Structural analysis of genetic variants of the human tumor suppressor PALB2 coiled-coil domain

Abstract

The tumor suppressor PALB2 is a key player in the homologous recombination (HR) pathway, functionally connecting BRCA proteins at the DNA damage site. PALB2 forms homodimers via its coiled-coil domain, and during HR, it forms a heterodimeric complex with BRCA1 using the same domain. However, the structural details of the human PALB2 coiled-coil domain are unknown. Several missense variants have been reported in the coiled-coil domain. The structure-function relationship of these variants is poorly understood, posing a challenge to genetic counseling. In this study, we present the solution structure of the human PALB2 coiled-coil domain, which forms an antiparallel homodimer. We then use this structure to investigate the impact of a few well-characterized missense mutations on the fold and interactions of the PALB2 coiled-coil domain. Our findings reveal a strong correlation between the structural impact of mutations and their efficiency in homologous recombination, suggesting that our approach can be applied to study other genetic variations in PALB2. These findings hold promise for improving genetic counseling and advancing cancer research.

Keywords: DNA repair; NMR spectroscopy; breast cancer; missense mutations; molecular dynamics simulations; tumor suppressor gene.

Figure 1:. Analysis of the PALB2 coiled-coil domain (PALB2cc) by biophysical techniques.

(A) Schematic representation of the PALB2 protein and its structural domains, along with the known binding sites of its interactors. The coiled-coil domain spans amino acids 6-42. (B) A few of the identified variants of the PALB2 in the coiled-coil domain were followed up in this study. The HR values are provided in a previous study by Boonen et al. .[13]. (C) The CD spectrum of the PALB2cc domain. The helical and parallel betaβ-sheet propensities obtained from fitting the CD data are mentioned at the top. (D) The size of the PALB2cc domain is determined by analytical size exclusion chromatography. The plot of molecular weight standards against the elution volume is on the left. Expected monomer, dimer molecular weights, and the measured molecular weight are given on the right.

Figure 2:. The structure of the PALB2cc domain was studied by NMR spectroscopy and MD simulations.

(A) The ¹H–-¹⁵N HSQC spectrum of PALB2cc domain (residues:6-41). The assigned backbone N–-H amide resonances are labeled. The amide peaks of the first two N-terminal residues (residue numbers 6 and 7) are broadened due to exchange, signifying a dynamic region. Residue number 8 is a proline whose amide peak is not visible in the HSQC. (B) The solution structure for the PALB2cc domain suggests an antiparallel orientation between the two helices. (C) The contacts between protomers in the PALB2cc dimer are highlighted. Green, black, and red lines connect the interacting residues. The green lines denote hydrophobic contacts. Black lines show the hydrogen bonds, and the red line shows the van der Waals interactions. (D) The angle between the aromatic rings of Y28 from each protomer is plotted against the distance between C_z- and C_z atom pairs. The cluster of structures with edge-to-face orientation is marked in green, and parallel orientation is marked in red. These were obtained from three (n = 3) independent 0.5 μs simulations. (E) Snapshots from the MD simulations show the edge-to-face and parallel orientation of the aromatic rings of Y28. (F) The number of inter-protomer contacts (normalized) in the wt and Y28A variants are plotted (P = 0.001, the number of samples is 3). Normalization was performed by calculating the number of contacts observed in wt-PALB2cc. The error bar represents the standard deviation of three (n = 3) independent 0.5 μs simulations.

Figure 3:. The surface charge distribution and stability of the PALB2cc domain.

(A) Charge distribution on the surface representation of PALB2cc monomer (Rred –—Nnegatively charged, Bblue –—Ppositively charged, and Ggrey- —uncharged). (B) Snapshots of a single helix from PALB2cc dimer simulation (blue) and monomer simulation (green). The snapshot of (C) Tthe number of backbone hydrogen bonds in the PALB2cc dimer and monomer. (D) Intra-protomer contacts are plotted for the dimer and monomer. The contacts are measured between all non-hydrogen atoms with a distance cutoff of 7 Å. (E) Percentage helicity.

Figure 4:. The structural and thermodynamic stability of the L35P-PALB2cc domain was studied by biophysical techniques and MD simulations.

(A) The CD spectrum of the L35P-PALB2cc domain. The % helicity is provided below. (B) A representative thermal melt of wt-PALB2cc and L35P-PALB2cc. The melts were performed twice, and the values obtained are provided in a table within the inset. (C) From MD simulations, snapshots of the wt-PALB2cc dimer (blue) and L35P-PALB2cc (green). (D) The number of backbone hydrogen bonds in wt-PALB2cc and L35P-PALB2cc observed in MD simulations is plotted (P < 0.01, the number of samples is 3). (E) The number of intra-protomer contacts between all non-hydrogen atoms within the distance of 7 Å in MD simulations is plotted for wt-PALB2cc and L35P-PALB2cc (P = 0.03,; the number of samples is 3). (F) The number of inter-protomer contacts detected between all non-hydrogen atoms in MD simulations is plotted for wt-PALB2cc and L35P-PALB2cc (P = 0.05,; the number of samples is 3). The y-axis in (D), (E), and (F) areis normalized by the number of contacts in the wt-PALB2cc. The error bars in (D), (E), and (F) represent the standard deviation of three (n = 3) independent 0.5 μs simulations.

Figure 5:. The PALB2-BRCA1 coiled-coil heterodimer was studied using MD simulations.

(A) Modeled with Swiss model after 100 ns simulations (Bblue-—PALB2cc, Ggreen-—BRCA1cc, and Rred-— dimerization interface residues). (B) PALB2cc-BRCA1cc contacts at the site of PALB2 mutation (L35P: P = 0.023, L24S: P < 0.001, R37H: P = 0.005, Y28C: P < 0.001, K18R: P = 0.002, T31I: P < 0.001., and Tthe number of samples is 3). Native contacts are calculated between non-hydrogen atoms with a pairwise distance cutoff of 10 Å. The number of contacts is normalized by the contacts observed in the wild-type PALB2. The error bar represents the standard deviation of three (n = 3) independent 0.5 μs simulations.