Conceived and designed the experiments: AS, TH, JR, ND. Performed the experiments: AS, TH, JR, ND. Analyzed the data: AS, TH, ND. Contributed reagents/materials/analysis tools: JR. Wrote the paper: AS, TH, JR, ND.
The authors have declared that no competing interests exist.
The absence of a functional ATP Binding Cassette (ABC) protein called the Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) from apical membranes of epithelial cells is responsible for cystic fibrosis (CF). Over 90% of CF patients carry at least one mutant allele with deletion of phenylalanine at position 508 located in the N-terminal nucleotide binding domain (NBD1). Biochemical and cell biological studies show that the ΔF508 mutant exhibits inefficient biosynthetic maturation and susceptibility to degradation probably due to misfolding of NBD1 and the resultant misassembly of other domains. However, little is known about the direct effect of the Phe508 deletion on the NBD1 folding, which is essential for rational design strategies of cystic fibrosis treatment. Here we show that the deletion of Phe508 alters the folding dynamics and kinetics of NBD1, thus possibly affecting the assembly of the complete CFTR. Using molecular dynamics simulations, we find that meta-stable intermediate states appearing on wild type and mutant folding pathways are populated differently and that their kinetic accessibilities are distinct. The structural basis of the increased misfolding propensity of the ΔF508 NBD1 mutant is the perturbation of interactions in residue pairs Q493/P574 and F575/F578 found in loop S7-H6. As a proof-of-principle that the S7-H6 loop conformation can modulate the folding kinetics of NBD1, we virtually design rescue mutations in the identified critical interactions to force the S7-H6 loop into the wild type conformation. Two redesigned NBD1-ΔF508 variants exhibited significantly higher folding probabilities than the original NBD1-ΔF508, thereby partially rescuing folding ability of the NBD1-ΔF508 mutant. We propose that these observed defects in folding kinetics of mutant NBD1 may also be modulated by structures separate from the 508 site. The identified structural determinants of increased misfolding propensity of NBD1-ΔF508 are essential information in correcting this pathogenic mutant.
Deletion of a single residue, phenylalanine at position 508, in the first nucleotide binding domain (NBD1) of the Cystic Fibrosis Transmembrane Conductance Regulator (CFTR) is present in approximately 90% of cystic fibrosis (CF) patients. Experiments show that this mutant protein exhibits inefficient biosynthetic maturation and susceptibility to degradation probably due to misfolding of NBD1 and the resultant incorrect interactions of other domains. However, little is known about the direct effect of the Phe508 deletion on NBD1 folding. Here, using molecular dynamics simulations of NBD1-WT, NBD1-F508A, and NBD1-ΔF508, we show that the deletion of Phe508 indeed alters the kinetics of NBD1 folding. We also find that the intermediate states appearing on wild type and mutant folding pathways are populated differently and that their kinetic accessibilities are distinct. Moreover, we identified critical interactions not necessarily localized near position 508, such as Q493/P574 and F575/F587, to be significant structural elements influencing the kinetic difference between wild type and mutant NBD1. We propose that these observed alterations in folding kinetics of mutant NBD1 result in misassembly of the whole multi-domain protein, thereby causing its premature degradation.
CF is the most common autosomal inherited disease with high morbidity among Caucasians. CF patients have altered epithelial ion transport that leads to decreased hydration of epithelial surfaces in the gut, kidney, pancreas, and airways
(A) CFTR protein consists of nucleotide-binding domains (NBD1 and NBD2), transmembrane domains (TMD1 and TMD2), and a regulatory domain (R domain). Deletion of F508 in NBD1 occurs in ∼90% of CF patients. (B) Superimposed NBD1-WT (PDB ID: 2BBO
Although more than 1,400 mutations are known in CFTR (
Nucleotide-binding domains of ABC proteins are highly conserved in sequence and structure. NBDs contain a typical F1 ATPase core subdomain, which consists of an α-helix surrounded by antiparallel β-sheets
Folding of multidomain proteins is aided by molecular chaperones to prevent and correct improper (non-native) associations between solvent-exposed hydrophobic regions. Smaller single-domain proteins correct and prevent formation of improper contacts through a sequence of partial folding-unfolding events en route to the native state. This sequence of partial folding-unfolding events reflects the ability of single-domain proteins to self-chaperone their folding. In NBD1, the attenuated refolding of the recombinant ΔF508 mutant is consistent with the notion that Phe508 reduces the activation energy of NBD1 folding
To investigate the effect of the Phe508 deletion on the stability, dynamics and kinetics of NBD1, we performed equilibrium dynamics simulations and folding simulations of NBD1-WT and NBD1-ΔF508. Our analysis shows that there is no significant difference in their stability and equilibrium dynamics, which agrees with experiments. However, even in the presence of correcting mutants (G550E, R553Q, and R555K)
To determine the equilibrium dynamics and stabilities of the wild type and mutant NBD1, we perform equilibrium simulations (106 time units∼0.5 millisecond
Energy is calculated from long equilibrium simulations (106 time units) of NBD1-WT, NBD1-F508A, and NBD1-ΔF508 crystal structures. Error bars represent±standard deviation. (Inset) The specific heat is calculated as
The slope at the transition temperature of the wild type (
Folding is a stochastic process, thus to investigate in detail the difference in folding kinetics and dynamics of NBD1-WT and NBD1-ΔF508, we perform 300 folding simulations on each of the structures. Starting from fully unfolded chains of NBD1-WTand NBD1-ΔF508, we progressively reduce the temperature of the system to simulate thermal folding (see
To investigate the molecular origin of the difference in folding yields and probabilities, we map the folding pathways of NBD1-WT, NBD1-F508A, and NBD1-ΔF508 by identifying their metastable folding intermediate states. The folding intermediate states of a folding trajectory are exhibited as peaks in the energy probability distributions (
We find that some intermediate states are accessible only by either NBD1-WT (S6 and S9) or NBD1-ΔF508 (S5 and S10), further suggesting that Phe508 deletion leads the mutant to off-folding pathways (see below). While states S2, S3, S4, S7, and S8 are both traversed by NBD1-ΔF508 and NBD1-WT, their time occupancies (length of time NBD1 spends in an intermediate state) are different (
We perform annealing simulations in which the temperature is decreased to facilitate folding of NBD1. Shown is a time evolution of energy starting from unfolded to native (N) states. As the protein proceeds towards it native state, it goes through metastable folding intermediate states (IS), which are observed as peaks in the energy probability distribution.
To determine the difference between the sequence of folding events of the wild type, ΔF508, and the F508A control, we estimate the probability of transitions between intermediate states (see
Transition probabilities between intermediate states are estimated by counting the number of trajectories that exhibit the transition (
We calculate the most dominant folding pathways in wild type and mutant NBD1. The most dominant path in wild type follows a sequence of transition Unfolded→S10→S8→S7→S5→S4→S1, while the dominant path in the mutant follows the sequence of transitions Unfolded→S9→S8→S7→S6→S4→S1. Thus, NBD1-WTand NBD1-ΔF508 undergo different sequences of folding events.
Because of the reduction in dimensionality of the folding process when energy is used as a reaction coordinate, each intermediate state represents an ensemble of NBD1 structures. To identify the primary structural characteristics of each intermediate state, we clustered structures in the corresponding state and calculated the frequency of contacts formed between pairs of residues (
To identify the structure most representative of an intermediate state, we cluster the structures within the folding intermediate. Shown are the centroids of the dominant clusters. Diameter of the backbone cartoons is proportional to the average per residue root-mean-square deviation (RMSD) of the structures within the intermediate state. Blue and red represent N- and C-termini, respectively.
Difference between average contact frequencies of structures within intermediate states shows malformed contacts in NBD1-ΔF508 (green) compared to NBD1-WT (blue). These identified malformed contacts in the mutants are critical determinants of NBD1 folding kinetics. In particular, P574 interacts with Q493 in wild type but not in the mutant. Also, F575 interacts with F587 in mutant but not in wild type. Redesigning these contacts to their wild type interactions in the ΔF508 background can potentially rescue NBD1-ΔF508.
We observe a number of folding trajectories reaching native energies (∼630 ε) and within a 2.5 Å root-mean-square deviation (RMSD) with respect to the native structure, but the resulting topological wiring of the secondary structures is incorrect. The “miswiring” consistently occurs in the H5-S6 loop. Interestingly, this H5-S6 loop is in the immediate neighbourhood of the loop containing Phe508. This suggests “weak” regions in NBD1 that are intrinsically prone to misfolding.
To verify that the identified contact pairs (Q493/P574 and F575/F587) found in the S7-H6 loop are indeed critical in the kinetics of NBD1, we revert their interactions in NBD1-ΔF508 to their interactions in NBD1-WT and perform folding simulations. In the case of the Q493/P574 pair, the residues are in close proximity in NBD1-WT but not in NBD1ΔF508, thus we changed the interaction between Q493 and P574 in NBD1ΔF508 from repulsive to attractive to mimic a possible rescuing mutation. Folding simulations of “rescued” NBD1-ΔF508 yield a folding probability of 19±2%. On the other hand, residues F575 and F508 are in close contact in NBD1-ΔF508 but not in NBD1-WT, thus we reverted their interaction in NBD1-ΔF508 from attractive to repulsive. Folding simulations of the second “rescued” NBD1-ΔF508 yield a folding probability of 20±2%. These folding probabilities of the two “rescued” NBD1-ΔF508s are higher than the 13±2% folding probability of the original NBD1-ΔF508, which supports our findings that the contacts between Q493 and P574 and between F575 and F587 are indeed critical to NBD1 folding.
Deletion of Phe508 in CFTR NBD1 is the most frequent mutation in patients with cystic fibrosis. Proteins with ΔF508 mutation in the first nucleotide binding domain NBD1 can not mature resulting in absence of functional CFTR from the plasma membrane. The molecular mechanism leading to this pathological situation is unknown. No significant difference in thermodynamic stabilities was experimentally observed between wild type and mutant NBD1
Our model agrees with the experimental studies where wild type and mutant NBD1 did not exhibit significant thermodynamic difference. Thermal denaturation curves of both wild type and mutant were calculated from long equilibrium dynamics simulations (∼0.5 millisecond) (
Interestingly, even a simplified protein model (4-bead representation for non-aromatic residues and 5-bead for aromatic residues) and a simplified potential (Go̅-type interaction) show a significant difference in NBD1-WT and NBD1ΔF08 kinetics. Indeed the multiple folding simulations of NBD1 models show significantly higher folding efficiency for the wild type than the ΔF508 mutant, which correlate with some experimental studies that found higher folding yield for wild type NBD1 than its ΔF508 mutant
We also find that the folding time of NBD1ΔF508 mutant is smaller than that of wild type, suggesting an increase in the effective folding rate upon Phe508 deletion. Consistent with the notion that the refolding of recombinant ΔF508 mutant reduces the activation energy of NBD1 folding
What is the origin of this loss in self-chaperoning capacity by the Phe508 mutant? To answer this question, we identify and compare the folding intermediates accessed by wild type and mutant
A drawback that may arise from using Go̅ is that the properties of a protein are determined solely by its geometry, an assumption that apparently deviates from the observation that sequence is also a key determinant of folding properties. However, this potential drawback is not limiting in our study of the folding kinetics of wild type NBD1 and its mutants. The nuanced effect of a mutation or deletion at position 508 is already reflected in the S7-H6 loop conformation of NBD1-WT, NBD1-F508A, and NBD1ΔF508 crystal structures.
Changes in NBD1 folding kinetics have been shown earlier experimentally. Qu
Our results reveal the intrinsic property of NBD1ΔF508 to fold improperly and raise the possibility of redesigning NBD1ΔF508 to rescue it from misfolding. In case of the contact that is found in wild type but not in the ΔF508 mutant (e.g., Q493/P574), one can find amino acid substitutions that promote interaction between this pair of residues (Q493/P574). On the other hand, for the contact found only in the ΔF508 mutant but not in wild type (e.g., F475/F587), candidate rescue mutants are those that destabilize the interaction between this residue pair (F475/F587). Knowing the molecular details of the altered folding in the case of the mutant domain also provides a basis for design of small molecules to correct the most prevalent and pathogenic mutation in CFTR.
To access time scales of NBD1 folding, we use a simplified protein model but still maintain important features of the protein such as side-chain packing. Amino acid residues were modelled as follows: (1) glycines are represented by three beads (-N, Cα, C′); (2) phenylalanine, tyrosine, tryptophan, and histidine by five beads (-N, Cα, C′, Cβ, Cγ), and (3) all other residues by four beads (-N, Cα, C′, Cβ)
Using discrete molecular dynamics
We perform 300 folding simulations for each NBD1-WT, NBD1-F508A, and NBD1-ΔF508. Starting from fully unfolded chains, the temperature of the system is progressively reduced to allow NBD1 to fold to its native structure. Folding simulations proceeded until τ
To identify the positions of intermediate states, a sum of multiple Gaussian curves
We estimate probability of transitions between states by counting the trajectories that underwent such transition. The sum of probabilities of paths emanating from a given state is normalized to 1, which physically means that the system always exits from its current intermediate state. The transition probabilities represent independent conditional probabilities, thus the most likely path from the unfolded state to the native is estimated by multiplying the probabilities of the traced edges.
We calculated a contact matrix for each structure in the intermediate state. An element of the contact matrix is 1 when two residues were within 4.5 Å or 0 otherwise. Dominant contacts between pairs of residues in NBD1 are determined from the average contact matrix of all the structures.
Positions of metastable intermediate states are identified by fitting a sum of gaussian distributions. Each gaussian curve corresponds to a folding intermediate state.
(0.57 MB TIF)
For a given state, we calculated the average fraction of native contacts (Q) coming from a particular folding trajectory. The normalized distribution of Q shows that the states defined using energy are structurally distinct.
(0.65 MB TIF)
Probability of kinetic transitions between intermediate states of NBD1-WT, NBD1-ΔF508, and NBD1-F508A. The probability of exiting a state is normalized to 1. The thickness and warmth of the transition edges are rendered proportional to the probability value.
(0.99 MB TIF)
Also shown are the average root-mean-square deviations (RMSD) of intermediate state structures with respect to the native structure.
(0.04 MB DOC)