APPLICATION OF CELL-FREE SYSTEMS FOR THE
DESIGN OF A GLUCOSE-RESPONSIVE
BIOSENSOR AND FOR STUDYING THE EFFECTS
OF KEY SARS-COV-2 PROTEINS
A Dissertation
Presented to the Faculty of the Graduate School
of Cornell University
in Partial Fulfillment of the Requirements for the Degree of
Master of Science
by
Anirudh Murali Narayanan
December 2021
© 2021 Anirudh Murali Narayanan
ALL RIGHTS RESERVED
ABSTRACT
Cell-free platforms have been key in the rapid advancement of synthetic biol-
ogy. They enable the design of metabolic pathways for the production of any
desired proteins. Cell-free systems also present many advantages over tradi-
tional in vivo systems such as higher synthesis rates, direct manipulation of the
chemical environment, and the ability to even produce proteins toxic to cells. In
this study, we made use of the PURE cell-free system to construct a genetic cir-
cuit to sense glucose by use of a transcription factor, GntR. Towards this aim, we
performed tests of its repression and de-repression characteristics on a reporter
with a promoter construct containing the GntR operator site. The Pareto Opti-
mal Ensemble Technique was used to determine the parameters of the model for
the simulation of this circuit. This work helps to provide a cell-free alternative
for rapid point-of-care detection of glucose for diabetics. Additionally, in light
of the recent Covid-19 pandemic, we made use of the myTXTL cell-free system
to study some key SARS-CoV-2 proteins. We synthesized the viral host transla-
tion inhibitor protein, nsp1 in cell-free and tested its effectiveness to inhibit the
translation of a reporter in the same cell-free system. Further, we developed a
model to simulate the effect of the RNA-dependent RNA Polymerase (RdRp)
on viral protein replication in the myTXTL cell-free system. This work paves a
way for simulating the effect of RdRp on viral infection in host cells.
BIOGRAPHICAL SKETCH
Anirudh is from Chennai, India. He graduated with a Bachelor of Technology
in Chemical Engineering from Anna University in 2019. In the fall of that year,
he joined the Master of Science Program at the Robert Frederick Smith School of
Chemical and Biomolecular Engineering at Cornell University and started his
research with Dr. Jeffrey D. Varner. It is difficult to find an Anirudh in the wild,
but he has occasionally been spotted haunting the halls of Olin or Kimball after
midnight.
iii
This work is dedicated to my awesome brother, Vignesh.
iv
ACKNOWLEDGEMENTS
I would like to thank Dr. Jeffrey D. Varner for giving me the opportunity to
work in his research group as well as the guidance and support he has given
me over the past two years. My gratitude also goes out to Dr. Matthew DeLisa
for his insightful comments and his amazing courses which have been instru-
mental in shaping my interests. I am grateful for the friendship and support I
received from all VarnerLab members. In particular, I would like to thank Ab-
hinav Adhikari for mentoring and training me in pretty much everything I can
do in the wet lab. Thank you Alagappan, Sachin, Vashist, and Vivek for those
weekly game nights that often ran into the mornings. They were the highlights
of each week during the Covid-19 pandemic. I would also like to thank the
amazing friends I have had in Ithaca, particularly Avinash, Arvind, Abhaiguru,
and Harish for being such amazing housemates. Most importantly, I would like
to express my gratitude to my parents for encouraging and supporting me all
these years. Finally, I would like to thank my brothers, sisters, grandparents,
great-grandparents, and every uncle and aunt of my amazing family for always
being there for me.
v
TABLE OF CONTENTS
Biographical Sketch . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iii
Dedication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iv
Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v
Table of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi
List of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viii
1 Introduction 1
1.1 Cell-free protein synthesis . . . . . . . . . . . . . . . . . . . . . . . 1
1.2 Types of Cell-free systems . . . . . . . . . . . . . . . . . . . . . . . 2
1.3 Applications of Cell-free platforms . . . . . . . . . . . . . . . . . . 5
1.3.1 Production of Biologics and Specialized Proteins . . . . . . 7
1.3.2 Metabolic Engineering . . . . . . . . . . . . . . . . . . . . . 8
1.4 Mathematical Modeling of Cell-free systems . . . . . . . . . . . . 8
1.4.1 Metabolic Modeling . . . . . . . . . . . . . . . . . . . . . . 10
1.4.2 Integrated cell-free models . . . . . . . . . . . . . . . . . . 13
2 Development of a gluconate-responsive transcription factor-based ge-
netic circuit in a reconstituted cell-free system 15
2.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15
2.2 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . 19
2.2.1 mP70 promoter construction . . . . . . . . . . . . . . . . . 19
2.2.2 Synthetic Circuits Architecture . . . . . . . . . . . . . . . . 19
2.2.3 Cell-Free Protein Synthesis Reactions . . . . . . . . . . . . 21
2.2.4 mRNA Quantification . . . . . . . . . . . . . . . . . . . . . 22
2.2.5 Protein Quantification . . . . . . . . . . . . . . . . . . . . . 22
2.2.6 Estimation of model parameters . . . . . . . . . . . . . . . 23
2.3 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . 26
2.3.1 Parameter Estimation . . . . . . . . . . . . . . . . . . . . . 27
2.3.2 Future Directions . . . . . . . . . . . . . . . . . . . . . . . . 29
3 I(nsp)ection of the SARS-CoV-2 nsp1’s host-translation effect in a cell-
free system 32
3.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32
3.2 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . 35
3.2.1 Cell-Free Protein Synthesis Reactions . . . . . . . . . . . . 35
3.2.2 mRNA Quantification . . . . . . . . . . . . . . . . . . . . . 35
3.2.3 Protein Quantification . . . . . . . . . . . . . . . . . . . . . 36
3.2.4 Coexpression of deGFP with nsp1 . . . . . . . . . . . . . . 37
3.2.5 Qualitative testing of nsp1 binding with E.coli Ribosomes 37
3.3 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.3.1 Future Directions . . . . . . . . . . . . . . . . . . . . . . . . 41
vi
4 Modeling the effect of the RNA-dependent RNA polymerase in vitro 43
4.1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 43
4.2 Materials and Methods . . . . . . . . . . . . . . . . . . . . . . . . . 46
4.2.1 Model Derivation . . . . . . . . . . . . . . . . . . . . . . . . 46
4.2.2 Control functions, Transcription and Translation Kinetic
Limits . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51
4.2.3 Building the model . . . . . . . . . . . . . . . . . . . . . . . 52
4.2.4 Cell-Free Protein Synthesis Reactions . . . . . . . . . . . . 53
4.2.5 mRNA Quantification . . . . . . . . . . . . . . . . . . . . . 54
4.2.6 Protein Quantification . . . . . . . . . . . . . . . . . . . . . 54
4.3 Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . 55
4.3.1 Future Direction . . . . . . . . . . . . . . . . . . . . . . . . 57
5 Conclusion 59
Bibliography 60
vii
LIST OF FIGURES
1.1 Cell-free system based on cell lysate. The Cell-free lysate is pre-
pared by the lysis of cells followed by removal of cellular debris
and chromosome DNA. The energy sources, amino acids, nu-
cleotides, and cofactors are added to this to prepare the cell-free
extract. The template DNA can be added to it immediately to ex-
press the target protein. Alternatively, the cell-free extract can be
freeze-dried as pellets for storage or transportation along with
lyophilized DNA constructs. These can be rehydrated through
the simple addition of water for the synthesis of target proteins.
[179] . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
1.2 Cell-free systems and their applications. The left-hand side of
the image indicates the components needed for a cell-free system
and the right-hand side indicates the various fields in which cell-
free systems can be applied. [170] . . . . . . . . . . . . . . . . . . 5
1.3 The integration of metabolism with transcription and translation
(TXTL) processes: The TXTL processes utilize the macromolecu-
lar precursors from metabolism (such as NTPs, amino acids, and
cofactors) for gene expression. The integrated framework can be
represented as a stoichiometric matrix of metabolites participat-
ing in certain reactions, along with a description of the metabolic
demands for protein expression. By applying a pseudo-steady-
state assumption, making use of various constraints and an ob-
jective function, the metabolic fluxes in the system can be esti-
mated. In the case of cell-free systems, the objective function is
the maximization of target protein yield. [177] . . . . . . . . . . . 13
2.1 In this image, the first box indicates the components involved in
an aTF regulated circuit: the RNA polymerases, aTFs, and DNA
transcription templates. The second box indicates the circuit in
the absence of the ligand. In the presence of the ligand, the aTF
will no longer bind to the operator sequence and allows expres-
sion of the downstream reporter as shown in the third box. [72] . 16
2.2 (a) C1 Circuit: σ70 induces Venus expression (b) C2 Circuit: σ70
induces the expression of GntR and Venus. The GntR protein
once expressed represses the mP70 promoter of Venus. . . . . . 20
2.3 C3 circuit: σ70 induces the expression of GntR and Venus.
The GntR protein once expressed represses the mP70 pro-
moter of Venus. In the presence of a high concentration of D-
Gluconate, the binding of GntR to mP70 gets repressed thereby
de-repressing Venus transcription . . . . . . . . . . . . . . . . . . 21
viii
2.4 Venus protein expression without GntR (C1 - Red), Venus repres-
sion by GntR (C2 - Blue), and de-repression by D-gluconate (C3
- Green). The shaded regions represent one standard deviation
of the experimental measurements . . . . . . . . . . . . . . . . . . 26
2.5 Model simulations versus experimental measurements for σ70
induced Venus expression for the C1 circuit. (a) Simulated and
measured Venus mRNA concentration versus time. (b) Simu-
lated and measured Venus protein concentration versus time.
The black points indicate the experimental values while the
black curve is the predicted curve. The shaded region denotes
the 95% confidence interval of the predicted curve. The error
bars denote one standard deviation of the experimental data . . . 27
2.6 Model simulations versus experimental measurements for the
C2 circuit. (a) Simulated and measured Venus mRNA concentra-
tion versus time. (b) Simulated and measured Venus protein con-
centration versus time. (c) Simulated and measured Gntr mRNA
concentration versus time. The black points indicate the experi-
mental values while the black curve is the predicted curve. The
shaded region denotes the 95% confidence interval of the pre-
dicted curve. The error bars denote one standard deviation of
the experimental data . . . . . . . . . . . . . . . . . . . . . . . . . 28
2.7 Schematic of a glucose biosensor: σ70 induces the expression of
GntR, σ28, insulin and α-σ28 which is downstream to insulin. σ28
induces expression of glucagon while α-σ28 inhibits the P28 pro-
moter thereby inhibiting the expression of glucagon. When the
gluconate levels are high, insulin will be expressed along with
σ28. At low gluconate levels, only glucagon will be expressed . . 31
3.1 Circuit representing coexpression of nsp1 along with a reporter
protein. The nsp1 binds to the ribosomes and selectively pre-
vents translation of the reporter protein. This circuit doesn’t in-
clude the potential mRNA degradation activity of nsp1 [61, 73,
165]. (Created with BioRender.com) . . . . . . . . . . . . . . . . . 34
3.2 Gene expression of nsp1. The points denote the mean, while the
error bars denote one standard deviation calculated from three
replicates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 38
3.3 Coexpression of deGFP with nsp1: (a) RFU measurements of
deGFP when coexpressed with nsp1 (Red). The controls: expres-
sion of deGFP alone (Blue), expression of deGFP along with nsp8
(Green). The shaded regions indicate one standard deviation (b)
SDS-PAGE Analysis of these samples to confirm the production
of all the proteins. The gels have been Coomassie-stained . . . . 39
ix
3.4 Rudimentary nsp1-70S Ribosome binding assay: The third chan-
nel contains pure nsp1. The first and second channels contain
nsp1 mixed with two times its volume of purified 70S ribosomes
solution. In other words, the concentration of nsp1 in the third
channel is thrice the concentration of nsp1 in the first two. Mea-
surement of the fluorescence of the nsp1 bands show that the
normalized concentration of nsp1 in all wells are approximately
the same indicating that there is negligible binding of nsp1 with
these ribosomes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 40
4.1 Replication strategy of a (+)ssRNA virus: The viral genome is
translated using cellular machinery. This genome also acts as
a template for the production of complementary -mRNA. This
complementary strand then acts as a template for the production
of a new viral genome (+mRNA). (Image adapted from [89]) . . . 45
4.2 Introduction of RdRp into a genetic circuit: RdRp uses the
+mRNA as a template for the production of a complementary
-mRNA and uses the -mRNA as a template for the production of
+mRNA. Meanwhile, the +mRNA is also translated to synthe-
size proteins (Created with BioRender.com) . . . . . . . . . . . . 48
4.3 Temporal variation of N mRNA and protein when expressed
alongside RdRp. The blue curves represent the expression when
the model doesn’t account for the effect of RdRp, while the or-
ange curves account for the RdRp feedback to amplify mRNA
and thereby amplify protein expression . . . . . . . . . . . . . . . 55
4.4 Gene expression of nsp7. The points denote the mean, while the
error bars denote one standard deviation calculated from three
replicates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
4.5 Gene expression of nsp8. The points denote the mean, while the
error bars denote one standard deviation calculated from three
replicates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
x
CHAPTER 1
INTRODUCTION
1.1 Cell-free protein synthesis
Cell-free systems (CFS) have recently evolved into key platforms for synthetic
biology applications. Cell-free systems present many advantages over tradi-
tional in vivo systems. Since cell-free systems don’t require cellular growth and
maintenance, they can allow for higher synthesis rates by allocating more re-
sources towards production of the desired product. The lack of cellular growth
also means that they can be used for the production of proteins that are oth-
erwise difficult to produce due to their toxicity to cells. It is also possible to
directly observe and manipulate the chemical environment of cell-free systems
allowing for rapid tuning of the reaction conditions [46, 92, 179].
This chapter is adapted with permission from the study of my colleagues
Vilkhovoy et al. at the VarnerLab [179]. Cell-free systems have historically been
used as investigative tools for Molecular Biology research. Some of the first
examples of their use were in the 1950s by Borsook [15] and Winnick [190] to
study the incorporation of amino acids into proteins. The work of Gale et al. to
study the effect of nucleic acids in protein synthesis made use of Staphylococcus
aureus extract [40]. Even the famous Nirenberg and Matthaei experiments in the
early 1960s to discover the genetic code was performed using E. coli cell-free
extracts [100, 117].
One of the first precursors to modern cell-free transcription and translation
(TXTL) platforms was developed by Lederman and Zubay [84] who developed
1
a coupled transcription-translation bacterial extract that allowed DNA to be
used as a template to directly synthesize proteins. In 1988, Spirin et al. de-
veloped a continuous cell-free system with a continuous flow of the feeding
buffer and a continuous removal of the protein product. They were able to run
their system for tens of hours [155]. The Swartz lab improved the energy effi-
ciency of E.coli based CFPS by generating ATP with substrate-level phosphory-
lation [77]. And this was further improved by oxidative phosphorylation in the
Cytomim system that they developed [67–69]. Then the Noireaux lab developed
the myTXTL platform that couples ATP regeneration with inorganic phosphate
recycling thereby further boosting the protein production [42].
Cell-free platforms can also be used to design synthetic genetic circuits to
control gene expression. Bacteriophage-encoded RNA polymerases (RNAP)
such as T7, T3, and SP6 RNAP are commonly used in CFPS for their higher tran-
scription efficiency [118,143]. The use of bacterial regulatory elements based on
the sigma factor family has allowed for the implementation of multi-layer ge-
netic cascades in which the protein produced in each stage is the input that is
required to promote or inhibit the following stage [2, 42, 118, 143, 144].
1.2 Types of Cell-free systems
Cell-free systems can be broadly classified into two major classes: crude cell
lysate-based systems and reconstituted systems. Cell lysate-based systems have
been in use for the major part of the history of cell-free systems. These are
prepared by the lysis of cells where the cell’s transcription and translation ma-
chinery is retained while cellular debris and chromosomal DNA are discarded.
2
This is followed by the addition of necessary components such as amino acids,
energy sources, salts, and buffers. They can be prepared from a variety of
cells such as cells from bacteriae, protozoa, plants, insects, and mammals. E.
coli , S. cerevisiae, HeLa, rabbit reticulocytes, wheat germ, and High Five in-
sect cells are commonly used cells for the preparation of such cell-free extracts
[9, 35, 42, 127, 132, 149, 193].
Figure 1.1: Cell-free system based on cell lysate. The Cell-free lysate is prepared
by the lysis of cells followed by removal of cellular debris and chromosome
DNA. The energy sources, amino acids, nucleotides, and cofactors are added
to this to prepare the cell-free extract. The template DNA can be added to it
immediately to express the target protein. Alternatively, the cell-free extract can
be freeze-dried as pellets for storage or transportation along with lyophilized
DNA constructs. These can be rehydrated through the simple addition of water
for the synthesis of target proteins. [179]
On the other hand, reconstituted systems are well defined, and are prepared
using a ”bottom-up” approach where the factors essential for protein synthe-
sis such as the purified enzymes, tRNAs, ribosomes (either prokaryotic or eu-
karyotic), amino acids, NTPs, and other energy molecules are reconstituted into
a system. The PURE (Protein synthesis Using Recombinant Elements) system
developed by Shimizu et al. was one of the first reconstituted cell-free sys-
tems [142].
3
Both classes of cell-free systems tend to have their own set of benefits and
drawbacks. Crude cell extracts tend to be less expensive and have higher yields
than reconstituted systems [46]. They also offer more complex metabolic capa-
bilities that can be exploited for energy regeneration, that allow for larger-scale
reactions and extend the duration of protein synthesis [56, 160].
Reconstituted systems also have their own advantages. For one, since the
reconstituted systems are prepared from individual components, their exact
composition is known and they can be used for studying biological processes
including protein expression and folding in the context of a completely defined
system. For instance, Li et al. showed that the efficiency of the PURE system
could be improved up to 5 fold by adding or adjusting a variety of factors that
affect transcription and translation, such as elongation factors, ribosome recy-
cling factors, release factors, chaperones, BSA and tRNAs. showed that the ef-
ficiency of protein synthesis was limited by translation elongation capacity, ri-
bosome release, and ribosome recycling [90]. Reconstituted systems such as the
PURE system can also be reconstituted in such a way that they do not contain
proteases and nucleases, to further improve protein production [142, 160].
4
1.3 Applications of Cell-free platforms
Figure 1.2: Cell-free systems and their applications. The left-hand side of the
image indicates the components needed for a cell-free system and the right-
hand side indicates the various fields in which cell-free systems can be applied.
[170]
Cell-free systems have found many roles in synthetic biology ranging from
portable diagnostics to fundamental discovery and prototyping. Cell-free sys-
tems have been very useful for the development of biosensors because of their
unique advantages over cell-based sensors. They can be used to detect cell wall-
impermeable or cytotoxic analytes, and they are immune to issues like muta-
tions and plasmid loss [149]. Cell-free biosensors can be lyophilized (freeze-
dried) which allows them to maintain their activity for months for use in
5
portable diagnostics. They can then be transported to the desired location, rehy-
drated with water for field use [121,160]. Low-cost cell-free paper-based biosen-
sors have been utilized for the detection of viruses such as norovirus, Ebola
virus, and Zika virus [97, 121, 122]. Hamada et al. made use of cell-free plat-
forms for the bottom-up construction of dynamic biomaterials with emergent
locomotion behavior powered by artificial metabolism. These could have appli-
cations in pathogen detection, and as scaffolds for hybrid nanomaterials [51].
Cell-free systems have a huge scope for both small-scale prototyping of bio-
logical processes and larger-scale bioengineering efforts. Because of their abil-
ity to allow for direct manipulation of the chemical environment, prototyping
is more efficient in cell-free systems. They can also be used to prototype in-
dividual genetic parts such as promoters or complex genetic designs in vitro
before implementing them in vivo. Moreover, the genetic constructs need not
be assembled into plasmids as cell-free systems even allow for the use of lin-
ear DNA [149]. Cell-free systems can also be used to analyze how individual
genetic parts function together in synthetic genetic circuits by considering each
individual part as a logic gate. A number of cell-free genetic circuits have been
assembled and prototyped using this strategy including sigma factor-based cas-
cades, RNA transcriptional cascades, RNA single input modules, and feed-
forward/feedback loops [2, 42, 118, 162, 163].
Developments in experimental setup and analysis techniques such as the
use of real-time fluorescent probes for mRNA measurement, automated liquid-
handling robotic systems for rapid screening of cell-free reactions, microflu-
idic devices, and cell-sized droplet-based expression have all improved the effi-
ciency of prototyping in cell-free systems [25, 106, 107, 115, 133, 149].
6
1.3.1 Production of Biologics and Specialized Proteins
CFPS can be utilized for the production of biologics and other specialized pro-
teins such as membrane proteins [44,93,113,145,160,185]. The ease of rapid par-
allel assay and screening with cell-free systems makes them important for the
production of libraries for protein engineering and post-genomic research [164].
Recently, cell-free systems have also been used to achieve proteins with complex
post-translational modifications such as N-linked glycoproteins [47]. A single-
pot glycoprotein synthesis system has also been developed by Jaroentomeechai
et al. that allows for on-demand biomanufacturing of glycoproteins [65]. Even
proteins incorporating noncanonical amino acids have been synthesized using
CFPS systems [6, 99].
By constructing cell-free systems in microfluidic devices, we can achieve
point-of-care (POC) production of proteins. Continuous flow microfluidic re-
actors offer more precise control over mixing when compared to batch reactors
[53,102]. Georgi et al. have developed an automated, microfluidic CFPS reactor
called the TRITT (Transcription - RNA Immobilization and Transfer - Transla-
tion) platform that runs both transcription and translation reactions in separate
compartments [43]. The quasi-continuous transfer of mRNA to the translation
chamber allows for longer CFPS reactions and better protein yield. Microfluidic
platforms can also be constructed with integrated purification methods such as
dialysis and affinity chromatography [107, 114].
7
1.3.2 Metabolic Engineering
Cell-free systems possess many advantages over in vivo systems for metabolic
engineering applications. The cells in vivo systems have their own objec-
tives such as growth or maintenance, which uses up more resources and
drives metabolic flux away from the desired pathways making it challenging
to achieve high flux in those pathways. On the other hand, in cell-free systems,
the allocation of resources to the production of desired products can be maxi-
mized [48]. The complexity of living cells also makes computational modeling
and optimization of metabolic flux difficult [32]. Cell-free systems, on the other
hand, can be accurately modeled and the reaction environment can be directly
manipulated and tuned according to the bio-synthetic needs.
Metabolic engineering in cell-free systems has been used to increase flux
through enzymatic pathways and improve product yield for the production
of high-value small molecule products such as cannabinoids, polyhydroxybu-
tyrate bioplastic, isobutanol, ethanol, limonene, etc. [8,33,49,82,120,174]. It has
also been used to handle bottlenecks in CFPS such as the need for energy and
cofactor regeneration in the system [14, 18, 68, 77].
1.4 Mathematical Modeling of Cell-free systems
There have been several mathematical models of CFPS. Most of these models
focus on the TXTL processes and are mostly systems of ordinary differential
equations (ODEs) based on saturation or Hill-like equations.
Karzbrun et al. derived a coarse-grained model comprising the entire en-
8
dogenous transcription, translation, as well as mRNA and protein degradation
machinery for an E.coli cell-free extract [74]. To simplify calculations, this model
was based on four enzymes and ten relevant rate constants. Transcription and
translation processes were assumed to follow Michaelis–Menten kinetics. They
assumed that the protein expression was governed by zeroth-order degradation
and noted that the expression followed a sharp transition from undetectable lev-
els to constant-rate accumulation, without reaching steady state. However, their
study only focused on the first hour of the reaction as the protein synthesis rate
of their system began to exponentially decay after that. This sudden decay was
assumed to be due to resource depletion and waste accumulation.
Stögbauer et al. developed a model that accounted for resource consumption
and degradation [157]. The model predicts that the protein yield depends on the
experiment timing in addition to template DNA concentration. They attempted
to use Hill functions to better predict the observed saturation effects of mRNA
and protein, but the optimized Hill coefficients were close to one, effectively
reducing the Hill equation to Michaelis–Menten equation. They identified ribo-
somal degradation as the cause for the cessation of cell-free protein synthesis in
the specific reconstituted system that they used.
More recently, Neiß et al. developed a comprehensive experimentally val-
idated model that was able to identify the major limiting factors of cell-free
protein synthesis - the supply of ternary complexes consisting of EFTu and
tRNA [116]. This model used an unusual hybrid black-box approach: Although
the translation processes were well detailed, the transcription processes were
simplified. The entire model was a large system of differential algebraic equa-
tions - a system of eight algebraic equations and over 400 ODEs. Gyorgy and
9
Murry’s model explicitly included competition for shared resources in genetic
networks [50]. This model also predicted possible product concentrations in
multiple-protein expression systems. The authors illustrated that resource com-
petition was a key consideration in genetic circuit design.
Genetic circuits based on regulation by RNA have also been explored and
mathematical models for such systems have been developed for cell-free plat-
forms. Transcriptional regulatory RNAs are of interest because they bypass the
need for regulatory proteins and they have been used to create various logic
gates and cascades [17, 21, 60, 94]. Hu et al. developed an experimentally val-
idated model of a synthetic RNA circuit that contained 8 ODEs and 13 previ-
ously unknown parameters which were estimated from sensitivity analysis [60].
Although these models of transcription, translation, resource competition, and
gene regulatory circuits have provided useful information for optimizing cell-
free biomanufacturing, more sophisticated models integrating metabolic path-
ways are necessary.
1.4.1 Metabolic Modeling
Traditional approaches to metabolic modeling, which were first developed to
describe living cells, could also be applied to cell-free systems. Mechanistic in
vivo metabolic models arose from the desire to predict microbial phenotypes
resulting from changes in intracellular or extracellular states [38]. The Shuler
lab constructed large-scale, dynamic metabolic models for single-cells that in-
corporated multiple regulated catabolic and anabolic pathways constrained by
experimentally determined kinetic parameters [31, 156, 192], minimal cell archi-
10
tectures [19], and DNA sequence-based whole-cell models of E. coli [10]. Next-
generation metabolic models described cellular processes such as RNA synthe-
sis, chromosome synthesis, and regulated catabolic and macromolecular syn-
thesis pathways in detail using ODEs [171]. Karr et al. even developed a whole-
cell computational model of the life cycle of the human pathogen Mycoplasma
genitalium that includes all of its molecular components and their interactions
inside the cell.
However, traditional metabolic modeling approaches are often complex and
nonlinear and require the estimation of a large number of unknown param-
eters. This is a difficult process because of the inherent noisiness of biolog-
ical data and the computational complexity involved in the repeated solving
of the model equations. To overcome such obstacles, constraint-based meth-
ods were developed to describe metabolic networks with only a limited need
for kinetic parameters [175]. Constraint-based approaches such as Flux Bal-
ance Analysis (FBA) have been used for the stoichiometric reconstructions of
metabolism [88]. FBA, Metabolic Flux Analysis (MFA), and convex network
decomposition approaches such as elementary modes and extreme pathways
have been developed to model intracellular metabolism using the biochemi-
cal stoichiometry and other constraints such as thermodynamic feasibility un-
der pseudo-steady-state conditions [52, 55, 137, 139, 189]. Such constraint-based
approaches use linear programming to predict productivity, mutant behavior,
and growth phenotypes for biochemical networks of even genome-scale net-
works [34, 119, 135, 175]. The objective function used in such constraint-based
approaches in a cell-free system is the maximization of the desired protein yield.
These stoichiometric reconstructions have been expanded to even include
11
the metabolic demands for protein synthesis. One of the earliest such mod-
els were the sequence-specific constraint-based models developed by Allen and
Palsson, which depended on the DNA and protein sequences of interest and
the TXTL processes were integrated with metabolism [7]. Since then, sequence-
specific constraint-based models have been expanded to the genome-scale with
detailed descriptions of gene expression (ME-model) and protein structures
(GEM-PRO) [7, 20, 88, 119, 201]. These developments have greatly expanded
the scope of stoichiometric reconstructions. Constraint-based methods can be
potentially used to predict non-intuitive strategies to optimize the interaction
between metabolism and gene expression in cell-free platforms. Thus, the
use of integrated constraint-based models for cell-free optimization studies is
a promising future research direction.
12
1.4.2 Integrated cell-free models
Figure 1.3: The integration of metabolism with transcription and translation
(TXTL) processes: The TXTL processes utilize the macromolecular precursors
from metabolism (such as NTPs, amino acids, and cofactors) for gene expres-
sion. The integrated framework can be represented as a stoichiometric matrix
of metabolites participating in certain reactions, along with a description of the
metabolic demands for protein expression. By applying a pseudo-steady-state
assumption, making use of various constraints and an objective function, the
metabolic fluxes in the system can be estimated. In the case of cell-free systems,
the objective function is the maximization of target protein yield. [177]
Although CFPS also depends on central carbon metabolism and other metabolic
pathways and doesn’t just rely on TXTL processes, very few mathematical mod-
els have managed to integrate cell-free TXTL processes with metabolic path-
ways [24, 59, 181]. Wayman et al. developed a hybrid cell-free modeling ap-
proach of Wayman that integrated kinetic modeling with a rule-based descrip-
tion of allosteric control [186]. Horvath et al. developed an ensemble of dynamic
E. coli CFPS models on top of this, using parameters estimated from measure-
ments of metabolite, amino acid, and protein concentrations from CFPS reac-
tions conducted using the PANOx-SP cell-free system [59]. By simulating re-
action group knockouts, they suggested that cell-free metabolism and protein
13
synthesis were strongly coupled with oxidative phosphorylation and glycolytic
flux. Vilkhovoy et al. developed an experimentally validated constraint-based
model of CFPS [181] to circumvent computationally expensive parameter esti-
mation. This model integrated the expression of a model protein product with
the supply of metabolic precursors and energy (Figure 1.3). This model coupled
TXTL processes with the available resources using only six adjustable parame-
ters. Model analysis suggested that protein expression in the PANOx-SP system
was translationally limited. Further, the same modeling approach was also able
to describe protein expression in the myTXTL system using only a limited num-
ber of experimentally derived parameters. Taken together, the incorporation of
complex metabolism with genetic regulatory networks using constraint-based
modeling is a promising approach to simulate cell-free systems.
14
CHAPTER 2
DEVELOPMENT OF A GLUCONATE-RESPONSIVE TRANSCRIPTION
FACTOR-BASED GENETIC CIRCUIT IN A RECONSTITUTED CELL-FREE
SYSTEM
2.1 Introduction
Diabetes mellitus is a group of metabolic diseases that can be characterized by
chronic hyperglycemia occurring due to imperfect insulin synthesis, insulin ac-
tion, or both [76]. Diabetes is increasing at an alarming rate globally. According
to International Diabetes Federation, about 463 million adults were living with
diabetes in 2019 and this number is expected to grow to 700 million by 2045 [1].
There are two main classes of the disease: (i) Type 1 diabetes occurs due to the
destruction of β cells of the pancreas causing the body to produce little to no
insulin [26, 30]. (ii) Type 2 diabetes is the more prevalent one where the body
doesn’t utilize the insulin well and may also not produce insulin [76].
Insulin is a hormone necessary to prevent blood sugar levels from rising too
high (hyperglycemia) by allowing adipose tissues to take up glucose. Glucagon
is a counter-regulatory hormone for insulin that stops blood sugar levels from
dropping too low (hypoglycemia) by breaking down glycogen to produce glu-
cose [70]. Cell-free systems can be engineered to create an input-responsive,
on-demand, protein synthesis system. In this work, we propose a genetic glu-
cose sensor circuit to direct the downstream synthesis of insulin or glucagon.
This circuit can be used for POC treatment of diabetes.
A number of effective metabolic engineering strategies have made use of
15
the advances in promoter engineering to dynamically regulate the expression
of pathway enzymes to optimize product yields [28, 58, 80, 91]. Allosteric tran-
scription factors (aTFs) are a core component of this. An aTF is a protein that
binds to the operator sequence of a DNA and prevents transcription. Upon
binding to a specific ligand, the aTF undergoes a conformational change that al-
ters its affinity for an operator DNA sequence [167]. Ligand binding leads to the
release of the aTF from the operator, allowing the RNA polymerase holoenzyme
to effectively bind to the promoter and increase the transcription rate. aTFs have
been co-opted for use as gene expression switches [96].
Figure 2.1: In this image, the first box indicates the components involved in
an aTF regulated circuit: the RNA polymerases, aTFs, and DNA transcription
templates. The second box indicates the circuit in the absence of the ligand. In
the presence of the ligand, the aTF will no longer bind to the operator sequence
and allows expression of the downstream reporter as shown in the third box.
[72]
More recently, aTFs have been increasingly leveraged for use in POC biosen-
sors. CFPS is a favorable platform for implementing biosensors because of its
portability, ability to manipulate the reaction environment in real-time, and tol-
erance to cytotoxic environments. aTFs can be used to ’sense’ a specific molecule
(the ligand that suppresses it) by altering the gene expression of a downstream
reporter protein. By monitoring the signal from the reporter, it would be pos-
sible to determine the amount of ligand present in the system. Such biosensors
have recently been used for detecting ions, biomarkers, and antibiotics (See Fig-
ure 2.1) [72, 153, 182].
16
The ability to sense glucose concentration in the bloodstream and direct the
synthesis of either insulin or glucagon (or a combination of both) via the use of
genetic circuits is very attractive, and given the advanced state of cell-free sys-
tems, it is an achievable feat. Toward this aim, it is important to first construct a
transcriptional unit that can effectively respond to varying levels of glucose or
its derivatives that can act as its proxy.
We aim to design a biosensing system that has the ability to sense glucose
concentration in the bloodstream and direct the synthesis of either insulin or
glucagon hormone (or a combination of both). In order to achieve this, we need
to construct a transcriptional unit that can effectively respond to variation in the
concentration of glucose or its derivatives (such as D-gluconate) that can act as
its proxy.
We plan to study the effectiveness of aTFs towards this purpose. An aTF
has a DNA binding region at the N-terminal and a ligand-binding region at the
C-terminal. Studies on the Gluconate operon transcriptional repressor (GntR)
family of E. coli and B. subtilis have shown that two derivatives of glucose, D-
glucono-1,5-lactone and D-gluconate, can bind to GntR protein at the C-terminal
and as an anti-repressor [39, 124, 130, 148, 173]. Since our biosensor should be
able to sense the presence of glucose in the system we decided to make use of
the GntR family in the biosensor.
In this work, we aimed to first construct a promoter (incorporating the op-
erator sequence of GntR) that can be efficiently used in an E. coli-based CFPS.
Toward this aim, we first cloned the GntR operator sequence to a bacteriophage
lambda promoter (P70) developed by Noireaux and coworkers [143]. Next, we
added a reporter protein gene downstream of this modified promoter to easily
17
characterize the promoter dynamics.
Earlier work done by my colleague, Abhinav Adhikari in the VarnerLab
made use of a deGFP reporter constructed with an mP70 promoter. This work
studied the repression and de-repression performance of GntR in the myTXTL
system. However, the D-gluconate was consumed by the metabolism in the
myTXTL system resulting in poor de-repression. So in this work, we plan to
make use of the PURE cell-free system which lacks the metabolic capability to
unintentionally consume D-gluconate. The PURExpress system also makes it
easier to purify the protein products. Since all the components of the system
(other than the ribosomes) are his-tagged, protein products in the native form
(with no His tag) can be purified easily and rapidly through ultrafiltration (to re-
move the ribosomes) followed by the use of Ni-NTA affinity column (to remove
the His-tagged components).
In addition to this, we also changed the reporter protein. We use the Yellow
Fluorescent Protein, Venus, instead of deGFP because Venus is a much brighter
reporter protein in this system, the time it took to reach maximum growth was
similar to different mutated GFP variants [87].
The CFPS reactions to express Venus were carried out either in the absence
or presence of the GntR repressor gene. We characterized the repression and de-
repression characteristics (using D-gluconate) in the PURExpress reconstituted
system based on PURE. Taken together, our results show that the GntR protein
effectively represses the expression of the Venus reporter protein, and that the
addition of D-gluconate is able to inhibit this repression.
18
2.2 Materials and Methods
2.2.1 mP70 promoter construction
The first step in constructing the promoter responsive to D-gluconate levels was
to include the operator sequence for the E.coli GntR in the promoter region.
The sequences of the operator and the repressor were obtained from the litera-
ture [124]. The 16 bp E. coli GntR operator sequence (ATGTTACCCGTATCAT)
was added between the -35 and -10 regions of mutated bacteriophage lambda
promoter (P70) developed by Noireaux and coworkers [143]. This modified P70
promoter with added GntR operator site was named mP70. The -10 binding site
of the RpoD holoenzyme was altered from GATAAT in the original promoter to
TATCAT in the mP70. The gene sequence of the reporter protein, Venus, was
added downstream of this modified promoter. On the other hand, the GntR
repressor gene was cloned downstream of the original P70 promoter. The ter-
minator and 5’ UTR for the DNA with the mP70 promoter were identical to that
of the DNA with the P70 promoter. The full constructs were ordered as linear
DNA fragments, with 150-200 bp flanker sequences on both ends, from Twist
Bioscience and Integrated DNA Technologies.
2.2.2 Synthetic Circuits Architecture
The three genetic circuits (C1, C2, and C3) used in this study were based upon
the bacterial sigma factor regulatory system. Sigma factor 70 (σ70) was the pri-
mary driver of each circuit. The E.coli RNA Polymerase Holoenzyme is the core
enzyme saturated withσ70 and allows for RNA synthesis fromσ70 specific pro-
19
moters. In C1, σ70 induced Venus (YFP) expression was explored in the absence
of additional regulators (Figure 2.2(a)). In C2, σ70 induced the expression of
GntR and Venus (Figure 2.2(b)). The GntR protein repressed the mP70 promoter
of Venus, thereby down-regulating Venus transcription. In C3, D-gluconate was
introduced into the circuit (Figure 2.3). D-gluconate suppresses the binding of
GntR to the mP70 promoter. This de-represses the Venus transcription. Study-
ing C1 allows us to estimate parameters governing the interaction of σ70 with
the mP70 promoter. Whereas, the C2 allows us to characterize the strength of
the transcriptional repression by GntR. Finally, C3 allows us to characterize the
de-repression effect D-gluconate has on GntR.
(a) (b)
Figure 2.2: (a) C1 Circuit: σ70 induces Venus expression (b) C2 Circuit: σ70
induces the expression of GntR and Venus. The GntR protein once expressed
represses the mP70 promoter of Venus.
20
Figure 2.3: C3 circuit: σ70 induces the expression of GntR and Venus. The
GntR protein once expressed represses the mP70 promoter of Venus. In the
presence of a high concentration of D-Gluconate, the binding of GntR to mP70
gets repressed thereby de-repressing Venus transcription
2.2.3 Cell-Free Protein Synthesis Reactions
The cell-free protein synthesis (CFPS) reactions were carried out using the PUR-
Express In Vitro Protein Synthesis Kit (New England Biolabs Inc) in 120 µL 384-
well plates (Thermofisher NUNC, flat-bottom) in Varioskan Lux plate reader
at 37°C. The working volume of all the reactions was 25 µL, composed of the
PURE solutions A and B (17.5 µL), RNase inhibitor, Murine (0.5 µL), NEB E.coli
RNA Polymerase holoenzyme (2 µL), the linear DNA: Venus (7 nM), GntR (10
nM). The E. coli RNA Polymerase, Holoenzyme is the core enzyme saturated
with sigma factor 70 and initiates RNA synthesis from sigma 70 specific pro-
moters. The full constructs were ordered as linear DNA fragments, with 150-
200 bp flanker sequences on both ends, from Twist Bioscience and Integrated
DNA Technologies. For the control reactions without GntR, an equal volume
of nuclease-free water was used in its stead. For the de-repression reactions,
21
D-gluconate (10 mM) was added while maintaining the same total volume.
2.2.4 mRNA Quantification
Following each CFPS run, the total RNA was extracted from 5 µL reaction mix-
ture using PureLink RNA Mini Kit (Thermo Fisher Scientific) and stored at
−80°C. The quantitative RT-PCR reactions were done using Applied Biosystems
TaqMan RNA-to-CT 1-Step Kit and Custom TaqMan Gene Expression Assays
(Thermo Fisher Scientific). An mRNA standard curve was used to determine
absolute mRNA concentrations for each of the samples. The mRNA standards
were prepared as follows: separate CFPS reactions for 5 nM of DNA (Venus and
GntR) were carried out for 2 hr. Total RNA was extracted using the full reac-
tion volume. This was followed by the removal of 16S and 23S rRNA using the
MICROBExpress Bacterial mRNA Enrichment Kit (Life Technologies Corpora-
tion). Lastly, the MEGAclear Kit (Life Technologies Corporation) was used to
further purify the mRNA. The mRNA concentrations were determined using
the Qubit™ RNA assay kit (ThermoFisher Scientific). At least three technical
replicates were performed for each standard.
2.2.5 Protein Quantification
Venus (Yellow Fluorescent Protein) fluorescence was measured using the Var-
ioskan Lux plate reader at 513 nm (excitation) and 531 nm (emission). The flu-
orescence was measured with 25 µL for each of the mixtures. For all measure-
ments, at least three biological replicates were performed. A protein standard
22
curve was used to determine the absolute protein concentrations for each of the
samples. The standard curve was prepared by using Venus standards of known
Relative Fluorescence Units (RFU). An SDS-PAGE of these standards along with
a Protein ladder of samples with known concentrations was performed at 300V
until the dye fronts reached the bottom of the gel. The SDS-PAGE was per-
formed in a Mini-PROTEAN Tetra Cell (Bio-Rad Laboratories) using Any kD™
Mini-PROTEAN® TGX Stain-Free™ Protein Gels, 15 well, 15 µl gels (Bio-Rad
Laboratories). This was followed by staining the gel with Bio-Safe™ Coomassie
Stain (Bio-Rad Laboratories). The stained gel was visualized using the Chemi-
Doc MP and then the concentrations were measured using Image Lab software.
2.2.6 Estimation of model parameters
To model the circuit, the mRNA and protein balances for mRNA mi and protein
pi are described using the following balance equations
ṁi = rX,iui − θm,imi (2.1)
ṗi = rL,iwi − θp,i pi (2.2)
where rX and rL are the transcription and translation rates and θm and θp are the
respective degradation rates. Here, wi and ui are the control functions modeled
as: ( )
− ln(2)ẇi =∑ wi (2.3)τL,1/2∑ j∈χ W j f j(...)ui = (2.4)
i∈C Wi fi(...)i
Wi denotes the weight of configuration i and fi is the binding function modeled
as a hill function. The rate of transcription rX, ju j was modeled as the product of a
23
kinetic limit, rX, j, and a control term u j ∈ [0, 1]. The kinetic limit of transcription
was derived from elementary reactions leading to the formation of m j, as previ-
ously done by McClure [101]. The kinetic limit of transcription was formulated
as:  G jr = Vmax  ( ) X, j X j 

, (2.5)
τX, jKX, j + τX, j + 1 G j
where the initiation time constant τX, j, and saturation constant KX, j were exper-
imentally measured by McClure [101]. The maximum transcription rate, VmaxX, j ,
was given by: [ ( )]
Vmax
v̇X
X, j = RX,T (2.6)lG, j
where RX,T is the RNA polymerase concentration (nM), v̇X is the RNA poly-
merase elongation rate (nt/h) and lG, j is the length of gene j in nucleotides (nt).
By analogy, the kinetic limit of translation was formulated as:
max  r V  x( mRNAL, j = L j ) , (2.7)
τL, jKL, j + τL, j + 1 xmRNA
where x is the mRNA concentration and VmaxmRNA L, j is the maximum translation
rate which was formulated as: [ ( )]
Vmax
v̇L
L, j = KPRL,T (2.8)lP, j
The model previously used by Adhikari [2] was extremely effective in capturing
the expression dynamics of mRNA and protein in synthetic cell-free circuits, but
it suffered from one major drawback. The parameters related to the Gibbs free
energy of binding between RNAP and other genes in the systems were over-
estimated. This was primarily due to the lack of literature that provides the
constraints for ∆G. Since the Gibbs free energy of binding drives the weight of
a particular configuration, any overestimation will lead to a high background
expression. This will affect the transcription rate which will thereby affect the
24
translation rate and the protein expression. Therefore, to address this issue,
we revised the model by changing the way the control function for transcrip-
tion is formulated. A similar modeling approach as previously mentioned was
adopted but the control functions w∑ere formulated as -∑ j∈χ W j f j(...)ui = (2.9)
i∈C W f (...)i i i
where W was formulated as- ( )
−∆E
W ii = exp (2.10)kbT
This had a two-fold advantage. Firstly, there were literature estimates available
for difference in binding energy (∆E) from Bintu and coworkers [12], in refer-
ence units, as multiples of Boltzmann constant (kb) and temperature (T). More-
over, this formulation was much closer to the statistical mechanics approach we
have undertaken to represent the configurations.
25
2.3 Results and Discussion
Figure 2.4: Venus protein expression without GntR (C1 - Red), Venus repression
by GntR (C2 - Blue), and de-repression by D-gluconate (C3 - Green). The shaded
regions represent one standard deviation of the experimental measurements
We first tested the performance of linear DNA with the mP70-Venus construct
(7 nM) in the PURExpress system (C1 circuit) in order to obtain a standard
for comparison for the repression and de-repression characteristics. In order
to characterize the repression by Gntr, we made use of the C2 circuit. We ex-
pressed two linear DNA fragments: E. coli GntR (10 nM) and mP70-Venus (7
nM). For the control case, the GntR DNA was excluded from the reaction mix-
ture. While the Venus concentration in the control case steadily increased, it
reached a steady plateau in approximately 4 hours in the +GntR reaction (well
before the -GntR control plateaus) and had a significant reduction in Venus ex-
pression. This showed that GntR was indeed repressing Venus expression.
26
In order to test the de-repression characteristic of GntR, we added 10mM
of D-gluconate (approximately 9 times the value of its KD established in litera-
ture [23]) along with GntR and Venus (C3 circuit). Although the Venus concen-
tration was not as high as in the case of the C1 circuit, it was appreciably higher
compared to when D-gluconate was not added (C2 circuit).
2.3.1 Parameter Estimation
(a) (b)
Figure 2.5: Model simulations versus experimental measurements for σ70 in-
duced Venus expression for the C1 circuit. (a) Simulated and measured Venus
mRNA concentration versus time. (b) Simulated and measured Venus protein
concentration versus time. The black points indicate the experimental values
while the black curve is the predicted curve. The shaded region denotes the
95% confidence interval of the predicted curve. The error bars denote one stan-
dard deviation of the experimental data
Parameter estimation for the C1 circuit was set up as a multi-objective opti-
mization problem with two objectives. One objective was the minimization of
the error between experimental and simulated values of Venus mRNA, and the
other for the corresponding Venus protein measurements. The algorithm was
run for 20 generations and produced an ensemble of 187 parameter sets, of the
27
11 unknown model parameters. The criterion for a parameter ensemble selec-
tion was a Pareto rank of 2 or below. The parameters generated for the C1 circuit
were able to create a good fit with the experimental data (Figure 2.5).
(a) (b) (c)
Figure 2.6: Model simulations versus experimental measurements for the C2
circuit. (a) Simulated and measured Venus mRNA concentration versus time.
(b) Simulated and measured Venus protein concentration versus time. (c) Sim-
ulated and measured Gntr mRNA concentration versus time. The black points
indicate the experimental values while the black curve is the predicted curve.
The shaded region denotes the 95% confidence interval of the predicted curve.
The error bars denote one standard deviation of the experimental data
In the case of the C2 circuit, the multi-objective optimization problem had
three objectives: Minimization of error between experimental and simulated
values of (i)Venus protein (ii) Venus mRNA, and (iii) GntR mRNA. The algo-
rithm was run for 10 generations to estimate 22 unknown model parameters
and estimated an ensemble of 565 parameter sets that had a Pareto rank of 2
or below. Although less effectively than for the C1 circuit, the model was able
to capture the mRNA concentration profile for Venus fairly well. The model
accurately captured the mRNA concentration profile of GntR and protein ex-
pression of Venus. Taken together, the model is able to simulate the repression
of mP70-Venus in the presence of GntR in a PURE system.
28
2.3.2 Future Directions
The D-gluconate levels at the different time points can be measured by Liq-
uid Chromatography/Mass Spectrometry (LCMS) using our lab’s published
derivatization protocol [180]. D-gluconate acts as a dynamic second messen-
ger for D-glucose concentration in myTXTL cell-free extracts. We used the
gluconate-responsive transcription factor GntR as the key transcription element
for the glucose sensor circuit. Thus, information about the background D-
gluconate concentration, and its temporal evolution in response to changes in
glucose concentration, is critical to the glucose-sensing capability of the pro-
posed circuit. We also need to construct a dose-response curve to study the
de-repression of GntR in the circuit at different concentrations of D-gluconate.
This would be necessary to better optimize the switch behavior of the circuit.
Secondly, in order to fully characterize repression and de-repression, we
plan to use different concentrations of the DNA construct with the mP70 pro-
moter and the GntR protein to perform Electrophoretic Mobility Shift Assays
(EMSA) in order to determine the binding parameters between Gntr protein and
mP70. We also plan to do end point measurements of the reporter fluorescence
for these concentrations. We also plan to implement real-time measurement of
mRNA to get more mRNA measurements making it easier to develop accurate
models [115].
Thirdly, we also plan to use different GntR constructs derived from P. aerug-
inosa and B. subtilis and test their effectiveness in repressing reporter fluores-
cence. The P. aeruginosa construct was shown to be partly responsive to D-
glucose, eliminating the need for a proxy, D-gluconate [23]. We are also in the
process of testing the effectiveness of a circuit which uses promoters recognized
29
by T7 polymerase instead of the E. coli RNAP. The T7 polymerase provides a
two-fold advantage over the E. coli RNAP: It synthesizes RNA at a much higher
rate [161]; It is inherently present in the PURE system and doesn’t need to be
added separately [142].
Fourthly, we plan to develop a mathematical model for the C3 circuit to take
into effect the presence of D-gluconate in the system to effectively characterize
transcription and translation behavior. We would have to estimate the model
parameters using multi-objective optimization similar to how they were deter-
mined for C1 and C2 circuits. Such a model would also help guide our future
experiments based on this system.
Lastly, we also plan to further engineer the GntR protein (and the promoter
region) to enable more efficient repression and derepression properties. Muta-
tions in the C-terminal region of GntR have been shown to affect its ability to
bind to D-gluconate while mutations in the N-terminal region affect operator
binding properties [196, 197]. On the basis of these details, we plan to gen-
erate a library of GntR protein with random mutations in the C-terminal and
N-terminal regions via error-prone PCR in order to discover GntR mutants that
have good operator binding strength, while still being effectively antagonized
by D-gluconate.
Once we have accomplished these, we can test the circuit with fluorescent
labeled insulin and glucagon. The ultimate aim is to make use of this transcrip-
tional unit to construct a viable glucose biosensor (Figure 2.7) that is capable
of producing insulin at high glucose (gluconate) levels and glucagon at high
glucose (gluconate) levels.
30
Figure 2.7: Schematic of a glucose biosensor: σ70 induces the expression of
GntR, σ28, insulin and α-σ28 which is downstream to insulin. σ28 induces ex-
pression of glucagon while α-σ28 inhibits the P28 promoter thereby inhibiting
the expression of glucagon. When the gluconate levels are high, insulin will
be expressed along with σ28. At low gluconate levels, only glucagon will be
expressed
31
CHAPTER 3
I(NSP)ECTION OF THE SARS-COV-2 NSP1’S HOST-TRANSLATION
EFFECT IN A CELL-FREE SYSTEM
3.1 Introduction
The SARS-CoV-2 virus which is the cause of the unprecedented Covid-19 pan-
demic belongs to the class of coronaviruses. Coronaviruses are enveloped
viruses with a positive-sense single-stranded RNA genome. This mRNA con-
tains a 5’ cap structure along with a 3’ poly (A) tail allowing it to serve as the
genetic material as well as the messenger RNA which translates to the pro-
teins [36, 183].
The SARS-CoV-2 has a genome of ∼30 kb. It possesses 14 Open Reading
Frames (ORFs) encoding for a total of 27 known proteins [75,78,131]. The over-
lapping ORFs ORF1a and ORF1ab located at the 5’ terminal of the genome en-
code the pp1a and pp1ab polyproteins respectively. The translation of ORF1ab
is mediated by a -1 ribosomal frameshift at the overlap of these two ORFs
[37, 183]. The other ORFs are located at the 3’ terminal and encode for the 4
structural proteins (Spike protein (S), Small Envelope protein (E), Matrix pro-
tein (M), Nucleocapsid (N)) and 8 accessory proteins (3a, 3b, p6, 7a, 7b, 8b, 9b,
ORF14) [183, 184, 191].
The pp1a and pp1ab polyproteins together contain 15 functional non-
structural proteins (nsp) (nsp11 is a 13 amino acid cleavage product that isn’t
known to have any function [109, 159]). These nsps are released from pp1a and
pp1ab upon proteolytic cleavage by nsp3 and nsp5 [37, 183].
32
The non-structural protein 1 (nsp1), also known as the host shutoff factor, is
encoded by the gene closest to the 5’ end of the viral genome and is one of the
first proteins to be expressed after the virus infects a host cell [138,198]. It is one
of the most important nsps that is necessary for the propagation of the virus.
It inhibits host gene expression and helps the virus suppress the host immune
response [110, 159, 169, 183].
In other coronaviruses like the SARS-CoV and MERS-CoV, the nsp1 inhibits
host translation by forming a complex with the 40S ribosomal unit which ren-
ders this inactive for host translation [66,108,110,111,138,165,169,198]. Further,
this also causes a selective endonucleolytic cleavage in the 5’ untranslated re-
gion (UTR) of the host mRNA [61, 108, 169]. This promotes degradation of the
host mRNA while the viral mRNA remains intact.
The SARS-CoV-2 nsp1 has ∼84% amino acid sequence similarity with its
SARS-CoV homolog [134, 150, 159, 169]. Such high sequence similarity sug-
gests that they share common properties and functions. The SARS-CoV-2 nsp1
binds to the human 40S ribosomal subunit to shut off host translation. The
protein inserts its C-terminal domain into the mRNA entry channel of the ri-
bosome (at the 18S ribosomal RNA), where it interferes with the host mRNA
binding [11, 83, 138, 150, 169]. A stem-loop structure of SARS-CoV-2 mRNA
5’UTR has been observed that is recognized by nsp1. This recognition en-
ables selective translation of the viral mRNA while still inhibiting host trans-
lation [134,138,150,169]. Thus, nsp1 also helps to boost viral protein production
by hijacking more of the cellular machinery for translation of viral proteins [11].
This recognition also protects the viral mRNA from the possible mRNA degra-
dation activity of the nsp1 [61, 73, 165].
33
Type I interferon (IFN) induction and signaling represents one of the major
innate antiviral defense mechanisms and many viruses have developed strate-
gies to evade this immune response [54,110]. The expression of these IFNs trig-
gers innate antiviral immune responses which help to suppress the viral repli-
cation. Coronavirus infections are sensed by RIG-I (Retinoic acid–Inducible
Gene I)-like receptors (RLRs), which activate this IFN-based defense mecha-
nism [129, 150, 154, 169]. nsp1’s translation inhibition functions interfere with
the production of these proteins by shutting down the host translation machin-
ery. Thus by affecting the translation of antiviral defense factors such as IFN-β
or RIG-I, nsp1 effectively weakens the innate immune response [166, 169].
Its central role in the suppression of host translation and the immune re-
sponse makes nsp1 a viable target for the development of anti-viral drugs and
further studies. In this work, we aimed to produce the SARS-CoV-2 nsp1 using
E.coli-based CFPS. Secondly, we wanted to test the potential of nsp1 to inhibit
host translation in the same cell-free system (See Figure 3.1). And then math-
ematically model the effect of nsp1 on the host translation if it turned out to
demonstrate statistically significant inhibition.
Figure 3.1: Circuit representing coexpression of nsp1 along with a reporter pro-
tein. The nsp1 binds to the ribosomes and selectively prevents translation of the
reporter protein. This circuit doesn’t include the potential mRNA degradation
activity of nsp1 [61, 73, 165]. (Created with BioRender.com)
34
To this purpose, we designed a DNA construct that was ”reverse-translated”
from the sequence of the protein [200]. We used a P70 promoter sequence, along
with its 5’ UTR, as it is optimized to maximize expression in the E. coli-based
myTXTL Cell-free system [42]. We also carried out an assay to test the ability of
nsp1 to inhibit translation of dual emission Green Fluorescent Protein (deGFP)
(which would take the place of a host protein) in myTXTL extract.
3.2 Materials and Methods
3.2.1 Cell-Free Protein Synthesis Reactions
The cell-free protein synthesis (CFPS) reactions were carried out using the
myTXTL Sigma 70 Master Mix (Arbor Biosciences) in 1.5 mL Eppendorf tubes.
The working volume of all the reactions was 12.8 µL, composed of the Sigma 70
Master Mix (9 µL), GamS nuclease inhibitor (0.8 µL), and the nsp1 linear DNA
fragments (3 µL - 9.38nM). The full constructs were ordered as linear DNA frag-
ments, with 150-200 bp flanker sequences on both ends, from Twist Bioscience
and Integrated DNA Technologies. The CFPS reactions were incubated at 29°C.
The CFPS runs were carried out in triplicate for 2hr, 4hr, 8hr, and 16hr time
points. Following each CFPS run, the samples were stored at -80°C.
3.2.2 mRNA Quantification
Following each CFPS run, the total RNA was extracted from 5 µL reaction mix-
ture using PureLink RNA Mini Kit (Thermo Fisher Scientific) and stored at
35
−80°C. The quantitative RT-PCR reactions were done using Applied Biosystems
TaqMan RNA-to-CT 1-Step Kit and Custom TaqMan Gene Expression Assays
(Thermo Fisher Scientific). An mRNA standard curve was used to determine
absolute mRNA concentrations for each of the samples. The mRNA standards
were prepared as follows: CFPS reaction for 5 nM of nsp1 DNA was carried out
for 2 hr. Total RNA was extracted using the full reaction volume. This was fol-
lowed by the removal of 16S and 23S rRNA using the MICROBExpress Bacterial
mRNA Enrichment Kit (Life Technologies Corporation). Lastly, the MEGAclear
Kit (Life Technologies Corporation) was used to further purify the mRNA. The
mRNA concentrations were determined using the Qubit™ RNA assay kit (Ther-
moFisher Scientific). At least three technical replicates were performed for each
standard.
3.2.3 Protein Quantification
Following each CFPS run, 2 µL of each sample was extracted and added to
4 µL of water. And an equal amount of 2x Laemmli sample buffer (with 2-
mercaptoethanol as a reducing agent) was added to these. An SDS-PAGE of
10µL of each of these diluted samples along with a Protein ladder of sam-
ples with known concentrations was performed at 300V until the dye fronts
reached the bottom of the gel. This was followed by staining the gel with Bio-
Safe™ Coomassie Stain (Bio-Rad Laboratories). The SDS-PAGE was performed
in a Mini-PROTEAN Tetra Cell (Bio-Rad Laboratories) using Any kD™ Mini-
PROTEAN® TGX Stain-Free™ Protein Gels, 15 well, 15 µl gels (Bio-Rad Labo-
ratories). The stained gel was visualized using the ChemiDoc MP and then the
concentrations were measured using Image Lab software.
36
3.2.4 Coexpression of deGFP with nsp1
The cell-free protein synthesis (CFPS) reactions were carried out using the
myTXTL Sigma 70 Master Mix (Arbor Biosciences) in 120 µL 384-well plates
(Thermofisher NUNC, flat-bottom) in Varioskan Lux plate reader at 29°C. The
working volume of all the reactions was 14.1 µL, composed of the myTXTl ex-
tract (9 µL), GamS nuclease inhibitor (0.6 µL), the linear DNA: deGFP (3 µL -
4.26 nM), nsp1 (1.5 µL - 4.26 nM). Two sets of control reactions were also run.
In one, instead of nsp1, an equal volume of water was used. And in the other,
equal volume and concentration of nsp8 linear DNA construct was used instead
of nsp1. nsp8 was chosen as a control as it has a size similar to that of nsp1 but
doesn’t inhibit translation unlike nsp1 (though it could affect protein trafficking
to the cell membrane) [11]. The Cell-free reactions were carried out in dupli-
cates for 16hr. deGFP fluorescence was measured using the Varioskan Lux plate
reader at 488 nm (excitation) and 535 nm (emission) at 5 min time intervals. In
order to confirm the expression of these proteins, following the CFPS run, 2 µL
of each sample was extracted and diluted with 3 µL of water. Following this 5
µL of 2x Laemmli sample buffer was added to these and an SDS-Page of these
samples was performed at 200V until the dye fronts reached the bottom of the
gel. This was followed by Coomassie staining of the gel and visualization using
the Image Lab software.
3.2.5 Qualitative testing of nsp1 binding with E.coli Ribosomes
We ran a CFPS reaction of His6-tagged nsp1 in myTXTL. The working volume
of the reaction was 12.5 µL, composed of the Sigma 70 Master Mix (9 µL), GamS
37
nuclease inhibitor (0.5 µL), and the nsp1 linear DNA fragments (3 µL - 9.6nM).
The CFPS reactions were incubated at 29°C for 12hr. Then we purified the His6-
tagged nsp1 using Ni-NTA affinity column. Meanwhile, we separated the E.coli
Ribosomes from the PURExpress In Vitro Protein Synthesis Kit (New England
Biolabs Inc). Then these ribosomes were added to half the volume of the puri-
fied nsp1 proteins and allowed to mix well at 29°C for 15 minutes. Then this
sample was loaded into a gel along with pure nsp1 protein samples and a Pro-
tein ladder for an SDS-PAGE run. The ribosomes and its subunits are expected
to be retained at the top while the nsp1 should be able to pass through. After
the SDS-PAGE run, the gel was stained with Bio-Safe™ Coomassie Stain (Bio-
Rad Laboratories) and the stained gel was visualized using the ChemiDoc MP.
The concentration of nsp1 in each well was measured using Image Lab software
(and adjusted for the dilution of nsp1 in the channel with the ribosomes).
3.3 Results and Discussion
(a) (b)
Figure 3.2: Gene expression of nsp1. The points denote the mean, while the
error bars denote one standard deviation calculated from three replicates
38
We were able to successfully synthesize SARS-CoV-2 nsp1 in myTXTL extract.
With a starting concentration of 9.38 nM, we were able to obtain up to an aver-
age of 2.18 µM of nsp1 protein after a CFPS run of 16hr. The mRNA concentra-
tion peaked at about 1010 nM at the 4hr measurement point and continued to
decline after that (Figure 3.2).
(b)
(a)
Figure 3.3: Coexpression of deGFP with nsp1: (a) RFU measurements of deGFP
when coexpressed with nsp1 (Red). The controls: expression of deGFP alone
(Blue), expression of deGFP along with nsp8 (Green). The shaded regions indi-
cate one standard deviation (b) SDS-PAGE Analysis of these samples to confirm
the production of all the proteins. The gels have been Coomassie-stained
From Figure 3.3, we can observe that although coexpressing nsp1 along with
deGFP does reduce the synthesis of deGFP, there isn’t a very significant differ-
ence from just expressing deGFP by itself. In addition to that, coexpressing
deGFP with a control gene coding for the nsp8 protein with a molecular weight
(∼ 22.8kDa) just slightly higher than nsp1 (∼ 19.77kDa) [200] also reduces the
production of deGFP. This is despite the fact that the nsp8 doesn’t inhibit host
translation [11]. This decrease in the production of deGFP in presence of nsp1
or nsp8 can be better explained by a competition for the metabolic resources
caused by the introduction of their genes.
39
This also goes to prove that nsp1 doesn’t affect the prokaryotic ribosomes.
Unlike the larger mammalian 80S ribosome made up of 60S and 40S subunits,
the prokaryotic 70S ribosome is made up of 50S and 30S subunits. The nsp1
which binds to the mRNA entry channel of the 40S subunit doesn’t bind to
the corresponding region in the prokaryotic 30S subunit. Therefore, it doesn’t
actually inhibit translation in the E.coli based myTXTL system.
Figure 3.4: Rudimentary nsp1-70S Ribosome binding assay: The third channel
contains pure nsp1. The first and second channels contain nsp1 mixed with
two times its volume of purified 70S ribosomes solution. In other words, the
concentration of nsp1 in the third channel is thrice the concentration of nsp1
in the first two. Measurement of the fluorescence of the nsp1 bands show that
the normalized concentration of nsp1 in all wells are approximately the same
indicating that there is negligible binding of nsp1 with these ribosomes
We used a rudimentary binding assay to test the binding of nsp1 with E.coli
ribosomes. Ribosomes treated with nsp1 proteins were loaded on two channels
of a gel alongside a different channel containing pure nsp1. The channels are
expected to retain the ribosomes, their subunits, and any nsp1 bound to it at
the top but allow the free unbound nsp1 to pass through [4]. The results of
the SDS page show a negligible difference between the channel with just nsp1
and the channels with nsp1 and the ribosome. The channel with only nsp1 had
116.4 ng/µL of nsp1 while the ones where it was treated with ribosome had
40
an average of 106 ng/µL of nsp1 (See Figure 3.4). If nsp1 had bound to the
ribosome, we would have seen a significantly reduced concentration.
These results combined confirm that nsp1 doesn’t bind/shows negligible
binding to E.coli ribosomes. Conversely, this does prove that we can still synthe-
size nsp1 using myTXTL cell-free extracts and that this synthesis can be mod-
eled.
3.3.1 Future Directions
There are several directions for this work that can be pursued. First, in order
to better characterize this system, we can use the obtained data to fit the pa-
rameters for nsp1 gene expression (transcription and translation behavior) in
myTXTL using the Pareto Optimal Ensemble Technique. Second, in order to
better determine the binding between E.coli ribosomes and nsp1, we can make
use of better assays such as Sucrose gradient centrifugation can be used. Lastly,
we can repeat these experiments with a mammalian cell-free extract. The nsp1
should be able to bind to the 40S ribosomal subunits of the 80S mammalian
ribosomes present in these.
We could also analyze nsp1’s binding to ribosomes in cell-free systems based
on different mammalian cells such as Rabbit reticulocytes, CHO cells, and HeLa
cells [46]. It would also be interesting to study nsp1’s inhibition effects in cell-
free extracts prepared from human cell lines of different systems/organs. Cells
from human lungs are primarily affected by the SARS-CoV-2 virus infection,
so human lung cell lines should theoretically provide the best environment for
the growth of the virus [136, 141, 172]. We could compare lung-based cell-free
41
extracts from Calu-3 cells, A549 cells, or HF10B4 cells [103, 188], with other hu-
man cell-free extracts like HeLa cells, HEK 293 cells, and K562 cells [16,187,199],
to observe if there are any noticeable differences between nsp1’s effect in these
systems.
With these cell-free systems, we can also study the host translation inhibition
of nsp1 and create a mathematical model for the system. Since the 5’UTR of the
viral genome helps it bypass nsp1’s translation inhibition and is necessary for
viral replication, the nsp1 gene can be designed with this 5’UTR. The nsp1 can
be used as a global translation inhibitor to terminate the production of proteins
in a suitable cell-free system.
42
CHAPTER 4
MODELING THE EFFECT OF THE RNA-DEPENDENT RNA
POLYMERASE IN VITRO
4.1 Introduction
The SARS-CoV-2 is a coronavirus and like all coronaviruses, it is a positive-sense
single-stranded RNA virus ((+)ssRNA). In other words, its genome consists of
a single strand of positive-sense RNA (which goes in the 5’ to 3’ direction) [36,
169,183,184]. This RNA genome serves as the template for both translation and
RNA replication.
Coronaviruses use an RNA-dependent RNA polymerase (RdRp) for the
replication of their genome. The SARS-CoV-2 RdRp forms a complex made of
multiple non-structural proteins (nsps). The core of the RdRp is the nsp12 cat-
alytic subunit [5]. The nsp12 contains an N-terminal nidovirus RdRp-associated
nucleotidyltransferase (NiRAN) domain, an interface domain, and a C-terminal
RdRp domain [57, 86, 125]. The RdRp domain consists of fingers, palm, and
thumb subdomains [41, 178]. However, nsp12 has little processivity by itself
and in order to function properly, it requires the accessory subunits, nsp7 and
nsp8 [41,45,79,158,195]. The nsp7 and nsp8 subunits bind to the thumb, and an
additional copy of nsp8 binds to the fingers domain [41,79]. The nsp7-nsp8 het-
erodimer is believed to be capable of de novo initiation and has been proposed
to function as an RNA primase to synthesize small primers that are further ex-
tended by the nsp12 [13, 62, 79, 168].
The RdRp also coordinates with many accessory factors such as the nsp13
43
and the nsp14 in order to carry out its functions [22, 98, 152, 194]. The nsp13
is a helicase that is essential for unwinding the RNA complex after replica-
tion in a 5’ −→ 3’ direction [63, 64, 85, 140]. Besides this, nsp13 also displays
5’-triphosphatase activity, which suggests a possibility that it is involved in
mRNA capping [22,63,64]. The nsp14 which possesses an N-terminal exonucle-
ase (ExoN) domain, together with its co-factor nsp10, forms plays an important
role as a proofreading machinery during viral replication [29, 105, 151, 152].
Once the virus enters the host cells, its positive-sense RNA (+mRNA) is able
to act as a template immediately without having to undergo any modification.
This allows the RdRp to use it to synthesize complementary negative-sense
RNA (-mRNA), immediately. The initiation of synthesis occurs at the 3’ end
of the template and proceeds in the 5’−→3’ direction [178]. This complementary
strand that is synthesized is then, itself, able to act as a template for the pro-
duction of new positive-stranded viral genomes. Meanwhile, the positive-sense
RNA is also translated using the cellular machinery. These newly synthesized
proteins and viral genome are further packaged and released from the cell ready
to infect more host cells. The -mRNA does not translate for proteins (or trans-
lates to garbage proteins) [89, 123, 128] (See Figure 4.1).
The key role that RdRp plays in viral replication makes it a promising an-
tiviral drug target [202]. Drugs such as Favipiravir, Remdesivir, and Galidesivir
are being developed to treat viral infection by inhibiting the effects of RdRp
[3, 81, 112]. For instance, Remdesivir is a nucleoside analog that gets incor-
porated into RNA when the viral genome is being replicated. This creates a
translocation barrier that the RdRp encounters after the addition of three more
nucleotides following remdesivir causing the RdRp to stall [45, 81].
44
Figure 4.1: Replication strategy of a (+)ssRNA virus: The viral genome is trans-
lated using cellular machinery. This genome also acts as a template for the pro-
duction of complementary -mRNA. This complementary strand then acts as a
template for the production of a new viral genome (+mRNA). (Image adapted
from [89])
The RdRp of some viruses have been studied to require the presence of cer-
tain sequences in the RNA analogous to the DNA promoter-specific transcrip-
tion by DNA-dependent RNA polymerases [147]. However, in the case of the
SARS-CoV-2, although nsp12 can distinguish RNA from DNA, no specific in-
teractions have been observed between base pairs of RNA and residues from
nsp12, which suggests that RdRp binding may be RNA sequence-independent
[71, 195]. In the event that the activity of the SARS-CoV-2 RdRp is sequence-
45
dependent, then the 5’UTR and 3’UTR of the genome must be highly con-
served in order for the RdRp to produce the complementary RNA strand from
the viral genome and also produce the viral genome from this complementary
strand [95]. This also indicates that the RdRp can be harnessed to boost the
production of any desired proteins.
In this work, we aimed to simulate the effect of RdRp in a cell-free system. To
this purpose, we introduce the RdRp into a simple genetic circuit in a myTXTL
based cell-free system and simulate its effect to amplify mRNA concentration
and thereby improving protein production. We also attempted to produce the
subunits of the RdRp complex using myTXTL in order to study its effects in
primer extension and expression of a reporter protein.
4.2 Materials and Methods
4.2.1 Model Derivation
In a typical gene expression system, the overall flow of information starts from
the coding region of DNA, which is read by the RNA polymerase, to produce a
messenger RNA molecule (Transcription - TX) which is then read by a ribosome
to produce a protein (Translation - TL). Along the way, both the mRNA and the
protein can degrade (at different rates) [177]
If we consider a system with N genes general, the gene expression of each
gene Gi is described by two differential equations: one for the rate of change
of mRNA ’i’ concentration (ṁi), and one for the rate of change of protein ’i’
46
concentration (ṗi):
ṁi = rX,iūi − (µ + θm,i)mi + λi (4.1)
ṗi = rL,iw̄i − (µ + θp,i)pi (4.2)
The term rX,iūi (nM/hr) denotes the regulated production rate of mRNA ’i’ by
transcription. It is the product of the kinetic limit of transcription rX,i (nM/hr)
and a transcriptional control term ūi Similarly, the term rL,iw̄i (nM/hr) denotes
the regulated production rate of protein ’i’ translation. It is the product of the
kinetic limit of translation rL,i (nM/hr) and a translational control term w̄i. The
term λi denotes the unregulated rate of transcription (the leak for gene Gi). The
ū and w̄ terms describe the control logic of the cell for transcription and transla-
tion respectively. These terms are dimensionless, and bounded between 0 and
1. The term µ denotes the dilution term (hr−1), while the terms θm,i and θp,i the
first-order mRNA degradation rate constant (hr−1) and protein degradation rate
constant (hr−1) respectively. [2, 177]
Since this reaction is in a cell-free system, the dilution (µ) term can be re-
moved and we assume no leak (λi). So these can be removed to modify the
equations to :
ṁi = rX,iūi − θm,imi (4.3)
ṗi = rL,iw̄i − θp,i pi (4.4)
47
Figure 4.2: Introduction of RdRp into a genetic circuit: RdRp uses the +mRNA
as a template for the production of a complementary -mRNA and uses the -
mRNA as a template for the production of +mRNA. Meanwhile, the +mRNA is
also translated to synthesize proteins (Created with BioRender.com)
In order to study the effect of the RdRp, we modified the mRNA rate equa-
tion to include the transcription by RdRp. In this case, we assumed the RdRp to
be a complex of nsp7, nsp8, nsp12, and nsp13. The changes brought by RdRp
can be applied for both +mRNA and -mRNA :
ṁ+i = rX,iūi − θm,im+i + f (m−i ,R − − +d) (4.5) ṁi = −θm,imi + f (mi ,Rd) (4.6)
For the +mRNA, this additional function f (m−i ,Rd) is a function of -mRNA con-
centration (m−i ) and RdRp concentration (Rd), while for the -mRNA, this function
f (m+i ,Rd) depends on +mRNA concentration (m
+
i ) and RdRp concentration. This
is because -mRNA acts as a template for the production of +mRNA using RdRp
and vice versa (See Figure 4.2).
In the case of the -mRNA production there is no term representing normal
transcription from a gene. This is because there is no gene that codes the -mRNA
here. The only source for -mRNA is from RdRp transcription with +mRNA as a
48
template.
We considered the transcription of the mRNA using RdRp to be analogous
to that of DNA-Dependent RNA Polymerase. The RdRp forms a complex with
the mRNA. However, unlike in the transcription involving RNAPII, there is no
open or closed complex here as the template here is an mRNA which is single-
stranded. This process also involves abortive initiations where the RdRp starts
transcription but aborts after a few nucleotides producing subgenomic RNA
(sgmRNA). [146] These can be represented by the following set of equations:
m+ + − −i + Rd ⇐⇒ (mi : Rd) mi + Rd ⇐⇒ (mi : Rd)
(m+ +i : Rd) −→ mi + Rd + sgmRNA− (m−i : Rd) −→ m− +i + Rd + sgmRNA
(m+i : Rd) −→ m+i + R − −d + mi (mi : Rd) −→ m−i + Rd + m+i
Consider the second set of reactions where +mRNA (m+i ) is produced from
-mRNA (m−i ) and RdRp (Rd). The rate constants for these reactions can be de-
noted by :
k∗E,i : Rate constant for transcription elongation
kA : Rate constant for abortive initiation
k+ & k− : ON/OFF rate constants for RdRp-mRNA complex.
We assume that these are going to be the same for the RdRp interactions
with both +mRNA and -mRNA. In this case, the rate of change of the complex
concentration (m−i : Rd) is given by the following equation :
d(m−i : Rd)
= k (m−)(R ) − k (m− : R ) − k (m− : R ) − k∗ (m− : R ) (4.7)
dt + i d − i d A i d E,i i d
k
At Steady State, (m− −1 − −1 +i : Rd) = K (mi )(Rd) where K = k− + kA + k∗E,i
49
In order to make this model more computationally simple, we are going to
assume some amount of compartmentalization in the model. We assume that
at pseudo-steady state conditions, RdRp is going to be equally divided between
binding to -mRNA and +mRNA. This is will indeed be valid if the rate constants
are the same for the interactions of RdRp with both RNAs. Assume that the total
RdRp concentration available for binding with either one of these at a given time
point = RTd , i.e., this is half the total amount of RdRp in the system.
T
RT − −1 −
R
d = Rd + (mi : Rd) = Rd + K (mi )(Rd) ⇒ R
d
d = −1 − (4.8)1 + K mi
( T )
f (m−,R ) r+ ∗ − ∗ −
Rd
i d = X,i = kE,i(mi : Rd) = kE,imi K + m−i
So, this new term can be replaced in equation (4.5() to give) :
RT
ṁ+i = r ū − θ m+ dX,i i m,i i + k∗ −E,imi − (4.9)K + mi
Similarly for -mRNA, ( T )
ṁ−
R
i = −θ − ∗ + dm,imi + kE,imi (4.10)K + m+i
By solving the ODEs (4.9) and (4.10), it is possible to obtain the concentration of
+mRNA ’i’ at any point of time.
At Steady State,
[ ( )] [ ( )]
1 RT 1 RT
m+∗i = rX,iū + k
∗ m−∗ d −∗ ∗ +∗ di
θ E,i i K + m−∗
mi = kθ E,i
mi +∗
m,i i m,i K + mi
Here m+∗i and m
−∗
i represent the steady state concentrations of +mRNA and -
mRNA respectively.
50
Combining these two equations, we get the following equation to denote the
steady state concentration of the +mRNA ’i’,
 ( [ ( )])∗ 1  T ∗ 1 ∗ ∗ Rd 
 

 ( [ RTm+ + di = rX,iūi + kθ E,i kθ E,imi ( )])K + m+∗ 1 ∗ ∗ RT 


m,i m,i i 1 + k + d
θm,i E,i
mi K+m+∗i
We can use equations (4.9) and (4.10) to modify the concentration of +mRNA
at any point of time. Then using this modified +mRNA concentration, we can
use the normal set of translation equations with the same translation parame-
ters to get the temporal variations in the concentration of the proteins. Since
only the +mRNA concentration has been affected by the presence RdRp (the -
mRNA doesn’t code for any protein), that is the only factor that will affect the
translation.
4.2.2 Control functions, Transcription and Translation Kinetic
Limits
The control functions, wi and ui are the(contr)ol functions modeled as:
− ln(2)ẇi = wi (4.11)
∑ τL,1/2
u ∑ j∈χ W j f j(...)i = (4.12)
i∈C W f (...)i i i
Wi denotes the weig∑ht of configuration i and fi is the binding function modeled
as a hill function. j∈χ W j f j(...) is the∑sum of all possible promoter configura-
tions leading to transcription, while i∈C Wi fi(...) is the sum of all possible con-i
figurations for the Gene ’i’. The weights Wi are related to the Gibbs energy of
51
configuration ’i’: ( )
−∆G
W ii = exp (4.13)RT
The rate of transcription rX, ju j was modeled as the product of a kinetic limit,
rX, j, and a control term u j ∈ [0, 1]. The kinetic limit of transcription was derived
from elementary reactions leading to the formation of m j, as previously done by
McClure [101]. The kinetic limit of transcription was formulated as:
rX, j = VmaxX j 
 G(j ) 

,  (4.14)
τX, jKX, j + τX, j + 1 G j
where the initiation time constant τX, j, and saturation constant KX, j were exper-
imentally measured by McClure [101]. The maximum transcription rate, VmaxX, j ,
was given by: [ ( )]
max v̇XVX, j = RX,T (4.15)lG, j
where RX,T is the RNA polymerase concentration (nM), v̇X is the RNA poly-
merase elongation rate (nt/h) and lG, j is the length of gene j in nucleotides (nt).
By analogy, the kinetic limit of translation was formulated as:
max 

x 
r V mRNA L, j = L , j ( )  (4.16)
τL, jKL, j + τL, j + 1 xmRNA
where xmRNA is the mRNA concentration and VmaxL, j is the maximum translation
rate which was formulated as: [ ( )]
max v̇LVL, j = KPRL,T (4.17)lP, j
4.2.3 Building the model
A preliminary sequence-specific model was generated on Julia using the Gene
Regulatory Network Generator - JuGRN [176]. This model takes into account
52
the transcription and translation parameters and solves for the protein expres-
sion as an FBA problem [181]. This model was modified to account for the ef-
fect of RdRp. The parameters for the model were obtained from BioNumbers or
Vilkhovoy, et al [104,181]. Arbor Biosciences’s myTXTL extract was chosen to be
the cell-free extract used in this model. These model equations were solved nu-
merically using the Rosenbrock23 routine of the DifferentialEquations.jl pack-
age [126]. The model code and parameters used in this chapter are available in
the following Github repository: https://github.coecis.cornell.edu/
am3246/RdRp
4.2.4 Cell-Free Protein Synthesis Reactions
The cell-free protein synthesis (CFPS) reactions were carried out using the
myTXTL Sigma 70 Master Mix (Arbor Biosciences) in 1.5 mL Eppendorf tubes.
The working volume of all the reactions to synthesize nsp12 or nsp13 was 17 µL,
composed of the Sigma 70 Master Mix (9 µL), T7 RNA polymerase DNA (1.5 µL
- 0.225 nM), and the plasmids (6.5 µL - 7.65 nM) which were constructed with
a T7 promoter. The working volume of all the reactions to synthesize nsp7 or
nsp8 was 12.8 µL, composed of the Sigma 70 Master Mix (9 µL), GamS nuclease
inhibitor (0.8 µL) and the linear DNA fragments (3 µL). These full constructs
were ordered as linear DNA fragments, with 150-200 bp flanker sequences on
both ends, from Twist Bioscience and Integrated DNA Technologies. The CFPS
reactions were incubated at 29°C. The CFPS runs were carried out in triplicate
for 2hr, 4hr, 8hr, and 16hr time points. Following each CFPS run, the samples
were stored at -80°C.
53
4.2.5 mRNA Quantification
Following each CFPS run, the total RNA was extracted from 5 µL reaction mix-
ture using PureLink RNA Mini Kit (Thermo Fisher Scientific) and stored at
−80°C. The quantitative RT-PCR reactions were done using Applied Biosystems
TaqMan RNA-to-CT 1-Step Kit and Custom TaqMan Gene Expression Assays
(Thermo Fisher Scientific). An mRNA standard curve was used to determine
absolute mRNA concentrations for each of the samples. The mRNA standards
were prepared as follows: separate CFPS reactions for 5 nM of DNA (nsp7,
nsp8) were carried out for 2 hr. Total RNA was extracted using the full reac-
tion volume. This was followed by the removal of 16S and 23S rRNA using the
MICROBExpress Bacterial mRNA Enrichment Kit (Life Technologies Corpora-
tion). Lastly, the MEGAclear Kit (Life Technologies Corporation) was used to
further purify the mRNA. The mRNA concentrations were determined using
the Qubit™ RNA assay kit (ThermoFisher Scientific). At least three technical
replicates were performed for each standard.
4.2.6 Protein Quantification
Following each CFPS run, 2 µL of each sample was extracted and added to
4 µL of water. And an equal amount of 2x Laemmli sample buffer (with 2-
mercaptoethanol as a reducing agent) was added to these. An SDS-PAGE of
10µL of each of these diluted samples along with a Protein ladder of sam-
ples with known concentrations was performed at 300V until the dye fronts
reached the bottom of the gel. This was followed by staining the gel with Bio-
Safe™ Coomassie Stain (Bio-Rad Laboratories). The SDS-PAGE was performed
54
in a Mini-PROTEAN Tetra Cell (Bio-Rad Laboratories) using Any kD™ Mini-
PROTEAN® TGX Stain-Free™ Protein Gels, 15 well, 15 µl gels (Bio-Rad Labo-
ratories). The stained gel was visualized using the ChemiDoc MP and then the
concentrations were measured using Image Lab software.
4.3 Results and Discussion
(a) (b)
Figure 4.3: Temporal variation of N mRNA and protein when expressed along-
side RdRp. The blue curves represent the expression when the model doesn’t
account for the effect of RdRp, while the orange curves account for the RdRp
feedback to amplify mRNA and thereby amplify protein expression
We used the developed model to simulate the effect of RdRp on an arbitrarily
chosen SARS-CoV-2 protein with a comparatively low half-life - the nucleocap-
sid (N) protein of the SARS-CoV-2. The simulation used 5 nM each of DNA
coding for nsp1 and RdRp. The simulation was run for a 30hr reaction time
with two different models - One assumed a normal reaction where the effect of
RdRp wasn’t accounted for, and the other accounted for the RdRp’s ability to
amplify mRNA and thereby amplify protein expression. From Figure 4.3, we
can observe that including the effect of RdRp on mRNA amplification does sig-
55
nificantly boost mRNA and thereby protein production. When we need to fit
the model with experimental measurements, this model can easily be verified
by replacing the chosen protein with a reporter protein such as deGFP.
We were able to synthesize the nsp7 (∼ 10.2 kDa) and nsp8 (∼ 22.8 kDa)
proteins using the myTXTL cell-free system. With a starting concentration of
9.38nM of the nsp7/nsp8 DNA, we were able to obtain up to an average of
29.98 µM of nsp7 protein and 6.96 µM of nsp8 protein after a CFPS run of 16hr.
The mRNA concentration peaked at the 4hr measurement point and continues
to decline after that for both (Figures 4.4 and 4.5).
(a) (b)
Figure 4.4: Gene expression of nsp7. The points denote the mean, while the
error bars denote one standard deviation calculated from three replicates
56
(a) (b)
Figure 4.5: Gene expression of nsp8. The points denote the mean, while the
error bars denote one standard deviation calculated from three replicates
However, we were not able to synthesize the larger nsp12 (∼ 106.2 kDa) and
nsp13 (∼ 67.8 kDa) subunits. Quantitative RT-PCR mRNA measurements of
nsp12 and nsp13 CFPS samples after a 12 hour run revealed that only 0.58 nM
of nsp12 mRNA and 2.97 nM of nsp13 mRNA was present. This explains the
absence of protein and helps us narrow down the cause. We suspect the issue
to be due to the presence of RNases (Ribonucleases) that are causing a higher
degradation rate for mRNA than expected.
4.3.1 Future Direction
The model is still in its nascent stages and requires a lot of tweaking in order
to make it more accurate. For one, currently, the model only takes into consid-
eration the effect of RdRp and doesn’t account for the formation of the RdRp
complex along with nsp7, nsp8, and nsp13. The model will need to account for
the parameters involved in the formation of this complex and the stability of
the complex in addition to the stability of the individual subunits. We could
57
also explore building a model using FBA (Flux Balance Analysis) instead of the
ODE-based model.
Secondly, several model parameters need to be determined using experi-
mental data. The viral non-structural proteins involved with RdRp have al-
ready been successfully synthesized in E. coli cells [22,27,194]; So, it follows that
it should still be theoretically possible to synthesize them in a E. coli-based cell-
free extract. However, since we were unable to synthesize the larger nsp12 and
nsp13 in our myTXTL system, we need to inspect other methods to synthesize
them. Towards this aim, we can explore the use of other systems using extracts
from other organisms such as the mammalian HeLa cell-free extract. The suc-
cessful production of certain therapeutic proteins is often based on a synthesis
setup that is closely related to in vivo conditions [199]. In the case it proves to
be futile to synthesize the proteins in vivo, we could make use of traditional in
vivo systems for the production of these proteins, but later assemble them for
study in a cell-free system.
Lastly, once the binding parameters for the system are determined, it would
be feasible to apply this model to cell-based systems to predict the spread of
viral infection in human cells. If this approach is successful, it will shed light on
the viral replication kinetics of all ssRNA viruses.
58
CHAPTER 5
CONCLUSION
We have constructed a promising transcriptional unit based on a gluconate-
responsive allosteric transcription factor, GntR, that can be utilized to ”sense”
glucose via its proxy, D-gluconate. The E. coli GntR combined with the mP70
promoter construct showed promising repression and de-repression results in
the PURE system, which did not have metabolic side reactions that consumed
D-gluconate. Taken together we have constructed a promising transcriptional
unit that can ”sense” glucose via its proxy, D-gluconate. We aim to integrate
this unit in a more complex genetic circuit to fully utilize its glucose-sensing po-
tential to direct downstream synthesis of proteins such as insulin and glucagon
for the management of diabetes. Second, we expressed the SARS-CoV-2 host
shutoff factor, nsp1 in the myTXTL cell-free system and demonstrated that it
does not bind the E.coli ribosomes. Thereby rendering its host translation inhi-
bition ineffective in the E.coli-based cell-free system. Lastly, we also developed
a mathematical model to study the effect of the SARS-CoV-2 RNA-dependent
RNA polymerase (RdRp) on the production of a desired protein in the myTXTL
cell-free system. Through simulations using the model built on a JuGRN back-
bone, we demonstrated that the presence of RdRp can amplify the mRNA in the
system, thereby boosting protein production. We also postulate that RdRp can
be harnessed to improve the production of desired proteins.
59
BIBLIOGRAPHY
[1] Diabetes facts and figures. https://idf.org/aboutdiabetes/
what-is-diabetes/facts-figures.html, 2019.
[2] Abhinav Adhikari, Michael Vilkhovoy, Sandra Vadhin, Ha Eun Lim, and
Jeffrey D Varner. Effective biophysical modeling of cell free transcription
and translation processes. Frontiers in bioengineering and biotechnology, 8,
2020.
[3] Syed Ovais Aftab, Muhammad Zubair Ghouri, Muhammad Umer Ma-
sood, Zeshan Haider, Zulqurnain Khan, Aftab Ahmad, and Nayla Mu-
nawar. Analysis of sars-cov-2 rna-dependent rna polymerase as a poten-
tial therapeutic drug target using a computational approach. Journal of
translational medicine, 18(1):1–15, 2020.
[4] Dmitry E Agafonov, Vyacheslav A Kolb, Igor V Nazimov, and Alexan-
der S Spirin. A protein residing at the subunit interface of the bacterial
ribosome. Proceedings of the National Academy of Sciences, 96(22):12345–
12349, 1999.
[5] Dae-Gyun Ahn, Jin-Kyu Choi, Deborah R Taylor, and Jong-Won Oh.
Biochemical characterization of a recombinant sars coronavirus nsp12
rna-dependent rna polymerase capable of copying viral rna templates.
Archives of virology, 157(11):2095–2104, 2012.
[6] Cem Albayrak and James R. Swartz. Cell-free co-production of an orthog-
onal transfer rna activates efficient site-specific non-natural amino acid
incorporation. Nucleic Acids Research, 41(11):5949–5963, 2013.
[7] Timothy E Allen and Bernhard Ø Palsson. Sequence-based analysis of
metabolic demands for protein synthesis in prokaryotes. Journal of theoret-
ical biology, 220(1):1–18, 2003.
[8] Jorge Alonso-Gutierrez, Eun-Mi Kim, Tanveer S. Batth, Nathan Cho, Qi-
jun Hu, Leanne Jade G. Chan, Christopher J. Petzold, Nathan J. Hillson,
Paul D. Adams, Jay D. Keasling, Hector Garcia Martin, and Taek Soon
Lee. Principal component analysis of proteomics (pcap) as a tool to direct
metabolic engineering. Metabolic Engineering, 28:123–133, 2015.
[9] Maria Anastasina, Ilya Terenin, Sarah J Butcher, and Denis E Kainov. A
technique to increase protein yield in a rabbit reticulocyte lysate transla-
tion system. Biotechniques, 56(1):36–39, 2014.
60
[10] JC Atlas, EV Nikolaev, ST Browning, and ML Shuler. Incorporating
genome-wide dna sequence information into a dynamic whole-cell model
of escherichia coli: application to dna replication. IET systems biology,
2(5):369–382, 2008.
[11] Abhik K Banerjee, Mario R Blanco, Emily A Bruce, Drew D Honson, Lin-
lin M Chen, Amy Chow, Prashant Bhat, Noah Ollikainen, Sofia A Quin-
odoz, Colin Loney, et al. Sars-cov-2 disrupts splicing, translation, and
protein trafficking to suppress host defenses. Cell, 183(5):1325–1339, 2020.
[12] Lacramioara Bintu, Nicolas E Buchler, Hernan G Garcia, Ulrich Gerland,
Terence Hwa, Jané Kondev, and Rob Phillips. Transcriptional regula-
tion by the numbers: models. Current opinion in genetics & development,
15(2):116–124, 2005.
[13] Mahamaya Biswal, Stephen Diggs, Duo Xu, Nelli Khudaverdyan, Jiuwei
Lu, Jian Fang, Gregor Blaha, Rong Hai, and Jikui Song. Two conserved
oligomer interfaces of nsp7 and nsp8 underpin the dynamic assembly of
sars-cov-2 rdrp. Nucleic acids research, 49(10):5956–5966, 2021.
[14] Igor W. Bogorad, Tzu-Shyang Lin, and James C. Liao. Synthetic non-
oxidative glycolysis enables complete carbon conservation. Nature,
502(7473):693–697, 2013.
[15] Henry Borsook. Protein turnover and incorporation of labeled amino
acids into tissue proteins in vivo and in vitro. Physiological reviews,
30(2):206–219, 1950.
[16] Shelton S Bradrick, Simardeep Nagyal, and Hilary Novatt. A mirna-
responsive cell-free translation system facilitates isolation of hepatitis c
virus mirnp complexes. Rna, 19(8):1159–1169, 2013.
[17] Sabine Brantl and E Gerhart H Wagner. Antisense rna-mediated transcrip-
tional attenuation: an in vitro study of plasmid pt181. Molecular microbiol-
ogy, 35(6):1469–1482, 2000.
[18] Kara A. Calhoun and James R. Swartz. Energizing cell-free protein
synthesis with glucose metabolism. Biotechnology and Bioengineering,
90(5):606–613, 2005.
[19] Mariajose Castellanos, David B Wilson, and Michael L Shuler. A modular
minimal cell model: purine and pyrimidine transport and metabolism.
Proceedings of the National Academy of Sciences, 101(17):6681–6686, 2004.
61
[20] Roger L Chang, Kathleen Andrews, Donghyuk Kim, Zhanwen Li, Adam
Godzik, and Bernhard O Palsson. Structural systems biology evaluation
of metabolic thermotolerance in escherichia coli. Science, 340(6137):1220–
1223, 2013.
[21] James Chappell, Melissa K Takahashi, and Julius B Lucks. Creating small
transcription activating rnas. Nature chemical biology, 11(3):214–220, 2015.
[22] James Chen, Brandon Malone, Eliza Llewellyn, Michael Grasso,
Patrick MM Shelton, Paul Dominic B Olinares, Kashyap Maruthi, Ed-
ward T Eng, Hasan Vatandaslar, Brian T Chait, et al. Structural basis for
helicase-polymerase coupling in the sars-cov-2 replication-transcription
complex. Cell, 182(6):1560–1573, 2020.
[23] A Daddaoua, A Corral-Lugo, J-L Ramos, and Tino Krell. Identification of
gntr as regulator of the glucose metabolism in pseudomonas aeruginosa.
Environmental microbiology, 19(9):3721–3733, 2017.
[24] David Dai, Nicholas Horvath, and Jeffrey Varner. Dynamic sequence spe-
cific constraint-based modeling of cell-free protein synthesis. Processes,
6(8):132, 2018.
[25] Samar Damiati, Rami Mhanna, Rimantas Kodzius, and Eva-Kathrin
Ehmoser. Cell-free approaches in synthetic biology utilizing microflu-
idics. Genes, 9(3):144, 2018.
[26] Denis Daneman. Type 1 diabetes. The Lancet, 367(9513):847–858, 2006.
[27] Tyler L Dangerfield, Nathan Z Huang, and Kenneth A Johnson. Expres-
sion and purification of tag-free sars-cov-2 rna-dependent rna polymerase
in escherichia coli. STAR protocols, 2(1):100357, 2021.
[28] Ulysses Amancio de Frias, Greicy Kelly Bonifacio Pereira, Marı́a-Eugenia
Guazzaroni, and Rafael Silva-Rocha. Boosting secondary metabolite pro-
duction and discovery through the engineering of novel microbial biosen-
sors. BioMed Research International, 2018:7021826–7021826, 2018.
[29] Mark R Denison, Rachel L Graham, Eric F Donaldson, Lance D Eckerle,
and Ralph S Baric. Coronaviruses: an rna proofreading machine regulates
replication fidelity and diversity. RNA biology, 8(2):270–279, 2011.
62
[30] Devasenan Devendra, Edwin Liu, and George S Eisenbarth. Type 1 dia-
betes: recent developments. Bmj, 328(7442):750–754, 2004.
[31] MM Domach, SK Leung, RE Cahn, GG Cocks, and ML Shuler. Computer
model for glucose-limited growth of a single cell of escherichia coli b/r-a.
Biotechnology and bioengineering, 26(9):1140–1140, 1984.
[32] Quentin M Dudley, Ashty S Karim, and Michael C Jewett. Cell-free
metabolic engineering: biomanufacturing beyond the cell. Biotechnology
journal, 10(1):69–82, 2015.
[33] Quentin M. Dudley, Connor J. Nash, and Michael Christopher Jewett.
Cell-free biosynthesis of limonene using enzyme-enriched escherichia coli
lysates. Synthetic Biology, 4(1), 2019.
[34] Jeremy S Edwards and Bernhard O Palsson. Metabolic flux balance anal-
ysis and the in silico analysis of escherichia coli k-12 gene deletions. BMC
bioinformatics, 1(1):1–10, 2000.
[35] Toru Ezure, Takashi Suzuki, Shoken Higashide, Eiichi Shintani, Kohki
Endo, Shin-ichiro Kobayashi, Masamitsu Shikata, Masaaki Ito, Koji Tani-
mizu, and Osamu Nishimura. Cell-free protein synthesis system prepared
from insect cells by freeze-thawing. Biotechnology progress, 22(6):1570–
1577, 2006.
[36] Anthony R Fehr and Stanley Perlman. Coronaviruses: an overview of
their replication and pathogenesis. Coronaviruses, pages 1–23, 2015.
[37] Yaara Finkel, Orel Mizrahi, Aharon Nachshon, Shira Weingarten-Gabbay,
David Morgenstern, Yfat Yahalom-Ronen, Hadas Tamir, Hagit Achdout,
Dana Stein, Ofir Israeli, et al. The coding capacity of sars-cov-2. Nature,
589(7840):125–130, 2021.
[38] AG Fredrickson. Formulation of structured growth models. Biotechnology
and bioengineering, 18(10):1481–1486, 1976.
[39] Y Fujita and Y Miwa. Identification of an operator sequence for the bacil-
lus subtilis gnt operon. Journal of Biological Chemistry, 264(7):4201–4206,
1989.
[40] Ernest F Gale, Joan P Folkes, et al. Effect of nucleic acids on protein syn-
63
thesis and amino-acid incorporation in disrupted staphylococcal cells. Na-
ture, 173:1223–7, 1954.
[41] Yan Gao, Liming Yan, Yucen Huang, Fengjiang Liu, Yao Zhao, Lin Cao,
Tao Wang, Qianqian Sun, Zhenhua Ming, Lianqi Zhang, et al. Struc-
ture of the rna-dependent rna polymerase from covid-19 virus. Science,
368(6492):779–782, 2020.
[42] Jonathan Garamella, Ryan Marshall, Mark Rustad, and Vincent Noireaux.
The all e. coli tx-tl toolbox 2.0: a platform for cell-free synthetic biology.
ACS synthetic biology, 5(4):344–355, 2016.
[43] V. Georgi, Leopold Georgi, Martin Blechert, Merlin Bergmeister, Michael
Zwanzig, Doreen A. Wüstenhagen, Frank F. Bier, Erik Jung, and Stefan
Kubick. On-chip automation of cell-free protein synthesis: new opportu-
nities due to a novel reaction mode. Lab on a Chip, 16(2):269–281, 2016.
[44] Aaron R Goerke and James R Swartz. Development of cell-free protein
synthesis platforms for disulfide bonded proteins. Biotechnology and bio-
engineering, 99(2):351–367, 2008.
[45] Calvin J Gordon, Egor P Tchesnokov, Joy Y Feng, Danielle P Porter, and
Matthias Götte. The antiviral compound remdesivir potently inhibits rna-
dependent rna polymerase from middle east respiratory syndrome coro-
navirus. Journal of Biological Chemistry, 295(15):4773–4779, 2020.
[46] Nicole E Gregorio, Max Z Levine, and Javin P Oza. A user’s guide to
cell-free protein synthesis. Methods and protocols, 2(1):24, 2019.
[47] Cassandra Guarino and Matthew P DeLisa. A prokaryote-based cell-free
translation system that efficiently synthesizes glycoproteins. Glycobiology,
22(5):596–601, 2012.
[48] Weihua Guo, Jiayuan Sheng, and Xueyang Feng. Mini-review: in
vitro metabolic engineering for biomanufacturing of high-value products.
Computational and Structural Biotechnology Journal, 15:161–167, 2017.
[49] Jan-Karl Guterl, Daniel Garbe, Jörg Carsten, Fabian Steffler, Bettina Som-
mer, Steven Reiße, Anja Philipp, Martina Haack, Broder Rühmann, Andre
Koltermann, Ulrich Kettling, Thomas Brück, and Volker Sieber. Cell-free
metabolic engineering: production of chemicals by minimized reaction
cascades. Chemsuschem, 5(11):2165–2172, 2012.
64
[50] Andras Gyorgy and Richard M Murray. Quantifying resource competi-
tion and its effects in the tx-tl system. In 2016 IEEE 55th Conference on
Decision and Control (CDC), pages 3363–3368. IEEE, 2016.
[51] Shogo Hamada, Kenneth Gene Yancey, Yehudah Pardo, Mingzhe Gan,
Max Vanatta, Duo An, Yue Hu, Thomas L. Derrien, Roanna Ruiz, Peifeng
Liu, Jenny Sabin, and Dan Luo. Dynamic dna material with emergent
locomotion behavior powered by artificial metabolism. Science Robotics,
4(29), 2019.
[52] Joshua J Hamilton, Vivek Dwivedi, and Jennifer L Reed. Quantitative as-
sessment of thermodynamic constraints on the solution space of genome-
scale metabolic models. Biophysical journal, 105(2):512–522, 2013.
[53] Ryan L. Hartman, Jonathan P. McMullen, and Klavs F. Jensen. Deciding
whether to go with the flow: evaluating the merits of flow reactors for
synthesis. Angewandte Chemie, 50(33):7502–7519, 2011.
[54] Hartmut Hengel, Ulrich H Koszinowski, and Karl-Klaus Conzelmann.
Viruses know it all: new insights into ifn networks. Trends in immunol-
ogy, 26(7):396–401, 2005.
[55] Christopher S Henry, Linda J Broadbelt, and Vassily Hatzimanikatis.
Thermodynamics-based metabolic flux analysis. Biophysical journal,
92(5):1792–1805, 2007.
[56] Jason R Hillebrecht and Shaorong Chong. A comparative study of protein
synthesis in in vitro systems: from the prokaryotic reconstituted to the
eukaryotic extract-based. BMC biotechnology, 8(1):1–9, 2008.
[57] Hauke S Hillen, Goran Kokic, Lucas Farnung, Christian Dienemann, Dim-
itry Tegunov, and Patrick Cramer. Structure of replicating sars-cov-2 poly-
merase. Nature, 584(7819):154–156, 2020.
[58] Yu HongFeng, Ding YongQian, Liu HaiTao, Zhu WenQian, Liu Guo-
Qiang, Fu XiuQing, and Ding WeiMin. Optimization design and appli-
cation of variable rate fertilization system for small-scaled fields. Transac-
tions of the Chinese Society of Agricultural Engineering, 34(3):35–41, 2018.
[59] Nicholas Horvath, Michael Vilkhovoy, Joseph A Wayman, Kara Calhoun,
James Swartz, and Jeffrey D Varner. Toward a genome scale sequence
specific dynamic model of cell-free protein synthesis in escherichia coli.
Metabolic engineering communications, 10:e00113, 2020.
65
[60] Chelsea Y Hu, Jeffrey D Varner, and Julius B Lucks. Generating effec-
tive models and parameters for rna genetic circuits. ACS synthetic biology,
4(8):914–926, 2015.
[61] Cheng Huang, Kumari G Lokugamage, Janet M Rozovics, Krishna
Narayanan, Bert L Semler, and Shinji Makino. Sars coronavirus nsp1
protein induces template-dependent endonucleolytic cleavage of mrnas:
viral mrnas are resistant to nsp1-induced rna cleavage. PLoS pathogens,
7(12):e1002433, 2011.
[62] Isabelle Imbert, Jean-Claude Guillemot, Jean-Marie Bourhis, Cécile Bus-
setta, Bruno Coutard, Marie-Pierre Egloff, François Ferron, Alexander E
Gorbalenya, and Bruno Canard. A second, non-canonical rna-dependent
rna polymerase in sars coronavirus. The EMBO journal, 25(20):4933–4942,
2006.
[63] Konstantin A Ivanov, Volker Thiel, Jessika C Dobbe, Yvonne Van
Der Meer, Eric J Snijder, and John Ziebuhr. Multiple enzymatic activities
associated with severe acute respiratory syndrome coronavirus helicase.
Journal of virology, 78(11):5619–5632, 2004.
[64] Konstantin A Ivanov and John Ziebuhr. Human coronavirus 229e non-
structural protein 13: characterization of duplex-unwinding, nucleoside
triphosphatase, and rna 5-triphosphatase activities. Journal of virology,
78(14):7833–7838, 2004.
[65] Thapakorn Jaroentomeechai, Jessica C Stark, Aravind Natarajan,
Cameron J Glasscock, Laura E Yates, Karen J Hsu, Milan Mrksich,
Michael C Jewett, and Matthew P DeLisa. Single-pot glycoprotein biosyn-
thesis using a cell-free transcription-translation system enriched with gly-
cosylation machinery. Nature communications, 9(1):1–11, 2018.
[66] Andrew R Jauregui, Dhruti Savalia, Virginia K Lowry, Cara M Farrell, and
Marc G Wathelet. Identification of residues of sars-cov nsp1 that differen-
tially affect inhibition of gene expression and antiviral signaling. PloS one,
8(4):e62416, 2013.
[67] Michael C Jewett, Kara A Calhoun, Alexei Voloshin, Jessica J Wuu, and
James R Swartz. An integrated cell-free metabolic platform for protein
production and synthetic biology. Molecular systems biology, 4(1):220, 2008.
[68] Michael C Jewett and James R Swartz. Mimicking the escherichia coli cy-
66
toplasmic environment activates long-lived and efficient cell-free protein
synthesis. Biotechnology and bioengineering, 86(1):19–26, 2004.
[69] Michael C Jewett and James R Swartz. Substrate replenishment extends
protein synthesis with an in vitro translation system designed to mimic
the cytoplasm. Biotechnology and bioengineering, 87(4):465–471, 2004.
[70] Guoqiang Jiang and Bei B Zhang. Glucagon and regulation of glucose
metabolism. American Journal of Physiology-Endocrinology And Metabolism,
284(4):E671–E678, 2003.
[71] Yi Jiang, Wanchao Yin, and H Eric Xu. Rna-dependent rna polymerase:
Structure, mechanism, and drug discovery for covid-19. Biochemical and
biophysical research communications, 538:47–53, 2021.
[72] Jaeyoung K Jung, Khalid K Alam, Matthew S Verosloff, Daiana A Capdev-
ila, Morgane Desmau, Phillip R Clauer, Jeong Wook Lee, Peter Q Nguyen,
Pablo A Pastén, Sandrine J Matiasek, et al. Cell-free biosensors for rapid
detection of water contaminants. Nature biotechnology, 38(12):1451–1459,
2020.
[73] Wataru Kamitani, Cheng Huang, Krishna Narayanan, Kumari G Lokuga-
mage, and Shinji Makino. A two-pronged strategy to suppress host pro-
tein synthesis by sars coronavirus nsp1 protein. Nature structural & molec-
ular biology, 16(11):1134–1140, 2009.
[74] Eyal Karzbrun, Jonghyeon Shin, Roy H Bar-Ziv, and Vincent Noireaux.
Coarse-grained dynamics of protein synthesis in a cell-free system. Phys-
ical review letters, 106(4):048104, 2011.
[75] Rozhgar A Khailany, Muhamad Safdar, and Mehmet Ozaslan. Genomic
characterization of a novel sars-cov-2. Gene reports, 19:100682, 2020.
[76] Akram T Kharroubi and Hisham M Darwish. Diabetes mellitus: The epi-
demic of the century. World Journal of Diabetes, 6(6):850–867, 2015.
[77] Dong-Myung Kim and James R Swartz. Regeneration of adenosine
triphosphate from glycolytic intermediates for cell-free protein synthesis.
Biotechnology and bioengineering, 74(4):309–316, 2001.
[78] Dongwan Kim, Joo-Yeon Lee, Jeong-Sun Yang, Jun Won Kim, V Narry
67
Kim, and Hyeshik Chang. The architecture of sars-cov-2 transcriptome.
Cell, 181(4):914–921, 2020.
[79] Robert N Kirchdoerfer and Andrew B Ward. Structure of the sars-cov
nsp12 polymerase bound to nsp7 and nsp8 co-factors. Nature communica-
tions, 10(1):1–9, 2019.
[80] Mathilde Koch, Amir Pandi, Olivier Borkowski, Angelo Cardoso Batista,
and Jean Loup Faulon. Custom-made transcriptional biosensors for
metabolic engineering. Current Opinion in Biotechnology, 59:78–84, 2019.
[81] Goran Kokic, Hauke S Hillen, Dimitry Tegunov, Christian Dienemann,
Florian Seitz, Jana Schmitzova, Lucas Farnung, Aaron Siewert, Claudia
Höbartner, and Patrick Cramer. Mechanism of sars-cov-2 polymerase
stalling by remdesivir. Nature communications, 12(1):1–7, 2021.
[82] Borimas Krutsakorn, Kohsuke Honda, Xiaoting Ye, Takashi Imagawa, Xi-
aoyu Bei, Kenji Okano, and Hisao Ohtake. In vitro production of n-
butanol from glucose. Metabolic Engineering, 20:84–91, 2013.
[83] Christopher P Lapointe, Rosslyn Grosely, Alex G Johnson, Jinfan Wang,
Israel S Fernández, and Joseph D Puglisi. Dynamic competition between
sars-cov-2 nsp1 and mrna on the human ribosome inhibits translation ini-
tiation. Proceedings of the National Academy of Sciences, 118(6), 2021.
[84] Muriel Lederman and Geoffrey Zubay. Dna-directed peptide synthesis i.
a comparison of t2 and escherichia coli dna-directed peptide synthesis in
two cell-free systems. Biochimica et Biophysica Acta (BBA)-Nucleic Acids and
Protein Synthesis, 149(1):253–258, 1967.
[85] Na-Ra Lee, Hyun-Mi Kwon, Kkothanahreum Park, Sangtaek Oh, Yong-
Joo Jeong, and Dong-Eun Kim. Cooperative translocation enhances the
unwinding of duplex dna by sars coronavirus helicase nsp13. Nucleic acids
research, 38(21):7626–7636, 2010.
[86] Kathleen C Lehmann, Anastasia Gulyaeva, Jessika C Zevenhoven-Dobbe,
George MC Janssen, Mark Ruben, Hermen S Overkleeft, Peter A Van Vee-
len, Dmitry V Samborskiy, Alexander A Kravchenko, Andrey M Leon-
tovich, et al. Discovery of an essential nucleotidylating activity associ-
ated with a newly delineated conserved domain in the rna polymerase-
containing protein of all nidoviruses. Nucleic acids research, 43(17):8416–
8434, 2015.
68
[87] Roberta Lentini, Michele Forlin, Laura Martini, Cristina Del Bianco,
Amy C Spencer, Domenica Torino, and Sheref S Mansy. Fluorescent pro-
teins and in vitro genetic organization for cell-free synthetic biology. ACS
synthetic biology, 2(9):482–489, 2013.
[88] Nathan E Lewis, Harish Nagarajan, and Bernhard O Palsson. Constrain-
ing the metabolic genotype–phenotype relationship using a phylogeny of
in silico methods. Nature Reviews Microbiology, 10(4):291–305, 2012.
[89] Dongsheng Li, Ting Wei, Catherine M Abbott, and David Harrich. The
unexpected roles of eukaryotic translation elongation factors in rna virus
replication and pathogenesis. Microbiol. Mol. Biol. Rev., 77(2):253–266,
2013.
[90] Jun Li, Liangcai Gu, John Aach, and George M Church. Improved cell-free
rna and protein synthesis system. PLoS One, 9(9):e106232, 2014.
[91] Di Liu, Trent Evans, and Fuzhong Zhang. Applications and advances of
metabolite biosensors for metabolic engineering. Metabolic Engineering,
31:35–43, 2015.
[92] Yuan Lu. Cell-free synthetic biology: Engineering in an open world. Syn-
thetic and systems biotechnology, 2(1):23–27, 2017.
[93] Yuan Lu, John P Welsh, and James R Swartz. Production and stabiliza-
tion of the trimeric influenza hemagglutinin stem domain for potentially
broadly protective influenza vaccines. Proceedings of the National Academy
of Sciences, 111(1):125–130, 2014.
[94] Julius B Lucks, Lei Qi, Vivek K Mutalik, Denise Wang, and Adam P Arkin.
Versatile rna-sensing transcriptional regulators for engineering genetic
networks. Proceedings of the National Academy of Sciences, 108(21):8617–
8622, 2011.
[95] Guangxiang Luo, Robert K Hamatake, Danielle M Mathis, Jason Racela,
Karen L Rigat, Julie Lemm, and Richard J Colonno. De novo initiation of
rna synthesis by the rna-dependent rna polymerase (ns5b) of hepatitis c
virus. Journal of virology, 74(2):851–863, 2000.
[96] Rolf Lutz and Hermann Bujard. Independent and tight regulation of tran-
scriptional units in escherichia coli via the lacr/o, the tetr/o and arac/i1-i2
regulatory elements. Nucleic acids research, 25(6):1203–1210, 1997.
69
[97] Duo Ma, Luhui Shen, Kaiyue Wu, Chris W Diehnelt, and Alexander A
Green. Low-cost detection of norovirus using paper-based cell-free sys-
tems and synbody-based viral enrichment. Synthetic Biology, 3(1), 2018.
[98] Brandon Malone, James Chen, Qi Wang, Eliza Llewellyn, Young Joo Choi,
Paul Dominic B Olinares, Xinyun Cao, Carolina Hernandez, Edward T
Eng, Brian T Chait, et al. Structural basis for backtracking by the sars-cov-
2 replication–transcription complex. Proceedings of the National Academy of
Sciences, 118(19), 2021.
[99] Rey W. Martin, Benjamin J. Des Soye, Yong Chan Kwon, Jennifer Kay,
Roderick G. Davis, Paul M. Thomas, Natalia I. Majewska, Cindy X. Chen,
Ryan D. Marcum, Mary Grace Weiss, Ashleigh E. Stoddart, Miriam Ami-
ram, Arnaz K. Ranji Charna, Jaymin R. Patel, Farren J. Isaacs, Neil L.
Kelleher, Seok Hoon Hong, and Michael C. Jewett. Cell-free protein syn-
thesis from genomically recoded bacteria enables multisite incorporation
of noncanonical amino acids. Nature Communications, 9(1):1203, 2018.
[100] J Heinrich Matthaei and Marshall W Nirenberg. Characteristics and sta-
bilization of dnaase-sensitive protein synthesis in e. coli extracts. Pro-
ceedings of the National Academy of Sciences of the United States of America,
47(10):1580, 1961.
[101] William R McClure. Rate-limiting steps in rna chain initiation. Proceedings
of the National Academy of Sciences, 77(10):5634–5638, 1980.
[102] D. T. McQuade and Peter H. Seeberger. Applying flow chemistry: Meth-
ods, materials, and multistep synthesis. Journal of Organic Chemistry,
78(13):6384–6389, 2013.
[103] Satoshi Mikami, Tominari Kobayashi, and Hiroaki Imataka. Cell-free pro-
tein synthesis systems with extracts from cultured human cells. In Cell-
Free Protein Production, pages 43–52. Springer, 2010.
[104] Ron Milo, Paul Jorgensen, Uri Moran, Griffin Weber, and Michael
Springer. Bionumbers—the database of key numbers in molecular and
cell biology. Nucleic acids research, 38(suppl 1):D750–D753, 2010.
[105] Ekaterina Minskaia, Tobias Hertzig, Alexander E Gorbalenya, Valérie
Campanacci, Christian Cambillau, Bruno Canard, and John Ziebuhr. Dis-
covery of an rna virus 3→ 5 exoribonuclease that is critically involved in
coronavirus rna synthesis. Proceedings of the National Academy of Sciences,
103(13):5108–5113, 2006.
70
[106] Simon J. Moore, James T. MacDonald, Sarah Wienecke, Alka Ishwarb-
hai, Argyro Tsipa, Rochelle Aw, Nicolas Kylilis, David J. Bell, David W.
McClymont, Kirsten Jensen, Karen M. Polizzi, Rebekka Biedendieck, and
Paul S. Freemont. Rapid acquisition and model-based analysis of cell-
free transcription–translation reactions from nonmodel bacteria. Proceed-
ings of the National Academy of Sciences of the United States of America,
115(19):201715806, 2018.
[107] Travis W. Murphy, Jiayuan Sheng, Lynette B. Naler, Xueyang Feng, and
Chang Lu. On-chip manufacturing of synthetic proteins for point-of-care
therapeutics. Microsystems Nanoengineering, 5(1):13, 2019.
[108] Keisuke Nakagawa, Krishna Narayanan, Masami Wada, Vsevolod L
Popov, Maria Cajimat, Ralph S Baric, and Shinji Makino. The endonucle-
olytic rna cleavage function of nsp1 of middle east respiratory syndrome
coronavirus promotes the production of infectious virus particles in spe-
cific human cell lines. Journal of virology, 92(21):e01157–18, 2018.
[109] Ahmad Abu Turab Naqvi, Kisa Fatima, Taj Mohammad, Urooj Fatima,
Indrakant K Singh, Archana Singh, Shaikh Muhammad Atif, Gururao
Hariprasad, Gulam Mustafa Hasan, and Md Imtaiyaz Hassan. Insights
into sars-cov-2 genome, structure, evolution, pathogenesis and thera-
pies: Structural genomics approach. Biochimica et Biophysica Acta (BBA)-
Molecular Basis of Disease, 1866(10):165878, 2020.
[110] Krishna Narayanan, Cheng Huang, Kumari Lokugamage, Wataru Kami-
tani, Tetsuro Ikegami, Chien-Te K Tseng, and Shinji Makino. Severe acute
respiratory syndrome coronavirus nsp1 suppresses host gene expression,
including that of type i interferon, in infected cells. Journal of virology,
82(9):4471–4479, 2008.
[111] Krishna Narayanan, Sydney I Ramirez, Kumari G Lokugamage, and
Shinji Makino. Coronavirus nonstructural protein 1: Common and dis-
tinct functions in the regulation of host and viral gene expression. Virus
research, 202:89–100, 2015.
[112] Katerina Naydenova, Kyle W Muir, Long-Fei Wu, Ziguo Zhang,
Francesca Coscia, Mathew J Peet, Pablo Castro-Hartmann, Pu Qian,
Kasim Sader, Kyle Dent, et al. Structure of the sars-cov-2 rna-dependent
rna polymerase in the presence of favipiravir-rtp. Proceedings of the Na-
tional Academy of Sciences, 118(7), 2021.
[113] Patrick P Ng, Ming Jia, Kedar G Patel, Joshua D Brody, James R Swartz,
71
Shoshana Levy, and Ronald Levy. A vaccine directed to b cells and pro-
duced by cell-free protein synthesis generates potent antilymphoma im-
munity. Proceedings of the National Academy of Sciences, 109(36):14526–
14531, 2012.
[114] Pamela N. Nge, Chad I. Rogers, and Adam T. Woolley. Advances in mi-
crofluidic materials, functions, integration, and applications. Chemical Re-
views, 113(4):2550–2583, 2013.
[115] Henrike Niederholtmeyer, Ling Xu, and Sebastian J Maerkl. Real-time
mrna measurement during an in vitro transcription and translation reac-
tion using binary probes. ACS synthetic biology, 2(8):411–417, 2013.
[116] Alexander Nieß, Jurek Failmezger, Maike Kuschel, Martin Siemann-
Herzberg, and Ralf Takors. Experimentally validated model enables de-
bottlenecking of in vitro protein synthesis and identifies a control shift
under in vivo conditions. ACS synthetic biology, 6(10):1913–1921, 2017.
[117] Marshall W Nirenberg and J Heinrich Matthaei. The dependence of
cell-free protein synthesis in e. coli upon naturally occurring or syn-
thetic polyribonucleotides. Proceedings of the National Academy of Sciences,
47(10):1588–1602, 1961.
[118] Vincent Noireaux, Roy Bar-Ziv, and Albert Libchaber. Principles of cell-
free genetic circuit assembly. Proceedings of the National Academy of Sciences,
100(22):12672–12677, 2003.
[119] Edward J O’brien, Joshua A Lerman, Roger L Chang, Daniel R Hyduke,
and Bernhard Ø Palsson. Genome-scale models of metabolism and gene
expression extend and refine growth phenotype prediction. Molecular sys-
tems biology, 9(1):693, 2013.
[120] Paul H Opgenorth, Tyler P Korman, and James U Bowie. A synthetic bio-
chemistry molecular purge valve module that maintains redox balance.
Nature communications, 5(1):1–8, 2014.
[121] Keith Pardee and James J. Collins. Paper-based synthetic gene networks,
2014.
[122] Keith Pardee, Alexander A. Green, Melissa K. Takahashi, Dana Braff,
Guillaume Lambert, Jeong Wook Lee, Tom Ferrante, Duo Ma, Nina
Donghia, Melina Fan, Nichole M. Daringer, Irene Bosch, Dawn M. Dud-
ley, David H. O’Connor, Lee Gehrke, and James J. Collins. Rapid, low-cost
72
detection of zika virus using programmable biomolecular components.
Cell, 165(5):1255–1266, 2016.
[123] Susan Payne. Introduction to rna viruses. Viruses, page 97, 2017.
[124] Norbert Peekhaus and T. Conway. Positive and negative transcriptional
regulation of the escherichia coli gluconate regulon gene gntt by gntr and
the cyclic amp (camp)-camp receptor protein complex. Journal of Bacteri-
ology, 180(7):1777–1785, 1998.
[125] Clara C Posthuma, Aartjan JW Te Velthuis, and Eric J Snijder. Nidovirus
rna polymerases: complex enzymes handling exceptional rna genomes.
Virus research, 234:58–73, 2017.
[126] Aditya Pratapa, Amogh P Jalihal, Jeffrey N Law, Aditya Bharadwaj, and
TM Murali. Benchmarking algorithms for gene regulatory network infer-
ence from single-cell transcriptomic data. Nature methods, 17(2):147–154,
2020.
[127] JM Pratt. Coupled transcription-translation in prokaryotic cell-free sys-
tems. Transcription and translation: a practical approach, 1984.
[128] Sephra Rampersad and Paula Tennant. Replication and expression strate-
gies of viruses. Viruses, page 55, 2018.
[129] Jan Rehwinkel and Michaela U Gack. Rig-i-like receptors: their regulation
and roles in rna sensing. Nature Reviews Immunology, 20(9):537–551, 2020.
[130] A. Reizer, J. Deutscher, M. H. Saier, and J. Reizer. Analysis of the glu-
conate (gnt) operon of bacillus subtilis. Molecular Microbiology, 5(5):1081–
1089, 1991.
[131] Maria Romano, Alessia Ruggiero, Flavia Squeglia, Giovanni Maga, and
Rita Berisio. A structural view of sars-cov-2 rna replication machinery:
Rna synthesis, proofreading and final capping. Cells, 9(5):1267, 2020.
[132] Gabriel Rosenblum and Barry S Cooperman. Engine out of the chassis:
cell-free protein synthesis and its uses. FEBS letters, 588(2):261–268, 2014.
[133] Ryota Sakamoto, Vincent Noireaux, and Yusuke T. Maeda. Anomalous
scaling of gene expression in confined cell-free reactions. Scientific Reports,
8(1):7364–7364, 2018.
73
[134] Shun Sakuraba, Xie Qilin, Kota Kasahara, Junichi Iwakiri, and Hidetoshi
Kono. Modeling sars-cov-2 nsp1-5’-utr complex via the extended ensem-
ble simulations. bioRxiv, 2021.
[135] Claudia Sánchez, Juan Carlos Quintero, and Silvia Ochoa. Flux bal-
ance analysis in the production of clavulanic acid by streptomyces
clavuligerus. Biotechnology progress, 31(5):1226–1236, 2015.
[136] Inga-Marie Schaefer, Robert F Padera, Isaac H Solomon, Sanjat Kanjilal,
Mark M Hammer, Jason L Hornick, and Lynette M Sholl. In situ detec-
tion of sars-cov-2 in lungs and airways of patients with covid-19. Modern
Pathology, 33(11):2104–2114, 2020.
[137] Christophe H Schilling, David Letscher, and Bernhard Ø Palsson. Theory
for the systemic definition of metabolic pathways and their use in inter-
preting metabolic function from a pathway-oriented perspective. Journal
of theoretical biology, 203(3):229–248, 2000.
[138] Katharina Schubert, Evangelos D Karousis, Ahmad Jomaa, Alain Scaiola,
Blanca Echeverria, Lukas-Adrian Gurzeler, Marc Leibundgut, Volker
Thiel, Oliver Mühlemann, and Nenad Ban. Sars-cov-2 nsp1 binds the ri-
bosomal mrna channel to inhibit translation. Nature structural & molecular
biology, 27(10):959–966, 2020.
[139] Stefan Schuster, David A Fell, and Thomas Dandekar. A general definition
of metabolic pathways useful for systematic organization and analysis of
complex metabolic networks. Nature biotechnology, 18(3):326–332, 2000.
[140] Anja Seybert, Annette Hegyi, Stuart G Siddell, and John Ziebuhr. The
human coronavirus 229e superfamily 1 helicase has rna and dna duplex-
unwinding activities with 5-to-3 polarity. Rna, 6(7):1056–1068, 2000.
[141] Jian Shang, Yushun Wan, Chuming Luo, Gang Ye, Qibin Geng, Ashley
Auerbach, and Fang Li. Cell entry mechanisms of sars-cov-2. Proceedings
of the National Academy of Sciences, 117(21):11727–11734, 2020.
[142] Yoshihiro Shimizu, Akio Inoue, Yukihide Tomari, Tsutomu Suzuki,
Takashi Yokogawa, Kazuya Nishikawa, and Takuya Ueda. Cell-free
translation reconstituted with purified components. Nature biotechnology,
19(8):751–755, 2001.
[143] Jonghyeon Shin and Vincent Noireaux. Efficient cell-free expression with
74
the endogenous e. coli rna polymerase and sigma factor 70. Journal of
biological engineering, 4(1):1–9, 2010.
[144] Jonghyeon Shin and Vincent Noireaux. An e. coli cell-free expression tool-
box: application to synthetic gene circuits and artificial cells. ACS synthetic
biology, 1(1):29–41, 2012.
[145] Takehiro Shinoda, Naoko Shinya, Kaori Ito, Yoshiko Ishizuka-Katsura,
Noboru Ohsawa, Takaho Terada, Kunio Hirata, Yoshiaki Kawano, Masaki
Yamamoto, Taisuke Tomita, Yohei Ishibashi, Yoshio Hirabayashi, Tomomi
Kimura-Someya, Mikako Shirouzu, and Shigeyuki Yokoyama. Cell-free
methods to produce structurally intact mammalian membrane proteins.
Scientific Reports, 6(1):30442, 2016.
[146] Bo Shu and Peng Gong. Structural basis of viral rna-dependent rna poly-
merase catalysis and translocation. Proceedings of the National Academy of
Sciences, 113(28):E4005–E4014, 2016.
[147] Robert W Siegel, Scott Adkins, and C Cheng Kao. Sequence-specific
recognition of a subgenomic rna promoter by a viral rna polymerase. Pro-
ceedings of the National Academy of Sciences, 94(21):11238–11243, 1997.
[148] Nicolas Sierro, Yuko Makita, Michiel J. L. de Hoon, and Kenta Nakai.
Dbtbs: a database of transcriptional regulation in bacillus subtilis contain-
ing upstream intergenic conservation information. Nucleic Acids Research,
36:93–96, 2008.
[149] Adam D. Silverman, Ashty S. Karim, and Michael C. Jewett. Cell-free
gene expression: an expanded repertoire of applications. Nature Reviews
Genetics, 21(3):151–170, 2020.
[150] Matthieu Simeoni, Théo Cavinato, Daniel Rodriguez, and David Gatfield.
I (nsp1) ecting sars-cov-2–ribosome interactions. Communications biology,
4(1):1–5, 2021.
[151] Everett Clinton Smith, Nicole R Sexton, and Mark R Denison. Thinking
outside the triangle: replication fidelity of the largest rna viruses. Annual
review of virology, 1:111–132, 2014.
[152] EJ Snijder, E Decroly, and J Ziebuhr. The nonstructural proteins direct-
ing coronavirus rna synthesis and processing. Advances in virus research,
96:59–126, 2016.
75
[153] Mehran Soltani, Brady R. Davis, Hayley Ford, J. Andrew D. Nelson, and
Bradley C. Bundy. Reengineering cell-free protein synthesis as a biosen-
sor: Biosensing with transcription, translation, and protein-folding. Bio-
chemical Engineering Journal, 138:165–171, 2018.
[154] Konstantin MJ Sparrer and Michaela U Gack. Intracellular detection of
viral nucleic acids. Current opinion in microbiology, 26:1–9, 2015.
[155] Alexander S Spirin, Vladimir I Baranov, Lubov A Ryabova, Sergey Yu
Ovodov, and Yuly B Alakhov. A continuous cell-free translation system
capable of producing polypeptides in high yield. Science, 242(4882):1162–
1164, 1988.
[156] DE Steinmeyer and ML Shuler. Structured model for saccharomyces cere-
visiae. Chemical engineering science, 44(9):2017–2030, 1989.
[157] Tobias Stögbauer, Lukas Windhager, Ralf Zimmer, and Joachim O Rädler.
Experiment and mathematical modeling of gene expression dynamics in
a cell-free system. Integrative Biology, 4(5):494–501, 2012.
[158] Lorenzo Subissi, Clara C Posthuma, Axelle Collet, Jessika C Zevenhoven-
Dobbe, Alexander E Gorbalenya, Etienne Decroly, Eric J Snijder, Bruno
Canard, and Isabelle Imbert. One severe acute respiratory syndrome
coronavirus protein complex integrates processive rna polymerase and
exonuclease activities. Proceedings of the National Academy of Sciences,
111(37):E3900–E3909, 2014.
[159] Rahul K Suryawanshi, Raghuram Koganti, Alex Agelidis, Chan-
drashekhar D Patil, and Deepak Shukla. Dysregulation of cell signaling
by sars-cov-2. Trends in microbiology, 2020.
[160] James R Swartz. Expanding biological applications using cell-free
metabolic engineering: an overview. Metabolic engineering, 50:156–172,
2018.
[161] Stanley Tabor. Expression using the t7 rna polymerase/promoter system.
Current protocols in molecular biology, 11(1):16–2, 1990.
[162] Melissa K Takahashi, James Chappell, Clarmyra A Hayes, Zachary Z
Sun, Jongmin Kim, Vipul Singhal, Kevin J Spring, Shaima Al-Khabouri,
Christopher P Fall, Vincent Noireaux, et al. Rapidly characterizing the fast
dynamics of rna genetic circuitry with cell-free transcription–translation
(tx-tl) systems. ACS synthetic biology, 4(5):503–515, 2015.
76
[163] Melissa K Takahashi, Clarmyra A Hayes, James Chappell, Zachary Z Sun,
Richard M Murray, Vincent Noireaux, and Julius B Lucks. Characterizing
and prototyping genetic networks with cell-free transcription–translation
reactions. Methods, 86:60–72, 2015.
[164] Kazuyuki Takai, Tatsuya Sawasaki, and Yaeta Endo. Practical cell-free
protein synthesis system using purified wheat embryos. Nature protocols,
5(2):227–238, 2010.
[165] Tomohisa Tanaka, Wataru Kamitani, Marta L DeDiego, Luis Enjuanes,
and Yoshiharu Matsuura. Severe acute respiratory syndrome coronavirus
nsp1 facilitates efficient propagation in cells through a specific transla-
tional shutoff of host mrna. Journal of virology, 86(20):11128–11137, 2012.
[166] Matthew Zirui Tay, Chek Meng Poh, Laurent Rénia, Paul A MacAry, and
Lisa FP Ng. The trinity of covid-19: immunity, inflammation and inter-
vention. Nature Reviews Immunology, 20(6):363–374, 2020.
[167] Noah D Taylor, Alexander S Garruss, Rocco Moretti, Sum Chan, Mark A
Arbing, Duilio Cascio, Jameson K Rogers, Farren J Isaacs, Sriram Kosuri,
David Baker, et al. Engineering an allosteric transcription factor to re-
spond to new ligands. Nature methods, 13(2):177–183, 2016.
[168] Aartjan JW Te Velthuis, Sjoerd HE Van Den Worm, and Eric J Snijder. The
sars-coronavirus nsp7+ nsp8 complex is a unique multimeric rna poly-
merase capable of both de novo initiation and primer extension. Nucleic
acids research, 40(4):1737–1747, 2012.
[169] Matthias Thoms, Robert Buschauer, Michael Ameismeier, Lennart
Koepke, Timo Denk, Maximilian Hirschenberger, Hanna Kratzat, Manuel
Hayn, Timur Mackens-Kiani, Jingdong Cheng, et al. Structural basis for
translational shutdown and immune evasion by the nsp1 protein of sars-
cov-2. Science, 369(6508):1249–1255, 2020.
[170] Aidan Tinafar, Katariina Jaenes, and Keith Pardee. Synthetic biology goes
cell-free. BMC biology, 17(1):1–14, 2019.
[171] Masaru Tomita, Kenta Hashimoto, Koichi Takahashi, Thomas Simon
Shimizu, Yuri Matsuzaki, Fumihiko Miyoshi, Kanako Saito, Sakura
Tanida, Katsuyuki Yugi, J Craig Venter, et al. E-cell: software environ-
ment for whole-cell simulation. Bioinformatics (Oxford, England), 15(1):72–
84, 1999.
77
[172] Chien-Te K Tseng, Jennifer Tseng, Lucy Perrone, Melissa Worthy,
Vsevolod Popov, and Clarence J Peters. Apical entry and release of severe
acute respiratory syndrome-associated coronavirus in polarized calu-3
lung epithelial cells. Journal of virology, 79(15):9470–9479, 2005.
[173] Olga Tsypik, Roman Makitrynskyy, Agnieszka Bera, Lijiang Song, Wolf-
gang Wohlleben, Victor Fedorenko, and Bohdan Ostash. Role of gntr
family regulatory gene sco1678 in gluconate metabolism in streptomyces
coelicolor m145. BioMed Research International, 2017:9529501–9529501,
2017.
[174] Meaghan A. Valliere, Tyler P. Korman, Nicholas B. Woodall, Gregory A.
Khitrov, Robert E. Taylor, David Baker, and James U. Bowie. A cell-
free platform for the prenylation of natural products and application to
cannabinoid production. Nature Communications, 10(1):565–565, 2019.
[175] Amit Varma and Bernhard O Palsson. Stoichiometric flux balance mod-
els quantitatively predict growth and metabolic by-product secretion in
wild-type escherichia coli w3110. Applied and environmental microbiology,
60(10):3724–3731, 1994.
[176] Jeffrey D. Varner. Gene Regulatory Network Generator in Julia (JuGRN),
April 2020.
[177] Jeffrey D Varner. Lectures notes in advanced biomolecular engineering,
2020.
[178] Sangita Venkataraman, Burra VLS Prasad, and Ramasamy Selvarajan.
Rna dependent rna polymerases: insights from structure, function and
evolution. Viruses, 10(2):76, 2018.
[179] Michael Vilkhovoy, Abhinav Adhikari, Sandra Vadhin, and Jeffrey D
Varner. The evolution of cell free biomanufacturing. Processes, 8(6):675,
2020.
[180] Michael Vilkhovoy, David Dai, Sandra Vadhin, Abhinav Adhikari, and
Jeffrey D Varner. Absolute quantification of cell-free protein synthesis
metabolism by reversed-phase liquid chromatography-mass spectrome-
try. JoVE (Journal of Visualized Experiments), (152):e60329, 2019.
[181] Michael Vilkhovoy, Nicholas Horvath, Che-Hsiao Shih, Joseph A Way-
man, Kara Calhoun, James Swartz, and Jeffrey D Varner. Sequence spe-
78
cific modeling of e. coli cell-free protein synthesis. ACS synthetic biology,
7(8):1844–1857, 2018.
[182] Peter L. Voyvodic and Jerome Bonnet. Cell-free biosensors for biomedical
applications. Current Opinion in Biomedical Engineering, 13:9–15, 2020.
[183] Philip V’kovski, Annika Kratzel, Silvio Steiner, Hanspeter Stalder, and
Volker Thiel. Coronavirus biology and replication: implications for sars-
cov-2. Nature Reviews Microbiology, 19(3):155–170, 2021.
[184] Huihui Wang, Xuemei Li, Tao Li, Shubing Zhang, Lianzi Wang, Xian Wu,
and Jiaqing Liu. The genetic sequence, origin, and diagnosis of sars-cov-
2. European Journal of Clinical Microbiology & Infectious Diseases, pages 1–7,
2020.
[185] Xiaoqiang Wang, Karolina Corin, Philipp Baaske, Christoph J Wienken,
Moran Jerabek-Willemsen, Stefan Duhr, Dieter Braun, and Shuguang
Zhang. Peptide surfactants for cell-free production of functional g
protein-coupled receptors. Proceedings of the National Academy of Sciences,
108(22):9049–9054, 2011.
[186] Joseph A Wayman, Adithya Sagar, and Jeffrey D Varner. Dynamic model-
ing of cell-free biochemical networks using effective kinetic models. Pro-
cesses, 3(1):138–160, 2015.
[187] LA Weber, ER Feman, and C Baglioni. Cell free system from hela cells
active in initiation of protein synthesis. Biochemistry, 14(24):5315–5321,
1975.
[188] Michael Weinfeld, Jane Lee, Gu Ruiqi, Feridoun Karimi-Busheri, Davis
Chen, and Joan Allalunis-Turner. Use of a postlabelling assay to examine
the removal of radiation-induced dna lesions by purified enzymes and
human cell extracts. Mutation Research/Fundamental and Molecular Mecha-
nisms of Mutagenesis, 378(1-2):127–137, 1997.
[189] Wolfgang Wiechert. 13c metabolic flux analysis. Metabolic engineering,
3(3):195–206, 2001.
[190] T Winnick et al. Studies on the mechanism of protein, synthesis in em-
bryonic and tumor tissues. 1. evidence relating to the incorporation of la-
beled amino acids into protein structure in homogenates. Arch. Biochem.,
27, 1950.
79
[191] Aiping Wu, Yousong Peng, Baoying Huang, Xiao Ding, Xianyue Wang,
Peihua Niu, Jing Meng, Zhaozhong Zhu, Zheng Zhang, Jiangyuan Wang,
et al. Genome composition and divergence of the novel coronavirus (2019-
ncov) originating in china. Cell host & microbe, 27(3):325–328, 2020.
[192] P Wu, NG Ray, and ML Shuler. A single-cell model for cho cells a. Annals
of the New York Academy of Sciences, 665(1):152–187, 1992.
[193] Raghavendra Yadavalli and Tobili Sam-Yellowe. Hela based cell free ex-
pression systems for expression of plasmodium rhoptry proteins. JoVE
(Journal of Visualized Experiments), (100):e52772, 2015.
[194] Liming Yan, Ying Zhang, Ji Ge, Litao Zheng, Yan Gao, Tao Wang, Zhihui
Jia, Haofeng Wang, Yucen Huang, Mingyu Li, et al. Architecture of a sars-
cov-2 mini replication and transcription complex. Nature communications,
11(1):1–6, 2020.
[195] Wanchao Yin, Chunyou Mao, Xiaodong Luan, Dan-Dan Shen, Qingya
Shen, Haixia Su, Xiaoxi Wang, Fulai Zhou, Wenfeng Zhao, Minqi Gao,
et al. Structural basis for inhibition of the rna-dependent rna polymerase
from sars-cov-2 by remdesivir. Science, 368(6498):1499–1504, 2020.
[196] Ken-Ichi Yoshida, Yasutaro Fujita, and Akinori Sarai. Missense mutations
in the bacillus subtilis gnt repressor that diminish operator binding ability,
1993.
[197] Ken-ichi Yoshida, Hisanobu Ohmori, Yasuhiko Miwa, and Yasutaro Fu-
jita. Bacillus subtilis gnt repressor mutants that diminish gluconate-
binding ability. Journal of bacteriology, 177(16):4813–4816, 1995.
[198] Francis K Yoshimoto. The proteins of severe acute respiratory syndrome
coronavirus-2 (sars cov-2 or n-cov19), the cause of covid-19. The protein
journal, 39:198–216, 2020.
[199] Anne Zemella, Lena Thoring, Christian Hoffmeister, and Stefan Kubick.
Cell-free protein synthesis: pros and cons of prokaryotic and eukaryotic
systems. ChemBioChem, 16(17):2420, 2015.
[200] Yang Zhang. Genome wide structure and function modeling of sars-cov-
2. https://zhanglab.ccmb.med.umich.edu/COVID-19/, 2020.
[201] Ying Zhang, Ines Thiele, Dana Weekes, Zhanwen Li, Lukasz Jaroszewski,
80
Krzysztof Ginalski, Ashley M Deacon, John Wooley, Scott A Lesley, Ian A
Wilson, et al. Three-dimensional structural view of the central metabolic
network of thermotoga maritima. Science, 325(5947):1544–1549, 2009.
[202] Wei Zhu, Catherine Z Chen, Kirill Gorshkov, Miao Xu, Donald C Lo, and
Wei Zheng. Rna-dependent rna polymerase as a target for covid-19 drug
discovery. SLAS DISCOVERY: Advancing the Science of Drug Discovery,
25(10):1141–1151, 2020.
81