Modeling

In our project, we are using aptamers to detect tau proteins in the tears. Aptamers are single-stranded nucleotide sequences that are artificially synthesized by Systematic Evolution of Ligands by Exponential Enrichment (SELEX).

We designed two aptamers based on an article by I-Ting Tang et al [1]. A table was provided in this article, and this article assisted us in determining which aptamers would be the most optimal for our experiment. We found that IT2e and IT2c would be ideal; one consisting of 30 base pairs (bp) and the other of 35 bp, respectively.

Aptamer:

After the identification of the aptamers, we proceeded to look at the LAMP method which requires a padlock probe. The padlock probe we used has a recognition site complementary to the p53 gene which binds to the aptamer with linker (ATAATATAATA). The linker was used to create some distance between the aptamer and padlock probe to avoid any binding interference that it may cause between tau protein and aptamer.

We further proceeded to confirm that these aptamers would be ideal for this project by modeling them. There were several steps/orders that were taken to obtain the results shown in the following diagram below.

First, secondary structures of the aptamers were generated by using the mFold web server [2]. The aptamer sequence was put straight into the programme. The programme would then generate an aptamer structure from that sequence and it was downloaded as a vienna file, containing the secondary structure and its sequence. Afterwards, the tertiary structures were generated by using the 3dRNA/DNA web server [3]. The Vienna file was pasted into 3dRNA/DNA. The result was then downloaded as a PDB file. Once the tertiary structure was obtained, we began docking in HADDOCK [4].

Docking

The entire protein sequence will not be used when docking or modeling, as the tears only contain fragments of tau and not the entire protein. This begins to raise some new questions, what fragments of tau should be modeled and why should these fragments be used?

our models, the following two tau fragments were used:

1. K V A V V R T P P K S P S

2. N I K H V P G G G S V Q I

From now on, tau fragment 1 will be referred to as KVAV and tau fragment 2 will be referred to as PGGG.

KVAV was used because it is one of the sequences that has been proposed as one of the microtubule binding motifs in the tau protein, and therefore has binding capabilities. The PGGG sequence is a sequence derived from 3 other similar sequences, and frequently appears in the tau protein. So by using PGGG as a binding site for the aptamer, we are covering more of the tau protein sequence in one docking run. Both of these tau sequences were found in the following paper by Cabrales et al [5].

As for the actual docking of the aptamer and the tau fragments, there were different combinations considered and were all run through the HADDOCK server. The following docking runs were made:

• Full aptamer + KVAV tau fragment
• Full aptamer + PGGG tau fragment
• Full aptamer with linker region + KVAV tau fragment
• Full aptamer with linker region + PGGG tau fragment
• Only loop-region + KVAV tau fragment
• Only loop-region + PGGG tau fragment

HADDOCK is a server to characterize biomolecule interactions that can compute/model molecules on an atomic level. The PDB file was uploaded to the server and binding sites were specified before running. The binding sites of the tau proteins were obtained from the article by I-Ting Tang et al [1], that was previously mentioned. Whereas, the binding site of the aptamer was chosen due to its presence of a loop that remained constant in all structures that were produced. Furthermore, when imputing the binding site for KVAV, the entire sequence would be selected. In contrast to PGGG fragment, where only PGGG is imputed as the binding site because that is the only part that constantly reappears in the tau protein's entirety.

HADDOCK produces clusters that provide us detailed information such as HADDOCK score, cluster size, RMSD, Van Der Waals energy, electrostatic energy, desolvation energy, restraints violation energy, buried surface area and z-score. Based on these results, our main focus was on buried surface area. A higher value of the buried surface area indicates that there is a stronger binding between the aptamer and tau protein since more of the molecules are interacting with each other [6]. However, HADDOCK has a limitation as there is a requirement to input the binding site before running. This would have a slight interference with our results as it does not provide us the energy level of a ‘natural’ binding energy between tau protein and aptamer. Despite that, the scores obtained by HADDOCK provided us with the information of the binding energy of those binding sites.

Interaction analysis

Based on the docking results, the interactions of the aptamer and the tau fragment were determined by using the PLIP web server [7], which allowed us to identify residues and type of interactions between the aptamer and the tau fragment. The interaction report from PLIP showed that both aptamer with and without linker have the same amount of hydrogen bonds after they bind to the tau fragment. Therefore, the linker has no effect on stabilizing the aptamer-protein interaction.

Furthermore, we have also compared the buried surface area value and the binding energy of the docking results from Haddock. The buried surface area is the estimate of the size of the interface between two macromolecules (in this case it is protein-DNA complex). According to the description provided by HADDOCK regarding the values used to calculate the buried surface area is done by calculating the sum difference of the accessible surface area for each molecule as well as the whole complex. This was done so by “1.4A water probe and has an accuracy of 0.075A” [8]. Based on the figure below, the value of the interaction of the aptamer with linker is higher and stronger than the interaction of the aptamer without linker.

Discussion

From the modeling experiments, we drew some conclusions and applied them towards our wet lab. This was done so by modeling the 2D-structure of the aptamer, we could tell that there was one structure in the aptamer that consistently appeared throughout all of the aptamers. This was the loop, which always consisted of the same nucleotides.

the aptamer is assumed to have some sort of specificity, then the structure would need to have a consistent recognition site for tau. Since the loop always appears in the structure, we can therefore assume that the loop-region is a recognition site of the aptamer. Not only does the loop appear in the 2D-structure, but it also retains the loop-structure when docking with the tau fragments in 3D with PyMol.

However, our aptamer would also contain a linker region which could lead to a possible interruption of the binding between the tau protein and aptamer. So, modeling was used to see how the linker can affect the binding. From the modeling experiments, we can deduce that the linker does not affect the binding of the aptamer to tau and should produce the same results as the binding of aptamer without linker. Based on these modeling results, we have ensured that the aptamer will work as intended for the aptamer with the linker in the wet lab.

Conclusion

With the help of modeling, we have researched whether or not the aptamer can bind to tau specifically. We can say that the aptamer seems to have a part that can bind to tau specifically, this being the loop. The linker also does not seem to have an effect on the binding, which was one of the main concerns for the wet lab. After acquiring the results from modeling, we were able to provide reassurance to the laboratory group that the aptamer does indeed bind to the tau protein and the linker does not interfere with the binding of these two molecules.