Bring In the Fluorines

One of the biggest differences between natural products and man-made pharmaceuticals is the presence of fluorine. Medicinal chemists love the odd things that a fluorine atom can do  –  it changes electron density drastically, battens down a C-H center to make it resistant to metabolism, alters solubility and other properties, and more. There’s nothing like a C-F bond, but handling fluorine chemistry is a real challenge for biochemistry. Fluorine itself is never found in nature as the pure element; it’s too insanely reactive for that. And when you do find it in nature, it’s often tied up in compounds that are so energetically favorable that it’s hard to pull the fluoride back out to do anything with it. The most common fluorine-containing mineral is calcium fluoride, known as fluorite or fluorspar. It’s soluble in water, although not very (that’s how you get those amazing crystals of it in geological formations). We make hydrofluoric acid from that stuff on an industrial scale by treating it with concentrated sulfuric acid, but that’s not exactly a friendly conversion for living cells. Meanwhile, fluoride is reasonably toxic to living cells all on its own. Toothpaste levels aren’t going to do much to you, but eating several grams of a fluoride salt is a really deadly idea, as is drinking water with a naturally high fluoride content. Sustained exposure to 10mg/day or more can lead to serious effects on the bones and other organs.

Even some of the (very few) fluorine-containing natural products can be toxic, with fluoroacetic acid at the top of that list. It’s produced in a number of tropical plants, and it stops the Krebs cycle in its tracks, which as you can imagine is not a good plan. With all this, the good part about the use of fluorine in drug structures is that we’re always looking at C-F bonds, which are quite stable and generally not sources of fluoride ion. You really want to avoid any structure that gives fluorine a chance to be a leaving group, such as putting it a benzylic carbon. And you also want to avoid putting a monofluoro on the end of an alkyl chain, for fear of further oxidative metabolism eventually taking that down to fluoroacetic acid. 1-fluorohexane, for example, is notably toxic for just that reason.

But when used judiciously, the pharmacological benefits of sprinkling in C-F bonds can be tremendous, so we’re always looking for better ways of doing that. The intersection of fluorine with natural products would be a particularly promising area if that were easier to accomplish, and this new paper has a way to try that idea out. They’ve engineered an enzyme in the biosynthetic path to polyketide natural products (malonyl-CoA transacylase) to be more of a fluoromalonyl-CoA transacylase. It’s mostly accomplished by making it less active in transferring the “standard” malonyl groups while maintaining activity for fluoromalonate, and the end results are polyketide products decorated with C-F bonds. The team behind this work has been active in this area for some years, and it’s not an easy project: the other enzymes in the biosynthetic pathway don’t always play well with the fluorinated intermediates they’re being handed, for example. 

They’ve shown enzymatic production of fluorinated analogs of erythronolide B (at right), which is a perfect example of a polyketide (and an example of fluorinations that would be very tedious (or not even possible with current chemistry). They can produce the 2-fluoro-2-desmethyl version as well as the 4-fluoro-4-desmethyl version, depending on the way they set up the in vitro enzyme system. In cells (E. coli), the system produced mg/liter amounts of various fluorinated triketide fragments. Yields were lower for the final products than the were with the in vitro enzyme system, but they were indeed produced. There’s going to be a lot more engineering needed to turn bacteria into fluorination machines, although this work does show that the goal should be possible. And it will produce compounds that we’ve never seen before – fluorinated antibiotics like erythromycin analogs and more, and their activity should be really interesting to explore!