Impact of the S20G Mutation on the Structural and Aggregation Properties of hIAPP Monomers

Protein conformational diseases are caused by protein misfolding and the resulting aggregates (such as amyloid deposits), including Alzheimer's disease (AD) involving β-amyloid peptide (Aβ), transmissible spongiform encephalopathies (TSEs) involving prion protein (PrP), and type II diabetes (T2D) involving human islet amyloid polypeptide (hIAPP). Currently, there is no specific highly effective treatment for these diseases. The reason for this is that our understanding of the "dynamic" aggregation structural characteristics and the driving forces of misfolded proteins in misfolding diseases is not deep enough, and we have not been able to understand the microscopic mechanisms of misfolded protein aggregation at the molecular level.

Diabetes mellitus (DM) is a common and prevalent metabolic disease associated with genetic factors. According to the classification and diagnostic criteria for diabetes published by the American Diabetes Association (ADA) in 2018, diabetes can be divided into the following types: type 1 diabetes, type 2 diabetes, and specific types of diabetes caused by other reasons. Research indicates that the most common type of DM is type 2 diabetes (T2D), with the majority of patients being between 35 and 40 years old, accounting for more than 90% of diabetes patients. However, in recent years, an increasing number of children have been diagnosed with this disease, and the rising incidence may be related to the increase in childhood obesity. Common symptoms of T2D include excessive thirst, frequent urination, and unexplained weight loss. Over the past 50 years, the number of patients has continued to grow, showing a trend of spreading from Europe to the Western Pacific regions such as Asia and Africa.¹ Additionally, according to statistics from the World Health Organization, by 2045, nearly 700 million people worldwide will be affected by this disease.¹

The characteristics of T2D include dysfunction of pancreatic β cells, abnormal function of pancreatic α cells, amyloid protein deposition in islet tissue, and insulin resistance in peripheral tissues. These changes lead to: (1) hyperglycemia due to impaired peripheral glucose uptake, (2) dyslipidemia (high triglycerides and low HDL cholesterol) due to impaired peripheral fat uptake, (3) impaired amino acid uptake and ATP production in peripheral tissues (such as skeletal muscle) due to impaired nutrient uptake, and (4) increased glucagon secretion, further exacerbating hyperglycemia and hyperlipidemia.^2-4 Statistics show that the etiology of T2D is complex, with varying degrees of pathological effects on different systems and organs of the human body.

The onset of T2D is determined by three factors: insufficient insulin secretion by pancreatic β cells, decreased insulin sensitivity in peripheral tissues, and hIAPP-derived aggregates or amyloid deposits.^5-8 Studies conducted in primates strongly support the view that islet amyloidosis and β cell apoptosis are two key determinants of islet dysfunction.^9-10 However, the cellular processes that regulate amylin conversion and hIAPP-induced β cell apoptosis in human islets are still poorly understood. At the same time, the pathogenesis of T2D is still not fully understood, and it is currently believed to be related to genetic factors,¹¹ environmental factors,^12-13 insulin resistance, and insulin secretion defects.^14-16

hIAPP, also known as amylin, is a neuroendocrine hormone that is stored in pancreatic β cells along with insulin.¹⁷ The islet amyloid deposits of hIAPP found in the pancreas of T2D patients are mainly composed of a 37-residue peptide. hIAPP can regulate blood glucose levels by inhibiting glucagon secretion and delaying gastric emptying. The monomeric form of hIAPP is modified at the C-terminus and has an intramolecular disulfide bond between Cys2 and Cys7 residues. Under physiological conditions, hIAPP mainly exists in an unstructured and soluble monomeric form.¹⁸ Further research has extended to other animal models, and studies on primates (rhesus monkeys, baboons) and cats have found that these animals also exhibit varying degrees of amyloid deposits in their islets when T2D characteristics are successfully induced.¹⁹ Further research shows that IAPP deposition occurs in three steps: linear IAPP first forms an α-helix structure, then a β-sheet, and finally oligomers and polymers, forming mature fibrils and leading to the formation of non-degradable amyloid deposits.²⁰ The amino-terminal region of IAPP, specifically residues 1-19, is crucial for forming the α-helix structure. Due to the formation of a disulfide bond, hIAPP is more prone to forming an α-helix structure compared to rIAPP. The 18th amino acid in hIAPP is histidine (His), while in rIAPP it is arginine (Arg). A series of research results indicate that whether IAPP can form amyloid deposits directly depends on its conformation, and hIAPP's tendency to form amyloid deposits is due to the instability of its peptide molecular structure, making it more prone to conformational changes.^9-10 There is a serine to glycine mutation at position 20 (S20G) in hIAPP was found in 4.1% of T2D patients and 10% of early-stage T2D patients in Japan, suggesting that the S20G mutation may play a partial role in the pathogenesis of T2D in the Japanese population.²¹ The S20G mutation found in the Asian population is associated with early-stage T2D onset.²² Besides, the behavior of histidine is highly susceptible to external factors such as pH environment, side chain interactions, solution concentration, and the orientation of N/N-H atoms in the imidazole ring. These factors can significantly alter histidine behavior in real-world conditions, leading to different structural properties and aggregation characteristics. Additionally, studies on artificial rodent islet amyloid polypeptide (rIAPP) indicate that H18 plays a crucial role in regulating hIAPP fibril formation.^23-24 However, under the combination of S20G and histidine behaviors, the structural properties and aggregation properties is still unclear. In current study, studying the impact of S20G on the conformational tendency of hIAPP's 18th histidine residue at the molecular level, as well as its effect on the structural tendency of hIAPP monomers and the molecular aggregation mechanism, may provide valuable insights into the role of hIAPP in the pathogenesis of T2D.

For construct the initial structural model of the system, we obtained the initial file of the hIAPP(1-37) monomer from the Protein Data Bank database (ID: 5MGQ) and select the amino acid sequence of hIAPP(37) for simulation (Fig. 2.1). Change the serine at position 20 to glycine (Ser→Gly) and set up three monomer systems for histidine in the following states: (ε), (δ), and (p) (Figure 2.2). The (ε) and (δ) states are tautomers of histidine (His), labeled as hIAPP(ε) and hIAPP(δ), while the (p) state represents the protonated form of His, labeled as hIAPP(p). Additionally, the Amber 18 software²⁵ with the ff99SB force field²⁶ was used. The SHAKE algorithm was employed to constrain bonds, including hydrogen bonds.²⁷ In the REMD simulation, 10 independent REMD trajectory simulations were conducted for each of the three molecular systems, with each replica being simulated for more than 900 ns. The temperatures were set from 310 K to 510 K in 10 groups. Structural changes are analyzed using Visual Molecular Dynamics.²⁸

Before conducting data analysis, it is essential to evaluate the equilibrium state of the Replica Exchange Molecular Dynamics (REMD) simulations. To this end, we statistically analyzed the clustering characteristics of three systems across different time scales and compared the conformational properties of the top 5 distributions in Fig. 1 with the 5 main distributions to confirm the equilibrium trajectories. In this study, each system underwent REMD simulations exceeding 800 ns. For the hIAPP(ε) system, the clustering distribution trends between 600-800 ns and 800-1000 ns are largely consistent, indicating that the hIAPP(ε) system has reached kinetic equilibrium. Therefore, the subsequent analysis can extract the 400 ns equilibrium trajectory of the hIAPP(ε) system within the 600-1000 ns time frame. Similarly, the hIAPP(δ) system is in kinetic equilibrium between 400-600 ns and 600-800 ns, allowing the use of the 400-800 ns equilibrium trajectory for further analysis. For the hIAPP(p) system, it is in kinetic equilibrium between 600-800 ns and 800-1000 ns, so the 600-1000 ns equilibrium trajectory should be used for subsequent analysis. Statistical tests of the clustering distributions of the three systems across different time scales indicate that the REMD simulations have converged well. The REMD exchange ratio is 26.7%-27.4%, specifically hIAPP(ε)-26.7%, hIAPP(δ)-27.2%, and hIAPP(p)-27.4%. The REMD exchange ratio also indicates that our systems explore a wider sampling space.

Figure1.

The cluster distributions in two adjacent 200 ns simulation time, (a) hIAPP(ε) S20G, (b) hIAPP(δ) S20G, and (c) hIAPP(p) S20G.

To thoroughly investigate the structural evolution characteristics of each system during the simulation, we employ the Definition of Secondary Structure of Proteins (DSSP) algorithm,²⁹ which assigns secondary structure types to residues based on the positions of backbone amide (N-H) and carbonyl (C=O) atoms, to conduct a detailed analysis of the secondary structures of each system. On the one hand, the analysis focuses on the content of parallel β-sheet, antiparallel β-sheet, and α-helix in the monomer structures, as illustrated in Fig. 2. Through systematic statistical analysis, organization, and in-depth examination of the data generated by the DSSP algorithm, we reveal that the β-sheet content in the hIAPP(ε), hIAPP(δ), and hIAPP(p) systems is relatively low, ranging from 2.4% to 4.6%. In stark contrast, the α-helix content in these three systems ranges from 10.4% to 11.9%, significantly higher than the β-sheet content. To further elucidate the specific characteristics of the residue structures in each system, we collect more detailed information on the secondary structures of the three systems, as shown in Fig. 2. From a holistic perspective, the β-sheet content in the three systems is as follows: 2.4% for the hIAPP(ε) system, 4.6% for the hIAPP(δ) system, and 2.4% for the hIAPP(p) system. Regarding parallel β-sheet structures, the content is 0.6% for the hIAPP(ε) system, 0.6% for the hIAPP(δ) system, and 0.3% for the hIAPP(p) system. The content of antiparallel β-sheet structures is 1.8% for the hIAPP(ε) system, 4.0% for the hIAPP(δ) system, and 2.0% for the hIAPP(p) system. It is evident that the content of parallel β-sheet structures is not only low but also very similar across the three systems. Additionally, the content of antiparallel β-sheet structures is slightly higher than that of parallel β-sheet structures in each system. A deeper comparison of the antiparallel β-sheet content reveals that in the hIAPP(ε) system, the content of antiparallel β-sheet structures is generally low, with all regions showing less than 5.0%. This strongly indicates that the hIAPP(ε) system struggles to form β-sheet structures. In the hIAPP(δ) system, the content of antiparallel β-sheet structures in the F15-H18 region (8.0%-19.1%) and the I26-S29 region (8.4%-17.9%) is significantly higher than in both the hIAPP(ε) and hIAPP(p) systems, and these high-content regions are relatively extensive. In contrast, in the hIAPP(p) system, antiparallel β-sheet structures are observed only in very small regions, such as the V17-S18 region (6.3%-6.7%) and the I26-L27 region (5.4%-7.1%). In summary, among the three systems, the hIAPP(δ) system has the highest β-sheet content, indicating the presence of a small number of relatively stable β-sheet structures. The hIAPP(p) system follows, suggesting that it may also exhibit a minimal number of β-sheet structures. The hIAPP(ε) system has the lowest β-sheet content, further confirming that this system struggles to form β-sheet structures.

Figure2.

The averaged secondary structure probabilities, (a) probability distributions of the parallel, antiparallel β-sheet and α-helix of each residue in hIAPP(ε) S20G, (b) hIAPP(δ) S20G, (c) and hIAPP(p) S20G (d) systems.

Additionally, the α-helix content in the three systems is as follows: hIAPP(ε) (11.1%), hIAPP(δ) (10.4%), and hIAPP(p) (11.9%). This indicates that the α-helix content order is hIAPP(p) > hIAPP(ε) > hIAPP(δ). However, as shown in Fig. 2, the α-helix content in specific regions of the three systems exhibits similarities, particularly in the A5-L12 (5.4%-29.2%), G20-L27 (6.5%-15.4%), and T30-V32 (5.8%-8.4%) regions. Notably, in the A13-V17 region, the α-helix content in the hIAPP(p) system (14.6%-41.9%) is significantly higher than in the hIAPP(ε) system (11.1%-30.5%) and the hIAPP(δ) system (15.7%-26.7%). These data suggest that all three systems have high α-helix content, indicating a strong propensity for forming helical structures.

From the overall, the hIAPP(p) system exhibits both α-helix structures and a small number of β-sheet structures. In contrast, the hIAPP(ε) system is predominantly characterized by helical structures, although the helical content is lower than that in the hIAPP(p) system. The hIAPP(δ) system has the lowest α-helix content and contains antiparallel β-sheet structures. Data analysis reveals that the β-sheet content in the hIAPP(δ) system is significantly higher than in the other two systems, suggesting that a portion of the α-helix structures in the hIAPP(δ) system has transformed into a small number of β-sheet structures, resulting in lower stability compared to the hIAPP(p) system. Reviewing the literature indicates that helical structures are considered a typical structural feature of oligomeric intermediates during hIAPP amyloid formation. Therefore, it can be concluded that the secondary structure of the hIAPP(p) system is the most stable, followed by the hIAPP(δ) system, and finally the hIAPP(ε) system.

On the other hand, surprisingly, we observe that the α-helix content in the N22-L27 region of the unmutated hIAPP(ε), hIAPP(δ), and hIAPP(p) systems ranges from 13.7%-35.0%, 11.9%-33.8%, and 6.4%-32.3%, respectively. In contrast, the α-helix content in the same region of the S20G mutated hIAPP(ε), hIAPP(δ), and hIAPP(p) systems are 10.4%-15.4%, 9.4%-14.7%, and 9.8%-14.5%, respectively. It is evident that the α-helix content in the N22-L27 region of the S20G mutant is significantly lower than that of the unmutated systems, indicating that the α-helix structures in this region have transformed into β-sheet structures. This suggests that the mutation promotes the formation of β-sheet structures, leading to a decrease in helical content at the C-terminus, and that the S20G mutation has a substantial impact on the C-terminal structure of the hIAPP monomer. In the hIAPP(δ) system, we also notice an anomaly. In the unmutated hIAPP(δ) system, the content of antiparallel β-sheet structures in the Q10-L12 region is 4.6%-6.4%, and in the V17-S20 region, it is 5.8%-7.7%. However, in the mutated hIAPP(δ) system, the content of antiparallel β-sheet structures in the F15-H18 region is 8.0%-19.1%. This data indicates that the mutation shifts the β-sheet structures closer to the N-terminus and increases their content. Additionally, we find that in the antiparallel β-sheet region of the hIAPP(δ) system, the content in the I26-S29 region increases from 2.3%-8.4% in the unmutated system to 8.4%-17.9% in the mutated system. This suggests that the mutation facilitates the formation of β-sheet structures in the hIAPP(δ) system. Notably, compared to the unmutated systems, the S20G mutation more effectively induces the formation of β-sheet structures in the hIAPP(δ) system.

Cluster analysis allows the numerous conformations obtained from the simulation to be divided into clusters, each representing a major conformational state or distribution region of the molecule. The probabilistic meaning of this analysis is the likelihood that a conformation or data point belongs to a particular cluster, indicating the degree to which each conformation tends to belong to each cluster. In current study, cluster analysis was employed to investigate conformational properties and distributions, as illustrated in Fig. 3. The primary conformations of each system are depicted in Fig. 3. The primary conformation possibilities for the hIAPP(ε), hIAPP(δ), and hIAPP(p) systems are 3.2%, 5.7%, and 18.7%, respectively. It is evident that all three systems exhibit β-bridge or β-sheet structures, although these do not dominate and are present in low amounts. Consequently, α-helix structures constitute the predominant conformation in all three systems, which is consistent with our previous analysis of secondary structure properties. Notably, the hIAPP(p) system contains a higher number of α-helix structures along with a small quantity of β-sheet structures, suggesting that the hIAPP(p) system is more stable compared to the hIAPP(ε) and hIAPP(δ) systems. However, regardless of whether the system is in a protonated state (hIAPP(p)) or deprotonated states (hIAPP(ε) and hIAPP(δ)), the most favorable conformation content in each system is relatively low. This indicates that the hIAPP monomers in all three systems tend towards a flexible structure.

Figure3.

The most abundant conformational states and its probability distributions in hIAPP(ε) S20G, hIAPP(δ) S20G, and hIAPP(p) S20G systems.

To further investigate the H-bonding interactions and functions of each residue, we analyze the probabilities of H-bonding networks in three systems (Table 1). In the hIAPP(ɛ) system, H-bonding interactions are primarily observed between N14/H18 (10.6%), H18/G20 (8.3%), H18/N21 (7.5%), and H18/N21 (13.0%). In the hIAPP(δ) system, the main H-bonding interactions are between N14/H18 (6.6%) and H18/N22 (5.3%). In the hIAPP(p) system, the main H-bonding interactions are between N14/H18 (5.4%) and F15/H18 (5.4%), which favorably maintain the stability of α-helix structure. The probabilities of H-bonding interactions in all three systems are relatively low, indicating structural instability, which is consistent with our conclusions from the secondary structure analysis. Additionally, we observe that the N14/H18 H-bonding interactions interaction is present in all three systems, suggesting that N14 plays a special role in the aggregation of hIAPP monomers.

To gain a deeper understanding of the contact properties in the three systems, we collect residue contact maps for the amino acids (Fig. 4). The contact maps clearly show that interactions in the hIAPP(ε), hIAPP(δ), and hIAPP(p) systems are limited to the N-terminus (K1-C2) and C-terminus (T36-Y37). This indicates that no specific residue contacts are formed regardless of whether histidine is in the (ε), (δ), or (p) state. Furthermore, the arrangement of residues around the carbon backbone further supports the conclusion that these three systems tend to form helical structures.

Figure4.

The contact map properties in hIAPP(ε) S20G, hIAPP(δ) S20G, and hIAPP(p) S20G systems.

In current study, we investigated the structural properties and aggregation characteristics of hIAPP S20G monomers to better understand the combination impact of S20G and histidine behaviors, including three histidine states: hIAPP(ε), hIAPP(δ), and hIAPP(p). Three independent REMD simulations were performed. In the analysis of secondary structure content, hIAPP(δ) exhibited the highest average β-sheet content. Small amounts of β-bridge and β-sheet structures were observed in both the hIAPP(δ) and hIAPP(p) systems, while no β-sheet structures were found in the hIAPP(ε) system. Conversely, the average α-helix content in all three systems was significantly higher than their average β-sheet content, with similar regional α-helix content observed across the systems (A5-L12, S20-L27, and T30-V32). Notably, the α-helix content in the L12-V17 region was higher in the hIAPP(p) system compared to the hIAPP(ε) and hIAPP(δ) systems. Interestingly, we found that the S20G mutation more effectively induced β-sheet formation in the hIAPP(δ) system, facilitating its aggregation compared to the unmutated form. Cluster analysis results were consistent with the secondary structure data. The dominant conformations in all three systems were helical structures, with the hIAPP(p) system showing the highest probability of favorable conformations and exhibiting small amounts of β-bridge and β-sheet structures. The hIAPP(p) system was also found to be more stable compared to the hIAPP(ε) and hIAPP(δ) systems. The presence of N14/H18 H-bonding interactions in all three systems underscores the critical role in hIAPP S20G monomer. Further analysis confirmed that the residues in all three histidine states are arranged around the carbon backbone, supporting the tendency of these systems to form helical structures. Our current study provides new insights into the conformational changes of hIAPP S20G monomers, which is helpful to understand the aggregation phenomena on T2D.