Structure of the HMG box motif in the B-domain of HMG1.

The conserved, abundant chromosomal protein HMG1 consists of two highly homologous, folded, basic DNA-binding domains, each of approximately 80 amino acid residues, and an acidic C-terminal tail. Each folded domain represents an 'HMG box', a sequence motif recently recognized in certain sequence-specific DNA-binding proteins and which also occurs in abundant HMG1-like proteins that bind to DNA without sequence specificity. The HMG box is defined by a set of highly conserved residues (most distinctively aromatic and basic) and appears to define a novel DNA-binding structural motif. We have expressed the HMG box region of the B-domain of rat HMG1 (residues 88-164 of the intact protein) in Escherichia coli and we describe here the determination of its structure by 2D 1H-NMR spectroscopy. There are three alpha-helices (residues 13-29, 34-48 and 50-74), which together account for approximately 75% of the total residues and contain many of the conserved basic and aromatic residues. Strikingly, the molecule is L-shaped, the angle of approximately 80 degrees between the two arms being defined by a cluster of conserved, predominantly aromatic, residues. The distinctive shape of the HMG box motif, which is distinct from hitherto characterized DNA-binding motifs, may be significant in relation to its recognition of four-way DNA junctions.     

DNA binding by single HMG box model proteins

The HMG1/2 family is a large group of proteins that share a conserved sequence of ~80 amino acids rich in basic, aromatic and proline side chains, referred to as an HMG box. Previous studies show that HMG boxes can bind to DNA in a structure-specific manner. To define the basis for DNA recognition by HMG boxes, we characterize the interaction of two model HMG boxes, one a structure-specific box, rHMGb from the rat HMG1 protein, the other a sequence-specific box, Rox1 from yeast, with oligodeoxynucleotide substrates. Both proteins interact with single-stranded oligonucleotides in this study to form 1:1 complexes. The stoichiometry of binding of rHMGb to duplex or branched DNAs differs: for a 16mer duplex we find a weak 2:1 complex, while a 4:1 protein:DNA complex is detected with a four-way DNA junction of 16mers in the presence of Mg²⁺. In the case of the sequence-specific Rox1 protein we find tight 1:1 and 2:1 complexes with its cognate duplex sequence and again a 4:1 complex with four-way branched DNA. If the DNA branching is reduced to three arms, both proteins form 3:1 complexes. We believe that these multimeric complexes are relevant for HMG1/2 proteins in vivo, since Mg²⁺ is present in the nucleus and these proteins are expressed at a very high level.     

Interactions between HMG boxes.

Many proteins consist of subdomains that can fold and function independently. We investigate here the interaction between the two high mobility group (HMG) box subdomains of the nuclear protein rHMG1. An HMG box is a conserved amino acid sequence of approximately 80 amino acids rich in basic, aromatic and proline side chains that is active in binding DNA in a sequence or structure-specific manner. In the case of HMG1, each box can bind structural DNA substrates including four-way junctions (4WJs) and branched or kinked DNA duplexes. Since proteins containing up to six HMG boxes are known, the question arises whether linking subdomains together influences the folding or function of individual boxes. In an effort to understand interactions between individual DNA-binding domains in HMG1, we created new fusion proteins: one is an inversion of the order of the AB di-domain in HMG1 (BA); in the second, we added a third A domain C-terminal to the AB di-domain (ABA). Pairs of boxes, AB or BA, behave similarly and are functionally active. By contrast, the ABA triple subdomain construct is partially unfolded and is less active than individual boxes or di-domains. Thus, long-range inter-domain effects can influence the activity of HMG boxes.     

Related function of mouse SOX3, SOX9, and SRY HMG domains assayed by male sex determination

Sox genes encode proteins related to each other, and to the sex determining gene Sry, by the presence of a DNA binding motif known as the HMG domain. Although HMG domains can bind to related DNA sequences, Sox gene products may achieve target gene specificity by binding to preferred target sequences or by interacting with specific partner proteins. To assess their functional similarities, we replaced the HMG box of Sry with the HMG box of Sox3 or Sox9 and tested whether these constructs caused sex reversal in XX mice. Our results indicate that such chimeric transgenes can functionally replace Sry and elicit development of testis cords, male patterns of gene expression, and elaboration of male secondary sexual characteristics. This implies that chimeric SRY proteins with SOX HMG domains can bind to and regulate SRY target genes and that potential SRY partner factor interactions are not disrupted by HMG domain substitutions. genesis 28:111-124, 2000.     

Structural requirements for cooperative binding of HMG1 to DNA minicircles

DNA minicircles, where the length of DNA is below the persistence length, are highly effective, preferred, ligands for HMG-box proteins. The proteins bind to them "structure-specifically" with affinities in the nanomolar range, presumably to an exposed widened minor groove. To understand better the basis of this preference, we have studied the binding of HMG1 (which has two tandem HMG boxes linked by a basic extension to a long acidic tail) and Drosophila HMG-D (one HMG box linked by a basic region to a short and less acidic tail), and their HMG-box domains, to 88 bp and 75 bp DNA minicircles. In some cases we see cooperative binding of two molecules to the circles. The requirements for strong cooperativity are two HMG boxes and the basic extension; the latter also appears to stabilize and constrain the complex, preventing binding of further protein molecules. HMG-D, with a single HMG box, does not bind cooperatively. In the case of HMG1, the acidic tail is not required for cooperativity and does not affect binding significantly, in contrast to a much greater effect with linear DNA, or even four-way junctions (another distorted DNA substrate). Such effects could be relevant in the hierarchy of binding of HMG-box proteins to DNA distortions in vivo, where both single-box and two-box proteins might co-exist, with or without basic extensions and acidic tails.     

The energetics of HMG box interactions with DNA: thermodynamics of the DNA binding of the HMG box from mouse sox-5

The energetics of the Sox-5 HMG box interaction with DNA duplexes, containing the recognition sequence AACAAT, were studied by fluorescence spectroscopy, isothermal titration calorimetry (ITC) and differential scanning calorimetry (DSC). Fluorescence titration showed that the association constant of this HMG box with the duplexes is of the order 4x10(7) M(-1), increasing somewhat with temperature rise, i.e. the Gibbs energy is -40 kJ mol(-1) at 5 degrees C, decreasing to -48 kJ mol(-1) at 32 degrees C. ITC measurements of the enthalpy of association over this temperature range showed an endothermic effect below 17 degrees C and an exothermic effect above, suggesting a heat capacity change on binding of about -4 kJ K(-1) mol(-1), a value twice larger than expected from structural considerations. A straightforward interpretation of ITC data in heat capacity terms assumes, however, that the heat capacities of all participants in the association reaction do not change over the considered temperature range. Our previous studies showed that over the temperature range of the ITC experiments the HMG box of Sox-5 starts to unfold, absorbing heat and the heat capacities of the DNA duplexes also increase significantly. These heat capacity effects differ from that of the DNA/Sox-5 complex. Correcting the ITC measured binding enthalpies for the heat capacity changes of the components and complex yielded the net enthalpies which exhibit a temperature dependence of about -2 kJ K(-1) mol(-1), in good agreement with that predicted on the basis of dehydration of the protein-DNA interface. Using the derived heat capacity change and the enthalpy and Gibbs energy of association measured at 5 degrees C, the net enthalpy and entropy of association of the fully folded HMG box with the target DNA duplexes was determined over a broad temperature range. These functions were compared with those for other known cases of sequence specific DNA/protein association. It appears that the enthalpy and entropy of association of minor groove binding proteins are more positive than for proteins binding in the major groove. The observed thermodynamic characteristics of protein binding to the A+T-rich minor groove of DNA might result from dehydration of both polar and non-polar groups at the interface and release of counterions. The expected entropy of dehydration was calculated and found to be too large to be compensated by the negative entropy of reduction of translational/rotational freedom. This implies that DNA/HMG box association proceeds with significant decrease of conformational entropy, i.e. reduction in conformational mobility.     

Four differently chromatin‐associated maize HMG domain proteins modulate DNA structure and act as architectural elements in nucleoprotein complexes

In contrast to other eukaryotes which usually express two closely related HMG1-like proteins, plant cells have multiple relatively variable proteins of this type. A systematic analysis of the DNA-binding properties of four chromosomal HMG domain proteins from maize revealed that they bind linear DNA with similar affinity. HMGa, HMGc1/2 and HMGd specifically recognise diverse DNA structures such as DNA mini-circles and supercoiled DNA. They induce DNA-bending, and constrain negative superhelical turns in DNA. In the presence of DNA, the HMG domain proteins can self-associate, whereas they are monomeric in solution. The maize HMG1-like proteins have the ability to facilitate the formation of nucleoprotein structures to different extents, since they can efficiently replace a bacterial chromatin-associated protein required for the site-specific β-mediated recombination. A variable function of the HMG1-like proteins is indicated by their differential association with maize chromatin, as judged by their ‘extractability’ from chromatin with spermine and ethidium bromide. Collectively, these findings suggest that the various plant chromosomal HMG domain proteins could be adapted to act in different nucleoprotein structures in vivo.     

Dual binding modes for an HMG domain from human HMGB2 on DNA

High mobility group B (HMGB) proteins contain two HMG box domains known to bind without sequence specificity into the DNA minor groove, slightly intercalating between basepairs and producing a strong bend in the DNA backbone. We use optical tweezers to measure the forces required to stretch single DNA molecules. Parameters describing DNA flexibility, including contour length and persistence length, are revealed. In the presence of nanomolar concentrations of isolated HMG box A from HMGB2, DNA shows a decrease in its persistence length, where the protein induces an average DNA bend angle of 114 +/- 21 degrees for 50 mM Na+, and 87 +/- 9 degrees for 100 mM Na+. The DNA contour length increases from 0.341 +/- 0.003 to 0.397 +/- 0.012 nm per basepair, independent of salt concentration. In 50 mM Na+, the protein does not unbind even at high DNA extension, whereas in 100 mM Na+, the protein appears to unbind only below concentrations of 2 nM. These observations support a flexible hinge model for noncooperative HMG binding at low protein concentrations. However, at higher protein concentrations, a cooperative filament mode is observed instead of the hinge binding. This mode may be uniquely characterized by this high-force optical tweezers experiment.     

Mutagenesis study on the zebra fish SOX9 high-mobility group: comparison of sequence and non-sequence specific HMG domains

A unique class of proteins, containing high-mobility group (HMG) domain(s), recognizes unusual DNA structures and/or bends specific to AT-rich linear double-stranded DNA. The DNA binding feature of these proteins is exhibited in the HMG domain(s). Although the sequence specific and non-sequence specific HMG domains exhibit very high degrees of sequence similarity, the reasons for the difference between their DNA recognition mechanisms are unclear. A series of zebra fish SOX9 HMG domain mutants was prepared in an effort to elucidate the importance of various residues on protein stability and DNA binding. This study is the first of a comprehensive mutagenesis study on a sequence specific HMG domain. Comparing how various residues influence sequence specific and non-sequence specific HMG domains helps us to rationalize their mode of action. Positively charged amino acids concentrated at the surface of sequence specific HMG domains recognize specific, linear AT-rich DNA segments. After the negative charges at the surface of the DNA are neutralized, the hydrophobic residues of the protein may intercalate DNA. Phenylalanine at position 12 plays a crucial role in the sequence specific HMG domain. The differences in pI values, the instability index, and DNA contact regions between sequence and non-sequence specific HMG domains are associated with their functional modes.     

The HMG domain of the ROX1 protein mediates repression of HEM13 through overlapping DNA binding and oligomerization functions.

The ROX1 gene of Saccharomyces cerevisiae encodes a protein required for the repression of genes expressed under anaerobic conditions. ROX1 belongs to a family of DNA binding proteins which contain the high mobility group motif (HMG domain). To ascertain whether the HMG domain of ROX1 is required for specific DNA binding we synthesized a series of ROX1 protein derivatives, either in vitro or in Escherichia coli as fusions to glutathione S-transferase (GST) protein, and tested them for their ability to bind to DNA. Both ROX1 proteins that were synthesized in vitro and GST-ROX1 fusion proteins containing the intact HMG domain were able to bind to specific target DNA sequences. In contrast, ROX1 proteins which contained deletions within the HMG domain were no longer capable of binding to DNA. The oligomerization of ROX1 in vitro was demonstrated using affinity-purified GST-ROXI protein and ROX1 labelled with [35S]methionine. Using various ROX1 protein derivatives we were able to demonstrate that the domain required for ROX1-ROX1 interaction resides within the N-terminal 100 amino acids which constitute the HMG domain. Therefore, the HMG domain is required for both DNA binding activity and oligomerization of ROX1.     

Hierarchy of binding sites for chromosomal proteins HMG 1 and 2 in supercoiled deoxyribonucleic acid

The interaction of chromosomal proteins HMG 1 and 2 with various DNA structures has been examined with plasmid pPst-0.9, which contains DNA sequences that can form the Z-DNA conformation and palindromic sequences that can form cruciform structures. Direct binding and competition experiments with 32P-labeled plasmid indicated that proteins HMG 1 and 2 preferentially bind to supercoiled form I DNA as compared to double-stranded linear DNA. The preferential binding to form I is due to the presence of single-stranded regions in this DNA. The binding of HMG 1 and 2 to the form I plasmid results in inhibition of S1 nuclease digestion in a selective manner. The B-Z junction is preferentially protected as compared to the cruciform, which in turn is more protected than other minor S1-sensitive structures present in pPst-0.9. Our results indicate that the binding of HMG 1 and 2 proteins to DNA is not random in that HMG 1 and 2 can distinguish between various S1 nuclease sensitive sites in the plasmid. The existence of a hierarchy of DNA binding sites for these proteins suggests that they can selectively affect the structure of distinct regions in the genome.     

A novel activity of HMG domains: promotion of the triple-stranded complex formation between DNA containing (GGA/TCC)11 and d(GGA)11 oligonucleotides.

The high mobility group protein (HMG)-box is a DNA-binding domain found in many proteins that bind preferentially to DNA of irregular structures in a sequence-independent manner and can bend the DNA. We show here that GST-fusion proteins of HMG domains from HMG1 and HMG2 promote a triple-stranded complex formation between DNA containing the (GGA/TCC)11 repeat and oligonucleotides of d(GGA)11 probably due to G:G base pairing. The activity is to reduce association time and requirements of Mg2+ and oligonucleotide concentrations. The HMG box of SRY, the protein determining male-sex differentiation, also has the activity, suggesting that it is not restricted to the HMG-box domains derived from HMG1/2 but is common to those from other members of the HMG-box family of proteins. Interestingly, the box-AB and box-B of HMG1 bend DNA containing the repeat, but SRY fails to bend in a circularization assay. The difference suggests that the two activities of association-promotion and DNA bending are distinct. These results suggest that the HMG-box domain has a novel activity of promoting the association between GGA repeats which might be involved in higher-order architecture of chromatin.     

DNA binding and bending by HMG boxes: energetic determinants of specificity

To clarify the physical basis of DNA binding specificity, the thermodynamic properties and DNA binding and bending abilities of the DNA binding domains (DBDs) of sequence-specific (SS) and non-sequence-specific (NSS) HMG box proteins were studied with various DNA recognition sequences using micro-calorimetric and optical methods. Temperature-induced unfolding of the free DBDs showed that their structure does not represent a single cooperative unit but is subdivided into two (in the case of NSS DBDs) or three (in the case of SS DBDs) sub-domains, which differ in stability. Both types of HMG box, most particularly SS, are partially unfolded even at room temperature but association with DNA results in stabilization and cooperation of all the sub-domains. Binding and bending measurements using fluorescence spectroscopy over a range of ionic strengths, combined with calorimetric data, allowed separation of the electrostatic and non-electrostatic components of the Gibbs energies of DNA binding, yielding their enthalpic and entropic terms and an estimate of their contributions to DNA binding and bending. In all cases electrostatic interactions dominate non-electrostatic in the association of a DBD with DNA. The main difference between SS and NSS complexes is that SS are formed with an enthalpy close to zero and a negative heat capacity effect, while NSS are formed with a very positive enthalpy and a positive heat capacity effect. This indicates that formation of SS HMG box-DNA complexes is specified by extensive van der Waals contacts between apolar groups, i.e. a more tightly packed interface forms than in NSS complexes. The other principal difference is that DNA bending by the NSS DBDs is driven almost entirely by the electrostatic component of the binding energy, while DNA bending by SS DBDs is driven mainly by the non-electrostatic component. The basic extensions of both categories of HMG box play a similar role in DNA binding and bending, making solely electrostatic interactions with the DNA.     

Diversification Pattern of the HMG and SOX Family Members During Evolution

From a database containing the published HMG protein sequences, we constructed an alignment of the HMG box functional domain based on sequence identity. Due to the large number of sequences (more than 250) and the short size of this domain, several data sets were used. This analysis reveals that the HMG box superfamily can be separated into two clearly defined subfamilies: (i) the SOX/MATA/TCF family, which clusters proteins able to bind to specific DNA sequences; and (ii) the HMG/UBF family, which clusters members which bind non specifically to DNA. The appearance and diversification of these subfamilies largely predate the split between the yeast and the metazoan lineages. Particular emphasis was placed on the analysis of the SOX subfamily. For the first time our analysis clearly identified the SOX subfamily as structured in six groups of genes named SOX5/6, SRY, SOX2/3, SOX14, SOX4/22, and SOX9/18. The validity of these gene clusters is confirmed by their functional characteristics and their sequences outside the HMG box. In sharp contrast, there are only a few robust branching patterns inside the UBF/HMG family, probably because of the much more ancient diversification of this family than the diversification of the SOX family. The only consistent groups that can be detected by our analysis are HMG box 1, vertebrate HMG box 2, insect SSRP, and plant HMG. The various UBF boxes cannot be clustered together and their diversification appears to be extremely ancient, probably before the appearance of metazoans. Received: 20 July 1998 / Accepted: 19 October 1998