Skip to Content

One Fragment to Bind Them All

A colleague put me onto this paper, which appears in a journal (IUCrJ) that it’s safe to say that I don’t keep up with very closely. The authors, from Rutgers, are looking at fragment screening, and have found that 4-bromopyrazole and 4-iodopyrazole are very odd in that regard. They show up as hits in fragment assays, and soaking them into protein crystals makes it clear that those results are not false positives. The halopyrazoles bind and spare not: in a crystal structure of the complex between HIV reverse transcriptase and the inhibitor rilpivirine, a soak of 4-bromopyrazole revealed fifteen separate binding sites. It alone recapitulated the majority of binding sites found during an entire fragment screening effort.

They tried a number of analogs to see how general this effect was, and only 4-iodopyrazole was comparable. Soaking these into other proteins entirely gave similar results – they didn’t get all the sites identified by a full fragment screen, but they did very well. What’s interesting is that these binding sites were of several different characters (electropositive, electronegative, polar, nonpolar), and the modes of binding were various as well:

We hypothesized that the promiscuous nature of 4-bromopyrazole and 4-iodopyrazole may be owing to the electrostatics and small size of these compounds. These compounds are capable of creating diverse interactions with protein and solvent atoms through a basic chemical scaffold consisting of six non-H atoms that form an aromatic ring with hydrogen-bond donor and acceptor functionality and a halogen, providing diverse contacts with protein atoms and solvent molecules. Density functional theory calculations revealed that the electronegative N atoms in the pyrazole ring enhance the polarization of the Br atom in both negative charge and the positive charge of the `sigma hole’ opposite to the C—Br bond when compared with less electronegative heterocyclic rings. . .

That small donor/acceptor pair motif is indeed found all over the biochemical world, probably because of the ubiquity of amides, esters, carboxylic acids and so on. And it may be that 4-halopyrazoles are just the size and shape to hit a lot of them, with (as mentioned above) the halogen contributing on its own as well.

That’s particularly fortunate for X-ray crystallography, since bromine and iodine atoms provide strong and utterly unmistakeable signs in X-ray data. They’re popular, of course, because if you pick the right wavelengths, you can get anomalous dispersion off of them (you want the math, you say?) that can zip up the phasing problem and (given the fearsome powers of modern software) quite possibly drop your structure right into your lap. Iodine’s especially nice for people who don’t have a synchotron down the block, because the Cu-based X-ray tubes on “home” diffractometers are a good fit for this technique. The authors tried, with some success, to see if a quick soak with these pyrazoles could be used to improve the structure solutions of their protein crystals, and it looks like this idea is well worth adding to the box of tricks. Most med-chem labs will have one or both of these compounds sitting around on the shelf – you do Suzuki couplings, I’ll bet – so the barrier for experimentation is very low indeed.

15 comments on “One Fragment to Bind Them All”

  1. Interesting paper – you can read a bit more about it, and the broader question of secondary binding sites in proteins, at Practical Fragments (link in name).

  2. Agent M says:

    I wonder if this means 4-bromopyrazole and the like will soon be listed as unhealthy food groups by Mothers Against Molecular Obesity (MAMO) – http://fbdd-lit.blogspot.com/search/label/MAMO

    1. von K says:

      Alternatively it might be denounced as a PRaT as was the unfortunate 4-phenyl-2-aminothiazole that I once had the temerity to include in a fragment library…

  3. Ash (Wavefunction) says:

    Interesting, and I am not too surprised since Br and I are very versatile atoms, sort of jack of all trades that can bind in a variety of binding pockets. They are at the same time lipophilic and polar, and sterically bulky without being overly so, partly because of their high polarizability. Br and I often show odd and counterintuitive effects in drug design, and this paper tells us that they should be exploited more.

    1. Peter Kenny says:

      I don’t think that it is accurate to describe Br and I (and even Cl) as ‘polar’ and one should probably not describe any atom as ‘lipophilic’ because liphophilicity is a bulk property. That said, I would agree that substitution with Br and I will increase the partition coefficient and it is reasonably clear what is meant by a ‘lipophilic atom’. One way that polarity of atoms can often be quantified experimentally (at least in model systems) by hydrogen bond basicity and by this measure Cl, Br, and I are weak HB acceptors and not especially polar. Halogen bonding means that Br and I can effectively mimic hydrogen bond donors although this doesn’t seem to do much for the aqueous solvation of halogenated species. Fluorine is often considered mysterious because it is so electronegative but yet a relatively weak HB acceptor. Some of the mystery goes away if you make a distinction between the electron distribution ‘within’ atoms and between atoms. I’ve linked a #realtimechem week post on halogen bonding and you’ll find a link to my 2008 EuroCUP talk in that post.

      1. Ed says:

        I don’t think it accurate to describe “substitution with Br and I will increase the partition coefficient” unless one first indicates exactly how one would define partition coefficient.

        1. Peter Kenny says:

          Concentration (or mole fraction) of compound in organic phase divided by concentration (or mole fraction) of compound in water with system at equilibrium.

      2. Ash (Wavefunction) says:

        By “polar” I meant in the Pauling (electronegativity) sense. As you indicate Br and I are of course non-polar but they are not your usual non-polar carbon substituents. I have yet to meet a medicinal chemist who hasn’t reported at least one contradictory observation in SAR effected by Br or I.

        1. partial agonist says:

          you mean polarizable, rather than polar, then

    2. CMCguy says:

      Ash perhaps current medchemists here can better answer this but I think most would routinely attempt to use Brominated and/or Iodinated analogs as part of their discovery efforts and would pursue them as warranted by the results. However as usual it can come down to a matter of how easily it is locating or varying them on your target structures. Certainly many many potential raw materials containing Br or I can be identified, to me it often seemed cursed that the ideal chemical on paper was not readily available, commercially or via simple procedures, especially with the orthogonal functionality could employ to incorporate in desired scaffolds. Also because of there are some reactivity concerns for side chemistry can occur that leads to instability after prep or again more difficult incorporation without alterations to normal routes. I may be wrong but do seem to recall learning halogenated analogs can show increased metabolism and therefore means more complex PK and other profiles as progress. Everything is a balance and while would not hesitate to include them have to be aware having Br and I carries baggage have to haul around, where sometimes can be well worth it (and never really know upfront)

  4. Peter Kenny says:

    In the past, I have used presence of an adjacent HB donor and acceptor as a criterion for selecting fragments. I would consider promiscuity to be more of a benefit (think searching in parallel) than a problem when using X-ray crystallography to screen fragments. I would be asking is the bromopyrazole particularly promiscuous or is it simply easier to detect? The atomic weight of Br (and I) is also an advantage when using SPR for detection. If claiming ‘magic bulletness’ for a neutral fragment, I would be checking to see that it also bound where charged species were known to bind.

  5. Barry H Levine says:

    There’s nothing magical about iodo (or bromo-)pyrrazole. We’ve been soaking Xenon gas into crystals for years, and using that co-crystal to solve the phasing problem. Xenon of course does not participate in H-bonds as donor or acceptor.
    It’s not uncommon that a well-folded protein-with its ordered waters–will still leave voids. We get to see the ones in which iodo (or bromo-)pyrrazole or Xenon have a lower free-energy than dis-ordered waters

  6. Hap says:

    In Pfizer, where the shadows lie?

  7. RedFiona says:

    For non-related reasons I really did want that maths so thanks 🙂

  8. Mandrake says:

    Reading this article reminded me of a drug I learned about 5 months ago – It works by stabilizing transthyretin tetramers – It’s quite a small molecule at only 308.116 daltons but it is a beast in it’s own right! https://en.wikipedia.org/wiki/Tafamidis

Comments are closed.