10.4 Potential Mechanisms for Transcript Control 205problematic (see Section 10.4.1 for more information). As more researchers use
these tools to better understand synthetic RNA function, the aggregated data
will lead to further insights and design strategies that can take advantage of sta-
bility or mitigate unwanted effects.
10.4 Potential Mechanisms for Transcript Control
10.4.1 Leveraging New Tools
The advent of recent technologies, such as high-throughput RNA secondary
structure elucidation, high-throughput RNA sequencing, and co-transcriptional
RNA folding simulations, provides new ways to investigate transcript control for
predictable gene expression engineering. Most artificial RNA-based control
strategies have yet to take full advantage of these technologies to more predict-
ably engineer synthetic systems.
High-throughput RNA structural sequencing [121, 122] presents a new way to
examine the structures of large numbers of RNA molecules. These techniques
use RNA cleavage events dependent on the absence or presence of secondary
structure, followed by high-throughput sequencing, to develop a map of base
pairing probabilities. This map can then be used to constrain models from struc-
ture prediction software [123–125]. If paired with half-life quantification, this
methodology could provide better understanding of the connection between
UTR structure and transcript stability, enabling more predictable introduction
of secondary structure into transcripts.
MFE simulations using tools like Mfold [114, 115] or RNAfold [116, 117] have
been a mainstay in RNA secondary structure prediction. These tools calculate
the lowest energy state of an RNA, which is interpreted to be the steady-state
conformation of the RNA. When attempting to predict RNA secondary struc-
tures inside cells, MFE calculations may be misleading, for at least two reasons.
First, mRNA folding is co-transcriptional, and thus the full transcript sequence
is not available for folding at all times. Second, the relatively short half-life of
most mRNAs in the cell [2] can preclude their reaching the MFE conformation
before degradation. To address these issues, there are software packages that
take co-transcriptional effects into account [118–120] and can thus be useful for
predicting UTR secondary structures more accurately in a cellular context. In
fact, the creation of a transcript design method built around kinetic co-tran-
scriptional RNA folding simulations was crucial for the rRED and aRED engi-
neering described in Section 10.2.4 [7]. In that work, custom software written to
implement kinefold [119] on a computational cluster enabled the design of
spacer sequences to allow assembly of individually generated and characterized
RNA parts into genetic devices with quantitatively predictable functions. There
was significant divergence between the transcript folds predicted with MFE
structure calculations and those obtained with kinetic simulations, underscoring
the importance of RNA sequence and structure design that explicitly considers
co-transcriptional folding. To extend those results, we are currently developing a
computational platform for designing RNA parts, devices, and transcripts with