Team:TUDelft/Partnership

Team

Team Members Collaborations Attributions Sponsors

Project

Project Description Design Engineering Success Results Hardware Contribution Sequencing analysis Partnership Sustainable Development

Lab

Protocols Notebook Parts

Model

Model

Human Practices

Human Practices Integrated Human Practices Education & Public Engagement Proposed Implementation Safety Entrepreneurship

Achievements

Judging Form Medal Criteria Awards and Nominations

Partnership

During our iGEM project, we partnered up with the Vilnius-Lithuania iGEM 2021 team. The aim of our partnership was to obtain a computationally predicted aptamer sequence that binds to retinol binding protein 4 (RBP4), the carrier protein of vitamin A. By doing so, we could expand the modularity of our diagnostic kit to include the detection of other protein-bound, non-water-soluble vitamins, in our diagnostic test. Simultaneously, we would experimentally validate the aptamer predicting software of the Vilnius-Lithuania team. We evaluated the binding of the predicted aptamer sequences to RBP4 by conducting an electrophoretic mobility shift assay (EMSA). Here, we report the development and the results of our partnership.

Introduction

The aim of AptaVita is to develop an accessible, quantitative, and modular rapid diagnostic test allowing for the detection of vitamin deficiencies. We evolved aptazymes to bind the water-soluble vitamins B1, B2, B6, B9, and B12 (for more details visit our Description page). Yet, we envision AptaVita as a tool to detect the complete spectrum of vitamins deficiencies to effectively help tackling hidden hunger. This includes the detection of protein-bound non-water-soluble vitamins such as vitamin A, D, E, and K, for which the development of new biosensors is required.

The iGEM Vilnius-Lithuania 2021 team created an aptamer prediction software based on a surface interaction model. This software generates sequences with potential affinity for a desired target. Such a tool would provide an advantageous starting point for the development of vitamin biosensors such as AptaVita.

Our partnership with Vilnius-Lithuania provided us the opportunity to explore the modular aspect of our test, by working together in the development of aptamer sequences with affinity to the vitamin A carrier, RBP4. In return, our experimental results could validate their predictive software and generate experimental data on affinity for its' future improvement.

June and July: The beginning of our partnership

Project introduction

During our project, we reached out to the iGEM community to find other teams working with aptamers. Fortunately, in June, we encountered the iGEM Vilnius-Lithuania 2021 team. This team has been working on an aptamer-based rapid diagnostic test for the detection of pyruvate phosphate dikinase (PPDK), as a biomarker for Entamoeba histolytica infection. Out of interest in each others’ projects, we decided to arrange a first meeting on the 2^nd of July (Fig. 1).

Fig. 1 First meeting with the iGEM Vilnius-Lithuania 2021 team on 02/07/2021. During this meeting, we presented our projects to each other.

Exploring the possibilities to work together

After our first meeting, we decided to explore the possibilities of collaborating. The Vilnius-Lithuania team used the Systematic Evolution of Ligands by Exponential Enrichment (SELEX) protocol for the evolution of aptamers, whereas we used the De Novo Rapid In Vitro Evolution of RNA biosensors (DRIVER) protocol [1, 2]. We suggested to create a review document covering the two different experimental methods. By doing so, we could provide a clear presentation of the advantages and disadvantages of both protocols in collaborative work as a guide for future iGEM teams. Unfortunately, it became evident that such a collaboration did not necessarily contribute to our own projects.

In early stages of our work, our team considered creating an aptamer predictive software. We therefore discussed working together on the improvement and adaptation of the Making Aptamers Without SELEX (MAWS) software from the iGEM Heidelberg 2015 team. However, our team was lacking the required resources to realize such a collaboration.

As we continued brainstorming, an opportunity for a partnership was created. Vilnius-Lithuania developed a software that predicts single stranded DNA (ssDNA) aptamer sequences based on a surface interaction model. During this stage of their project, Vilnius-Lithuania had gone through one engineering cycle. But, for them to continue with their engineering process they required experimental data that validated their software. Here, we saw an opportunity for Vilnius-Lithuania to generate a sequence for a vitamin deficiency biomarker that could allow us to expand the application of AptaVita to detect other vitamin deficiencies.

August: The target of our partnership

Aptamer prediction for B vitamins

Our team had the idea of comparing computationally predicted aptamer sequences binding to B vitamins with those obtained through our in vitro evolution process. Vilnius-Lithuania considered vitamins as viable targets for their software and therefore agreed on predicting aptamers for three ligands. These were vitamin B9, due to its relevance in public health [3], B12, in hope that its larger surface area was more likely to generate a functional aptamer [4], and B1, as an extra alternative. Unfortunately, Vilnius-Lithuania concluded that the surface area of vitamins is too small for their software to successfully predict aptamer sequences. This represented an unexpected limitation in their software. After efforts to correct this limitation, they concluded that, in the present stage, their software is limited to the prediction of aptamers for proteins, such as their initial target PPDK. Consequently, they proposed to generate an aptamer sequence with affinity for a protein that could be of our interest. For us, this was an opportunity to expand the scope of AptaVita towards the detection of protein-bound, non-water-soluble vitamins.

Aptamer prediction for retinol binding protein 4

We developed aptazymes to bind to B vitamins: a group of small water-soluble vitamins that are freely available in blood. In contrast, other relevant vitamins, such as vitamin A, D, E, and K, have a hydrophobic nature. These vitamins are transported throughout the body by binding to lipoproteins or carrier proteins [5]. Therefore, these vitamins are not freely in circulation but rather shielded by their carrier, making them inaccessible for direct binding to an aptazyme. A strategy to detect these vitamins is then the detection of the proteins they are bound to.

An example of a carrier protein as a potential biomarker for vitamin deficiency is RBP4. RBP4 is a 21 kDa protein that binds retinol, a species of the hydrophobic vitamin A class. Upon binding retinol, RBP4 shields the hydrophobic molecule and enables its transportation in the blood [6]. RBP4 is prone to renal filtration when it is not bound to retinol. Therefore, only 15% of the unbound RBP4 (apo-RBP4) remains in circulation, whilst 85% is in its bound conformation (holo-RBP4). Thus, the blood concentration of apo-RBP4 is correlated to the concentration of retinol, making RBP4 a potential biomarker for vitamin A deficiency [7, 8].

Diagnosing vitamin A deficiency

Vitamin A is involved in fetal growth, the immune system, and eye development. Deficiency of vitamin A may cause a series of symptoms, including (night) blindness, infertility, delayed growth, and poor wound healing, and occurs most common in children and women of reproductive age in underdeveloped countries. Current detection of vitamin A deficiencies is conducted by enzyme immunoassays, requiring expensive analytical instruments [3, 9, 10].

We asked Vilnius-Lithuania to generate an aptamer with affinity for RBP4. They confirmed that this protein could be used to generate aptamer sequences and validate their software. Our initial expectation was that the aptamer prediction could directly include the constant regions of our aptazyme. Nonetheless, we were informed that, at the current stage of the predictive software, including such a constraint was not possible. We then decided to continue with the prediction of just the aptamer and subsequently chimerize it into our aptazyme.

Together, we established the following cooperative strategy:

1. Adapting software prediction from ssDNA to RNA aptamers as we worked with RNA aptamers
2. Generating random control sequences for our validation experiments
3. Calculating HDOCK & MFold scores to confirm the stability of all sequences
4. Ordering sequences and proteins
5. Establishing experimental binding conditions (buffer composition, oligo concentrations, incubations parameters, etc.)
6. Performing an EMSA to analyze the binding affinity of the aptamer to RBP4

Steps 1-3 were carried out by Vilnius-Lithuania and steps 4-6 were performed by our team.

Results from the aptamer sequence prediction software

Running the software for our target protein without any constraints resulted in the sequence RBP4_21_DNA (Tab. 1), being 21 nucleotides long. To provide sequences directly related and applicable to our system, our colleagues expanded the reach of their software by modifying it to work with RNA sequences too. Moreover, they constrained the predictions to exactly 30 nucleotides sequences so that we could include the aptamer sequences as the target binding loop in our RNA aptazyme. This adjustment was applied to both ssDNA and RNA prediction, resulting in sequence RBP4_30_DNA and RBP4_30_RNA (Tab. 1). Subsequently, the Vilnius team generated a random sequence for each species of nucleotides to be used as controls in our experiments (Random_21_DNA, Random_30_DNA, Random_30_RNA_1, and Random_30_RNA_2).

Tab. 1 Sequences and corresponding MFold and HDOCK scores generated by the Vilnius-Lithuania 2021 iGEM teams’ aptamer sequence prediction software. Shown are the sequences, their MFold and HDOCK scores, and their names.

Name	Sequence (5’ → 3’)	MFold score	HDOCK score
RBP4_21_DNA	GTTGATTGTTATGTTTAGTGA	1.25	-317.59
Random_21_DNA	GGCAGGTCAATTCGCACTGTG	-0.40	-320.05
RBP4_30_DNA	GTTGATTGTTATGTTTAGTGACGGGTTCCC	0.78	-363.45
Random_30_DNA	AGGGTCACATGGGCGTTTGGCACTACCGAC	-1.22	-356.26
RBP4_30_RNA	GUCCCCCGCCCGUGUCCCGCUAGCCCCGCG	-1.6	-376.82
Random_30_RNA_1	CUGUUUUCGAAAUUACCCUUUAAGCGCGGG	-2.20	-305.81
Random_30_RNA_2	AGCAUUCUAUCACGUCGGCGACCACUAGUG	-0.60	-339.68

N.B.: From here on, we will refer to these highlighted sequences as RBP4_ssDNA, rand_ssDNA, RBP4_RNA, and rand_RNA, respectively.

September: Determination of experimental conditions & computational analysis

Experimental conditions and software limitations

For the aptamers to work at our desired AptaVita conditions (Design page), their tertiary structure should be compatible with a cell-free system functioning at 37 °C and under physiological conditions such as pH, ionic strength, and buffer composition. To ensure this, our experimental design was supported by literature (more details on the conducted experimental work can be consulted in our notebook and protocol). We formulated our binding buffer using phosphate buffered saline, aiming to mimic physiological ion concentrations, and supplementing it with salts as recommended [11, 12, 13].

We consulted Vilnius-Lithuania on the conditions used for the predictions. We learned that their software is not designed to consider these physical parameters. They did inform us that these physical parameters were used for the calculation of the MFold and HDOCK scoring. We identified this as an important consideration and decided to investigate the following possible limitations:

(i) The predicted aptamers were generated and the EMSAs were going to be performed using the apo-RBP4 form. ​​Since we are interested in the holo-RBP4 form of the protein (vitamin bound), can an apo-form binding aptamer also bind to the holo-form?

(ii) The prediction software uses x-ray crystallography structures of the target protein to generate aptamers sequences. Crystallographic structures reveal only a static picture of the protein and can impose non-realistic conditions that are absent in biological environments. We therefore disputed whether the protein allows for a proper aptamer binding under more realistic biological conditions.

Molecular dynamics of retinol binding protein 4

Before experimentally testing the aptamer sequences, we performed a Nanoscale Molecular Dynamics (NAMD) computational analysis of our target protein to provide answers on the above mentioned considerations. Molecular dynamics (MD) is a valuable and sophisticated computational tool to probe the dynamic evolution of molecular systems, providing a time-dependent picture that emerges from interatomic interactions and accounts for the influence of external effects such as the presence of ligands.

Molecular dynamic simulations

MD simulations were performed with NAMD [14] over simulation times of 50 ns using the CHARMM force field. Apo- and holo-forms of the human RBP4 protein (Protein Data Bank (PDB) ID 5NU7) were prepared (and the ligand, retinol, parameterized) for their MD study using the PDB Reader service of CHARMM-GUI (http://www.charmm-gui.org/) [15]. Periodic solvation boxes were constructed with 14 Å spacing and water molecules according to the TIP3P model [16]. Sodium and chloride ions were added to counter the total charges of the protein systems setting a 0.150 M salt concentration, resembling the ionic composition of the phosphate buffer saline used in our experimental binding assay. The particle-mesh Ewald summation method [17] was used for long-range electrostatics and a 10 Å cutoff was set for short-range non-bonded interactions. Initial geometries were first minimised at 3,000 conjugate-gradient steps, water was then equilibrated at 298 K and 1 atm for 100 ps at 2 fs time steps, and production runs were then performed for 50 ns at 2 fs time steps (25 million steps per calculation) in the NPT ensemble at 1 atm and 298 K. Langevin dynamics for T control and the Nosé-Hoover Langevin piston method for P control were employed. NAMD output was stored every 12,500 steps, giving trajectories composed of 2,000 frames that were processed and analysed with VMD 1.9 [18]. Root mean square distances (RMSD) computed with Cα-atoms were obtained for structural superpositions with the CEALIGN method [19] implemented in PyMOL 1.4 (pymol.org).

The 50-ns MD simulations showed very low mobility of both the apo- and holo-forms of RBP4 (Fig. 2). To confirm this, the RMSD, used to measure the difference between the structural conformation of the starting point of the simulation and all succeeding frames, was computed (Fig. 3A). The mobility of the protein was determined by the deviations produced during the course of the simulation. The average RMSD computed for the apo-form was 1.301 ± 0.195 Å (mean ± standard deviation), and 1.106 ± 0.195 Å for the holo-form. These low RMSD values (< 2 Å) confirm the low mobility of the apo- and holo-RBP4 forms as observed in Fig. 2. This suggests that having generated the aptamer sequence using the crystallographic RBP4 structure was not a critical limitation for this specific protein. Additionally, the final geometries of the apo- and holo-forms obtained after completion of MD simulations were superimposed (Fig. 3B) with a computed RMSD of 1.56 Å (< 2 Å). It was thus conjecturable that the structure of the holo-RBP4 presents little changes with respect to the apo-RPB4 form, suggesting that an aptamer predicted, or experimentally confirmed to bind to the apo-form, was likely to bind the holo-form too.

Fig. 2 50-ns all-atom MD simulations. Apo-RBP4 (left, green), holo-RBP4 (right, orange), and retinol (right, blue) are shown. Because Vilnius-Lithuania's prediction software uses the molecular interactions at the surface of the protein to predict a specific aptamer sequence, a simplified surface representation is rendered together with the ribbon representation. MD simulations show an overall low mobility of the apo- and holo-forms of the RBP4 protein.

Fig. 3 RMSD values and superimposition of the apo- and holo-RBP4 form. (A) RMSD values of non-hydrogen atoms (backbone and side chain) of the apo- and holo-forms of RBP4 in 50 ns MD simulations. (B) Structural superposition of apo- and holo-forms of RBP4 after 50 ns MD simulations.

After defining the experimental approach to be undertaken and exploring the dynamic characteristics of RBP4 as a target protein, the potential ssDNA and RNA aptamer sequences were ordered, along with a random sequence of each nucleic acid. The decision was based on the best MFold and HDOCK scores (Tab. 1), resulting in RBP4_ssDNA, rand_ssDNA, RBP4_RNA, and rand_RNA.

October: Electrophoretic mobility shift assay

Following the computational analysis of RBP4 and the planning of experimental work, the designed protocols were conducted in the laboratory.

From the native polyacrylamide gel for ssDNA we found that there is no clear binding between RBP4 and the predicted sequences, under the experimental conditions in the EMSA. Fig. 4 shows the resulting native polyacrylamide gel for ssDNA sequences. Lanes 2 and 5 show the expected migration patterns for the free RBP4_ssDNA and rand_ssDNA oligos, respectively, in the absence of target protein. It can be seen that all samples containing RBP4_ssDNA migrate the same distance as the free oligos, at the mark of 40 nucleotides, while no evidence of a motility shift is visible. This suggests a lack of oligo:protein interaction. The same results are visible for rand_ssDNA, sitting at a height of 25 nucleotides. The difference in the migration patterns of RBP4_ssDNA and rand_ssDNA, both 30 nucleotides long, is attributed to their different sequence-dependent tertiary structures in the non-denaturing gel.

Fig. 4 Native-polyacrylamide gel of the EMSA conducted with ssDNA and RBP4 at different oligo:protein molar ratios. (Ladder) ssDNA oligo ladder, (NC) Negative control: no oligos nor target protein, (tracking oligos) show migration of free oligos through gel, (RPB4_ssDNA) potential RBP4 single stranded DNA aptamers, (rand_ssDNA) random single stranded DNA oligos. Lane 3 was skipped due to overflow of sample from lane 2. Lane 6 was used as a spacer. All ratios refer to RBP4_ssDNA oligos:protein ratios except for 1rand:1, where it refers to rand_ssDNA. Ratios based on an oligo sample of 50 ng.

Similar results were observed in the native polyacrylamide gel for the RNA sequences (Fig. 5). The potential RNA aptamer sequences did not show affinity towards RBP4, under the experimental conditions in the EMSA. All samples migrated below the 50 nucleotides marker, with no visible shift for any of the ratios. Moreover, rand_RNA oligos were barely perceptible. This could be due to an improper handling of the oligos before addition to the reaction mix.

Fig. 5 Native-polyacrylamide gel of the EMSA conducted with RNA and RBP4 at different oligos:protein molar ratios. (Ladder) Low Range ssRNA ladder, (NC) Negative control: no oligos and no protein, (tracking oligos) show migration of free oligos through gel, (RPB4_RNA) potential RBP4 RNA aptamers, (rand_RNA) random RNA oligos. All ratios refer to RBP4 RNA oligos:protein ratios except for 1rand:1, where it refers to rand_RNA. Ratios based on an oligo sample of 50 ng.

To confirm the presence of the target protein in the gel, we decided to perform a protein stain of the gels using SimplyBlue™ SafeStain. In each case, the presence of RBP4 or bovine serum albumin (BSA) proteins was confirmed in the upper area of the gels.

Fig. 6 Protein staining of the ssDNA EMSA gel. The top red arrow shows migration of proteins in gel. The red and yellow lines indicate the distance reached by rand_ssDNA and RBP4_ssDNA oligo sequences, respectively, after the electrophoretic procedure. White stain on lanes 5, 6, 7, 8, and 9 is due to an imperfection on the lens of the visualization equipment.

Fig. 6 shows the presence of proteins in the upper area of the gel, particularly in lanes 5, 7, 8, 9, and 10. These lanes correspond to the ones containing RBP4. A change in the intensity can also be seen, which matches the protein ratios loaded in each of the lanes.

Fig. 7 Protein staining of the RNA EMSA gel. The top red arrow shows migration of proteins in gel. The red line indicates the distance reached by the oligo sequences after the electrophoretic procedure. White stain on lanes 5, 6, 7, and 8 is due to an imperfection on the lens of the visualization equipment.

The same results occurred for the RNA gel after staining. Fig. 7 reveals the presence of proteins inside of the gel for all loaded wells except for the ladder. Particularly, a higher intensity is perceptible in lanes 4, 5, 6, 7, and 8. These lanes correspond to the samples containing RBP4 and their change of intensity matches the protein ratios used. The evidence of proteins in the negative control (lane 1) and tracking lanes (lane 2 and 3) is due to the presence of BSA contained in the binding buffer. Given these results, we concluded that, under the used conditions, the aptamer sequences predicted by the Vilnius-Lithuania team are not capable of binding to RBP4 in the EMSA. However, it is important to note that EMSA protocols are highly protein and sequence specific and it is suggested to explore different conditions to find optimum binding parameters.

Our work focused on physiologically compatible conditions, using a phosphate buffered saline based binding buffer, to mimic blood’s ionic composition, and an incubation temperature of 37 °C. This was done in an effort to ensure that a native protein conformation was maintained as simulated for the predicted sequences. Different temperatures and ionic compositions should be explored to further confirm that there exist no oligos:protein interactions. It is also possible that the affinity between oligos and protein is too low to generate visible bands. Alternatively, the Surface Plasmon Resonance-based assay could be used to detect lower affinity interactions [20].

After discussion with Vilnius-Lithuania, these results raised the question whether, even within proteins, there is a limitation regarding the molecular size for which the predictive software can effectively generate aptamer sequences for, as the size is related to the surface area (the fundamental basis with which the software predicts the outcome).

Conclusion

The experimental validation of the aptamer design software from the Vilnius-Lithuania 2021 iGEM team suggest that the predicted aptamer sequences do not appear to have affinity for RBP4. However, throughout the course of this partnership, both teams gained valuable insights to be considered for future research. We found that up until now, the prediction software was unable to generate aptamer sequences for small molecules and it seemed to be constrained to proteins. Nevertheless, taking this limitation into account, the Vilnius-Lithuania team applied the software to RBP4, and generated potential sequences to further extend the modularity of our application.

Moreover, we were able to gain additional insights on the behaviour of the target protein under a more realistic scenario by molecular dynamic analysis. Through this, we confirmed that, for RBP4, the use of crystallographic structures as an input for the predictive software is representative enough to be compatible with the real conditions of our application. This analysis provides further considerations that could be taken into account to evaluate whether predicted aptamer-protein interactions represent a realistic scenario.

In conclusion, even though our collaborative effort to find novel aptamers for protein-bound vitamins did not retrieve a potential AptaVita candidate, we believe that bioinformatic tools, such as Vilnius-Lithuania’s predictive software, can accelerate the development and evolution of biosensors. We recommend the use, when possible, of molecular dynamic simulations to provide a deeper understanding of the target protein’s behaviour and interactions. As a contribution for future work in the field, we established an EMSA methodology for the evaluation of potential aptamers, designed for proteins under physiological conditions. Yet, we recommend to future users of this technique to explore the use of different binding conditions in order to find the optimal conditions for their application.

References

1. Tuerk, C., & Gold, L. (1990). Systematic evolution of ligands by exponential enrichment: RNA ligands to bacteriophage T4 DNA polymerase. Science, 249(4968), 505-510. https://doi.org/10.1126/science.2200121
2. Townshend, B., Xiang, J.S., Manzanarez, G., Hayden, E.J., & Smolke, C.D. (2021). A multiplexed, automated evolution pipeline enables scalable discovery and characterization of biosensors. Nature Communications, 12(1). https://doi.org/10.1038/s41467-021-21716-0
3. Centers for Disease Control and Prevention, World Health Organization, Nutrition International, & UNICEF. (2020). Micronutrient Survey Manual. Geneva: World Health Organization.
4. Pubchem (2021). Vitamin-B12. Retrieved from https://pubchem.ncbi.nlm.nih.gov/compound/Vitamin-B12 on 20/10/2021.
5. Kono, N., & Arai, H. (2014). Intracellular transport of fat‐soluble vitamins A and E. Traffic, 16(1), 19-34 https://doi.org/10.1111/tra.12231
6. Steinhoff, J.S., Lass, A. & Schupp, M. (2021). Biological functions of RBP4 and its relevance for human diseases. Frontiers in Physiology, 11, 294. https://doi.org/10.3389/fphys.2021.659977
7. Frey, S.K., Spranger, J., Henze, A., Pfeiffer, A.F.H., Schweigert, F.J., & Raila, J. (2009). Factors that influence retinol-binding protein 4–transthyretin interaction are not altered in overweight subjects and overweight subjects with type 2 diabetes mellitus. Metabolism, 58(10), 1386–1392. https://doi.org/10.1016/j.metabol.2009.05.003
8. Tanumihardjo, S.A., Russell, R.M., Stephensen, C.B., Gannon, B.M., Craft, N. E., Haskell, M.J., Lietz, G., Schulze, K., & Raiten, D.J. (2016). Biomarkers of Nutrition for Development (BOND)—Vitamin A Review. The Journal of Nutrition, 146(9), 1816S–1848S. https://doi.org/10.3945/jn.115.229708
9. Oregon State University (n.d.). Vitamin A. Recovered from https://lpi.oregonstate.edu/mic/vitamins/vitamin-A#introduction on 19/10/2021.
10. World Health Organization (n.d.). Vitamin A deficiency. Recovered from https://www.who.int/data/nutrition/nlis/info/vitamin-a-deficiency on 19/10/2021.
11. Rio, D.C. (2014). Electrophoretic mobility shift assays for RNA-protein complexes. Cold Spring Harbor Protocols, 4, pdb.prot080721–pdb.prot080721. http://doi.org/10.1101/pdb.prot080721
12. Ream, J.A., Lewis, L.K., & Lewis, K.A. (2016). Rapid agarose gel electrophoretic mobility shift assay for quantitating protein: RNA interactions. Analytical Biochemistry, 511, 36–41. https://doi.org/10.1016/j.ab.2016.07.027
13. Hellman, L.M., & Fried, M.G. (2007). Electrophoretic mobility shift assay (EMSA) for detecting protein–nucleic acid interactions. Nature Protocols, 2(8), 1849–1861. https://doi.org/10.1038/nprot.2007.249
14. Phillips, J.C., Braun, R., Wang, W., Gumbart, J., Tajkhorshid, E., Villa, E., Chipot, C., Skeel, R.D., Kalé, L., & Schulten, K. (2005). Scalable molecular dynamics with NAMD. Journal of Computational Chemistry, 26(16), 1781–1802. https://doi.org/10.1002/jcc.20289
15. Jo, S., Cheng, X., Islam, S.M., Huang, L., Rui, H., Zhu, A., Lee, H.S., Qi, Y., Han, W., Vanommeslaeghe, K., MacKerell, A.D., Roux, B., & Im, W. (2014). CHARMM-GUI PDB manipulator for advanced modeling and simulations of proteins containing nonstandard residues. Advances in Protein Chemistry and Structural Biology, 96, 235–265. https://doi.org/10.1016/bs.apcsb.2014.06.002
16. Jorgensen, W., Chandrasekhar, J., Madura, J., Impey, R. & Klein, M. (1983). Comparison of simple potential functions for simulating liquid water. Journal of Chemical Physics, 79, 926-935. https://doi.org/10.1063/1.445869
17. Darden, T., York, D., & Pedersen, L. (1993). Particle mesh Ewald: An N⋅log(N) method for Ewald sums in large systems. Journal of Chemical Physics, 98, 10089-10092. https://doi.org/10.1063/1.464397
18. Humphrey, W., Dalke, A., & Schulten, K. (1996). VMD: visual molecular dynamics. Journal of Molecular Graphics, 14(1), 33–28. https://doi.org/10.1016/0263-7855(96)00018-5
19. Shindyalov, I.N., & Bourne, P.E. (1998). Protein structure alignment by incremental combinatorial extension (CE) of the optimal path. Protein Engineering, 11(9), 739–747. https://doi.org/10.1093/protein/11.9.739
20. Chang, A.L., McKeague, M., Liang, J.C., & Smolke, C.D. (2014). Kinetic and equilibrium binding characterization of aptamers to small molecules using a label-free, sensitive, and scalable platform. Analytical Chemistry, 86(7), 3273–3278. https://doi.org/10.1021/ac5001527

A big thank you to our sponsors!