Structural insights into the role of domain flexibility in human DNA ligase IV

Abstract

Knowledge of the architecture of DNA ligase IV (LigIV) and interactions with XRCC4 and XLF-Cernunnos is necessary for understanding its role in the ligation of double-strand breaks during nonhomologous end joining. Here we report the structure of a subdomain of the nucleotidyltrasferase domain of human LigIV and provide insights into the residues associated with LIG4 syndrome. We use this structural information together with the known structures of the BRCT/XRCC4 complex and those of LigIV orthologs to interpret small-angle X-ray scattering of LigIV in complex with XRCC4 and size exclusion chromatography of LigIV, XRCC4, and XLF-Cernunnos. Our results suggest that the flexibility of the catalytic region is limited in a manner that affects the formation of the LigIV/XRCC4/XLF-Cernunnos complex.

Graphical abstract

Figure 1

Gel Filtration Chromatography Studies of Complex Formation of LigIV, XRCC4^ΔCTD;CtoA, and XLF^ΔCTD (A) Profiles of the UV absorbance at 280 nm during gel filtration chromatography. Colors of profiles and their corresponding constructs are shown at the bottom of the figure. Gray arrows indicate peak positions of protein standards, void, ferritin (440 kDa), aldolase (158 kDa), conalbumin (75 kDa), ovalbumin (43 kDa), carbonic anhydrase (29 kDa), and RNase A (13.7 kDa). (B) SDS-PAGE of LX4^ΔCTD;CtoA and XLF^ΔCTD fractions eluted from a Superdex 200 10/30 column. The molecular weight markers are in column “M” column and their molecular weights (kDa) are shown on the left of the gel. The fraction ranges used for the SDS-PAGE are indicated using a blue arrow in (A). Each fraction contained 250 μl of the eluted sample. (C) SDS-PAGE of L^ΔcatX4^ΔCTD;CtoA and XLF^ΔCTD eluted from the gel filtration column. The fractions used for SDS-PAGE are indicated alphabetically (green a-n) both in (A) and in the gel. (D) Schematic representation of the constructs used in the gel filtration experiment. The domain names and boundaries are shown in LX4^ΔCTD;CtoA and XLF^ΔCTD. In XRCC4^ΔCTD;CtoA and XLF^ΔCTD, HD, CC, and FB represent head, coiled-coil, and fold-back domains, respectively. See also Figure S1.

Figure 2

SAXS Studies of LX4 (A) Experimental scattering curves of LX4 constructs. The scattering intensities (log I versus s-value) with error bars (gray) of LX4 (blue), LX4^ΔCTD;CtoA (green) and L^ΔcatX4^ΔCTD;CtoA (blue) are displaced by factor of 100 for clarity. The scattering curves of the latter two constructs were modified after Ochi et al., 2010. (B) Distance distributions of LX4 constructs. The same color scheme as in (A) is used in this figure. The error bars are represented with gray. (C) Shape reconstruction of LX4^ΔCTD;CtoA. The molecular envelope of LX4^ΔCTD;CtoA is shown in two perpendicular orientations, which derived from an averaging process of several, individually restored 3D shapes. The structure of a LX4 construct (PDB code: 3II6; Wu et al., 2009). The structure was fitted into the envelope manually and refined using Chimera (Pettersen et al., 2004). The two structural superimpositions providing the highest correlation correlations coefficients are illustrated to highlight additional molecular density not present in the crystal structure. See also Figure S2.

Figure 3

Structure of NTase-3 (A) Overall architecture of NTase-3. Conserved motifs I, III and IIIa are shown in purple, pink and yellow. Dotted lines represent missing loops (gray and pink) connecting DNA-binding regions D1 and D2 (pink). (B) A schematic presentation of secondary structure elements of NTase-3. See also Figure S3.

Figure 4

Rigid-Body Modeling and Protein-Protein Binding Assays of Human DNA Ligase IV/XRCC4 Complex (A) Rigid-body modeling of LX4^ΔCTD;CtoA using BUNCH. Ten individual rigid-body models were superposed on the structure of the L^ΔcatX4^ΔCTD;CtoA region. The models with the three highest χ² values are shown in a cartoon representation and the others are shown as their Cα traces. (B) Left: EMSAs of individual catalytic domains and L^ΔcatX4. The proteins used are indicated with “+.” Right: GST pull-down assays of OBD and L^ΔcatX4. The upper and lower figures show the results of the assays using GST-OBD-620 and GST-OBD-653, respectively. The first lane protein markers (M) are followed by unbound proteins (U) and bound proteins (B) to GST affinity resin.

Figure 5

Comparison of the DNA-Binding Loop D1 of NTase-3 The structure of NTase-3 of LigIV (cyan) is shown together with that of LigI (I, pink; PDB code: 1X9N) and LigIII (III, blue; PDB code: 3L2P), and DNA (PDB code: 1X9N). Backbone phosphates of DNA are labeled as 12 and 13. The pink-dotted lines represent hydrogen bonds between LigI and the DNA. The original residue numbers of the phosphates shown in the PDB file are used here. See also Figure S4.

Figure 6

Interactions of Residues that Cause LIG4 Syndrome that Are Close to the Catalytic Residues of Human LigIV (A) Hydrophobic core of NTase-3 of LigIV around Y288 (magenta). Since the electron density for the side chains of K283 and I315 was not observed, the amino acids were represented as alanines. (B) Interactions in the model between R278 (magenta) and surrounding residues. (C) Interactions between Q280/H282 and surrounding residues. Salt bridges and hydrogen bonds are represented using black dotted lines. See also Figure S5.