In its crystallized form, ADAM10 is a cyclically-symmetrical heterohexamer of three unique subunits. Each cyclically-symmetrical half consists of one of two subunits (A and D) of ADAM10 itself and two of four subunits (B, C, E, and F) representing the heavy and light chains of the mAb 8C7 antibody required for the crystallization procedure (3). ADAM10 is derived from Bos taurus, has a molecular weight of 59,247.56 Da, and a theoretical isoelectric point of 8.15, implying the relative alkalinity of this protein (4). However, within the entire crystallized complex with the mAb 8C7 antibody, the molecular weight becomes 141029.31 Da, with an isoelectric point of approximately 6.4, which is substantially more acidic than the ADAM10 protein on its own. At a secondary structural level, the identical A and D subunits of the crystallized complex, which are both ADAM10, are constituted of approximately 7% helical structure, with 4 helices in the subunit that involve 15 residues, and of approximately 17% beta-sheet behavior, consisting of 21 strands involving 36 residues, implying that the vast majority of the structure is comprised of random coil behavior (3). The light and heavy chain subunits of mAb 8C7 have comparable proportions of helical structure, with 7% and 5% composition, respectively, but these subunits possess substantially more beta-sheet structures, at respective levels of 50% and 51%. Lastly, because ADAM10 needs to interact with both the cell membrane and intracellular components, it needs to possess both nonpolar and polar segments, which it does. 

     The Bos taurus-derived ADAM10 protein is not known to have any particularly notable alternate confirmations, nor any specific binding behavior with any pharmaceuticals or drug therapies. Also, due to its relatively small size and lack of abundance of well-defined tertiary structure, hydrophobic interactions and hydrogen bonding interactions between residues are also not highly prevalent. However, two key intramolecular interactions exist in ADAM10, both of which are critical to the function of the protein. In the complex with mAb 8C7, ADAM10 has three glutamic acid residues at positions 573, 578, and 579, and these three residues constitute a negatively-charged binding pocket within ADAM10 that serves two particularly important functions. The first of which is that this pocket binds with mAb 8C7 for crystallization, and the second is that the negative charge of this pocket is crucial for binding to transmembrane Eph/Ephrin complexes, which are members of another subfamily of receptor tyrosine kinases, like the EGF receptor (5). This negatively-charged domain is stabilized via interactions with two different cysteine-rich regions of the proteins, which have been shown to form intramolecular disulfide bonds. These disulfide bridges in cysteine-rich domains span residues Cys-594 to Cys-639 and Cys-632 to Cys-645, demonstrating a pseudo-interlinkage of these two bridges. These regions were initially thought to be the binding site for antibodies like mAb 8C7, due to their sheltering of the negatively-charged pocket described above, but instead it was found that they stabilize this substrate-binding pocket to optimize its interactions with transmembrane receptor proteins (6).

     ADAM10 is not rife with associated compounds and ions, but there are three molecules in particular that have been shown to exist in crystallized forms of ADAM10. Magnesium and sulfate ions have been observed in the crystallized structure, but sulfate is not shown to have any specific biological significance for ADAM10 function; in fact, sulfate anions are used for crystallization itself (6). This lack of necessity for sulfate anion is better highlighted by comparison of the ligands of ADAM10 to proteins of highly similar structure. Zinc metalloproteinase-disintegrin bothropasin (PBD ID: 3DSL) from Bothrops jararaca and atragin (PBD ID: 3K7L) from Naja atra are two proteins with a high degree of similarity to ADAM10, both structurally and in sequence. Through the use of PSI-BLAST, a program used to find proteins of similar primary structure as that of a query protein by assigning "gaps" in the comparison sequences, and the Dali server, which is a method for finding proteins that possess comparable tertiary structures to those of an input protein by a "sums of pairs" method comparing intramolecular differences, the degree of similarity between these two proteins and ADAM 10 were quantified; for PSI-BLAST results, lower E-values imply a higher degree of commonality, but for Dali server results, higher Z-scores imply higher similarity. For reference, an E-value of less than 0.05 is the threshold for significance for PSI-BLAST results, and the threshold for Dali server results is a Z-score above 2. When compared to ADAM10, zinc metalloproteinase-disintegrin bothropasin had an E-value of 9 x 10-11 and a Z-score of 10.0, and atragin had an E-value of 2 x 10-9 and a Z-score of 11.7, showing incredibly significant degrees of commonality between both these proteins and ADAM 10 (7, 8). Like ADAM10, these two proteins possess both a cation ligand (Mg2+ in ADAM10, Ca2+ in atragin and zinc metalloproteinase-disintegrin bothropasin) and N-acetyl-D-glucosamine (NAG), demonstrating that the similar structures of the proteins attract similar ligands. While the importance of NAG as a prosthetic group is more ambiguous and requires more research, the metal cations of all three proteins are known to be directly related to the catalytic function of the protein itself. (6). 

     One notable aspect of these comparison proteins is that both of them have much higher concentrations of helical structures than does ADAM10, with 23% abundance for zinc metalloproteinase-disintegrin bothropasin and 26% abundance for atragin (3). However, these two comparison proteins are also shown to possess both disintegrin or distintegrin-like domains, as well as cysteine-rich domains, much like those located in ADAM10 (9, 10). Given that these proteins also function as metalloproteases, it is clear that metalloprotease function is less dependent upon specific tertiary structure and more dependent upon the existence of these cysteine-rich regions that can interact intramolecularly. That said, atragin functions in almost the opposite way as ADAM10 in biological systems. Atragin is a snake venom and functions to limit cell migration, which is why it has a conserved cysteine-rich domain, so that it can have a protected substrate-binding site needed to interact with the membrane-bound proteins associated with cell migration (9). However, the negatively-charged binding pocket of ADAM10 is what allows it to bind to Eph/Ephrin complexes, which are critical for tasks such as cellular migration (5). Moreover, seeing as the targets for these metalloproteases are generally transmembrane proteins, existence of both alpha-helical and beta-sheet structures are necessary to allow the protein to traverse the membrane as needed, but an overabundance is not needed, as the catalytic domains of these proteins is much more contingent upon charged interactions between receptor proteins and the metalloproteases. For example, atragin and zinc metalloproteinase-disintegrin bothropasin specifically contain catalytic zinc cations that have to be stabilized by negatively-charged binding pockets, much like that found in ADAM10 (9, 10). 

     In summation, the major takeaway of the structure of ADAM10 is that the protein?s function is more dependent upon cysteine-rich domains than specific tertiary structures. However, this is not to say that ADAM10?s functionality is solely dependent upon these cysteine residues, as the negatively-charged pocket created by residues Glu-573, Glu-578, and Glu-579 is also immensely important, and mutations to this region as well would cause a loss of acute functionality.