Human CaaX protease ZMPSTE24 expressed in yeast: Structure and inhibition by HIV protease inhibitors

Abstract

The function and localization of proteins and peptides containing C-terminal "CaaX" (Cys-aliphatic-aliphatic-anything) sequence motifs are modulated by post-translational attachment of isoprenyl groups to the cysteine sulfhydryl, followed by proteolytic cleavage of the aaX amino acids. The zinc metalloprotease ZMPSTE24 is one of two enzymes known to catalyze this cleavage. The only identified target of mammalian ZMPSTE24 is prelamin A, the precursor to the nuclear scaffold protein lamin A. ZMPSTE24 also cleaves prelamin A at a second site 15 residues upstream from the CaaX site. Mutations in ZMPSTE24 result in premature-aging diseases and inhibition of ZMPSTE24 activity has been reported to be an off-target effect of HIV protease inhibitors. We report here the expression (in yeast), purification, and crystallization of human ZMPSTE24 allowing determination of the structure to 2.0 Å resolution. Compared to previous lower resolution structures, the enhanced resolution provides: (1) a detailed view of the active site of ZMPSTE24, including water coordinating the catalytic zinc; (2) enhanced visualization of fenestrations providing access from the exterior to the interior cavity of the protein; (3) a view of the C-terminus extending away from the main body of the protein; (4) localization of ordered lipid and detergent molecules at internal and external surfaces and also projecting through fenestrations; (5) identification of water molecules associated with the surface of the internal cavity. We also used a fluorogenic assay of the activity of purified ZMPSTE24 to demonstrate that HIV protease inhibitors directly inhibit the human enzyme in a manner indicative of a competitive mechanism.

Keywords: CaaX protease; HIV protease inhibitor; Isoprenoid; X-ray crystallography; ZMPSTE24; enzyme inhibitor; membrane protein; metalloprotease; progeria; zinc.

Figure 1

Schematic of prelamin A processing. The C‐terminal sequence of prelamin A is indicated with processing steps. CaaX proteolysis and the final endoproteolytic step are catalyzed by ZMPSTE24, although the alternative protease RCE1 may also be capable of catalyzing the CaaX cleavage step.

Figure 2

Expression and purification of ZMPSTE24. A. Aligned size exclusion chromatogram following affinity purification and SDS polyacrylamide gel of relevant fractions. The fraction numbers refer to 0.8 mL fractions (column bed volume 120 mL). The first lane (His6‐3C) contained 2.5 μg of purified His6‐tagged rhinovirus 3C protease as a reference. B. Quantitation of purified ZMPSTE24 relative to BSA.

Figure 3

Overall structure of ZMPSTE24. The angles of view and coloring of secondary structure elements are as presented for yeast SmSte24p by Pryor et al.17 i.e. helix I, purple; helix 2, blue; helix 3, cyan; helix IV, green, helix V, yellow, helix Vi, orange; helix VII, red; cytosolic loop 5 domain, brown, C‐terminus, pink. a) Cartoon view. b) Surface representation. The zinc atom is shown in light green. Presumed lipid, sulfate, and polyethylene glycol or detergent molecules are shown as stick representations.

Figure 4

ZMPSTE24 active site. a) The positions of three residues in the closest proximity to the zinc are shown. Zinc is shown as an orange sphere. The water molecule in closest proximity to the zinc is shown as a small red ball. Bond angles and distances (Å) are indicated. b) Map of electron density (2F_o‐F_c, contoured at 1σ) serving as the basis for modeling two water molecules immediately adjacent to the zinc. c) and d) Two views of a comparison of the active sites of thermolysin (PDB:8TLN) and ZMPSTE24. Atoms from thermolysin are shown in cyan, atoms from ZMPSTE24 are shown in magenta. Labels indicate the position of the indicated residues in thermolysin/ZMPSTE24). Zinc is shown as a large sphere. Water molecules closes to the zinc are shown as small balls with appropriate coloring. The structures were aligned using the PDB Superpose function of the Phenix suite based on the three residues closest to the zinc (H142/335, H146/339, and E166/415). The positions of eight residues from thermolysin that were previously found to adopt similar positions in yeast Ste24p17 are indicated. Only seven of these positions are found in the ordered regions of ZMPSTE24.

Figure 5

Large fenestration. a. The figure shows the largest fenestration in ZMPSTE24 (surface representation), along with a detergent molecule modeled into density protruding through the opening. b. Smaller fenestrations in ZMPSTE24 also shown with modeled protruding detergent molecules. Surfaces of modeled carbon atoms are shown in green, hydrogen in grey, nitrogen in blue, oxygen in red, and sulfur in yellow.

Figure 6

Lipid‐like densities. Electron density and modeled detergent and phospholipid molecules are represented by the mesh and the molecular stick representations. The figure was generated as a 2F_o‐F_c map using mesh at a contour level of 1.0σ showing all electron density within a distance of 2.0 Å of the modeled lipids and detergents.

Figure 7

Cavity. Views of the interior of the ZMPSTE24 cavity. Modeled lipids, detergents and small molecules are shown as stick representations. Surfaces of modeled carbon atoms are shown in green, hydrogen in grey, nitrogen in blue, oxygen in red, and sulfur in yellow. The positions of modeled waters are shown as red spheres.

Figure 8

Kinetics of proteolysis by purified ZMPSTE24. Initial velocity, assayed based on cleavage of fluorogenic prenylated peptide is displayed arbitrary fluorescent units. Nonlinear least squares fitting to the data yielded an estimate of the K _m of 57 ± 12 µM as described in the text.

Figure 9

Inhibition of the activity of purified ZMPSTE24 activity by HIV protease inhibitors. Activity in the presence indicated concentrations of (a) atazanavir, (b) lopinavir, (c) ritanovir, and (d) tipranovir is presented as a fraction of activity in the absence of inhibitor. Inhibition constants derived from the indicated nonlinear least squares fits based on a competitive inhibition model are listed in Table II.

Figure 10

Individual and global analysis of inhibition of purified ZMPSTE24 by lopinavir. Activity was assayed in the presence of the indicated substrate and inhibitor concentrations. The same experimental data is displayed in all three panels. The concentration of ZMPSTE24 was 60 nM. (a) Individual fits to Michaelis Menten kinetics of ZMPSTE24 activity in the presence of four different concentrations of lopinavir (see Table III). (b) Global fit of competitive inhibition model to ZMPSTE24 activity. Parameters derived from the fit are displayed in Table III. (c) Global fit of noncompetitive inhibition model to ZMPSTE24 activity.