Pivotal to our project is the construction of the Troygenics.
For this we constructed two plasmids with genes mainly from the M13 bacteriophage.
One plasmid - the Assembly Plasmid - carries all genes necessary to build the
protein coat of our Troygenics. The other plasmid - the Application Plasmid - bears a packaging sequence for
incorporation into the Troygenics and codes for the endocytosis inducing
ligands as well as for the gene of interest. Both plasmids were co-transformed into an E. coli
cell and the assembly of the Troygenics was verified by DNA and protein based methods.
The assembly of the Troygenics was demonstrated with a PCR on the supernatant of a producing E. coli culture, as well as with
a fluorescence spectroscopy. We analyzed the Application Plasmid by ddPCR and nanopore sequencing.
Apart from the DNA we also analyzed protein properties using standard methods such as Bradford
assay and SDS-PAGE. Furthermore, we were able to visualize our Troygenics using atomic force microscopy.
We cloned two different plasmids to assemble our Troygenics (Fig.1). The first one is the Assembly Plasmid. It
contains most of the genes coding for the Troygenic coat. Only two coat coding genes are placed on the second
plasmid - the Application Plasmid. This plasmid codes for the two coatprotein coding genes VIII, which is fused to mCherry,
and truncated gene III. This is also the plasmid that will make up the single stranded DNA packed in the Troygenics later.
We constructed the Assembly Plasmid (Fig. 2, BBa_K2926228)
as the main provider of Troygenics structure and assembly proteins. The plasmid carries most of the genes from the M13K07 helperphage:
The coat protein genes VI, VII, IX and VIII, the assembly genes I, IV and XI and the genes coding for the control proteins pII,
pX and pV (Smeal et al. 2017). We amplified the genes II-VIII (BBa_K2926023),
the fd terminator (BBa_K2926024) and genes III-IV
PCR and cloned these three
fragments into the iGEM backbone pSB3K3 using
A short C-terminal sequence of gene III containing the promoter region for gene IV was cloned into the Assembly Plasmid upstream of gene IV.
This short sequence of gene III does not express a functional protein III. A larger, functional part of the gene III sequence was expressed
on our other plasmid, the Application Plasmid. As a result we could ensure that the Troygenics are only built, if the producing cell contains
both plasmids. E. coli DH5α was
transformed with Application Plasmid and positive transformants were selected on an agar plate containing kanamycin.
The Assembly Plasmid codes for the protein coat of our Troygenics which is similar to the coat of the M13K07 helperphage.
Major coat protein pVIII
The most notable gene on the Assembly Plasmid is gene VIII. It codes for
protein pVIII, which is the major coat protein of our Troygenics. In addition to the
wild-type coding-sequence of gene VIII on the Assembly Plasmid our Application Plasmid
also carries a modified gene VIII version. The Application Plasmid version of gene VIII is
fused with a peptide ligand to facilitate Troygenics affinity to target cells surface proteins.
The resulting mixture of wild-type protein VIII and ligand-fused protein VIII is especially
important to ensure stable Troygenics coat assembly. In the case of exclusively ligand-fused
protein VIII the coat would not assemble correctly due to steric hindrance.
The second plasmid we built is the Application Plasmid (Fig. 4, BBa_K2926229). This plasmid is derived from the iGEM backbone pSB1C3 and carries a
f1 origin of replication (BBa_K314110), truncated gene III (BBa_K2926026) from the M13K07
helperphage and a fusion gene consisting of the fluorescence marker mCherry and gene VIII (BBa_K2926027) from the M13K07 helperphage. We built the fusion gene of mCherry and gene VIII by inserting the mCherry (BBa_J06504)
coding sequence upstream of gene VIII’s coding sequence, removing the mCherry stop codon and adding the cmyc linker sequence between both coding sequences. The endogenous promoter and ribosome binding site of gene VIII are located upstream of the resulting fusion construct.
The expressed protein is mCherry fused to the N-terminus of protein VIII. This protein is then incorporated into the Troygenics coat which could be verified by fluorescence spectroscopy, we were
able to determine correct assembly of the Troygenics.
This plasmid expresses the single stranded DNA which is packed into our Troygenics. In the
following experiments and throughout our project this DNA is called Application Plasmid.
All following experiments concerning Troygenics assembly and purification were carried out with this plasmid. Also, this
plasmid is the starting point for the cloning of more sophisticated Application Plasmid variants. This includes the replacement of mCherry
to cell-specific ligands to enable the specific endocytosis as well as the addition of our gene of interest in this plasmid later.
f1 origin of replication
The f1 origin of replication is an important feature of the Application Plasmid because it enables replication of Application Plasmid in E. coli as well as packaging of the Application Plasmid into the Troygenics by enabling the cells to produce single stranded DNA.
mCherry - pVIII fusion protein
The fluorescence marker mCherry is cloned into our plasmid to act as a verification for the
presence of our Troygenics. We fused mCherry to gene VIII, encoding the
major coat protein, which represents one of our fusion parts in the iGEM parts
registry. To achieve this, we deleted the stop codon at the end of mCherry and set mCherry
in frame with gene VIII, seperated by the cmyc linker. The resulting mCherry-pVIII fusion protein is incorporated into the
Troygenics coat (in a mixture with wild-type pVIII, see above) in such a way that mCherry is
presented on the outside of our Troygenics. This enables us to detect that our Troygenics were
expressed and correctly assembled by measuring mCherry fluorescence.
The described fusion between gene VIII and mCherry is conceived as an initial
validation mechanism for Troygenics expression and assembly. Over the course of our project we added different surface-receptor ligands as fusion partners to gene VIII and mCherry to
to enable target cell specific endocytosis.
Truncated protein pIII
The variant of gene III we used here is not the wild-type gene III, but a modified and truncated version.
We placed gene III on this plasmid to ensure that one plasmid on its own cannot build a functional Troygenic.
We constructed our version of gene III by removing the wild-type protein’s N-terminus. In detail we removed the N1 and N2 regions
as well as the linkers of the gene III from the wild-type gene III from the M13K07 helperphage We kept the residues
275 to 425 which express the C-terminus of the pIII protein.
The Application Plasmid is constructed by PCR amplifying the truncated gene III and VIII from the M13K07 phage genome, the f1 origin of replication and
the iGEM parts mCherry and pSB1C3 as backbone vector. The fragments were combined by Gibson assembly.
The assembly product was transformed into competent cells
of E. coli DH5α and plated on agar with chloramphenicol.
Co-Transformation of both plasmids into E. coli
Due to various safety as well as biological reasons, we separated the necessary components for our Troygenics on two plasmids. To produce Troygenics,
we had to co-transform both of our plasmids into one E. coli cell thereby supplying all proteins needed for coat formation and genome packaging to yield a functional
Troygenic. In practice, both plasmids were co-transformed in E. coli ER2566. The cells were regenerated for 1.5 instead of only
1 hour because of the stress to take up two plasmids each with differnt antibiotics, and were plated on agar containing both antibiotics – kanamycin and chloramphenicol.
For general application of two or more different plasmids in one producing
organism, we designed split antibiotic resistances as our basic
and improved part.
With such a split resistance, the organism is only resistant to the antibiotic if both plasmids and therefore both parts of the split-resistance protein are present in the same cell.
Several split antibiotics were provided to the iGEM community.
Purification of the Troygenics and their DNA
For testing the correct assembly of our functional Troygenics, we first had to purify the Troygenics from the E. coli culture supernatant.
To purify the Troygenics we removed the
E. coli cells from the expression culture by centrifugation, thus separating them from the Troygenics which remain in the supernatant. Afterwards, we precipitated the Troygenics with PEG/NaCl and
collected them by centrifugation. Doing so we got the Troygenics in a pellet and could resuspend and dilute them in PBS buffer for further use.
To purify the DNA, we used the same protocol with some additional steps. First, we added DNase to the supernatant
after the centrifugation of the culture to destroy any remaining E. coli DNA.
The functional Troygenics should be intact during this step and should therefore keep their Application Plasmid safe from the DNase.
The second addition to the protocol is that the Troygenics were resuspended in lysis buffer instead of PBS buffer. This step
destroyed the protein coat of the Troygenics and enabled DNA purification using a plasmid isolation column.
Although we purified the Troygenics with the described, typical phage purification protocol,
there was still a lot of E. coli DNA in the sample. Consequently, we searched for another
protocol and found the protocol NEB uses for phages. This protocol includes additional centrifugation
and precipitation steps to remove the remaining E. coli DNA. To further analyze the
Troygenics purified with both methods, we performed multiple experiments including atomic force microscopy, Nanopore sequencing and droplet-digital PCR.
Demonstrating the correct assembly
We demonstrated the correct assembly of the Troygenics depending on their DNA on the one and on their
protein structure on the other hand.
First we took a look on the DNA of the Troygenics, which was tested with ddPCR and Oxford Nanopore Sequencing.
The experiments based on the proteins and functions of the Troygenics were Bradford Assay followed by SDS-PAGE,
a Phage plaque assay and a fluorescence test. Furthermore, we took a picture of our Troygenics with an atomic
Demonstrating the presence of intact Troygenics
E. coli producing our Troygenics in culture will secrete them into the medium. Therefore, we could detect the Application
plasmid, as it is the DNA of the Troygenics, in the medium of the cultuivation. So, we performed a PCR with primers binding on the
Application plasmid on the cultivation medium.
We cultivated two different cultures of E. coli. One culture contained only the Application Plasmid and the other culture
contained the Application Plasmid as well as the Assembly Plasmid. The second approach produced our Troygenics and secreted them into
the medium. Both cultures grew for one day before we took a sample, centrifuged and used the supernatant as PCR template. The
PCR was performed with
primers which specifically amplify a region on the Application Plasmid (Fig. 11).
The PCR yielded a product only on the supernatant of the culture containing E. coli transformed with the Application Plasmid and the Assembly Plasmid. In the culture of E. coli only with the Application Plasmid no PCR product is visible.
This strongly indicates, that E. coli with both plasmids produces our Troygenics and secretes them
into the supernatant. Thus, their DNA is detectable in the media. The missing PCR product in the sample from the culture with only the Application Plasmid shows that this plasmid alone is, as expected, not sufficient to produce the Troygenics. Furthermore, this result shows, that occurrence of the PCR target inside E. coli cells does not interfere in this assay and produce false positive results. Therefore, this assay allows for an easy detection of Troygenics production, assembly, and export by our two-plasmid system.
Growth experiments with the Troygenic producing E. coli strains DH5α and ER2566
As E. coli will produce our Troygenics, we first had to improve the growing of the bacteria.
We wanted to optimize the expression parameters for our Troygenics. This included the determination of
the E. coli strain with the highest expression level. We tried E. coli DH5α and ER2566,
each transformed with the Application Plasmid and the Assembly Plasmid. We cultivated each strain in three different media, each with
two different concentrations of the antibiotics kanamycin and chloramphenicol.
We recorded the optical density after eight hours of the twelve different cultures to decide, which parameters to use for further growth experiments.
Optical density (OD600) after 8 hours
E. coli DH5α LB
E. coli DH5α SOC
E. coli DH5α 2YT
E. coli ER2566 LB
E. coli ER2566 SOC
E. coli ER2566 2YT
25 µg/ml Kan 30 µg/ml Cm
25 µg/ml Kan 15 µg/ml Cm
The results show that E. coli grows best in SOC
medium with 25 µg/ml kanamycin and about 15 µl/ml chloramphenicol. We used these parameters for determination of a growth
curve of E. coli DH5α or ER2566 with the Application Plasmid and the Assembly Plasmid in triplicate.
The growth curves of E. coli DH5α and ER2566 transformed with our Troygenics plasmids look similar. To get a
further hint which strain produces more Troygenics in both cultures, we measured the amount of Troygenics produced. Therefore,
Application Plasmid was purified as described in all six samples and the DNA concentrations
were measured by fluorometric quantification with the ThermoFisher Qubit System. In two out of three E. coli DH5α cultures, no Troygenics were detected. The E. coli ER2566 strain on the other hand produced Troygenics in all three cultures. As the E. coli ER2566 strain produces a larger amount of Troygenics with a higher reliability, we worked with ER2566 in the following experiments.
Application Plasmid concentration [ng/µl]
Fluorescence spectroscopy to demonstrate that mCherry is on the Troygenic coat
Our basic version of the Troygenics presents mCherry on their coat. This means that we can detect correct Troygenics assembly
by measuring the fluorescence of the incorporated mCherry. We measured the fluorescence on the TECAN Infinite M200 plate reader,
for which we used Texas Red 2.5 µM as reference dye for our Troygenics. In a first experiment
the excitation wavelength was 570 nm and the emission wavelength 610 nm. We
were able to detect fluorescence as a result of our Troygenics.
In a further experiment we recorded the emission spectrum of our Troygenics (Fig. 13). As a result, we have seen that
there was a typical emission graph with a peak around 630 nm, which refers to a red fluorescent protein.
We compared the emission spectrum of our Troygenics with the emission spectrum of mCherry, because this is the fluorescence
marker cloned on the coat of the Troygenics. We have seen a difference in the emission peaks of the Troygenics compared to mCherry of about 20 nm. This is probably
because the mCherry on our Troygenics is fused to the whole Troygenic. Because of this, there might be a different folding
resulting in a shift of the emission spectrum.
Detecting the amount of Application Plasmid trough ddPCR
To measure the concentration of Application Plasmid in our samples and to detect if e.g. E. coli DNA
remained in our sample even after purification we used droplet digital polymerase chain reaction
The droplet digital PCR (ddPCR) is a method for measuring the absolute amount of
DNA in a sample. We used it for the
absolute quantification of our Application Plasmid. It is very precise and has additional advantages as you do not need a
standard curve to quantify your samples and the measurement is partly independent on the occurrence of
PCR efficiency modifying components (e.g. PCR inhibitors) in the sample of interest.
The ddPCR works by partitioning very diluted DNA into small droplets containing a PCR mix and and dsDNA-specific dye
(here EvaGreen). This partitioning is done in a droplet generator, a microfluidic device, where two oil streams
splits the water phase stream (containing the DNA and the PCR components) into little water-in-oil-droplets. In this
emulsion a PCR is carried out and droplets containing DNA with the intended target produce a PCR product which creates
together with the dsDNA specific dye a fluorescence readout. In the end the droplets are counted and the fluorescence
is measured. As the initial DNA is very diluted most droplets are only occupied by one or zero DNA molecules. This
information together with the fraction of positive droplets enables the calculation of the initial DNA concentration
as molecules per microliter by the software belonging to the ddPCR machine.
For the droplet digital PCR, the sample is serially diluted to very less DNA as
predicted by the measured DNA concentration and the assumption that the sample only contains single-stranded
Application Plasmid. All dilutions were used for ddPCR with primer pairs specific for the
Application Plasmid, the Assembly Plasmid and for E. coli. With this experimental setup we could
measure the amount of Assembly Plasmid DNA and E. coli DNA contaminating our purified Troygenics. In the ddPCR we
show that Application Plasmid is the main component of our sample. In addition there was also a small amount of
Assembly Plasmid DNA and genomic E. coli DNA detectable. The amount of Application Plasmid is 3-times the
amount of Assembly Plasmid DNA and 11.7-times E. coli genomic DNA.
The droplet digital PCR demonstrated the assembly of our Troygenics as it showed that the Application Plasmid was precent
in a much higher concentration than the other two plasmids. But there was also to see, that there still was an amount of
E. coli DNA, so we desided to purify the Troygenics another time.
After we redid the Troygenic purification with the expanded protocol from NEB, we did the ddPCR again.
This time there was even less E. coli DNA in the sample than before. The Application Plasmid is 4.1- and 21.4-times
more abundant than the Assembly Plasmid or genomic E. coli DNA respectively.
This experiment demonstrates, that the new protocol resulted in Troygenics with a higher purity. Additionally, the new protocol
yielded a higher amount of purified Application Plasmid.
Oxford Nanopore-Sequencing to characterize the Application Plasmid
We performed nanopore sequencing of our purified Application Plasmid. This showed if the assembled plasmids contain the correct DNA sequence or if there are any mutations that would prevent the assembly of the Troygenics. Additionally, Nanopore sequencing shows the amount of DNA from other sources, e.g. from E. coli, in the sample and can be used as a quality control.
For nanopore sequencing we used Oxford Nonopore Technologies (ONT) flow cells (R9.4.1 and R10) with many little protein nanopores on it to analyze
the sequence of DNA samples. For the library preparation we used the rapid barcoding kit (RBK004) with a
TN5 transposase that randomly cleaves DNA and adds barcoded adapter sequences. Due to the small size of the genome of our Troygenics, sequencing was performed together with other samples thus requiring barcodes. The TN5 transposase linearizes the circular plasmids so that it can be pulled through a nanopore later. After the ligation
of the adapter the transposase is inactivated by heat.
The barcoded and adapter-attached DNA was bound to magnetic beads. This binding is dependent on
pH, ion strength and buffer conditions. For our fragments which should be longer than 4,000 bp we
used a volume of beads that matched the volume of our samples. The fragments bound to magnetic beads were washed using a magnetic rack which
attracts the beads with our DNA. The liquid was removed and the beads were
washed gently with 70% ethanol. The ethanol was removed and the beads were allowed to dry briefly. Afterwards the DNA was eluted from the beads by the addition of water. After a short incubation
time we added rapid sequencing adapters which bind to the barcoded adapter and incubated this mix for ten
During the incubation time the flow cell was primed. For this flush tether (FLT) and flush buffer (FB) were mixed in a tube. Then the Priming Port was opened and
air bubbles were removed from the channel leading from the priming port to the flow cell. This is very important because there must not be
air inside the flow cell. Air bubbles break down nanopores on contact. Therefore,
we removed the air using a pipet. You can control if there is air inside by the
color. If the channel is bright there is air, if it is dark there is liquid in it.
When the air was removed, we put the flush tether/flush buffer mix into the flow cell.
After incubation of our sample, we opened the Spot On and loaded flush buffer to the Priming Port to create
a water bridge between flow cell and Spot On. This prevents that air gets inside the flow cell during loading of the sample.
We loaded the flow cell by putting our sample on the Spot On, closed the ports and
started our sequencing run with about one hour delay to allow the DNA to sink down and reach nanopores.
When the sample is placed on the nanopore, the read starts at the adapter. Then the
barcode and afterwards the DNA sequence of the fragment is read. The read ends a few nucleotides
before the end because the helicase cannot hold the strain anymore. Normally the
helicase holds the DNA strain and slows it down. At the end the helicase cannot bind
anymore, and the strain runs through the nanopore.
For the nanopore sequencing we used samples of our Application Plasmid from four different cultures and aligned them against our Application Plasmid sequence. By comparing
the results of the four samples, we could see if there were any mutations in the Troygenic
DNA or if it was built correctly. We assembled the reads with the assembler canu v1.8 (Koren, 2017) and mapped them against our Application Plasmid map.
We detected only a few point mutations. The mutations are:
This deletion is not critical to the assembly of the Troygenics because it is not placed on a coding sequence.
These two mutations are not critical because in the f1 origin of replication there is no frameshift, which could lead to a different amino acid sequence and therefore to a wrong folding of a protein.
The f1 ori is only marked by the base sequence, but not by the amino acid sequence.
This mutation leads to a different amino acid. Instead of serine now a glycine is incorporated in pIII at position 379. This could influence protein folding and lead to a
different 3D-structure of pIII. But as shown with the AFM picture our Troygenic assembles correctly. Therefore, we assume that this mutation does not inhibit the pIII folding.
We also used the sequencing data to calculate the fraction of Application Plasmid in the sample and eventual contaminations for each of the four
samples. The data shown in table 3 come from Troygenics purified with the initial protocol. For each index the total number of reads, the number of reads mapped to the Troygenic genome sequence and the number of reads mapped to the E. coli
ER2566 genome sequence we used to express the Troygenics could be compared.
reads from nanopore sequencing mapped to the Troygenic and E. coli genome sequences, respectively
In our samples purified with our initial protocol the Application Plasmid was detectable but made up only 5.14 % of the reads. In contrast the genomic E. coli DNA represents 97.63 % of the reads. Due to this high amount of E. coli DNA we started using the extended purification protocol and repeated this experiment with DNA from this second purification process.
reads from nanopore sequencing mapped to the Troygenic and E. coli genome sequences, respectively
Using the extended purification protocol the amount of Application Plasmid was increased by a factor of three from 5.14 % to 14.43 %. This showed that our second purification protocol removed much more of non-Troygenics material and is therefore much better suited for purification processes for future work with our Troygenics. When normalizing the number of mapped reads to the genome size, we can see that the relative abundance of the Application Plasmid is about 100 X fold higher than the E. coli DNA.
Demonstrating the Troygenics do not attack bacteria with phage plaque assay
Because of the truncated pIII on the Application Plasmid, our Troygenics cannot affect bacteria anymore. To demonstrate this, we did a phage plaque assay with our Troygenics and another phage as a reference on two different strains of E. coli, one expressing an F-pilus and one not expressing an F-pilus.
We prepared three different approaches and tested each approach with various concentrations of phages. In summary we had 3 approaches and 6 dilutions resulting in 18 agar plates.
The first approach was E. coli strain S2060 with our Troygenics. This E. coli strain does not have a F-pilus, on which the natural pIII could dock. Because it is not possible for phages to attack these E. coli strain, this approach was our negative control.
The second approach acted as the positive control. We plated the E. coli strain S2060 + pJC175e, which expresses F-pili, with M13TyrRS phages, which generate plaques in E. coli.
The third approach was the important one for our project. We used E. coli strain S2060 + pJC175e expressing F-pili and added our Troygenics. If they were native pIII on the Application Plasmid, our Troygenics would infest bacteria and plaques would be built. We truncated pIII because we did not want the Troygenics to infect bacteria, so there must not be plaques in the E. coli culture.
Since no plaques were observed, we concluded that the truncation was successful.
Troygenic protein characterization with Bradford and SDS-Page
We wanted to quantify the protein content in our purified Troygenics using a Bradford assay. Therefore, we measured a BSA standard curve.
We measured our Troygenics with the same Bradford assay to compare the results with the BSA standard curve. The results showed that the produced Troygenics have a protein concentration of 44.58 µg/ml.
We analyzed our purified Troygenics on an SDS-PAGE to check the size of the individual proteins. A lot of different bands were visible.
The strongest band was at about 30 kDa, which might indicate the presence of pVIII fused to mCherry.
Another strong band was at about 5 kDa, which probably corresponds to the native pVIII. The remaining bands were at 32, 50, 70
and 75 kDa. These bands could correspond to the remaining proteins of our Troygenics. A confirmation of these assumtions by MALDI-ToF MS was planned but execution fell short due to lack of time.
Imaging the Troygenics with Atomic Force Microscopy
To demonstrate that Troygenics were correctly assembled, we did atomic force microscopy (AFM) with them. On Fig. 31 a Troygenic can be seen. Compared to the original M13K07 helperphage (Fig. 32), our Troygenics are visibly smaller which results from the fact that the M13K07 bacteriophage coat assembles around the integrated DNA. If this plasmid is smaller, the phages will be smaller, too.
This difference can be explained by the shorter DNA molecule inside our Troygenic. The different size of M13-derived phage-like particles dependent on their genome size is well described in the literature.
As a reference we have taken pictures of the original M13K07 helperphage with the AFM and with the TEM.
Our Troygenics have the same structure but they are smaller, because of less DNA inside.
BIO-RAD. n.d. Droplet Digital PCR Application Guide. Link
Compton, S. J., and C. G. Jones. 1985. “Mechanism of Dye Response and Interference in the Bradford Protein Assay.” Analytical Biochemistry 151 (2): 369–74. Link
Manns, Joanne M. 2011. “SDS-Polyacrylamide Gel Electrophoresis (SDS-PAGE) of Proteins.” Current Protocols in Microbiology 22 (1): A.3M.1-A.3M.13. Link
Marvin, DA. 1998. “Filamentous Phage Structure, Infection and Assembly.” Current Opinion in Structural Biology 8 (2): 150–58. Link
Smeal, Steven W., Margaret A. Schmitt, Ronnie Rodrigues Pereira, Ashok Prasad, and John D. Fisk. 2017. “Simulation of the M13 Life Cycle I: Assembly of a Genetically-Structured Deterministic Chemical Kinetic Simulation.” Virology 500 (January): 259–74. Link