Engineering Success

Engineering design cycle seen through our project as a whole and for a specific part.

Broad (Overall Project)

Define:

The issue we are addressing is that there are not enough cheap, at-home iron level testing kits that don’t involve drawing blood.

Ask:

Approximately 25% of the world population is iron deficient (StatPearls) with women, children, and people in lower socioeconomic settings being more susceptible (and having less access to good healthcare on the global scale). This is a problem because many people might not have funds or insurance, blood drawing is dangerous for anemic people, and the current devices being hard-to-use makes them less accessible and reliable. There are also many countries where political and human rights issues prevent women, who are more likely to suffer from anemia, from having easy access to a doctor. This issue is compounded by the fact that people in lower socioeconomic classes are often the ones most affected by anemia and are the least likely to test their iron levels as it usually requires insurance or easy access to a free health clinic that tests these diagnostics.

Imagine:

Our goal is to create a cheap, electrical, biosensor that detects one specific iron level marker found in saliva. We believe it is realistic to start off with one marker as we can develop the test to be more accurate and it will increase an individual's ability to understand their iron levels and make informed decisions. A more long-term goal would be to potentially detect more iron level markers depending on the data supporting our ability to do so.

Plan:

Our plan is to:

  • Find this target marker - done - it’s salivary ferritin!
  • Gain understanding of different biosensor designs - in progress
    • There is so much to learn about biosensors! One of our captains (also wet lab lead) is taking a course right now on biosensors and she’s learning so much from it!
  • Find aptamers capable of binding to our marker (ferritin) - done
    • After much literature review and computational docking (will go into detail below), we found two!!
  • Survey the public to get a better idea of this project’s importance - in progress
    • We are trying our best to reach out to more women, people of color, and people from lower socioeconomic backgrounds as their perspective is especially important to making our biosensor accessible.
  • Test the aptamers to see what (binding, detection, selectivity, etc) works best - in progress
    • We are currently doing direct biotin-streptavidin ELISA to test the aptamers against native ferritin samples.
  • Develop biosensor design - in progress
    • We can confirm that our goal to make the biosensor cheaper is to make it arduino based.
    • We will potentially be doing a fluorescent signal for detection as the devices necessary to do this are inexpensive.
  • CAD Modelling the biosensor - not started
  • Building the biosensor and testing the aptamer’s stability within in - not started

Prototype:

Our prototypes include the two aptamers (one peptide, one DNA) we designed through literature review and computational docking (more below on this). Our biosensor prototype is still a work in progress as we are still researching different designs and it is partially dependent on the stability of whichever aptamer we end up selecting after testing.

Test:

  • Before finalizing our aptamer designs and ordering them, we tested their binding affinity to salivary ferritin using computational docking. After confirming that their binding affinities were reasonably high, we are now testing the physical aptamers against native ferritin (liver specifically - has similar composition to salivary ferritin) through ELISA and DNA shift assay. More details on this are provided below and on our experiments and results pages!
  • As our biosensor is still being researched, we are not yet at this prototype’s testing stage. Our tests will mainly consist of usability, how stable the aptamer is within our biosensor, and the accuracy of the data produced.

Improve:

We will continue to improve our overall project’s design as we get more results from our aptamer testing, feedback from our surveying, and later on, results from our biosensor testing. We recognize that a project like this is definitely not finished in the first attempt and within a year. It is a strong cause that our team is passionate about, so we plan to continue it to completion and run through multiple iterations until the final product fits our goals.

Specific (Peptide Aptamer Design and Testing)

Define:

The next step after identifying the iron level marker (ferritin) we wanted to detect was finding an aptamer capable of binding to it. An issue we ran into is that there aren’t many commercially available and there is not much literature/research on small proteins or oligos that can bind to ferritin.

Ask:

We decided on a peptide or DNA aptamer for our project as we believe aptamers are generally quicker and more selective with binding. Generally, antibodies are used to detect markers like ferritin, but they are much slower and less accurate. We’ve included resources that we used to learn more about the benefits of aptamers over antibodies (ie. Bauer et al. and Novaptech) in the references section.

Imagine:

We decided to turn to proteins and DNA sequences that already exist and are capable of binding to human ferritin. Afterall, the human body is full of them. From this, we could identify the specific residues responsible for binding and design a smaller aptamer.

Plan/Sequence of Events:

What’s been done:

    We started off with an extensive literature review to find proteins and DNA in nature that bind to ferritin. We found that H Kininogen and Human Transferrin Receptor 1 are two proteins in mammals that are capable of binding to ferritin. We originally looked into the binding residues in transferrin receptor 1, however there was not much literature or research already done on this chain’s binding affinity with ferritin. When looking into H kininogen, we found that there was a certain type known as high molecular weight kininogen (HKa) that is capable of binding to ferritin and has a lot of research backing this binding ability.

    HKa is an anti-angiogenic protein, so when ferritin is in excess (correlates to high iron levels), it binds to HKa and prevents this anti-angiogenic activity. We found that there were already a few papers looking into finding a more specific ferritin binding region on HKa. Through different experiments testing how different peptide sequences on HKa inhibit HKa’s binding with ferritin when added to the mix, Coffman et al. were able to find that the 22aa sequence HGHGHGKHKNKGKKNGKHNGWK inhibited HKa-ferritin binding the most. In other words, this 22aa sequence has a very high binding affinity to ferritin which was then proven through other experiments detailed in that paper. We believed that this was a good amount of evidence to move onto the computational docking stage with this sequence.

    We then spent June and July doing computational docking to test the peptide against ferritin. We specifically used Autodock Vina to accomplish this. Our first hurdle was that there were no pdb files available of the entire ferritin structure. Ferritin itself is just made up of 24 subunits of two repeating proteins (light chain and heavy chain ferritin). So, our solution to not having a full ferritin pdb was to test the peptide against just light chain ferritin and heavy chain ferritin and use those results as partial confirmation on the overall binding ability of our peptide. We then used Chemdraw and Pymol to draw our peptide aptamer, clean it (remove water, polarize, etc), and convert it to a pdb file. After multiple experiments of the peptide in different conformations with the light and heavy chains, we got some of the data shown below.

    A goal for the future if we use computational docking again is to find if there is a way to use Autodock on our HPC. We are very lucky at NYU to have an HPC. We want to utilize it for tests like this as each Autodock experiment usually took a day to run and that slowed our progress down a lot. We found that the binding affinities we got from the peptide-ferritin tests were much lower than the ones found in the research papers. We believe that the most likely cause for this is that we were testing the peptide against one light chain or heavy chain ferritin instead of multiple chains.

    Our plan was to use AlphaFold or a similar software to create a multimer with multiple light and heavy chains. However, we were having issues with accessing the programming necessary to do this and it was already reaching the end of July. We decided to move forward with ordering the peptide from Genscript based on the computational docking data we already had and our PI agreed with this decision.Below, the first image is the computational docking results from one of our first Autodock tests between the peptide and heavy chain ferritin. After noticing that the magnitude of the binding affinity was lower than expected, we energy minimized the peptide structure using another software and tested it again with heavy chain ferritin (results in second image below). Unfortunately, the images we collected of the tests run with light chain ferritin were not properly saved. The affinity (kcal/mol) of the mode 1 peptide conformation with light chain ferritin was -4.8 (so around the same as heavy chain). We will be running more autodock tests along with the ELISA to recollect this data.

    no_em_peptide_heavy_chain

    Autodock Binding Affinity Results of Non-energy Minimized Peptide Aptamer with Heavy Chain Ferritin

    em_peptide_heavy_chain

    Autodock Binding Affinity Results of Energy Minimized Peptide Aptamer with Heavy Chain Ferritin

To be done:

    We are now testing our peptide aptamer physically in the lab. We ordered the peptide with one end biotinylated from Genscript. We are doing direct biotin-streptavidin ELISA to test the peptide’s binding ability with ferritin.

Prototype:

After our computational docking tests, we determined that the peptide aptamer is viable enough to go onto the next stage - physical testing. We ordered the peptide, along with a DNA aptamer (more details on contribution page), from genscript to test its binding affinity with ferritin in the lab. We ordered the peptide to have the N-terminal biotinylated so that we could use it in direct biotin-streptavidin ELISA. The peptide came lyophilized and was found to be soluble in PBS, DPBS, tris-HCl, ultrapure water, and saline. Below is an image of an energy minimized conformation of the peptide on Chem3D.

em_peptide_aptamer_3d

Energy Minimized 3D Structure of Peptide Aptamer in Chem3D

Test:

We are now doing direct biotin-streptavidin ELISA to test the binding affinity of the peptide aptamer with native ferritin. Our standards are recombinant ferritin from a Thermo Fisher ELISA kit. We are testing the peptide aptamer’s binding ability against a control antibody known to bind to ferritin that came with the kit. After getting a standard curve from the recombinant ferritin standards given in the first experiment, we will start using the native ferritin (purified ferritin from human liver) that we ordered as both standards and samples. More details on our protocol and any results can be found in the experiments and results pages. Here is a short summary of the plates we’ve done so far to give more context for the improvements we’re making.

First experiment:

  • Plate 1:
    • recombinant ferritin standards with antibodies (both are from kit)
    • Native ferritin samples at decreasing concentrations with three different peptide dilutions – we’re finding the ideal peptide dilution while testing its binding ability
    • Blanks along the bottom of the plate
  • Plate 2:
    • native ferritin standards with antibodies (antibodies from kit)
    • Native ferritin samples at decreasing concentrations with three different peptide dilutions (same concentrations as plate 1) - we’re finding the ideal peptide dilution while testing its binding ability
    • Blanks along the bottom of the plate

Improve:

Problems encountered:

  • The Thermo Fisher ferritin ELISA kit we ordered didn’t come with a specific blocking buffer - we assumed that the ferritin sample diluent contained BSA blocking buffer based on what minimal information we could find on the reagents. It must have been enough, however, as there was a lot of background noise in the first two plates we ran.
  • The native ferritin and antibody did not bind in the standard rows of the second plate even though the kit said its antibody was capable of binding to purified native ferritin. Because of this, we won’t be able to use the native ferritin and antibody as a positive control.

Solutions:

  • We have ordered Casein blocking buffer to properly block the plate moving forward.
  • Potential positive control: Order more recombinant DNA and antibody as those two definitely bind together. This option is expensive, however.
  • Potential negative control: Having blanks in the plate (which we have been doing until now)
    • This isn’t as helpful as the positive control but it can be used to mitigate background noise.

References

Bauer, M. J., Strom, M., Hammond, D. S., & Shigdar, S. (2019). Anything you can do, I can do better: Can aptamers replace antibodies in clinical diagnostic applications? Molecules, 24(23), 4377. https://doi.org/10.3390/molecules24234377

Coffman, L. G., Parsonage, D., D'Agostino, R., Jr, Torti, F. M., & Torti, S. V. (2009). Regulatory effects of ferritin on angiogenesis. Proceedings of the National Academy of Sciences of the United States of America, 106(2), 570–575. https://doi.org/10.1073/pnas.0812010106

Dhar, P., Samarasinghe, R. M., & Shigdar, S. (2020). Antibodies, nanobodies, or Aptamers—Which is best for deciphering the proteomes of Non-Model species? International Journal of Molecular Sciences, 21(7), 2485. https://doi.org/10.3390/ijms21072485

Di Rienzo, L., Milanetti, E., Testi, C., Montemiglio, L. C., Baiocco, P., Boffi, A., & Ruocco, G. (2020). A novel strategy for molecular interfaces optimization: The case of Ferritin-Transferrin receptor interaction. Computational and Structural Biotechnology Journal, 18, 2678–2686. https://doi.org/10.1016/j.csbj.2020.09.020

Eberhardt, J., Santos-Martins, D., Tillack, A.F., Forli, S. (2021). AutoDock Vina 1.2.0: New Docking Methods, Expanded Force Field, and Python Bindings. Journal of Chemical Information and Modeling.

Montemiglio, L. C., Testi, C., Ceci, P., Falvo, E., Pitea, M., Savino, C., Arcovito, A., Peruzzi, G., Baiocco, P., Mancia, F., Boffi, A., Georges, A. D., & Vallone, B. (2019). Cryo-EM structure of the human ferritin–transferrin receptor 1 complex. Nature Communications, 10(1). https://doi.org/10.1038/s41467-019-09098-w

NOVAPTECH. (n.d.). Aptamers vs antibodies I Key advantages of aptamers. https://novaptech.com/aptamers-vs-antibodies-advantages

Ponczek M. B. (2021). High Molecular Weight Kininogen: A Review of the Structural Literature. International journal of molecular sciences, 22(24), 13370. https://doi.org/10.3390/ijms222413370

The PyMOL Molecular Graphics System, Version 2.0 Schrödinger, LLC.

Tesfay, L., Huhn, A. J., Hatcher, H., Torti, F. M., & Torti, S. V. (2012). Ferritin Blocks Inhibitory Effects of Two-Chain High Molecular Weight Kininogen (HKa) on Adhesion and Survival Signaling in Endothelial Cells. PLOS ONE, 7(7), e40030. https://doi.org/10.1371/journal.pone.0040030

Torti, S. V., & Torti, F. M. (1998). Human H-kininogen is a ferritin-binding protein. The Journal of biological chemistry, 273(22), 13630–13635. https://doi.org/10.1074/jbc.273.22.13630

Warner MJ, Kamran MT. Iron Deficiency Anemia. [Updated 2023 Aug 7]. In: StatPearls [Internet]. Treasure Island (FL): StatPearls Publishing; 2023 Jan-. Available from: https://www.ncbi.nlm.nih.gov/books/NBK448065/