicated that SecY functioned at
a later stage in the export pathway than SecE. Recent work has suggested
that a truncated version of SecE is functional (Schatz et al., 1991). The
SecE truncate had only one of three membrane-spanning domains remaining
and this was sufficient for its function.
In addition to these proteinaceous factors, ATP (Lill et al., 1989) and
the proton motive force (Schiebel et al., 1991) are also required for protein
export across the _E. coll inner-membrane (IM ).