Antimicrobial activity of histones from hemocytes of the Pacific white shrimp Se ´ verine A. Patat 1 , Ryan B. Carnegie 1, *, Celia Kingsbury 1, , Paul S. Gross 2,3 , Robert Chapman 3,4 and Kevin L. Schey 1,3 1 Department of Cell and Molecular Pharmacology, 2 Department of Biochemistry and 3 Marine Biomedicine & Environmental Sciences Center, Medical University of South Carolina, Charleston, SC, USA; 4 Hollings Marine Laboratory, Marine Resources Institute, SCDNR, Charleston, SC, USA The role of vertebrate histone proteins or histone derived peptides as innate immune effectors h as only recently been appreciated. In this study, high levels of core histone proteins H2A, H2B, H3 and H4 were found in hemocytes from the Pacific white shrimp, Litopenaeus vannamei. The proteins were identified by in-gel digestion, mass spectrometry ana- lysis, and homology searching. The L. vannamei histone proteins were found to be highly h omologous to histones o f other species. Based on this homology, histone H2A was cloned a nd its N-terminus was found to resemble the known antimicrobial histon e p eptides buforin I, parasin, and hip- posin. Consequently, a 38 amino acid synthetic peptide identical to the N-terminus of shrimp H2A was synthesized and assayed, along with endogenous histones H2A, H2B, and H4, for growth inhibition against Micrococcus luteus. Histone H2A, p urified to homogeneity, completely inhibited growth of the Gram-positive bacterium at 4.5 l M while a mixture of histones H2B and H4 was active at 3 l M .In addition, a f raction containing a f ragment of h istone H1 was also found to be active. The synthetic peptide similar to buforin was active at s ubmicromolar concentrations. These data indicate, for the first time, t hat shrimp hemocyte histone proteins possess antimicrobial activity and represent a d ef- ense mechanism previously unreported in an invertebrate. Histones may be a component of innate immunity more widely conserved, and of earlier origin, than previously thought. Keywords: a ntimicrobial peptide; histone; invertebrate; mass spectrome try; shrimp. Marine organisms, such as the shrimp Litopenaeus vanna- mei, have developed efficient methods to survive and prosper in a microbe-rich oceanic environment. Hemocytes, the principal e ffector cells in shrimp defenses, a re involved in nonself recognition, phagocytosis, melanization, cytotoxic- ity, and cell–cell communication [1]. There are three classes of hemocytes: the hyaline, semigranular, and granular cells. The hyaline cells are associated with coagulation a nd phagocytosis while the semigranular and granular cells are involved in phagocytosis, release of p roteins in the prophe- noloxidase cascade, and release o f antimicrobial peptides. The first stage of the immune response is recognition of nonself. Microbial cell wall components [e.g. lipopolysaccharides (LPS), b-1,3-glucans, or peptidogly- cans] are recognized by specific proteins in the hemolymph, namely LPS-binding proteins, b-glucan b inding proteins [2], or lectins [3]. b-1,3-Glucan binding proteins have been identified in Penaeus californiensis and in L. vannamei hemolymph [4,5]. These molecules probably interact with the hemocytes to initiate defense responses. The prophe- noloxidase cascade leading to melanization has been one of the most studied immune reactions in crustaceans [6]. Melanin and its intermediates have been shown to be toxic to microbes [7]. Antimicrobial peptides represent an essential alternative first line of defense. These molecules, universally distributed in metazoans, are usually amphipathic, carry a net positive charge, a nd can form a-helical or b-sheet structures in membrane-like environments [8]. Antimicrobial peptides can be constitutively or inducibly expressed and have been found in different cell types including epithelial cells and phagocytes. Some antimicrobial peptides are also derived from larger proteins by proteolysis. For example, bovine lactoferricin is derived from lactoferrin, an iron bin ding protein, by pepsin cleavage of its N-terminus [9]. In crustaceans, antimicrobial peptides have been described in the c rab Carcinus maenas [10], in the crayfish Pacifastacus Correspondence to K. L. Schey, Medical University of South Carolina, Department of Cell and Molecular Ph armacology, 173 Ashley Avenue, PO Box 250505, Charleston, SC 29425, USA . Fax: +1 843 792 2475, Tel.: +1 843 792 2471, E-mail: scheykl@musc.edu Abbreviations: LPS, lipopolysaccharide; NET, neutrophil extracellular trap; MAS, modified alsever s olution; TFA, tr ifluoroacetic acid; MIC, minimum inhibitory concentration. *Present add ress: Virginia Institute of Marine Science, PO Box 1346, Gloucester Pt., VA 23062, USA. Present add ress: Pharmacopeia, Inc., PO Box 5350, Princeton, NJ 08543-5350, USA. Note: The protein sequence data reported in t his paper will appear in the SwissProt and TrEMBL knowledgebase und er the accessio n numbers P83841, P83863, P83864 and P83865, and the nucleotide sequence data has been submitted to GenBank unde r the accession numbers AY576482 and AY576483. (Received 20 August 2004, r evised 29 September 2004, accepted 21 October 2004) Eur. J. Biochem. 271, 4825–4833 (2004) Ó FEBS 2004 doi:10.1111/j.1432-1033.2004.04448.x leniusculus [11], and in several penaeid shrimp. I n penaeid shrimp, the most studied family of antimicrobial peptides is the penaeidins. These peptides range from 5.5 to 6.6 kDa and are mostly active against Gram-positive bacteria [12,13]. They are stored in hemocyte granules and released at the site of infection by lysis of the cells [14,15]. Two other antimicrobial peptides have been described in the shrimp: crustin and hemocyanin-derived peptides. Crustin, an 11.5 kDa antimicrobial peptide a lso found in the shore crab Carcinus maenas [16,17], has not been fully character- ized. Finally, C-terminal hemocyanin fragments are active against fungi but the mechanism by which hemocyanin is cleaved and activated is still unclear [18]. Histone proteins or derived fragments have antimicrobial activity in vertebrates r anging from fish to humans. Histone antimicrobial activity was first demonstrated in 1958 for histones A and B purified from calf thymus, which exhibited antibacterial activity against various Gram-positive and Gram-negative bacteria [19]. It was not until the 1990s that other groups described such activity. In several fish, antimicrobial histone proteins have been detected in skin mucus or liver tissue: H2B-like proteins in catfish skin [20], H2A in trout skin [21], and H1 in the atlantic salmon liver [22]. In humans, histone H1 and its fragments derived from epithelial cells in the gastrointestinal tract were active against Salmonella typhimurium [23], and histones H2A and H2B expressed on the surface of and secreted from amnion epithelial cells (placenta) contributed to the antimicrobial activity of amniotic fluid [24]. A recent report identified histone containing NETs (n eutrophil extracellular t raps) in human neutrophils as a novel antimicrobial mechanism [25]. Three peptides derived from the N-terminus domain of histone H2A: buforin I (39 amino acids), parasin I (21 amino acids), and hipposin (51 amino acids ) from the Asian toad stomach and catfish and halibut skin mucus, respect- ively, have been described as active against bacteria [26–28]. Active antimicrobial fragments of histone H1 have been observed in the rainb ow trout, On corh ynchus my kiss [29], a s well as in stimulated human granulocytes [30] and in the serum and mucus of LPS-challenged Coho salmon, Onco- rhyncus kisutch [31]. Here we present the results of a proteomic investigation of hemocyte antimicrobial peptides of the P acific white shrimp, L. vannamei, and the r ole of histones as potentially important components of their immune system. Materials and methods Animal handling L. vannamei individuals were obtained from Waddell Mariculture Center, Bluffton, SC, USA. They were trans- ported to the laboratory in oxygen-saturated water and bled within 6 h of collection. Sample collection and preparation of cells Hemolymph was taken from the ven tral sinus of the a nimal under a n equal volume of modified alsever solution (MAS: 27 m M sodium citrate, 336 m M sodium chloride, 115 m M glucose, 9 m M EDTA, pH 7; according to Rodriguez et al. [32]) with 3 mL syringes and 25 gauge needles. Hemolymph from 3 to 4 animals was pooled in 1.5 mL Eppendorf tubes and immediately centrifuged at 800 g for 15 min (4 °C) to separate the hemocytes from the plasma. After removal of the plasma from the cell p ellets, 600 lLof MAS was added to the eppendorf tubes, which were then vortexed gently to wash the cells. Cells were centrifuged at 800 g for 15 min (4 °C) and t he buffer was removed. The washing a nd centrifugation steps w ere r epeated with 200 lL of MAS. After r emoval of the buffer, 100 lLofwaterwere added t o t he cell pellets, followed by ultrasonication for 30 s (Branson SONIFIER 450, power 2, duty cycle 20; Branson, Danbury, CT, USA) to lyse the cells. After lysis, the lysates were stored at )20 °C until needed. The protein concentrations of the cells were estimated by the Bradford method (Bio-Rad, H ercules, CA, USA) with commercial BSA standards (2 mgÆmL )1 ,Pierce,Rockford, IL, USA). 1D gel electrophoresis and in-gel digestion Precast 10–20% linear gradient Tris/HCl Criterion gels (Bio-Rad) were used. Approximately 10 lgofhemocyte lysate proteins were mixed with r educing sample buffer [2· sample buffer (Invitrogen, Carlsbad, CA, USA)/water/ 2-mercaptoethanol; 50 : 50 : 2.5; v/v/v], boiled in a water bath for 10 min, and run on the gel with running buffer [24 m M Tris base, 192 m M glycine, 0.1% (w/v) SDS] for 60 min at 200 V. The gel was stained with copper. Briefly, the gel was washed for 5 min with washing buffer [200 m M Tris, pH 8.8, 0 .1% ( w/v) SDS] and staine d with 0.3 M copper chloride for 20 min. The gel was stored in water prior to image analysis and in-gel d igestion. Bands of interest were cut out of the g el and sliced into 1 m m p ieces. The copper stained pieces were d estained with 500 lL of copper destain solution for 20 min. After destaining, the gel pieces were washed three times for 20 min w ith 100 m M ammonium bicarbonate, once f or 15 min with acetonitrile/100 m M ammonium bicarbonate (50 : 50, v/v), and once for 15 min with acetonitrile, before being dried in a spee d vac ( 5 min). Dried gel pieces were rehydrated with 10 lL of trypsin or chymotrypsin (100 ng in 100 m M ammonium bicarbonate, pH 7.8) and then covered with ammonium bicarbonate (100 m M , 30 lL). Digestion was carried out overnight at 37 °C. Supernatants were removed and the same volume of acetonitrile/water/formic acid (50 : 45 : 5, v/ v/v) was added to the gel pieces followed by sonication for 20 min. Supernatants were added to the previous super- natants. These steps were repeated with acetonitrile/water/ formic acid (85 : 10 : 5, v/v/v). Supernatants were then dried in a speed vac and r esuspended in acetonitrile/water (10 : 90, v/v). Zip tips (C18 Millipore, Billerica, MA, USA) were used to desalt the samples before analysis . Elution from the Zip tips was accomplished with 3.5 lL of acetonitrile/wat er/acetic acid (49 : 49 : 2, v/v/v). Protein purification and quantification After thawing, 300 lL of h emocyte lysate supernatant was subjected to HPLC by injection onto a C18 column (Alltech, C18 Prosphere, 4.6 · 250 mm, 300 A ˚ ,5lm; Alltech, Deerfield, I L, USA). T he solvent s ystem included 4826 S. A. Patat et al.(Eur. J. Biochem. 271) Ó FEBS 2004 0.1% trifluoroacetic acid (TFA) in water (solvent A) and 0.08% TFA in ace tonitrile (solvent B). Proteins were eluted from the column with a gradient of 5–30% solvent B o ver 13 min followed by 30–60% solvent B over 67 min at a flow rate of 0.7 mLÆmin )1 . Protein elution was monito red b y U V absorbance at 225 nm and peak fractions were collected. Fractions were subsequently lyophilized, reconstituted in 100 lL water, and analyzed by MALDI MS. Histones H 4 and H2B co eluted and were not further separated prior to protein assay and antimicrobial assay. Fractions containing only histone H2A were pooled. F ractions containing impure histone H2A were repurified over the same C 18 colum n and solvent system but with a shallower gradient (40–60% solvent B over 60 min). Pure histone H2A fractions from multiple runs were pooled as well as H2B/H4 fractions and the resulting samples were lyophilized, and reconstituted in 100 lL of deionized water for protein assay and liquid growth inhibition assay (see below). The concentrations of the endogenous histone protein fractions were estima ted b y t he Bradford method (Bio-Rad) with commercial BSA standards (2 mgÆmL )1 , Pierce). Proteolytic treatments After HPLC fractionation, histone proteins were digested by trypsin (Promega, Madison, WI, USA), chymotrypsin (Roche Applied Science, Indianapolis, IN, USA), or endoproteinase Glu-C (Roche A pplied Science). T ryptic or chymotryptic digestions were carried out overnight at 37 °Cin25m M ammonium bicarbonate/10% (v/v) aceto- nitrile, pH 7.8 with 200 ng of trypsin or 1 lgofchymo- trypsin. Glu-C digests were carried out overnight at room temperature in 25 m M ammonium bicarbonate/10% (v/v) acetonitrile, pH 7.8 with 150 ng of enzyme. Mass spectrometry analysis and peptide sequencing Matrix assisted laser desorption ionization mass spectro- metry (MALDI M S) was carried out on an Applied Biosystems Voyager-DE STR (Applied Biosystems, Foster City, CA, USA). The matrix a-cyano-4-hydroxycinnamic acid (10 mgÆmL )1 ) in 7 0% (v/v ) acetonit rile/0.1% ( v/v) TFA was used. Peptides and matrix solutions were mixed (1 : 3 lL) and 0.5 lL of t he mixture was placed on top of 0.5 lL of dried matrix on the sample plate and allowed to dry. Typically, 250 laser shots were averaged t o produce a mass spectrum. Nanospray tandem mass spectrometry w as carried out on a quadrupole/time of flight instrument (QSTAR, Applied Biosystems) or on an ion trap intrument (LCQ Classic, Finnigan West Palm Beach, FL, USA) using a custom built nanospray source (LCQ) or a Protana source (QSTAR). Two microliters of sample were loaded into the nanospray tip f or analysis. Precursor ions were selected from the MS survey scan. About 20 mass spectra were averaged to increase the signal to noise ratio on the LCQ. On the QSTAR, tandem mass spectra were acquired and averaged for 3 min a nd, when necessary, the enhancer mode was used to increase intensities of specific regions of the spectrum. MS/MS was also acquired on a MALDI-TOF-TOF instrument (4700 Proteomics Analyzer, Applied Biosys- tems) with matrix a-cyano-4-hydroxycinnamic acid (10 mgÆmL )1 ) in 70% (v/v) acetonitrile/0.1% (v/v) TFA in a 1 : 1 ratio sample t o m atrix. Sequences from tandem m ass spectrometry were determined manually and homologies to known proteins were searched using a BLAST search for short, nearly exact matches available online from the NCBI server (http://www.ncbi.nlm.nih.gov/blast) with the non- redundant database. Histone H2A N-terminal cDNA cloning Two degenerate forward primers (F1: 5¢-AACMGKGCM GGACTCCAG-3¢;F2:5¢-TMCGYAARGGMAACTA TG-3¢) and three degenerate reverse primers (R1: 5¢-TTCG TCRTTMCKGATGGC-3; R2: 5¢-CACCTCCTTGRG CRA-3¢;R3:5¢-CTTKGGDAGRACTGC-3¢)were designed based on a CLUSTALW alignment (http://www. ebi.ac.uk/clustalw/) of histone H2A DNA sequences from the crustaceans Tigriopus californicus (GenBank acc. no. S49144), Artemia (X14815), a nd Asellus aqu aticus (AJ238321). Several combinations of forward and reverse primers were used in a ttempts to amplify a 3¢ histone H2A gene fragment f rom L. vannamei gill cDNA library. Reac- tion volumes of 20 lL i ncluded PCR buffer, nucleotides at 1.25 m M , primers at 100 ngÆlL )1 , Taq DNA polymerase (Advantage kit, Clontech, Palo Alto, CA, USA), and 0.4 lL o f g ill library cDNA. Reactions were carried out in a RoboCycler Gradient 96 thermocycler (Stratagene, La Jolla, CA, USA). The reaction profile began with an initial denaturation for 3 min, followed by 30 cycles of 94 °Cfor 1min,48°C for 30 s and 72 °C for 30 s, and ended with a 5 m in final extension at 72 °C. DNA products were separated on a 1.2% agarose g el and stained with ethidium bromide. Bands of interest were excised and extracted u sing a Nucleospin k it (Clontech), and then cloned into Escheri- chia coli XL1 blue w ith ampicillin. Plate colonies were picked randomly and grown in 2 mL of Luria broth with 200 lg of ampicillin overnight at 37 °C. Plasmid DNA was extracted using NucleoSpin Plus Minipreps (Clontech), and inserts were sequenced at the MUSC DNA Sequencing Facility (Medical U niversity of South Carolina, Charleston, SC, USA). Based on the initial sequence, a new reverse primer (5¢-GGATGGCCAGCTGCAAGTGACGG-3¢) was designed. To determine the 5¢ histone H2 A gene sequence, the new reverse primer, a vector-specific fo rward primer, an d hemocyte cDNA library were used in a second PCR. Briefly, 50 ng plasmid DNA was used in a 20 lL reaction volume as above. The reaction profile began with an initial denaturation for 3 min, followed by 3 0 cycle s o f 94 °Cfor1min,68°C for 30 s, and 72 °Cfor30s,and ended w ith a 5 min final exten sion at 72 °C. Products were cloned and sequenced as above. DNA sequences were translated for comparison with described histone sequences using the ExPASy translate tool (http://www.expasy.org/). Peptide synthesis A peptide similar in length to buforin [26] was s ynthesized at the MUSC P eptide Synthesis Facility. The sequence was based o n the N-terminus of the shrimp H 2A DNA sequence (SGRGKGGKVKGKSKSRSSRAGLQFPVGRIHRLL RKGNY). After synthesis, the peptide, named H2A 2–39, was purifiedby HPLC.Approximately 0 .2 mg ofpeptide was Ó FEBS 2004 Antimicrobial activity of shrimp histone proteins (Eur. J. Biochem. 271) 4827 loaded onto a C18 column ( Alltech, C18 Prosphere, 4.6 · 25 mm, 300 A ˚ ,5lm). The solvent system included 0.1% (v/v) TFA in water (Solvent A) and 0.08% (v/v) TFA in acetonitrile (Solvent B). Peptide was eluted o ff the c olumn with a gradient of 5–75% B in 150 min at a flow rate of 0.7 mLÆmin )1 . T he purified synthetic peptide’s concentra- tion was measured by a mino acid analysis by C. Schwabe, MUSC Biochemistry Department, Charleston, SC, U SA. Liquid growth inhibition assays Antibacterial activity was tested against the Gram-positive bacterium Micrococcus l uteus. Stocks of bacteria were kept at )80 °C in 50% (v/v) glycerol ( 200 lL). One stock of bacteria was a dded to 5 mL of Luria–Bertani (LB) media andgrownovernightat37 °C with s haking, and was used to spread a LB agar p late. A fter overnight i ncubation at 37 °C, theplatewaskeptat4°C and served as a stock for the experiments. One colony was picked and incubated in 5 mL of LB media overnight at 37 °C with shaking. After incubation, bacteria were diluted 1 : 100 or 1 : 50 and incubated until the attenuance at 595 nm was about 0.1. The dilution medium was either LB broth or 1% (w/v) bacto- tryptone/1% (w/v) sodium chloride at pH 7.2. When an attenuance (D) of 0.1 was reached, bacteria were further diluted with poor broth [1% (w/v) bactotryptone/0.5% (w/v) sodium chloride, pH 7.2] to a D of 0.01 at 595 nm. Ninety microliters of bacteria (D ¼ 0.01) were added to each we ll of a 96 well plate with 10 lL of the appropriate control or sample. Negative controls were water or a synthetic peptide (MIP peptide: CVTGEPVELD TQAL, 10 l M ) chosen from a pool of synthetic peptides used for vision research and without known antimicrobial activity, and t he positive c ontrol was Ala-magainin (Sigma, S t Louis, MO, USA, 5 l M ). Samples were fractionated histone proteins or the N-terminal H 2A synthetic peptide (H2A 2– 39) at various concentrations (less than 10 l M ) diluted in water. The plate was then i ncubated at 37 °Covernightwith shaking and the D at 595 nm or 570 nm was read in a plate reader (M olecular Devices). Results Identification of histone proteins in the hemocytes In the course of an initial proteomic inve stigation of shrimp immune function, simple one-dimensional SDS/PAGE analysis of the hemocytes soluble proteins was carried out. The analysis (Fig. 1) revealed four intense bands in the region of 17 kDa. They were cut out of the gel and digested by trypsin or c hymotrypsin, and the resulting p eptides were sequenced by nanospray tandem mass spectrometry. The peptides were identified using a BLAST homology search of the obtained sequences. The lower band yielded peptides similar to histone H4, the upper b ands contained peptides from histones H2A and H2B, and the highest molecular mass band produced peptides homologous to histone H3 (Table 1). After fractionation of the hemocyte soluble proteins by C18 reverse phase H PLC and enzyme digestion, additional peptides were sequenced from the histone proteins (Table 1). Furthermore, intact protein molecular masses were determined by MALDI MS of the histone containing fractions. Direct MALDI MS analysis of the unfractionated hemocyte supernatant shows the molecular masses of the histone proteins: histone H4 is 11.3 kDa, H2A 13.2 kDa, H2B 13.5 kDa, and histone H3 is 15.3 kDa (Fig. 2 ). Histone proteins have a h igh number o f basic residues, which make them run more slowly on a SDS/ PAGE gel thus explaining the molecular mass differences observed between the 1D gel and the MALDI MS data. Because histone sequences are highly conserved between species, the identification of these proteins was straightfor- ward, based on primary amino acid sequences and the measured molecular m asses. Also, t he H2A N-terminal cDNA cloning (see below) and an expressed sequence tag from L. vannamei similar to histone H4 (http://www. marinegenomics.org; EST #7381) confirmed our assign- ments to histone proteins. Both the 1D gel and the direct MALDI MS spectrum show very strong signals for the hemocyte core histone proteins, suggesting a high abun- dance of histones in these cells. Histone H2A N-terminal cDNA cloning and C-terminal mass spectrometry determination The forward primer F2 and the reverse primers R1 and R2 successfully amplified a 161 base DNA sequence (F2 and R1) corresponding to 53 amino a cids, and a 207 base sequence (F2 and R2) corresponding to 69 amino acids, of histone H2A. From the nucleic acid sequence one new reverse primer was generated to obtain the 5¢ end of the cDNA sequence. Figure 3A shows the translated amino acid sequence using the translating tool of the ExPASy website. The C-terminal sequence was determined by tandem mass spectrometry of chymotryptic p eptides. Purification of histone proteins for liquid growth inhibition assays The hemocyte lysate supernatants wer e s eparated b y r everse phase HPLC on a C18 column. Histone proteins eluted between 40 and 5 5% solvent B. More specifically, a histone H1 fragment eluted at 42.5% solvent B, histones H2B and H4 coeluted at 51% solvent B, and histone H2A eluted at 52% solvent B as depicted in Fig. 4 . Histone H3 coeluted Fig. 1. 1D SDS/PAGE of hemocyte solu ble prote ins. Copp er stained 1D SDS 10–20% Tris/HCl gel of h emocyte l ysate p roteins. Molec ular markers are indicated on the left in kDa. Histone proteins and hemocyanin protein bands indicated on the right. 4828 S. A. Patat et al.(Eur. J. Biochem. 271) Ó FEBS 2004 with hemocyanin, the abundant oxygen carrier p rotein, at 56% solvent B (not shown). Approximately 61 lgof histone H2A proteins and 152 lg of mixed histones H2B/ H4 proteins were purified from 300 shrimp. Histone H2A protein was purified close to homogeneity, and histones H2B a nd H4 are the two major components in t heir fraction as evidenced by MALDI MS (Fig. 5A,B) and gel electro- phoresis (Fig. 5D). The fraction containing histone H2A shows two peaks by MALDI MS analysis (Fig. 5B) with a 78 Da difference. The two forms were separated by HPLC and both were s hown to be histone H2A by n anospray mass spectrometry sequencing a fter tryptic d igestion. Liquid growth inhibition assays with endogenous histone proteins purified from the hemocytes When tested for antimicrobial activity against the Gram- positive bacterium M. luteus , the lowest concentration of histone H2A that c ompletely inhibited growth was 4.5 l M and the lowest concentration of H2B/H4 fraction that completely inhibited growth was 3 l M (Fig. 6). Partial inhibition of growth was observed w ith 2 l M of the H2B/H4 fraction and no inhibition was observed with 0.5 l M of Table 1. Litopenaeus vannamei histone peptides sequenced by mass spectrometry. Predicted m/z is monoisotopic except for peptides from MS on MALDI-STR (average). Underlined sequence indicates sequenc e determined from MS/MS data. Enzyme indicates digestion with trypsin (T), chymotrypsin (C) or Glu-C (G). Amino acid positions of histones H2A and H4 are reported based on the L. vannamei cDNA sequences. Observed m/z (charge state) Predicted m/z (charge state) Instrument, Dm/z Sequence Enzyme Identification 425.73 (2 + ) 425.77 (2 + ) QSTAR, 0.04 HLQLAIR T H2A 82–88 454.21 (2 + ) 454.24 (2 + ) QSTAR, 0.03 YLAAEVLE G H2A 57–64 472.73 (2 + ) 472.77 (2 + ) QSTAR, 0.04 AGLQFPVGR T H2A 21–29 545.30 (2 + ) 545.29 (2 + ) LCQ, 0.01 AERVGAGAPVY C H2A 40–50 1092.58 (1 + ) 1092.68 (1 + ) LCQ, 0.1 PNIQAVLLPK T H2A 109–118 476.90 (3 + ) 476.60 (3 + ) LCQ, 0.3 AIRNDEELNKLL C H2A 86–97 760.34 (2 + ) 760.40 (2 + ) QSTAR, 0.06 RVGAGAPVYLAAVMoxE b G H2A 42–56 767.74 (3 + ) 767.80 (3 + ) QSTAR, 0.06 LLSGVTIAQGGVLPNIQAVLLPK T H2A 96–118 3146.99 (1 + ) 3147.67 (1 + ) MALDI-STR, 0.68, QSTAR a AIRNDEELNKLLSGVTIAQGGVLPNIQAVL C H2A 86–115 3388.22 (1 + ) 3388.96 (1 + ) MALDI-STR, 0.74, QSTAR a QLAIRNDEELNKLLSGVTIAQGGVLPNIQAVL C H2A 84–115 4099.91 (1 + ) 4100.87 (1 + ) MALDI-STR, 0.96, QSTAR a AIRNDEELNKLLSGVTIAQGGVLPNIQAVLLPKKTEKK C H2A 86–123 953.58 (1 + ) 953.60 (1 + ) LCQ, 0.02 LLLPGELAK T H2B 795.61 (2 + ) 795.44 (2 + ) LCQ, 0.17 AKHAVSEGTKAVTKY C H2B 895.68 (2 + ) 895.42 (2 + ) LCQ, 0.26 AMoxSIMNSFVNDIFER b T H2B 993.70 (2 + ) 993.54 (2 + ) LCQ, 0.16 PGELAKHAVSEGTKAVTKY C H2B 1106.91 (2 + ) 1106.62 (2 + ) LCQ, 0.29 LLPGELAKHAVSEGTKAVTKY C H2B 537.24 (1 + ) 537.33 (1 + ) QSTAR, 0.09 KPHR T H3 330.67 (2 + ) 330.70 (2 + ) QSTAR, 0.03 LPFQR T H3 358.18 (2 + ) 358.21 (2 + ) QSTAR, 0.03 DIQLAR T H3 394.71 (2 + ) 394.74 (2 + ) QSTAR, 0.03 KLPFQR T H3 416.21 (2 + ) 416.25 (2 + ) QSTAR, 0.04 STELLIR T H3 425.68 (2 + ) 425.72 (2 + ) QSTAR, 0.04 EIAQDFK T H3 516.76 (2 + ) 516.80 (2 + ) QSTAR, 0.04 YRPGTVALR T H3 1335.65 (1 + ) 1335.69 (1 + ) 4700, 0.04 EIAQDFKTDLR T H3 989.54 (1 + ) 989.58 (1 + ) 4700, 0.04 VFLENVIR T H4 61–68 1180.58 (1 + ) 1180.62 (1 + ) 4700, 0.04 ISGLIYEETR T H4 47–56 647.14 (2 + ) 646.85 (2 + ) LCQ, 0.29 LENVIRDAVTY C H4 63–73 1325.71 (1 + ) 1325.75 (1 + ) 4700, 0.04 DNIQGITKPAIR T H4 25–36 734.22 (2 + ) 733.91 (2 + ) LCQ, 0.31 TVTAMDVVYALKR T H4 81–93 584.65 (3 + ) 584.35 (3 + ) LCQ, 0.3 RDNIQGITKPAIRRL C H4 24–38 a MS determined on MALDI-STR, MS/MS acquired on QSTAR. b Mox indicates an oxidized methionine. Fig. 2. MALDI spectrum of hemocyte soluble proteins. Histone pro- teins are indicated by their abbreviations: H2A, H2B, H3 and H4. Ó FEBS 2004 Antimicrobial activity of shrimp histone proteins (Eur. J. Biochem. 271) 4829 histone H2A (data no t shown). Also, a fraction containing a histone H1 fragment inhibited growth of M. luteus (data not shown), but it is not known if the H1 fragment is the active antimicrobial component in the fraction. Note that the liquid growth inhibition assay does not distinguish between bacteriocidal or bacteriostatic mechanisms. Generation of the synthetic peptide H2A 2–39 and liquid growth inhibition assays Based o n t he N-ter minal sequence of H 2A from L. vannamei, a synthetic peptide (H2 A 2–39) s imilar i n length to buforin I w as synthesized (Fig. 3B) and pu rified by HPLC on a C18 column for antimicrobial assays. The MALDI MS of the tested fraction is depicted in Fig. 5C. The synthetic peptide H2A 2–39 was tested for anti- microbial activity against the Gram-positive bacterium M. luteus and has a minimum inhib itory concentration (MIC) value (inhibition of 50% growth compared to control) in the range of 0.5–1.0 l M (Fig. 7). Peptide antimicrobial activity was also observed against the Gram-positive bacteria Bacillus subtilis and Bacillus mega- terium at concentrations between 1.5 and 5 l M (data not shown). Discussion Histone protein s are primarily involved in DNA pack- aging and regulation of DNA replication and transcrip- tion. These proteins form the basic building blocks of chromatin structure when the four core histone proteins, H2A, H2B, H3 and H4, come together as heterodimers to constitute the nucleosome. The c ore histone proteins are highly conserved between species. Histone H1 is the linker t hat condenses the nucleosomes and exhibits greater sequence variability. Many reports have shown that histone proteins or histone-derived peptides from various vertebrates possess antimicrobial activity [19– 21,24,26,28]. Our results indicate that histones in L. van- namei have antimicrobial activity in vitro. Their high abundance in hemocytes suggests that the shrimp may be using histones i n antimicrobial defense. To the best o f our knowledge, this would be unprecedented in an invertebrate. Shrimp histone H2A protein completely inhibited growth of the test bacterium M. luteus at a c oncentration of 4.5 l M . Histone H2A from a variety of v ertebrate species has been shown to poss ess an timicrobial activity, but its activity appears to be variable. Fernandes et al.[21]showedthatthe trout histone H2A protein was active against several G ram- positive bacteria (including M. luteus) w ith M IC value s between 0.08 and 1.2 l M , but activity against Gram- negative bacteria was not observed at those concentrations. However, Kim et al. [24] tested the human histone H2A protein against the Gram-negative E. coli bacteria and Fig. 3. Histone H2A alignments and synthetic peptide sequence. (A) Litopenaeus vannamei histone H2A alignment with h istone H2A from other species. The shrimp histone H2A sequence was determined by cloning of the cDNA (bold sequence) and by mass spectro- metry sequencing (underlined sequence). (B) Synthetic peptide H2A 2–39, based on the length of buforin I and the shrimp N-terminus sequence. Fig. 4. Chromatogram at 22 5 nm of hemocyte soluble proteins on a C18 column. The acetonitrile concentration gradient is indicated by the dashed line. The peaks containing t he histone prote ins are indicated with the arrows. 4830 S. A. Patat et al.(Eur. J. Biochem. 271) Ó FEBS 2004 observed a complete inhibition of growth at 10 lgÆmL )1 ( 1.2 l M ). Peptides derived from the N-terminus of histone H2A of different organisms have activity against Gram-positive, Gram-negative b acteria a nd fungi at concentrations ranging from 0.5 to 5.0 l M [26–28]. In the toad, t he intact protein is secreted into the s tomach [33] and is then cleaved by p epsin. A similar m echanism w as demo nstrated in the catfish where the protein was secreted in the skin mucus and then cleaved by cathepsin D [34] to produce the active peptides. These data suggest the p ossibility that the N-terminus of shrimp histone H2A could be an active antimicrobial peptide. At this point in time however, there is no evidence that the shrimp H2A fragment 2–39 exists in vivo; but, a synthetic peptide identical to the first 38 amino acids of the N-terminus of H2A is active at submicromolar range against M. luteus and at concentrations lower than 5 l M against B. subtilis and B. megaterium. Examination of shrimp hemolymph for histone peptides is an obvious extension of this work. The shrimp histone H2B/H4 fraction completely inhibited growth of the test bacterium M. luteus at 3 l M . Clearly the shrimp fraction H2B/H4 needs to be further fractionated to determine if the active component is H2B, H4, or both. Histone H2B protein has been reported as having antimicrobial activity in the c hannel catfish skin [20], in human placenta [24], and in murine macrophages [35]. Also, an active fragment of histone H2B is e xpressed in T cells and natural killer cells [36]. Histone H4-derived fragments were reported to be present in an active fraction against B. megaterium, however, the peptides were not purified to homogeneity (histone H1 fragments coeluted), so it could not be concluded that they were the active components [30]. The L. vannamei H1 fragment found in an active fraction is smaller than 11 kDa and contains peptides similar to the central domain o f t he Drosophila histone H1 protein. It is necessary to purify this peptide to homogeneity to confirm its activity. A histone H1 protein fragment from rainbow trout skin secretions [37] has recently been described to inhibit g rowth of v arious Gram-positive and Gram-negative bacteria including M. luteus.Thisfragmentis69amino acids l ong and i s derived from the C-terminus of histone H1. Histone proteins undergo s everal post-translational mod- ifications (acetylation at s everal lysines, at their N-terminus or methylation) during regulation of DNA transcription and packaging. Kim et al. demonstrated by immunocyto- Fig. 6. Liquid growth inhibition assay of endogenous histones H2A and H2B/H4 against M. luteus. *Significant difference (P < 0.001) in growth compared to the water control growth using a Student’s t-test assuming equal variances. n ¼ 4 for each treatment except for H2A (n ¼ 2). M IP, CVTGEPVELD TQAL. Fig. 7. Growth inhibitio n assa y of M. luteus after 25 h incubation with the synthetic peptide H2A 2–39 at 37 °C, n ¼ 4, D 59 5 nm. A B C D Fig. 5. MALDI spectra (A–C) a n d 1D SDS/PAGE gel (D) of fraction s tested for inhibition of g rowth of M. luteus. (A) Histones H 2B and H4. (B) Histone H2A. (C) Synthetic peptide H2A 2–39. (D) Silver stained 1D SDS 10–20% Tris/HCl gel. D1, histones H2B and H4 ( 80 ng loaded); D2, histone H2A fraction ( 60 ng loaded). Ó FEBS 2004 Antimicrobial activity of shrimp histone proteins (Eur. J. Biochem. 271) 4831 chemistry t hat buforin I is unacetylated and that it is derived from a cytoplasmic unacetylated histone H2A p rotein [33]. However, in the rainbow trout, the histon e H2A active protein is a cetylated at its N-terminus [21]. The N-terminal 20 amino acids of shrimp his tone H2A have not been observed by mass spectrometry, and therefore the status of the N-terminus is unknown. Based on the cDNA partial sequence and the m ass spectrometry sequencing o f the C-terminal, the predicted molecular mass of histone H2A is 13 266 Da. If Met1 is cleaved and the protein acetylated on the N-terminus then the predicted molecular mass is 13 177 Da. Taking into account the oxidized methionine at position 55, the predicted molecular mass becomes 13 193 Da. The observed molecular mass for histone H2A by MALDI MS is 13 189 Da, which is within the experi- mental error of the instrument. Furthermore, two forms of histone H2A were identified in the shrimp. It is possible t hat the protein with a molecular m ass higher b y 7 8 Da is phosphorylated. The role of modification on the antimicro- bial properties remains to be elucidated. Many questions remain regarding the in vivo mechanism of antimicrobial histone action, particularly in the shrimp. Originally, histone proteins were thought to be localized in the nucleus; however, they have been shown to occur in the cytoplasm of various cell types [24,35] as well as, most recently, in NETs from neutrophils [25]. Histone proteins or histone-derived peptides can also be secreted, as demonstrated by a fragment of histone H2B found in the secretions of natu ral killer cells when stimulated with interleukin-2 [36]. I n both the toad and catfish, histone H2A is secreted prior to enzymatic cleavage to active peptides [33,34]. We hypothesize that the shrimp histone proteins from the hemocytes are localized in the cytoplasm and more specifically in the granules along with other antimicrobial peptides. Antimicrobial action could take place after release into the hemolymph upon infection or inside the cell phagosome. The well characterized antimi- crobial peptides in shrimp, the penaeidins, are produced and stored in the granular and semigranular hemocytes [14]. T heir release into the hemolymph is not thought to be through exocytosis but rather through a release of the granular content into the cytoplasm followed by hemocyte lysis [15]. In summary, the present study demonstrates that the core histone proteins H2A, H2B, H3 and H 4, identified by m ass spectrometry sequencing, are abundant proteins in L. vannamei hemocytes. Histone H2A as well as a mixture of histones H2B and H4 p revented growth of the t est bacterium M. luteus at concentrat ions in the same range as measured for vertebrate histones. A synthetic peptide containing the shrimp H2A sequence homologous to the known antimicrobial peptide buforin was antimicrobial with a MIC of 0.5–1.0 l M . Intact histone H1 protein was not found in this study; however, an H1-derived fragment was found in an HPLC fraction that was active against M. lute us. Further work is necessary to deter mine the in vivo defense mechanisms of shrimp hemocyte histone proteins. The present report suggests that multifunctional histone proteins are a conserved feature of innate immunity, probably not limited to vertebrates, such that all organisms in which histones are present are potentially able to utilize them as antimicrobial agents. Acknowledgements The authors t hank Dr Gregory Warr for helpful and insightful discussions; Dr Craig Browdy, Sarah Prior and Adrienne Metz for their help with the live shrimp; Javier Robalino and Brandon Cuthbertson for their en couraging discussi ons; the MUSC Mass Spectrometry facility f or the use of their instruments; D r Christian Schwabe from t he MUSC Biochemistry and Molecular Biology Department, and the MUSC Biotechnology f acility f or D NA sequencing a nd peptide synthesis. This study was funded by NSF grants IBN 0317303 to K.L.S. and EPS 0083102. Any opinions, findings, and conclusions or recommendations expressed in this material are those of t he authors and do not necessarily reflect the views of the National Scienc e Foundation. This work was also supported in part by grant NA03NMF4720362 from the National Marine Fisheries Service and is contribution number 550 to the Marine Resources Div ision of the South Carolina Department o f Natural Resources and num ber 9 to the MUSC Marine Biomedicine & Environmental Sciences Center. References 1. Johansson, M.W., Keyser, P., Sritunyalucksana, K. & Soderhall, K. (2000) Crustacean haemocytes and haematopoiesis. Aqua- culture 191, 45–52. 2. Vargas-Albores, F. & Yepiz-Plascencia, G. (2000) Beta glucan binding protein and its role in shrimp immune response. Aqua- culture 191, 13–21. 3. Marques, M.R.F. & Barracco, M.A. (2000) Lectins, as n on-self- recognition factors, in crustaceans. Aquaculture 191, 23–44. 4. Vargas-Albores, F., Jimenez-Vega, F. & Soderhall, K. (1996) A plasma protein isolated from brown shrimp (Penaeus californiensis) which enhances the activation of prophenoloxidase system by beta-1,3-glucan. Dev. Comp. Immunol. 20, 299–306. 5. Vargas-Albores, F., Jimenez-Vega, F. & Yepiz-Plascencia, G.M. (1997) Purification and comparison of beta-1,3-glucan binding protein from white shrim p (Penaeus vannamei). Comp. Biochem. Physiol. B, Biochem. Mol. Biol. 116, 453–458. 6. Sritunyalucksana, K. & Soderhall, K. (2000) The proPO and clotting system in crustaceans. Aquaculture 191, 53–69. 7. Soderhall, K. & Cerenius, L. (1998) Role of the p rophenoloxidase- activating system in invertebrate immunity. Curr. Opin. Immunol. 10, 23–28. 8. Zasloff, M. (2002) Antimicrobial peptides of multicellular organ- isms. Nature 415, 389–395. 9.Lehrer,R.I.&Ganz,T.(1999)Antimicrobialpeptidesin mammalian and insec t host def enc e. Curr. Opin. Immunol. 11,23– 27. 10. Schnapp, D., Kemp, G.D. & Smith, V.J. (1996) Purification and characterization of a proline-rich antibacterial peptide, with sequence similarity to bactenecin-7, from the haemocytes of the shore crab, Carcinus maenas. Eur. J. Bioc hem. 240, 532– 539. 11. Lee, S.Y., Lee, B.L. & Soderhall, K. (2003) Processing of an antibacterial peptide from hem ocyanin of the fresh water crayfish Pacifastacus leniusculus. J. Biol. Chem. 278, 7927–7933. 12. Destoumieux, D., Bulet, P., Loew, D., Van Dorsselaer, A., Rodriguez, J. & B achere, E. (1997) Penaeidins, a new family of antimicrobial peptides isolated from the shrimp Pena eus vannamei (Decapoda). J. Biol. Chem. 272, 28398–28406. 13. Destoumieux, D., Bulet, P., Strub, J.M., Van Dorsselaer, A. & Bachere, E. (1999) Recombinant expression and range of activity of penaeidins, antimicrobial peptides from p en aeid shrimp. Eur. J. Biochem. 266, 335–346. 14. Destoumieu x, D., Munoz, M., Cosse au, C., Rodriguez, J., Bulet, P., Comps, M. & Bachere, E. (2000) Penaeidins, antimicrobial peptides with chitin-binding activity, are produced and stored i n 4832 S. A. Patat et al.(Eur. J. Biochem. 271) Ó FEBS 2004 shrimp granulocytes and r eleased after m icrobial challenge. J. Cell Sci. 113, 461–469. 15. Munoz, M., Vandenbulcke, F., Saulnier, D. & Bachere, E. (2002) Expression and distribution of pen aeidin antimicrobial peptides are regulated by haemocyte reactions in microbial challenged shrimp. Eur. J. Biochem. 269, 2678–2689. 16.Gross,P.S.,Bartlett,T.C.,Browdy,C.L.,Chapman,R.W.& Warr, G .W. (2001) Im mune gene discove ry by e xpressed sequence tag analysis o f hemocytes and hepatopancreas in the Pac ific White Shrimp, Litopenaeus v annamei, and the Atlantic White Sh rimp, L. setiferus. Dev. Comp . Immu nol. 25, 565–577. 17. Relf, J.M., Chisholm, J.R., Kemp, G.D. & Smith, V.J. (1999) Purification a nd characterization of a c ysteine-rich 11.5-kDa antibacterial protein from the granularhaemocytesoftheshore crab, Carcinus maenas. Eur. J. Biochem. 264, 350–357. 18. Destoumieux-Garzon, D., Saulnier, D., Garnier, J., Jouffrey, C., Bulet, P. & Bachere, E. (2001) Crustacean immunity. Antifungal peptides are generated from the C terminus of shrimp hemocyanin in response to microbial challenge. J. Biol. Chem. 276, 47070– 47077. 19. Hirsch, J.G. (1958) Bactericidal action of hist ones. J. Exp. Med. 108, 925–944. 20. Robinette, D., Wada, S., Arroll, T., Levy, M.G., Miller, W.L. & Noga, E .J. (1998) Antimicrobial activity in the skin of the channel catfish Ictalurus punctatus: characterization of broad-spectrum histone-like antimicrobial proteins. Cell. Mol. Life Sci. 54,467– 475. 21. Fernandes,J.M.,Kemp,G.D.,Molle,M.G.&Smith,V.J.(2002) Anti-microbial properties of histone H2A from skin secretions of rainbow trout, Oncorhynchus mykiss. Biochem. J. 368,611– 620. 22. Richards, R.C., O’Neil, D.B., Thibault, P. & Ewart, K .V. (2001) Histone H1: an antimicrobial protein of Atlantic salmon (Salmo salar). Biochem. Biophys. Res. Commun. 284, 549–555. 23. Rose, F .R., Bailey, K., Keyte, J .W., Chan, W.C., Greenwood, D. & Mahida, Y.R. (1998) Potential role of epithelial cell-derived histone H1 proteins i n innate a ntimicrobial d efense in t he h uman gastrointestinal tract. Infect. Immun. 66, 3255–3263. 24. Kim, H.S., Cho, J.H., Park, H.W., Yoon, H., Kim, M.S. & K im, S.C. (2002) Endotoxin-neutralizin g antimicro bial proteins of the human placenta. J. Immunol. 168, 2356–2364. 25. Brinkmann, V., Reichard, U., G oosmann, C., Fauler, B., Uhl- emann, Y., Weiss, D.S., Weinrauch, Y. & Zychlinsky, A. (2004) Neutrophil extracellular traps kill bacteria. Science 303, 1532– 1535. 26. Park, C.B., Kim, M.S. & K im, S.C. ( 1996) A novel antimicrobial peptide f ro m Bufo b ufo gargarizans. Bi ochem. Biop hys. R es. Commun. 218, 4 08–413. 27. Park, I.Y., Park, C.B., Kim, M.S.&Kim,S.C.(1998)ParasinI,an antimicrobial peptide derived from histone H2A in t he catfish, Parasilurus asotus. FEBS Lett. 437, 258–262. 28. Birkemo, G.A., Luders, T., Andersen, O., Nes, I.F. & Nissen- Meyer, J. (2003) Hipp osin, a histone-d erived ant imicrobial pe ptide in Atlantic halibut ( Hippoglossus hippoglossus L.). Bioc him. Bio- phys. Acta 1646, 207–215. 29.Fernandes,J.M.,Saint,N.,Kemp,G.D.&Smith,V.J.(2003) Oncorhyncin III: a potent antimicrobial pe ptide d erive d from the non-histone chromosomal prote in H6 o f rainbow trout, Oncor- hynchus mykiss. Biochem. J. 373 , 621–628. 30.Wang,Y.,Griffiths,W.J.,Jornvall,H.,Agerberth,B.& Johansson, J. (2002) Antibacterial peptides in stimulated human granulocytes: characterization of ubiquitinated h i stone H1A. Eur. J. Biochem. 269, 512–518. 31. Patrzykat, A., Zhang, L., Mendoza, V., Iwama, G.K. & Hancock, R.E. (2001) Synergy of histone-derived peptides of coho s almon with lysozyme and fl ounde r pleur ocidin. Antimicrob. Agents Chemother. 45, 1337–1342. 32. Rodriguez, J., B oulo, V., Mialhe , E. & Bach  ere, E. (1995) Char- acterization of shrimp haemocytes and plasma components by monoclonal antibodies. J. Cell Sci. 108, 1043–1050. 33. Kim,H.S.,Yoon,H.,Minn,I.,Park,C.B.,Lee,W.T.,Zasloff,M. & Kim, S.C. (2000) Pepsin-mediated processing of the cytoplasm ic histone H2A to strong antimicrobial peptide buforin I. J. Immu- nol. 165, 3268–3274. 34. Cho, J.H., Park, I.Y., K im, H.S., Lee, W.T., Kim, M.S. & Kim, S.C. (2002) Cathepsin D produces antimicrobial peptide parasin I from histone H2A in th e s kin m ucosa of fi sh. FASE B J. 16 ,429– 431. 35. Hiemstra, P.S., Eisenhauer, P.B., Harwig, S .S., van den Barselaar, M.T., v an Furth, R. & Lehrer, R.I. (1993) A ntimicrobial proteins of murine macrophages. Infect. Immun. 61, 3038–3046. 36. Agerberth, B., Charo, J ., Werr, J., O lsson, B., Idali, F., Lind bo m, L., Kiessling, R., Jornvall, H., Wigzell, H . & Gudmundsson, G.H. (2000) The human antimicrobial and c hemotactic peptides LL-37 and alpha-defensins are expressed by specific lymphocyte and monocyte populations. Blood 96, 3086–3093. 37. Fernandes, J.M., Molle, G., Kemp, G.D. & Smith, V.J. (2004) Isolation and characterisation of oncorhyncin II, a histone H1-derived antimic robial peptide from s kin secretions of rainbow trout, Oncorhynchus mykiss. Dev. Comp. Immunol. 28, 127–138. Ó FEBS 2004 Antimicrobial activity of shrimp histone proteins (Eur. J. Biochem. 271) 4833 . Antimicrobial activity of histones from hemocytes of the Pacific white shrimp Se ´ verine A. Patat 1 , Ryan B. Carnegie 1, *,. hypothesize that the shrimp histone proteins from the hemocytes are localized in the cytoplasm and more specifically in the granules along with other antimicrobial