Posted by: Dan | July 12, 2006

Modeling Evolution of Protein-DNA Interactions

Much like Bridgham et al.‘s reconstruction of hormone-receptor evolution by molecular exploitation, and in contrast to Dembski’s bad math, PLoS Computational Biology has an article out on the Evolutionary potential of a duplicated repressor-operator pair: simulating pathways using mutation data.

The presumption is that duplication events are important for the acquisition of novelty and complexity in the evolution of regulatory pathways, and that following duplication, selection and the fitness landscape can easily facilitate coevolutionary divergence in protein-DNA interactions. Poelwijk et al. state that “Our results contrast with the notion that a number of neutral or even deleterious mutations have to accumulate before a new function can develop.” Let’s see how they came to that conclusion (below the fold):

The authors focused on the creation of new and unique protein-DNA recognition, starting from two identical repressors and two identical operators, and consider selective conditions that favor the evolution toward independent regulation.

At a higher level, features of the network topology shape the landscape surface and divergence potential. We found that the tightly interconnected topology, as present after the duplication, does not frustrate divergence but instead promotes it. In contrast to an isolated repressor-operator pair, where a drop in the binding strength decreases the fitness, the same mutation can be neutral in the interconnected topology. Compensation for the decrease in binding strength can be attributed to two features of the topology. First, there is the characteristic pressure to not bind the rival operator: when a mutation decreases an interaction that should be maximized, this negative effect on the fitness is partly balanced by the decrease of an unwanted cross-interaction. A second mechanism is a coevolutionary twist on Ohno’s original idea, in which one repressor-operator pair can search for a new recognition, while the other repressor maintains repression on both operators in the very early stages. As we have observed that a drop in the binding strength is necessary for efficient divergence, the ability to compensate for its negative contribution to the fitness is crucial for funneling.

This nicely reflects a finding of Evo-Devo and duplications of the Hox genes in comparative development studies of vertebrate and nonvertebrate animals – at some point over 550 million years ago, several duplication events occurred within the developmental “toolkit” genes, enabling greater anatomical and physiological complexity to arise. Similarly, duplication of eukaryotic regulatory pathways has clearly occurred at a few stages in the last 2 billion years, as is evidenced by the expansion of paralogs (e.g. the eukaryotic kinome).

Interestingly, Poelwijk et al. consider further experiments:

Similarly, one could attempt to evolve a duplicate lac repressor/operator copy towards the independent regulation of a second operon. However, this more complex assay does require key modifications: (1) growth and selection of the mutants should occur in alternating media, in analogy to our discussion of multiple input conditions, and (2) a starting network must be engineered that satisfies the conditions for DNA-binding divergence: a duplicate repressor/operator and a selective pressure for tight and independent binding.

The first point is critical, because evolution of novel regulatory pathways requires novel functions, such as specialization for survival in alternative conditions. Again, this has a parallel with animal evolution and the acquisition of complexity: many of the regulatory pathways novel to vertebrates are associated with the acquisition of new physiological functions (e.g. a greater variety of tissue types, appearance of an adaptive immune system, etc.).


  • Evolutionary potential of a duplicated repressor-operator pair: simulating pathways using mutation data. Poelwijk FJ, Kiviet DJ, Tans SJ. PLoS Comput Biol. 2006 May; 2(5):e58. Epub 2006 May 26. Pubmed.
  • Evolution of hormone-receptor complexity by molecular exploitation. Bridgham JT, Carroll SM, Thornton JW. Science. 2006 Apr 7; 312(5770):97-101. Pubmed.

Related posts:


  1. […] This diversification from a single enzyme fits extremely well with previous models of evolution of signaling complexity, as put forth most recently on this blog with Modeling Evolution of Protein-DNA Interactions. […]

  2. […] As his alternative, Lynch proposes that random genetic drift and mutational pressures account for the acquired robustness and diversity of storage of genetic information – something very much echoed by observations on gene duplication in divergence of protein-DNA interactions, and evolution of digital organisms and computer viruses. […]


%d bloggers like this: