Exploring alternative expression systems for expression of heterologous proteins and metabolic pathways

Abstract

Our next goal was to use chromosomal integration for metabolic engineering. Multiple studies, from our laboratory and others, have found that modulatin of gene expression is key to pathway optimization. Generally, this has been achieved using promoters of varying strengths on plasmids of variable copy number. However, chromosomal integration offers advantages for metabolic engineering that could resilt in increased production of high-value chemical products. Here, we have shown, usnig a model five-gene pathway for the production of the purple pigment violacein from tryptophan, again that the location of the integration matters. We found that of the four genomic locations used, only one showed any level of violacein production, and it was surprisingly not the same location that resulted in increased levels of mCherry expression. Interestingly, when the one-gene TAL pathway, which converts phenylalanine to trans-cinnamic acid, was integrated into the same loci, very little difference in production was observed between the locations. These results indicate that optimization of production may not only be location dependent, but it may also depend on the genes being expressed.

Next, we attempted to find an improved expression system for the heparin biosynthetic enzymes. To this end, two species of Bacillus and the yeast Pichia pastoris were tested for their ability to secrete the heparin biosynthetic enzymes. While intracellular expression of enzymatically active 3-OST-1 was achieved in the Bacillus species, no level of secretion was measured using a small library of signal peptides in Bacillus megaterium. Successful secretion of all of the sulfotransferases required for the chemoenzymatic synthesis of heparin was achieved in P. pastoris. Here, we hae shown that when expressed as glycoproteins in P. pastoris, these enzymes show favorable sulfotransferase activity compared to their E. coli-expressed counterparts. These results indicate that the post-translational modification of protein glycosylation plays an important role in the activity of the heparin biosynthetic enzymes and that the yeast P. pastoris may be a good host for the expression and secretion of the heparin biosynthetic enzymes.

Heparin is the most widely used anticoagulant drug in the world. Current production of heparin from porchine intestinal tissue led to a contamination of the heparin supply in 2008 that resulted in over 100 deaths. This, coupled with the demand heparin puts on the worldwide pig population, has led to the need of non-animal sourced heparin. A chemeoenzymatic synthetic scheme, developed in our laboratory, has previously been used to successfully produce anticoagulant heparin. However, the enzymes for this synthesis are currently expressed and purified from Escherchia coli utilizing plasmid-based expression vectors. This Gram-negative organism is widely used for heterologous protein expression because of its fast growth and the vast molecular toolkit avilable. The work here aimed to alleviate problems with plasmid-based expression in E. coli.

Heterologous protein expression has long been carried out using plasmids as expression vectors in E. coli. Problems with plasmid-based expression, primarily the necessity for selection using antibiotics that can be expensive and deleterious to the environment, led to the desire to integrate the genes for the chemoenzymatic synthesis of heparin on the genomic chromosome of E. coli. While the expression of these enzymes was unsuccessful from the genome, it became apparent that chromosomal integration had other uses. Here, we have used model reporter proteins and pathways to show that there is much to learn about chromosomal integration. By integrating the gene encoding the fluorescent reporter protein mCherry into various loci throughout the genome of E. coli, we found that expression level was location-dependent. Further, we found that integration of a single copy of the gene into one particular locus resulted in over 5-fold higher mCherry expression than a high-copy plasmid expressing the same gene. We had hypothesized that to achieve the same level of expression found in plasmid-based systems, we would need to increase the copy number of the gene. Surprisingly, we found that when we increased the copy number of the gene in a single genomic locus, there was actually a decrease in mCherry expression.