About this site
Here we'll review recent developments in drug discovery and medicine and the IP issues and financial implications they have, along with general thoughts about research. Also likely to make an appearance: occasional digressions into useful topics like which lab reagents smell the worst.
About this author
Derek Lowe, an Arkansan by birth, got his BA from Hendrix College and his PhD in organic chemistry from Duke before spending time in Germany on a Humboldt Fellowship on his post-doc. He's worked for several major pharmaceutical companies since 1989 on drug discovery projects against schizophrenia, Alzheimer's, diabetes, osteoporosis and other diseases.
To contact Derek email him directly.
The post below will be the last for another day or so. I'll be gone at a short conference, and won't be posting anything tomorrow night. See everyone on Thursday!
Next, the Magic 8-Ball. . .
So, what keeps drug company researchers up at night? Well, plenty of things. Wondering if their latest lead compound is going to get through its two-week toxicity testing - that's always good for a pillow-pounding interval. Wondering if they're going to even find a compound that's worth a two-week tox in the first place (there's a lot of that around, too.) But the last few years have seen one particular worry, and a close cousin, grow larger and larger: How do we know if the target we're working on has anything to do with the disease? And where is the next target going to come from if we get rid of this one?
That gets back to the question of where targets come from, in general. Well, in general, they come from
reading the Journal of Biological Chemistry. Whoops. I mean reading Science or Nature. Ahem. Well, actually, that is where some of them come from. When a really good paper comes out in a big journal, it's like a starter's gun going off, and the race is on. But we're also supposed to come up with some of our own. And we've come up with so many over the years, that the fear of diminishing returns has well and truly set in.
One solution drug companies have been trying is to invest heavily in genomic research. It's been several years now since the genomic gold rush was in full swing. And unfortunately, unless I'm missing something (and my friends around the industry assure me that I'm not,) no one has exactly struck a mighty vein of new drug targets in the genome yet. Actually, we all began to suspect that we were in trouble when the Human Genome Project estimated that we have a mere 30,000 or so genes. That meant that a lot of the interesting stuff was happening post-genomically, in the ways that genes are transcribed and the ways that proteins are synthesized and modified. The idea of "One gene - one target" is wrong more often than it's right.
But the genomic work still generates a lot of ideas. They're just not very well-worked-out ones. You can see various genes that seem to be associated with different diseases, but they're things that no one understands. They code for enzymes that might as well be called "Whateverase." It can be very hard to figure out what these proteins do, and what relationship they might have to the disease you're trying to treat. (At a bare minimum, you have to wonder if the changes you're seeing are a cause of the disease, or just an effect of it.)
So a hot topic these days is target validation. Everyone's looking for ways to sort things out, to cut through the clutter and find the good stuff (which had better be in there somewhere.) These could mean tracking subtle changes in cell cultures, or trying to mess with protein expression via things like RNA interference, or even finding new ways to monitor the physiology of whole animals. The same techniques might be useful for finding new targets without even going to the genome
So there are companies, more companies than you can count, selling RNA tools. And there are outfits springing up that advocate animal models beyond the simple mouse and rat - how about zebrafish? They grow faster, and they're transparent; you can see exactly what's going on. Others are developing methods to track all the changes in some particular class of cell constituents (lipids or the like,) or to ever-more-closely analyze blood and urine changes in live animals.
Do any of these things actually work well enough to help us out? I've no idea, and (frankly) neither do the people at all these small companies. But everyone's in there pitching. There's a lot of money waiting for whoever can come up with a Drug Target Validat-o-Matic, that's for sure. I'm watching to see if anything looks promising, but it's too soon to say about most of the really interesting stuff. If anyone has some favorite approaches to this problem, I'd be more than happy to hear about them. (I can see the brochures hitting my mailbox now!)
Big As Life and Twice As Ugly
I called the Roche/Trimeris drug Fuzeon (T-20) "one insane molecule" last week, but perhaps I shouldn't have been so hard on it. True, it's a whopper, but it's made out of amino acids, which are pretty well-defined linkers. Making T-20 is just ("just," he says) a matter of hooking them together. Of course, if you do that one after the other, you're in for a long haul. One hundred and six steps worth, to be precise, which means that even if each one of them works great, you're still going to take terrible losses over the course of the synthesis. (You're a lot better off making it as independent subunits that you stitch together at the end. Not that it's still any fun whatsoever.)
So T-20 has its crazy aspects. But there are plenty of compounds out there that are crazier. Natural products, for example - that's a catch-all term for molecules that are made by living systems, but aren't things that are essential to their metabolism. It includes a lot of antibiotics, famously the penicillin types and things like erythromycin. It also accounts for a lot of anticancer agents, as well as a rogue's gallery of poisons. (Those two categories overlap a bit more than everyone would like for them to, actually. . .)
The last bunch, the toxins, can be especially weird. These are things like strychnine, and things that make strychnine look like hot cocoa. Some of them are small and fairly simple (like anatoxin and muscarine) but the structures can really get out of hand. Marine natural products are famous for their head-banging complexity: take brevetoxin, for example - there's a nightmare for you. A even more hideous example is palytoxin.
As unlikely as it seems, both of those compounds have been synthesized from scratch. Of course, these syntheses are about as uneconomical as it's possible for chemistry to be. They're massive, complicated one-offs, assembled by teams of post-docs and graduate students, assisted by a swarm of sweating lab-Sherpas. If the survival of the human race depended on a steady supply of palytoxin, we'd be in the soup for sure.
The closest example of a compound like that which is actually on the market is taxol, or paclitaxel, as Bristol-Meyers Squibb would prefer that I call it. That's another natural product, this time (as most folks know) from the yew tree. Unfortunately, the yew does not produce enough of the compound to assure any kind of supply (especially when you start peeling off the bark to get the taxol!) So a synthetic approach was intensely sought.
Not that it can't be made. These days, we can make just about anything, given enough time and money. But those two quantities, which are equivalent in much the same way that mass and energy are, are both in short supply in the drug development world. Taxol was synthesized, sure, but in the Cecil B. DeMille cast-of-thousands style, which wasn't of much practical use. But the real breakthrough came from Robert Holton's group at Florida State, who perfected a route that starts with another yew-tree natural product. This one comes from the needles, though, and that's a renewable source with a higher yield. The tree does a lot of the really funky chemistry on its own, and my fellow chemists take it from there.
I'll have more to say about the utility (or lack of same) of synthesizing these monster molecules from the ground up. I speak from experience, since that's the kind of chemistry I did my PhD on. But things have changed a bit since the 1980s, though. Haven't they? At some point this week we'll take a look and see.
Dosing Schedules, For Fun and Profit
I mentioned the problems with having a drug that has to be injected, but there are problems with drugs that have to be taken orally, too. What if your drug is taken twice a day, and someone else comes up with a similar thing that only has to be taken once? You can guess who's going to end up with more of the market. Patients like it because of the convenience, and physicians like it because patient compliance is better.
That means that, conceivably, someone could come up with another peptide drug like Fuzeon, one that attacks the same protein it does, but only has to be taken once a day. That would probably take over whatever market share Roche and Trimeris end up enjoying, but I don't see it happening.
For one thing, I'm not aware of anything like that coming up in the clinic. For another, you can bet that the two companies put time, effort, and money into trying to get something like Fuzeon to work just that way. There are a number of tricks used to make peptide drugs last longer in the blood, and the folks at Roche know them just as well as anyone else. What we're seeing is probably about the best they could do. (Another sign of trouble is how large the dose of Fuzeon is - twice a day, at 90mg a dose. That's pretty substantial for any compound (although, to be fair, antivirals and antibacterials typically run pretty high.) But it's a gigantic amount for a synthetic peptide.
Getting back to oral dosing, it's widely held that once a day is the best. Twice a day will work if the drug can be taken morning and night, and three times a day will work if it can be taken with meals. Above three times a day, you can forget it. I can't think of a drug off the top of my head with an oral dosing schedule like that, but if there is, it must have some major problems.
Weirdly, taking something less often than once a day can be trouble. Patients lose track of what days they're taking the stuff. The best solution is to package the pills in some sort of calendar-associated way, as are done with birth control steroids, but that's extra trouble and expense that a company will avoid if it can.
All these dosing schedules are determined by how well the compound gets taken up into the body, where it circulates, and how long it hangs around. The industry acronym for all this is ADME, for Absorption, Distribution, Metabolism, and Excretion. Enough drugs have come to grief on these rocks, and at expensively late stages in their development, to stimulate a whole side industry of people trying to predict and understand what's involved. It's safe to say that anyone who manages to do that reliably will do very well for themselves indeed. And, not coincidentally, very well for the people who need medicines that actually work. It's Adam Smith in action - hey, if you look sharp, you can almost see the Invisible Hand. . .
What Things Have Come To
Fuzeon (T-20,) a new anti-HIV drug from Trimeris and Roche, is one insane molecule. As I detailed a year ago on Lagniappe (see the August 8 post here), it's a huge peptide. Instead of making it through a biotech approach (engineered bacteria,) they're making it the way someone like me would have to: sheer organic chemistry. It's surely the largest synthesized compound ever to hit the market as a drug. The synthesis is one hundred and six steps long. I have to admire the nerve of the people who signed off on this method.
But audacity might not be enough. Since it was approved back in the spring, Trimeris hasn't been selling as much of the compound as they'd hoped for. For one thing, they'd been having supply problems, which didn't surprise anyone who'd looked at the synthetic plan. But the worry is now that they may end up having a bit more of the compound than anyone wants.
Another problem is the price. Whenever we're developing a new medicine, the cost-of-goods analysis is a key step. How much is it going to cost to make the stuff? That includes the starting materials, all the solvents and reagents along the way, the costs of running the plant that makes it, etc. This is where process chemists earn their pay, finding new and cheaper ways to produce the molecules. (Generally, the way that folks in discovery research (like me) make them is completely uneconomical.) But there's no getting around the horrendous cost of making Fuzeon. The size of it, the loads of starting materials required, the number of chemical steps - it's a nightmare. Roche and Trimeris built a whole specialized factory in Colorado just to do it. All this has helped to make a year's worth of treatment go for at least $20,000. That's complicated things on the managed-care end.
But what could be the biggest problem is the way the drug has to be given. The trade newsletter BioCentury, in a recent piece on the drug, said "TRMS is not planning to develop an oral version,. . ." a statement that qualifies as very dry humor indeed. No peptide that size could be dosed orally; your digestive tract would treat it just the way it treats a steak. So it's intravenous dosing all the way, and twice a day, at that. Now, ten years ago, that might have not meant quite as much. But now there are a lot of orally administered HIV therapies, and more are on the way.
It's well-accepted in the industry: the only areas where an injection-only drug can reliably succeed are the ones with no other treatment options. Type I Diabetes, for example - if your body has stopped making insulin, you have no choice but to take it, and there's no way to take it other than by injection. (Pfizer and Inhale Therapeutics have been trying to change that, but it's been a very tough process) Many kinds of cancer fall into the same category. An effective treatment for, say, pancreatic cancer could still make it even if the only way to dose it was on the end of a sword. There just aren't any other options.
But HIV, fortunately, isn't quite in that no-hope category any more. It's been a remarkable stuggle to get to that point, and there's still a lot more to be accomplished, but I think that Fuzeon's difficulties are illustrating just how far things have come. Just imagine going back to the mid-1980s and trying to convince people that a new therapy for HIV infection would someday be having trouble because it's more difficult to take. All the competition on the market, you know. They'd have put a net over you and hauled you off.
Geron has been around a while. But they haven't had many days like they've had today. They presented at an investment conference, gave an update on some interesting cancer therapy, and the stock went up like a startled frog. Hey, if it keeps this up, it'll get back to where I first bought it. (Upfront note: Yep, I'm a shareholder, and have been for several years now. I first bought into the company in the mid-$30s, watched it zip up in the craziness of early 2000, and watched it zip right back down again. I bought more on the way down. They call me a long-term investor, among other things.)
But fifteen dollars a share is pretty good going for a stock that was under two back in the spring. I can't resist doing the idiotic calculations about where I'd be if I'd put all my liquid net worth into the stock back in March. Of course, anyone who puts all his liquid net worth into a single stock, and especially a single biotech, needs to be taken down with a tranquilizer dart and banded with a radio transmitter for his own protection. But still.
So from whence flows all the market craziness? It's Geron's telomerase program, which they've been banging away on for some time now. Telomeres, as most everyone will have heard about in some form, are structures located at the ends of chromosomes. They provide a way for the enzymes of transcription to avoid dealing with the problems of coming to that region, among other functions, and (famously) they're degraded with each cell division. They get inexorably shorter. This discovery solved an old mystery - how cells seemed to have a counter built into them which would let them divide only a set number of times (the Hayflick limit.)
There's a way out. The enzyme called telomerase replenishes the telemores, but it's not active in most cells (germ cells and tumor cells are notable exceptions, along with a few other high proliferators.) The initial hope was that some sort of telomerase activator might be an anti-aging therapy. Introducing telomerase activity into a cell line can keep it dividing for much longer than the Hayflick limit, true. But aging is due to a lot more than the sum of your cell divisions, so attention turned several years ago to the flip side, some sort of telomerase inhibition as a treatment for cancer.
It's a worthy idea. Telomerase is present and very active in many tumor cell lines, and it gives them a real leg up on growth. The cells divide and divide some more, never taking any clicks on their chromosomal odometers. And the near-absence of the enzyme in normal cells would seem to provide an opportunity for selective therapy. It's been uncertain if telomerase inhibition could be a front-line therapy or not, but it could certainly be useful in slowing down metastatic and fast-growing tumors. And it would seem to be a useful add-on to existing therapies as well.
Geron (and others) have been attacking the problem several ways. The most advanced is an attempt to raise an immune response to telomerase via a vaccine. This vaccine program caused a flurry of interest a few years ago, which was probably premature. The company's CEO, Thomas Okarma, presented the latest data in person, and it can be summed up as "so far, so good."His job is to persuade everyone that the idea has matured.
Back in April, the company reported low-dose results from adminstering the vaccine to prostate cancer patients, and the results looked encouraging (enough to start the long climb awy from penny-stockhood.) The latest results come from further dosing, and they are indeed seeing an enhanced immune response, and apparently still no side effects. Okarma claims that they're seeing the same level of T-cell response that is seen in something like measles vaccine, which is a record for a cancer vaccine.
What we don't know, yet, is how these patients are doing. That veil flutters to the floor later this year, and there will be a fine frenzy leading up to it, clearly. The earlier low-dose group showed some strong responses (almost completely knockdown of circulating metastatic cells, for example,) so this is going to be very interesting indeed.
Another approach is to directly inhibit telomerase. Many drug companies have looked for small-molecule inhibitors, but the search hasn't been easy. (Finding good inhibitors of enzymes that deal directly with nucleic acids is always a challenge.) Geron has an oligonucleotide (basically a short stretch of DNA, 13 bases long in this case) called 163, which seems to be a good inhibitor. (This makes sense, given that that's what the enzyme's active site is built to hold.)
Okarma referred to this program several times as their "home run." Is it? They've shown that direct injection of 163 into tumors has a dramatic effect in animal models, which is a good start. But an oligonucleotide is a tough candidate as a drug. Okarma knows this, of course, and he knows that investors have been burned by companies that develop them. He told this conference that 163 isn't an antisense compound, in an effort to reassure investors, but that's rather disingenuous. Of course it isn't - he's already stated that it's an enzyme inhibitor, not something that inhibits the enzyme from being made. But it's going to have many of the same problems as an antisense drug, from the standpoint of drug delivery, so the comparison is still valid.
Geron realizes this, of course, and they've developed a modified version, 163L, which is more stable in vivo. Details are scant, but I'd assume that there's some long greasy tail that's been added, similar to the polyethylene glycol "PEGylation" technique used for peptides. This compound shows better activity than its precursor. Okarma showed some data in a multiple myeloma animal model, but it's not clear how the compound was dosed - I'd assume intravenously, not that that's an easy thing to do in mice, but it's certainly not oral. They also have two-week toxicology data, which are clean.
This is very interesting, and well worth watching. I think it's a bit foolhardy to use the phrase "home run," though, because this compound is still a good way from getting into the clinic (as far as I can tell.) For one thing, you need a lot more tox data to even start talking about human trials. Geron has a history of aggressive self-promotion, which I don't think does them favors at this point. And keep in mind that others are hot on the same trail. But perhaps there really is something here worth the hype.
A lot of people who piled into the stock today sure seem to think so. Much as my portfolio would like it, I have to think that some of this is going to hiss right back out of the price in the coming weeks. The "home run" inhibitor probably isn't going to be the big news for a bit, if it ever is. But that clinical update on the vaccine later this year is going to either send the company's stock up to places it hasn't been in years, or it's going to pitch it right back to the low single digits. Get your popcorn ready.
Just Who's The Target Audience Here?
Today's Wall Street Journal has an interesting article about a theme that I've sounded several times on this site: the drug-safety argument against reimportation. The headline gives a good sense of the article's stance: "Drug Makers Cry 'Danger' Over Imports."
I've long argued the drug safety tactic is a loser, because it dodges the more important reasons why drug reimportation is a bad idea. And by leaning on this point so heavily, I worry that the industry will be left without an argument at all if safety inspections are stepped up in response to all the public hand-wringing. "See, we made the drug supply safer! Now why can't we import these wonderful cheap drugs?"
As the article shows, it was an unusual step for the industry to even bring up the possibility of unsafe or contaminated pharmaceuticals. Usually it's not the sort of thought you want your customers to have. But PrRMA, the industry trade group, seems to have decided that it's a better one than the thought of cheap goods from Toronto. Like so many other political maneuvers, this one was focus-grouped. Edelman, a large PR firm, seems to have been contracted to find what technique would shake up audiences the most, and this was the winner.
Scott Hensley, the Journal's reporter, has it exactly right when he says ". . .so far, the documented risks pale next to the rhetoric where imports are concerned." So now the whole foreign-drug-safety tactic, which PhRMA seems to be relying on as some sort of magic talisman, is being seen as just a pile of public relations spin. Right in front of our eyes, it's losing any value it might have had. And we have what, exactly, to fall back on once it's gone?
Remember, this is in the Wall Street Journal, not some anti-corporate fishwrap handed out by people in black bandannas. This kind of story is bad news for the industry. The real problem is that PhRMA is putting the industry in a situation where the truth is bad news. We need to start making the real arguments against reimportation, or we're not going to have any arguments to make at all.
We Have Ways of Making You React
Mentioning that ice-expansion pressure reaction the other day brings up a thought that I have pretty often: the artificiality of organic chemistry.
By that, I mean that the ways that people like me form molecules have almost nothing to do with the ways they'd get formed without us. They're the same molecules, all right (no vitalism here!) but we sure do take different tactics to make them. In some cases, that's because we don't have the time to do it by Nature's way, and in other cases it's because we don't have the talent.
The first case applies to most natural inorganic processes. When I go to a museum's mineral collection and see the spectacular sulfur crystals, giant tourmalines and so on, I can't help but think of the largest crystals I've grown in the lab - smaller than my finger. But all you need to make these huge mineral specimens is constant heat and pressure for thousands of years. It's simple. I'm sure that there are a lot of reactions that we could make use of in the lab if we just had those time scales to work in. As it is, a multi-day reaction is considered a long haul. If you have one of those in your synthetic plan, the search starts immediately to find a replacement.
The second case is more of a biochemical matter. Any organic chemist stands in awe of biochemistry. I've mentioned that my PhD work involved the (attempted) total synthesis of an antibiotic molecule (rosaramycin.) It was a beast, but it's far from the most beastly molecule out there. I struggled with it from all directions; the synthesis seemed to fight back at every opportunity. But as I sat there at two in the morning, up to my clavicles in dirty flasks, I knew that out in Texas, in the dirt of a golf course, the bacteria that made rosaramycin naturally were handily splicing it together. At room temperature, in water, in their spare time. It was a humbling thought, and it's not like I particularly needed more humbling at that point.
So we cheat, by the standards of Nature. We jam molecules together at huge concentrations, with functional groups trip-wired to react as quickly as possible. We make bubbles of the most artificial conditions imaginable (known as round-bottom flasks.) Baths of pure ether, of pure dichloromethane: try finding pools of those things in a rock formation. And we toss in every catalyst we know to try to get the reaction rates up to a useful level - a few hours, a couple of days, we tell them, or you go in the red waste can over there in the corner. The third degree is pretty much the only technique we know.
So, Is Silence Golden Or Not?
Note: In this post, I've tried to follow through on what seemed like a reasonable request - to bold-face some key points in my posts. Let me know if you find the idea useful, or if it's a distraction. I think the bubbling river of my prose can stand it, but aesthetes may disagree.
I've spoken a few times about RNA interference technology, mostly to point out that it's going to be the subject of at least one Nobel prize, and probably more. It's just a matter of figuring out how to divide it up, and when the committee wants to hand it out. There's no doubt in my mind. Just look at the explosion of research that's going on in the area; it's staggering. But the technique is of much more than academic interest. In this age, almost everything in molecular biology is of more than academic interest.
For those who don't keep up with the field - and there are, conceivably, better things to do with one's life - RNAi is a method that allows you to block a given gene from being transcribed into protein. There are other ways to do that, but they all have severe limitations. For example, you can knock out a gene entirely to keep it from being read, but we only know how to do it well in mice. Or you can come in with antisense DNA and try to mess up the gene that way, which is a technique that caused a worrying similar storm of interest when it was first reduced to practice.
There are differences. Antisense is a completely artificial phenomenon, while RNAi is a system that's native to the cell, not that we knew that before about 1998. It's everywhere, showing up in plants, insects, worms, rodents, and humans. It works a bit downstream of the DNA involved in antisense, bollixing up the later messenger RNA on its way to becoming protein. Here's a reasonably technical lecture on the subject.
The pathway seems to be largely a defense against viruses, which often come into the cell with double-stranded RNA (not normally something you have floating around.) The process ends up stopping any of the messenger RNA from the virus from being read out into protein, and the same system can be hijacked into wiping out normal messenger RNA as well. It's strange to think of such an important and interesting process that we didn't know anything about until a few years ago. Makes you wonder what other key things we're missing. . .
So the potential use of RNAi as an antiviral therapy is pretty self-evident - HIV is number one on the list and there are plenty of other candidates (RSV, influenza, and viral hepatitis for starters.) And if you can go after those, then you can presumably go on to target some of the things you'd use gene therapy for (when the problem is over-production of some protein) and all the things you'd think of using antisense for. Those two technologies are cousins, in many ways. And just like in the antisense field a few years back, RNAi companies have sprouted up all over the landscape, trying to get something into the clinic.
The hope is that antisense will prove to be the underachiever of the two; the fear is that it'll be the template. It's never lived up to its nearly boundless potential, despite a huge amount of work, time, and money (all of which are still being spent, mind you.) Some of the same problems it's faced are inherent to RNAi - the biggest one, from a medicinal chemist's point of view, is how to get the needed short chains of nucleic acid where they're needed. The natural tendency is for them to fall apart quickly, and even if they don't, the natural tendency is for them to bounce uselessly off the outside of the cells. Like the antisense folks, the RNAi crowd is turning to viruses to deliver the payload, but there are risks. Other technologies (ways to stabilize and package the RNA, mostly) are still in the shakedown stage.
It's full speed ahead, though, as a recent Nature roundup (425, 12) makes clear. But there's a new set of results that must have made for some increased coffee consumption at all these small outfits. It looks like even the small, seemingly innocuous RNA molecules that people are using can be enough to triggered an unwanted antiviral response. Genes in the interferon pathway are being set off, which, as a commentary in Nature Cell Biology put it, "(poses) a complication that had been widely dismissed."
It's not being dismissed now, I can tell you. The legions of people who are using RNAi in vitro now have to think about whether their experimental results have been compromised. And the companies pushing for clinical agents now have to worry about some potentially nasty side effects. There may well be a way around this - it's not clear, for one thing, if it's a general effect, or just found with some RNAi techniques or sequences. And there may be ways to compensate for it, even if it is a broad problem. But it's a clear sign that this new magic wand is going shoot sparks out in unexpected directions.
Why Didn't I Think of That?
We organic chemists try all sorts of voodoo to get our reactions to work. The standard way to kick-start one is to heat it up, naturally. Problem is, that tends to accelerate everything. So if your reaction isn't the most favorable pathway available, then you get just get the wrong thing to happen faster.
And we throw in acid catalysts, and basic catalysts, and sometimes both at the same time. (I'm not kidding. There's one called the Knoevenagel condensation that works pretty well that way. Plus, you're a member of the gang if you can say the name right.) We throw in transition metals, finely ground carbon, whatever. If Aldrich sold 1 molar eye-of-newt solution, they'd get customers. But there's one method that seems too simple to work, but it does. We squeeze 'em. And I mean squeeze. Bottom-of-the-ocean pressure is about right.
This only works if your reaction is of a particular type, though. Every chemical reaction goes through a high-energy state, a hump that it has to get over on its way to completion. The way you speed up your reaction is to make the gap between your starting materials and that transition state as small as possible - give it the smallest hurdle to jump over. (As for the difference between the transition state and your products, that determines how much heat and energy is given off when the reaction happens. A small barrier plus a big drop at the other end means you'd better make a break for it: that's the profile of a touchy explosive. But there are plenty of reactions that have a potential big drop in energy, but have such a huge climb up to their transition state that they just don't happen. Burning nitrogen gas is a good example, and a good thing, too.
So, if your reaction has a transition state that takes up less volume than the starting materials do when they're ready to react, then you can make that more favorable by applying pressure. A lot of ring-forming cycloaddition reactions fall into that category, and people have used all sort of apparatus to get recalcitrant ones to go. It's one of those promising techniques, though, that's remained promising for about forty years now, because not many people have the special equipment. (And many that do have no desire to use it.)
But a Japanese team (from the Tokyo University of Science and RIKEN) might be about to change that. In the last year or so, they've reported a really elegant way to do this stuff. Their latest is in the Journal of the American Chemical Society, page 11208. (Yeah, I know. It's only September, and JACS is up to page eleven thousand. Subscribe to it at your peril. But it's better than the Journal of Biological Chemistry, which will push you out into your yard in about a year and a half if you try to keep the hard copies.)
Here's what these folks do: they take their starting materials, and put them in a Teflon tube, and seal it up. Then they put that thing in a metal reactor, the sort that chemists know as a "Parr bomb," after a popular model. It's a thick-walled metal cylinder, with a cap that screws on with bolts. They fill the thing up with water, all the way to the top, and bolt the top down. But they don't heat it. They put it in the freezer.
And the water freezes, and it tries to expand as water (and pretty much only water) wants to do. But it's stuck in a screwed-down metal case. . .so the pressure on that Teflon tube climbs up to about 2000 atmospheres while the guys in the lab go out to take in a movie or something. They leave the thing in the freezer overnight, and then just take it out and let the water melt the next morning. And the reaction's done.
I really have to take my hat off to the person who thought this one up. It's the neatest trick I've seen in a long time. Everyone in my research department I've shown this to has looked at the paper for a moment, and grinned when it hits them. Congratulations!
New Ways to Erase Your Credit Cards
There's a paper in a recent Proceedings of the National Academy of Sciences that might lead to some interesting new experiments. These folks have worked out a technique that makes NMR about 10,000 times more sensitive than usual.
There's room for it. NMR is a wonderful technique (PDF file) that modern chemistry absolutely cannot live without. You get vast amounts of information about the different atoms in a sample - what all the hydrogens are attached to and what angles those make with the next atom over, how many hydrogens the backbone carbon atoms have bonded to them and which ones they are, which atoms are close to each other in space but not actually bonded - oh, you can wring so much information out of one sample, and it doesn't tear up any of it. You just get it back in the same tube you started with and go on your way. NMR destroyed the old ways of figuring out the structures of organic molecules, and good riddance to most of them, too.
But our best friend has some weaknesses. And one of the big ones is sensitivity, and that applies to both NMR's good old-fashioned forms and to its flashy upscale turn as magnetic resonance imaging. Compared to a technique like mass spectrometry, which can pick up really miniscule amounts of a given substance, NMR is a pretty blunt tool. The problem is that NMR depends on an energy gap between nuclear spin states, and that gap isn't anything to write home about. It isn't too wide, although the larger the magnet you have, the wider it gets (which is why we use the most insanely huge superconducting magnets we can possibly afford.) And in any sample, there aren't that many nuclei that are flipped up to the higher-energy state, and those are the only ones you can detect.
(That's all I'm going to say about NMR theory. I could go on about it for another couple of days if I had to, and in the process send my readership to new lows. See some of the stuff in that link two paragraphs above if you feel the urge for more, though. Reminds me of the other week when I started a discussion of chirality, which I still need to get back to. I told my wife about writing that one, and she said "Oh yeah, that'll bring 'em in.")
Anyway, what this new technique does is take your sample and pump up the atoms to where a huge number of them are in that higher spin state. When you put that revved-up sample into the standard NMR machine, you get this roaring signal. Potentially useful but painful nuclei like nitrogen-15 start shouting back thousands of times more strongly, and that's going to open up a lot of experiments that just aren't feasible under standard conditions. The effect isn't permanent, but it lasts long enough to get the spectra.
NMR imaging in people and animals is an obvious application. This sort of thing has already been used with xenon gas to image lung tissue, and now that other nuclei can be enhanced, there should be some very interesting data coming. It's always been a dream of medicinal chemists to use NMR techniques to follow compounds around when they're dosed in vivo. But the sensitivity problem, among others, has kept this from being realized most of the time. Maybe we can take a crack at it now.
I'm back to what passes for normal - just a couple of spectacular bruises. Thanks to everyone who wrote in wishing the best for my left hand!
I mentioned a palladium reaction last week, the one where the authors of a new paper found that the less catalyst they added, the better the reaction. That makes a person wonder: what happens if you take that to the reducio ad absurdum?
We're in the process of finding out. There's another research group that's working on what's become a classic palladium-catalyzed reaction - the Suzuki coupling. That one was discovered in the early 1980s, and is a terrific way to bond one aromatic ring to another. I wouldn't even want to think about how many papers have been published in the last twenty years on variations and improvements of that chemistry. And the number of sheer examples of its use is surely climbing into the tens of thousands.
So it comes as a bit of a shock to see these folks, from King's College, London, claiming that the reaction actually works fine without any palladium catalyst at all. It's like hearing that someone has found a way to hit home runs without using a bat. The researchers say that the reaction works as long as you do it in water at 150 degrees C (in a sealed tube, obviously) and that microwave heating works best. There are some other limitations - a narrower range of coupling partners can be used as compared to the standard Suzuki reaction, for example. You can find this, if you're curious, in a recent issue of the Journal of Organic Chemistry, page 5660.
Naturally, there is some scepticism. I had the opportunity not long ago to hear a lecture from Professor Suzuki himself. Aldrich, the reagent supply house, is squiring him around the country at the moment, while he speaks to universities and industrial sites. When asked about this work, Suzuki seems politely disbelieving, as you'd expect. After all, this reaction is his child.
I'm not sure where I stand on the issue, myself. It's certainly possible that the uncatalyzed reaction works as advertised: the King's group has done a number of control experiments to address the usual questions. But it's also easy for me to think that a few atoms of palladium here or there in the lab might be enough to sneak into the flask and get things to happen. No matter what, it appears that the catalyzed reaction will continue to be more in demand, except for applications where the lack of metal and the simple conditions make a difference (on large scale, for example.)
Interestingly, Suzuki himself had trouble getting his initial reaction report published. As he tells the story, it was rejected by Tetrahedron Letters, which was the premier place to send short communications at the time. He ended up placing it in Synthetic Communications, which was a relatively young journal at the time, lacking prestige. (Now it's an older journal that lacks prestige.) Suzuki's 1981 paper has to be the most-cited thing ever to appear there.
Blogging will be light to nonexistent until my post for this coming Monday. I took a good fall onto some blacktop last night, and one hand is sprained and bruised enough to make typing notably less joyful. I should be healed up in time for next week!
Backing Down, in Public
There's been a very public retraction of a controversial paper from last year, one which linked MDMA (Ecstasy) to dopamine-linked neuronal damage (and thus possible Parkinson's disease.) The researchers (at Johns Hopkins) injected monkeys with what they thought was a sample of MDMA and saw clear signs of neurotoxicity.
But the next study, this time via oral dosing with another sample of MDMA, failed to show the effect. So they went back and tried the injection route again, and this time it showed no dopamine-neuron toxicity, either. The discrepancy in the oral versus i.v. dosing would be unexpected, but you could find ways to explain it, if it were a real and reproducible effect. But when the i.v. experiment failed that second time, the authors must have known that they had a real problem.
The vial that the MDMA came from in the first experiment had long since been discarded, but they had a sample of methamphetamine that had been received from the same source at the same time - and it proved, on examination, to contain MDMA instead. That led to suspicions of a label-swapping mistake, and sure enough, examination of the brain samples from the original monkey experiment showed the presence of methamphetamine, but not MDMA. It all fits, and it's all wrong.
Doubts were expressed, very loud doubts, about the results of the study when it first came out. Other research groups in the field were sceptical, pointing out that there should be a lot more cases of Parkinson's showing up in the ecstasy-using population, especially considering the doses that some of these folks were exposing themselves to. Those sceptics have been vindicated more thoroughly than they ever could have hoped.
I have several comments on all this. The first thing I need to do is commend the Johns Hopkins people for doing the right thing and retracting their paper. It must have been mighty quiet around the lab for a while after they got the results from that repeat i.v. study. The thought of something like this happening can really keep you up at night.
I'm sure that some people are going to point the finger at this group for not checking the samples of MDMA and methamphetamine. But I can't fault them so much on that point. In vivo pharmacologists are not chemists, and aren't expected to assay the samples that they're dosing. In every drug research project I've been on, the animal folks make it clear that they depend on compounds being what the label says they are. They have no way to confirm it themselves. (In this case, Research Triangle Institute, the source of the samples, says that things were fine on their end, as you'd figure they would. Depends on where the label came from on that remaining methamphetamine sample, doesn't it?)
But all that said, I have to then turn around and wonder why the original paper was published at all. I was surprised to learn that their results hadn't been repeated beforehand. You'd think that this would be necessary, given the public health implications of the work and its variance with the results of others in the field. I can't help but think that the researchers got their original data, thought they had a hot result that would make everyone sit up straight, and got it into publication as fast as they could.
I'm really taken aback to learn that they hadn't looked at the original monkeys for MDMA levels before. Getting blood samples from monkeys is no easy task, but why wait until there's a problem to do the post-mortem brain levels? Those numbers really would have helped to shore up the original results - and would have immediately shown that there was a problem, long before the paper was even published. I don't like to sound this way, but it's true: in the drug industry, we consider pharmacokinetic data like this to be essential when interpreting an animal study.
It's even more vital when you're trying to figure out brain effects, since the levels of compounds in the CNS can only be determined by specifically checking there. So, you see brain damage? The next question is "How much compound was there in the brain?" And you don't go on until you've answered it. Perhaps the authors decided to rely on known brain exposures when they ran the study. But those known exposures would have been from studies that didn't show the neuronal damage. I just can't find a good excuse here.
The further experiments that disproved the original results are the sort of thing that should have been done before the first paper was published, frankly. Like anyone else in the drug industry, I understand that monkey studies are slow, costly, and hard to get started. But this was an extraordinary claim, and should have been held to a higher standard of evidence. If you live by making a big splash in Science, you may die by the same route.
I Think I Need to Sit Down For a While
Last week I told my story about spending weeks trying to get a multistep route to work, when I could have made the stuff in one step by ignoring what I was told to do. I have to make sure that anyone who enjoyed that one sees this from Greg Hlatky at Borzoiblog. I've seen minor examples of what he's talking about, but this one really gave me the shivers. . .
And fear not, I have some stories up my sleeve that I don't think he'll be able to top. Or if he can, they'll certainly be worth hearing!
I didn't get around to commenting on the recent PYY news, the peptide that shows appetite-suppressing effects. This looks like a very real and robust effect, as Richard pointed out over at Living Code.
The big hurdle will be the fact that it's a peptide. You're not going to be able to take this orally. It's going to be i.v. or bust for some time, I'm afraid. There's some hope for taking peptides of this size nasally, although it's tricky, and there's an outside chance of getting them through via a transdermal patch, but I wouldn't bet on that one. If this ever makes it, it's going to be an injectable, with all the disadvantages of that dosing method. Another way to attack the problem is to make your proteins more stable, but the best anyone's done is to make them last longer in the bloodstream. Getting a protein intact past the gut is like trying to drag a prime rib through a piranha pond. Technically speaking.
An alternative would be to find an orally acting drug that works the same way, but don't stand on one leg waiting for that, either. Conserving the activity of a protein as you change over to a small molecule has proven to be very, very difficult. No doubt drug companies will be taking (another) pass through their screening files to see if they have anything, but the odds aren't good.
And these are early days. The obesity field is cluttered up with promising ideas that have deflated. One of the comments on this news over at DB's Medical Rants brought up a good point:
Every year or so I hear about these wonderful new treatments for obesity in some press release and then never hear about them again. Leptin, hello? OEA, where are you? And of course there's Phen-Fen, which everyone wishes they'd never heard of. Somehow the reality never matches the initial hype.
Tell me about it. It's just as bad when you're working in the field. We get better information than what shows up in the newspapers, which is a good thing. But at least newspaper readers don't turn around and work on the stuff for a few years before finding out that it's not the answer. The history of leptin is a good example; the fellow who posted that comment really should have heard about that one by now. See my October 24 posting here for a brief rundown.
But as for OEA, take a look at this, which is also commented on in the Living Code post. It's still around, and the receptor it hits is very much alive as a target in the drug industry. In fact, a drug is in the clinic.
I've always been a bit sceptical that pharmacological treatment of obesity will work very well. Eating is a highly protected biochemical pathway - every animal whose hunger reflex was easy to shut down is long gone from the earth. But damping down appetite (or, equivalently, enhancing satiety) probably has as good a chance as anything does. It's either that or rev up metabolism, and it's easy to think of a dozen reasons why neither of them will work. I would enjoy being proven wrong!
Got My Attention, Anyway
The latest issue of the chemistry journal Organic Letters has an interesting paper (p. 3285) from a group in the Netherlands. The title stopped me in my tracks: "Homeopathic Ligand-Free Palladium as a Catalyst in the Heck Reaction."
Well, that's fortunately not too accurate. Homeopathic "medicines," as everyone should know, are sometimes diluted to the point that not one molecule of the putatively active substance is likely left in the vial. The dilution is the whole point, according to homeopaths, and from there you wander off into a swamp of hoo-hah about how dilution makes them more potent, latent memory of water molecules, energy fields and it just gets worse.
I'm relieved to say that these researchers still have active catalyst present in their reactions. But once you get past the cute use of "homeopathic," they have an interesting point to make. The Heck reaction is one of a large family (getting larger every month) of reactions catalyzed by metals like palladium. All sorts of carbon-carbon bonds (and bonds to heteroatoms, too) can be made this way, and the field is nowhere near played out. We use this stuff constantly in the drug industry - I'd hate to have to get along without it.
The big disadvantage to palladium reactions is the voodoo factor. There are an insane number of variations on even the simplest palladium coupling. Temperature, solvent(s), base, catalyst - you can vary all of them out the wazoo, and sometimes you have to. It's a truism in the field that any metal-catalyzed coupling can be made to work well. Eventually. (That's kind of like the low-tech weather forecast: "Rain. At some point.") But if you're on a deadline, and we often are, you can go nuts working through all the possibilities.
These folks have found that no matter what brew people use to get the Heck coupling to work, the real catalyst is probably a very small amount of extremely active palladium that's in solution. If you run the reaction with more catalyst, to try to get it to go faster, then you actually end up putting it in reverse. The palladium gets a chance to fall out of solution in clumps when there's too much of it around. But as you run the reactions with less and less palladium, the thing actually speeds up. Whatever the active metal is, (single atoms of palladium? nanoscale clusters?) it has a chance to do its thing without precipitating out. Thus the "homeopathy," but homeopathic medicines should have something as active as monomeric palladium in them. And they don't.
I actually ran a Heck the other day, and scooped in my customary I'm-not-paying-for-it amount of palladium acetate. And most of it precipitated out, as a metallic mirror on the inside of the flask. Very pretty, but it's clear that that stuff was no longer doing my reaction any good. Next time, I'll just dink it with a microscopic amount, and we'll find out who's kidding whom.
Everything You Know is Wrong
I mentioned that I've had to learn the lesson, more than once, that you shouldn't trust other people to know more about your research than you do. That's one of my "Lowe's Laws of the Lab," actually: Your Project Is In Big Trouble If Other People Know More About It Than You Do.
There was an experience early in my research career that got me on that path. I had been asked, as a student, to make a cyclic compound as a test system to look at some unusual reactions. There was a synthesis worked out for me; I looked it over and it didn't seem obviously crazy, so I went with it. (Keep in mind that I knew some organic chemistry, but certainly not as much as I could have used.) These compounds hadn't been made before, so I was willing to take a crack at them.
First step, not too bad. Second step, no problem. Third step. . .needed to be fixed a bit. It was a reduction of a triple bond to a double bond, which calls for the more mystical end of the hydrogenation catalyst field (one whose mainstream was accurately described to me in my first organic course as "witchcraft.") But I got it to work: an old-fashioned Lindlar catalyst, plus some quinoline to make it even less reactive. It's the sort of reaction that a chemist in 1952 would have felt very comfortable with. But those reaction still work just fine - it's one of the things I like about chemistry, that good reactions never go out of style.
So far, so OK. Now it came time for a ring closure, making a cyclic ether, and the way that my betters had mapped out for me was to use something called the Mitsunobu reaction. My fellow organic chemists will start to hear the faint sound of alarm bells right about now, because that's not the most reliable method for that sort of thing. For many kinds of rings, there isn't a most reliable method for that sort of thing, but you can usually hack something out.
Except in this case, and for good reason. Shop talk for my colleagues: you'll have to remember that I was very early in my career when I tell you that one of the reacting centers was neopentyl. For those outside the field, that means that the Mitsunobu reaction, and a host of others that depend on the same sort of attack, was doomed from the start. The two end of the chain I was trying to link together were too lumpy and crowded; I'd have been lucky to get it to work in a diamond anvil under geologic pressures.
Well, it took a little while for this realization to creep up on me, but it finally got my attention, and I sat down to think about why I was wasting my time like this. Surely someone had made ugly hindered cyclic ethers before? I went down to the bound volumes of Chemical Abstracts, and this was before the days of access to computer-based searching of that mighty database. Not long before - by the next year, I was sitting in front of a terminal doing ASCII-based structure searching - but there was no way at the time to do it than by hand.
It's a lost art, and it deserves to be a lost art. One semi-effective way was to find the name of your parent system of interest, and to cruise down the columns in the Name Index, looking for likely compounds in their alphabetic ranks. My cyclic ether was easy to localize, and I started down the list. . .and it wasn't long before, to my complete consternation, I found the exact compound I was trying to make.
But hold on - this was a new compound, right? What was it doing in Chemical Abstracts? Well, it was presented to me as a new compound, because (as it turned out) no one had bothered to see if it had ever been made before. I jotted down the reference number, hopped immediately over to the abstract volume (you used to have to do that, folks,) and from there to the original paper itself. I still have the copy of it that I made that night.
It's in my files, turning a bit yellow around the edges after twenty years or so. I'll never throw it away. These folks had made the compound I needed, in one step, from two things that you could buy from the Aldrich catalog. They used a reaction I'd actually never heard of, which made me feel a bit better (at least I hadn't missed something I supposedly knew.) But there was my last few weeks of on-and-off work (I was taking classes, too,) which could have been accomplished in an afternoon if I'd had the sense to look the thing up before I started. If I'd had the sense not to just take what was handed to me, that is. Believe me, you deserve what you get. My career since then has been an effort to get as little of it as I can.
'Tis Folly To Be Wise
I came across an article in my files today that I thought I'd share. It's by the late Calvin Mooers, an information scientist. He addressed his colleagues on the question of why some information systems got so much more use than others - often with no correlation between the amount of use and how useful the tools actually were.
"It is my considered opinion, from long experience, that our customers will continue to be reluctant to use information systems - however well designed - so long as one feature of our present intellectual and engineering climate prevails. This feature - and its relevance is all to commonplace in many companies, laboratories, and agencies - is that for many people it is more painful and troublesome to have information than for them not to have it."
When I first read this, I experienced that quick shock of encountering something that you feel as if you'd known all along, without realizing that you knew it. Of course. It's not a new idea, but we keep having to learn it over and over. Mooers again:
"Thus not having and not using information can often lead to less trouble and pain than having and using it. Let me explain this further. In many work environments, the penalties for not being diligent in the finding and use of information are minor, if they exist at all. In fact, such lack of diligence tens often to be rewarded. The man who does not fuss with information is seen at his bench, plainly at work, getting the job done. Approval goes to projects where things are happening. One must be courageous or imprudent, or both, to point out from the literature that a current laboratory project which has had an extensive history and full backing of the management was futile from the outset."
Oh, yes. Yes, indeed. I've seen these examples made real right in front of my eyes, and more than once. Have I mentioned that Mooers wrote all this in 1959? The problem has not lessened one bit since then. If anything, our vast information resources and the powerful tools we have to dig for it have made things worse. Just try being the person who finds a patent claim that stops a project in its tracks, one that was missed while the work went on for months. Or find out that a close analog of the lead compound was found to be toxic twenty years ago.
We're supposed to be able to find these sorts of things. But everyone assumes that because it's possible to do it, that it's been done. Taken care of: "Didn't we see that paper before? I thought we'd already evaluated that patent - isn't that one one that so-and-so found? It can't be right, anyway. We wouldn't have gone this far if there were a problem like that out there, clearly."
My rule, which I learned in graduate school and have had to relearn a few times since, is to never take anything on faith when you join a new project. Go back and read the papers. Root through the primary literature. Look at the data and see if you believe it. If you let other people tell you what you should believe, then you deserve what you get when it comes down around your ears.
Mercury is in the news a bit this week. It's an element that we don't see much of in the pharmaceutical industry, for some pretty obvious reasons. There's no way that we'd be able to get a pure mercury compound to work as a drug - well, maybe if it evaporated pancreatic cancer in one dose, but don't hold your breath waiting for that one. Mercury accumulates, notoriously, and it's not something you want steady doses of. We don't even use mercury-containing reagents, except on very rare occasions when there's no other choice. So far, in my coming-up-on-fourteen-years in the business, there's always been one. I haven't used a mercury compound since I was in graduate school.
But there's toxic, and there's toxic. Mercury, as the elemental metal, isn't too bad taken orally - although why anyone would gulp liquid mercury is another question. But it doesn't get oxidized (much) in the intestinal tract, and it doesn't get absorbed as the metal. It just sort of rolls right on through, from what I've heard. It's certainly not going to improve you, but it won't kill you.
But the metal is truly dangerous if it's spread around a poorly ventilated area. It has a small, but real vapor pressure, and it's evaporating off slowly all the time. And breathing in mercury vapors is a very different thing than swallowing a ball of the stuff - that gives it a chance to be absorbed into the bloodstream, and it starts in doing its slow damage.
And soluble mercury compounds are very bad news indeed. I wrote some time ago about the terrible case a few years back of Karen Wetterhahn, who died from a small amount of dimethylmercury, which passed right through the material of her glove. It took months for her to realize that she'd been killed that afternoon. But there was nothing that could be done, once the compound had been dispersed through her system. A terrible case. I cannot imagine what would induce me to handle dimethylmercury. Perhaps wearing a hybrid outfit of body armor and technical-grade scuba gear I might consent to lift an unopened bottle. But probably not.
One thing that tames mercury a bit is bonding to sulfur; it's a relatively strong bond, and keeps the mercury from going out and finding worse things to bind to (like the thiol groups on proteins in your neurons.) An old-fashioned way of cleaning up a mercury spill, actually, is to toss powdered sulfur on it and let it combine. Inefficient, but perhaps useful in an emergency. The reason I bring this up is that it now appears that the mercury found in fish is bonded to sulfur, from cysteine, to be more precise. The standard estimate of mercury toxicity in fish is based on it being in a more reactive form, so this is good news.
There's another sulfur-mercury bond making headlines. Back about a year ago, I wrote a number of posts about thimerosal, the now-removed organomercury vaccine additive. I was sceptical of its proposed link to autism, and now there is, thankfully, a bit more evidence to back that up. Autism levels in Denmark appear to have been unaffected by the thimerosal ban (which went into effect there several years before the additive was removed in the US.)
I'm sure that there are ways to attack this study, and there are people who have a strong attachment to the thimerosal hypothesis who will find a way to do it. But in the next few years, we're going to be getting more and more data of this sort, which I strongly believe will continue to show that thimerosal is uncorrelated with autism. Now we just need to find out what is correlated with autism. . .
Clean This Up For Me, My Good Man
Do you know what I'd consider a completely luxurious laboratory? The lab of my dreams, the one that exists only in my head? No, it doesn't have platinum flasks, or my own private NMR, or the entire Aldrich catalog on shelves in the back room. What it has, what it would have if I could have it, would be someone to finish up my reactions for me.
I'm not the only synthetic chemist with this dream. For many of us, setting up reactions is a lot more fun than working them up. That's because the situation is a bit like Tolstoy's line in Anna Karenina about happy families being all alike: there are unlimited reactions to set up, but they all tend to get finished off in the same fashion. And after a few thousand of them, you long to have someone take the job off your hands.
Organic chemists spend an awful lot of time cleaning their compounds up. Rare is the reaction that gives you material that's obviously good enough for the next step. There's usually some side product that you fear will hang around and be even harder to get rid of later, or failing that, there's almost always some colorful gunk, the sort of stuff that sticks to the top of a chromatography column. All of it has to be gone by the time your synthesis is finished, and the longer you wait, the worse the problem will generally get.
So it's time, once again, to partition your reaction between some organic solvent and some aqueous washes - acid, base, brine. Time to dry the organic layers over sodium sulfate (or magnesium sulfate, depending on your religious affiliation, because that's about the only way to characterize a liking for one over the other.) Time to rota-vap off the organic solvent, which almost always takes longer than you'd think, even after twenty years of doing it. And it's time for yet another chromatography column.
That's when I want some dedicated flunkies to step in. I'll just think up the reactions, the weird ideas, and I'll go to the hood and get them going. Then when I give the word, the cleanup crew will arrive and do the grunt work. Call me when the stuff's pure; don't trouble me with the details. I'll be setting up another reaction by then.
Now, it's true that I have people reporting to me. But I'm not going to use them in that capacity. For one thing, they wouldn't stand for it (and they shouldn't!) For another, they're in the same boat I am: they want someone to come clean up their products, too. They don't find chromatographies any more exciting than I do, and they run a lot more of them than I do.
So I don't know who we'd get to do this sort of thing, but I continue to hope. It would be like working down at the bottling plant, but it would be steady employment, that's for sure.
Archives, RSS & Email
Click here for access to our archives.
Dictionary of Scientific Quotations
If the Universe is Teeming with Aliens
(Then where is everybody?)
The Night is Large
A Martin Gardner collection
Eurekas and Euphorias