Tag Archives: biomimicry in reverse

Making an enzyme do a carbene reaction

The P450 cytochrome enzymes are the chemical workhorses of biosynthesis and xenobiotic detoxification.  As a neurologist I always had to worry about the way adding a new anticonvulsant to a an existing set would alter the body levels of all of them.  The new anticonvulsant would cause the P450 cytochrome enzyme metabolizing it to increase, and the new P450 might also chew up some of the others.  Phenobarbital was a particular problem this way.

Over 14,500 P450 genes have been found in various organisms. FYI the 450 in P450 comes from the absorption at 4500 Angstroms of the Fe++ in the heme group when complexed to carbon monoxide.  This is called the Soret band.  The P450 cytochromes are a superfamily comprising 36 known gene families, each consisting of 2 -20 discrete cytochromes.

The P450 cytochromes take 2 electrons from a reduced cofactor (NADPH) which itself gets the electrons from another molecule.  The cytochrome then binds an oxygen molecule splitting it so an oxygen atom is bound to the Fe++  (the other oxygen goes to water).  The oxygen is then used by the various enzymes to insert into C-H bonds with an incredible specificity (this produced by the conformation of the protein around the heme-iron-oxygen).

In Science vol. 339 pp. 283 – 284, 307 – 310 ’13 a group at Caltech which has been using direct evolution (randomly mutate a protein and select for the reaction you want) used a P450 cytochrome to accept diazo esters as reagents.  This formed a carbene linked to the Fe.  They found that it would add to olefins.  Further engineering (random selection actually) produced variants of the P450 to produce diastereoselective and enantioselective cyclopropanation of styrene.

It would be nice to see the crystal structures of the engineered enzymes carrying out the reactions, to see how they constrain the styrene to react so specifically.

So here we have chemists acting like the blind watchmaker of Dawkins.  It would be nice if theory was good enough to predict and design the P450 variants produced here a prioi, but it isn’t.  The editorialist calls it biomimicry in reverse.