CH391L/S12/MAGE lycopene production, CAGE "Amberless" E. coli
This article is currently under construction, any missing citations will be filled in upon completion.
MAGE Lycopene Production
Mulitplex Automated Genome EngineeringMAGE or multiplex automated genome engineering is a technique developed in George Church's lab at Harvard that can be used for large-scale programming and evolution of cells. Many directed evolution and in vitro technologies are limited to manipulations of single genes, making it slow if trying to alter much of the chromosome. MAGE has advantages over these techniques because it is able to simultaneously target many locations on chromosome for modification. Modification can be in the form of mismatch mutations, insertions, or deletions. Through the use of oligos with well-defined sequences, predictable modification can arise. However, through degenerate nucleotides, high-diversity chromosome modifications occur, producing a large variety of genetic variants.
Modification in MAGE is done through oligo-mediated allelic replacement, which is controlled by the λ-Red single-stranded DNA binding protein β. This protein works by binding the single-stranded oligo and helping it to displace the okazaki fragment on the lagging strand. Normally, the cell's repair proteins would spot the mismatch, but one of the key genes for the repair mechanism has been knocked out. (More details) Upon the next round of DNA replication, the introduced oligo is copied and becomes part of genome. It is also possible that the fragments to be replaced by other near matching oligos, generating huge diversity. The Church group found using 90-mer oligos were the most efficient for replacement. This is likely due to the λ-Red protein needing at least 30 bps to complex DNA, and that 90bp presents a good chance of homology to the target.. As well, oligos larger than 90-mer have a higher chance of forming secondary structure, greatly reducing replacement effieciency.
Efficiency of MAGE
Through oligo-mediated allelic replacement with the λ-Red single-stranded DNA binding protein β, mismatch mutations up to 30bp, insertions up to 30bp, and deletions of up to 45 kbp were introduced to cell populations. (Image Figure 2) The efficiency of a mismatch or insertion is based on its homology to the genome target, and deletion efficiency is based on size. Since oligos with more homology to target sites are incorporated at a higher frequency, MAGE can be tuned to provide desired evolution.
Measuring Sequence Diversity Rate
Finding the rate of genomic diversification by MAGE was done by making mismatch changes using three different 90-mer oligos to target a part of the lacZ gene. Olgis cN6 and cN30 contained 6 and 30 sequential degenerate bases, respectively. While, iN6 oligos had 6 degenerate bases spread out over a 30 bp region.  The data comes from 96 random clonal isolates after MAGE cycles of 2, 5, 10, or 15, which gives a good idea of variation in the cell populations. In the cN6 population, more than 4.3 billion variants were produced each day. After 15 cycles, all genotype combinations of N6 cell populations were created from either the cN6 or iN6. While only 21.8% of the cN30 population had accomplished allelic replacement in 15 cycles, likely because it is more difficult to match 30 consecutive degenerate bases. From this result, many have concluded that MAGE is a fairly inefficient technique. The Church group determined that MAGE diversity is dependent on: "the degree of sequence variation desired at each locus, the number of loci targeted, and the number of MAGE cycles performed". (Image)
Lycopene is a carotenoid pigment found in tomatoes and other red fruits and vegetables, because it is non-toxic and has antioxidant properties it is a good food coloring agent. It is also an important intermediate in the synthesis of many other carotenoids. Screening lycopene production is simple because colonies producing it show intense red pigmentation.
Exhibiting that MAGE can be used to target specific sequences with well-defined oligos, the 1-deoxy-D-xylulose-5-phosphate or DXP synthesis pathway responsible for lycopene production was targeted. To increase lycopene production, Wang et al. sought to modify 20 endogenous genes known to increase lycopene yield as well as 4 secondary genes responsible for decreasing yield. For the lycopene increasing genes, 90-mer oligos with degenerate RBS (DDRRRRDDDD, D=A,G,T; R=A,G)  sequences with some homologous regions on the sides, were used for genes increasing lycopene production. Because the replaced RBS sites were more similar to the Shine-Dalgarno sequence (TAAGGAGGT), translation efficiency increased. For the remaining four genes, two nonsense mutations were inserted into the open reading frame via oligos, inactivating these genes and increasing lycopene yield. Screening ~10^5 colonies after 5-35 MAGE cycles resulted in cell populations increasing lycopene production five fold relative to the ancestral strain. Sequencing variants showed that RBS convergence toward the consensus Shine-Dalgarno sequence.
Specificity of MAGE
Through the lycopene production pathway, the Church group showed that MAGE could be extremely specific depending on what oligos are introduced. From the DXP pathway, translation optimization of lycopene production genes such as idi alone (EcHW2a) increased lycopene production 40%, while optimizing dxs and idi increased production by 390% (ExHW2e).  It was also shown that in the secondary pathway, inactivation of gdhA increases lycopene production but lowers growth rate in EcHW2b by32%. The specificity possible by MAGE use was expanded by the Church group to other projects.
CAGE "Amberless" E. Coli
Hierarchical Conjugative Assembly (CAGE) was developed well in George Church's lab as a means to merge sets of codon modifications from MAGE into genomes with 80 precise changes. It was demonstrated that synonymous codon changes can be combined into strains without lethal effects on the cell population. E. coli has three stop codons and two release factors. Release factor 1 (RF1) recognizes UAA and UAG, while RF2 recognizes UAA and UGA. The hypothesis was that replacing all TAG codons with TAA codons, the genetic dependence on RF1 would be abolished and the newly introduced TAA codons would be recognized by RF2. The group sought to test if E. coli that had replaced all 314 TAG stop codons with TAA codons would be viable. (MG1655 genome) An interesting prospect of this technique is that if successful, the TGA stop codon may be recaptured into the genomic code for other purposes such as incorporation of unnatural amino acids.
Codon Modification Strategy
The genome of MG1655 (a mismatch repair-deficient strain) with 314 TAG stop codons ( at least 43 essential genes and 39 TAG codon overlaps of ORF) was split up into 32 regions, 31 of which had 10 TAG stop codons and one with four. This strategy was selected because pools of at least 10 oligos have been shown to have high replacement efficiency and that the total number of cell divisons to achieve replacement. MAGE was used to introduce all 10 TAG=>TAA codon modifications. The 314 oligos encoding these specific mutations were designed computationally on the basis of previous MAGE experiments. After 18 cycles of MAGE allelic replacement frequences were analyzed in 1504 clones (47 clones for each 32 segments). The average replacement frequency was 37 +/- 19% after 18 cycles, with 42% of the population unconverted. Two types of cells were shown to have been evolving: one that ready permits replacement and one largely resistant. It was also shown that no TAG stop codons were essentil for survival or robust growth.
Assembling Stop Codon Modifications
Through the use of Hierarchical Conjugative Assembly (CAGE), merging the modifications from MAGE was accomplished. This technique is dependent upon bacterial conjugation to transfer the modified segments. The oriT sequence typically used for conjugation is fused with kanR for integration into E. coli genome by the λ-Red mediated dsDNA recombination. This makes for precise control of conjugation initiation location use of a ~ 2-kb casette in place of the 30-kb Hfr fragment.  (Image Figure 4 A) The 32 strains were converted to 16 pairs for conjugation, with a donor strain transferring its genomic region to a recipient. Selectable makers control placement of transfered DNA, in the donor strain the recoded region was flanked upstream by the oriT-Kan cassete and downstream by a positive selectable marker (ie: antibiotic resistance). The recipient strain contained a different positive selectable marker and a positive-negative selectable marker (tolC), about the recoded region. Of the 81 integration sites tested, 12 gave no recombination, 23 sites had a recombination frequency of ~10^7, 38 sites with ~10^-6, and 8 sites with ~10^5 recombination frequencies.
The "amberless" E. Coli
The original 32 recoded strains were turned into 8 strains, each with 1/8 of the genome recoded. Two of these strains had a dysfunctional tolC phenotype, meaning that it passed the positive-negative control selections. MAGE was used to reconstruct these strains from the ancestral strain (also had mutation). In the end, 28 of the 31 conjugations were accomplished, still falling short of providing entire-genome modification.
His-Tagging with MAGE
In 2010 the Venter group of JCVI synthesized a 1.08–mega–base pair genome and transplanted it to M. capricolum to create the new M. mycoides. (Cite Venter) This synthetic biology achievement took 400 scientists year to create, along with a whopping price tag of $40 million. So unless you are Craig Venter, this approach to genome synthesis is likely unrealistic, but this is where MAGE and CAGE may shine. The idea is that you add only DNA with the specific changes to the genome desired. Even if you are trying to change hundreds or thousands of genes at once, after a few cycles in the machine, a good proportion of the cells should have all the desired changes. This can be checked by sequencing. (Cite http://b-giar.blogspot.com/2011/07/evolution-machine-genetic-engineering.html) However, due to the very low frequency of both oligo transformation and recombination, it does not yet seem that the Church groups' techniques provide a definite alternative to whole-genome synthesis.
- Wang HH, Isaacs FJ, Carr PA, Sun ZZ, Xu G, Forest CR, and Church GM. . pmid:19633652.
- Isaacs FJ, Carr PA, Wang HH, Lajoie MJ, Sterling B, Kraal L, Tolonen AC, Gianoulis TA, Goodman DB, Reppas NB, Emig CJ, Bang D, Hwang SJ, Jewett MC, Jacobson JM, and Church GM. . pmid:21764749.