Emergence of Ultra-Conserved Protein Domains and Amino Acid Repeats: Adaptation, Competition and Thresholds

Some proteins, such as homeodomain transcription factors, contain highly conserved regions of sequence that cannot be attributed to the constrains imposed by any single function. It has recently been suggested that multiple conserved functional domains overlap and together explain the high conservation of these regions. However, because these highly conserved domains are part of much larger proteins, we are still left with the question why so many functional domains cluster together. Here we have modeled an evolutionary mechanism that can produce this kind of clustering. Due to adaptive competition between different protein functions for control over amino acid residue identity, conserved functional domains get displaced from regions undergoing adaptive evolution. At first they undergo a steady random walk within the sequence for an indefinite amount of time; however, a threshold is reached when two functional domains happen to come into contact, at which point there is a dramatic shift in the adaptive dynamics such that the domains rapidly converge, lengthen, and evolve overlap — stabilizing at a fully overlapped state. We also studied the evolution of single amino acid tandem repeats (a.k.a. homopeptides), which are especially prevalent in transcription factors. Homopeptides that are encoded by nonhomogenous mixtures of synonymous codons cannot be explained by the neutral process of replication slippage. Our model provides two ways to explain the origin and maintenance of such repeats, and their over-representation in highly conserved proteins: competition between multiple functional domains for space within a sequence, or reuse of a sequence for many functions over time. Both processes depend on reaching certain critical thresholds, however they both deterministically cause the evolution of repeats once these thresholds are reached. Further, both of these processes are characteristic of multi-functional proteins such as homeodomain transcription factors. We conclude that our model can explain two widely recognized features of transcription factor proteins: conserved domains and a tendency to accumulate homopeptides.