ABSTRACT
DNA shape recognition determines the preferred binding sites for sequence-independent DNA binding proteins, and here we document that archaeal histones assemble archaeal nucleosomes in vitro centered preferentially within (CTG)6 and (CTG)8repeats, close to junctions with flanking mixed-sequence DNA. Archaeal nucleosomes were not positioned by (CTG)4-, (CTG)5-, or (CTG)3AA(CTG)3-containing DNA sequences. The features of CTG repeat-containing sequences that direct eucaryal nucleosome positioning may also be similarly recognized by archaeal histones.
Several different families of small, abundant, positively charged DNA binding proteins in prokaryotes have been described (13). DNA binding by these proteins is usually described as sequence independent, although binding to genomic DNA in vivo is unlikely to be completely random and some DNA sequences have been shown to act as preferred binding sites. Members of the bacterial HU family bend DNA (15), and the archaeal histones wrap DNA (10); therefore these proteins are expected to assemble complexes preferentially at locations where the DNA sequence inherently has, or readily adopts, a shape that facilitates bending or wrapping. Consistent with this, a consensus sequence has been established for integration host factor (IHF) binding (5), even though as a member of the HU family IHF does not bind to DNA through specific base pair contacts (11). The consensus sequence actually reflects the ease with which certain dinucleotides accept the helix distortions that result from IHF binding (11). By exploiting this shape-determined localized binding, IHF plays a direct role in regulating specific gene expression, and similarly, in Eucarya, the positioned assembly of histones into nucleosomes is a major factor in regulating gene expression. It has been established that it is the histone (H3+H4)2tetramer that recognizes nucleosome positioning signals (3, 7, 14), and a hierarchy of eucaryal nucleosome positioning sequences has been established. High on this list are tandem repeats of the trinucleotide CTG (4), and here we demonstrate that CTG repeat-containing sequences similarly direct the localized assembly of archaeal histones into archaeal nucleosomes. We have previously established that archaeal nucleosomes have a histone tetramer-containing structure very similar to that formed by the eucaryal histone (H3+H4)2 tetramer at the center of the eucaryal nucleosome (8, 9), and the results reported here extend this similarity to the recognition of positioning signals in the DNA.
The templates used to map archaeal nucleosome assembly sites were generated by the hybridization of complementary oligonucleotides that formed double-stranded DNA molecules with single-strand extensions complementary to EcoRI and BamHI half-sites. These molecules were ligated with EcoRI- andBamHI-digested pLITMUS28 (New England Biolabs, Mass.), and the constructions confirmed by DNA sequencing were found to contain (CTG)4, (CTG)5, (CTG)6, (CTG)8, and (CTG)3AA(CTG)3 inserts. With these plasmids as templates and the primers indicated in Fig.1A, Taq-based PCRs in the presence of [α-32P]dATP generated the [32P]DNA molecules used for archaeal nucleosome assembly.
(A) Sequence of the multiple cloning site region of pLITMUS28 with the (CTG)6 insert indicated. Lowercase letters identify nucleotides in the pLITMUS28 vector that are replaced in the insert-containing DNA. The positions and sequences of the oligonucleotides used as primers in PCRs to generate the substrate DNAs are indicated by the converging arrows. Asterisks identify the midpoints of the MN-protected fragments of the (CTG)6-containing substrate. The RE sites used as positioning references are shown beneath their recognition sequences. (B) Products of MN digestion of rHMfB-containing nucleosomes assembled on the 105-bp (CTG)6-containing substrate (left) and on the 96-bp substrate from the pLITMUS28 vector (right). Aliquots were taken before MN treatment (0) and then at 4-min intervals. (C) Products of MN digestion of (CTG)6-containing nucleosomes assembled by using rHMfB (left) and rHMfA (right). Aliquots were taken before MN digestion (0) and then at 10-min intervals. M, DNA size standards that differ by 10 bp, with the intense triplet band centered at 100 bp.
DNA and recombinant archaeal histones, purified from Escherichia coli (12), were mixed at a 1:3 mass ratio in 50 mM Tris-acetate (pH 8.8)–1 mM CaCl2 and incubated briefly at 37°C before the addition of 0.003 U of micrococcal nuclease (MN)/μg of DNA. MN digestion was terminated by addition of EDTA (10 mM final concentration; pH 8) and an equal volume of 50% phenol–50% chloroform. The aqueous phase was removed and extracted with chloroform, and DNA molecules were precipitated by the addition of 3 volumes of ice-cold ethanol, 0.3 M sodium acetate, and 1 mM MgCl2.
The DNA fragments which were protected from MN digestion [by incorporation into archaeal nucleosomes formed by recombinant HMfB (rHMfB) by using the 105-bp (CTG)6-containing substrate DNA] decreased in length with increasing time of MN digestion, and two populations of kinetic intermediates with lengths of ∼84 and ∼70 bp accumulated (Fig. 1B). Accumulation of these shorter MN-protected fragments was also observed when the assembled archaeal nucleosomes contained rHMfA (Fig. 1C), rHFoB, or rHPyA1 (results not shown), recombinant versions of related histones from Methanothermus fervidus, Methanobacterium formicicum, andPyrococcus strain GB-3a (10), respectively. In contrast, MN digestions of archaeal nucleosomes assembled by archaeal histones with a 96-bp DNA substrate amplified from pLITMUS28 (lacking a CTG repeat insert) (Fig. 1B), or with the substrates that contained (CTG)4, (CTG)5, or (CTG)3AA(CTG)3 inserts (data not shown), resulted in a gradual decrease in the abundance of starting-length DNA molecules but no accumulation of shorter molecules. In the absence of archaeal histones, incubation of the DNA with MN resulted in rapid and complete digestion of the DNA with no detectable accumulation of intermediates. MN digestion of archaeal nucleosomes assembled with the (CTG)8-containing substrate did result in the protection and accumulation of shorter kinetic intermediates (data not shown). MN-protected fragments derived from mixed-sequence DNA by incorporation into archaeal nucleosomes average ∼60 bp in length; however, cloning and sequencing of such protected fragments have revealed a range of lengths, including molecules close to 70 bp (6), as is the case for the CTG repeat-containing protected fragments. These differences presumably reflect inherent structural properties, unique to individual DNA sequences, that influence archaeal histone interactions.
To determine if rHMfB-containing nucleosomes were positioned, unlabeled populations of the ∼70-bp MN-protected fragments originating from the (CTG)6- and (CTG)8-containing substrates were purified from nondenaturing 8% (wt/vol) polyacrylamide gels, labeled at the 5′ end with 32P, and subjected to restriction endonuclease (RE) digestions. High-resolution electrophoresis through 10% polyacrylamide–8.3 M urea gels in 0.5× Tris-borate-EDTA buffer revealed that the majority of the protected fragments were 66 to 71 bp in length, with a small minority ranging from 55 to 75 bp (Fig.2, lane 2). With random nucleosome positioning, all possible ∼70-bp sequences from the substrate DNA molecule would be equally represented in the population of protected fragments, whereas localized positioning would protect defined sequences and sequences not incorporated into nucleosomes would be lost. Following random positioning, RE digestion of the protected fragments would generate restriction fragments with all lengths from 1 to 70 bp, whereas RE digestion of fragments protected by localized assembly would result in defined fragments or no cutting if the RE site was not within the protected region. As illustrated in Fig. 2, the MN-protected fragments from archaeal nucleosomes, assembled on the (CTG)6-containing molecule, were shown by RE digestion to be predominantly mixtures of two populations of discrete molecules: one of fragments 66 bp in length and the other of molecules ranging from 69 to 71 bp in length. BsiWI, XhoI,EcoRI, and HindIII digestions generated restriction fragments with lengths of <10 and 54 bp, 15 and 47 bp, 21 and 41 bp, and 58 bp, respectively (Fig. 2, lanes 3 to 6). This is consistent with a substrate molecule 66 bp in length (the sums of the two fragment sizes plus the restriction sites) that formed the region of the (CTG)6-containing molecule covered by the leftmost oval of the diagram in Fig. 2. There is no NcoI site in this region, and NcoI digestion did not cleave the 66-bp MN-protected fragments (Fig. 2, lane 7). XhoI,EcoRI, and HindIII digestions of the starting mixture of molecules also generated restriction fragments of 61 to 63 bp, 55 to 57 and 9 to 11 bp, and 18 to 20 and 46 to 48 bp, respectively, and NcoI digestion generated 11- to 13- and 53- to 55-bp fragments (Fig. 2, lanes 4 to 7). Based on these products, a second population of MN-protected fragments that overlapped, differed in length by 1 or 2 bp, and originated from the region of the (CTG)6-containing molecule covered by the three rightmost ovals in Fig. 2 was present. The two predominant positions at which archaeal nucleosomes assembled on the (CTG)6-containing substrate are centered within the (CTG)6 sequence but close to the ends of this sequence (Fig. 1). RE digestions of MN-protected fragments from the (CTG)8-containing insert confirmed that archaeal nucleosomes assembled on this substrate similarly centered near the insert-vector junctions (data not shown). AsHindIII and NcoI digestions of the (CTG)6- and (CTG)8-containing fragments also generated restriction fragments with lengths of 20 to 50 bp (Fig. 2, lanes 6 and 7), a minority of archaeal nucleosomes must therefore have also assembled, and protected DNA, to the right of the two major positions indicated in Fig. 2.
High-resolution electrophoretic separation of RE-generated fragments. RE digestion products of ∼70-bp MN-protected fragments of the (CTG)6-containing substrate, separated by electrophoresis, are shown adjacent to a ladder of nucleotide size standards in lane 1 and the undigested (U) MN-protected fragments in lane 2. In lanes 3 to 7, the restriction fragments were generated by digestion with BsiWI (B), XhoI (X),EcoRI (E), HindIII (H), and NcoI (N), respectively. On the map of the (CTG)6-containing substrate shown above the electropherogram, ovals identify the preferred locations for archaeal nucleosome assembly, with RE cleavage sites and the (CTG)6 sequence shown as boxes.
CTG repeats are strong eucaryal nucleosome positioning elements (4), and we have shown that they also localize the assembly of archaeal nucleosomes. As with eucaryal nucleosomes, a minimum of six CTG tandem repeats was required for archaeal nucleosome positioning. However, whereas only one preferred site was identified for eucaryal nucleosome positioning, centered on the last G of the CTG repeat (4), archaeal nucleosomes assembled at two preferred positions, centered within the repeats but near the junctions with the flanking DNA (Fig. 1). The site at which eucaryal nucleosomes assembled preferentially was established by reference to only one RE site in a natural DNA sequence, and with this limitation, it was noted that additional preferred assembly sites might not have been detected (4).
The basis for nucleosome positioning by CTG repeats is unclear. CTG repeats are more flexible and exhibit more writhe than mixed-sequence B-form DNA (1, 2), and therefore the juxtaposition of such a flexible region to a less flexible region may generate a structure intrinsically favorable for histone wrapping of DNA into nucleosomes.
ACKNOWLEDGMENTS
We thank Suzette Pereira for advice and discussion and Ekta Sirohi for assistance in plasmid constructions.
This work was supported by NIH grant GM53185.
FOOTNOTES
- Received 19 October 1998.
- Accepted 13 November 1998.
- Copyright © 1999 American Society for Microbiology
