PREreview del Conserved energetic changes drive function in an ancient protein fold

Review for: Conserved energetic changes drive function in an ancient protein fold

https://www.biorxiv.org/content/10.1101/2025.04.02.646877v1.full.pdf 

Summary:

Protein folds and their subsequent dynamics drive function and confer specificity. To understand how protein families with conserved topologies can carry out distinct functions, it is essential to examine not just their static structures but also their conformational ensembles and allosteric regulation. This paper explores the energetic relationships of functional specificity in the LacI/GalR family of transcription factors (TFs) and periplasmic binding proteins (PBPs), which share the Venus flytrap fold (VFT) but perform different biological roles.

Through leveraging evolutionary conservation analysis, hydrogen-deuterium exchange mass spectrometry (HX/MS), and molecular dynamics simulations, the authors explore how energetic landscapes and local dynamics are modulated by ligand binding. Their HX/MS pipeline provides a residue-level resolution in the dynamics of protein conformational changes upon ligand binding. The results suggest that despite the conserved structural motif, the energetic blueprints of TFs and PBPs have evolved distinctly to confer their unique functions. They elucidate the unique mechanisms of ligand binding in the different families, with TFs experiencing subtler allosteric structuring and PBPs undergoing a hinge-mediated conformational change. Using molecular dynamics simulation, the authors uncovered crucial localizations of water molecules that drive the allosteric behavior within the TF ensembles.

Major Points:

• The paper could benefit from a bit more explanation of the ∆G_op and list the assumptions that were made to infer these values. In particular, although Figure 2C and its caption visualize the experimental workflow and results of HX/MS using PIGEON/FEATHER, they rely heavily on citation to prior work and a more intuitive explanation without digging back into prior literature would be helpful.

  •  The “HX/MS experiment” diagram introduces the kinetic constants of k_op and k_cl as the protein transitions from the open and closed states. An addition of a simple visualization of the structural differences of the two states could aid in highlighting the relevance of these constants.

  • Figure 1C details out the equations that calculate the exchange rate and protection factor, which in turn derives ∆G_op. In the equation for the exchange rate, k_op and k_cl are shown within panel “HX/MS experiment”, but k_int is not shown in this figure. It could be helpful to explicitly convey k_int as the exchange of the “open” state hydrogen to the “open” state deuterium, as it is done in Ref 17’s overview of the PIGEON-FEATHER workflow.

• Figure 2D presents an ambitious and comprehensive summary of residue-level ∆G_op across multiple proteins, mapped alongside SASA, sequence conservation, and structural topology. This figure is currently difficult to interpret due to its dense presentation of the data, but still contains valuable insights that can become even more impactful with more guidance on navigating the figure.

  • The relationship between the SASA trace, sequence conservation, and the ∆G_op should be established. Highlighting the regions of particular interest between the relationships between these traces would be helpful for the reader in navigating the information-rich figure.

  • The authors highlight several regions of interest in the changes of ∆G_op in different states across different proteins, but these patterns are currently not immediately apparent. Because the supplemental table S8 lists quantitative values for each of the residue’s ∆G_op, the figure could be easier to digest if the ∆∆G_op (between the apo and bound states) are plotted to eliminate an additional row of data visualized in the figure.

• Throughout the paper, the authors present numerous compelling arguments based on the comparisons of ∆G_op profiles. We interpret the context in this paper and prior papers that the authors suggest that the absolute ∆G_op values are only directly comparable when the underlying closed-state baselines are equivalent, which is an assumption that may not be true across different proteins. Since most of the insights stem from the intra-protein ∆∆G_op patterns, noting this point will provide a more thorough context while still maintaining the study’s conclusions.

Minor Points:

• In Figure 1B, the coloring of both the conservation score plot and the annotation of the protein sequence could be misleading, especially in the PBP plot, as the colors of the N-term VFT designation and the bar chart appear identical.

  • The parallel usage of the colors in the TF plot in Figure 1B that highlights low conservation score to orange and high score blue to the scale used in Figure 1D could also be employed to improve readability.

• In Figure 1B and 1D, the authors note that the N-term VFT domain of the TF family exhibits lower sequence conservation as compared to the DNA-binding and the C-term VFT domain. While Figure 1B illustrates this well, Figure 1D is less convincing without careful inspection.

  • Adding annotations of the N-term and C-term domains, similar to Figure 1E, would help in interpreting this result along with the alternative view already provided.

• In Figure 1D, the authors point to the sequence variability in the TF dimer interface. Currently, the site of the greatest sequence diversity seems to appear at the solvent-exposed surface of the protein. The figure could be simpler to interpret if it directly indicated the interface of interest.

  • Highlighting the residues that are part of this dimerization interface in figure 1B could be helpful to demonstrate a meaningful difference in the conservation score between the dimer interface and other parts of the protein, such as the DBD.

• Figure 3D demonstrates how the operator binding stabilizes the different helix interfaces. The bar chart of the ∆G_op is difficult to interpret due to the overlapping gray and light green bars, indicating the APO and DNA state. The overlay creates a shade of green that seems to more closely match the color of the DNA state in the legend. Thus, it makes the figure seem like there is an alternative state of light green that is not yet described. Using different color schemes or even redesigning the figure to highlight the ∆∆G_op could easily eliminate some of the confusion.

• Improve the overall readability of Figure 4 to strengthen the claims about sugar-binding induced structural responses.

  • Re-highlight the regions of interest that are specifically called out, such as the hinge loops and transducer helix, in Figure 4B and 4E. 

  • While matching the spatial information of the residues in the binding pocket to the radar plot in figure 4C/D is helpful in highlighting the region of interest as the ligand binding pocket, it is still difficult to fully comprehend what they are describing. Clearly distinguishing the difference between the lines and the boxes at the edge of the radar plot could alleviate some of the confusion of this figure. The caption tries to explain these shapes, but does not explicitly discuss what the additional elements on the outside of the radar plot is. Additionally, the data presented in the radar plot could be reformatted to demonstrate the ∆∆G between the two states to reduce the complexity of the figure.

• In the section “Two VFT-fold protein families respond differently to binding the same sugars”, the authors claim that sugar binding triggers global stabilization in TFs through allosteric restructuring, pointing to a beta strand slippage as a mechanism. A more explicit connection between the stability data and these claims could be helpful for the reader since this claim suggests interesting and exciting mechanisms of sugar binding in PBPs and TFs.

  • Since the authors have already gathered and presented X-ray crystallography data in Figure S4, including a detailed caption or text in the figure to explicitly highlight the structural data that they have already connected will further bolster their claims about their hypothesized mechanism through the stability measurements.

  • If HX/MS data is sufficient to gather mechanistic details on these structural changes, an explanation of interpreting ∆∆G_op to different secondary structure formation, such as a “super-sheet” will aid the readers’ understanding of these claims.

• Figure 5A is not referenced in the text– it appears that it should have been referenced in place of Figure S4-2 during the section “Water molecules structure the induced states of LacI/GalR TFs”.

Competing interests

The authors declare that they have no competing interests.