Barrier-to-autointegration factor: major roles in chromatin decondensation and nuclear assembly

Abstract

Barrier-to-autointegration factor (BAF) is a DNA-bridging protein, highly conserved in metazoans. BAF binds directly to LEM (LAP2, emerin, MAN1) domain nuclear membrane proteins, including LAP2 and emerin. We used site-directed mutagenesis and biochemical analysis to map functionally important residues in human BAF, including those required for direct binding to DNA or emerin. We also tested wild-type BAF and 25 point mutants for their effects on nuclear assembly in Xenopus egg extracts, which contain approximately 12 microM endogenous BAF dimers. Exogenous BAF caused two distinct effects: at low added concentrations, wild-type BAF enhanced chromatin decondensation and nuclear growth; at higher added concentrations, wild-type BAF completely blocked chromatin decondensation and nuclear growth. Mutants fell into four classes, including one that defines a novel functional surface on the BAF dimer. Our results suggest that BAF, unregulated, potently compresses chromatin structure, and that BAF interactions with both DNA and LEM proteins are critical for membrane recruitment and chromatin decondensation during nuclear assembly.

Figure 1.

BAF mutagenesis. Residues that comprise the five α-helices in BAF are indicated by bars above the amino acid sequence of human BAF (Umland et al., 2000; Cai et al., 2001). Point mutations are indicated by E (glutamic acid), Q (glutamine), and A (alanine), and numbered. Each BAF mutant contained one substituted residue. A few residues were changed to either of two mutant residues.

Figure 2.

Binding of mutant hBAF proteins to blot-immobilized emerin. Blots bearing human emerin protein (residues 1–222) were cut into strips. Each strip was probed with ³⁵S-labeled wild-type or mutant BAF, numbered as in Fig. 1. Radiolabeled wild-type human BAF (WT) served as a positive control. Binding of each mutant BAF to emerin was scored relative to the amount of each input probe (unpublished data), and summarized in Table I.

Figure 3.

DNA binding activity of BAF mutants. Each ³⁵S-labeled wild-type or mutant BAF protein was incubated with (+) or without (−) native DNA cellulose beads, then pelleted, washed, separated on SDS-PAGE, and detected by autoradiography.

Figure 4.

Subunit exchange assay. Each His-tagged mutant BAF protein was incubated with ³⁵S-labeled wild-type BAF, and immunoprecipitated with anti-His antibody. As positive and negative controls, ³⁵S-wild-type BAF was incubated with (WT) or without (−) His-tagged wild-type BAF, respectively, before immunoprecipitation and SDS-PAGE. The left and right panels are from different experiments, and had different amounts of input ³⁵S-wild type BAF. (A) Autoradiographs showing the ³⁵S-labeled wild-type BAF that coimmunoprecipitated with wild-type BAF (WT), or each BAF mutant (numbered as in Fig. 1). (B) Parallel Western blots probed with anti-His antibody, showing the amount of His-tagged BAF present in each reaction. All recombinant proteins migrated at their expected mass of 10 kD. (C) Densitrometric ratios of signals shown in A and B. Graphs show relative amounts of ³⁵S-wild-type BAF that interacted with each input His-tagged BAF.

Figure 5.

Cloning, expression, and localization of Xenopus BAF. (A) Human (top) and Xenopus (bottom) BAF are 84% identical (red) and 91% similar (blue). (B) Affinity-purified rabbit antibodies (serum 3710) recognized both recombinant (R) and endogenous (E) Xenopus BAF. Recombinant BAF (calculated mass, 10.2 kD) migrated at 10 kD on SDS-PAGE, whereas endogenous BAF migrated at 40 kD. (C) Western blot of the soluble fraction of Xenopus egg extracts showing that recognition of endogenous BAF by affinity-purified 3710 antibody was specifically competed by pretreatment (+) with antigenic peptide. Pre, preimmune antibody; pep, antigenic peptide. (D and E) Indirect immunofluorescent staining of endogenous BAF in cultured Xenopus A6 cells (D) and XLK-WG cells (E) using affinity-purified immune (Imm) or preimmune (pre) 3710 antibody. xBAF localizes predominantly at the nuclear rim, but is also found in the nuclear interior and cytosol (D and E, right). Left panels show DNA in the same cells, stained by Hoechst 33258.

Figure 6.

Exogenous BAF has two distinct effects on chromatin when added to Xenopus nuclear assembly reactions. Purified recombinant Xenopus BAF (A) or human BAF (B) were added to Xenopus nuclear assembly reactions at time zero, at concentrations of 0, 0.5, 2.5, or 5 μM recombinant BAF dimers. (A–C) Upper panels show nuclei by phase contrast microscopy; corresponding lower panels show same nuclei stained for DNA with Hoechst 33258. Nuclei were imaged after 2 h of assembly. (C) Timecourse (20-min intervals) of nuclear assembly without (no addition), or with 5 μM exogenous xBAF dimers (xBAF). Bars: (A and B) 10 μm; (C) 30 μM.

Figure 7.

Transmission EM of control and wild-type xBAF-inhibited nuclei. Nuclei were assembled for 2 h with no BAF added (A) or 5 μM added xBAF dimers (B), and visualized by TEM. (A) Asterisks indicate nuclear pore complexes. (B) Nuclei assembled in 5 μM xBAF had patches of membranes at the chromatin surface. Arrow indicates chromatin emerging between membrane patches. Paired arrowheads bracket the electron-dense outer shell of chromatin in inhibited nuclei. (C) TEM cross-section of sperm chromatin before addition to assembly extracts. Bars, 500 nm.

Figure 8.

Effects of mutant hBAF proteins on nuclear assembly in Xenopus egg extracts. Mutant BAF proteins were added at time zero to the indicated final concentrations, and imaged after 2 h of assembly. Mutants fell into four phenotypes by light microscopy: (A) wild-type (decondensed-to-condensed), (B) inactive, (C) always condensed, and (D) inactive-to-condensed. The representative mutant shown for each class is bolded. Mutants are numbered according to Fig. 1. Mutant 75E behaved like wild-type BAF, but with clumps of DNA at the nuclear poles. Bar, 10 μm.

Figure 9.

TEM analysis of nuclei assembled in Xenopus extracts for 2 h with the indicated hBAF mutant. (A and B) Nuclei assembled in 0.5 μM (A) or 5 μM (B) always condensed mutant 14A. Arrows indicate the thin (A) and thick (B) shell of condensed chromatin caused by this mutant. (C and D) Nuclei assembled in 0.5 μM (C) or 5 μM (D) always condensed mutant 47E; note the normal chromatin and pore-less double membrane in C, and uniformly condensed chromatin in D. (E) Nucleus assembled in 5 μM inactive-to-condensed mutant 53E. (F–I) Higher magnifications of panels A, B, D, and E. (F and G) Mutant 14A at 0.5 μM (F) and 5 μM (G). (H) Mutant 47E at 5 μM. (I) Mutant 53E at 5 μM. Bars: (A–E) 500 nm; (F–I) 200 nm.

Figure 10.

Functional residues on the BAF dimer surface. (A–C) Corresponding ribbon diagram (A) and surface structure representation (B) of the wild-type human BAF dimer (left). In the front orientation, dsDNA molecules bind to the left and right ends, and the LEM-binding domain faces the reader. The BAF dimer at right is rotated down 90° to show the top surface. Unprimed and primed numbers (e.g., 27 and 27') indicate residues in the left and right monomers, respectively. (B) Residues essential for emerin binding are light blue: surface-exposed residues 51, 53, and 54 cluster in the valley that spans both monomers. Residues 46 and 47 (Umland et al., 2000) are buried at the dimer interface. Surface-exposed residues 6, 9, 27, and 75, in which mutations reduced (but did not eliminate) binding to emerin, are dark blue. (C) BAF residues essential for DNA binding are light blue: residues 6, 25, and 27 map to the left and right of the dimer; residue 46 is buried. Residue 25 is not visible in this front view. Surface-exposed residues 9, 51, 54, and 75, in which mutations reduced DNA binding, are dark blue. (D) Always condensed mutants mapped to the top of the dimer (residues 14 and 18), and the dimer interface (residue 47; buried). (E) Inactive mutants mapped to the left and right ends of the dimer (residues 25 and 27) and the dimer interface (residue 46; buried), as viewed from the side and bottom.