Team:BGU Israel/Results
1. Delivery of Bacteria- Inspect if Serratia indeed integrates into the mosquito's microbiome and pass on to the eggs.
2. Plasmids Construction- The plan and construct of our polycistronic plasmid.
3. Toxicity Assays- Check if our initial plasmids containing the Bti subunits indeed kill mosquito larvae.
4. Bonuses- You will have to read further in order to find out!
Introduction:
With the knowledge that Serratia marcescens makes up about fifty percent of the Aedes aegypti mosquito's gut microbiome, we decided to use Serratia bacteria in our project[1]. As we aimed to have engineered bacteria on or in the mosquito eggs, we thought about a simple and easy way to deliver Serratia from the adult mosquito to the eggs. We decided to use the bacteria gut microbiome and to feed the adult male mosquitoes with the Serratia. Our delivery approach is based on the mosquito anatomy, where the gut is structured in a way that the reproductive glands and the colon share a common exit. Therefore, eggs could be infected through the gut, moreover bacteria might be delivered during sexual reproduction.
The general idea was that if we feed the mosquitoes with the bacteria, and it will survive in the gut, it may be transmitted to the reproductive glands and then to the eggs. This assumption was based on a recently published paper[2].
To test whether an engineered Serratia could be delivered from the gut to the eggs, we used a GFP-containing plasmid that was inserted via electroporation into the Serratia bacteria (Please see "Plasmids Construction" section for more information). We fed the mosquitoes with the engineered Serratia to test if bacteria containing the GFP gene would pass on to the eggs, then the we could confirm that our transfer strategy works.
Integration and survival of Serratia in the mosquito's gut:
To verify the survival of Serratia in the Aedes aegypti mosquito's gut, we feed mosquitos with our bacteria and track for GFP-positive bacteria in the eggs. We allowed one-day-old mosquitoes feed on a concentration of 10^7 bacteria/ml mixed in a 5% sucrose solution. After one day, we replaced the solution with a sterile 10% sucrose solution. We replaced the sucrose in order to validate that our Serratia was survived and integrated in the mosquito's gut microbiome. We observed the mosquito's gut at two extreme time points: three- and eight-days post bacteria feeding. We washed the mosquito twice from any bacterium on the surface of the mosquito as we were interested to test only internal bacterium that was exist in the mosquito’s gut. Next, we crushed and plated the gut with an appropriate antibiotic selection that exist on the plasmid and incubated with suitable conditions to test for GFP-positive Serratia. Our results show that the bacteria indeed grew on the plates. Thus, we concluded that the Serratia integrated and survived in the gut at these two time points.
Vertical transfer of the bacteria: From female mosquitoes to their eggs:
After two repeats of the above experiment, we decided to continue to test the transfer of the Serratia from the adult mosquito to the eggs. Based on the literature, we designed an experiment to examine if the bacteria fed to the mosquitoes would pass to the laid eggs[2]. We allowed one-day-old mosquitoes feed on a concentration of 10^7 bacteria/ml in a 5% sucrose solution. The following day, we replaced the feeding to 10% sucrose solution for one more day. The next¬ day, we let the mosquitoes feed on a blood meal and lay eggs on a moist Whatman paper. We made a bacterial culture from the eggs and incubated them overnight in order to culture the bacteria. Then, we measured the fluorescent of the samples. Using a t test, a significant difference was viewed between the untreated eggs and the infected eggs. Results can be shown in Figure 3. We also examined the eggs using a confocal microscope and discovered that the eggs presented auto-fluorescence, which prevented us from detecting if the signal was from our bacteria (Figure 4).
At this point, we verified the fact that our Serratia integrated and survived in the mosquito's gut and passed to the eggs.
After consultations with professionals, we realized that dispersing feeding stations or releasing infected female mosquitoes with our transgenic bacteria are not acceptable in some aspects (See "Integrated Human Practices" for more information). Therefore, we decided to take our project in a different direction, where our final product will be releasing infected male mosquitoes. However, we must verify that the Serratia could be transferred from infected males to females that eventually will lay infected eggs.
Horizontal transfer of bacteria: From male to female mosquitoes:
Females mate only once in their lifetime and store the sperm that fertilized the eggs they will lay, whereas males can mate several times[3]. Mosquitoes are ready to mate a few days after leaving their pupal casings. The pupae stage is the immature stage before becoming adults in the mosquito life cycle. To make sure that the female mosquitoes are virgins, we had to separate them before adulthood at their pupal stage. The separation was done by size. Female pupae are larger that male pupae.
After separating the mosquitoes into two vessels, males and females, we placed each vessel in a separate cage and waited until they reached adulthood, about one or two days later. We allowed the females to feed on a 10% sucrose solution. For the horizontal transfer experiments, we had males that were fed on 109 bacteria/mL in a 5% sucrose solution and control males that were fed on a 10% sucrose solution (the "untreated" control). After two days of feeding, we transferred the same number of non-infected female mosquitoes to the male cages. We allowed mating for a day, and then fed on a blood meal and waited for eggs to be laid on a moist Whatman paper. We made a bacterial culture for bacterial growth from the eggs and incubated them overnight in order to culture the bacteria. Then, we measured the fluorescence of the samples. We found that the bacteria indeed trasnfers to the eggs after only males are being fed (Figure 5). Using a t test, a significant difference was viewed between the untreated eggs and the infected eggs. We also looked at an untreated and an infected male mosquito using a confocal microscope to see if there is fluorescence in its gut (Figure 6 and 7, respectively). The results show that indeed an infected male mosquito's gut shows a fluorescent signal, while the untreated male mosquito's gut does not.
Difficulties we encountered during the experiments:
Our initial goal in the integration and survival of Serratia in the mosquito's gut experiment was to quantify the Serratia in the mosquito's gut. But since we had technical difficulties in plating and quantifying the growing bacteria, we decided to change our course of action and instead of quantifying the bacteria, we have conducted a qualitative experiment. Instead, we showed that the Serratia taken from the mosquito's gut grew on a selective plate with the proper antibiotics (Carbenicillin X5) resistance that is encoded in our plasmid only.
After the vertical transfer succeeded, we attempted the same conditions for the horizontal transfer, using the same concentration of bacteria 10^7 bacteria/mL, and the same number of feeding days (one day). But during the first attempt of horizontal transfer, from male to female mosquitoes, we observed little to no fluorescent signal in the laid eggs. Hence, we decided to change the concentration of bacteria in the feeding vessels to 109 bacteria/mL and to the feeding days from one to two days of feeding. As seen in the results, the reading was acceptable (Figure 5).
Conclusions:
We showed that transgenic Serratia can integrate and survive in the mosquito's gut. Furthermore, we demonstrated a vertical and horizontal transfer of the Serratia, meaning that after feeding of the bacteria, to both female and male mosquitoes, the Serratia transfer to the laid eggs (through an unknown mechanism which we intend to further explore in the future). Hence, the bacteria can be delivered to the eggs not only by feeding the female mosquitoes but rather by mosquito males through their sexual reproduction system or through physical contact and transfer mechanism. Thus, we were able to find a simple and effective delivery system of bacteria - from the adult mosquitoes to their offspring.
References
[1] Gusmão, Desiely S., et al. "Culture-dependent and culture-independent characterization of microorganisms associated with Aedes aegypti (Diptera: Culicidae)(L.) and dynamics of bacterial colonization in the midgut." Acta tropica 115.3 (2010): 275-281.
[2] Wang, Sibao, et al. "Driving mosquito refractoriness to Plasmodium falciparum with engineered symbiotic bacteria." Science 357.6358 (2017): 1399-1402.
[3] Boyer, Sebastien, et al. "Sexual performance of male mosquito Aedes albopictus." Medical and Veterinary Entomology 25.4 (2011): 454-459.
[4] H.Weber, Grundriss der insektenkunde (1996); Gustav Fischer Verlag.
[5] Hodapp, Cyril J., and Jack Colvard Jones. "The anatomy of the adult male reproduction system of Aedes aegypti (Linnaeus)(Diptera, Culicidae)." Annals of the Entomological Society of America 54.6 (1961): 832-844.
Our design considerations are detailed herein:
1. Choosing a promoter for transcription initiation in the host bacteria:
The pBEST plasmid is a synthetic high copy number plasmid containing Ampicillin resistance and deGFP insert. The recommended growth temperature of bacteria harboring the plasmid is 30⁰C, the same temperature that we grow the Serratia marcescens bacteria in our project.
The original promoter and ribosome binding sequence (RBS) of the pBEST plasmid (developed by Vincent Noiraux from the University of Minnesota) were designed for cell free protein transcription and translation systems and are the strongest promoter and RBS ever reported for bacterial protein expression.
The PR1 promoter is a mutant of the PR promoter which expresses high levels of GFP in vivo, hence, we have built our system using the pBEST backbone that contained a modified PR1 promoter.
The sequence of the modified PR1 promoter is detailed herein :
(5'-TGAGCTAACACCGTGCGTGTAGACAATTTTACCTCTGGCGGTGATAATGGTTGCA-3')
and deGFP with a HIS tag.
2. Construction of a polycistronic plasmid instead of several plasmids with a promoter & terminator for each gene:
In order to make sure that the transgenic bacteria will contain and express every subunit in high efficiency, the gaps that contain RBSs and a restriction site between each gene were based on the polycistronic pET[4]. For our polycistronic plasmid, we chose the subunits that are most toxic to Aedes Aegypti mosquitoes, on which we preform our project[3].
The order of the Bacillus thuringiensis subsp. israelensis (Bti) toxin subunits downstream of the promoter was determined by the importance of the subunits to the overall toxicity of the Bti toxin. As this plasmid contains one promoter only, the transcribed mRNAs that encodes for the last subunits are transcribed and expressed less efficiently[4][5].
Therefore, we designed a polycistronic plasmid where the last gene is deGFP without a tag. Thus, a fluorescent signal will indicate the expression and the expression efficiency of all the subunits in the plasmid and will allow us to detect the transgenic bacteria in live infected mosquitoes.
3. Tags for the cloned Bti toxin subunits for Western blot analyses:
As the our selected Bti subunits have similar protein weight, it is not possible to distinguish between them by their size on an SDS-PAGE. Therefore, we selected three different tags that will be used for the three Bti subunits:
HIS tag – 5' GGCAGCAGCCATCATCATCATCATCACAGCTCT 3'
Strep tag – 5' TGGAGTCATCCTCAATTTGAAAAA 3'
HA tag – 5' TACCCATACGATGTTCCAGATTACGCT 3'
The tag fusion location within the protein was chosen based on the literature - with the following considerations: the location should be one that will not interfere with the protein function and will exist in the protein after its proteolytic activation[1][2][3].
4. Amplification of the Bti toxin subunits:
We have received (with thanks) the ADRC plasmid from Dr. Vadim Khasdan (Ben-Gurion University), which contained the Cry11aA, Cyt1aA and the p20 genes. The Cry4Ba subunit was amplified from the pBtoxis plasmid of Bacillus thuringiensis subsp. israelensis (Bti) bacteria that we received from the student's lab course of the Life Sciences Department at our university[5].
The pBtoxis is a toxin-encoding plasmid that originates from the Bti which encodes to four subunits of Cry toxins (Cry4Aa, Cry4Ba, Cry10Aa and Cry11Aa) and three subunits of Cyt toxins (Cyt1Aa, Cyt2Ba and Cyt1Ca).
The primers for amplifying those genes were designed according to the sequences of the plasmids we received and after a confirmation of the genes sequenced by BLAST.
Using long primers (80 bp), we amplified the necessary genes with extra nucleotides upstream and downstream. The flanking regions of the amplified genes were homologous to the promoter and RBS sequence upstream of the target gene and to the GFP downstream of it. It allowed a Gibson assembly of DNA fragments and the addition of tags to the C or N termini of each subunit.
The downstream primers contained a sequence that did not exist and allowed the insertion of an RBS upstream of the GFP gene. It also allowed the deletion of the N-terminal HIS tag sequence by using nucleotides from the first codon of the GFP.
Restriction sites were inserted upstream of each RBS, which will allow future changes in the identity of the expressed genes (different subunits for different mosquito pest control), thus we created a modular system that could be modified in the future to be directed to specific mosquito secrecies.
Results of Sections 1-4:
Amplification of the different subunits from the ADCR plasmid.
The constructions of our plasmids.
5. Final Polycistronic Plasmid Construction:
The pBEST Bti toxin plasmid that we are constructing is a toxin-encoding plasmid that originates from the Bacillus thuringiensis subsp. israelensis (Bti) toxic subunits. It is planned to encode five genes overall: three toxin subunits (Cry4Ba, Cry11Aa, Cyt1Aa), one chaperone (p20) and one marker (deGFP). Each of the toxin subunits and the chaperone will contain a tagin order to be able to see the expression of those genes in Western blot analysis (tags are not shown in the figure bellow).
The final plasmid has not been constructed yet due to a problem in the assembly of the Cyt1Aa subunit. The assembly will be performed by restriction ligation of Cry4Ba and its upstream RBS to the three subunits plasmids/Gibson assembly with an insertion of restriction sites. The final assembly step can be determined only after the previous plasmids will be completed.
Every step of the initial 3 plasmids was confirmed by colony PCR and next generation sequencing using short primers.
Challenges we faced:
The biggest challenge in the assembly of a polycistronic plasmid is the similarity between homologous parts – every fragment had a similar sequence gap between the stop codon and the start codon of the next gene. Every gene needed an RBS and extra nucleotides for the best re-initiation of the ribosome, the only difference was the restriction sites' sequences. This issue caused many problems and we had to preform several Gibson assembly attempts with different sets of primers to succeed in the assembly of the larger plasmids.
References:
[1] Florez, A. M., Suarez-barrera, M. O., Morales, G. M., Rivera, K. V., Orduz, S., Ochoa, R., … Muskus, C. (2018). Toxic Activity , Molecular Modeling and Docking Simulations of Bacillus thuringiensis Cry11 Toxin Variants Obtained via DNA Shuffling, 9(October), 1–14.
[2] Anchor, M. (n.d.). The C-Terminal Domain of the Bacillus thuringiensis, (Diii). https://doi.org/10.3390/toxins11020062.
[3] Ben-dov, E. (2014). Bacillus thuringiensis subsp. israelensis and Its Dipteran-Specific Toxins, 1222–1243.
[4] Tan, S. (2001). A Modular Polycistronic Expression System for Overexpressing Protein Complexes in Escherichia coli, 234, 224–234.
[5] Khasdan, V. PhD Thesis: Cloning combinations of four genes from Bacillus thuringiensis subsp. israelensis and tmf for expression in various bacteria to enhance mosquito biocontrol. Supervisors: Prof. A. Zaritsky, Prof. S. Boussiba and Prof. D. Borovsky.
After we constructed our initial plasmids (Cry4Ba, Cry11Aa, and Cry11Aa + p20), we needed to check the toxicity of the plasmids on live larvae to see if our constructs express well and are toxic to larvae of Aedes aegypti.
We decided to separate a hundred, four-days-old larvae into different containers (triplicates for robustness) and infect them with our genetically engineered Serratia marscecens to check how they react and if they die.
To get the same concentration of bacteria, we made bacterial cultures with antibiotic for selection, then a day later we made subcultures out of the bacterial cultures clean of antibiotic, so it won't interfere with the natural microbiome of the larvae. The subcultures were measured in a plate-reader at an optical density (OD) OD600nm to indicate bacterial concentrations.
We poured a concentration of 10^7 (between 0.3-0.4 O.D.600 ) of the bacteria into the containers of the live larvae and counted every 24 hours the number of living larvae per container. We also checked in comparison to “Baktush”, which is the known treatment for mosquito larvae world-wide, and contains Bacillus thuringiensis subsp. israelensis (Bti) spores. The graphs below show the experimental results.
After we saw that our plasmids show low toxicity properties, we realized that the larvae could be quite old and turn into pupae in the middle of the experiment, which could cause the bacteria to become nonlethal to them. So, to overcome this problem, we altered the protocol again to use two-days-old larvae instead of four-days-old larvae. Also, we decided to place 50 larvae instead of 100 in each container in order to be able to monitor and count the live larvae more easily. The graphs beneath show the experimental results.
As indicated in the results above, all our genetically engineered bacteria show toxicity to larvae to some degree. During the third repetition of the experiment, lethality could be seen also in the larvae that were only treated with Serratia containing only a GFP marker. These results hint that there may have been a contamination in the experiment, and so a fourth repetition took place. From the results, we could observe that the plasmids are more lethal to two-days-old larvae rather than the four-days-old ones. Also, the results show that the plasmids harboring Cry11Aa and Cry11Aa+p20 are the deadliest and show the greatest death rates. This was confirmed in an ANOVA test, with a p value < 0.001.
These results are partial while we still haven't constructed our final polycistronic plasmid yet, containing all 3 toxic subunits of Bti that were chosen (please see "Plasmids Construction" in our Results to learn more).
Another feature we observed during the toxicity assays was that dead-infected larvae turns red (due to the color of the infected bacteria when exposed to oxygen), which could be seen with the naked eye. The control, which contained live larvae that were not treated, remained white throughout the experiments. The difference can be viewed in the figures below.
Conclusions from the Toxicity Assays:
From the toxicity assays, we can conclude that our ‘minimal plasmids’ harboring only one subunit of the toxin show a certain toxicity level, some more than others. In terms of larvicidal death, we could conclude that our plasmids indeed work well. The results prove that the plasmid containing Cry11Aa and Cry11Aa + p20 show the lowest survival rate for larvae, as predicted in the literature indicating the most toxic Bti subunits specifically to the Aedes aegypti mosquito[1].
These results are very beneficial to our project, as we can conclude that our constructs are indeed lethal to the larval stage of the mosquito. Bear in mind that we have yet to construct our final plasmid, containing 3 toxin subunits of Bti, and test them on live larvae.
From the observation of the dead larvae turning red, we can assume that there is a phenotype that indicates the Serratia's native behavior, that while bacteria prosper, some metabolic pathways are initiated, and a byproduct of this initiation is the prodigiosin pigment (red)[2]. This phenotype reconfirms our analysis to immediately detect dead larvae, and that the cause of death is our engineered bacteria. Since the larvae that were not treated (the control in these experiments) did not possess such a phenotype and remained white.
References
[1] Ben-Dov, E. Bacillus thuringiensis subsp. israelensis and Its Dipteran-Specific Toxins. Toxins 2014, 6, 1222-1243.
[2] Tanaka, Y. Temperature-Dependent Bacteriostatic Activity of Serratia marcescens. J-Stage 2004, Volume 19.3, 236-240.
Examine the presence of bacteria in hatched larvae from infected eggs:
After we examined the transfer of Serratia marscecens bacteria with a plasmid containing a GFP gene from the adult mosquito to the eggs in two mechanisms (vertical and horizontal), and the toxicity of our designed plasmids, we decided to explore the transfer of bacteria from the eggs to the hatched larvae.
We hatched larvae from the vertical experiment in the transfer section of our designed project, while we left the egg casisngs and vatman paper in the hatched container. We assumed that, naturally, the egg casings are being left in the water source and serve as an organic food source for the hatched larvae, while also serving for distribution of our transgenic bacteria.
We allowed the larvae to grow for four days, and afterwards washed them externally with 70% ethanol and 1% PBS, in order to make sure the source of the bacteria is internal. Then we crushed them and plated the homogenate with an appropriate selection (Carbenicillin [CRB] X5). After incubation in the suited conditions, we measured the fluorescent signal in a plate reader and discovered that the Serratia indeed transfer from the adult mosquito to the eggs and to the hatched larvae(Figure 1). Using a t test, a significant difference was viewed between the untreated larvae and the infected larvae.
Examine the distribution of the transgenic bacteria through a water source:
We did not settle for the infection mechanism of the hatched larvae from the infected eggs only. We wanted to examine the infection of the hatched larvae on regular larvae, and on a water source.
To do this, we built a specific container. In its center we hatched infected eggs with Serratia, and in its walls we hatched non- infected eggs. The container was built in a way that the larvae could not mix, while the bacteria did. We allowed the larvae to grow for four days, crushed them with the external water, and plated the homogenate with an appropriate selection (CRB X5). After incubation in the suited conditions, we measured the fluorescent signal in a plate reader and discovered that the Serratia indeed transfer from the hatched larvae from the infected eggs to the water source and to the regular larvae(Figure 2). Using a t test, a significant difference was viewed between the untreated larvae and the infected larvae.
Behavior tests of mosquito larvae:
During the toxicity assays, we filmed videos of the infected larvae in the containers with our transgenic bacteria and analyzed their behavior. We took a video of the live larvae in order to analyze their movement ability in the water. Using a pipette, we touched the water near the surface where they were floating in order to stimulate the larvae to dive and "avoid" predators encounter and to analyze their behavior reaction.
During the videos, we noticed a unique behavior of the infected larvae with our transgenic Serratia expressing the toxin subunits. While the control larvae showed great movement ability and responded to our stimulation by diving rapidly into the bottom of the container, the infected larvae had a less and weaker movement ability in general, when some infected larvae barely moved at all. When we repeated our stimulation by tapping the water surface with a pipette, the infected larvae had passive responses, and sometimes did not move at all. To make sure they aren't dead, we picked the infected larvae up using the pipette. Only then did they start showing a bit of movement, which confirmed they were still alive.
We confirmed that this behavior did not appear in the control experiments of uninfected larvae and in the larvae infected with wild-type Serratia. In general, most of the infected larvae died, and we assumed that this behavior was a result of the toxin subunit expression by the Serratia.
We could conclude that, although a few infected larvae could potentially recover to normal, the ones displaying this behavior would have a harder time surviving in nature, in terms food gathering and avoiding predators, and in general would not be able to survive in more stressful environments than in the laboratory conditions.
Live untreated larvae feeding on dead infected larvae:
During our toxicity assays, we noticed a mysterious thing. Every day we came to count the dead larvae and every day we counted fewer dead larvae than the day before.
We placed a camera to see what happens and we found that the live larvae eat the dead larvae. We looked at the larvae that ate an infected larva under a confocal microscope and detected a fluorescence signal in the gut. We realized that we got an even better distribution then we thought. Not only the larvae that will hatch from infected eggs will eat the toxin and die, but also larvae that hatched from non-infected eggs but were laid in the same small water source, can eat those larvae and die.
Below are photos of dead, infected larva, under a confocal microscope (x10). Fluorescence could be seen throughout the larva (Figure 4). Next, a live, untreated larva which fed on the dead, infected larva. Under a confocal microscope (x10), a fluorescnce signal is noticeable inside the live larva as well (Figure 5).
Difficulties we encountered during the experiments:
First, it is important to state that throughout our project, due to technical difficulties, we were unable to detect whether the larvae hatch infected from the eggs, or whether the hatch uninfected and become infected later from the casings of the egg itself.
While we watched the eggs under a confocal microscope they hatched (OMG!) And we tried to examine the larvae themselves. Because the eggs and larvae themselves showed very high autofluorescence, we were unable to detect a fluorescent protein signal, and we tried to plate them individually in bacterial cultures. We had technical difficulties and were looking for another way to examine the tranfer of the Serratia to the larvae. However, we did not formally conclude that the larvae hatch as they carry the bacteria on their bodies, or whether they infect them in some other way.
Conclusions:
We have shown that the Serratia can pass from the eggs to the larvae originating from those eggs. In addition, we have shown that the Serratia thrives in a water source, which may indicate an excellent infection mechanism to other larvae in the water sources. We have also seen that live, untreated larvae feeds on their fellow dead, infected larvae, and get infected as well. This could also account for better distribution of our product. We were able to find a simple and effective delivery system of bacteria - from the infected eggs to a water source.