Yes, there’s finally a new post on this blog. But this is not the post you’re looking for…

It’s time to wrap things up with this blog, which is a really tough thing to say. After 365 posts, and more than 8000 (mostly on-topic) comments, this a chapter of my life I don’t really want to give up. But the practicalities life and my career don’t really permit it. This website was really active for about four years, and I’m proud to say that we (and this really is a “we”) did some amazing things on here. We witnessed the rise of a new generation of synthetic pioneers (Rychnovsky, Baran, Shair and so many more), and unfortunately lost some before their time (Gin [1]). We got stuck into some really great science, and frankly, bitched like we were on the 200th fraction of the column that wouldn’t die. But more than all of that, we changed the face of scientific debate [2].

Another generation of bloggers [3] are taking the baton and running with it, pushing the debate further and faster. Traditional academic publishing is on a serious catch-up mission, and we’re all benefiting from this. Power to the people, and all that!

So why close the blog? Well, even though I work for a huge chemical company, it’s been over a year since I went into an actual lab. I don’t own a labcoat, nor lab-glasses. And even though I do read the odd paper, I simply don’t have the fluency in organic chemistry anymore [4]. I still work in R&D, and still need to think like chemist (occasionally), and certainly don’t feel I’ve disowned my education [5]. But it doesn’t feel right.

So what to do with the blog? Amazingly, the traffic to this site isn’t decreasing. It seems that the content lives on, and that my witterings are still informative and helpful. So I’m not planning to pull the plug – at least not in the near future. Any suggestions on how to preserve the content – please do add a comment.

Oh – and the photo is of my brother and I doing science, 1986-style.

[1] I’m a huge admirer of David’s work. I came this close to Post-doc’ing with him…
[2] We also gave one scientist in China a really awful week. I do still feel quite guilty about that.
[3] Sheesh I’m old…
[4] I’m blaming learning a new language (Dutch) on that. Limited brain capacity and all that…
[5] I’m lucky to be the manager of the small but talented informatics department, based in Holland, near Leiden.

Zhang, Chen, Chen, Xie. ACIEE, 2011, 51, 1024. DOI: 10.1002/anie.201106587  

Second time around for Stenine on TotallySynthetic – the first synthesis I blogged takes us back to 2008, and Aube’s neat work focusing around a tandem Diels-Alder / Schmidt reaction. This latest publication moves the research to Kunming, China, the home of it’s use as part of Chinese herbal medicine.  That doesn’t alter the core of the synthetic strategy, though, as Hongbin Zhang seems to agree with Aube that building the cyclohexane core first is the key to this target.

Zhang, however, prefers the use of a double Michael addition, enhanced with a little catalyst control to engender asymmetry in the system.  Using Evans’ work, a catalyst derived from (1R,2R)-(-)-1,2-diaminocyclohexane was employed in conjunction with some silica-bound KOH as well as ultrasound.  These conditions aren’t exactly typical, and Zhang doesn’t postulate much in the way of reason.  My feel is that the sonication is to break-up silica-KOH particles to provide maximum surface area for the heterogeneous reaction conditions, but I’m not speculating any further than that!

A strange-brew it may be, but it delivers the goods in excellent yield and control of asymmetry.  Four new stereocenters constructed with (apparent) complete diastereomeric control is no mean feat, but the next reaction proves their strategy to be a winner.  Under reductive conditions, the nitro-group is  activated and forms not just the azepanone but the pyrrolidine too, completing the core of the target in two steps.  The change in stereoelectronics also seems to cause a tautomeric shift, as the enol now seems to prefer a beta-keto-ester existance.

So often this is the point in an otherwise tasty synthesis where the wheels come-off as the synthetic team have to faff-about with functional group transformations and carbon chain homologations to get to their target.  However, Zhang gets to stenine quite neatly, firstly by nuking the remaining methyl ester under Krapcho conditions.  An alkylation with ethyl bromoacetate adds the carbon required for the required gamma-lactone, and a bit of gentle reduction delivered the final ring.

Of course, they’re not quite there – alkylation of the lactone with methyl iodide provided the carbon skeleton in reasonable yield, whilst the extraneous oxidation on their medium ring with dealt with by firstly forming the thioamide using Lawesson’s reagent, and reduction with Raney nickel delivered the product.

Neat stuff, but I have to say, awkwardly written.  Of course, my Mandarin (or Cantonese for that matter) is non-existent, but I expect better in Angewantde.  Perhaps, though, it’s because the writing in Aube’s paper is just so nice!

Li, Deng, Zhu, Lu, Yu. JACS, 2011, ASAP. DOI: 10.1038/nchem.1196  

Just take a quick look at that target and consider that the synthetic route I’m about to summarise took thirteen steps.  I’m fairly stunned – I read the top-line number in the abstract, and immediately thought that they must have started with an advanced intermediate or degredation product, but no – the synthetic action (like many natural product syntheses) begins with Citronellal.  Sure, only three of the rings are carbocyclic, but there’s a lot going on here, so lets get into it.

As I said, the synthetic starting material is a fairly large pot of tasty-smelling citronellal.  It’s not exactly cheap, (about £100 for 5g from S-A), there are more expensive entries to this terpenoid skeleton. They spend the first four steps (which they somewhat cheekily don’t count as it’s all known chemistry) mashing the isoprene group into a protected hydroxyl group, and then do a pair of Horner-Wadsworth-Emmons reactions to firstly build the triene system, and latterly bolt-on the 1,3-dicarbonyl moiety.  Interestingly, you’ll note the thioester group – this was to make the prior carbonyl chemistry a bit less dramatic, and make the subsequent chemistry very neat.

A little cooling and addition of everyone’s favourite Lewis acid with weird punctuation resulted in a very neat intramolecular Diels-Alder, building four stereocenters, three tertiary and one quaternary in a yield anyone would be happy with.

Whilst that is a very neat reaction, Diels-Alder cycloadditions are so 19th-century; what happens has to be at least early 20th century!  The 1,3-dicarbonyl moiety is, of course, quite acidic, so an allylic displacement reaction could be encouraged by addition of Ag[TFA] and a little base (transesterification conditions).  The reaction initially occurs by O-alkylation, forming a furan system, but the equilibrium of C-alkylation to O-alkylation could be transformed by treatment with Pd(OAc)2 and electron-rich ligand PBu3.  I’m guessing a little, but I suppose that formation of a Pd-[allyl] intermediate allows this facile interconversion to occur.

Okay, the three carbocyclic rings are in, but we’ve still got two heterocycles to build.  The teams next step towards that was formation of an amide by displacement of the fluorinated ester chain with a derivative of D-serine.  The remaining terminal olefin was oxidised (after a little experimentation) under Wacker conditions, and the carbonyl reduced with some selectivity to the secondary alcohol.  Lastly, this final intermediate was treated with sodium methoxide, promoting a Dieckmann condensation – the base again deprotonates the remaining proton on the 1,3-dicarbonyl, which attacks the methyl ester.  The immediate product is the cyclic keto-lactam, but the ketone is attacked by the proximal secondary alcohol, neatly forming the last ring and completing the target.

There was one catch, though – which you may have noticed.  The team unfortunately made the enantiomer of the target, but as the spectral data matched with opposing optical rotation, I think we should let them have their prize!

Baran, Mendoza, Ishihara. Nature Chem., 2011, EarlyView. DOI: 10.1038/nchem.1196   

Hmm… I’m still not convinced about whether I should really post on this paper, but I’m interested in it, so to hell with the rules!  What we’ve got here is a sort-of meta-synthesis; Baran doesn’t actually make any of the taxane natural products, but demonstrates an extremely neat synthesis of a potential common precursor.  The theory behind the work is related to previous Baran diatribes, where he points out that too many steps in syntheses simply oxidise and reduce the same carbons.  However, in this case Baran describes the Taxanes (which are of course heavily oxidised) as being parented by a simple (almost) un-oxidised ‘precursor’.

This beastie is known as ‘Taxadienone’ – and as the name suggests, it contains a pair of alkene groups and a sole carbonyl as its only oxidation. However, deeming it a ‘precursor’ is perhaps a little ambitious, as to get to taxol or related natural products, a lot of oxidation of unactivated C-H bonds is required.  In the hands of your average chemist, this seems a pretty insurmountable challenge, but with Baran, perhaps…

Baran’s approach to the complex 6,8,6-ring system, as ever, appears fairly simple.  Start with a cyclohexenone, do a pair of carbon-carbon bond formations about the alkene, and then unite these sidechains in a neat cycloaddition.  And the interesting thing is that Baran makes it work just as easily as that.

The first, and more complex, sidechain was appended using an organocopper 1,6 addition.  This combination of diene with dienone looks set to generate a black tar in my hands, but team Baran got a remarkable 86%.  Then using some of Alexakis’ chemistry, they methylated this product in a 1,4 addition to introduce asymmetry.  Using a chiral phosphoramidite ligand, this was achieved in a very reasonable 89% yield and 93% e.e., using pretty low loading of the copper thiophenecarboxylate and ligand.  I really like the simplicity of this approach – keeping the system symmetrical until this point reduces complexity considerably.

Having trapped the product of the methylation as it’s TMS ether, the group then did a Mukaiyama-type aldol coupling with Acrolein.  This was a little awkward, as the group often found the ketone product of desilylation rather than their desired aldol product, even when rigorously excluding water.  Bizarely, adding water helped – and by raking to the back of the Lewis acid cupboard in the lab, they found success with the unusual gadolinium triflate.  These are apparently Kobayashi’s conditions, and gave them a yield of over 85%, but as a 2:1 mixture of diastereomers.  However, they didn’t separate the isomers – rather, they chucked the lot into the Diels-Alder below.

Under more typical reaction conditions, these reaction fairly nicely, giving a reasonable yield of product, but perhaps more importantly (in a sense), the ability to separate the isomers with ease.  They weren’t quite done – a triflation of the cyclohexenone carbonyl at the only acidic methylene position in the molecule generated a enol triflate.  Reaction of this under Negishi conditions with dimethyl zinc provided them with the required methyl cyclohexene moiety – and in a satisfying 84% yield.

Synthesis done, and like me, you may well be wondering what the excitement is about.  However, continue reading the paper into the discussion, and the chemistry takes on a different note, as Baran describes various strategies that were uncooperative.  These included a fail aldol approach to closing the medium ring as a later stage, and a stereochemical nightmare involving a Shapiro reaction.

More importantly, Baran believes that site-selective oxidations of taxadienone are possible, and may lead to a synthesis of taxol form this intermediate.  Now that’s a paper worth getting excited about!

Carreira, Weiss. ACIEE, 2011, EarlyView. DOI: 10.1002/anie.201104681   

Now that’s a busy polycyclic ring-system!!  Three five-membered rings, two six-membered rings and one seven (depending, of course, on how one counts the ring sizes.  Anyone read the IUPAC Gold Book last night?) – adding up regardless of maths to one hell of a synthetic challenge.  And there’s even some token biological activity to aim for – this bad-boy has some moderate vasorelexant acitivity next to rat aorta.  (When the paper mentions ‘moderate’ activity, we can be pretty sure it does very little indeed…) But who knows – perhaps one of it’s derivatives or intermediates could be more biologically interesting.

What we’re interested it is of course the synthesis – and it’s a one-man job in this case. That man by the fumehood is Matthias Weiss, and he started the chemoenzymatic or chromatographic separation of a diethyl succinate derivative.  With no Supporting Information to hand, I’m unsure how he went about that – it’s always a disappointment when leading PIs like Erick Carriera neglect the scientific process… Lets just hope that it was high-yielding and amenable to scale!  A bit of mono-protection of this C2 symmetric intermediate, followed by enol triflation gave them a partner for a rather neat B-alkyl Suzuki cross-coupling, appending the silyl-protected propanol side-chain.  Unfortunately this reaction require the use of Triphenylarsine – not the most pleasent of additives – required to suppress reductive detriflation.

The product was then treated with an array of Group-13 organometallics, firstly diastereoselectively hydroborating the alkene, and then reducing virtually everything to give a triol product in excellent yield.  What I like about this procedure is the apparent simplicity in functional group transformation.  A few steps manipulating protecting groups lead them to an intermedate that was O-alkylated with the corresponding tosylate then took them to the precursor to two Claisen rearrangements.  The first is perhaps the more obvious – the alkenes are sitting the perfect place to be tickled with a little heat to encourage carbon-carbon bond formation, and diastereoselective formation of a congested tertiary – tertiary system.  Next, they treated the resulting ketone with base and 18-crown-6 to O-alkylate again, this time with allyl bromide.  Again, heating was used to simulate rearrangement, neatly forming the quaternary centre in three steps.

Eight steps further on (including a neat Henry condensation and a stereoselective methylation), and the team were ready for a spectacular formation of the octahydroazulene domain.  Working from an alkyl-iodide and looking for coupling with the cyclopentenone, the group had a lot of possible conditions to consider.  However, what they went forward with was rather unusual – a cobaloxime-mediated Heck cyclization, requiring irradiation to proceed in good yield.  Carreria states that this method development will be published in another paper, but I haven’t seen anything just yet.  Looks very interesting though –  and delivers a stunning yield in this challenging system.

With the medium ring closed, the (one-man) group were pretty close to the finish.  An intramolecular aldol codensation was used to install the remaining cyclopentene (requiring two reaction cycles to achieve a 77% yield), whilst de-BOCing the amine with TFA was enough to enough to prompt imine formation, and completion of the final ring. Removal of final protecting group completed the synthesis – one that cleverly balances new methods with classical techniques.

A bit of history…

At the beginning of 2007, a start-up (which was actually almost ten years old) called Arrow Therapeutics was purchased by AstraZeneca, the Anglo-Swedish BigPharma multinational. Arrow’s focus was on anti-infectives, primarily anti-virals, complementing AZ’s purchase of Cambridge AntiBody and US-based biologics giant, Medimmune. Arrow was definitely pocket-money in comparison to Medimmune, but still a significant purchase. Not long after the purchase, Arrow went-a-recruiting, and, amazingly, took on me along with another young chemist. I must say, it was a bit weird that many of the staff ‘knew’ me through these pages before I’d even interviewed, but the process was mercifully short in comparison to some of the multi-round affairs that many BigPharmas operate.

Less than three years later, things at AZ weren’t looking too hot [1, 2, 3 – I shan’t blame the messenger, Derek!]. The AZ Strategic Review was announced – but I have to say that I felt confident that they wouldn’t drop Arrow, as we’d bested all targets and delivered some tasty compounds. Also, buying a company and then selling them on three years later just looks dumb. But I was concerned about my career. Lay-offs were certain, adding to the pool of qualified and experienced medicinal chemists looking for work in the UK, especially after GSK had more-or-less closed their Harlow site. No-one else in pharma seemed to be recruiting, and certainly not at the team-leader level, which is where I wanted to be. Looking around, I really couldn’t see this happening for years – and frankly, that wasn’t what I signed up for. Having a job is great, but a career is what I wanted.

Come Christmas 2009…

… and my feeling of malaise about the Pharma industry reached breaking-point. Surfing the usual recruitment website – as one does just to see who’s hiring – had shown bugger-all in med-chem for months. So I decided to try again – career 2.0 – and applied to work alongside friends from my undergraduate days, in a different industry. And that’s where you’ll find me now – Project Leader with AkzoNobel in Global R&D. It’s not pharma, and I don’t wear a labcoat very often these days – I’ve got a team and a robot for that – but thinking like a chemist is still extremely important. The door to pharma isn’t closed for me, but I’m happy to sit the carnage out. Perhaps I’ll reconsider in a decade or so, but I’m enjoying my job and career with AkzoNobel too much to care for now.

So, Total Synthesis? Some might think that I’m now too far away to still be interested – but it’s no different to working in med chem. Sure, I did some pretty interesting reactions, but a lot of it was the Suzuki and amide bashing it’s thought to be. Total synthesis – for good and bad – is another world entirely. One that I still find fascinating and that keeps my brain (as an organic chemist) firmly in the fume-hood (HSE aside).

And Arrow?

Well, we can still count on Big Pharma to make the wrong move. Yep, AZ decided to ‘divest Arrow Therapeutics as a Going Concern‘. In other words, they were shifting the company and hoping not to look too bad about it. I was still there for part of the ‘divestment efforts’ (whilst my move to AkzoNobel was going on in the background), and it was frankly torturous. AZ couldn’t decided which parts of Arrow’s IP were to move with Arrow, or remain with AZ. Months of shuffling paperwork ensued, and even though there were many interested parties, the doors closed on New Year’s eve for good. Bastards. Many of my former collegues have found employment; other are working their entrepreneurial arses-off, looking towards new horizons. And as these are some of the smartest people I’ve had the pleasure to meet, I’m sure it’ll work out well.

What did I learn?

I guess no-one reading this would disagree with me in my feeling that Pharma is on a serious down-turn just now. The reasons for this are multiple, but for me it all comes down to accountability.

In every R&D sector, projects & programmes move from early ‘I wonder what happens when…‘ science, through to identifying a commercial basis for a period of development to finally launching a product or service. What makes Pharma stand out to me is the length of that process – it’s quite routine these days for time-frame from idea to pill to be ten to fifteen years. So who is accountable for the NCE development over that timespan? Normally, there’ll be a Biological team, who try to develop some sort of assay that can be used for in vitro testing, giving the chemists those IC50s (or similar). Once there’s some sort of robust efficacy test, it’s time to do a bit of screening (perhaps a few million small molecules, or ten-thousand fragments) to give the Lead Identification team something to work with. After those chemists have played with the results for a bit, it’s off to the Lead Optimisation team, whose goal is to refine the structure into a Candidate Drug. If that looks good, and doesn’t kill too many things that go squeak, the Development and Formulation teams take what was a chemical sample into a drug, through the familiar Pre-Clinical and Clinical trials, FDA sign-off and $$$.

The point is, the project is passed from team-to-team over many years, each of which accepts the project from the previous team in a different state. Of course, there’s also an Exec-team (perhaps a Review Board) who are following the progress, and analysing results to determine which projects get to pass through the stage-gates and onto the next step in development. Sounds great, but there are some pretty large elephants in the room at this point. Firstly, none of these Execs is hanging around for fifteen years – five sounds a lot more reasonable. So it’s not too surprising that when a new project comes in from biologists and LI team, it looks all shiny and amazing. Perhaps three years later, it gets passed to the LO team, and by now some of that shine has worn-off, “…but it’s still viable, right? Sure, there are probably some other players in the field by now, but we’re going to be better and faster, right?” So the LI team does it’s job, and gets us 100g or so of something pretty decent. Fluorine atoms hanging-off it in random places, morpholine groups, benzimidazoles and quinolines have been tacked-on, “…but it’s got a great IC90. Who cares if the mice died? The dog didn’t…” So it’s off to the pre-clinical team, and it now looks like Rival A and Rival B have options that are both better and further ahead. But “We’ll push this through, after-all, we don’t have anything else to take it’s place…” And maybe it makes it to the end, and gets launched, but only in a few markets, and it’s neither first nor best in class. More donkey-derby runner-up than prize-winning thorough bred. But who can be held accountable? There won’t be anyone left on the Review Board who remembers the project in Lead Identification. Ten-fifteen years is simply too long for any person or group to be responsible.

And it’s not like these Review Boards are effective anyway. Helping lame-duck projects through the stage-gates is all too common – “after all, we need to keep the number of CDs, Phase I and Phase II candidates up, right“? Most of the review meetings are exercises in Group Think, as no-one wants to be the person to kill a project that might have legs. And they don’t care if it all goes wrong, because they’ll be in a new job, at another company, long before the chickens come home.

It’s a number game

That actually touches on one of the more ridiculous aspects of Big Pharma – setting Targets for R&D teams. It’s fairly common for Med-Chem teams to have a quota of CD’s that they should produce every 1/2 years. But it’s nonsense, as there’s no way to quantify the quality of those results – so any old shit can get shiny’d-up so that it’ll pass the Stage Gate, even though most of the people involved know that’s not quite there. And this attitude gets all the way through to candidates entering Phase III trials that aren’t anywhere near ready. Remember Torcetrapib? Surely someone knew that this wasn’t a winner – at least before Pfizer spent all of that $1B + on it’s development. And this is why I think attrition rates keep increasing.

Of course, the Execs running the show aren’t idiots – there are a lot of very smart and sharp people in those boardrooms. But the problem is again numbers. How do you quantify quality? After all, if the quality can be kept high, and the quantity can be worked on, then Board’s on to a winner. This is more-or-less the idea behind the (Lean)6-Sigma, and AstraZeneca tried to apply it to varying degrees of success. The problem is, how does one define or quantify the quality of a candidate? Guidelines, such as Lipinski’s rules, can be used to judge the ‘crazyness’ of an idea, but can also limit the creativity of the medicinal chemist. But do they really help, or do they mean that most LO sets are average compounds, with average properties?  I know a fair bit about 6Sigma – I’m hopefully going to get my Green Belt this week, and train for a Black Belt next month.  But it applies really well to our R&D processes, unlike those of Pharma.

However, if there’s one thing worse that spending a billion on developing on a few tonnes of solid organic waste, it’s being sold a lemon. In particular, paying way too much in M&A (mergers and acquisitions) – reading an article about M&A failure is a history of big Pharma over the last few decades. Just look at Pfizer’s stock over the last ten years – it’s halved in that time from over $40 to $18.15 today. Buying-a-Biological-company was a trend that I think all small-molecule chemists watched with some concern, as some of those pipelines looked a little light for the price. Looking back, it’s now clear that there was a degree of desperation to fill product pipelines with something potentially generic resistant, and perhaps to keep-up with the Jones’…

Silver Bullets

Everyone in Pharma is well aware of the fact that there’s no such thing as a ‘Silver Bullet’ – a drug or therapy that has no side-effects and targets the source of disease infallibly.  But for some reason, the management teams seem to think that the process of drug discover can have it’s own silver bullets – new ideas or approaches that will magically solve the pipeline problems and save them from the patent cliff. There was combinatorial chemistry in the 1980’s, diversity-oriented synthesis in the 2000’s along with fragment-based drug-discovery, in-silico and virtual screening…

All are fantastic ideas, and have their relative merits, but for some reason Big Pharma leaps upon these ideas and chucks the baby out with with bathwater.  Surely the appropriate way to introduce a new technology or ideology is to adopt it portion-wise, rather than jumping in with both feet?

So, Pharma Quo Vadis?

I’m a long way from being the first person to voice these opinions, but seeing some of these issues first-hand was both enlightening and traumatising. So how does Pharma fix itself?  I’ve no idea – but from my point of view, these can’t hurt:

1. Content Quality is king – but relying on a spreadsheet to tell you where the good stuff lies will always result in an average compound.  And an average compound is never going to be best or first in class.
2. You can’t buy a pipeline – the overheads in M&A and the loss of productivity will always sabotage the ‘synergies’ that were desired.
3. Separating the compound design chemists from the synthetic chemists isn’t a great idea – you need time for ideas to evolve, and running another column is good time to do that. Plus, it really doesn’t inspire high morale in the split =-team.
4. Make sure the project Review Board are actually reviewing something.  Endless pre-meetings and pre-pre-meetings will only reduce the opportunity for a worthwhile decision making process.  And make sure the board are packing collective balls.

That’s it.  It certainly made me feel better.  But my opinion is neither broad nor particularly deep – so let me know what you think!

Reisman, Cha, Yeoman. J. Am. Chem. Soc, 2011, ASAP. DOI: 10.1021/ja2073356 

A busy couple of weeks in the Reisman lab, it appears – really nice syntheses of the Cepharatines and 8-Demethoxyrunanine in Angewandte (doi: 10.1002/anie.201104487), but I decided to blog this tasty synthesis of Maoecrystal Z.  Of course, this isn’t the first member of that family that I’e blogged – the seminal synthesis of Maoecrystal V made my December 2010 column in Chemistry World.  I remember when I wrote that piece that the pseudo-3D structure used to represent the target annoyed me – in far too many papers, the author uses such a structure when a normal 2D sketch would do.  However, trying to ‘flatten’ Maoecrystal Z or it’s family members is an exercise in mental self-harm – the best I could do is shown below, and is frankly ugly.

So what’ll do is stick to Reisman’s figures and hope that you can follow along…

The synthesis starts with γ-cyclogeraniol, initially TBS protecting the primary alcohol, and then epoxidising the exo-cyclic alkene using mCPBA.  This takes them to the substrate for a rather ‘witches-brew’ type reaction, coupling the epoxide with methyl acrylate.  The conditions were inherited from Gansauer, first published in his 1998 paper on reductive epoxide couplings (look at the hyperlink to see a reminder that Wiley used to be crap with DOI codes!), requiring the zinc as the terminal reductant.  Unfortunately, the initial results were relatively poor – only a 28% yield was possible.  However, by moving to 2,2,2-trifluoroethylacrylate rather than methyl acrylate, and adjusting the rate of addition of catalyst, the group were able to boost this to a very respectable 74%.  Especially nice when you consider how direct this reaction is!

Working directly from the product of that reaction, the group coupled on a sidechain by alkylation, and added a pair of stereocenters.  One came with the sidechain, and was it’s self constructed using a pseudo-ephidrine controlled alkylation (and I’ve got a feeling I’ve made this substrate myself, many years ago…).  The other wasn’t controlled, and was later removed by oxidation of the lactone to an α,β-unsaturated lactone using the tried-and-tested phenyl-selenation and oxidising elimination.

After this, it was time to move to the pièce de résistance, a samarium (II) iodide mediated radical coupling.  Treating the advanced intermediate with SmI2 initially resulted in disappointment, as only trace quantities of the product were identified.  However, utilising lithium chloride and tert-butanol as additives saved the day, allowing the group a very reasonable 45% yield, with cracking stereoselectivity.

Bouyed with this success, the group then went made a similar substrate, but this time with an aldehyde instead of the protected primary alcohol.  When this substrate was treated in the same manner, not only did a further ring formation occur, but the yield was improved to an even more impressive 54%.

More importantly, though – this pair of C-C bond formations and ring formations took the group remarkably close to the target in a single step. From here, the acetate group (actually, two…) was installed, followed by ozonolysis and treatment of the aldehyde with Eschenmoser’s salt and triethylamine to get to an enal.  They the (slightly) selectively removed the erroneous actetate, completing the target in twelve steps.

Has to be strong candidate for synthesis of the year!

Baran, Su, Rodriguez. J. Am. Chem. Soc, 2011, ASAP. DOI: 10.1021/ja206191g

When does a person or groups work in a particular area of total synthesis become pedestrian, or even dull? This may sound harsh (especially as he’s a really nice guy), but Paterson’s (at Cambridge) work on macrolides isn’t doing it for me any more. Conversely, Baran’s work in the area of pyrrole-imidazole alkaloids is still facinating, even though I’ve blogged about it more than a few times! What we’re looking at here, though, isn’t quite newly conquered territory; rather, it’s an efficient smoothing-over of some the bumps along the way to Baran’s previous synthesis of the Axinellamines.

Have a look back at my previous post on the Axinellamines (can’t believe that was more than three years ago!!), and perhaps the post on Palau’amine. Although the chemistry is pretty amazing stuff, the step count definitely leans towards the arse-end of the alphabet. This has clearly stuck in the collective throat of the Baran group, and they address the their strategy towards a key intermediate in this paper. The key precursor in these syntheses is a spriocycle, produced with no control of stereochemistry at that central position. The initial work, which can be followed in my post from 2009, or in more detail in this 2010 J. Org. Chem paper, takes twenty steps to get to that intermediate, and not without a lot of chromatography.

So, in other words, the group were set the reasonable challenge of improving upon that situation. In the latest paper, the synthetic action begins with a rather neat Pauson – Khand cycloaddition of a bis-allylic TMS ether and Boc-protected propargylamine. Except I don’t think you can buy 2-Butene-1,4-diol bis(trimethylsilyl) ether – I think one would have to make it. And since trans-2-Butene-1,4-diol is really expensive, I expect that they had to make it from the corresponding acetylene. Now, this might not look like an issue, but this reduction is a pig of a reaction – I should know, as it was the starting point for a lot of the work I did in my DPhil. Basically, there isn’t a good solvent for this reaction – the chemistry only happens to the small portion of the SM that goes into the THF. I used start with about 30g of the SM, then syringe in three 100mL bottles of LAH in THF, and then heat the crap out of it over the weekend. Getting the product out of the aqueous layer was quite a battle, I remember, even after drying out the reaction.

However, (and if you’ll permit me to drone on like this), I do remember one weekend where I didn’t have to vac the reaction down at all. Y’see, I’d bunged up the three-necked flask I was using with septae, syringed in the LAH and set the heater stirrer, and then buggered off to the pub. When I came in on the Saturday, I couldn’t see into the flask… and on further examination, I realised that I’d left the damn septae in, which had swollen and then popped-off the flask. The THF, of course, was gone, leaving about 12g of still-quite-active LAH caked onto the sides of the flask. Not a fun day…

Any-ho, I survived, and the Baran group have got their 2-Butene-1,4-diol bis(trimethylsilyl) ether from somewhere. What’s important is that the Pauson – Khand they do is really nice, setting up that trans-bis-ethanol type system required of their target intermediate. However, getting this reaction to behave took quite a bit of effort, as both ethylene glycol and NMO were required to get a sensible yield. My reading of the paper suggests that the group aren’t entirely sure of why these conditions are so effective, but based on past efforts from the team, I expect they’ll publish a neat explanation at some point.

Next up, they did a Luche reduction, which also stripped the TMS protecting groups, giving them a triol. This was then treated with N-chlorosuccinimide and a little triphenylphosphine to effect a substitution of all three alcohols with chloride. The group then planned a desymmetrizing Barbier coupling of an aldehyde to install a sidechain, but again came unstuck. Using typical conditions, only a very modest yield could be acheived, but moving a slightly bizzare combination of Zinc and Indium with ammonium chloride gave them the goods in great yield. They’ve done all the right control experiments – it’s not indium chloride, and they do need both metals – but they’re again still figuring this one out in the lab. Impressive result, though – setting two stereocenters and working very directly.

Treating the bis-chloride with sodium azide (nasty!) did the expected displacement, whilst deboccing with a little TFA followed by guanidine installation. This took the group to the pivotal spirocycle formation – which went without stereocontrol in their previous route. However, they them stumbled upon yet another interesting set of conditions, as the chlorination reaction was strongly encouraged by trace trifluoromethanesulfonamide remaining from the previous reaction. Thus, employing a little TfNH2 as catalyst, and tert-butyl hypochlorite as the chlorinating agent, the spirocycle formed in excellent yield, and crucially, as a single diastereoisomer. NMR studies performed by the group suggest that the trifluoromethanesulfonamide isn’t acting on the t-butyl hypochlorite, but on the substrate…

Oxidation of the initial spirocyclic intermediate, and a little more TFA, took the team to the key intermediate targeted at the outset, but this time in eight steps from the starting materials described. Not only was the yield increased markedly, the amount of chromatography required was significantly reduced. Neat. The paper then describes the remaining steps to get to the natural products (see that earlier blog post), but from what I read, nothing has changed significantly (not that it needed to).

Great work – anyone looking for a couple of great Process chemists should call the chemists above!

Garg, Huters, Quasdorf, Styduhar. JACS, 2011, ASAP. DOI: 10.1021/ja206538k

Technically beaten to the finish-line by Rawal (JACS in March), but still the first asymmetric synthesis of the Welwitindolinone family, this synthesis is one of many contributing to a hell of a year for Neil Garg.  I think the key to the synthesis was picking the perfect starting point – and dealing with poor yields that get you seriously further forwards.

That critical starting point is  fantastically smelling (S)-Carvone.  Not only does this make the lab (and presumably the chemist, his notebook, laptop, wallet…) smell great, but it also gives an asymmetric entry into the synthetic campaign.  Working from a previous synthesis of Hapalindole O (by Natsume in 1994 – PDF), a referenced synthesis appends a vinyl group and adds a stereocenter.  This was neatly done over five step (but tracing the yields was an exercise in exasperation, so I apologise!)

Cleavage of the pivolate group with a little base set them up for a major fragment coupling – appending that indole group.  The team did this using an iodine promoted alkylation, presumably by formation of the halonium ion followed by attack by the bromo-indole.  The yield might not be great, but this is a fantasticly succinct coupling strategy.

TBS protection (why didn’t they just keep the pivolate?  Did it prevent the coupling reaction from going well?) then set them up for the next critical coupling.  The medium ring was installed using the surprisingly simple approach of an indolyne cyclisation (have a look here to read more about indolyne, but think of it as a benzyne derivative), using just a little sodium amide.  So the base does two things – it forms the reactive aryne intermediate, and it forms the proximal enolate by deprotonation.  However, this kind of reaction always works better on paper than in practice, as the two different processes probably have different thermal requirements.  Attack of the O-enolate to form an ether was a competing reaction in what wasn’t the highest yielding process, limiting the team to a 33% overall yield – but yet again, this a was an impressive step-forward.

The latter intermediate is now looking pretty much like the target, but two key features are missing – the vinyl-chloride group, and the bridgehead nitrogen.  Let’s start with the earlier and easier transformation, the chlorination.  Taking the protected alcohol, deprotecting and oxidising with DMP took them to the corresponding cyclohexanone, which was deprotonated and treated with Comins’ reagent.  This generated a vinyl triflate, which was then converted to a vinyl stannane via some palladium catalysis.  Lastly, treatment with copper chloride (as below), gave them there vinyl chloride.  Nice.

Now for what looked like the more challanging move – converting the bridgehead methine into an amine.  Reduction of the ketone with DiBAL-H and formation of a carbamate was the easy bit, but necessary, as intermolecular amidation was impossible.  The group then intended to cyclise onto bridgehead, but of course, there are two bridgeheads, both neopentyl, so a selectivity issue appears to beckon. Looking at the (ultimately successful) reaction conditions shows a real witches brew of stuff – but we’re definately heading in an oxidation direction.  What acually happens under these conditions is formation of a nitrene, followed by C-H insertion – smart.

Quite a tour-de-force of chemistry here, using an impressive manner of techniques.  Congrats to the group!

Romo, Liu. ACIEE, 2011, 50, 7537-7540. DOI: 10.1002/anie.201102289

Guest Blogger: SPF

What if I’d ask you to synthesize this little molecule with six neighbouring stereocenters from biologically available material? Oh, by the way you can’t use any protecting groups and keep it short! Sounds impossible? Not according to Romo et al.. They’ve achieved the synthesis of this stereo-dense molecule in 10 steps from (R)-carvone (Mmmhhh spearmint). (+)-Omphadiol belongs to the africanane (guess where they grow) family of sesquiterpenes. Its brothers and sisters show some significant bioactivity, but due to insufficient amounts from extraction of natural sources, no further study was possible. There’s hope to change that with some synthetic effort. The synthesis starts off with a previously reported, Mn^+III catalyzed, chemo- and regioselective, formal hydration of the enone in (-)-carvone. The resultant mixture of α-hydroxy ketone diastereoisomers was then cleaved to ketoacid with periodic acid.

Activation of carboxylic acid with tosylchloride and use of 4-pyrrolidinopyridine as nucleophilic promoter enabled aldol lactonization to a bicyclic β-lactone. The high diastereocontrol is explained by a chair-like transition state, where pseudo-equatorial is the best, no-clash, position for the isoproprenyl substitutent.

They then solve the problem of the required four carbon homologation quite elegantly, by reducing the β-lactone to the diol, converting it to the bromide, employing an intramolecular alkylation and quenching it with methyl iodide (Puuhhh…, that must have taken some time to figure out).

The original idea was to reduce the lactone, add vinyl magnesium bromide and use of a metathesis reaction to couple both olefins together. A subsequent Simmons-Smith would then give them their product. But surprise, surprise their product didn’t match NMR data. A crystal structure confirmed that they’ve actually made the 5-epi-product. This is due to a magnesium alkoxide on C9 forming an 8-membered metallocycle with the C5 Aldehyde, giving the undesired stereochemistry.

So, don’t use the diol? That proved to be troublesome and was only possible by finding a way around the DIBAL reduction/Grignard sequence above. The solution was the addition of an allyllithium which they had to make from transmetallation with allyltriphenyltin before adding the lactone. High yielding tandem isomerization/RCM of the diene, without any unfavoured 8-membered ring formation, afforded the enone.

After that, they regio- and stereoselectively reduced this enone to the right allylic alcohol and a Simmons-Smith cycolopropanation later they finally were holding (+)-omphadiol in their hands.

All in all, they’ve finished this sweet total synthesis in 10 steps, 5 of which are C-C bond forming, from (R)-carvone in 18% yield. Plus, they didn’t use a single protecting group to construct the 6 adjacent stereocenters.

(Editorial – another cracking Guest-Post, this time by SPF. Remember, if you’d like to give it a go, just contact me!)