Download Targeting of interleukin-2 to the periplasm of

Journal of General Microbiology (1993), 139, 2465-2473.
Printed in Great Britain
2465
Targeting of interleukin-2 to the periplasm of Escherichia coli
GABRIELE
HALFMANN,'t HERVEBRAILLY,~
ALAINBERNADAC,
FELIX
A. MONTEROJULIAN,~
CLAUDE
LAZDUNSKI'
and DANIELBATY~*
'Laboratoire dlngknierie et Dynamique des S y s t h e s Mernbranaires du CNRS, 31 chemin J. Aiguier, BP 71,
13402 Marseille Cedex 20, France
'lmmunotech SA, 130 Avenue De Lattre De Tassigny, B.P. 177F-13276 Marseille Cedex 9, France
(Received 10 February 1993; revised 27 April 1993; accepted 10 May 1993)
A synthetic gene coding for interleukin-2 (IL-2) was used to produce large amounts of recombinant IL-2 (met-IL2) in Escherichia cofi. Met-IL-2 was found to accumulate in the cytoplasm in an insoluble, aggregated form.
Inclusion bodies located at the pole caps of cells were detected using immunogoldlabelling. Constructs were designed
to fuse the IL-2 gene to DNA fragments encoding signal peptides for an outer-membrane protein (OmpA) or for
a periplasmic protein (PhoA) of E. coli. No significant maturation was observed with these fusion proteins which
were found in an insoluble form in the cytoplasm. The influence of charge disposition at the N-terminus of the
mature portion of the protein was investigated by replacing positively charged amino acids with glutamic acid.
None of the introduced substitutions had any effect. Various factors that might affect expression, secretion and
folding were examined in an attempt to obtain secretion. By fusing IL-2 to the precursor maltose-binding protein
(preMBP) a large fraction of the preMBP-IL-2 protein was correctly processed and transported to the periplasmic
space. IL-2 derived from MBP-IL-2 after FXa cleavage possessed similar specific activity to recombinant IL-2
produced in Chinese Hamster ovary cells.
Introduction
The lymphokine IL-2 plays a central role in the
development of the immune response. It is an inducible
protein synthesized and secreted by activated T lymphocytes with the N-terminal extension of a signal peptide.
As a growth factor, IL-2 acts via a specific receptor and
stimulates the proliferation of IL-2-dependent T cells
(Taniguchi et al., 1983; Smith, 1988). Interest in IL-2 has
increased with the finding that this lymphokine can be
used in the treatment of solid tumours (Rosenberg et al.,
1986) and for partial restoration of T-cell function
(Pahwa et al., 1989).
The cloning and synthesis of human IL-2 in E. coli was
first described by Devos et al. (1983), but most of the
* Author for correspondence. Tel. + 33 91 164117; fax + 33 9 1712124;
e-mail [email protected].
t Present address : Universitat Wien, Institut fur Mikrobiologie und
Genetik, Althanstr. 24, A- 1090 Wien, Austria.
Abbreviations : IL-2, interleukin-2 ; met-IL-2, recombinant interleukin-2 ; MBP, maltose-binding protein.
protein produced remained insoluble in the cytoplasm.
Biological activity therefore could not be restored
completely and was not proportional to total synthesis.
Expression of recombinant proteins in E. coli has become
a valuable tool for basic research and industrial
applications. However, high levels of expression of
eukaryotic proteins lead in most cases to the formation
of insoluble inclusion bodies in E. coli (Marston, 1986).
The extraction, denaturation and renaturation of these
proteins in their active form are time-consuming and
expensive. The efficiency of renaturation and the activity
of the renaturated protein depend on the correct
formation of its disulphide bridges (Marston, 1986).
Expression via secretion offers several advantages over
intracellular expression. If the signal sequence is processed correctly, the N-terminus of the recombinant
protein will be identical to the authentic product. Export
of proteins to the periplasm can prevent the degradation
of the polypeptide. Disulphide bond formation in E. coli
has been shown to occur concomitantly with export from
the cytoplasm (Pollit & Zalkin, 1983; Derman &
Beckwith, 1991). Folded eukaryotic proteins with correct
disulphide bonds can therefore be produced (Oka et al.,
1985; Parker & Wiley, 1989),which is a major advantage.
0001-8143 0 1993 SGM
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 03 Aug 2017 19:13:35
2466
G . Halfmann and others
In some cases the export of heterologous proteins to
the periplasm of E. coli requires only the correct fusion
between a prokaryotic signal peptide and the mature
portion of the foreign protein provided that the protein
itself is normally secreted (Ghrayeb et al., 1984; Duffaud
et al., 1987; Better et al., 1988; Parker & Wiley, 1989).
Some eukaryotic proteins, such as interleukin-lp, can be
recovered by simple osmotic shock (Joseph-Liauzun et
al., 1990); others, however, are unable to enter the export
pathway. For these proteins, additional steps must be
taken to obtain their secretion.
IL-2 was used as a model protein to investigate the
possible export of a protein known to form insoluble
particles in the cytoplasm. This study of the export of IL2 in E. coli represents a continuation of a systematic
analysis of different proteins to determine the prerequisites for the export of heterologous proteins. The
influence of prokaryotic signal sequences has been
investigated for several proteins (Denkfle et al., 1989;
D. Baty, unpublished results).
In this study, different factors influencing protein
maturation and export, such as (i) the nature of the
signal peptide per se (OmpA or PhoA), (ii) the role of
charge distribution near the leader peptidase cleavage
site, (iii) the influence of chaperones (GroES and
GroEL), (iv) the incubation temperature and inducer
concentration, (v) the use of lysis proteins and (vi) fusion
to a known, secreted protein (preMBP) efficiently
targeted by SecB to the export apparatus, were investigated.
Methods
promotor regulated by the lacIQrepressor. The 420 bp EcoRI-Hind111
fragment of BBG3 replicative form (BBL) was inserted between the
EcoRI and HindIII sites of pJF118EH to create pGF25 (Fig. 1). The
ssOmpA-IL-2 fusion was created by linking the synthetic IL-2 gene to
the signal peptide of OmpA by inserting the XhoI-BglII fragment of
BBG3 between the BamHI and HindIII sites of PIN-111-ompA-Hind
(Rentier-Delrue et al., 1988) connected by the oligonucleotides 5'AGCTGCTCCTACTAGC-3' and 5'-TCGAGCTAGTAGGAGC-3'
to create pFQ (Fig. 2). The ssPhoA-IL-2 fusion in pFP6 was
constructed by inserting the XhoI-Hind111 fragment of BBG3 between
the HindIII sites of plasmid pBX4.1 (D. Baty, unpublished result)
connected by the oligonucleotide cassette used for the previous
construction (Fig. 2). The XhoI-Hind111 fragment of BBG3 was
inserted between the StuI and HindIII sites of pMal-p (New England
Biolabs) using the oligonucleotides 5'-GCTCCTACTAGC-3' and 5'TCGAGCTAGTAGGAGC-3' to give pLS28.
Gene expression andprotein characterization.The plasmid-containing
strains were grown in rich medium at the indicated temperatures. Gene
expression was induced by the addition of 01-1 m~ -IPT G . Cell
fractionation (Neu & Heppel, 1965) was performed as follows : cells in
late exponential growth were harvested and washed in TE (10 mMTris/HCl pH 7.5, 1 mM-EDTA). After centrifugation, the cell pellet
was gently resuspended in 30 mM-Tris/HCl, pH 7.5, 20 YO (w/v)
sucrose, 0.5 mM-PMSF, 1 mM-EDTA to a density of 10" bacteria ml-'.
After 10 min at room temperature, the suspension was centrifuged for
5 min at 6000g. The sucrose solution was carefully drained from the
tube and the pellet was resuspended in a same volume of cold water.
The resuspended cells remained on ice for 10 min and were centrifuged
at 15000g for 10 min at 4 "C. The supernatant is referred to as osmotic
shock fluid (periplasmic fraction). The pellet was resuspended in
30 mM-Tris pH 7.5, 0.5 mM-PMSF, 1-25mM-MgCl,, 25 pg DNAase I
ml-l, freeze-thawed six times and centrifuged for 5 min at 15000g.
The pellet and the supernatant are referred to as the membrane and
cytoplasmic fraction, respectively. All samples were resuspended in
SDS loading buffer [60 mM-Tris/HCl pH 6.8, 1 % (w/v) SDS, 10%
Hinm
HindIII
Strains and media. E. coli TB1 araA(1ac proAB) rpsL (#80 lac2
AM15) (Johnston et al., 1986) was used as host for the expression of the
MBP-IL-2 fusion protein. In all other experiments, the plasmids were
expressed in E. coli W3110 (Lloub6s et al., 1986). M9 minimal medium
and LB media have been described by Miller (1972). Antibiotic
resistant plasmids were maintained in transformed strains by addition
of 100 pg ampicillin ml-' and, when required, chloramphenicol
(50 pg ml-') or kanamycin (50 pg I&') to the growth medium.
DNA manipulation. The synthetic gene encoding IL-2 was provided
by BBL. Oligomers were synthesized using a model 380A automatic
DNA synthesizer (Applied Biosystem). The nucleotide sequence
analysis by the T7 polymerase chain termination method was carried
out using bacteriophage M 13 or NaOH-denaturated plasmid templates
[lo80 Ci
(Messing, 1983; Hattori & Sakaki, 1986) and [cz-~~PI~ATP
(40 MBq) mmol-' ; Amersham]. Site-directed mutagenesis was performed as described by Baty et al. (1987b). To exchange lysines (K)
either at position K8 and/or K9 of IL-2 for glutamic acid (E), the
following oligonucleotides were used : K8E, 5'-GAGTTTTCTCAGTGCTCG-3'; K9E, 5'-GTTGAGTTTCCTTAGTGCT-3'; K8/9E, 5'GTTGAGTITCCTCAGTGCT-3'. The substitution of cysteine 125
(C125) by alanine 125 (A125) was performed with the oligonucleotide
C125A, 5'-CGATTGGGAAAGGTG-3'.
Plasmid construction. The plasmid pJFl18EH (Furste et al., 1986)
was used as a source to express IL-2 under the control of the tuc
EcoRI
v
EcoIWHindIII
HindUl
'
5680 bp
Fig. 1. Construction of pGF25. Plasmid GF25 was constructed by
inserting the EcoRI-Hind111 fragment (black) containing the coding
sequence for IL-2 into pJFl18EH. The arrow indicates the direction of
transcription from the tac promotor. The bla (white) and the lucP
(stippled) genes are indicated.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 03 Aug 2017 19:13:35
Targeting of IL-2 to periplasm of E. coli
2467
HindIII
I
\r/
HindIIVEamHI
Oligonucleotides forjunction
pBX4- I
\/
Hind111
XhoI
tac
[
XhoI
PFP6
5370bp
MKQSXIALUWLLTTPVTKAAPTSSSTKKT...
Fig. 2. Construction of the ssOmpA-IL-2 and ssPhoA-IL-2 fusion proteins. The sequences coding for the signal sequence of OmpA
(bold dashed) and PhoA (bold dotted) are indicated. pBX4.1 contains the growth release factor (GRF) gene (dashed) which is removed
during the construction. The other genes are as described in Fig. 1 except that lacl is used in place of lac14. The amino acid sequences
of the signal sequences preceding the IL-2 sequence are indicated in bold at the bottom of the figure.
(v/v) glycerol, 2 % (v/v) /3-mercaptoethanol] heated to 95 "C for 5 min,
analysed by 14 YOSDS-PAGE and immunoblotted on nitrocellulose as
described by Towbin et al. (1979). The met-IL-2 and the IL-2 fusion
proteins were visualized by immunodetection with monoclonal antibodies directed against IL-2.
Pulse-chase experiments. E. coli W3110 harbouring pFQ (preOmpAIL-2) or pFP6 (prePhoA-IL-2) were co-transformed with plasmid
groES/EL (Goloubinoff et al., 1989). Pulse-chase experiments were
performed by diluting overnight cultures of the samples 100-fold in M9
medium. The presence of both plasmids was maintained by the
addition of ampicillin (100 pg ml-') and chloramphenicol(50 pg ml-I).
At an OD, of 0 4 simultaneous induction of the IL-2 fusions and
GroES/GroEL was obtained by the addition of 1 mM-IPTG. After
10 min, 10 pCi (370 kBq) [35S]methionine
was added. After a 20 s pulse
label, the culture was chased for varying periods of time with 1 %
unlabelled methionine. Aliquots corresponding to 60000 c.p.m. were
loaded on a 14% SDS-polyacrylamidegel.
IL-2 bioassay. The biological activity of human IL-2 was tested using
the IL-2-dependentmurine T lymphocytecell line CTLL-2 (Gillis et al.,
1978). The CTLL cell line was cultured in RPMI 1640 (Gibco)
containing 1 mwpyruvate, 2 a-L-glutamine, 50 pg streptomycin and
penicillin ml-' (all chemicals were from Flow), 10 % (v/v) foetal calf
serum (Gibco) and supplemented with 1 ng rhIL-2 ml-' (Sanofi). The
I L 2 bioassay was then performed as described previously. Briefly,
100 pl per well of the cell suspension at lo5 cells ml-' giving lo4 cells per
well was added to 96-well culture plates together with 100 p1 per well of
the IL-2 containing samples. The cells were cultured for 48 h at 37 "C
with 5 % (v/v) CO, in a humidified incubator and were then pulsed for
8 h with 0-5 pCi (18.5 kBq) [3H]thymidine(Amersham) per well. The
cells were then harvested and the radioactivity incorporated in the cell
pellets was determined in a beta counter (Kontron). All determinations
were performed in triplicate. A standard range of serial dilutions of
calibrated IL-2 (Sanofi)was included in every experiment. One Unit of
activity was defined as the amount of IL-2 that leads to half maximum
proliferation of CTLL cells. Concentrations of IL-2 in the samples were
then expressed in Units. In parallel with bioactivity determinations,IL2 concentrations were determined using a commercial immunoassay
(Immunotech) calibrated against the international standard NIBSC
86/564. From these determinations specific activities, expressed as
CTLL Units per ng IL-2, were derived for the different preparations.
Electron microscopy. The detailed procedure for fixation, embedding,
cryosectioning and staining has been described previously (Bernadac 8z
Lazdunski, 1981). The only difference was the use of the monoclonal
antibody directed against IL-2 and of gold-coated protein A instead of
ferritin-labelled anti-rabbit antibodies.
MBP-IG2 puriJication. The protocol used for the purification of
MBP-IL-2 and its cleavage by FXa was that described by New England
Biolabs with some modifications. The amylose column was washed
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 03 Aug 2017 19:13:35
2468
G. Halfmann and others
with 50 mM-Tris/HCl, pH 7.5, 150 mM-NaC1 (TNA buffer). The
periplasmic shock fluid was circulated through the column overnight at
60 ml h-' with TNA. The column was then washed with TNA until the
A,,, reached the base line. The MBP-IL-2 was eluted with TNA
containing 5 mmmaltose.
maturation were examined by analytical cell fractionation. The OmpA-IL-2 and PhoA-IL-2 constructs
showed a level of synthesis similar to that of the wild-
Results
High-level expression of mature IL-2 from pGF25 (Figs
1 and 3) resulted in its aggregation into inclusion bodies
that were located either in the middle of the cells or at the
polar caps (Fig. 5). This confirmed previous findings by
Devos et al. (1983) that IL-2 aggregates in the cytoplasm
of E. coli. To overcome this, we have investigated the
targeting of IL-2 to the periplasm of E. coli by fusing it
either to bacterial signal peptides or to the periplasmically-located maltose-binding protein. IL-2 was fused
to two different E. coli signal sequences, one from the
outer membrane protein (OmpA) and the other from the
periplasmic alkaline phosphatase (PhoA) in pFQ and
pFP6, respectively (Fig. 2). The synthesis of IL-2 and of
the fusion proteins was compared in the same E. coli
W3 110 strain under identical experimental conditions.
About 5 YOof total cellular protein was expressed as metIL-2 or ssOmpA-IL-2 and 15% as ssPhoA-IL-2 (Fig. 3).
The induced E. coli pGF25 (met-IL-2), pFQ (ssOmpAIL-2) and pFP6 (ssPhoA-IL-2) were fractionated and
separated into spheroplasts, periplasmic and membrane
fractions (Fig. 4). The hybrid proteins, ssOmpA-IL-2
(16.9 kDa) or ssPhoA-IL-2 (16.7 kDa), were neither
properly processed to the mature form of IL-2 (14.5 kDa)
nor secreted to the periplasm of E. coli (Fig. 4). The
majority of the fusion proteins remained unprocessed in
the membrane fraction. Only a weak processing of the
precursor to the mature form could be observed (Fig. 4,
arrows in lanes T). To localize the fusion proteins found
in the membrane fraction, immunogold labelling was
performed. Immunogold labelling of E. coli W3 110
revealed no difference in localization of the proteins of
the cytoplasmic version or those derived from the
ssPhoA-IL-2 (Fig. 5) or ssOmpA-IL-2 fusion proteins
(not shown). These observations were consistent with the
absence of processing of the fusion proteins and clearly
demonstrated that they did not even enter the secretion
pathway.
The N-terminus of mature IL-2 has two consecutive
positively charged lysines at positions 8 and 9 that may
perturb the export process. To test this possibility, in
vitro mutagenesis was performed to substitute the lysine
residues (K) at positions 8 and 9 by glutamic acid
residues (E) which should not influence the biological
activity of IL-2 (Ju et al., 1987). K8, K9 or K8/9 were
replaced by E8, E9 or E8/9, respectively. The mutant IL2 gene sequences were fused to the signal sequence of
either OmpA or PhoA. Secretion competence and protein
Fig. 3. Expression of met-IL-2, ssOmpA-IL-2 and ssPhoA-IL-2. IPTG
(1 mM) was added to exponentially growing cultures of E. coli W3110
(pGF25, pFP6 or pFQ) at 37 "C when the OD, reached 1. At 0 , l and
3 h, samples were removed, centrifuged, resuspended in sample buffer,
boiled for 5 min and aliquots corresponding to 0.2 OD,,, units, were
run on 14% SDS-polyacrylamide gels. Proteins were blotted on
nitrocellulose and immunoblot analysis was carried out using an IL-2
monoclonal antibody. The positions of prestained molecular mass
standards (Gibco-BRL) are shown on the left.
Fig. 4. Localization of the recombinant IL-2 and the IL-Zfusion
polypeptides. The expression of cytoplasmic met-IL-2 (pGF25),
ssOmpA-IL-2 (pFQ) and ssPhoA-IL-2 (pFP6) was induced in an
exponentially growing culture at an OD,, of 1.0 by the addition of
1 mM-IPTG at 37 "C. Two hours after induction, samples were taken
and cell fractionation was performed as described in Methods. Protein
fractions corresponding to 0.2 OD,,, units were run on 14% SDSpolyacrylamide gels and immuno-detected with IL-2 antibodies. T
(total cells), sl (supernatant of culture medium), s2 (sucrose fraction
after osmotic shock), P (periplasmic fraction), C (cytoplasmic fraction)
and M (membrane fraction). Arrows indicate the migration of mature
IL-2. The positions of prestained molecular mass standards (GibcoBRL) are shown on the right.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 03 Aug 2017 19:13:35
Targeting of IL-2 to periplasm of E. coli
2469
Fig. 5. Localization of the cytoplasmic version of
IL-2 (pGF25) and the prePhoA-IL-2 fusion by
immuno-microscopy. E. coli W3 1lO(pGF25) and
W3110(pFP6) were cultured in M9 medium at
37 "C to an ODao of 0.4.Cells were induced with
IPTG for 4 h and then fixed with paraformaldehyde (2%, v/v) and glutaraldehyde (0.2%) for
1 h. Ultrathin sections were incubated with 20 pg
IL-2 monoclonal antibodies ml-' for 60 min and
labelled with rabbit anti-mouse serum and protein
A-gold for 30 min. Uninduced and induced cells
of pGF25 (a, b) and of pFP6 (c, d) are shown,
respectively. Intracellular inclusion bodies could
be seen in the cells even without any specific
labelling of the proteins. Bar, 0.1 pm.
type IL-2, as detected by immunoblotting (data not
shown). None of the substitutions of lysine by glutamic
acid (E8, E9 and E8/9) improved precursor maturation
or translocation of the fusion proteins as compared to
the wild-type (data not shown).
IL-2 contains three cysteine residues, two forming a
disulphide bridge. To assess whether the free cysteine
residue 125 (C125), which is not essential for the
biological activity of IL-2 (Ju et al., 1987) but may be
responsible for the inter- and/or intramolecular formation of disulphide bridges, might be involved in the
insolubility of the protein (and therefore increased
protein aggregation), C125 was substituted by an alanine
residue in the plasmids containing ssPhoA-IL-2 and
ssOmpA-IL-2. C125 was also mutated, in addition to the
replacement of the lysine residues by glutamic acid in the
same plasmids. The expression was comparable to that
of the wild type. No difference in protein maturation or
translocation could be detected in the IL-2-Al25 mutants
and their fusion derivatives (data not shown). C125
therefore seemed not to be the dominant factor in the
formation of intracellular aggregates.
We attempted to suppress premature folding and
aggregation of the ssOmpA-IL-2 and ssPhoA-IL-2 fusion
proteins that may have prevented translocation by
increasing the intracellular level of the chaperones
GroEL and GroES. Plasmids encoding groES and groEL
(Goloubinoff et al., 1989) were introduced into E. coli
W3 110 harbouring pFQ (ssOmpA-IL-2) or pFP6
(ssPhoA-IL-2). Simultaneous induction of the IL-2
fusions (ssOmpA-IL-2 or ssPhoA-IL-2) with groES and
groEL did not result in any significant increase in the
maturation of these protein fusions (data not shown).
The pMal-p vector was used to construct a maltosebinding protein (MBP)-IL-2 fusion protein. This contains the sequence IEGR encoding the recognition site of
the specific protease factor Xa, located inside the
polylinker insertion site of the vector (Fig. 6a). After
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 03 Aug 2017 19:13:35
2470
G.HaZfmann and others
Table 1. IL-2 activities in CTLL assay
(4
SP
FXa
-)c
lael
wlE-IL-2
-+J,
rmB bla ori
Sample
IL-2
bioactivity
(CTLL;
Units m1-l)
IL-2
concn
(ng m1-l)
Specific
activity
(CTLL;
Units ng-')
1
412
660
0
0.07
92
62
0
14-29
4.48
10.65
0
rhIL-2*
MBP-IL-2t
IL-2$
Bufferg
* Recombinant Chinese Hamster ovary-derived IL-2.
t MBP-IL-2
fusion protein purified by maltose affinity chromatography.
$ IL-2 derived from MBP-IL-2 by FXa treatment.
Q 10 mM-Tris/HCl, pH 7.5.
proteolysis with factor Xa, the MBP-IL-2 fusion protein
should be cleaved to give authentic IL-2 protein.
Induction resulted in about 5 % of the total cellular
protein being preMBP-IL-2 and MBP-IL-2 fusion
proteins (Fig. 6 b, lanes T). Analytical cell fractionation
revealed that most of the MBP-IL-2 hybrid protein was
secreted to the periplasmic space of E. coli (Fig. 6 b, lanes
P). The fusion protein was degraded during purification
(Fig. 6 c , lane 0). After incubation with FXa, the hybrid
proteins yielded MBP and IL-2 (Fig. 6c, lanes 4 and 8).
The MBP-IL-2 fusion protein and the IL-2 derived from
FXa cleavage were tested for their ability to sustain the
proliferation of the IL-2-dependent cell line CTLL-2
(Table 1). Both recombinant proteins were found to be
active in the assay. Interestingly, FXa cleavage yielded a
protein with higher specific activity which was very
similar to that of a preparation of Chinese Hamster
ovary-derived recombinant IL-2 (rhIL-2).
Discussion
Fig. 6. Expression, secretion and purification of the MBP-IL-2 hybrid
protein. (a) The IL-2 gene was inserted in the pMal vector that encodes
the maltose-binding protein (MBP), the signal sequence (SS) of
preMBP and the factor Xa cleavage site (IEGR). The vertical arrow
heads indicate the sites of cleavage by the signal peptidase (SP) and the
factor Xa (FXa), respectively. The horizontal arrows indicate the
different relevant genes and promotors of the plasmid pLS28. (b) The
preMBP-IL-2 hybrid was induced with 1 mM-IPTG in LB at 37 "C at
an OD,, of 0-5.Two hours after induction, samples were taken and cell
fractionation was performed as described in Methods. Protein samples
corresponding to 0.5 OD,,, units were applied to a 14% SDSpolyacrylamide gel. After immunoblotting, the MBP-IL-2 fusion was
visualized with anti-IL-2 and anti-MBP antibodies. Anti-/3-lactamase
(anti-Bla) was used to control the quality of the periplasmic fraction.
MBP-IL-2* represents a degradation product of MBP-IL-2. Lanes T
(total cells), P (periplasmic fraction), C (cytoplasmic fraction) and M
(n..mbrane fraction). The positions of prestained molecular mass
standards (Gibco-BRL) are shown on the right. (c) After purification
the preMBP-IL-2 protein was treated for 0, 4 or 8 h with FXa. After
immunoblotting, the proteins were visualized with anti-IL-2. The
positions of prestained molecular mass standards (Gibco-BRL) are
shown on the left.
In general, recombinant polypeptides can be accumulated to higher levels when retained intracellularly than
when secreted. However, export of a polypeptide may be
preferred for several reasons: (i) to allow formation of
disulphide bridges required for function; (ii) to decrease
proteolytic degradation, although severe proteolysis of
heterologous proteins may occur in the periplasmic
space ; (iii) to facilitate the purification of recombinant
proteins. For these reasons, we tried to obtain an E. coli
construct able to produce and export IL-2. The fusion of
a prokaryotic signal peptide to the mature IL-2 did not
promote export which contrasted with results observed
for other recombinant proteins produced in E. coli
(Ghrayeb et al., 1984; Better et al., 1988; Parker &
Wiley, 1989; Denkfle et al., 1989). Not only, were the
ssOmpA-IL-2 and the ssPhoA-IL-2 not exported but
they did not even enter the export pathway and it is
therefore likely that they aggregated very early after
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 03 Aug 2017 19:13:35
Targeting of IL-2 to periplasm of E. coli
synthesis. The signal peptides may have become buried
very soon after synthesis and were unable to direct the
protein to the export machinery. By using IL-2 as a
model polypeptide, we have investigated the effects of
different factors that may interfere with the export of
recombinant proteins.
First we altered the charge distribution near the
cleavage site since it had been demonstrated that by
replacing basic residues in this region, defective export
could be corrected (Summers & Knowles, 1989; Parker
& Wiley, 1989). Although the mechanism of export in
prokaryotes and eukaryotes appears to be similar, the
presence of basic residues near the cleavage site is more
frequent in eukaryotic secretory proteins (Watson, 1984;
von Heijne, 1986). Whether this reflects some difference
between early events in the secretory process of prokaryotes and eukaryotes remains an intriguing question.
Expression and secretion of IL-2 in Streptomyces lividans
resulted in active IL-2 being secreted to the culture
supernatant with, however, inefficient processing and
translocation of the precursor (Bender et al., 1990). The
inability of the IL-2 fusions to enter the export pathway
in E. coli may be explained mainly by thermodynamic
considerations of the conformation of the mature region,
independently of the signal sequence used. By contrast,
experiments on the export of human interleukin-18
showed a remarkable influence of the bacterial peptide
on the secretion of this eukaryotic protein (Denifle et al.,
1989).
Observations concerning the nature of the N-terminal
residues of a protein as a key factor in secretion (for
review, see Boyd & Beckwith, 1990) led us to introduce
mutations in the non-secretable hybrid proteins that
comprise the signal peptides of PhoA and OmpA fused
to the complete IL-2 gene. Some of the mutations were
designed to change the net charge near the N-terminus of
the mature form of IL-2, while other alterations were
made to avoid the incorrect formation of disulphide
bridges (leaving the N-terminal sequences unchanged).
We found that the substitution of the basic lysine
residues by glutamic acid did not improve the secretion
competence of the fusion proteins, regardless of the
replacement of one or the other (K8 or K9) or both
(K8/9) residues. Results obtained in studies of the
influence of the net charge at the N-terminus of the
mature form of prokaryotic proteins on protein export
(MacIntyre et al., 1990; Yamane & Mizushima, 1988)
therefore do not apply to the secretion of heterologous
proteins when these proteins tend to aggregate spont aneously .
For export, the precursor must be maintained in a
translocation-competent conformation and the signal
peptide must remain exposed to the secretory machinery.
Early folding or aggregation of the precursor leads to
2471
loss of translocation. Bacterial chaperone factors, including trigger factor, GroEL, GroES, DnaK and SecB
appear to be required to prevent early protein folding
and to keep the precursor in a translocation-competent
conformation (for review, see Saier et al., 1989; Lee &
Olins, 1992). However, the rapid folding of IL-2 in the
cytoplasm could not be inhibited by increasing the
intracellular amount of GroEL and GroES.
Two general problems encountered in the synthesis of
eukaryotic proteins in E. coli at a high level are protein
aggregation and/or incomplete processing by bacterial
peptidase activity. Altering growth conditions has been
shown to influence favourably the yield of correctly
processed material, e.g. by lowering the temperature
(Schein & Noteborn, 1988) or by limited induction
(Ghrayeb et al., 1984). Because induction with 1 mMIPTG leads to a high production level of the ssOmpAIL-2 and ssPhoA-IL-2 fusions, we reduced the level of
IPTG to 1 0 p in
~ LB and/or decreased the growth
temperature to 28 "C. We could not observe any
significant difference in the extent of maturation (precursor stability and processing competence) under
different conditions (data not shown).
Since we could not promote export by the different
techniques described above, we tried to apply our
experience in the use of lysis proteins to promote nonspecific release of proteins to the extracellular medium.
Neither the colicin A lysis protein (Baty et al., 1987a),
nor the bacteriophage 4x174 lysis protein E (Witte et al.,
1989) promoted the extracellular release of putatively
soluble forms of IL-2.
IL-2 in its original environment may have some
additional factors which influence its expression and
secretion, either as directed by its original signal peptide
or by other factors such as eukaryotic chaperones. In the
case of the maltose-binding protein, attainment of a
stable tertiary structure in the cytoplasm correlates with
the loss of competence for secretion (Randall & Hardy,
1986).Thus, once a bacterial protein is properly folded in
the cytoplasm, it seems that there is no mechanism of
repackaging it for export. Binding proteins and chaperones appear to prevent the nascent polypeptide chains
from crossing the energy barrier to a folded stable state,
either stabilizing the partially folded structure or protecting these incompletely folded, flexible molecules from
premature aggregation (Fischer & Schmid, 1990).
In a further attempt to promote secretion of IL-2, the
mature protein was fused to the C-terminus of the
maltose-binding protein. The export of the latter has
been studied extensively and plasmid vectors have been
designed based on these studies to promote high levels of
expression and secretion of recombinant proteins (for
review, Bankaitis et al., 1985; Randall et al., 1987). This
system allows the protein to fold by itself without
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 03 Aug 2017 19:13:35
2472
G . Halfmann and others
interference of an adjacent signal sequence and has been
used successfully with various proteins to obtain export
in E. coli (Martineau et al., 1992;Bregkgkre & Bedouelle,
1992). This approach was successful and allowed export,
processing of the MBP-FX-IL-2 fusion protein and
production of active IL-2. Export of preMBP-IL-2
presumably occurs because export begins cotranslationally before completion of the synthesis of the IL-2
moiety. However, we cannot rule out the hypothesis that
interaction of the hybrid protein with the SecB chaperone
plays a part in preventing misfolding and aggregation.
We are grateful to M. Delaage for helpful discussions. We thank
J. M. Clement for the anti-maltose-binding protein. We also thank
M. Green and H. Rickenberg for careful reading. This work was
supported by the Commission of the European Communities, the
Ministkre de la Recherche et de la Technologie and the Fondation pour
la Recherche Medicale. G. H. was the recipient of an EC postdoctoral
fellowship.
References
BANKAITIS,
V. A., RYAN,J. P., RASMUSSEN,
B. A. & BASSFORD,
P. J., JR
(1985). The use of genetic techniques to analyse protein export in
Escherichia coli. Current Topics in Membrane Transport 24, 105-1 50.
BAR, D., LLOUBES,
R., GELI,V., LAZDUNSKI,
C. & HOWARD,P. S.
(1987a). Extracellular release of colicin A is non-specific. EMBO
Journal 6, 2463-2468.
BATY,D., KNIBIEHLER,
M., VERHEU,H., PATTUS,F., SHIRE, D.,
BERNADAC,
A. & LAZDUNSKI,
C. (1987 b). Site-directed mutagenesis
of the COOH-terminal region of colicin A: effect on secretion and
voltage-dependent channel activity. Proceedings of the National
Academy of Sciences of the United States of America 84, 1152-1 156.
BENDER,
E., KOLLER,
K. P. & ENGELS,
J. W. (1990). Secretory synthesis
of human interleukin-2 by Streptomyces lividans. Gene 86, 227-232.
BERNADAC,
A. & LAZDUNSKI,
C. (1981). Immunoferritin labelling of
ultrathin frozen sections of Gram-negative bacterial cells. Biology of
the Cell 41, 21 1-216.
BETTER,
M., CHANG,
C. P., ROBINSON,
R. R. & HORWITZ,
A. H. (1988).
Escherichia coli secretion of an active chimeric antibody fragment.
Science 240, 1041-1043.
BOYD,D. & BECKWITH,
J. (1990). The role of charged amino acids in
the localization of secreted and membrane proteins. Cell 62,
103 1-1033.
B ~ G E G ~ RF.E ,& BEDOUELLE,
H. (1992). Expression, export and
purification of antibody fragments fused to the maltose-binding
protein of Escherichia coli. Compte rendu de PAcademie des Sciences,
Paris 314, serie 111, 527-532.
DEN~FLE,
P., KORVARIK,
S . , CIORA,T., GOSSELET,
N., BENICHOU,
J. C.,
LATTA,M., GUINET,F., RYTER,A. & MAYAUX,J. F. (1989).
Heterologous protein export in Escherichia coli: influence of bacterial
signal peptides on the export of human interleukin 1B. Gene 85,
499-5 10.
DERMAN,A. I. & BECKWITH,
J. (1991). Escherichia coli alkaline
phosphatase fails to acquire disulfide bonds when retained in the
cytoplasm. Journal of Bacteriology 173, 7719-7722.
D~vosR.,
, PLAETINCK,
G., CHEROUTRE,
H., SIMONS,
G., DEGRAVE,
W.,
TAVERNIER,
J., REMAUT,
E. & FIERS,W. (1983). Molecular cloning of
human interleukin 2 cDNA and its expression in E. coli. Nucleic
Acids Research 11, 4307-4323.
DUFFAUD, G. D., MARCH,P. E. & INOUYE,
M. (1987). Expression and
secretion of foreign proteins in Escherichia coli. Methods in
Enzymology 153,492-502.
FISCHER,G. & SCHMID,F.X. (1990). The mechanism of protein
folding. Implications of in vitro refolding models for de novo protein
folding and translocation in the cell. Biochemistry 29, 2205-2212.
FURSTE,J. P., PANSEGRAU,
W., FRANK,R., BLOCKER,
H., SCHOLZ,
P.,
BAGDASARIAN,
M. & LANKA,E. (1986). Molecular cloning of the
plasmid RP4 primase region in a multi-host-range tacP expression
vector. Gene 48, 119-131.
GHRAYEB,
J., KIMURA,
H., TAKAHARA,
M., HSIUNG,H., MASUI,Y . &
INOUYE,
M. (1984). Secretion cloning vectors in Escherichia coli.
EMBO Journal 3, 2437-2442.
GILLIS,S., FERM,
M. M., Ou, W. & SMITH,K. A. (1978). T-cell growth
factor : parameters of production and a quantitative microassay for
activity. Journal of Immunology 120, 2027-2032.
GOLOUBINOFF,
P., GATENBY,
A. A. & LORIMER,G. H. (1989). GroE
heat-shock proteins promote assembly of foreign prokaryotic
ribulose bisphosphate carboxylase oligomers in Escherichia coli.
Nature, London 337, 44-47.
HATTORI,M. & SAKAKI,
Y. (1986). Dideoxy sequencing method using
denatured plasmid templates. Analytical Biochemistry 152, 232-238.
VON HEIJNE,
G. (1986). Net N-C charge imbalance may be important
for signal sequence function in bacteria. Journal of Molecular Biology
192, 287-290.
JOHNSTON,
T. C., THOMPSON,
R. B. & BALDWIN,
T. 0. (1986). Nucleotide sequence of the IexB gene of Vibrio harveyi and the complete
amino acid sequence of the beta subunit of bacterial luciferase.
Journal of Biological Chemistry 261, 480548 11.
JOSEPH-LIAUZUN,
E., LEPLATOIS,P., LEGOUX,R., GUERVENO,
V.,
MARCHESE,
E. & FERRARA,
P. (1990). Human recombinant interleukin-1 beta isolated from Escherichia coli by simple osmotic shock.
Gene 86, 291-295.
Ju, G., COLLINS,L., KAFFKA,K. L., TSIEN,W. H., CHIZZONITE,
R.,
CROW,R., BHATT,R. & KILIAN,P. L. (1987). Structure-function
analysis of human interleukin-2. Identification of amino acid residues
required for biological activity. Journal of Biological Chemistry 262,
5723-573 1.
LEE, S . C. & OLINS,P. 0. (1992). Effect of overproduction of heat
shock chaperones GroES and DnaK on human procollagenase
production in Escherichia coli. Journal of Biological Chemistry 267,
2849-28 52.
LLOUB$S,
R., BATY,D. & LAZDUNSKI,
C. (1986). The promoters of the
genes for colicin production, release and immunity in the ColA
plasmid : effects of convergent transcription and lexA protein.
Nucleic Acids Research 14, 2621-2636.
MACINTYRE,
S., ESCHBACH,
M. L. & MUTSCHER,
B. (1990). Export
incompatibility of N-terminal basic residues in a mature polypeptide
of Escherichia coli can be alleviated by optimising the signal peptide.
Molecular and General Genetics 221, 466-474.
MARSTON,
F. A. 0. (1986). The purification of eukaryotic polypeptides
synthesized in Escherichia coli. Biochemical Journal 240, 1-12.
MARTINEAU,
P., GUILLET,
J.-G., LECLERC,
C. & HOFNUNG,
M. (1992).
Expression of heterologous peptides at two permissive sites of the
MalE protein : antigenicity and immunogenicity of foreign B-cell and
T-cell epitopes. Gene 113, 3546.
MESSING,J. (1983). New M13 vectors for cloning. Methods in
Enzymology 101, 20-78.
MILLER,J. H. (1972). Experiments in Molecular Genetics. Cold Spring
Harbor Laboratory, NY: Cold Spring Harbor Laboratory.
NEU, H. C. & HEPPEL,L. A. (1965). The release of enzymes from
Escherichia coli by osmotic shock and during formation of
spheroplasts. Journal of Biological Chemistry 240, 3685-3692.
OKA, T., SAKAMOTO,
S., MIYOSHI,K. I., FUWA, T., YODA, K.,
YAMASAKI,
M., TAMURA,
S . & MIAYAKE,
T. (1985). Synthesis and
secretion of human epidermal growth factor by Escherichia coli.
Proceedings of the National Academy of Sciences of the United States
of America 82, 72 12-72 16.
PAHWA,R., CHATILA,
T., PAHWA,S., PARADISE,
C., DAY,N. K., GEHA,
N. & GOOD,R. A. (1989).
R., SCWARTZ,S . A., SLADE,H., OYAIZU,
Recombinant interleukin 2 therapy in severe combined immunodeficiency disease. Proceedings of the National Academy of Sciences
of the United States of America 86, 5069-5073.
PARKER,
K. C. & WILEY,D. C. (1989). Overexpression of native human
fl-microglobulin in Escherichia coli and its purification. Gene 83,
117-124.
POLLIT,
S . & ZALKIN,
H. (1983). Role of primary structure and disulfide
bond formation in B-lactamase secretion. Journal of Bacteriology
153, 27-32.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 03 Aug 2017 19:13:35
Targeting of IL-2 to periplasm of E. coli
RANDALL,
L. L. & HARDY,S . J. S . (1986). Correlation of competence
for export with lack of tertiary structure of the mature species: a
study in vivo of maltose-binding protein in E. coli. Cell 46, 921-928.
RANDALL,L. L., HARDY,
S. J. S. & THOM,J. R. (1987). Export of
protein : a biochemical view. Annual Review of Microbiology 41,
507-541.
RENTIER-DELRUE,
F., SWENNEN,D. & MARTIAL,J. (1988). PIN-111
ompA secretion vectors: modification of the ompA signal peptide
sequence for easier insert cloning. Nucleic Acids Research 16, 8726.
ROSENBERG,
S. A., SPIESS,
P. & LAFRENIERE,
R. (1986). A new approach
to the adoptive immunotherapy of cancer with tumor-infiltrating
lymphocytes. Science 233, 1318-1 321.
SAIER,M. H., WERNER,P. K. & MWLLER,M. (1989). Insertion of
proteins into bacterial membranes : mechanism, characteristics, and
comparisons with the eucaryotic process. Microbiological Reviews
53, 333-366.
SCHEIN,C. H. & NOTEBORN,
M. H. M. (1988). Formation of soluble
proteins in Escherichia coli is favored by lower growth temperature.
Biol Technology 6, 29 1-294.
SMITH,K. (1988). Interleukin-2 : inception, impact, and implication.
Science 240, 1169-1 176.
2473
SUMMERS,R. G . & KNOWLES,J. R. (1989). Illicit secretion of a
cytoplasmic protein into the periplasm of Escherichiu coli requires a
signal peptide plus a portion of the cognate secreted protein. Journal
of Biological Chemistry 264, 2007&2008 1.
TANIGUCHI,
T., MATSUI,H., FUJITA,T., TAKAOKA,
C., KASHIMA,
N.,
YOSHIMOTO,
R. & HAMURO,
J. (1983). Structure and expression of a
cloned cDNA for human interleukin-2. Nature, London 302,305-3 10.
TOWBIN,H., STAEHELIN,
T. & GORDON,J. (1979). Electrophoretic
transfer of proteins from polyacrylamide gels to nitrocellulose
sheets : procedure and some applications. Proceedings of the National
Academy of Sciences of the United States of America 76,4350-4354.
WATSON,M. E. E. (1984). Compilation of published signal sequences.
Nucleic Acids Research 12, 5 145-5 1 64.
WITTE,A., BLASI,U., HALFMANN,
G., SZOSTAK,
M., WANNER,G. &
LUBITZ,W. (1989). Phi X174 protein E-mediated lysis of Escherichiu
coli. Biochimie 72, 191-200.
YAMANE,
K. & MIZUSHIMA,
S. (1988). Introduction of basic amino acid
residues after the signal peptide inhibits protein translocation across
the cytoplasmic membrane of Escherichiu coli. Relation to the
orientation of membrane proteins. Journal of Biological Chemistry
263, 19690-19696.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 03 Aug 2017 19:13:35