The N-terminal tail of histone H2A binds to two distinct sites within the nucleosome core

Abstract

Each of the core histone proteins within the nucleosome has a central “structured” domain that comprises the spool onto which the DNA superhelix is wrapped and an N-terminal “tail” domain in which the structure and molecular interactions have not been rigorously defined. Recent studies have shown that the N-terminal domains of core histones probably contact both DNA and proteins within the nucleus and that these interactions play key roles in the regulation of nuclear processes (such as transcription and replication) and are critical in the formation of the chromatin fiber. An understanding of these complex mechanisms awaits identification of the DNA or protein sites within chromatin contacted by the tail domains. To this end, we have developed a site-specific histone protein–DNA photocross-linking method to identify the DNA binding sites of the N-terminal domains within chromatin complexes. With this approach, we demonstrate that the N-terminal tail of H2A binds DNA at two defined locations within isolated nucleosome cores centered around a position ≈40 bp from the nucleosomal dyad and that this tail probably adopts a defined structure when bound to DNA.

Approximately 75–80% of the mass of the core histone proteins is contained within a structured, primarily α-helical “histone fold” domain (1). These domains make multiple interlocking heterotypic interactions to form the spool onto which DNA is wrapped twice around within the nucleosome (1–3). The structures and interactions of the histone fold domains have been well documented (3). The remaining mass of each of the core histone proteins is contained within “tail” regions located primarily at the N-termini of these proteins. Although they exist outside of the canonical histone fold domain and are not required for the correct assembly of the histone octamer in vitro, the tail domains are highly evolutionarily conserved (4). However, the structures and molecular contacts made by the tail domains within the nucleosome are not well defined.

The histone tail domains contain a high proportion of basic residues, and thus it is assumed that they bind to DNA within chromatin (5). A free peptide representing the N-terminal tail domain of histone H4 binds DNA with high affinity (6). Indeed, the tail domains are bound within isolated nucleosome core particles (7–9) or chromatin (10) under low ionic strength conditions and are released and mobile in solutions containing moderate concentrations of salt (0.3–0.4 M) (10). Upon release, tail domain residues adopt random coil conformations and become extremely sensitive to proteolysis (8, 11). Although the tail domains contribute to the thermal stability of the nucleosome probably by providing additional countercharge to the polyanionic backbone of DNA (12), salt-dependent tail release or proteolytic removal of the tails results in only marginal changes in the conformation and hydrodynamic properties of the nucleosome core (13–17).

The core histone tail domains are necessary to direct the assembly of an extended nucleosomal array into the more compact ≈30-nm diameter chromatin fiber (18, 19). Although tail contributions to fiber compaction are partially due to electrostatic interactions between the tails and DNA (20, 21), recent evidence suggests that some of these tails may participate in protein–protein interactions within the condensed fiber (22, 23). In addition, the tails are thought to mediate associations of 30-nm fibers into higher order complexes perhaps related to the chromonema structures observed by light microscopy (24, 25). However, little is known about the specific nature of these important tail–protein or tail–DNA interactions or how these interactions may rearrange upon condensation of the chromatin fiber (22, 23).

Remodeling of chromatin structures and the associated changes in the accessibility of the underlying DNA are thought to be integral parts of nuclear processes, such as transcription and replication (26–28). Such remodeling is probably mediated in part through mechanisms involving the multiple posttranslational modifications that occur within the core histone tail domains (28–31). These modifications may alter the strength and locations of tail domain interactions with nucleosomal DNA. Tail functions also involve direct interactions with other proteins that have been shown to affect the transcriptional state of the underlying DNA sequence (32–34). Such interactions also may affect the efficiency of chromatin assembly in vivo or play a role in regulating progression through the cell cycle (35).

Clearly, an understanding of the complex mechanisms involving the core histone tails will require a precise accounting of the specific DNA or protein interactions made by these domains in all relevant chromatin contexts. Unfortunately, the positions of the tail domains are not defined in crystallographic studies of the nucleosome core or the histone octamer (1, 2). Furthermore, biochemical determination of the DNA binding sites of these domains has been difficult, and only partial characterization of a few tail domains has been accomplished. It has been established that a portion of the carboxyl-terminal tail of H2A can be cross-linked to the DNA near the dyad axis of symmetry within nucleosome core particles and near the periphery of core DNA within chromatin complexes that contain linker DNA (36, 37). In addition, histidine-18 within the N-terminal tail of H4 can be cross-linked to DNA ≈1.5 helical turns to either side of the dyad axis of symmetry within the nucleosome (38). However, definitive characterization of the sites of interaction for the remainder of these or other N-terminal tails is not yet available.

To directly assess the disposition of the N-terminal tail domains in a variety of chromatin contexts, we have developed an approach combining recombinant histone proteins with a site-specific photoinducible cross-linking technique (39–41). Here we investigate the location of the N-terminal tail of core histone H2A within an isolated mononucleosome in low ionic strength solution conditions. We find that this N-terminal tail makes two apparently distinct and precise sets of contacts with nucleosomal DNA.

MATERIALS AND METHODS

DNA Fragments.

Either the 154-bp EcoRI-RsaI fragment or the 215 bp EcoRI-DdeI fragment, which contain a Xenopus borealis somatic 5S RNA gene, was derived from the plasmid pXP10 and was used in the reconstitutions as indicated in the figure legends (16). Fragments were radioactively labeled at the 5′ or 3′ ends of individual strands as noted in the figure legends by standard techniques. Identical results were obtained with both fragments.

Preparation of Core Histones.

Recombinant Xenopus H2A and H2B were prepared and purified as pre-formed dimers as described (42). Coding sequences for these proteins containing cysteine substitutions were prepared by standard PCR techniques. Dimers containing H2AA12C or H2AG2C (see Fig. 1A) were reduced by incubation in 50 mM DTT in elution buffer (20 mM Tris⋅HCl, pH 8.0/2 mM EDTA/1.0 M NaCl) for 1–2 h at 25°C. Reduced proteins were diluted 2-fold with TE (10 mM Tris·HCl, pH 8.0/1 mM EDTA) and then bound to BioRex 70 (100≈200 mesh, Bio-Rad), and the column was washed with 10–15 bed volumes of 20 mM Tris⋅HCl, pH 8.0/2 mM EDTA/0.5 M NaCl before elution with the same buffer containing 1.0 M NaCl. Proteins were stored in elution buffer at −80°C.

Figure 1.

Schematic of APB modification. (A) Position of cysteine substitutions within the N-terminal domain of H2A. (B) Chemical structure of APB.

Modification of H2A with 4-Azidophenacylbromide (APB).

Approximately 5 nmols of fully reduced H2AA12C or H2AG2C was reacted with a 2-fold molar excess of APB (Sigma) for 1 h in the dark at room temperature. The degree of modification was monitored by subsequent reaction of a portion of the reaction with a molar excess of ¹⁴C-labeled N-ethylmaleimide to label any remaining free sulfhydryls. The ¹⁴C N-ethylmaleimide labeling reaction was quenched by adjusting the solution to 5 mM DTT, and the samples were loaded onto SDS/PAGE gels. After staining with Coomassie blue the gels were dried and exposed in Kodak MR5 autoradiographic film.

Reconstitution of Nucleosomes.

Native core histones H2A/H2B and H3/H4 were prepared from chicken erythrocyte nuclei, nucleosomes were reconstituted, and products were analyzed as described (43). Reconstitutions with the DNA fragments used in this study yielded nucleosomes in which the dyad axis of symmetry was located at position −3 relative to the start site for transcription of the 5S gene (+1).

Photocross-linking and Alkali Cleavage.

Approximately 100 pmols of reconstituted nucleosomes were separated on preparative 0.7% agarose gels, and octamer complexes were visualized by exposure of the wet gel to film for 3 h. Gel slices containing nucleosome complexes were then irradiated for 20 s with a VWR Scientific LM20E Transilluminator set at 365 nm. Irradiated complexes were loaded onto preparative SDS/PAGE gels and run at 80 mA for 4 h. The gel was exposed to Kodak X-Omat AR film for 3 h, and the cross-linked complexes were gel isolated. DNA from irradiated complexes was purified and ethanol-precipitated, the DNA was resuspended in 25 mM NHOAc/0.1M EDTA/2% SDS, and the samples were incubated at 90°C for 30 min. Then, 5 μl of 2 M NaOH was added into the samples and incubated at 90°C for 1 h. The alkali treatment was stopped with the addition of 6.5 μl of 2 M HCl and 100 μl of 20 mM Tris⋅HCl (pH 8.0). Samples were then ethanol-precipitated, dried, and resuspended in gel loading dye (0.02% bromophenol blue/0.02% xylene cyanol/99% formamide). Samples were loaded onto 6% polyacrylamide sequencing gels and run at 55 watts for 1 h. Gels were then dried and autoradiographed on Kodak X-Omat AR film.

Cross-linking Within Nucleosome Core Particles.

Approximately 5 μg of calf thymus DNA was reconstituted with native histones and the APB-modified H2A mutants. Reconstituted material was then digested with 0.02 units of micrococcal nuclease in the presence of 2 mM CaCl₂ at 37°C for 10 min. The reaction was stopped with 25 mM EDTA and 2.5% glycerol and directly loaded onto a 0.7% agarose gel. After separation, the entire gel was then irradiated with UV light as above. The core particle band was identified by ethidium bromide staining and brief UV illumination. The cross-linked core DNA was purified, 5′ end-labeled, and treated with 0.1% SDS and 1.5 μg of proteinase K. The 147-bp core particle DNA was isolated from an 8% nondenaturing acrylamide gel, and sites of cross-linking were detected as above.

RESULTS

To attach a photoactivatable cross-linking reagent to a particular site within the N-terminal tail of histone H2A, we generated two mutant proteins in which either the 2nd or 12th amino acid residues were substituted for cysteine to create proteins H2AG2C and H2AA12C, respectively (Fig. 1A). The positions of modification were chosen because they lie at opposite ends of the protease-sensitive N-terminal domain of H2A. Native histone H2A from Xenopus borealis lacks cysteine residues, and thus this residue is a convenient target for site-specific modifications. Sulfhydryl groups within cysteine-substituted H2As were modified with the bifunctional cross-linking reagent, APB (Fig. 1B) (38–41). Upon UV irradiation, the photoactivatable phenylazide moiety formed a highly reactive nitrene intermediate, resulting in covalent bond formation with any protein or DNA in close vicinity (≈10 Å) to the modified cysteine. The level of APB modification was determined by subsequent reaction with ¹⁴C N-ethylmaleimide and reveals that at least 95% of the H2A mutants are modified with APB (Fig. 2A).

Figure 2.

(A) Extent of APB modification of cysteine-substituted H2A. H2AA12C/H2Bwt dimers were incubated in the presence or absence of APB as indicated and subsequently reacted with ¹⁴C-labeled N-ethylmaleimide . Proteins used for the reconstitutions were separated on SDS/PAGE, and the gel was stained with Coomassie blue and then autoradiographed, as indicated. (B) Nucleoprotein gel analysis of irradiated nucleosomes. Nucleosomes containing APB-modified H2A were either loaded onto the gel (lane 2) or first irradiated and then loaded (lane 3). The positions of naked DNA (lane 1) and the nucleosome complex within the gel are indicated. (C) Protein–DNA cross-linking is dependent on APB modification and UV irradiation. Nucleosomes containing labeled DNA and either APB-modified or unmodified H2A were irradiated with UV light, products were separated by SDS/PAGE, and the gel was autoradiographed. Lanes: 1, UV-irradiated wild-type nucleosomes; 2, unirradiated APB-modified nucleosomes; 3, APB-modified and UV-irradiated nucleosomes. The positions of uncross-linked naked DNA and protein–DNA cross-linked products are indicated.

APB-modified H2A was used for nucleosome reconstitution with wild-type histones H2B, H3, H4, and 5S DNA radiolabeled on the 5′ end of the bottom strand at the EcoRI site at +75 relative to the transcription start site. A portion of the reconstitution was run on a 0.7% agarose nucleoprotein gel to confirm the integrity of the nucleosomes. Reconstitutions with recombinant histones generate nucleosomes indistinguishable from those containing native histones isolated from chicken erythrocyte nuclei (42). Furthermore, APB modification and UV irradiation of APB-modified proteins have no effect on the mobility of the nucleosome complex within a nucleoprotein gel, suggesting that no large changes in conformation or loss of proteins from the complex is induced by the cross-linking procedure (Fig. 2B). In addition, hydroxyl radical footprinting reveals no change in the structure of the DNA within nucleosomes containing modified H2A mutant proteins, either before or after irradiation (results not shown).

To assess the appearance of protein–DNA cross-linking within the nucleosome complex, we irradiated reconstitutes with UV light, added SDS, and separated the constituents by SDS/PAGE. Nucleosomes reconstituted with radioactively end-labeled DNA and wild-type histones yielded only a single band on an autoradiograph of the gel corresponding to naked DNA, even after UV irradiation of these complexes (Fig. 2C, lane 1, and results not shown). However, nucleosomes containing an APB-modified H2A yielded a higher M_r complex on the gel that was dependent on UV irradiation of the sample (Fig. 2C, lanes 2 and 3). Furthermore, irradiation of a nucleosome containing a wild-type H2A that had been treated with APB in the same manner as the mutant proteins did not yield a higher M_r product on the gel (results not shown). Densitometric analysis revealed that the efficiency of cross-linking was ≈5%.

To identify the location of the cross-links, UV-irradiated nucleosomes were isolated from preparative nucleoprotein gels, and then cross-linked products were isolated by SDS/PAGE denaturing gels. DNA was recovered from the gel and incubated in high heat (90°C) and alkali to induce strand breaks at the site of the cross-linking. Analysis of the cleavage pattern from irradiated nucleosomes containing APB-modified H2AA12C when the top strand of 5S DNA was 5′ end-labeled revealed a single strong signal at ≈+39, ≈40 bp away from the dyad axis of symmetry within the 5S nucleosome (located near the start site for transcription at +1). (Fig. 3, Top, lane 3). A control lane derived from nucleosomes containing only wild-type histones treated in the same way revealed only low intensity background signals (Fig. 3, Top, lane 2). Cross-links made to the bottom strand of 5S DNA by nucleosomes containing this mutant are shown in the bottom gel in Fig. 3. Clearly, a strong signal at approximately position −36 within 5S DNA is observed. Note that the two sets of cross-links detected on complementary strands are approximately symmetry-related about the dyad axis of the nucleosome (see Discussion). Controls in which modified H2AA12C/H2B dimer was added directly to the labeled DNA and irradiated show that cross-links are scattered throughout the 5S DNA fragment (not shown).

Figure 3.

Location of cross-links between APB-modified H2AA12C and nucleosome DNA. Nucleosomes were reconstituted with the 215-bp EcoRI-DdeI fragment 5S DNA fragment and core histones including either APB-modified H2AA12C or wild-type H2A. Cross-linking was carried out and the DNA analyzed on sequencing gels as described in Materials and Methods. (Top) Samples from nucleosomes reconstituted with the EcoRI-DdeI DNA fragment 5′-end labeled at the EcoRI (labels the top strand). Lanes: 1, G-specific sequencing reaction markers; 2 and 3, cross-linking within nucleosomes containing wild-type H2A and APB-modified H2AA12C, respectively. (Bottom) As in the top gel except nucleosomes were reconstituted with the EcoRI-DdeI DNA fragment 5′-end labeled at the DdeI site (labels the bottom strand). Lanes: 1, G-specific markers; 2, cross-linking within nucleosomes containing APB-modified H2AA12C.

To identify the location(s) of the N-terminal-most portion of the H2A N-terminal tail domain, we assembled nucleosomes containing APB-modified H2AG2C (Fig. 1) and treated them in the same manner as above. The cleavage from such complexes when the top strand of 5S DNA was 5′ end-labeled revealed a general enhanced cross-linking reactivity to DNA relative to the H2AA12C (Fig. 4). This is probably because of the greater motility of the outer end of the tail domain. Specifically, a strong signal is observed at position +34, a weaker signal is observed at +43, and additional signals are observed at positions −32 and −49/−50/−51 (Fig. 4, Top). Signals on the complementary strand of 5S DNA reveal strong signals at +45 +36, −31, and −42 (Fig. 4, Bottom). Some minor bands that vary in intensity from experiment to experiment are also observed in the control lanes and are thus not due to specific cross-link formation. On average, cross-links due to APB-modified H2AG2C were found on both strands of DNA ≈35 and 45 bp away from the dyad axis. These sites are located ≈5 bases to either side of the sites of cross-linking identified for H2AA12C.

Figure 4.

Location of cross-links between APB-modified H2AG2C and nucleosome DNA. Samples were reconstituted with the 154 bp EcoRI-RsaI 5S DNA fragment and prepared as in Fig. 3. (Top) samples from nucleosomes reconstituted with the EcoRI-RsaI DNA fragment labeled at the 5′ end of the top strand at the EcoRI site. Lanes: 1, G-specific sequencing reaction markers; 2, hydroxyl radical footprint of the nucleosomes; 3 and 4, cross-linking within nucleosomes containing wild-type H2A and APB-modified H2AG2C, respectively. (Bottom) As in the top gel except the DNA was labeled at 3′ end of the bottom strand of the EcoRI-RsaI DNA fragment at the EcoRI site. Lanes 1 and 2, products from nucleosomes containing APB-modified H2AG2C and wild-type H2A, respectively.

To insure that the results obtained with our 5S nucleosome complexes were not influenced by sequence selectivity of the cross-linking reagent or any sequence-dependent structural variations within the 5S nucleosome, we repeated the cross-linking experiments with a standard preparation of nucleosome core particles containing random sequences. Nucleosomes containing either wtH2A, H2AG2C, or H2AA12C were reconstituted onto bulk DNA isolated from calf thymus and core particles prepared by careful digestion with micrococcal nuclease (4). Cores were isolated by preparative nucleoprotein gel electrophoresis and irradiated, and the positions of cross-links were identified as above. A clear, single site of cross-linking was found in cores containing APB-modified H2AA12C at a position ≈40 bp away from the dyad axis of symmetry (Fig. 5, lane 3), in excellent agreement with the results obtained above. When cores containing APB-modified H2AG2C are used in the experiment, cross-links are found at positions ≈35 and 45 bp distant from the dyad, with a clear preference of the latter (Fig. 5, lane 2). Note that, because of the size heterogeneity associated with the core particle preparation, the cross-linking signal is distributed over a larger range of DNA sizes than is observed in previous experiments using 5S nucleosomes. Furthermore, because of the random sequence nature of the DNA fragments, bands differing by 1 nt in length are not resolved on the sequencing gel in samples subjected to the base-elimination chemistry (Fig. 5, lanes 1–3).

Figure 5.

Location of cross-linked sites with random sequence core particles containing APB-modified H2AG2C or H2AA12C. Cross-linked products from core particles were prepared and analyzed by sequencing gel electrophoresis and autoradiography as described in Materials and Methods. (A) Autoradiograph of products from core particles containing wild-type H2A (lane 1), H2AG2C (lane 2), and H2AA12C (lane 3). The position of bands is indicated relative to the distance from the dyad axis of symmetry in nucleotides. Only one symmetrical half of the nucleosome core is shown as indicated by the schematic (Left). The hydroxyl radical footprint of nucleosome core particles is shown in lane 4 for reference. (B) Densitometric scans of gel in A. Distance from dyad axis of symmetry is indicated as in A.

DISCUSSION

We used a site-specific cross-linking approach to identify regions of nucleosomal DNA contacted by opposite ends of the N-terminal tail domain of histone H2A. Our approach allowed the position of cross-links due to a probe attached to specific residues within the tail domain to be identified. The H2A tail domain spans approximately residues 1–12 within the protein, as defined by sensitivity to trypsin proteolysis (11). In addition, H2A residues 1–14 apparently are disordered in the crystal structure of the core histone octamer, which was solved in the absence of DNA (1). Results of our cross-linking experiments clearly show two groups of signals, symmetrically related to each other about the dyad axis of symmetry within the nucleosome. Because there are two symmetrically related N-terminal tails of H2A within each nucleosome, each situated near one set of contacts, we assume that each group of cross-links is due to contacts by a single tail of H2A. A summary of the cross-links attributed to a single tail within the nucleosome is shown in Fig. 6A.

Figure 6.

Summary of cross-links. (A) Linear representation of cross-linking by one H2A N-terminal tail domain within 5S (Upper) or random sequence (Lower) nucleosome cores. Sites of cross-linking from cross-linking probe positioned at the second (H2AG2C) or 12th amino acid residue (H2AA12C) are indicated by the filled or open arrows, respectively. Distance in base pairs from the center (dyad) of the nucleosome is indicated. (B) Positions of cross-links by one H2A N-terminal tail domain within one superhelical turn of nucleosomal DNA. Top view of the nucleosome showing the locations of cross-links as in A. Only protein and DNA in the top half of the nucleosome are shown. Core histone α-helices (2) are represented by columns and other secondary structures as thin tubes as in Felsenfeld (27). The mobile histone tail domains are indicated by the dashed lines, and residues within the N-terminal tail of H2A are indicated by 3.5-Å spheres. The positions of the 2nd and 12 residues are indicated by the closed and open spheres, respectively. Histones H3, H4, H2B, and H2A are shaded dark to light gray, respectively. Note that a small portion of H3 from the bottom half of the octamer is shown. Numbers indicate distance from nucleosomal dyad as in A.

The cross-linking probe was attached at either the very beginning of the tail (H2AG2C) or at the position in the tail closest to the structured domain (H2AA12C). When the probe was attached to the 12th position, a single site of cross-linking was detected ≈40 bp to either side of the dyad, primarily to one of the two strands of DNA. In contrast, when the probe was attached to the beginning of the tail (H2AG2C), we observed two sets of cross-links, ≈5 bp symmetrically disposed to either side of the single cross-linking site identified with H2AA12C. The data suggest that the innermost portion of the tail is sufficiently anchored by the structured domain such that DNA contacts are restricted to a single location. As might be expected, the free end of the N-terminal clearly has much greater mobility. Surprisingly, however, the cross-linking experiments indicate that the conformational space occupied by the free end of the tail when bound to DNA is also somewhat restricted such that this end of the tail adopts either of two defined positions located ≈9–10 bp apart (Fig. 6).

The location of the cross-links agrees well with the expected location of the H2A N-terminal tail based on the location of DNA fitted onto the surface of the histone octamer structure determined by x-ray crystallography (3). Arginine 15 is the first residue in the N-terminal region of H2A imaged in the structure of the octamer and is located very near the cross-links found when APB is attached to the 12th residue within this tail (H2AA12C) (Fig. 6B). The tail clearly emanates outward from a single fixed location near position 40 and contacts five additional bases to either side of this point.

The specificity and pattern of cross-links suggest that the N-terminal tail of H2A adopts a unique structure in either of two orientations when bound to DNA. In a fully extended conformation of the N-terminal tail, the distance between position H2AA12C and H2AG2C is ≈35 Å, enough to cover ≈10 bp of DNA (Fig. 6B). The actual distance between cross-links at the extremes of the N-terminal tail of H2A is ≈5 bp, suggesting that these residues are organized into a structure containing ≈1.5 Å per residue when bound to DNA. Additionally, previous studies have shown that the core histone N-terminal tails adopt random coil conformations when released from DNA (9). If the tails were to bind DNA in a similar random fashion or with multiple conformations, a much less localized distribution of cross-links would be expected in the above experiments.

Previous studies are consistent with the interpretation that the N-terminal tail of H2A can equilibrate between two bound positions within the nucleosome core. Methylation protection experiments by Hill and Thomas (44) were unable to observe any protection of lysines in the N-terminal tail of H2A, suggesting that this tail undergoes a facile equilibrium between states, which include a solution-exposed state (44). In addition, NMR spectra of nucleosome core particles at low salt exhibit a residual intensity that suggests that a portion of the N-terminal tail of H2A is relatively mobile (7, 9). Thus, this tail may equilibrate between two fixed locations via a solvent-exposed state.

The location of the N-terminal tail of H2A we have identified corresponds to the location of the H2A and H2B tails predicted from thermal denaturation studies. Ausio et al. (12) determined that these tails influence the thermal stability of ≈10 bp of DNA located ≈40–50 bp from the dyad axis of symmetry, in excellent agreement with our results. Of interest, Mirzabekov and coworkers identified a protein–DNA cross-link ≈58 bp distant from the dyad axis as caused by the extended N-terminal tail of a sea urchin sperm-specific histone H2B variant (45). These results suggest that the H2A tail contacts DNA between +35 and +45 and H2B interacts with DNA between +45 and +55 from the dyad, in good agreement with current models of the nucleosome core (Fig. 6B). Experiments to determine if the shorter N-terminal tail of the major mammalian H2B species contacts DNA in the same area as the tail of the sperm variant are in progress.

Recent studies have revealed that the histone tail domains are not only involved in mediating folding of the chromatin fiber but are also key components in the regulation of cellular events such as transcription and replication (26, 28). Tails undergo numerous posttranslational modifications, making them potentially important regulatory switches of cell processes that probably involve modulation of the molecular contacts made by the tail domains (23). It will be interesting to see if the DNA contacts we find for the N-terminal tail of H2A within a mononucleosome at low salt are maintained or changed within an array of nucleosomes when an array is condensed to form a solenoid structure or affected by the binding of linker histones (46).

ABBREVIATION

APB

4-azidophenacylbromide