Mol Gen Genet (1991) 227:40-48
002689259100141F
MGG
<rJ Springer-Verlag
Signal
export
peptidase loverproduction
and maturation
of hybrid
1991
results in increased efficiencies
of
secretory proteins in Escherichia
coli
Jan Maarten van Dijl, Anne de Jong, Rilde Srnith, Sierd Bron and Gerard Venerna
Departmentof Genetics,Centreof BiologicalSciences,
NL-9751 NN Haren (Gn), The Netherlands
Received January 3, 1991
Summary. The effects of 25-fold overproduction
of
Escherichia coli signal peptidase I (SPase I) on the processing killetics of various (hybrid) secretory proteins,
comprising fusions between signal sequence functions
selected from the Bacillus subtilis chromosome and the
mature part of TEM-p-Iactamase, were studied in E.
coli. One precursor (pre[A2d]-p-Iactamase) showed an
enhanced processing rate, and consequently, a highly
improved release of the mature enzyme into the periplasm. A minor fraction of a second hybrid precursor
(pre[A13i]-p-Iactamase), which was not processed under
standard conditions of SPase I synthesis, was shown to
be processed under conditions of SPase I overproduction. However, this did not result in efficient release of
the mature p-Iactamase into the periplasm. In contrast,
the processing rates of wild-type pre-p-Iactamase and
pre(A2)-p-Iactamase, al ready high under standard conditions, were not detectably altered by SPase I overproduction. These results demonstrate that the availability
of SPase I can be a limiting factor in protein export
in E. coli, in particular with respect to (hybrid) precursor
proteins showing low (SPase I) processing efficiencies.
Key words: Precursor conformation -Protein
Signal peptidase I -TEM-p-lactamase
export
Introduetion
Export of proteins by bacteria depends on the properties
of both the exported protein and the export machinery
of the host. The exported proteins are usually synthesized as precursors with an NH2-terminal extension, the
signal peptide, which is essential for export (Pollit and
Inouye 1987; Watson 1984).
From both a fundamental and an applied point of
view it is important to elucidate the mechanisms underlying protein export. In studies on this topic, two lines
Offprint
requests ta: J.M. van Dijl
of investigation have mainly been followed. The first
involved studies on the properties of the exported proteins, including both the signal peptides, the mature
parts and the folding properties of the precursors (for
reviews, see Bankaitis et al. 1987; Saier et al. 1989). The
second major line of investigation has dealt with the
analysis of the cellular components of the protein export
machinery.
Genetically and biochemically the best characterised
bacterial export machinery is that of Escherichia coli
(for review, see Saier et al. 1989). It consists of cytoplasmic and membrane-bound components. Cytoplasmic
proteins like SecB and GroEL, denoted as chaperones,
are essential for the export competence of the precursors,
either by stabilizing an unfolded conformation (Lecker
et al. 1989; Saier et al. 1989), or by preventing their aggregation (Lecker et al. 1990; Mitraki and King 1989).
The soluble, peripheral cytoplasmic membrane protein
SecA (Cunningham et al. 1989; Lill et al. 1989; Schmidt
et al. 1988), and the integral cytoplasmic membrane proteins SecE (Schatz et al. 1989) and SecY (Ito 1984; Watanabe and Blobel 1989) form a complex, which functions as a translocase (Bieker and Silhavy 1990; Brundage et al. 1990; Hartl et al. 1990). The possible functions of other components of the export machinery, like
PrlC (Trun and Silhavy 1989) and SecD (Gardel et al.
1987, 1990), are not known. Finally, the signal peptidases (SPases), which are integral cytoplasmic membrane proteins, catalyse the proteolytic removal of the
signal peptide from the precursor (processing; for review, see Ray et al. 1986). SPase I, also known as leader
peptidase, has been shown to process bacteriophage
M13 procoat and several periplasmic and outer membrane proteins (Dalbey and Wickner 1985; Wolfe et al.
1982). SPase II, also known as prolipoprotein signal peptidase, has been shown to process glyceride-modified lipoproteins (Tokunaga et al. 1982; Yamada et al. 1984;
Yamagata et al. 1982).
A common approach to the analysis of the structurefunction relationship of signal peptides has been the introduction of mutations. Such studies have demon-
41
strated the importance of different domains of the signal
peptide: the basic NH2-terminal region, which is followed by a hydrophobic core, and an SPase cleavage
site (for review, see Pollitt and Inouye 1987). In addition,
the availability of mutant signal peptides, which frequently re sult in reduced or impaired export efficiencies,
has allowed the selection of a variet y of suppressor mutations, both intragenic (Bankaitis et al. 1987; Cover
et al. 1987) and extragenic. Extragenic suppressor mutations were found in several components of the export
machinery like SecA, SecE, SecY and PrlC (for reviews,
see Bankaitis et al. 1987; Beckwith and Perro- Novick
1986; Stader et al. 1989).
We have folIowed a different approach to study signal
sequence functions in Bacillus subtilis and E. coli (Smith
et al. 1987). We selected a large number of sequences
from the B. subtilis chromosome, which can functionally
replace natural signal sequences of exported bacterial
proteins. The selected signals encoded typical signal peptides, which were able to direct Bacillus licheniformis (Xamylase and E. coli TEM-p-Iactamase secretion in B.
subtilis and export into the periplasm in E. coli (Smith
et al. 1988). Some of these signals encoded large polypeptide stretches, reminiscent of the pro-regions of exoproteases (Stahl and Perrari 1984; Vasantha et al. 1984),
between the putative signal peptide and the fusion point
with the target protein. The efficiencies of processing
of these hybrid precursor proteins varied greatly and
appeared to depend on (1) the signal peptide, (2) the
target protein ((X-amylase or p-Iactamase) and (3) the
host strain (B. subtilis or E. coli; Smith et al. 1989a,
b; H. Smith et al., in preparation).
Little is known about the components of the protein
export machinery that may become rate limiting under
certain conditions. In the present investigation we asked
whether the export efficiency of poorly processed proteins can be increased by overproduction of E. coli SPase
I. Por this purpose, the signal sequences selected from
the B. subtilis chromosome, referred to above, appeared
extremely valuable. These signals, when fused to p-Iactamase, directed translocation of the precursor across the
cytoplasmic membrane, but in several cases processing
occurred very slowly (Smith et al. 1989a, b; H. Smith
et al., in preparation). This suggests that the processing
of these translocated proteins by SPase I was inefficient
and was the limiting step in the export. This hypothesis
was examined in the present study. The results showed
that for two poorly processed precursors, the efficiency
of processing could be significantly increased by SPase
loverproduction.
Materials and methods
Bacteria and plasmids. Table 1 lists the bacterial strains
and plasmids used.
Media and plates. TY medium contained 10 g/l Bacto
tryptone, 5 g/l Bacto yeast extract and 10 g/l NaCl. M9
medium (MilIer 1972) for E. coli contained 0.4% glucose, 15 Ilg/ml CaCI2, 250 Ilg/ml MgSO4.7H20, 0.02%
Casamino acids, 1 Ilg/ml thiamine and 2 Ilg/ml thymi-
Table
I. Bacterial
Plasmid ar
strain
strains
and plasmids
Properties and genotype
Saurce ar
reference
pSC101 derivative carrying
the lambda cJ857 gene;
7.3 kb; Km'
Van Dijl et al. (1988)
pKK223-3
pBR322-derived expression
vector carrying the tac
promoter (de Boer et al.
1983); 4.6 kb; Ap'
Pharmacia,
Uppsala,
Sweden
pTDiOi
pBR322 carrying the
E. coli Iep operon;
8.9 kb; Ap'
Date and Wickner
(1981)
pGDL2
pBS61t1pL carrying the
This paper
E. coli Iep gene under the
control of the tac promoter ;
9.8 kb; Km'
pGPB14
Signal sequenceselection
vector
Smith et al. (1987)
pGB25
pGPB14 carrying the intact
TEM-p-lactamase gene;
Ap'; Em'
H. Smith et al.
in preparation
pSPB-A2
pGPB14 carrying signal
sequenceA2; Ap'; Em'
Smith et al. (1988)
pSPB-A2d
pGPB14 carrying signal
sequenceA2d; Ap'; Em'
H. Smith
Plasmids
pBS61L1pl
pSPB-A13i
pGPB14 carrying signal
sequenceA13i ; Ap' ; Em'
et al.
in preparation
H. Smith
et al.
in preparation
Escherichia coli
C600
thr leu thi lac y tonA phx
supE vtr
Phabagen collection,
State University
Utrecht,
The Netherlands
dine. M9 medium-2 was a methionine- and cysteine-free
medium that differed from M9 medium-1 in that
MgSO4.7H2O was replaced by 250 I.Lg/ml MgCl2 and
the Casamino acids by a solution of all (20) amino acids
(250 I.Lg/ml) except methionine and cysteine. If required,
100 I.Lg/ml erythromycin and 20 I.Lg/ml kanamycin were
added.
DNA techniques. Procedures for DNA purification, restriction, ligation, agarose gel electrophoresis, and transformation of competent E. coli cells were carried out
as described by Maniatis et al. (1982). Enzymes were
from Boehringer (Mannheim, FRG).
Western blot analysis. The expression of SPase I was
assayed by Western blotting (Towbin et al. 1979) on nitrocellulose membranes (BA 85; Schleicher and Schuell,
Dassel, FRG). SPase I production was monitored with
specific antibodies and subsequent tracing of bound antibodies with alkaline phosphatase (AP)-anti-rabbit IgG
conjugates (Protoblot@, Western Blot AP system, Promega Biotec, Madison, USA). Reference SPase I was
purified from an overproducing strain as described by
Wolfe et al. (1983b).
42
Protein
metbod
content. Protein content
of Bradford (1976).
was determined
by tbe
ultroscan X L enhanced laser densitometer (LKB,
den).
Pulse-chase protein labelling. Pulse-chase labelling for
studying the kinetics of processing of exported proteins
was carried out essentially as described by Minsky et al.
(1986). Exponentially growing cells in M9 medium-1,
supplemented with 0.2 mM isopropyl-p-D-thiogalactopyranoside (IPTG), were washed once with M9 medium2 and incubated for 45 min in this methionine- and cysteine-free medium, which was also supplemented with
IPTG. Labelling with [3SS]methionine (15 ~Ci/ml;
1330 Ci/mmol; Radiochemical Centre, Amersham, UK)
for the times indicated in the respective experiments,
chasing with excess (2.5 mg/ml) non-radioactive methionine and cysteine, and sampling, folIowed by immediate
precipitation with trichloroacetic acid (TCA; 0° C), were
performed as described previously (Van Dijl et al. 1988).
In vitro transcription, translation and processing. 35S-labelled precursors of exported proteins were synthesized
in vitro and their processing by purified SPase I at 37° C
was studied as described by de Vrije et al. (1987). In
co- and post-translational processing assays SPase I was
added to the translation mixture 5 min and 25 min af ter
the start of the translation, respectively. Incubation with
SPase I was continued for 30 min.
Results
Overproduction of signal peptidase I in E. co li
The E. coli Iep gene encoding SPase I (Wolfe et al. 1983a)
was inserted into plasmid pBS61L1pL, and placed under
the control ofthe repressible tac promoter, derived from
pKK223-3 (Fig. 1). This resulted in plasmid pGDL2.
In E. coli C600 the expression of the pGDL2-encoded
Iep gene resulted in approximately 25-fold overproduction of SPase I, as estimated from Western blots (Fig. 2).
In this strain the tac promoter appeared not to be detec-
Spheroplasting. To study the intracellular localization of
precursors and mature proteins pulse-chase labelling experiments were performed as described above. However,
samples were not immediately precipitated with TCA,
but incubated for 30 min at 0° C with spheroplast buffer
(100 mM TRIS-HC1, pH 8.0, 0.5 M sucrose, 10 mM
EDTA, 0.1 mg/mllysozyme). Spheroplasts and periplasmic contents were then separated by centrifugation prior
to precipitation of the proteins with TCA.
-SPase
Immunaprecipitatian, palyacrylamide gel electrapharesis
and fluaragraphy .Immunoprecipitation
was carried out
as described by Edens et al. (1982) with specific antisera.
Sodium dodecyl sulphate-polyacrylamide
gel electrophoresis (SDS-PAGE) was performed according to
Laemmli (1970). (14C]Methylated molecular weight reference markers were from Amersham (Radiochemical
Centre, Amersham, UK). Fluorography was performed
as described by Skinner and Griswold (1983). Relative
amounts of radioactivity (pulse-chase experiments), or
of alkaline phosphatase staining (Western blot analysis)
were estimated by densitometer scanning with an LKB
2
Hind III Bgl II
3
Hind
Smal
BamH
III
PstI
I
H I
<I
EcoRI
C1I
I
Fig. 2. Production of SPase I. Escherichia coli C600 cells, transformed with pBS61L1pL or pGDL2, were grown in M9 minimal
medium. Exponentially growing cells were lysed in buffer containing 0.1 M potassium phosphate, pH 7.2 and 0.2 mg/mllysozyme.
Similar amounts (approximately 0.02 mg) oftotal protein were subjected to SDS-polyacrylamide gel electrophoresis (SDS-PAGE) and
Western blotting. SPaseI was detected with specific antisera. Lane
1, E. coli C600(pBS61L1pL); lane 2, E. coli C600(pGDL2); lane
3, reference SPaseI
Bglll/
~
Swe-
c1857
PstI
EcoRI
BamHI
Sal I
~R
pBS61 ~ pL Km
7300 bps
pBS61 ~pL.Bgl
Hind III
.Sma
I
7.3 kb
II
~
BamH I
pKK 223-3. BamH I
250 bp fragment
( Ptoc }
~\~
Hind
m
Km
~SmaI
Bglll/ Ps
pGDL2
9800 bps
Hind
pTD101. Bg(ll/BamH
I
2.3kb fragment ((ep)
Fig. I. Schematic presentation ofthe construction ofpGDL2. Only
restriction sites relevant for the construction and properties of the
plasmids are shown. The pSC101-derived plasmid pBS61LJpL(Van
Dijl et al. 1988) was linearized with Bg/II and ligated to a 2.3 kb
BamHI-Bg/II fragment from pTD101 (Date and Wickner 1981)
Bc~
Ba
/
containing the promoterless Escherichia coli Iep gene and a BamHI
fragment (approximately 250 hp) from pKK223-3 containing the
tac promoter (de Boer et al. 1983). This yielded plasmid pGDL2
in which the Iep gene is placed under the control ofthe tac promoter
43
tably repressed, since the levelof SPase loverproduction
was similar in the presence or absence of IPTG (data
not shown). Moreover, SPase loverproduction
did not
appear to be affected by the growth phase of the cells,
or by the growth medium (TY or M9; data not shown).
For the present studies plasmid pGDL2 offered two advantages over plasmid pTD101, which causes SPase I
overproduction to a similar level (Date and Wickner
1981). First, pGDL2 is compatible with the plasmids
encoding the hybrid pre-p-lactamases, which are based
on pBR322. Second, pGDL2 does not encode (wildtype) p-lactamase, which would interfere with measurement of the processing kinetics of the hybrid pre-p-lactamases.
Effects of SPase [overproduction
ics of hybrid precursor proteins
on the processing kinet-
In order to study the effects of SPase loverproduction
in E. coli C600 on the processing kinetics of various
hybrids of mature p-lactamase, fused to signal peptides
randomly selected from the B. subtilis chromosome,
pulse-chase labelling experiments were performed. Por
this purpose three hybrid pre-p-lactamases were chosen
that showed varying efficiencies of processing in E. coli
under standard conditions of SPase I production.
Pre(A2)-p-lactamase was one of the most efficient signals for processing (Smith et al. 1989a; H. Smith et al.,
in preparation). A precursor that was processed extreme-
A
0
30"
60"
120" B
0
30"
region of the selected export function had been deleted.
Processing of a third precursor, pre(A13i)-/3-lactamase,
which was obtained by extending the hydrophobic core
(nine amino acids) of the relatively inefficient export
signal A13 with ten amino acids, was totally inhibited.
However, translocation of this precursor appeared not
to be affected since it was exposed to the periplasm,
giving rise to high levels of ampicillin resistance, although it remained attached to the membrane (H. Srnith
et al., in preparation). As a control, the effects of SPase
loverproduction
on the processing kinetics of wild-type
TEM-/3-lactamase were measured. Strains of E. coli
C600 producing the various plasmid-encoded hybrid /3lactamases were transformed with plasmids pGDL2
(overproduction of SPase I), or pBS61L1pL (standard
production of SPase I). The latter plasmids, based on
pSC101, were present in all test strains.
The results of the pulse-chase labelling experiments
with the various precursors are shown in Figs. 3 to 6.
Figure 3 shows that at 25° C the processing rate of wildtype /3-lactamase, which is very high compared with
those of the other (hybrid) /3-lactamases, was not measureably affected by SPase loverproduction.
The time
necessary to process 50% of the precursor (tso) was approximately
10 s both under standard conditions
(Fig. 3A and C), and when SPase I was overproduced
(Fig. 3 B and C). At 37° C processing was too fast to
allow detection of the precursor (data not shown).
Pre(A2)-/3-lactamase was processed more slowly than
wild-type /3-lactamase (Fig. 4). The tso, measured at
60" 120.'
,"-~
~-
ly slowly in E. coli was pre(A2d)-/3-lactamase, a derivative of pre(A2)-/3-lactamase, from which the "pro-like "
-p
0
A
---m
30"
2'
5'
B o
30
l'
2
m-
C100
I
5'
-p
p-
--
--
I
-rn
C100
o
~l60~ ~ ~
20 ~
I
0
16° I
~
I
~
,
,
60
,
sec
,
120
Fig.3A-C. Pulse-chase analysis of wild-type TEM-p-Iactamase
processing. The processing of wild-type pre-p-Iactamase by
Escherichia coli C600(pGB25, pBS61L1pL) (A) and E. coli
C600(pGB25, pGDL2) (B) was analysed by pulse-chase labelling
at 25° C and subsequent immunoprecipitation, SDS-PAGE and
fluorography as described in Materials and methods. Cells were
labelled for 30 s and samples were withdrawn af ter the chase (t=o)
at the times indicated. The kinetics of processing are plotted as
the percentage of the total p-Iactamase protein (precursor plus mature) that is still present in the precursor form at the time of sampling. C (.) E. coli C600(pGB25, pBS61L1pL); (0) E. coli
C600(pGB25, pGDL2). p, precursor; rn, mature
20
I,
O
1
2
,
3
,
I
4 min 5
Fig. 4A-C. Pulse-chaseanalysis ofpre(A2)-p-lactamase processing.
The processing of pre(A2)-p-lactamase by Escherichia coli
C600(pSPB-A2, pBS61ApL) (A), and E. coli C600(pSPB-A2,
pGDL2) (B) was analysed by pulse-chase labelling at 25° C and
subsequent immunoprecipitation, SDS-PAGE and f1uorography as
described in Materials and methods. Cells were labelled for 60 s
and samples were withdrawn after the chase (t=o) at the times
indicated. The kinetics of processing are plotted as in Fig. 3. C
(.) E. coli C600(pSPB-A2, pBS61ApL); (0) E. coli C600(pSPB-A2,
pGDL2). p, precursor; rn, mature
44
25° C, was about 150 s under standard conditions
(Fig. 4A and C). Under conditions of SPase I overproduction (Fig. 4B and C) the tso was not significantly
reduced (data not shown). These results suggest that
SPase loverproduction
has no effect on the rates of
processing of p-lactamase precursors that are already
efficiently processed under standard conditions of SPase
I production.
In contrast to the wild-type p-lactamase and pre(A2)p-lactamase, the processing kinetics of the other two
pre-p-lactamases studied were clearly affected by SPase
loverproduction.
A drastic effect was observed with
pre(A2d)-p-lactamase. This precursor, which in E. coli
C600 is hardly processed under standard conditions
(Fig. 5A; H. Smith et al., in preparation), showed a dramatic increase in the rate of processing under conditions
of SPase loverproduction
(Fig. 5 B) : about 65% of the
total (A2d)-p-lactamase was already processed 1 min
after the start of labelling (Fig. 5C; tso < 1 min). The
possibility that an unidentified suppressor mutation had
occurred in the signal sequence of pre(A2d)-p-lactamase,
which might explain the increased processing efficiencies, could be ruled out. Transformation of E. coli C600
with pSPB-A2d extracted from the SPase I overproducing strain, and subsequent pulse-chase analysis (under
non-overproducing
conditions) revealed the original
slow processing kinetics, characteristic of pre(A2d)-Plactamase (data not shown).
The processing of pre(A13i)-p-lactamase was also
clearly affected by SPase loverproduction.
In contrast
to the other two hybrid p-Iactamases described above,
A
0
10' 20'
30'
B
0
10
20'
30'
-p
-rn
c 100t=
A
0
10
20
B
30
0
10'
20
30
c 100 ~
~
!6,0
r
~
r
I,
0
10
,
20
I
m in 30
Fig. 6A-C. Pulse-chaseanalysis of pre(AI3i)-p-lactamase processing. The processing of pre(AI3i)-p-lactamase by Escherichia coli
C600(pSPB-AI3i, pBS61ApL) (A) and E. coli C600(pSPB-AI3i,
pGDL2) (B) was analysed by pulse-chase labelling at 37° C and
subsequent immunoprecipitation, SDS-PAGE and fluorography is
described in Materials and methods. Cells were labelled for 60 s
and samples were withdrawn after the chase (t=o) at the times
indicated. The kinetics of processing are plotted as in Fig. 3. C
(.) E. coli C600(pSPB-AI3i, pBS61ApL); (0) E. coli C600(pSPBA13i, pGDL2). p, precursor; rn, mature
no processing at all of this precursor could be detected
under standard conditions (Fig. 6A; H. Smith et al., in
preparation). However, when SPase I was overproduced
a small amount of mature enzyme was detectable
(Fig. 6B), which was already present immediately af ter
the chase. lts relative amount (6% to 7% of the total
amount of [A13i)-f3-lactamase synthesized) remained almost unaltered during the period of sampling (Fig. 6 C),
indicating that its rate of processing was very high.
::j::
-i
Localization of precursors and mature enzymes
L
O
~
::J
u
Q)
L
0-
f
0
~
0}
-0u
-<>--- -0
LU 111111
JV
Fig. 5A-C. Pulse-chase analysis of pre(A2d)-p-lactamase processing. The processing of pre(A2d)-p-lactamase by Escherichia coli
C600(pSPB-A2d, pBS61ApL) (A) and E. coli C600(pSPB-A2d,
pGDL2) (B) was analysed by pulse-chase labelling at 37° C and
subsequent immunoprecipitation, SDS-PAGE and fluorography as
described in Materials and methods. Cells were labelled for 60 s
and samples were withdrawn after the chase (t=o) at the times
indicated. The kinetics of processing are plotted as in Fig. 3. C
(.) E. coli C600(pSPB-A2d, pBS61ApL); (0) E. coli C600(pSPBA2d, pGDL2). p, precursor, rn, mature
Proteolytic processing of the precursors of translocated
proteins by SPases in E. coli is considered to be a prerequisite for the release of the mature enzyme into the
periplasm (Dalbey and Wickner 1985). It was therefore
of interest to study the localization of the slowly processed export proteins used in these investigations (pre[A2d]-f3-lactamase and pre[A13i]-f3-lactamase), which
were sensitive to SPase loverproduction.
The localization of precursors and mature products was determined
by pulse-chase labelling of proteins, followed by spheroplasting, and subsequent separation of spheroplasts
and periplasmic contents by centrifugation. As controls,
the localizations of wild-type f3-lactamase and (A2)-f3lactamase were determined at 25° C. Apparently, processing of wild-type f3-lactamase (the samples were taken
90 s af ter the chase) could not be stopped instantaneously by chilling on ice, as revealed by the fact that no
precursor could be detected. As expected, the mature
45
B
A
A
B
(
m-
1
2
3
1
+
3
2
3
R
1
2
3
F
E
p-
-p
-rn
2
3
+
R
+
Fig. 8A-C. In vitro processing. In vitro synthesized precursors were
incubated post-translationally
for 30 min at 37° C with ( + ) or
without
( -) purified SPase I. A Wild-type
pre-p-lactamase;
B
pre(A2)-p-lactamase;
C pre(A2d)-p-lactamase.
p, precursor;
rn,
mature; R, molecular weight reference (30 kDa)
D
(
1
2
2
3
Fig. 7 A-F. Localization of precursors and mature proteins. Cells
producing the respective precursors were pulse-labelled and chased
under the conditions indicated in the legends to Figs. 3-6. Samples
were withdrawn at various times after the chase (t=o) and the
cells were spheroplasted at 00 C. Total samples (1, spheroplasts
plus periplasmic content, 2, spheroplasts, and 3, periplasmic contents) were precipitated with trichloroacetic acid (TCA), and anaIysed by immunoprecipitation, SDS-PAGE and f1uorography as
indicated in Materials and methods. A Escherichia co/i
C600(pGB25, pBS61,1pL), t=90 s; B E. co/i C600(pSPB-A2,
pBS61,1pL), t=2 min; C E. co/i C600(pSPB-A2d, pBS61,1pL), t=
5min; D E. co/i C600(pSPB-A2d, pGDL2), t=5min;
E E. co/i
C600(pSPB-A13i, pBS61,1pL), t=30min; F E. co/i C600(pSPBA13i, pGDL2), t=30min. p, precursor; rn, mature; R, molecular
weight reference (30 kDa)
tant increase in the rate of release of the mature enzyme
into the periplasm. These results are in full agreement
with the idea (Dalbey and Wickner 1985) that proteolytic processing of the precursor is required for the release
of the mature protein into the periplasm.
Fractionation of cells producing (A13i)-p-lactamase
showed a different pattem. Under all conditions tested,
pre(A13i)-p-lactamase
remained associated with the
spheroplasts (Fig. 7E and F). However, unexpectedly,
a major fraction of the mature enzyme, detectable only
under conditions of SPase loverproduction
(Fig. 7 F),
was also associated with the spheroplast fraction. Only
about 20% of the mature enzyme, representing less than
2% of the total amount of (A13i)-p-lactamase synthesized, was detectable in the periplasmic fraction. It remains uncertain whether this small periplasmic fraction
resulted from true release into the periplasm as a consequence of SPase I processing, or from lysis of a small
fraction of the spheroplasts.
In vitro processing
p-lactamase fractionated with the periplasmic contents
(Fig. 7 A), which was not affected by SPase I overproduction (data not shown). The localization ofpre(A2)-plactamase and its mature product was determined in
samples taken 2 min af ter the chase. Although, as with
wild-type pre-p-lactamase, processing was not stopped
instantaneously, a significant amount of pre(A2)-p-lactamase was detectable (Fig. 7 B), presumably as a consequence of the fact that pre(A2)-p-lactamase is processed
more slowly than the wild-type pre-p-lactamase (Figs. 3
and 4). The pre(A2)-p-lactamase appeared to be exclusively associated with the spheroplast fraction, whereas
most of the mature product (94%) was present in the
periplasmic fraction (Fig. 7 B). These results were not
altered by SPase loverproduction
(data not shown).
In contrast, the levelof SPase I production clearly
affected the localization of (A2d)-p-lactamase in samples
taken 5 min af ter the chase (37° C). Although, as with
pre(A2)-p-lactamase, pre(A2d)-p-lactamase was associated exclusively with the spheroplasts (Fig. 7C and D),
as a result of the increased rate of pre(A2d)-p-lactamase
processing under conditions of SPase loverproduction
(Fig. 5), a markedly increased amount of mature enzyme
could be detected in the periplasmic fraction (Fig. 7D).
These results indicate that overproduction of SPase I
not only resulted in a drastic increase in the rate of
pre(A2d)-p-lactamase processing, but also in a concomi-
In order to determine wbetber tbe precursors studied
bere were appropriate substrates for SPase I, in vitro
syntbesized 35S-labelled precursors were incubated posttranslationally for 30 min witb purified SPase I. Tbe
amount of SPase I added corresponded to tbat present
in approximately 1.5 ml of exponentially growing cultures (M9 medium-1 ; approximately 0.4 mg/ml protein)
of cells producing standard amounts of SPase I. For
unknown reasons pre(A13i)-p-lactamase could not be
syntbesized in vitro. Tbe results for tbe otber precursors
are sbown in Fig. 8. In agreement witb tbe results obtained in vivo from pulse-cbase labelling experiments,
wild-type pre-p-lactamase and pre(A2)-p-lactamase appeared to be efficiently processed: af ter treatment witb
SPase I 76% of tbe wild-type p-lactamase and 78% of
tbe (A2)-p-lactamase were mature. In contrast, pre(A2d)-p-lactamase was not processed in vitro by tbe
SPase I added. Similar results were obtained wben tbe
same amounts of SPase I were present during tbe syntbesis of tbe precursors in a co-translational assay (data
not sbown).
Discussion
Recently we have shown that the efficiencies of translocation and processing of different hybrid precursor proteins containing signal peptides, selected from the B. sub-
46
ti/is chromosome, depend on (1) the export signal, (2)
the mature protein (tJ-lactamase and (X-amylase were
compared) and (3) the orgranism that expressed the precursors (B. subti/is and E. co/i were compared) (Smith
et al. 1989a, b; H. Smith et al., in preparation). These
effects might reflect either the folding properties of the
different precursors, or their interaction with particular
components of the bacterial export machinery, or both.
As a first approach to improve the export of two
slowly processed hybrid tJ-lactamases, we examined
whether the availability of SPase I might be a limiting
factor. Other studies have demonstrated that the reduced
availability of SecA, SecB, SecD, SecE, SecY (see Saier
et al. 1989), SPase I (Dalbey and Wickner 1985; Van
Dijl et al. 1988) and SPase II (Yamagata et al. 1982)
as compared with the wild-type situation results in impaired processing and export of periplasmic E. co/i proteins. For our purpose, to study whether SPase I overproduction might result in increased rates of precursor
processing and protein export, we constructed the low
copy number plasmid pGDL2, carrying the E. co/i /ep
gene, encoding SPase I under the control of the strong
tac promoter (de Boer et al. 1983). E. co/i cells harbouring pGDL2 overproduced SPase I about 25-fold. The
results showed that this levelof SPase loverproduction
was sufficient to alleviate several export defects in two
different hybrid tJ-lactamase precursors to different extents.
The stimulatory effect on export was most drastic
with pre(A2d)-tJ-lactamase. This precursor was obtained
by deletion of 37 amino acid residues from the signal
of the hybrid pre(A2)-tJ-lactamase, which is processed
efficiently (Fig.4; Smith et al. 1989a; H. Smith et al.,
in preparation). As compared with pre(A2)-tJ-lactamase,
processing of pre(A2d)-tJ-lactamase in E. co/i was severely retarded (Fig. 5; H. Smith et al., in preparation). This
retardation of processing could be largely abolished by
SPase loverproduction
(Fig. 5). Moreover, the processed (A2d)-tJ-lactamase was present in the periplasm,
which is in agreement with earlier experiments by Dalbey
and Wickner (1985) indicating that processing is a prerequisite for the release of mature proteins into the periplasm.
In contrast to pre(A2d)-tJ-lactamase, pre(A13i)-tJ-lactamase was, under standard conditions, not processed
at all (Fig. 6; H. Smith et al., in preparation). Nevertheless, pre(A13i)-tJ-lactamase was translocated across the
cytoplasmic membrane (H. Smith et al., in preparation).
This indicates that pre(A13i)-tJ-lactamase is a poor substrate for SPase I in vivo. Since the original pre(A13)-tJlactamase was processed efficiently (Smith et al. 1989a,
b; H. Smith et al. in preparation) and the insertion of
ten additional hydrophobic amino acids in the hydrophobic core of signal A13 lef t the putative SPase I processing site unaltered, we speculate that the conformation of pre(A13i)-tJ-lactamase was such that the interaction with SPase I was severely affected or, alternatively,
that the SPase I cleavage site was not properly exposed.
The observation that overproduction of SPase I caused
some cleavage of pre(A13i)-tJ-lactamase (6% to 7%)
shortly af ter its synthesis suggests that it had, at least
for a short period of time, a processing-competent conformation. Apparently, the increased availability of
SPase I allowed the processing of (part of) that fraction
of processing-competent pre(A13i)-p-lactallase
molecules. Illproper folding might also explain the poor release of the mature (A13i)-p-lactallase
into the periplasm. Minsky et al. (1986) have dellonstrated that the
release of mature wild-type p-lactallase occurs concomitantly with a conformational change. This change might
be inefficient in (A13i)-p-lactallase, thereby preventing
efficient release.
A sillilar explanation to that given for the processing
of a fraction of pre(A13i)-p-lactallase
might be entertained for the effect of SPase loverproduction
on the
processing of pre(A2d)-p-lactallase.
In this case also,
processing of a certain precursor fraction occurred very
rapidly: about 65% of the total (A2d)-p-lactallase was
processed during the period of labelling (1 min). This
suggests that pre(A2d)-p-lactallase was processing competent for a short period of time, and that illproved
availability of SPase I allowed the rapid processing of
a substantial fraction of the processing-competent precursor molecules, before they became a poor substrate
for SPase I. The in vitro processing experiments provide
additional support for the idea that pre(A2d)-p-lactamase may assume a processing-incompetent conformation: while the in vitro synthesized wild-type pre-p-lactamase and pre(A2)-p-lactallase
were readily processed
by purified SPase I, the pre(A2d)-p-lactallase
was not
processed by SPase I. The observation that even the
presence of SPase I during synthesis did not re sult in
processing of the latter precursor indicates that it assulled the processing-incompetent conformation very
rapidly.
The results obtained with pre(A2d)-p-lactallase
under conditions of SPase loverproduction
are relliniscent of those obtained with the precursor of coat protein
of the mutant bacteriophage M13am8H1R6. The procoat-R6, which differs in its mature part in three allino
acids froll the wild-type protein (Boeke et al. 1980), is
processed considerably more slowly (half-time approximately 60 s) than the wild-type procoat (half-time <2 s)
in infected cells (Russel and Model 1981). Date and
Wickner (1981) have dellonstrated that the rate of procoat-R6 processing could be greatly increased by 30-fold
overproduction of SPase I. Thus, it would appear that
SPase loverproduction
not only has a favourable effect
on the processing rate of this membrane spanning protein, but also on the processing rates of hybrid pre-plactallases, exported into the periplasm.
Interestingly, pre(A2d)- and pre(A13i)-p-lactallase
are processed in B. subtilis under standard conditions,
although with a lower efficiency than the corresponding
non-llutated precursors pre(A2)- and pre(A13i)-p-lactamase (H. Smith et al., in preparation). This suggests that
in B. subtilis either more SPase is available than in E.
coli, or that the interaction of these precursors with the
B. subtilis SPase is more efficient. Alternatively, B. subtilis might be able to control the folding of these precursors more effectively than E. coli, resulting in prolonged
periods of processing competence.
Both the in vivo and the in vitro experiments indicated that wild-type and pre(A2)-f3-lactamase are appropriate substrates for SPase I. Overproduction of SPase
I had no, or only limited effect on the processing kinetics
of these precursors. This suggests that the availability
ofSPase I is not limiting for the export ofthese efficiently processed precursors. Similar results have been obtained for the in vivo processing rates of wild-type pre-f3lactamase and pre-PhoE (Anba et al. 1986) and pre-maltose-binding protein (Randall and Hardy 1984).
In summary our results indicate that the export of
proteins that are processed with low efficiencies under
standard conditions in particular may be considerably
improved by overproduction of SPase I. We conceive
that this could be of considerable help in improving the
export of heterologous proteins, for instance from eukaryotic sources, which are frequently exported with low
efficiencies in E. coli and other prokaryotes.
Acknowledgements. We thank Dr. W. Wickner for providing
pTD10l, Dr. R. Zimmermann for providing sera directed against
SPase I, and Henk Mulder for preparing the figures. We thank
Drs. H. de Cock and Dr. J. Tommassen for making available the
in vitro processing system. Funding for the project of which this
work is a part was provided by Gist-brocades N .V. Delft, the Netherlands.
References
Anba J, Lazdunski C, Pages J-M (1986) Lack of effect of leader
peptidase overproduction on the processing in vivo of exported
proteins in Escherichia coli. J Gen Microbiol132 : 689-696
Bankaitis VA, Altman E, Emr SD (1987) Export and localization
of Escherichia coli envelope proteins. In: Inouye M (ed) Bacterial outer membranes as model systems. Wiley & Sons, USA,
pp 75-116
&ckwith J, Ferro-Novick S (1986) Genetic studies on protein export in bacteria. Curr Top Microbiol ImmunoI125:5-27
Bieker KL, Silhavy T (1990) PrlA (SecY) and PRIG (SecE) interact
directly and function sequentially during protein translocation
in Escherichia coli. Cell 61: 833-842
Boeke JD, Russel M, Model P (1980) Processing of filamentous
phage pre-coat protein. J Mol Biol144 : 103-116
Bradford MM (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle
ofprotein dye binding. Anal Biochem 72:248-254
Brundage L, Hendrick JP, Schiebel E, Driessen AJM, Wickner
W (1990) The purified Escherichia coli intergral membrane protein SecY/E is sufficient for reconstitituon of SecA-dependent
precursor protein translocation. Cell 62: 649-657
Cover WH, Ryan PJ, Bassford PJ, Walsh KA, Bollinger J, Randall
LL (1987) Suppression of a signal sequence mutation by an
amino acid substitution in the mature portion of the maltosebinding protein. J Bacteriol169 : 1794-1800
Cunningham K, Lill R, Crooke E, Rice M, Moore K, Wickner
W, Oliver D (1989) SecA protein, a peripheral protein of the
Escherichia coli plasma membrane, is essential for the functional binding and translocation ofproOmpA. EMBO J 8:955-959
Dalbey RE, Wickner W (1985) Leader peptidase catalyzes the release of exported proteins from the outer surface of the
Escherichia coli plasma membrane. J Biol Chem 260: 1592515931
Date T, Wickner W (1981) Isolation of the Escherichia coli leader
peptidase gene and effects of leader peptidase overproduction
in vivo. Proc Natl Acad Sci USA 78:6106-6110
De Boer HA, Comstock LJ, Vasser M (1983) The tac promoter:
a functional hybrid derived from the trp and lac promoters.
Proc Natl Acad Sci USA 80:21-25
De Vrije T, Tommassen J, de Kruijff B (1987) Optimal posttranslational translocation of the precursor of PhoE protein across
Escherichia coli membrane vesicles requires both A TP and the
proton motive force. Biochim Biophys Acta 900: 63- 72
Edens L, Heslinga L, Klok R, Ledeboer AM, Maat J, Toonen
MY, Visser C, Verrips CT (1982) Cloning ofthe cDNA encoding the sweet-tasting plant protein thaumatin and its expression
in Escherichia coli. Gene 18:1-12
Gardel C, Benson S, Hunt J, Michaelis S, Beckwith J (1987) SecD,
a new gene involved in protein export in Escherichia coli. J
BacterioI169:1286-1290
Gardel C, Johnson K, Jacq A, Beckwith J (1990) The secD locus
of Escherichia coli codes for two membrane proteins required
for protein export. EMBO J 9:3209-3216
Hartl F-U, Lecker SL, Schiebel E, Hendrick JP, Wickner W (1990)
The binding cascade of SecB to SecA to SecY/E mediates preprotein targeting to the Escherichia coli plasma membrane. Cell
63:269-279
Ito K ( 1984) Identification of the secy (prIA) gene product involved
in protein export in Escherichia coli. Mol Gen Genet 197: 204208
Laemmli UK (1970) Cleavage of structural proteins during the
assembly ofthe head ofbacteriophage T4. Nature 227 :680-685
Lecker SL, Lill R, Ziegelhoffer T, Georgopoulos C, Bassford Jr
PJ, Kumamoto CA, Wickner W (1989) Three pure chaperone
proteins of Escherichia coli -SecB, trigger factor and GroEL
-form
soluble complexes with precursor proteins in vitro.
EMBO J 8:2703-2709
Lecker SL, Driessen AJM, Wickner W (1990) ProOmpA contains
secondary structure and tertiary structure prior to translocation
and is shielded from aggregation by association with SecB protein. EMBO J 9:2309-2314
Lill R, Cunningham K, Brundage LA, Ito K, Oliver D, Wickner
W (1989) SecA protein hydrolyzes A TP and is an essential
component of the protein translocation A TPase of Escherichia
coli. EMBO J 8:961-966
Maniatis T, Fritsch EF, Sambrook J (1982) Molecular cloning:
a laboratory manual. Cold Spring Harbor Laboratory Press,
Cold Spring Harbor, New York
Milier JH (1972) Experiments in molecular genetics. Cold Spring
Harbor Laboratory Press, Cold Spring Harbor, New York
Minsky A, Summers RG, Knowles JR (1986) Secretion of p-Iactamase into the periplasm of Escherichia coli: Evidence for a
distinct release step associated with a conformational change.
Proc Natl Acad Sci USA 83:4180-4184
Mitraki A, King J (1989) Protein folding intermediates and inclusion body formation. Biotechnology 7: 690-697
Pollit S, Inouye M (1987) Structure and functions of the signal
peptide. In: Inouye M (ed) Bacterial outer membranes as model
systems. Wiley & Sons, USA, pp 117-139
Randall LL, Hardy SJS (1984) Export of protein in bacteria. Microbiol Rev 48: 290-298
Ray P, Dev I, MacGregor C, Bassford Jr P (1986) Signal peptidases. Curr T9P Microbiol ImmunoI125:75-102
Russel M, Model P (1981) A mutation downstream from the signal
peptidase cleavage site affects cleavage but not membrane insertion of phage coat protein. Proc Natl Acad Sci USA 78: 17171721
Saier MH, Werner PK, Miiller M (1989) Insertion of proteins into
bacterial membranes: mechanism, characteristics, comparisons
with the eucaryotic process. Microbiol Rev 53: 333-366
Schatz PJ, Riggs PD, Jacq A, Fath MJ, Beckwith J (1989) The
secE gene encodes an integral membrane protein required for
protein export in Escherichia coli. Genes Dev 3: 1035-1044
Schmidt MG, Rollo EE, Grodberg J, Oliver DB (1988) Nucleotide
sequence of the secA and secA(Ts) mutations preventing protein export in Escherichia coli. J Bacteriol170: 3404-3414
Skinner MK, Griswold MD (1983) Fluorographic detection of radioactivity in polyacrylamide gels with 2,5-diphenyloxazole in
48
acetic acid and its comparison with existing procedures. Biochern J 209:281-284
Smith H, Bron S, van Ee J, Venerna G (1987) Construction and
use of signal sequence selection vectors in Escherichia coli and
Bacillus subtilis. J Bacteriol169 : 3321-3328
Smith H, de Jong A, Bron S, Venerna G (1988) Characterization
of signal-sequence-coding regions selected frorn the Bacillus
subtilis chromosome. Gene 70: 351-361
Smith H, de Jong A, van Dijl JM, Bron S, Venerna G (1989a)
Protein secretion in Bacillus subtilis: Characterization of randomly selected signal-sequence-coding regions. In: Butler LO,
Harwood C, Moseley BEB (eds) Genetic transformation and
expression. Intercept, Andover, UK, pp 519-527
Smith H, de Jong A, van Dijl JM, Bron S, Venerna G (1989b)
Characterization of signal sequence coding regions selected
frorn the Bacillus subtilis chromosome. Fifth international conference on genetics and biotechnology of Bacilli. Asilomar, Pacific Grove, USA
Stader J, Gansheroff LJ, Silhavy TJ (1989) New suppressors of
signal-sequence mutations, prlG, are linked tightly to the secE
gene of Escherichia coli. Genes Dev 3: 1045-1052
Stahl ML, Ferrari E (1984) Replacement of the Bacillus subtilis
subtilisin structural gene with an in vitro-derived deletion system. J BacterioI158:411-418
Tokunaga M, Loranger JM, Wolfe PB, Wu H C (1982) Prolipoprotein signal peptidase in Escherichia coli is distinct frorn the M13
procoat protein signal peptidase. J Biol Chern 257: 9922-9925
Towbin H, Staehelin T, Gordon J (1979) Electrophoretic transfer
of proteins frorn polyacrylamide gels to nitrocellulose sheets:
procedures and some applications. Proc Natl Acad Sci USA
76:4350-4354
Trun NJ, Silhavy TJ (1989) PrIC, a suppressor of signal sequence
mutations in Escherichia coli, can direct the insertion of the
signal sequenceinto the mernbrane. J Mol Biol 205: 665-676
Van Dijl JM, Smith H, Bron S, Venerna G (1988) Synthesis and
processing of Escherichia coli TEM-/3-lactamase and Bacillus
licheniformis IX-amylasein Escherichia coli: The role of signal
peptidase. I. Mol Gen Genet 214: 55-61
Vasantha N, Thompson LD, Rhodes <;:,Banner C, Nagle J, Filpula
D (1984) Genes for alkaline protease and neutral protease from
Bacillus amyloliquefaciens contain a large open reading frame
between the regions coding for signal sequenceand mature protein. J BacterioI159:811-819
Watanabe M, Blobel G (1989) Site-specific antibodies against the
PrlA (secY) protein of Escherichia coli inhibit protein export
by interfering with plasma membrane binding of preportions.
Proc Natl Acad Sci USA 86: 1895-1899
Watson MEE (1984) Compilation of published signal sequences.
Nucleic Acids Res 12: 5145-5164
Wolfe PB, Silver P, Wickner W (1982) The isolation of homogeneous leader peptidase from a strain of Escherichia coli which
overproduces the enzyme. J Biol Chem 257: 7898- 7902
Wolfe PB, Wickner W, Goodman JM (1983a) Sequence of the
leader peptidase gene of Escherichia coli and the orientation
of leader peptidase in the bacterial envelope. J Biol Chem
258:12073-12080
Wolfe PB, Zwizinsky C, Wickner W (1983b) Purification and characterization of leader peptidase from Escherichia coli. In: Fleischer S, Fleischer B (eds) Methods Enzymol97 :40--46
Yamada H, Yamagata H, Mizushima S (1984) The major outer
membrane lipoprotein and new lipoproteins share a common
signal peptidase that exists in the cytoplasmic membrane of
Escherichiacoli. FEBS Lett 166:179-182
Yamagata H, Ippolito C, Inukai M, Inouye M (1982) Temperaturesensitive processing of outer membrane lipoprotein in an
Escherichia coli mutant. J Bacteriol152 : 1163-1168
Communicated by H. Bohme