β-Secretase Complex with Oxazine Inhibitor (PDB ID: 4J0T) from Homo sapiens

Created by: Megan Ha 

         β-Secretase (PDB ID: 2ZHV, BACE1) is an aspartic protease found in Homo sapiens. This protein is made up of an extracellular domain containing a signal peptide, a catalytic domain, a trans-membrane domain, and a short cytoplasmic stretch (1). BACE1 is highly expressed in the brain and has been implicated in Alzheimer’s Disease (AD) as a possible target for drug inhibition (1). Alzheimer’s Disease is marked by the aggregation of amyloid-β peptides, believed to be one of its primary causes. BACE1 generates these peptides by cleaving the β-amyloid precursor protein (APP) (2). This suggests that an effective therapy for AD would be through inhibition of the protease. Therefore, many studies analyze the efficacy of oxazines as treatment, particularly 5-Ethoxy-pyridine-2-carboxylic acid [3-((R)-2-amino-5,5-difluoro-4-methyl-5,6-dihydro-4H-[1,3]oxazin-4-yl)-4-fluoro-phenyl]-amide in complex with BACE1 (PBD ID: 4J0T).

         The crystal structure of the BACE1 complex with an oxazine inhibitor was obtained using vapor diffusion sitting drop, and the data was obtained using X-ray diffraction (3). The primary structure of the protein is composed of 501 residues, and the segments that were crystallized were residues 57-218, 228-306, and 446-453. There is a single mutation: Lys-307 to Ala-307. The protein’s theoretical molecular weight of 45514.26 Da and isoelectric point of 4.83 was found using Expasy, a bioinformatics resource portal (4). This indicates that optimal protein function occurs in an acidic environment, validated by studies that the preferred pH range is from 4.5-5.3 (2).

The secondary structure of BACE1 is 13% alpha helices, 38% beta strands, and remaining random coils, 3/10-helices, beta bends, beta bridges, and beta turns (3). The majority of alpha helices exist in the trans-membrane region, where the internal H-bonding accomplishes a lower thermodynamic cost in the hydrophobic bilayers of the cell.

Classified as a type-1 membrane protein, BACE1 consists of only one subunit, responsible for binding with the substrate at the active site as well as cleaving at specific aspartate residues (1). The globular protein has three binding pockets, known as S1, S2, and S3 (1). The region of the active site contains two conserved aspartic acid residues: Asp-93 and Asp 289, located between N- and C-terminal lobes of the protein within the binding cleft. A flexible antiparallel hairpin loop in the ectodomain called the “flap” covers the central recognition motif in the ligand-binding site, consisting of residues Tyr-129 to Glu-138 of the sequence (1). The different flap conformations of BACE1 can be seen by the apo structure (PBD ID: 1W50) and the bound structure. In the “open” position, the hairpin loop is further from the binding site; in the “closed” position, it is closer (1). The flap plays a large role in ligand recognition process. When bound and active, the flap becomes stabilized around the substrate (5). The hydrogen bond between the phenolic ring of the conserved residue Tyr-132 with the indole nitrogen of the residue Trp-137 on the protease contributes to this stabilization (6). The catalytic aspartate residues are co-planar and coordinate a single water molecule (Wat1) that initiates the nucleophilic attack on the peptide carbonyl group once the substrate is bound (7). The peptide bond is hydrolyzed, resulting in a proton transfer from the aspartate to the amino group of the substrate (6). It is also important to note that other hydrogen bonding exists to stabilize the closed conformation of the protein, ultimately influencing the shape and flexibility of the active site (7).

Studies have shown that all BACE1 inhibitors interact with the aspartic acid residues through hydrogen bonds (8). It has been observed that the majority of water molecules present in the active site of the protein prefer to act as hydrogen bond acceptors with the ligand. Many regions of BACE1 are polar and favorably interact with the polar groups of ligands. In the BACE1 complex with an oxadiazoyl tertiary carbinamine inhibitor (PBD ID: 2IS0), Gln-134 is an active site residue located in the flap region that forms H-bonds with the ligand (8). In the BACE1 complex with a carboxamide inhibitor (PBD ID: 2VKM), hydrophobic amino acids in the substrate pockets form backbone-mediated hydrogen bonds with ligands, like Thr-293, creating water bridge H-bonds (8). The side chains of the amino acids in these BACE1 complexes are more capable of forming ligand interactions than the backbone, due to the rigidity of planar peptide units (8). Other residues that have been found as crucial for the activity of the protease are due to disulfide bond formation between the thiol groups of the amino acids Cys-277 and Cys-481, Cys-339 and Cys-504, and Cys-391 and Cys-441 (2).

Oxazine compounds are currently being developed as possible inhibitors for the protease. 5-Ethoxy-pyridine-2-carboxylic acid [3-((R)-2-amino-5,5-difluoro-4-methyl-5,6-dihydro-4H-[1,3]oxazin-4-yl)-4-fluoro-phenyl]-amide is such a compound that aims to act as a competitive inhibitor by forming a complex with BACE 1 (PBD ID: 4J0T). The nitrogen atom in the ortho position of the oxazine ring directs the aromatic structure into the S3 pocket, displacing bound water molecules (1). There is a network of van der Waal interactions between the fluorine atoms and residues in the active site, and an H-bond forms between the amide of the ligand and the backbone carbonyl oxygen of Gly-291 (1). Most importantly, the protonated amidine present in the oxazine head group forms tight hydrogen bonds with the two catalytic aspartates (1).

To further understand the dynamics of BACE1, comparisons with other proteins were done. PSI-BLAST is a program that compares a given protein sequence with those in protein databases. It produces an algorithm to make gapped alignments between primary structures, ultimately creating a position-specific score matrix (9). This substitution matrix compares the segments using a score called the E-value to evaluate similarity. The E-value must be lower than the threshold of 0.05, or it is considered insignificant. Gaps, the amino acids present in the given protein but absent in the query protein when alignments are made, would increase the E value. Ultimately, β-Secretase 2 (PBD ID: 3ZKM, BACE2) was found to have a high sequence homology with BACE1 and have an E value of 3e-138 (10). The use of the Dali server further confirmed this similarity. The Dali Server compares the tertiary structures of proteins and calculates the differences in intramolecular distances. It uses a sum-of-pairs method to analyze structural similarity, ultimately evaluating homology with a Z-score that is considered significant if above 2. The Z-score of BACE2 was 52.0 (11). BACE2 is a homolog of BACE1 and is also an aspartyl protease. While it also cleaves APP, BACE2 prefers a position downstream of the BACE1 cleavage site (Asp-672), specifically Phe-691 (1). BACE2 cleavage does not produce amyloid-β peptides (1). There is a flexible region called the “10s loop” common to both structures. This loop is from residues Gly-69 to Gly-74 in BACE1, making up a portion of the S3 binding pocket (1). While in BACE1, the loop has an up or down conformation, in BACE2, the loop is always in a down conformation. Additionally, the S3 pocket of BACE2 is smaller, causing larger conformational strain and reduced effectiveness of inhibitors (1). Evidence has shown BACE2 can compensate for BACE1 loss-of-function (12). Therefore, current studies on Alzheimer’s disease treatment focus on analyzing the differences between the β-secretase structures to specifically inhibit BACE1.

         Researchers are currently debating the targets of BACE1. BACE1 has a relatively weak affinity for APP, suggesting that APP is not its primary endogenous substrate (2). Low-density lipoprotein receptor related proteins (LRPs) have been found as important substrates that interact with BACE1 through their C-terminal domains. One important note is that LRP may have an effect on the ability of BACE1 to interact with APP, as LRP acts as a scaffold (2). Other substrates of BACE1 include the sodium-gated channels, ST6Gall and PSGL-1 (2). While there has been evidence that BACE1 plays a role in axonal growth, brain development, and inflammation, the majority of current studies with the protein continue to focus on BACE1 inhibition as a possible treatment for Alzheimer’s disease (2). The primary hypothesis for the role of BACE1 in AD is called the amyloid hypothesis, which hypothesizes the amyloid-β peptide as the beginning of a cascade that ultimately leads to amyloid plaques, causing neurofibrillary tangle formation, neuronal problems, and finally dementia (12). Ultimately, more research needs to be done on the dynamics and structure of β-secretase to produce effective treatments for Alzheimer’s disease.