The influenza pandemic of 1918 caused the deaths of many tens of millions of people worldwide. Later pandemics in 1957 and 1968 were only somewhat less catastrophic, causing around a million deaths. In more recent years, the emergence of the influenza variant H5N1 (“bird flu”) has highlighted the potential for future pandemics, and the need for effective therapeutics to treat influenza. The influenza virus is composed of a protein “envelope” surrounding a payload of viral RNA that encodes for around a dozen viral proteins. One of these proteins is the pH-activated proton channel M2 that we discussed earlier (Chapter 2, Box 2.1) and is inhibited by the anti-influenza drug amantadine. Amantadine and a close analog were the only options for treating flu prior to the approval in the late 1990s of the neuraminidase inhibitor oseltamivir (Tamiflu®), the topic of this case study (Figure 4.16).
Figure 4.16 Chemical structures of sialic acid, the neuraminidase inhibitors oseltamivir (in active and prodrug forms), and zanamivir. The active form of oseltamivir is shown in its chair-like conformation.
Hemagglutinin and neuraminidase comprise the main components of the viral envelope. Hemagglutinin is a glycoprotein that recognizes sialic acid groups displayed on host cells, and thus plays an important role in viral infection. Neuraminidase is an enzyme that removes sialic acid from the surface of the host cell and virus particle, thus facilitating egress of newly formed viral particles. Enzymatically speaking, neuraminidase is a glycoside hydrolase—an enzyme that cleaves the glycosidic C–O bonds between sialic acid and other sugars in glycoproteins. Neuraminidase inhibitors such as oseltamivir and zanamivir were conceived as transition-state analogs—structural mimics of the transition-state intermediate formed during glycan hydrolysis in the active site of neuraminidase.
The glycolysis reaction performed by neuraminidase involves breaking the glycosidic C–OR bond and forming a new C–OH bond (Figure 4.17). The reaction proceeds through a carbocation intermediate in which the positive charge is stabilized by the neighboring oxygen atom. This intermediate has significant double bond character, resulting in partial flattening of the six-membered ring, as shown. Neuraminidase accelerates the glycolysis reaction by binding to the flattened transition-state intermediate with greater affinity than either the substrate or product (both of which have chair-like conformations). For example, binding will be tighter when the carboxylate side chain lies in the same plane as the six-membered ring (as in the transition-state intermediate) and weaker when it is in an axial position (as in substrate and product). This is why in the drugs oseltamivir and zanamivir the carboxylate side chain is made to project from an sp2-hybridized carbon atom and thus lie in the plane of the ring. These drugs are designed to conformationally mimic the transition-state intermediate. Unlike the transition-state intermediate however, the drug molecules cannot undergo further reaction, and instead remain tightly bound within the active site, inhibiting the enzyme.
Comparing the structures of oseltamivir and zanamivir to sialic acid reveals some additional changes made by the medicinal chemists who developed these compounds. Both oseltamivir and zanamivir retain the carboxylate and N-acetyl (–NHAc) side chains of sialic acid, which were found to be optimal substituents at their respective positions. The hydroxyl substituent of sialic acid however was replaced with a primary amine or guanidine function in oseltamivir and zanamivir, respectively. These basic groups afford a stronger interaction with neuraminidase; the X-ray crystal structure of oseltamivir bound to neuraminidase reveals an ionic/hydrogen bonding interaction between the amine function and Glu119 and Asp151. A notable difference between the two drug structures is that zanamivir retains the glycerol (trihydroxypropyl) side chain of sialic acid whereas oseltamivir bears a much more hydrophobic 3-pentyloxy group at the same position. Interestingly, the pentyloxy group appears to interact with the hydrophobic π face of an Asp-Arg hydrogen bonding pair in neuraminidase. Recall that Asp-Arg pairs are known to stack on the hydrophobic face of Tyr and Phe side chains in protein structures (Chapter 2, Figure 2.11). A similar motif is apparently behind the ability of neuraminidase to recognize both hydrophilic and hydrophobic side chains on sialic acid -inspired inhibitors.
Figure 4.17 Partial structure of a sialic acid glycan with the glycosidic bond shown in blue. Neuraminidase accelerates glycosidic bond cleavage by binding to the transition-state intermediate with higher affinity than to the glycan substrate or glycolysis product. Drugs like oseltamivir were designed to mimic the transition-state intermediate and thereby inhibit the enzyme.