Identification of Electrostatic Hotspots at the Binding Interface of Amylin and Insulin-Degrading Enzyme: A Structural and Biophysical Investigation

1. Introduction

Amylin is a 37-amino-acid pancreatic hormone acting to control energy homeostasis and body weight [1,2,3,4]. Physiologically, amylin regulates glucose homeostasis by inhibiting insulin and glucagon secretion [5,6,7]. Furthermore, amylin modulates satiety and inhibits gastric emptying via the central nervous system [8,9,10,11]. Produced and released by pancreatic cells, amylin shares a common secretion pathway with insulin, collectively orchestrating the postprandial control of glucose homeostasis [12,13,14,15]. The intricate regulatory network governing blood sugar levels involves a delicate interplay between hormones and enzymes, among which amylin (islet amyloid polypeptide, IAPP) and insulin-degrading enzyme (IDE) stand as key players [12,16,17,18,19,20,21], where IDE plays a crucial role in the degradation and clearance of amylin from the bloodstream [22,23,24]. The cooperation between amylin and IDE, therefore, is pivotal in maintaining the delicate balance of glucose homeostasis and averting the detrimental consequences of amyloid deposition [25,26,27,28,29]. For instance, IDE defects are linked to the development of type 2 diabetes mellitus (T2DM) and Alzheimer’s disease (AD) [13,18,30].

2. Motivation

Thanks to the continued development of experimental structural biology and the half-a-century old Protein Data Bank (PDB) [31,32], a comprehensive structural biophysical (CSB) analysis becomes possible [33,34,35] for specific ligand-receptor [36,37,38], antigen-antibody [39] or enzyme-substrate [40,41,42] complex structures deposited in PDB, expanding our understanding of the structural and biophysical basis of their interfacial structural stability, and facilitating the design of drug analogues with improved affinity to their interacting partners [43,44].

As a matter of fact, structural and biochemical analyses [45,46,47,48] have already revealed the binding mode and pattern for the formation of the IDE-amylin complex structure, and this experimental information is useful but insufficient for the development of promising inhibitors (e.g., small molecules) of IDE-amylin interaction to improve glucose homeostasis. This manuscript, therefore, seeks to delve into the structural and biophysical aspects of the interaction between amylin and IDE, with a specific focus on identifying electrostatic hotspots at their binding interface. In case these hotspots are able to act as potential binding sites for drug-like small molecules, the aim of this article is to provide a precise and targeted approach for the development of small molecules to disrupt the amylin-IDE interaction.

3. Materials and Methods

As of February 1, 2024, a total of 101 experimental structures have been deposited in Protein Data Bank (PDB) [31] as listed in Table 1, according to a text query: QUERY: Full Text = "amylin" of the Protein Data Bank (PDB) [31]. Among them, only two experimental structures represent the amylin-IDE complex, with PDB IDs: 2G48 [45,46] and 3HGZ [47,48], respectively, providing an accurate structural basis of the IDE-amylin interaction specificity for subsequent comprehensive structural biophysical (CSB) analysis of the two structural models (two yellow rows in Table 1).

First, after the atomic coordinates file for PDB IDs: 2G48 [45,46] and 3HGZ [47,48] were downloaded from the PDB [31] website, Chimera [49] was employed to manually add hydrogen atoms to the structural model of the two structural models representing IDE-amylin complex. Afterwards, the two hydrogen-added structural models were subject to a set of comprehensive structural biophysical (CSB) analysis as described in [33] to identify key residue-specific interactions at the amylin-IDE binding interface and uncover the interstructural biophysics underlying the IDE-amylin complex structure.

Specifically, the CSB analysis here [33] consists of the structural identification of salt bridges and side chain hydrogen bonds at the binding interface of amylin and IDE. Given the fact that native proteins are in dynamic equilibrium with their less-structured, partially folded and/or unfolded states [50], and that, according to an NMR structure of human amylin bound to model membranes, the helix structure of amylin itself is also dynamic, this article uses two sets of screening criteria for the structural identification of potential hotspots at the IDE-amylin binding interface in the two structural models i.e., PDB IDs: 2G48 [45,46] and 3HGZ [47,48].

First, the same set of criteria as in [33] was used, i.e., the interfacial salt bridge analysis was conducted with an in-house python script only for titrateable residues (Asp, Glu, Lys, Arg and His), 4.0 Å was used as the cutoff distance for the two oppositely charged groups [33,51]. The hydrogen bond analysis was also conducted for only side chain nuclei with an in-house python script, and employed two geometric criteria: (a) a cutoff value of the angle formed by acceptor (A), donor (D) and hydrogen (H) ($\angle ADH$) of 30; (b) a cutoff value of donor-acceptor distance at 3.0 Å. That is, a hydrogen bond is only considered to be formed if $\angle ADH$ is not larger than 30 and the donor-acceptor distance is not larger than 3.0 Å [33,51].

Afterwards, a new set of criteria was used to account for the dynamic -helix structure of amylin itself, i.e., the interfacial salt bridge analysis was conducted with an in-house python script only for titrateable residues (Asp, Glu, Lys, Arg and His), 6.0 Å was used as the cutoff distance for the two oppositely charged groups [33,51]. The hydrogen bond analysis was also conducted for only side chain nuclei with an in-house python script, and employed two geometric criteria: (a) a cutoff value of the angle formed by acceptor (A), donor (D) and hydrogen (H) ($\angle ADH$) of 50; (b) a cutoff value of donor-acceptor distance at 5.0 Å. That is, a hydrogen bond is only considered to be formed if $\angle ADH$ is not larger than 30 and the donor-acceptor distance is not larger than 5.0 Å [33,51].

Here, the in-house python scripts essentially are the same as those used in [52], except for the differences in three key parameters related to the screening criteria, i.e., the salt bridge distance cutoff in Å, cutoff angle $\angle ADH$ in for hydrogen bonding, and the cutoff distance (in Å) of donor-acceptor for hydrogen bonding.

4. Results

4.1. Characterization of residue-specific electrostatic interactions at the amylin-IDE binding interface

As of February 1, 2024, there are only two experimental structures representing the amylin-IDE complex, with PDB IDs: 2G48 [45,46] and 3HGZ [47,48], respectively (Table 1). As defined in the PDB format content of PDB IDs: 2G48 [45,46] and 3HGZ [47,48], chains A and B represent IDE in 3HGZ, chains D and E represent amylin in 3HGZ, while chains A and B represent IDE in 2G48, and chains C and D represent amylin in 2G48, respectively. For the two structural models (2G48 [45,46] and 3HGZ [47,48]), this article reports a set of comprehensive structural biophysical (CSB) analysis as described in [33], which lead to a set of residue-specific electrostatic interactions at the binding interface of amylin-IDE, as listed in Table 2, Table 3, Table 4, Table 5, Table 6 and Table 7. Specifically,

• 10 interfacial salt bridges (Table 2) were structurally identified for the amylin-IDE complex structure PDB ID 2G48 [45,46], according to the old set of criteria as in [33].
• of the 10, only 1 interfacial salt bridge (Table 2) was structurally identified at the amylin-IDE binding interface for the amylin-IDE complex structure PDB ID 2G48 [45,46], according to the old set of criteria as in [33].
• 6 interfacial salt bridges (Table 2) were structurally identified for the amylin-IDE complex structure PDB ID 3HGZ [47,48], according to the old set of criteria as in [33].
• of the 6, 2 interfacial salt bridges (Table 2) were structurally identified at the amylin-IDE binding interface for the amylin-IDE complex structure PDB ID 3HGZ [47,48], according to the old set of criteria as in [33].
• 14 interfacial side chain and main chain hydrogen bonds (Table 3) were structurally identified for the amylin-IDE complex structure PDB ID 2G48 [45,46], according to the old set of criteria as in [33].
• no interfacial side chain or main chain hydrogen bond was structurally identified for the amylin-IDE complex structure PDB ID 3HGZ [47,48], according to the old set of criteria as in [33].
• 4 interfacial side chain hydrogen bonds (Table 4) were structurally identified for the amylin-IDE complex structure PDB ID 2G48 [45,46], according to the old set of criteria as in [33].
• no interfacial side chain was structurally identified for the amylin-IDE complex structure PDB ID 3HGZ [47,48], according to the old set of criteria as in [33].

As discussed above, to account for the dynamic helix structure and less-structured or random coil region of amylin, this article uses two sets of screening criteria for structural identification of potential electrostatic hotspots at the IDE-amylin binding interface in the two structural models i.e., PDB IDs: 2G48 [45,46] and 3HGZ [47,48]. Specifically,

• a total of 22 interfacial salt bridges (Table 5) were structurally identified for the amylin-IDE complex structure PDB ID 2G48 [45,46], according to a new set of criteria [33] as defined in the section of Materials and Methods.
• among the 22, only 3 interfacial salt bridges (Table 5) were structurally identified between Lys1 and Glu341 (Figure 2 and Figure 3) at the binding interface of IDE and amylin for the amylin-IDE complex structure PDB ID 2G48 [45,46], according to a new set of criteria [33] as defined in the section of Materials and Methods.
• a total of 16 interfacial salt bridges (Table 5) were structurally identified for the amylin-IDE complex structure PDB ID 3HGZ [47,48], according to a new set of criteria [33] as defined in the section of Materials and Methods.
• among the 16, only 4 interfacial salt bridges (Table 5) were structurally identified between Lys1 and Glu341 (Figure 2 and Figure 3) at the binding interface of IDE and amylin for the amylin-IDE complex structure PDB ID 3HGZ [47,48], according to a new set of criteria [33] as defined in the section of Materials and Methods.
• a total of 39 (Table 6) interfacial side chain hydrogen bonds were structurally identified for the amylin-IDE complex structure PDB ID 2G48 [45,46], according to a new set of criteria [33] as defined in the section of Materials and Methods.
• among the 39, 9 (Table 6) interfacial side chain hydrogen bonds were structurally identified at the binding interface of IDE and amylin for the amylin-IDE complex structure PDB ID 2G48 [45,46], according to a new set of criteria [33] as defined in the section of Materials and Methods.
• a total of 15 (Table 7) interfacial side chain hydrogen bonds were structurally identified for the amylin-IDE complex structure PDB ID 3HGZ [47,48], according to a new set of criteria [33] as defined in the section of Materials and Methods.
• among the 15, only four (Table 7) interfacial side chain hydrogen bonds were structurally identified at the binding interface of IDE and amylin for the amylin-IDE complex structure PDB ID 3HGZ [47,48], according to a new set of criteria [33] as defined in the section of Materials and Methods.

4.2. Structural identification of an electrostatic hotspot at amylin-IDE binding interface

Among the residue-specific electrostatic interactions at the amylin-IDE binding interface described above, one extraordinary pair of oppositely charged residues appear rather outstanding: lysine (Lys1, K1) residue of amylin and Glu341 (E341) of IDE, as this residue pair is the only one where interfacial electrostatic interactions were structurally identified at the amylin-IDE binding interface for two experimental structures representing the amylin-IDE complex, with PDB IDs: 2G48 [45,46] and 3HGZ [47,48] as listed in Table 1.

Take 2G48 [45,46] for example, one 3.514 Å interfacial salt bridge (Figurer̃efpymol2) was found to be formed between the oppositely charged side chains of Lys1 of amylin and Glu341 of IDE, while two interfacial salt bridges (2.441 and 3.179 Å) were also found to be formed between the oppositely charged side chains of Lys1 of amylin and Glu341 of IDE for 3HGZ [47,48]. Moreover, while no further salt bridges were found to be formed according to the old set of criteria as defined in [33], they are still quite close to each other for the side chains of Lys1 of amylin and Glu341 of IDE. Take 2G48 [45,46] for example, apart from the salt bridge, other positively charged side chain atoms of Lys1 of amylin are only 4.876 and 5.297 Å away from the negatively charged side chain oxygens of Glu341 of IDE, as shown in Figuresr̃efpymol2 and Figure 3.

In addition to interfacial salt bridges, the one extraordinary pair, i.e., Lys1 of amylin and Glu341 of IDE, were also found to be involved in a set of interfacial side chain hydrogen bonds, as shown by two yellow rows in Table 7 and also two yellow rows in Table 6, according to a new set of criteria [33] as defined in the section of Materials and Methods. Finally, what even more interesting is one interfacial side chain bridges formed between Lys1 of amylin and Glu341 of IDE, as shown in Figuresr̃efpymol2 and Figure 3, the main chain amide nitrogen (positively charged) of Lys1 of amylin is only 3.4 Å away from the negatively charged side chain oxygen of Glu341 of IDE, making the residue pair close to each other enough to form a strong main chain-side chain interfacial salt bridge, further strengthening the binding between amylin and IDE.

Taken together, these three sets of interfacial electrostatic interactions between Lys1 of amylin and Glu341 of IDE, i.e., side chain-side chain interfacial salt bridges, side chain-side chain interfacial hydrogen bonds, main chain-side chain interfacial salt bridges, highlights an extraordinary electrostatic hotspot at the amylin-IDE binding interface, making it an attractive precise target for small molecule inhibitor to reach inside IDE (the cage, Figurer̃efpymol1) and disrupt the amylin-IDE interaction for improved metabolic homeostasis.

5. Conclusion

Starting from two experimental structures representing the amylin-IDE complex, with PDB IDs: 2G48 [45,46] and 3HGZ [47,48] as listed in Table 1, this article puts forward a set of structural characterization for residue-specific electrostatic interactions at the amylin-IDE binding interface, as listed in Table 2, Table 3, Table 4, Table 5, Table 6 and Table 7. Moreover, this article also highlights one intriguing electrostatic hotspot (Figuresr̃efpymol2 and Figure 3) between the first N-terminal lysine (Lys1, K1) residue of amylin and Glu341 (E341) of IDE, with both interfacial salt bridges and side chain hydrogen bonds formed between the two oppositely charged residues sitting at the binding interface of amylin-IDE.

To sum up,

• the structural identification of electrostatic hotspots at the binding interface of amylin and IDE presents a promising avenue for the development of small molecules capable of disrupting this crucial interaction [54,55,56].
• This finding also contributes to the growing body of knowledge aimed at unraveling the intricacies of protein–protein interactions and provides a foundation for future research endeavors in the development of targeted therapeutics for metabolic disorders, particularly diabetes (T2DM) and obesity [57,58].
• The rationale for targeting electrostatic hotspots in the amylin-IDE interaction lies in the role of such sites as preferred binding locations for drug-like small molecules. By pinpointing these hotspots, we aim to provide a precise target for the development of small molecules capable of disrupting the amylin-IDE interaction. The potential therapeutic implications of such disruptors extend to modulating glucose homeostasis and mitigating the risk of amyloid formation, which is particularly relevant in the context of type 2 diabetes.

6. Discussion

6.1. Disrupting the amylin-IDE interaction: a drug discovery and design perspective

In drug discovery and design, targeting the interaction between amylin and IDE could be a potential strategy for therapeutic intervention. As amylin is involved in the regulation of blood glucose levels, manipulating its interaction with IDE might be explored as a way to modulate glucose homeostasis. While this article identified an intriguing electrostatic hotspots at the amylin-IDE binding interface, offering precise targets for therapeutic intervention and the development of small molecule inhibitors targeting the amylin-IDE interaction, the inhibition of the interaction between amylin and IDE can also lead to side effects such as glucose intolerance [59], as the intricate interplay between amylin and IDE constitutes a pivotal aspect of glucose homeostasis, with implications for the prevention of amyloid formation and maintenance of metabolic health [60,61,62]. Moreover, disrupting the interaction between amylin and IDE could have unintended consequences, as IDE is involved in the degradation of various peptides, and altering its interaction with amylin might affect the levels of other important regulatory molecules, selectivity is a key challenge in the design of small molecule inhibitor or even other molecular modalities such as small peptide of peptide-small molecule conjugate. Finally, still essential is a thorough understanding of the biological consequences of disrupting the amylin-IDE interaction, while disrupting the interaction between amylin and IDE could be a potential avenue for drug development, especially in the context of diabetes, obesity or even neurodegenerative diseases.

6.2. High-throughput comprehensive structural biophysical analysis: a methological perspective

In 2017, a comprehensive structural biophysical (CSB) approach was for the first time used in the analysis of experimental complex structures deposited in PDB [31] to address this question: how do SMA-linked mutations of SMN1 lead to structural/functional deficiency of the SMA protein i.e., the survivor motor neuron protein [33]? Here, the same structural biophysical approach was used here for the analysis of experimental complex structures of amylin and IDE, allowing for the precise identification of electrostatic hotspots at the binding interface of amylin and insulin-degrading enzyme. This level of precision aids in targeting specific regions for the development of small molecules. Moreover, By exploring the structural details and biophysical characteristics of the interaction, this approach provides valuable insights into the molecular interactions between amylin and insulin-degrading enzyme. Understanding these interactions is crucial for rational design of effective small molecule disruptors with high specificity, i.e., without affecting other biological processes. This specificity is crucial for minimizing off-target effects and enhancing the safety profile of potential therapeutic agents.

In addition, the information obtained through this approach contributes to rational drug design by providing a solid foundation for the development of small molecules. The identified hotspots serve as rational targets for disrupting the interaction, potentially enhancing the success rate of drug development efforts. Furthermore, in light of the increasingly large size of the Protein Data Bank, this CSB approach technically is applicable for a exhaustive analysis of the entire PDB for high-throughput extraction of structural and biophysical features [63,64] and continued generation and accumulation of synthetic structural and biophysical data with reasonable accuracy to support the development of machine learning-models such as GIBAC [65]. With this respect, however, in original PDB-format data, the experimentally determined atomic coordinates are presented in the ATOM records, and chances are that they do not exactly match the sequence (consisting of nucleic and/or amino acid residues) of the experimental sample per se, be it protein, DNA, RNA or their complexes with drugs and/or other small molecules [66,67].

Take amylin and IDE for example, the fasta format sequence of amylin in PDB ID 2G48 [45,46] is as below:

>2G48_2|Chains C, D|Islet amyloid polypeptide|null [45,46]

KCNTATCATQRLANFLVHSSNNFGAILSSTNVGSNTY

while the fasta format sequence of amylin in PDB ID 3HGZ [47,48] is as below:

>3HGZ_2|Chains C[auth D], D[auth E]|Islet amyloid polypeptide|null (9606) [47,48]

KCNTATCATQRLANFLVHSSNNFGAILSSTNVGSNTY

As listed in Table 8, in PDB ID 2G48 [45,46], there are a total of 37 residues for amylin in the experimental sample, with atomic coordinate information of 19 missing for chain C (amylin in PDB ID 2G48) and 20 missing for chain D (amylin in PDB ID 2G48), while in PDB ID 3HGZ [47,48], there are also a total of 37 residues for amylin in the experimental sample, with atomic coordinate information of 30 missing for chain D (amylin in PDB ID 3HGZ) and 27 missing for chain D (amylin in PDB ID 3HGZ).

As previously described in [68] in 2017, the past 53 years of experimental structure deposition in PDB also saw continued accumulation of experimentally uncharted territories (EUTs) inside it, reaching a point already where it is increasingly pressing for biomolecular structures (especially membrane proteins like Ca²⁺ channel [69,70]) to be experimentally determined in an EUT-less manner, as exemplified here again by the experimentally uncharted territories (EUTs) in two structural models representing IDE-amylin complex (PDB IDs: 2G48 [45,46] and 3HGZ [47,48]).

7. Ethical statement

No ethical approval is required.

8. 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.