Background: Nowadays, the focus in metabolic engineering research is shifting from massive overexpression and inactivation of genes towards the model-based fine tuning of gene expression. In this context, the construction of a library of synthetic promoters of Escherichia coli as a useful tool for fine tuning gene expression is discussed here.

E. coli cells produce acetate as an extracellular coproduct of aerobic cultures. Acetate is undesirable because it retards growth and inhibits protein formation. Most process designs or genetic modifications to minimize acetate formation aim at balancing growth rate and oxygen consumption. In this research, three genetic approaches to reduce acetate formation were investigated: (1) direct reduction of the carbon flow to acetate (ackA-pta, poxB knock-out); (2) anticipation on the underlying metabolic and regulatory mechanisms that lead to acetate (constitutive ppc expression mutant); and (3) both (1) and (2). Initially, these mutants were compared to the wild-type E. coli via batch cultures under aerobic conditions. Subsequently, these mutants were further characterized using metabolic flux analysis on continuous cultures. It is concluded that a combination of directly reducing the carbon flow to acetate and anticipating on the underlying metabolic and regulatory mechanism that lead to acetate, is the most promising approach to overcome acetate formation and improve recombinant protein production. These genetic modifications have no significant influence on the metabolism when growing the micro-organisms under steady state at relatively low dilution rates (less than 0.4 h(-1)).

Recent developments in cellular and molecular biology require the accurate quantification of DNA and RNA in large numbers of samples at a sensitivity that enables determination on small quantities. Five current methods for nucleic acid quantification have been compared: 1) UV absorbance spectroscopy at 260nm, 2) Colorimetric reaction with the orcinol reagent, 3)Colorimetric reaction based on diphenylamine, 4) Fluorescence detection with reagent Hoechst 33258 and 5) Fluorescence detection with thiazole orange reagent. Genomic DNA of three different microbial species (with widely different G+C content) was used as well as two different types of yeast RNA and a mixture of equal quantities of DNA and RNA.We can conclude that for nucleic acid quantification a standard curve with DNA of the microbial strain under study is the best reference. Fluorescence detection with reagent Hoechst 33258 ® is a sensitive and precise method for DNA quantification if the G+C content is lower than 50%. In addition this method allows quantification of very low levels of DNA (ng-scale). Moreover, the samples can be crude cell extracts. Also UV absorbance at 260nm and fluorescence detection with thiazole orange reagent are sensitive methods for nucleic acid detection, but only if purified nucleic acids have to be measured.

1Many different extraction and analysis methods exist to determine the protein fraction of 2 microbial cells. For metabolic engineering purposes it is important to have precise and 3 accurate measurements. Therefore six different protein extraction protocols and seven protein 4 quantification methods were tested and compared. Comparison was based on the reliability of 5 the methods and boxplots of the normalized residuals. 6Some extraction techniques (SDS/chloroform and toluene) should never be used: the 7 measurements are neither precise nor accurate. Bugbuster extraction combined with UV280 8 quantification gives the best results, followed by the combinations sonication-UV280 and 9EasyLyse-UV280. However, if one does not want to use the quantification method UV280, 10 one can opt to use Bugbuster, EasyLyse or sonication extraction combined with any 11 quantification method with exception of the EasyLyse-BCA_P and sonication-BCA_P 12 combinations. 13 14

The present contribution focuses on the mathematical techniques used to solve steady state metabolic models for the case of an overdetermined system. Even when parts of the system are underdetermined it is possible to solve the model partially and obtain statistically meaningful results. This is illustrated with data gathered from a set of E. coli W3110.shik1 phosphate- or carbon-limited continuous cultures. It is shown that the low yield in shikimate for C-limited cultures is not due to a lower flux going to the shikimate pathway but is caused by a high secretion of byproducts. Carbon-limited cultures could be better for shikimate production than carbon-abundant cultures provided the byproduct secretion is reduced. Finally, flux calculations are compared with RNA expression data.