B.R.S.M. The road to Tet. Lett. Is paved with good intentions



Enantioselective Total Synthesis of (-)-Acetylaranotin, a Dihydrooxepine Epidithiodiketopiperazine

Reisman et al., J. Am. Chem. Soc., 2011, Early View; [PDF][SI][GROUP]

DOI: 10.1021/ja209354e

This week saw another brilliant synthesis from the still fairly new Reisman group over at Caltech, this time an epidithiodiketopiperazine (ETP),[1] the group's first. These interesting secondary metabolites have so far only been isolated from fungi, and owe their toxicity to their disulfide bridges that generate reactive oxygen species by redox cycling. Although ETPs have been popular targets for the last 40 years, and the field has seen some impressive chemistry from Kishi in the 70s to more recent efforts by Movassaghi, Overman and Nicolaou. More challenging still are the dihydrooxepine containing ETPs such as acetylaranotin, which has stood unconquered since its isolation 1968, and in fact no synthesis of such a compound has been reported, until now.

As the disulfide bridge was judged to be (redox) sensitive, its installation was envisaged near the end of the route. The C2-symmetric diketopiperazine onto which it would be attached could be formed by dimerisation of a suitable amino acid that was to be constructed using a 1,3-dipolar cycloaddition to form the pyrrolidine and a 7-endo cycloisomerisation to access the dihydrooxepine.

The route began with the 1,3-dipolar cycloaddition between t-butyl acrylate and the azomethine ylide of the cinnamaldimine shown (itself easily prepared from the condensation of cinnamaldehyde and ethyl glycinate). The use of catalytic copper along with the brucine-derrived ligand gave the expected endo-pyrrolidine in an excellent 96% ee, although only 50% yield. Cleavage of the t-butyl ester using TFA, followed by Teoc protection of the amine gave the carbamate. Ozonolysis of the styryl group and subsequent treatment with ethynylmagnesium bromide unfortunately formed the lactone with the undesired configuration at the new stereocentre. To correct this, instead of allowing cyclisation to occur under the reaction conditions the hydroxyacid intermediate was isolated and cyclised to the lactone under Mitsunobu conditions with inversion at the alcohol.

This lactone was then reduced to the diol using sodium borohydride leaving the ethyl ester unaffected. Next, it was necessary to protect the more hindered secondary alcohol over the nearby primary alcohol. This was accomplished by protection of both, followed by cleavage of the primary TBS ether using a mixture of acetic acid, THF and water. In order to install a halogen which could later be eliminated to  introduce the second dihydrooxepine olefin, the primary alcohol was then oxidised to the aldehyde, α-chlorinated and then reduced back to the alcohol. Although a little redox inefficient, the yields for these steps were all near-perfect, and the group was then in a position to examine the key cycloisomerisation they'd planned to use to install the oxepine. Thus, after a bit of experimentation, the group found that heating the alkynol with catalytic [Rh(cod)Cl]2 with tris(4-fluorophenyl)phosphine as a ligand in DMF gave the desired tetrahydrooxepine in great yield.

The intial plan for construction of the diketopiperazine core involved elimination of the chloride followed by deprotection and dimerisation. However, due to problems with selective deprotection and then with the diketopiperazine formation itself this approach had to be abandoned in favour of a stepwise union of two different fragments. Thus, the acid containing fragment was obtained by a slightly unusual hydrolysis of the elimination product with trimethyltin hydroxide in 1,2-dichloroethane. Interestingly, selective cleavage of the Teoc group in the amine containing partner could only be smoothly carried out on the chloride, meaning that elimination had be performed second. These two halves were then united without incident by amide coupling using BOP-Cl.[2]

Global desilylation was then achieved using TBAF•(t-BuOH)4 in acetonitrile. This also had the effect of causing both spontaneous cyclisation and a double epimerisation of both diketopiperazine methine positions. This reaction had to be conducted under an inert atmosphere using degassed solvents as the presence of oxygen resulted in oxidation of the two amide α-positions to give the syn diol. Investigation of this unexpected process is apparently ongoing, although the fact that it could be effected by the modified TBAF is interesting, as complexation to t-BuOH ligands significantly reduces its basicity. Fortunately, this process did seem to indicate that reactions of the diketopiperazine occured with high diastereoselectivity under simple subtrate control, offering hope for stereoselective dithiolation. This was done under slightly unusual conditions based on those used recently by Nicolaou.[3] The yield for this step was modest, but given the numerous acidic protons and the sensitive dihydrooxepines successful functionalisation at this late stage was still impressive. Finally, the two hydroxyl groups were acetylated, and the tetrasulfide was reduced to the dithiol, which reoxidised to the disulfide in air, completing the natural product.

1. ETP review in Microbiology for free here.

2. Not to be confused with BOP, which is entirely different, and bit less good in that it produces HMPA.

3. This certainly took someone a lot of optimisation! From the SI:

To a stirred suspension of S8 (22.4 mg, 87.3 μmol) in THF (5.8 mL) under a N2 atmosphere was added a solution of NaHMDS in PhMe (0.6 M, 437 μL, 0.262 mmol) over two minutes. The S8 solids dissolved over the course of the addition, giving initially a dark green solution that evolved to an orange-brown color towards the end of the addition. While this solution was allowed to stir for an additional one minute, a solution of NaHMDS in PhMe (0.6 M, 145 μL, 87.3 μmol) was added to a stirred solution of diketopiperazine 26 (3.13 mg, 8.73 μmol) in THF (1.0 mL) under a N2 atmosphere. The resulting pale brown solution was then added dropwise via syringe to the previously described solution of reagents over one minute, rinsing once with THF (0.25 mL) (rinsings transferred over an additional one minute) to give a slightly opaque yellow/orange solution. After an additional one minute, a solution of NaHMDS in PhMe (0.6 M, 292μL, 0.175 mmol) was added dropwise to the reaction mixture. After 50 min, the reaction was quenched with sat. aq. NaHCO3 solution and extracted with EtOAc.


Comments (12) Trackbacks (0)
  1. In the ozonolysis the DCM went in at the beginning, and DCM was also included in the chlorination step.
    Isn’t DCE dichloroethAne?
    Footnote 4 should be 3, &, as stated in that footnote, the THF went in at the beginning of the thiolation.
    The diketopiperazine is definitely not flat as shown (both in your entry & the paper!).
    Why was triethylsilane included in the de-t-butylation step?

    • Typically one includes Et3SiH (or anisole or thioanisole, sometimes) to trap carbocations generated in deprotection processes (e.g. the PMB cation, or in this case, the tert-butyl cation).

    • Sorry for a slightly slow response – very busy trying to finish some work I’m presenting at a conference next month. Those are all fair points, I’ll fix the schemes now. HPCC is spot on for your question; couldn’t’ve explained it better myself

  2. that thiolation procedure is nothing more than intellectual plagiarism…looks just like Nicolaou’s prep in his synthesis of epicoccin G (below). It probably took the student a few days to tweak the conditions, nothing too fancy. Unimpressive.

    To a suspension of sulfur (436 mg, 13.58 mmol, 8.0 equiv) in THF (8.5 mL) at 25 °C under argon was added NaHMDS (0.6 M in PhMe, 8.49 mL, 5.09 mmol, 3.0 equiv) dropwise over 2 min. During the addition the insoluble yellow S8, turned to homogeneous dark blue, then dark orange, and finally light orange solution. This solution was stirred for an additional 1 min and bis-diene 7 (500 mg, 1.69 mmol, 1.0 equiv) dissolved in THF (8.5 mL) was added dropwise at ambient temperature over 2 min, at which time the reaction mixture turned light brown. The mixture was stirred for an additional 1 min, then more NaHMDS (0.6 M in PhMe, 5.66 mL, 2.0 equiv) was added and the resulting mixture was stirred 0.5 h at 25 °C. The reaction mixture was quenched with sat. aq. NH4Cl solution (50 mL). The mixture was extracted with CH2Cl2 (3 × 25 mL) and the combined organic layers were dried over MgSO4, filtered and concentrated. The resulting brownish residue was used for the next step without further purification.

    • I always wondered how one can measure 8.49 or 5.66 mL when a normal syringe of that size has ticks every 0.1-0.2 mL only…

    • Hey Bengu how many times did Nicolaou have to bang you to get you to write that? Huh? The molecule is different man!

    • I’m not really seeing that intellectual plagiarism as Reisman explicitly states

      ‘These conditions were adapted from Nicolaou’s recent modification of Schmidt’s sulfenylation procedure: Nicolaou, K. C.; Totokotsopoulos, S.; Giguere, D.; Sun, Y.-P.; Sarlah, D. J. Am. Chem. Soc. 2011, 133, 8150.’

      and references the Nicolaou paper. I mean, yeah the experimental write up is pretty similar, but it’s the same reaction. Reisman and coworkers are just being good scientists and recording what they did; they’re not trying in any way to take credit for the method. Is it plagiarism every time I do a named reaction, or repeat a literature prep and write it up in my lab book in a similar way? After all, why do we even publish experimental?

      P.s. Also notice in my post I credit Nicolaou and say it took someone some time to work out.

    • As a by-the-way, Nicolaou’s now published a paper on this sulfenylation in angewandte: http://dx.doi.org/10.1002/anie.201107623

  3. I am a little late to the party but I hope I can still get a response.
    Can someone explain to me why the reaction in the third step with the grignard does not result in an addition to the lactone?

    • Late response: the substrate contains a hemiacetal group, which is in equilibrium with the free aldehyde and alcohol. At lower temperatures, reactions are slower and therefore more selective. This means that the acetylide can preferentially add to the aldehyde over the ester and lactone, since it’s more electrophilic.

      • Sorry for the spam, an additional note: I meant to say that this particular hemiacetal is in equilibrium with the aldehyde and carboxylic acid. The acetylide deprotonates the acid and makes it much less electrophilic.

Leave a comment

No trackbacks yet.