Skip to main content

Analytical Chemistry

Natural Product Artifacts

Like many organic chemists, I find natural products very interesting, since their structures are often things that I would never imagine making (and in some cases have trouble imagining how to make at all!) But there’s a feature of the literature in that area that not everyone appreciates: the fact that a reasonable number of structures are, in fact, artifacts.

That’s reviewed in this article. These artifacts occur during the isolation process, and can come from a variety of sources. Carboxylic acids can get turned into esters on exposure to methanol or ethanol, for example, a situation that is not made any clearer by the fact that both methyl and ethyl ester structures can be legitimate natural products themselves. Lactones can get opened up, and similarly, aminals and acetals can undergo exchange reactions with such solvents. There’s been quite a bit of confusion over the years with all of these, and by no means is it all cleared up. There are many structures whose status is still open for debate.

Another complication is the set of natural products that feature reactive groups in their structures – sometimes what you isolate is the “sprung trap” version of the real compound. In all cases, you have to be alert to what conditions the extracts and isolates have been exposed to, and be wary of solvolysis, etc. Not only alcohols, but things like acetone and dichloromethane can get involved, even under what seems like very gentle handling. Chromatography, of course, is a big source of solvolysis artifacts, as is storage in solution (even at low temperatures), but there are cases where just transferring a compound from one flask to another and evaporating it down again is more than enough to cause trouble.

The same goes for oxidation – there are plenty of organisms that live (and thrive) in low-oxygen or flat-out anaerobic environments, and they can produce some very interesting compounds. Unfortunately, those compounds are not necessarily built to last out in our oxygenated world, and can thus spit out a whole range of artifactual structures.

But perhaps the biggest section in the paper is on artifacts produced by pH changes, both acidic and basic. That’s similar to the oxidation problem – natural products are generally produced in (and for) environments whose pH is rather tightly regulated, so exposure to out-of-range conditions can set off all sorts of chemistry. As the paper notes, though, “Unfortunately, the natural products literature has many accounts where the impact of acid has not been constrained” Rearrangements, dehydrations, condensations, and cycloadditions have all been documented. And on the basic side, Michael additions, epimerizations, aldols, and decarboxylations all feature. Some of these processes can be quite elaborate.

The gold standard is to identify proposed natural products in as crude an extract as possible, saved before too much handling too place, but too many papers don’t seem to bother with that step. This new review notes that the literature clearly under-reports artifact formation, but also details how it can be a way to recognize the actual compounds when they’re too sensitive to be easily isolated. But that only happens when you’re alert to the idea that what you’re seeing may not be the real natural product, of course.

Handling samples as gently as possible (heat, light, exposure to oxygen, to pH changes, and to potentially reactive solvents) is clearly the way to go, but that’s not always compatible with actually purifying and isolating all the components of a mixture, either. So the literature itself has to be handled with care, too. . .

8 comments on “Natural Product Artifacts”

  1. Adonis says:

    I always viewed bench phytochemistry as survival of the fittest. The idea that you will preserve endogenous chemistry after all that rotavaping, silica gel, light, solvents, fractions sitting on the bench sometimes for days is just silly. What ends up being isolated are either stable endogenous compounds or their stable derivatives. Glad that the issue is receiving its due. In the plant world in particular, where new chemical scaffolds are becoming rare, artifacts often come handy to many investigators as a source of “new” compounds.

  2. Anon says:

    If someone quickly wants to learn reaction mechanisms or arrow pushing exercises, this review from down-under, I reckon is the one of the best I have seen in a long time. Also, can be a part of many pop-quiz questions as well for educators. Great job by the author and kudos to him.

  3. A Nonny Mouse says:

    It’s (or rather, was) also possible to turn the X-ray picture around and get the wrong absolute configuration as happened with the clerodane series, which lead to a lot of horrible naming conventions.

  4. Mark says:

    There are probably quite a few dodgy natural products out there, harvested from the interface of chemistry and biology, where biologists are forced to go without a guide into territory that for them is a bit wobbly, and where identification has to be based on a mass spec coupled with imagination, a trailing wind, and maybe 1ug of product at best (on the plus side, in a crude mixture that probably hasn’t been handled much – before a chromatographer got their paws on it, anyway). Getting 1mg (or at least 0.1mg) of pure compound for a friendly NMR spectroscopist is a complete impossibility for people looking at interesting-but-dilute natural products in small, intractable source-organisms. If something is reported as a novel product, it will probably get at least some chemical attention from a reviewer, but if it appears as a side-line in a paper that’s primarily biological, it may not get the critical appraisal it really needs. Unfortunately there is no way to publish a structure (or enter it in a database) with an appended probability-value of its being correct! Seriously, there should be. Sometimes it’s good to know that someone guessed that this is a methylated version of X, with some decent evidence, but took a complete random punt on where the methyl group is (methylation alone is useful information even if you don’t know where it’s gone), but it’s hard to keep track of this sort of uncertainty.

  5. Anonymous says:

    Here’s an example from “Perspectives in Organic Chemistry”, 1956, Chapter “Synthesis”, by RBW, pp. 155-184:
    “Cyclopentane-1,3-dione provides an instructive elementary lesson in this respect. After the compound had eluded a number of plausible synthetic attempts, many chemists were tempted to consider that it must be an unstable substance — that, for example, it might be susceptible to very ready cleavage. Yet when it appeared on the chemical scene recently for the first time, it was produced by the action of boiling hydriodic acid and red phosphorus on a degradation product of aureomycin!”
    There are natural products that, upon isolation, are known to be unstable to their own functionality. E.g., prephenate (salt) is stable. Upon attempts to capture prephenic ACID, it undergoes self-catalyzed acid dehydration to phenyl pyruvate, essentially dissolving in its own tears. It is one of many chemical squonks.

    From wikipedia (link in my handle):
    “Hunters who have attempted to catch squonks have found that the creature is capable of evading capture by dissolving completely into a pool of tears and bubbles when cornered. A certain J.P. Wentling is supposed to have coaxed one into a bag, which, while he was carrying it home suddenly lightened. On inspection, he found that the bag contained only the liquid remains of the sad animal. … The “scientific name” of the squonk, Lacrimacorpus dissolvens, comes from Latin words meaning “tear”, “body”, and “dissolve”.

    The Capon article cited by Derek is behind a paywall (aaaaarrrrgh!). There are probably more chemical squonks therein.

  6. Curt F. says:

    These days, LC-MS is good enough that after traditional “grind-and-find” fractionation and structure solving, it should be almost routine to look for the molecule you’ve purified and characterized in crude extracts. If you want to be extra sure, perhaps you make a few different crude extracts, using different solvents/formulations/etc…

  7. JCR says:

    Just making a crude extract of an organism will release many enzymes that can alter other compounds unless steps are taken to block the activity. Simple examples are proteases and nucleases that can hydrolyze the polymeric compound one is looking for. The first report of DNA polymerase from Thermus aquaticus was later shown to be a proteolysed portion of the full length molecule. Both it and the full length molecule could extend DNA but only the full length molecule had an associated 5′ to 3′ nuclease activity useful for an in-process detection of PCR product production.

  8. István Ujváry says:

    Take, for example the ‘phytocannabinoids’ THC and CBD formed upon decarboxylation of THC-COOH and CBD-COOH, the genuine biosynthetic products produced by hemp (Cannabis sativa). Yet (almost) everybody considers them ‘natural products’.

Comments are closed.