New Synthetic Methodology for the Construction of 7-Substituted Farnesyl Diphosphate Analogs

Abstract

Through the use of a 1,2-metalate rearrangement, six 7-substituted farnesol analogs were generated in a concise manner. This new synthetic route allowed us to quickly prepare several diverse farnesyl diphosphate analogs with interesting biological activities against mammalian protein-farnesyl transferase.

Isoprenoid diphosphates play vital roles in many cellular processes, such as cholesterol biosynthesis, protein prenylation, and as substrates for sesquiterpene cyclases in plants. The post-translational prenylation of proteins is a required event in the targeting and function of many biologically important proteins. Farnesyl protein transferase (FTase) catalyzes the attachment of a 15 carbon isoprenoid moiety from (farnesyl diphosphate) onto over 60 targets bearing a carboxyl-terminal Ca₁a₂X sequence, including many signalling proteins that play a key role in cancer.¹ With the high prevalence of farnesylated proteins found in cancer, there is a need to characterize the biological activities of farnesylated proteins. The overall goal of this project is to devise new chemical tools that will enable a method to distinguish between the biological activities of different farnesylated proteins. The ability to differentiate the biological impact of one farnesylated protein over another may lead to the development and design of more powerful and robust anti-cancer therapeutics.

The crystal structure of FTase reveals that FPP adopts a conformation in the enzyme that leads to an interaction between the 7 position of the isoprenoid and the a₂ residue of the incoming peptide substrate.² Based on the crystal structure and the known ¹conformational changes in the FTase active site, we evaluated several 7-substituted FPP compounds against a library of CaaX-containing peptides.^3,4 The biochemical screening revealed several 7-substituted FPP analogs that selectively farnesylated certain CaaX-box containing peptides but not others.⁴ Based on these previous results, there is a need for a larger and more diverse library of 7-substituted FPP compounds.

Previously we used the synthetic methodology for the synthesis of 7-substituted FPP analogs developed by Rawat and Gibbs.³ This synthesis was highlighted by two consecutive rounds of vinyl triflate-mediated chain-elongation sequences to successfully complete 7-substituted FPP compounds in 10 linear steps. While successful, thus route has several drawbacks. It is linear, with the diversity element (the 7-substituent) installed early in the route. Secondly, the triflimide reagent needed for each isoprenoid homologation step is expensive and utilized in excess. To facilitate an increase in the size and diversity of the 7-substituted FPP library we developed a new approach that would reduce the number of linear steps. This methodology would also allow us to access other centrally modified farnesyl diphosphate analogs.

After an extensive investigation of various synthetic approaches that could lead to the synthesis of 7-substituted FPP compounds, we focused on a route that utilizes substituted dihydrofuran molecules for installing the 7-substituents into the farnesyl structure. The synthesis of tri-substituted olefins from a Ni-(0)-catalyzed coupling of 2,3-dihydrofurans with Grignard reagents was first reported by Wenkert and colleagues⁵ and more extensively studied by Kocienski.⁶ Kocienski and colleagues reported a copper (I)-catalyzed coupling of Grignard reagents and organolithiums with 5-lithio-2,3-dihydrofuran results in trisubstituted olefins ^7,8 in a straightforward and stereoselective manner. Therefore, we applied the synthesis of 4-homogeraniol derivatives to the generation of substituted farnesyl analogs.

To begin the synthesis of 4-homogeraniol derivatives (Scheme 1), we first prepared homoprenyl iodide (2) from cyclopropyl methyl ketone in a 75% yield.⁹ Subsequent lithium-halogen exchange, followed by the addition of CuCN, resulted in 3. With 3 in hand, we generated 5-lithio-2,3-dihydrofuran (5) from the action of t-BuLi on 2,3-dihydrofuran (4). Attempts to replicate the 1,2-metalate rearrangement resulting from the addition of 3 to 5, as reported by Kocienski and colleagues⁸ for the synthesis of 4-homogeraniol derivatives, were largely unsuccessful because of the decomposition of organocuprate 3. The problem was resolved when dimethylsulfide was added as a cosolvent, which presumably stabilizes the organocuprate, and as a result the expected 1,2-metalate rearrangement took place.¹⁰ The 1,2-metalate rearrangement led to the production of the higher order alkenylcuprate 6, a versatile intermediate for the synthesis of 7-substituted FPP analogs. The coupling of 6 with a variety of electrophiles (SnBu₃Cl, I₂, TMS-propargyl bromide, and allyl bromide) was achieved by re-cooling the solution of alkenylcuprate (6) to 0 °C and adding in the appropriate electrophiles. This led to the formation of 4-substituted homogeraniol derivatives (7a–e) in moderate yields (42–62%). Despite the modest yields, utilization of this synthetic transformation is beneficial in the synthesis of farnesol derivatives because it allows for the transformation of readily available starting materials into advanced synthetic intermediates in one step. We next focused on introducing various substituents into iodide 7a and stannane 7b that would eventually become the 7-substituent on the corresponding farnesyl diphosphate analog.

Scheme 1.

Synthesis of 4-homogeraniol derivatives.

The facile nature of the Suzuki-Miyaura reaction and the commercial availability of a large library of organoboranes prompted us to examine their cross coupling of organoboranes with 7a. We used the standard Suzuki-Miyaura coupling conditions¹¹ in which 2-thienyllboronic acid (8) was successfully coupled to vinyl iodide 7a to result in 9a in a 56% yield (Scheme 2). Similar Suzuki-Miyaura couplings should allow for the future installment of diverse aromatic and vinyl substituents at the 7-position of FPP. Once the methodology was developed to enable the synthesis of 7-substituted FPP compounds with aryl and vinyl moieties at the 7-position, we next examined methodology that would enable us to install alkyl moieties at the 7-position. A Negishi coupling was envisioned to replace the iodide of 7a with a variety of commercially available or readily prepared alkyl zinc reagents.¹² The organozinc reagent 10 was prepared through a lithium-halogen exchange of 1-iodo-4-trimethylsily-1-butyne with t-BuLi, followed by the addition of a zinc chloride solution in THF. With 13 in hand, we employed a Negishi cross coupling procedure,¹² to the join 7a and the organozinc 10 resulting in the synthesis of the alkyl substituted 4-homogeraniol derivative 9b. The vinyl stannane 7b would serve to provide 4-substituted homogeraniol derivatives via the Stille reaction with commercially available vinyl and aryl halides. One substituent we wished to install at the 7-position of FPP was the 1,3-dimethylpropenyl moiety. Using CsF as an additive, we successfully coupled 7b to 11 using Pd(PPh₃)₄, in DMF to produce 9a in a 50% yield (Scheme 2).

Scheme 2.

Pd-catalyzed coupling reactions with 4-substituted homogeraniol derivatives.

Once the synthesis of the 4-substituted homogeraniol derivatives (7c–e and 9a–c (Schemes 1 and 2)) was complete, we performed the final steps necessary to generate 7-substituted FPP analogs (Scheme 3). Transformation of the homogeraniol derivatives 10c–e and 11a–c into the resulting iodides 13a–f was accomplished in good yields.¹³ With the alkyl iodides 13a–f in hand, we next focused our attention on converting the alkyl iodides into their corresponding organozinc reagents. To achieve this, we employed two different procedures depending on the substituent at the four position of our homogeraniol derivatives. We successfully employed the same strategy to prepare the the organozinc derivatives 14a–c, e–f as were generated as previously described for the construction of 10 (Scheme 2). Preparation of organozinc derivative 14d was unsuccessful with the harsh conditions employed for constructing 10. The slightly acidic proton or the conjugation of the thiophene to the double bond of the homogeraniol backbone may have caused unwanted side reactions with t-BuLi, thus preventing the preparation of 14d. In order to circumvent this problem, we followed a procedure developed by Huo to convert unactivated halides into the corresponding organozincs without the use of strong bases,¹⁴ which resulted in the preparation of the organozinc intermediate 14d. The rest of the organozinc derivatives (14a–c and 14e–f) were generated as previously described for the construction of 10 (Scheme 2). Vinyl iodide 15^15,16 was then added to the organozinc solution of 14a–f, followed by the addition of Pd(PPh₃)₄ at room temperature resulted in the synthesis of 7-substituted-farnesol derivatives (16a–f). With the 7-substituted farnesol derivatives synthesized we next converted them into their corresponding diphosphates (17a–f) by employing a Corey-Kim chlorination¹⁷ of the allylic alcohol, followed by displacement of the resulting chloride with tetrabutylammonium diphosphate (Scheme 3).¹⁸ Previously, our laboratory has shown that 7-substituted FPP analogs can display diverse biochemical activity with FTase and a variety of CaaX-peptides.⁴ The 7-substituted FPP compounds synthesized in this study (17a–f, Figure 1a) were initially evaluated for their biological activity in an in vitro fluorescence-based assay⁴ with FTase and the CaaX-peptide, dansyl-GCVLS (H-Ras). Among the 7-substituted compounds assayed for their in vitro biological activity, compounds 17 a,b,e,f were found to behave as efficient substrates (Figure 1b). In addition to the substrate activity of the newly synthesized 7-substituted compounds, one compound (17c) was found to act as a weak inhibitor with an IC₅₀ of 25 µM against FTase. Of particular importance are the substrate activities of the two alkynyl FPP compounds 17e and 17f with dansyl-GCVLS. Both compounds are efficient substrates as demonstrated in the fluorescence assay and in an HPLC confirmation assay (data not shown). These compounds were designed with the goal to be used as chemical tools for the tagging of prenylated proteins via “click chemistry”.^19,20 This strategy has proved valuable for the dissection of the cellular biology of various ligands with protein targets,^21–24 including FTase.^{25, 26}

Scheme 3.

Conversion of 4-homogeraniol derivatives into their corresponding 7-substituted FPP analogs

Figure 1.

Structures and substrate activity of 7-substituted FPP analogs. (a) Structures of synthesized 7-substituted diphosphates. (b) Substrate activity of 7-substituted FPP analogs with dansyl-GCVLS.

The synthetic approaches developed in this study for the synthesis of 7-substituted FPP analogs have been used to develop several diverse FPP analogs with biological activities. The original synthetic plan³ for these compounds required eleven steps for the completion of 7-allylfarnesyl diphosphate (17a; overall yield = 7%). The newly developed strategy allows for the rapid and efficient synthesis of 7-substituted FPP analogs; specifically, 7-allyl analog 17a was prepared in six steps with a 12% overall yield. In addition to significantly shortening the route, we have also provided methodology that allows for the construction of FPP analogs unachievable by previous methods.