Plasmids have been the preferred method for pathway engineering in prokaryotes due to the ease of genetic manipulation and strong expression conferred by high gene dose.  However, fundamental flaws in plasmid propagation prohibit their usefulness in long term expression.  As low value-added products (e.g. biofuels) become a major focus in biotech applications, the prospect of low-cost continuous processes utilizing recombinant metabolic pathways would be useful but cannot be achieved using plasmid-based expression.  Genomic integration, while genetically stable, is cumbersome for the integration of multiple copies.  It is therefore imperative to develop new expression systems that will achieve high recombinant expression for long periods of growth.

Expression loss in plasmids can proceed by a mechanism we have termed ‘allele segregation,' which is not impeded by plasmid maintenance loci, such as antibiotic resistance/auxotrophic markers or par.  A mutation that results in a loss of pathway expression in one plasmid (with the remaining plasmids in the cell being functional) can be propagated to all copies of the plasmid, while still having a functional plasmid maintenance allele.  We have generated a subpopulation balance model that shows that the emergence of cells that have lost productivity is driven by (a) the possibility that one daughter cell can inherit both copies of duplicated mutant plasmid and (b) a growth advantage for fewer copies of the non-mutated allele.  Mutation rates, which previously have been the major focus of improving genetic stability, had little effect on the timescale of productivity loss.  Realizing the most relevant process was driven by the random distribution of plasmids to daughter cells, a new genomic integration approach was devised.

Tandem gene duplication is as a method to avoid the possibility of (a) by guaranteeing that each daughter cell will inherit only one copy of the mutant allele.  By this, it is predicted that the time to productivity loss can be increased by 10 fold.  Using homologous recombination, head to tails copies of an expression cassette can be created to any copy number up to 50.  The copy number is completely stabilized by recA deletion and by having all copies on one continuous strand of DNA, (a) cannot occur.  In addition, tandem gene constructs require no maintenance loci to stabilize the gene dose and have no variation in copy number in a population nor is copy number dependent on growth phase.

A description of the allele segregation subpopulation model, methods for creating tandem gene constructs, and an analysis of the stability and usefulness for metabolic engineering will be presented.  A case study for the production of poly-3-hydroxyburate (PHB) in E. coli showed that 23 copies of the PHB operon could be generated.  The PHB TGD construct maintained expression for over 70 generations with a specific PHB production of 40 mgPHB gbiomass^-1 h^-1, compared to only 12 generations for an analogous plasmid/antibiotics system.    This method is generally applicable for high gene dose expression over long periods of time without antibiotics or other plasmid maintenance strategies, and should be extendable to many other organisms, including minimal genome cells.