SAF-Box, a conserved protein domain that specifically recognizes scaffold attachment region DNA

SARs (scaffold attachment regions) are candidate DNA elements for partitioning eukaryotic genomes into independent chromatin loops by attaching DNA to proteins of a nuclear scaffold or matrix. The interaction of SARs with the nuclear scaffold is evolutionarily conserved and appears to be due to specific DNA binding proteins that recognize SARs by a mechanism not yet understood. We describe a novel, evolutionarily conserved protein domain that specifically binds to SARs but is not related to SAR binding motifs of other proteins. This domain was first identified in human scaffold attachment factor A (SAF-A) and was thus designated SAF-Box. The SAF-Box is present in many different proteins ranging from yeast to human in origin and appears to be structurally related to a homeodomain. We show here that SAF-Boxes from four different origins, as well as a synthetic SAF-Box peptide, bind to natural and artificial SARs with high specificity. Specific SAR binding of the novel domain is achieved by an unusual mass binding mode, is sensitive to distamycin but not to chromomycin, and displays a clear preference for long DNA fragments. This is the first characterization of a specific SAR binding domain that is conserved throughout evolution and has DNA binding properties that closely resemble that of the unfractionated nuclear scaffold.     

The novel SAR-binding domain of scaffold attachment factor A (SAF-A) is a target in apoptotic nuclear breakdown.

The scaffold attachment factor A (SAF-A) is an abundant component of the nuclear scaffold and of chromatin, and also occurs in heterogeneous nuclear ribonucleoprotein (hnRNP) complexes. Evidence from previous experiments had suggested that SAF-A most likely has at least two different functions, being involved both in nuclear architecture and RNA metabolism. We now show that the protein has a novel scaffold-associated region (SAR)-specific bipartite DNA-binding domain which is independent from the previously identified RNA-binding domain, the RGG box. During apoptosis, but not during necrosis, SAF-A is cleaved in a caspase-dependent way. Cleavage occurs within the bipartite DNA-binding domain, resulting in a loss of DNA-binding activity and a concomitant detachment of SAF-A from nuclear structural sites. On the other hand, cleavage does not compromise the association of SAF-A with hnRNP complexes, indicating that the function of SAF-A in RNA metabolism is not affected in apoptosis. Our results suggest that detachment of SAF-A from SARs, caused by apoptotic proteolysis of its DNA-binding domain, is linked to the formation of oligonucleosomal-sized DNA fragments and could therefore contribute to nuclear breakdown in apoptotic cells.     

Localization and dynamics of small circular DNA in live mammalian nuclei

While genomic DNA, packaged into chromatin, is known to be locally constrained but highly dynamic in the nuclei of living cells, little is known about the localization and dynamics of small circular DNA molecules that invade cells by virus infection, application of gene therapy vectors or experimental transfection. To address this point, we have created traceable model substrates by direct labeling of plasmid DNA with fluorescent peptide nucleic acids, and have investigated their fate after microinjection into living cells. Here, we report that foreign DNA rapidly undergoes interactions with intranuclear structural sites that strongly reduce its mobility and restrict the DNA to regions excluding nucleoli and nuclear bodies such as PML bodies. The labeled plasmids partially co-localize with SAF-A, a well characterized marker protein for the nuclear ‘scaffold’ or ‘matrix’, and are resistant towards extraction by detergent and, in part, elevated salt concentrations. We show that the localization and the low mobility of plasmids is independent of the plasmid sequence, and does not require the presence of either a scaffold attachment region (SAR) DNA element or a functional promoter.     

Interaction of DNA with nuclear scaffolds in vitro

We have previously identified a number of specific DNA fragments called SARs (scaffold-associated regions) that are associated with the nuclear scaffold and define the basis of DNA loops. We demonstrate that cloned DNA fragments containing SAR sequences bind to nuclear scaffolds in vitro with the same specificity as have genomic SAR fragments. This specific interaction is observed with the biochemically complex type I scaffolds. These scaffolds are composed of the nuclear lamina proteins and a set of other proteins that forms the internal network of these structures. So-called type II scaffolds, which are composed primarily of the lamina proteins and lack the proteins of the internal network, do not bind the SAR fragments at a detectable level. Competition experiments show that different SARs share common structural elements and can bind to the same sites on the nuclear scaffold, although with different affinities. Moreover, the SAR binding sites appear to be evolutionarily conserved, as all the Drosophila SARs also bind with identical specificity to nuclear scaffolds derived from rat liver nuclei. These Sar interaction studies were carried out with lithium 3,5-diiodosalicylate-extracted nuclei. Interestingly, scaffolds prepared by high-salt extraction also bind the genomic and exogenously added SAR fragments specifically. However, the endogenous transcribed sequences, as opposed to the same fragments added as purified DNA, associate randomly with these scaffolds.     

Specificity of SAF-A and lamin B binding in vitro correlates with the satellite DNA bending state

There is evidence that Matrix Attachment Region (MAR)-binding proteins also bind satellite DNA (satDNA). The aim of the current work was to determine whether the major nuclear matrix (NM) MAR-binding proteins are able to recognize satDNAs of different locations and what DNA structural features are important for the recognition. In nuclei and NM, a number of the same polypeptides were recognized on a southwestern blot when MAR of immunoglobulin kappa gene (Ig kappa MAR) and pericentromeric (periCEN) satDNA fragments were used. However, the binding decreased dramatically when human and mouse CEN satDNA were used for the probes. After an NM extract was subjected to ion exchange chromatography, the main DNA-binding proteins were identified as SAF-A (scaffold attachment factor A) and lamin B. It was not possible to test the binding of lamin B by gel mobility shift assay (GMSA), but SAF-A showed an ability to distinguish CEN and periCEN satDNA fragments in GMSA. While periCEN fragments have an abnormally slow mobility on electrophoresis, which is a hallmark of bent DNA, CEN satDNA fragments have a normal mobility. A computer analysis was done using the wedge model (Ulanovsky and Trifonov [1987] Nature 326:720-722), which describes how the curved state depends on particular nucleotide sequences. The curved states of the fragments predicted by the model are in good agreement with their ability to be recognized by NM proteins. Thus SAF-A and lamin B are able to recognize conserved structural features of satDNA in the same way that MAR-binding proteins recognize MARs in spite of a lack of a consensus sequence. CEN and periCEN satDNAs are distinguished by proteins in correlation with the helical curvature of these fragments.     

hnRNP-U is a specific DNA-dependent protein kinase substrate phosphorylated in response to DNA double-strand breaks

Cellular responses to DNA damage are orchestrated by the large phosphoinositol-3-kinase related kinases ATM, ATR and DNA-PK. We have developed a cell-free system to dissect the biochemical mechanisms of these kinases. Using this system, we identify heterogeneous nuclear ribonucleoprotein U (hnRNP-U), also termed scaffold attachment factor A (SAF-A), as a specific substrate for DNA-PK. We show that hnRNP-U is phosphorylated at Ser59 by DNA-PK in vitro and in cells in response to DNA double-strand breaks. Phosphorylation of hnRNP-U suggests novel functions for DNA-PK in the response to DNA damage.     

Binding of 5-bromouracil-containing S/MAR DNA to the nuclear matrix.

Substitution of thymine with 5-bromouracil in DNA is known to change interaction between DNA and proteins, thereby inducing various biological phenomena. We hypothesize that A/T-rich scaffold/nuclear matrix attachment region (S/MAR) sequences are involved in the effects of 5-bromodeoxyuridine. We examined an interaction between DNA containing an intronic S/MAR sequence of the immunoglobulin heavy chain gene and nuclear halos prepared from HeLa cells. Upon substitution with 5-bromouracil, the S/MAR DNA bound more tightly to the nuclear halos. The multi-functional nuclear matrix protein YY1 was also found to bind more strongly to 5-bromouracil-substituted DNA containing its recognition motif. These results are consistent with the above hypothesis.     

Scaffold attachment of DNA loops in metaphase chromosomes

We have examined the higher-order loop organization of DNA in interphase nuclei and metaphase chromosomes from Drosophila Kc cells, and we detect no changes in the distribution of scaffold-attached regions (SARs) between these two phases of the cell cycle. The SARs, previously defined from experiments with interphase nuclei, not only are bound to the metaphase scaffold when endogenous DNA is probed but also rebind specifically to metaphase scaffolds when added exogenously as cloned, end-labeled fragments. Since metaphase scaffolds have a simpler protein pattern than interphase nuclear scaffolds, and both have a similar binding capacity, it appears that the population of proteins required for the specific scaffold-DNA interaction is limited to those found in metaphase scaffolds. Surprisingly, metaphase scaffolds isolated from Drosophila Kc cells contain both the lamin protein and a pore-complex protein, glycoprotein (gp) 188. To study whether lamin contributes to the SAR-scaffold interaction, we have carried out comparative binding studies with scaffolds from HeLa metaphase chromosomes, which are free of lamina, and from HeLa interphase nuclei. All Drosophila SAR fragments tested bind with excellent specificity to HeLa interphase scaffolds, whereas a subset of them bind to HeLa metaphase scaffolds. The maintenance of the scaffold-DNA interaction in metaphase indicates that lamin proteins are not involved in the attachment site for at least a subset of Drosophila SARs. This evolutionary and cell-cycle conservation of scaffold binding sites is consistent with a fundamental role for these fragments in the organization of the genome into looped domains.     

The nuclear scaffold exhibits DNA-binding sites selective for supercoiled DNA

Binding of exogenous DNA to the nuclear scaffold was investigated using a plasmid DNA (pBR322, EcoRI site deleted) of various topological forms and nuclear subfractions with different levels of nuclear DNA depletion. When supercoiled DNA was incubated with histone-depleted nuclei (nuclear halo), a dose-dependent binding of the DNA occurred, whereas no binding was observed with relaxed and linear forms of DNA. The bound DNA was released upon linearization with BamHI or digestion of the scaffolding structure with proteinase K. Extensive digestion of the halo with micrococcal nuclease generated additional sites which bind both relaxed and linear DNA. In the presence of a large excess of calf thymus DNA, these sites were effectively blocked and the specificity to supercoiled DNA was restored. The binding of all forms of DNA was abolished by heat-denatured DNA. There was no detectable change in linking number of the scaffold-associated supercoils. Competitive binding was observed between supercoiled DNAs with unrelated sequences, indicating that no specific nucleotide sequence is required for the binding. RNA was found to be a weak competitor. A DNA binding assay performed on electrophoretic blots of solubilized nuclear scaffold revealed a protein component with apparent molecular weight of 120,000 which retained selective binding to supercoils. These results suggest that the nuclear scaffold possesses DNA-binding sites for torsionally strained domains of chromatin and that an integral protein factor is involved in the binding. Implications of the findings are discussed in connection with proposed functions of the nuclear scaffold and topoisomerase II.     

Scaffold attachment regions in centromere-associated DNA

Due to indications that kinetochore proteins are an integral part of the protein scaffold component of the chromosome (Earnshaw et al. 1984), we chose to map the distribution of scaffold attachment regions (SARs) at centromeres. Using the SAR mapping assay of Mirkovitch et al., Southern blots were prepared and probed with ³²P-labeled fragments from the human 1.9 kb centromeric α-satellite repeat unit of chromosome 1 or the 1.7 kb centromeric α-satellite repeat unit of chromosome 16. Our results demonstrated the presence of one SAR site per 1.9 kb repeat unit in chromosome 1, and every 1.7 kb repeat unit in chromosome 16, separated by regions of small DNA loops over the length of the α-satellite regions. We also identified several in vitro vertebrate topoisomerase II and cenP-B consensus sequences throughout the chromosome 1 α-satellite region using computer and base ratio analysis, to address the question as to why some α-satellite regions are SAR related and others are not. To provide in situ indications of SAR localization in the human genome, SAR DNA and non-SAR DNA were prepared following lithium 3,5-di-iodosalicylate extraction. Sequences protected from DNAse I digestion by SAR proteins, as compared with unprotected DNA that was digested by the enzyme, was labeled with biotin-UTP, hybridized to chromosomal DNA in situ, and then detected with fluorescein-avidin-DCS. Both SAR and non-SAR DNA selectively labeled virtually all centromeric regions of the human metaphase karyotype. Chromosomal arms were less strongly bound by SAR DNA, with a pattern that followed the chromosomal axis. In the more condensed chromosomes an R-banding pattern was evident. In general, labeling patterns produced by both SAR and non-SAR fractions were similar, as expected from the indications that SAR DNAs are heterogenous in sequence and do not form a specific class of sequences. We conclude that centromeric regions of several, possibly all, human metaphase chromosomes are also regions where the chromosomal axis contains loops, smaller in size than in the arms and where attachment sites are concentrated. This clustering of SARs may be responsible in part for the tight chromatin packing associated with the primary constriction of the centromeric region. Received: 10 October 1995; in revised form: 10 May 1996 / Accepted: 13 May 1996     

Phosphorylation of SAF-A/hnRNP-U Serine 59 by Polo-Like Kinase 1 Is Required for Mitosis

Scaffold attachment factor A (SAF-A), also called heterogenous nuclear ribonuclear protein U (hnRNP-U), is phosphorylated on serine 59 by the DNA-dependent protein kinase (DNA-PK) in response to DNA damage. Since SAF-A, DNA-PK catalytic subunit (DNA-PKcs), and protein phosphatase 6 (PP6), which interacts with DNA-PKcs, have all been shown to have roles in mitosis, we asked whether DNA-PKcs phosphorylates SAF-A in mitosis. We show that SAF-A is phosphorylated on serine 59 in mitosis, that phosphorylation requires polo-like kinase 1 (PLK1) rather than DNA-PKcs, that SAF-A interacts with PLK1 in nocodazole-treated cells, and that serine 59 is dephosphorylated by protein phosphatase 2A (PP2A) in mitosis. Moreover, cells expressing SAF-A in which serine 59 is mutated to alanine have multiple characteristics of aberrant mitoses, including misaligned chromosomes, lagging chromosomes, polylobed nuclei, and delayed passage through mitosis. Our findings identify serine 59 of SAF-A as a new target of both PLK1 and PP2A in mitosis and reveal that both phosphorylation and dephosphorylation of SAF-A serine 59 by PLK1 and PP2A, respectively, are required for accurate and timely exit from mitosis.     

Nuclear scaffold attachment stimulates, but is not essential for ARS activity in Saccharomyces cerevisiae: analysis of the Drosophila ftz SAR

Nuclei isolated from eukaryotic cells can be depleted of histones and most soluble nuclear proteins to isolate a structural framework called the nuclear scaffold. This structure maintains specific interactions with genomic DNA at sites known as scaffold attached regions (SARs), which are thought to be the bases of DNA loops. In both Saccharomyces cerevisiae and Schizosaccharomyces pombe, genomic ARS elements are recovered as SARs. In addition, SARs from Drosophila melanogaster bind to yeast nuclear scaffolds in vitro and a subclass of these promotes autonomous replication of plasmids in yeast. In the present report, we present fine mapping studies of the Drosophila ftz SAR, which has both SAR and ARS activities in yeast. The data establish a close relationship between the sequences involved in ARS activity and scaffold binding: ARS elements that can bind the nuclear scaffold in vitro promote more efficient plasmid replication in vivo, but scaffold association is not a strict prerequisite for ARS function. Efficient interaction with nuclear scaffolds from both yeast and Drosophila requires a minimal length of SAR DNA that contains reiteration of a narrow minor groove structure of the double helix.     

Characterization of a plant scaffold attachment region in a DNA fragment that normalizes transgene expression in tobacco.

Using a low-salt extraction procedure, we isolated nuclear scaffolds from tobacco that bind specific plant DNA fragments in vitro. One of these fragments was characterized in more detail; this characterization showed that it contains sequences with structural properties analogous to animal scaffold attachment regions (SARs). We showed that scaffold attachment is evolutionarily conserved between plants and animals, although different SARs have different binding affinities. Furthermore, we demonstrated that flanking a chimeric transgene with the characterized SAR-containing fragment reduces significantly the variation in expression in series of transformants with an active insertion, whereas a SAR fragment from the human beta-globin locus does not. Moreover, the frequency distribution patterns of transgene activities showed that most of the transformants containing the plant SAR fragment had expression levels clustered around the mean. These data suggest that the particular plant DNA fragment can insulate the reporter gene from expression-influencing effects exerted from the host chromatin.     

DNA-PK-dependent binding of DNA ends to plasmids containing nuclear matrix attachment region DNA sequences: evidence for assembly of a repair complex

We find that nuclear protein extracts from mammalian cells contain an activity that allows DNA ends to associate with circular pUC18 plasmid DNA. This activity requires the catalytic subunit of DNA-PK (DNA-PKcs) and Ku since it was not observed in mutants lacking Ku or DNA-PKcs but was observed when purified Ku/DNA-PKcs was added to these mutant extracts. Purified Ku/DNA-PKcs alone did not produce association of DNA ends with plasmid DNA suggesting that additional factors in the nuclear extract are necessary for this activity. Competition experiments between pUC18 and pUC18 plasmids containing various nuclear matrix attachment region (MAR) sequences suggest that DNA ends preferentially associate with plasmids containing MAR DNA sequences. At a 1:5 mass ratio of MAR to pUC18, approximately equal amounts of DNA end binding to the two plasmids were observed, while at a 1:1 ratio no pUC18 end binding was observed. Calculation of relative binding activities indicates that DNA end-binding activities to MAR sequences was 7–21-fold higher than pUC18. Western analysis of proteins bound to pUC18 and MAR plasmids indicates that XRCC4, DNA ligase IV and scaffold attachment factor A preferentially associate with the MAR plasmid in the absence or presence of DNA ends. In contrast, Ku and DNA-PKcs were found on the MAR plasmid only in the presence of DNA ends suggesting that binding of these proteins to DNA ends is necessary for their association with MAR DNA. The ability of DNA-PKcs/Ku to direct DNA ends to MAR and pUC18 plasmid DNA is a new activity for DNA-PK and may be important for its function in double-strand break repair. A model for DNA repair based on these observations is presented.