In Vitro Acetylation of HMGB-1 and -2 Proteins by CBP: the Role of the Acidic Tail ² Evdokia Pasheva, Mihail Sarov, § Kiril Bidjekov, | Iva Ugrinova, Bettina Sarg, Herbert Lindner, and Iliya G. Pashev* ,‡ Institute of Molecular Biology, Bulgarian Academy of Sciences, 1113 Sofia, Bulgaria, and Department of Medical Chemistry and Biochemistry, UniVersity of Innsbruck, Fritz-Pregl-Strasse 3, Innsbruck, Austria ReceiVed September 8, 2003; ReVised Manuscript ReceiVed December 8, 2003 ABSTRACT: Histone acetyltransferases CBP, PCAF, and Tip60 have been tested for their ability to in vitro acetylate HMGB-1 and -2 proteins and their truncated forms lacking the C-terminal tail. It was found that these proteins were substrates for CBP only. Analyses of modified proteins by electrophoresis, amino acid sequencing, and mass spectrometry showed that full-length HMGB-1 and -2 were monoacetylated at Lys2. Removal of the C terminus resulted in (i) an increased incorporation of radiolabeled acetate within the proteins to a level close to that observed with histones H3/H4 and (ii) creation of a novel target site at Lys81. Acetylated and nonmodified HMGB-1 and -2 protein lacking the acidic tail were compared relative to their binding affinity to distorted DNA and the ability to bend linear DNA. Both proteins showed similar affinities to cisplatin-damaged DNA; the acetylated protein, however, was 3-fold more effective in inducing ligase-mediated circularization of a 111-bp DNA fragment. The alterations in the acetylation pattern of HMGB-1 and -2 upon removal of the C-terminal tail are regarded as a means by which the acidic domain modulates some properties of these proteins. High-mobility group (HMG) 1 proteins 1 and 2 (recently renamed HMGB-1 and -2, see ref 1) are the most thoroughly studied subgroup of HMG chromosomal proteins, the func- tions of which are still obscure. Their property to bend DNA and to bind preferentially to distorted DNA structures has led to the view that these non-sequence-specific DNA binding proteins act primarily as architectural elements, promoting the assembly of nucleoprotein complexes (2, 3) and facilitating nucleosome disruption and remodeling (4). Surprisingly, work over the past few years demonstrated that these proteins also have functions outside the cell, related to inflammation and tumor growth and metastasis (5). The extracellular protein is regarded as a signal that com- municates the death of the cell to its neighbors (5). The problems about the functions of HMGB-1 and -2 are further extended by the existence of postsynthetic modifications. Phosphorylation of some counterparts of these proteins from insects (6, 7) and from higher plants (8) has been found to be essential for their proper folding and DNA binding specificity. Another modification of HMGB proteins is the reversible postsynthetic acetylation, demonstrated so far in vertebrate proteins only (9). Although this modification has been known for 25 years, the properties of the acetylated protein were studied very recently. We demonstrated that in vivo acetylation of HMGB-1 at lysine 2 significantly enhanced its affinity to distorted DNA structures (10). Acetylation in the presence of butyrate was generally attributed to histone acetyltransferase (HAT) activity ( 9). The discovery in the mid 1990s that this activity is an intrinsic property of some transcriptional adaptors led to the identi- fication of many nuclear HATs, some of them acting also on non-histone proteins (11-13), including HMG-14 and -17 (14, 15). Thus, our next aim was to study some known HATs for their ability to acetylate HMGB-1 and -2 in vitro, to map the target lysines, and to analyze the consequences for their DNA binding properties. Meanwhile, by testing a large panel of proteins as substrates for recombinant CBP, HMGB-1 was reported among the few that showed a positive result (16), but neither have the target lysines been mapped nor the properties of the modified protein studied. Recently, CBP was reported to acetylate the HMG-box containing architectural factor UBF both in vitro and in vivo (17). This paper presents our data on in vitro acetylation of rat HMGB-1 and -2 with CBP, the role of the C-terminal domain on the pattern of acetylation as well as some properties of the modified truncated protein. EXPERIMENTAL PROCEDURES Preparation of HMGB-1 and -2. Nuclei from Guerin ascites tumor cells were used to isolate the two proteins by a non-denaturing salt extraction procedure (18). Removal of ² This work was partially supported by Grant K1101/01 from the National Science Fund, Ministry of Education and Science, Bulgaria. * Author to whom correspondence should be addressed [telephone (359) 2 72 02 38; fax (359) 2 72 35 07; e-mail igp@obzor.bio21.bas.bg]. Bulgarian Academy of Sciences. § Present address: Max Planck Institute of Molecular Cell 108, 01307 Dresden, Germany. | Present address: Biology and Genetics, Pfotenhauerstrasse Institut fu ¨r Klinische Chemie, Klinikum Grosshadern, Marchionini str. 15, Grosshadern, Mu ¨nchen 81377, Germany. University of Innsbruck. 1 Abbreviations: CBP, CREB-binding protein; HAT, histone acetyl- transferase; HMG, high-mobility group; HMGB-1 and -2, high-mobility group box-containing proteins 1 and 2; PAGE, polyacrylamide gel electrophoresis; PCAF, p300/CBP-associated factor; 0.5×TBE, 0.045M Tris-borate, 1 mM EDTA; Tip60, Tat-interactive protein, 60 kDa; trHMGB-1 and -2, truncated HMGB-1 and -2 proteins, lacking the C-terminal tail. 2935 Biochemistry 2004, 43, 2935-2940 10.1021/bi035615y CCC: $27.50 © 2004 American Chemical Society Published on Web 02/18/2004