Total synthesis of Lys³ tamandarin M. A potential affinity ligand

Abstract

The synthesis of Lys³ tamandarin M is described. This analogue can be used as a protein affinity ligand to probe the mechanism of action of this unique class of molecules.

The didemnins are a family of marine natural products that were isolated by Rinehart in 1984.¹ They consist of a 23-membered depsipeptide macrocycle with a side chain attached to the threonine nitrogen of the macrocycle. The didemnins (Figure 1) have been the subject of many syntheses.² Didemnin B (2) was the first marine natural product to enter into phase II clinical trials, but the trials were terminated due to toxicity issues associated with dosing levels,³ and didemnin M (3) was later reported by Rinehart in 1994.⁴ The difference between didemnins B and M is that the side chain of didemnin M is extended by two additional amino acid residues. It is one of the most potent immunosuppressive reagents known. In 2000, Fenical and Vervoort reported the structure of the structurally related tamandarins A and B⁵ (Figure 1) and both molecules were subsequently synthesized.⁶ The main difference between didemnins and tamandarins is that tamandarins A and B posses a structurally simpler hydroxyisovaleryl (Hiv) fragment within a 21-membered ring instead of the α–(α-hydroxyisovaleryl)propionyl fragment within a 23-membered ring. This structural distinction results in minor differences in biological activity, but the absence of the highly epimerizable stereocenter present in the didemnins makes the tamandarins a more accessible synthetic target. Although these compounds have been extensively studied, their mechanism of action remains unclear.

Figure 1.

Structures of the didemnins and tamandarins and Lys³ tamandarin M

Many of the unresolved questions in the didemnin field are related to its mechanism(s) of action. Previous affinity studies have focused exclusively on the macrocycle by attaching a linker to the N-Me-D-Leu residue of didemnin A (1).⁷ These studies revealed two binding proteins; elongation factor 1α (EF-1α) and palmitoyl protein thioesterase 1 (PPT1). Although the first study showed that an EF-1α-mediated effect upon protein translation explained certain biochemical effects of didemnins, their potent antiproliferative effect remained unexplained.⁸ The relevance of PPT1 in mediating antiprolifrerative activities of didemnins was questioned.⁹ Since the residues in both didemnins and tamandarins are the major contributors to the biological activity, it is possible that the molecular targets mediating the activity did not recognize the immobilized side chain. A probe with a fully elaborated side chain would be optimal for affinity studies since the side chain is vital for biological activity.³ Lys³ Tamandarin M (6) was designed with the replacement of the leucine residue with a lysine residue to allow for attachment of a biological probe far removed from the intact side chain (Figure 1).

The first step in the retrosynthetic analysis was disconnection of the side chain from the macrocycle to afford 7 and 8 (Scheme 1). The macrocycle (7) was then opened at the Pro⁴, Me₂-Tyr⁵ junction to form linear precursor 9. This site of closure was chosen due to previous reports showing that this position for cyclization has consistently given the highest yields.^{2c, 10} This linear precursor was then divided into three fragments; the Pro-Lys-Hiv fragment (10), the norstatine fragment (11) which has been previously synthesized¹¹, and the Me₂Tyr-Thr fragment (12).

Scheme 1.

Retrosynthetic analysis

The synthesis began with the formation of the Pro-Lys-Hiv fragment 10 (Scheme 2). Boc-L-proline (13) was esterified with 2-(trimethylsilyl)ethanol to form the TMSE ester. Removal of the Boc group followed by coupling to 14 with diethyl cyanophosphonate (DEPC) gave dipeptide 15 in good yield.^2c Boc removal followed by a BOP mediated coupling with 2-hydroxy-3-methylbutanoic acid (16) yielded the Pro-Lys-Hiv fragment (10).

Scheme 2.

The Tyr-Thr fragment was prepared next (Scheme 3). L-Tyrosine (17) was first protected as its 2-(trimethylsilyl) ethoxycarbonyl (Teoc) carbamate. Dimethylation then subsequent hydrolysis yielded acid 18. The resulting acid was esterified with N-Boc-Thr-OBn using EDCI to furnish the desired benzyl ester. Hydrogenolysis led to acid 12 in quantitative yield.

Scheme 3.

With all fragments in hand the linear precursor was assembled (Scheme 4). Esterification of acid 11 with alcohol 10 in the presence of EDCI and DMAP furnished 19. Treatment of 19 with HCl removed the the Boc group and fragmented the oxazolidine to yield a salt which was immediately coupled to acid 12 in the presence of HATU to yield the desired linear precursor 9.

Scheme 4.

Unfortunately, all attempts to remove the protecting groups of 9 failed to give the desired product (9a). Initial treatment with TBAF resulted in hydrolysis of the Tyr-Thr ester (Table 1, entry 1). Treatment with other fluoride sources also did not yield the free linear precursor. TAS-F¹² and TBAF on silica gel¹³ removed the TMSE ester only under a variety of conditions to yield 9b (Table 1, entry 3–6). Treatment with potassium fluoride at elevated temperatures only resulted in decomposition (Table 1, entry 7). Finally treatment with HF-pyridine resulted in recovered starting material. We were able to remove the TMSE group under fluoride ion conditions, but the Teoc group was unnaffected.

Table 1.

Screening different conditions to effect protecting group removal.

entry	reagent and solvent	temperature (°C)	result
1	TBAF, THF	0 or 25	Hydrolysis of Tyr-Thr ester
2	TBAF, THF AcOH	25	No reaction
3	TBAF on SiO2	25	9b
4	TAS-F, DMF	25 or 40	9b
5	TAS-F, CH3CN	25, 40, or 80	9b
6	TAS-F, CH2Cl2	25, 40, or 50	9b
7	KF, CH3CN 18-crown-6	70	Decomposition
8	HF·Py	25	No reaction

A change in protecting group strategy was necessary. A linear precursor was designed in which the Me₂Tyr⁵ nitrogen was protected as a Boc carbamate, and the Thr⁶ nitrogen was protected as a Troc carbamate (20) (Figure 3). This strategy would allow for a one step deprotection of the linear precursor followed by cyclization. The TMSE group was kept the same since it could be removed with fluoride sources, and also could be removed along with the Boc group in one step using TFA.¹⁴ The only fragment of the molecule that needed revision was the Me₂Tyr⁵-Thr⁵ portion. With this change in protecting groups all retrosynthetic disconnections remained unchanged.

Benzyl ester 21 was treated with acid to remove the Boc group and the free amine was subsequently protected as its Troc carbamate (Scheme 5). Esterifcation of tyrosine derivative¹⁵ with this N-Troc-Thr-OBn and hydrogenolysis yielded acid 22. An HATU-mediated coupling of acid 22 with 23 led to the a linear precursor. Treatment of this linear precursor with 10% TFA led to removal of both terminal protecting groups in quantitative yield. Cyclization with BOPCl led to product 24 in modest yield. Unfortunately the Troc group could not be removed to yield 25. A variety of reducing conditions including Zn,¹⁶ Cd, ¹⁷ Cd/Pb,¹⁸ Zn/Pb,¹⁹ or SmI₂ ²⁰ would not promote the removal of this protecting group.

Scheme 5.

At this point we surmised that a protecting group on the tyrosine nitrogen might not be needed (Scheme 6). Tyrosine derivative 26^3d was esterified with Boc-Thr-OBn (21). This product was exposed to hydrogenolysis conditions to remove the Cbz and benzyl esters to form 27.

Scheme 6.

Coupling of acid 27 to amine salt 23 resulted in the desired product. Treatment of this linear precursor with TASF removed the desired silyl protecting group. Cyclization with BOPCl provided macrocycle 7 in modest yield. The Boc group of 7 was removed with HCl and the product coupled to the side chain acid 8 to yield Lys³ Tamandarin M (6).²¹ At this point the lysine residue will stay protected as Cbz carbamate until the biological probe is ready to be attached.

In conclusion a tamandarin/didemnin M analogue has been designed and synthesized. This product will be used as a biological probe in affinity studies to clarify the mode of action of this unique class of natural products. This is the first time a tamandarin analogue with an intact side chain that allows for attachment to a tag on a remote location on the macrocycle has been synthesized, leaving an intact side chain for biological studies. Results of this study will be reported in due course.

Figure 2.

Modifications of protecting groups