PREreview del Distinct energetic blueprints diversify function of conserved protein folds

This pre-review stems from a journal club discussion of the preprint and reflects the contributions of Julia Zhang, Daniel Keedy, and several anonymous lab members.

This study investigates a fundamental question in structural biology: how proteins with highly conserved folds can evolve distinct functions. Using an integrative approach combining HDX/MS, X-ray crystallography, bioinformatics, and molecular dynamics, the authors compare LacI/GaIR transcription factors (TFs) and periplasmic binding proteins (PBPs), which share the Venus flytrap (VFT) fold but perform different biological roles. They demonstrate that these protein families encode distinct “energetic blueprints”, reflected in differences in residue-level free energies of opening (∆Gop) despite structural similarity and shared ligand binding. These energetic differences are revealed to have functional consequences, as validated through rational design of LacI variants with tunable ligand sensitivity. This work suggests that evolution sculpts energetic networks rather than structure alone, reframing how we interpret structure-function relationships.

We particularly appreciated the conceptual clarity and visual representation of the work. The comparison between parallel systems (TFs vs. PBPs) is especially effective in highlighting how conserved folds can encode divergent energetic behaviors. Figures are generally well-designed and intuitive. Figure 2 clearly separates qualitative structural information from quantitative ∆Gop data. Figure 3 effectively uses color to map secondary structures. Figure 4 presents engineered variants with clearly labeled designs and outputs. Figure 6 succinctly summarizes localised vs. distributed energetic effects in LacI/GaIR TFs, reinforcing the central message. Beyond the study itself, the implications are exciting. Extending this framework to other allosteric systems — such as other metabolite binding proteins or enzymes — could reveal how energetic landscapes regulate catalytic dynamics. Integrating HDX/MS, sequence analysis, and simulations may also enable prediction of cryptic allosteric sites for drug design, and provide insight into how disease-associated mutations disrupt energetic regulation.

Despite these significant strengths, some aspects of the data presentation and interpretation could be clarified. Several figures lack clear definitions of how ∆∆Gop values are combined across proteins (e.g., Figures 2A and 3B). In addition, labeling inconsistencies — particularly in Figure 3C, where the meaning of the “30/60” values is unclear and the font appears visually inconsistent — reduce interpretability. In Figure 2C-1, it is also unclear whether the use of an AlphaFold model of GalR (?) is appropriate to support conclusions about conserved allosteric effects of DNA binding. These effects are less convincing for GalR, as the highlighted regions that appear shared between LacI and RbsR are not clearly recapitulated. Relatedly, visual overlays in Extended Data Figure 4 do not always clearly align across the family, making it difficult to assess whether DNA binding induces similar energetic changes across all three TFs. In Figure 3F, the conclusion that similar TFs maintain stabilized secondary structural elements around ligand-binding regions is not fully supported by the visual overlays, where the purple and green regions show only minimal or inconsistent overlap; it is therefore unclear whether this panel adds substantial evidence beyond earlier analyses. In Figure 5A, GalR is categorized under X-ray-derived water molecules despite being based on an AlphaFold model, which may introduce confusion regarding the structural basis of these observations. Furthermore, in Figure 5C, while high occupancy of specific water molecules is reported across TFs, the relatively short maximum durations observed for GalR and RbsR raise questions about the functional significance of these interactions. This discrepancy makes it difficult to reconcile occupancy with stability, and it is unclear whether inclusion of the “max duration” metric strengthens the authors’ conclusions or instead introduces ambiguity. In general the figure captions are relatively short and sparse, and could benefit from additional detail to clarify some points like the ones raised above.

A few additional points are as follows:

• Regarding this sentence: “Interestingly, the designed variants included several mutations that ablate LacI function when introduced individually." Which mutations were these? Were they in the higher EC50 designs?

• Regarding this sentence: "The water does not appear in operator-bound TFs (Fig. S5-2)29, and we expect only transient waters in the binding pockets of apo TFs." What do RosettaECO and/or MD have to say about that expectation? Not a huge priority, but this might benefit from more support to confirm that the water is only there when sugars bind.

• Figure 1 caption: "Right: VFT-fold structure schematic." We think this should say "Top".

• “b-factors” should have capital B, by convention.

• “APO” is not an acronym; would be better as “Apo” or “apo”.

Despite these points, the overall framework presented is compelling and has strong potential for broader impact. With clearer data presentation and refinement of key interpretations, this work could serve as a powerful foundation for extending energetic landscape analysis to other allosteric protein families and for guiding future efforts in protein engineering and mechanistic understanding.

Competing interests

The authors declare that they have no competing interests.

Use of Artificial Intelligence (AI)

The authors declare that they did not use generative AI to come up with new ideas for their review.