Conformation and domain structure of the non-histone chromosomal proteins, HMG 1 and 2. Isolation of two folded fragments from HMG 1 and 2

The human RNA polymerase I structure reveals an HMG-like docking domain specific to metazoans

We characterize the human RNA polymerase I by evolutionary biochemistry and cryo-EM revealing a built-in structural domain that apparently serves as transcription factor–binding platform in metazoans.

Transcription of the ribosomal RNA precursor by RNA polymerase (Pol) I is a major determinant of cellular growth, and dysregulation is observed in many cancer types. Here, we present the purification of human Pol I from cells carrying a genomic GFP fusion on the largest subunit allowing the structural and functional analysis of the enzyme across species. In contrast to yeast, human Pol I carries a single-subunit stalk, and in vitro transcription indicates a reduced proofreading activity. Determination of the human Pol I cryo-EM reconstruction in a close-to-native state rationalizes the effects of disease-associated mutations and uncovers an additional domain that is built into the sequence of Pol I subunit RPA1. This “dock II” domain resembles a truncated HMG box incapable of DNA binding which may serve as a downstream transcription factor–binding platform in metazoans. Biochemical analysis, in situ modelling, and ChIP data indicate that Topoisomerase 2a can be recruited to Pol I via the domain and cooperates with the HMG box domain–containing factor UBF. These adaptations of the metazoan Pol I transcription system may allow efficient release of positive DNA supercoils accumulating downstream of the transcription bubble.

The interaction of HMGB1 and linker histones occurs through their acidic and basic tails

H1 and HMGB1 bind to linker DNA in chromatin, in the vicinity of the nucleosome dyad. They appear to have opposing effects on the nucleosome, H1 stabilising it by "sealing" two turns of DNA around the octamer, and HMGB1 destabilising it, probably by bending the adjacent DNA. Their presence in chromatin might be mutually exclusive. Displacement/replacement of one by the other as a result of their highly dynamic binding in vivo might, in principle, involve interactions between them. Chemical cross-linking and gel-filtration show that a 1:1 linker histone/HMGB1 complex is formed, which persists at physiological ionic strength, and that complex formation requires the acidic tail of HMGB1. NMR spectroscopy shows that the linker histone binds, predominantly through its basic C-terminal domain, to the acidic tail of HMGB1, thereby disrupting the interaction of the tail with the DNA-binding faces of the HMG boxes. A potential consequence of this interaction is enhanced DNA binding by HMGB1, and concomitantly lowered affinity of H1 for DNA. In a chromatin context, this might facilitate displacement of H1 by HMGB1.

Dual roles for HMGB1: DNA binding and cytokine

Effective therapies against overwhelming Gram-negative bacteremia, or sepsis, have eluded successful development. The discovery that tumor necrosis factor (TNF), a host-derived inflammatory mediator, was both necessary and sufficient to recapitulate Gram-negative sepsis raised cautious optimism for developing a targeted therapeutic. However, the rapid kinetics of the TNF response to infection defined an extremely narrow window of opportunity during which anti-TNF therapeutics could be successfully administered. HMGB1 was previously studied as a DNA-binding protein involved in DNA replication, repair, and transcription; and as a membrane-associated protein that mediates neurite outgrowth. A decade-long search has culminated in our identification of HMGB1 as a late mediator of endotoxemia. HMGB1 is released by macrophages upon exposure to endotoxin, activates many other pro-inflammatory mediators, and is lethal to otherwise healthy animals. Elevated levels of HMGB1 are observed in the serum of patients with sepsis, and the highest levels were found in those patients that died. The delayed kinetics of HMGB1 release indicate that it may be useful to target this toxic cytokine in the development of future therapies.

Solution structure and backbone dynamics of the DNA-binding domain of mouse Sox-5

The fold of the murine Sox-5 (mSox-5) HMG box in free solution has been determined by multidimensional NMR using (15)N-labeled protein and has been found to adopt the characteristic twisted L-shape made up of two wings: the major wing comprising helix 1 (F10--F25) and helix 2 (N32--A43), the minor wing comprising helix 3 (P51--Y67) in weak antiparallel association with the N-terminal extended segment. (15)N relaxation measurements show considerable mobility (reduced order parameter, S(2)) in the minor wing that increases toward the amino and carboxy termini of the chain. The mobility of residues C-terminal to Q62 is significantly greater than the equivalent residues of non-sequence-specific boxes, and these residues show a weaker association with the extended N-terminal segment than in non-sequence boxes. Comparison with previously determined structures of HMG boxes both in free solution and complexed with DNA shows close similarity in the packing of the hydrophobic cores and the relative disposition of the three helices. Only in hSRY/DNA does the arrangement of aromatic sidechains differ significantly from that of mSox-5, and only in rHMG1 box 1 bound to cisplatinated DNA does helix 1 have no kink. Helix 3 in mSox-5 is terminated by P68, a conserved residue in DNA sequence-specific HMG boxes, which results in the chain turning through approximately 90 degrees.

The energetics of HMG box interactions with DNA. Thermodynamic description of the box from mouse Sox-5

The structural energetics of the HMG box from the DNA-binding protein mouse Sox-5 were examined calorimetrically. It was found that this box, notwithstanding its small size (molecular mass about 10 kDa), does not behave as a single cooperative unit and, on heating, the box reversibly unfolds in two separate stages. The first transition (tt approximately 34 degrees C) involves about 40% of the total enthalpy and the second (tt approximately 46 degrees C) the remainder. Both transitions proceed with significant heat capacity increment, showing that they are associated with the unfolding of two sub-domains having non-polar cores. According to heat capacity, ellipticity, fluorescence and NMR criteria, this HMG box is in a fully compact native state only below 5 degrees C. HMG boxes consist of two approximately orthogonal wings: the minor wing comprises helix 3 and its associated antiparallel N-terminal strand, whilst the major wing is composed of helices I and II. Analysis of the fluorescence and NMR spectra for this box obtained at different temperatures shows that the lower melting transition can be assigned to the minor wing and the upper transition to the major wing. Under physiological conditions (37 degrees C), the minor wing is considerably unfolded, whilst the major wing is essentially fully folded. DNA binding in vivo therefore involves refolding of the minor wing.

Molecular cloning of polybromo, a nuclear protein containing multiple domains including five bromodomains, a truncated HMG-box, and two repeats of a novel domain

A number of transcription factors that act as adaptor proteins have been found to contain an 87 amino acid domain called the bromodomain. In a study to identify and characterise bromodomain proteins expressed in chicken cells, a novel gene has been isolated which encodes five repeats of the bromodomain. In addition, the encoded protein, termed polybromo, contains four other domains: an unusual truncated HMG box, two repeats of a novel domain which we term the BAH domain and a sequence related to a region within the regulatory domain of the DNA cytosine-5 methyltransferase enzyme. Polybromo was found to be related to a yeast protein U19102 which has two bromo domains, a BAH domain and the DNA methyltransferase-related sequence. Antibodies that were raised against polybromo were used in confocal microscopy analysis to show that the 180-kDa polybromo protein is located within the nucleus but excluded from the nucleolus. Gel filtration analysis of nuclear extracts demonstrate that polybromo is part of a large complex with a mass of approximately 2 million dalton.

The HMG-1 box protein family: classification and functional relationships

The abundant and highly-conserved nucleoproteins comprising the high mobility group-1/2 (HMG-1/2) family contains two homologous basic domains of about 75 amino acids. These basic domains, termed HMG-1 boxes, are highly structured and facilitate HMG-DNA interactions. Many proteins that regulate various cellular functions involving DNA binding and whose target DNA sequences share common structural characteristics have been identified as having an HMG-1 box; these proteins include the RNA polymerase I transcription factor UBF, the mammalian testis-determining factor SRY and the mitochondrial transcription factors ABF2 and mtTF1, among others. The sequences of 121 HMG-1 boxes have been compiled and aligned in accordance with thermodynamic results from homology model building (threading) experiments, basing the alignment on structure rather than by using traditional sequence homology methods. The classification of a representative subset of these proteins was then determined using standard least-squares distance methods. The proteins segregate into two groups, the first consisting of HMG-1/2 proteins and the second consisting of proteins containing the HMG-1 box but which are not canonical HMG proteins. The proteins in the second group further segregate based on their function, their ability to bind specific sequences of DNA, or their ability to recognize discrete non-B-DNA structures. The HMG-1 box provides an excellent example of how a specific protein motif, with slight alteration, can be used to recognize DNA in a variety of functional contexts.

Homology model building of the HMG-1 box structural domain

Nucleoproteins belonging to the HMG-1/2 family possess homologous domains approximately 75 amino acids in length. These domains, termed HMG-1 boxes, are highly structured, compact, and mediate the interaction between HMG-1 box-containing proteins and DNA in a variety of biological contexts. Homology model building experiments on HMG-1 box sequences 'threaded' through the 1H-NMR structure of an HMG-1 box from rat indicate that the domain does not have rigid sequence requirements for its formation. Energy calculations indicate that the structure of all HMG-1 box domains is stabilized primarily through hydrophobic interactions. We have found structural relationships in the absence of statistically significant sequence similarity, identifying several candidate proteins which could possibly assume the same three-dimensional conformation as the rat HMG-1 box motif. The threading technique provides a method by which significant structural similarities in a diverse protein family can be efficiently detected, and the 'structural alignment' derived by this method provides a rational basis through which phylogenetic relationships and the precise sites of interaction between HMG-1 box proteins and DNA can be deduced.

cDNA sequence and structure of a gene encoding trout testis high-mobility-group-1 protein

Perchloric acid extraction of trout testis nuclei revealed the presence of two large high-mobility-group (HMG) proteins, HMG-T1 and HMG-T2. The sequence of a complete cDNA (1407 bp) for trout testis HMG-1 protein (referred as to HMG-T1) has been determined. The deduced HMG-T1 protein contains 203 amino acids with more than 86% similarity to mammalian HMG-1 proteins. A single-sized mRNA for HMG-T1 has been detected by Northern-blot analysis consistent with the size derived from the HMG-T1 cDNA. Amplification of human and trout genomic DNAs by polymerase chain reaction using primers specific for trout and human HMG-1 cDNAs revealed that unlike the human genome, which contains predominantly intronless HMG-1 sequences, intronless HMG-T1 sequences were not found in the fish genome. Southern-blot analysis suggested that the trout testis HMG-1 gene is encoded by at least two sequences with high similarity. A gene encoding HMG-T1 protein has been isolated from a trout testis genomic library and by PCR of trout genomic DNA (3879 bp). The trout testis HMG-1 gene is organized into five exons (four exons corresponding to the protein-coding region) and its exon/intron boundaries are identical to those of the human HMG-2 gene [Shirakawa, H. & Yoshida, M. (1992). J. Biol. Chem. 267, 6641-6645] suggesting the evolution of HMG-1 and HMG-2 genes from a common ancestor.

DNA looping by the HMG-box domains of HMG1 and modulation of DNA binding by the acidic C-terminal domain

We have compared HMG1 with the product of tryptic removal of its acidic C-terminal domain termed HMG3, which contains two 'HMG-box' DNA-binding domains. (i) HMG3 has a higher affinity for DNA than HMG1. (ii) Both HMG1 and HMG3 supercoil circular DNA in the presence of topoisomerase I. Supercoiling by HMG3 is the same at approximately 50 mM and approximately 150 mM ionic strength, as is its affinity for DNA, whereas supercoiling by HMG1 is less at 150 mM than at 50 mM ionic strength although its affinity for DNA is unchanged, showing that the acidic C-terminal tail represses supercoiling at the higher ionic strength. (iii) Electron microscopy shows that HMG3 at a low protein:DNA input ratio (1:1 w/w; r = 1), and HMG1 at a 6-fold higher ratio, cause looping of relaxed circular DNA at 150 mM ionic strength. Oligomeric protein 'beads' are apparent at the bases of the loops and at cross-overs of DNA duplexes. (iv) HMG3 at high input ratios (r = 6), but not HMG1, causes DNA compaction without distortion of the B-form. The two HMG-box domains of HMG1 are thus capable of manipulating DNA by looping, compaction and changes in topology. The acidic C-tail down-regulates these effects by modulation of the DNA-binding properties.

Molecular cloning, expression analysis, and chromosomal localization of mouse Hmg1-containing sequences

We isolated clones encoding the mouse high-mobility-group (Hmg) chromatin protein, Hmg1, from a 7.5-day mouse embryo cDNA library. The translated amino acid sequence encodes a protein of 24,890 daltons and is identical to previously characterized mouse, rat, and hamster Hmg1. However, comparison of the two mouse Hmg1 cDNA sequences revealed nine sequence alterations. This observation, together with the finding of a complex pattern of hybridizing bands in genomic Southern analysis, suggests that mouse Hmg1 is encoded by a multigene family. The expression of Hmg1 was examined by Northern analysis of RNA isolated from the early mouse embryo and revealed a predominant 1.5-kb transcript in conjunction with low levels of a 2.5-kb transcript. Further analysis of mouse embryos by in situ hybridization showed that Hmg1 transcripts are expressed in high abundance during early mouse embryogenesis. As development progresses, Hmg1 transcript abundance is modulated in a spatially restricted and developmentally regulated manner. Chromosomal localization with recombinant inbred strains revealed that Hmg1-related sequences are widely dispersed in the mouse genome. Here we also report the mapping of six Hmg1 loci to mouse Chromosomes (Chrs) 10, 13, 16, and 17.

Solution structure of a DNA-binding domain from HMG1

We have determined the tertiary structure of box 2 from hamster HMG1 using bacterial expression and 3D NMR. The all alpha-helical fold is in the form of a V-shaped arrowhead with helices along two edges and one rather flat face. This architecture is not related to any of the known DNA binding motifs. Inspection of the fold shows that the majority of conserved residue positions in the HMG box family are those involved in maintaining the tertiary structure and thus all homologous HMG boxes probably have essentially the same fold. Knowledge of the tertiary structure permits an interpretation of the mutations in HMG boxes known to abrogate DNA binding and suggests a mode of interaction with bent and 4-way junction DNA.

Mapping of a sequence essential for the nuclear transport of the Xenopus ribosomal transcription factor xUBF using a simple coupled translation-transport and acid extraction approach

The amino acid sequences directing the nuclear transport of the RNA polymerase I transcription factor xUBF have been studied by a novel combination of in oocyte-coupled translation-nuclear transport and selective HCl extraction. Synthetic mRNA was used to direct the translation of labeled xUBF and its mutants in microinjected oocytes. After manual dissection of nuclei and cytoplasm, labeled xUBF and mutants were isolated essentially pure by HCl extraction. Using deletion mutations, a sequence essential, but not necessarily sufficient, for nuclear transport was mapped to a 29-amino-acid segment lying between the most carboxy-terminal putative HMG-box DNA-binding domain, HMG-box 5, and the highly acidic carboxy-terminal domain. It was shown that deletion of only 5 amino acids from this segment eliminated xUBF transport, and it could be deduced that at least 11 of the 29 amino acids were essential for nuclear transport. The segment of xUBF necessary for nuclear transport contains a sequence conforming to the bipartite nuclear transport motif consensus, but this sequence in itself was insufficient for nuclear transport.

The RNA polymerase I transcription factor xUBF contains 5 tandemly repeated HMG homology boxes

The RNA polymerase I transcription factor UBF has been identified in human, mouse, rat and Xenopus and the primary structure of the human protein has been determined. Human UBF was shown to contain four tandem homologies to the folding domains of the HMG1 and 2 proteins and hence to belong to a previously unrecognised family of 'HMG-box' transcription factors. Here, cDNA clones encoding the Xenopus laevis UBF (xUBF) have been isolated and sequenced. Northern and Southern blots revealed that in tissue culture cells, xUBF is coded on a single major mRNA size species by a small number of genes. The deduced primary structure of xUBF is highly homologous with the human protein except for a central deletion which removes most of one HMG-box. This explains the major size difference between the X. laevis and human proteins and may well explain their different transcriptional specificities. It is shown that xUBF contains 5 tandemly repeated HMG-boxes and that by analogy the human protein contains 6.

Expression of P30, a protein with adhesive properties, in Schwann cells and neurons of the developing and regenerating peripheral nerve

P30 is a heparin-binding protein with adhesive and neurite outgrowth-promoting properties present at high levels in the developing rat central nervous system (Rauvala, H., and R. Pihlaskari. 1987 J. Biol. Chem. 262:16625-16635). Partial sequencing of p30 has revealed homology or identity with HMG-1 (Rauvala, H., J. Merenmies, R. Pihlaskari, M. Korkolainen, M.-L. Huhtala, and P. Panula. 1988. J. Cell Biol. 107:2292-2305), a 28-kD protein that was originally purified from the thymus (Goodwin, G.H., C. Sanders, and E. W. Johns. 1973. Eur. J. Biochem. 38:14-19) which binds DNA in vitro. We have analyzed the distribution of p30 in the developing rat peripheral nervous system (PNS). P30 was detected by immunohistochemistry and Western blot analysis using antibodies raised against intact p30 and against a synthetic peptide corresponding to the amino terminus of the p30 molecule. P30 was localized to nonnuclear compartments of neurons and peripheral glial cells (Schwann cells). P30 immunoreactivity of PNS neurons persisted into adulthood. In contrast, Schwann cell staining decreased after the second postnatal week and was not detectable in adult animals. Neuron-Schwann cell contact was correlated with diminished p30 levels in Schwann cells. Schwann cells of the normal adult sciatic nerve did not express p30; however, when deprived of axonal contact by nerve transection, the Schwann cells of the distal nerve stained intensely for p30. In addition, when Schwann cells and dorsal root ganglion neurons were grown in coculture, Schwann cells that were associated with neurites were not as intensely stained by anti-p30 as Schwann cells that were not in contact with neurons. The pattern of p30 expression during development and regeneration, and its apparent regulation by cell-cell contact suggests that p30 plays a role in the interaction between neurons and Schwann cells during morphogenesis of peripheral nerves.

Isolation and characterization of folded fragments released by Staphylococcal aureus proteinase from the non-histone chromosomal protein HMG-1

HMG-1 was isolated from newborn calf thymus without exposure to overt denaturing conditions. The purified protein was digested under several solvent conditions with the proteinase (endoproteinase GluC) from Staphylococcus aureus strain V8. We found that the preferred site of attack by the enzyme on HMG-1 was influenced markedly by ionic strength and temperature. In 0.35 M NaCl/50 mM Tris-phosphate (pH 7.8) at 37 degrees C, cleavage near the junction between the A and B domains is predominant, as previously reported by Carballo et al. (EMBO J. 2 (1983) 1759-1764). However, in 50 mM Tris-phosphate (pH 7.8) lacking NaCl and at 0 degrees C, cleavage between the B and C domains strongly predominates. Three major products of the digestions were purified and characterized. The fragment consisting of domains B and C was found by circular dichroism to contain a substantial amount of helix. This re-emphasizes the importance of avoiding overt denaturing conditions when working with members of the HMG-1 family.

Prothymosin alpha is a nuclear protein

The cellular location of the so-called 'thymic hormone' prothymosin alpha has been studied by microinjection into the cytoplasm of Xenopus oocytes, followed by separate monitoring of nuclear and cytoplasmic concentrations. It is shown that prothymosin alpha migrates to the nucleus at a rate comparable to that of histone H1. Prothymosin alpha cannot therefore be a hormone in the usual sense of the word.

Specific recognition of cruciform DNA by nuclear protein HMG1

Cruciform DNA, a non-double helix form of DNA, can be generated as an intermediate in genetic recombination as well as from palindromic sequences under the effect of supercoiling. Eukaryotic cells are equipped with a DNA-binding protein that selectively recognizes cruciform DNA. Biochemical and immunological data showed that this protein is HMG1, an evolutionarily conserved, essential, and abundant component of the nucleus. The interaction with a ubiquitous protein points to a critical role for cruciform DNA conformations.

Identification of the core-histone-binding domains of HMG1 and HMG2

High mobility group (HMG) nonhistone chromosomal proteins are a group of abundant, conservative and highly charged nuclear proteins whose physiological role in chromatin is still unknown. To gain insight into the interactions of HMG1 and HMG2 with the fundamental components of chromatin we have introduced the methodology of photochemical crosslinking. This technique has allowed us to study the interaction of HMG1 and HMG2 with the core histones, in the form of an H2A X H2B dimer and an (H3 X H4)2 tetramer, for an effective time of crosslinking of less than 1 ms and under very mild conditions. This is achieved by using flash photolysis. With this procedure we found that both HMG1 and HMG2 interact with H2A X H2B and also with (H3 X H4)2. In the second case, they seem to do this through histone H3. To obtain more information about the interactions, we split HMG1 and HMG2 into their peptides using staphylococcal proteinase. The peptides obtained, which reflect the domain distribution of these proteins, were then used along with the histone oligomers to elucidate their interactions by means of photochemical crosslinking. Results obtained indicate that the domain of HMG1 and HMG2 involved in the interaction with H2A X H2B histones is the highly acidic C-terminal, whereas the N-terminal is involved in the interactions with (H3 X H4)2 histones. In all cases, the interactions found appear appreciably strong. Along with other data published in the literature, these proteins appear to have at least one binding site per domain for the chromatin components.

Differential binding of chromosomal proteins HMG1 and HMG2 to superhelical DNA

The binding of chromosomal proteins HMG1 and HMG2 to various DNA structures was examined by a nitrocellulose filter binding assay using a 32P labelled supercoiled plasmid. Binding assays and competition experiments indicated that HMG2 has a higher affinity than HMG1 for supercoiled DNA. Studies at various ionic strengths and pH values reveal differences in the interaction of the two proteins with DNA. The results suggest that HMG1 and HMG2 are involved in distinguishable cellular functions.

Conformation and domain structure of the non-histone chromosomal proteins HMG 1 and 2. Domain interactions

The sequence of the 224 residues of HMG 1 suggests it consists of three domains. We have previously proposed [Cary et al. (1980 Eur. J. Biochem. 131, 367-374] that the A and B domains can fold autonomously and that there is also a small N domain. Several proteases are now found to cut at the end of the B domain (at or close to residue 184). It is shown that the A + B-domain fragment also folds and probably contains all the helix of intact HMG 1. The stability of the B domain is enhanced by the presence of the A domain. The acidic C domain undergoes a coil----helix transition on lowering the pH. Several peptides have been prepared by cleavage at tryptophan. Peptide 57--C-terminus contains complete B and C domains but does not fold. In the absence of the A domain the C domain is thus able to destabilise the B domain. It is concluded that the stability of the B domain in HMG 1 is due to interaction with the A domain and the C domain has a separate function from the other domains.

Purification of non-histone acceptor proteins for ADP-ribose from mouse testis nuclei

Acceptor proteins for poly(ADP-ribose) have been purified from mouse testis nuclei. Nuclear proteins were labelled in vitro with [14C]ribose and [3H]adenine, extracted with 5% (v/v) HClO4 and 0.25 M-HCl and separated by ion-exchange chromatography. Non-histone proteins were found to be the major acceptors in both the 5% (w/v)-HClO4-soluble and 5%-HClO4-insoluble HCl-extractable fractions. Of the two groups of non-histone proteins associated with chromatin, the LMG (low-mobility-group) proteins were preferentially ADP-ribosylated. HMG (high-mobility group) proteins were labelled to lower specific radioactivity. Six LMG proteins were purified to approx. 90% homogeneity and were identified from their mobility on polyacrylamide gels at pH 2.9 and from their amino acid composition. The average length of the poly(ADP-ribose) chain was estimated to be four to six repeating ADP-ribose units. It is suggested that ADP-ribosylation of LMG proteins, a long-neglected group of chromatin-associated proteins, is important during spermatogenesis for the production of spermatozoa with intact and competent DNA.

Production of HMG-3 by limited trypsin digestion of purified high-mobility-group nonhistone chromatin proteins

Three isolated nonhistone proteins (HMG-1, HMG-2 and HMG-E) have been purified from chicken erythrocyte chromatin without exposure to overt denaturing conditions, and subjected to limited proteolysis. When treated with trypsin, the three proteins exhibited similar patterns of degradation, as judged by SDS and acid/urea gel electrophoresis. In particular, the first product, P1 (a relatively stable intermediate in each digestion), was a protein analogous to HMG-3, a principal degradation product in preparations of calf thymus high-mobility-group proteins. At least in the case of HMG-E, the products formed by tryptic attack on P1 are the two individual DNA binding domains of HMG-E. P1 derived from HMG-E and one of the individual DNA binding domains of HMG-E were purified by chromatography on columns containing DNA-cellulose or phosphocellulose. The properties of these two portions of HMG-E are consistent with our recently postulated three-domain structure for HMG-1 and its homologs (Reeck, G.R., Isackson, P.J. and Teller, D.C. (1982) Nature 300, 76-78). Thus, P1 consists of two DNA-binding domains of approximately equal molecular weight covalently linked together. From chromatography on DNA-cellulose columns, it is clear that P1 binds to DNA more tightly than does HMG-E. The highly acidic C-terminal domain of HMG-E (which is removed by trypsin in generating P1) thus counteracts the DNA binding of the two other domains of HMG-E (at least in the protein's interaction with purified DNA).