When [Carl] Pabo went back to refine this structure, of course he found this. It turned out that he'd actually seen this interesting crystallography, he's a very pure crystallographer and if he doesn't understand the result he doesn't publish it. When he looked at his map, and he told me this, and we could see it in his map anyway, he could see that there were interactions with the second strand but it was from a... it was from a glutamate to a cytosine. Now the glutamate is negatively charged, now it interacts... and so... now we, I knew this but I also knew that in other crystal structures of certain... what do you call them? Restriction endonucleases you've sometimes got this interaction and I don't know how it came about because it seemed to defy the laws of chemistry, but Pabo saw it and he didn't put... he didn't say anything about it, just omitted to mention it. But when you refine the structure to 1.98 resolution he saw very clearly what... well, that was done some years later. It's very cunning, what happens is that the glutamate, I'm sorry to go into such detailed stereochemistry but it's one of the things you really have to understand deeply if you're going to work in this field, the glutamate has a couple of aliphatic residues before you get to the acidic residue. And what Pabo didn't know was that the five six bond of cytosine is hydrophobic. And so, these aliphatics... the binding is not to the... it's not hydrogen bonding at all, which as most of the other interactions are, its hydrophobic bonding and it involves interaction between the aliphatic residues of glutamate and the five six bond of cytosine. And I suspected that because I knew about the five six bond of cytosine from our tRNA works. And then the acidic part bends backwards and binds to the backbone of the protein... so then Pabo published that later. So... but there are many different ways you can recognise a cytosine. Valine will do it because it's a methyl group, so will alanine and so on. So in the end we did work out a stereochemical code but it's very complex so you... so in the end you would have to... I couldn't write down by that time, write down by eye the sequence of zinc fingers which will bind the DNA but it won't be the highest binding. To find the highest binding you've got to make combinations of those different sites, that's really... that's really what it amounted to. That's why the selection... so thought we used to write, particularly Yen Choo, we always talk about tailor-made proteins... they were selected proteins, selected zinc fingers.