Nucleic Acids Research

Directional binding of HMG-I(Y) on 4H DNA

OP-Cu is an efficient chemical nuclease that cleaves the phosphodiester backbone of DNA or RNA under physiological conditions by oxidation of deoxyribose or ribose sugars (45). Covalent attachment of an OP-Cu moiety to a specific amino acid residue at a known position in a protein followed by binding of the derivatized product to a DNA substrate under reactive conditions results in strand scission in the local vicinity of the OP-Cu attachment site. Analysis of the resulting pattern of DNA cleavage products allows determination of both the position and orientation of the bound protein (43,46,47). Unfortunately, such a position-specific cleavage strategy using OP-Cu is not readily applicable for directly determining whether the HMG-1 protein binds to 4H DNA with a distinctly polar orientation, as the asymmetric footprint in Figure 5 suggests, since all known vertebrate HMG-1 proteins contain numerous cysteine residues (48) and thus cannot be easily derivatized at only a single position with the chemical cleavage reagent.

Using hydroxyl radical footprinting we have recently demonstrated (11) that a member of a different HMG protein family, HMG-I(Y), also binds to the arms of 4H DNA in a distinctly asymmetric manner (Fig. 5), suggesting that it, like HMG-1, might also bind to this substrate with a specific directional orientation. The possibility that both of these proteins may be behaving similarly in this respect is not entirely surprising considering the suggestion that the DNA-binding domains of the HMG-1 and HMG-I(Y) proteins are likely evolutionarily related (15) owing to the fact that the A·T-hook DNA-binding motif of the HMG-I(Y) protein has both extensive amino acid sequence identity with and structural similarity to the extended N-terminal peptide segments of the HMG-1 boxes present in the wild-type HMG-1 protein (36,49-51). Results of peptide domain swap experiments between the DNA-binding domains of the HMG-1 and HMG-I(Y) proteins are entirely consistent with the evolutionary relatedness of these two proteins (52). For these reasons, and due to the fortuitous lack of cysteine residues in the wild-type HMG-I(Y) protein, we used a site-specifically mutated and OP-Cu-derivatized HMG-I(Y) protein to determine whether this protein binds to 4H DNA with a preferred directional orientation. In these experiments an OP-Cu cleavage moiety was covalently attached to a unique cysteine residue at the C-terminal end of a truncated recombinant HMG-I(Y) protein, known as HMG-I[Delta]E91, whose negatively charged C-terminal tail had been removed by in vitro mutagenesis (20). Important for the interpretation of these experiments, we have previously demonstrated that the HMG-I[Delta]E91 protein specifically binds to both naked DNA (20,36) and to isolated nucleosomal core particles (53) in the same manner and with the same substrate specificity as the full-length wild-type recombinant HMG-I(Y) protein.

In a series of footprinting experiments, the HMG-I[Delta]E91-OP-Cu protein adduct was bound in a 1:1 molar complex (see for example Fig. 2C) to various 4H DNA molecules containing different 5[prime]-radiolabeled oligonucleotides and the resulting protein-DNA complexes were subjected to chemical cleavage reactions. Figure 2A (lane 4) shows that HMG-I[Delta]E91-OP-Cu cleavage of 4H DNA (radiolabeled on the 5[prime]-end of oligonucleotide 4) results in the generation of a limited number of specific and prominent scission products (identified by solid circles) that flank the branch point of the molecule (indicated by the arrowhead). In contrast, in the absence of any bound protein, exposure of this same 4H DNA to either free OP-Cu reagent (lane 3) or to free hydroxyl radicals (lane 2) results in the generation of numerous cleavage bands that are spread throughout the length of the gel lanes. These cleavage bands arise from scission of each nucleotide within the DNA substrate by the free chemical reagents. These control, ‘free in solution’ cleavage reactions demonstrate two important points. First, the pattern of cleavage bands generated by the freely diffusable hydroxyl radicals (lane 2) differs markedly from that produced by free OP-Cu (lane 3). As expected, free hydroxyl radicals cleave the substrate DNA at each nucleotide in a more or less uniform manner whereas the free OP-Cu reagent, whose intensity of cutting (i.e. rate of cleavage) depends on local triplet nucleotide sequences with minor grooves that vary in their degree of complementarity to the structure of the bound tetrahedral [OP]₂Cu⁺ complex (46,47), shows much more variability in cleavage along the length of the DNA. Second, the bands that are prominent in lane 4 also show slightly enhanced cutting in lane 3. Nevertheless, when intensities of bands are compared, it is obvious that the patterns of cleavage generated by the free reagents are quite different from the distinct, and regionally localized, major DNA cleavages generated by the OP-Cu moiety when it is tethered to the HMG-I[Delta]E91 protein (lane 4). The background cleavage bands seen in Figure 2A (lane 4) result from either low levels of free hydroxyl radicals generated in the reaction by trace contaminants (46,47) and/or are caused by the presence of low amounts of non-conjugated free OP-Cu reagent in the solution. If the reaction buffer solution was completely devoid of these trace contaminants, only the site-specific cleavages resulting from the bound OP-Cu-conjugated protein would be observed (46,47). Importantly, the specific DNA cleavages associated with the protein-tethered OP-Cu reagent are still readily distinguishable above this non-specific background. Since the OP-Cu reagent is covalently attached to the C-terminus of the HMG-I[Delta]E91 protein in this experiment, we interpret this non-random, spatially restricted distribution of prominent cleavage bands (lane 4) on the labeled oligonucleotide as reflecting the fact that HMG-I(Y) binds to the 4H DNA substrate with a distinct directionality.

A - B

C

Figure 2. Chemical cleavage experiments with OP-Cu-derivatized proteins demonstrate asymmetric association of HMG-I(Y) with 4H DNA. (A) In control experiments, a solution of protein-free 4H DNA, radiolabeled on the 5[prime]-end of oligonucleotide 4, was exposed to either free hydroxyl radicals (lane 2) or to free OP-Cu chemical cleavage reagent in solution (lane 3). In contrast, lane 4 shows the chemical cleavage pattern obtained when the OP-Cu chemical cleavage moiety is covalently attached to the C-terminal end of a specifically mutated form of HMG-I and the resulting derivatized protein, designated HMG-I[Delta]E91-OP-Cu, bound in an approximately 1:1 complex with the same radiolabeled 4H DNA. Lane 1 contains the products of a Maxam-Gilbert G sequencing reaction of radiolabeled oligonucleotide 4. (B) Free 4H DNA, radiolabeled on the 5[prime]-end of oligonucleotide 1, was cleaved with either free hydroxyl radicals (lane 2), free OP-Cu reagent (lane 3) or derivatized HMG-I[Delta]E91-OP-Cu protein bound in an approximately 1:1 complex (lane 4). Lane 1 contains the products of a Maxam-Gilbert G sequencing reaction of radiolabeled oligonucleotide 1. In both panels the chemical cleavage reactions shown in lane 4 were performed on protein:DNA complexes in which >90% of the 4H DNA was complexed with a single molecule of HMG-I[Delta]E91-OP-Cu protein. The solid dots next to lane 4 in each panel show the sites of preferred cleavage by the bound HMG-I[Delta]E91-OP-Cu protein on the particular radiolabeled oligonucleotides and the solid arrowheads show the position of the branch, or crossover, points of those oligonucleotides in the 4H DNA. (C) This control EMSA demonstrates that under the conditions of the chemical cleavage experiments, a single protein:DNA complex formed. Lane 1 contains free migrating 4H DNA and the following lanes contain increasing amounts of HMG-I[Delta]E91-OP-Cu protein. The lane containing the same protein concentration used in the cleavages in Figure 2A and B is marked with an arrow.

To further investigate the apparent directional specificity of binding of the HMG-I(Y) protein, additional cleavage experiments were performed in which HMG-I[Delta]E91-OP-Cu was bound to 4H DNA substrates that had been individually radiolabeled at the 5[prime]-ends of the other three oligonucleotides. Figure 2B shows the results of one such experiment involving labeled oligonucleotide 1. Lanes 2 and 3 of this figure show the results of control experiments in which 4H DNA alone has been exposed to either free hydroxyl radicals (lane 2) or to free OP-Cu (lane 3) and indicate the patterns of background cleavages observed with these reagents on this labeled substrate. Lane 4, on the other hand, shows the cleavage patterns observed when the HMG-I[Delta]E91-OP-Cu protein is bound to the same 4H DNA substrate. The results in lane 4 demonstrate that when the OP-Cu-derivatized protein is bound to 4H DNA, a number of distinct and spatially restricted cleavage bands (indicated by solid circles) are observed superimposed on a background pattern of non-specific cleavages, again showing that HMG-I(Y) is binding to 4H DNA with a specific polar orientation.

In contrast to the distinct OP-Cu cleavage patterns obtained when derivatized HMG-I(Y) protein was bound to 4H DNA labeled on either oligonucleotide 4 (Fig. 2A) or 1 (Fig. 2B), when the HMG-I[Delta]E91-OP-Cu protein was bound to 4H DNA substrates containing either radiolabeled oligonucleotide 2 or 3 only non-specific cleavage bands were observed (data not shown). These negative results serve to reinforce the significance of the specific cleavage patterns observed with oligonucleotides 1 and 4. In addition, these results also strongly suggest that the C-terminal end of the HMG-I[Delta]E91-OP-Cu protein, when bound to 4H DNAs containing either labeled oligonucleotide 2 or 3, was not in close enough proximity to induce specific cleavages in either of these labeled DNA strands (data not shown).

Figure 5 diagrams a planar representation of the 4H DNA oligonucleotide sequence and shows the points of preferential chemical cleavage (indicated by lines emanating from stars) observed when the C-terminal derivatized HMG-I[Delta]E91-OP-Cu protein is bound to this substrate compared to the previously identified sites (gray shaded bars) of physical interaction of the wild-type HMG-I(Y) protein bound to the same 4H DNA as determined by hydroxyl radical footprinting (11). This composite figure clearly illustrates the fact that the major sites of chemical cleavage are asymmetrically located on 4H DNA, occurring almost exclusively near the base of arm (1:4) with relatively little, if any, specific cleavage occurring elsewhere on the 4H DNA molecule. We interpret these data as indicating that the major cleavages on the 4H DNA occur at the nucleotides in closest proximity to the C-terminal tail of the derivatized HMG-I[Delta]E91-OP-Cu protein with its reactive chemical moiety and therefore conclude that the HMG-I(Y) protein binds to 4H DNA with one strongly preferred directional orientation (see below).

Hydroxyl radical footprinting of the globular domain of histone H1 (GH1) on 4H DNA

Previous reports indicate that the centrally located globular domain of histone H1 specifically interacts with 4H DNA in the same manner as the intact histone H1 protein (7,54). To verify that the H1 globular domain (GH1, derived from Xenopus) used in the present experiments also behaves similarly to intact H1 proteins, we performed EMSAs in which increasing concentrations of recombinant GH1 were added to labeled 4H DNA followed by electrophoretic separation of the resulting protein:DNA complexes (Fig. 3). The results of these experiments demonstrate that the GH1 peptide, like full-length histone H1 protein (11), forms only a single protein complex with 4H DNA under our conditions. In addition, the K_d for binding of the GH1 peptide to 4H DNA was determined to be ~13 nM (data not shown), similar to values reported for the binding of both full-length H1 protein (~16 nM) (11) and the globular domains of both histones H1 and H5 (7) from different species to 4H DNA.

Figure 3. Titration EMSA of GH1 on 4H DNA. Labeled 4H DNA was titrated with increasing concentrations of GH1. Lane 1 is free migrating 4H DNA (F) and lanes 2-6 contain 13, 26, 40, 52 and 65 nM GH1, which formed only a single 4H DNA-protein complex (C). The arrow indicates the direction of electrophoresis.

It has been reported that the intact histone H1 protein, as well as the isolated GH1 globular domain, has two independent peptide surfaces for contacting both B-form duplex (17,54) and 4H (35) DNAs. Little information is available, however, concerning the actual points of physical contact that these two peptide binding surfaces make with 4H DNA substrates. To obtain more information about these molecular contacts, we used hydroxyl radical footprinting to map the binding sites of GH1 on the open conformation of 4H DNA at single nucleotide resolution.

Hydroxyl radical footprinting reactions with GH1 bound to 4H DNA substrates that had each of the four different oligonucleotides individually radiolabeled were performed and the resulting cleavage patterns were quantitatively analyzed as described above for the HMG-1 protein. Figure 4 shows representative results obtained when either oligonucleotide 1 (Fig. 4A) or oligonucleotide 2 (Fig. 4B) was end-labeled, assembled into 4H DNA structures, bound by one GH1 molecule and subjected to hydroxyl radical footprinting cleavage reactions. For comparative purposes, and to demonstrate the regions of the 4H DNA which are protected from hydroxyl radical cleavage by the GH1 protein, individual scans of both naked DNA (dashed lines) and protein-bound DNA (solid lines) were overlaid in both Figure 4A and B. In this figure the regions of the 4H DNA that consistentlyshow the greatest differences between free and protein-bound molecules in various independent experiments are indicated by a solid bar below the scans. The arrows in the figure indicate the location of the branch points of the 4H DNA. In a similar manner, each of the other two strands of the 4H DNA were likewise independently radiolabeled, assembled into 4H DNA structures, bound to GH1 peptide and analyzed by hydroxyl radical footprinting. The composite results of these footprinting experiments are shown in Figure 5, with the protected nucleotides indicated by cross-hatched bars. As can be seen in this figure, the footprint of GH1 covers regions of each of the four arms of 4H DNA starting about 1-2 nt from the branch point and extending outward along each arm. Importantly, in none of our numerous experiments did we observe data suggesting contacts of GH1 with any of the branch point nucleotides, indicating that the globular domain of H1 binds exclusively to the arms of the 4H DNA molecule near to, but not including, the crossover region.

Figure 4. Densitometry scans of the footprint of GH1 on 4H DNA. 4H DNA substrates containing a single 5[prime]-end-labeled oligonucleotide were subjected to hydroxyl radical cleavage in the presence (footprinted) or absence (naked) of GH1 protein. The cleavage products in each case were separated on a sequencing-type gel. Naked and footprinted lanes from each gel were scanned and then overlaid to illustrate the regions of the 4H DNA that are protected from hydroxyl radical cleavage by GH1 (solid lines under the scans). Only those regions of each radiolabeled oligonucleotide that were consistently observed to be protected by the binding of GH1 protein in numerous repeat experiments are indicated by the solid lines. (A) Results obtained with labeled oligonucleotide 1. (B) Results obtained with labeled oligonucleotide 2. In both panels the dashed lines indicate naked DNA and the solid lines indicate GH1 protein-bound (i.e. footprinted) DNA. The arrow indicates the position of the branch point.

Figure 5. Footprint of HMG-1, HMG-I(Y) and GH1 on 4H DNA. Composite figure showing the footprint of HMG-1 (black bars), HMG-I(Y) (gray shaded bars) and GH1 (cross-hatched bars) on 4H DNA. Regions that are protected from hydroxyl radical cleavage by all three proteins are indicated by white letters in black boxes and bases which show enhanced cleavage with OP-Cu-derivativzed HMG-I(Y) are connected by lines to the stars. The dot marks the 5[prime]-end of oligonucleotide 1.

Comparison between the 4H DNA footprints of HMG-1, GH1 and HMG-I(Y)

Figure 5 shows a composite diagram indicating the binding sites of HMG-1, GH1 and HMG-I(Y) (11) on 4H DNA as determined by hydroxyl radical footprinting. The white letters in black boxes indicate the nucleotides of the 4H DNA that are bound in common by HMG-1 (black bars), HMG-I(Y) (gray shaded bars) and GH1 (cross-hatched bars). Our previous competition analyses demonstrated that HMG-I(Y), HMG-1 and H1 compete for binding to 4H DNA in a mutually exclusive manner, suggesting that these three proteins share common binding sites on the 4H DNA substrate (11). However, simply because different proteins directly compete with each other for binding to a substrate does not necessarily indicate that they must share common binding sites on the substrate. Other possible explanations (e.g. steric hindrance considerations and/or protein-induced changes in substrate structure) might also account for such mutually exclusive binding properties. Thus, the footprinting results compiled in Figure 5 are significant because they unequivocally demonstrate that HMG-I(Y), HMG-1 and GH1 all bind to a limited number of common nucleotide binding sites present in each of the four arms of 4H DNA. The presence of such common binding sites, therefore, offers at least a partial explanation for the previously reported competition studies. These data likewise emphasize the important point that in 1:1 protein:DNA complexes, HMG-I(Y), HMG-1 and GH1 all bind in a unique manner at, or near, the branch point of 4H DNA in the open conformation and do not make any contacts near the distal ends of the arms. Furthermore, Figure 5 also illustrates the fact that whereas HMG-1 and HMG-I(Y) both bind asymmetrically to the arms of 4H DNA and make contacts with nucleotides across the branch point, the globular domain of histone H1 binds to each of the four arms of the 4H structure in a much more symmetrical fashion and does not make any contact with branch point nucleotides.

DISCUSSION

Over the last decade the conformation of 4H DNA has been rigorously characterized by a number of different experimental techniques (reviewed in 55-57). These studies indicate that in the absence of bound protein, the overall structure of 4H DNA in solution is dynamic and shows a marked dependence on divalent cation concentrations. In the absence of Mg²⁺, 4H DNA adopts an ‘open’ conformation where the four duplex B-form arms are in a maximally extended configuration with both the major and minor grooves widening somewhat as they enter the open branch point. Recent X-ray crystallographic studies have demonstrated that both the [lambda] phage Cre recombinase (58,59) and the bacterial RuvA (60,61) proteins form co-complexes exclusively with the open form of 4H DNA. In contrast to the open configuration, in the presence of Mg²⁺ the branch point is closed and the arms of the junction co-axially ‘stack’ into an X-shaped structure creating two continuous strands and two exchanging, or non-continuous, strands. The arms of 4H DNA in both the stacked X (55) and the open (58-61) configurations exhibit the minor and major groove characteristics of B-form DNA.

Recently, Pohler et al. (44) demonstrated that the affinity of HMG-1 for 4H DNA is reduced, or even abolished, by titrating the protein-DNA binding reactions with increasing concentrations of Mg²⁺, indicating that HMG-1, like other proteins (58-61), preferentially binds to the ‘open’ conformation of 4H DNA. These same authors also reported that HMG-1 has the ability to alter the conformation of 4H DNA from the stacked X conformer to the ‘open’ conformer upon binding (44). The footprinting data of HMG-1 on 4H DNA in the ‘open’ conformation shown in Figures 1 and 5 are entirely consistent with these observations, but also indicate that not all of the arms of the 4H DNA are equally protected from hydroxyl radical cleavage by the HMG-1 protein. This asymmetric footprint suggests that the wild-type HMG-1 protein, which has two independent HMG-1 ‘box’ DNA-binding domains, preferentially binds to 4H DNA in one predominant orientation. Since all HMG-1 box-containing proteins bind to 4H DNA with high affinity, but without apparent substrate sequence specificity (14,15,36,37,62), it seems likely that such biased binding results from preferential recognition by the HMG-1 protein of some (as yet unidentified) asymmetric structural feature of the 4H DNA itself. Nevertheless, some insight into how the HMG-1 protein might be binding to 4H DNA comes from studies of a co-complex of the human testis-determining protein SRY (40), which contains a single HMG-1 box, with a DNA substrate (63). When the HMG-1 box of SRY binds to the minor groove of B-form DNA, it partially unwinds the DNA strands, widens the groove and concomitantly induces a pronounced bend in the substrate. It has been proposed that the branch point nucleotides of 4H DNA resemble this bent duplex DNA structure induced by binding of HMG-1 box proteins (63). The hydroxyl radical footprint data summarized in Figure 5, demonstrating that most of the branch point nucleotides of the 4H DNA are protected by HMG-1 protein, are entirely consistent with this suggestion. However, since the wild-type HMG-1 protein contains two independent DNA-binding domains, its mode of interaction with 4H DNA to produce the asymmetric protection pattern shown in Figure 5 is considerably more complex than previously studied examples and the precise details of this association remain to be elucidated.

Strong indirect support for polar binding by wild-type HMG-1 protein to 4H DNA comes from studies of HMG-I(Y), a protein that also binds to 4H DNA with a distinctly asymmetric footprint and which does, indeed, bind with a preferred directional orientation (Figs 2, 5 and 6A; see below). As noted previously, the extended N-terminal peptide segments of the HMG boxes of the HMG-1/-2 class of proteins have both extensive amino acid sequence identities and structural similarities to the A·T-hook DNA-binding motif of the HMG-I(Y) protein (36,49-51). Given such close similarities it is not surprising that the HMG-1 protein, like HMG-I(Y), binds asymmetrically to 4H DNA, probably as a result of preferred directional binding.

Figure 6. (A) Modeling HMG-I(Y) on 4H DNA. Molecular scale model showing the preferred direction of binding of the last two DNA-binding domains of the HMG-I(Y) protein (BDII and BDIII) on 4H DNA that is consistent with current data. The dot marks the radiolabeled 5[prime]-end of oligonucleotide 1 and the * indicates the site of covalent attachment of the OP-Cu chemical cleavage moiety at the C-terminal end of the HMG-I[Delta]E91protein. The first 49 amino acid residues of the HMG-I(Y) protein, including the first putative DNA-binding domain (BDI), are not shown in the figure since they do not bind to the 4H DNA under the experimental conditions employed. (B) Molecular model of one possible configuration of GH1 on 4H DNA. A model of the GH1 peptide (based on the X-ray crystal structure of GH5; 30) shown oriented between two arms of 4H DNA to illustrate one possible way (out of several) in which the globular domain could ‘dock’ with the 4H DNA and account for the hydroxyl footprint on those two arms. The composite footprint of GH1 on 4H DNA shown in Figure 5 is proposed to result from a mixed population of 1:1 GH1:DNA complexes in which individual GH1 peptides randomly bind to different pairs of arms in a mutually exclusive manner. Such mutually exclusive binding could result from GH1-induced structural changes induced in the 4H DNA substrate (for simplicity, these alterations in the 4H DNA are not shown in this diagram) that both maximize the strength of protein:DNA interaction and also prevent subsequent binding by a second GH1 peptide to the complex (see text). The dot marks the 5[prime]-end of oligonucleotide 1.

The structure of the A·T-hook DNA-binding domain of HMG-I(Y) bound to a DNA substrate was recently solved by solution NMR (36) and indicates that individual DNA-binding peptides interact in the minor groove with ~4-6 bp of substrate DNA. Approximately the same number of base pairs are protected on arms (1:4) and (2:3) on either side of the branch point of 4H DNA by a single bound HMG-I(Y) molecule (Fig. 5; 11). The length of these footprinted regions on 4H DNA suggests that only two (out of three) of the A·T-hooks of the HMG-I(Y) protein are contacting the arms of this substrate, one motif binding on each of the arms on opposite sides of the branch point. Several additional independent lines of evidence also indicate that the substrate association of only two A·T-hooks of HMG-I(Y) is responsible for binding of the protein to 4H DNA. For example, a subfragment of the HMG-I(Y) protein [designated HMG-I(2/3); 36] that contains only the second (BDII) and third (BDIII) A·T-hook motifs (plus the intervening peptide backbone) binds to 4H DNA with about the same affinity as the full-length wild-type protein containing all three potential DNA-binding domains (data not shown). Likewise, under appropriate conditions where wild-type protein binds, a full-length HMG-I(Y) that contains specific point mutations in only BDII and BDIII does not bind at all to 4H DNA, indicating that the last two A·T-hook motifs of the protein are responsible for binding to this substrate (data not shown). These observations, taken together with the data in Figures 2 and 5 demonstrating directional binding of HMG-I(Y) to this substrate, lead to the scale model diagram in Figure 6A, which is consistent with all of the presently available hydroxyl radical footprinting (11) and HMG-I[Delta]E91-OP-Cu protein chemical cleavage results.

Although a detailed map of the interaction of the globular domain of histone H1 with the DNA on a monomer nucleosome has been reported (40), to our knowledge, the footprints of the globular domain of H1 on 4H DNA (Figs 4 and 5) represent the first documentation of the precise contact sites of either histone H1 or its globular domain with this or any other free DNA substrate with single nucleotide resolution. It is readily apparent from Figure 5 that, in contrast to the asymmetric and directional binding of HMG-I(Y) and HMG-1, the globular domain of histone H1 does not interact with the crossover nucleotides of the 4H DNA molecule but, rather, binds to the arms of 4H DNA with a fairly symmetrical arrangement around the branch point. Research by others has demonstrated that histone H1 has a high affinity for DNA crossover structures like those found in supercoiled molecules (17,54) and in 4H DNA (7,11). Furthermore, electron microscopy of the globular domain of H1 incubated in the presence of duplex DNA reveals tramline-like structures of proteins sandwiched between two DNA strands, suggesting that the globular domain of H1 is able to simultaneously contact two DNA strands via two separate binding surfaces (64). Mutation of the putative second DNA-binding surface of the globular domain abolished the formation of the tramline-like structures as well as its affinity for 4H DNA (35). Together these results indicate that both of the DNA-binding surfaces of the globular domain are necessary for H1 to bind simultaneously to two duplex DNA strands or to 4H DNA. Another relevant observation is that histone H1 has a lower affinity for 4H DNA in the presence of Mg²⁺ than in its absence (7), suggesting that H1, like HMG-1, prefers to bind to the open conformation of 4H DNA rather than to the closed, or stacked X, conformer. Based on this and other information, Zlatanova and van Holde (13) have proposed that histone H1 exhibits an ‘interior mode’ of binding to 4H DNA, i.e. that the two DNA-binding domains of the globular domain of the H1 protein contact the concave, or ‘inside’, surface of 4H DNA molecules. Such an interior mode of binding of H1 on 4H DNA predicts that GH1 will make contacts with regions of DNA on adjacent arms of the substrate. The footprinting data summarized in Figure 5 for GH1 are consistent with this proposal and clearly demonstrate that the globular domain of H1 binds to the arms of the 4H DNA very near, but not including, the branch point nucleotides. Such dual site binding is in agreement with a previous report demonstrating that both DNA-binding surfaces must be intact in order for GH1 to bind to 4H DNA (30). Since [alpha]-helix III is the primary major groove DNA-binding surface of other winged helix proteins (reviewed in 31), it is likely that [alpha]-helix III of GH1 also binds in the major groove of 4H DNA.

The model shown in Figure 6B shows one possible docking of the H1 globular domain in which a major and minor groove of 4H are contacted; other interactions, including contacts with only major grooves, are possible. Given such an association, an attractive explanation for the nearly symmetrical footprinting pattern on the arms of 4H is as follows: the protected regions are, in reality, the composite result obtained when individual GH1 molecules randomly bind to different pairs of arms in a mutually exclusive manner resulting in a mixed population of GH1:DNA complexes in a 1:1 molar association (Fig. 3). Such mutually exclusive binding could result from structural changes induced in the 4H DNA substrate by binding of the GH1 peptide. For example, as schematically shown by the ‘to scale’ molecular diagram in Figure 6B, the GH1 peptide could bind between two adjacent arms of the 4H DNA molecule resulting in an induced structural change in the substrate that both maximizes the strength of this protein:DNA interaction and also prevents subsequent binding by a second GH1 peptide to the complex (e.g. by bringing the two adjacent protein-bound arms together and widening the distance, or distorting the angle, between the other two, distal, non-bound arms; 65). Such a structurally dynamic situation is not adequately illustrated by the static diagram in Figure 6B, which illustrates only one of several possible modes of binding by a single GH1 peptide and also fails to indicate possible structural changes in the 4H DNA induced by peptide binding. Direct support for binding-induced structural alterations comes from EMSA experiments which demonstrate that 4H DNA:histone H1 complexes exhibit an anomalous electrophoretic migration behavior on gels, consistent with protein-induced changes in the DNA substrate conformation (65).

The combined footprinting results shown in Figure 5 indicate that HMG-I(Y), HMG-1 and H1 share common binding sites on 4H DNA, thus partially explaining the observed mutually exclusive binding and competition by these proteins for this substrate. Nevertheless, the nature of each footprint indicates not only that each protein binds to 4H DNA in a different way but also suggests that the proteins recognize slightly different structural features of the 4H DNA and/or introduce distinctly different structural alterations in their respective protein:DNA complexes.

The present findings have important implications concerning the possible in vivo roles of HMG-I(Y), histone H1 and HMG-1 in various biological processes. For example, the uniqueness of each protein’s interactions with 4H DNA suggests that the individual proteins likely serve different structural roles and/or perform different functions in cellular events such as genetic recombination, immunoglobulin V(D)J gene rearrangements or integration of retroviruses into host cell genomes. Experiments are currently in progress to investigate these possibilities.

ACKNOWLEDGEMENTS

We thank Dr Jeffrey Hayes for his generous gift of the recombinant Xenopus GH1 (also known as NG-H1°) protein, Dr A.R. Srinivasan and Dr Wilma K. Olsen for providing PDB structural files of the four-way junction in several different conformations. We also thank Susan Johns of the Washington State University VADMS laboratory for generating the computer models of four-way junctions used for illustrative purposes. This work was supported by National Science Foundation grant MCB-9506878 and National Institute of Health grant GM-46352 (both to R.R.) and, in part, by a Washington State University Summer Graduate Research Assistantship Program (to D.A.H).

REFERENCES

1. Lilley,D.M., Clegg,R.M., Diekmann,S., Seeman,N.C., von Kitzing,E. and Hagerman,P.J. (1995) Nucleic Acids Res., 23, 3363-3364.

2. Holliday,R. (1964) Genet. Res., 5, 282-304.

3. White,M.F. and Lilley,D.M. (1997) J. Mol. Biol., 266, 122-134. MEDLINE Abstract

4. van Gent,D.C., Mizuuchi,K. and Gellert,M. (1996) Science, 271, 1592-1594. MEDLINE Abstract

5. Lilley,D.M. (1992) Nature, 357, 282-283. MEDLINE Abstract

6. Varga-Weisz,P., van Holde,K. and Zlatanova,J. (1994) Biochem. Biophys. Res. Commun., 203, 1904-1911. MEDLINE Abstract

7. Varga-Weisz,P., Zlatanova,J., Leuba,S.H., Schroth,G.P. and van Holde,K. (1994) Proc. Natl Acad. Sci. USA, 91, 3525-3529. MEDLINE Abstract

8. Zlatanova,J. and van Holde,K. (1996) Prog. Nucleic Acid Res. Mol. Biol., 52, 217-259. MEDLINE Abstract

9. Bianchi,M.E., Beltrame,M. and Paonessa,G. (1989) Science, 243, 1056-1059. MEDLINE Abstract

10. Claus,P., Schulze,E. and Wi'sniewski,J.R. (1994) J. Biol. Chem., 269, 33042-33048.

11. Hill,D.A. and Reeves,R. (1997) Nucleic Acids Res., 25, 3523-3531. MEDLINE Abstract

12. Varga-Weisz,P., van Holde,K. and Zlatanova,J. (1993) J. Biol. Chem., 268, 20699-20700.

13. Zlatanova,J. and van Holde,K.E. (1998) FASEB J., 12, 421-431. MEDLINE Abstract

14. Landsman,D. and Bustin,M. (1993) Bioessays, 15, 539-546. MEDLINE Abstract

15. Bustin,M. and Reeves,R. (1996) Prog. Nucleic Acid Res. Mol. Biol., 54, 35-100. MEDLINE Abstract

16. Bustin,M., Lehn,D.A. and Landsman,D. (1990) Biochim. Biophys. Acta, 1049, 231-243. MEDLINE Abstract

17. Krylov,D., Leuba,S., van Holde,K. and Zlatanova,J. (1993) Proc. Natl Acad. Sci. USA, 90, 5052-5056. MEDLINE Abstract

18. Ivanchenko,M., Hassan,A., van Holde,K.E. and Zlatanova,J. (1996) J. Biol. Chem., 271, 32580-32585.

19. Ivanchenko,M., Zlatanova,J., Varga,W., Hassan,A. and van Holde,K. (1996) Proc. Natl Acad. Sci. USA, 93, 6970-6974. MEDLINE Abstract

20. Nissen,M.S. and Reeves,R. (1995) J. Biol. Chem., 270, 4355-4360.

21. Aiyar,A., Hindmarsh,P., Skalka,A.M. and Leis,J. (1996) J. Virol., 70, 3571-3580. MEDLINE Abstract

22. Agrawal,A. and Schatz,D.G. (1997) Cell, 89, 43-53. MEDLINE Abstract

23. van Gent,D.C., Hiom,K., Paull,T.T. and Gellert,M. (1997) EMBO J., 16, 2665-2670. MEDLINE Abstract

24. Sawchuk,D.J., Weis,G., Malik,S., Besmer,E., Bustin,M., Nussenzweig,M.C. and Cortes,P. (1997) J. Exp. Med., 185, 2025-2032. MEDLINE Abstract

25. Costello,E., Saudan,P., Winocour,E., Pizer,L. and Beard,P. (1997) EMBO J., 16, 5943-5954. MEDLINE Abstract

26. Farnet,C. and Bushman,F. (1997) Cell, 88, 483-492. MEDLINE Abstract

27. Li,L., Farnet,C.M., Anderson,W.F. and Bushman,F.D. (1998) J. Virol., 72, 2125-2131. MEDLINE Abstract

28. Clore,G.M., Gronenborn,A.M., Nilges,M., Sukumaran,D.K. and Zarbock,J. (1987) EMBO J., 6, 1833-1842. MEDLINE Abstract

29. Cerf,C., Lippens,G., Muyldermans,S., Segers,A., Ramakrishnan,V., Wodak,S.J., Hallenga,K. and Wyns,L. (1993) Biochemistry, 32, 11345-11351. MEDLINE Abstract

30. Ramakrishnan,V., Finch,J.T., Graziano,V., Lee,P.L. and Sweet,R.M. (1993) Nature, 362, 219-223. MEDLINE Abstract

31. Brennan,R.G. (1993) Cell, 74, 773-776. MEDLINE Abstract

32. Kodandapani,R., Pio,F., Ni,C.Z., Piccialli,G., Klemsz,M., McKercher,S., Maki,R.A. and Ely,K.R. (1996) Nature, 380, 456-460. MEDLINE Abstract

33. Pemberton,I.K., Buc,H. and Buckle,M. (1998) Biochemistry, 37, 2682-2690. MEDLINE Abstract

34. Thomas,J.O. and Wilson,C.M. (1986) EMBO J., 5, 3531-3537. MEDLINE Abstract

35. Goytisolo,F.A., Gerchman,S.E., Yu,X., Rees,C., Graziano,V., Ramakrishnan,V. and Thomas,J.O. (1996) EMBO J., 15, 3421-3429. MEDLINE Abstract

36. Huth,J.R., Bewley,C.A., Nissen,M.S., Evans,J.N., Reeves,R., Gronenborn,A.M. and Clore,G.M. (1997) Nature Struct. Biol., 4, 657-665.

37. Bianchi,M.E. (1988) EMBO J., 7, 843-849. MEDLINE Abstract

38. Ausubel,F.M., Brent,R., Kingston,R., Seidman,J.G., Smith,J.A. and Struhl,K. (1988) Current Protocols in Molecular Biology. John Wiley & Sons, New York, NY.

39. Adachi,Y., Mizuno,S. and Yoshida,M. (1990) J. Chromatogr., 530, 39-46. MEDLINE Abstract

40. Hayes,J.J. (1996) Biochemistry, 35, 11931-11937. MEDLINE Abstract

41. Nissen,M.S., Langan,T.A. and Reeves,R. (1991) J. Biol. Chem., 266, 19945-19952.

42. Tullius,T.D., Dombroski,B.A., Churchill,M.E. and Kam,L. (1987) Methods Enzymol., 155, 537-558. MEDLINE Abstract

43. Pan,C.Q., Feng,J.A., Finkel,S.E., Landgraf,R., Sigman,D. and Johnson,R.C. (1994) Proc. Natl Acad. Sci. USA, 91, 1721-1725. MEDLINE Abstract

44. Pohler,J.R., Norman,D.G., Bramham,J., Bianchi,M.E. and Lilley,D.M. (1998) EMBO J., 17, 817-826. MEDLINE Abstract

45. Sigman,S.S. (1986) Acc. Chem. Res., 19, 180-186.

46. Sigman,D.S., Kuwabara,M.D., Chen,C.H. and Bruice,T.W. (1991) Methods Enzymol., 208, 414-433. MEDLINE Abstract

47. Sigman,D.S., Landgraf,R., Perrin,D.M. and Pearson,L. (1996) Metal Ions Biol. Syst., 33, 485-513.

48. Nightingale,K., Dimitrov,S., Reeves,R. and Wolffe,A.P. (1996) EMBO J., 15, 548-561. MEDLINE Abstract

49. Weir,H.M., Kraulis,P.J., Hill,C.S., Raine,A.R., Laue,E.D. and Thomas,J.O. (1993) EMBO J., 12, 1311. MEDLINE Abstract

50. Read,C.M., Cary,P.D., Crane-Robinson,C., Driscoll,P.C. and Norman,D.G. (1993) Nucleic Acids Res., 21, 3427-3436. MEDLINE Abstract

51. Reeves,R. and Nissen,M.S. (1990) J. Biol. Chem., 265, 8573-8582.

52. Banks,G.C., Mohr,B. and Reeves,R. (1999) J. Biol. Chem., in press.

53. Reeves,R. and Nissen,M.S. (1993) J. Biol. Chem., 268, 21137-21146.

54. Ivanchenko,M., Zlatanova,J. and van Holde,K. (1997) Biophys. J., 72, 1388-1395. MEDLINE Abstract

55. Lilley,D.M. and Clegg,R.M. (1993) Annu. Rev. Biophys. Biomol. Struct., 22, 299-328. MEDLINE Abstract

56. Seeman,N.C. and Kallenbach,N.R. (1994) Annu. Rev. Biophys. Biomol. Struct., 23, 53-86. MEDLINE Abstract

57. Lilley,D.M. (1997) Proc. Natl Acad. Sci. USA, 94, 9513-9515. MEDLINE Abstract

58. Guo,F., Gopaul,D.N. and van Duyne,G.D. (1997) Nature, 389, 40-46. MEDLINE Abstract

59. Gopaul,D.N., Guo,F. and van Duyne,G.D. (1998) EMBO J., 17, 4175-4187. MEDLINE Abstract

60. Roe,S.M., Barlow,T., Brown,T., Oram,M., Keeley,A., Tsaneva,I.R. and Pearl,L.H. (1998) Mol. Cell, 2, 361-372. MEDLINE Abstract

61. Hargreaves,D., Rice,D.W., Sedelnikova,S.E., Artymiuk,P.J., Lloyd,R.G. and Rafferty,J.B. (1998) Nature Struct. Biol., 5, 441-446.

62. Ferrari,S., Harley,V.R., Pontiggia,A., Goodfellow,P.N., Lovell-Badge,R. and Bianchi,M.E. (1992) EMBO J., 11, 4497-4506. MEDLINE Abstract

63. Werner,M.H., Huth,J.R., Gronenborn,A.M. and Clore,G.M. (1995) Cell, 81, 705-714. MEDLINE Abstract

64. Thomas,J.O., Rees,C. and Finch,J.T. (1992) Nucleic Acids Res., 20, 187-194. MEDLINE Abstract

65. Hill,D.A. (1998) PhD dissertation, Washington State University.

---

*To whom correspondence should be addressed. Tel: +1 509 335 1948; Fax: +1 509 335 9688; Email: reevesr@mail.wsu.edu
⁺Present address: Department of Cell Biology, University of Massachusetts Medical School, Worcester, MA 01655, USA

---

This page is run by Oxford University Press, Great Clarendon Street, Oxford OX2 6DP, as part of the OUP Journals
Comments and feedback: jnl.info{at}oup.co.uk
Last modification: 27 Apr 1999
Copyright©Oxford University Press, 1999.
CiteULike    Connotea    Del.icio.us    What's this?

This article has been cited by other articles:

				J. E. Adair, Y. Kwon, G. A. Dement, M. J. Smerdon, and R. Reeves Inhibition of Nucleotide Excision Repair by High Mobility Group Protein HMGA1 J. Biol. Chem., September 16, 2005; 280(37): 32184 - 32192. [Abstract] [Full Text] [PDF]

				M. Novoseler, G. Hershkovits, and D. J. Katcoff Functional Domains of the Yeast Chromatin Protein Sin1p/Spt2p Can Bind Four-way Junction and Crossing DNA Structures J. Biol. Chem., February 18, 2005; 280(7): 5169 - 5177. [Abstract] [Full Text] [PDF]

				C. E. Jones, E. M. Green, J. A. Stephens, T. C. Mueser, and N. G. Nossal Mutations of Bacteriophage T4 59 Helicase Loader Defective in Binding Fork DNA and in Interactions with T4 32 Single-stranded DNA-binding Protein J. Biol. Chem., June 11, 2004; 279(24): 25721 - 25728. [Abstract] [Full Text] [PDF]

				N. Takaha, A. L. Hawkins, C. A. Griffin, W. B. Isaacs, and D. S. Coffey High Mobility Group Protein I(Y): A Candidate Architectural Protein for Chromosomal Rearrangements in Prostate Cancer Cells Cancer Res., February 1, 2002; 62(3): 647 - 651. [Abstract] [Full Text] [PDF]

				H. WANG, H. YANG, C. J. CZURA, A. E. SAMA, and K. J. TRACEY HMGB1 as a Late Mediator of Lethal Systemic Inflammation Am. J. Respir. Crit. Care Med., November 15, 2001; 164(10): 1768 - 1773. [Full Text] [PDF]

				M. Fedele, S. Battista, G. Manfioletti, C. M. Croce, V. Giancotti, and A. Fusco Role of the high mobility group A proteins in human lipomas Carcinogenesis, October 1, 2001; 22(10): 1583 - 1591. [Abstract] [Full Text] [PDF]

				C. E. Jones, T. C. Mueser, K. C. Dudas, K. N. Kreuzer, and N. G. Nossal Bacteriophage T4 gene 41 helicase and gene 59 helicase-loading protein: A versatile couple with roles in replication and recombination PNAS, July 17, 2001; 98(15): 8312 - 8318. [Abstract] [Full Text] [PDF]

				R. Reeves, D. D. Edberg, and Y. Li Architectural Transcription Factor HMGI(Y) Promotes Tumor Progression and Mesenchymal Transition of Human Epithelial Cells Mol. Cell. Biol., January 15, 2001; 21(2): 575 - 594. [Abstract] [Full Text]

				M. F. Shannon, L. S. Coles, J. Attema, and P. Diamond The role of architectural transcription factors in cytokine gene transcription J. Leukoc. Biol., January 1, 2001; 69(1): 21 - 32. [Abstract] [Full Text]

				H. Xin, S. Taudte, N. R. Kallenbach, M. P. Limbach, and R. S. Zitomer DNA binding by single HMG box model proteins Nucleic Acids Res., October 15, 2000; 28(20): 4044 - 4050. [Abstract] [Full Text] [PDF]

R. Reeves, W. J. Leonard, and M. S. Nissen Binding of HMG-I(Y) Imparts Architectural Specificity to a Positioned Nucleosome on the Promoter of the Human Interleukin-2 Receptor alpha Gene Mol. Cell. Biol., July 1, 2000; 20(13): 4666 - 4679. [Abstract] [Full Text]