Exploring the limits of the xylS/Pm expression cassette for recombinant protein production in Escherichia coli:

Abstract

Recombinant protein production is applied for a wide range of applications, however, so far no universal system exists which allows for production of sufficient amounts of any protein of interest. Thus improvement of existing systems and better understanding of their molecular function is still as important as the discovery and development of new expression tools. Particularly expression systems which allow a certain degree of flexibility and adaptation concerning growth conditions, expression levels, and choice of expression host appear promising. The work described in this thesis concentrates on vectors with these characteristics. They combine the positively regulated XylS/Pm promoter system with the broad-host-range RK2 mini-replicon. They are well established and have been shown to be capable of industrial-level production of several medically important proteins. Intensive studies, especially with bla (coding for β-lactamase) as reporter gene, and mutagenesis of several of the control elements in the applied system have resulted in identification of variants that increase expression levels also for other reporter genes tested. In this thesis it was aimed to assess, whether there was potential for further improvements of the vectors, and indeed, a combination of several high-expression variants in one cassette resulted in higher expression levels than those achieved by single variants. With the reporter gene bla about 75 times as much protein as with the wild type system could be produced when variants for Pm, the DNA-region corresponding to the 5’-untranslated region of mRNA (5′-UTR) upstream of the reporter gene, and the gene of the positive regulator of the system, XylS, were combined. As a result of this combination corresponding transcript levels could also be increased up to 63-fold.

Testing of these new vectors with several other reporter genes demonstrated that expression of them in general could be increased, however, only within the limits of gene- and protein-specific boundaries. This was particularly investigated in the case of celB (coding for phosphoglucomutase), and a new role for the plasmid copy number was discovered: an increase in plasmid copy number does not only lead to higher expression levels due to increased production of transcript, but also facilitates better exploitation of the resources due to the larger distribution of mRNA across the cell.

The transcription factor XylS is of central importance for the here applied vectors and a study of its expression revealed new insights into limitations and mechanisms by which it activates Pm. Its expression could not be increased by codon optimization, which was partly, but not only, caused by secondary structure formation in its translation initiation region. However, XylS expression could be improved either by exchange of its 5’-UTR and/or the promoter from which it was expressed. The choice of the positively regulated promoter system ChnR/Pb for transcription of xylS allowed variation of XylS expression levels with concurrent investigation of its mechanisms of activation of Pm. A synthetic operon construct permitted to detect XylS at physiological concentrations. This revealed the existence of a roof for induced expression from Pm depending on XylS expression levels, which probably is caused by oligomerization of the transcription factor at high levels inside the cell. This indicates that it is redundant to increase XylS expression levels beyond the saturating concentration, and the maximum induction ratio at Pm with the wild type system was found to be around 700-fold, with bla as reporter gene.

Regulation of XylS expression was also demonstrated to be useful in reduction of background expression from Pm, which often is increased, when induced expression is improved. This might provide benefits for expressing host-toxic proteins. Also background levels of the vectors which combine several high-expression variants, could be reduced by XylS regulation.

Another way of background reduction and enhancement of the induction ratio was demonstrated to be the usage of low-expression 5′-UTR variants, which turned out to be a promising approach for metabolic engineering purposes. This also revealed Rho-dependent termination of celB-transcription at low expression levels.

The combination of high-expression variants turned out to be useful in reduction of the metabolic load, which is imposed onto cells by recombinant protein production. The high expression levels which were achieved by this approach allowed expression from one single copy on the chromosome at levels higher than that of the wild type plasmid system.

The findings described here demonstrate that there is potential to further improve expression from vectors that combine the XylS/Pm expression system with the RK2 mini-replicon. In addition there already exist variants that cause varying phenotypes for several of the elements in these vectors, and it offers a large toolkit with many possibilities for combination of these features. Thus the system probably can be adapted to expression requirements of many proteins. Some of the reported findings can also be seen in a broader context and potentially adapted to other systems, as for example the combination of variants with defined phenotype. Low-expression variants might be helpful to reduce background levels of strong promoters like in the largely applied T7-system. The discovery of the new role in plasmid copy number might be of importance for every expression system and the mechanisms which were revealed for XylS may potentially also exist for other and related transcription factors.