1
Introduction

1.1 Protein-Protein Interactions and Their Small-Molecule Modulators

1.1.1 Characteristics of Protein-Protein Interactions

Proteins that work and degrade in highly congested and complex environments must be found by their partners in a large number of non-partners. It is estimated that human beings have 650?000 different pairs of interactions, which are responsible for a number of key biomolecular processes [1]. The surface of soluble proteins is covered by hydrophobic and hydrophilic residues, as well as by hydrophilic backbone. The highly specific physical contact between two or more protein molecules is mainly related to hydrophobic interactions, salt, and hydrogen bonds.

Protein-protein interactions have different affinity and longevity. Some complexes are weakly and instantaneously clustered; some may continue to form part of a larger protein complex, stabilized through multiple interactions; some reversible signal complexes have high pairing affinity, but only limited time; some complexes are stable, but have built-in timers; the presence of antibodies and antigens and protease and inhibitor complexes can take up to a day, some of which may be categorized as irreversible [2].

In addition, protein-protein interactions can be categorized according to the structural characteristics (Figure 1.1) [3]: the interaction between globular protein pairs, the interactions between globular proteins and individual peptide chains with continuous or discontinuous table position, and the interaction between two segments of peptide chains. Correspondingly, the polypeptide that participates in protein-protein interactions may adopt a combination of structures: the extension structure in the groove, ß-sheet, a-helix, and even the poly-proline helix.

There is certain regularity in the presence of amino acid residues in proteins [3a]. In the general interface, leucine is the most common residue, followed by arginine. Furthermore, charged residues are more common than polar residues, and both, except for arginine and histidine, are generally abundant on the surface. Aromatic amino acids, except for tryptophan, have a very low abundance on the surface but have a high abundance at the interface. As is mentioned above, the frequency of occurrence of hydrophobic residues is generally high at the interface and is low on the surface. Cysteine is particularly rare both on the surface and at the interface.

Figure 1.1 Classification of protein-protein interactions and examples [3b].

Source: Scott et al. [3b]. © 2016, Springer Nature.

In addition, based on the results of alanine scanning mutagenesis, the residual base that has a great influence on the binding affinity is called "hot spot" [4]. Hot spots are almost always buried in the center of the core, not in contact with solvents. The hot spot processes the highest sequence conservation [5]. Tryptophan, arginine, and tyrosine are the most common, accounting for more than half of the total, as hot spots. These three versatile residues were able to form hydrophobic, aromatic, and polar interactions, all of which can be wrapped in complementary surfaces to meet unpaired hydrogen-bonded donors and receptors. In addition, the polar "p-cation" bond between arginine and tryptophan or tyrosine was found in more than 50% hot spots [6]. Apart from a "p-cation" bond with arginine, the traditional side chain interaction is more common for tyrosine. By contrast, the most common residual at the interface, leucine, is rarely found in hot spots, while isoleucine is rich.

In the complexes in the protein database, 62% has a helix on the interface [7]. However, the presence of a helix at the interface does not mean that the helix plays a key role. Analysis shows that in about 60% of the interface, the hot residue is located on one side of the helix, one-third of the complexes with the hot spots on two faces of the helix, and about 10% of the complex with all three faces participating in the interaction with the target protein. In the protein database, the first four major types of function of protein-protein interactions, where helices are involved, are gene regulation, enzyme function, cell cycle, and signal transduction.

Analysis of the contribution of each helix residue to the interaction shows that leucine appears most in the interface area. This is not surprising, because in general, leucine is also the most common residue in proteins. After the normalization of natural abundance, aromatic amino acids, arginine, and leucine are of the highest frequencies at the helix interface as compared with polar residues [4, 8]. In addition, polar and charged residues are also important contributors to the interface.

1.1.2 Intervention of Protein-Protein Interactions Using Small Molecules

Abnormal protein-protein interactions are the basis of multiple diseases, and an increasing number of researchers are committed to developing molecules to modulate protein interactions for therapeutic purposes. Small molecule is a class of entity with potentially ideal therapeutic potentials. However, the contact surface of some of the protein interactions is large and shallow (about 1000-6000?Å2), especially those featured by a linear peptide epitope 1-4 amino acids long, compared to the traditionally small and deep small-molecule binding pockets [9] (Figure 1.2). Therefore, the interface between proteins is sometimes regarded as a target of "undruggable." In establishing guidelines for the discovery of protein-protein interaction (PPI) inhibitors, clinical success cases should be considered in the context of the type of interface.

Work in recent years has begun to show that some protein-protein interactions are able to be suppressed by small molecules. Most of the developed inhibitors target PPIs, where hot spot residues are restricted to small binding pockets (250-900?Å2) [11]. Some small-molecule inhibitors disrupt the interaction between a globular protein and a single peptide chain with a secondary or tertiary structure, through binding to the pocket on the globular protein. It is noteworthy that the secondary structural features processed by the peptide chain, such as a-helices and ß-strands, have important implications for the design of inhibitors that mimic and replace these peptides. With a better understanding of the structural biology of the protein-protein interactions, it seems more promising and reasonable to discover drugs targeting protein-protein interactions with defined structures. In addition, the hot spots of the interaction interface can be targeted by inhibitors of protein-protein interactions. The interaction of the rigid globular protein with a polypeptide may be more suitable for small-molecule interruption because the polypeptide can contribute more to the binding energy and be replaced by the small molecule with good design. At present, there are many strategies to discover hits or leads that interfere with protein-protein interactions, the most notable of which is high-throughput screening, fragment screening, and optimization.

Figure 1.2 The complexity of the PPI interface affects druggability PPIs can be classified by whether one side of the interface consists of a primary (linear) protein sequence (green), a single region of secondary structure (such as an a-helix, yellow), or multiple sequences requiring tertiary structure (red). There are fewer examples of small-molecule inhibitors of PPIs as the interface becomes more complex (from primary to secondary to tertiary epitopes). Structures shown are BRDt/histone (green; Protein Data Bank [PDB]: 2WP1), MDM2/p53 (yellow; PDB: 1YCR), and IL-2/IL-2Ra (red; PDB: 1Z92) [10].

Source: Arkin et al. [10]. © 2014, Elsevier.

High-throughput screening is an effective way to find a hit in a traditional drug target. Most of the high-throughput screening strategies rely on assays such as fluorescence resonance energy transfer, amplified luminescent proximity homogeneous assay screen, surface plasmon resonance, or fluorescence polarization because they are highly efficient, sensitive, and reagent-available [12]. However, these methods can usually disrupt enzyme activity and lead to more false-positive signals. Another method is based on the label-free strategy, including the refractive index properties and mass spectrometry [12b]. Their applications may be more extensive, more quickly developed, and robust because they eliminate the steps associated with introducing and observing tags. Despite these established methods, it is still difficult to effectively generate protein-protein interaction inhibitors through high-throughput screening since the compounds used for screening are mainly targeting traditional drug targets. Traditional high-throughput screening faces some challenges in dealing with protein-protein interactions - low hit ratio, low activity, and hard to eliminate false positives [12b]. However, high-throughput screening has been successfully applied in the discovery of the analog of discontinuous epitope on an a-helix.

The fragment-based drug discovery is a strategy to discover molecules from smaller fragment of drugs or functional groups with low affinity, which can...