The nomenclature has evolved over time. The first edition used roman numerals for family names, and that was dropped quickly. Newer developments are requiring adequate solutions, such as how do alternative splicing variants get named? We have not answered that one yet. The definition of families and subfamilies has shifted toward lower cutoffs for inclusion. A more realistic view today is that a cutoff window is better than an absolute value like the original 40% identity or higher for family inclusion. More problematic is the merging or blurring of subfamilies that used to be distinct. The CYP4 family is a prime example. It is now so full of subfamilies that they are hard to keep from overlapping. This makes naming new members very difficult.

My history in this area goes back to my first post-doc with Henry Strobel at the University of Texas Medical School in Houston. I joined Henry's lab in 1985 when the first P450 sequences were just being published. I remember writing these sequences out by hand on graph paper to make alignments of about a dozen sequences. These were color coded with highlighters and posted on the wall in the hall outside the lab. They took up about eight feet of wall. When the alignment reached 34 sequences I sought some help from the Center for Demographic and Population Genetics at UT. Masatoshi Nei, who was chairman there, introduced me to Joel Claiborne "Clay" Stephens who gave me a FORTRAN program for doing UPGMA trees. I converted it to BASIC to run on an IBM PC and made my first P450 tree. That was the basis for my first P450 publication "Evolution of Cytochrome P-450 Proteins"³. That paper prompted Dan Nebert to ask me to join the nomenclature committee, and I have been doing P450 naming ever since.

Where will this lead? The immediate goal is to find every member of the P450 superfamily in the completed genome sequences from as many species as possible. The ability to point to complete sets of genes is powerful. It defines history by saying we know absolutely what is here now in all these different lineages, so how did we get here? That is the more interesting question. I am now caught up with the genome projects on Dictyostelium (45 P450 genes and counting), Neurospora crassa (38 P450s), Phanerochaete chrysosporium (white rot fungus, about 150 P450 genes), Fugu and Tetraodon (pufferfishes), rice (about 290 P450 contigs) and soon Ciona intestinalis (a tunicate). Human, fly, worm, and Arabidopsis are already done, with mouse to follow soon, so the next phase will be comparative genomics of all these P450s. I am optimistic that the evolutionary holes in the data set will be filled. It would be grand to have an echinoderm set and a sponge, as well as many protist genomes, especially a Stramenopile and a non-parasitic alveolate (like Tetrahymena). It was a bit of a shock this summer when the Plasmodium falciparum genome was searchable and I found no P450s in this aerobic parasite, not even a CYP51.

I would like to praise all the genome sequencers who have worked so hard to fill gaps in genomes and put the data out in a timely fashion, sometimes daily. I use it every day and I appreciate it being there to search, even if it is not at Genbank, but on private genome project servers. One of my biggest frustrations is knowing that genomes are done already for rice and mouse and these are not publicly available. I have signed private legal agreements to see some of this (rice) data, but I doubt the utility of holding this back and imposing restrictions. It will all come out in the end, so why hold it hostage now? Fortunately, there is no shortage of blast-searching to be done at NCBI and elsewhere. I wish they all had faster computers.

David R. Nelson, Ph.D.
Department of Molecular Sciences
University of Tennessee
Memphis, TN, USA