HIV protease inhibitors and nuclear lamin processing: getting the right bells and whistles.

O ne of the most notable successes of rational pharmaceutical design has been the development of drugs that inhibit the protease cleaving the polyproteins of the HIV into their structural and catalytic protein components (1, 2). Several of these inhibitors are currently used as components of ‘‘highly active antiretroviral therapy’’ for treating HIV infection and AIDS (2, 3). Given the similar chemistry involved in all peptide bond hydrolysis reactions, as well as the different sequence contexts of the polyprotein cleavage sites, the ability of these drugs to specifically target the HIV protease and not the hundreds of other proteases required for human health—digestive enzymes, enzymes of the coagulation pathway, enzymes of the complement cascade, proteases involved in protein maturation and transport—is truly remarkable. However, in a recent issue of PNAS, Coffinier et al. (4) demonstrate that these drugs do inhibit at least one unrelated protease and that this inhibition could conceivably account for some of the side effects of these pharmaceuticals. They show that HIV protease inhibitors, including lopinavir, a first-line widely used member of the saquinavir family, can also inhibit ZMPSTE24, a distinct protease involved in the conversion of farnesylprelamin A to lamin A, a key structural component of the nuclear lamina (Fig. 1). In recent years, genetic defects in the conversion of farnesylprelamin A to lamin A in humans have been shown to cause severe progeroid disorders (e.g., restrictive dermopathy, Hutchinson–Gilford progeria syndrome) (6–8). These findings suggest that the inhibition of prelamin A processing by HIV protease inhibitors may result in some of the same disease phenotypes observed in the genetic prelamin Aprocessing disorders. The prelamin A translation product is modified by an intriguing series of enzymatic reactions that transiently form a short isoprenylated and methyl-esterified C-terminal tail that is subsequently clipped off to generate mature lamin A (8). These reactions include the Sisoprenylation of the C-terminal–CSIM sequence by protein farnesyltransferase, the cleavage of the peptide bond between the farnesylcysteine and serine residues, the methylation of the C-terminal carboxylate group of the farnesylcysteine residue, and finally the upstream cleavage of a peptide bond to release a 15-aa farnesylated C-terminal fragment. It is unclear why such a complicated path is taken; it would be possible to encode the same mature lamin A protein simply with an earlier stop codon. Presumably, the farnesylated and methylated tail of prelamin A assists in the proper targeting of lamin A to the nuclear lamina or functions in the assembly of the complexes of lamins A, C, B1, and B2 (6, 8). However, failure to remove the tail from prelamin A can compromise the integrity of the nuclear lamina, leading to ‘‘blebbing’’ and folds in the nuclear envelope (9). The structurally abnormal nuclear envelope may then send cells on a downward course leading to the multiple disease phenotypes characteristic of the progeroid aging syndromes. An additional puzzling feature of lamin biology remains to be explained: mice entirely lacking lamin A and prelamin A proteins are quite healthy and manifest only trivial structural abnormalities in the nuclear envelope (10). The link between HIV protease inhibitors and lamin function was first proposed by Caron et al. (11). One of the side effects of HIV protease inhibitors is a type of partial lipodystrophy characterized by a redistribution of adipose tissue resulting from loss of fat in the face, arms, and legs, and gain of fat in the trunk, particularly a characteristic ‘‘buffalo hump’’ in