The allosteric properties of hexameric insulin have been used in the formulation of pharmaceutical preparations to treat diabetes mellitus. Historically, this occurred through the use of anti-microbial agents (phenol, m- cresol, and methylparaben) and anions like chloride as isotonic agents. These compounds have been shown to influence the protein more than what was thought initially. Through structural and functional studies (X-ray crystallography and various solution studies) these interactions and their mechanisms have been further characterized. 
The T to R allosteric transition involves the interconversion of three allosteric states of the protein, the T6, T3R3, and R6. On going from the T-state to the R-state, residues B1-8 form a more extended a-helical conformation. This displaces PheB1 approximately 30 Å which exposes an amphipathic pocket at the dimer-dimer interface and further creates a narrow tunnel (~12 Å) down from the surface of the hexamer to the HisB10 residues (these are the metal binding sites that bind anion). Furthermore, this allosteric transition also converts the coordination geometry of the HisB10 metal sites from octahedral to tetrahedral coordination geometry. The allosteric transition of the insulin hexamer can be quantitatively described by the Seydoux, Malhotra, and Bernhard (SMB) allosteric model for half-site reactivity and cooperativity.

To view the insulin hexamer with 4H3N bound to the HisB10 sites, click on the image to the left.  To view this structure, however, you must have the free Chime plugin.  Chime runs in both Internet Explorer and Netscape.

The HisB10 metal-binding sites of the Co(II) substituted R3 units of the T3R3 and R6 hexamers show relatively high affinity for monovalent anions and carboxylate ions. In fact, certain carboxylates have an affinity for the hexamer that is sufficient to stabilize the R-state complex in the absence of phenolic ligands. One carboxylate in particular, 4-Hydroxy-3-Nitrobenzoic acid (4H3N) has been shown to bind the Co(II) substituted hexamer with a relatively high affinity (KDapp. " 0.08 mM ). Since carboxylates like 4H3N combine coordination to the HisB10 metal ion and weak polar and nonpolar interactions with protein surface, manipulation of the structure of 4H3N in such a way as to maximize these ligand-protein interactions will be effective in stabilizing the R-hexamer. The design of a more stable R-hexamer through rational drug design could lead to the elimination of toxic phenolic ligands in insulin preparations and the development of improved insulin formulations.