Catching Rabies by the Toe:
An Investigation into the Toehold Switch as a Sensor for Rabies virus.
by
Nadja Hedrich

A THESIS SUBMITTED IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
BACHELOR OF SCIENCE (HONS.)
in the
Departments of Biological and
Physical Sciences
(Biology)

This thesis has been accepted as conforming to the required standards by:
Donald Nelson (Ph.D.), Thesis Supervisor, Dept. Biological Sciences
Bruno Cinel (Ph.D.), Co-Supervisor, Dept. Physical Sciences
Jonathan Van Hamme (Ph.D.), Examining Committee member, Dept. Biological Sciences

Dated this 28th​ ​ day of April, 2016, in Kamloops, British Columbia, Canada

i

ABSTRACT

Rabies is a disease of the central nervous system caused by the Rabies virus. Current methods to
test for Rabies are costly, time consuming, and not sensitive enough to identify the disease early
enough for treatment. A novel mechanism of virus identification that has shown promise for
Ebola and Zika virus detection is the toehold switch. Toehold switches are RNA-based
constructs that allow the visual identification of genomic targets of interest. They are composed
of a switch RNA and a trigger RNA. In the presence of the trigger, a change in secondary
structure causes the reporter gene following the switch to be expressed. When the trigger is not
present, the gene is repressed. This project designed a toehold switch for use as a Rabies virus
sensor. First, a region of the Rabies genome was identified as a promising trigger sequence.
From this sequence, a complementary switch was designed with GFP as a reporter gene. The
designed switch was constructed and inserted into a pBluescript plasmid backbone. The next
steps of this project would be to develop an assay to test the effectiveness of the switch to
identify Rabies virus.
Thesis Supervisor: Dr. Don Nelson

1

ACKNOWLEDGEMENTS
Thank you to my supervisors for their guidance and honesty, and making sure this project was
very much my own, and to Thompson Rivers University for the labs and facilities. Thank you to
the UREAP for funding this project, and to Jason, for proofreading all of these pages many times
over. And finally, thank you to my family and Jon for all of their support and understanding.

2

TABLE OF CONTENTS
ABSTRACT

1

ACKNOWLEDGEMENTS
LIST OF FIGURES

2
6

INTRODUCTION

8

MATERIALS AND METHODS
I.​Identifying Targets in the Rabies Genome for Trigger Sequence
II. Design of Complementary Toehold Switches
III. Design of Final Toehold Construct
IV. Plasmid Production
A. Transformation
B. Restriction Digests and Fragment Isolation
C. Ligations and Visualization
D. Determination of Plasmid Identity

14
14
15
16
17
17
18
18
19

RESULTS AND DISCUSSION
I. ​Identifying Targets in the Rabies Genome for Trigger Sequence
II. Design of Corresponding Toehold Switches
III. Design of Final Toehold Construct

20
20
26
29

CONCLUSIONS AND FUTURE WORK

39

REFERENCES

42

APPENDIX
44
A: Sequences, locations and penalties assigned to the top 8 trigger sequences.
44
B. Sequences of the top 8 switches corresponding to 8 top triggers.
45
C: Results of transformation of DH5-α E. coli with three different plasmids (pBS, pGFP,
and pToe, as well as a no DNA control.
46
D: Agarose gel quantification of plasmids isolated using the Qiagen maxiprep kit and
protocol. Plasmids were linearized with Kpn1 and 1 µL of DNA was loaded and 500 ng
of lambda HindIII marker. The gel was run for 1.5 hrs at 65 V.
46
E. Results of transformations done to test ligation reaction efficiency. Competent E. coli
were heat transformed and grown on Lb agar plates overnight.
47
F: Annotated partial sequence of the constructed positive control plasmid.
48
G​: Annotated partial sequence of the constructed experimental toehold plasmid.
50

3

LIST OF FIGURES
Figure 1. Schematic of the general Rabies virus lifecycle. Boxes indicate genes in the Rabies genome.
Figure 2. The mechanism of action of toehold switches. The toehold without the trigger does not express
the gene as translation is blocked by the hairpin loop. When the trigger is added to the switch, the hairpin
is disrupted, opening up the ribosome binding site and allowing translation to occur and the gene to be
expressed.
Figure 3. Secondary structure at 37oC of the eight top trigger candidate sequences visualized using
NUPACK software. Thermodynamic stability of the structure is shown below each trigger as free energy
of secondary structure.
Figure 4. Trigger candidate dimer formation at 37oC visualized using NUPACK software. The free
energy of the secondary structure is shown below each dimer.
Figure 5. Designed switch sequences corresponding to the top eight identified triggers. Switches were
visualized using NUPACK software.
Figure 6. Secondary structure of the chosen trigger sequence, corresponding switch sequence and the
trigger switch complex, visualized using NUPACK software.
Figure 7. A. General schematic of the ordered toehold switch construct as optimized by Green et al. 2014.
B. Actual purchased nucleotide sequence containing the switch, and GFPuv as a reporter gene. Colours
represent different portions of the sequence. Beige=stuffer nucleotides, blue=restriction site, orange=
GGG site, grey=switch sequence, cyan=ribosome binding site, purple=start sequence, red=linker region,
green=GFPuv, magenta=terminator loop.

4

Figure 8. Plasmid maps of the three plasmids used in subcloning experiments, pBS, pGFPuv, and pToe.
Important features and their location are shown on each map.
Figure 9. Cloning experiment to create the control plasmid. The plasmid was created by combining the
pBS backbone and the GFP gene from pGFPuv using a restriction digest, fragment isolation, and ligation.
Figure 10. Cloning experiment to create the experimental toehold plasmid. The plasmid was created by
combining the pBS backbone and the designed pToe insert using a restriction digest, fragment isolation,
and ligation.
Figure 11. Restriction digests performed to support the identity of the constructed experimental toehold
plasmid.

5

INTRODUCTION
Rabies is a disease of the central nervous system caused by the Rabies virus​. ​In humans,
it is almost always transmitted by the bite of an infected animal. Rabies is most commonly
transmitted by dogs, and is fatal unless treated immediately after infection. There are tens of
thousands of human deaths worldwide annually due to Rabies, most in developing countries
(Rupprecht). Once the virus enters the body, it infects the peripheral nervous system, then
progressively travels through the nervous system cells to the central nervous system. It
aggregates in the central nervous system at first, and then spreads to other organs late in the
disease (Galveston, 1996). Symptoms are not seen until after the incubation period, which
generally lasts from 30 to 60 days, but varies depending on the location and severity of the
infection site (Rupprecht). Symptoms come in two forms, consisting of either paralysis or
hyperactivity, but both forms result in comas and ultimately respiratory arrest Galveston, 1996).
The virus itself is a member of the ​Lyssavirus ​genus and family ​Rhabdoviridae. ​It is
rod-shaped and consists of a nucleocapsid (containing proteins and the RNA genome) and an
outer envelope consisting of a lipid bilayer with glycoprotein spikes. The viral genome consists
of single stranded, nonsegmented negative sense RNA; this means that it must be transcribed
into positive mRNA before it can be translated into proteins. The genome is about 12kB in
length, and codes for five proteins: L(RNA-polymerase), N(nucleoprotein),
NS(RNA-polymerase associated) aka P (phosphase), M(matrix), and G(glycoprotein)
(​www.cdc.gov​). Figure 1 shows the positions of these genes in the genome and the changes to
the Rabies genome throughout its lifecycle. The G protein forms the glycoprotein spikes on the

6

surface of the virus, which allow the virus to adsorb onto and penetrate other cells. After the
virus has entered a host cell, it releases its nucleocapsid and the L protein transcribes the
genomic RNA to make mRNAs. These mRNAs are then translated into proteins using the host
cell's ribosomes. Once a certain amount of viral proteins are produced, the viral polymerase starts
making full length positive copies of the RNA genome, from which more negative RNA strands
are made. Then the proteins and RNA aggregate and form new viruses with the help of the N
protein, which leave the cell and infect neighbouring cells. Once the virus reaches the brain,
aggregation of the virus in brain tissue causes inflammation of the tissue, resulting in the
psychological symptoms associated with Rabies virus infection. However, the exact
physiological mechanism that causes this inflammation is not yet known (​www.cdc.gov​ ,
www.who.int​).

7

Figure 1: Schematic of the general Rabies virus lifecycle. Boxes indicate genes in the Rabies genome.

A vaccine does exist, and a dog vaccination program has been successful in drastically lowering
the cases of rabies infection in developed countries; however, the virus is difficult to detect, as
symptoms only appear late in the disease (Rupprecht). Traditional methods of screening for
rabies include direct immunofluorescence tests on skin biopsies to look for rabies antigen.

8

Another method of testing is culturing the virus from saliva samples in neuroblastoma cells or
rats, but this method is expensive. Screening for rabies antibody in cerebrospinal fluid is also
used. Finally, the most common testing method for Rabies in animals is to examine a post
mortem brain tissue sample (​www.who.int​). However, all of these tests are limited, as they are
not very sensitive, and may not detect the virus early on in the disease. As such, a method to
screen for Rabies virus that can identify low virus concentrations, and does not require the death
of the host organism is desirable. One method that seems like a promising solution to this
problem due to its sensitivity and affordability is the toehold switch.
Toehold switches are a new form of translational gene regulators (Church et al., 2014). This
means that they can regulate when proteins are being expressed by inhibiting or promoting
translation of a certain gene, normally through changes in secondary structure. Toehold switches
consist of two strands of RNA; a switch and a trigger. The switch RNA consists of a specific
hairpin loop secondary structure that contains the ribosome binding site at the top of the loop and
the start sequence, as well as a toehold region at the base of the hairpin loop. The trigger RNA
sequence is complementary to this toehold sequence and part of the hairpin loop stem. When the
switch is present in a cell without the trigger, the ribosome binding site is hidden and translation
does not occur. However, when the trigger is present, it will bind to the toehold site and disrupt
the hairpin, making the site accessible, thus allowing translation to occur (see Figure 2).
The principle behind this mechanism is DNA strand displacement. When two strands are
complementary to the same sequence, they will compete with each other to bind to that
sequence. However, if one of the free strands is bound to a region adjacent to the contested
portion (a toehold region), then it will have an advantage. The double stranded toehold portion

9

will increase the local concentration of the toehold mediated strand, allowing it to outcompete
the other strand. Even when one strand is already bound, as in a toehold switch, the second
strand will displace the first through branch migration. The opposite reaction is much slower
because the lack of a toehold section does not allow for colocalization of the strand therefore the
toehold strand will displace the incumbent strand (Zhang and Winfree, 2009).
These `toehold switches have a wide dynamic range and low off target effects (Green et al.
2014). Dynamic range refers to the difference between the smallest, and the largest visible
signal, dependent on concentration of trigger. A wide dynamic range is favourable so that small
amounts of viral genome will be identified as having turned the switch “on”, and the amount of
virus can be determined, giving an idea of disease progression. Off target effects refer to the
ability of similar triggers to turn on switches other than the one they were designed to bind with.
This is unfavourable because it can lead to the formation of false positive.

10

Figure 2. The mechanism of action of toehold switches. The toehold without the trigger does not express
the gene as translation is blocked by the hairpin loop. When the trigger is added to the switch, the hairpin
is disrupted, opening up the ribosome binding site and allowing translation to occur and the gene to be
expressed (Green et al., 2014).

These switches have the potential to be modified to act as sensors to detect specific viral
genomes (Pardee et al., 2014). To do this, the gene being repressed would be a reporter gene that
codes for proteins whose presence is easily visualized, and the trigger used would be a piece of
endogenous RNA, such as a part of a viral genome. When these modifications are made, this
mechanism can be used to identify the presence of specific genetic sequences. Recently, toehold
switches have been used to sense Ebola virus and Zika virus (Pardee et al., 2014, 2015). Virus
sensing was done by using a portion of the viral genomes as the trigger and designing switches to
report the presence of the viral genome using green fluorescent protein (GFP) or B-galactosidase
as reporters. Toehold switches are sensitive assays that can also be applied to cell free systems
which makes them quick and easy to use outside of a laboratory environment.

For my Honours project, I designed a toehold switch to detect Rabies RNA. This involved
finding a portion of the Rabies genome to use as a trigger, and developing a switch that would
express a reporter protein in the presence of Rabies virus genome, but would not express this
reporter when Rabies virus is not present. To do this, I first identified potential targets in the
Rabies viral genome for use as a unique trigger sequence, then designed a toehold switch and
trigger sequence ​in silico​ from the chosen target. After the design was complete, I purchased the
designed construct, and inserted it into a pBluescript plasmid backbone, through a series of
transformations, restriction enzyme digests, and ligation reactions.
11

MATERIALS AND METHODS

I.

Identifying Targets in the Rabies Genome for Trigger Sequence

To identify targets for the trigger sequence in the Rabies genome, the nucleotide sequence for the
SAD B19 strain of Rabies from NCBI was used. The VIPR primer search engine was then used
to search for targets with the following specifications (Pardee et al., 2016):

1. In L gene (5414 to 11797)
2. Length 30 nucleotides
3. 40-60% GC content
4. Melting temperature >41C
5. No runs of 4 or more of the same nucleotides
6. Sequence of SSW (or WSS, SWS) at the 13-15th nucleotides
(where S=strong=G or C, and W=weak=A or T)

7. Minimal secondary structure
8. Minimal dimer formation
9. Higher AT at 3’ end, higher GC at 5’end (6 nucleotides)
Criteria 1, 2, and 3 were entered into the primer engine search criteria, returning only primers
containing those specifications. Fifty primers were returned at a time, and manually examined
for criteria 4.

12

The returned sequences were then visualized using the NUPACK software system and
examined for criteria 6 and 7 (​http://www.nupack.org​). NUPACK software is a free online
platform that allows for the easy visualization of RNA secondary structure; a single sequence can
be inputted to examine intramolecular structure, or several sequences can be inputted to examine
binding between and within sequences. Several parameters can be set, including temperature, ion
concentrations, and strand concentrations. In the software, the parameters were set to a
temperature of 37°C, 1.0 M Na​+​, and 0.0 M Mg​2+​ (Serra and Turner, 1995), and assumed strand
concentrations of 100 nM. The sequences were also visually examined and ranked using criteria
8 and 9.

II. Design of Complementary Toehold Switches

Toehold switches complementary to the resulting trigger sequences were then designed,
following Green et al. 2014. The switch contained a GGG sequence, complementary sequence to
the trigger, a ribosome binding site, a linker sequence, and GFPuv as a reporter gene. The GFPuv
gene has a known sequence (​https://www.ncbi.nlm.nih.gov/nuccore/1490531​) , so no extra
design was required. These designed switches were then examined using NUPACK software to
look for unwanted secondary structures​ and unwanted stop codons. Unwanted secondary
structure would include any binding of the toehold region to itself, or to the linker region,
thereby blocking trigger binding. Stop codons would interfere with the expression of the GFP by
prematurely terminating the ribosome. The binding of the trigger and sequences was examined in
NUPACK using a ​temperature of 37°C, 1.0 M Na​+​, 0.0 M Mg​2+​, and assumed strand

13

concentrations of 100 nM.

The designed switch and trigger pairs were then ranked according to the selection criteria
mentioned above, with an emphasis on favourable secondary structure (no trigger secondary
structure, minimal dimer formation, and no switch secondary structure in the toehold region).
The most promising switch/trigger combination was then examined using NCBI BLAST
(​https://blast.ncbi.nlm.nih.gov​) to determine uniqueness. The trigger sequence was searched
against all sequences in the database excluding only Rabies virus. resulting sequences were then
returned along with their corresponding similarities to the trigger sequence.

III. Design of Final Toehold Construct

Once a suitable trigger/switch combination was determined using the previously
mentioned steps, the linker sequence, GFPuv sequence, and a terminator loop sequence was
added, according to Green et al. 2104 (See Figure 7). Restriction enzyme sites corresponding the
EcorI and KpnI were added, EcoR1 on the 3’ end, and KpnI at the 5’ end, as well as six
nucleotides at the terminal ends of the sequence to increase the restriction enzyme activity.
Synthesis of the entire construct, now consisting of 878 nucleotides was ordered from Thermo
Fisher Scientific Invitrogen (​https://www.thermofisher.com​) (full sequence in Appendix C).

14

IV. Plasmid Production

Two plasmids; a positive control plasmid and an experimental plasmid, were constructed once
the ordered switch sequence arrived. The plasmids were constructed from previously isolated
stock solutions of pBluescript, pGFPuv, and the ordered switch plasmid, which will be
referenced to as pToe. The two plasmids were constructed using a series of transformation,
fermentation, restriction reactions, fragment isolations, and ligation reactions according to
standard molecular biological approaches. First, a control plasmid consisting of a pBluescript
backbone with the GFP gene from the GFPuv plasmid was created. The GFP gene was inserted
into the multiple cloning site using EcoRI and KpnI restriction sites. The experimental plasmid
was constructed in the same way as the control, but the insert consisted of the ordered switch
sequence and the GFP reporter gene.

A. Transformation
First, previously isolated pGFP and PBS as well as the purchased toehold plasmid (pToe)
were transformed into DH5-α competent ​E. coli​ cells using standard heat transformation
protocols (Fasman 1989). The transformed bacteria were plated on Lb +Amp plates and
incubated overnight at 37 °C. The next day 5 mL of LB broth containing ampicillin was
inoculated with one colony. The 5 mL broth was incubated overnight at 37 °C and then used to
inoculate a 100 mL broth which was incubated overnight. A maxiprep using the Qiagen plasmid
maxiprep kit and protocol was performed to isolate the plasmids. Following this, samples were
15

linearized using EcoR1 and visualized on a 1% agarose gel to determine concentration (Qiagen
Kits).

B. Restriction Digests and Fragment Isolation
The isolated pBS, pToe, and pGFP were then both cut using a two-step restriction
digestion. The first digestion was performed using either EcoRI or KpnI (Fasman 1989). The
reactions were incubated at 37 °C for 1 hour. The linearized plasmid DNA was then isolated
using isopropanol, spun down at 14,000 rpm for one minute, then washed with 70% ethanol,
spun twice at 14000 rpm, and dried. A second restriction enzyme digestion was then performed
using the either EcoRI or KpnI (the opposite as previously used).
The cut plasmids were then visualized on a 1% agarose gel, and the pBS backbone
fragment, and GFP gene fragment were isolated using the QiaEx II fragment isolation kit and
protocol (QIAEX® II Handbook). The isolated fragments were then visualized on a 1% agarose
gel to determine efficiency of isolation and DNA concentration.

C. Ligations and Visualization
Next the isolated fragments were combined in a ligation reaction (Fasman 1989). The
reactions were incubated overnight at room temperature. The ligated plasmids were then heat
transformed into competent DH5-α ​E. coli​ cells. The plates were incubated at 37°C overnight,
and then examined the following day for the presence of white colonies. Several controls were
performed testing for linearization efficiency, and transformation efficiency.

16

D. Determination of Plasmid Identity
In order to determine whether the intended plasmids were successfully produced, white
colonies (that do not have a working lacZ gene) indicating successful insertion into the pBS
multiple cloning site from the ligated plasmid plates were restreaked, and then used to inoculate
5mL Lb + Amp broth. After being incubated at 37°C overnight, a miniprep was performed on 1
mL of the 5mL culture, and 1 mL was used to inoculate 100 mL of Lb +Amp broth which was
incubated at 37°C overnight. The miniprep was performed using Qiagen Plasmid prep kit buffers
and standard protocol (Fasman 1989).
A plasmid maxiprep was then performed using the 100 mL culture, following the
protocol and using the contents of the Qiagen Plasmid maxiprep kit (Qiagen kit). The isolated
plasmids from the maxiprep were then visualized on a 1% agarose gel to determine a rough
estimate of yield.

A series of restriction digests were then performed on the isolated plasmids following the
previously discussed protocol. The fragments were then visualized on a 1% agarose gel with
ethidium bromide under UV light.

17

RESULTS AND DISCUSSION
The purpose of this project was to design and create a novel toehold switch to be used as a sensor
for the Rabies virus.

I.

Identifying Targets in the Rabies Genome for Trigger Sequence

The first step in this project was to find targets in the Rabies genome that would be promising for
use as trigger sequences. Several criteria that were known to optimize the trigger binding to and
“turning on” the switch sequence were used as parameters to identify promising targets (Green et
al,. 2013).
The first of these parameters was that it was found in the L gene. The L gene was chosen because
it is highly conserved between Rabies species, suggesting that the trigger would we common to
multiple strains of Rabies (Tordo et al. 2017). The length of 30 nucleotides was found to produce
the best strand displacement, as shorter toeholds showed less successful strand displacement, and
longer toeholds did not significantly increase displacement. The GC content is important because
GC bonds are stronger than AT bonds. Stronger binding to the toehold increases strand
displacement (Xu et al., 2013). Similarly the melting temperature optimizes strand binding and
displacement (Green et al, 2014).
After inputting these initial criteria into the ViPR primer search engine, approximately 300
primers were examined. Nucleotide runs of four or more were eliminated because they are
unstable and increase transcription and translation errors (Ackermann et al., 2006). Next the
sequence of two strong and one weak base at the 13th-15th nucleotides were examined. This
sequence has strong enough binding for the stopp loop structure to be stable on its own, but weak
18

enough to allow for optimal strand displacement when the trigger binds the toehold. Using those
criteria to eliminate primers, the top eight primers were examined using NUPACK software.
NUPACK allows for easy visualization of RNA secondary structure, so it was used to examine
secondary structure of the eight trigger sequences.

19

1

2

3

4

5

6

7

8

Figure 3. Secondary structure at 37oC of the eight top trigger candidate sequences visualized using
NUPACK software. Thermodynamic stability of the structure is shown below each trigger as free energy
of secondary structure.

20

The results from the NUPACK software inputting only the trigger sequences are seen in Figure 3
above. Secondary structure varied from no nucleotide pairs in trigger 4, to eight nucleotide pairs
in trigger 1. From this initial visualization, trigger 4 seemed most promising, as it contained no
secondary structure that would interfere with the binding of the trigger to the switch. Triggers 3
and 5 also showed some promise, as they had minimal secondary structure (only two paired
nucleotides), and these paired nucleotides were not close to the start of the sequence, so the
trigger would still potentially bind with the switch RNA. The other switches showed increased
secondary structure; and in some cases, as in trigger 7, had large stem loop structures that would
potentially inhibit binding of the trigger to the toehold switch.
These eight trigger sequences were then examined again using NUPACK software but this time
for dimer formation, the binding of one trigger sequence to another identical trigger molecule.
This is important to consider because an ideal trigger will bind only to its corresponding switch.
If the trigger binds to another copy of itself, forming a dimer, there will be less binding of the
trigger to the switch, thereby making the sensor less sensitive. In the context of the entire Rabies
genome, this is less of a concern, however, dimer formation could lead to problems when testing
the switches with shorter sequences (oligonucleotides), giving potential false negatives.

21

1

2

3

n/a

4

5

6

n/a

7

8

Figure 4. Trigger candidate dimer formation at 37oC visualized using NUPACK software. The free
energy of the secondary structure is shown below each dimer.

22

When the trigger sequences were examined for dimerization, there was a range of bonding seen,
from no dimerization, to 22 nucleotide pairs (See Figure 4). As dimerization tendency increases,
the availability of the trigger sequence to bind to the switch sequence decreases, as some of the
trigger molecules are bonded to each other. Therefore, a trigger sequence that shows no dimer
formation would be ideal.
As the trigger is a portion of the Rabies genome, this is not a large concern in practical
applications, as dimerization of the entire Rabies virus genome is unlikely, however, it may be
significant in a testing assay using oligonucleotides, where dimerization is more probable.
Triggers 2 and 4 showed no dimer formation, suggesting that the trigger would preferentially
bind to the switch sequence than another trigger. Trigger 5 and 6 both had fourteen bonds in their
dimers, with several bubbles where no bonds were present, and no bonds near the ends,
suggesting that binding with the switch sequence may still be preferred over dimerization.
Trigger 8 had only six bonds in its dimer form, but the stem loops seen in the secondary structure
of the trigger sequence were also present. The other three triggers had more bonds in the dimers,
up to 22 bonds in trigger 1. These would not be as promising as trigger sequences because there
may be a significant decrease in trigger/switch binding due to high levels of dimerization.
From the examination of trigger sequence secondary structure and dimerization tendency, trigger
4 seemed the most promising, as it contained no secondary structure, and did not form dimers.
Trigger 4 also had no dimer formation, but contained significant secondary structure, a stem-loop
of seven nucleotide pairs that would potentially inhibit its binding to the switch sequence.
Triggers 3 and 5 had minimal secondary structure (only two bonds), but both showed some
dimer formation, with 18 and 14 bonds in the dimer form, respectively. These are still promising

23

trigger sequences, but the possibility of dimerization may decrease switch interactions in a
testing assay.
Next, the switch sequences corresponding to these triggers were designed according to layouts
determined by previous research (Green et al., 2014), and the switch sequences were examined
NUPACK software.

II. Design of Corresponding Toehold Switches
Toehold switches corresponding to the top eight trigger sequences were designed according to
previous specifications (Green et al. 2014). These switches were then examined using NUPACK
software for ideal stem-loop structure, and any in frame stop sequences.

24

1

2

3

4

5

6

7

8

Figure 5. Designed switch sequences corresponding to the top eight identified triggers. Switches were
visualized using NUPACK software.

25

When the switch sequences corresponding to the top eight trigger sequences were visually
examined, several different RNA secondary structures were seen (see Figure 5). Switches 4 and
7 showed a secondary structure consisting of the ideal stem-loop structure with 15 nucleotides
bonded, a three nucleotide bubble at the start site, and no secondary structure in the toehold
section before the loop (see Figure 7). The small stem-loop in the linker sequence would not
inhibit trigger binding, or subsequent ribosome action. Switches 1 and 8 had small stem loops in
the toehold regions, but the small amount of bonding would likely still allow for successful
trigger binding. Some of the other switch sequences had less favourable secondary structures
however, with switch 2 showing the toehold region binding to the linker region, which would
likely decrease the ability of the trigger to bind. Switch 6 has a large stem-loop structure in the
toehold section, which again would likely decrease trigger bonding, making it an unfavourable
candidate. Finally, switches 3 and 5 had shifted stem loops and some unwanted secondary
structure in their toehold regions, also making them less than ideal candidates for a successful
trigger/switch combination.
Using this information, along with trigger secondary structure and dimer formation, trigger 4 and
its corresponding switch was chosen as the best candidate for use as a Rabies virus sensor. The
free energies of the secondary structure, indicating the stability of secondary enthalpies of the
dimers and trigger sequences were also compared (see Figures 3, 4, and 5).
The chosen trigger sequence was then examined in NCBI BLAST, and no other sequences
outside of the Rabies genome were found to match the sequence 100% when queried against the
entire NCBI database excluding other Rabies viruses. The closest matches had <50% similarity,
and none had the toehold region the same as the trigger sequence. Within different Rabies virus
26

strains, the sequence was conserved, indicating that many different strains of Rabies virus could
be detected with this trigger.

III. Design of Final Toehold Construct

Using the sequence of trigger 4 and its corresponding switch, which were found to be unique
using NCBI blast, contained no unwanted secondary structure, and formed no dimers, a final
toehold construct was designed following a previously optimized design (Green et al., 2014).

27

Figure 6. Secondary structure of the chosen trigger sequence, corresponding switch sequence and the
trigger switch complex, visualized using NUPACK software.

Figure 6 (above) shows the secondary structure of trigger 4, its corresponding switch, and the
trigger/switch complex more closely. As was mentioned before, the trigger shows no secondary
structure and no dimer formation, making it an ideal candidate because there was nothing present
28

that would inhibit the trigger bonding to the switch. The switch sequence also does not contain
any unwanted secondary structure in the toehold region, again making it ideal for binding to the
trigger; additionally, the stem loop structure follows the optimized design discussed by Green et
al. 2014. Finally, the trigger switch combination shows the trigger fully binding to the switch,
and linearizing the stem loop structure. There is some secondary structure seen in the area behind
the trigger binding, but it does not seem to be likely to inhibit ribosome binding as it consists of
very few nucleotides, and will not be very stable. There are also no stop sequences present in the
switch sequence, making this a very promising candidate for continuing the project.
Using this switch sequence as a starting point, a construct was designed to sense Rabies virus.
Behind the linker sequence, the sequence for GFP was added, and a terminator loop was
included behind the GFP. GFP is an easily visualized reporter gene. It does not require anything
more than a UV light to be visualized, and is a protein, so no other substrates are necessary,
unlike some other reporters such as LacZ. Also, because GFP is a protein the amount of green
fluorescence is proportional to the amount of GFP produced, and to the amount of trigger present
(​Feilmeier)​. The terminator loop will end translation by the RNA polymerase after the GFP is
expressed. At the 3’ end of the sequence, an EcoRI restriction site was added, and on the 5’ end a
KpnI restriction site was added, along with 6 more stuffer nucleotides at each end (see Figure 7).
The restriction sites allow for the sequence to be easily moved between plasmids using standard
molecular cloning techniques. Two different restriction sites were used so that the sequence can
only be inserted in one orientation, eliminating the possibility of a reversed insert. The stuffer
nucleotides are necessary for the restriction enzyme to effectively cut the DNA at the restriction
site.

29

30

The switch construct, as seen in Figure 7 was designed and then ordered from Thermo Fisher
Scientific. There, they combined pre-prepared synthetic oligonucleotides into the desired
nucleotide sequence and inserted it into a plasmid containing ampicillin resistance and a ColE1
promoter. The sequence was examined to ensure that it was the same as the ordered sequence,
and the construct was delivered as an insert in a plasmid termed pToe (see Figure 8).

31

IV. Plasmid Production

Figure 8. Plasmid maps of the three plasmids used in subcloning experiments, pBS, pGFPuv, and pToe.
Important features and their location are shown on each map.

Figure 8 above shows the main elements of each of the three plasmids used in the subcloning

32

experiment. Each plasmid has a ColE1 origin of replication, and ampicillin resistance. The pBS
plasmid also contains a LacZ gene spanning a multiple cloning site (MCs), which contains EcorI
and KpnI restriction enzyme sites. Normally, the LacZ gene will cause the bacterial colonies to
appear blue when plated with IPTG and X-gal, however, when a fragment is inserted in the
MCS, the LacZ gene will not be functional, and any colonies plates on IPTG and X-gal will not
be blue. The pBS also contains a T7 RNA polymerase promoter region, allowing for selected
promotion of the region following. This is why pBS was chosen as a backbone for the
constructed plasmids.
pGFP has a ColE1 site, Amp resistance, and the GFP gene flanked by KpnI and EcoRI restriction
sites, and pToe is very similar, but it contains the designed toehold construct between the
restriction sites.

33

Figure 9. Cloning experiment to create the control plasmid. The plasmid was created by combining the pBS
backbone and the GFP gene from pGFPuv using a restriction digest, fragment isolation, and ligation.

Figure 9 shows the cloning experiment to create the positive control plasmid. Both pBS and

34

pGFPuv were cut with EcoRI and KpnI. Following this, the backbone of pBS, and the GFp insert
of pGFPuv were isolated and ligated. The positive control contains all of the elements of pBS
such as Amp resistance, ColE1 origin, and the T7 promoter. However, the LacZ gene has been
disrupted, and GFP has been inserted. This plasmid works as a positive control because when
induced by the T7 promoter, GFP should be expressed, just as the activated trigger and switch
complex should express GFP.

35

Figure 10. Cloning experiment to create the experimental toehold plasmid. The plasmid was created by combining
the pBS backbone and the designed pToe insert using a restriction digest, fragment isolation, and ligation.

Figure 10 shows the cloning experiment to create the experimental plasmid. This time, pBS and

36

pToe were cut with EcoRI and KpnI. Following this, the backbone of pBS, and the toehold
construct insert of pToe were isolated and ligated. Again. the experimental plasmid contains all
of the elements of pBS such as Amp resistance, ColE1 origin, and the T7 promoter and a
disrupted LacZ gene. However, this time the toehold construct has been inserted into the MCS.

Figure 11. Restriction digests performed to support the identity of the constructed experimental toehold
plasmid.

To determine whether the constructed experimental toehold plasmid was successfully created, a
series of restriction digests was performed: KpnI, EcoRI, ClaI, HindIII, EcoRI and KpnI, EcoRI
and ClaI, KpnI and ClaI, and BglII. The results of these restriction enzyme reactions are seen in
Figure 11 above. The HindIII and BglII lanes show only closed circular plasmid, as expected
because there should be no BglII or HindIII sites present in the plasmid. The EcoRI, KpnI, and
ClaI all linearized the plasmid, as expected for enzymes with only a single restriction site present
in the plasmid. The single band seen in the EcoRI/ClaI digest is also as we would expected, with
a single large fragment seen, very slightly smaller than the linearized plasmid. The second
fragment would only consist of approximately 10 nucleotides, so we do not expect to see that
37

fragment. The KpnI/ClaI and KpnI/EcoRI lanes look very similar, with two bands, one
corresponding to the toehold and GFP insert, and the other, larger band corresponding to the pBS
backbone. However, we expect the KpnI/ClaI was expected to have a slightly larger GFP insert,
as was seen. Therefore, these restriction enzyme reaction provide support that the toehold
plasmid was successfully created.

CONCLUSIONS AND FUTURE WORK
Due to the shortcomings of present forms of Rabies virus identification, the development
of a novel detection method is of considerable interest. Toehold switches are a promising novel
mechanism for virus identification that have recently been used in the Ebola and Zika virus
epidemics with success. Not only are they sensitive and affordable, they can be portable, which
makes them especially convenient for third world countries. This is important for Rabies virus,
because it is third world countries, lacking in proper vaccination programs, that would most
benefit from this sensor. This project made strides to developing a new way of sensing Rabies
virus using toehold switch technology.
Following the guidelines set out by Green et al. 2014, a sequence was selected from the
Rabies genome for use as a trigger. The chosen sequence was theoretically ideal, lacking in
secondary structure, and unable to dimerize, ensuring that the trigger and switch binding would
not be hindered by unwanted secondary structure. The sequence was also examined using NCBI
BLAST, and found to be unique to Rabies virus, though conserved between different Rabies
virus strains. Due to these qualities the sequence was chosen as a trigger.

38

From this, a corresponding switch was designed ​in silico​, and it was found to lack
unwanted secondary structure in its toehold region and any in-frame stop codons, making it a
promising candidate for a switch sequence. GFPuv was chosen to be a reporter gene because it is
easily visualized, and has previously been used as a reporter for toehold switches with success.
Restriction enzyme sites were added to the terminal ends of the trigger sequence in order to make
subsequent ligation into a pBS backbone easier, and the whole construct was purchased from
Invitrogen.
A series of transformation, restriction, and ligation experiments successfully inserted the
purchased toehold construct into a pBS backbone. The pBS was chosen as a backbone because of
the presence of a T7 RNA polymerase promoter region, allowing for selective expression of the
toehold sequence.
There were some issues with DNA amount due to low yield returned from the plasmid
preps, making some visualization of the restriction and ligation reactions unsuccessful. However,
there was sufficient DNA for the reactions to occur. Redoing the plasmid prep early in the
experiment, or doing two preps and concentrating the DNA may have avoided these problems.
In order to test whether the toehold switch successfully senses the Rabies sequence, a
series of testing assays will then be performed to examine the ability of the trigger to induce
expression of GFP by binding with the switch. Initially, an assay using oligonucleotide
sequences of the 30 nucleotide trigger sequence will be performed. In order to test out toehold
function with an oligonucleotide, DNA-RNA bonding in a cell free system will be used. Cell free
systems will be created through lysis of E. coli cells through sonication, and then centrifugation
to remove cell membranes and organelles, leaving only cytoplasm, and cell machinery related to

39

transcription and translation. In this way, the trigger oligonucleotide can be introduced directly to
the cell free extract, also containing the switch RNA and ribosomes, and the extract can then be
tested for the presence of GFP with UV light. This test would be performed using the
experimental toehold insert, and the control insert, once it is successfully produced. Following
this test, another test will be performed examining RNA-RNA interactions and switch function.

40

REFERENCES
Ackermann, Martin, and Lin Chao. "DNA Sequences Shaped By Selection For Stability". ​PLoS
Genetics​ 2.2 (2006): e22. Web. 27 Apr. 2017.
"BLAST: Basic Local Alignment Search Tool". ​Blast.ncbi.nlm.nih.gov​. N.p., 2017. Web. 27
Apr. 2017.

Center for Disease Control. “The Rabies Virus.” Accessed at
http://www.cdc.gov/rabies/transmission/virus.html​ Sept. 22, 2016
Church G. M., Elowitz M. B., Smolke, C. D., Voigt C. A., and R. Weiss. VIEWPOINT
“Realizing the potential of synthetic biology.” ​Nature Reviews. Molecular Cell Biology.
15 (2014): 289-294. Web.
"Cloning Vector Pbad-Gfpuv, Complete Sequence - Nucleotide - NCBI". ​Ncbi.nlm.nih.gov​. N.p.,
2017. Web. 27 Apr. 2017.

Fasman, Gerald D. ​Practical Handbook Of Biochemistry And Molecular Biology​. 1st ed. Boca
Raton, Fla.: CRC Press, 1989. Print.

Feilmeier, Bradley J. et al. “Green Fluorescent Protein Functions as a Reporter for Protein
Localization in ​Escherichia Coli​.” ​Journal of Bacteriology​ 182.14 (2000): 4068–4076.
Print.
"Geneart Gene Synthesis And Services | Thermo Fisher Scientific". ​Thermofisher.com​. N.p.,
2017. Web. 27 Apr. 2017.

Green A.G. et al. “Toehold Switches: De-Novo-Designed Regulators of Gene Expression.” ​Cell
159 (2014): 1–15. Web

"QIAGEN Plasmid Mini, Midi, And Maxi Kits - (EN)". ​Qiagen.com​. N.p., 2017. Web. 25 Apr.
2017.

41

Pardee K. et al. "Paper-Based Synthetic Gene Networks." ​Cell​ 159.4 (2014): 940-54. Web.
Pardee, K. et al. "Rapid, Low-Cost Detection of Zika Virus Using Programmable Biomolecular
Components."​Cell​ 165.5 (2016): 1255-266. Web.
Rupprecht C. E. “Rhabdoviruses: Rabies Virus.” In: Baron S, editor. Medical Microbiology. 4th
edition. Galveston (TX): University of Texas Medical Branch at Galveston; 1996.
Chapter 61. Available from: ​https://www.ncbi.nlm.nih.gov/books/NBK8618/
Tordo, Noël et al. "Completion Of The Rabies Virus Genome Sequence Determination: Highly
Conserved Domains Among The L (Polymerase) Proteins Of Unsegmented Negative-Strand
RNA Viruses". ​Virology​ 165.2 (1988): 565-576. Web. 27 Apr. 2017.
Xu, Pingping, Fujian Huang, and Haojun Liang. "Real-Time Study Of A DNA Strand
Displacement Reaction Using Dual Polarization Interferometry". ​Biosensors and
Bioelectronics​ 41 (2013): 505-510. Web.
World Health Organization. “Rabies”. Accessed at
http://www.who.int/mediacentre/factsheets/fs099/en/​ Sept 22, 2016
Zhang, David Yu, and Erik Winfree. "Control Of DNA Strand Displacement Kinetics Using
Toehold Exchange". ​Journal of the American Chemical Society​ 131.47 (2009):
17303-17314. Web. 27 Apr. 2017.

42

APPENDIX

Appendix A: Sequences, locations and penalties assigned to the top 8 trigger sequences.
#

Trigger sequence

Start in the
genome

Concentrations of strands
for dimerization

1

CTGGAGTCTAGTCAGCTTATTGATGATAG
A

7519

0.081mM single
0.0096 mM double

2

GGATCTATTATTCAGACAGATCAGACCTC
A

7389

n/a

3

GAATAGACCTTATCGGATGACTCTAACAG
A

5584

0.075mM single
0.013mM double

4

CTCAATACTTACCAGTCTCATCTTCTACTC

10037

n/a

5

ATATATTCTCCTGAGGCTAGATAACCATCC 9778

0.075mM single
0.012mM double

6

TCTCATATATTCTCCTGAGGCTAGATAACC

9774

0.097mM single
0.0015mM double

7

CTCACTGGATCAGGTTGATTTACAAGATA
G

11632

0.096mM single
0.0016mM double

8

GGTTGAGAGAAACCTATCTAAGAGTATGA
G

10072

0.093mM single
0.0035mM double

43

Appendix B. Sequences of the top 8 switches corresponding to 8 top triggers.
Number

Switch Sequence

1

GGGTCTATCATCAATAAGCTGACTAGACTCCAGATACAGAAACAGAGG
AGATATCTGAUGTCTAGTCAGAACCUGGCGGCAGCGCAAAG

2

GGGTGAGGTCTGATCTGTCTGAATAATAGATCCATACAGAAACAGAGG
AGATATGGAAUGATTATTCAGAACCUGGCGGCAGCGCAAAG

3

GGGTCTGTTAGAGTCATCCGATAAGGTCTATTCATACAGAAACAGAGG
AGATATGAAAUGACCTTATCGAACCTGGCGGCAGCGCAAAG

4

GGGGAGTAGAAGATGAGACTGGTAAGTAATGAGATACAGAAACAGAG
GAGATATCTCAUGACTTACCAGAACCUGGCGGCAGCGCAAAG

5

GGGGGATGGTTATCTAGCCTCAGGAGAATATATATACAGAAACAGAG
GAGATATATAAUGTCTCC​TGA​GAACCUGGCGGCAGCGCAAAG

6

GGGGGTTATCTAGCCTCAGGAGAATATATGAGAATACAGAAACAGAG
GAGATATTCTAUGATATTCTCCAACCUGGCGGCAGCGCAAAG

7

GGGCTATCTTGTAAATCAACCTGATCCAGTGAGATACAGAAACAGAGG
AGATATCTCAUGGGATCAGGTAACCUGGCGGCAGCGCAAAG

8

GGGCTCATACTC​TTAGATAGGTTTCTCTCAACCATACAGAAACAGAGG
AGATATGGTAUGGAGAAACCTAACCUG​GCGGCAGCGCAAAG

44

Appendix C: Results of transformation of DH5-α E. coli with three different plasmids (pBS,
pGFP, and pToe, as well as a no DNA control.
Sample

Volume plated

Number of Colonies

Ampicillin

Control

100uL

carpet

N

Control

100uL

none

Y

pBS

100uL

carpet

N

pBS

100uL

carpet

Y

pBS

10uL

>400

Y

pGFPuv

100uL

~100

Y

pGFPuv

10uL

9

Y

pToe

100uL

~100

Y

pToe

10uL

~30

Y

Appendix D: Agarose gel quantification of plasmids isolated using the Qiagen maxiprep kit and
protocol. Plasmids were linearized with Kpn1 and 1 µL of DNA was loaded and 500 ng of
lambda HindIII marker. The gel was run for 1.5 hrs at 65 V.

45

Appendix E. Results of transformations done to test ligation reaction efficiency. Competent E.
coli were heat transformed and grown on Lb agar plates overnight.
Sample

Amount of DNA (ng)

# Colonies

Ampicillin

CCC pBS +lig

2.8

Carpet

N

CCC pBS +lig

2.8

~300

Y

CCC pBS -lig

2.8

0

Y

CCC pGFP +lig

2.8

Carpet

N

CCC pGFP +lig

2.8

>400

Y

CCC pGFP - lig

2.8

0

Y

pBS e-k +lig

4.7

Carpet

N

pBS e-k +lig

4.7

20

Y

pBS e-k -lig

4.7

0

Y

pBS k-e +lig

4.7

50

Y

pBS k-e -lig

4.7

0

Y

pGFP e-k +lig

4.7

9

Y

pGFP e-k -lig

4.7

0

Y

pGFP k-e +lig

4.7

~100

Y

pGFP k-e -lig

4.7

0

Y

Construct

28.0

Carpet

N

Construct

2.8

40

Y

Construct

28.0

~300 ~50 white

Y

Construct

28.0

~300

Y

Construct

28.0

~300

Y

46

Appendix F: Annotated partial sequence of the constructed positive control plasmid.

47

48

Appendix F: Annotated partial sequence of the constructed positive control plasmid.

49

50