Report

Chromatin Docking and Exchange Activity Enhancement of RCC1 by Histones H2A and H2B

See allHide authors and affiliations

Science  25 May 2001:
Vol. 292, Issue 5521, pp. 1540-1543
DOI: 10.1126/science.292.5521.1540

Abstract

The Ran guanosine triphosphatase (GTPase) controls nucleocytoplasmic transport, mitotic spindle formation, and nuclear envelope assembly. These functions rely on the association of the Ran-specific exchange factor, RCC1 (regulator of chromosome condensation 1), with chromatin. We find that RCC1 binds directly to mononucleosomes and to histones H2A and H2B. RCC1 utilizes these histones to bind Xenopus sperm chromatin, and the binding of RCC1 to nucleosomes or histones stimulates the catalytic activity of RCC1. We propose that the docking of RCC1 to H2A/H2B establishes the polarity of the Ran-GTP gradient that drives nuclear envelope assembly, nuclear transport, and other nuclear events.

RCC1 can be considered as a chromatin marker. Catalysis of guanine nucleotide exchange on Ran by RCC1 to produce Ran-GTP is essential for mitotic spindle assembly and nuclear envelope formation (1–4). Once enclosed by the envelope, chromatin-bound RCC1 generates a Ran-GTP gradient across nuclear pores that permits vectorial nucleocytoplasmic transport (4). The docking mechanism for RCC1 onto chromatin is unknown. RCC1 may bind DNA in vitro, but removal of the NH2-terminal domain of RCC1 abrogates DNA binding without loss of in vivo chromatin association (5).

To identify the mechanism for chromatin binding, HeLa nuclei were partially digested with micrococcal nuclease and the solubilized chromatin was separated by centrifugation (6). The majority of RCC1 cosedimented with nucleosomal fractions (Fig. 1A), suggesting that endogenous RCC1 associates with nucleosomes. Glutathione S-transferase (GST)–tagged RCC1 (GST-RCC1), but not GST, bound H1-depleted mononucleosomes (6), demonstrating a direct interaction (Fig. 1B) (7). The core histones have folded regions that participate in octamer assembly and DNA binding as well as highly modified NH2-terminal tails (8). Brief treatment with trypsin digests these NH2-terminal tails while leaving the folded domains intact. GST-RCC1 bound both intact and trypsinized mononucleosomes (Fig. 1B) (7). Binding of GST-RCC1 to recombinant histone tails from H2B and H3 was barely detectable (9). Thus, the histone tails may contribute to, but are not necessary for, the binding of RCC1 to chromatin.

Figure 1

RCC1 binds mononucleosomes. (A) HeLa nuclei were digested with micrococcal nuclease and centrifuged through a linear 8 to 20% sucrose gradient. Samples of individual fractions were electrophoresed through a tris-borate EDTA-agarose gel and visualized by ethidium bromide staining (top) or precipitated with trichloroacetic acid, subjected to SDS-PAGE, and immunoblotted (N-19, Santa Cruz) for endogenous RCC1 (bottom). (B) Immobilized GST, GST-RCC1, or GST-RCC1(23-421) was incubated with intact or trypsinized H1-depleted mononucleosomes. After washing, proteins were eluted, subjected to SDS-PAGE, and stained with Coomassie.

Association of RCC1 with nucleosomes could be mediated by binding to DNA, histones, or both components; however, a mutant RCC1 lacking its DNA binding region [GST-RCC1(23-421)] retained its ability to associate with trypsinized nucleosomes (Fig. 1B). To determine whether RCC1 binds directly to histones, immobilized GST or GST-RCC1 was incubated with histones from chicken erythrocyte nuclei (7). In this experiment, we presume that H2A/H2B assemble heterodimers that are not associated with the H3/H4 heterotetramers (10). GST-RCC1 depleted all four histones from the supernatant (Fig. 2A); however, H3/H4 were removed during the washes and only H2A/H2B remained bound. Thus, RCC1 binds directly to histones, and preferentially to H2A/H2B.

Figure 2

Histones bind directly to RCC1 and enhance its catalytic activity. (A) Immobilized GST or GST-RCC1 (1 μM) was incubated with chicken histones (1 μM) with (+) or without (–) the indicated form of Ran (10 μM). Histones in the unbound fraction were isolated with SP-Sephadex beads (Pharmacia). Bound and unbound fractions were subjected to SDS-PAGE and stained with Coomassie. The asterisk indicates a degradation product from Ran(Q69L), not histone H4. (B) As in (A), except that H2A or H2B was used. (C) Exchange reactions were performed with or without H2A/H2B (333 nM). Samples were analyzed by binding to nitrocellulose and scintillation counting. (D) As in (C), except that a fixed concentration of RCC1 (0.9 nM) was used with various concentrations of H2A/H2B or H3/H4 (Roche). Reactions were performed with (E) or without (F) RCC1 (0.9 nM), with or without DNA, BIB domain, nucleosomes, or trypsinized nucleosomes (all 250 nM). Dissociation rate constants were calculated assuming first-order kinetics (24). Error bars represent the range from the mean (n = 2). Data are representative of at least three independent experiments.

Ran-GTP regulates the interaction of nuclear transport receptors with their cargoes. To determine whether Ran also regulates the histone-RCC1 interaction, the binding assay was repeated with an excess of wild-type or one of two mutant forms of Ran (7). Ran(T24N) has a reduced ability to bind nucleotide but increased affinity for RCC1; Ran(Q69L) is constitutively GTP-bound (11, 12). These forms of Ran did not decrease RCC1 binding to H2A/H2B (Fig. 2A), suggesting that Ran does not disrupt the binding of RCC1 to chromatin.

Because H2A and H2B heterodimerize, RCC1 could interact with either protein alone, or only with the heterodimer. GST-RCC1 bound H2A or H2B alone even when their concentration was only 50 nM (Fig. 2B), suggesting that RCC1 associates tightly with either histone in the absence of DNA.

Because the interaction between RCC1 and H2A/H2B was not competitive with Ran binding, we expected that chromatin-bound RCC1 would catalyze nucleotide exchange. Therefore, exchange reactions were performed (13) with or without H2A/H2B or H3/H4. The addition of H2A/H2B stimulated the catalytic activity of RCC1 (Fig. 2, C and D), increasing the dissociation rate constant,k off, for nucleotide by ∼twofold (Fig. 2C). The effect was saturable, with an apparent dissociation equilibrium constant K 1/2, of ∼25 nM. H2A and H2B stimulated RCC1 individually and enhanced exchange of guanosine diphosphate (GDP) for GTP or vice versa; the NH2-terminal tails from H2A/H2B did not stimulate RCC1 (9). The beta-like import receptor-binding (BIB) domain, a basic protein fragment derived from ribosomal protein L23a (4), also did not stimulate RCC1, demonstrating that the effect was histone-specific and not merely a consequence of positive charge (Fig. 2E). In addition, the histones H3/H4 only stimulated RCC1 weakly and at high concentrations (Fig. 2D). Intact and trypsinized RCC1-depleted mononucleosomes (Fig. 2F) stimulated exchange similarly to H2A/H2B (Fig. 2, D and E), with aK 1/2 of ∼100 nM (14), whereas DNA itself had no effect (Fig. 2E). Taken together, these data suggest that H2A/H2B are cofactors for RCC1 at the chromatin surface.

It was important to determine whether RCC1 utilizes H2A/H2B for chromatin binding in vivo. Therefore, we used chromatin isolated from the sperm of Xenopus laevis that are deficient in histones H2A/H2B (15–17). Xenopussperm are the only known cells in which RCC1 is not predominantly chromatin-bound (18).

When added to egg lysate, Xenopus sperm undergo two decondensation reactions (19). During stage I decondensation (∼10 min), nucleoplasmin replaces sperm-specific basic proteins (X and Y) with H2A/H2B (16, 17). The addition of a membrane fraction permits stage II decondensation which includes pronucleus formation. Because RCC1 is required for nuclear envelope assembly (1, 2), it must mobilize to chromatin during either stage I or early stage II decondensation. The level of endogenous RCC1 on demembranated sperm chromatin was too low to be detectable; however, within 10 min after the addition of egg extract, RCC1 relocalized to the nuclear pellet (Fig. 3A, top) (20). To observe chromatin binding of RCC1, we used RCC1 tagged with green fluorescent protein (RCC1-GFP). Binding of RCC1-GFP to sperm chromatin was detectable within 10 min after addition of the egg extract, but not the buffer alone (20). These data demonstrate that mobilization of RCC1 to chromatin occurs during stage I decondensation.

Figure 3

Binding of RCC1 to chromatin uses histones H2A/H2B. (A) Xenopus sperm nuclei were incubated with either buffer or a high-speed supernatant from aXenopus egg lysate for 10 min. After washing, nuclei (∼10,000 per lane) were subjected to SDS-PAGE and immunoblotting against endogenous RCC1 (top). Alternatively, RCC1-GFP was added. After staining with DAPI, samples were analyzed by fluorescence microscopy. (B) Sperm nuclei were incubated with or without NplC with or without human histones. Scale bar, 5 μm. (C) Total nuclear fluorescence was quantified with Openlab software (Improvision). Error bars represent ±1 SD from the mean (n = 20 to 25). (D) Sperm nuclei (∼250,000) were treated as above and subjected to a two-dimensional gel electrophoresis (17). The first dimension (left to right) consisted of a 17.5% acrylamide Triton X-100/acetic acid/urea gel; the second (top to bottom) was a 15% SDS-PAGE gel. Protein was stained with Coomassie.

Decondensation may allow access to preexisting RCC1-binding sites on DNA; alternatively, addition of egg lysate may create RCC1-binding sites by depositing H2A/H2B onto chromatin. To differentiate between these possibilities, we separated chromatin decondensation from H2A/H2B incorporation and observed RCC1-GFP binding (21). When treated with the buffer alone, sperm remained condensed (Fig. 3B). However, addition of nucleoplasmin core (NplC) decondensed nuclei without promoting RCC1-GFP binding (Fig. 3, B and C). This evidence supports our conclusion that neither DNA nor endogenous H3/H4 are sufficient to dock RCC1 to chromatin (Fig. 3, B through D). When histones were added to condensed chromatin, a small increase of RCC1-GFP binding was observed. Fluorescence appeared in discrete foci along the chromatin periphery, suggesting that RCC1 could access only a fraction of the condensed chromatin. In contrast, treatment with NplC plus histones produced a ∼25-fold increase in RCC1-GFP binding. Purified H2A/H2B stimulated chromatin binding to a similar extent as the histone extract (14).

We next compared protein constituents of the chromatin under each experimental condition. By two-dimensional gel electrophoresis (17), the level of H2A/H2B deposition reflected the level of RCC1-GFP binding. In addition, binding of RCC1 occurred during the same time frame as H2A/H2B incorporation. Together, these data suggest that the interaction between RCC1 and chromatin requires the integration of histones H2A/H2B. This observation, together with both the binding exhibited by RCC1 for H2A/H2B in vitro and the stimulation of RCC1 by nucleosomes, strongly suggests that RCC1 binds H2A/H2B on chromatin, possibly at surfaces exposed on the faces of each nucleosome.

  • * To whom correspondence should be addressed. E-mail: men5w{at}virginia.edu

  • Present address: University of Virginia, Hospital West RM 7191, Post Office Box 800577 HSC, Charlottesville, VA 22908.

REFERENCES AND NOTES

View Abstract

Navigate This Article