Monday, July 20, 2020

Tagetitoxin: Total Synthesis with Friends

Our recently completed synthesis of tagetitoxin was just published in JACS. I think it’s safe to say that this work is unlike anything we’ve done before in our lab. This is true not only from the standpoint of the chemical problems presented to us (see the paper, the SI, and some of the stories below), but also from a team standpoint, for everyone on this project is actually very good friends, both in and out of the lab. We therefore decided to use the space below for each friend on the team to write a little bit about their experience working with this beast of a natural product. Warning: long blog post ahead!

Hang Chu (Oct. 2016 – May 2018)
While searching for another molecule to pursue after the thapsigargins, Phil sent me Porter’s paper, which disclosed the most recently revised structure of tagetitoxin (tgt) at the time. Having primarily worked on terpenes at that point in my Ph. D., tgt seemed like a suitable molecule. As I recall, I wanted to work with something that had a nitrogen or two. Tgt certainly satisfied that requirement. As a bonus, throw in some oxygens, sulfur and phosphorous. Sure, why not? 

One of the toughest challenges initially was coming up with a reasonable retrosynthetic analysis for tagetitoxin. While there were not that many logical ways to take apart the molecule, finding an efficient and creative strategy to build the densely functionalized cyclopentane posed a challenge. In the SI you’ll find the evolution of our strategy as problems arose, so I will not belabor an extensive explanation of our logic. Our initial retrosynthetic analysis is shown below. Long story short, we foresaw three distinct challenges at the start of the program: 1. Core construction, 2. Cyclopentane functionalization and 3. Cysteine incorporation. 

 While there were many interesting snippets of problem solving in the first generation synthesis as disclosed in the SI, I’ll focus the two key reactions I had the pleasure to work on – 1) the oxidative furan rearrangement and 2) the thio-Claisen (see David’s section). 
To the best of our knowledge, there was no direct way in the literature to construct this type of amino hydroxy substituted cyclopentenone at the time (about 4 years ago..time flies). As the scheme illustrates, we had considered a couple of other ways to build this unusual cyclopentenone (double Claisen from a symmetrical aldehyde surrogate or an aziridine opening). These approaches were plagued with stereochemical issues and potentially harsh reaction conditions. We focused in on the β-hydroxy cyclopentenone portion of this building block and recalled that the Piancatelli rearrangement was an effective sequence to access these types of structure. Adding in the remaining functionalities took us back to a furanyl amino ester intermediate, which required the introduction of an equivalent of oxygen to get to the desired oxidation state. The problem-solving sequence to get this reaction to work initial was disclosed in the SI. Long story short, a non-basic reductant (dimethyl sulfide) was critical for this highly sensitive reaction to work. It has been a number of years since the discovery of this sequence so I’m a little cloudy on the details. I do recall isolating a wide range different retro-aldol initiated products along with a number of colorful decompositions. Also the use of the trifluoroethyl ester was essential to promote 1) good stability of the amino ester starting material and 2) the formation of the desired enone.


David Kossler (Mar. 2018 – Jan. 2019)
Since the synthesis of tagetitoxin was just published, I have the opportunity to reflect on my time as part of the team. After a short stint in the lab, working on method development, I joined Hang at the beginning of 2018. As we were office mates, I already had a sneak view on what was going on. By that time, he had found a neat way to build up the cyclopentane core via an oxidative furan rearrangement. As the stereoselective installation of the sulfur was problematic, he envisioned a relay via the allylic alcohol. It was clear from the beginning that this project was technically challenging in terms of sensitivity of the substrates, requiring careful adjustments of reaction conditions and workup procedures. Small deviations could lead to disastrous outcomes, and this happened. After we fixed a couple of issues in the first steps, the precursor for the [3,3]-rearrangement could be obtained in high purity. To our favor, once running the thio-CDI step in acetonitrile, the minor diastereomer from the Luche reduction reacted first with TCDI, and then cyclized intramolecularly, which made it easy to remove by column.

 Subsequently the rearrangement became a robust and clean transformation. This period was one of the most enjoyable, with both of us as a team trying to get as far as possible with the project. The dihydroxylation proved to be tricky as well, as the osmate cleavage led either to decomposition or protecting group swap without unveiling of the free alcohols. So we decided to install the sidearm first and then conduct the osmylation afterwards. All the details are in the comprehensive SI. Hang wrote his thesis alongside the lab work and graduated in mid-2018, upon which I continued with the project. Eventually, the free diol could be accessed, and the crystal structure confirmed the correct stereochemical alignment of all substituents. Big shout out at this point to the X-ray facility at UCSD, who perfectly resolved all structures I brought over during my postdoc, which was really helpful. 

A special moment was the first assembly of tagetitoxin’s skeleton. When analyzing the crude NMR, and in between all signal peaks, locating a set of multiplets that made sense for what we were trying to synthesize. Moreover, the comparison with the reported coupling constants for the natural product matched beautifully. With the trans-6,5 bicycle in hand, this was also the first proper evaluation point to see that the structure proposed by Aliev and co-workers was indeed the correct one. If the Js would have not correlated at this stage, it would have raised a big question mark that we were following the real structure of tagetitoxin.
 But in synthesis, for every problem you solve, you directly have the next step ahead of you, with the additional hurdle to bring material even further to the frontline. At this stage of the project limited material throughput became a real concern and hampered the advancement tremendously. I spent time on having a first look into the phosphorylation and the enantioselectivity challenge, but at some point the end of my Postdoc approached and I decided to go back from San Diego to Europe to finally enjoy some rain. Phil decided to hand the project over to Chi and Tom, who were about to finish the Herqulines. In order to not lose valuable knowledge and gained experimental details in this transfer, I assembled all available data and infos. Finally, in my last weeks we ran through the sequence together once. From then on, this well-oiled machine went on to make to target. Over the course of the last year I could still follow the updates and see the project evolving, one obstacle after another being removed. Looking on the finished sequence, it’s great to see how everything fell into place after extensive experimentation and many learnt lessons. I have nothing but huge respect for the final team in this relay run.

Tom Stratton (Jan. 2019 – Jun. 2020)
Fresh from our recent herquline victory lap, Chi and I were confident we would quickly shut the door on this project. We were dead wrong and spun our wheels for the better part of 18 months trying to understand what makes these molecules tick. I personally spent much of this time optimizing the key bromocyclization step, a reaction that will haunt me for the rest of my days. Since that story is (mostly) told in the SI, I will instead use this space to talk about how much I’ve learned from my good friends.

When I was a (highly) clueless first year grad student, Hang took notice. Right when things were starting to get truly ugly, Hang swooped in like a mama bird, brought me to his nest in BCC-439, and fed me tidbits of knowledge directly from his chemistry beak on a daily basis. He was a true mentor to me then and now, and a is also really strong dude.
I was assigned to be David Kossler’s (or Herr Doktor… show some respect) “temporary lab buddy” when he first arrived at Scripps from Europe. Lucky for us, the exact time component of “temporary” is ill-defined, as we still share cold beers together to this day, albeit from across the pond. The fact that we got to overlap (for one week) on this project was incredible as I got to see David’s remarkable precision, organization, attention to detail, and sense of team work first hand. We were lucky to have the wealth of knowledge gained by Hang and David at our disposable, mostly to the incredible work of my temporary buddy. Prost!
Kelly joined the project for a relatively short time, but in doing so, solved a problem that was a “non-starter” in terms of wrapping this project up. That is, stoichiometric osmium was required to effect dihydroxylation, and Kelly figured out the solution in literally a couple of weeks. This is no surprise. When Kelly was interviewing at Scripps, we recruited her hard. Not only did she possess several years of experience as a process chemist, she displayed the poise, focus, and kindness that is required to be a successful student and excellent teammate. Thus when my new hoodmate Kelly was recruited onto this project, Chi and I were absolutely elated and she stepped up to the plate big time. 
Dillon M. Flood has saved my ass more than once. Whether it be stuck on a 50sketchy left traverse at Big Sky, or at the outer reef of Teresa’s when my leash snapped and the waves were the size of a small cathedral, Dillon has been there to bail me out. Thus, it should be of no surprise that he rose to the occasion and contributed so much to this project in its final stages. We had intended to use a CRO to run the RNA polymerase assay, and when I asked Dillon for recommendations, he said “dude we just got a new plate reader and I can do this for you next week.” Having already showed Chi and I the ins and outs of anion exchange chromatography, he decided to bring his impact on this project to the next level. Sure enough, one week later he was sending us beautiful data depicting that (+)-tagetitoxin is the sole enantiomer active against E. coli RNAP. What a guy! I’ll see you up north, my friend.
There are simply no words that can describe how I feel about my partner Chi, but I’ll do my best. I doubt there is a more patient, humble, brilliant scientist out there than my friend Chi. Thank you for everything you have given me. I hope I have been able to return even a fraction of it. 
Finally, Phil is the best mentor I could have ever asked for. I truly feel like I won the PI lottery. Phil has a passion for exploring the unknown that is contagious and these ideas transformed the way I have thought about solving problems in chemistry forever. 

Kelly Eberle (Sep. 2019 – Jan. 2020)
I certainly feel lucky to have started in the lab at a time when there was still work to be done on Tagetitoxin. I knew that I wanted to do total synthesis, was interested in alkaloids, and that ideally I would work on an existing project to ‘get my feet wet’. When Phil suggested I join this team it seemed like a perfect opportunity. As a first year who had worked in pharma before coming to grad school, I thought I would know what I was doing. I can assure you that feeling went away very quickly. Thankfully I had joined a team of very intelligent chemists that were kind enough to mentor me and show me the ropes. Tom’s down-to-earth realness and Chi’s endless positivity helped get me through some of my darkest and most frustrating days as a first year. Working on this project also exposed me to several reactions I hadn’t run or even heard of before coming to grad school. Spending my first five months on this team taught me so much about technique, time-management, and chemistry in general; I’m very fortunate to have this foundation to build off of for the rest of my time in the lab.
Some key takeaways from working on TGT: Always be bringing material forward while you’re working at the front line. Compared to my past life of working in a process lab, scale up in academia is tough and the glassware is heavy. Ask your teammates questions (especially when things aren’t working) because there’s always more to learn.

Chi He (Jan. 2019 – Jun. 2020)
After finishing the herquline project, Tom and I decided to keep working on total synthesis. At that moment, our friend Herr Doktor (David) was going back to Germany to start his professional career. Tom and I took over the tagetitoxin project from Herr Doktor, not only because of its intriguing structure, but also our friendship with Hang and David. 

Hang and David had worked on this project for over 1 year. Moreover, David had confirmed the desired stereochemistry of bromocyclization product by X-ray in the last week before he left (see above). To be honest, this good news made me and Tom very excited to think we could put a bullet to this project soon. In two short weeks, however, we would come to understand just how many more problems awaited us. In general, we modified and optimized the whole route in the past 1.5 years (for more experimental details and logics, please see SI). Here, I would like to share my experiences and feelings from 2 late-stage procedures: phosphorylation and the final purification. 
 Des-P tagetitoxin was easily generated by convenient workup after exhaustive hydydrolysis by Ba(OH)2. We were excited to find the chemical shifts and J-values in 1H NMR of des-P tagetitoxin were significantly similar or identical to the original data in the isolation paper, which indicated the structure we were working on was promising (non-trivial given the structure of tagetitoxin had yet to be confirmed). It seemed we could put a bullet in this project  in just one more step! However, the final selective phosphorylation of C8 alcohol failed after several attempts with chemical conditions. The failure was probably due to the poor solubility of des-P tagetitoxin in organic solvents. Although it might be possible to achieve the phosphorylation with the help of enzymes, we decided to pursue this final step using chemical methods. 
 To improve the solubility in organic solvents, we protected the free amino alcohol as the hemiaminal. After methanolysis, 18’ was obtained as an ideal substrate for phosphorylation, with two masked methyl esters and hemiaminal to improve the solubility and stability. To our surprise, a lot of conventional phosphorylation conditions did not work on this substrate due to either poor chemoselectivity or low reactivity. Fortunately, our lab had already developed a practical method to achieve the chiral phosphorylation by a talented graduate student Kyle. With the help of Kyle, we successfully put on P(V) with highly reactive (+)-Ψ reagent in excellent yield. The reaction condition is mild and highly reproducible. I want to highlight that the advantages of this method were not limited in phosphate installation, but also helpful in crystallization and chiral resolution. We confirmed the structure and absolute configuration through the X-ray of 19a and obtained both (+)-tagetitoxin and (–)-tagetitoxin in two more steps.
 I still remember the first time I got tagetitoxin. I saw a huge rainbow outside of window in early morning of 2019 Thanksgiving. After taking a 1H-NMR of this crude mixture, I was excited to tell Yuzuru I might have made tagetitoxin! One day later, I emailed Phil and my teammates about this exciting news after confirming by 13C-NMR. That moment, I thought we could finally put an end to this project and enjoy the Christmas holidays. However, it was just a beginning of the most challenging problem I experienced during this project: the final purification. 

The natural product contains so many polar functional groups on a tiny skeleton, which makes it much more polar than peptides or even palau’amine. Attempts to purify with preparative HPLC were proved intractable as the highly-polar mixture eluted at very beginning. Luckily, our friend Dillon, another talented graduate student, gave me a hand at that moment. After trying several different purification methods, we found that anion exchange chromatography with DEAE resin was optimal. However, the only way to confirm the purity was by taking NMR of every fraction after lyopholization (LC/MS, for example, was not sufficient to determine purity). It was common that a single purification procedure could take several days, only to realize at the very end of the sequence that we had again failed. I forgot how many anion exchange columns I had tried, definitely >15. In most cases, the impurities came from the product and they were co-polar. Further, it was impossible to attempt re-purification from this mixture. The only solution was to carefully scale up a new batch of tagetitoxin and try a new purification. After several failures, we realized tagetitoxin is extremely sensitive to acid, as decomposition was observed at pH 4 for 30 min. After fully understanding this nuance, and combined with lots of effort and advice from Dillon, we finally collected clean NMR spectra of tagetitoxin! All told, the final purification took us almost half a year, finally enabling completion of this project. It was very challenging and at times frustrating for me to spend such a long time to deeply understand how fragile and tricky this molecule is. On the other hand, I am also very proud of our courage to overcome the array of challenges that were presented to us!

Finally, the most impressive thing for me in this project is the great cooperation between friends. I would like to thank the support and help from Hang, David, Tom, Kelly, Dillon and Kyle. The project would not be done without any one of you!! 🍻

Dillon Flood (Jan. 2020 – Jun. 2020)
Over the past few years, Tom and I have had countless conversations which generally revolved around south pacific storm cycles, where we were chasing swell that weekend, and who would pick up the Tecate before we left. Although never at the forefront, chemistry would bubble up through these interactions. The world of total synthesis inhabited by Tom and Chi felt foreign to me, so the usual griping about failed experiments, unrealistic deadlines, and long hours always seemed to need a bit of translation.

I’m not sure what initially provoked Tom and Chi to ask for my help with an anion exchange purification but it very well could’ve been during a collegial discussion on right way to drink one’s preferred Mexican beer. However it happened, Tom and Chi were showed up on the first floor of the Beckmann, bright eyed and bushy tailed, taking copious notes, and very ready to purify their compound that I didn’t want to ask about how long it took them to make… Although this should’ve been a routine method, it never is. This led to Chi spending better parts of some days in the Dawson lab perfecting his anion exchange chromatography, while always expounding on the genius of LeBron James, just so I was sure. 
At some point I made the mistake to ask Tom (while surfing at Blacks), what this compound even does? After a long story about phytotoxins, RNA polymerase, optical rotations, isolation chemists without compound, and a half-baked idea to spray plants (?!), all I could think at the time was that I needed another wave. But out of this conversation, and a little googling, came the idea of performing the in vitro assay on both enantiomers. At first, the pair was not so enthused when I mentioned the assay. But when I mentioned that I could run it for them, everyone was on board. And then again, Tom and Chi were back down on the first floor of Beckmann, compounds in hand, ready for me to tell them which enantiomer was active. At this point I had to ask them to leave, unless they wanted me to watch me mess this thing up. So after loading the plate to the Hamilton soundtrack, running the assay and reading the results, I was shocked the experiment had worked so well. One compound was active and the other was dead. But this was the beauty of working at Scripps, friends in totally different fields could halfheartedly bounce ideas off one another that may help resolve and unanswered question. 



2 comments:

  1. I saw your HNMRs in D2O and think that NMR technic WET for solvent peak reduction will be usefull for you.

    ReplyDelete
  2. I enjoyed reading the Total Synthesis of Tagetitoxin. What an accomplishment!. Congratulations! I have worked on a similar densely functionalized cyclopentane contained natural product Pactamycin.

    ReplyDelete