Lipoxygenases (LOXs) are dioxygenases that play a key role in the metabolism of polyunsaturated fatty acids (PUFAs)—among which is arachidonic (eicosatetraenoic) acid—in a large variety of living cells [1] ranging from microorganisms [2] to plants [3] and mammals [4]. Such a ubiquitous distribution is suggestive of relevant biological roles conferred to LOXs by natural evolution ^[5][6][5,6]. The reaction catalyzed by LOXs is the oxygenation of PUFAs, i.e., the insertion of molecular oxygen (O₂) in their acyl chain with the generation of hydroperoxyl (HOO-) groups at different positions. Insertion of O₂ is specific for each LOX isoform that is indeed named with the number of the carbon atom where O₂ has been bound. LOXs products initiate crucial biosynthetic pathways in living organisms. In mammals, 5- and 15-LOXs induce the synthesis of important signaling molecules ^[1][7][1,7], such as leukotrienes (5-LOX) and lipoxins (15-LOX), that play a crucial role in inflammation and immunity [8]. They are also involved in the development of pathological atherosclerosis [9], as a direct consequence of their ability to bind low density lipoproteins [10]. Plant lipoxygenases (8-, 13-LOXs) are instead involved in germination and growth, as well as in pathogen resistance [3]. Additionally, some LOXs can modify the structure of lipid bilayers to reach specific metabolic goals: for example, in plants, oxidation of membranes by LOXs drives leaf senescence or lipid mobilization during the germination phase [3]; in animals, 15-LOX of reticulocytes attack the mitochondria envelope and promote the maturation of red ^[11][12][11,12].

The first biochemical characterization of a (plant) LOX dates back to the 1970s ^[13][14][13,14] but, despite the widespread presence of LOXs in the seeds of legumes [15], the first tridimensional structure of a soybean 15-LOX (also known as LOX-1) was obtained only 20 years later ^[16][17][16,17]. The main 3D features of soybean 15-LOX are: (i) the existence of two rather distinct structural domains, and (ii) the presence of a non-heme catalytic iron located in the C-terminal domain of the polypeptide chain ^[16][17][16,17]. For several years, only soybean 15-LOX 3D structure was available, and thus was used as a template to model human 5-, 12-, and 15-LOXs [18]. A more reliable model for mammalian (and in particular human) LOXs was possible when rabbit reticulocyte 15-LOX was crystallized [19] and its preliminary structure was determined [20] and then refined [21]. The further characterization of human ^[22][23][22,23], porcine [24], and coral [25] LOXs has provided evidence that a common 3D architecture does exist in animal LOXs.

Starting in the 1970s, the attention on LOXs has increased considerably, as it appears from the number of papers published since then (Figure 1A). Human 5-LOX and 15-LOX are clearly the most studied members (Figure 1A) due to their major impact on health and disease conditions ^{[26][27][28][29][30]}[26,27,28,29,30]. Of note, only a small number of studies have interrogated the interaction of LOXs with bio-membranes (Figure 1B), despite two fundamental questions arising from this event: (i) How do soluble enzymes (such as LOXs) search and find their substrates in a rather peculiar environment like lipid bilayers? (ii) How does the interaction with bio-membranes modulate the activity of LOXs? The first issue is obviously not specific for LOXs, yet these enzymes may represent a paradigmatic example to shed light on other membrane-interacting proteins [31].

The second issue seems to be of great relevance for LOXs, especially for human 5-LOX, whose activity leads to the synthesis of bioactive compounds—leukotrienes—from arachidonic acid ^[26][29][26,29]. Human 5-LOX is, indeed, the only member of the LOXs family that is present both in the cytoplasm and in the nucleus of a cell and that is able to bind different types of membranes (plasma membranes, nuclear membranes) both directly or through the specific 5-LOX activating protein, FLAP [27].

2. Insights into the Architecture of LOXs

Despite the huge amount of data available on phylogenetics, biological activity, design and action of inhibitors, and in vivo localization of LOXs, these proteins remain quite elusive from the structural point of view. In fact, a limited number of 3D structures are as yet available because of crystallization problems due to unstable segments in the protein sequence. The first two characterized LOXs are a plant enzyme, namely soybean 15-LOX (the first ever to be crystallized), and its mammalian counterpart, 12/15-LOX from rabbit reticulocyte. These two proteins share a limited sequence homology (<24%, Figure 2) and have different molecular weights (94,000 and 77,000, respectively); nonetheless, they display the same 3D organization. In particular, X-ray measurements ^[16][20][16,20] revealed the presence of two different domains, namely an N-terminal β-barrel PLAT (Polycystin-1, Lipoxygenase, Alpha-Toxin) domain, and a larger C-terminal domain that is mainly characterized by α-helices (Figure 3 and Figure 4). The two domains are connected by a short flexible linker and play different functional roles: the C-terminal contains the catalytic site ^[16][20][16,20], while the N-terminal has regulatory functions and, for instance, influences the membrane binding ability of mammalian enzymes [32]. This general structural organization is also highly conserved in other human variants, such as 5-, 12-, and 15-LOX ^[22][23][33][22,23,33], and in coral 8-LOX [25], despite the degree of sequence identity with both plant and rabbit LOXs being on the average quite low (Figure 2). The composition of the two N- and C-terminal domains is significantly different in all the above-mentioned LOX structures. The smaller β-barrel N-terminal is rather dense, mainly containing tightly packed hydrophobic residues; the C-terminal core is instead characterized by large cavities suitable for oxygen transit and substrate ^{[34][35][36][37]}[35,36,37,38]. The two domains are separated by a large gap, yet they are not fully independent. In Figure 3 and Figure 4, the analysis of the domain–domain interface is reported in terms of “contact maps”, which indicate the closest points of contact between the two protein sections. Plant and mammal lipoxygenases share a similar architecture, in which a few (≈4) main groups of contacts characterize the interaction between the two domains (Figure 3 and Figure 4). However, the positions of such contacts along the polypeptide chain of soybean 15-LOX are not the same reported for mammalian enzymes or coral 8-LOX, due to the larger size of the plant enzyme. Indeed, except for soybean 15-LOX, the other 5 enzymes considered in this study exhibit the most relevant contacts in similar locations, namely at positions 15, 25, 60, and 100 in the N-terminal domain, and approximately at positions 160, 390, and 620 in the C-terminal domain (Figure 3 and Figure 4). A more detailed analysis of these regions reveals a high homology score among the five LOX polypeptide chains (Figure 5), suggesting that such amino acids might play an important role in the protein dynamics. In particular, it could be speculated that these regions communicate movements of the N-terminal to the C-terminal and vice versa, in analogy to the “hot spots” found in the contact networks of oligomeric protein inter-subunit surfaces ^[38][39]. Furthermore, it should be stressed that many of the conserved or semi-conserved residues are histidines and aromatic amino acids (Figure 5), whose large side chain fills the gap at the domain–domain interface.