in a right-handed superhelix resulting in a crescent shaped subunit that envelops part of the P-subunit. This unusual structure has been seen previously in the crystal structures of lipovitellin-phosvitin and bacterial muramidase (10,11). Twelve a helices of the P-subunit are folded into an a-a barrel (Fig. 2), similar to those found in bacterial cellulase, endoglucanase CelA, and glycoamylase (12-15). Six parallel helices (3P, 5P, 7P, 9P, 11P, and 13P) form the core of the barrel. Six additional helices (2P, 4P, 6P, 8P, 10P, and 12P) form the outside of the barrel. These peripheral helices are parallel to each other but anti-parallel to the core helices. One end of the barrel is blocked off by residues 399P to 402P. The opposite end is open to the solvent, forming a deep funnel-shaped cavity in the center of the barrel. This cavity is hydrophobic in nature and is lined with conserved aromatic residues (Fig. 2). The enzymatic active sites of other a-a barrel proteins are located in such cavities.
The N-terminal proline-rich domain of the a-subunit (residues 1-54) is disordered in the crystal structure. Deletion ofthis domain does not affect the catalytic activity ofFTase (16). Taken together, these observations suggest that the proline-rich domain may interact with other factors in the cell, perhaps serving a role in enzyme localization. The crystal structure of a truncated form of rat FTase that lacks the proline-rich domain has been determined to 2.75A resolution (17). Not surprisingly, the deletion of this domain had no significant effect on the structure of the rest of the protein.
Multiple-sequence alignments ofmammalian and yeast a-subunits reveal five tandem sequence repeats (18). Each repeat consists of two highly conserved regions separated by a divergent region of fixed length. These motifs appear in the first five "helical hairpins" of the FTase structure (Fig. 1). The second a helix of each helical pair contains an invariant Trp residue which, together with other hydrophobic residues, forms the hydrophobic core of the hairpin. The conserved sequence motif Pro-X-Asn-Tyr (where X is any amino acid) (18) is found in the turns connecting two helices of the coiled-coil. These turns form part of the interface with the P-subunit. Internal repeats of glycine-rich sequences also have been identified in the P-subunits of other protein prenyltransferases (18). These repeats correspond to the loop regions that connect the C-termini of the peripheral helices with the N-termini of the core helices in the barrel.
Fig. 2. The a-a barrel of the P-subunit, showing the aromatic resides (yellow) that line the interior creating a hydrophobic cavity. This view is a 90° clockwise rotation relative to Fig. 1A. The zinc ion is shown as a magenta sphere. Only helices 2P to 13P are shown.
Fig. 3. Zinc binding site. The zinc ion is coordinated by side-chains of Asp297P, Cys299P, His362P, and a well-ordered water molecule. Carbon is shown in coral; oxygen in red; nitrogen in blue; sulfur in green; zinc in magenta; polypeptide chain in cyan. H-bonds are indicated in black.
Fig. 4. Stereoview of the active site with bound FPP. Portions of the a- and P-subunits are drawn as ribbons in red and blue, respectively. The isoprenoid moiety of the FPP molecule binds to the hydrophobic cavity inside the a-a barrel of the P-subunit. Residues colored in green line this hydrophobic cavity. The diphosphate group of the FPP molecule binds in a positively charged cleft near the subunit interface formed by residues colored in pink, and is adjacent to the catalytic zinc ion whose ligands are colored in gray. Confirming the observed location of the FPP molecule, mutation of P-subunit residues His 248P, Arg 291P, Lys 294P, Tyr 300P, or Trp 303P results in anincrease inKd(FPP) (27). Lys 164a—which, whenmutated toAsndramaticallyreduces catalytic turnover (16)—is in close proximity to the C1 atom of the FPP molecule and may be directly involved in catalysis.
The FTase subunit interface is quite extensive, burying3322 A2 or 19.5% ofaccessible surface area of the a-subunit and 3220 A2 or 17.2% of accessible surface area of the P-subunit (1). Although the size ofthis subunit interface is typical for an oligomeric protein, there are nearly double the normal number of hydrogen bonds. The number of hydrogen bonds found in the FTase subunit interface may explain the unusual stability ofthe FTase heterodimer, which cannot be dissociated unless denatured (6).
A striking feature of the FTase structure is a deep, funnel-shaped cleft formed by the central cavity of the a-a barrel. The cleft is lined with highly conserved hydrophobic residues (Fig. 2). A second cleft runs parallel to the rim ofthe a-a barrel and is hydrophil-lic in nature. These two clefts intersect at the site ofa zinc ion that is required for catalysis.
FTase is a zinc metalloenzyme that contains one zinc atom per protein dimer (19,20). Experimental evidence indicates that the zinc ion is required for catalytic activity and important forthe binding ofpeptide, butnot isoprenoid, substrates (19). A direct involvement ofzinc in catalysis is supportedby several studies (19,21-23). Themost compelling is that the zinc ion coordinates the thiol of the CAAX cysteine residue in the ternary complex (24). In the FTase crystal structure, there is a single zinc ion bound to the P-sub-unit, near the subunit interface, that marks the location of the active site. The zinc is coordinated by P-subunit residues D297P, C299P, H362P, and a well-ordered water molecule (Fig. 3). The cysteine thiol of the CAAX protein substrate coordinates the zinc ion, displacing this water molecule in a ternary enzyme complex (3,4). D297P forms a bidentate ligand, resulting in a distorted penta-coordinate geometry. All three protein ligands are conserved in the P-subunits ofprotein prenyltransferase enzymes. C299P had previously been identified from mutational analysis and biochemical studies to affect zinc binding and abolish catalytic activity (25).
Several site-directed mutagenesis studies confirm that D297P, C299P and H362P are ligands for the zinc ion (26-28), and additionally suggest that D359P has a role in zinc binding (27). In the crystal structure, D359P forms a hydrogen bond to H362P, possibly stabilizing a required conformation for binding zinc.
Was this article helpful?