The sequence PPGPXPXPXXGN is highly conserved in some cytochromes P450 across species from humans to plants. Within species, however, this specific sequence is not conserved although a proline-rich sequence is present. In CYP2C proteins, mutagenesis of the prolines and glycine in the PPGPXPXP part of the sequence results in less efficient assembly of native P450 molecules, but those molecules that are formed properly have specific activities similar to wild type suggesting that this region plays a role in folding of the protein. Further, the pattern of requirements for prolines was consistent with a left-handed polyproline II (PPII) helix structure in the PPGP segment. The recent determinations of the structures of CYP2C proteins permit a reinterpretation of these earlier experimental studies. In CYP2C5, the PPGPXPXP part of the sequence is in a left-handed polyproline II-like sequence with a 90 degrees bend at the glycine residue. The PXPXXGN part forms a hairpin structure with the remaining sequence protruding at a right angle. The structure forms tertiary interactions with protein segments centered on Tyr-61, Phe-219, and Tyr-376. van der Waals contacts of the rings of prolines with those of the equally highly conserved tyrosine residues may be particularly important. The proper positioning of the N-terminal segment containing Tyr-61 and the C-terminal segment containing Tyr-376 by interactions with the proline-rich region may be important for proper folding because these residues are in loops extending from four strands of a beta-sheet structure. A schematic model of the sequential folding interactions is presented, and although speculative, it is proposed that constraints required for the folding of a hydrophobic knob that will be inserted into the membrane contribute to the high conservation of the PPGPXPXPXXGN sequence in a subset of cytochromes P450.