Fluid protein fold space and its implications

Abstract

Fold-switching proteins, which remodel their secondary and tertiary structures in response to cellular stimuli, suggest a new view of protein fold space. For decades, experimental evidence has indicated that protein fold space is discrete: dissimilar folds are encoded by dissimilar amino acid sequences. Challenging this assumption, fold-switching proteins interconnect discrete groups of dissimilar protein folds, making protein fold space fluid. Three recent observations support the concept of fluid fold space: (1) some amino acid sequences interconvert between folds with distinct secondary structures, (2) some naturally occurring sequences have switched folds by stepwise mutation, and (3) fold switching is evolutionarily selected and likely confers advantage. These observations indicate that minor amino acid sequence modifications can transform protein structure and function. Consequently, proteomic structural and functional diversity may be expanded by alternative splicing, small nucleotide polymorphisms, post-translational modifications, and modified translation rates.

Keywords: alternative splicing; metamorphic proteins; protein evolution; protein fold switching; protein folding.

Figure 1.

The single-fold hypothesis and fold-switching observations suggest different models of protein fold space. a) The classical single-fold hypothesis suggests a discrete fold space in which homologous sequences encode very similar fold architectures, and different folds are encoded by disparate protein sequences. Consistent with this assumption, structural basins (contours) encoding distinct fold architectures (green: α+β, yellow: mainly β, blue: mainly α, and red: α/β) contain homologous sequences that each assume the same architecture (dots within the contours). Architectures are represented by XCL1 (chemokine fold, green, PDB ID: 1J8I), the C-terminal domain of RfaH (β-roll fold, yellow; its N-terminal domain is gray, PDB ID: 6C6S, chain D), the human-serum-albumin-binding streptococcal protein G_A (3-α-helix bundle, blue, PDB ID: 2FS1, chain A), and the immunoglobulin G (IgG)-binding streptococcal protein G_B (α/β-grasp fold, red, PDB ID: 1GB1, chain A). b) Fold switching suggests that fold space is more fluid than previously realized because homologous amino acid sequences can assume multiple folds with different secondary structure arrangements. Contours from the all β-sheet alternative fold of XCL1 (green, PDB ID: 2N54) overlap with the all β-sheet fold of RfaH (yellow, PDB ID: 6C6S, chain D). These overlapping contours are shaded light green since they contain folds with the same major architecture. Furthermore, contours from the all α-helical alternative fold of RfaH (yellow, PDB ID: 5OND, chain A) overlap with contours of the 3-α-helical bundle fold, shaded purple to represent overlap between all-α and α/β folds resulting from engineered G_A and G_B variants with extremely high levels of sequence identity but different folds. Accordingly, the α/β-grasp fold is shaded magenta also to reflect this overlap. Both plots were calculated by performing multidimensional scaling [22] on pairwise structural differences [23] between all folds and plotting the resulting contour maps and datapoints (Supplementary Text). Structures in all figures made with PyMOL [24].

Figure 2.

Fold-switching proteins can have different mechanisms and energy landscapes. (a) The mycobacterial protein PimA populates two distinct folds in nearly equal proportions under physiological conditions. Fold-switching regions of PimA are colored dark and light green (PDB IDs: 4N9W, chain A and 4NC9, chain A), respectively. Regions undergoing the largest conformational changes are enlarged below. In the dark green conformation (left), α₂ and β₃ switch to β₂ and α₃ in the light green conformation (right), respectively, and the orientations of the β-sheets change from parallel (dark green) to antiparallel (light green). (b) The bacterial plant pathogen PopP2 switches folds in response to the small molecule trigger InsP6. The fold-switching regions of its apo structure (cyan, left, PDB ID: 7F3N, chain A) switch to the dark blue structures upon binding Insp6 (right, PDB ID: 5W3T, chain A). Most notably, its long helix in the apo form switches to a β-sheet upon binding InsP6 (orange). This structural change allows its regulatory domain (pink) to assume a more ordered conformation and carry out its acetyltransferase function. (c) Most single-fold proteins, such as ubiquitin (Red, PDB ID: 1D3Z), have one deep energy well that allows for minor conformational fluctuations while the same fold is maintained. The energy landscapes of equilibrium fold switchers, such as PimA, have wells with nearly equal depths separated by a larger energy barrier. Finally, the energy landscapes of triggered fold switchers change in response to environment. Without Insp6, PopP2’s Apo conformation was calculated to be ~3 kcal/mol more favorable than its Insp6-bound fold [56]. Thus, its two energy wells have unequal depths. Their relative depths shift, however, upon binding Insp6, and the alternative fold becomes more favorable. All energy wells are for illustrative purposes only and do not quantitatively represent the true energy landscapes of these proteins.

Figure 3.

Biological processes that could trigger fold switching. Alternative splicing. A β-strand (yellow) of the short C₂A Piccolo domain is displaced with an alternatively spliced exon (red) in the long form (PDB ID: 1RH8), changing its calcium binding mechanism and affinity. The short form is represented by synaptotagmin (PDB ID: 1RHY), a homolog with calcium binding properties similar to the short form of C₂A Piccolo. Single nucleotide polymorphisms. The C-terminal helix of human MEF2b is switched to a b-strand by D83V, the most common non-Hodgkin-lymphoma-associated SNP in this protein. N- and C-termini are labeled for clarity. Post-translational modifications. Could hypothetically induce fold switching. A potential modification (yellow circle) is shown on the right figure. Altered cotranslational folding rates could also lead to changes in protein folds, as shown with the ω subunit of RNA polymerase, which is disordered when translated quickly but folds when translated slowly (black, PDB ID: 3EQL, chain O). N- and C-termini are labeled for clarity. Proposed triggers of fold switching with gray background are indirectly supported and hypothetical.

Figure 4.

Examples of experimental methods that have identified or could identify fold switching. (a). Circular dichroism spectroscopy was recently used to distinguish between single-folding (salmon) and fold-switching (teal) NusG proteins with different ground state folds [32]. This approach was successful because the secondary structures of their C-terminal domains have very different ground states: all β-sheet (single fold, salmon) and all α-helix (fold switch, teal). Structurally conserved N-terminal domains are gray. (b) Hydrogen exchange monitored by mass spectrometry was recently used to distinguish between two distinct conformations of the SARS-CoV-2 spike protein [106]. It has the potential to also distinguish between fold-switching proteins such as a recently engineered protein that switches from an α/β-plait (blue) to an all α-helical (red) fold as temperature decreases [76].