Molecular mechanisms of functional rescue mediated by P53 tumor suppressor mutations

Abstract

We have utilized both molecular dynamics simulations and solution biophysical measurements to investigate the rescue mechanism of mutation N235K, which plays a key role in the recently identified global suppressor motif of K235/Y239/R240 in the human p53 DNA-binding domain (DBD). Previous genetic analysis indicates that N235K alone rescues five out of six destabilized cancer mutants. However, the solution biophysical measurement shows that N235K generates only a slight increase to the stability of DBD, implying a rescue mechanism that is not a simple additive contribution to thermodynamic stability. Our molecular simulations show that the N235K substitution generates two non-native salt bridges with residues D186 and E198. We find that the nonnative salt bridges, D186-K235 and E198-K235, and a native salt bridge, E171-R249, are mutually exclusive, thus resulting in only a marginal increase in stability as compared to the wild type protein. When a destabilized V157F is paired with N235K, the native salt bridge E171-R249 is retained. In this context, the non-native salt bridges, D186-K235 and E198-K235, produce a net increase in stability as compared to V157F alone. A similar rescue mechanism may explain how N235K stabilize other highly unstable β-sandwich cancer mutants.

Introduction

Premalignant cells must overcome the physiological regulation of growth and programmed cell death (or apoptosis) in order to turn into aggressive and metastatic cancers. Therefore, inactivation of tumor suppressor proteins through the accumulation of gene mutations is a central transforming event. Prominent among these tumor suppressor genes is p53; mutant p53 genes have been found in about 50% of human tumors, [1] About 75% of these are missense mutations that express full-length, but functionally inactive, protein [1], [2]. These mutations are concentrated within the DNA-binding core domain (DBD) of approximately 200 amino acids, a domain that is central to the function of p53 as a transcription factor [1].

The significance of p53 mutation to cancer biology and the predominance of missense mutations in DBD have motivated searches for potential second-site suppressor mutations. Three classes of missense cancer mutations have been identified based on their inactivation mechanisms, i.e.: 1) loss of DNA contacts, 2) local structural perturbation, and 3) unfolding of DBD [3]. Accordingly, rescue of cancer mutants in these three classes is expected to require three distinct classes of suppressor mutations, i.e.: 1) mutations that introduce additional DNA contacts, 2) mutations that correct local structural distortion, and 3) mutations that refold DBD [4]. For an example of the first of these three classes, rational design identified a suppressor mutation, T284R, that increases the DNA binding affinity of the particular contact mutants R273C and R273H by providing a new DNA contact [5]. Of course, the limitation of this approach and of rescue of DNA contact mutants in general is that each particular mutant is likely to require an individually tailored suppressor. The same limitation might be expected to extend to cancer mutants that disrupt local structure in the DNA-binding region.

In contrast, missense cancer mutations that unfold DBD but do not perturb local structures at the DNA binding interface extend the promise of a global rescue mechanism based on refolding of DBD. Since DBD denatures in a two-state process [6], [7] a suppressor located anywhere in the structure that pushes the equilibrium in the direction of the folded state should contribute to the refolding of all destabilized cancer mutants. Prominent among these destabilized cancer mutants are the β-sandwich mutants for which thermodynamic destabilization is indicated to be the only inactivation mechanism [3].

Brachmann and co-workers demonstrated the feasibility of a genetic approach to select suppressors of p53 cancer mutations [8], [9], [10]. Initial screens identified several candidates [8]. Two of these, N268D and N239Y, represent potential examples of more broadly targeted refolding suppressors. The DBD of N268D is more stable than that of the wild type and V143A, a strongly destabilized β-sandwich mutant, is rescued by N268D due to additive energetic effects on stability [4], [11]. N239Y was also found to increase the stability of DBD, [4], [11] a result explained by the recent crystallographic structure of a super stable quadruple mutant [12]. Subsequent PCR mutagenesis suggests a suppressor motif of K235/Y239/R240 [10]. Sixteen of the 30 tested cancer mutants were rescued by combinations of these suppressor mutations [10]. The group rescued includes six mutations located in the β-sandwich where the hydrophobic core is located, and three each in the L2/L3 loops and loop-sheet-helix. Of seven β-sandwich cancer mutants tested, only one, Y163C, is not rescued, which suggests a mechanism involving the refolding of the DNA-binding core domain by global stabilization [10].

As described above, the thermodynamic investigations of N239Y [4] indicate that it operates by this refolding mechanism. The proximity of R240 to the DNA binding interface indicates that it may function by introducing additional DNA contacts. However, the rescue mechanism of N235K, located in S8 of the β-sandwich, [13] is less clear. This is an important suppressor, particularly for destabilized β-sandwich mutants. N235K alone rescues five of six β-sandwich cancer mutants among the top 30 cancer mutants (Danziger et al., in preparation). Residue 232 in the mouse p53 DNA-binding core domain is lysine, i.e, the equivalent to K235 in human p53. In the crystallographic structure of the mouse DBD (2IOI) [14] shown in Fig. 1, K232 forms a salt bridge with E195, which corresponds to E198 in human p53. A potential salt bridge interaction with D183, which corresponds to D186 in human p53, is also noticeable. This is similar to the situation in N268D, which was shown to rescue certain human p53 cancer mutants due to formation of an additional salt bridge interaction. However in contrast to N268D which increases the stability of human p53 DBD by 1.2 kcal/mol, [4] N235K generates only a slight increase to the stability of the human p53 DBD, i.e. by 0.3 kcal/mol (Table 1). This implies a rescue mechanism that is not a simple additive contribution to thermodynamic stability. Due to the potential significance of N235K as a global cancer suppressor, the mechanism for its rescue of destabilized β-sandwich mutants, if not by simple contribution to global stability, poses a significant question.

In this study, we have used solution biophysical measurements and molecular simulations in explicit solvent to investigate the likely rescue mechanisms of N235K to restore unstable cancer mutants. Our focus on protein stability is based on the available crystal structures that show virtually no major difference between cancer mutants and the wild type DBD whether they are stable or not [15], [16], [17]. Our simulation design is also to bypass the difficulty in the stability analysis through direct simulation of protein folding/unfolding at room temperature. The first MD simulations of the p53 DNA-binding domain were reported as early as the beginning of this century, [18], [19], [20], [21], [22], [23], [24], [25], [26] though none of these previous MD studies focus on direct folding/unfolding simulations of the protein, which are clearly beyond the limit of current computational power. We have chosen two β-sandwich cancer mutants as test cases in our analysis of the rescue mechanisms of N235K. These are V157F, one of the most destabilizing cancer mutations, but which is refolded by N235K alone and Y163C, a similarly destabilized cancer mutant that is not refolded by N235K alone, though it can be in combination with other elements of the 235/239/240 suppressor motif.