196 10 Programming Gene Expression by Engineering Transcript Stability Control and Processing in Bacteria
powerful platform for meeting system design goals for applications, such as met-
abolic pathway engineering or biosensing, that require the ability to generate
specific levels of static gene expression or dynamic genetic outputs that change
as the function of a targeted molecule. Many of the naturally occurring mecha-
nisms can be tuned to program static levels of gene expression [67], and dynamic
control [11] is possible if these static mechanisms are regulated by the binding
activities of functional RNA structures evolved with in vitro selection to bind
specific metabolites (e.g., RNA aptamers, or RNA aptamer-regulated ribozymes,
aptazymes) [7].
Several TSC mechanisms have been used over the past 15 years in synthetic
genetic systems. A small number of TSC mechanisms, namely, 5′ and 3′ UTR
hairpins [67–72], 5′ UTR cleavage [7], and antisense RNA (asRNA)/sRNA bind-
ing [10, 53, 73–76], have been used to explicitly control transcript stability. (The
systems developed using these mechanisms are discussed in detail in the follow-
ing subsections.) Others were not explicit attempts to alter transcript stability
[9, 10, 77, 78]. Rather, by changing ribosome binding and UTR structure, there
were likely changes in RNA degradation, even though altered stability was not
the chief actuator of control. Nevertheless, this substantial body of work has
significantly advanced knowledge of RNA engineering that will undoubtedly be
important in creating novel genetic control systems based on tuning mRNA
stability. Moreover, this work has helped identify RNA components that are most
easily engineered and understood and has reinforced the many strengths of
RNA-based technologies, namely, low host metabolic burden [75], inherent
orthogonality [76], and the evolvability [79, 80] of new components. With this
work and advancing knowledge of degradation processes, transcript stability
is poised to become a powerful means of genetic control, either on its own or as
part of a larger control scheme.
Moving forward with increasing understanding of RNA device design princi-
ples and mechanistic understanding of degradation processes, it should be pos-
sible to formulate model-driven frameworks based on TSC mechanisms. Casting
biochemical, mechanistic understanding of transcript degradation in terms of
measurable and tunable design variables will enable us to take advantage of com-
putational techniques to increase the speed of design, predictability, and scale of
synthetic biological systems [7].10.2.2 Secondary Structure at the 5′ and 3′ Ends
The earliest attempts to engineer the stability of transcripts in bacteria involved
hindering ribonuclease’ entry by adding stable hairpin secondary structures
to the 5′ end of transcripts (Figure 10.3a) [67–69, 81, 134] or to the 3′ end [70, 71],
followed by hairpins at both termini [72]. When a hairpin from the T7 gene10
leader sequence was added to the 5′ end of lacZ in E. coli, β-galactosidase activity
increased threefold, but only when RNase E was present, suggesting that the
hairpin increased transcript half-life by reducing RNase E binding and cleavage
rates [69]. A similar experiment also saw a threefold improvement in half-life
after a 5′ hairpin addition [67]. Carrier and Keasling built a small library of 5′
hairpins that conferred an order-of-magnitude range in half-life, from 2 min up