Rational redesign of bacterial signaling proteins based on amino acid co-evolution

Whether independent and unicellular or nestled within multicellular organisms, cells must have the ability to respond to their environment in an appropriate fashion. Responding to physical and chemical aspects of the environment is crucial, but so is responding to neighboring cells that also inhabit that environment. In multicellular animals a surprisingly small set of molecules coordinate the complex communication between cells to direct important cellular processes, such as cell growth, division, and migration, that ultimately translate into tissue-level and organismal properties.

In bacteria, the situation is perhaps even more extreme. Most environmental signals are relayed through the action of a single type of system, the two-component signal transduction system. The first component is a histidine kinase (HK) that acts as the sensor for the environmental stimulus. When is detects the stimulus, it enzymatically tacks on a phosphate group to itself at a specific histidine amino acid. It then transfers this phosphate to the second component of the two-component system, the response regulator (RR). This transfer activates the response regulator, which can trigger changes in gene expression or other aspects of bacterial physiology. Each HK interacts exclusively with a matching (or “cognate”) RR.

Most bacteria have between 20 and 30 matching HK-RR pairs, though some can have as many as 300. The HKs are fairly similar to one another, as are the RRs. How, then, is specificity maintained so that each HK activates only the RR it’s supposed to and not any of the others? Whether in vivo or in vitro, HKs are incredibly specific, so this information to avoid crosstalk must be contained within the amino acid sequence.

Michael Laub’s lab at MIT reasoned that the amino acids important for the interactions between HKs and RRs should co-vary with one another. That is, if an amino acid is important for a specific interaction, and a mutation alters that amino acid, then to maintain the interaction that amino acid must revert or there must be a corresponding change in the other partner. Thus, the amino acids in physical locations most crucial for specific interactions should co-evolve with one another.

Skerker et al. identified nearly 1300 HK-RR pairs from nearly 200 sequenced bacterial genomes and, because all these proteins are sufficiently similar, lined up all their sequences against one another to compare. They found 43 pairs of residues that their algorithm suggested co-varied more often than expected by chance alone. Thirty-six of these pairs were between the RR and a region of the HK called the DHp domain, suggesting that this region was most important for the HK-RR interactions.

Next, they swapped the DHp domain from one protein onto the remaining portion of a different HK to make a chimeric protein. This switch caused the HK to interact with the RR corresponding to the DHp domain, not the rest of the protein. Narrowing down further, they switched just a small part of the DHp domain, and observed the same result—whatever the identity of the DHp subdomain, that’s the RR with which the chimera interacted in a highly specific manner. Even though it was a combination protein, it exclusively interacted with its new, DHp-defined partner, and not at all with its original partner.

These DHp subdomains contained seven of the amino acids identified from the computational analysis that suggested they were involved in determining specificity. By changing a small number of these amino acids the authors attempted to convert the EnvZ histidine kinase to act on the RR of the RstB kinase instead. Altering just one amino acid resulted in a protein that could interact with both RRs, altering two amino acids produced an enzyme that preferred the RstB substrate but could still interact a bit with the EnvZ substrate. Changing three amino acids, though, was enough to completely convert the specificity of EnvZ to that of RstB. Finally, the authors demonstrated that these changes could also rewire the flow of signaling information in vivo, not just in vitro.

Based on the evolutionary prediction of co-evolution of amino acids that are crucial for the specific interaction of two proteins, these researchers were able to identify exactly which amino acids those were to such a degree of certainty that they could predictably change the specificity of a bacterial histidine kinase by changing a very small number of amino acids. The authors of this study were interested in how similar proteins can interact in precise specific ways, not really in evolution. But they used concepts of evolution to crack this problem. So, the next time an ID creationist suggests that molecular biology research has no need for evolution and would proceed no differently if evolution was discarded or ignored, we can point to this paper as another fantastic counter-example.

SKERKER, J., PERCHUK, B., SIRYAPORN, A., LUBIN, E., ASHENBERG, O., GOULIAN, M., LAUB, M. (2008). Rewiring the Specificity of Two-Component Signal Transduction Systems. Cell, 133(6), 1043-1054. DOI: 10.1016/j.cell.2008.04.040