(A-32) Modeling bivalent avidity of GAB1-PTPN11 binding in SpringSaLaD

The Src homology 2 (SH2) domain is a prevalent protein domain that operates in signaling pathways by binding to phosphorylated tyrosine (pTyr) residues (1). Several proteins, such as PTPN11, Syk, and ZAP-70, contain two SH2 domains which can bind to two pTyr residues on other proteins via their tandem SH2 domains, yielding a bivalent avidity effect and contributing to high specificity in their signaling pathways (2). An avidity effect between multivalent proteins occurs when the individual affinities between the proteins' binding sites enhance the overall affinity between the proteins themselves, due to the proximity between binding sites after an initial bond is formed (3). A target of the PTPN11 tandem SH2 domains is GRB2-associated-binder 1 (GAB1), which is recruited to the membrane via its PH domain and is phosphorylated by activated growth factor receptors (4). PTPN11 subsequently binds to GAB1, activating downstream pathways and stimulating several cellular processes (4). PTPN11 SH2 domains have distinct affinities, wherein the N-SH2 domain binds to pTyr-627 and the C-SH2 domain binds to pTyr-659 (5). We sought to better understand when and how the bivalent avidity effect influences GAB1-mediated recruitment of PTPN11 to the membrane using a stochastic, particle-based modeling software called SpringSaLaD (6).

SpringSaLaD models have been used to simulate multivalent protein-protein interactions such as actin signaling protein clustering (7,8). We present a SpringSaLaD model of GAB1-PTPN11 binding for the purpose of simulating the avidity effect produced by tandem SH2 domains and its response to changes to the system and the binding parameters.

We used SpringSaLaD to create a spatial stochastic model of GAB1-PTPN11 binding at the membrane (6, Fig. 1). GAB1 was simulated as an intracellular membrane-bound molecule, consisting of a membrane-bound PH domain (radius=1.0 nm) connected with a 10.0 nm link to an N-pTyr site (radius=1.0 nm), which was connected with a 5.0 nm link to a C-pTyr site (radius=1.0 nm). PTPN11 was simulated as an intracellular molecule, consisting of an N-SH2 domain (radius=1.8 nm) connected with a 2.0 nm link to a C-SH2 domain (radius=1.8 nm). All sites have a diffusion constant=2.0 μM2s-1. N-SH2 and N-pTyr sites (N-termini) bound at KD=0.287 μM (kon=3.484 μM-1s-1, koff=1 s-1) (9). C-SH2 and C-pTyr sites (C-termini) bound at KD=20.0 (kon=0.05 μM-1s-1, koff=1 s-1). Simulations were run in a 500×500×500 nm volume, with total time=1 s and time step=100 ns.

To calculate bivalent effective KD, we assumed equilibrium at 1 s and calculated the fraction of GAB1 bound with varying [PTPN11]total at fixed [GAB1]total. We plotted fraction bound against [PTPN11]free. The following equation (10) was used to generate a binding curve, and KD was approximated with Microsoft Excel Solver: 

Fraction bound=[PTPN11]free/([PTPN11]free+KD)

Effective KD was also calculated for N-termini and C-termini binding individually.

To monitor avidity, we created two monovalent versions of our model with either the N-termini or the C-termini binding reaction deleted. We then found the fraction of GAB1 bound at 1 s with varying C-termini binding KD. Fraction values from the monovalent simulations were added and compared to fractions in the bivalent model.

The effective bivalent KD of GAB1-PTPN11 binding and the effective KD of N-termini binding are both ≅0.210 μM (Fig. 2A, B). This KD is a higher affinity than the highest monovalent affinity (0.287 μM), indicating an avidity effect. Effective KD of C-termini binding is ≅0.236 μM, which is markedly lower than its monovalent KD (20.0 μM) (Fig. 2C). Affinity between C-termini is enhanced in the bivalent interaction, indicating that strong N-termini binding subsequently positions C-termini in proximity to one another.

When comparing the bivalent interaction to the two monovalent interactions, the fraction of GAB1 bound in the bivalent model was greater than the sum of the fractions of GAB1 bound in the two monovalent models when C-termini binding KD was ≥5 μM (Fig. 3). Thus, the bivalent affinity is greater than would be predicted by adding the monovalent interactions alone. However, the difference between these values decreased as C-termini binding KD increased from 5 μM to 40 μM, demonstrating a limit on the extent of the avidity effect.

Our model exhibits aspects of the avidity effect. Our modeling approach can be expanded to simulate additional molecules in the GAB1-PTPN11 system in order to observe how changes in the model’s parameters perturb the downstream recruitment of signaling molecules. Changes in model parameters can be used to reflect mutations in the pathway, allowing a better understanding of the relationship between protein binding kinetics, avidity, and signal transduction.

Alekhya Kandoor provided the measurements for the diameters of PTPN11 SH2 domains.

1. Machida, K., & Mayer, B. J. (2005). The SH2 domain: versatile signaling module and pharmaceutical target. Biochimica et biophysica acta, 1747(1), 1–25. https://doi.org/10.1016/j.bbapap.2004.10.005

2. Ottinger, E. A., Botfield, M. C., & Shoelson, S. E. (1998). Tandem SH2 domains confer high specificity in tyrosine kinase signaling. The Journal of biological chemistry, 273(2), 729–735. https://doi.org/10.1074/jbc.273.2.729

3. Vauquelin, G., & Charlton, S. J. (2013). Exploring avidity: understanding the potential gains in functional affinity and target residence time of bivalent and heterobivalent ligands. British journal of pharmacology, 168(8), 1771–1785. https://doi.org/10.1111/bph.12106

4. Wang, W., Xu, S., Yin, M., & Jin, Z. G. (2015). Essential roles of Gab1 tyrosine phosphorylation in growth factor-mediated signaling and angiogenesis. International journal of cardiology, 181, 180–184. https://doi.org/10.1016/j.ijcard.2014.10.148

5. Cunnick, J. M., Mei, L., Doupnik, C. A., & Wu, J. (2001). Phosphotyrosines 627 and 659 of Gab1 constitute a bisphosphoryl tyrosine-based activation motif (BTAM) conferring binding and activation of SHP2. The Journal of biological chemistry, 276(26), 24380–24387. https://doi.org/10.1074/jbc.M010275200

6. Michalski, P. J., & Loew, L. M. (2016). SpringSaLaD: A Spatial, Particle-Based Biochemical Simulation Platform with Excluded Volume. Biophysical journal, 110(3), 523–529. https://doi.org/10.1016/j.bpj.2015.12.026

7. Case, L. B., Zhang, X., Ditlev, J. A., & Rosen, M. K. (2019). Stoichiometry controls activity of phase-separated clusters of actin signaling proteins. Science (New York, N.Y.), 363(6431), 1093–1097. https://doi.org/10.1126/science.aau6313

8. Chattaraj, A., Youngstrom, M., & Loew, L. M. (2019). The Interplay of Structural and Cellular Biophysics Controls Clustering of Multivalent Molecules. Biophysical journal, 116(3), 560–572. https://doi.org/10.1016/j.bpj.2019.01.001

9. Ronan, T., Garnett, R., & Naegle, K. M. (2020). New analysis pipeline for high-throughput domain-peptide affinity experiments improves SH2 interaction data. The Journal of biological chemistry, 295(32), 11346–11363. https://doi.org/10.1074/jbc.RA120.012503

10. Jarmoskaite, I., AlSadhan, I., Vaidyanathan, P. P., & Herschlag, D. (2020). How to measure and evaluate binding affinities. eLife, 9, e57264. https://doi.org/10.7554/eLife.57264

11. Jumper, J., Evans, R., Pritzel, A., Green, T., Figurnov, M., Ronneberger, O., Tunyasuvunakool, K., Bates, R., Žídek, A., Potapenko, A., Bridgland, A., Meyer, C., Kohl, S. A. A., Ballard, A. J., Cowie, A., Romera-Paredes, B., Nikolov, S., Jain, R., Adler, J., Back, T., … Hassabis, D. (2021). Highly accurate protein structure prediction with AlphaFold. Nature, 596(7873), 583–589. https://doi.org/10.1038/s41586-021-03819-2

12. Varadi, M., Anyango, S., Deshpande, M., Nair, S., Natassia, C., Yordanova, G., Yuan, D., Stroe, O., Wood, G., Laydon, A., Žídek, A., Green, T., Tunyasuvunakool, K., Petersen, S., Jumper, J., Clancy, E., Green, R., Vora, A., Lutfi, M., Figurnov, M., … Velankar, S. (2022). AlphaFold Protein Structure Database: massively expanding the structural coverage of protein-sequence space with high-accuracy models. Nucleic acids research, 50(D1), D439–D444. https://doi.org/10.1093/nar/gkab1061

13. Berman, H. M., Westbrook, J., Feng, Z., Gilliland, G., Bhat, T. N., Weissig, H., Shindyalov, I. N., & Bourne, P. E. (2000). The Protein Data Bank. Nucleic acids research, 28(1), 235–242. https://doi.org/10.1093/nar/28.1.235