Overall structure of the human GINS complex. Three different views are shown. On top left, Sld5 (blue), Psf1 (orange), Psf3 (green), and Psf2 (magenta) are arranged in a clockwise direction. The pseudo-twofold symmetry axis is shown in red arrows or a red circle between the Psf2–Sld5 and Psf3–Psf1 complexes. In front of the secondary structure, labels (H or S) 1, 2, 3, and 5 are added for Psf1, Psf2, Psf3, and Sld5, respectively, in all figures.

Structures of the four proteins of the GINS complex. (A) Structural similarities between Sld5 (blue) and Psf1 (orange). The N-terminal domain forms an arch shape, and the C-terminal domain is on top of the arch. The structure of Psf1 resembles that of Sld5. (B) Structural similarities between Psf2 (magenta) and Psf3 (green). (C) The four proteins of GINS share four arch-forming helices (top) and a jelly-roll fold (bottom). The N-terminal and C-terminal domains of Sld5/Psf1 are permuted in Psf2/Psf3.

Interfaces of the GINS complex. (A) A close-up view of the intermolecular four-helical bundle formed by helices H3 and H5 of Sld5 (blue) and helices H3 and H5 of Psf1 (orange). Residues that are mutated in our analysis are boxed with red or blue. In addition to the extensive hydrophobic interactions, the two ion pairs and the five H-bonds further strengthen this interface (see Supplementary Fig. S3 for details). (B) Psf2 (magenta) and Psf3 (green) dimerize to form an intermolecular four-helical bundle. Helices H5 and H6 of Psf2 interact with helices H8 and H10 of Psf3 primarily through hydrophobic interactions. (C) The atomic interactions between the N-terminal domain of Sld5 (blue) and the N-terminal domain of Psf2 (magenta). Helices H1 and H3 of Sld5 interact with helices H1 and H2 and with strands S2 and S4 of Psf2. (D) The atomic interactions between the N-terminal domain of Psf1 (orange) and the N-terminal domain of Psf3 (green). This interface is very similar to that of Sld5–Psf2 shown in C.

Mutant complementation analysis of yeast GINS. (A) A surface representation of the human GINS structure with the degree of sequence conservation encoded by the red color. Invariant residues among all five species are colored red, and less conserved residues (80%) are colored orange. Fifteen surface patches for mutations are circled. Surface patches that affect cell viability are marked with a box, and distances between these patches are shown. (B) The lethality of mutant strains (P_GAL1-SLD5 and P_GAL1-PSF1) transformed with a control vector or with corresponding yeast GINS surface mutant plasmids was tested on SC-Ura media containing different carbon sources, such as galactose (left) and glucose (right). (C) A complementation test of a mutant strain, P_GAL1-PSF2, that was transformed with a control vector or with various corresponding yeast Psf2 surface mutant plasmids was performed on SC-Ura + Gal (left) and SC-Ura_Glu (right) plates. All plates were incubated at 30°C for 2–3 d. (D) Gel filtration analysis showed that the surface mutant proteins of the human GINS complex, which affects the cell growth, form a tetramer. For clarity, only gel filtration profiles of Sld5-1, Psf2-2, and Psf2-5 are shown. On the right side, an SDS gel is shown for the wild-type (lane 1) and Sld5-1 mutant (lane 2) GINS complexes. Other surface mutants showed the same results and are omitted. (E, top) CD analysis showed that the surface mutations on the Sld5-1 patch (or other Psf2 mutants) did not affect the overall conformation of the GINS complex. For clarity, only Sld5-1, Psf2-2, and Psf2-5 profiles are shown. (Bottom) Temperature scanning analysis showed that the Tm values of the wild-type GINS complex and the six surface mutant GINS proteins are basically the same.