Molecular and Computational Characterization of GLP-1 Receptor–Agonist Binding Mechanisms for Rational Therapeutic Design

ABSTRACT

Obesity is one of the fastest-growing global health challenges, driving diabetes, cardiovascular disease, and reduced life expectancy. Glucagon-like peptide-1 (GLP-1) receptor agonists—including liraglutide, semaglutide, exenatide, and lixisenatide—have revolutionized treatment; however, their injection-based delivery, high cost, and side effects limit widespread accessibility. To guide the development of next-generation therapies, this project established a computational pipeline to investigate GLP-1R interactions at the structural level. Ligand and receptor sequences were retrieved from UniProt and modeled using AlphaFold. Predicted receptor structures were analyzed for ligand-binding pockets, followed by molecular docking of peptide agonists to the GLP-1R using HDOCK. Resulting complexes were ranked and evaluated through binding free-energy estimation with PRODIGY and interaction mapping with PLIP. Comparative analysis across agonists revealed conserved interaction motifs and key residues that stabilize receptor engagement, forming drug-specific “interaction fingerprints.” Docking identified a common extracellular pocket for all ligands, but distinct stabilization patterns determined their relative affinities. Liraglutide showed the strongest predicted binding (−14.5 kcal/mol), enhanced by hydrophobic contacts from its fatty-acid tail, whereas semaglutide, exenatide, and lixisenatide exhibited fewer interactions. These findings provide a structural basis for clinical differences in potency and durability, while offering novel design principles to engineer safer, and effective, GLP-1R agonists.

INTRODUCTION.

Obesity is a major global health challenge that increases the risk of diabetic issues, cardiovascular disease, and a reduced quality of life.[1] While weight loss through nutrition, physical activity, and behavioral interventions remains the cornerstone of treatment, many patients struggle to achieve or sustain meaningful results. Pharmacological options are increasingly used to support weight management, but traditional medications such as orlistat or phentermine often provide only minimal benefits while causing significant gastrointestinal, cardiovascular, or psychological side effects. Surgical approaches like gastric bypass or gastrectomies achieve more dramatic outcomes but require invasive procedures and long recovery periods.

As the limitations of conventional weight-loss methods have become increasingly evident, there has been a growing focus on novel molecular and pharmacological strategies to address obesity more effectively. In recent years, glucagon-like peptide-1 (GLP-1) receptor agonists have transformed the landscape of obesity treatment. Compounds such as semaglutide, liraglutide, and the dual Gastric Inhibitory Polypeptide (GIP)/GLP-1 agonist tirzepatide mimic satiety hormones to suppress appetite, slow gastric emptying, and improve glycemic control (Figure 1).[2] Despite their clinical success, these drugs are expensive, raise safety concerns, and their molecular mechanisms of receptor binding are not fully understood. Since pharmaceutical companies rarely make detailed structural data publicly available, independent computational studies are necessary to clarify drug–receptor interactions. This study aims to investigate the molecular binding interactions between GLP-1R agonists and their targets using computational tools, artificial intelligence, and chemical analysis. Protein structures were modeled using AlphaFold, and ligand docking was performed with HDOCK, with binding affinity and interaction profiling assessed through PRODIGY and PLIP. By identifying binding patterns and key residues, this research provides insights that may inform the design of safer and more effective weight management therapies.

Given that most GLP-1R agonists are peptidyl in nature, understanding the biochemical and structural characteristics of peptide-based drugs is essential for optimizing their safety, stability, and therapeutic performance. Peptidyl drugs are therapeutic agents made of short chains of amino acids (peptides) that mimic natural biological molecules to regulate physiological processes. These drugs bridge the gap between small-molecule drugs and large biologics (like antibodies) — offering high specificity, potency, and lower toxicity. In this research, we employed computational simulations to characterize the interaction between peptidyl drugs and the GLP-1R, focusing on their inhibitory and activation mechanisms. Based on these interaction insights, we performed rational chemical mutations on the peptide drugs to improve their ability to modulate and potentially inhibit receptor activation. This work contributes to the development of novel therapeutic strategies targeting GLP-1R.

METHODS

Protein Sequence Retrieval.

The amino acid sequences of tirzepatide, exenatide, liraglutide, lixisenatide, and semaglutide were obtained from PubChem[3], which provides detailed molecular structures and physicochemical data for small peptide drugs. (UniProt ID of human GLP-1R: P43220), were retrieved from UniProt, a comprehensive database providing reliable protein sequences and associated functional annotations.[4] These curated sources ensured accuracy in molecular modeling and structural validation.

Protein Structure Prediction and Visualization.

Protein structures were modeled using AlphaFold, a deep learning–based tool that predicts three-dimensional protein conformations from amino acid sequences.[5] This server was used to generate 3D models for agonist. The resulting structures were saved in PDB format for subsequent docking analysis. PROTTER, an interactive protein visualization tool, was used to explore structural features of GLP-1R, including surface properties and functional annotations.[6] In addition, PrankWeb was used to predict and visualize putative binding pockets by uploading the modeled protein structures.[7]

Molecular Docking.

Molecular docking simulations between the receptor and the peptides were performed using the HDOCK server, which predicts receptor-ligand binding by generating three-dimensional docking poses and corresponding affinity scores.[8] Docking was conducted for tirzepatide, exenatide, liraglutide, lixisenatide, and semaglutide with the GLP-1R. Predicted complexes were further refined and compared across ligands to evaluate their relative binding conformations.

Binding Energy and Interaction Analysis.

Binding affinity values were calculated using the PRODIGY webserver, which estimates free energy of binding (ΔG) based on receptor–ligand interface characteristics.[9] To further characterize molecular interactions, the Protein–Ligand Interaction Profiler (PLIP) was employed.[10] PLIP automatically identified hydrogen bonds, hydrophobic contacts, salt bridges, π–stacking, and other noncovalent interactions within each docking pose. These results allowed for the identification of key amino acid residues involved in stabilizing receptor binding and facilitated comparisons between different GLP-1R agonists.

Mutations Studies.

Among the GLP-1R agonists—liraglutide demonstrated the strongest binding affinity to the GLP-1R based on molecular docking simulations. To further enhance this binding efficiency, several site-directed chemical mutations were introduced into the liraglutide sequence, inspired by observed mutations in related analogs. These included F6R, V20H, G16D, S8D, K28R, A18D, D9R, and Y13E substitutions. Each mutation was designed to increase receptor interaction stability and improve peptide resistance to enzymatic degradation.

RESULTS

The computational pipeline integrating AlphaFold, PrankWeb, HDOCK, PLIP, and PRODIGY produced a detailed structural understanding of how different GLP-1R agonists interact with GLP-1R. Each modeled complex revealed distinct binding orientations, noncovalent contact networks, and binding free energies that together explained differences in pharmacological performance among the drugs.

Predicted Binding Site.

The specific binding pocket positions shown in Figures 2a and 2b were predicted using PrankWeb to identify potential binding sites. These locations were then annotated and visualized using PROTTER to show structural features and ChimeraX for a surface-level depiction. These results will be used to validate the molecular docking results as discussed in the next section. As illustrated in Figure 2c, GLP-1R contains a clearly defined binding cavity for GLP-1 located in the extracellular vestibule.

Molecular Docking Simulations.

We then docked each peptide to the GLP-1R, selecting the highest-ranked pose for each (Figure 3). The resulting complexes were retrieved from HDOCK, opened in ChimeraX, and color-coded—purple for the receptor and pink for the peptide ligands. All four agonists were found to occupy the cavity predicted by PrankWeb, but each adopted a slightly different orientation and depth. These subtle pose variations indicate how each peptide forms initial contacts with different areas of the pocket rim and interior, which is consistent with the patterns in interaction types and binding energies discussed in the following sections.

Molecular Docking Analysis.

Using PLIP, we analyzed the noncovalent interactions in each peptide-receptor complex, including hydrogen bonds, hydrophobic interactions, salt bridges, and π-interactions. As detailed in Figure 4, liraglutide stood out with the highest total number of contacts. It established numerous hydrogen bonds within the receptor cavity and demonstrated strong hydrophobic anchoring, largely thanks to its fatty-acid tail interacting along the receptor’s surface. In contrast, exenatide, semaglutide, and lixisenatide formed fewer stabilizing interactions and had less dense hydrogen-bonding networks, resulting in weaker overall interaction profiles. These structural differences help explain the affinity hierarchy observed in our binding energy results.

Binding Energy Profiling.

The binding free energy estimates calculated using PRODIGY support the observed interaction profiles (see Figure 5). The GLP-1R complexes showed ΔG values of −14.5 kcal/mol for liraglutide, −11.6 kcal/mol for semaglutide, −10.4 kcal/mol for exenatide, and −9.8 kcal/mol for lixisenatide. Liraglutide’s notably strong predicted affinity further confirms its extensive interaction network and firm engagement with the binding pocket, supporting its selection for additional mutational studies. To investigate which specific residues contribute most to binding stability, we designed and analyzed several sequence variants (including F6R, V20H, G16D, G16D, S8D, K28R, A18D, D9R, Y13E) aiming to define key positions for optimizing future analogs. This bar graph illustrates the comparative number of molecular interactions formed between various GLP-1R agonists [exenatide, liraglutide, lixisenatide, and semaglutide] and different receptor mutants. Among the drugs, liraglutide and the A18D mutant exhibited the highest number of hydrophobic interactions, suggesting strong nonpolar stabilization within the receptor-binding pocket. These two also showed a greater number of hydrogen bonds, indicating enhanced binding specificity and affinity. In contrast, salt bridges were fewer across all drugs, reflecting their relatively smaller contribution to overall binding. Mutants such as A18D and the combined mutation group (“All Mutations”) demonstrated the greatest total interaction counts, implying that specific amino acid substitutions can improve receptor–ligand binding strength. Meanwhile, lixisenatide and the Y13E mutant displayed the lowest number of interactions, suggesting weaker or less favorable receptor engagement compared to other complexes.

Figure 5 reports the quantitative outcomes (ΔG, kcal/mol) for all peptides and all liraglutide edits: exenatide ≈ −10.6 kcal/mol; lixisenatide ≈ −10.0 kcal/mol; semaglutide ≈ −11.6 kcal/mol; liraglutide −14.5 kcal/mol; F6R −14.6 kcal/mol; A18D −14.8 kcal/mol; D9R −13.4 kcal/mol; Y13E −11.1 kcal/mol; G16D, S8D, K28R −12.1 kcal/mol; F6R, V20H, G16D −13.0 kcal/mol; F6R, V26H, G16D −15.3 kcal/mol; F6R, V26H −15.6 kcal/mol. PLIP interaction results for these designs correspond to the same binding pocket used for the original peptide and include the hydrogen bonds and hydrophobic contacts referenced above.

Finally, we evaluated the potential for amyloid formation in the peptide ligands using AmyloGram.[11] Figure 6 displays per-residue amyloidogenicity scores plotted against a blue threshold line; any segment surpassing this line is flagged as likely to form amyloids. This screening helps identify specific sequence regions that may pose a risk while confirming that the majority of residues remain below the threshold. These insights are used to pinpoint residues that should be altered or avoided when designing optimized analogs including in the liraglutide variant set, helping to maintain strong binding affinity while minimizing aggregation risks before moving to experimental testing.

DISCUSSION

Drug-based methods for weight management involve FDA-approved medications that help reduce body weight by suppressing appetite, increasing satiety, reducing fat absorption, or regulating metabolic pathways. Commonly prescribed agents include GLP-1 receptor agonists such as semaglutide and liraglutide, which slow gastric emptying and enhance satiety, leading to significant weight loss.[12] Dual GIP/GLP-1 agonists like tirzepatide further amplify appetite control and metabolic benefits, showing even greater efficacy in recent clinical trials.[13] Centrally acting appetite suppressants, including phentermine, phentermine/topiramate, and naltrexone/bupropion, modulate neurotransmitters involved in hunger and reward pathways.[14] Another option is orlistat, a gastrointestinal lipase inhibitor that reduces fat absorption by approximately 30%.[15] These medications are typically recommended for individuals with a BMI ≥30 kg/m², or ≥27 kg/m² with obesity-related comorbidities, and are most effective when combined with lifestyle modification.[2] However, long-term use requires monitoring for side effects such as gastrointestinal disturbances, nausea, mood changes, or cardiovascular risks, depending on the drug class.[16] Across current guidelines, pharmacotherapy is considered an essential second-line treatment for obesity when behavioral interventions alone do not produce adequate weight loss.

Peptide-based therapeutics are inherently susceptible to aggregation and amyloid-like fibril formation, particularly under high concentration or prolonged storage conditions. Since GLP-1 receptor agonists are administered chronically for metabolic disease, minimizing aggregation risk is critical for maintaining stability, bioavailability, and safety. Aggregation can reduce effective drug concentration, increase immunogenicity, and in rare cases contribute to tissue-level deposition. Therefore, amyloidogenicity screening using AmyloGram was incorporated into this study to evaluate whether sequence modifications designed to enhance receptor binding inadvertently increase aggregation propensity.

Application of this research.

One primary application is in type 2 diabetes, since GLP-1R agonists also help regulate insulin and glucose levels. Additionally, this research can support drug design and optimization by utilizing computational modeling to develop new GLP-1 analogs with improved stability, reduced side effects, and lower costs. It can also be applied to metabolic syndrome research, since many of the same pathways affect blood pressure, cholesterol, and energy metabolism.

CONCLUSION

This analysis examined GLP-1R binding to guide next-generation peptide design for weight management, comparing approved agonists and then modifying liraglutide within the same GLP-1R pocket using a consistent docking, PLIP, and binding-energy workflow. Three modified liraglutide variants—F6R, V26H, F6R, V26H, G16D, and F6R—demonstrated stronger GLP-1R binding than the parent peptide, indicating that targeted sequence substitutions can enhance receptor affinity and guide next-generation therapeutic design.

REFERENCES

1. P. G. Kopelman, Obesity as a medical problem. Nature 404, 635-643 (2000).
2. X. Pi-Sunyer et al., A Randomized, Controlled Trial of 3.0 mg of Liraglutide in Weight Management. New England Journal of Medicine 373, 11-22 (2015).
3. S. Kim et al., PubChem 2023 update. Nucleic Acids Research 51, D1373-D1380 (2022).
4. T. U. Consortium, UniProt: a hub for protein information. Nucleic Acids Research 43, D204-D212 (2014).
5. J. Abramson et al., Accurate structure prediction of biomolecular interactions with AlphaFold 3. Nature 630, 493-500 (2024).
6. U. Omasits, C. H. Ahrens, S. Müller, B. Wollscheid, Protter: interactive protein feature visualization and integration with experimental proteomic data. Bioinformatics 30, 884-886 (2013).
7. R. Krivák, D. Hoksza, P2Rank: machine learning based tool for rapid and accurate prediction of ligand binding sites from protein structure. Journal of Cheminformatics 10, 39 (2018).
8. Y. Yan, D. Zhang, P. Zhou, B. Li, S. Y. Huang, HDOCK: a web server for protein-protein and protein-DNA/RNA docking based on a hybrid strategy. Nucleic Acids Res 45, W365-w373 (2017).
9. C. L. Xue, P. J. Rodrigues, L. P. Kastritis, M. A. Bonvin, A. Vangone, PRODIGY: a web server for predicting the binding affinity of protein–protein complexes. Bioinformatics 32, 3676-3678 (2016).
10. S. Salentin, S. Schreiber, V. J. Haupt, M. F. Adasme, M. Schroeder, PLIP: fully automated protein–ligand interaction profiler. Nucleic Acids Research 43, W443-W447 (2015).
11. D. B. Kell et al., AmyloGram Reveals Amyloidogenic Potential in Stroke Thrombus Proteomes. Biochemical Journal, (2025).
12. J. P. H. Wilding et al., Once-Weekly Semaglutide in Adults with Overweight or Obesity. New England Journal of Medicine 384, 989-1002 (2021).
13. T. Min, S. C. Bain, The Role of Tirzepatide, Dual GIP and GLP-1 Receptor Agonist, in the Management of Type 2 Diabetes: The SURPASS Clinical Trials. Diabetes Therapy 12, 143-157 (2021).
14. A. A. Coulter, C. J. Rebello, F. L. Greenway, Centrally Acting Agents for Obesity: Past, Present, and Future. Drugs 78, 1113-1132 (2018).
15. S. Henness, C. M. Perry, Orlistat. Drugs 66, 1625-1656 (2006).
16. I. Potočnjak, V. Degoricija, D. Vukičević Baudoin, J. Čulig, M. Jakovljević, Cardiovascular side effects of psychopharmacologic therapy. International Journal of Cardiology 219, 367-372 (2016).

---

---

Posted by buchanle on Tuesday, June 2, 2026 in May 2026.

Tags: AlphaFold, GLP-1 Receptor, Liraglutide, Weight Management