Living Materials as Platform for New Technological Innovations

The fundamental problems impeding the successful realisation of living materials are the demanding requirements for an environment in which living organisms can be maintained in a viable, functional way and biosafety concerns of the escape of genetically modified organisms into the environment (1). This lead us to establish design requirements, preferences and constraints that apply to our living material. These are general concepts, as the living material we propose can be used for many different applications. For each application, other specific criteria apply as can be seen in our application scenario of a living material as a patch for burns and chronic wounds (see below).

Requirements:
• Modularity: The living material must be capable of performing different functions by integrating genetically engineered E. coli with different functionalities.
• (Bio)safety: The living material must be designed in such a way that bacterial escape from the material is prevented and the material is safe to use.
• Viability: The living material must be a permissive environment for E. coli, allowing their survival and functioning in the material for a prolonged period.
• Robustness: The living material must have appropriate mechanical properties to yield a living device that can perform in the desired situation.

Preferences:
• Affordability: The production and maintenance of the living material is not too expensive for the application.
• Ease of manufacturing & scalability: To allow widespread use of this new technological innovation, production should be straightforward and allow for upscaling of the manufacturing procedure to industrial scale.
• Sustainable: The living material and its production process must be environmentally-friendly.
• User friendliness: The material should be easy to use by the targeted user for a certain application.

Constraints:
• We have to use E. coli as living organism in our material as we were limited to bacteria available at our university.

Our material of choice for a living material is a hydrogel, a polymer network with water as dispersion medium. Hydrogels are inherently biocompatible, permeable and have a high water content. This makes them an ideal matrix for living materials (1). Their use as cell delivery vehicles (2) and scaffolds for tissue engineering (3) proves their suitability as viable environment for living cells and organisms. We quickly realized that our hydrogel has conflicting demands. It should prevent bacterial escape but also have an open porous structure to allow a viable environment with free diffusion of oxygen, nutrients, signalling molecules and proteins. Simple physical entrapment of the bacteria does therefore not suffice. We envisioned that the solution to this conflict is to actively tether E. coli tightly to our material. In this way, we prevent the diffusion of E. coli out of our gel while allowing the free transport of solutes. Our hydrogel-anchor combination should thus act as a sort of prison, keeping the E. coli bacteria alive and happy, but restricting their movement. The concept of Gelcatraz was born.

Figure 1. Adhesin structure.

Adhesin

The first step to realizing Gelcatraz started with the identification of a suitable bacterial adhesin to anchor the E. coli bacteria to the hydrogel. Recently, a large 1.5 MDa adhesin was found in the Antarctic bacterium Marinomonas primoryensis. This large adhesin, named M. primoryensis ice-binding protein (MpIBP), is 0.6 µm long and consists of more than 130 domains. Its structure is shown on the right, click here for background information about the adhesin. 120 of these domains encode for the same immunoglobin-like β-sandwich that rigidifies upon binding of Ca²⁺ and acts as an extender to project the binding domains positioned on top of this extender chain into the medium (4).

In our design process, we envisioned that we could use the sugar binding domain of MpIBP, from which we created a new basic part , to bind to a sugar-based polymer. On the other parts of the wiki, we will refer to the shortened version of MpIBP combined with the sugar binding domain as ‘adhesin’ . In this way, we can tether the E. coli bacteria to our hydrogel.

In its native host, the MpIBP adhesin is exported out of the cell using the single-step type I secretion machinery. The type I secretion of the adhesin requires four protein domains. Hemolysin A (HlyA) is a protein domain C-terminal to the adhesin protein, signalling for type I secretion. Hemolysin B (HlyB), Hemolysin D (HlyD) and TolC are the other proteins which are responsible for forming a trans-periplasmic export channel upon interaction with the HlyA domain, as shown in the figure. In type I secretion, all proteins passing the conserved the β-barrel pore of TolC, must remain unfolded. The RIM and RIN domains of the adhesin (see Figure. 1) are the only domains that fold 'independent of calcium. Calcium is naturally present in millimolar concentrations in sea water, the natural environment of M. primoryensis. The entire adhesin is translocated to the extracellular space where the domains fold, but the RIM and RIN domain are already folded and prevent translocation through the TolC protein. The RIM protein fits inside the TolC pore, with the RIN being a bulky anchor preventing total release of the adhesin. The RIM & RIN domain act as a TolC β-barrel plug (4). In our design, we want to use this ingenious system to express the adhesin on the surface of E. coli. However, the native adhesin is too large to be produced in E. coli. To avoid problems with expressing the adhesin in E. coli, we reduced the amount of RII repeats in the gene encoding the adhesin to fifteen.

Figure 2. MpIBP Type I secretion system.

Plasmid design

As we adhere to the engineering principle of modularity, we want all our individual components of our project to be able to be altered or optimized separately. Therefore, we cloned the shortened MpIBP adhesin into a PET24 vector (kanamycin resistance) under control of a T7 promotor inducible by adding isopropyl β-D-1-thiogalactopyranoside (IPTG). TolC is already constitutively expressed in E. coli (6). To enable expression of HlyB and HlyD, both proteins were cloned into a pSTV28 vector (chloramphenicol resistance) under control of of a lac promotor inducible with IPTG.

Hydrogel

To find an appropriate hydrogel for our adhesin to bind to, we used the fact that this sugar binding domain has its highest affinity for glucose (7). In our design of a living material platform, the following requirements were established to find a gel:

• The hydrogel must be biocompatible (non-toxic to E. coli).
• The hydrogel must have an open porous structure for efficient and fast nutrient, oxygen and solute exchange for survival and functioning of the E. coli bacteria.
• The hydrogel must be mechanically stable to prevent bacterial leakage and to be easy to handle.
• The hydrogel must not degrade for a sustained period of time.
• The hydrogel synthesis must be relatively straightforward and reproducible.

Dextran is a polymer that consists of α-1,6-linked D-glucopyranose (glucose) residues. This glucose-based nature makes dextran an ideal polymer for our MpIBP adhesin to bind to. The polymer is non-toxic and has been investigated for tissue engineering applications and is already used in the clinic as blood plasma substitutes (8). A hydrogel from dextran polymer chains can be synthesized using radical polymerisation of glycidyl methacrylated dextran (9). By adapting their protocol, we designed a supermacroporous dextran hydrogel with pores ranging from 1 to 100 micrometer that suits our needs. This unique morphology allows for the rapid and non-restricted mass-transport of solutes of virtually any size (9). Combining all elements described above, we envision our living material to be as shown in the figure below:

Testing Our Living Material Platform

To demonstrate the performance and behaviour of our designed living material platform, we designed and carried out multiple experiments of which a more detailed analysis of the results can be seen here. Scanning Electron Microscopy experiments were performed to find the optimal hydrogel polymerisation conditions to produce the desired microporous structure. When this yielded mechanically unstable hydrogels, shaking experiments with different hydrogel synthesis conditions were performed to find the appropriate conditions for stable hydrogels.

For the adhesin, we first investigated the dextran binding capability of the isolated MpIBP sugar binding domain. In this experiment, the MpIBP sugar binding domain is incubated with dextran beads and washed, after which we could show that the protein remains bound to dextran using SDS-Page and that binding occurs in a calcium-dependent fashion. Next, we used Fluorescence Activated Cell Sorting (FACS) and flow cytometry experiments to investigate the binding of red fluorescent E. coli bacteria expressing the MpIBP adhesin to dextran beads and could confirm binding of the adhesin to dextran.

Integrating both the adhesin and hydrogel together as a complete platform, a hydrogel cell seeding and washing experiment was devised using small gels casts in 96-well plates. After overnight seeding of the gels, washing and subsequent crushing of the gel was used. Plating these fractions, we could demonstrate that the adhesin causes a three-to-four fold higher retention of bacteria inside the hydrogel.

Unfortunately, we came to the realization that seeding our gels this way resulted in very inconsistent and difficult to reproduce results. That is, the results where difficult to standardize or accurately quantify. This goes against multiple of our formulated engineering principles and requirements for the living material platform. We therefore had to completely redesign both our hydrogel and the seeding process. Using Polydimethylsiloxane (PDMS) microfabrication, we designed a cubic mold in which a plastic pin can be inserted. When we cast the hydrogel in this mold, we create a hydrogel with a chamber in the center. This chamber can be loaded with bacteria to seed the hydrogel. After seeding, a plastic plunger can be used to close the gel. With this reinvented prototype, seeding experiments yielded more consistent results, observing a difference in retention of several orders of magnitude betweenbacteria with and without adhesin.

Figure 3. Seeded hydrogel with chamber.

Fighting against Infections of Burns and Chronic Wounds

As a team, we felt that convincing the public about the potential of living materials is difficult without a clear, concrete and appealing application scenario. To illustrate the potential of living materials, we decided to develop it into a living skin patch capable of the continuous secretion of bacteriocins to combat bacteria responsible for infections in patients suffering from burns and chronic wounds. For a detailed analysis of the background of the problem and our interviews with experts and stakeholders please see our Human Practices section and Applied Design.

Figure 4. The illustration of burns and chronic wounds healing patch.

As mentioned above, this specific application scenario leads to additional requirements, preferences and constraints. As a team we formulated the following:

Requirements:
• Safety & Biocompatibility: The bacteriocin secreting patch should be safe to use for patients and for healthcare professionals and should be easy to sterilize before seeding with bacteria to avoid the introduction of other bacteria near the wound.
• Effectivity: The effect of the patch should be limited to offending bacteria, with Staphylococcus aureus as our primary target
• User friendliness: The patch must be easy to use for healthcare professionals such as doctors and nurses. It must also be comfortable for patients. In contrast to the living material, user friendliness is for wound healing no longer a preference, but a requirement.

Preferences:
• Modularity: Ideally our patch can target different bacterial strains by simply changing the bacteriocin encoding gene.
• Longevity: Our patch should be able to perform its function for a prolonged period, maximizing the use of continuous secretion and limiting the amount of patch changes patients experience.
• Storability: The patch must have a suitable shelf-life and should be straightforward to store to compete with already existing treatments and to avoid availability issues.
• Approval: The base material of the patch (the hydrogel) should ideally be already approved by the U.S. Food and Drug Administration (FDA) and European Medicines Agency such that only the integration of living organisms requires investigation and approval by regulatory institutions to accelerate implementation.

Constraints:
• The bacteriocins secreted by the patch must not be toxic to E. coli and must only be active against the specific strains of bacteria causing problems.

Our living dextran hydrogel is very suitable for wound healing. Our hydrogel is stable at physiological conditions for months and can be stored in a dried state and re-swollen prior to use, avoiding availability problems (9). Dextran is already approved by the FDA for clinical use for plasma volume expansion, antithrombotic agent and peripheral flow promotion (10). Furthermore, the hydrogel can be easily sterilized using radiation (11).

To target S. aureus, we used lysostaphin, a zinc endopeptidase that cleaves the pentaglycine cross-bridges present in the bacterial cell wall of S. aureus. Lysostaphin exhibits its lytic effect not only against regular S. aureus, but also against methicillin-resistant S. aureus (MRSA), vancomycin-resistant S. aureus and S. epidermidis. Recently, it was demonstrated that hydrogels delivering lysostaphin outperform prophylactic antibiotic treatment and soluble lysostaphin treatment in murine femur fracture models (12). The iGEM teams HIT-Harbin 2012 and Stockholm 2016 have used lysostaphin before in their projects to target S. aureus. However, both teams relied on cell lysis to free the lysostaphin. For our application, this approach is useless as releasing bacterial cell lysate into the wound will only act as food for the offending bacteria causing the infection. Moreover, cell lysis does not offer a solution for sustained release of lysostaphin. The solution to this problem is to make lysostaphin secretable. This will allow sustained release of the active enzyme and ensures that the lysostaphin reaching the wound is of high purity.

To minimize the burden on our modified E. coli bacteria already expressing the MpIBP adhesin, we make use of the same type I secretion system. This can be achieved by fusing the HlyA domain C-terminal to the lysostaphin construct. To optimize expression, we first codon optimized the HIT-Harbin 2012 truncated lysostaphin construct . Fusion to HlyA will enable secretion, circumventing the use of cell lysis, improving the biokiller device developed by HIT-Harbin 2012 . With other iGEM projects in mind, we chose a thrombin linker to fuse HlyA to lysostaphin as it can be readily cleaved with the thrombin enzyme and added a His tag to the HlyA domain for protein detection and purification purposes. If all systems work as designed, culturing of induced E. coli transfected with our designed construct and the HlyB/HlyD secretion proteins results in medium filled with our lysostaphin-HlyA fusion protein. Running the medium over a nickel affinity column isolates the fusion protein and the thrombin linker can be cleaved with the thrombin enzyme, resulting in pure truncated lysostaphin without the HlyA domain. Purification of a recombinantly expressed protein from a medium sample has important advantages over intracellular expression followed by cell lysis and purification from the lysate. The need of cell lysis is eliminated, aggregation in inclusion bodies in the cytosol is prevented, toxic effect of overexpressed proteins are reduced and downstream purification and processing steps are reduced significantly. (13) Our construct can benefit future teams by making purification of lysostaphin straightforward. For example, Stockholm 2016 intended to functionalize spider silk with lysostaphin but did not manage to successfully purify lysostaphin. We also created a different lysostaphin construct already containing an IPTG inducible T7 promotor allowing other teams to start right away.

Plasmid design

To adhere to the engineering principle of modularity, the lysostaphin component has to be able to be controlled individually. Therefore, we cloned our lysostaphin-HlyA fusion protein into a pBAD vector, under the control of an AraC promotor inducible with arabinose.

Adapting Our Patch Design

After several conversations with experts on the topic of burn and chronic wounds, we found out that besides S. aureus, P. aeruginosais also a major cause of problems in wound care of patients with burns and chronic wounds. As our lysostaphin-HlyA has a modular design, we changed the coding sequence of lysostaphin to that of pyocin S5generating a new BioBrick in addition to lysostaphin. Pyocin S5 is a bacteriocin active against P. aeruginosa(14), enabling our patch to combat both simultaneously. With the conversations we had with experts, we also learned about the importance of a kill switch besides our adhesin to neutralize the danger posed to the environment if bacteria would be able to escape despite our adhesin. Therefore, we also worked on this in our design. Please see our Applied Design section for the kill switch and for an analysis of the potential impact of our bacteriocin secreting patch.

Putting Our Patch to the Test

To test the design of our bacteriocin secreting patch, we first carried out protein expression experiments to show that our modified E. coli can express the lysostaphin. Next, we co-transformed our pBAD-lysostaphin part with the HlyB/D secretion system and showed that secretion of lysostaphin works . To prove the functionality of our construct, multiple experiments were carried out in collaboration with the PAMM foundation. We designed the experiments, but PAMM carried them out for us as we were not allowed to work with S. aureus due to biosafety regulations. The PAMM foundation is the regional center for infectious diseases and pathology of South East Brabant, the Netherlands. The experimental plan of showing the functionality of the patch is as follows:

• Show that lysate of our induced genetically modified E. coli is toxic against S. aureus, but not against Streptococcusagalactiae (a gram positive bacterial train used as negative control) by pipetting cell lysate onto Streptococcusculture and show that uninduced E. coli is not toxic.
• Show that medium (purified, no more antibiotic present) of our induced genetically modified E. coli is toxic to S. aureus, but not against Streptococcus agalactiae and show that uninduced secretion or secretion of only HlyA is not toxic against S. aureus to prove secretion of our construct.
• Show that a co-culture of our induced genetically modified E. coli is toxic against S. aureus, but not against Streptococcusagalactiae and show that a co-culture of uninduced modified E. coli lacks this toxic effect to demonstrate our E. coli results in the lytic effect and that this effect is not due to something else present in the medium (e.g. remnants of antibiotics).
• Show that purified medium of our induced genetically modified E. coli loaded into the hydrogel cube with chamber kills S. aureus to show that diffusion of lysostaphin through the gel works.

We were able to perform all the experiments described above successfully and could demonstrate the functionality of our lysostaphin-HlyA construct and the diffusion through the gel. For the results of the experiments in collaboration with PAMM, please click here. For Pyocin S5, we envisioned a similar strategy but were not able to complete it within time.

The last experiment is a complete proof-of-principle: Combining both the adhesin component of our living material platform to retain the E. coli bacteria inside the hydrogel continuously secreting lysostaphin to combat S. aureus. Unfortunately, we did not have time to finish this experiment.

Summary

Looking back at the initial goals of our project, we can conclude that we managed to develop a hydrogel with suitable mechanical and chemical stability with the right supermacroporous microstructure for our envisioned application. The effect of the adhesin to entrap the E. coli in our hydrogel and both the secretion and effectivity of our designed lysostaphin construct has been successfully demonstrated. All individual components of our living bacteriocin secreting patch work as intended. Therefore, we are convinced that the integrated wound healing patch will function as expected.

After finishing our project, we can review our engineering principles formulated before we started the design of our project. We achieved modularity by using unique promotors and vectors for each component to allow individual manipulation to optimize the integrated system. We used prototyping to generate different shapes and sizes of our hydrogel and used PDMS microfabrication to improve our design of the hydrogel. By performing all experiments at least in duplo and repeating them multiple times we ensured the reproducibility of our results. Via integration of our conversations we had with experts for Human Practice and Applied Design parts of our project, we implemented safety as a key feature in our design.

References

1. Liu X, Tang T-C, Tham E, Yuk H, Lin S, Lu TK, et al. Stretchable living materials and devices with hydrogel–elastomer hybrids hosting programmed cells. Proc Natl Acad Sci [Internet]. 2017;114(9):2200–5. Available from: http://www.pnas.org/lookup/doi/10.1073/pnas.1618307114
2. Zhao X, Kim J, Cezar CA, Huebsch N, Lee K, Bouhadir K, et al. Active scaffolds for on-demand drug and cell delivery. Proc Natl Acad Sci [Internet]. 2011;108(1):67–72. Available from: http://www.pnas.org/cgi/doi/10.1073/pnas.1007862108
3. Lee KY, Mooney DJ. Hydrogels for tissue engineering. Chem Rev. 2001;101(7):1869–79.
4. Guo S, Stevens CA, Vance TDR, Olijve LLC, Graham LA, Campbell RL, et al. Structure of a 1.5-MDa adhesin that binds its Antarctic bacterium to diatoms and ice. Sci Adv. 2017;3(8).
5. Thomas S, Holland IB, Schmitt L. The Type 1 secretion pathway - The hemolysin system and beyond. Biochim Biophys Acta - Mol Cell Res [Internet]. 2014;1843(8):1629–41. Available from: http://dx.doi.org/10.1016/j.bbamcr.2013.09.017
6. Zgurskaya HI, Krishnamoorthy G, Ntreh A, Lu S. Mechanism and function of the outer membrane channel TolC in multidrug resistance and physiology of enterobacteria. Front Microbiol. 2011;2(SEP):1–13.
7. Guo S. Unpublished.
8. Lévesque SG, Lim RM, Shoichet MS. Macroporous interconnected dextran scaffolds of controlled porosity for tissue-engineering applications. Biomaterials. 2005;26(35):7436–46.
9. Plieva F, Oknianska A, Degerman E, Galaev IY, Mattiasson B. Journal of Biomaterials Science , Novel supermacroporous dextran gels. 2012;(November):37–41.
10. Mehvar R. Dextrans for targeted and sustained delivery of therapeutic and imaging agents. J Control Release. 2000;69(1):1–25.
11. Szafulera K, Wach RA, Olejnik AK, Rosiak JM, Ulański P. Radiation synthesis of biocompatible hydrogels of dextran methacrylate. Radiat Phys Chem. 2018;142(October 2016):115–20.
12. Johnson CT, Wroe JA, Agarwal R, Martin KE, Guldberg RE, Donlan RM, et al. Hydrogel delivery of lysostaphin eliminates orthopedic implant infection by Staphylococcus aureus and supports fracture healing. Proc Natl Acad Sci U S A [Internet]. 2018;(12):201801013. Available from: http://www.ncbi.nlm.nih.gov/pubmed/29760099
13. Freudl R. Signal peptides for recombinant protein secretion in bacterial expression systems. Microb Cell Fact [Internet]. 2018;17(1):1–10. Available from: https://doi.org/10.1186/s12934-018-0901-3
14. Elfarash A, Dingemans J, Ye L, Hassan AA, Craggs M, Reimmann C, et al. Pore-forming pyocin S5 utilizes the FptA ferripyochelin receptor to kill Pseudomonas aeruginosa. Microbiol (United Kingdom). 2014;160(PART 2):261–9.