Total Synthesis of Carolacton, a Highly Potent Biofilm Inhibitor

Thomas Schmidt and Andreas Kirschning

DOI: http://dx.doi.org/10.1002/anie.201106762

Happy new year everybody… I start this year with a nice synthesis from my university done by a very smart student. In this paper a lot of cool metal mediated chemistry is employed which reminds me a bit of the work done by the Fürstner group at the MPI.

Some methodology needs as usual special attention and will be examined at the end of this short write-up.

As the title says Carolacton is a potent biofilm inhibitor which is nowadays a very interesting field to establish new drugs and natural products as biofilms play a major role in clinical treatments.

Retrosynthetically Carolacton was divided in two fragments of almost the same complexity. Beside the crucial macrolactonization step the linear precursor was envisaged to be formed by a stereoselective Nozaki-Hiyama-Kishi reaction. The key steps in the formation of fragment 1 contain an alkyl Negishi coupling and a cool aldol reaction developed by the Ley group. The second fragment was formed among others by a Marshall reaction and an underrepresented Duthaler-Hafner acetate aldol reaction.

Scheme 1

Starting from commercially available lactic acid 3 which was transformed into triflate 4 in a four step sequence through acetylation, ester formation, deacetylation, and triflate formation a zinc mediated SN2 coupling of triflate 4 with homoallyl Grignard formed ester 5. Thorough reduction of the ester group furnished after TIPS protection silyl ether 6. Ozonolysis of the terminal double bond was followed by reductive work-up and Appel reaction to give bromide 7.

Scheme 2

Ester 8, which was obtained by a short protocol developed by Fu et al., was coupled with the corresponding zinc organyl of 7 in the presence of PyBOX-ligand A and a pinch of nickel. Ester 9 is obtained with great diastereoselectivity regarding the red colored methyl group. It was found that it was necessary to grind the added sodium chloride to get good results. [1]

After reduction and manganese dioxide oxidation to aldehyde 10 the above mentioned Ley aldol reaction with ester 15 was employed. Aldol product 11 was obtained in good yield and excellent diastereoselectivity.

Scheme 3

Fragment 15 can easily be prepared through a three step sequence which is shown in detail below. Commercially available diol 12 was first protected as the dioxane ether 13. The 1,2-dimethoxy protecting group is used as a chiral memory unit because the stereochemical information at position 2 is lost during enolization. Chloride elimination and Ozonolysis of the resulting double bond gave after neutral workup ester 15.

Enolization with LHMDS then produces an enolate which reacts in a Zimmermann-Traxler boat conformation transition state with aldehyde 10. The whole story of this methodology can be looked up in the article which is cited above.

Scheme 4

The second half of the synthesis started from propargyl alcohol 16. Oxidation and hydride-transfer reduction with isopropanol catalyzed by Noyori’s catalyst gave enantiomerically pure alcohol 17. Mesylation formed the first partner for the Marshall reaction. The aldehyde partner was obtained from Roche ester 19 which was protected, excessively reduced and again oxidized. Next an anti-selective Marshall reaction mediated by indium(I)iodide and catalyzed by Pd(0) was used to give homo-propargyl alcohol 21. Desilylation with TBAF, acetal formation and chemoselective reduction then furnished after Swern oxidation aldehyde 22. This underwent another interesting reaction: a Duthaler-Hafner titanium mediated aldol reaction with t-butylacetate. This step will be detailed at the end of this post.

After reaction with Meerwein’s salt fragment 2 was almost ready for the Nozaki-Hiyama-Kishi reaction.

Scheme 5

First fragment 11 was freed from the chiral memory unit and the resulting diol protected as the acetonide. During protection the TIPS group was lost and the free alcohol then oxidized to give aldehyde 23.

On the other hand fragment 2 was exposed to Schwartz’s reagent and quenched with iodine to give vinyl iodide 24. In the presence of Cr(II), in situ formed Ni(0), and ligand B both halves were combined to give 25 and 26 as a mixture of diastereomers (~ 1 : 5). [2] Besides carrying the synthesis on with 26, it was tried using the wrong diastereomer 25 in a Mitsunobu process to close the macrolactone. Instead of closing the ring a SN2’ reaction was observed to give the vinylogous macrolactone.

Scheme 6

At last the acyclic precursor 26 was converted into the 12-membered lactone after saponification of the methyl ester and reaction with MNBA. Because of the delicate framework the t-butyl ester was transesterified with TESOTf and desilylated to give 27. PMB removal was followed by Dess-Martin oxidation and acetonide cleavage to yield Carolacton in good yield. [3]

Scheme 7

And as promised here is the mechanistic rationale for the Duthaler-Hafner reaction. Under standard aldol conditions i.e. LDA deprotonation of the acetate and subsequent reaction with aldehyde the diastereoselectivity can be predicted using the Felkin-Anh model. Here the wrong diastereomer is preferred. To overcome this substrate controlled reaction, the enolate is made chiral by using a stoichiometric amount of this furanose modified titanium reagent. After transmetallation of the enolate a Zimmermann-Traxler transition state can be formulated which is inherently controlled by the chiral ligands. This time the correct diastereomer is formed. [4]

Scheme 8

Nice work and worth a read.
[1] Interestingly the Negishi cross coupling can be conducted employing a racemic mixture of the chloride. Only one diastereomer is formed. The authors of this article (DOI: http://dx.doio.org/10.1039/B805648J) therefore propose a radical pathway.

[2] More about this cool stereoselective Nozaki-Hiyama-Kishi reaction can be found here: DOI: http://dx.doi.org/10.1021/ol0269805

[3] Interestingly the last deprotection step took 6 days. Pretty tough this acetonide…

[4] If you want to read more about this have a look in here,
DOI: http://dx.doi.org/10.1002/anie.198904951
Because the paper did not state exactly how the stereochemistry observed can be explained I tried it by myself. If anyone has a better explanation I really would like to hear it.

(+)-Sorangicin A: evolution of a viable synthetic strategy

Amos B. Smith III., Shuzhi Dong, Richard J. Fox, Jehrod B. Brenneman, John A. Vanecko, Tomohiro Maegawa

 DOI: http://dx.doi.org/10.1016/j.tet.2011.09.035

Sorangicin A is an extremely potent antibiotic acting against Gram-positive and Gram-negative bacteria with a MIC ranging from 0.3 to 25 µg/ml, respectively. It inhibits the RNA polymerase of bacterial cells leaving eukaryotic cells unaffected.

Besides a dioxabicyclo[3.2.1]octane, 15 stereogenic centers, and a tetrahydro- and dihydropyran ring, this 31-membered macrocycle contains a E,Z,Z-ester motif. The whole story of structure elucidation (and correction) is featured in the cited article. Because the whole is beyond the scope of this write-up interested readers should really have a look at it.

The completion of the synthesis was accomplished in 2009 but recently this full account appeared containing a lot of improvements. So here it is:

Scheme 1

 

This giant beastie is cut into four fragments. The two main fragments, blue and green, are of almost the same and much higher complexity compared to red and orange. It was planned to connect all fragments via two Julia-Kocienski, one Stille, and one Mukaiyama macrolactonization reaction. Synthesis of fragments 7, 21, and 33 will be presented separately before all were connected to form 36.

At the outset of the synthesis the acetonide of glyceraldehyde underwent a Cr-catalyzed Diels-Alder reaction with Danishefsky’s diene to give after addition of TFA dihydropyranone 2. Enantioselective Michael addition of in situ formed styryl cuprate and quench with methyl iodide furnished tetrahydropyranone 3. The temporary transmetallation with dimethyl zinc was necessary to suppress double ortho methylation of the resulting enolate. Selective ketone reduction from the less hindered face was followed by acetonide cleavage and trisylation to form alcohol 4.

 Scheme 2

 

In the presence of KHMDS epoxide formation was induced. Subsequent intramolecular epoxide opening produced the delicate dioxabicyclo[3.2.1]octane in high yield. Parikh-Doering oxidation and Takai-Utimoto olefination gave vinyl iodide 6 which was transformed into 7 by means of a dihydroxylation/periodate diol cleavage. Dihydroxylation is chemoselective with respect to the more electron rich double bond.

Scheme 3

 

The synthesis of the blue fragment started off with commercially available lactone 8 which was benzylated to give fragment 9. After elective monoreduction transformation of the resulting lactol into alkyne 10 was accomplished with TMS-diazomethane. The alkyne was used to introduce selectively a trans double bond through hydrozirconation. After some experimentation the group found that prior to in situ formation of Cp₂Zr(H)Cl (Schwartz’s reagent) in the presence of lithium triethylborohydride the latter one should be added first to form the alcoholate which no longer interferes with the hydrometallation step. Quenching of the metallate species with NIS then gave iodide 11. Parikh-Doering oxidation, reaction of methyl Grignard under Luche conditions, and another oxidation then produced ketone 12.

Scheme 4

 

The coupling partner of 12 was synthesized in four steps from amide 13. Aldol reaction of the boron enolate of 13 with the aldehyde shown gave aldol 14 which was transformed into Weinreb amide 15 prior to protection of the free alcohol with TBS and monoreduction with DiBAl-H to give 16.

 Scheme 5

 

Fragments 12 and 16 were combined through boron enolate chemistry to give aldol 17 in high yield but as a 1 : 1 mixture of diastereomers. Because separation of the diastereomers was easily accomplished both were transformed into pyran 19 through different routes. Desilylation with buffered HF was followed by BF₃ promoted thioacetal formation in the presence of ethyl mercaptan to give from diol 18a/b pyran 19 in high yield. In the case of 18b a two step oxidation/reduction protocol was necessary to invert the stereochemistry of the free alcohol.

 Scheme 6

Moving on with the synthesis vinyl iodide 19 was coupled in a neat alkyl Suzuki reaction with a 9-BBN boronate derivative to give olefin 20. Desulfuration of the thioacetal was followed by MOM protection of the free alcohol; triple debenzylation and selective mesylation of the resulting primary alcohol then gave mesylate 21. Acetonide protection of the vicinal diol and replacement of the mesylate with phenyltetrazolyl mercaptane produced after oxidation fragment 22.

 Scheme 7

 

The third and last main fragment was synthesized from known ephedrine derivative 23. Under Myers conditions amide 24 was produced in almost quantitative yield and excellent diastereomeric ratio. Reductive removal of the auxiliary and subsequent alkynylation under Corey-Fuchs conditions with concomitant methylation gave alkyne 25. Hydrometallation and bromination then furnished vinyl bromide 26 which was combined with dihydropyranone 27. The latter one was synthesized using again the chromium catalyst previously shown in scheme 2 through a Diels-Alder reaction with Danishefsky’s diene. In situ protection of the enolate with TES provided dihydropyran 28.

 Scheme 8

The TES enolate was used in a Rubottom oxidation to yield after deprotection/TBS protection ketone 29. Removal of the unwanted oxygen function was done through triflate formation with Comins reagent and palladium mediated reduction to give dihydropyran 30. PMB removal was then followed by two step oxidation of the free alcohol to the acid and tert-butylation to yield ester 31. Later on the group found out that the stereochemistry determined for the highlighted alcohol functionality was wrong. So an intermezzo of deprotection, oxidation/reduction, and global TBS protection was necessary to get the diastereomeric product. Selective TBS removal, Mitsunobu thiolation and oxidation then gave at last fragment 33.

Scheme 9

 

Assembling of fragments 7 and 22 via Julia-Kocienski olefination proved to be problematic. Considerable efforts were needed to get the reaction done. Optimal results were obtained by using tBuLi in the presence of HMPA as the base to give after several cycles product 34 in useful yield. Desilylation with buffered HF was followed by Dess-Martin oxidation to aldehyde 35 which was subjected to another Julia-Kocienski olefination. This time optimum results were obtained with KHMDS as a base to give after TBS removal ester 36 in good yield.

Scheme 10

 

Interestingly the group first planned to connect fragments 34 and 33 in the reverse direction by means of reacting the aldehyde of 33 with the sulfone of 34. Using the latter starting materials the group got besides a yield of 30 % a 1 : 1 mixture of olefins. [1]

Stille coupling of vinyl iodide 36 with stannane 37 which was obtained in one step following a known protocol gave after chemoselective saponification acid 38.

 Scheme 11

 

Macrolactonization proved difficult but after some experimentation the group found that good yield could be achieved using Mukaiyama’s modified conditions. [2] Deprotection of the tert-butyl ester was also problematic so the group first transformed it into a TBS ester. This was necessary because under standard acidic or Lewis acidic conditions the trienoate linkage was harmed. Global deprotection was done with hydrochloric acid to furnish in the end Sorangicin A as a single diastereomer.

 Scheme 12

 

This is really a huge achievement and with a yield of 3.2 % over 30 steps in the longest linear sequence not so bad J. It features some interesting chemistry namely the cool Diels-Alder reaction, the neat formation of the dioxabicyclo[3.2.1]octane, and the alkyl Suzuki coupling. Although the group first got the wrong diastereomer and proved later on that the original published structure was wrong the synthesis is straightforward.

THX for reading and comments are as usual welcome.

[1] I have never run a Julia-Kocienski olefination to date but I am sure this is only a matter of trial and error. Or does anyone have a good explanation concerning these results i.e. 30 % y, 1 : 1 versus 86 % y, only E?

[2] By employing standard Yamaguchi or modified conditions the sensitive trienoate linkage isomerized to some extent by a Michael addition of DMAP. Using a Mukaiyama reagent modified by Evans, the group worried about the iodide counterion which might react in a similar manner. Did anyone of you ran into similar problems during macrolactonization chemistry?

[3] Hopefully B.R.S.M. takes it as a compliment that I copy his style of putting marks in my text to direct you to the comments above…

And big big thanks to Bobby for reading all this more than twice. I promise this will be the last big one this year…

Identification of an unexpected 2-oxonia[3,3]sigmatropic rearrangement/aldol pathway in the formation of oxacyclic rings – Total synthesis of (+)-Aspergillin PZ

 Stephen M. Canham, Larry E. Overman, Paul S. Tanis

DOI: http://dx.doi.org/10.1016/j.tet.2011.09.079

Aspergillin PZ was isolated in 2002 by a Chinese group from a soil fungus and contains an interesting 12-oxatricyclo[6.3.1.0]dodecane ring system. Its structure is very similar to that of Aspochalasin D and Cytochalasin D. Aspergillin PZ displays some potential against various cancer cell lines but due to low supply of material an evaluation of the biological profile was not possible.

The retro scheme is shown below:

An obvious disconnection is the opening of the C-ring via an intramolecular Diels-Alder reaction (IMDA). The unsaturated side chain might be installed through a Suzuki cross coupling of a vinyl iodide which arises from Takai-Utimoto olefination of the corresponding aldehyde. Disconnection of the lactam then furnishes an aldehyde which could be produced through a sigmatropic rearrangement/aldol process.

 Scheme 1

The first few steps are only described in the supporting information. Starting from methyl pyruvate the first chiral center was produced using an asymmetric allylation reaction of in situ formed allylstannane. Allylation of the free alcohol in 2 gave ester 3 which was cyclized in the presence of Grubbs I. The resulting double bond was isomerized in the same pot. I am not entirely sure why they add CuCl. Epoxidation and regioselective epoxide opening with methanol was followed by TBS protection and ester reduction to give aldehyde 5.

If anyone has references to the first allylation step I would really like to see them. There is no comment to this reaction in the paper or in the SI. A quick Reaxys® search did not provide any useful information…

 Scheme 2

 

Stereoselective alkynylation under Carreira’s conditions (I think it is used in every synthesis I posted the last months…) gave alcohol 6 in very good yield. TBS protection and cis selective reduction of the triple bond reproducibly furnished pyran 7 in over 85 % yield. Sigmatropic rearrangement/aldol reaction was accomplished with SnCl₄ to give aldehyde 8. Unfortunately the reaction furnished the wrong diastereomer with respect to the aldehyde motif. It was not possible to identify a useful protocol to invert this stereocenter so the group went through a 7 (!) step reaction sequence.

TBDPS removal in the presence of a TBS group was done with TBAF. Pinnick oxidation of the aldehyde to the acid and esterification under standard conditions led to lactone 9. Having the internal anchor in place it was possible to get the inversion done with DBU as the base. Lactone reduction, chemoselective TBDPS protection and Ley oxidation then gave aldehyde 11 with the correct stereochemistry.

 Scheme 3

 

The group first planned to install the highly unsaturated side chain via some Wittig chemistry but all attempts were unsuccessful. So Takai-Utimoto olefination was chosen and followed by Suzuki cross coupling to give 13. Again TBDPS removal and Swern oxidation gave aldehyde 14 which was reacted with a lactam made from leucine and meldrum’s acid. DMP oxidation of the resulting alcohol furnished 16.

Scheme 4

What remained was the installation of the double bond which was accomplished through selenation/oxidation/elimination under standard conditions. Upon heating the IMDA proceeded to give pentacycle 18 in good yield. Debenzoylation and TBS removal gave at last Aspergillin PZ.

Scheme 5

 

The crucial rearrangement step is shown below:

Under Lewis acid catalysis the methanol is eliminated to give an oxocarbenium ion. A sigmatropic rearrangement then gave a silyl enol ether which undergoes an aldol reaction to give after hydrolysis aldehyde 8. The stereochemistry of this step should be a consequence of steric hindrance between the bulky protecting groups.

 Scheme 6

 

All in all a nice synthesis except the unproductive inversion protocol. The authors state that further studies are on the way to get the right stereochemistry in place.

And big thanx to Bobby for proofreading.