A novel strategy for the periplasmic production of heterologous proteins in E. coli

فرضية

2007

Malik, Ajamaluddin . 2007

Since the advent of recombinant DNA technology, application of proteins in pharmaceuticals
was remarkably changed. All these proteins can not be obtained from natural sources because
many of them are present in extremely low amounts. Genetically engineered proteins having
special advantages (e.g. Insulin analogs) are as such artificial molecules and can therefore only
be obtained recombinantly. Escherichia coli offer a means for the rapid and economical
production of recombinant proteins. These advantages are coupled with a wealth of
biochemical and genetic knowledge. Although significant progress has been made in
improvement of transcription, translation and secretion, obtaining the product in a soluble and
bioactive form is still a major challenge.

Many naturally secreted proteins such as hormones, growth factors, antibodies etc. are used for
diagnostic and therapeutic applications. In general, if secreted proteins contain two or more
cysteines then they form disulfide bonds that are essential for structure formation and function.
Production of these proteins in the cytoplasm of E. coli usually gives inclusion bodies due to a
reducing environment. In vitro oxidative refolding can be quite difficult. Secretion of such
proteins into the periplasm of E. coli provides a better chance of oxidative folding due to the
presence of oxidative folding and disulfide isomerization machinery. Besides the formation of
correct disulfide bonds, production in the periplasm can also reduce the proteolysis and can
ease the purification. For the translocation of proteins to the periplasm, a prokaryotic signal
peptide is required, but the presence of this signal sequence does not always ensure efficient
protein translocation. The sequence next to the signal peptide cleavage site in the mature part of
the protein and other region of mature part play an important role in translocation. If this is the
case, fusion to a full length periplasmic protein that is well produced and properly folded is
more promising.

In this study, human pepsinogen and human proinsulin were used as model proteins to establish
an expression system for the production of disulfide bonded proteins in native form in the
periplasm of E. coli. Both, pepsinogen and proinsulin contain three disulfide bonds. For the
production of pepsinogen, three different signal sequences (pelB, ompT and dsbA) were fused
to its N-terminus for translocation. The genes were expressed from the T7 promoter in pET vectors, but no pepsin activity could be determined in the periplasm. The expression level was
very high, so that pepsinogen remained in the cytosol along with the signal sequence. Next,
pelB-pepsinogen was cloned into pTrc99a and pBAD22 to replace the T7 promoter by the
hybrid trc and by the weak araBAD promoter, respectively. Using the trc promoter, about 16
μg pepsinogen per liter OD1 was determined in the periplasm. However, production of
pepsinogen was not reproducible even though different strains and culture conditions were
tested. In case of the araBAD promoter, expression level was very poor and no significant
pepsin activity could be determined.

As a new approach, human pepsinogen was fused to the C-terminus of ecotin, E. coli trypsin
inhibitor, which is a homodimeric periplasmic protein (16 kDa). Each subunit contains one
disulfide bond. It is a highly stable protein and withstands even 100 °C and pH 1.0 for 30 min.
The ecotin-pepsinogen fusion was expressed in pTrc99a and was translocated into the
periplasm with the help of the ecotin signal peptide. When the periplasmic extract was
acidified, the ecotin-pepsinogen fusion was converted into pepsin, indicating that pepsinogen
was produced in its native form. In shake flask experiments, the amount of native ecotinpepsinogen
present in the periplasm was 100 μg per liter OD1 that corresponds to 70 μg
pepsinogen. After large scale cultivation, the native fusion protein was purified to homogeneity
in three step of purification, Ni-NTA, ion exchange and gel filtration chromatography with a
yield of 23%. From 30 g wet biomass, 5.2 mg ecotin-pepsinogen corresponding to 3.6 mg
pepsinogen was obtained. This corresponded to 7.6 mg native pure pepsinogen per liter
fermentation broth.
To identify and quantify pepsinogen in periplasmic samples, a highly sensitive assay method
was needed due to the low amount of protein present in the periplasmic samples. A
fluorometric assay for pepsin and pepsinogen was developed using enhanced green fluorescent
protein (EGFP) as a substrate. Acid denaturation of EGFP resulted in a complete loss of
fluorescence that was completely reversible on neutralization. In the proteolytic assay
procedure, acid-denatured EGFP was digested by pepsin or activated pepsinogen. After
neutralization, the remaining amount of undigested EGFP refolded and was quantified by
fluorescence. The sensitivity of the proteolytic assay was dependent on the incubation time and
temperature. If digestion of EGFP was done for 3 hours at 37 °C, even 50 pg pepsin were sufficient to give a reasonable signal. Under standard digestion conditions at 20 °C for 10 min,
the sensitivity of pepsin or activated pepsinogen was in the range of 0-30 ng. Using porcine
pepsin, the specific digestion rate of EGFP under standard condition was 38 ± 6.7 ng EGFP ng-1
pepsin min-1. Acid treated, activated porcine pepsinogen revealed a similar specific digestion
rate (37.2 ± 5.2 ng EGFP ng-1 activated pepsinogen min-1). The pepsin-catalyzed EGFP
digestion showed typical Michaelis–Menten kinetics.

To evaluate the applicability of ecotin as a periplasmic fusion tag, a second human protein from
a diverse family, proinsulin, was chosen. Proinsulin contains three non-consecutive disulfide
bonds. It was genetically fused to the C-terminus of ecotin. The ecotin-proinsulin fusion was
cloned downstream of the trc promoter in pTrc99a and the fusion protein was produced in E.
coli BL21(DE3)Gold. Parameters were optimized for the improvement of ecotin-proinsulin
production. In high cell density cultivation, 153 mg ecotin-proinsulin per liter broth was
produced. Downstream processing was done in one step using a newly established affinity
purification method. Since ecotin is a trypsin inhibitor, trypsinogen or inactive trypsin variants
immobilized to a column can serve as a highly selective affinity material. Native human
proinsulin was purified to homogeneity, estimated by ELISA and characterized by mass
spectrometry. To evaluate the effect of proteolysis in the periplasm of E. coli, the amount of
ecotin-proinsulin was determined in a wild-type strain, E. coli BL21(DE3)Gold, and in a strain
deficient in several periplasmic protease, E. coli SF120. At the shake flask level, the specific
yield of ecotin-proinsulin was 3-4 fold higher in E. coli SF120 than in a wild-type strain, E. coli
BL21(DE3)Gold. In summary, the ecotin fusion protein system is a novel and useful tool to
efficiently produce recombinant proteins in the bacterial periplasm.
 

• نوع عمل المنشور: PhD