Axel P. Mathieu

Phosphorylation and function of the hamster adrenal steroidogenic acute regulatory protein (StAR)

In order to study the effect of phosphorylation on the function of the steroidogenic acute regulatory protein (StAR), 10 putative phosphorylation sites were mutated in the hamster StAR. In pcDNA3.1-StAR transfected COS-1 cells, decreases in basal activity were found for the mutants S55A, S185A and S194A. Substitution of S185 by D or E to mimic phosphorylation resulted in decreased activity for all mutants; we concluded that S185 was not a phosphorylation site and we hypothesized that mutations on S185 created StAR conformational changes resulting in a decrease in its binding affinity for cholesterol. In contrast, the mutation S194D resulted in an increase in StAR activity. We have calculated the relative rate of pregnenolone formation (App. V(max)) in transfected COS-1 cells with wild type (WT) and mutant StAR-pcDNA3.1 under control and (Bu)(2)-cAMP stimulation. The App. V(max) values refer to the rate of cholesterol transported and metabolized by the cytochrome P450scc enzyme present in the inner mitochondrial membrane. The App. V(max) was 1.61 +/- 0.28 for control (Ctr) WT StAR and this value was significantly increased to 4.72 +/- 0.09 for (Bu)(2)-cAMP stimulated preparations. App. V(max) of 5.53 (Ctr) and 4.82 ((Bu)(2)-cAMP) found for S194D StAR preparations were similar to that of the WT StAR stimulated preparations. At equal StAR quantity, an anti-phospho-(S/T) PKA substrate antibody revealed four times more phospho-(S/T) in (Bu)(2)-cAMP than in control preparations. The intensity of phosphorylated bands was decreased for the S55A, S56A and S194A mutants and it was completely abolished for the S55A/S56A/S194A mutant. StAR activity of control and stimulated preparations were diminished by 73 and 72% for the mutant S194A compared to 77 and 83% for the mutant S55A/S56A/S194A. The remaining activity appears to be independent of phosphorylation at PKA sites and could be due to the intrinsic activity of non-phosphorylated StAR or to an artefact due to the pharmacological quantity of StAR expressed in COS-1. In conclusion we have shown that (Bu)(2)-cAMP provokes an augmentation of both the quantity and activity of StAR, and that an enhancement in StAR phosphorylation increases its activity. The increased quantity of StAR upon (Bu)(2)-cAMP stimulation could be due to an augmentation of its mRNA or protein synthesis stability, or both; this is yet to be determined.

Efficient parallel synthesis of macrocyclic peptidomimetics.

A new method for solid phase parallel synthesis of chemically and conformationally diverse macrocyclic peptidomimetics is reported. A key feature of the method is access to broad chemical and conformational diversity. Synthesis and mechanistic studies on the macrocyclization step are reported.

Comparison of the hamster and human adrenal P450c17 (17α-hydroxylase/17,20-lyase) using site-directed mutagenesis and molecular modeling

In order to understand the activity specificity of the hamster cytochrome P450 17 alpha-hydroxylase/17,20-lyase (P450c17), we have studied its structure/activity using three hamster P450c17 recombinant mutants (T202N/D240N/D407H). In transiently transfected COS-1 cells, the mutation T202N reduced 17 alpha-hydroxylation of pregnenolone and progesterone to 24 and 44% of wild type (WT), respectively, followed by reduced 17,20-cleavage to 71 and 67%, respectively. On the other hand, the mutation D240N decreased specifically 17,20-lyase activity to 61% of WT when incubated with pregnenolone while the mutation D407H only decreased 17 alpha-hydroxylation to 46% when incubated with progesterone.To comprehend the altered activity profiles of these hamster P450c17 mutants, we have elaborated a 3D model of the hamster P450c17 and compared it to our preceding model of the human P450c17. Analysis of the mutants with this model showed that, without direct contact to the substrates, these mutations transmit structural changes to the active site. By analogy, these results support the concept that any cellular changes modifying the external structure of P450c17, such as phosphorylation, could have influence on its active site and enzymatic activities.

Molecular modeling and structure-based thermodynamic analysis of the StAR protein

Although much progress has been achieved in the study of the steroidogenic acute regulatory protein (StAR) dependent cholesterol transfer inside mitochondria, not one mechanism can account for all experimental data obtained to date. We have thus investigated the possibility that molecular modeling and structure-based thermodynamic calculations (STC) could enlighten these discrepancies. Starting from the crystallographic data of the human MLN64, a StAR homology model was generated and subjected to STC to verify the importance of StAR structural alterations for proper function. As expected, the model resembled the MLN64 crystal, although no binding site “tunnel” was obtained. Instead, a closed cavity was discovered, approximately the size and shape of cholesterol. This suggests that StAR does indeed require structural alterations to allow cholesterol binding, most evidently by the C-terminal α-helix above the U-shaped β-barrel. Through STC, it is shown that unfolding of this helix is probable and leads to a 2% subpopulation of partially unfolded StAR, supportive of both the intermembrane shuttle and the molten globule hypotheses.

Molecular dynamics of substrate complexes with hamster cytochrome P450c17 (CYP17): mechanistic approach to understanding substrate binding and activities

The cytochrome P450c17 isoforms from various animal species have different substrate selectivity, especially for 17,20-lyase activity. In particular, the human P450c17 selectively produces dehydroepiandrosterone with little androstenedione (AD). Hamster P450c17, on the other hand, produces both of these steroids at comparable rates. We thus investigated if computational analysis could explain the difference in activity profiles. Therefore, we inserted the four P450c17 substrates-pregnenolone, progesterone, and their 17alpha-hydroxylated forms-inside our hamster P450c17 model, which we derived from our human P450c17 model based on the crystal structure of P450BMP. We performed molecular dynamics (MD) simulations on the complexes and analyzed the resultant trajectories to identify amino acids that interact with substrates. Starting with substrates in two different orientations, we obtained two sets of binding trajectories in each case. The first set of trajectories reveal structural rearrangements that occur during binding, whereas the second set of trajectories reflects substrate orientations during catalysis. Our modeling suggests that three distinct steps are required for substrate selectivity and binding to the hamster P450c17: (1) recognition of the substrate at the putative substrate entrance, characterized by a pocket at the surface of the hamster P450c17 containing charged residues R96 and D116; (2) entry of the substrate into the active site, in an intermediate position directed by possible hydrogen bonding of the substrates with the heme D-ring propionate group, R96, R440, and T306; followed by (3) 90 degrees counterclockwise rotation of the substrates, positioning them in optimal position for reactivity, a process that may be directed by hydrogen bonding to the 110-112 region of the hamster P450c17. With some substrates, we obtained trajectories which suggest that major distortions in the I-helix and opening of the H-I loop occur during substrate binding. In conclusion, these modeling exercises provide insight to possible structural reorganizations that occur during substrate binding and suggest that amino acids that participate in three distinct steps of this process may all contribute to substrate binding and activity.

Molecular modeling of the hamster adrenal P450C17.

The cytochrome P450C17 (C17) is the steroidogenic enzyme responsible for the conversion of pregnenolone and progesterone to dehydroepiandrosterone (DHEA) and 4-androstenedione (AD) respectively. This conversion is achieved by two enzymatic activities, 17α-hydroxylase and 17,20-lyase, located at the same active site. In man, the adrenal C17 basically only produces DHEA. We have shown that the hamster adrenal C17 produces DHEA as well as AD. Moreover, the hamster like man produces cortisol as its major glucocorticoid. We can thus compare the hamster and human adrenal C17, and use their differences in order to elaborate a strategy for structure-function studies. We have thus engineered hamster adrenal C17 mutants which possess modified enzymatic activities. We also proceeded to elaborate a three-dimensional model of the hamster C17 to visualise the structural impact of these mutations. This model demonstrates that the mutations created are not localised at the active site, but rather in surrounding regions. These could affect the conformation of the active site, in turn, modulating the 17α-hydroxylase and 17,20-lyase activities. For example, the mutation T202N is located next to Val 482 and Val 483 which compose the roof of the active site. This mutation decreased both 17α-hydroxylase and 17,20-lyase activities, indicating the importance of the roof of the active site for general functionality of the C17.