Sequence‐encoded dynamics diversify allostery within a conserved protein scaffold

Abstract

Proteins evolve allosteric mechanisms through sequence variation while preserving a conserved structural fold. How these substitutions reshape the protein's conformational landscape to drive functional diversification remains poorly understood. Members of the cAMP receptor protein family (CRP), a ubiquitous bacterial transcription factor, share a common architecture yet differentially couple cAMP binding to DNA interactions. Here, our results support a model in which allosteric diversification in CRP arises from the tuning of sequence‐encoded protein's conformational dynamics. Comparative analysis of sequence conservation and helical propensity across proteobacterial and actinobacterial CRPs reveals lineage‐specific distributions of conformational heterogeneity within the dimerization C‐helix. These evolutionary patterns map onto the proteobacterial Escherichia coli (E. coli) CRP and the actinobacterial Mycobacterium tuberculosis (MTB) CRP, two protein homologs that respond to cAMP through distinct allosteric mechanisms. Guided by functional and evolutionary differences, we introduced CRP _{E. coli} ‐inspired C‐helix substitutions into CRP _MTB to identify sequence, structural and conformational features that contribute to the divergent allosteric behavior in CRP _MTB . The resulting C‐helix variants preserved the global fold, but displayed altered thermodynamic stability, cAMP binding affinity and cooperativity, DNA recognition properties, and promoter‐associated oligomerization. Among these, the C‐helix hinge residue, N135, emerges as an allosteric node that redistributes cAMP‐induced conformational responses, modulating DNA binding. Together, our results identify the C‐helix as an evolvable hotspot where sequence changes reweight conformational ensembles and generate distinct allosteric outputs within a conserved protein scaffold.

1. INTRODUCTION

Proteins often evolve new regulatory mechanisms without abandoning the ancestral fold (Cooper & Dryden, 1984; Popovych et al., ²⁰⁰⁶). In many cases, allostery is tuned by mutations that reshape the energetic landscape and redistribute the conformational ensembles coupling ligand binding to function (Millard et al., 2025; Nussinov, ²⁰²⁵). Members of the cAMP receptor protein–fumarate and nitrate reductase (CRP‐FNR) family exemplify this principle, with homologs that retain a conserved structural scaffold yet exhibit distinct allosteric responses and regulatory outputs (Blake et al., 2002; Frendorf et al., ²⁰¹⁹; Körner et al., ²⁰⁰³; Zhou et al., ¹⁹⁹⁴).

In Escherichia coli (E. coli), cAMP receptor protein (CRP) is a homodimeric transcriptional regulator with each subunit containing an N‐terminal cyclic nucleotide‐binding domain (cNBD), linked by a long dimerization C‐helix to a C‐terminal DNA binding domain (DNABD) (Figure 1a). cAMP binding drives reorientation of the DNABDs, enabling high‐affinity sequence‐specific DNA binding (Dong et al., 2002; McKay et al., ¹⁹⁸²; Tomlinson et al., ²⁰⁰⁶). In contrast, the Mycobacterium tuberculosis (MTB) homolog, CRP _MTB , binds DNA with equally high affinity in the apo‐ and cAMP‐bound states (Figure 1b) (Gallagher et al., 2009; Reddy et al., ²⁰⁰⁹). Instead, cAMP binding strongly suppresses high‐order protein oligomerization at the promoter site (Dokas et al., 2025; Gárate et al., ²⁰²¹). Thus, despite the high structural conservation (~53% sequence similarity and a root‐mean‐square deviation (RMSD) of 2.5 Å of the cAMP‐bound structures) these homologs convert the same cAMP signal into distinct regulatory outputs (Kumar et al., 2010; Townsend et al., ²⁰¹⁴).

FIGURE 1.

Comparison of structures and C‐helices hinge‐associated hydrogen‐bonding networks between CRP _{E. coli} and CRP _MTB . (a, b) Crystal structures of CRP _{E. coli} and CRP _MTB highlighting differences within the dimerization C‐helices hinge region. In both CRPs, the N‐terminal cAMP binding domains are in blue (CRP _{E. coli} ) or beige (CRP _MTB ), while the C‐terminal DNA binding domains are shown in orange (CRP _{E. coli} ) or light blue (CRP _MTB ). The dashed red circles mark the C‐helices hinge region that links the two functional domains. Enlarged views (right panels) show the conformational changes upon binding cAMP (magenta spheres). In apo‐CRP _{E. coli} , the hinge adopts a coiled conformation, whereas in CRP _MTB the C‐helices remain helical, forming a pre‐assembled dimer interface. (c, d) Top and side view of CRP _{E. coli} . cAMP (magenta) forms hydrogen bonds (yellow dashes) involving residues T127 and S128 from the adjacent subunit. Specifically, T127 of subunit A (T127^A) and S128 of subunit B (S128^B) accept hydrogen bonds from the bound cAMP of subunit A (cAMP^A), and vice versa. These residues, in turn, donate hydrogen bonds to R123^A and L124^B, respectively, drawing the C‐helices closer together to propagate the coil‐to‐helix transition. (e, f) Top and side view of CRP _MTB . The homologous residues T134^A and N135^B, also accept hydrogen bonds from the bound cAMP^A. However, N135^B is proposed to donate an additional hydrogen bond to cAMP^A, rather than to L131^B. See Materials and Methods for Protein Data Bank codes.

A key structural distinction resides within the hinge region of the dimerization C‐helices (Figure 1a,b, dashed red circle) and their corresponding hydrogen‐bonding network (Figure 1c–f) (Chu et al., 2001; Dokas et al., ²⁰²⁵; Gárate et al., ²⁰²¹; Kumar et al., ²⁰¹⁰; Townsend et al., ²⁰¹⁴). In CRP _{E. coli} , residues T127 and S128 define a hinge that couples cAMP binding to redistribution of the ensemble, favoring conformations that stabilize the coil‐to‐helix transition at the dimer interface (Figure 1c,d) (Harman, 2001; Won et al., ²⁰⁰⁰; Youn et al., ²⁰⁰⁶). Conversely, while the corresponding residues in CRP _MTB (T134 and N135) occupy the same spatial region, they appear to encode a distinct conformational landscape (Figure 1e,f) (Stapleton et al., 2010). Moreover, N135 of subunit B (N135^B) of apo‐CRP _MTB forms a hydrogen bond with T134 of subunit A (T134^A), an interaction absent in CRP _{E. coli} (Figure S1) (Stapleton et al., 2010). It is possible that contribution from this intersubunit interaction may stabilize the C‐helix dimer interface, favoring retention of the pre‐organized C‐helix hinge in the apo‐state. This divergence suggests that the hinge may serve as a key site for sequence‐encoded tuning of allosteric mechanisms.

Here, we investigated whether lineage‐specific sequence features within the C‐helix hinge contribute to the divergent allosteric outputs of the two homologs. Using a combination of helical propensity and sequence conservation analysis, targeted mutagenesis, and biophysical characterizations, we assessed whether introducing CRP _{E. coli} ‐like hinge features into the CRP _MTB scaffold is sufficient to redistribute its conformational ensemble and reconfigure coupling among cAMP binding, DNA recognition, and protein oligomerization. Our results identify the hinge as an evolvable regulatory element, where modest sequence changes tune allosteric output without disrupting the conserved CRP fold, with N135 emerging as a context‐dependent control point of hinge dynamics in CRP _MTB .

2. RESULTS

2.1. Evolutionary signatures identify the C‐helix hinge as a candidate site of allosteric tuning

Comparative analysis of the calculated helical propensities revealed that proteobacterial and actinobacterial CRPs distribute conformational variability differently across the dimerization C‐helices (Figure 2). In proteobacteria, which includes CRP _{E. coli} , helical variability is largely concentrated outside the hinge, whereas the hinge itself remains comparatively constrained (Figure 2a). In actinobacteria, to which CRP _MTB belongs to, helical variability extends more broadly across the C‐helix, including the hinge (Figure 2b). These lineage‐specific patterns suggest that the hinge may serve as a potential site for allosteric tuning, rather than neutral sequence drift.

FIGURE 2.

Helical propensity and sequence analyses of C‐helix sequence properties in proteobacterial (top row) and actinobacterial (bottom row) cAMP receptor protein (CRP) homologs. (a, b) Residue‐specific helical propensities calculated for proteobacteria (top; CRP _{E. coli} ‐like) and actinobacteria (bottom; CRP _MTB ‐like). Individual homologs are shown as gray curves, lineage‐average profiles are shown as red curves with the red error bars representing the standard deviation (SD) at each residue position, and blue curves represent CRP _{E. coli} and CRP _MTB within their respective lineages. The dashed vertical gray line marks the boundary of the C‐helix hinge region. (c, d) Sequence logos showing residue conservation across the C‐helix of proteobacterial and actinobacterial CRPs. Letter height corresponds to positional conservation. (e, f) Scatter plots comparing conservation scores with the mean helical propensity at each residue position. No statistically significant relationship was observed in either lineage (proteobacteria, p‐value = 0.50; actinobacteria, p‐value = 0.51). (g, h) Scatter plots showing the correlation between positional conservation scores and the SD of helical propensity across each lineage. The green and magenta points represent residues in the pre‐hinge and hinge of the C‐helix, respectively. A negative correlation indicates that highly conserved residues exhibit reduced variability in helical propensity. This relationship is statistically significant in both proteobacteria (p‐value = 0.03) and actinobacteria (p‐value = 1.17 × 10⁻³).

We next asked whether these patterns track with amino acid conservation (Figure 2c,d). In both lineages, positions with lower variability in helical propensity are generally more conserved, however, the spatial distribution differed. Proteobacterial CRPs showed strong conservation across the hinge and adjacent positions, consistent with restricted local flexibility, whereas actinobacterial CRPs tolerated broader variations across the C‐helix. Thus, both lineages may preserve function through sequence constraint, but partition this constraint across different regions within the same structural element.

Quantitative comparison across both bacterial groups supported this interpretation. The average helical propensity showed little relationship with sequence conservation in proteobacteria (slope = 0.08, p‐value = 0.50) and actinobacteria (slope = 0.07, p‐value = 0.51).

The standard deviation (SD) of helical propensity, however, correlated more strongly with sequence conservation in both lineages. This relationship was stronger in actinobacteria (slope = −0.13, p‐value = 1.17 × 10⁻³), than in proteobacteria actinobacteria (slope = −0.07, p‐value = 0.03). Therefore, highly conserved residue positions exhibited less variability in helical propensity.

Although helical propensity does not directly measure protein dynamics, positional variation in propensity may reflect differences in the range of conformational flexibility permitted by the protein sequence. Consistent with this interpretation, the observed relationship suggests that evolutionary pressures may help define where conformational variability is tolerated along the C‐helix (Figure 2e–h). Because lineage‐specific differences are most pronounced at the hinge, we focused subsequent experiments on this region. Specifically, we introduced residues from CRP _{E. coli} (T127–L137) into the homologous positions of CRP _MTB (T134–T144) to examine whether sequence‐encoded differences can reprogram allosteric behavior. Swap 1 included all eight substitutions, whereas Swap 2 retained the native N135 and incorporated the remaining seven. The differential functional behavior of the two chimeras motivated further investigation of the single‐point mutant, N135S.

2.2. Hinge substitutions preserve the global fold but alter apo‐state energetics

All three engineered variants remained dimeric in solution, consistent with wildtype CRP _MTB (Figure 3a and S2). Differences in elution volume and peak broadness were observed among the variants, suggesting modest changes in hydrodynamic behavior while retaining the dimeric state (Figure S2; Table S1). The variants also retained similar overall secondary structure content despite differences in circular dichroism (CD) signal intensity, which may suggest altered packing or organization of secondary structure elements, rather than substantial loss of structure (Figure 3b; Table S2). Thus, while the hinge substitutions preserved the dimeric architecture and overall secondary structure of CRP _MTB , they do not inform on the thermodynamic stability of the variants. Chemical denaturation revealed distinct differences in the unfolding behaviors of the variants, as denoted by ∆G°_{(Unf_App)} (Figure 3c; Table 1). Swap 1 remained comparably stable to wildtype, as reflected by its similar apparent stability. Swap 2 was substantially more stable despite sharing most of the same substitutions with Swap 1. In contrast, N135S was destabilized, supporting a disproportionate contribution of position 135 to the hinge energetics. Together, these results show that the mutations preserve the mean structure while measurably reshaping the apparent stability of the apo‐state ensemble.

FIGURE 3.

Structural and biophysical characterization of wildtype and mutant CRP _MTB . (a) Representative size‐exclusion profiles of wildtype (orange), Swap 1 (blue), Swap 2 (green), and N135S (magenta). Independent replicate runs produced comparable elution profiles. All variants elute as single, symmetric peaks, consistent with comparable dimeric assemblies. (b) Circular dichroism (200–260 nm) showing that wildtype and mutants retain similar secondary structure profiles characteristic of alpha‐helical conformations. (c) Chemical denaturation curves monitored by tryptophan fluorescence in the presence of increasing concentrations of guanidine hydrochloride. The normalized fluorescence signal (FS) was fit to a two‐state unfolding model to obtain apparent thermodynamic stabilities summarized in Table 1. The error bars reflect the standard deviation (SD) from at least three experimental repeats.

TABLE 1.

Apparent thermodynamic stability, cAMP binding, and DNA binding affinities of wildtype and mutant CRP _MTB .

	Thermodynamic stability a			cAMP binding affinity and cooperativity b			DNA binding c
	∆G°(Unf_App)	m‐value	c 1/2	k 1	k 2	c	k DNA (apo)	k DNA (cAMP)
WT	2.9 ± 0.2	−1.8 ± 0.1	1.6 ± 0.1	3.5 ± 0.2	2.2 ± 0.1	0.6	2.9 ± 0.5	4.4 ± 0.7
Swap 1	2.8 ± 0.1	−1.9 ± 0.1	1.5 ± 0.1	1.2 ± 1.2	0.6 ± 0.2	0.5	0.9 ± 0.3	7.9 ± 3.3
Swap 2	7.4 ± 1.1	−3.6 ± 0.5	2.1 ± 0.4	0.1 ± <0.1	3.6 ± 1.7	36.0	3.8 ± 1.9	5.9 ± 3.6
N135S	2.3 ± 0.2	−1.6 ± 0.1	1.4 ± 0.2	0.4 ± <0.1	3.6 ± 0.4	9.0	0.9 ± 0.2	1.9 ± 0.4

^a Apparent thermodynamic stability derived from unfolding experiments. The apparent free energy of change of unfolding (∆G°_{(Unf_App)}), denaturant dependence (m‐value), and the midpoint of the transition (c _1/2) are reported for wildtype and mutants, with units kcal·mol⁻¹, kcal·mol⁻¹·M⁻¹, and M, respectively.

^b cAMP binding constants derived from fits to a two‐site binding model. k ₁ and k ₂ represent the affinity constants for the first and second cAMP binding events, respectively. The cooperativity factor is c = k ₂ /k ₁. The units of k ₁ and k ₂ are 10⁴ M⁻¹.

^c DNA binding affinities in the absence (k _DNA(apo)) and presence of cAMP (k _DNA(cAMP)) are reported. The units of k _DNA(apo) and k _DNA(cAMP) are 10⁸ M⁻¹. All values represent the mean ± standard deviation (SD) obtained from nonlinear fitted parameters as described in Materials and Methods.

2.3. Position 135 strongly modulates cAMP binding affinity and cooperativity

Hinge substitutions strongly altered the thermodynamic coupling between the two cAMP‐binding sites (Figure 4; Table 1). Wildtype displayed moderate negative cooperativity (k ₂ /k ₁ = c = 0.6) (Dokas et al., 2025; Gárate et al., ²⁰²¹). Swap 1 preserved that negative cooperativity, though with reduced affinity at both sites, whereas Swap 2 and N135S exhibited positive cooperativity (Figure 4b–d). In the latter two variants, weakened first‐site binding was followed by relatively favored second‐site binding, indicating that hinge perturbations can fully reverse the sign of cooperativity without altering the global fold. The 8‐anilino‐1‐naphthalenesulfonic acid (ANS) fluorescence signal (FS) further showed that the mutants do not share identical cAMP‐bound surface properties (Singh et al., 2019). Wildtype, Swap 1, and N135S exhibited decreased ANS fluorescence upon cAMP binding, whereas Swap 2 showed the opposite trend (Figure 4d). This distinct and reproducible response suggests that cAMP binding significantly redistributes the ANS‐accessible environments differently in this variant, relative to the other constructs.

FIGURE 4.

cAMP binding of wildtype and mutant CRP _MTB . (a–d) cAMP binding to wildtype and mutants was monitored using ANS fluorescence, in which the normalized ANS signal is plotted as a function of cAMP concentration. Binding curves are shown for wildtype (black), Swap 1 (blue), Swap 2 (green), and N135S (maroon). Solid lines reflect fits to a two‐site binding model.

2.4. Hinge rewiring differentially redistributes DNA binding and protein oligomerization

Binding of wildtype CRP _MTB to the fluorescein‐tagged serC promoter displayed two separable phases (Figure 5a; Table 1): a high‐affinity phase attributed to sequence‐specific promoter recognition and a low‐affinity phase associated with protein oligomerization at the promoter site (Dokas et al., 2025; Gárate et al., ²⁰²¹). Wildtype bound DNA tightly in the apo‐state and formed oligomeric assemblies that were strongly reduced by cAMP, with k _DNA(apo) = (2.9 ± 0.5) × 10⁸ M⁻¹ and k _DNA(cAMP) = (4.4 ± 0.7) × 10⁸ M⁻¹ (Figure 5a,b). Stoichiometric binding assays showed at least four apo‐CRP _MTB bound per serC promoter (Figure 5i). Upon cAMP binding, oligomerization was suppressed, as evidenced by the plateaued anisotropy signal beyond the one‐to‐one stoichiometric complex. This behavior in wildtype establishes a reference state against which the mutants are compared.

FIGURE 5.

DNA binding of wildtype and mutant CRP _MTB . (a–h) Non‐stoichiometric binding experiments using fluorescence anisotropy. Normalized anisotropy is plotted as a function of cAMP receptor protein (CRP) for wildtype (orange), Swap 1 (blue), Swap 2 (green), and N135S (magenta). Dashed gray lines indicate the transition region corresponding to free apo‐CRP oligomerization to sequence‐specific promoter sites at high (CRP). Solid lines reflect fits to a single‐site binding model. (i–l) Stoichiometric binding experiments plotted as the molar ratio (CRP/DNA), where the dashed gray lines mark the stoichiometry of one CRP dimer per one DNA molecule. The empty and filled circles correspond to the apo‐ and cAMP‐bound states, respectively. All of the data points represent mean values with error bars indicating standard deviation (SD) from at least three experimental repeats.

The variants separated distinct components of CRP _MTB regulation (Figure 5c–h; Table 1). Swap 1 showed reduced apo‐state DNA affinity, k _DNA(apo) = (0.9 ± 0.3) × 10⁸ M⁻¹, but cAMP substantially enhanced its binding, k _DNA(cAMP) = (7.9 ± 3.3) × 10⁸ M⁻¹, producing a more CRP _{E. coli} ‐like response, while still allowing apo‐state oligomerization (Figure 5c,d,j) (Heyduk & Lee, 1990; Lanfranco et al., ²⁰¹⁷). Swap 2 retained near‐wildtype DNA binding and oligomerization behavior (Figure 5e,f,k). N135S also preserved the qualitative CRP _MTB pattern of cAMP‐independent DNA recognition and cAMP‐suppressed oligomerization, but with weaker DNA affinities with k _DNA(apo) = (0.9 ± 0.2) × 10⁸ M⁻¹ and k _DNA(cAMP) = (1.9 ± 0.4) × 10⁸ M⁻¹ (Figure 5g,h,l). Thus, the contrast between Swap 1 and Swap 2 identifies residue 135 as a critical hinge determinant.

2.5. Fluorescence quenching reveals altered cAMP‐induced conformational redistribution

To probe how hinge substitutions reshape the conformational ensemble upon cAMP binding, we measured tryptophan fluorescence quenching using neutral and charged quenchers which report on the tryptophan solvent accessibility and local surface electrostatics, respectively (Figure 6a–l; Tables 2 and S3) (Maillard et al., 2008). Because the variants retain similar global architectures, these measurements are interpreted as readouts of altered accessibility and local surface charge redistribution surrounding the tryptophan residues, rather than changes in the global fold (Lakowicz, 2010; Maillard et al., ²⁰⁰⁸).

FIGURE 6.

Tryptophan fluorescence quenching using multiple quenchers. Stern‐Volmer analysis of intrinsic Trp fluorescence for wildtype and mutant CRP _MTB in the absence (empty circles) and presence (filled circles) of cAMP. Quenching was measured using three quenchers with distinct physicochemical properties: Acrylamide (a–d), potassium iodide, KI (e–h), and thallium acetate, ThAc (i–l). Data are shown for wildtype (black), Swap 1 (blue), Swap 2 (green), and N135S (maroon). Linear fits to the Stern‐Volmer relationship were used to determine the Stern‐Volmer constants (K _SV) and data points represent mean values with error bars indicating standard deviation (SD) from at least three experimental repeats.

TABLE 2.

Stern‐Volmer constants and relative quenching efficiencies for wildtype and mutant CRP _MTB with and without cAMP.

	Acrylamidea		Potassium iodideb		Thallium acetatec
	K SV	K SV	𝜸 KI/ 𝜸 acryl	𝜸 KI/ 𝜸 acryl	𝜸 ThAc/ 𝜸 acryl	𝜸 ThAc/ 𝜸 acryl
	(apo)	(cAMP)	(apo)	(cAMP)	(apo)	(cAMP)
WT	11.2 ± <0.1	7.1 ± <0.1	0.85 ± 0.02	1.54 ± 0.03	0.60 ± 0.02	1.04 ± 0.04
Swap 1	8.7 ± 0.3	8.6 ± 0.3	1.18 ± 0.06	0.84 ± 0.05	0.80 ± 0.05	1.05 ± 0.06
Swap 2	9.0 ± 0.3	7.9 ± 0.2	0.91 ± 0.05	1.18 ± 0.05	0.93 ± 0.05	1.16 ± 0.06
N135S	8.8 ± 0.3	10.3 ± 0.2	0.62 ± 0.06	0.90 ± 0.04	0.53 ± 0.03	0.97 ± 0.05

Note: Stern‐Volmer constants (K _SV) determined from fluorescence quenching with and without cAMP, using acrylamide^a, potassium iodide^b, and thallium acetate^c. Relative quenching efficiencies (𝛾 _KI/𝛾 _acryl and 𝛾 _ThAc/𝛾 _acryl) are reported to compare the effectiveness of charged quenchers relative to acrylamide. K _SV values represent the mean ± standard deviation (SD) from fitted parameters as described in Materials and Methods. The units for K _SV are M⁻¹.

Using acrylamide, wildtype CRP _MTB showed reduced quenching upon cAMP binding, consistent with reduced tryptophan solvent accessibility in the cAMP‐bound state, with K _SV(apo) = 11.2 ± <0.1 M⁻¹ and K _SV(cAMP) = 7.1 ± <0.1 M⁻¹, respectively (Figure 6a; Table 2). Swap 1 showed no difference in acrylamide quenching, indicating little change in the overall tryptophan accessibility upon cAMP binding (Figure 6b). Swap 2 showed a modest decrease in quenching, consistent with a reduction in tryptophan accessibility, though to a lesser degree than observed for wildtype (Figure 6c). N135S displayed an opposite trend, showing increased acrylamide quenching upon cAMP binding. This suggests greater accessibility and a cAMP‐induced response distinct from the other protein constructs (Figure 6d). These interpretations are based on the absolute acrylamide K _SV constants.

Because the intrinsic tryptophan fluorescence reflects the combined and potentially unequal contributions of all four tryptophan residues in the dimer, these measurements cannot be assigned to specific domains. Differences in both the local environment and quencher accessibility of each tryptophan could contribute differently to the observed quenching responses (Figure S3). Charged quenchers were therefore used to provide complementary results to the cAMP‐dependent changes in the electrostatic environments surrounding the cNBD‐ and DNABD‐associated tryptophans (Figure 6e–l). The comparisons discussed below refer primarily to the acrylamide‐normalized quenching efficiencies, defined as 𝛾 ₁/𝛾 ₂ (Table 2). The ratio 𝛾 ₁/𝛾 ₂ describes the accessibility of the charged quencher relative to acrylamide, which serves as a reference for the overall tryptophan solvent accessibility.

Potassium iodide (KI), a negatively‐charged quencher, preferentially targets tryptophan residues surrounded by positively‐charged side chains, as seen for the residues W203 in the DNABDs (Figure S3). KI showed the strongest cAMP‐induced increase in wildtype CRP _MTB , suggesting a substantial cAMP‐induced redistribution in the electrostatic environments surrounding the DNABD‐associated tryptophan residues, with 𝛾 _KI/𝛾 _acryl = 0.85 ± 0.02 and 𝛾 _KI/𝛾 _acryl = 1.54 ± 0.03 in the apo‐ and cAMP‐bound states, respectively (Figure 6e; Table 2). In contrast, Swap 1 showed a modest decrease in relative KI quenching efficiency, whereas Swap 2 and N135S increased modestly upon cAMP binding, preserving the direction but not the magnitude of the wildtype response (Figure 6f–h).

Thallium acetate (ThAc), a positively‐charged quencher, favors tryptophan residues surrounded by negatively‐charged side chains, as seen for the residues W112 in the cNBDs (Figure S3). Wildtype displayed increased ThAc sensitivity upon cAMP binding, with 𝛾 _ThAc/𝛾 _acryl = 0.60 ± 0.02 and 𝛾 _ThAc/𝛾 _acryl = 1.04 ± 0.04 in the apo‐ and cAMP‐bound states, respectively (Figure 6i; Table 2). Similarly, Swap 1 and Swap 2 also increased in ThAc efficiency upon cAMP binding, but to a lesser degree (Figure 6j,k). In contrast, N135S showed a larger relative increase in ThAc quenching efficiency in the cAMP‐bound state, comparable in magnitude to the wildtype response (Figure 6l; Table S3).

Taken together, these measurements show that hinge substitutions alter tryptophan accessibility and the local surface electrostatics at distantly regions of the protein, resulting in distinct cAMP‐induced quenching responses across the CRP _MTB variants. These differences provide a possible mechanistic link between the apo‐state energetic differences and the divergent functional behaviors described previously.

3. DISCUSSION

This study examines whether residue‐level variation within the dimerization C‐helix hinge can explain how homologous CRPs achieve divergent allosteric behaviors while preserving a conserved global fold. Using CRP _{E. coli} and CRP _MTB as a comparative framework, we show that sequence variation across the C‐helices reveals lineage‐specific patterns of conformational constraint across the C‐helices, suggesting distinct evolutionary strategies to preserve function (Figure 2). Targeted substitutions in CRP _MTB further show that perturbations within the hinge are sufficient to reshape cAMP binding thermodynamics, DNA binding, and protein oligomerization (Figures 4 and 5; Table 1). Across these analyses, N135 emerges as a key determinant of hinge dynamics and allosteric output.

3.1. Evolution constrains where conformational variability is tolerated

Both proteobacterial and actinobacterial CRPs show a correlation between sequence conservation and reduced variability in helical propensity. However, the location of this constraint differs between lineages (Figure 2). In proteobacteria, the hinge region is comparatively restricted, consistent with preserving the coil‐to‐helix conformational transition that couples cAMP binding to productive DNA binding. In actinobacteria, helical variability extends more broadly across the C‐helix, suggesting a redistribution of structural conservation and local flexibility. Thus, the key evolutionary difference is not merely sequence divergence, but possibly the spatial allocation of conformational heterogeneity within the same structural element.

3.2. N135 operates as a context‐dependent hinge regulatory node

Mutational analyses show that the hinge can be substantially perturbed without disrupting the global CRP architecture (Figure 3a,b). All variants remain dimeric and retain secondary structure, yet exhibit measurable differences in apo‐state stability (Figure 3c; Table 1). N135S is the least stable, whereas Swap 2 is the most stable, highlighting position 135 as a key site for localized energetic control. This effect is highlighted from the comparison of Swap 1 and Swap 2, which differ only at residue 135. Retention of the native residue N135 in Swap 2 is reflected by a substantial increase in apparent stability relative to Swap 1, which contains S135. A possible structural basis for this difference can be observed in the apo‐CRP _MTB structure, in which N135 of subunit B forms a cross‐subunit hydrogen bond with T134 of subunit A (Figure S1) (Stapleton et al., 2010). Swap 2, which retains N135, may preserve this interaction that potentially contributes to the stabilization of the dimer interface. Notably, the effect of substitution at this position is also sequence context‐dependent. When S135 is introduced within a hinge that more closely resembles CRP _{E. coli} , as in Swap 1, the destabilizing effect is mitigated. Conversely, retention of the native CRP _MTB hinge environment amplifies this effect, as observed for N135S. The separation between preserved fold and altered energetics is central to our proposed mechanism, in which hinge substitutions reshape how the protein samples conformational states through its local interaction network, rather than its ability to adopt a folded CRP‐like dimer.

3.3. Energetic tuning redirects cAMP‐induced ensemble redistribution

Substitutions within the hinge alter both cAMP binding affinity and cooperativity, showing that local sequence changes can alter the energetic coupling between the two cAMP binding pockets (Figure 4a–d; Table 1). Wildtype CRP _MTB and Swap 1 exhibit negative cooperativity, whereas Swap 2 and N135S display positive cooperativity. The key contrast between Swap 1 and Swap 2, which differ at residue 135, identifies N135 as a critical contributor to hinge‐mediated energetic tuning whose effects depend on the surrounding sequence.

Fluorescence quenching further suggests that cAMP does not drive all variants toward a common end‐state with different amplitudes (Figure 6; Table 2). Instead, cAMP binding may redistribute each variant's conformational ensemble differently, characterized by differences in solvent accessibility and local surface electrostatics sampled by the native tryptophan residues. Although these measurements do not resolve a unique structural state, they support altered conformational sampling as a plausible mechanistic link between hinge sequence variation and allosteric behavior.

3.4. cAMP shifts CRP _MTB toward a constrained DNA‐bound ensemble, disfavoring protein oligomerization

Wildtype CRP _MTB binds DNA with high affinity in the apo‐state, k _DNA(apo) = (2.9 ± 0.5) × 10⁸ M⁻¹, and forms higher‐order oligomers at specific promoter sites (Figure 5a). cAMP binding allosterically dissociates these oligomers and marginally enhances DNA affinity, with k _DNA(cAMP) = (4.4 ± 0.7) × 10⁸ M⁻¹ (Figure 5b; Table 1). This behavior suggests a model in which cAMP binding shifts the protein toward a more conformationally constrained ensemble that preserves site‐specific DNA contacts. Simultaneously, cAMP‐induced redistribution of surface electrostatics may alter protein–protein interactions associated with the oligomeric state (Figure 6e,i). Taken together, these findings suggest an allosteric mechanism where cAMP binding suppresses oligomerization through coordinated structural rigidification and surface electrostatics reorganization without compromising DNA recognition.

3.5. Hinge‐mediated energetic tuning linked conserved structure to divergent regulatory output

Functional assays show that these energetic differences propagate unevenly to downstream outputs. Although DNA binding and protein oligomerization are coupled in wildtype, the mutants reveal that these outputs can be redistributed independently. Swap 1 may shift the ensemble toward a CRP _{E. coli} ‐like DNA binding response, while still retaining apo‐state oligomerization (Figure 5c,d). This indicates that cAMP‐dependent amplification of DNA recognition can be redirected from oligomer suppression. In contrast, Swap 2 largely preserves wildtype CRP _MTB function despite sharing seven of eight substitutions with Swap 1, highlighting N135 as a key determinant of how hinge energetics are propagated (Figure 5e,f). Consistent with this interpretation, N135S retains the qualitative CRP _MTB behavior, though with weakened DNA interactions (Figure 5g,h).

Taken together, we propose a model in which the C‐helix hinge may govern both the magnitude and direction of allosteric response, thereby affecting DNA recognition and protein oligomerization. Under this model, cAMP binding does not appear to converge the variants into a common end‐state. Rather, wildtype CRP _MTB and the C‐helix hinge variants are proposed to sample an ensemble of conserved folded conformations, with differences in hinge sequence shifting the relative populations and energetic coupling among the substates associated with cAMP binding, DNA recognition, and oligomerization (Figure 7). More broadly, CRP _{E. coli} and CRP _MTB illustrate a general principle in the evolution of homologous transcriptional regulators, where distinct signaling behaviors can emerge through redistribution of conformational ensembles without changes to the global fold (Nussinov et al., 2023; Onuchic et al., ¹⁹⁹⁷). Within this framework, the dimerization C‐helix hinge functions as an evolvable hotspot, in which sequence changes reshape the energetic landscape (Mallamace et al., 2016; Onuchic et al., ¹⁹⁹⁷), alter coupling between cAMP binding and domain reorganization, and redirect functional output toward distinct regulatory strategies (Fieweger et al., 2019; Wilburn et al., ²⁰²²). The engineered CRP _MTB variants support this proposed model by showing that localized sequence changes are sufficient to diversify regulatory behavior without altering the conserved CRP fold.

FIGURE 7.

Hinge substitutions reshape the conformational energy landscape and redirect allosteric coupling between cAMP binding, DNA recognition, and oligomerization. (a) Apo‐state energy landscapes, illustrating conformational ensembles sampled by wildtype and mutant CRP _MTB . (b) Apo‐state ensemble favors oligomerization. The landscape is distributed across multiple basins corresponding to DNA binding capabilities and oligomeric‐prone conformational states. WT and Swap 2 occupy a relatively stabilized minimum, and bind DNA with high affinity. Swap 1 occupies a broadened basin, consistent with greater conformational heterogeneity. N135S occupies a distinct basin that reflects a less stable conformational ensemble. (c) cAMP binding reshapes the energy landscapes distinctly for each mutant. (d) cAMP shifts the ensemble toward conformational states with reduced oligomers. In wildtype, cAMP induces minimal redistribution of the ensemble. Swap 1 undergoes a larger redistribution that favors DNA binding‐competent states and partially reduces oligomerization. Swap 2 shows limited redistribution, consistent with a more constrained conformational ensemble. In N135S, cAMP redistributes and stabilizes the ensemble but does not enhance DNA binding. The basin positions, widths, and depths within these landscapes are qualitative and do not represent experimentally determined free‐energy profiles or state populations. Created in BioRender. Maillard, R. (2026) https://BioRender.com/sllf7j5

4. CONCLUDING REMARKS

This study supports a model in which CRP homologs diversify their regulatory mechanisms through evolutionary tuning of the dimerization C‐helix hinge. By reshaping where conformational variability is permitted within a conserved scaffold, sequence differences alter the energetic coupling between cAMP binding, DNA recognition, and protein oligomerization, enabling homologous transcription factors to evolve distinct regulatory behaviors without structural divergence. Direct characterization of the wildtype CRP _{E. coli} and reciprocal introduction of the CRP _MTB hinge into the CRP _{E. coli} scaffold would further test the generality of this model across the CRP homologs. Although these comparative experiments extend beyond the scope of the present study, they provide important future directions to further investigate the contributions of lineage‐specific hinge sequences in tuning allosteric regulation, while defining how these effects are shaped by the broader sequence context of each CRP scaffold.

5. MATERIALS AND METHODS

5.1. CRP _MTB and CRP _{E. coli} structures

Structural analyses were performed using the PyMol Molecular Graphics System (Version 2.0, Schrödinger, LLC). Structural superpositions and RMSD values were calculated using PyMol's alignment feature. Alignment of the cAMP‐bound domain (residues 21–104 in CRP _{E. coli} and residues 28–110 in CRP _MTB ) served as the reference for all RMSD calculations. The following Protein Data Bank (PDB) entries were used: CRP _MTB , 3d0s (apo‐state) and 3i54 (cAMP‐bound); CRP _{E. coli} , 2wc2 (apo‐state) and 1g6n (cAMP‐bound).

5.2. Helical propensity analysis

Helical propensities for 100 proteobacterial and actinobacterial CRPs were calculated using residue‐level α‐helical propensity parameters, as described previously (Deléage & Roux, 1987). For each sequence, residue‐level propensities were mapped along the primary sequence to generate position‐dependent profiles across the dimerization C‐helices, including the hinge region. Local helical propensity was evaluated within a sliding window of three residues to estimate cumulative propensity across neighboring positions. Because this analysis solely relies on sequence‐derived statistical parameters, the resulting profiles reflect intrinsic helical propensity of the backbone rather than tertiary quaternary structural constraints.

5.3. Sequence conservation and statistical analysis

Sequence conservation across aligned C‐helix positions was determined from multiple‐sequence alignments of proteobacterial and actinobacterial CRPs generated using the NCBI Constraint‐based Multiple Alignment Tool (COBALT). Shannon entropy was calculated at each aligned position and normalized to the theoretical maximum entropy for 20 amino acids. Normalized entropy values were converted into conservation scores, where higher scores indicate greater conservation. For each aligned position, the average helical propensity and SD were calculated and compared with conservation scores to evaluate how residue conservation related to mean intrinsic helicity and positional variability. Linear regression analysis was performed using Wolfram Mathematica to assess these relationships, and R ² values were used to summarize the correlation strengths.

5.4. Cloning, expression, and purification of CRP _MTB wildtype and mutants

The wildtype DNA sequence of CRP _MTB , flanked by NdeI and BamHI sites, was amplified by PCR using PfuUltra polymerase, digested with NdeI and BamHI, and cloned into a C‐terminal histidine‐tagged (His₆‐tag) pET‐3a expression vector (Addgene). The His₆‐tag was retained throughout purification and in all subsequent binding experiments for both wildtype and mutant CRP _MTB . Previous characterization of CRP _MTB comparing His₆‐tagged and tag‐cleaved protein found no measurable differences in cAMP and DNA binding behavior, supporting the use of the retained tag in the present study (Stapleton et al., 2010). The CRP _MTB mutant constructs (Swap 1, Swap 2, and N135S) were obtained from GenScript, and expressed in E. coli T7 Express pLysY cells (NEB). Cells were grown overnight on ampicillin‐resistant agar, inoculated into ampicillin‐supplemented LB media, and induced with 1 mM IPTG for 2 h at 37°C.

Cell pellets were resuspended at 10 mL/g of wet weight in lysis buffer (20 mM Tris, 200 mM NaCl, 1 mM EDTA) containing protease inhibitors (10 mM benzamidine, 0.4 mM AEBSF, 1 μM pepstatin, 1 μM leupeptin, 28 μM TPCK/TLCK, 10 μM IBMX, 1 mM PMSF), lysed by microfluidization at 10,000 psi. The resulting lysate was centrifuged (JA 25.5 rotor, 45 min, 15,000 rpm, 4°C), and the supernatant was incubated overnight at 4°C with His60 Ni Superflow Resin (Takara Bio) in the presence of 30 mM imidazole. The resin‐supernatant mixture was washed thrice with 40 mL of lysis buffer supplemented with 30 mM imidazole and eluted with lysis buffer containing 500 mM imidazole. CRP _MTB elutions were further purified by size‐exclusion chromatography (SEC) and stored at −80°C in a storage buffer (50 mM HEPES, 150 mM KCl, 1 mM EDTA, pH 7.2). The SEC profiles shown are representative chromatograms, with independent replicates producing comparable elution volumes and peak behaviors. The SEC column was calibrated using five molecular weight protein standards, and the resulting curve was used to confirm the molecular masses of the wildtype and mutant CRP _MTB proteins (Figure S2). The wildtype and mutant proteins eluted at ~89 mL, consistent with a dimeric assembly. Apo‐state protein concentrations were determined at 280 nm (ε = 24,980 cm⁻¹ M⁻¹), with “apo‐state” referring to purified protein without the addition of exogenous cAMP. Final measured concentrations showed A ₂₆₀/A ₂₈₀ ratios of ~0.5, consistent with minimal nucleic acid contamination. The first variant, Swap 1, included eight mutations: N135S, N136E, N137K, L138V, A139G, D140N, I142A, and T144L. Swap 2 retained the original N135 residue and incorporated the seven remaining mutations present in Swap 1.

5.5. CD and fluorescence measurements

CD spectra were collected from 195 nm to 260 nm to monitor the secondary structure of the CRP _MTB wildtype and mutants (Aviv Model 202‐01 spectrometer) using 12 μM of protein. Secondary structure content was estimated using the BeStSel web server (Micsonai et al., 2015; Micsonai et al., ²⁰¹⁸). Spectra were analyzed over the 200–260 nm wavelength range using a scale factor of 1, and the resulting secondary structure estimates were used for comparative analysis among the CRP _MTB constructs (Table S2). Intrinsic tryptophan fluorescence was measured at λ _ex = 285 nm and λ _em = 335 nm (PTI QM40 fluorimeter) using 5 μM of protein.

5.6. Chemical denaturation using guanidine hydrochloride (GdnHCl)

Protein unfolding was monitored by intrinsic tryptophan fluorescence at λ _ex = 285 nm and λ _em = 335 nm (PTI QM40 fluorimeter) using 5 μM of protein in storage buffer (50 mM HEPES, 150 mM KCl, 1 mM EDTA, pH 7.6). Three independent titrations were performed for each construct, and signals were corrected for buffer contributions. The data was analyzed using the linear extrapolation method and fitted to a two‐state unfolding model (Santoro & Bolen, 1988)

$$S_T=S_F{f}_F+S_U{S}_U$$ (1)

where S _T reflects the total observed signal, S _F and S _U represent the folded and unfolded state signals, respectively, and f _F and f _U are the fractional populations of the folded and unfolded protein, respectively. f _F and f _U are associated with the equilibrium constant of the folded and unfolded states as follows:

$$\begin{array}{c}{f}_F=\frac{1}{1+K}\end{array}$$ (2)

$$\begin{array}{c}{f}_U=\frac{K}{1+K}\end{array}$$ (3)

in which,

$$\begin{array}{c}K=e^{\frac{-∆G^{°}}{\mathrm{RT}}}\end{array}$$ (4)

and,

$$\begin{array}{c}∆G^{°}=∆G{°}_{{\mathrm{H}}_2\mathrm{O}}+m\left[d\right]\end{array}$$ (5)

$∆G{°}_{{\mathrm{H}}_2\mathrm{O}}$ is the apparent free energy of unfolding without a denaturant, m is the slope of the linear dependence of the apparent ∆G° on the denaturant concentration as illustrated by the linear extrapolation method (Santoro & Bolen, 1988), and d is the denaturant concentration. Combining Equations (1), (2), (3), (4), (5) yields the fitting equation,

$$\begin{array}{c}{S}_T=\frac{S_F+S_U{e}^{-\left(\frac{∆G{°}_{{\mathrm{H}}_2\mathrm{O}}+m\left[d\right]}{\mathrm{RT}}\right)}}{1+e^{-\left(\frac{∆G{°}_{{\mathrm{H}}_2\mathrm{O}}+m\left[d\right]}{\mathrm{RT}}\right)}}\end{array}$$ (6)

The denaturation data were analyzed using a two‐state model because the unfolding curves were dominated by a single transition. A three‐state model (F ⇄ I ⇄ U) was also tested, but the parameters associated with the intermediate state were poorly resolved and could not be reliably estimated from the data, particularly for Swap 2. Therefore, the two‐state model was retained as the simplest description of the data and is not intended to imply a physically two‐state unfolding pathway. The resulting stabilities are reported as apparent ∆G° values to allow for comparison among wildtype and the C‐helix hinge variants.

5.7. cAMP binding by ANS fluorescence

The fluorescence properties of 8‐anilino‐1‐napthalenesulfonic acid (ANS) (λ _ex = 350 nm and λ _em = 480 nm) were used to monitor cAMP binding to CRP _MTB (Dokas et al., 2025; Gárate et al., ²⁰²¹; Lanfranco et al., ²⁰¹⁷). Samples contained 47.7 μM ANS and 3.6 μM protein in ANS buffer (50 mM Tris, 50 mM KCl, 1 mM EDTA, pH 7.8) in a final volume of 800 μL. Measurements were collected on a PTI QM40 fluorimeter, and at least three independent titrations were performed. The data was fitted to three binding models: a cooperative two‐site (Lanfranco et al., 2017), an independent two‐site, or a single‐site binding model. The fitting equation for the cooperative two‐site binding model is (Dokas et al., 2025; Gárate et al., ²⁰²¹; Lanfranco et al., ²⁰¹⁷)

$$\begin{array}{c}{F}_{480\mathrm{nm}}=\frac{F_0+2F_1{k}_1\left[x\right]+F_2{k}_1{k}_2{\left[x\right]}^2}{1+2k_1\left[x\right]+k_1{k}_2{\left[x\right]}^2}\end{array}$$ (7)

where F _{480 nm} is the observed fluorescent signal, F ₀, F ₁, and F ₂ reflect the fluorescent signals of the apo‐, singly‐, and doubly‐bound states of the protein, respectively. k ₁ and k ₂ are the microscopic binding affinity constants for the first and second cAMP molecule, respectively, while x is the cAMP concentration. The cooperativity factor is c = k ₂ /k ₁, where values <1 indicate negative cooperativity and values >1 indicate positive cooperativity. In the independent two‐site binding model, k ₁ = k ₂ . Therefore, the cAMP binding sites are not allosterically linked and there is no cooperativity. Control measurements were used to correct the collected data for non‐binding effects. A no‐titrant control monitored the loss of ANS fluorescence caused by photobleaching. A buffer‐only titration replicated the volume changes associated with cAMP addition to correct for sample dilution. Finally, cAMP was also titrated into ANS without CRP _MTB to monitor changes that may arise from ANS–cAMP interactions.

5.8. DNA binding by fluorescence anisotropy

DNA binding to the specific sequence target site was monitored using a 32‐bp (λ _ex = 480 nm and λ _em = 518 nm) serC fluorescein‐tagged DNA promoter (5′–GCGCGTAGTGTGAACAAGCTCACATGCAAGCC–3′) (Dokas et al., 2025; Gárate et al., ²⁰²¹). Samples contained 3 nM DNA with either 0 mM or 1 mM of cAMP in DNA binding buffer (75 mM KCl, 50 mM HEPES, 1 mM EDTA, pH 7.6) in a final volume of 2 mL. Measurements were collected on a PTI QM40 fluorimeter, and at least three independent titrations were conducted. The data was baseline‐corrected, and the raw anisotropy at each protein concentration was normalized to the first experimental value measured in the absence of the protein:

$$\begin{array}{c}{r}_{\mathrm{norm}}=\frac{r_i}{r_0}\end{array}$$ (8)

where r _i is the raw anisotropy at each CRP concentration and r ₀ is the anisotropy measured in the absence of protein. The data was then fitted to the equation below to determine the DNA binding affinity (Dokas et al., 2025; Gárate et al., ²⁰²¹; Heyduk & Lee, ¹⁹⁹⁰; Lanfranco et al., ²⁰¹⁷)

$$A_{\left(\mathit{obs}\right)}=A_{{\left(\mathit{DNA}\right)}_F}+\left(A_{\left(P-\mathit{DNA}\right)}-A_{{\left(\mathit{DNA}\right)}_F}\right)\left(\frac{K\left[\mathit{DN}{A}_T\right]+K\left[P_T\right]+1-\sqrt{\left({\left(K\left[\mathit{DN}{A}_T\right]+K\left[P_T\right]+1\right)}^2+4K^2\left[\mathit{DN}{A}_T\right]\left[P_T\right]\right)}}{2K\left[\mathit{DN}{A}_T\right]}\right)$$ (9)

where $A_{\left(\mathit{obs}\right)}$ is the observed anisotropy signal, $A_{{\left(\mathit{DNA}\right)}_F}$ and $A_{\left(P-\mathit{DNA}\right)}$ are the anisotropy signals for the unbound DNA and protein–DNA complex, respectively. $\left[\mathit{DN}{A}_T\right]$ is the total DNA concentration, $\left[P_T\right]$ is the total protein concentration, and $K$ reflects the association constant between the protein and DNA promoter sequences, with and without cAMP. The above experimental protocol was repeated for stoichiometric binding titrations using 200 nM of DNA.

5.9. Tryptophan fluorescence quenching to monitor conformational dynamics

Conformational changes of wildtype and mutant CRP _MTB , with and without cAMP, were examined by tryptophan fluorescence quenching (λ _ex = 295 nm and λ _em = 335 nm) using acrylamide, KI, and ThAc (Maillard et al., 2008). Samples contained 2 μM protein, with either 0 mM or 1 mM of cAMP in DNA binding buffer (75 mM KCl, 50 mM HEPES, 1 mM EDTA, pH 7.6) to a final volume of 2 mL. For ThAc measurements, 75 mM NaOAc replaced KCl due to chloride interference with thallium. Quenchers were added to a final concentration of 30 mM. Data was collected on a PTI QM40 fluorimeter, and at least three independent titrations were conducted. The data was baseline‐corrected, normalized to the first experimental value, and analyzed using the Stern‐Volmer equation (Stern & Volmer, 1919; Lakowicz, ²⁰¹⁰)

$$\begin{array}{c}\frac{I_0}{I}=1+K_{\left(\mathrm{SV}\right)}\left[Q\right]\end{array}$$ (10)

where I ₀ and I are the fluorescence intensities without and with a quencher, respectively. K _SV is the Stern‐Volmer constant, reflecting the quencher accessibility for a given quencher, and [Q] is the quencher concentration.

In addition, control experiments were conducted to probe whether the quenchers themselves induce significant structural perturbation under the experimental conditions used. The intrinsic tryptophan emission spectra were collected for apo‐state CRP _MTB in the absence and presence of 30 mM acrylamide and 30 mM KI, the highest quencher concentrations used. Apo‐state CRP _MTB in 0 M and 4 M GdnHCl were included as the folded and unfolded controls, respectively. The spectra were examined for changes in peak intensity and peak shift (Figure S4). Acrylamide and KI affected the fluorescence intensities, but did not produce a detectable shift in the emission wavelength associated with protein unfolding. Therefore, neither quencher induces a global unfolding under these experimental conditions.

5.10. Tryptophan fluorescence quenching efficiency by iodide and thallium

Absolute K _SV values of acrylamide were used to evaluate changes in the combined solvent accessibility of the tryptophan residues because acrylamide is a neutral quencher. In contrast, quenching by KI and ThAc is influenced by the local electrostatic environment surrounding the tryptophan residues. Their effects were therefore evaluated using relative quenching efficiencies. Relative quenching efficiencies of KI and ThAc were determined by normalizing charged‐quencher K _SV constants to acrylamide (Berman & Taylor, 1978; Maillard et al., ²⁰⁰⁸)

$$\begin{array}{c}\frac{{\gamma }_1}{{\gamma }_2}=\frac{{K_{\left(\mathrm{SV}\right)}}_1^b}{{K_{\left(\mathrm{SV}\right)}}_1^f}\frac{{K_{\left(\mathrm{SV}\right)}}_2^f}{{K_{\left(\mathrm{SV}\right)}}_2^b}\end{array}$$ (11)

where subscripts 1 and 2 define the charged quenchers (Q ₁ = iodide or thallium) and the reference quencher (Q ₂ = acrylamide), respectively. Superscripts b and f refer to the fluorophore in the bound (i.e., the native state of CRP _MTB ) and free (free N‐AcTrpNH₂ in solution) states, respectively. The K _SV values of free N‐AcTrpNH₂ used for acrylamide, iodide, and thallium were 17.0 ± 0.1 M⁻¹, 10.6 ± 0.1 M⁻¹, and 20.7 ± 0.7 M⁻¹, respectively (Maillard et al., 2008). Thus, 𝛾 ₁/𝛾 ₂ represents the relative accessibility of the charged quencher normalized to acrylamide after accounting for differences in their quenching properties. Uncertainties in relative quenching efficiencies were propagated from the fitted K _SV values of CRP _MTB (apo‐ and cAMP‐bound states) and free N‐AcTrpNH₂, using standard error propagation:

$${\sigma }_{\gamma }=\left(\frac{{\gamma }_1}{{\gamma }_2}\right)\sqrt{{\left(\frac{{\sigma }_{{K_{\left(\mathrm{SV}\right)}}_1^b}}{{K_{\left(\mathrm{SV}\right)}}_1^b}\right)}^2+{\left(\frac{{\sigma }_{{K_{\left(\mathrm{SV}\right)}}_1^f}}{{K_{\left(\mathrm{SV}\right)}}_1^f}\right)}^2+{\left(\frac{{\sigma }_{{K_{\left(\mathrm{SV}\right)}}_2^f}}{{K_{\left(\mathrm{SV}\right)}}_2^f}\right)}^2+{\left(\frac{{\sigma }_{{K_{\left(\mathrm{SV}\right)}}_2^b}}{{K_{\left(\mathrm{SV}\right)}}_2^b}\right)}^2}$$ (12)

AUTHOR CONTRIBUTIONS

Sanuja Mohanaraj: Conceptualization; investigation; writing – original draft; methodology; visualization; writing – review and editing; formal analysis; data curation. Miray Samuel: Methodology; investigation. Eva Siminiceanu: Methodology; investigation. Nazgol Missaghi: Methodology; investigation. Rodrigo A. Maillard: Conceptualization; investigation; funding acquisition; validation; writing – review and editing; formal analysis; data curation; supervision; resources; methodology; visualization.

CONFLICT OF INTEREST STATEMENT

The authors declare no conflicts of interest.

Supporting information

Figure S1. Comparison of C‐helix hinge residues S128 and N135 in apo‐CRP _{E. coli} and apo‐CRP _MTB .

Figure S2. Size‐exclusion chromatography calibration of the column.

Figure S3. Native tryptophan residues occupy distinct electrostatic environments in CRP _MTB .

Figure S4. Intrinsic tryptophan fluorescence spectra of apo‐CRP _MTB in the presence of quenchers.

Table S1. Size‐exclusion calibration standards and estimated molecular masses of wildtype and mutant CRP _MTB .

Table S2. BeStSel‐derived secondary‐structure estimates for wildtype and C‐helix hinge variants of CRP _MTB .

Table S3. Stern–Volmer constants for wildtype and mutant CRP _MTB with and without cAMP.

Mohanaraj S, Samuel M, Siminiceanu E, Missaghi N, Maillard RA. Sequence‐encoded dynamics diversify allostery within a conserved protein scaffold. Protein Science. 2026;35(11):e70821. 10.1002/pro.70821

DATA AVAILABILITY STATEMENT

The underlying data will be deposited in a publicly accessible repository and made available upon publication. Repository information and accession details will be provided once available.