1. Bacteriophages as delivery vectors
PHOCUS is a biopesticide consisting of a cocktail of bacteriophages that will be engineered to express molecules toxic to locusts after infection of their target locust gut bacteria. We have used the T7 bacteriophage as the phage model and Escherichia coli (E. coli) BL21 (DE3) as the host model for phage engineering, as bacteria present in the locust gut belong to the same genus as E. coli, the Enterobacter genus [Dr. O. Lavy, personal interview]. We engineered the T7 bacteriophage replacing a phage non-essential gene with a GFP reporter gene, that would ultimately be substituted by our desired locust toxic molecule.
To engineer a phage to encode toxic molecules to locusts, we first investigated whether we could successfully engineer bacteriophages by replacing non-essential genes. We identified the three non-essential T7 genes as possible candidates: 0.6A (early gene), 1.1 (early gene) and 4.3 (middle gene) . In order to determine which of these genes should be substituted with our recombinant gene, we performed proof of concept experiments using enhanced GFP del6 (229) (eGFP) as a reporter , substituting each of the target T7 non-essential genes separately.
The non-essential genes were replaced by a double-stranded DNA (dsDNA) construct, coding for enhanced GFP, through Bacteriophage Recombineering of Electroporated DNA (BRED), designed by Marinelli et al. . This construct was designed with 100 nucleotides homologous to flanking regions on each side of these aforementioned non-essential genes. As the first product of BRED results in a mixture of wild-type and engineered phages , a CRISPR/Cas9 system was designed and expressed in E. coli BL21 (DE3) as described by Kiro et al  to speed up the selection process for the engineered phage. Cas9 is an endonuclease guided by single-guide RNA (sgRNA) that can be encoded on a plasmid . The plasmid encoding the sgRNA was designed to contain a protospacer of the wild type T7 genes 0.6A, 1.1 and 4.3. Guided by this sgRNA, Cas9 binds to the target sequence and makes a double stranded break at this certain position, resulting in the cleavage of the T7 wild-type phage DNA.
The dsDNA substrates encoding eGFP were constructed by PCR amplification of plasmid pUC57-OriLR-deGFP. The expected size of the amplified fragment was 878 bp; composed of 678bp from eGFP and 200bp from the homologous regions. Figure 1 shows the amplified PCR fragments, all of them at the expected 878bp size.
Confirmation of the gene replacement
The phage samples obtained after BRED were screened by PCR for the presence of the insert, using both internal and external primers (Figure 2A). The external PCR products of the BRED samples were of the same size as the product from the wildtype phage, indicating that the wild-type phages were still abundant in the phage samples after BRED . The internal PCR products did show amplification of DNA at each position, because the internal primer only anneals to GFP, therefore showing that there were engineered phages present in each sample after BRED. A second band was shown at position 0.6A and 1.1 of unexpected size. Due to these results, it was decided to continue working with the sample containing T7 engineered at position 4.3. (Figure 2B). Additionally, the iGEM team from Austin advised us to continue experiments with T7 engineered at gene 4.3, as, in the T7 phage, early genes do not have a strong promoter and that upstream terminators can hinder expression of later genes downstream (see Partnership).
To select for the engineered phage, the BRED sample was plated on E. coli BL21 (DE3) with an induced CRISPR/Cas9 system targeting the wild-type gene 4.3 expecting a higher concentration of the engineered phage in the resulting plaques. This transformed host bacteria contained the pKDsgRNA_4.3 plasmid containing the sgRNA targeting the T7 gene 4.3 (BBa_K3407025), as well as the Cas9 plasmid pCas9-CR4 . Plaques were screened with primer binding inside the eGFP gene (Figure 2A, subsection I), resulting in one positive plaque (Figure 2C, line 4). The positive plaque was replated on E. coli BL21 (DE3) with an induced CRISPR/Cas9 system targeting the wild-type gene 4.3 and picked, as well as lysed for screening with primer binding inside the eGFP gene (Figure 2A, subsection I) (Figure 2D). Amplification of most plaques with those primers showed DNA bands, however, not at the expected size of 727 bp, but around 300 bp (Figure 2D). Some of the positive plaques obtained were screened for the presence of wild-type T7 phage by using both external and internal primers (Figure 2E). This amplification showed high concentrations of amplified DNA of approximately 600 bp, the expected size for the wild type phage. This indicates that the secondary plaques still contained high concentration of the wild type phage. We were not able to confirm the presence of the eGFP by sequencing.
BRED was performed to replace non-essential genes 0.6A, 1.1 and 4.3 of the wild type T7 phage by eGFP, successfully obtaining positive engineered phages in each of the positions after screening by PCR. T7 phages engineered in the 4.3 gene position were re-plated further in order to obtain isolated engineered phages, as the wild type phage was still present in the samples. Further PCR screening should be done to isolate the T7 phage engineered at position 4.3, using the constructed E. coli BL21 (DE3) host strain containing the CRISPR/Cas9 system targeting the 4.3 gene of the wild-type T7 phage.
Future experiments will include the characterisation of GFP expression of an isolated T7 bacteriophage engineered containing eGFP in the 4.3 gene position. This would provide information about the expression levels of a gene replacing the non-essential native 4.3 gene. The ultimate goal would include engineering the T7 bacteriophage containing GFP, substituting the eGFP with our desired locust toxic molecules.
Validation of the Phage and Toxin Production model
In Model 1- Phage and Toxin Production, the change of the number of bacteria, phages and toxin production over time were modeled. The objective of this model was to investigate if enough toxin could be produced by the infected gut bacteria to kill the locust in less than 7 days. To validate one part of the model, the biomass (OD600) of susceptible bacteria upon infection as well as the phage titer (P) were measured over time. After cells had grown to an OD of 0.3, measurements were taken every 5 minutes. This experiment was done in singlet. The obtained experimental profiles (Figure 3) show clear similarities to the predicted concentration profiles obtained with the model in terms of trend and value. This demonstrates that our equations and parameters are a fair representation of reality. An elaborate analysis of the data obtained from these experiments can be found here.
Partnership iGEM Vienna
iGEM team of BOKU-Vienna has also worked with T7 bacteriophage for their project. Their aim was also to engineer the T7 bacteriophage using BRED, but replacing another gene: the g10 gene. We decided to design experiments to compare the effect of the replacement of these different genes and the use of these different phages on the infectivity of the phage.
The experiments were performed using the wild-type phages of each team as no engineered phages were created and isolated in time. Both teams conducted experiments determining the phage titer needed to clear a liquid culture of susceptible bacteria (infection dose assay) to determine the infectivity of the phages. In addition, a plaque assay was done wherein the diameter of the plaque size was measured over time to compare the diffusion speed and virulence of the phages. We obtained different results for both experiments and we concluded that the experiments should be repeated to find out what caused this difference. More elaborate results from both teams can be found in the Partnership page.
2. Production and characterization of locust toxic molecules
Our approach to killing locusts is based on a unique combination of the insecticidal activity of the crystal toxin Cry7Ca1 and the gene silencing effect of RNA interference (RNAi). Cry7Ca1 punctures the locust gut lining, harming the locust and allowing the RNAi precursors to reach the hemolymph, where they mature and silence the expression of vital locust genes.
The Cry7Ca1 protein is a crystal protein from Bacillus thuringiensis (Bt) strain BTH-13 that has been reported to have activity against the locust species Locusta migratoria manilensis . This protein turns into an activated toxin when cleaved by insect gut proteases. The Cry7Ca1 activated toxin is more efficient against L. migratoria than the protoxin . In our study, An expression cassette was designed for the Cry7Ca1 activated toxin, using the viral promoter and terminator from T7, the model phage used in phage engineering experiments. The goal of these experiments was to confirm the expression of our composite part BBa_K3407017, that allows expression of the Cry7Ca1 under viral T7 promoter and terminator. The toxin was designed to be expressed with an N-terminal His-tag with a thrombin cutting site, giving a protein molecular weight of 68.5 kDa. This would allow the further removal of the His-tag for characterization experiments, as it would be inserted in the engineered phage without this tag.
Optimal conditions for Cry7Ca1 toxin overexpression
In order to check the overexpression of the Cry7Ca1 toxin, we incubated the E. coli BL21 (DE3) (Negative control) and E. coli BL21 (DE3) transformed with the plasmid pTWIST_Cry7Ca1 containing the Biobrick BBa_K3407017 at 37ºC until the OD600 reached approximately 0.6A. In order to find the best induction conditions, we added different IPTG concentrations (0.5 and 1 mM) and incubated the cultures for 4h and overnight. The total protein content of the cells and the soluble proteins obtained after cell lysis were analysed by SDS-PAGE electrophoresis (Figure 4).
Figure 4 shows successful overexpression of Cry7Ca1 toxin with all IPTG concentrations used, as a band corresponding to the molecular weight of the Cry7Ca1 toxin can be observed in samples from E. coliBL21 (DE3) Cry7Ca1 (Figure 4A). After lysis with FastBreak™ Cell Lysis Reagent (Promega), bands corresponding to the size of the Cry7Ca1 toxin are also observed (Figure 4B), although at lower intensity. These results indicated that a final IPTG concentration of 0.5 mM is enough for the overexpression of our BioBrick. They also suggest that our protein may not be completely soluble after lysis, as the fraction of soluble protein recovered seems lower than that of the total overexpressed amount. Moreover, we noticed a leaky expression of our protein, as a band was observed after an overnight culture of E. coli BL21 (DE3) pTWIST_Cry7Ca1 without inducer. This may be due to the absence of a repressor in our constructed plasmid. In order to avoid leaky expression, another host could be used such as E. coli BL21 (DE3) pLysS. The BBa_K3407017 could also be modified to include a repressor such as the lacI gene.
Purification of Cry7Ca1 toxin
We purified the Cry7Ca1 using the identified IPTG concentration for overexpression. As the recombinantly expressed Cry7Ca1 toxin contains a His-tag at the N-terminal part, we purified it using affinity chromatography with the HisLink protein purification system (Promega). All samples obtained from the purification process were analysed by SDS-PAGE electrophoresis (Figure 5).
A clear band of the expected size was obtained in the elution sample (Figure 5), indicating that the Cry7Ca1 toxin has successfully been purified. The fact that the toxin protein is also present in the flow through sample suggests that further optimization is possible.
From these results, we conclude that the Cry7Ca1 toxin is overexpressed in E. coli cells after induction with 0.5 mM IPTG (although there is some leaky expression). In this expression conditions, the Cry7Ca1 toxin can be further successfully purified by affinity chromatography thanks to its N-terminal Histag.
The next step in our project would be to encode the Cry7Ca1 toxin in the T7 phage genome to see whether this protein can be expressed upon infection of the target bacteria. As the toxin can be successfully purified, it could also be used for in vitro assays to examine its toxicity against different insect gut cells, including gut bacteria of our target locust species, the desert locust (Schistocerca gregaria).
3. RNA interference (RNAi)
RNA interference (RNAi) is a eukaryotic intracellular defence against viral infections and transposable elements triggered by double-stranded RNA (dsRNA) . RNAi will be used against locusts swarms by designing short hairpin RNA (shRNAs) containing a short sequence of dsRNA targeting locust vital genes. Dicer is the cytoplasmic enzyme in charge of producing siRNA from the shRNA, the first step of the siRNA-mediated gene silencing.
The components of a theoretical fusion protein (see Design - Future perspectives) were designed to deliver the shRNA produced in the locust gut into the hemolymph where dsRNA causes a strong gene silencing. A component of this cargo is Fox-1 RNA binding domain (Fox-1 RBD), which was chosen to bind specifically our shRNAs by the loop of the hairpins. Theoretically provided with a fused transporter peptide (TMOF), Fox-1 RBD will deliver the shRNA across the locust gut’s epithelium into the hemolymph, triggering ubiquitous RNAi response. In principle, shRNAs would be transcribed in tandem (tshRNA) and mini-3, the dsRNA counterpart of classical dsDNA restriction enzymes, will cut its target sequence located at the base of each individual hairpin, detaching them from the transcript. This cleavage produces a GG overhang in the 3’ end.
We aim to prove our concepts by:
- Testing whether in vitro transcription by T7 RNA polymerase can efficiently generate shRNAs with strong secondary structures. A shRNA targeting GFP was used as a model (BBa_K3407022), and the same shRNA with two point mutations in the loop (G6A and C3U) was used as a control of later experiments (BBa_K3407023).
- Expressing and purifying RNA binding domains (RBD) from Fox-1: wild type (BBa_K3407020) and mutated less active version (BBa_K3407024). With this we can further test the binding activity of Fox-1 RBD when compared to a mutated inactive version.
- Testing whether the Fox-1 RBD domain (BBa_K3407020) binds specifically to our shRNA.
- Showing that our designed shRNAs (BBa_K3407022) with GG 3’ overhang is a good substrate for Dicer, therefore being able to perform the first enzymatic conversion from the RNAi pathway.
Fox-1 RBD and RBD* (mutated) overexpression and purification
In order to check the overexpression of Fox-1 RBD and its mutated version, we incubated the E. coli BL21 (DE3) (Negative control) and E. coli BL21 (DE3) transformed either with the pBbB7a_Fox-1_RBD (BBa_K3407020) or pBbB7a_Fox-1_RBD* (BBa_K3407024) at 37ºC until the OD600 reached ~ 0.6. We induced the cultures with a final IPTG concentration of 1 mM and incubated them 4h at 37ºC and overnight at 30ºC. The total protein content of the cells was analysed by SDS-PAGE electrophoresis (Figure 6A). As the recombinantly expressed Fox-1 RBD and RBD* (mutated) contain a His-tag at their N-terminal part, we purified them using affinity chromatography with the HisLink protein purification system (Promega). Samples obtained from the different purification steps were analysed by SDS-PAGE electrophoresis (Figure 6B).
Additional protein bands corresponding to molecular weights of 13.6 kDa (Fox-1 RBD) and 13.4 kDa (Fox-1 RBD*, mutated) were observed after induction (Figure 6A), indicating that both proteins were successfully overexpressed. Moreover, a clear band is obtained in the elution samples (E) (Figure 6B), demonstrating that both Fox-1 RBD and RBD* were successfully purified. These two purified cargo proteins were subsequently used in Electrophoretic Mobility Shift Assay (EMSA) to verify binding with in vitro produced shRNA.
shRNA can be transcribed with T7 RNA polymerase
Hairpin structures can hamper transcription, causing premature termination. This is the case for the T7 viral RNA polymerase, whose commonly used terminator is a hairpin itself . To test whether T7 RNA polymerase is able to transcribe the designed shRNAs, two types of hairpins with GG overhangs were produced: one with the Fox-1 binding sequence in its loop, denoted as shRNA (BBa_K3407022), and a second one, named shRNA* (BBa_K3407023) that contains two point mutations in the loop (G6A and C3U), which reduces its affinity with Fox-1 RBD .
For each short hairpin, a pair of complementary primers were annealed to form a DNA template for transcription (see Design page). The DNA template was designed to possess a T7 promoter followed by a 27nt inverted repeat sequence taken from the eGFP gene (nt 78 to 105), and linked together by the 9nt sequence containing the target of Fox-1 RBD. On the 3’ termini, two GG were added to produce the desired overhang.
The commercial kit T7 RiboMAX™ Express RNAi System (Promega) containing the T7 RNA polymerase was used to produce the shRNA hairpins (Figure 7). The expected sizes of the transcript are ~65bp for ssRNA when denatured, and ~32,5bp for dsRNA with its native hairpin structure if the loop is assumed to be part of the dsRNA (which represents an approximation). We would expect the dsRNA ladder to be denatured into ssRNA in the conditions of the gel. On the other hand, shRNA bands are located near 30nt, which indicates the gel denaturing conditions might not have been enough to undo the strong secondary structures of the hairpins, making them unable to move in the gel as their expected ~65 bp ssRNA but rather as ~32,5bp. This phenomenon is commonly seen when studying hairpins in Urea-PAGE assays . We conclude both shRNA variants have successfully been transcribed in vitro and they can be used for binding assays with the Fox-1 RBD protein and Dicer processing experiments.
Fox-1 RBD binds to shRNA
To study RNA-protein interaction, an Electrophoretic mobility shift assay (EMSA) was performed . When a protein binds to RNA, the complex migrates slower in the gel, shifting the RNA band to a higher position. shRNA (BBa_K3407022) and Fox-1 RBD (BBa_K3407020) were first incubated in 1:1 and 1:2 molar ratios shRNA:Fox in binding buffer. 400ng of RNA were loaded in each lane corresponding to 10 fmol, along with 10 fmol of Fox-1 RBD (1:1 ratio) or 20 fmol (1:2 ratio) in a total of 15 μL reaction. To demonstrate binding specificity, the same experiment was performed with Fox-1 RBD* (BBa_K3407024), a mutated version of Fox-1 with a reduced affinity for its binding sequence. In addition, the assays were also carried out with shRNA* (BBa_K3407023), a version of the shRNA with two mutations in the loop, which inhibits binding to Fox-1 RBD  (Figure 8). Samples were then loaded in a native polyacrylamide gel to avoid protein denaturation (Figure 9).
The non-bound shRNA and shRNA* appear on the gel as ~30bp dsRNA (negative control, Figure 9). Incubation of shRNA and ahRNA* with wild-type Fox-1 RBD leads to a notable shift of the RNA band at both 1:1 and 1:2 molar ratios. This upper band corresponds to the shRNA-Fox-1 RBD complex, showing that both purified compounds are active. The lower fraction of free shRNA in the 1:2 molar ratio indicates that increasing the concentration of Fox-1 RBD yields a higher amount of complex formed.
Interestingly, incubation of shRNA with mutated Fox* RBD does not lead to complex formation. This results corroborates literature data showing that the introduced mutations strongly reduce the affinity of the protein to the target RNA sequence. Surprisingly, Fox-1 RBD (denoted as Fox), but not Fox-1 RBD* (denoted as Fox*), is also able to bind to mutated shRNA*, suggesting that under the studied conditions the mutations in the RNA loop sequence do not abolish complex formation (Figure 10). Different results were previously obtained using surface plasmon resonance , which may be explained by differences in the experimental conditions.
Hairpins with GG 3’ overhangs are suitable Dicer substrates
Dicer processing is the first step of the RNAi-mediated gene silencing where small interfering RNAs (siRNAs) are produced. Dicer-2 isoform products are later used to cleave messenger RNAs (mRNAs) with a 100% complementarity with one of the siRNA’s strand . Not all dsRNA structures are processed with the same efficiency due to the preferences Dicer shows towards certain substrates. For instance, dsRNA sequences embedded in a longer RNA transcript or with capped ends have been reported to show less processability than those at the termini, which generally show an increased siRNA production rate when ending with 3’ overhangs in absence of ATP . In particular, GG overhangs have shown a medium level of processivity among all possible combinations in humans , but it has not been proven in insects yet. A scheme about how shRNAs would be processed by Dicer-2 is shown in Figure 11.
With this in mind, we decided to study whether GG protruding 3’ overhangs emulating those left by mini-3 cleavage are suitable for the production of free, mature shRNA with RNAi triggering potential. shRNAs were designed with the aforementioned overhang and their processivity was tested with a Dicer cleaving assay using purified Dicer-2 from the insect species Drosophila melanogaster (Figure 12), kindly provided by TU Delft BN department.
The unprocessed shRNA gives a clear band on the gel, slightly above the 30nt size marker of the dsRNA ladder (Figure 12). Incubation with Dicer yields two lower bands, likely corresponding to the two cleavage products: the siRNA and the loop. Increasing the reaction duration from 1 hour to longer times is accompanied by a higher fraction of both enzymatic products. These results demonstrate that the in vitro transcribed shRNA containing GG 3’ overhangs can be processed by Dicer from the insect Drosophila melanogaster, adding evidence that our shRNAs have potential to silence any targeted gene from the locusts by means of RNAi.
T7 viral RNA polymerase is a suitable candidate to transcribe our designed shRNAs shRNA (BBa_K3407022 and BBa_K3407023).successfully. The combination of the selected promoter and inverted repeats linked with a loop leads to the transcription of the desired shRNAs. As in vitro production has already been proven, the designed template DNA can further be brought to the next level and engineer the phage to contain it.
Although mobility in Urea-PAGE is unpredictable, results show Dicer is able to process the shRNA to form siRNA, constituting the first step for a successful RNAi-mediated gene silencing. The designed overhangs seem to yield good processivity with Dicer-2 from the insect D. melanogaster, setting the first evidence that mini-3 is an acceptable candidate for cleaving the shRNAs from the tandem transcript, producing free and mature shRNA with RNAi triggering potential in insects.
From the hypothetical cargo that would deliver shRNAs into the hemolymph, at least Fox-1 RBD can bind most of the shRNA in 1:2 molar ratio. This finding suggests the domain can be used as a part of a fusion protein, and further research should be done to elucidate whether the other domains of the cargo are also suitable for our purpose. Although the binding takes place, it does not show the expected specificity towards its target sequence under our conditions. This result makes evident the necessity of a deeper study of the interaction between the loop structure and Fox-1 RBD, where the dsRNA part or the secondary structure might influence the specificity.
The theoretical delivery molecule requires further research to prove its functionality, such as the influence of the hairpin structure in the specific binding to Fox-1 RBD, or ultimately testing the ability to cross the gut barrier. Phages seem to be a realistic candidate to produce shRNAs in gut bacteria to target insects such as Schistocerca gregaria.
In order to produce multiple shRNAs and maximise the chances of silencing vital functions of the locusts, the shRNAs can be produced as a single tandem transcript RNA (tshRNA). Each shRNA could be generated after cleavage by an enzyme such as Mini-3 from Bacillus subtilis. Mini-3 shows a high sequence associated cleavage, cutting dsRNA containing “ACCU” with little off-targets , leaving 3’ protruding overhangs, as depicted in Figure 13.
As the designed overhangs in our shRNA (BBa_K3407022) seem to yield good processivity with Dicer-2 from the insect D. melanogaster, mini-3 would represent an acceptable candidate for cleaving the shRNAs from the tshRNA, producing free and mature shRNA with RNAi triggering potential in insects.
In order to protect shRNAs from degradation by RNAses, the protein YmdB could be co-expressed. This enzyme has shown to inhibit the RNAse III from E. coli , which we expected to help in production of shRNA in vivo. In this project, both YmdB and mini-3 have been cloned and overexpressed in compatible plasmid backbones for their characterization (see Design for more information). Preliminary experiments have been conducted in this direction and are described below.
Cloning of Mini-3 and YmdB from bacterial genome
We obtained the DNA sequences encoding for the enzyme Mini-3 and the protein YmdB by genome amplification from Bacillus subtilis and E. coli BL21 (DE3) respectively (Figure 14).
As it can be seen in Figure 14, both genes were successfully amplified by PCR as bands corresponding to their molecular weights can be observed. The amplified sequences were purified from the gel and cloned by Gibson Isothermal Assembly in the correspondent amplified backbone vector. Mini-3 was cloned into the vector pBbE8c, which contains an araBAD promoter and a T7Te terminator and is induced by anhydrotetracycline (aTc), thus obtaining a new composite part (BBa_K3407018). YmdB was cloned into the vector pBbA2k, which contains a tetR/tetA promoter and a T7Te terminator and can be induced by L-arabinose, thus obtaining a new composite part (BBa_K3407019). The assembled new plasmids were transformed into E. coli DH5-α. Plasmids purified from the transformed colonies were sequenced and transformed in E. coli BL21 (DE3) for protein expression.
Overexpression of Mini-3 and YmdB.
To overexpress Mini-3 and YmdB, we incubated at 37ºC E. coli BL21 (DE3) (negative control) and E. coli BL21 (DE3) transformed with the plasmids pBbE8c_mini3 or pBbA2k_YmdB. When the cultures reached OD600 ≈ 0.6, they were induced with 20 mM L-arabinose and 400 mM anhydrotetracycline respectively, at 37ºC for 4 hours and overnight at 20ºC. The total protein content of the cells was analysed by SDS-PAGE electrophoresis (Figure 15).
As seen in the SDS-PAGE of the total protein content (Figure 15), both Mini-3 and YmdB proteins were successfully overexpressed after induction with 20 mM L-arabinose and 400 mM anhydrotetracycline, respectively. No additional bands can be observed in E. coli BL21 (DE3) pBbE8c_mini3 or pBbA2k_YmdB without induction, indicating that there is no leaky expression.
Future experiments will include purification of the overexpressed Mini-3 to use it for in vitro activity assays. If the tshRNA is produced, then the purified Mini-3 could be used for its cleavage. Furthermore, co-expression of tshRNA, Mini-3 and YmdB experiments could also be performed, therefore enabling the production of multiple shRNAs in vivo.
- Bacteriophages: Biology and Applications - Google Books. (n.d.). Retrieved October 24, 2020, from https://books.google.nl/books?hl=nl&lr=&id=6sTBd2EJpmYC&oi=fnd&pg=PA29&dq=phages+infect+bacteria&ots=j0fY0fm6fs&sig=tIckCDuOjW0sZwV13HHBBTZNtI8&redir_esc=y#v=onepage&q=phages%20infect%20bacteria&f=false
- Marinelli, L. J., Piuri, M., Swigoňová, Z., Balachandran, A., Oldfield, L. M., van Kessel, J. C., & Hatfull, G. F. (2008). BRED: A Simple and Powerful Tool for Constructing Mutant and Recombinant Bacteriophage Genomes. PLoS ONE, 3(12), e3957. https://doi.org/10.1371/journal.pone.0003957
- Kiro, R., Shitrit, D., & Qimron, U. (2014). Efficient engineering of a bacteriophage genome using the type I-E CRISPR-Cas system. RNA Biology. https://doi.org/10.4161/rna.27766
- Jinek, M., Chylinski, K., Fonfara, I., Hauer, M., Doudna, J. A., & Charpentier, E. (2012). A programmable dual-RNA-guided DNA endonuclease in adaptive bacterial immunity. Science. https://doi.org/10.1126/science.1225829
- Reisch, C. R., & Prather, K. L. J. (2015). The no-SCAR (Scarless Cas9 Assisted Recombineering) system for genome editing in Escherichia coli. Scientific Reports. https://doi.org/10.1038/srep15096
- Wu, Y., Lei, C. F., Yi, D., Liu, P. M., & Gao, M. Y. (2011). Novel Bacillus thuringiensis δ-endotoxin active against Locusta migratoria manilensis. Applied and Environmental Microbiology. https://doi.org/10.1128/AEM.02462-10
- Obbard, D. J., Gordon, K. H. J., Buck, A. H., & Jiggins, F. M. (2009). The evolution of RNAi as a defence against viruses and transposable elements. Philosophical Transactions of the Royal Society B: Biological Sciences. https://doi.org/10.1098/rstb.2008.0168
- Schwartz, A., Rahmouni, A. R., & Boudvillain, M. (2003). The functional anatomy of an intrinsic transcription terminator. EMBO Journal. https://doi.org/10.1093/emboj/cdg310
- Auweter, S. D., Fasan, R., Reymond, L., Underwood, J. G., Black, D. L., Pitsch, S., & Allain, F. H. T. (2006). Molecular basis of RNA recognition by the human alternative splicing factor Fox-1. EMBO Journal. https://doi.org/10.1038/sj.emboj.7600918
- Soler Canton, A. (2019) Synthetic biology meets liposome-based drug delivery. Casimir PhD series, Delft-Leiden 2019-35. https://doi.org/10.4233/uuid:b46cc724-c6cf-4282-b3ba-3817e8bebf87
- Lam, J. K. W., Chow, M. Y. T., Zhang, Y., & Leung, S. W. S. (2015). siRNA versus miRNA as therapeutics for gene silencing. Molecular Therapy - Nucleic Acids. https://doi.org/10.1038/mtna.2015.23
- Sinha, N. K., Trettin, K. D., Aruscavage, P. J., & Bass, B. L. (2015). Drosophila dicer-2 cleavage is mediated by helicase- and dsrna termini-dependent states that are modulated by loquacious-PD. Molecular Cell. https://doi.org/10.1016/j.molcel.2015.03.012
- Vermeulen, A., Behlen, L., Reynolds, A., Wolfson, A., Marshall, W. S., Karpilow, J., & Khvorova, A. (2005). The contributions of dsRNA structure to Dicer specificity and efficiency. RNA. https://doi.org/10.1261/rna.7272305
- Głów, D., Pianka, D., Sulej, A. A., Kozłowski, Ł. P., Czarnecka, J., Chojnowski, G., … Bujnicki, J. M. (2015). Sequence-specific cleavage of dsRNA by Mini-III RNase. Nucleic Acids Research. https://doi.org/10.1093/nar/gkv009
- Głów, D., Pianka, D., Sulej, A. A., Kozłowski, Ł. P., Czarnecka, J., Chojnowski, G., … Bujnicki, J. M. (2015). Sequence-specific cleavage of dsRNA by Mini-III RNase. Nucleic Acids Research. https://doi.org/10.1093/nar/gkv009