- Open Access
Antibiotic-free segregational plasmid stabilization in Escherichia coli owing to the knockout of triosephosphate isomerase (tpiA)
© Velur Selvamani et al.; licensee BioMed Central Ltd. 2014
- Received: 23 January 2014
- Accepted: 10 April 2014
- Published: 21 April 2014
Segregational stability of plasmids is of major concern for recombinant bacterial production strains. One of the best strategies to counteract plasmid loss is the use of auxotrophic mutants which are complemented with the lacking gene along with the product-relevant ones. However, these knockout mutants often show unwanted growth in complex standard media or no growth at all under uncomplemented conditions. This led to the choice of a gene for knockout that only connects two essential arms of an essential metabolic pathway – the glycolysis.
Triosephosphate isomerase was chosen because its knockout will have a tremendous effect on growth on glucose as well as on glycerol. On glycerol the effect is almost absolute whereas on glucose growth is still possible, but with considerably lower rate than usual. This feature is essential because it may render cloning easier. This enzymatic activity was successfully tested as an alternative to antibiotic-based plasmid selection. Expression of a model recombinant β-glucanase in continuous cultivation was possible with stable maintenance of the plasmid. In addition, the complementation of tpiA knockout strains by the corresponding plasmids and their growth characteristics were tested on a series of complex and synthetic media. The accumulation of methylglyoxal during the growth of tpiA-deficient strains was shown to be a possible cause for the growth disadvantage of these strains in comparison to the parent strain for the Keio Collection strain or the complemented knock-out strain.
Through the use of this new auxotrophic complementation system, antibiotic-free cloning and selection of recombinant plasmid were possible. Continuous cultivation and recombinant protein expression with high segregational stability over an extended time period was also demonstrated.
- Segregational stability
- Triosephosphate isomerase
- Keio collection
- Knockout strain
- Continuous cultivation
In 2009 almost one third (45 out of 151) of commercial biopharmaceutical recombinant proteins licensed by FDA and EMEA were produced by cultivation of Escherichia coli. This meant that nearly 30% of high-value recombinant proteins were produced by means of a plasmid-based expression system. Various other commercial recombinant proteins (e.g. enzymes) or non-protein products (e.g. amino acids) were also obtained with strains of E. coli harboring plasmids. Searching the Science Citation Index database for “coli recombinant protein” would lead to more than 26,000 results. These facts show the high importance of E. coli as an expression system for commercial as well as scientific use.
Expression systems based on E. coli depend in almost all cases on the presence of at least one plasmid, which has to be segregated into dividing cells during growth. Insufficient segregational plasmid stability would render plasmid-based expression systems useless. Therefore, a variety of methods have been developed in order to achieve plasmid stability . The commonly used method consists in adding antibiotics into the cultivation medium and placing genes for antibiotic resistance on the plasmid carrying the target gene. This strategy is widely used in research, since most often only low working volumes have to be applied. In industrial biotechnology, the addition of antibiotics can be generally excluded – not only with respect to economic reasons. The elimination of antibiotics from media and waste streams maybe required during downstream processing. Therefore, other methods for plasmid stabilization are in high demand.
A simple strategy in trying to keep a high segregational plasmid stability without adding antibiotics consists in using plasmids of high copy number and expecting that the statistical distribution of plasmids during cell division may always yield cells with at least some plasmids. This works fine as long as the average plasmid copy number is homogeneous . As soon as cells with very low plasmid copy numbers show up preferentially during high growth rates, cells might have not enough time to synthesize plasmids in high copy numbers. Caused by the additional metabolic burden, plasmids may be lost completely, since plasmid-free cells would gain a growth advantage in comparison to plasmid-bearing cells. This would have a drastic influence especially on fed-batch- and continuous cultivation processes.
High metabolic burden, basal transcription levels and possible toxic effects of recombinant proteins ask for other methods for stabilizing plasmids. In some cases, changes in the cultivation strategies, mainly by decreasing the specific growth rate, may lead to sufficient plasmid stability. This can be achieved by reducing the cultivation temperature or changing the carbon source of the medium. However, most methods are based on adding some stabilizing elements by genetic engineering.
Some lessons may be learned from nature’s strategies to ensure plasmid segregation. Thus, single copy plasmids often show sophisticated stabilizing mechanisms like that based on the par-system which leads to a controlled distribution of plasmids, similar to the highly organized and controlled chromosome distribution in higher organisms. The par-system consists of at least two protein-coding genes and one special site on the plasmid for controlled distribution in the dividing cell .
Other systems lead to the post-segregational killing of plasmid-free cells and need the genetic information of a toxin and its corresponding antidote, with the antidote on the target plasmid . Such a combination was recently applied for the development of a Streptomyces based protein expression system .
In all these cases, the plasmids may become quite large, leading to a higher metabolic burden. A larger size may also lead to another problem - decreased structural plasmid stability due to stochastic mutations, which would lead to a reduction in productivity, as well. In this regard, the ccdB/ccdA poison-antidote system, modified as a separate-component-stabilization system provides an alternative using very small genetic constructs to provide efficient antibiotic-free maintenance of plasmid in E. coli.
Many studies on DNA vaccine production have particularly been interested in alternative plasmid selection mechanisms, due to the need to avoid all kinds of resistance genes or proteins in therapeutic DNA according to safety standards. An RNA based method, using constitutive expression of sacB as a counter-selectable marker during growth on sucrose was reported to be able to bring about antibiotic-free selection and highly productive fermentation while not being restricted to ColE1 vectors .
Another perhaps more elegant way to stabilize the propagation of plasmids, is to destroy the function of an essential gene on the chromosome and place this gene on the plasmid carrying the target genes. This approach requires a strategy to generate competent cells of the now auxotrophic strain. One of the first methods for commercial use was the valS-system . The wild-type gene for valyl-tRNA synthetase (valS) of the host carries a temperature-sensitive mutation, whereas the gene without mutation is placed on a plasmid. For transformation the host is grown at 30°C. During cultivations at 37°C the mutated synthetase on the chromosome loses its function and only cells carrying the valS-harboring plasmid can survive and grow. However, the valS gene is still around on the chromosome and the selection pressure favors a recombination of the mutated valS on the chromosome and the wild type valS on the plasmid. Such a recombination produces revertants with no selection pressure by valS and leads to plasmid instability. One way to reduce the probability of recombination is the complete knockout of the chromosomal gene. In the case of valS this would not allow to obtain viable cells for transformation and, therefore, would not be practicable.
The Operator-Repressor-Titration (ORT) strategy is based on negative regulation of an essential chromosomal gene by an operator sequence allowing the binding of a constitutively expressed repressor protein. In order to allow expression of the essential gene and survive, the cell has to titrate the repressor molecules against a similar operator sequence that may be present in multiple copies on the target plasmid to be maintained . The essential function complementation strategy could also be effective at the RNA level. Recently, an amber nonsense mutation introduced into the essential thyA gene in the chromosome causing thymidine auxotrophy, was overcome by recombinant plasmids carrying a suppressor tRNA, which allowed antibiotic-free plasmid selection and also recombinant luciferase reporter expression in eukaryotic tissues and in tumour cells . The fabI - triclosan system, which is an essential gene- gene product inhibitor combination, was a completely different alternative plasmid selection concept. However, the risks associated with the biocide triclosan, the requirement of the selection agent due to an addictive effect, the need to induce and over-express the essential gene marker and plasmid instability in absence of selection were also noted . Another example of antibiotic-free plasmid selection is the strategy involving the essential infA gene coding for translation initiation factor .
If an auxotrophy can be overcome by supplements in the medium, competent cells can be prepared and transformed using such a supplemented medium. This may be achieved by knockout of an essential gene for e.g. the synthesis of an amino acid like glycine . In this case, the glyA gene is knocked out in E. coli M15 leading to a auxotrophic strain which can be cultivated on glycine containing media. The glyA gene is cloned on an expression vector under the control of a constitutive weak promoter. This system has one disadvantage in that glycine-containing media may lead to plasmid-free cells, and many complex industrial media contain glycine.
One solution of this problem would consist in the construction of a knockout strain, the chromosomal knockout of which would still allow at least some growth on complex media. The specific growth rate for a strain harbouring the plasmid containing the knockout gene could generate a selection advantage high enough to keep the plasmid in the cell population.
During the search for a gene which would be appropriate for the application of the auxotrophy-based strategy for plasmid stabilization we have focused on tpiA, the gene for triosephosphate isomerase, a central enzyme in the glycolysis pathway. One of recent great scientific achievements in E. coli research was the establishment of the so-called Keio collection of knockout strains . Using the Keio tpiA knockout strain we present the construction of tpiA-harbouring plasmids leading to E. coli strains bearing plasmids of high segregational stability.
Screening for an appropriate knockout gene
As a consequence of biodiesel business, glycerol has evolved as an interesting carbon source for fermentation processes. A closer look on the glycerol metabolism of microorganisms, therefore, is of particular importance. Escherichia coli can utilize glycerol as the sole carbon and energy source. After having been imported through the cytoplasmic membrane by a facilitator protein (GlpF), glycerol can be metabolised on two alternative pathways depending on the growth conditions. One consists of a phosphorylation step by a glycerol kinase (GlpK) to yield L-glycerol-3-phosphate followed by an oxidation step due to the appropriate dehydrogenase (GlpD under aerobic and GlpABC under anoxic conditions) leading to dihydroxyacetone phosphate (DHAP). The alternative pathway consists of an oxidation step by glycerol dehydrogenase (GldA) to yield dihydroxyacetone (DHA) followed by phosphorylation by DHA kinase (DhaK) to give DHAP as well. The kinase-dehydrogenase (GlpK-GlpD/GlpABC) route is the preferred one in E. coli. Accordingly, different genes and their products come into play as there are: glpF glpK glpD glpABC gldA and dhaK. In addition, the resulting intermediate metabolite DHAP must be fed into the general glycolytic pathway through isomerisation by triosephosphate isomerase (TpiA) as glyceraldehyde-3-phosphate (GA3P). In the absence of this enzyme, DHAP is converted to the toxic compound methylglyoxal .
The knockout of glpF, the gene for the glycerol uptake facilitator protein, has no significant effect on growth. The normal permeability of both plasma and outer membrane may be sufficient for glycerol uptake . When gldA, the glycerol dehydrogenase, is knocked out growth on LB medium is not disturbed, either. Compared with the growth characteristics of the E. coli MG1655 wild type strain, these knockouts seem to be even slightly beneficial. Keio strain ΔglpK, in which the glycerol kinase is knocked out, does not grow at all, although its effect might be partly circumvented . The second strain showing useful characteristics is that in which the triosephosphate isomerase (tpiA) has been knocked out. In consequence, this gene is considered even more appropriate than glpK for developing a strategy for plasmid stabilization based on the auxotrophy principle since it represents an enzymatic activity essential also for the glycolysis pathway and not only for glycerol metabolism.
Cloning of tpiA
For the first construct of the tpiA gene for complementation of tpiA knockout, the wild type promoter- and terminator sequences were chosen. Thus, chromosomal DNA of E. coli MG1655 was isolated and the tpiA region was amplified by PCR using 5′ · taagctggcgctatctgaatcg · 3′ and 5′ · gatggtacggcagagtgataac · 3′ as forward and reverse primer, respectively. The amplified tpiA fragment started 150 bp upstream of the ATG start codon and ended 172 bp downstream of the TAA stop codon of the tpiA structural gene. The upstream region started in the yiiQ gene, coding for an unknown conserved protein, including the predicted tpiA promoter . The downstream region reached into the cdh gene including the predicted Rho-independent tpiA terminator . The resulting PCR fragment of 1090 bp length was cloned into plasmid pJET (Fermentas) and subsequently verified by sequencing. This construct called pJET-tpiA was transformed into the auxotrophic host strain E. coli JW3890-2, CGSC#: 10805, Keio Collection (ΔtpiA).
All Keio knockout strains show a special structure in the genome. The knockout locations are found directly behind the ATG start codon followed by the gene for the kanamycin resistance flanked by FRT-sites. The knockouts stop 18 bases in front of the stop codon of the knocked out gene.
In case of the cloned fragment of the tpiA region, there was still an overlap of 153 bp upstream and 195 bp downstream of the structural target gene. This arrangement still involved the risk that a recombination between the knockout location on the chromosome and the tpiA region on a plasmid could occur.
Application of the tpiA knockout in a recombinant production system
Previous work described the extracellular production of a hybrid bacterial β-glucanase using plasmid p582 having a size of 6 kb and a pUC19 origin of replication . A strong constitutive synthetic promoter (CP7) controlled the expression of the β-glucanase gene . It was followed by a Bacillus-derived signal sequence for periplasmic targeting. At the carboxy terminus a hexahistidine tag was placed for facilitating protein capture and detection. The other important gene on the plasmid p582 was kil coding for the bacteriocin release protein of ColE1 the expression of which would initiate the release of periplasmic proteins into the extracellular space. This gene was under the control of the weak stationary phase promoter of the gene fic. The plasmid contained antibiotic resistance genes against both ampicillin and kanamycin.
As expected, the original Keio tpiA knockout strain (JW3890-2) did not grow in the presence of glycerol as the only carbon source (SGA-Kan). If a complex carbon source was present (LB-Kan), however, some growth was observed. Only the strains complemented with plasmids pFC1 (JW3890-2-pFC1) and pFC4 (JW3890-2-pFC4) showed normal growth behaviour in the presence of glycerol as the only carbon source (SGA-Kan). The strain JW3890-2-pFC1 showed a shorter lag phase while the strain with construct pFC4 had a relatively longer lag phase. In fact, this phenomenon could be reproduced over multiple trials. Both strains reached equal maximum optical densities (600 nm) of about 8.0. The media had been supplemented with kanamycin in all three cases. Although, both clones were able to complement the auxotrophy of the tpiA Keio knockout strain, their product expression capabilities differed considerably. Both constructs were studied under batch cultivation conditions in appropriate bioreactors.
Batch fermentation with strain JW3890-2-pFC1
Batch fermentation with strain JW3890-2-pFC4
Since the plasmid pFC4 showed higher extracellular enzyme activity compared to pFC1, it was retained for further studies.
Test of plasmid stabilization with tpiA in continuous culture
Before starting the chemostat study, the plasmid stability was tested in shake flasks by repeated subculture for over 70 generations and was found to be adequate (data not shown). As a decisive test, a chemostat was set up consisting of a small reactor of 2 L total volume in which a working volume of 1 L was maintained. The SGA minimal medium based on glycerol as the only C-source was used. The feed solution contained 20 g L−1 glycerol. No antibiotics were added to the media.
Construction of a tpiA vector with altered downstream overlap
Primers for cloning an artificial terminator behind the tpiA gene on pJET-tpiA
tpiA Term. lacUV
tpiA Term. direct
Growth of the tpiA knockout strain on different media in the absence or presence of the complementing plasmid pJET-tpiA-Tart
The growth rates of the parent strain for the Keio Collection and the plasmid carrying strain were quite similar. The plasmid supplemented strain was able to grow in synthetic media, whereas the plasmid-free tpiA deletion strain JW3890-2 was not able to grow on synthetic media.
Analysis of plasmid stability
Because of the fact that the plasmid free tpiA deletion strain could not grow on synthetic media, the experiments for antibiotic free selection were performed in SGA medium. For this purpose shake-flask cultivations were carried out in a repetitive fashion by reintroducing 100 μL of a final cultivation broth into 50 mL of fresh medium in shake flasks of 500 mL for a cultivation period of 1 day. The same procedure was applied to the parent strain for the Keio Collection as well as to the mutant ΔtpiA Keio strain in the presence of either pJET-tpiA or pJET-tpiA-Tart as complementing plasmid. For each series 13 repeated cultures were followed over a total period of 17 d.
For a cultivation time of 1 d, the parent strain for the Keio Collection strain grew to an optical density of 8.1 ± 0.25 whereas the pJET-tpiA complemented mutant strain reached 12.1 ± 0.53, and that complemented with pJET-tpiA-Tart grew to an OD600 of 11.3 ± 0.57. The data are given with their standard deviations collected over 10 samples. The segregational plasmid stability was tested at the same time and was found to be stable at 98 to 100%.
In addition, the plasmid stability was tested for cultivations in the presence of different media in the absence of ampicillin. Thus, the plasmid maintenance was found to be absolute for the growth of the Keio mutant strain complemented with pJET-tpiA-Tart in the case of the media HSG, LB, LB supplemented with glucose, LB supplemented with glycerol, SGA, SGA supplemented with glucose, SOC as well as TB.
Because of these data it can be expected, that the TpiA deficiency can be used as a selection marker for plasmid stabilization.
Accumulation of methylglyoxal may lead to a toxic effect
The reason why the TpiA deficient strain stopped growing at rather low biomass concentrations compared with the wild type or the plasmid-supplemented strain during cultivation on different media as shown in Figure 9 could be a consequence of DHAP accumulating in the cells. This could in turn lead to methylglyoxal by enzymatic conversion of DHAP by MgsA (methylglyoxal synthase), which should be toxic for the cells. Some experiments were performed in order to test this assumption.
The data show that methylglyoxal accumulated during the cultivation of the tpiA deficient strain and that it was consumed during the second cultivation on the spent medium, and fell from 1.1 mM to 0.4 mM. A methylglyoxal concentration of 0.6 mM is assumed to be toxic for E. coli. Thus, the toxicity of methylglyoxal generated from DHAP may force an auxotrophic effect even in the presence of glucose as the main carbon source.
Antibiotic free cloning
To test the possibility of antibiotic free cloning with tpiA deficient strains, SGA agar plates were created by adding 15 g L−1 agar agar into SGA medium and pouring into petri dishes. The tpiA deficient strain was exposed on these plates and showed no growth. After transforming the plasmid pJET-tpiA-Tart in a chemical competent JW3890-2 strain, colonies appeared in the antibiotic free SGA plates containing the correct plasmid.
The constructs pFC1 and pFC4 contain the auxotrophy-complementing tpiA gene in opposite orientations and some consequences of this difference could be observed from Figures 4 and 5. With respect to growth, it was seen over multiple trials that the strain JW3890-2-pFC1 grew with a slightly shorter lag phase and achieved higher maximum biomass concentrations than JW3890-2-pFC4. Whereas, with respect to the extracellular recombinant β-glucanase activity, the construct pFC4 clearly returned higher volumetric activity than pFC1 and this difference could also be reproduced over multiple trials. It is common for expression levels to vary between different clones and in this case one possible scenario is that in the case of pFC1, the tpiA gene could be read into the downstream product gene bgl, which has the same orientation.
Other experiments conducted to check the secretion achieved with these constructs showed that both the clones pFC1 and pFC4 were poor in secretion into extracellular space. The analysis of samples from the stationary phase showed that, the activities in the periplasmic fraction were higher than those in the extracellular fraction. This was in clear contrast to the control strain E. coli JM109-p582, where the reverse was to be seen and the equilibrium was clearly in favour of the extracellular fraction (data not shown). In general, the control strain showed up to 10 times the activity achieved with the auxotrophic strains and thus represented an ideal case. However, since the auxotrophic system offers antibiotic-free cultivation, it deserves further development. The question of what could be causing the suboptimal activity of the BRP in the Keio strains remains open at this moment. The general low efficiency of these auxotrophic hosts could be one possible explanation. In this regard, tpiA gene knockouts from the control strain are being developed which should then be able to serve as efficient hosts for the expression and secretion of the recombinant product.
As mentioned under the section Results, the target gene deletion in the Keio strains is restricted to the region between the start codon and the last six codons. The gene deletion in the control strain would be wider and more extensive than in the Keio strain and is thus designed to overcome the risk of recombination of the plasmid-borne tpiA copy with the host genome and loss of auxotrophy. Another potential route to better productivity from the clones would be the expression of tpiA at the minimum required level. This would avoid unnecessarily high rates of transcription and stress in the cells. Therefore, the native P1/P2 promoters for tpiA are being replaced with artificial weak promoters from a synthetic promoter library .
Using the Keio knockout strains, a flux from DHAP into the xylulose pathway is not possible due to the deletion of rhamnose metabolism genes . A bypass of DHAP to lactate and pyruvate through methylglyoxal is possible through the glyoxalase I-II pathway . This should also explain the limited growth of cells without the complementation plasmid in complex medium. However, when glycerol is the sole carbon and energy source, the toxicity of methylglyoxal should render this route unfavourable. Moreover, this would only be a partial metabolism and not an energy-efficient pathway to give a growth rate advantage.
Antibiotic-free recombinant plasmid selection and stabilization in E. coli based on the auxotrophy complementation of the activity of TpiA has been demonstrated. From our observations, we have been able to achieve antibiotic-free cloning, selection, expression of a model recombinant product and long-term stability of the plasmid in continuous culture. The growth advantage shown by the plasmid-complemented strain even under non-selective conditions makes this system particularly attractive for large-scale industrial processes.
List of Escherichia coli strains used
E. coli strain
BW25113 (parent strain for the Keio Collection)
Δ(araD-araB)567, ΔlacZ4787(::rrnB-3), lambda−, rph-1, Δ(rhaD-rhaB)568, hsdR514
CGSC, Yale University
Δ(araD-araB)567, ΔlacZ4787(::rrnB-3), lambda−, rph-1, Δ(rhaD-rhaB)568, hsdR514 ΔtpiA778::kan
CGSC, Yale University
F−, lambda−, rph-1
endA1, recA1, gyrA96, thi, hsdR17 (rk–, mk+), relA1, supE44, Δ(lac-proAB), [F’ traD36, proAB, laqIqZΔM15]
All the modifications in the knockout strain other than the tpiA deletion are already present on the parent strain E. coli BW25113. The modifications in the arabinose, lactose or rhamnose metabolic genes do not have any consequences for the current experiments reported with this strain.
List of plasmids used
rep (pUC19), bla (AmpR), npt (KanR), PCP7-bgl, Pfic-kil, T7 terminator, multiple cloning site
rep (pMB1), bla (AmpR), eco47IR, PlacUV5, T7 promoter, multiple cloning site, insertion site
rep (pMB1), bla (AmpR), eco47IR, PlacUV5, T7 promoter, residual multiple cloning site, tpiA expression cassette
rep (pUC19), bla (AmpR), npt (KanR), PCP7-bgl, Pfic-kil, T7 terminator, tpiA expression cassette in forward orientation
rep (pUC19), bla (AmpR), npt (KanR), PCP7-bgl, Pfic-kil, T7 terminator, tpiA expression cassette in reverse orientation
rep (pMB1), bla (AmpR), eco47IR, T7 promoter, residual multiple cloning site, tpiA expression cassette, artificial terminator
Composition of the SGA medium
Solution concentration/g L−1
Medium concentration/g L−1
Micronutrient solution 500×
FeCl3 · 6H20
ZnSO4 · 7H2O
MnSO4 · H20
CoSO4 · 7H20
Na2MoO4 · 2H2O
Citric acid hydrate
Salt solution 10×
Citric acid hydrate
Ammonium sulphate solution
Composition of the HSG medium
Medium concentration/g L−1
MgSO4 · H20
SGA media are fully synthetic ones. All solutions were autoclaved separately and combined as explained in Table 4. Depending on the preferred carbon source either glycerol or glucose was added. For batch fermentations with SGA, water and carbon source were autoclaved in situ, while the other components were sterilized separately and added to the reactor. When the medium was complete, the pH was found to be around 6.5. This was then corrected to a pH of 7 through addition of sterile NaOH before inoculation. For the strain JM109, sterile-filtered thiamine solution was added at a final concentration of 0.01 g L−1 due to its auxotrophy as given in the genotype.
The HSG medium is a buffered complex rich medium with glycerol as carbon source. The K2HPO4/KH2PO4 solution was autoclaved separately. The pH was adjusted to 7.4. Its composition is explained in Table 5.
Other media like LB (Lysogeny Broth), TB (Terrific Broth) and SOC (Super Optimal broth with Catabolite repression) were prepared according to standard laboratory protocols.
Shake flask cultivation
All cultivations were performed in a working volume of 50 mL in shake flasks of a total volume of 500 mL equipped with baffles (Schott, Germany). The rotary shaker LS-X (Kühner, Switzerland) had a rotating frequency of 120 min−1 and an eccentricity (diameter) of 50 mm. The cultivation was carried out at a temperature of 30°C.
The bioreactor cultivations were carried out in a 2 L in-house fermenter with a working volume of 1 L. The reactor had a height of 280 mm and a diameter of 94.4 mm. The impeller diameter measured 46 mm and the impeller blades measured 12 mm on each side. A total of 3 impellers with 6 blades each ensured good mixing for homogeneous conditions. Moreover, the vessel contained 4 baffles with a height of 260 mm and width of 8 mm each.
Water and glycerol for the SGA medium were sterilized in situ along with the fermenter, while the remaining components were sterilized separately and added under sterile conditions. After adjusting initial pH and saturating the medium with oxygen, the fermenter was inoculated from an overnight shake flask culture. Agitation was carried out at a stirrer speed of 800 rpm and aeration was achieved with an air space velocity of 1 vvm. The process temperature was maintained at 37°C for the control strain E. coli JM109-p582 and 30°C for the auxotrophic strains, since the latter showed inclusion body formation due to stress (data not shown) when grown at 37°C. The pH was automatically maintained at 7.0 using 10% orthophosphoric acid and 2 M sodium hydroxide as correction solutions. The process was monitored by a ADAM-4060 relay output module (Advantech Ltd., USA) and run by using DASYLAB 6.0 software (National Instruments Service GmbH, Germany).
The chemostat process was started as a batch fermentation, and upon reaching the end of exponential phase, the feed and outlet pumps were switched on. The desired space velocity (dilution rate) was set by the flow rate of the medium. The feed medium supplied from a 40 L reservoir had the same composition as the medium for batch but with 20 g L−1 glycerol. The culture volume in the reactor was maintained at a constant level by operating the inlet pump at the exact flow rate required whereas the outlet pump was operated at twice the flow rate, with the suction pipe entrance positioned at a fixed depth corresponding to a liquid working volume of 1 L. Stably maintained levels of dissolved oxygen (DO) in the reactor medium, CO2 content in the exit air and optical density of the culture indicated a steady state. Any change in the feed flow rate was followed by allowing the system to stabilize by passing at least three to four culture volumes through the reactor before a new steady state was assumed.
Measurement of growth
Prior to each cultivation experiment a single colony was transferred from an agar plate into the appropriate medium and cultivated overnight. The pre-culture media were exactly the same as the media used in the subsequent cultivation experiments.
For the main cultivation the initial optical density at a wavelength of 600 nm (OD 600) was adjusted to 0.1. Over the cultivation process samples were taken and the optical density was measured against the cultivation medium. If required, samples were diluted with fresh cultivation medium.
Determination of the specific growth rate
The specific growth rate (μ) was determined by means of the standard method .
The plasmid stability was determined by plating out appropriately diluted samples on normal LB agar plates prepared without antibiotics and incubating at 37°C for at least 15 h. The colonies were then transferred onto an LB agar plate containing 200 mg L−1 ampicillin followed by incubation at 37°C for at least 15 h. The ratio of the number of colonies counted on plates prepared in the presence of antibiotics to that observed in the absence of antibiotics was interpreted as plasmid stability.
For measurements of the plasmid content in the samples, a fixed volume of culture broth (2 mL) was used for plasmid isolation by the Wizard Plus SV DNA Purification System (Promega, USA). The elution of the plasmid from the column was carried out into a constant volume (50 μL) of nuclease-free water (Promega, USA). The plasmid DNA concentration was measured by means of the Nanodrop spectrophotometer (Peqlab Biotechnologie GmbH, Germany), and the value was normalised to the optical density of the culture sample.
Determination of methylglyoxal concentration
Methylglyoxal was quantitatively determined according to the method described by Cooper .
Determination of glycerol concentration
Culture samples were centrifuged to separate the cells. Glycerol concentration in the supernatant was measured by HPLC using a Nucleogel sugar 810H cation exchange column (Macherey-Nagel GmbH, Germany).
Assay for β-Glucanase activity
The recombinant protein expression was analysed with an activity test for endo-1,3-1,4-β-glucan-glucanohydrolases. Substrates used were either lichenan (Roth, Germany) or barley β-glucan (Megazyme, Ireland) both of which have very similar properties. The substrate contains 70% β-1,4- and 30% β-1,3-bonds.
The substrate solution in 40 mM sodium acetate buffer  was incubated with samples containing the enzyme to be assayed for 20 min. The reducing groups released were analysed by the method based on dinitrosalicylic acid by following the absorbance of light at a wavelength of 530 nm (OD 530). This was compared to the maximum value expected for the particular substrate at complete conversion, which can be obtained practically by applying an excess enzyme activity and incubating for 2 h. The final enzyme activity was obtained as a volumetric activity taking into account the particular enzyme kinetics . One unit is defined as the enzyme activity that releases an equivalent amount of 1 μmol glucose residues per min at a temperature of 50°C and a pH 5.6.
All cloning experiments were carried out according to standard methods .
Potential clones were verified by sequencing the cloned region using corresponding sequencing primers. Cycle-sequencing was performed at the Centre for Biotechnology (CeBiTec) (Bielefeld University, Germany). A BigDye® terminator version 3.1 chemistry was used for the PCR in a GeneAmp® PCR System 9700 (both Applied Biosystems, Germany). The fragments of various lengths were then separated by capillary electrophoresis and the fluorescence signal of each base converted to their corresponding digital data by a 96-well 3730xl DNA Analyzer (Applied Biosystems, Germany).
The authors would like to thank Eberhard “Ebson” Wünsch for technical assistance in the lab. RSVS thanks the Deutscher Akademischer Austauschdienst (DAAD) for financial support through providing a PhD scholarship.
We acknowledge support of the publication fee by Deutsche Forschungsgemeinschaft and the Open Access Publication Funds of Bielefeld University.
- Ferrer-Miralles N, Domingo-Espín J, Corchero JL, Vázquez E, Villaverde A: Microbial factories for recombinant pharmaceuticals. Microb Cell Fact. 2009, 8: 17-View ArticleGoogle Scholar
- Friehs K: Plasmid copy number and plasmid stability. New Trends and Developments in Biochemical Engineering, Volume 86. Edited by: Scheper T. 2004, 47-82. Berlin Heidelberg: Springer-VerlagView ArticleGoogle Scholar
- Summers D: The Biology of Plasmids. 1996, Oxford: Blackwell-ScienceView ArticleGoogle Scholar
- Easter CL, Sobecky PA, Helinski DR: Contribution of different segments of the par region to stable maintenance of the broad-host-range plasmid RK2. J Bacteriol. 1997, 179: 6472-6479.Google Scholar
- Zielenkiewicz U, Ceglowski P: Mechanisms of plasmid stable maintenance with special focus on plasmid addiction systems. Acta Biochim Pol. 2001, 48: 1003-1023.Google Scholar
- Sevillano L, Diaz M, Santamaria RI: Stable expression plasmids for Streptomyces based on a toxin-antitoxin system. Microb Cell Fact. 2013, 12: 39-View ArticleGoogle Scholar
- Szpirer CY, Milinkovitch MC: Separate-component-stabilization system for protein and DNA production without the use of antibiotics. Biotechniques. 2005, 38: 775-781.View ArticleGoogle Scholar
- Luke J, Carnes AE, Hodgson CP, Williams JA: Improved antibiotic-free DNA vaccine vectors utilizing a novel RNA based plasmid selection system. Vaccine. 2009, 27: 6454-6459.View ArticleGoogle Scholar
- Nilsson J, Skogman SG: Stabilization of Escherichia coli tryptophan–production vectors in continuous cultures: a comparison of three different systems. Nat Biotechnol. 1986, 4: 901-903. 10.1038/nbt1086-901.View ArticleGoogle Scholar
- Cranenburgh RM, Lewis KS, Hanak JAJ: Effect of plasmid copy number and lac operator sequence on antibiotic-free plasmid selection by operator-repressor titration in Escherichia coli. J Mol Microbiol Biotechnol. 2004, 7: 197-203.View ArticleGoogle Scholar
- Marie C, Vandermeulen G, Quiviger M, Richard M, Préat V, Scherman D: pFARs, Plasmids free of antibiotic resistance markers, display high-level transgene expression in muscle, skin and tumour cells. J Gene Med. 2010, 12: 323-332. doi:10.1002/jgm.1441View ArticleGoogle Scholar
- Goh S, Good L: Plasmid selection in Escherichia coli using an endogenous essential gene marker. BMC Biotechnol. 2008, 8: 61-doi:10.1186/1472-6750-8-61, View ArticleGoogle Scholar
- Hägg P, de Pohl JW, Abdulkarim F, Isaksson LA: A host/plasmid system that is not dependent on antibiotics and antibiotic resistance genes for stable plasmid maintenance in Escherichia coli. J Biotechnol. 2004, 111: 17-30.View ArticleGoogle Scholar
- Vidal L, Pinsach J, Striedner G, Caminal G, Ferrer P: Development of an antibiotic-free plasmid selection system based on glycine auxotrophy for recombinant protein overproduction in Escherichia coli. J Biotechnol. 2008, 134: 127-136.View ArticleGoogle Scholar
- Baba T, Ara T, Hasegawa M, Takai Y, Okumura Y, Baba M, Datsenko KA, Tomita M, Wanner BL, Mori H: Construction of Escherichia coli K-12 in-frame, single-gene knockout mutants: the Keio collection. Mol Syst Biol. 2006, 2: 2006.0008-doi:10.1038/msb4100050, View ArticleGoogle Scholar
- Tötemeyer S, Booth NA, Nichols WW, Dunbar B, Booth IR: From famine to feast: the role of methylglyoxal production in Escherichia coli. Mol Microbiol. 1998, 27: 553-562.View ArticleGoogle Scholar
- Truniger V, Boos W: Glycerol uptake in Escherichia coli is sensitive to membrane lipid composition. Res Microbiol. 1993, 144: 565-574.View ArticleGoogle Scholar
- Applebee MK, Joyce AR, Conrad TM, Pettigrew DW, Palsson BØ: Functional and metabolic effects of adaptive glycerol kinase (GLPK) mutants in Escherichia coli. J Biol Chem. 2011, 286: 23150-23159.View ArticleGoogle Scholar
- Mendoza-Vargas A, Olvera L, Olvera M, Grande R, Vega-Alvarado L, Taboada B, Jimenez-Jacinto V, Salgado H, Juárez K, Contreras-Moreira B, Huerta AM, Collado-Vides J, Morett E: Genome-wide identification of transcription start sites, promoters and transcription factor binding sites in E. coli. PLoS One. 2009, 4 (10): e7526-doi:10.1371/journal.pone.0007526, View ArticleGoogle Scholar
- Keseler IM, Collado-Vides J, Santos-Zavaleta A, Peralta-Gil M, Gama-Castro S, Muñiz-Rascado L, Bonavides-Martinez C, Paley S, Krummenacker M, Altman T, Kaipa P, Spaulding A, Pacheco J, Latendresse M, Fulcher C, Sarker M, Shearer AG, Mackie A, Paulsen I, Gunsalus RP, Karp PD: EcoCyc: a comprehensive database of Escherichia coli biology. Nucleic Acids Res. 2011, 39: D583-D590.View ArticleGoogle Scholar
- Beshay U, Miksch G, Friehs K, Flaschel E: Integrated bioprocess for the production and purification of recombinant proteins by affinity chromatography in Escherichia coli. Bioproc Biosyst Eng. 2009, 32: 149-158. 10.1007/s00449-008-0227-3.View ArticleGoogle Scholar
- Borriss R, Olsen O, Thomsen KK, von Wettstein D: Hybrid bacillus endo-(1–3, 1-4)-beta-glucanases: construction of recombinant genes and molecular properties of the gene products. Carlsberg Res Commun. 1989, 54: 41-54.View ArticleGoogle Scholar
- Jensen PR, Hammer K: The sequence of spacers between the consensus sequences modulates the strength of prokaryotic promoters. Appl Environ Microbiol. 1998, 64: 82-87.Google Scholar
- Ferguson GP, Tötemeyer S, MacLean MJ, Booth IR: Methylglyoxal production in bacteria: suicide or survival?. Arch Microbiol. 1998, 170: 209-219.View ArticleGoogle Scholar
- Stanbury PF, Whitaker A, Hall SJ: Principles of Fermentation Technology. 1995, Oxford: Butterworth HeinemannGoogle Scholar
- Cooper RA, Anderson A: The formation and catabolism of methylglyoxal during glycolysis in Escherichia coli. FEBS Lett. 1970, 11: 273-276.View ArticleGoogle Scholar
- Beshay U, Miksch G, Flaschel E: Improvement of a beta-glucanase activity test by taking into account the batch reactor balance of the test system. Bioproc Biosyst Eng. 2007, 30: 251-259. 10.1007/s00449-007-0121-4.View ArticleGoogle Scholar
- Sambrook J, Fritsch EF, Maniatis T: Molecular Cloning: A Laboratory Manual. 1989, Cold Spring Harbor: Cold Spring Harbor Laboratory PressGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly credited. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.