In a previous blog post, we talked about the role of metabolic engineering and synthetic biology in facilitating a return to natural products as drug candidates in drug discovery and development. In the August 13 issue of Nature, George Church (Harvard Medical School) and his colleagues reported on their new method for accelerating the optimization of metabolic pathways to produce medically and industrially useful natural products.
The Church group calls its technology Multiplex Automated Genome Engineering (MAGE). MAGE is an efficient, inexpensive, automated system to simultaneously modify many targeted chromosomal locations (such as genes or regulatory elements) across a large population of cells, through the repeated introduction of synthetic oligonucleotides. A bacteriophage-mediated homologous recombination system is used to replace the targeted sequences with sequences of the introduced oligonucleotides. As the result of this process, researchers obtain a heterogeneous, highly diverse population of cells. Researchers may subject this population to selection for a desirable property, such as more efficient production of a desired product. The selected cells may then be subjected to additional rounds of MAGE, followed by additional rounds of selection. The result is the evolution of strains that efficiently produce the desired product. These strains may be scaled up to produce the product for research or commercial purposes.
The Church group chose to demonstrate their MAGE technology by optimizing a pathway in Escherichia coli for production of the carotenoid lycopene (the red pigment found in tomatoes and watermelons, which is valued as a nutraceutical). These researchers’ approach to utilizing and optimizing this pathway builds upon the work of leading metabolic engineers Jay Keasling (University of California at Berkeley) and Gregory Stephanopoulos (MIT).
Carotenoids such as lycopene are members of a larger class of compounds called isoprenoids. Another class of isoprenoids is the terpenoids. As discussed in our previous blog post, terpenoids include numerous marketed natural product drugs, and this class of compounds is also of interest to researchers interested in discovering novel drugs. Because of common pathways for biosynthesis of precursors of carotenoids and terpenoids, Church’s work on optimizing production of lycopene in E. coli is relevant to researchers interested in applying synthetic biology to the synthesis and study of terpenoid drugs.
The pathway in E. coli (and in other prokaryotes) for synthesis of isoprenoids is known as the DXP (deoxyxylulose-5-phosphate) pathway. This is in contrast to the better-known mevalonate pathway, which is found principally in eukaryotes and in archaea. We discussed Dr. Keasling’s engineering of the mevalonate pathway in yeast and in E. coli (the latter of which was engineered to express this exogenous pathway) to produce terpenoid drugs in our 2007 synthetic biology report. A review of work on metabolic engineering of both the mevalonate pathway and the DXP pathway by the Keasling group and by others can also be found in a 2007 paper by Drs. Withers and Keasling.
In order to utilize the E. coli DXP pathway to produce lycopene, researchers must engineer the bacteria to express the enzymes that catalyze the final steps in lycopene biosynthesis (i.e., the three enzymes that convert the final product of the DXP pathway to lycopene). The Church group transfected their starting E. coli strain with a plasmid containing the genes (derived from another species of bacterium) for these three enzymes. The resulting E. coli strain produced lycopene at a basal level. It was that strain that the researchers subjected to MAGE.
The researchers used the MAGE system to target each of 20 endogenous E. coli genes in the DXP pathway. For each gene, they designed 90-mer oligonucleotides that contained variants of the gene’s ribosome binding site (RBS), in order to replace the endogenous RBS with one that would give more efficient translation of mRNA into protein. They also designed oligonucleotides to knock out four endogenous genes that encode enzymes that siphon off intermediates from the DXP pathway, in order to increase the flux through the DXP pathway to improve lycopene production. The total pool of oligonucleotides was in the hundreds of thousands. The goal was to optimize 24 genes simultaneously in order to achieve maximal lycopene production.
The researchers added the cells and oligonucleotides to the MAGE system, cycling the cells through oligonucleotide delivery (via electroporation), growth, and washing cycles, yielding billions of genetic variants per day. Every 24 hours, the researchers selected the variants that produced the reddest colonies, and thus the most lycopene. After only three days, the procedure yielded variants that exhibited a fivefold greater lycopene production than the starting strain, with a greater yield (approximately 9,000 micrograms per gram dry cell weight) than previously documented.
E. coli strains with an optimized DXP pathway, as developed by the Church group, could in principle be used to produce other isoprenoid compounds, including terpenoid therapeutics. In order to do so, researchers would need to transfect specific sets of genes to carry out the final steps of the biosynthesis of their desired compounds into the strains, instead of the specific lycopene biosynthesis genes used by the Church group. They might also use methods such as the “designed divergent evolution” technology developed by the Keasling group, to develop variants of enzymes that carry out the final steps of the biosynthesis of terpenoids, in order to discover novel terpenoid drugs that are not found in nature.
MAGE, which allows researchers to simultaneously optimize the expression of large sets of genes in a metabolic pathway, contrasts with traditional metabolic engineering, which is typically a slow process in which genetic constructs are introduced into cells one at a time. It thus represents a potential advance. However, as in the above MAGE-based optimization of lycopene production, applications of MAGE to natural product drug discovery and production will build on the work of metabolic engineers who use more conventional methods.