Skip to main content
NCBI home page
ACS AuthorChoice logoLink to ACS AuthorChoice
. 2025 Feb 5;147(7):5714–5724. doi: 10.1021/jacs.4c13462

Proline cis/trans Conformational Selection Controls 14–3–3 Binding

Frederik F Theisen 1,2, Andreas Prestel 1, Nina L Jacobsen 1, Oline K Nyhegn-Eriksen 1, Johan G Olsen 1, Birthe B Kragelund 1,*
PMCID: PMC11848828  PMID: 39909402

Abstract

graphic file with name ja4c13462_0007.jpg

Intrinsically disordered protein regions (IDRs) are structurally dynamic yet functional, often interacting with other proteins through short linear motifs (SLiMs). Proline residues in IDRs introduce conformational heterogeneity on a uniquely slow time scale arising from cis/trans isomerization of the Xaa-Pro peptide bond. Here, we explore the role of proline isomerization in the interaction between the prolactin receptor (PRLR) and 14–3–3. Using NMR spectroscopy, thermodynamic profiling, and molecular dynamics (MD) simulations, we uncover a unique proline isomer-dependent binding, with a cis conformation affinity 3 orders of magnitude higher than the trans. MD simulations identify structural constraints in the narrow 14–3–3 binding groove that provide an explanation for the observed isomer selectivity. The cis preference of WT PRLR introduces a slow kinetic component relevant to signal propagation and a steric component that impacts chain direction. Proline isomerization constitutes a previously unrecognized selective component relevant to the ubiquitous 14–3–3 interactome. Given the prevalence of prolines in IDRs and SLiMs, our study highlights the importance of considering the distinct properties of proline isomers in experimental design and data interpretation to fully comprehend IDR functionality.

Introduction

Intrinsically disordered proteins (IDPs) are unique in the proteomic landscape, characterized by their lack of stable tertiary structure.1 Structural flexibility allows IDPs to engage in a broad spectrum of protein–protein interactions. Central to many interactions involving IDPs are short linear motifs (SLiMs), typically defined as conserved sequence stretches of fewer than ten residues.2,3 SLiMs constitute critical functional units often incorporating hydrophobic and aromatic residues essential for protein interaction.2

A distinctive feature of SLiMs is the frequent inclusion of proline residues.4 Proline has traditionally been viewed as a rigidifying element due to its unique cyclic structure.5 Paradoxically, the unique backbone chemistry of prolines allows highly populated cis peptidyl-prolyl peptide bond conformations. Thus, proline may introduce specific structural heterogeneity, with the trans conformation resulting in an extended structure while the cis conformation produces a tight turn.6 In structured proteins the peptide bond conformation is often dictated by the requirements of the protein fold and thus in most cases, only one isomer will typically be observed. For high-resolution crystal structures, the frequency of cis proline peptide bonds is around 6%.7 For IDPs, no such structural constraints exist, and thus the proline peptide bond populates an equilibrium between cis and trans, governed not by the structural context but rather by the local sequence context.8,9 The equilibrium between the cis and trans isomers strongly depends on the residue preceding the proline, with aromatic residues enabling up to 40% cis occupancy,8 although large cis populations can also be obtained without preceding aromatics.10,11 Despite the small free energy difference between the two proline isomers, their interconversion is slow, with exchange rate constants typically on the order of 10–3 s–1 at 4 °C8. For many experimental methods, the minute time scale means that the cis and trans isomers effectively exist as two distinct molecules with potentially different structural and thermodynamic properties, affecting protein function and interactions.10,12−17 As differences originating from cis and trans conformers can introduce lengthy equilibration times, challenges in both data analysis and modeling of biological systems arise. Nuclear magnetic resonance (NMR) spectroscopy is often a key method used to characterize proline isomer specific properties due to the cis and trans signals of prolines being directly accessible.18−20 Understanding the nuances of proline isomerization is therefore crucial, not only for accurate experimental interpretation but also for the broader comprehension of IDP functionality where the slow component of conformational exchange may impact the ensemble of states and hence IDP specificity and selectivity.16

In this work we investigate how proline isomerization in disorder-based protein interactions imposes conformational selection in binding with a kinetic component relevant for biological responses. Using the 14–3–3 interaction with the human prolactin receptor (PRLR) intracellular domain as a model, we study the unique thermodynamic, kinetic, and structural properties that govern proline isomer-dependent interactions. Biophysical studies of a 21-residue phosphorylated PRLR fragment using isothermal titration calorimetry (ITC) and NMR spectroscopy show an interaction with a kinetic phase on the minute time scale. In addition, real-time NMR experiments reveal that 14–3–3 binds PRLR with a unique preference for the cis proline conformation. Our findings show that SLiM-based interactions can be proline isomer specific, and that this selectivity introduces a significant kinetic barrier with the isomer selectivity being tunable by altering the motif sequence. The results have impact on the selectivity and timing in biological systems and are thus relevant to drug targeting and development against complexes involving IDPs, particularly those involving 14–3–3 proteins.

Results and Discussion

Structural Analysis Reveals High cis Proline Frequency in 14–3–3 Binding Peptides

While proline residues in IDRs can sample both the cis and trans conformations in their free state, no studies have investigated the frequency of cis proline in ID-based interactions. To examine the prevalence of cis proline in ID-based interactions we searched the protein data bank (PDB) for crystal structures of protein complexes where one partner was shorter than 30 residues and the resolution better than 2.0 Å. This data set contained 9181 entries, which were analyzed to determine if the peptide component contained any prolines and whether they were in either cis or trans conformation. The analysis revealed 3528 structures where the peptide component contained a proline residue, of which 193 structures (5.5%) had a proline in the cis conformation. Duplicate entries were removed based on sequence alignments of proteins and peptides, resulting in 124 (6.4%) unique cis and 1804 unique trans proline entries (Figure 1A).

Figure 1.

Figure 1

Open in a new tab

Structural and bioinformatic investigation of proline peptide bond conformations in protein:peptide complexes. (A) Number of unique peptides in high-resolution structures containing a proline in cis (yellow) or trans (red) conformation. Frequency of specific structure keywords and terms within each isomer group and the number of unique 14–3–3 complexes containing a proline in the +2 position relative to the phosphoresidue. (B) Xaa-Pro peptide bond angles in 14–3–3:peptide complexes, with the number of unique peptides and unique structures indicated along with the expected free state peptide cis proline population.8 *Structure contains a large ligand molecule. **Electron density suggests binding in both cis and trans proline conformations. (C) Superimposed 14–3–3 complex structures. The cutouts show the phosphoresidue, the peptide C-terminus and the +2 proline residue side chain position. 14–3–3 is shown as a monomer for simplicity. Cis proline PDB IDs: 1QJB, 1YWT, 6BCR, 6NV2, 6YLU, 6Y3R, 6ZVD, 6ZVE, 5EXA, 5EWZ, 6BCY, 8C43, 8A62, 6QZR, 4QLI, 6YO8, 8A9G, 5XY9. (D) Overview of the human PRLR structure and 14–3–3 binding motif. Schematic representation (top) and ensemble model (based on homology model by Pogozheva et al.21) of hormone (prolactin, PRL) bound PRLR dimer with bound JAK2 kinase (left). The phosphorylated threonine is indicated using dark gray spheres. Selection of aligned PRLR homologue sequences and motif logo (right). Human PRLR residues are colored blue in the sequence logo. The human FOXO-1 motif sequence is shown for comparison. *Phosphorylated threonine residue.

Examining the annotated functions of the complexes revealed that immune system- and hydrolase-related complexes were most abundant, but also that neither class showed significant isomer preference. As expected, isomerases were prominent with many cis proline peptides; however, 14–3–3 complexes were also notable, with 12 out of 59 unique peptides bound in the cis proline conformation. The high frequency of cis proline in 14–3–3 bound peptides has not been discussed in detail before. We therefore set out to explore 14–3–3 complexes and elucidate the structural and functional consequences of cis proline peptide binding.

Distinct Bound State Conformations of cis and trans Proline Peptides in 14–3–3 Complexes

Due to their ubiquitous presence in eukaryotic organisms and involvement in many diseases,22 14–3–3 proteins are widely studied. 14–3–3 proteins are folded and feature a binding groove which typically binds Ser/Thr phosphorylated peptides through a SLiM loosely defined as [RK]XXp[ST]XP, although different motifs exist.23−25 Several isoforms of 14–3–3 proteins exist, and they generally form homo or heterodimers. The PDB contains more than 500 structures of 14–3–3 proteins in complex with a partner and we identify 471 structures containing a 14–3–3 protein in complex with a phosphorylated peptide ligand. Around half of the published 14–3–3:peptide complex structures feature a proline residue in the +2 position relative to the phosphoresidue. Our analysis of the structures shows that around 20% of these crystallize in their cis isomer, an overrepresentation compared to the general cis proline frequency (Figure 1A). Structurally, the trans proline bound peptides bind along the groove while the cis proline peptides extend outward toward the opposing dimer subunit, displaying distinct binding solutions in the conformational space (Figure 1C). This conformational variation may have functional implications.

Analysis of the proline-containing 14–3–3 motifs and their bound state isomer suggests that the phosphoresidue type strongly influences conformational preference, with phosphothreonine promoting cis and phosphoserine promoting the trans proline bound conformation (Figure 1B). However, crystal structure bias does not necessarily reflect affinity difference between cis and trans isomers and thus does not allow quantification of bound state populations. In addition, the residue following the phosphoresidue and binding of small molecules also appear to influence the bound state isomer.26

The FOXO-1 transcription factor binds 14–3–3 through a p[T]WP motif,27,28 for which the expected free state cis population is around 38%.8 Three crystal structures of the complex have been solved, all showing the peptide proline in the cis conformation; however, it is also the only Trp-Pro motif for which a 14–3–3 complex structure has been published. To investigate the 14–3–3 proline isomer effect, we searched for other Trp-Pro containing 14–3–3 motifs and selected the one from the PRLR intracellular domain, a 300-residue IDR that contains a 14–3–3 SLiM very similar to that of FOXO-1 (Figure 1D).29−101 Alignment of homologous proteins shows that the Trp392 and Pro393 residues of the PRLR 14–3–3 SLiM are highly conserved relative to surrounding positions, suggesting biological importance (Figure 1D).

PRLR:14–3–3 Binding Depends on the Proline Conformation

The Trp-Pro peptide bond in the 14–3–3 SLiM is expected to have a large population in the cis isomer.8 To verify this, we recorded and assigned NMR spectra of a 21-residue PRLR peptide harboring the phosphorylated 14–3–3 SLiM. These spectra revealed that several protons including Trp392Hε1, pThr391HN and pThr391Hγ produced two large peaks corresponding to cis and trans Pro393 populations (Figure 2A). Peak-shape fitting of isolated proton peaks indicated a cis population of 38.0 ± 0.3%, in accordance with expectations (Figure S1).

Figure 2.

Figure 2

Open in a new tab

Thermodynamic and kinetic analyses of the PRLR interaction with 14–3–3. (A) Sequence and 1D 1H NMR spectra of free PRLR WT peptide showing the existence of at least two species. Going right, the cutouts show the amide region, the Trp392Hε1, and the Thr391HN regions, respectively. Horizontal brackets indicate the regions shown. Triangles indicate peaks corresponding to the cis conformation of Pro393 (filled) and Pro395 (hollow). *Phosphorylated threonine. (B) ITC experiment using high protein concentrations to demonstrate the slow postinjection equilibration time. COPASI based isomerization model fit is shown in red. (C) NMR-based time-dependent concentrations of free PRLR cis (yellow) and trans (red) isomer following addition of 14–3–3. The data is fitted using a single-exponential decay function and the horizontal dashed lines indicate the starting concentration and the extrapolated t0 concentration. The amount of instant PRLR binding is indicated outside the plot area. (D) Equilibrium model of the PRLR interaction with 14–3–3 with the apparent Kd, as determined by one-set-of-sites model fitting, along with deconvoluted affinities determined by modeling the system using COPASI.

To determine the 14–3–3 binding affinity of PRLR, we performed ITC experiments at 37 °C, titrating the pThr391 PRLR peptide into 14–3–3ζ (Figure S2A). In initial experiments we observed large exothermic peaks, followed by a slow equilibration on the time scale of several minutes, indicating an initial binding phase followed by a slower phase of unknown origin, but consistent with cis/trans proline isomerization. To capture this slower phase and ensure proper equilibration, we repeated the experiments using a 600 s delay between injections (Figures 2B and Figure S2B). The longer experiments confirmed the biphasic peak behavior and enabled us to obtain a reliable baseline, which allowed analysis using the one-set-of-sites model (Table 1). The observed two-phase kinetics likely reflects differences in binding affinities of the cis and trans proline isomers.

Table 1. Affinities and Thermodynamic Parameters of 14-3-3:PRLR Variant Interactionsa.

Peptide Kd (μM) ΔH (kJ/mol) ΔG (kJ/mol)
WTb 0.23 ± 0.03 –23.5 ± 0.3 –39.4 ± 0.3
cisc 0.13 ± 0.04 –23 ± 2 –41.0 ± 0.8
transc,d 80 ± 20 –30 ± 10 –24.3 ± 0.8
pT391pSb 2.7 ± 0.4 –18.0 ± 0.9 –33.1 ± 0.4
cisc 1.5 ± 0.2 –21 ± 3 –34.7 ± 0.4
transc 11 ± 2 –17 ± 1 –29.4 ± 0.4
P393V 28 ± 4 –12.9 ± 0.1 –27.0 ± 0.3
transc 29 ± 3 –13.0 ± 0.6 –27.0 ± 0.3
W392Yb 2.8 ± 0.8 –20.4 ± 0.9 –32.9 ± 0.7
cisc 0.43 ± 0.06 –19.4 ± 0.3 –37.8 ± 0.3
transc 8 ± 2 –22.6 ± 0.4 –30.2 ± 0.5
W392Lb 13 ± 2 –27 ± 4 –29.0 ± 0.4
W392Ab 4.0 ± 0.1 –34.2 ± 0.2 –32.1 ± 0.1
Open in a new tab
a

All parameters were obtained using either a one-set-of-sites model or a COPASI simulation modeling combined cis and trans isomer binding in addition to time-dependent isomerization. Reported errors are standard deviations based on three replicates. All isotherms are shown in Figures S2–S6.

b

Apparent parameters based on the one-set-of-sites model not accounting for difference between proline isomers.

c

Values derived from COPASI ITC experiment simulation.

d

Values are strongly dependent on the fixed, real-time NMR derived, Kd ratio.

To confirm and decompose the binding preference between cis and trans, we turned to real-time 1D 1H NMR, allowing us to distinguish and rapidly quantify the cis and trans populations free in solution. Since the 14–3–3-bound peptide is invisible due to the large size of the 14–3–3 dimer and consequent fast transverse relaxation of the complex, the decrease in signal intensity upon addition of 14–3–3 is directly reporting on the extent of binding of the respective isomer. While maintaining low temperature, we added an amount of 14–3–3 corresponding to ∼90% of the cis isomer and recorded 1D 1H NMR spectra continuously for 3 h. The first spectrum, obtained ∼5 min after introduction of 14–3–3, showed a significant intensity loss of the cis population peak as well as a minor decrease of the trans population peak. The initial changes were followed by a slow increase in cis and further decrease in trans peak intensities in subsequent spectra (Figure S2C). The behavior suggested a preferential binding of the cis isomer and a subsequent re-equilibration of the unbound PRLR cis and trans proline populations. To quantify the binding preference, we analyzed the time-dependence of the unbound cis and trans concentrations and extrapolated a single-exponential decay function to obtain the t0 concentrations directly after addition of 14–3–3 (Figure 2C and Figure S2D). The model fit indicated an isomerization exchange rate constant of (9.7 ± 0.7) × 10–4 s–1, very similar to proline isomerization rates reported for model peptides.8 Based on the y-axis intercept, the addition of 14–3–3 into PRLR resulted in instant binding of 101 ± 7 μM cis proline PRLR, while the trans concentration was not significantly affected at t0, with quantification suggesting binding of 1 ± 3 μM. Accounting for the isomer concentrations, this corresponds to an affinity ratio of Kd,trans/Kd,cis = 700 ± 200 (n = 2). Our real-time NMR experiment thus enabled us to quantify the unique isomer specificity of the PRLR interaction with 14–3–3.

Since the PRLR interaction with 14–3–3 is highly isomer specific, the equilibrium between cis and trans proline must influence the affinity observed by ITC. Therefore, the fitted one-set-of-sites dissociation constant of 230 ± 30 nM represents an apparent Kd with individual underlying Kds for the cis and trans isomers. To obtain the individual Kds, we used the COmplex PAthway Simulator (COPASI)30 program to construct a kinetic model of the ITC experiment, implementing the experimentally determined proline isomer equilibrium in addition to dissociation constants and binding enthalpies of both cis and trans isomers. Using this model, the thermodynamic parameters could be fitted to obtain a Kd,cis = 130 ± 40 nM and Kd,trans = 80 ± 20 μM based on the dissociation constant ratio (Figure 2D and Figure S2E, Table 1). Thus, the cis isomer affinity is almost 3 orders of magnitude higher than the trans isomer affinity and substantially higher than the apparent value determined using only ITC.

PRLR P393V trans Only Variant Has Low Affinity

Deconvolution of the ITC experiment suggested that the PRLR trans isomer bound 14–3–3 weakly with a Kd of ∼80 μM. To probe the trans affinity directly, we synthesized a PRLR P393V variant peptide, which has no significant cis conformation population of the Trp392-Val393 peptide bond (Figure 3A). In contrast to the WT peptide, the ITC experiments of the P393V variant showed rapid postinjection equilibration, indicative of a process lacking the slow peptide bond isomerization.31 The isotherm was analyzed using the one-set-of-sites model, producing a Kd of 28 ± 4 μM (Figure 3B), resembling the low affinity of the WT trans isomer. To validate our COPASI ITC modeling, we performed the fitting procedure on the P393V data and obtained parameters closely resembling those of the one-set-of-sites model fit (Figure 3C, Table 1).

Figure 3.

Figure 3

Open in a new tab

Analysis of the PRLR P393V variant peptide using NMR and ITC. (A) Peptide sequence and 1D 1H NMR spectrum of the PRLR P393V peptide. *Phosphorylated threonine. (B) Example isotherm of binding experiment titrating PRLR P393V into 14–3–3 analyzed using the one-set-of-sites model. COPASI model fit is shown in red. (C) Schematic model of the P393V variant interaction with 14–3–3 with no isomerization of residue 393. The apparent dissociation constant was determined from one-set-of-sites model fitting, while the trans specific constant was obtained using a COPASI ITC model simulation.

Phosphothreonine Conveys Strong cis Binding Preference

The bioinformatic analysis suggested a considerable effect of the phosphoresidue type on the proline isomer binding preference in 14–3–3:peptide complexes (Figure 4A). To investigate this bias, we synthesized a variant PRLR peptide substituting the phosphorylated threonine with a phosphorylated serine and characterized the binding using ITC and NMR. Interestingly, the cis proline content of the p[S]WP motif in the free state was only 27.9%, considerably less than expected for a Trp-Pro peptide bond and surprising considering the 38.0% measured for the p[T]WP motif. ITC experiments characterizing the pT391pS PRLR variant exhibited the same slow equilibration that was seen for the WT (Figure 4B), again indicating different binding affinities of the two isomers with slow interconversion.

Figure 4.

Figure 4

Open in a new tab

Phosphoresidue type influences isomer preference in 14–3–3 binding. (A) Crystal structure proline isomer frequencies for phosphoserine and phosphothreonine peptides and effect on bound state peptide phosphorus and Cα atom positions. pThr and pSer data show the distribution of distances from phosphoresidue group center to each member. The “cross” category shows the distribution of distances from each group member to the center of the other group. Distributions are shown as mean and standard deviation. (B) NMR-based time-dependent concentration analysis of PRLR pT391pS variant following addition of 14–3–3.

To quantify the isomer-specific affinities of the phosphoserine variant, we performed real-time NMR which revealed behavior similar to that of the WT peptide, with populations of both isomers initially decreasing and the cis population partially recovering over time (Figure 4B). However, in contrast to the WT peptide, extrapolating back to t0 revealed that both the cis and the trans isomer concentrations were reduced instantly upon addition of 14–3–3, indicating that both isomers bound. Quantification showed that the cis affinity was only around an order of magnitude higher than the trans affinity (Kd,trans/Kd,cis = 7.8 ± 0.6, n = 1) (Table 1). Based on the crystal structure bias (Figure 4A) we expected the isomer selectivity to be strongly dependent on the phosphoresidue and the 100-fold greater isomer sensitivity of the WT pThr motif relative to the pSer motif confirms this. A detailed examination of the crystal structure data set revealed that the position of the phosphoryl group is independent of phosphoresidue type, while the additional methyl group of threonine leads to a subtle dislocation of the peptide backbone (Figure 4A). This backbone dislocation might explain the effect on isomer binding preference.

14–3–3:PRLR cis Proline Selectivity Is Highly Conserved

While 14–3–3 appears effectively unable to bind the trans conformation of WT PRLR, numerous peptides do bind 14–3–3 in their trans conformation (Figure 2B). To elucidate the factors influencing this binding preference, we evaluated a series of PRLR peptide variants, spanning cis populations from 0% in the previously introduced P393V variant, to 38.0% in the WT peptide. For this, we included three additional variants with intermediate cis populations of 23.5% for W392Y, 12.5% for W392L, and 6.6% for W392A as determined using 1D 1H NMR (Figure S1).

ITC experiments using the PRLR W392Y variant revealed two notable features: the absence of an observable slow kinetic phase and a reproduceable, consistent, deviation from the standard one-set-of-sites binding model (Figure 5A and Figure S5A). Even with the apparent fast equilibration, the poor model fit suggested that both the cis and trans conformations of the W392Y peptide bind 14–3–3 but do so with a measurable difference in affinity. Fitting the ITC data to a two-sets-of-sites model yielded Kds of 400 ± 200 nM and 2.3 ± 0.2 μM, although the interpretation is ambiguous since this model does not describe isomerization.

Figure 5.

Figure 5

Open in a new tab

PRLR variant affinities and selectivity. (A) ITC data with parameters of one-set (dashed line) and two-sets-of-sites (solid line) model fits. COPASI based isomerization model fit is shown in red. (B) NMR-based time-dependent concentration of free PRLR cis and trans isomers following addition of 14–3–3. (C) The apparent and deconvoluted isomer affinities as a function of PRLR variant peptide residue number 392 side chain volume. (D) Bound equilibrium population bias of proteins capable of binding both cis and trans proline conformations with quantified affinities. The black diamonds indicate free state equilibrium populations. Data based on the Med25-ACID:DREB2A,13 ACTR:NCBD,10 and MDM2:p5333 interactions. *Isomer affinities based on molecular dynamics simulations. (E) MD simulation RMSD trajectories of PRLR in complex with 14–3–3. RMSDs were calculated for the PRLR C-terminus Cα atoms. Alignments were made based on 14–3–3 helical regions. The black line shows smoothed RMSD values. (F) Example structures extracted from MD simulations of 14–3–3:PRLR complexes in cis and trans proline isomers. Both trans panels show the same frame. (G) Analysis of PRLR peptide bond strain based on the cosine of the peptide bond angle. Reference strain values, obtained from a 14–3–3 simulation, are shown using a boxplot. The strained Pro393-Leu394 peptide bond is highlighted using a dotted outline in panels F and G.

To account for isomerization and accurately quantify the isomer affinities, we performed real-time NMR experiments using the W392Y variant (Figure 5B and Figure S5B). This peptide displayed similar time-dependence as the phosphoserine variant; with both the cis and the trans isomer concentrations reduced instantly upon addition of 14–3–3. This corroborates the ITC data indicating that both bind with comparable affinity. Comparing the fitted t0 concentrations with total isomer concentration suggested an affinity ratio of Kd,trans/Kd,cis = 19 ± 2 (n = 2). To elucidate the individual affinities, we simulated the ITC binding curve using COPASI and fitted the affinities and enthalpies of both the cis and trans interactions. The model revealed that the observed binding curve could be obtained with a Kd,cis of 430 ± 60 nM and a Kd,trans of 8 ± 2 μM (Table 1), only slightly deviating from the initial analysis using the simple two-sets-of-sites model. Thus, the W392Y variant exhibited weaker cis affinity compared to the WT, whereas its trans isomer bound considerably stronger than the WT.

Fast postinjection ITC equilibration was also observed for the W392L and W392A variants, likely indicating significant binding of both proline isomers. However, in contrast to the W392Y variant, ITC data for the W392L and W392A variants could be described using a one-set-of-sites model with apparent Kd values of 13 ± 2 μM and 4.0 ± 0.1 μM, respectively (Table 1, Figure S6). Agreement with the one-site model suggests that there is no strong isomer binding preference. The analysis of the published crystal structures, however, suggests that p[T]LP motifs preferentially populate the cis bound conformation, at least in other sequence contexts (Figure 1B). Unfortunately, real-time NMR experiments of the W392L and W392A variants were not possible due to overlap of the available cis and trans population probe peaks. Consequently, we were not able to determine isomer specific affinities and the reported Kds must be taken as apparent affinities.

Analysis of binding affinities and the side chain volume32 of the residue following the phosphoresidue suggested a dependence where larger side chains result in lower trans affinities but higher cis affinities (Figure 5C). Although we did not explore motif context in detail, the highly conserved PRLR p[T]WP motif results in a unique cis proline selectivity and our experiments show that deviation from the WT motif sequence enables significant trans proline binding. The strong conservation of the sequence that leads to almost binary proline selectivity suggests a biological relevance. To our knowledge, this degree of proline isomer selectivity has not previously been quantified for 14–3–3 or ID-based interactions. However, isomer specific affinities have been experimentally quantified for several interactions including ACTR:NCBD10 and Med25:DREB2A13 as well as computationally for MDM2:p5333 (Figure 5D). A study by Fassolari et al. identified a cis proline-selective interaction between a monoclonal antibody and the E7 protein of human papillomavirus; however no direct observations of the individual isomer interactions were reported.15

Limited Space in the Narrow Binding Groove Dictates Isomer Specificity

To investigate the determinants of the 14–3–3:PRLR isomer preference, we performed molecular dynamics simulations of the complex in both cis and trans proline isomers. Cis proline PRLR retained the initial peptide structure predicted by AlphaFold3,34 closely resembling the crystal structure of the 14–3–3:FOXO-1 complex, with an RMSD of less than 2 Å (Figure 5E and Figure S7A). However, the PRLR trans proline simulation showed increased fluctuations in the initial trans-crystal-like configuration (Figure 5E and Figure S7B), finally dislocating the Trp392 side chain out of the binding groove and undergoing significant restructuring of the C-terminal region after around 600 ns. Following this restructuring the system stabilized in an alternative trans conformation (Figure S7C). The diversion of the trans peptide configuration from the initial canonical conformation seen for almost all crystallized trans p[ST]XP peptides alludes to the incompatibility of the WT trans interaction.

Examining the final peptide structures revealed that the trans proline peptide adopts a Trp-Pro motif configuration similar to the one seen in the cis bound peptide. However, while the cis proline isomer enables the peptide chain to exit the binding groove, the trans proline forces the chain into the groove (Figure 5F). This configuration differs from that of typical trans proline bound peptides, where the backbone loops out of the groove before re-entering (Figure 1C and Figure S8). Analysis of dihedral angles indicates that the peptide bond following the trans proline is severely strained in this alternative configuration (Figure 5G). This strain is likely due to steric clashes between PRLR and 14–3–3 residues Asn42, Ser45, and Val46 and offers a structure-based rationale for the poor trans isomer binding (Figure 5F and Figure S7D).

Simulations of variant PRLR trans conformation peptides indicated similar final structures for WT PRLR and the W392Y variant, while the W392A and P393V variants retained the trans-crystal-like configuration (Figure S9). The PRLR phosphoserine and W392L variants sampled both the trans-crystal-like configuration and the alternative configuration adopted by the WT trans peptide, with the side chain of residue 392 being displaced outside the groove. The heterogeneity of the phosphoserine variant supports the bioinformatic analysis indicating that the threonine methyl group increases the isomer selectivity by subtly displacing the peptide backbone to prevent the canonical trans proline bound conformation.

Thus, based on the structures observed in the trans proline MD simulations, we argue that the Trp and Tyr side chains are poorly accommodated in the typical trans binding position, and their displacement from the groove, as seen for the cis peptides, results in clashes between 14–3–3 and residues following Pro393.

Implications of Isomer-Selective Steps in Signal Transduction

Protein–protein interactions typically occur on the millisecond time scale and in cases where no conformational selection is involved, responses are observed within this rapid window (Figure 6A). However, proline dependent conformational selection, such as in the 14–3–3:PRLR interaction, can introduce complex, multiphase response curves. In such cases, the available compatible isomer binds on the millisecond time scale, resulting in rapid but limited response, while the full response requires several minutes (Figure 6B). In multistep pathways, such as phosphorylation and binding, with subsequent steps requiring the alternate isomer, significant lag times between the signal and the response may occur (Figure 6C). Such scenarios are not unlikely due to the prevalence of proline directed kinases and the known isomeric selectivity of e.g., the ERK2 kinase35 and the Ssu72 phosphatase.36 These response delays may be modulated by proline cis/trans isomerases which can increase the isomerization rate by several orders of magnitude.37 Interestingly, an isomer selective binding partner, like 14–3–3, has the opposite effect compared to a proline cis/trans isomerase: the equilibrium is shifted to the preferred isomer and the isomerization kinetics are dampened in the bound state. Consequently, competing reactions like i.e., dephosphorylation can be effectively suppressed for time scales up to several hours (Figure 6D).

Figure 6.

Figure 6

Open in a new tab

Interaction schemes (left) and associated response curves on a logarithmic time scale (right) of various possible pathways. The response (concentration) curves are colored according to the protein state they represent. (A) Interaction with no conformational selection. (B) Protein in cis and trans proline equilibrium with only the cis proline capable of binding. Two-phase response curve with initial binding of available cis isomer, followed by trans to cis isomerization and binding. (C) Multistep pathway with specific, opposing, isomer-selective steps. Here shown with trans isomer phosphorylation and cis isomer specific binding. (D) 14–3–3:PRLR trans specific dephosphorylation response.

It is evident from the response curves (Figure 6) that the length and amplitude of an isomer-dependent signal will influence the total proline isomer population. An implication of this is that the ensemble may gain a memory-like functionality, maintaining a specific state for a substantial length of time following dissociation.17,38 Thus, proline isomerization can act as a timed molecular switch controlling the progression of signaling events, as seen e.g., in the interaction of the circadian clock related proteins BMAL1 and CLOCK.12

Conclusion

In this study, we explored the effect of proline isomerization on interactions between 14–3–3 and their phosphorylated disordered binding partners through a combination of bioinformatics and experimental approaches supported by MD simulations. We present a novel experimental method for real-time NMR combined with ITC that enables straightforward quantification of isomer affinities. This approach may be of general relevance for characterizing interactions involving proline-containing SLiMs.

Our investigation of the PRLR interaction with 14–3–3 reveals a strong preference for the cis proline isomer of a canonical 14–3–3 binding SLiM. Notably, the 14–3–3 motif of PRLR exhibits near-binary selectivity, with negligible binding of the trans isomer. This results in a binding with high apparent affinity, but slow effective association due to the requirement for upstream proline isomerization. Crystal structure analysis, supported by variant peptide experiments, demonstrated that 14–3–3 isomer selectivity is highly dependent on the phosphoresidue type, with threonine strongly favoring cis proline specificity. Our mutational studies also show that the WT sequence has a unique isomer preference through a dual effect of high cis affinity, while at the same time having a low trans isomer affinity. MD simulations provided the structural basis for this selectivity, showing that the large tryptophan side chain of the WT motif requires an alternative, strained, peptide backbone conformation to accommodate the trans proline peptide C-terminus. This highlights the functional conservation of the PRLR 14–3–3-binding SLiM and suggests a key role for proline conformational selection in modulating steric or kinetic aspects of the PRLR signaling pathway.

The observed isomer selectivity has important implications for drug development. Several 14–3–3 interaction partners bind both subunits of the 14–3–3 dimer simultaneously via closely spaced SLiMs.39−41 The specific spacing necessitates a sharp turn of the peptide backbone similar to that seen for cis proline bound peptides. The ability of certain small molecules to promote cis proline binding could specifically enhance these bivalent interactions. Additionally, therapeutic strategies relying on small molecule binding to 14–3–3 may affect cis and trans proline binding affinities differently, possibly introducing isomer preference and thus a kinetic component. Targeting isomer-specific interactions may enable therapeutic strategies that direct signaling pathways toward specific outcomes, offering a novel approach for disease treatment.

Finally, we highlight that proline isomer-selective steps, when generalized to simplified signal transduction pathways, can introduce substantial delays, occurring on time scales rarely encountered in structural biology. This delay mechanism may confer unique memory-like functionality in biological systems, offering new perspectives on the temporal regulation of signaling processes.

Methods

Analysis of Protein:Peptide Complexes

14–3–3:peptide complex entries were collected by querying the RCSB database for crystal structures containing at least two different protein polymers, with one being no more than 30 residues long and the overall resolution being better than 2.0 Å. These entries were analyzed using a PyMOL based script (available on GitHub) to identify peptide proline residues and for each proline the peptide bond angle of the preceding peptide bond was determined. Cis and trans proline conformations were classified from bond angles of 0 ± 30 and 180 ± 30 degrees, respectively. Separately analyzing the two conformations, each entry was grouped based on sequence identity of the larger folded protein by performing a multiple sequence alignment of all entries. To remove duplicate and very similar peptides within each group, we computed the Levenshtein distance between each peptide. Peptides with a Levenshtein distance less than 20% of the length of the peptide were considered identical and not counted as a separate unique entry. Complex class keywords were extracted from crystallographic information files and each entry was attributed one or more classes depending on the content of the file. One entry can have several classes, eg., hydrolase inhibitor. To count the number of entries in each class while avoiding duplicate counting, each group was processed individually with each keyword or class receiving a number of counts proportional to the number of unique entries in the group and the fraction of entries in the group with the specific class attributed.

Analysis of 14–3–3 Complexes

14–3–3 complexes were analyzed using a PyMOL-based script (available on GitHub) which identified phosphorylated residues and measured the peptide bond angle of the +1 to +2 peptide bond. Each complex within the published crystal asymmetric unit were analyzed separately. Peptides were grouped based on the phosphoresidue and the following two residues and for the purpose of this work, entries without proline in the +2 position were discarded.

Protein Expression and Purification

Human 14–3–3ζ protein was expressed with a 6xHis-SUMO tag in Escherichia coli NiCo21(DE3) cells (New England BioLabs). Following transformation, an overnight culture was initiated by inoculating 10 mL of LB medium containing 50 μg/mL kanamycin with a single colony. This culture was then used to inoculate 1 L of LB medium supplemented with 50 μg/mL kanamycin. The cells were grown at 37 °C to an OD600 of 0.8–0.9, and protein expression was induced with 0.5 mM IPTG for 4–5 h at 37 °C. Cells were harvested by centrifugation at 5,000 × g for 20 min at 4 °C. The pellet was resuspended in 30–35 mL of lysis buffer (20 mM Bis-Tris, pH 7.2, 150 mM NaCl, 10 mM imidazole, 5 mM β-mercaptoethanol) and lysed using a French press (Constant Systems Ltd., Daventry, UK) at 25,000 psi or sonication on ice (10 min of 0.5 s pulses, per 20 mL sample). The lysate was clarified by centrifugation at 20,000 × g for 45 min at 4 °C, and the supernatant was incubated with 5 mL of Ni-NTA beads for 20–30 min. The Ni-NTA resin was washed with 50 mL of a high-salt buffer (20 mM Bis-Tris, pH 7.2, 1 M NaCl, 10 mM imidazole, 5 mM β-mercaptoethanol), followed by a wash with 50 mL of the lysis buffer. The 14–3–3ζ protein was eluted with 15 mL of elution buffer (20 mM Bis-Tris, pH 7.2, 150 mM NaCl, 250 mM imidazole, 5 mM β-mercaptoethanol). The eluted protein was dialyzed against 20 mM Bis-Tris (pH 6.5) to remove imidazole and reduce the salt concentration. Retaining the SUMO tag at this stage was crucial, as it facilitates binding to a heparin column, which was employed to separate 14–3–3ζ from a specific impurity. The dialyzed protein was filtered through a 0.45 μm filter and loaded onto a 1 mL heparin column equilibrated with buffer (20 mM Bis-Tris, pH 6.5, 5 mM β-mercaptoethanol). The column was washed with 15 column volumes (CV) of the same buffer, and the protein was eluted using a two-step gradient: 0% to 30% of a high-salt buffer (20 mM Bis-Tris, pH 6.5, 1 M NaCl, 5 mM β-mercaptoethanol) over 3 CV, followed by 30% to 100% of the same buffer over 20 CV. Fractions containing the purified 14–3–3ζ protein, free from the impurity, were pooled and dialyzed to prepare the sample for SUMO tag cleavage. ULP1 protease (∼0.1 mg) and 2 mM dithiothreitol (DTT) were added, and the sample was incubated overnight at room temperature. The cleaved protein was then passed through a reverse Ni-NTA column to remove the 6xHis-SUMO tag, and the flow-through containing the untagged 14–3–3ζ was collected. Final purification and buffer exchange was done by size-exclusion chromatography using a HiLoad Superdex 200 16/60 column.

Synthetic phosphorylated human PRLR peptides were purchased from Biosynth (Biosynth B.V., Lelystad, The Netherlands) as lyophilized powder at >95% purity according to reversed phase and mass spectrometry analysis (Table S1). No additional peaks from major contaminants were detected during NMR analysis of the peptides. Peptides were dissolved in water containing 1 mM β-mercaptoethanol and adjusted to pH ∼ 7 prior to aliquoting and lyophilizing. All peptides were analyzed by 13C NMR to confirm complete phosphorylation (Figure S10).

Isothermal Titration Calorimetry

ITC was performed on either a MicroCal ITC200 (MicroCal Instruments, Northampton, United Kingdom) or a Malvern Panalytical PEAQ-ITC (Malvern Panalytical, Malvern, United Kingdom). All ITC experiments were performed using a 50 mM HEPES pH 7.4, 100 mM NaCl and 0.5 mM tris(2-carboxyethyl)phosphine (TCEP) buffer. PRLR peptides were buffer exchanged in 3 kDa cutoff spinfilters to ensure identical buffer conditions for cell and syringe samples. Concentrations were determined using the absorbance at 280 nm measured using a NanoDrop 1000 spectrophotometer. All experiments were performed using a ∼ 1:10 cell to syringe protein concentration ratio. PRLR WT and pT391pS experiments were performed initially using a 180 s injection delay which did not equilibrate between injections. Experiments were redone using 600 s injection delay which resulted in adequate postinjection equilibration. Other variant peptides did not display slow equilibration and were thus studied using shorter 180 s injection delays.

Samples were degassed by centrifugation at 20,000 × g for 20 min at experimental temperature. Reported thermodynamic parameters are given at 37 °C. Data were analyzed using Origin and in-house scripts using the one-set-of-sites or two-sets-of-sites models as documented by MicroCal and Origin. Dilution heats were fitted using an offset parameter. All experiments were performed using a 0.5 μL initial injection, which was discarded in the analysis, followed by PRLR injections between 2.0 and 2.4 μL.

NMR Spectroscopy

NMR data was recorded on Bruker Avance 750 or 800 MHz (1H) spectrometers equipped with cryogenic probes. 1D proton spectra were analyzed using TopSpin v4.2.0. Two-dimensional spectra were processed using TopSpin and analyzed using CcpNmr Analysis. All NMR samples were prepared in NMR buffer containing 10% (v/v) D2O, 0.02% (w/v) NaN3, 0.2 mM 4,4-dimethyl-4-silapen-tane-1-sulfonic acid (DSS), 20 mM Na2HPO4/NaH2PO4 pH 6.5, 100 mM NaCl and 10 mM DTT. 1D 1H NMR spectra were processed using an exponential decay window function.

NMR Time-Dependent Intensity

Time-dependent NMR peak integrals were obtained for PRLR WT, pT391pS and W392Y peptides on a Bruker Avance 750 MHz (1H) spectrometer at 5 °C. For the PRLR WT data, three samples were prepared, with one used as a starting point reference by adding NMR buffer to obtain spectra of the unperturbed PRLR peptide at the experimental peptide concentration. For the pT391pS and W392Y samples, several spectra were recorded on the samples before adding 14–3–3 in NMR buffer to start the experiment. Final concentrations of the samples were 350 μM PRLR and 113 μM 14–3–3 for WT, 315 μM PRLR and 66 μM 14–3–3 for W392Y, and 1031 μM PRLR and 155 μM 14–3–3 for pT391pS. For WT and pT391pS samples, the Trp392 side chain indole 1Hε1 peak was observed. For the W392Y sample, the pThr391 backbone 1HN peak was used. The experiments were started by transferring the PRLR sample from the NMR tube to a tube containing a specific volume and concentration of 14–3–3 in matched NMR buffer. The mixed sample was then transferred back to the NMR tube and lowered into the magnet. Upon stabilization of the lock signal, the standard 1D experiment preparations were performed using automatic procedures. The first spectra were obtained around 5 min after mixing the PRLR sample with 14–3–3. Spectra were processed using TopSpin and exported using the “totxt” command (batch version available on GitHub). Baseline correction and peak integration was performed using a python script (available on GitHub) using the same width integration regions for cis and trans population peaks. Signal-to-concentration conversion factors were established using the premixing spectra, assuming the measured concentration was correct. The premixing sample spectra were obtained using the same pulse sequence with the same parameters as the time trace recordings. The concentration time dependence was modeled using an single phase exponential decay function [PRLR] = ([PRLR]0 – [PRLR]final) × exp(-kex × t) + [PRLR]final, where [PRLR]0 corresponds to the concentration immediately following addition of 14–3–3. Cis and trans PRLR traces were fitted individually and shown kex values are mean ± standard deviations of all available traces.

ITC Simulation Model Fitting

ITC experiments were modeled using COPASI by building a model of the interaction which included both injection and binding of cis and trans peptide. The model included variables parameters for the affinity of the cis isomer, the Kd ratio determined from the NMR time-course experiments, enthalpies of cis and trans binding, the proline isomerization equilibrium constant, and the isomerization rate. Time-course simulations were performed using events to trigger injection-based dilution of cell species and addition of free PRLR peptide according to the concentration expected in the syringe. Concentrations of various species were extracted prior to injections and released heat was calculated based on changes in concentration, binding enthalpies, and the volume of the reaction cell. In addition to the thermodynamic parameters, the fitting routine implemented an N-value which was implemented as a change in the concentration of the cell protein at the beginning of the time-course simulation. Due to the particular way species in the reaction cell are diluted upon and during injections, a concentration error of 0.2% accumulated during the simulations. This concentration error was disregarded, and injections were compared directly. The code and a sample COPASI project are available on GitHub.

Bound State Equilibrium

Bound state populations were calculated using COPASI steady-state simulations and published free state cis/trans populations along with quantified isomer binding affinities. Errors were obtained by performing 1000 simulations with normal distributed parameters according to reported errors.

Molecular Dynamics Simulations

Molecular dynamics was performed using AMBER99SB-ILDN42 with parametrized phosphorylated residues in TIP3P water43 with 100 mM NaCl and neutral overall charge. Starting structures were obtained from AlphaFold334 predictions using 14–3–3ζ and a short PRLR peptide corresponding to what is typically resolved in crystal structures of 14–3–3 complexes. The systems were equilibrated using NPT and NVT protocols at 300 K. The trans proline isomer of the PRLR WT complex was obtained by performing a 10 ns simulation with a dihedral restraint on the Trp-Pro peptide bond corresponding to 60 kJ/mol/rad2 with a threshold of ±10 degrees. The WT PRLR cis isomer was simulated unrestrained for 1 μs while the unrestrained trans isomer was simulated for 2 μs. To evaluate peptide fluctuations and rearrangement, the structured parts of 14–3–3 were aligned and RMSDs relative to the starting structure calculated for the peptide. Variant complexes were obtained similarly by restraining the Xaa-Pro peptide bond to obtain the opposite isomer and simulating for 1 μs. Three replicates of the phosphoserine variant simulation were performed. Starting structures and representative snapshots are available on GitHub (doi.org/10.5281/zenodo.13692244). Peptide bond strain was calculated as the average cosine of two times the bond angle.

Signal Response Curve Simulations

Response curves were obtained using COPASI time-course simulations of the systems starting from an equilibrium and then introducing binding partner, kinase, or phosphatase at t = 0. Simulations were performed using a kon of 500 μM–1 s–1, an isomer exchange rate of ∼8 × 10–4 s–1, and an irreversible (de)phosphorylation rate of 1 s–1. The proline cis/trans equilibrium was modeled according to a Trp-Pro motif at 4 °C.

Acknowledgments

The authors want to thank Giulio Tesei for the course-grained model of the PRLR:PRL:JAK2 membrane assembly shown in Figure 1D, F. Emil Thomasen for discussions on all-atom molecular dynamics simulations, Signe A. Sjørup for technical support, and Abdel Hafid El Bouyahyaoui for initial work on 14–3–3 purification. This work was supported by the Novo Nordisk Foundation challenge grant REPIN, rethinking protein interactions (NNF18OC0033926 to B.B.K.) and Novo Nordisk Foundation postdoc fellowship grant (NNF24OC0088767 to F.F.T.). NMR spectra were recorded at cOpenNMR, an infrastructure facility funded in part by the Novo Nordisk Foundation (NNF18OC0032996). We thank Villumfonden for supporting the NMR facility. Molecular dynamics simulations were performed on the UCloud interactive HPC system managed by the eScience Center at the University of Southern Denmark.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/jacs.4c13462.

  • Raw or minimally processed data deposited on GitHub and archived using Zenodo (doi.org/10.5281/zenodo.13692244), peptides used in this study (Table S1), population analyses of PRLR peptide cis and trans isomers (Figure S1), ITC and NMR results for the WT PRLR peptide (Figure S2), ITC and NMR results for the PRLR pT391pS variant peptide (Figure S3), ITC results for the PRLR P393V variant peptide interaction with 14–3–3 (Figure S4), ITC and NMR results for the PRLR W392Y variant peptide (Figure S5), ITC result for the PRLR W392L (A) and W392A (B) peptides (Figure S6), molecular dynamics simulation and analysis of cis and trans Pro393 PRLR WT in complex with 14–3–3 (Figure S7), crystal and MD simulation derived structures of peptides in complex with 14–3–3 (Figure S8), molecular dynamics simulation of PRLR variant peptides in complex with 14-3-3 (Figure S9), NMR analysis of peptide phosphorylation level (Figure S10) (PDF)

This work was supported by the Novo Nordisk Foundation challenge grant REPIN, rethinking protein interactions (NNF18OC0033926 to B.B.K.) and Novo Nordisk Foundation postdoc fellowship grant (NNF24OC0088767 to F.F.T.). NMR spectra were recorded at cOpenNMR, an infrastructure facility funded by the Novo Nordisk Foundation (NNF18OC0032996). Villumfonden supported the NMR Facility.

The authors declare no competing financial interest.

Supplementary Material

ja4c13462_si_001.pdf (31.3MB, pdf)

References

  1. Holehouse A. S.; Kragelund B. B. The Molecular Basis for Cellular Function of Intrinsically Disordered Protein Regions. Nat. Rev. Mol. Cell Biol. 2024, 1, 187–211. 10.1038/s41580-023-00673-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  2. Davey N. E.; Van Roey K.; Weatheritt R. J.; Toedt G.; Uyar B.; Altenberg B.; Budd A.; Diella F.; Dinkel H.; Gibson T. J. Attributes of Short Linear Motifs. Mol. Biosyst 2012, 8 (1), 268–281. 10.1039/C1MB05231D. [DOI] [PubMed] [Google Scholar]
  3. Ivarsson Y.; Jemth P. Affinity and Specificity of Motif-Based Protein–Protein Interactions. Curr. Opin Struct Biol. 2019, 54, 26–33. 10.1016/j.sbi.2018.09.009. [DOI] [PubMed] [Google Scholar]
  4. Theillet F.-X.; Kalmar L.; Tompa P.; Han K.-H.; Selenko P.; Dunker A. K.; Daughdrill G. W.; Uversky V. N. The Alphabet of Intrinsic Disorder. Intrinsically Disord Proteins 2013, 1 (1), e24360 10.4161/idp.24360. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Morris A. L.; MacArthur M. W.; Hutchinson E. G.; Thornton J. M. Stereochemical Quality of Protein Structure Coordinates. Proteins: Struct., Funct., Bioinf. 1992, 12 (4), 345–364. 10.1002/prot.340120407. [DOI] [PubMed] [Google Scholar]
  6. Wedemeyer W. J.; Welker E.; Scheraga H. A. Proline Cis–Trans Isomerization and Protein Folding. Biochemistry 2002, 41 (50), 14637–14644. 10.1021/bi020574b. [DOI] [PubMed] [Google Scholar]
  7. Weiss M. S.; Jabs A.; Hilgenfeld R. Peptide Bonds Revisited. Nat. Struct Mol. Biol. 1998, 5 (8), 676–676. 10.1038/1368. [DOI] [PubMed] [Google Scholar]
  8. Reimer U.; Scherer G.; Drewello M.; Kruber S.; Schutkowski M.; Fischer G. Side-Chain Effects on Peptidyl-Prolyl Cis/Trans Isomerisation. J. Mol. Biol. 1998, 279 (2), 449–460. 10.1006/jmbi.1998.1770. [DOI] [PubMed] [Google Scholar]
  9. Sebák F.; Szolomájer J.; Papp N.; Tóth G. K.; Bodor A. Proline Cis/Trans Isomerization in Intrinsically Disordered Proteins and Peptides. Front. Biosci.-Landmark 2023, 28 (6), 127. 10.31083/j.fbl2806127. [DOI] [PubMed] [Google Scholar]
  10. Zosel F.; Mercadante D.; Nettels D.; Schuler B. A Proline Switch Explains Kinetic Heterogeneity in a Coupled Folding and Binding Reaction. Nat. Commun. 2018, 9 (1), 3332. 10.1038/s41467-018-05725-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Gibbs E. B.; Lu F.; Portz B.; Fisher M. J.; Medellin B. P.; Laremore T. N.; Zhang Y. J.; Gilmour D. S.; Showalter S. A. Phosphorylation Induces Sequence-Specific Conformational Switches in the RNA Polymerase II C-Terminal Domain. Nat. Commun. 2017, 8, 15233. 10.1038/ncomms15233. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Gustafson C. L.; Parsley N. C.; Asimgil H.; Lee H. W.; Ahlbach C.; Michael A. K.; Xu H.; Williams O. L.; Davis T. L.; Liu A. C.; Partch C. L. A Slow Conformational Switch in the BMAL1 Transactivation Domain Modulates Circadian Rhythms. Mol. Cell 2017, 66 (4), 447–457. 10.1016/j.molcel.2017.04.011. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Theisen F. F.; Prestel A.; Elkjaer S.; Leurs Y. H. A; Morffy N.; O’Shea C.; Teilum K.; Kragelund B. B.; Skriver K. Molecular Switching in Transcription through Splicing and Proline-Isomerization Regulates Stress Responses in Plants. Nat. Commun. 2024, 15, 592. 10.1038/s41467-024-44859-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. Favretto F.; Flores D.; Baker J. D.; Strohäker T.; Andreas L. B.; Blair L. J.; Becker S.; Zweckstetter M. Catalysis of Proline Isomerization and Molecular Chaperone Activity in a Tug-of-War. Nat. Commun. 2020, 11 (1), 6046. 10.1038/s41467-020-19844-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Fassolari M.; Chemes L. B.; Gallo M.; Smal C.; Sánchez I. E.; De Prat-Gay G. Minute Time Scale Prolyl Isomerization Governs Antibody Recognition of an Intrinsically Disordered Immunodominant Epitope. J. Biol. Chem. 2013, 288 (18), 13110–13123. 10.1074/jbc.M112.444554. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Mateos B.; Conrad-Billroth C.; Schiavina M.; Beier A.; Kontaxis G.; Konrat R.; Felli I. C.; Pierattelli R. The Ambivalent Role of Proline Residues in an Intrinsically Disordered Protein: From Disorder Promoters to Compaction Facilitators. J. Mol. Biol. 2020, 432 (9), 3093–3111. 10.1016/j.jmb.2019.11.015. [DOI] [PubMed] [Google Scholar]
  17. Sarkar P.; Reichman C.; Saleh T.; Birge R. B.; Kalodimos C. G. Proline Cis-Trans Isomerization Controls Autoinhibition of a Signaling Protein. Mol. Cell 2007, 25 (3), 413–426. 10.1016/j.molcel.2007.01.004. [DOI] [PMC free article] [PubMed] [Google Scholar]
  18. Follis A. V.; Llambi F.; Merritt P.; Chipuk J. E.; Green D. R.; Kriwacki R. W. Pin1-Induced Proline Isomerization in Cytosolic P53 Mediates BAX Activation and Apoptosis. Mol. Cell 2015, 59 (4), 677–684. 10.1016/j.molcel.2015.06.029. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Eckert B.; Martin A.; Balbach J.; Schmid F. X. Prolyl Isomerization as a Molecular Timer in Phage Infection. Nat. Struct Mol. Biol. 2005, 12 (7), 619–623. 10.1038/nsmb946. [DOI] [PubMed] [Google Scholar]
  20. Alderson T. R.; Lee J. H.; Charlier C.; Ying J.; Bax A. Propensity for Cis-Proline Formation in Unfolded Proteins. ChemBioChem. 2018, 19 (1), 37–42. 10.1002/cbic.201700548. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Pogozheva I. D.; Cherepanov S.; Park S.-J.; Raghavan M.; Im W.; Lomize A. L. Structural Modeling of Cytokine-Receptor-JAK2 Signaling Complexes Using AlphaFold Multimer. J. Chem. Inf Model 2023, 63 (18), 5874–5895. 10.1021/acs.jcim.3c00926. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Foote M.; Zhou Y. 14–3-3 Proteins in Neurological Disorders. Int. J. Biochem. Mol. Biol. 2012, 3 (2), 152–164. [PMC free article] [PubMed] [Google Scholar]
  23. Yaffe M. B.; Rittinger K.; Volinia S.; Caron P. R.; Aitken A.; Leffers H.; Gamblin S. J.; Smerdon S. J.; Cantley L. C. The Structural Basis for 14–3-3:Phosphopeptide Binding Specificity. Cell 1997, 91 (7), 961–971. 10.1016/S0092-8674(00)80487-0. [DOI] [PubMed] [Google Scholar]
  24. Pennington K.; Chan T.; Torres M.; Andersen J. The Dynamic and Stress-Adaptive Signaling Hub of 14–3-3: Emerging Mechanisms of Regulation and Context-Dependent Protein–Protein Interactions. Oncogene 2018, 37 (42), 5587–5604. 10.1038/s41388-018-0348-3. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. Johnson C.; Crowther S.; Stafford M. J.; Campbell D. G.; Toth R.; MacKintosh C. Bioinformatic and Experimental Survey of 14–3-3-Binding Sites. Biochem. J. 2010, 427 (1), 69–78. 10.1042/BJ20091834. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Somsen B. A.; Sijbesma E.; Leysen S.; Honzejkova K.; Visser E. J.; Cossar P. J.; Obšil T.; Brunsveld L.; Ottmann C. Molecular Basis and Dual Ligand Regulation of Tetrameric Estrogen Receptor α/14–3-3ζ Protein Complex. J. Biol. Chem. 2023, 299 (7), 104855 10.1016/j.jbc.2023.104855. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Tzivion G.; Dobson M.; Ramakrishnan G. FoxO Transcription Factors; Regulation by AKT and 14–3-3 Proteins. Biochimica et Biophysica Acta (BBA) - Molecular. Cell Research 2011, 1813 (11), 1938–1945. 10.1016/j.bbamcr.2011.06.002. [DOI] [PubMed] [Google Scholar]
  28. Saline M.; Badertscher L.; Wolter M.; Lau R.; Gunnarsson A.; Jacso T.; Norris T.; Ottmann C.; Snijder A. AMPK and AKT Protein Kinases Hierarchically Phosphorylate the N-Terminus of the FOXO1 Transcription Factor, Modulating Interactions with 14–3-3 Proteins. J. Biol. Chem. 2019, 294 (35), 13106–13116. 10.1074/jbc.RA119.008649. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Olayioye M. A.; Guthridge M. A.; Stomski F. C.; Lopez A. F.; Visvader J. E.; Lindeman G. J. Threonine 391 Phosphorylation of the Human Prolactin Receptor Mediates a Novel Interaction with 14–3-3 Proteins. J. Biol. Chem. 2003, 278 (35), 32929–32935. 10.1074/jbc.M302910200. [DOI] [PubMed] [Google Scholar]
  30. Haxholm G. W.; Nikolajsen L. F.; Olsen J. G.; Fredsted J.; Larsen F. H.; Goffin V.; Pedersen S. F.; Brooks A. J.; Waters M. J.; Kragelund B. B. Intrinsically disordered cytoplasmic domains of two cytokine receptors mediate conserved interactions with membranes. Biochem J. 2015, 468 (3), 495–506. 10.1042/BJ20141243. [DOI] [PubMed] [Google Scholar]
  31. Araya-Secchi R.; Bugge K.; Seiffert P.; Petry A.; Haxholm G. W.; Lindorff-Larsen K.; Pedersen S. F.; Arleth L.; Kragelund B. B. The prolactin receptor scaffolds Janus kinase 2 via co-structure formation with phosphoinositide-4,5-bisphosphate. Elife 2023, 12, e84645. 10.7554/eLife.84645. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. Hoops S.; Sahle S.; Gauges R.; Lee C.; Pahle J.; Simus N.; Singhal M.; Xu L.; Mendes P.; Kummer U. COPASI—a COmplex PAthway SImulator. Bioinformatics 2006, 22 (24), 3067–3074. 10.1093/bioinformatics/btl485. [DOI] [PubMed] [Google Scholar]
  33. Odefey C.; Mayr L. M.; Schmid F. X. Non-Prolyl Cis-Trans Peptide Bond Isomerization as a Rate-Determining Step in Protein Unfolding and Refolding. J. Mol. Biol. 1995, 245 (1), 69–78. 10.1016/S0022-2836(95)80039-5. [DOI] [PubMed] [Google Scholar]
  34. Zamyatnin A. A. Amino Acid, Peptide, and Protein Volume in Solution. Annu. Rev. Biophys. Bioeng. 1984, 13 (1), 145–165. 10.1146/annurev.bb.13.060184.001045. [DOI] [PubMed] [Google Scholar]
  35. Zhan Y. A.; Ytreberg F. M. The Cis Conformation of Proline Leads to Weaker Binding of a P53 Peptide to MDM2 Compared to Trans. Arch. Biochem. Biophys. 2015, 575, 22–29. 10.1016/j.abb.2015.03.021. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Abramson J.; Adler J.; Dunger J.; Evans R.; Green T.; Pritzel A.; Ronneberger O.; Willmore L.; Ballard A. J.; Bambrick J.; Bodenstein S. W.; Evans D. A.; Hung C. C.; O’Neill M.; Reiman D.; Tunyasuvunakool K.; Wu Z.; Žemgulytė A.; Arvaniti E.; Beattie C.; Bertolli O.; Bridgland A.; Cherepanov A.; Congreve M.; Cowen-Rivers A. I.; Cowie A.; Figurnov M.; Fuchs F. B.; Gladman H.; Jain R.; Khan Y. A.; Low C. M. R.; Perlin K.; Potapenko A.; Savy P.; Singh S.; Stecula A.; Thillaisundaram A.; Tong C.; Yakneen S.; Zhong E. D.; Zielinski M.; Žídek A.; Bapst V.; Kohli P.; Jaderberg M.; Hassabis D.; Jumper J. M. Accurate Structure Prediction of Biomolecular Interactions with AlphaFold 3. Nature 2024, 630 (8016), 493–500. 10.1038/s41586-024-07487-w. [DOI] [PMC free article] [PubMed] [Google Scholar]
  37. Weiwad M.; Küllertz G.; Schutkowski M.; Fischer G. Evidence That the Substrate Backbone Conformation Is Critical to Phosphorylation by P42 MAP Kinase. FEBS Lett. 2000, 478 (1–2), 39–42. 10.1016/S0014-5793(00)01794-4. [DOI] [PubMed] [Google Scholar]
  38. Werner-Allen J. W.; Lee C. J.; Liu P.; Nicely N. I.; Wang S.; Greenleaf A. L.; Zhou P. Cis-Proline-Mediated Ser(P)5 Dephosphorylation by the RNA Polymerase II C-Terminal Domain Phosphatase Ssu72. J. Biol. Chem. 2011, 286 (7), 5717–5726. 10.1074/jbc.M110.197129. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Babu M.; Favretto F.; Rankovic M.; Zweckstetter M. Peptidyl Prolyl Isomerase A Modulates the Liquid-Liquid Phase Separation of Proline-Rich IDPs. J. Am. Chem. Soc. 2022, 144 (35), 16157–16163. 10.1021/jacs.2c07149. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Schmidpeter P. A. M.; Koch J. R.; Schmid F. X. Control of Protein Function by Prolyl Isomerization. Biochimica et Biophysica Acta - General Subjects. 2015, 1850, 1973–1982. 10.1016/j.bbagen.2014.12.019. [DOI] [PubMed] [Google Scholar]
  41. Kostelecky B.; Saurin A. T.; Purkiss A.; Parker P. J.; McDonald N. Q. Recognition of an Intra-Chain Tandem 14–3-3 Binding Site within PKCe. EMBO Rep 2009, 10 (9), 983–989. 10.1038/embor.2009.150. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Stevers L. M.; Lam C. V.; Leysen S. F. R.; Meijer F. A.; Van Scheppingen D. S.; De Vries R. M. J. M.; Carlile G. W.; Milroy L. G.; Thomas D. Y.; Brunsveld L.; Ottmann C. Characterization and Small-Molecule Stabilization of the Multisite Tandem Binding between 14–3-3 and the R Domain of CFTR. Proc. Natl. Acad. Sci. U. S. A. 2016, 113 (9), E1152–E1161. 10.1073/pnas.1516631113. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Chen X.; Liu Z.; Shan Z.; Yao W.; Gu A.; Wen W. Structural Determinants Controlling 14–3-3 Recruitment to the Endocytic Adaptor Numb and Dissociation of the Numb·α-Adaptin Complex. J. Biol. Chem. 2018, 293 (11), 4149–4158. 10.1074/jbc.RA117.000897. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Lindorff-Larsen K.; Piana S.; Palmo K.; Maragakis P.; Klepeis J. L.; Dror R. O.; Shaw D. E. Improved Side-Chain Torsion Potentials for the Amber Ff99SB Protein Force Field. Proteins 2010, 78 (8), 1950–1958. 10.1002/prot.22711. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Jorgensen W. L.; Chandrasekhar J.; Madura J. D.; Impey R. W.; Klein M. L. Comparison of Simple Potential Functions for Simulating Liquid Water. J. Chem. Phys. 1983, 79 (2), 926–935. 10.1063/1.445869. [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

ja4c13462_si_001.pdf (31.3MB, pdf)

Articles from Journal of the American Chemical Society are provided here courtesy of American Chemical Society

RESOURCES