Download Molecular and General Genetics

Mol Gen Genet (1988) 215:181 183
© Springer-Verlag 1988
Expression of a cloned p-glucanase gene
from Bacillus am yloliquefaciens
in an Escherichia coli relA strain after plasmid amplification
Michael Hecker 1, Sabine Riethdorf I, Christiane Bauer 1, Andreas Schroeter 1, and Rainer Borriss 2
1 Department of General Microbiology, Ernst-Moritz-Arndt-University, DDR-2200 Greifswald
2 Zentralinstitut ffir Genetik und Kulturpflanzenforschung, DDR-4325 Gatersleben
Summary. Amino acid starvation of cells of the Escherichia
coli relA strain, CP79, which cannot accumulate guanosine
tetraphosphate (ppGpp) in response to amino acid limitation, increased the pEG1 plasmid content about 5- to 7-fold
in comparison with exponentially growing cells
(pEG1 : pBR322 with an insertion of Bacillus amyloliquefaciens D N A coding for fl-glucanase). In contrast, no pEG1
amplification occurred in E. coli CP78, the stringently controlled counterpart, after amino acid starvation. In order
to verify these results, the plasmid D N A content was monitored by measuring the expression of pEGl-encoded fl-glucanase from B. amyloliquefaciens both before and after plasmid amplification. When amino acid starved CP79 cells
were given an additional dose of amino acids, a more than
10-fold increase in pEGl-encoded fl-glucanase activity (per
cell mass) was measured. This increase in enzyme activity
correlates with pEG1 amplification during amino acid limitation. Under comparable conditions the activity of fl-glucanase was not increased in strain CP78, which did not
amplify the plasmid. We suggest that the replication of
pEG1 in amino acid starved E. coli cells is somehow under
negative control by ppGpp. Moreover, we found the Bacillus fl-glucanase in E. coli relA cells to be excreted into the
growth medium after starvation and overexpression.
Key words: Plasmid pEG1 - relA -fl-glucanase overexpression/excretion - Plasmid amplification
While in E. coli inhibition of protein synthesis by chloramphenicol causes a rapid amplification of ColEl-related plasraids, a comparably low increase of plasmid D N A follows
amino acid starvation (Clewell and Helinski 1972). This
amino acid starvation is known to be accompanied by the
accumulation of ppGpp, whereas chloramphenicol treatment prevents the formation of p p G p p (Cashel 1975).
Therefore we asked the question whether p p G p p might interfere with plasmid amplification in amino acid starved
cells.
In order to elucidate this problem, we studied plasmid
D N A amplification in a pair of E. coli relA and relA +
strains after amino acid exhaustion. In the E. coli relA
strain, CP79, which cannot accumulate p p G p p in response
to amino acid limitation, the synthesis of pBR322 D N A
continued under this condition, leading to an increase of
Offprint requests to .' M. Hecker
the plasmid D N A content per cell mass. In the stringently
controlled isogenic strain CP78, however, p p G p p was produced in response to amino acid limitation and plasmid
amplification did not occur. Therefore we suggested that
the replication of ColEl-related plasmids might be negatively controlled by p p G p p (Hecker et al. 1983). Lin-Chao and
Bremer (1986), who found an even higher pBR322 D N A
content in amino acid starved cells of an E. coli B relA +
strain than in its relaxed counterpart, did not agree with
our conclusion.
In order to exclude artifacts due to different efficiencies
of plasmid recovery, we decided to determine the activity
of the plasmid encoded fl-glucanase in E. coli cells harboring
the plasmid pEG1 both before and after plasmid amplification. The plasmid pEG1 was constructed by inserting a
3.6 kb EcoRI fragment of Bacillus amyloliquefaciens D N A
into the EcoRI site of pBR322 (Borriss et al. 1985). This
fragment contains the bgl gene coding for the endo-fl-l,31,4-glucanase. The aim of these experiments was:
1. To study whether pBR322 derived recombinant plasmids
can also be amplified by amino acid starvation in E. coli
relA.
2. To gain further support for plasmid amplification by
measuring a plasmid encoded enzyme activity which is supposed to be correlated with gene dosage.
3. To demonstrate that the amplification regime described
in this paper can be used for the overexpression of plasmid
encoded genes.
Materials and methods
The E. coli strains CP78 (arg, thr, leu, his, thi) and CP79
(relA, arg, thr, leu, his, thi), each harboring the plasmid
pEG1, were employed. Bacteria were grown at 30 ° C under
continuous shaking in a synthetic medium (Hecker et al.
1983) supplemented with thiamine (10 rag/l), glucose (6 g/l),
NazHPO4" 12 H 2 0 (322 mg/1), NHgC1 (1 g/l) and threonine,
arginine, leucine, and histidine (each 50 rag/l). Cells entered
the stationary growth phase as a result of arginine exhaustion. Ten hours later, amino acids (threonine, arginine, leucine and histidine, each 50 rag/l) were added again and
growth resumed in CP78.
The plasmid D N A content was determined as described
earlier (Hecker et al. 1985). Extracellular and cellular flglucanase activities were determined according to Borriss
(1981 ; for experimental details see Fig. la and b).
182
Results and discussion
Wherreas in arginine-starved cells of E. coli relA strain
CP79 pEG1 plasmid content per cell mass increased 5- to
7-fold, the plasmid content did not significantly change in
cells of the relA ÷ strain, CP78 (Fig. la and b).
The activity of the plasmid encoded fl-glucanase was
predicted to be an indicator for bgI gene dosage. In amino
acid starved E. coli relA cells the activity of fl-glucanase
remained unchanged in spite of plasmid amplification. An
approximately 10-fold increase in fl-glucanase activity per
cell mass was observed after the addition of the required
amino acids to amino acid starved ceils. This increase in
enzyme activity correlates with the preceding plasmid am-
+ r.
CP79pEGI
o
~
,?, gl
.~--800-~
[
-600 - ~
7_]
1~200 a~ I
a,
_
5
I
z_
// ~
:
+1
-1
_B-~
l
CP78pEGI
zc+
~_
13[-
._
>. 11p
/
×
/
/
7t-
3-+00 +L /
2-400~ I /
/ ~ 3F/
° o
o
~---Y~
Fig. la and b. pEG1 plasmid content and #-glucanase activity during growth and amino acid starvation of the Eseherichia coli relA
strain CP79 (Fig. la) and of the E. coli relA + strain CP78 (Fig. lb).
Cells were grown to stationary phase at 30° C. After t0 h amino
acids were added again (arrow). The plasmid content was determined as described earlier (Hecker et al. t985; ng plasmid DNA
per unit of OD at 500 nm). #-glucanase activity was assayed at
37o C in cell extracts (periplasmic activity) and in supernatants
(extracellular activity) by determination of liberated sugars from
lichenan by the dinitrosalicyclic acid method (see Borriss 1981).
All enzyme activities (cellular and extracellular) were expressed
in optical density units (OD 530 nm) per cell mass (unit of OD
at 500 nm). The #-glucanase activity of exponentially growing cells
was taken as 1.0. Growth was followed by measuring the OD
at 500 nm
plification during the starvation phase (Fig. la). Earlier experiments showed that Bacillus fl-glucanase was mainly accumulated within the periplasmic space of E. coli cells
(Cantwell and McConnell 1983; Borriss et al. 1985). We
found that the overexpression of otherwise periplasmic flglucanase in E. coli relA led to excretion of this enzyme
into the extracellular medium (Fig. la).
In E. coli relA + cells starved for arginine, plasmid amplification did not occur. There was no increase offl-glucanase
activity after the resumption of growth triggered by the
addition of amino acids. Furthermore, only a low extracellular enzyme activity was measured (Fig. lb).
Surprisingly, there was an approximately 3-fold increase
in fl-glueanase activity during stationary growth phase of
E. coli reIA + cells which accumulated ppGpp. It can be
supposed that in the presence of ppGpp, R N A polymerase
might prefer genes, such as the bgl gene, that are not under
negative control by ppGpp (see Gallant 1979; Ryals et al.
1982). Moreover, growth of both strains exhibited striking
differences. On one hand, in amino acid starved cells of
E. coli relA harboring amplified pEGI DNA, there was
a low increase of cell mass after readdition of amino acids.
On the other hand, growth resumed in starved cells of E.
coli relA + when amino acids were added. This result raises
the question whether, in the relA strain, growth inhibition
might be due to an overexpression of plasmid encoded
genes.
Thus, our results led us to suppose that the amplification regime described here could be useful in obtaining high
levels of expression of foreign genes in E. coli. Alterations
in outer membrane composition of starved E. coli reIA cells
might be responsible for the excretion of otherwise periplasmic proteins (see Hancock 1987). Altered outer membrane
proteins might arise from mistranslation when ppGpp is
absent (Gallant 1979).
Additionally, we found that the excretion is related to
the rate of overexpression. We could further increase the
synthesis of #-glucanase by fusion of the bgl gene to strong
promoters. In these experiments we obtained an enhanced
excretion offl-glucanase (not shown here). In electrophoretic analysis of cellular and extracellular proteins we have
shown that this enzyme excretion was not caused by cell
lysis, at least in the early stages of high level expression.
Furthermore, we could not find any extracellular activities
of cytoplasmic enzymes measured in our studies (not shown
here). A similar excretion of an otherwise periplasmic protein was shown by Pages et al. (1987).
Summarizing our results, we conclude that amplification
of ColEl-related plasmids occurs in amino acid starved cells
of E. coli relA (CP79). Because our findings did not correspond with results published by Lin-Chao and Bremer
(1986), we have studied the plasmid replication in other
relA +/reIA strains ofE. coli. In all pairs tested so far amplification was only established in the relA strain (in preparation). Unfortunately, the discrepancy between our results
and those of Lin-Chao and Bremer (1986) still remains to
be explained. Thus, the mechanism of plasmid amplification
described in this study has to be investigated in more detail
in order to define the target for the ppGpp-mediated regulation of pEG1 replication.
Acknowledgements. We thank Dr. Jiirgen Hofemeister (Gatersle-
ben) for critical reading of the manuscript.
183
References
Borriss R (1981) Purification and isolation of an extracellular flglucanase from Bacillus IMET B376. Z Allg Mikrobiol
22:293-298
Borriss R, B/iumlein H, Hofemeister J (1985) Expression in
Escherichia coli of a cloned fl-glucanase gene from Bacillus amyloliquefaciens. Appl Microbiol Biotechnol 22:63-71
Cantwell BA, MeConnell DJ (1983) Molecular cloning and expression of a Bacillus subtilis fl-glucanase gene in Escherichia coli.
Gene 23:211-219
Cashel M (1975) Regulation of bacterial p ~ i p p and pppGpp.
Annu Rev Microbiol 29:301-318
~
Clewell DB, Helinski DR (1972) Effect of growth conditions on
the formation of the relaxation complex of supercoiled ColE1
deoxyribonucleic acid and protein in Escherichia coli. J Baeteriol 110:1135-1146
Gallant JA (1979) Stringent control in E. coli. Annu Rev Genet
13 : 393415
Hancock REW (1987) Role of porines in outer membrane permeability. J Bacteriol 169 : 929-933
Hecker M, Schroeter A, Mach F (1983) Replication of pBR322
DNA in stringent and relaxed strains of Escherichia coll. Mol
Gen Genet 190: 355-357
Hecker M, Schroeter A, Mach F (1985) Escherichia coli relA strains
as hosts for amplification of pBR322 plasmid DNA. FEMS
Microbiol Lett 29:331-334
Lin-Chao S, Bremer H (1986) Effect of relA function on the replication of plasmid pBR322 in Escherichia coli. Mol Gen Genet
203:150-153
Pages J-M, Anba J, Lazdunski C (1987) Conditions leading to
secretion of a normally periplasmic protein in Escherichia coli.
J Bacteriol 169:1386-1390
Ryals J, Little R, Bremer H (1982) Control of rRNA and tRNA
syntheses in Escherichia coli by guanosine tetraphosphate. J
Bacteriol 151 : 1261-1268
C o m m u n i c a t e d by R. D e v o r e t
Received April 6, 1988