Biology Reference
In-Depth Information
As an example of such an approach, Salis et al., using the RBS calculator tool described
before, aimed to optimize the function of an AND-gate circuit that induces the expression of
GFP under the control of a T7 promoter only when two inputs, one activating the
expression of an amber suppressor tRNA and the other activating the expression of a
T7 polymerase amber mutant, are both present.
36
The accuracy of the AND gate, however,
depends on how strongly the T7 polymerase mutant is expressed, since weak expression
results in little expression of GFP under all conditions and overexpression results in
leakiness that produces GFP even when an input is absent. Using a quantitative model for
this specific AND-gate circuit,
46
the authors simulated the ideal level of T7 polymerase
expression and then used their predictive design of RBS program to generate an RBS
sequence targeting that ideal regime of expression, resulting in optimal AND gate function.
Semirational Directed Evolution
Nonetheless, models that describe circuit dynamics are not trivial to obtain and, when
absent, it becomes difficult to exactly estimate the circuit parameters that optimize
functionality. Directed evolution techniques offer a very flexible framework to optimize
small-scale circuits with a somewhat high probability of success.
47
The challenge with using
directed evolution for circuit design, however, is that regulatory circuits contain many parts
and thus are much further removed from the level of point mutations than traditional
directed evolution targets such as enzyme-binding sites. Therefore, it is necessary to use
rational approaches in conjunction with directed evolution techniques to reduce the search
space and increase the efficiency of the screening methodology.
48
This semirational directed evolution approach has been quite valuable to the field of
metabolic engineering, where the simple overexpression of genes does not always result in
high production of the target compound, as overexpression places metabolic burdens on
the cell.
49,50
As a result, expression and activity of enzymes participating in a metabolic
pathway need to be properly balanced to maximize titers for the desired final compound,
avoiding the accumulation of toxic intermediates.
1
Therefore, methods that rationally target
combinatorial diversity to multienzyme expression levels are useful.
73
For example, Pfleger et al. designed a library of tunable intergenic regions (TIGR) that can
be placed in operons to generate differential gene expression for each gene within the
operon.
17
These elements are composed of multiple regulatory parts that include mRNA
secondary structures, RNase cleavage sites, and RBS sequestering sequences. Upon
transcription, mRNA is cleaved at the RNase sites, thereby generating multiple transcripts
with various secondary structures that have variable RNA stabilities and translation
efficiencies. Using libraries of TIGRs inserted between the three genes of the operon for
mevalonate biosynthesis, the authors were able to optimize, through screening, the flux of
the mevalonate pathway, which produces a precursor to the antimalarial drug artemisinin.
Because TIGRs rationally varied the performance of key parts that control expression
variables for key enzymes in the pathway, screening of only
600 different genetic circuits
yielded strains that had up to seven-fold greater production than the original expression
system. Further characterization of these strains revealed that balance between multiple
enzymes was significantly different from the starting regulatory circuit.
B
On a larger scale, Church and colleagues have developed a platform for multiplex automated
genome engineering (MAGE).
18
This method utilizes oligonucleotide-directed mutagenesis,
mediated by an optimized system for Lambda red recombination, to modify multiple loci in
the genome of
E. coli
in parallel. The result is a highly scalable method for generating and
screening/selecting multiloci phenotypes, especially since the authors have created a setup for
automating the MAGE process. In short, MAGE facilitates the ability to generate variation at
scales of biological organization much larger than single nucleotides, allowing the
application of directed evolution to multicomponent genetic circuits. To demonstrate the
