As you may have read before on our website: we wanted to fix the hard water issues in Brussels (and around the world) and were inspired by the fact that chickens can obtain calcium from their food and turn this into calcium carbonate eggshells. We looked around nature to look for more animals made from calcium carbonate found things such as mollusks, sponges, corals and more. Then, we looked at papers that described proteins responsible for this process of calcium carbonate crystallization. Additionally, our supervisors noticed that when trying to image curli protein fibers, crystals formed spontaneously around it, so we also used some loops of those fibers to use as a construct: couldn’t hurt right?

Turns out that the most important parts of these proteins were often peptides rich in glutamates and aspartates. This is probably because these amino acids can bind calcium, causing very high local concentrations that could induce nucleation.

However, as we were working with peptides, we started to realize that we would need a good scaffold. That’s where ENA comes in. ENA stands for ‘endospore appendage’. It is a fibrous protein produced in different Bacillus species. However, it has some incredible benefits:

• It’s easy to produce in Escherichia coli (E.coli). Indeed, under an inducible promotor E.coli can produce multiple grams of protein! We usually measured around 5 grams of protein for a one liter culture.


• It’s super stable; ENA had previously been cooked in SDS to remove any other proteins while ENA remains intact!

Here are a few papers on ENA:

• Pradhan, B., Liedtke, J., Sleutel, M., Lindbäck, T., Zegeye, E. D., O Sullivan, K., Llarena, A. K., Brynildsrud, O., Aspholm, M., & Remaut, H. (2021). Endospore Appendages: a novel pilus superfamily from the endospores of pathogenic Bacilli. The EMBO journal, 40(17), e106887. https://doi.org/10.15252/embj.2020106887


• Zegeye, E. D., Pradhan, B., Llarena, A. K., & Aspholm, M. (2021). Enigmatic Pilus-Like Endospore Appendages of Bacillus cereus Group Species. International journal of molecular sciences, 22(22), 12367. https://doi.org/10.3390/ijms222212367

Also, it looks really cool.

However, by using this scaffold we restricted our design to a certain extent: the peptides needed to have an N and C terminus that were relatively close to each other. However, most of our peptides were disordered so that didn’t pose too much of a problem, and they also couldn’t be too large. Another, unrelated restriction was that the peptides had to be suitable for Gibson assembly (our preferred cloning type).

We then placed our proteins in the following loop down in ENA. First in silico by using alphafold to check everything was in order and to get an idea of what it would look like, then we went to building. Here you can see an example of how that workflow went for one of our 19 constructs.

Here's a useful overview of our building (and a part of the testing) stage. The building process was quite straightforward; we started from vectors with the sequence coding for ENA under an inducible promotor and linearized these vectors by a PCR reaction at the desired site. We ordered our 19 constructs and put them into this linearized plasmid by using Gibson cloning. After that, we did a heat shock transformations to put these constructs in E.coli. We then shuttled our vector with construct on it to a production strain. One IPTG induction later we had cells that had produced our protein. By a lysozyme treatment and cooking the proteins in a 1% SDS solution, we purified our constructs, and we were ready for testing!

We opted for the Eriochrome Black T titration because of the cost, speed and scalability.

We could finally see if our constructs work! For that we did the following experiment. At the start of the summer and before all the experimental work we went to the store and bought a lot of different water types. We did this because on the label of a water bottle you can see quite exact measurements on all the ions that the water contains. We tested all these different water types to test the accuracy of our titration and saw that the titration worked well. Then we started making proteins. Once we had the proteins, we could start having fun. We put all of the proteins in the different water types and saw that both Diactinin (from Sycon ciliatum) (CalcifEna 2) and one of the loops from the curli fibers (CalcifEna 1) reduced the calcium concentration by up to 25%! Here is an overview of the decreases caused in the different commercial water types.

Bingo, we identified two ENA fusions that reduced calcium concentrations in our tested waters, within minutes of addition. So what about the mechanism of this water softening activity?

Basically, the way calcium carbonate comes into water is by the following reaction mechanism:

Based on theoretical grounds, we reasoned two possible modes of action for our water softening ENA fusion. In a first, which we refer to as calcium sequestration (i), calcium ions are stoichiometrically bound to calcium-binding sites in the ENA fusion proteins, resulting in a depletion of free calcium from the solution. In a second mechanism, referred to as calcium mineralization (ii) , calcium carbonate is bound to ENA fusions in a structured fashion, resembling calcite or other calcium carbonate crystal lattices, thus adopting a nucleating activity on the crystallization and deposition of calcium salts otherwise seen as catalysing the orange reaction. These are testable hypotheses that will show a different concentration dependency in the water softening activity of our ENA fusions.

We believe our protein to be following the second way of action, because of the following observations:

• Adding more protein does not further decrease the calcium content


• Our serial dilution also did not have an effect on the free calcium


• Our protein does not produce a decrease in soft water


• Our decrease is strengthened by temperature


• Our TEM images (you can see them in our proof of concept where we also talk about our mineralisation) strongly suggest the formation of nuclei

The way we see our protein work in practice would be working with a temperature increase in combination with our protein to soften water, since increasing temperature increases the amount of CaCO₃ (in solution). But there are other ways to increase the CaCO₃ in solution, and these things can be designed into cells! A major example of this is carbonic anhydrase. Carbonic anhydrase catalyses the first reaction, which we believe could also increase the working of our protein. Another way we could increase the working of our protein is by artificially creating a local environment with an increased pH, which would allow for more H⁺ in solution, also propelling our reactions forward and increasing the working. We would like to implement these ideas in our future designs. We also want to look into extracellular production of our protein and production in Lactococcus lactis, so that we can fully switch to GRAS-bacteria to increase safety of using our constructs.