Terminating Pt-Initiated Cation-Olefin Reactions with Simple Alkenes

Cyclase Enzyme Mimics

The en masse cyclization of polyolefins into polycyclic terpenoids by cyclase enzymes (e.g. squalene to hopene), is a biosynthetic reaction of particular fascination to chemists.^[1] Noteworthy recent additions to synthetic mimics^[2] of the cyclase enzymes are asymmetric methods that include Brønsted Lewis Acids (BLA),^[3] masked equivalents of Br⁺ and I⁺,^[4] organocatalysts,^[5] and electrophilic metal catalysts.^[6] With the exception of Hg(II) reagents,^[7] few electrophilic metal catalysts cyclize poly-enes with bio-like alkene terminators.^[8] The development of methods whose catalysts can initiate, cyclise, and terminate poly-enes under ligand control would significantly advance the state of the art.

Herein we describe the development of an alkene terminated cation-olefin cascade reaction that is initiated by the dicationic platinum complex [(PPP)Pt][BF₄]₂ (PPP = bis(2-diphenylphosphanylethyl) phenylphosphane), 1.^[9] Compound 1 is especially efficient at initiating cyclizations wherein the polyene carries a mono-substituted alkene terminus.^[10] In addition to diastereoselectively forming polycyclic products with a broad variety of terminating alkenes, the reactions described herein contrast Hg(II) reagents by the lack of premature termination processes.^[11]

We previously reported that L_nPt²⁺ sources will initiate the cation-olefin cascade with subsequent termination by the intramolecular addition of a protic trap (alcohol, phenol, or sulfonamide, e.g. eq 1).^[12] Computational analysis showed that when a base was H-bonded to the protic terminus and the alkene was in a suitable geometry, the cyclization was highly favourable and virtually barrierless.^[13] In contrast, base-free calculations were characterized by high energy intermediates and significantly less favourable thermodynamics.

(1)

This latter scenario most likely describes the early stages of a poly-ene cascade that terminates with a non-protic group, which in the case of an alkene is not even acidic until the cation is fully formed. The difficulty of productively engaging a Brønsted base at an alkene terminus thus likely explains the paucity of synthetic examples.^{[14, 15]}

The combination of a polar solvent (EtNO₂) and either Ph₂NMe or, more conveniently, a resin N-bound piperidine base led to an efficient and highly diastereoselective cyclization of triene 2 to 3 (eq 2). In contrast to protic terminators however, the reaction proceeds much more slowly (minutes for eq. 1 vs. 36 h for eq. 2), a difference which we interpret as reflecting the kinetic cost of generating a discrete tertiary cation.

(2)

Crystallographic characterization of 3^[16] pointed to a predictable initiation at the least substituted alkene, a chair-chair cyclization conformer, and the intermediacy of an exocyclic tertiary cation that eliminates to the isopropenyl group. Several features are notable in the solid state structure of 3. The first is the Pt-CH orientation, which positions the C-H vector in the square plane to minimize steric congestion. This rotamer positions the angular CH₃ group near the face of one P-Ph group, which causes an upfield shifting of this CH₃ group in the ¹H NMR (to ~0.1 ppm). This resonance proved to be diagnostic and was observed in each of the described structures (vide infra).

A number of poly-enes with terminating tertiary carbocations were examined that varied in the number of rings formed (two-three), the arrangement of the terminating alkene (endo versus exo-cyclic), and the ring size (Table 1). Even more facile than the 6-exo termini were reactions wherein the terminating alkene was arranged to react with the 6-endo geometry.^[17] These reactions were 2–4 times faster than the 6-exo analog 2, and provided a number of carbon skeletons. In the case of 5, the putative tertiary cation, formed from a chair/chair/chair transition state, eliminates to give the more stable C12/13 alkene isomer (Scheme 1). Products that would have arisen from premature quenching of a putative cation at C5 or C9 were not observed (<5%). In most cases, the structure of the resulting Pt-complex was confirmed by X-ray methods (see SI).

Table 1.

Yields and scope of polyolefin cyclizations initiated by dicationic platinum complex 1.

Entry	Substrate[a]	Product[b]	Isolated Yield (%)
1			80
	2	3 X-ray
2			74
	4	5
3	6	7	89
4	8	9 X-ray	97
5			95
	10	11 X-ray
6			76
	12	13 X-ray
7			80
	14	15 X-ray

^[a]Conditions: (PPP)PtI₂, 2 eq substrate, 2.5 eq AgBF₄, 3 eq piperidine resin base, and EtNO₂.

^[b][Pt]⁺ = [(PPP)Pt]⁺

Scheme 1.

Chair-chair-chair cyclization with 6-endo termination

Even more reactive were conformationally constrained dihydro-naphthalene terminating groups (6, 8, and 10), which efficiently converted to the tetra- and pentacyclic products (entries 3–5).^[8] The conversion of 8 to 9 was ~4 fold faster than the non-methoxy substituted example (entry 3), suggesting that the nucleophilicity^[18] and/or cation stability of the terminus plays a significant role in the reaction kinetics. As judged by comparing the cyclization rates of 8 and 10, an additional isoprene unit in the main chain does not significantly affect the reaction barrier.

In the case of 6, extended reaction times led to a partial conversion to the tetra substituted isomer at the B/C ring junction.^[19] As reported by Corey,^[8] this isomerization could be accelerated by acids, though the sulfonic acids also caused partial protodemetallation of the Pt.^[20] By contrast, the tertiary cation formed on cyclizing 12 preferentially eliminates to the more stable tetrasubstituted alkene product 13.^[21]

An entirely different path was followed when terminating the cascade to a 3° carbenium ion required a 5-exo geometry. In these cases a clean Wagner-Meerwein rearrangement converted the tertiary cation to the rearranged carbon skeleton of 15,^[22] which was confirmed by X-ray analysis.^[16]

To gain insight into the diverging behaviour of 6-exo and 5-exo terminated reactions, a computational analysis (DFT B3LYP/6-31G*)^[23] of the key 1,2-shifts were carried out on simplified model systems (Scheme 3). Revealing was the differential activation energy for the initiating 1,2-hydride shift, which was 7.3 kcal/mol more favourable for the 5-exo terminated ring systems than for the 6-exo. The subsequent steps were lower in energy suggesting that it is a slower initiating 6-6 1,2-H-transfer that diverts the reaction towards a competitive base induced elimination.

Scheme 3.

A comparison of the 6,5 (a) and 6,6 (b) energies (kcal/mol) along the Wagner-Meerwein reaction coordinate

Compound 1 was additionally investigated for its ability to cyclise a squalene analog that lacks the terminal methyl groups (eq. 3). Although the spectral complexity was significant and more than one isomer was formed, similarities to 15 suggested that the cyclization followed a 6-6-5-exo pathway to a C14 cation, which non-selectively rearranged akin to 4. Unlike cyclase enzymes, the environment of the terminating cation is not conducive to ring expansion/D-ring annulation.^{[24, 25]} van Tamelen made similar observations in Brønsted-mediated reactions on squalene oxide.^[26]

(3)

The viability of performing an asymmetric cascade cyclization was investigated using the chiral P₂PPt²⁺ complex (P₂ = DTBM-SEGPHOS, P = PMe₃) 19. The combination of a chiral P₂ ligand and an achiral monodentate phosphine has been previously shown to catalyze cyclorearrangment reactions with high ee’s.^[27] When 19 was reacted with 8 under the standard conditions (Table 1), NMR spectroscopy indicated that a single stereoisomer was obtained (¹H, ³¹P), ie. the chiral initiator efficiently and diastereoselectively activates a single olefin face.

In summary, we report the results of a Pt(II) mediated cyclization methodology that explores the boundaries of poly-alkene cation-olefin reactions. These data reinforce the notion that the nucleophilicity/cation stability of the terminating alkene is of paramount importance and the termination outcomes depend on structure. Electrophilic Pt-dications are also shown to be unique in their ability to activate and mediate the cascade reactivity of poly-ene reactants. The results pave the way to as of yet unknown catalytic asymmetric cation-olefin cyclizations of poly-alkenes.

Experimental Section

Standard Cyclization Reaction. To 30 mg of (PPP)PtI₂ was added 15 mg of AgBF₄ followed by 0.75 mL EtNO₂. The mixture was then stirred for 1 h in the dark. The contents were filtered through a 0.2 μm PTFE syringe filter, washing out the flask and syringe with 0.25 mL EtNO₂, into a flask containing 2 equiv of substrate and 3 equiv of piperidine resin. The reaction mixture was stirred in the dark until the reaction was complete (3–48 h, verified by ³¹P NMR). The reaction mixture was passed through a 0.2 μm PTFE syringe filter, washing out the flask and syringe filter with 0.25 mL EtNO₂. Solvent was then removed under a stream of N₂. The complex was twice reconstituted in a minimum amount of CH₂Cl₂ and force precipitated with cold tBuOMe. The mixture was centrifuged and the solvent was decanted off. The crude was purified by flash column chromatography.

Figure 1.

X-ray structure of 3 showing the diagnostic interaction between the axial-methyl group and the P-Ph ring.

Scheme 2.

Reaction termination by selective deprotonation