Immunochemical analysis of the exposure of high mobility group protein 14 and 17 surfaces in chromatin.

Antisera were elicited against synthetic peptides corresponding either to regions common to all members of the high mobility group 14 and 17 protein family protein or to distinct domains of the HMG-14 or HMG-17 subgroup. The antisera were used to probe the accessibility of various HMG domains in chromatin. Competitive enzyme-linked immunosorbent assays indicate that the central region of the proteins, which contains their DNA binding domain and is positively charged, is exposed to a smaller degree than the C-terminal region of the proteins, which has a net negative charge. The C-terminal regions of the HMG-14 and HMG-17 proteins are exposed and available to interact with other proteins.


Antisera
were elicited against synthetic peptides corresponding either to regions common to all members of the high mobility group 14 and 17 protein family protein or to distinct domains of the HMG-14 or HMG-17 subgroup.
The antisera were used to probe the accessibility of various HMG domains in chromatin. Competitive enzyme-linked immunosorbent assays indicate that the central region of the proteins, which contains their DNA binding domain and is positively charged, is exposed to a smaller degree than the Cterminal region of the proteins, which has a net negative charge.
The C-terminal regions of the HMG-14 and HMG-17 proteins are exposed and available to interact with other proteins.
Chromosomal proteins HMG-14' and HMG-17 are ubiquitous nuclear proteins which may be involved in the generation, or in the maintenance, of structural features characteristic of transcriptionally active chromatin (l-3). Reconstutition experiments with salt-stripped nucleosomes indicate that each nucleosome contains two potential binding sites for either 5). The proteins bind to the end of the core DNA (5) and may alter the interaction between the histone octamer and the DNA (6). The binding of HMG-14/ -17 to the nucleosome is accompanied by a small increase in its radius of gyration (7), suggesting that the proteins do not disrupt major contacts between histones and DNA. In fact, the binding of the proteins stabilizes the nucleosomal structure, elevates its thermal denaturation, and protects it from DNase I digestion (8,9). The binding of  to core particles is salt-dependent; at low ionic strengths the entire molecule is bound to the core particle while at higher ionic strengths only the central positively charged region remains bound (10).
The structure of the HMG-14/-17 chromosomal proteins is evolutionarily conserved and similar to that of certain transcriptional activators such as 12). The proteins have a modular structure and an uneven charge distribution along the polypeptide chains. The N-terminal region of the molecule has a slight net positive charge; the * The costs of publication of' this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
' The abbreviations used are: HMG, high mobility group; EDC, 1. ethyl-3,3-(3-dimethylaminopropyl)carbodiimi&; ELI&&, enzymelinked immunosorbent assay. central region, which contains the DNA binding domain, has a high positive charge while the C-terminal region has a net negative charge. Furthermore, helical wheel projections of the C-terminal regions reveal that the negative charges are clustered on one surface of the helix (3).
One of the unanswered questions pertinent to the organization of HMG proteins in chromatin, and their possible function, is whether their surfaces are accessible and free to interact with other molecules. Recent studies indicate that cooperative and synergistic protein interactions play an important role in transcriptional regulation (13,14). If HMG proteins participate in such interactions it would be expected that part of the protein surface would be exposed and capable of interacting with other macromolecules.
In the present study we use antibodies elicited by peptides derived from various regions of the molecules to probe the accessibility of HMG domains in chromatin.

MATERIALS AND METHODS
Preparation of Antigens and Antisera-The preparation and characterization of antisera to calf  and  (16) has been previously described.
The various peptides used in this study (see Fig. 1   to human, an evolutionary span of over 400 million years. Furthermore, the invariant residues are clustered producing evolutionary conserved domains. In addition, the charge distribution along the polypeptide chain is asymmetric. The N-terminal region has a slight positive charge, the central region a high positive charge, and the C-terminal region is negatively charged (3). The peptides used as immunogens were selected either from regions common to all the members of the HMG-141-17 protein family or from distinct domains of either the HMG-14 or the HMG-17 subgroups. Fig. 1 shows an alignment of all the known members of the HMG-14/-17 protein family and the regions chosen as immunogens.
Peptide 2 spans the entire DNA binding domain (22) of human HMG-17. The sequence of the first 14 amino acids in this peptide is virtually invariant among all HMGs, from trout to human. Antibodies elicited by this peptide would be expected to bind to all the proteins of the HMG-14/-17 family. The C-terminal half of peptide 2 contains a proline-rich region which is common to all members of the HMG-17 group and is not present in the HMG-14 group. This 13-amino acid-long peptide, containing 6 proline and 4 lysine residues, is named peptide 1. Antibodies elicited by this region should recognize members of the HMG-17 group. Peptide 3 corresponds to the C-terminal region of HMG-17 which is conserved and has a net charge of (-)3. Peptide 4 is derived from the DNA binding domain of HMG-14 (23) and is homologous to peptide 2. The N-terminal half of this peptide is exactly the same as that of peptide 2 while the C-terminal half contains sequences specific for HMG-14. Peptide 5 contains regions which are specific for human and calf HMG-14; however, the last 6 amino acids are invariant among all the HMG proteins. Peptide 6 corresponds to the C-terminal amino acids of the HMG-14 proteins, which is distinct from the C-terminal of the HMG-17 proteins, and has a net charge of -8.
For a more quantitative analysis, the specificity was tested by ELISA, at several sera dilutions, with various concentrations of each antigen. Table I summarizes the antisera specificity as detected by ELBA.
The 100% reaction was taken as the value obtained by reacting the antisera with the immunogen (i.e. antipeptide 1 with peptide 1) at a 1:500 dilution of both the first and second antibody (see "Materials and Methods").
The concentration of the antigen added to the well was 5 Kg/ml peptide or 1 pg/ml protein, and the reaction was stopped when the Ako5 reached 0.8-1.0. Antisera to peptide 1 reacted specifically with peptide 1 and, as expected, also recognized peptide 2. The weak reactivity with  is due to the molar concentration of the protein which is about 50 times smaller than the peptide concentration. At higher HMG-17 concentrations the reaction between antipeptide 1 and HMG-17 was similar to that obtained with the peptide. Antipeptide 2 reacts with both HMG-14 and HMG-17 as well as with peptide 2. Peptide 3 did not bind to the ELISA plate; therefore, the reaction with HMG-17 was taken as 100%. The specificity with the peptides derived from HMG-14 was similar to those derived from HMG-17 in that antipeptide 4 reacted with both HMG-14 and HMG-17 as well as with peptide 2 but not with peptide 1. Antipeptide 6 reacted better with HMG-14 than HMG-17.
Accessibility of HMG Antigenic Determinants in Chromatin-The availability of antigenic determinants was assessed by competitive ELISA assays (20, 21). Antisera to HMG proteins were preincubated with chromatin for 1 h, and the incubation mixture was centrifuged to remove the chromatin and the antibodies bound to it. The unadsorbed antibodies, remaining in the supernatant, were then tested by ELBA. The ability of chromatin to inhibit the reaction of anti-HMG-17 with HMG-17 is shown in Fig. 2 (3), did not adsorb antibodies to HMG proteins. Control antibodies did not interact with chromatin (21). We conclude therefore that the interaction of anti-HMG-17 with chromatin is specific and that only a fraction of the antibodies, generated by HMG-17, binds to chromatin.
The lack of complete binding could be due either to steric hindrance or conformational change in the protein upon binding to chromatin.
Binding of Antipeptide Sera to Chromatin-Preincubation of antibodies to peptide 1 with chromatin inhibited the reaction with HMG-17 by 85% suggesting that this region of the protein is exposed and available to interact with antibodies (Fig. 3A). Preincubation of antisera elicited by peptide 2 with chromatin inhibited its reaction with HMG-17 to only about 25% (Fig. 3B). Peptide 2 encompasses the entire sequence of peptide 1, and therefore a fraction of the antibodies elicited by peptide 2 cross-reacts with peptide 1. Since antibodies to peptide 1 bind to chromatin, we assume that at least part of the antipeptide 2 antibodies that bind to chromatin were elicited by the portion corresponding to peptide 1. Since the overall binding of antipeptide 2 is significantly lower than that of antipeptide 1, we conclude that antibodies elicited against the region specific to peptide 2 (i.e. absent from peptide 1, see Fig. 1) bind poorly to chromatin. These findings suggest that access to that region is sterically hindered or that the conformation of that region in chromatin is significantly different from that of the protein free in solution. Antibodies to peptide 3 encompassing the C-terminal of the molecule were efficiently adsorbed by chromatin so that the reaction with HMG-17 was inhibited by about 80% (Fig. 3C), suggesting that this region of the protein is exposed and able to interact with other molecules.
As an additional test for the specificity of the sera, we examined the ability of heterologous peptides to inhibit the reaction between a peptide and its antibody under the same  conditions as the inhibition studies described above. As shown in Fig. 30, the interaction of antipeptide 1 with HMG-17 was fully inhibited with peptide 1, only 40% inhibited by preincubation with peptide 2, and not inhibited by peptide 4. Likewise, the binding of antipeptide 4 to HMG-14 was fully inhibited by peptide 4, partially inhibited by peptide 2, and not inhibited by peptide 1 (not shown). These results verified that the antibodies elicited by the peptides recognize different regions of the proteins. Differential Exposure of HMG Protein Surfaces in Chromatin-The ability of chromatin to inhibit the reaction between an antibody and its antigen depends on the exposure of the antigenic site in chromatin.
The data shown in Table  II indicate that the ability of chromatin to inhibit the interaction between an antibody and its antigen varied among the various antibodies. The table documents the ability of chromatin to inhibit the reaction between various antisera and HMG-14 or HMG-17. The percent inhibition is reported for two points in the ELISA assay, the midpoint of the maximal reaction and at 100% reaction obtained between the antisera and the protein. The 100% reaction is an arbitrary point taken 30-60 min after the addition of substrate when the Abo5 reached 0.8-1.0. For example, at 50% of the maximal reaction chromatin inhibited only 4% of the reaction between antipep-tide 2 and HMG-17 while at the end point (100% reaction) chromatin inhibited the ELISA by 14%. The data indicate that antisera to peptides 1, 3, and 6 bind to chromatin to a significantly larger degree than antisera to peptides 2 and 4. Peptide 1 has partial homology with a region of histone Hl, and although the reaction with histone Hl is only 5% of that obtained with HMG-17 (not shown), it is feasible that part of the reaction is due to binding to this histone. The weak binding of the antibody to peptides 2 and 4 suggests that this region is not accessible to antibody binding, either because of steric hindrance or conformational change. This region constitutes the DNA binding domain of the proteins (22, 23); therefore we assume that it is closely associated with DNA. This association may cause steric hindrance or conformational changes, both of which would decrease the binding of antibodies to this protein domain. The negatively charged Cterminal region of the proteins is recognized by the antisera, suggesting that this region is free to interact with other components.
The accessibility of the various regions of the molecule to antibody binding parallels the interaction of the molecule with nucleosomes. Cook et al. (10) reported that the basic central region of the molecule binds more strongly to the core DNA than the acidic C-terminal region. The model presented inFig. 4 incorporates the information obtained here with that available from the literature. The proteins are located near the entry/exit points of the core DNA, the central region of the molecule is near histone H2A, and the Cterminal region is relatively exposed.
The transcription activating domain of certain regulatory proteins resides in a negatively charged region of the molecule (13, 14). By analogy, the C termini of the HMG-14/-17 proteins may be involved in a similar function. A minimum requirement for this function is that this region should be exposed and available for interactions with other molecules. The results presented here indicate that the C terminus of the proteins is exposed and available to participate in reactions which require contact with other proteins. ;: