Ca(2+)-calmodulin-dependent phosphorylation of arginine in histone 3 by a nuclear kinase from mouse leukemia cells.

A Ca(2+)-calmodulin dependent histone 3 kinase was partially purified from a low salt (150 mM NaCl) nuclear extract of mouse leukemia cells by calmodulin-Sepharose affinity chromatography. In vitro, the kinase activity transferred gamma-phosphate from ATP to histone 3 to form an acid-labile and alkaline-stable linkage. Under the assay conditions 1.8 mol of phosphate are incorporated per mol of histone 3. Upon modification of arginine residues with phenylglyoxal prior to phosphorylation, a considerable decrease in the amount of phosphate transferred to histone 3 was observed. Amino acid analysis revealed that H3 was phosphorylated on arginine residues. To identify the phosphorylated peptide(s), histone 3 was cleaved with cyanogen bromide prior to phosphorylation. The phosphorylated mixture was then separated by gel filtration high-performance liquid chromatography under denaturing conditions. Fragments I (N-terminal 10.3-kDa peptide) and III (C-terminal 1.7-kDa peptide) were both phosphorylated. Amino acid sequencing further revealed that the molar yields of 3 of the 4 arginines present in the phosphorylated cyanogen bromide fragment III were reduced by a factor of about 10 compared with the corresponding arginines from the unphosphorylated fragment. In the case of fragment I, 25 cycles of Edman degradation revealed that the recovery of only arginine 2 was reduced by a factor of 20. The putative phosphorylation sites are arginines 2, 128, 129, and 131. The sequence information offered an indirect evidence that these arginines were the sites of phosphorylation. The kinase described in this report represents a first member of a potentially important new class of kinases which are Ca(2+)-calmodulin dependent and which phosphorylate arginine.

A Cas*-calmodulin dependent histone 3 kinase was partially purified from a low salt (160 mM NaCl) nuclear extract of mouse leukemia cells by calmodulin-Sepharose affinity chromatography. In vitro, the kinase activity transferred y-phosphate from ATP to histone 3 to form an acid-labile and alkaline-stable linkage. Under the assay conditions 1.8 mol of phosphate are incorporated per mol of histone 3. Upon modification of arginine residues with phenylglyoxal prior to phosphorylation, a considerable decrease in the amount of phosphate transferred to histone 3 was observed. Amino acid analysis revealed that H3 was phosphorylated on arginine residues. To identify the phosphorylated peptide(s), histone 3 was cleaved with cyanogen bromide prior to phosphorylation. The phosphorylated mixture was then separated by gel filtration high-performance liquid chromatography under denaturing conditions. Fragments I (N-terminal 10.3-kDa peptide) and I11 (C-terminal 1.7-kDa peptide) were both phosphorylated. Amino acid sequencing further revealed that the molar yields of 3 of the 4 arginines present in the phosphorylated cyanogen bromide fragment I11 were reduced by a factor of about 10 compared with the corresponding arginines from the unphosphorylated fragment. In the case of fragment I, 26 cycles of Edman degradation revealed that the recovery of only arginine 2 was reduced by a factor of 20. The putative phosphorylation sites are arginines 2, 128, 129, and 131. The sequence information offered an indirect evidence that these arginines were the sites of phosphorylation. The kinase described in this report represents a first member of a potentially important new class of kinases which are Ca2'-calmodulin dependent and which phosphorylate arginine.
Protein phosphorylation plays an important role in a number of cellular activities, including initiation of mitosis (reviewed by Nurse, 1990) and regulation of transcription (reviewed by Bohmann, 1990). Among the proteins phosphorylated during these processes are transcription factors (Jackson et al., 1990;Nygard et al., 1991), topoisomerase I1 (Kroll and Rowe, 1991), and histones 1 and 3 (Gurley et al., 1978;Matthews and Bradbury, 1978). Phosphorylation of H3 is a mitotic event (Gurley et al., 1978;Marks et aL, 1973;Gurley et al., 1974) and has been suggested to be closely associated with chromatin condensation (Shibata et al., 1990). Stimula-*This work was supported by the Potts Estate. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement'' in accordance with 18 U.S.C Section 1834 solely to indicate this fact.
II To whom correspondence should be addressed. tion of quiescent cells by growth factors, phorbol esters, okadaic acid, and protein synthesis inhibitors also leads to rapidphosphorylation of H3 (Mahadevan et aL, 1991). Within minutes of stimulation, the proto-oncogenes c-fos and c-jun were transcriptionally activated and structural changes in chromatin was observed (Mahadevan et al., 1991). Mahadevan et al. (1991) have suggested that the rapid phosphorylation of H3 modulates nucleosomal characteristics and potentially regulates early gene expression at the structural level. Our present knowledge of the kinases that phosphorylate H3 is limited to an H3 kinase identified as a component of HeLa nucleosomes (Simpson, 1978), a 38-kDa chromatinbound H3 kinase from calf thymus Chalkley, 1978, 1980), and the catalytic subunit of CAMP-dependent kinase used by Shibata et al. (1990). Early reports indicated that the phosphorylation of H3 is regulated by physiological concentrations of Ca2+ in butyrate-treated HeLa cells (Whitlock et al. , 1983. Recent reports demonstrate that the level of intracellular Ca2+ is related to such events as nuclear envelop breakdown and chromatin condensation (Keith et al., 1985;Kao et al., 1990). In addition, calmodulin (CaM)' (primary mediator of Ca2+-dependent signaling in eucaryotic nonmuscle cells) together with Caz+ was shown to play an important role in regulating cell cycle-related events, including G2/ M transition (Lu et al., 1993 and for review see Means et al., 1991). In a previous report from our laboratory, a Ca2+-CaMactivated H3 kinase was identified as a component of calf thymus chromatin (Wakim et al., 1990). In this report a similar and probably identical kinase complex was partially purified from mouse leukemia nuclear extracts and used as the source of H3 kinase in the identification of the phosphorylated amino acids and their sites in H3. The data presented indicate that H3 is phosphorylated in a Ca2+-CaMdependent manner on arginine residues at potentially four different sites, three of which are in the C-terminal tail of the protein, Phosphorylation of basic amino acids including histidine, lysine, and arginine lead to the formation of acid labile phosphoramidate linkages as summarized by the following reaction.
Protein-NH + ATP c* protein-N-PO, + ADP (Reaction 1) The formation of phosphoramidate linkages has been reviewed by Matthews and Huebner (1984) and Huebner and Matthews (1985). The existence of N-linked phosphate in proteins in vivo has been reported in a variety of cell types and subcellular compartments (Huang et al., 1991). In vitro, histone 4 has been shown to be a good substrate for a histidine The abbreviations used are: CaM, calmodulin; TEA, triethylamine; TEAA, triethylamine acetate; HPLC, high-performance liquid chromatography. kinase identified in Physarum polycephalum nuclear extracts and purified from Saccharomyces cereuisiae whole cell extracts (Huang et al., 1991). Phospholysine was found in H1 and phosphoarginine in myelin basic protein (Chen et al., 1977;Steiner et al., 1980). An arginine-specific protein kinase has been isolated from the pellet of a 450 mM NaCl extract of rat liver chromatin and phosphorylated an 11-kDa protein (Levy-Favatier et al., 1987).
Kinases that generate P-N linkages have received much less attention that those generating P-0 linkages primarily because of the technical difficulties related to the identification of the phosphorylated proteins and the phosphorylated amino acids. For example, acid precipitation of phosphoproteins with trichloroacetic acid, fixation of polyacrylamide gels with acetic acid, or HPLC chromatography in the presence of trifluoroacetic acid or any other method utilizing acidic conditions, commonly used in the case of acid stable phosphoproteins, must be excluded because they lead to a rapid reversal of the phosphorylation reaction (Hultquist, 1968;Wei and Matthews, 1990a, 199Ob). In this report we demonstrate the existence of a kinase in the nucleus of mouse leukemia cells which phosphorylated arginine residues primarily in the Cterminal tail of H3.

EXPERIMENTAL PROCEDURES
Materials-Calf thymus histones were prepared according to Bonner et d. (1968). Calmodulin was purified by the method of Gopalakrishna and Anderson (1982). CaM-Sepharose was prepared from purified CaM and cyanogen bromide-activated Sepharose 4B (Pharmacia LKB Biotechnology Inc.) according to the manufacturer's instructions. Phosphoarginine was purchased from Sigma. [y3'P] ATP was purchased from Amersham Corp. Purification of H3-H3 was purified in a single step by reversed phase HPLC using a Vydac C-18 column and an acetonitrile gradient in the presence of 0.1% trifluoroacetic acid. Calf thymus histones (200 pg) were dissolved in 100 pl of 6 M guanidine hydrochloride containing 10% #?-mercaptoethanol, heated at 55 "C for 20 min, and applied to the column. The column was eluted at 1 ml/min with an increasing acetonitrile gradient.
Mouse Leukemia Cell Culture-Murine L1210 cells were grown in 1630 media supplemented with 10% heat-inactivated bovine calf serum (HyClone, Inc.) and maintained in logarithmic state by serial passes in the same media. Cells were harvested by centrifugation at 1,000 X g and stored at -70 "C.
Isolation of the Ca'+-CaM-dependent Histone 3 Kinase Activity-The mouse leukemia L1210 cell pellet was thawed, suspended in 50 volumes of homogenization buffer (10 mM potassium phosphate, 320 mM sucrose, 1 mM MgClz, and 10 mM NaHS03, pH 6.9) and homogenized in a glass homogenizer at 4 "C. All steps that followed were performed at 4 "C. Nuclear material was precipitated by centrifugation at 1,000 X g for 5 min. The nuclear pellet was washed twice in the same buffer. It was then suspended in 30 ml of extraction buffer(50 mM Tris-HC1, 150 mM NaCl, 1 mM M&12, 1 mM CaC12, 1 mM dithiothreitol, pH 7.5) and extracted while stirring for 30 min. The mixture was then centrifuged at 10,000 X g for 30 min. The supernatant (nuclear extract) was applied to a CaM-Sepharose affinity column (15 X 1 cm). The column was washed with extraction buffer containing 500 mM NaCl instead of 150 mM and 0.2 mM CaCIZ instead of 1 mM until the absorbance at 280 was less than 0.005. The column was then eluted with wash buffer except that Ca2C1 was replaced with 2 mM EGTA. The protein peak was collected and used as the source of the Cal+-CaM-dependent kinase activity.
Histone 3 Kinase Assay-Histone 3 phosphorylation was performed as described previously (Wakim et al., 1990) with little modification. The assay was carried out in a volume of 100 pl of 100 mM Tris, 1 mM magnesium acetate, 0.5 mM B-mercaptoethanol, 2 mM CaC12, or 2 mM EGTA (pH 6.9) in the presence or absence of CaM. The assay contained either 100 pg of a histone mixture or purified H3. The assay was started by adding ATP to a final concentration of 0.1 mM and 10' dpm of [T-~'P] ATP. The reaction was carried out for 30 rnin at 30 "C and stopped either by adding 20 p1 of SDS buffer containing 10% 8-mercaptoethanol, when the phosphorylated proteins were to be separated by SDS-polyacrylamide gel electrophoresis or by adding an equal volume of 6 M guanidinium hydrochloride in 10 mM potassium phosphate (pH 6.9) when re-purification of the phosphorylated protein/peptides were to be performed. The SDS-polyacrylamide gels were stained and destained in the presence of 50% methanol and 0.1% Coomassie Blue in the absence of acetic acid.
Cyanogen Bromide Cleavage and Acid/Base Treatment of Histone 3-Cyanogen bromide cleavage was performed by incubating the protein with 10 pg of cyanogen bromide in 72%, v/v, formic acid for 24 h at room temperature. Phosphorylated H3 was either incubated in 0.1 ml of 3 N HCl or 3 N NaOH for 2 h at 37 "C.
Purification of Phosphorylated H3 and Its Cyanogen Bromide Fragments-Purification was performed on a tandem pair of Bio-Si1 SEC 250 and 125 HPLC gel filtration HPLC columns in the presence of 6 M guanidine hydrochloride and 10 mM potassium phosphate (pH 6.9) at 1 ml/min. Phosphorylated CB 111 was desalted by reversed phase HPLC at pH 6.9 in the presence of 40 mM triethylamine acetate (TEAA) (pH 6.9) and an acetonitrile gradient using a Vydac C-4 column.
Base Hydrolysis and Identification of Phosphoarginine-Samples were hydrolyzed in 100 p1 of 3 M KOH under argon for 2 h at 125 "C.
Samples were placed on ice, and an equal volume of 3 M perchloric acid was added in order to neutralize the KOH. The salt was precipitated by centrifugation at 10,000 X g for 30 min. The supernatant containing the hydrolyzed protein was dried down and derivatized using the phenylisothiocyanate method described by Bidlingmeyer et al. (1984). Phosphohistidine, -lysine, and -arginine standards were separated on a PicoTag column by isocratic elution in the presence of sodium acetate containing Triethylamine (TEA) (pH 5.0) at 50 'C. Phosphoarginine from phosphorylated H3 was identified from the retention time of the radioactivity. Amino Acid Sequencing-Amino acid sequencing was performed using an Applied Biosystems 477A sequencer with a model 120A online HPLC system.

RESULTS
H3 Is the Preferred Substrate, among the Five Histones, for the Ca'+-CaM-dependent Arginine Kinuse-The kinase was prepared as described under "Experimental Procedures." Briefly, nuclear material of mouse leukemia cells were extracted with 150 mM NaCl and centrifuged at 10,000 X g. The supernatant (nuclear extract) was applied to a CaM-Sepharose affinity column in the presence of Ca2+ and eluted with a buffer containing EGTA instead of Ca2+. The EGTA eluate contained Ca2+-CaM-binding proteins or protein complex. Sucrose density gradient centrifugation showed that the EGTA eluate consisted of a complex but did not add to the purity of the kinase (data not shown). Consequently, we chose to use the EGTA eluate as the source of the phosphorylating activity. Equal amounts (1 pg of protein) of enzyme were used per assay. The nuclear extract and the EGTA eluate were assayed for possible endogenous phosphorylating activities in the absence of histone substrates. In the case of the nuclear extract, three proteins with apparent molecular masses of 105, 65, and 41 kDa were phosphorylated in a Ca2+-CaM independent manner (Fig. 1, lanes 1 and 2). The EGTA eluate contained an endogenous protein with an apparent molecular weight of 55 kDa that was phosphorylated in a Ca2+-CaMdependent manner (Fig. 1, lanes 3 and 4 ) . This either represented autophosphorylation of the kinase(s) of interest or one of its subunits or simply a protein that co-migrated with the EGTA eluate and which happened to be phosphorylated by that kinase in a Ca2+-CaM-dependent manner. It also seems possible that the 55-kDa protein could be the CaM-kinase I1 (a subunit). Phosphorylation was also performed in the presence of a mixture of all five calf thymus histones. H1 and H2B were the preferred substrates for the nuclear extract, and their phosphorylation was not activated in the presence of Ca2+-CaM (Fig. 1, lanes 5 and 6). The EGTA eluate, on the other hand, contained a kinase that specifically phosphorylated H3 in a Ca2+-CaM-dependent manner (Fig. 1, lanes  7 and 8). Phosphorylation of Purified H3 and Its Cyanogen Bromide Peptides-Histone 3 was purified by reversed phase HPLC (Fig. 2). In fact, by this method all five histones can be purified in a single step. The purity of the histone fractions was confirmed by SDS-polyacrylamide gel electrophoresis (Fig. 2,  inset). Purified H3 was phosphorylated by the EGTA eluate and repurified by gel filtration HPLC (Fig. 3A). Phosphorylated H3 was then desalted by ultrafiltration using a Centricon 3 microconcentrator. The maximum stoichiometry of phosphorylated H3 was 1.8 mol of phosphate/mol of H3. The purified phospho-H3 was then divided into two fractions and dried. One of the fraction was treated with base and the other with acid as described under "Experimental Procedures.'' The samples were then neutralized and H3 repurified as above. It was observed that acid treatment led to complete loss of the radioactivity from co-eluting with the protein peak (Fig. 3B). Treatment with base, on the other hand, led to no loss of radioactivity (Fig. 3C). This indicated that the y-phosphate bound to H3 was acid-labile and base-stable which in turn indicated that one or more basic amino acid, including histidine, lysine, and/or arginine was the likely phosphorylated amino acid. Upon modification of arginine residues with phenylglyoxal prior to phosphorylation, a considerable decrease in the phosphorylated H3 was observed (Fig. 4). Lane 1 represents unmodified H3, lanes 2 and 3 represent H3 that was modified in the presence of 0.5 mM and 5 mM phenylglyoxal, respectively, prior to phosphorylation. The data presented in Fig. 4 is consistent with the other observations shown in this paper that arginine is the likely phosphorylated amino acid.
Identification of Phosphoarginine in Phosphorylated H3-Phosphorylated H3 was base-hydrolyzed and the resulting amino acids derivatized and analyzed as described under "Experimental Procedures." As can be observed from Fig. 5A, phosphoarginine can be separated from phospholysine and phosphohistidine, all of which were elute isocratically under the analysis condition. Fig. 5B represent the isocratic part of the amino acid analysis of phosphorylated H3. The 32P counts eluted at identically the same retention time as phosphoarginine. This strongly indicated that arginine is the likely phosphorylated amino acid.
Cyanogen bromide cleavage of H3 was performed prior to phosphorylation of the peptide mixture. CB I, representing the N-terminal 10.3-kDa peptide, and CB 111, representing the C-terminal peptide were both phosphorylated (Fig. 6). CB I11 contained about three times the amount of the bound 32P compared to CB I. CB I11 lacked any serine, threonine, or tyrosine. Thus these amino acids could be excluded from being the phosphorylated residues at least in this peptide. For the purpose of microsequencing, fragment I was desalted by repeated washes through a Centricon 3 and fragment I11 by reversed phase HPLC at pH 6.9.
Amino Acid Seqwncing of 25 Edman Degradation Cycles of CB I and of All CB 111-Amino acid sequencing revealed that in the case of fragment I, among the 25 amino acids identified, the recovery of only arginine 2 dropped by a factor of about 20 (Fig. 7A). In the case of CB I11 the molar recovery of 3 of the 4 arginines present in that fragment dropped by a factor of about 10 in the case of the phosphorylated compared with the unphosphorylated fragment (Fig. 7B). These results were taken as further evidence that arginine is indeed the phosphorylated amino acid and that 3 of the 4 arginine residues present among the C-terminal 15 amino acids as well as arginine 2 from the N terminus of H3 were the primary sites of phosphorylation. The putative sites of phosphorylation based on the identification of the phosphorylation sites in the cyanogen bromide fragments are thus arginines 2, 128, 129, and 131.

DISCUSSION
In this report we demonstrate that nuclear extracts from mouse leukemia cells contain a Ca2+-CaM-dependent kinase that is capable of phosphorylating histone 3 (H3) on arginine residues. Based on the amino acid sequence information obtained using a cyanogen bromide digest of H3 prepared prior to phosphorylation, we suggest that 3 out of the 4 arginine residues present in the 15 amino acid C terminus of H3 and arginine 2 at its N terminus are the likely phosphorylation sites. This is the first such activity to be identified which is Ca2+-CaM-dependent and which phosphorylates the basic amino acid arginine. A similar kinase was identified from rat cm). Phosphorylated H3 was purified from the assay mixture by gel filtration HPLC in the presence of 6 M guanidine hydrochloride and 10 mM potassium phosphate at pH 6.9 at 1 ml/min ( A ) . Purified phospho-H3 was either treated with acid or base as described under "Experimental Procedures" and repurified by gel filtration as above. E and C, represent the repurification of phospho-H3 after acid treatment ( E ) and base treatment ( C ) . As can be observed from B, acid treatment led to loss of the radioactivity from co-eluting with the protein peak. Base treatment radioactivity (C). t .": --0 1 " _" " " " + " " " " " " " " " " " " "". " 4. Autoradiogram of phenylglyoxal-treated phospho-H3. H3 was treated with phenylglyoxal prior to phosphorylation in the presence of CaM followed by SDS-polyacrylamide gel electrophoresis and autoradiography. Equal amounts of protein were loaded in each lane. Lane I represents the phosphorylation of untreated H3. Lunes 2 and 3 represent the phosphorylation of H3 that have been treated with 0.5 and 5 mM phenylglyoxal, respectively, prior to phosphorylation. liver chromatin which also phosphorylated arginine (Levy -Favatier et al., 1987). The kinase described by Levy-Favatier et al. (1987) consists of a single subunit of an apparent molecular mass of 34 kDa capable of autophosphorylation and phosphorylates a single chromosomal protein of 11 kDa. The kinase presented in this paper phosphorylated H3 preferentially and did not show any phosphorylated band a t 34 kDa. Also the rat liver kinase was isolated from the pellet of a 0.45 M NaCl chromatin extraction. The kinase presented in this paper is isolated from a low salt (150 mM NaCI) extract of chromatin. We believe that the two kinases are different. The maximum stoichiometry of phosphorylation of H3 was found to be 1.8 moles phosphate per mole H3. This was less than the expected 4 moles which could be either due to unequal phosphorylation of the four arginines in question or due to partial loss of the phosphate during the purification of the phosphorylated H3. The fact that more than one mole phosphate per mole H3 was observed, indicates multiplicity of phosphorylation sites which is in agreement with the rest of our data.
Phosphorylation of H3 is believed to be involved with chromatin condensation (Shibata et al., 1990). The mitoticspecific sites of H3 phosphorylation according to these authors were serines 10 and 28 both of which are present in the N terminus of H3. In their experiments they used the catalytic subunit of CAMP-dependent kinase to phosphorylate H3 and precipitated the phosphorylated protein with trichloroacetic acid, thus eliminating any potential phosphoramidate linkages. Our finding that H3 was phosphorylated on arginine residues in a Ca"-CaM-dependent manner neither contradicts nor supports their observations, since a different enzyme was used, but demonstrate that H3 was phosphorylated on arginine residues primarily present in the C-terminal tail of the protein. Based on earlier reports indicating that the Cterminal short tail of H3 extended from the globular central domain of the protein (Bradbury, 1983), any change in the charge of that region of the protein will be suspected of modifying the binding of H3 to DNA. Certainly, phosphorylation of arginine will change the overall charge of the Cterminal tail and is likely to alter the association of H3 with DNA during nucleosome assembly.
Phosphorylation of H3 is known to be stimulated by Ca2' Over 95% of the radioactivity was eluted within the first 10 min of isocratic separation of the derivatized amino acids. A, His(P), Lys(P) and Arg(P) standards. Tyr(P) was also included in the standard prior to base hydrolysis but showed no peak in the first 10 min of separation. B, elution profile of amino acids from phosphorylated H3 (solid line) and the radioactivity (dotted line). at physiological concentrations (Whitlock et aZ., 1980(Whitlock et aZ., , 1983. Calmodulin, which is the primary mediator of Ca2+-dependent signaling in eucaryotic none muscle cells, has been shown to regulate different stages of the cell cycle, including G2/M transition (Lu et al., 1993). The existence of a Ca2+-CaMdependent nuclear kinase which specifically phosphorylated H3 is no surprise. Such an activity was first observed to be a component of a complex purified from calf thymus nuclear 20 ji 100 B % = F 20 1  1~1  1  1  1~1~1  1  (  I  ,  ,  material (Wakim et al., 1990). The activity described in the present report was extracted from nuclear material of mouse leukemia cells and is believed to be identical to that from calf thymus but is over 100 times more abundant than in thymus tissue. It was a surprise to find out that the phosphate transferred to H3 is acid-labile and base-stable and explained the difficulties that we initially encountered in purifying the phosphorylated peptide by standard reversed phase HPLC in the presence of trifluoroacetic acid. Modification of arginine residues using phenylglyoxal prior to phosphorylation led to a considerable decrease in the phosphorylated H3. These results were consistent with the other data presented in this paper that arginine is the likely amino acid to be phosphorylated. It is also possible, however, that the modification with phenylglyoxal altered the structure of H3 thus making it inaccessible to phosphorylation. The strongest evidence that arginine was the phosphorylated amino acid in H3 came from the data obtained by amino acid identification of the phosphorylated residues where the radioactivity and phosphoarginine eluted at identical retention times. The gel filtration data revealed that CB I and CB I11 were phosphorylated and that the 32P counts that migrated with CB I11 were about three times than those migrating with CB I. This indicated either that CB I11 is about three times more phosphorylated than CB I or that CB I11 is phosphorylated at a three time faster rate that CB I under the assay conditions. Amino acid sequencing of phosphorylated CB I11 revealed a considerable decrease in the molar recoveries of 3 of the 4 arginines compared with their recoveries in the case of the unphosphorylated fragment. This either means that phosphorylated arginines elute at a position other than arginine, which we were not able to identify, or bind to the polybrene filter in a manner similar to the acid stable phosphoamino acids, including phosphoserine and phosphothreonine. Upon applying equal amounts of either arginine or phosphoarginine to the sequencer, it was observed that phosphoarginine eluted at exactly the same retention time as arginine. However, the recovery of phosphoarginine was about 20 times less than arginine. The amino acid sequence information offers indirect evidence about the sites of H3 phosphorylation. Taken together, the data presented in this paper strongly indicated that the C-terminal tail of H3 was phosphorylated on arginine residues, thus altering the overall charge of that region and potentially regulating the binding of H3 to DNA during nucleosome assembly/disassembly.