G protein-coupled receptors (GPCRs) are regarded to take up more than one-fourth of marketed human medicines [1, 2]. Drugs targeting GPCRs account for the majority of the best-selling drugs and about 40% of all prescription pharmaceuticals in the marketplace [3].

Histamine receptors belong to one family of rhodopsin-like class A GPCRs, and four subtypes are named in chronological order as H1, H2, H3, and H4. Histamine exerts its effects through the activation of these four histamine receptors. Each type of histamine receptor reacts to subtype-specific ligands into an active or inactive form.

GPCRs are integral membrane proteins that consist of 7 transmembrane segments connected by 3 intracellular and 3 extracellular loops of variable length. The crystal structures of GPCRs with their binding ligands have revealed the features of the ligand binding pockets and extracellular loops [4-6]. A ligand of GPCR activates the receptor by changing the receptor structure to the active form.

The biogenic amine histamine [2-(1H-imidazol-5-yl) ethanamine] (Fig. 1) is produced by decarboxylation of L-histidine and acts as a chemical mediator and neurotransmitter in central and peripheral tissues. It acts as an important pharmacological modulator involved in the processes of allergy, inflammation, neurophysiology, and cancer [7-10].

The molecular understanding of ligand-receptor interactions of GPCR remains unclear and is still the subject of investigations. There are several theories explaining the ligand-receptor interaction mechanism, such as shape theory, binding theory, and vibration theory [11, 12]. The structures of histamine receptor ligands are so variable that we can not easily classify the pharmacological function of the ligand. To find any other characteristic in the molecular patterns between agonists and antagonists of histamine receptors, a computational approach to molecular vibration was carried out in an attempt to find a bit of the molecular interaction mechanisms.

The Tanimoto coefficients between histamine and other ligand molecules are shown in Table 2. Typically, a Tanimoto coefficient > 0.85 is considered highly similar, and a coefficient > 0.75 is considered similar for the purpose of clustering molecules that may have similar biological activity profiles [15]. The highest Tanimoto coefficient among agonists was 0.45 (imetit), and this is not high enough to be considered a molecule that has potential agonist efficacy. Moreover, the lowest value of Tanimoto coefficient among agonists was 0.08 (dimaprit and SKF91488) and is also a lower value as an antagonist. It seems that there is no related pattern between Tanimoto coefficients and the functional types of molecules in the case of histamine receptor agonists or antagonists.

To search for a novel characteristic for the classification of histamine receptor ligands, a kind of molecular calculation using agglomerative hierarchical clustering was adopted in this work. The result of the hierarchical clustering of the similarity matrix from CIMVF is shown in Fig. 2. As shown in the figure, eight agonists were located nearby (the part in the dotted circle), except impromidine, and all antagonists were clustered close to each other in the radial tree. We can tell the regional difference between agonists and antagonists in the tree and also find that the information from the molecular vibrational frequency may play a role in the classification of agonists/antagonists for histamine receptor as a possible molecular descriptor. For these methods, clustering with CIMVF shows the more proper result in the case of histamine receptor ligands. With a more concentrated study on the relationship between the molecular vibrational frequency and pharmacological function of a ligand, the vibrational spectrum of a molecule may shed light on the field of ligand-receptor interaction mechanisms.