Hybrid homology modeling and mutational analysis of cytochrome P450C24A1 (CYP24A1) of the Vitamin D pathway: insights into substrate specificity and membrane bound structure-function

Abstract

Cytochrome P450C24A1 (CYP24A1), a peripheral inner mitochondrial membrane hemoprotein and candidate oncogene, regulates the side-chain metabolism and biological function of vitamin D and many of its related analog drugs. Rational mutational analysis of rat CYP24A1 based on hybrid (2C5/BM-3) homology modeling and affinity labeling studies clarified the role of key domains (N-terminus, A', A, and F-helices, beta3a strand, and beta5 hairpin) in substrate binding and catalysis. The scope of our study was limited by an inability to purify stable mutant enzyme targeting soluble domains (B', G, and I-helices) and suggested greater conformational flexibility among CYP24A1's membrane-associated domains. The most notable mutants developed by modeling were V391T and I500A, which displayed defective-binding function and profound metabolic defects for 25-hydroxylated vitamin D3 substrates similar to a non-functional F-helix mutant (F249T) that we previously reported. Val-391 (beta3a strand) and Ile-500 (beta5 hairpin) are modeled to interact with Phe-249 (F-helix) in a hydrophobic cluster that directs substrate-binding events through interactions with the vitamin D cis-triene moiety. Prior affinity labeling studies identified an amino-terminal residue (Ser-57) as a putative active-site residue that interacts with the 3beta-OH group of the vitamin D A-ring. Studies with 3-epi and 3-deoxy-1,25(OH)2D3 analogs confirmed interactions between the 3beta-OH group and Ser-57 effect substrate recognition and trafficking while establishing that the trans conformation of A-ring hydroxyl groups (1alpha and 3beta) is obligate for high-affinity binding to rat CYP24A1. Our work suggests that CYP24A1's amphipathic nature allows for monotopic membrane insertion, whereby a pw2d-like substrate access channel is formed to shuttle secosteroid substrate from the membrane to the active-site. We hypothesize that CYP24A1 has evolved a unique amino-terminal membrane-binding motif that contributes to substrate specificity and docking through coordinated interactions with the vitamin D A-ring.

Figure 1

A 2C5/BM-3 hybrid homology-model based representation of 1,25(OH)₂D₃ bound in the rat CYP24A1 active-site. Calcitriol is seen docked in the rat CYP24A1 active site with the proximity of putative binding residues (M246, F249, V391, T394 and I500) shown. Also shown are influential residues from the A' (W75) and A-helices (Y89), and the conserved catalytic threonine (T330) in the I-helix. The K'-helix residue Thr-416, not studied in the current work is also shown. Thr-416 along with Ile-500, have been reported to influence the catalytic regiospecificity of rat CYP24A1 [20]. Image was generated using VMD [22] .

Figure 2

The side-chain metabolism of 1,25(OH)₂D₃ by wild-type CYP24A1 and site-directed mutants. Wild-type (WT) CYP24A1 multi-catalytic activity with 1,25(OH)₂D₃ substrate is compared to a series of model-derived mutant enzymes using enzyme reconstitution analysis (10 min.@ 37C, 0.1 μM ADX, 0.1 μM ADR ). (Abbreviations are: 1,24,25 = 1,24,25(OH) ₃D₃; 24-oxo-1,25 = 24-oxo-1,25(OH) ₂D₃; 24-oxo-1,23,25 = 24-oxo-1,23,25(OH) ₃D₃; 1,23 = 24,25,26,27-tetranor-1,23(OH) ₂D₃.)

Figure 3

Altered metabolism of 1,25(OH)₂D₃ by CYP24A1 Ser-57 mutants. Purified enzymes were reconstituted with ADX (0.1 (1X) and 0.2 μM (2X)) and ADR (0.1 μM) then incubated with 1,25(OH)₂D₃ for 5 min. prior to metabolite extraction. S57D displays a diminished capacity to catalyze early C24 oxidation reactions. The rate of electron flux from ADX does not appear to be involved in this phenomenon.

Figure 4

Chemical structures of calcitriol (1α,25-dihydroxyvitamin D₃) and the 3-OH analogs (3α-epi- and 3-deoxy-1α,25(OH)₂D₃) used in the present study.

Figure 5

Schematic representations of the A.) 2C5/BM-3 Hybrid and B) CYP2B4-based homology models of rat CYP24A1. Calcitriol (yellow) is seen docked over the heme center (black) in both models, with the active site adopting unique positions based on the positioning of individual helices (A=Gold, A'=Turquoise, B'=Orange F=Magenta; G=Blue, I=White). Both models depict close interaction among the A'-, B'-, F- and G-helices, β3a strand and β5 hairpin that form the top of the active-site near the opening of pw2a-like substrate access channel. Interactions between the B', F, G and I' helices appear to control the size and shape of the active-site and may control the pw2a channel that functions to shuttle polar products to the matrix. W75 is seen at the junction of the pw2a channel and a pw2d-like channel predicted to pass out the A helix (Y89) towards the enzyme's N-terminus and the putative active-site residue Ser-57, which is fully modeled in panel B. Images in this figure were created using PyMol (www.pymol.org).

6

A model for CYP24A1 membrane binding and the proposed amino-terminal binding site. Serine-57 in rat CYP24A1 may form a contact point with the 3β-hydroxyl group of calcitriol and that functions in substrate recruitment, trafficking and active-site docking events. Noted are the locations of viable and non-viable rat CYP24A1 mutants developed for these studies; in short, functional enzyme was limited to mutations made within membrane-associated domains implying a higher degree of conformational flexibility associated with membrane binding. The location of these flexible domains guided the hypothetical membrane insertion presented here. An N-terminal binding site, shown with a key contact point (S57), may function to recruit and guide substrate through a pw2d-like access channel. Together with portions of the F-helix, this N-terminal binding site is hypothesized to provide important spatial cues to the vitamin D A-ring that allow for efficient trafficking and docking of substrate. The CYP image was generated using VMD [22].