As the fossil fuel supply continues to deplete novel
methods of energy and valuable product production are required. Microbes can
harness renewable resources to produce desired products and chemicals, however
often at low yields. This lab project is aiming to build a synthetic biology
toolkit using NeoFlex (an in house modular system based on EcoFlex) to alter
the expression of endogenous genes of the central metabolism in microorganisms
to divert carbon flux toward the necessary intermediates to increase yield.
This will be tested for the manipulation of product formation mainly in Escherichia coli. We will synthetically
assemble transcription units that can affect expression of endogenous genes by
methods reported in literature such as antisense RNA, CRISPR interference and
gene deletion and replacement. The effect of the inserted transcription units
can be tested both individually and as combined constructs to optimize the
target pathway, such as that for ethanol production.
and Importance of project
Since the early days of fermentative brewing mankind has
exploited the metabolism of microorganisms for numerous biotechnological
applications. Now microbes play a crucial role in many other industries to
produce antibiotics, vaccines, amino acids, organic acids, enzymes, vitamins,
and pharmaceutical drugs/precursors (Ran et
al., 2008). The irrevocable consumption of the planets fossil fuel supply,
coupled with the resultant effect of climate change has initiated the beginning
of a global effort to shift toward sustainable approaches to fuel our world. Microbes
are part of this shift as they can convert carbon based compounds into biofuels
(such as ethanol, bio-butanol), as well as produce many other chemicals and
polymers used industrially (for example 1,4-Butanediol used in plastic
production and Isopropanol used in personal care products) (Yim et al., 2011. Hanai et al., 2007).
As exemplified by the increase in Penicillin yields from
3mg/l (Alexander Fleming’s strain of Penicillium
notatum) to now 20g/l, product yields can be maximised through random
mutation, screening, selection and mutating feedback inhibition systems.
However, within a timeframe yield can only be increased so much. Other yield
enhancing methods such as fusing enzymes, as exemplified by the work of Lewicka
et al. (2014) fusing pyruvate
decarboxylase and alcohol dehydrogenase to increase ethanol production in Escherichia coli offer some novel
alternative approaches to increase yield.
Synthetic biology can introduce new metabolic pathways
into microorganisms for the production of valuable products but also can provide
us with the ability to alter the expression of endogenous genes of the
organism’s central metabolism to divert carbon flux toward the necessary
intermediates to quickly enhance yield of the desired product whilst maintain
the redox balance within the cell. Whereas genetic engineering may be
considered small scale manipulation, synthetic biology seeks to engineer entire
genetic code for a desired product/outcome through using toolkits made up of standardization, abstraction,
and de novo DNA synthesis.
Microorganism metabolism and diverting flux toward
The use of ethanol as a sustainable biofuel is helping to
reduce consumption of fossil fuels and is produced as an end product of the fermentation
of cellulosic and lignocellulosic biomass by microorganisms but the widely used
fermenting yeast in the beverage industry, Sacharomyces
cerevisiae, cannot utilize monosacharides generated from hemicellulose to
produce ethanol. For ethanol to be
considered one of the serious alternatives to fossil fuels we must maximise the
yield of ethanol produced per microorganism per unit of biomass and so the
challenge of developing of the most efficient microorganisms at converting
biomass-derived hexoses and pentose’s to ethanol is presented (Ragauskas et
al., 2006). With the sole aim of maximising yield in optimal
conditions in an ethanologenic strain of E.
coli (pathway originally from Zymomonas mobilis), Trinh et al. (2008) used
8 gene knockout mutations (zwf, ndh, sfcA, maeB, ldhA,
frdA, poxB, and pta) as shown in Fig. 1, by P1
transduction selected by elementary mode (EM) analysis – a more thorough
approach than flux balance analysis methods such as OptKnock and MOMA) – to
reduce the metabolic pathway options (which have evolved to allow the cell to
survive under changing environments such as decrease ethanol production during
growth) by 99.9% leaving only 6 crucial pathway possibilities that maximise the
conversion of sugars into ethanol for example by removing
catabolite repression allowing simultaneous utilization of glucose and xylose.
In reality the genomes of the
minimal strains contain most genes present in wild-type cells but the
combination of expressed gene products supports only a few pathways as their
findings indicated a reduction in genome size wouldn’t drastically alter the
phenotype of these cells. This study is of added interest because the
methodology used is general and the stains with simple metabolic pathways could
designed for many biotechnological applications.
Many of the enzymes in secondary metabolic pathways have
lower activity compared to those involved in the central carbon metabolism
making it harder to overproduce secondary metabolites and considering generally high-throughput screening tools are
not available for metabolic pathway-engineering Fang et al. (2016) developed a biosensor assisted method to sequentially
optimize the deoxyviolacein biosynthetic pathway in E. coli names intermediate sensor-assisted push-pull strategy (InterSPPS).
The pathway was split into two modules with L-Trp (the biosensor) being the
product of the upstream module and the substrate of the downstream module for
deoxyviolacein synthesis. The result was an increase of 4.4-fold (1.92 g/L) in
the deoxyviolacein titer (the highest deoxyviolacein production reported).
Their report suggests biosynthetic pathways can be enhanced with biosensors to
produce desired products without end-product screening.
In order to increase succinate yield and productivity in the
anaerobic central metabolic pathway of E. coli, Sánchez et al., (2005) made 4 deletions and 1
overexpression by ?Red recombinase method of chromosomal disruption and
P1-phage transduction to knockout NADH competing pathways and divert carbon
flux in the necessary direction (Fig 2).
Non-coding RNA: A natural regulatory molecule
There are many known RNA regulators of gene expression,
acting via different mechanisms in different organisms which has spurred
interest to create RNA bases synthetic control systems. Firstly, structural
mechanisms: Kozak (2005) discovered that secondary structure of mRNA is able to
shield access to the ribosome binding site (RBS) thus inhibiting translation.
Prokaryote mRNA transcripts can also form tight hairpins pausing translation thus
further regulating gene expression, and are shown to alter translation to in
response to temperature shock (Yanofsky, 1981. Kozak, 2005). Secondary
structures on the 5? and 3? ends of a mRNA are also provide protection from
ribonucleases resulting in increased transcript half-lives and subsequently
translation (Alifano et al., 1994).
RNA also acts catalytically in the form of ribozymes to
cleave/ligate the RNA backbone via phosphodiester bonds. This is used by
prokaryotes and eukaryotes for alternative splicing, RNA replication,
translation, riboswitches, and transcript stability and function. RNase P RNA
acts in trans to execute multiple turnover cleavage events when processing 5′ leader
sequences from tRNA, showing that one catalytic RNA can regulate many different
genes in a system (Serganov and Patel, 2007).
Perhaps the most important method of RNA protein
synthesis regulation is antisense-mediated regulation. In prokaryotes, small
(50-514 nucleotide), trans-acting RNAs (sRNAs) modulate target
gene expression by base-pairing with their target mRNA (earning them the name
antisense RNAs (asRNAs)) preventing translation by changing mRNA accessibility
to the translation machinery or increasing the rate of transcript degradation (Schluter et al., 2010). In
eukaryotes, the RNA interference (RNAi) pathways are extensively used to
regulate expression via small interfering RNAs (siRNA) and microRNAs (miRNA) to
silence transcripts via complementary binding (Carthew and Sontheimer, 2009).
Synthetic engineering of non-coding RNA
ncRNA has mostly been designed by
Watson-Crick base pairing with the 5′ UTR of the target mRNA or DNA (as well as
secondary structure prediction programs e.g NUPACK and Mfold, which estimate
the Gibbs free energy of RNAs) to allow the rational design of synthetic ncRNA
needed for regulation of gene expression (Lee and Moon, 2018). Numerous types
of synthetic ncRNAs have been rationally engineered to regulate gene expression
synthetic ncRNAs have been rationally
engineered to regulate gene expression.
We will cover the major examples of synthetic ncRNAs which have
been developed mainly in bacteria to regulate gene expression at the post
transcriptional or translation level.
Synthetic antisense RNA (asRNA)
Until recently RNA interference was only
being used as a tool for gene silencing in eukaryotes, but using the method as
above Man et al., (2011) developed
the first artificial trans-encoded prokaryotic sRNAs (atsRNAs) able
to suppress the expression of exogenous EGFP gene and endogenous uidA gene
in E. coli, which also
revealed atsRNA-mediated gene silencing was Hfq (AU rich binding site)
dependent and asRNA’s action can be significantly improved by adding an Hfq binding site on the
3?-end of target binding region (TBR) of a synthetic asRNA. Naturally occurring asRNAs regularly target multiple
mRNAs but only recently has multi-targeting RNA in E. coli been possible synthetically thanks to Lahiry
et al (2017)
modifying native antisense fingerloop motifs in DsrA.
When developing synthetic asRNA one
must consider many design factors that affect silencing as shown in Fig. 3 including how
changing one of these factors may influence the other (e.g changing asRNA
length will change the thermodynamics of the interaction with mRNA. To design a synthetic
asRNA, a 3′ Hfq binding site is needed on the TBR to improve the asRNA’s gene
silencing and decrease off target effects (Lee and Moon, 2018). The
thermodynamics (?G) of the asRNA-mRNA interaction must also be considered as
research suggests ?G can show a strong correlation between the ?G of the
asRNA-mRNA complex and repression (Na et
al., 2013). As asRNA is likely to be degraded once complexed with mRNA the
transcription rates of asRNA and mRNA should also be considered for creating an
asRNA gene expression control system. So to silence an mRNA it is recommended
to ensure asRNAs are in stoichiometric excess of the target mRNAs (achieved
using a higher copy number plasmid or a stronger promoter for asRNA
expression) (Hoynes-O’Connor and Moon, 2016).
Hoynes-O’Connor and Moon (2016) also found a liner correlation in the increase
in TBR length (thus length of dsRNA) and increase in repression and a negative
correlation between the number of mismatches in the TBR and the level of
To research the possibilities of
metabolic engineering in E. coli, Na et al., (2013) created synthetic sRNAs
to modulate target gene expression through combinatorial knockdown of 4 genes: tyrR (tyrosine repressor), csrA (regulates
the expression of enzyme genes in glycolysis), pgi (phosphoglucose
isomerase, converts glucose-6-phosphate to fructose-6-phosphate) and ppc (phosphoenolpyruvate
carboxylase, converts phosphoenolpyruvate to oxaloacetate). They produced
strains capable of producing 2 g/l tyrosine and identified endogenous gene targets
such as murE that when repressed
upregulated cadaverine production by 55%. They built sRNAs out of a MicC scaffold
sequence (for Hfq protein recruitment) from naturally occurring sRNAs in E. coli as this has the best
repression capability and a target binding sequence through site-directed
mutagenesis. Their work is generalizable to other bacteria and more appealing
than conventional gene-knockout strategies as it’s easy to implement, ability
to control chromosomal gene expression without modifying the actual genes thus
not relying on pre-constructed strain libraries.
Small transcription activating
RNA (STAR) and Attenuator system
Although there are ncRNAs which can do
so, there are no naturally occurring sRNAs which can directly activate
transcription. However, Chappell et al.,
(2015) recently developed an sRNA-mediated, transcription activation system
(STAR) which consists of 2 RNA elements: an intrinsic transcription terminator
(pT181 terminator) between the promotor and RBS of the target gene and an sRNA
transcriptional activator which prevents the terminator hairpin formation by
RNA-RNA interactions permitting transcription (Fig. 4). With the STAR method,
they achieved 9000-fold activation of target gene expression providing an
exciting system for regulating gene expression levels. The intrinsic terminator
is a central feature to the system and functions in E. coli by its short RNA hairpin and a U-rich sequence (U-tract) which
cause dissociation of the ternary elongation complex thus terminating
transcription. Work by Chen et al.
(2013) characterising a library of natural and synthetic terminators revealed
that a perfect U-tract, a loop consisting of GAAA and an 8-bp stem gave the
strongest terminator strength.
The attenuator system (Fig. 4) comprises
of an attenuator and an asRNA. The attenuator is in the 5? UTR of mRNA to
regulate gene transcription downstream gene via RNA structural changes. Only in
the presence of asRNA does the terminator hairpin form due to complementary
base pairing. The issue with this system is only a limited number of orthogonal
attenuators exist in nature so to overcome this Takahashi and Lucks (2013) created
chimeric attenuators using parts from translational regulators and then used
mutagenesis to expand the number of chimeric attenuators thus increasing the
potential to use transcriptional attenuators as orthogonal regulatory tools.
CRISPR: Interference, Activation and
interference and asRNA are effective at RNA knockdown, however it is prone to
off target effects and to visualise RNAs exogenous tags are required (Abudayyeh
et al., 2017). The development of the
CRISPR interference (CRISPRi) system is based on the CRISPR-Cas system (that
binds and cleaves sequence specific foreign DNA) with mutated RuvC1 and HNH Cas
nuclease domains (dCas) to represses target gene expression in a controllable
manner without cleavage (Lee and Moon, 2018).
Single guide (sg)RNA (Fig. 5) guides the dCas9 to its specific DNA target and
the riboprotein complex interference represses transcription through
interfering with RNA polymerase binding and transcriptional elongation (Lee
and Moon, 2018). Many different systems
including type II CRISPR-Cas from S.
pyogenes, CRISPR-Cas9 from S.
thermophilus and type I CRISPR-Cas from E.
coli have all been used for programmable gene repression.
dCas9 can also be manipulated to become
a transcriptional activator by joining an effector (such as the p65 activation
domain in eukaryotes) to create a CRISPR activation (CRISPRa) (Fig. 5) system,
however just one effector (the ? subunit of RNA polymerase) fused to dCas9 has
activated gene expression in bacteria and was dependant on the target location
of the sgRNA-dCas9-? riboprotein complex. In order to induce maximum gene
expression, the optimal distance from the promotor should be found (Bikard et
CRISPRi is of particular interest
because when combined with asRNA (Fig. 5) it allows for dynamic regulation of
gene expression within a cell in a reversible manner unlike irreversible gene
knockouts. This is possible by expressing a synthetic asRNA which targets an
artificial linker between dCas9 and the binding site of sgRNA thus when
targeted prevents sgRNA from binding the target DNA and allows for flexible
sequence selection of asRNA as its binding sequence is engineered based on the
inserted linker region of sgRNA creating an asRNA with a high de-repression
efficiency (up to 95%) (Lee et al.,
on therapeutic applications Abudayyeh et
al. (2017) engineered an RNA-guided RNA-targeting CRISPR–Cas effector
Cas13a8 for endogenous mammalian cell RNA knockdown and binding, however their
initial screening of orthologues to identify the most effective in an
interference assay (Cas13a from L. wadei) was done in E. coli giving
us great insight into using CRISPR in our project. They tested knockdown for 3
endogenous genes (KRAS, CXCR4 and
PPIB) and their results were knockdown levels comparable to those
by RNA interference but with improved specificity (40.4% for PPIB, 83.9% for CXCR4, 57.5% for KRAS).
A comparison of reporter gene (Gluc and Cluc) knockdown by Cas13a and RNAi
showed 5/6 of the top performing guides achieved significantly higher levels of
knock-down (P < 0.05) than the matched shRNAs but only 2/6 guides showed significantly better (P < 0.01) knock-down of endogenous genes. Their results also suggest that target accessibility plays a large role in the targeting efficiency of the Cas13a. Gene replacement Gene replacement has been used fairly extensively in metabolic engineering in order to increase product formation. For example Yuan et al., (2006) used ?-Red recombinase to increase ?-carotene production in E. coli to 6 mg/g of dry cell weight (much higher than the previously proposed maximum production limit by the cell) by switching the native promotor for isoprenoid genes with a stronger bacteriophage T5 promotor. This idea can be used to replace endogenous ncRNA regulators with synthetically developed ones. Modular Systems Modular systems and toolkits allow simple generation of constructs which can be used to target normal cellular processes by the variety of different mechanisms as mentioned above. There are many successfully made toolkits such as that created by Lee et al, (2015) which contains a basic set of characterized parts allowing for rapid, modular, multipart assembly in Yeast, as well as complex Synthetic Transcriptional Programs with CRISPR RNA Scaffolds which encode target locus and regulatory action allowing for flexible redirection of flux through a complex metabolic pathway in yeast (Zalatan et al., 2015). Synthetic biology is a relatively new field and advances have been restricted due to lack of interoperable parts, techniques for dynamically probing biological systems and frameworks for the reliable construction and operation of complex networks (Lu et al., 2009). Until recently, golden gate cloning methods have been overlooked for synthetic biology in bacterial systems but the recent arrival of the CIDAR kit by Iverson et al. (2015) and the Ecoflex (Fig. 6) toolkit by Moore et al. (2016) allows for customisable and well characterized, one-pot MoClo assembly of constructs up to 20 genes long, offer exciting potential for altering endogenous genes for example through the assembly of synthetic CRISPR interference and antisense RNA gene constructs especially when individual transcription units are combined into level 2 and 3 modules to see the overall effect on the target pathway.