An Exhaustive Exploration of the Semaglutide-GLP-1R Sequence Space towards the Design of Semaglutide Analogues with Elevated Binding Affinity to GLP-1R

1. Introduction

Semaglutide is a synthetic glucagon-like peptide-1 (GLP-1) receptor agonist that has garnered significant attention in the field of diabetes management [1,2,3]. Structurally, it is a synthetic peptide consisting of 39 amino acids. Semaglutide shares 94% sequence homology with natural human GLP-1 [2,3,4,5,6,7,8,9]. At the molecular level, semaglutide binds and activates the GLP-1 receptor, promoting insulin secretion and inhibiting glucagon release from pancreatic beta and alpha cells, respectively [10,11,12].

Originally developed by Novo Nordisk, semaglutide was approved by regulatory agencies for the treatment of type 2 diabetes mellitus (T2DM) due to its potent glucose-lowering effects and additional benefits such as weight loss and cardiovascular risk reduction [13,14,15,16]. Semaglutide is available in both injectable and oral formulations, with the injectable form typically administered once weekly and the oral form taken once daily [17,18,19]. The therapeutic efficacy of semaglutide arises from its ability to activate GLP-1 receptors located on pancreatic beta cells, leading to increased insulin secretion in a glucose-dependent manner. Additionally, semaglutide slows gastric emptying, suppresses appetite, and promotes satiety, contributing to its effects on weight loss and glycemic control [20,21,22].

Ligand-receptor binding affinity is an essential parameter in computer-assisted drug discovery and structure-based drug design [23]. Thanks to the continued development of experimental structural biology and the half-a-century old Protein Data Bank (PDB) [24,25,26,27,28], a comprehensive structural biophysical analysis becomes possible [29,30] for specific ligand-receptor complex structures deposited in PDB, such that our understanding of the structural and biophysical basis of their interfacial stability is able to help us modify the binding affinity of certain drug target and its interacting partners [31,32,33,34,35]. Take semaglutide for instance. On July 26, 2021, with a manually defined set of computational structural and biophysical analysis, a simple Val27-Arg28 exchange was for the first time introduced in the backbone of semaglutide to strengthen the semaglutide-GLP-1R binding affinity [7,9,36].

Figure 1. Strengthening semaglutide-GLP-1R binding affinity via a Val27-Arg28 exchange in the peptide backbone of semaglutide [9]. This figure was prepared with PyMol [37].

Figure 1. Strengthening semaglutide-GLP-1R binding affinity via a Val27-Arg28 exchange in the peptide backbone of semaglutide [9]. This figure was prepared with PyMol [37].

2. Motivation

The development of semaglutide analogues with increased GLP-1R binding affinity holds significant clinical relevance, offering the potential for enhanced glucose control, weight loss, and cardiovascular benefits in patients with type 2 diabetes and obesity [3,38,39]. By leveraging insights from structural biology and computational modeling and biophysics, this article seeks to design semaglutide derivatives that exhibit tighter interactions with the GLP-1R binding site, thereby improving receptor activation and downstream signaling pathways. These analogues may represent a new class of GLP-1R agonists with superior therapeutic efficacy and reduced dosing frequency, addressing current limitations in the management of metabolic disorders [40,41].

3. Materials and Methods

As listed in Table 1, there is one structure (determined by Cryo-EM) of Semaglutide-bound Glucagon-Like Peptide-1 (GLP-1) Receptor in Complex with Gs protein (PDB ID: 7KI0 [42]) as of 2024/05/14 14:02:52.

However, with a QUERY code: QUERY: Full Text = "Semaglutide", a total of three experimental structures related to semaglutide were found in the Protein Data Bank (PDB [24]), as listed in Table 2.

Among the three, there is one structure (determined by X-ray diffraction) of the semaglutide peptide backbone in complex with the extracellular domain of GLP-1R (PDB ID: 4ZGM [39]). Briefly, the amino acid sequences of the two chains of semaglutide and GLP-1R (according to PDB entry 4ZGM [39]) are listed in italics in fasta format as below,

>4ZGM_1|Chain A|Glucagon-like peptide 1 receptor|Homo sapiens (9606)

RPQGATVSLWETVQKWREYRRQCQRSLTEDPPPATDLFCNRTFDEYACWPDGEPGSFVNVSCPWYLPWASSVPQGHVYRFCTAEGLWLQKDNSSLPWRDLSECEESKRGERSSPEEQLLFLY

>4ZGM_2|Chain B|Semaglutide peptide backbone; 8Aib,34R-GLP-1(7-37)-OH|Homo sapiens (9606)

HAEGTFTSDVSSYLEGQAAKEFIAWLVRGRG

In combination with the comprehensive structural and biophysical analysis [29], the key amino acid residues at the semaglutide-GLP-1R complex binding interface (PDB ID: 4ZGM) were examined carefully in PyMol [9,37,39], and the inter-residue distances were calculated by PyMol [37] to identify potential neighbouring residue pair(s) to modulate the structural stability of the semaglutide-GLP-1R complex structure, leading to the design of a set of semaglutide variants with enhanced binding affinity.

In general, after homology structural modeling of semaglutide variants with Modeller [43], the binding affinity between semaglutide and GLP-1R was calculated using Prodigy [44,45]. Specifically, a total of ($s=g(28,4)=\frac{28!}{4!\left(24\right)!}\times {20}^4$ [46]) set of semaglutide analogues were generated with in-house Python script with four site-specific missense mutations introduced into native semaglutide sequences as listed above. Afterwards, homology structural modeling was carried out using Modeller [43] with PDB entry 4ZGM [39] as the structural template. Finally, the binding affinity between semaglutide and GLP-1R was calculated using Prodigy [44,45] for native semaglutide (10000 times) and for $s=g(28,4)=\frac{28!}{4!\left(24\right)!}\times {20}^4$ [46] semaglutide analogues (twenty times each).

4. Results

First off, with the X-ray structure of the semaglutide peptide backbone in complex with the extracellular domain of GLP-1R (PDB ID: 4ZGM [39]) in place, Modeller [43] was employed to build 10000 structural models with 100% homology to PDB ID: 4ZGM [39], and the binding affinity between semaglutide and GLP-1R was calculated using Prodigy [44,45] for native semaglutide (10000 times). As shown in Figure 2, most of the K_d values are located between 2.5 × 10^-6 M and 4.0 × 10^-6 M, with an average at 3.278 × 10^-6 M, which is rather close to the one K_d (3.4 × 10^-6 M) as reported in [9].

Secondly, with the X-ray structure of the semaglutide peptide backbone in complex with the extracellular domain of GLP-1R (PDB ID: 4ZGM [39]) as the structural template, a total of $s=g(28,4)=\frac{28!}{4!\left(24\right)!}\times {20}^4$ [46] semaglutide variants’ sequence were generated, and plugged into Modeller [43] to build 20 structural models for each semaglutide analogue, and the binding affinity between semaglutide and GLP-1R was calculated using Prodigy [44,45]. In total, the binding affinities of 100 semaglutide analogues to GLP-1R are included in Table 3, including their minimum, maximum, average and standard deviation of the K_d values calculated using Prodigy [44,45] for the semaglutide analogues, each 20 times of homology structural modeling using Modeller [43] In supplementary file supps.pdf, Table 1 includes a total of 8915 semaglutide analogues, including their minimum, maximum, average and standard deviation of the K_d values calculated using Prodigy [44,45] for the semaglutide analogues, each 20 times of homology structural modeling using Modeller [43].

Among the 100 semaglutide analogues included in Table 3, one particular semaglutide analogue stood out, named here as semaglutideX, where the semaglutideX-GLP-1R structural model is reaching a K_d value of 3.0 × 10^-8 M, while the K_d is 3.4 × 10^-6 M for the binding of native semaglutide to GLP-1 [9]. The amino acid sequence of semaglutideX is listed in italics in fasta format as below,

>semaglutideX (supplementary file semx.pdb)

HAEGTFTSDVSSYLEAQAAKEFQAWRNRGRG

For a close comparison, the amino acid sequence of semaglutide (PDB ID: 4ZGM [39]) is listed in italics in fasta format as below,

>4ZGM_2|Chain B|Semaglutide peptide backbone; 8Aib,34R-GLP-1(7-37)-OH|Homo sapiens (9606)

HAEGTFTSDVSSYLEGQAAKEFIAWLVRGRG

and the amino acid sequence of semaglutide with a Val27-Arg28 exchange [9] is listed in italics in fasta format as below,

>Val27-Arg28 exchange

HAEGTFTSDVSSYLEGQAAKEFIAWLRVGRG

With the X-ray structure of the semaglutide peptide backbone in complex with the extracellular domain of GLP-1R (PDB ID: 4ZGM [39]) in place, Modeller [43] was employed to build 20 structural models for each one of the 8915 semaglutide variants (supplementary file semx.pdb), and the binding affinity between semaglutide variants and GLP-1R were calculated using Prodigy [44,45] for 20 times. As shown in Figure 3, most of the K_d values are located between 1.0 × 10^-7 M and 2.0 × 10^-7 M, with an average at 1.437 × 10^-7 M, which is at least one order of degree lower that the K_d (3.4 × 10^-6 M) as reported in [9].

5. Conclusion

To sum up, this article reports a set of computationally designed semaglutide analogues with elevated binding affinity to the GLP-1 receptor compared to native semaglutide and also to a previous semaglutide with a Val27-Arg28 exchange [9]. Overall, these analogues hold promise for improving the treatment of type 2 diabetes and obesity by enhancing receptor activation and downstream signaling cascades. Future preclinical and clinical studies are needed to evaluate the pharmacological properties and therapeutic potential of these novel semaglutide derivatives. Moreover, continued optimization of GLP-1R agonists through structure-based drug design approaches could lead to further advancements in the field, ultimately benefiting patients with diabetes.

6. Discussion

Pharmacologically, there are reasons to design semaglutide analogues with elevated binding affinity to the glucagon-like peptide-1 receptor (GLP-1R) compared to semaglutide itself. However, designing semaglutide analogues with heightened binding affinity to the glucagon-like peptide-1 receptor (GLP-1R) presents both opportunities and challenges in the field of metabolic therapeutics. On the positive side, analogues with enhanced receptor affinity hold the potential to improve therapeutic outcomes by maximizing receptor activation and downstream signaling pathways, leading to more robust glucose control, weight loss, and potentially cardioprotective effects [47]. Additionally, these analogues may offer the advantage of reduced dosing frequency, enhancing patient compliance and convenience [48]. However, there are also drawbacks to consider, including the risk of off-target effects [49] and increased receptor desensitization with prolonged exposure to highly potent semaglutide analogues. Furthermore, the development of analogues with elevated binding affinity necessitates careful optimization to maintain selectivity and minimize adverse reactions [50,51]. Balancing these factors is crucial for realizing the full therapeutic potential of semaglutide analogues with enhanced GLP-1R binding affinity while mitigating potential risks, which is where a truly general intermolecular binding affinity calculator [46,52,53] will be useful to accomodate off-target effects and drug-drug interactions [49,50,51].

Finally, while this study represents a series of semaglutide analogues with elevated binding affinity to the GLP-1 receptor compared to native semaglutide and also to a previous semaglutide with a Val27-Arg28 exchange [9], the entire process of the design, along with the structural biophysics-based strategy for the molecular design [31,32,33,34,54], is essentially also a process of the construction of a semaglutide-GLP-1R based mini general intermolecular binding affinity calculator (GIBAC) [46,52,53] based on the structure of semaglutide peptide backbone in complex with the GLP-1 receptor extracellular domain determined by X-ray diffraction [39]. With a series of semaglutide analogues with elevated binding affinity to the GLP-1R available and this semaglutide-GLP-1R based mini GIBAC in place, here, this articles calls again for the construction of a truly general intermolecular binding affinity calculator to be listed on the agenda of the entire community of drug discovery and design [46,52,53].

8. Ethical Statement

No ethical approval is required.

9. Declaration of Generative AI and AI-Assisted Technologies in the Writing Process

During the preparation of this work, the author used OpenAI’s ChatGPT in order to improve the readability of the manuscript, and to make it as concise and short as possible. After using this tool, the author reviewed and edited the content as needed and takes full responsibility for the content of the publication.