Original domain for the serum albumin family arose from repeated sequences.

The characteristic three-domain structure has been conserved throughout mammalian evolution by serum albumin and its fetal counterpart, alpha-fetoprotein. Thus, one still detects 35.2% amino acid sequence homology between bovine serum albumin and murine alpha-fetoprotein. Yet, natural selection cannot be invoked as the major factor responsible for the observed conservation of these sequences, for the simple reason of their dispensability. Inherited analbuminemia is apparently a harmless trait in man and the rat. The conservation appears inherent in their repetitious origin. Each protein is made of triplicate copies of the ancestral domain. Furthermore, analysis of the published sequence data suggests that the original coding sequence for the ancestral domain arose as repeats of the 18-base-long primordial building block sequence TTC-ACA-GAG-GAG-CAG-CTG specifying Phe-Thr-Glu-Glu-Gln-Leu and its shorter subsidiary TTC-ATG-GAG-GAG specifying Phe-Met-Glu-Glu. Consequently, the homology between bovine serum albumin and alpha-fetoprotein is mostly confined to small segments still specified by recognizable descendants of these building block sequences. The point to be made here is that evolutionary conservation of coding sequences can be an inherent property; natural selection need not be invoked.