Design

Classic molecular biology relies on established processes and design patterns that work well in laboratory and industrial applications. Both hardware and DNA constructs are optimized for maximum yield and efficiency, always striving towards more product at less costs. In environmental synthetic biology, this approach is bound to fail.

Environmental applications require a paradigm change

Optimizing biologic systems for yield works under tightly controlled conditions. Cells need to be under constant surveillance, and slightest changes quickly lead to a total failure of the production chain. If we decided to use classic molecular biology approaches to solve environmental problems by releasing GMOs which serve a designed function in the lab, probably not much would happen. Over years of optimization, we disregarded the self-regulating abilities of biological systems to a point at which we call cell growth regulators “toxins”¹.

We need to drastically change our view of a well-designed genetic circuit. Instead of using constitutive or strong inducible promoters, we need transcriptional regulators which respond to the environment. Instead of the strongest ribosome binding sites we need carefully selected combinations to minimize metabolic burden in energy-demanding systems. Instead of bioreactors we supply with virtually limitless energy, we need elaborate hardware that can protect and sustain environmental synthetic biology applications, even if they are not watched day and night.

At iGEM Hamburg, we made sustainability of environmental synthetic biology an emerging reality. Here, we present design patterns we applied to the S.H.I.E.L.D., as a proof of concept we are confident many future environmental synthetic biology applications can use as a foundation towards a future in which we do not need to change our environment to meet our inabilities, but let our environment shape the abilities of our intelligent biological systems.

Basic Parts as building blocks for sustainable circuits

We created a set of basic parts to both monitor and respond to environmental changes. Our environmental change sensors include glucose-repressed promoter MlcRE (BBa_K2588000), improved cold-shock promoter HybB (BBa_K2588001), and RNA-based transcription inverter icsA 5’ UTR/RnaG120 (BBa_K2588003). Using these, as well as combinations of them and Parts of the registry, simple ways to monitor changes in nutrient availability and temperature can be monitored and converted into transcriptional signals.

How an appropriate respond to these signals looks like heavily depends on the application. A common response which we employ in the S.H.I.E.L.D. is repression of bacterial growth. There are many applications that already have implement to limit the number of bacteria in a system. These applications most often rely on killing surplus bacteria using a kill switch. While this certainly meets its goal, it is an unreasonable waste of energy and supplies. Instead, we propose to use regulator proteins from E. coli genome that already fulfil the task of responding to limiting factors with stalling cell growth and division. We identified key regulators of all important steps of cell growth: cspD (BBa_K2588005) inhibits DNA replication², mraZ (BBa_K2588007) inhibits peptidoglycan synthesis³, and, cbtA (BBa_K2588006) and sulA (BBa_K2588008) inhibit cell elongation and division^4,5. Using a carefully selected combination of these, E. coli growth should be stalled or even completely inhibited without wasting resources.

Assembling sustainable constructs

While providing basic parts is good, the largest energy and resource savings can be made with the largest and most complex constructs that fulfil difficult tasks. For the S.H.I.E.L.D., we employ multiple highly complex partly synthetic, partly natural metabolic pathways for odour bait- and insecticide synthesis. Just overexpressing each enzyme of each pathway would certainly lead to production of each molecule, but in the S.H.I.E.L.D., E. coli needs to be able to maintain production without maintenance over extended periods of time. This requires special design considerations when assembling each pathway, which may impact efficiency and yield, but allow for a prolonged functionality.

Making a trap attractive to its victim

The ability to attract malaria mosquitoes is one of the most important functions of the S.H.I.E.L.D. With the goal of creating a lure that would imitate the mixture of volatiles on the human skin, attractive to malaria mosquitoes, we searched through generations of academic and industrial research. As it turns out, bacteria present on the human skin are mainly responsible for the production of these volatiles and even more intriguing: volatiles produced by skin bacteria are still more attractive to mosquitoes than the same volatiles produced synthetically (source: Smallegange et al 2010).

Rather than applying one attractant alone, the combination of many attractants has been shown to lure human targeting mosquitoes the most effectively (Verhulst, Okumu, Smallegange). We therefore chose lactic acid, myristic acid and 3-methyl-1-butanol as lures to implement into the S.H.I.E.L.D., a blend that has previously been shown to be a very effective lure (Verhulst).

Each of these synthesis pathways has been already implemented in E.coli bacteria, for various industrial applications. As we aim for sustainability of the S.H.I.E.L.D., we designed constructs for the production of each compound. Because of the large metabolic load, constructs would be implemented only separately in designated cultures within the E.coli compartment.

The metabolic load is additionally reduced by the glucose-dependent promoter BBa_K2588002. It ensures lures are only produced in the presence of glucose and therefore, the cells are not additionally burdened.

Lactate

Lactate dehydrogenase A (ldhA) is an enzyme catalyzing the synthesis of lactate from

pyruvate, as part of the fermentation metabolism usually activated upon anaerobic conditions. Lactate is used as part of the so called “basic blend” (Mathew, Verhulst), which is a combination of compounds that have been proven to be reliable attractors in the context of research of Anopheles attractants. Therefore we introduced ldhA (BBa_K2588009) as novel basic part and characterized it using the characterization construct pSB1C3-BBa-K206000-BBa_0034-ldhA (BBa_K2588041).

Myristic acid

Like lactate myristic acid has been used as part of the Anopheles mosquito attracting basic blend (Mathew, Verhulst). We therefore designed a module, where we implemented a pathway that leads to a significantly higher production of C-14 fatty acids, including myristic acid. For this module we submitted 5 novel genes that encode 4 subunits: accA (BBa_K2588010), accB (BBa_K2588011), accC (BBa_K2588012), accD (BBa_K2588013),

tesA (BBa_K2588014). We assembled all genes into one module, to be regulated by glucose through the NOT-MlcRE promoter (BBa_K2588022). For characterization a construct was designed with an arabinose induced promoter (BBa_K2588023), with the goal of verifying myristic acid by HPLC (high performance liquid chromatography).

3-Methyl-1-butanol (3MB) is an organic compound that has been isolated as one of fourteen volatiles produced by human skin bacteria, that showed attracting effects on A. gambiae. Tested on its own 3MB was shown to be the best attractant that was able to increase the effectiveness of the basic blend three fold in a lab and field test setup. We chose key enzymes from an engineered pathway that catalyzed the reaction from pyruvate to 3MB. Those enzymes are encoded by LeuABCD Operon (BBa_K2588015), kivD (BBa_K2588016), ADH2 (BBa_K2588017) and were designed to be assembled to a glucose controlled (NOT-MlcRE) module (BBa_K2588024) to be implemented in the S.H.I.E.L.D. To ensure, the expression of ADH2 and kivD, they are located upstream of the LeuABCD gene. To characterize the module, an arabinose induced promoter was chosen (BBa_K2588025), again with the goal of verifying 3MB by HPLC.

However, mosquitoes are not only attracted by volatiles alone. Factors like heat and moisture play a role in the attraction of mosquitoes (Olanga) too. Thus, we searched for ways to implement heatproduction into our design.

We decided to employ BBa_K410000, a registry Part submitted by GeorgiaTech in 2010.

BBa_K410000 is composed of the cold-shock promoter HybB (BBa_J45503) and the alternative oxidase a1 from Nelumbo nucifera (NnAOX1a), which we submitted as a basic part as BBa_K2588018 since it was not individually submitted by GeorgiaTech themselves. NnAOX1 salvages electrons from the respiratory chain for heat generation. Combined in BBa_K410000, they have the function of inducing heat production at low temperatures.

The S.H.I.E.L.D. is heavily focused on sustainability. We liked GeorgiaTech’s approach of coupling heat production to a negative feedback loop to ensure E. coli do not take damage from overexpression of a transmembrane protein constantly salvaging electrons from their respiration chain. That HybB is active only at low temperatures, and only shows low expression at higher temperatures, compliments our philosophy of only using moderate expression to ensure viability of cells over long periods of time. HybB is active only at low temperatures, and only shows low expression at higher temperatures, which complements our philosophy of only using moderate expression to ensure viability of cells over long periods of time.

We expect the intrinsic feedback loop of BBa_K410000 to work even at higher temperatures, despite its overall lower expression. When we analysed BBa_K410000 to confirm that it was properly designed, HybB caught our attention. Despite claiming to be a promoter sequence, an alignment against the E. coli genome revealed that BBa_J45503 not only comprises of the clean promoter sequence, but also contains the Hyb0 5’ UTR and the first 100 bp of Hyb0 coding sequence. When using BBa_J45503 in a standard assembly with a downstream gene of interest, we expect this to result in a conflict in translation of Hyb0 with the gene of interest encoded on the mRNA, which we expect to reduce the expression of the gene of interest for the benefit of truncated Hyb0 as unwanted side-product.

Soft Growth Inhibition

We employ multiple proteins that are commonly referred to as toxins. This connotation is based on their biology to inhibit E. coli cell growth, completely disregarding their evolutionary function: To respond to environmental changes to save and protect cells from overexpending on energy and resources. Overexpressing these genes is a difficult endeavour. Too little, and they are silenced by numerous regulatory systems without impact on growth. Too much, and inhibition of only one part of cell growth leads to certain death by unbalance of cellular systems.
To circumvent both issues, a careful regulation of multiple simultaneously expressed growth inhibitors is needed. We have four basic parts to choose from: cbtA, cspD, mraZ, and sulA, each repressing a crucial step of cell division. Furthermore, numerous combinations of ribosome binding sites and promoters are available to build a regulation that fits all needs. But:
We do not know anything about optimal concentrations of any of the growth inhibitors.
If we were to select ribosome binding sites by intelligent design, we would be bound to fail. The only way to find a working growth inhibition module is to trust in biology to find the optimal way itself.
For each growth inhibitor, we designed multiple GoldenGate-compatible primers each containing a ribosome binding site of arbitrary strength from the registry. We amplified each gene by PCR using all its respective primers, and performed a GoldenGate Assembly employing equal amounts of each gene with each ribosome binding site. This way, we produced a mixture of all RBSs with all genes, assuming that a few viable combinations could be among them. Since any working growth inhibition module would not result in a colony on a transformation plate, we employed DAP-repressible promoter dapAP (BBa_I718018), and brought transformed E. coli onto plates containing diaminopimelic acid.
We picked colonies and sent them for sequencing with insightful results: The only growing colonies contained strongly mutated constructs, in which only few or none of the growth inhibitors could be expressed, due to nonsense mutations or inactivation of RBSs. We identified one of these constructs which we could rescue by carefully designed PCR with primers which repaired two of these nonsense mutations. We brought the resulting combination of BBa_B0034-cbtA-BBa_B0032-cspD-BBa_B0032-mraZ into pSB1C3 and characterized the resulting construct.

Putting the E in S.H.I.E.L.D.

Black Scorpion alpha Insect Toxin BjaIT (not registered as basic part) is a highly potent biological insecticide without toxicity to mammals⁶. Recombinant production in E. coli has been described before for industrial applications, in which BjaIT is produced intracellularly, and purified using an affinity tag after lysis⁷. In the S.H.I.E.L.D., E. coli needs to export BjaIT itself. We employ the type II secretion system composed of HlyB, HlyD, and TolC to export BjaIT from the cell. The secretion system was first registered at iGEM by UNICAMP EMSE Brazil as BBa_K554013. It can be used to export any protein with a HlyA signal peptide (BBa_K554002) at its C terminus. Since we were unsure whether BjaIT is active with the HlyA signal present, we incorporated an outer membrane protease T (OmpT, BBa_K2588020) site into the linker between BjaIT and HlyA. We expanded BBa_K554013 functionality by exchanging its promoter to more general use cases using pBAD promoter BBa_K206000, and expanded the mRNA by OmpT coding sequence. We registered the resulting composite part as BBa_K2588030. Since all proteins expressed by this parts are transmembrane proteins and thus exert significant stress on cells expressing them, we suggest not fully inducing the pBAD promoter.

We registered BjaIT with linker and OmpT site as basic part BBa_K2588019. While BjaIT is harmless to mammals, there are similar proteins which could be used for harmful causes. Note that we do not annotate or explain its sequence in detail due to biosecurity concerns.

Complementing sustainable synthetic biology with sustaining hardware

Using genetically engineered organisms in unmaintained, self-controlling field applications is not a prospect we greet light-heartedly. On the other hand, our bacteria would probably not survive for a long time due to the heavy duty synthetic biology they have to perform. With both problems in mind, we do not think that there is a biological solution to cater sustainability, biosafety, social acceptance, and regulation at the same time.

That is how the Sustainable Human-Imitating Elimination and Lure Device was born.

We began our hardware design process by defining all necessary functions:

Producability

The S.H.I.E.L.D. has a high task: Preventing malaria where all other methods have failed. Economic restrictions are a major reason for that failure. Therefore, any new solution needs to cost less for the same or a better effect. While we are positive that synthetic biology will produce better solutions, producing fancy hardware can quickly produce costs that render an application not feasible for mass production.

To produce the S.H.I.E.L.D.’s prototype, we employed 3D printing. However, for mass production, this method is far too time consuming and expensive. Designing the S.H.I.E.L.D., we therefore employed simple shapes that could also be produced with less elaborate and cheaper production methods like injection moulding. Injection moulding requires producing parts which do not have internal overhangs, so they can be taken out of their mould. Since the S.H.I.E.L.D. requires complex shapes, this means that it has to be made of several separate parts. Our prototype consists of six individual parts that have to be assembled prior to use. Between every two parts, there are seams. This brings up the problem of biosafety.

Biosafety

The S.H.I.E.L.D. needs to ensure that no GMOs can exit, and no biological agents can enter from outside. We employ durable NDR-70 sealing rings between each two separate parts to prevent release and contamination.

Odour bait molecules, the insecticide, oxygen and CO₂ need to be able to enter and exit the S.H.I.E.L.D. Therefore, the S.H.I.E.L.D. cannot be sealed entirely. As interface between S.H.I.E.L.D. interior and environment we selected a commercially available nitrocellulose nano filter. Usually used to filter sterilize water in the laboratory, it is the most economic, most established, and most fit alternative for this task.

Odour bait release

Myristic acid, 3-methylbutan-1-ol, and lactic acid have low solubility in water. They can be attractive to mosquitoes due to their volatile nature, as they quickly evaporate from the human skin. To emulate this effect, we needed a surface from which odour bait molecules can evaporate evenly.

Kill visiting mosquitoes

Once attracted by odour baits, mosquitoes need to take up the insecticide from the S.H.I.E.L.D. In the laboratory, mosquitoes are fed with human blood in a flat bowl spanned with parafilm. Mosquitoes can sting through the parafilm to take up the blood. The parafilm is destroyed in the process. For the S.H.I.E.L.D., we needed a solution that enables mosquitoes to take up insecticides without impairing successive function.

To emulate human skin for both odour bait release and mosquito behaviour, we employ a self-healing poly ammonium salt hydrogel. It simultaneously serves as a reservoir for the insecticide, the lures and as a surface, on which the mosquitoes can land and sting into. How self-healing?! The hardware design of the trap enables condensation to be caught in the morning. The hydrogel is designed to be stable all year long in rain - and dry seasons in the malaria regions.

Sustain a bacteria culture

The S.H.I.E.L.D. is designed to be a long-lasting mosquito trap used in the warm malaria regions all over the world. To control the amount of bacteria we could have used a kill switch system. Instead of this, we decided to transform a growth inhibition module in our E. coli, which just interrupts cell division, but not the functionality of the bacterium. So the bacterium will live for a long period of time and doesn't grow until death phase. Usually for a sustainable bacteria culture there has to be a constant addition of medium, as the iGEM team Düsseldorf, we collaborated with, examined. By creating a co-culture of E. coli and cyanobacteria we were able to eliminate the necessity for additional medium. The cyanobacteria, placed in the transparent top of the S.H.I.E.L.D., will provide glucose by photosynthesis and other nutrients by its basic metabolic reactions. Thus the E. coli culture can last for extended periods of time.

Perform without human interference

With the growth inhibition module the S.H.I.E.L.D. will never be overgrown with E. coli. The device can be used by everybody, even if they are untrained with bacterial handling. Also there is no need for maintenance of the S.H.I.E.L.D. With the co-culture of E. coli and cyanobacteria the medium does not have to be refilled as the cyanobacteria provide all needed nutrients for E. coli.

Team:Hamburg/Design