On Biosynthetic and Biomimetic Steroid Chemistry: The Polyene Cyclization Mechanism of Lanosterol

            The biogenetic mechanism for the formation of the 6-6-6-5 tetracyclic steroid lanosterol (outlined below), an intermediate in the biosynthesis of cholesterol, has been studied by chemists for over half a century. Enzyme crystallographic data, together with other bioorganic and mutational studies, have yielded a biomolecular mechanism for enzymatic triterpene cyclization starting from acyclic 2,3-oxidosqualene. The mechanism involves initiation through protonation by a specific aspartic acid residue. The polycyclization cascade then proceeds away from the catalytic acid, with intermediates shielded by aromatic amino acid residues that line the active site cavity with p-electron-rich faces. However, this biosynthetic conversion is incredibly complicated and many of its precise details have not been fully resolved. One key question that has remained unanswered until recently concerns the extent to which ring formation is concerted with the initial protonation step. Earlier this year, in an effort to sort out uncertainties associated with this complex mechanism, Ruibo Wu’s group in Guangzhou, China, in collaboration with a team at Vanderbilt University led by the theoretical computational chemist Andes Hess, successfully characterized the cyclization of squalene oxide at the highest computational level carried out to date.

            The investigators used quantum mechanics/molecular mechanics molecular dynamics (QM/MM MD) simulations to construct the overall cyclization free-energy profile (reproduced below) for this fascinating biosynthetic transformation. The study shows that the entirety of the process (acyclic substrate à protosterol cation) entails two concerted, asynchonous reactions which proceed through two metastable states (the A and B states) and one stable intermediate, under oxidosqualene cyclase (OSC) enzyme catalysis. The overall cyclization can be considered a carbocation migration and is exergonic by ~4.4 kcal/mol. The rate-limiting step is the concerted formation of the A-ring, with opening of the epoxide ring. The stable intermediate (the C’ state) is a tricyclic 6-6-5 species that is subsequently converted to the tetracyclic protosterol cation (the D state) by a ring expansion with concomitant (concerted) formation of the D-ring. Importantly, all previous experimental mutagenic data is consistent with the identified reaction mechanism.

Hess and Wu. ACIE 2015, 54, 8693.

            The 2015 computational mechanistic study of Hess and Wu will no doubt inspire the synthetic enthusiasts among us to reflect on the myriad of creative ways in which organic chemists have tried to mimic the action of the enzyme OSC in a laboratory flask or reactor. This area of so-called ‘biomimetic’ research was pioneered by W. S. Johnson at Stanford University in the 1960’s and 70’s. Johnson’s efforts were later extended by countless others. In his 1976 Review Article in Angewandte Chemie, Johnson elegantly summarizes the synthetic strategy: “Certain polyenic substances having trans olefinic bonds in a 1,5 relationship can be induced to undergo stereospecific, non-enzymic, cationic cyclization to give polycyclic products with all-trans (“natural”) configuration. These transformations appear to mimic in principle the biogenetic conversion of squalene into polycyclic triterpenoids…” A highlight from Johnson’s impressive body of research is a remarkably short synthesis of racemic progesterone, starting from the cyclopentenol shown below. Cyclization under Brønsted acidic conditions produces a crystalline tetracyclic product in 71% yield. Subsequent zonolysis followed by intramolecular aldol condensation then furnishes synthetic progesterone. A related process (lower panel below) was also conceived by Johnson to generate a valuable 11a-substituted intermediate that intersects with a known commercial route for the production of hydrocortisone acetate. More recently, ‘biomimetic’ polyolefin carbocyclization has been elegantly applied by E. J. Corey’s laboratory to the chemical synthesis of complex limonoid steroidal systems (highlighted here).

Aminosterols from the Dogfish Shark: The Discovery of the Allosteric Phosphatase Inhibitor, Trodusquemine

            In 1993, as part of an effort to search for antibiotic host defense agents in the gut of various animals including the dogfish shark (spiny dogfish) Squalus acanthias, the unique antimicrobial aminosterol squalamine (structure shown above) was discovered. Squalamine is essentially the condensation adduct between a C24-sulfated bile salt derivative and the polyamine, spermidine. The steroidal portion of the squalamine structure bears the trans configuration at the A/B ring junction and is additionally hydroxylated at C7 in the a-configuration. It was initially shown that stomach extracts of the dogfish shark exhibited potent antimicrobial activity. Further efforts to purify and identify the bioactive molecule responsible for the observed activity led to the isolation and structure determination of squalamine. The aminosterol was initially identified as a broad-spectrum antimicrobial, as it was found to exhibit potent activity against fungi, protozoa, and both Gram-negative and Gram-positive bacteria. The antimicrobial activity was attributed to squalamine’s ability to modify membrane integrity by increasing permeability. Interestingly, it was later shown that squalamine also possessed antiangiogenic and antitumor properties and the molecule was eventually advanced to Phase II clinical trials for the treatment of patients with advanced nonsmall cell lung cancer.

            Attempts to procure larger amounts of squalamine from Squalus acanthias resulted in the discovery and isolation of a related aminosterol MSI-1436, later dubbed trodusquemine. Trodusquemine, as compared to squalamine, features an invariant steroid skeleton but is conjugated at C3 to spermine (as opposed to spermidine), an elongated tetrabasic polyamine. Rather unexpectedly, trodusquemine was found to induce profound appetite suppression in vivo in mammals (For another example of a natural appetite-suppressant steroid, see here). As a result, it has been speculated that trodusquemine is responsible for the sporadic feeding behavior of the dogfish, which eats only once every two weeks. In vitro screening against a panel of potential cellular targets revealed protein-tyrosine phosphatase 1B (PTP1B) inhibitory activity in cell-free and cell-based assays. PTP1B is a negative regulator of the effects of insulin and leptin signaling through dephosphorylation of the insulin receptor (IR) and IR Substrate 1, thereby inactivating the insulin pathway. Thus, inhibition of PTP1B maintains the insulin and leptin pathways in active, phosphorylated states, which triggers appetite suppression. Indeed, PTP1B expression and activity is increased in obese and insulin-resistant humans and neuronal-specific PTP1B knockout mice have markedly reduced weight.

            The inhibitory activity of trodusquemine at the cellular target PTP1B is significant because it has been notoriously challenging for medicinal chemists to develop effective small molecule inhibitors targeting the active site of this enzyme. One problem is a high degree of active site protein sequence homology, leading to difficulties in achieving selectivity over other off-target phosphatases. Moreover, potent active site tyrosine phosphatase inhibitors were designed to mimic phosphotyrosine (see examples shown above). Consequently, the inhibitor ligands that were developed were highly charged and thus had limited membrane permeability and drug development potential. A known inhibitor of striatal-enriched protein tyrosine phosphatase (or STEP), a potential Alzheimer’s target, is TC-2153, a synthetic molecule that features a rarely encountered heterocyclic ring comprised of five contiguously attached sulfur atoms. Current interest in a structure like TC-2153 reinforces the point that it has been historically difficult to find small molecule phosphatase inhibitors with conventional ‘druglike’ substructures and physicochemical properties. Trodusquemine avoids many of these pitfalls by binding to an inhibitory allosteric site within an intrinsically disordered segment of the C-terminal, noncatalytic region of PTP1B. Many of the previous high-throughput screening campaigns conducted against PTP1B used a recombinant, truncated form of the enzyme lacking this C-terminal segment and, therefore, failed to identify selective allosteric inhibitors such as trodusquemine. As a selective, non-competitive inhibitor of PT1B, trodusquemine has tremendous potential as an anti-obesity and anti-diabetic therapeutic agent. The molecule has also been shown to be able to cross the blood brain barrier (BBB), and thus may be centrally active, opening a new range of potential indications. Trodusquemine has been well tolerated in dose escalation and dose ranging clinical studies completed to date in over 65 patients.

            The antimicrobial natural product squalamine has been synthesized by processes starting from b-stigmasterol, as well as chenodeoxycholic acid (CDCA, see Scheme above). One interesting and critical transformation that has been utilized to generate gram-quantities of squalamine is a regio- and stereoselective reductive amination with a spermidine equivalent, wherein the distal primary amine is masked as a non-nucleophilic nitrile. In this conversion, the primary amine reacts preferentially over the internal secondary amine to generate an intermediary imine species, upon elimination of water with assistance from trimethyl orthoformate. Sodium borohydride is then used to reduce the imine in a stereoselective fashion from the bottom face of the molecule to give, predominantly, the desired b-orientation at the C3 position (d.r. 6:1). The nitrile is then easily converted to the requisite amine (spermidine) by catalytic hydrogenation, with both operations performed in the presence of the apparently unreactive C24-sulfate. The hydrogenation must be conducted under acidic conditions (TFA) to avoid cyclization of the internal secondary amine onto the pendant nitrile to give a cyclic amidine impurity. In this example, the semisynthetic squalamine was purified by HPLC and isolated as the trifluoroacetate salt with a final purity of 97%. One presumes that a similar process is currently used to manufacture the related clinical candidate, trodusquemine.