Snapshots of urea‐induced early structural changes and unfolding of an ankyrin repeat protein at atomic resolution

Abstract

Protein folding and unfolding is a complex process, underscored by the many proteotoxic diseases associated with misfolded proteins. Mapping pathways from a native structure to an unfolded protein or vice versa, identifying the intermediates, and defining the role of sequence and structure en route remain outstanding problems in the field. It is even more challenging to capture the events at atomistic resolution. X‐ray diffraction has so far been used to understand how urea interacts with and unfolds two stable globular proteins. Here, we present the case study on PSMD10^Gankyrin, a prototype for a moderately stable, non‐globular repeat protein, long and rigid, with its termini located at either end.   We define structural changes in the time dimension using low urea concentrations to arrive at the following conclusions. (a) Unfolding is rapidly initiated at the C‐terminus, slowly at the N‐terminus, and proceeds inwards from both ends. (b) C‐terminus undergoes an α to 3₁₀ helix transition, representing the structure of a potential early unfolding intermediate before disorder sets in. (c) Distinct and progressive changes in the electrostatic landscape of PSMD10^Gankyrin were observed, indicative of conformational changes in the seemingly inflexible motif involved in protein–protein interaction. We believe this unique study will open up the field for better and bolder queries and increase the choice of model proteins for a better understanding of the challenging problems of protein folding, protein interactions, protein degradation, and diseases associated with misfolding.

Short abstract

PDB Code(s): 7VXV, 7VXW, 7VY4 and 7VY7;

1. INTRODUCTION

Proteins maintain a native fold due to the intricate balance between intra‐protein, inter‐subunit interactions, and their interactions with the environment. Multiple factors can disrupt this delicate balance leading to protein unfolding. ¹ Premature unfolding or incomplete folding can result in protein aggregation, which is responsible for a group of diseases called protein misfolding diseases. ² Using in vitro steady state and kinetic experiments and, of late, in cellular biophysical studies, substantial progress has been made in the understanding of the role of sequence and structure in the formation of protein aggregates and on and off‐pathway intermediates. All of these can impact the protein's final yield, stability, and functions. It is also clear from these studies that the folding and unfolding pathways traced by proteins obey some common rules but cannot be described by a “one size fit all” model. Besides these relatively large‐scale changes, there are subtler and transient changes in the structure of proteins that are important for several other functions, such as protein degradation, protein translocation, post‐translational modification, and protein interactions. Multiple techniques, such as stop‐flow assays, NMR, hydrogen–deuterium exchange experiments, and molecular dynamics simulations, have been used to characterize such partial, local, or early events of structural transitions and unfolding. ³ , ⁴ , ⁵ In these experiments, chaotropic agents such as urea and guanidinium chloride (GnHCl), changes in temperature, or mechanical stress have been used to study reversible protein folding. Lower concentrations of denaturants or minor changes in temperature can induce mild perturbations in the structure without inducing large‐scale changes in the protein. ¹ , ⁶ Hydrogen–deuterium exchange experiments using NMR or mass spectrometry have evolved to capture these dynamic and early structural changes in the proteins at high resolution.

Protein unfolding by X‐ray crystallography became a reality in 1968, when crosslinked monoclinic lysozyme crystals were reversibly denatured and refolded in situ (inside the crystals). ⁷ Instead of soaking preformed crystals in urea, in 1974, Snape et.al., co‐crystallized lysozyme with urea. Urea causes extensive structural alterations in lysozyme without loss of crystallinity. ⁸ In 1977, for the first time, Yonath et al. captured two renaturation intermediates using the SDS‐denatured lysozyme crystals. They reported the presence of SDS molecules in these crystal structures. ⁹ In 2016, Raskar et al., diffracted crystals of hen egg white lysozyme (HEWL) soaked in urea at different time points. They captured various denaturation intermediates where urea had replaced water molecules. By combining diffraction data with molecular dynamics simulations, ¹⁰ they showed both direct (perturbation of hydrophobic interaction) and indirect (disruption of the water network) mechanisms of urea‐induced protein denaturation. They also investigated the mechanism of HEWL unfolding by GnHCl. They showed that the water network around the protein was disrupted by GnHCl, which interacts with hydrophobic residues, thereby increasing the solubility of the unfolded proteins. ¹¹

In contrast to globular proteins such as the lysozyme and RNase, the mechanism of the unfolding of repeat proteins remains largely unexplored. Repeat proteins comprise a tandem array of repetitive structural motifs of 30–50 amino acid residues. This class of proteins forms an elongated structure and is stabilized by inter‐repeat interactions such as the hexapeptide repeats, leucine‐rich repeat proteins, and ankyrin‐repeat‐containing proteins. ¹² Here, we use PSMD10^Gankyrin an ankyrin repeat protein, as a model system to study early events in protein unfolding in the crystalline form. PSMD10^Gankyrin functions as a molecular chaperone during proteasome assembly and is a well‐known oncoprotein. PSMD10^Gankyrin is a bean‐shaped, non‐globular protein formed by seven repeat motifs called the ankyrin repeats (Figure 1a). These repeats help interact with multiple proteins, a feature important for the oncogenicity of PSMD10^Gankyrin. ¹³ , ¹⁴ , ¹⁵ , ¹⁶ More importantly, unlike the globular proteins used as model systems for studying unfolding by X‐ray diffraction, the N and C termini of PSMD10^Gankyrin are spatially isolated (like other repeat proteins) and located at either end of the bean‐shaped molecule. This spatial isolation allows one to look at the origin of unfolding with respect to the protein termini. Taking advantage of this unique structural arrangement, using time‐lapsed crystallography, we report on how structural transitions/unfolding is initiated at the termini, follow the direction of unfolding, and identify a stable core structure. We can also visualize an early unfolding intermediate formed due to changes in the secondary structure at the C‐terminal region of PSMD10^Gankyrin. Because such elaborate details at low urea concentrations, where the majority of the structure was intact, were available, we speculated on the relevance of these findings in the context of PSMD10^Gankyrin structure, stability, and function.

FIGURE 1.

(a) Ankyrin repeat architecture of PSMD10^Gankyrin. Chemical denaturation of PSMD10^Gankyrin by urea. (b) Intrinsic tryptophan fluorescence of PSMD10^Gankyrin as a function of urea concentration obtained using the nanoDSF instrument was converted to a fraction unfolded. Blue line indicates the urea concentration (3.4 M urea) at which 50% of the protein is unfolded. Red line indicates the concentration of urea used in the experiments where PSMD10^Gankyrin crystals were soaked in urea; at this concentration, ~30% of the protein is unfolded. (c) Crystal structures of PSMD10^Gankyrin soaked in urea were aligned to the 0 hr structure and represented as cartoons (0 hr––blue, 1 hr––green, 2 hr––orange, and 3 hr––red). A few residues from N‐terminal, C‐terminal, and middle ankyrin repeats are represented as sticks. The boxes highlight the changes to sidechain conformations in the N‐ and C‐terminal regions. (d) RMSD of all the structures calculated with 0 hr as reference

2. RESULTS

2.1. Urea‐soaked PSMD10^Gankyrin crystals are stable at low urea concentrations

We first checked the denaturation profile of purified PSMD10^Gankyrin in urea by monitoring its intrinsic tryptophan fluorescence (nanoDSF; ChemPR). Upon incubation with ~3.5 M of urea for 15 min, 50% of PSMD10^Gankyrin is denatured (Figure 1b), and this value did not change after 1 hr of incubation (Figure S1). The protein seems to follow a simple two‐state transition under these conditions. The stability of PSMD10^Gankyrin crystals upon treatment with urea is unknown. Therefore, an initial screen was performed to determine the integrity of PSMD10^Gankyrin crystals in urea. Crystals were incubated in 9, 6, and 4.5 M as well as 3 M urea. At 3M urea, ~30% of the protein in solution is unfolded (Figure 1b). These crystals were visually inspected for up to 12 hr. The crystals disintegrated when soaked for longer than 12 hr in 3 M urea or for 1 hr at urea concentrations above 3 M. Therefore, 3 M urea was used to study early events of structural changes/unfolding of PSMD10^Gankyrin in its crystalline form.

Data collected from “native or 0 hr”, “1 hr”, “2 hr,” and “3 hr” urea‐incubated PSMD10^Gankyrin crystals were used to solve the structures (PDB IDs in Table 1). The refinement statistics for all four crystal structures are presented in Table 1. The four crystals are isomorphous, belong to the trigonal space group P3₂21 and diffract X‐rays at a resolution of approximately 2.2 Å. The average B‐factor for the protein gradually increases from 0 to 3 hr of incubation in urea, indicating that these crystal structures represent the early events of protein unfolding. Crystals incubated for 4 hr or up to 12 hr in 3 M urea remained intact but did not diffract (data not shown), and therefore data could not be collected for these crystals.

TABLE 1.

Data collection and refinement statistics

	Gankyrin 0 hr PDB ID – 7VXV	Gankyrin 1 hr PDB ID – 7VXW	Gankyrin 2 hr PDB ID – 7VY4	Gankyrin 3 hr PDB ID – 7VY7
Wavelength	1.54 Å	1.54 Å	1.54 Å	1.54 Å
Resolution range	47.82–2.23 (2.31–2.23)	52.13–2.224 (2.303–2.224)	48.03–2.224 (2.303–2.224)	47.93–2.23 (2.31–2.23)
Space group	P 32 2 1	P 32 2 1	P 32 2 1	P 32 2 1
Unit cell	59.9899 59.9899 122.31 90 90,120	60.1898 60.1898 122.29 90 90,120	60.2998 60.2998 122.31 90 90,120	60.1398 60.1398 122.43 90 90,120
Total reflections	25,250 (2,425)	25,598 (2,366)	26,200 (2,435)	26,058 (2,510)
Unique reflections	12,720 (1,225)	12,878 (1,203)	13,187 (1,248)	13,038 (1,263)
Multiplicity	2.0 (2.0)	2.0 (2.0)	2.0 (2.0)	2.0 (2.0)
Completeness (%)	97.98 (96.31)	97.71 (93.11)	99.67 (96.97)	99.95 (99.76)
Mean I/sigma (I)	16.06 (4.00)	7.71 (1.27)	7.52 (1.20)	7.40 (1.43)
Wilson B‐factor	24.12	34.10	34.40	32.36
R‐merge	0.04393 (0.1963)	0.09417 (0.5859)	0.09629 (0.6649)	0.09256 (0.5095)
R‐meas	0.06213 (0.2776)	0.1332 (0.8286)	0.1362 (0.9404)	0.1309 (0.7206)
R‐pim	0.04393 (0.1963)	0.09417 (0.5859)	0.09629 (0.6649)	0.09256 (0.5095)
CC1/2	0.998 (0.871)	0.993 (0.132)	0.992 (0.136)	0.993 (0.204)
CC	0.999 (0.965)	0.998 (0.484)	0.998 (0.489)	0.998 (0.583)
Reflections used in refinement	12,718 (1,225)	12,878 (1,203)	13,184 (1,248)	13,036 (1,263)
Reflections used for R‐free	1,264 (122)	1,291 (121)	1,308 (123)	1,306 (129)
R‐work	0.1999 (0.2353)	0.2475 (0.3921)	0.2405 (0.3972)	0.2431 (0.4130)
R‐free	0.2421 (0.3254)	0.3218 (0.4547)	0.3028 (0.3728)	0.3126 (0.4402)
CC (work)	0.947 (0.832)	0.948 (0.258)	0.944 (0.231)	0.940 (0.225)
CC (free)	0.926 (0.683)	0.840 (0.151)	0.881 (0.210)	0.895 (0.231)
Number of non‐hydrogen atoms	1,742	1,739	1,751	1,747
Macromolecules	1,694	1,693	1,697	1,693
Ligands	0	12	20	12
Solvent	48	34	34	42
Protein residues	224	224	224	224
RMS (bonds)	0.019	0.024	0.029	0.020
RMS (angles)	1.88	2.33	2.08	1.75
Ramachandran favored (%)	96.85	91.44	90.99	90.54
Ramachandran allowed (%)	3.15	7.66	7.66	7.21
Ramachandran outliers (%)	0.00	0.90	1.35	2.25
Rotamer outliers (%)	0.56	4.49	6.70	8.43
Clashscore	3.24	9.67	10.77	19.34
Average B‐factor	25.97	46.24	43.77	44.11
Macromolecules	26.10	46.40	43.88	44.47
Ligands		67.36	54.92	39.60
Solvent	21.39	30.85	31.57	31.24

Note: Statistics for the highest‐resolution shell are shown in parentheses.

2.2. Urea‐induced structural changes in PSMD10^Gankyrin in its crystalline state

To monitor the changes in the crystalline form of the protein upon incubation with urea, we aligned the data obtained from the urea‐exposed crystal structures to the native structure of PSMD10^Gankyrin. ¹⁷ As seen from the RMSD values, PSMD10^Gankyrin undergoes global structural changes as a function of time. The sidechain conformations of most of the residues undergo a small but detectable change. In contrast, those of residues at the N‐ and C‐terminal regions undergo significant changes (Figure 1c). Using the 0 hr structure as the reference, we calculated the RMSD values for Cα atoms and the entire residues in 1, 2, and 3 hr crystal structures. The Cα RMSD and residue‐RMSD values indicate that the C‐terminal region is the most labile region in the protein (Figures 1d and S2a). Furthermore, with the increase in time, there is a rapid loss of sidechain electron density at the C‐terminal region, while such loss is more gradual at the N‐terminal region (Figure 2a–h).

FIGURE 2.

Initial conformational changes in PSMD10^Gankyrin due to urea‐induced denaturation are observed at the C‐terminal region. (a) Plot depicting average B‐factor versus residue number indicates the dynamic nature of residues at the termini. (b) Representation of the secondary structural changes (order to disorder transition) observed at the C‐terminal region upon incubation with urea for 0–3 hr. Ramachandran distance was calculated between (c) 0 and 1 hr, (d) 0 and 2 hr. (e) Ramachandran distance of the C‐terminus region (201–226) of the 1 and 2 hr structures. (f) Cartoon representation of the time‐dependent unfolding at the N‐ and C‐termini of PSMD10^Gankyrin

2.3. The C‐terminal region of PSMD10^Gankyrin is flexible and dynamic

Since the N‐ and C‐terminal regions of the protein showed substantial changes in the RMSD values, we monitored changes in the B‐factor of individual residues. B‐factor or temperature factor is the atomic displacement parameter, which describes the vibration of an atom around its mean position specified by the atomic coordinates. Well‐ordered atoms tend to have lower B‐factor values, and therefore atoms with comparatively higher B‐factor values are considered dynamic. ¹⁸ B‐factor values of all four structures (0, 1, 2, and 3 hr) were compared using the structure comparison module in Phenix. Consistent with the changes in RMSD values, residues 1–40 (N‐terminal region) and 160–226 (C‐terminal region) show significant changes in their B‐factor values (main chain, side chain, and overall). The results indicate that the terminal regions of the protein undergo maximum changes upon incubation with urea. Strikingly, even as early as 1 hr of incubation in urea, changes in the B‐factor values were evident in the C‐terminal region of PSMD10^Gankyrin, implying that structural changes leading to protein unfolding originate here. The next set of changes, at 2 and 3 hr of incubation in urea, are seen at the farther end of the protein, namely, the N‐terminal region (Figures 3a and S2b,c).

FIGURE 3.

Progressive loss of electron density in urea‐incubated PSMD10^Gankyrin crystal structures. (a, c, e, and g) represent the N‐terminal density of 0, 1, 2, and 3 hr urea incubated crystal structures of PSMD10^Gankyrin, and (b, d, f, and h) represent the C‐terminal density of 0, 1, 2, and 3 hr urea incubated crystal structures of PSMD10^Gankyrin

Structural changes in different conformations of the same protein can also be measured by the “Ramachandran distance.” Although not widely used, this is a useful parameter to follow secondary structural transitions. Ramachandran distance is the difference in the ɸ and ψ angles of a residue in two different conformations of the protein. Since the ɸ and ψ are different for different secondary structural elements, one can trace the conformational changes by calculating the Ramachandran distance. With the 0 hr structure acting as the reference point, we calculated the Ramachandran distance ¹⁹ of all the residues in the crystal structures of PSMD10^Gankyrin obtained after 1 or 2 hr of incubation in urea. We observed significant changes in the region between 200 and 224 residues at the C‐terminal segment of the protein. At 1 hr, the region undergoes a distinct structural transition from an α helix to a 3₁₀ helix, and the ɸ and ψ values change for all the residues in this segment (200–224). At 2 hr, the structure in this region completely melts and becomes disordered (Figure 3b–f). We did not see such contiguous changes in any other part of the structure.

One of the expected changes, when such structural destabilization occurs in the presence of urea, is the change in electrostatics as the intra and interchain bonds are broken, and the surface exposure of residues is altered. To map these changes, we used the Adaptive Poisson–Boltzmann Solver plugin from PyMOL to calculate and visualize the electrostatic charge distribution. ²⁰ We observed changes in the electrostatic landscape, and among them two were prominent and interesting. One, the pocket formed by R41 and K116, which is predominantly positively charged and binds to negatively charged residues of the interacting partner, gradually became less positive during the 3 hr urea incubation. Two, the pocket formed by residues LACDE (178–182), which is predominantly negatively charged and is a conserved motif known for interaction with the Retinoblastoma protein, gradually becomes more negative during the 3 hr urea incubation (Figure 4).

FIGURE 4.

Change in the electrostatic landscape of PSMD10^Gankyrin upon treatment with urea. (a) Surface electrostatics of PSMD10^Gankyrin‐S6C structure (PDB ID‐2DVW) (without the S6C chain). I (R41), II (K116), and III (178LACDE182) indicate the regions important for the interaction with S6C. (b) Comparative representation of surface electrostatics for 0, 1, 2, and 3 hr structures. The blue and red arrows indicate the regions undergoing changes to the surface electrostatics

2.4. Mechanism of urea‐induced local structural changes in PSMD10^Gankyrin

To further confirm the order of structural changes and understand the mechanism by which urea induces such changes, we first identified the number of urea molecules and their position in the structure. There are 3, 5, and 3 urea molecules in the PSMD10^Gankyrin crystal structures obtained after 1, 2, and 3 hr incubation, respectively. The interactions of urea molecules with protein residues for all the structures were extracted using NCONT in the CCP4 package. High B‐factor of the region and loss of electron density are the probable reasons why no urea molecules were found in the C‐terminal region (Ank7) (Figures 3a, S2b,c, 2b,d,f,h, and S3). For those urea molecules for which electron density is observed (Figure S3), we describe their interaction with residues in PSMD10^Gankyrin (Figure 5 and supplementary document 1). For each time point, the urea molecules detected are numbered as Ure1, Ure2, and so forth.

FIGURE 5.

Presence and interaction of urea with PSMD10^Gankyrin residues in the crystal. Representation of interactions between urea (0 hr––cyan and 1–3 hr––green) and (a–c) 1 hr, (d–f) 2 hr, and (g–i) 3 hr urea incubated PSMD10^Gankyrin crystal structures. (j) A “topological” map of urea interactions

Analysis of the interactions made by urea with the protein indicates that most are hydrophobic. These are further supported by hydrogen bonds formed with adjacent polar residues. U1 (U = Ure), in the 1 hr structure (Ank5), makes a hydrogen bond with the backbone oxygen of Y160. The ‐NH of U1 also makes an NH–π interaction with Y160 (Figure 5a). U2 of the 1 hr structure (Ank4) makes a hydrophobic contact with the Cß of H119 and hydrogen bonds with N117, R118, and E120 (Figure 5b). U3 in the 1 hr (Ank1 and Ank2) makes a hydrophobic contact with the Cß of Q38 and a hydrogen bond with the Nδ of N7 (Figure 5c).

At 2 hr, U1 is at the same site as at 1 hr (Ank5), interacting with the same residues. U5 of 2 hr (Ank5 and Ank6) makes hydrophobic contact with the sidechain of L159 and hydrogen bonds with the backbone oxygen of Q194 and A163 (Figure 5d). U4 of 2 hr (Ank1) is present near the N‐terminus and makes hydrophobic contacts with Sγ of C4 and a hydrogen bond with Nδ of N12. U2 of 2 hr (Ank1 and Ank2) interacts with W46 through a hydrophobic contact with Cζ3 of the W46 sidechain, OH of Y15 and S49 are present in the first and second ankyrin repeats (Figure 5e). U3 of 2 hr (Ank2) makes hydrophobic contact with Cα of P65 and hydrogen bonds with the backbone nitrogen of P65 and the backbone oxygen of G63 (Figure 5f). U2, U3, and U4 of 2 hr interact predominantly with the N‐terminal region (Ank1 and Ank2) of the protein, correlating with the increase in the B‐factor at the N‐terminal region at 1 and 2 hr.

U3 of 3 hr and U5 of 2 hr (Ank5 and Ank6) are present at the same site and interact with the identical residues (Figure 5d,g), while U2 in 3 hr (Ank1) is present near the N‐terminus and makes a hydrophobic bond with Cδ1 of L8 and hydrogen bonds with backbone oxygen and Oδ of N7 (Figure 5i). Finally, U4 of 3 hr (Ank2 and Ank3) interacts with G51, T53, G84, R85, D86, E87, and I88 along with U2 and U3 (Figure 5h). These interactions result in progressive changes in structure that are initiated at the C‐terminal region, followed by changes in the N‐terminal region, ending up in the structural changes at the second and third ankyrin repeats.

These interactions between urea and residues in PSMD10^Gankyrin are summarized in the Figure 5j. Interestingly, although urea molecules can be found in the core region of the protein at all time points, they do not cause any structural changes in this region as observed by the B‐factor or the RMSD values. These results support the hypothesis that the core must undergo structural changes after the C‐ and N‐terminal regions unfold. This would eventually lead to crystals that cannot diffract and are not amenable for structural characterization.

To check if crystal contacts play a role in the proposed order of unfolding (structural changes in the C‐terminal region, followed by the N and finally the core), we mapped the loss/gain of crystal contacts in all four crystal structures. These changes are represented in the form of a heatmap (Figure S4 and Supplementary document 2). The analyses show that the symmetry‐related crystal contacts are present over the entire molecule: at the N‐terminal region, the middle region, and the C‐terminal region of the protein. Except for a small region at Y160, the rest of the contacts are uniformly weak. The N‐terminal region has relatively stronger crystal contacts than the core region of the protein, and yet, the repeats in the N‐terminal region undergo substantial structural changes early during incubation compared to the core region (as seen from RMSD and B‐factor values in Figures 1d and 3a). Therefore, taken together with the fact that we observe density for urea in all ankyrin repeats, indicating its accessibility, we conclude that there is no particular bias enforced by crystal contacts that directs unfolding. Further changes were observed in the C‐terminal region of PSMD10^Gankyrin crystal structure are specific to urea, since we do not observe such changes when crystals are incubated for 24 hr with doxorubicin, a drug that binds to PSMD10^Gankyrin (Figure S5). ²¹

3. DISCUSSION

Urea denatures proteins, both by direct and indirect mechanisms, by weakening the hydrophobicity of the protein and increasing its solubility. The direct mechanism involves interaction between non‐polar amino acids and urea, whereas the indirect mechanism involves alteration in the water network around hydrophobic amino acids. ²² , ²³ , ²⁴ We do not have enough resolution in the crystallographic data to obtain a water map and, therefore cannot comment on the mechanism involving urea‐induced indirect changes in the water network. However, we can see density for urea molecules at various ankyrin repeats and, therefore can deduce a mechanism that involves direct interaction of urea with residues in the protein.

The binding of urea occurs in regions of high aliphatic index and hydrophobic residues. ²⁵ , ²⁶ The C‐terminal hydrophobic region in PSMD10^Gankyrin has the highest aliphatic index (117) and is also rich in glycine, an amino acid although not aliphatic, is known to be one of the highly hydrophobic residues (Figure 3a). ¹⁸ This is followed by the N‐terminal region (aliphatic index 96) and the core region (aliphatic index 89) of the protein. Thus, the initiation of structural transitions in the C‐terminal region of PSMD10^Gankyrin follows the well‐known principles by which urea interacts with and unfolds the protein.

Although urea molecules can be found at the inner ankyrin repeats as early as 1 hr of incubation, no substantial changes in structure are seen. In the core region of the protein, a urea molecule was found to interact with Y160 and Y161 through an NH–π interaction, with the nitrogen of urea perpendicular to the aromatic ring of Y160 at an approximate distance of ~4.5 Å. ²⁷ The NH–π interaction is supported by hydrogen bonds with adjacent atoms. The presence of polar residues around a hydrophobic residue presumably acts like a harness or yoke to hold urea within the polar environment. At the same time, this allows the weakening of the hydrophobic interactions by urea rendering the aliphatic region more polar and thus promoting protein unfolding by increasing the aqueous solubility of the more open structure.

Prior to our investigations, PSMD10^Gankyrin was used as a model system to study unfolding of repeat proteins using classical techniques, such as molecular dynamic simulations (MD), stop‐flow experiments, and atomic force microscopy. ²⁸ , ²⁹ , ³⁰ Serquera et al., in 2010, demonstrated that mechanical stress initiated unfolding from the C‐terminus of PSMD10^Gankyrin by molecular simulations. ³⁰ Hutton et al., in 2015, using MD and simulation‐guided mutagenesis, showed that PSMD10^Gankyrin unfolding follows two different pathways depending on the urea concentration and the incubation time. They observed that the N‐ and C‐termini undergo a “contraction and relaxation cycle” before unfolding, with the C‐terminal region more prone to such oscillatory motion. ²⁸

With these mechanistic insights in place and a fair understanding of the structural changes at atomistic details during PSMD10^Gankyrin (protein) unfolding, we asked whether we can draw better insights and speculate on the role of these conformational changes in the functions or the fate of the protein. For example, PSMD10^Gankyrin is a protein involved in multiple oncogenic interactions in many cancers. ¹³ , ¹⁵ , ³¹ , ³² The ankyrin repeats, and the loops are said to be involved in these interactions. Often the entire protein along its length seems to be important for some of the interactions, and in some cases, very distinct residues within a hot spot region provide the bulk of the binding energy. There are two such regions/pockets on PSMD10^Gankyrin. One is the basic pocket made of R41 and K116 involved in interactions with the negatively charged amino acid residues EEVD in S6ATPase of the proteasome and presumably other proteins that carry the motif EEVD like that of the S6ATPase. ¹³ , ³³ Because of the relatively lower flexibility observed at the N‐terminal region of the PSMD10^Gankyrin pocket formed by R41 and K116, binding at this site may involve initial docking via surface and charge complementarity. Such docking may then  reinforce interactions involving other ankyrin repeats. The second interacting region is the negatively charged, conserved LACDE (178–182) motif that binds to the Retinoblastoma protein. This region is located within the penultimate ankyrin repeat (Ank6) in the C‐terminal region and is adjacent to the inherently dynamic and most labile C‐terminal Ank repeat (Ank7) within, which undergoes the earliest structural transition in the presence of urea. We hypothesize that the dynamic C‐terminal region will help accommodate the incoming Rb protein and/or help in dissociation. A similar but more dramatic mechanism were previously observed with IκBα, an ankyrin repeat protein with an unstructured C‐terminal region or intrinsically disordered region. IκBα undergoes a large‐scale structural change to a folded structure upon binding to NFκB. ³⁴

There may be other roles for the fragile C‐terminal region of PSMD10^Gankyrin. The early unfolding intermediate formed by the α helix to 3₁₀ helix transition may represent the origin of the degradation signal required for proteasomal degradation. This speculation is based on the following observations in the literature. Besides the polyubiquitin chain that binds to receptors on the proteasome, a long stretch of disordered region (30–40 aa long) that can reach the active site residues within the channel of the 20 S proteolytic chamber is mandatory for the degradation of proteins. ³⁵ , ³⁶ The origin of such degradation signals can often be rate‐limiting, and how such disordered regions originate in proteins remains an outstanding question in the field. ³⁵ , ³⁷ In one of our early studies, we demonstrated how removal of a ligand can create a segment that is long and disordered/floppy, creating a degron. ³⁸ The local structural changes observed at the C‐terminus of PSMD10^Gankyrin, results in an unfolding intermediate poised to become disordered, can be such a degradation signal. Although speculative, we believe these hypotheses are testable and can provide insights.

Collectively, our data suggests that the process of unfolding is initiated by the structural changes and destabilization of the C‐terminal region as observed at 1 hr (Ank7), followed by increased urea binding at the N‐terminal region at 2 hr leading to its destabilization (Ank1). As is demonstrated by the movement of urea molecules at 3 hr, unfolding progresses toward the relatively stable inner ankyrin repeats (Ank 3, 4, and 5) from both ends. Figure 6 represents such a model of PSMD10^Gankyrin unfolding. Thus, the current study, for the first time provides atomic details of the dynamics of PSMD10^Gankyrin structure during unfolding, speculates on the structure of an unfolding intermediate, and possible implications of these events in protein–protein interaction and degradation.

FIGURE 6.

Model for urea induced structural transition of PSMD10^Gankyrin during its unfolding

4. MATERIALS AND METHODS

4.1. Strains, plasmids and protein purification

Escherichia coli DH5α and Rosetta (2DE3) were used as cloning and expression host, respectively. PSMD10^Gankyrin WT was cloned into a pRSETA plasmid with a 6X HIS tag in the N‐terminus, Ampicillin at 100 μg ml⁻¹ and chloramphenicol at 34 μg ml⁻¹ were used for the selection and maintenance of transformed cells. ¹³ PSMD10^Gankyrin WT was expressed and purified as described earlier. ³⁹

4.2. Chemical denaturation

A series of dilutions of urea (concentration range of 0–7 M) were prepared in protein buffer (50 mM Tris pH 7.5, 150 mM NaCl). Protein was incubated with urea for 15 min, then loaded onto a capillary and placed in the loading tray of Prometheus NT.48. Intrinsic tryptophan fluorescence of PSMD10^Gankyrin was collected using PR.ChemControl software of Nanotemper Technologies to monitor chemical denaturation. We calculated the slopes of the pre‐transition and post‐transition data points and used the equations as described earlier to determine fraction unfolded. ⁴⁰ The data points were further fitted using four parametric equations in Graphpad V8.0.

4.3. Crystallization

PSMD10^Gankyrin WT crystallization was setup in‐hand using Art Robbins 48‐well Intelliplate (3 drop – sitting drop). A total of 20 mg ml⁻¹ of PSMD10^Gankyrin WT protein was used to setup crystal trial (1.5 μl protein +1 μl buffer +1 μl seed) in 60% Tacsimate pH 7.0 (Hampton Research) at 22°C. ³⁹

Time of exposure and the concentration of urea to obtain intact diffraction quality crystals were determined by soaking PSMD10^Gankyrin WT Crystals in different concentrations of urea and visually observing them at different time intervals. The crystals were cryoprotected using mother liquor mixed with 25% glycerol for PSMD10^Gankyrin WT and 25% glycerol plus 3 M urea for the urea‐soaked crystal.

4.4. Data collection and crystal solution

A single wavelength (λ = 1.54 Å) data set for urea‐soaked crystals were collected at home source––Microstar microfocus rotating anode X‐ray (Bruker) Cu anode with MAR detector. Data were processed using iMosflm and Scala in CCP4i suite. ⁴¹ , ⁴²

The structure was determined using Molecular Replacement with Phaser‐MR in Phenix suite, ⁴³ PDB entry 1UOH. (Since our high resolution crystal structure was solved in tandem with this study we used 1UOH rather than 7VO6 as the search model.) Refinement was performed using Phenix.refine ⁴³ and model building was done using the program Coot. ⁴⁴ Omit map was made using CCP4. ⁴⁵ , ⁴⁶ The models for representation were made using PyMOL and CCP4MG. ¹⁷ , ⁴⁷

4.5. Ramachandran distance

Ramchandran distance was calculated using the formula provided in Reference 19. The following equation was used to calculate Ramachandran distance (D):

$$D=\sqrt{{\left({\varphi }_0-{\varphi }_n\right)}^2+{\left({\psi }_0-{\psi }_n\right)}^2}$$

the ϕ₀ and ψ₀ angles of 0 hr structure, ϕ _n and ψ _n angles from each time point individually.

AUTHOR CONTRIBUTIONS

M G Mukund Sudharsan: Conceptualization (supporting); data curation (equal); formal analysis (equal); investigation (equal); methodology (equal); software (equal); validation (equal); visualization (equal); writing – original draft (equal); writing – review and editing (supporting). Somavally Dalvi: Conceptualization (supporting); data curation (equal); formal analysis (equal); investigation (equal); methodology (equal); software (equal); validation (equal); visualization (equal); writing – original draft (equal). Prasanna Venkatraman: Conceptualization (lead); data curation (equal); funding acquisition (lead); investigation (lead); methodology (equal); project administration (lead); resources (lead); software (supporting); supervision (lead); validation (supporting); visualization (supporting); writing – original draft (supporting); writing – review and editing (lead).

CONFLICT OF INTEREST

Authors have no conflict of interest.

Supporting information

SUPPLEMENTARY DOCUMENT 1 is an excel file with the list of all the possible interactions (distance <5 Å) between urea and PSMD10^Gankyrin in 1, 2, and 3 hr crystal structures as determined using NCONT (CCP4).

SUPPLEMENTARY DOCUMENT 2 is an excel file with the list of all possible symmetry related crystal contacts (distance <4 Å) in 1, 2, and 3 hr crystal structures as determined using Contact (CCP4).

FIGURE S1 Denaturation profile of PSMD10^Gankyrin in urea. Data are plotted for PSMD10^Gankyrin incubated with urea for 15, 30, 45, and 60 min. Fraction unfolded is plotted in the Y‐axis, and the concentration of urea is plotted in the X‐axis.

FIGURE S2 (a) C‐alpha RMSD values calculated with 0 hr as reference. Plot depicting B‐factor versus residue number (b) main chain and (c) side chain indicate the dynamic nature of residues at the termini.

FIGURE S3 OMIT map for urea in PSMD10^Gankyrin crystal structures determined after incubation of crystals in urea for 1 hr (a–c), 2 hr (b–f), and 3 hr (g–h).

FIGURE S4 Heat map depicting the change in symmetry related crystal contacts in PSMD10^Gankyrin crystal structures determined after incubation of crystals in urea for 0, 1, 2, and 3 hr.

FIGURE S5 Crystal structure of doxorubicin bound PSMD10^Gankyrin. Plot depicting RMSD values between doxorubicin bound PSMD10^Gankyrin crystal structure (obtained after 24 hr of soaking) and native PSMD10^Gankyrin structure. There is no significant difference between the two structures unlike the structures obtained by soaking crystals in urea.

Medur Gurushankar MS, Dalvi S, Venkatraman P. Snapshots of urea‐induced early structural changes and unfolding of an ankyrin repeat protein at atomic resolution. Protein Science. 2022;31(12):e4515. 10.1002/pro.4515

DATA AVAILABILITY STATEMENT

Data supporting these findings will be made available in the PDB database. However, we have attached the PDB files for the benefit of the reviewers.