Since many basic principles of genetics have been discovered using E. coli as
a model organism, today its genetics is well characterized, and many
details about the regulation of gene expression are described. This
knowledge made it possible to use its promoters in a cassette fashion,
unaffected by the surrounding nucleotide context. Additionally, a large
number of cloning vectors are available, which provides a long-range
repertory of components out of which the most appropriate can be
chosen. The productivity of the expression system is influenced by many
parameters, and besides the basic choice of an available expression
system, these parameters can be tuned by the insertion or deletion of
genetic modules.
In all organisms gene expression is regulated on various levels, at
which mRNA stability provides a key control device. As several RNases
are involved in mRNA degradation, including endonucleases and
3'exonucleases, many sequences that act as stabilizers have been
identified. Some of them only work under specific conditions, others
stabilize mRNAs of miscellanous sources under normal conditions and at
high growth rates. One well characterized example of a such a
stabilizing element is a 21-bp fragment of the φ10 promoter of bacteriophage T7 which can potentially form an 8-bp stem loop [28,29].
In our assays, the insertion of this stem loop quadrupled the amount of
GFP that is produced in 24 h and fortunately had no measurable effect
on plasmid maintenance. This advantage is preserved even if different
combinations of other modules are added, and it exhibits the most
explicit improvement achieved by the modulation that were carried out
in this study. Though we have not tested the actual stability of the
mRNA produced by our system, one can presume that this characteristic
has been improved considerably and marks a significant step.
Besides the stem-loop insertion into the rhaP region we
have tried to improve expression by modification of the -10 promotor
region. The sequence 'TAGACT' in pJOE4056.2 has been mutated into
'TATAAT', and the resulting plasmid pJOE5115.1 has been examined in an eGFP expression
assay as described herein. Unfortunately, this modulation resulted in a
elevated basal level which under non-induced conditions attained the
output of induced pJOE4056.2 cells, though it evidently fell behind in
expression compared to pJOE5058.1 (data not shown). Hence, this
approach reveals that additional alterations of the already efficient rhaPBAD expression
vectors are crucial, not only because they possibly do not advance
yields, but also may disturb a well-balanced system by making it less
regulable.
Adjacent to plasmid inherent regulation mechanisms, the current
metabolic status of the cells has an exceeding impact on the production
of recombinant proteins as well. For example, some promotors act mainly
in the exponential phase of cell growth, whereas others are activated
primarily in steady state cultures, and in addition the intracellular
availability of nutrients further affects the activity of a given
promotor. Generally, promotors assigned to operons associated with
carbohydrate catabolism are controlled by CRP via the intracellular
cAMP levels. As D-glucose lowers cAMP levels and thus inactivates
binding of CRP, but usually is added as a carbon source in minimal
medium used for fermentations, it would be of interest to make such a
promotor less sensitive towards low cAMP levels. The approach tested in
this study was to enhance the affinity of the CRP-binding site.
Unfortunately, the adjustment of the DNA sequence according to the
consensus sequence had no measurable effects. The mutated plasmid still
was inhibited by glucose in the same extend as the unmodified plasmid.
Presumedly, the original binding site already shows a good affinity to
CRP and could not be improved by an adjustment with the consensus
sequence. On the other hand, the plasmid pWA64.1 with the one base pair
deletion in the CRP binding site was clearly inferior. It still was
susceptible to glucose addition but obtained only about 10% of the
yield compared to the original plasmid. These results precisely show,
that it is difficult to elevate protein production concomitantly with a
low basal level and a tight regulation.
Another way to elevate protein production is to increase the
corresponding gene copies in the cells. This was acchieved in this
study by deleting the rop gene which is involved in
controlling plasmid copy numbers. The deletion increased the content of
plasmid DNA in the cells at least threefold as expected but there was
no concomitant increase of eGFP production. This is not surprising
since the eGFP production in cells with pWA21 already amounts to about
20% of the total protein and even with the strongest expression systems
there is a limit at about 30% of the total protein whatever gene is
used. To see the high-copy plasmid effect a reporter gene would be
needed which is expressed at a lower level. This might also answer the
question if there is a sufficient supply of the activator RhaS for
high-copy plasmids from the single rhaS gene copy on the chromosome. When the rhaRS genes were introduced on a compatible plasmid with a moderate copy number (pBBR1MCSII, [30])
in cells containing already pWA21, pWA23 or pWA28, no significant
increase in eGFP production was observed upon induction with rhamnose
(data not shown).
Previous results reported by Wilms et al.[31] showed that the introduction of the cer site of ColE1 to plasmids reduces the appearance of multimers drastically. This is especially true for recA-proficient strains like W3110, whereas the multimerization is inhibited in recA-deficient strains like JM109. Actually, the insertion of cer led
to a tremendous stabilization. Over 90% of the cells were still
carrying the plasmid, while less than 50% of the cells kept the
plasmids without the cer-site when ampicillin was absent. Accidentally, in these cases two cer-sites
in tandem had been ligated into the vectors. The removal of one of them
led to an increased eGFP expression while the plasmid stability was
only negligibly influenced. The accompanying decrease in plasmid DNA
apparently did not adversely affect the protein yields, probably
because it is the duplication of cer that causes problems to the cell division which are solved by removing one. Furthermore, one cer site sufficiently ensured the plasmid maintenance.
Additionally, the bacterial plasmid addiction system ccdAB was provided to some of the vectors in order to test its influence on the stability. Unexpectedly, ccdAB did
not deliver any improvement, as it did not stabilize the plasmid but in
contrast increased plasmid loss at 37°C. As the toxin CcdB kills cells
that have lost their plasmids and therefore can not produce the
unstable antidote CcdA any more, apparently cells still carrying ccdAB plasmids
are at a serious disadvantage in cell growth. It amounts to a situation
where few cells that have successfully eliminated the plasmids start to
overgrow cells that still express ccdAB. Presumedly, the
continous production of the toxin becomes such a great metabolic burden
to the cells, that the evasion by plasmid loss outweighs the initial
toxication. Thus it appears that the insertion of the addiction module
in this case represents a set-back.
Answers to all your Biology Questions
